HVHan LIBRARIES SMITHSONIAN INSTITUTION NOlifUIlSNI NVINOSH1I1MS, S3 I a VM a II _ L I B R A|

ITUTION ^NOlinillSNI^NVINOSHimS^Sai avaan^LlBRARI ES^SMITHSONIAN^INSTITUTION NOlinj

***£&/ o '<$p&£y o

Z -I z _J z

avaan libraries Smithsonian institution NoiinmsNi nvinoshjuins sitavaan^uBRA

z | Z \ m CD m

(j> = CD ± CD X ± CD

ITUTION NOlinJLIJLSNI NVINOSHillNS S3 lavaa II LIBRARIES SMITHSONIAN INSTITUTION NOIlflJ

</) 2; CD Z C/5 2 av CD

H /Kg

j- •/,^5r /-'> /ii'v -r /,yy. /'s/ss o j®n;5n . Xv-- **• d?*

8

z Z NftS _

> 5 5 v’^ > XguiiSSX 2 >

avaa n2 L1BRAR I ES^SMITHSONIAN INSTITUTION NOIlfUIJLSNI NVINOSHIIWS^SS I a Va 3 11 LIBRA

c/> 2

■ • Xaso

cpmsSX o XgNpcJZ " O x”* d O xv Z q

riTUTION^NOlinillSNI^NVINOSHilWS^Saiyvaan^LIBRARIES^SMITHSONIAN^INSTITUTION^^OlinJ
2 r- z _ r- z r z

H Va 8 n ~LI B R AR I ES^SMITHSONIAN'lNSTITUTION^NOlinillSNI-NVINOSHlIWS^Sa I avaa n ZLI B RA

<r \ Z ^<STlrrr^v 5 2" 'iSSSit <C 2 . *c

v

5 M

3

z wj _

> XJVAst^ ^ > s XiVOTv^ >

•ITUTION NOUnillSNS^MVINOSHilWS^Saiavaan^LIBRARIES^SMITHSONIAN INSTITUTION NOliOJ

CD _ — CD 5 CD

^Stin/X ^ ui Xdda^x cu

H “ VV

_ m vi-

O ' W Z o NigSsWX Z ■ O

Z _J Z _j Z

avaan libraries smithsonian institution noiiiihisni nvinoshiivms saiavaan libra

z r- z r* ... z

MNOSH±mh_^ J I a v a d I
O

1U. J D r\ M r\ I u o QIVJ I I nOUM in

-rr 1*1 PI

~ CO ^ ^ ? CO ~ ^

1ITHS0NIAN INSTITUTION NOUnillSN! NVIN0SH1MS S3iavaai1 LI B RAR I ES SMITHSONIAN INSTI

t . <n z to z > . 2:

< \v 2 < V 5 <

— VW? ,

O >s^yosv^>

MITHSON IAN INSTITUTION NOliniUSNI NVIN0SH1IINS SBIBVUail LIBRARIES SMITHSONIAN INSTI
z r- z r z

ro ^v<v 2 oj 0

3

CO

tflN0SHlMS~S3 I a vy a I1~L I B R AR I ES^SMITHSONIAN “INSTITUTION " NOliniUSNI NVINOSHIIWS S3 I *

CO z ... CO 2: co z ^ ^ w

^jg^X s ,. < s s

' z —I z

X . ''/#■ -Aw O X O s, X

»« ^ ,n l<i?K m WN*%St \ --

3 CO

x

Siy.DC/ 5^ w" s '» § xgmissx 2 >

rfITHSONIAN INSTITUTION NOliniUSNI NVINOSHIIWS^S 3 I a V 8 8 1 1 LI B RAR I ES"SMITHSONIAN_INST

< (of

^ ^ ^ ^ a:

5 5 V ® 5 X2\dcX ~ W*- o

tflN0SHimSZS3l yvaan^LIBRAR IESZSMITHSONIAN-J|NSTITUTION N0linillSNI^NVIN0SHimS^S3 I *

v. Z 1 Z c- Z

to

||J ^

MITHS0NIAN-|NSTITUTI0NWN0linillSNl”NVIN0SHimSWS3 I a V8 a 11 _LI B R AR I ES SMITHSONIAN INST
_ z •. co z co z > co ^ ^ z

;•

S ' IsF' > ■ 2 Vfoos*/ >

VIN0SHimSWS3 I a V8 a I1ZLI B RAR 1 ES^SMITHSONIAN INSTITUTION N0UniUSNI_NVIN0SHlllNS'nS3 I
— co — CO z \ w

z ^ /ogySx ^ CO ^ ^

?/ c

ssr ~ 5 “,r O o

-J Z — _j z -J 2

MITHSONIAN INSTITUTION NOliniUSNI NVIN0SH1UMS SBIHVaan LIBRARIES SMITHSONIAN INST
r- _ z r- z 5“ X ~

VOLUME 22

Palaeontology

1979

PUBLISHED BY THE
PALAEONTOLOGICAL ASSOCIATION
LONDON

Dates of publication of parts of Volume 22

Part 1, pp. 1-272, pis. 1-27
Part 2, pp. 273-512, pis. 28-55
Part 3, pp. 513-746, pis. 57-97
Part 4, pp. 747-994, pis. 98-130

1 February 1979
21 May 1979
27 July 1979
5 November 1979

THIS VOLUME EDITED BY J. W. MURRAY, C. B. COX, M. G. BASSETT, AND K. C. ALLEN

Dates of publication of Special Papers in Palaeontology

Special Paper No. 22 June 1979

Special Paper No. 23 August 1979

© The Palaeontological Association , 1979

Printed in Great Britain
at the University Press, Oxford
by Eric Buckley
Printer to the University

CONTENTS

Part

Adams, C. G. and Belford, D. J. A new foraminifer from the Middle Eocene of

Papua New Guinea 1

Anstey, R. L. See Podell, M. E. and Anstey, R. L.

Appleby, R. M. The affinities of Liassic and later ichthyosaurs 4

Bakr, A. See Macfadden, B. J. and Bakr, A.

Baxendale, R. W. Plant-bearing coprolites from North American Pennsylvanian coal
balls 3

Belford, D. J. See Adams, C. G. and Belford, D. J.

Bockelie, J. F. Taxonomy, functional morphology and palaeoecology of the Ordovician
cystoid family Hemicosmitidae 2

Bosence, D. W. J. Live and dead faunas from coralline algal gravels, Co. Galway 2

Brannan, J. See Briggs, D. E. G., Rolfe, W. D. I. and Brannan, J.

Brenchley, P. J. See Pickerill, R. K. and Brenchley, P. J.

Briggs, D. E. G. Anomalocaris, the largest known Cambrian arthropod 3

Briggs, D. E. G., Bruton, D. L. and Whittington, H. B. Appendages of the
arthropod Aglaspis spinifer (Upper Cambrian, Wisconsin) and their significance 1

Briggs, D. E. G., Rolfe, W. D. I. and Brannan, J. A giant myriapod trail from the
Namurian of Arran, Scotland

Bruton, D. L. See Briggs, D. E. G., Bruton, D. L. and Whittington, El. B.
Buffetaut, E. and Ford, R. L. E. The crocodilian Bernissartia in the Wealden of the

Isle of Wight 4

Callomon, J. H. See Sykes, R. M. and Callomon, J. H.

Campbell, K. S. W. See Chatterton, B. D. E., Johnson, B. D. and Campbell, K. S. W.
Chatterton, B. D. E., Johnson, B. D. and Campbell, K. S. W. Silicified Lower
Devonian trilobites from New South Wales 4

Chatterton, B. D. E. See also Perry, D. G. and Chatterton, B. D. E.

Clarke, M. R. and Fitch, J. E. Statoliths of Cenozoic teuthoid cephalopods from
North America 2

Clarkson, E. N. K. The visual system of trilobites 1

Cocks, L. R. M. New acrotretacean brachiopods from the Palaeozoic of Britain and
Austria 1

Cooper, R. A. and Stewart, 1. R. The Tremadoc graptolite sequence of Lancefield,
Victoria 4

Crick, R. E. and Teichert, C. Siphuncular structures in the Devonian nautiloid
Archiacoceras from the Eifel of West Germany 4

Davey, R. J. Two new early Cretaceous dinocyst species from the northern North Sea 2

De Deckker, P. The Middle Pleistocene ostracod fauna of the West Runton fresh-
water bed, Norfolk 2

Drewry, G. E. See Hughes, N. F., Drewry, G. E. and Laing, J. F.

Edwards, D. A late Silurian flora from the Lower Old Red Sandstone of south-west
Dyfed 1

Elliott, G. F. Taxonomy and opercular function of the Jurassic alga Stichoporella 2
Fitch, J. E. See Clarke, M. R. and Fitch, J. E.

Ford, R. L. E. See Buffetaut, E. and Ford, R. L. E.

Freeman, E. F. A Middle Jurassic mammal bed from Oxfordshire 1

Haile, N. S. See Metcalfe, I., Koike, T., Rafek, M. B. and Haile, N. S.

Hughes, N. F., Drewry, G. E. and Laing, J. F. Barremian earliest angiosperm pollen 3

Page

181

921

537

363

449

631

167

273

905

799

479

1

93

767

747

427

293

23

407

135

513

CONTENTS

Part

Jennings, J. R. and Millay, M. A. Morphology of fertile Pecopteris unita from the
Middle Pennsylvanian of Illinois 4

Johnson, B. D. See Chatterton, B. D. E., Johnson, B. D. and Campbell, K. S. W.
Johnson, M. E. Evolutionary brachiopod lineages from the Llandovery Series of
eastern Iowa 3

Kennedy, W. J. and Wright, C. W. Vascoceratid ammonites from the type Turonian 3
Kennedy, W. J. See also Wright, C. W. and Kennedy, W. J.

Koike, T. See Metcalfe, I., Koike, T., Rafek, M. B. and Haile, N. S.

Laing, J. F. See Hughes, N. F., Drewry, G. E. and Laing, J. F.

Lehmann, U. The jaws and radula of the Jurassic ammonite Dactylioceras 1

Lofaldli, M. and Thusu, B. Micropalaeontological studies of the Upper Jurassic and
Lower Cretaceous of Andoya, northern Norway 2

Ludvigsen, R. See Von Bitter, P. H. and Ludvigsen, R.

Macfadden, B. J. and Bakr, A. The horse Cormohipparion theobaldi from the
Neogene of Pakistan, with comments on Siwalik hipparions 2

Mackay, S. See Williams, A. and Mackay, S.

McNamara, K. J. Trilobites from the Coniston Limestone Group (Ashgill Series) of
the Lake District, England 1

Mancenido, M. O. and Walley, C. D. Functional morphology and ontogenetic
variation in the Callovian brachiopod Septirhynchia from Tunisia 2

Metcalfe, I., Koike, T., Rafek, M. B. and Haile, N. S. Triassic conodonts from
Sumatra 3

Millay, M. A. See Jennings, J. R. and Millay, M. A.

Palmer, T. J. The Hampen Marly and White Limestone formations: Florida-type
carbonate lagoons in the Jurassic of central England 1

Perry, D. G. and Chatterton, B. D. E. Wenlock trilobites and brachiopods from the
Mackenzie Mountains, north-western Canada 3

Petters, S. W. Maastrichtian arenaceous Foraminifera from north-western Nigeria 4
Pickerill, R. K. and Brenchley, P. J. Caradoc marine benthic communities of the
south Berwyn Hills, North Wales 1

Podell, M. E. and Anstey, R. L. The interrelationship of early colony development,
monticules and branches in Palaeozoic bryozoans 4

Rafek, M. B. See Metcalfe, I., Koike, T., Rafek, M. B. and Haile, N. S.

Rolfe, W. D. I. See Briggs, D. E. G., Rolfe, W. D I. and Brannan, J.

Steele-Petrovic, H. M. The physiological differences between articulate brachiopods
and filter-feeding bivalves as a factor in the evolution of marine level-bottom
communities 1

Stewart, I. R. See Cooper, R. A. and Stewart, I. R.

Sykes, R. M. and Callomon, J. H. The Amoeboceras zonation of the Oxfordian 4

Tanabe, K. Palaeoecological analysis of ammonoid assemblages in the Turonian
Scaphites facies of Hokkaido, Japan 3

Teichert, C. See Crick, R. E. and Teichert, C.

Thusu, B. See Lofaldli, M. and Thusu, B.

Tripp, R. P. Trilobites from the Ordovician Auchensoul and Stinchar limestones of the
Girvan District, Strathclyde 2

Von Bitter, P. H. and Ludvigsen, R. Formation and function of protegular pitting
in some North American acrotretid brachiopods 3

Walley, C. See Mancenido, M. O. and Walley, C. D.

Whittington, H. B. See Briggs, D. E. G., Bruton, D. L. and Whittington, H. B.
Williams, A. and Mackay, S. Differentiation of the brachiopod periostracum 3

Wright, C. W. and Kennedy, W. J. Origin and evolution of the Cretaceous
micromorph ammonite family Flickiidae 3

Wright, C. W. See also Kennedy, W. J. and Wright, C. W.

Page

913

549

665

265

413

439

53

317

737

189

569

947

229

965

101

839

609

339

705

721

685

Palaeontology

VOLUME 22 • PART 1 FEBRUARY 1979

Published by

The Palaeontological Association • London

Price £1 2

THE PALAEONTOLOGICAL ASSOCIATION

The Association publishes Palaeontology and Special Papers in Palaeontology. Details of member-
ship and subscription rates may be found inside the back cover.

The journal Palaeontology is devoted to the publication oi papers on all aspects of palaeontology.
Review articles lare 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 on all aspects of palaeontology and strati-
graphical palaeontology are invited. They should conform in style to those already published in
this journal, and should be sent to Dr. R. A. Fortey, Palaeontological Association, Department of
Palaeontology, British Museum (Natural History), Cromwell Road, London SW7 5BD, England,
who will supply detailed instructions for authors on request (these were published in Palaeon-
tology. 20, pp. 921-929).

Special Papers in Palaeontology is a series of substantial separate works. Members may sub-
scribe to the Series ; alternatively, Ordinary and Student members only may obtain individual
copies at reduced rates. The following Special Papers are available :

1. (for 1967): Miospores in the Coal Seams of the Carboniferous of Great Britain, by a. h. v. smith and
m. a. butterworth. 324 pp., 72 text-figs., 27 plates. Price £8 (U.S. $16.00), post free.

2. (for 1968): Evolution of the Shell Structure of Articulate Brachiopods, by A. williams. 55 pp., 27 text-
figs., 24 plates. Price £5 (U.S. $10.00).

3. (for 1968): Upper Maestrichtian Radiolaria of California, by Helen p. foreman. 82 pp., 8 plates.
Price £3 (U.S. $6.00).

4. (for 1969): Lower Turonian Ammonites from Israel, by R. freund and m. raab. 83 pp., 15 text-figs.,
10 plates. Price £3 (U.S. $6.00).

5. (for 1969): Chitinozoa from the Ordovician Viola and Fernvale Limestones of the Arbuckle Moun-
tains, Oklahoma, by w. a. m. jenkins. 44 pp., 10 text-figs., 9 plates. Price £2 (U.S. $4.00).

6. (for 1969): Ammonoidea from the Mata Series (Santonian-Maastrichtian) of New Zealand, by
r. a. HENDERSON. 82 pp., 13 text-figs., 15 plates. Price £3 (U.S. $6.00).

7. (for 1970): Shell Structure of the Craniacea and other Calcareous Inarticulate Brachiopoda, by
A. williams and A. D. Wright. 51 pp., 17 text-figs., 15 plates. Price £1-50 (U.S. $3.00).

8. (for 1970): Cenomanian Ammonites from Southern England, by w. J. Kennedy. 272 pp., 5 tables,
64 plates. Price £8 (U.S. $16.00).

9. (for 1971): Fish from the Freshwater Lower Cretaceous of Victoria, Australia, with Comments on
the Palaeo-environment, by m. waldman. 130 pp., 37 text-figs., 18 plates. Price £5 (U.S. $10.00).

10. (for 1971) : Upper Cretaceous Ostracoda from the Carnarvon Basin, Western Australia, by R. h. bate.
148 pp., 43 text-figs., 27 plates. Price £5 (U.S. $10.00).

11. (for 1972): Stromatolites and the Biostratigraphy of the Australian Precambrian and Cambrian, by
m. r. Walter. 268 pp., 55 text-figs., 34 plates. Price £10 (U.S. $20.00).

12. (for 1973) : Organisms and Continents through Time. A Symposium of 23 papers edited by n. f. hughes.
340 pp., 132 text-figs. Price £10 (U.S. $20.00) (published with the Systematics Association).

13. (for 1974): Graptolite studies in honour of O. M. B. Bulman. Edited by r. b. rickards, d. e. jackson,
and c. p. hughes. 261 pp., 26 plates. Price £10 (U.S. $20.00).

14. (for 1974): Palaeogene Foraminiferida and Palaeoecology, Hampshire and Paris Basins and the
English Channel, by J. w. Murray and c. a. wright. 171 pp., 45 text-figs., 20 plates. Price £8
(U.S. $16.00).

15. (for 1975): Lower and Middle Devonian Conodonts from the Broken River Embayment, North
Queensland, Australia, by p. g. telford. 100 pp., 9 text-fig's., 16 plates. Price £5-50 (U.S. $1 1.00).

16. (for 1975): The Ostracod Fauna from the Santpnian Chalk (Upper Cretaceous) of Gingin, Western
Australia, by J. w. neale. 131 pp., 40 text-figs., 22 plates. Price £6-50 (U.S. $13.00).

17. (for 1976): Aspects of Ammonite Biology, Biogeography, and Biostratigraphy, by w. j. Kennedy
and w. a. cobban. 94 pp., 24 text-figs., 11 plates. Price £6 (U.S. $12.00).

18. (for 1976): Ostracoderm Faunas of the Delorme and Associated Siluro-Devonian Formations,
North West Territories, Canada, by d. l. dineley and e. j. loeffler. 218 pp., 78 text-figs., 33 plates.
Price £20 (U.S. $40.00).

19. (for 1977) : The Palynology of Early Tertiary Sediments, Ninetyeast Ridge, Indian Ocean, by E. M. kemp
and w. k. Harris. 74 pp., 2 text-figs., 8 plates. Price £7 (U.S. $14.00).

20. (for 1977): Fossil Priapulid Worms, by s. c. morris. 159 pp., 99 text-figs., 30 plates. Price £16
(U.S. $32.00).

21 . (for 1978) : Devonian Ammonoids from the Appalachians and their bearing on International Zonation
and Correlation, by M. r. house. Price £12 (U.S. $24.00).

© The Palaeontological Association, 1979

Cover. The lobopod animal Aysheaia pedunculata Walcott, 1911, from the Burgess Shale, Middle
Cambrian, British Columbia, • 4. Specimen in National Museum of Natural History, Smithsonian
Institution, Washington D.C., U.S.A. Photograph by H. B. Whittington, in ultraviolet radiation reflected
from surface (see Phil. Trans. R. Soc., London, B < 4).

THE VISUAL SYSTEM OF TRILOBITES

by E. N. K. CLARKSON
( Twenty-first Annual Address, delivered 8 March 1978)

Abstract. The compound eyes of trilobites are the oldest of all known visual systems and their evolution can be
traced over 350 million years. Only the lentiferous surface is preserved, however, since the lenses alone are calcitic.

Holochroal eyes have many lenses closely packed together. From a study of their evolution, the morphology and
optics of their lenses, various systems of lens packing, and the relationships between lens-thickness and that of the
cuticle, it is possible to disentangle those features of the lentiferous surface which result purely from geometrical
growth constraints from those which may have been of physiological significance. Holochroal eyes probably func-
tioned in a manner analogous to that of many modern insect and crustacean eyes.

Schizochroal eyes, unique to the animal kingdom, have large separated lenses. They probably were derived by
paedomorphosis from a holochroal precursor. The complex internal structures of these lenses has been investigated
using optical and scanning electron microscopy, as well as cathodoluminescence, which has enabled primary to be
distinguished from secondary structures.

Each lens, like those of holochroal eyes, when sectioned parallel with its principal plane shows calcite fibres arranged
in lamellae radiating from the central axis. Sections cut along the axis, however, show first how the lower part of
the lens contains an intralensar bowl of different texture to the rest of the lens, and secondly that the radial lamellae
are constructed of calcite fibres (trabeculae) diverging fanwise from the axis in the upper part of the lens, to abut the
upper convex surface near normally.

The complex internal structure of the schizochroal lens seems both to minimize birefringence, and to correct for
spherical aberration. Such high-quality lenses must have been linked to a photoreceptive system capable of making
use of their sharp images; in this and other contexts various theories of optical function in schizochroal eyes are
discussed and analysed.

Trilobites are amongst the most ancient of all Phanerozoic marine invertebrates,
but from their first appearance in the fossil record, they come equipped with remark-
ably elaborate sensory organs. The relics of these can be seen in the sensillar pits,
terraces, pore canals, and tubercles which may cover the cuticle (Miller 1976), and
most prominently in the paired compound eyes. These are the oldest of known
sensors and this alone would seem to render them a viable subject for study, but if
thereby the biological quality of the animals that bore them is illuminated, then their
interest is all the greater.

Compound eyes are typically arthropodan, but since the arthropods are probably
polyphyletic (Manton 1977), different kinds of compound eyes may have arisen in
more than one evolving stock. Thus it is hard to know what kind of functional
equivalence might be expected in fossilized eyes where only the outer lentiferous
surfaces are preserved, the soft parts having gone with little trace. The eyes of tri-
lobites are not the only fossilized compound eyes; they are found in ancient mero-
stomes as well in their modern relative Limulus (Eldredge 1974), in eurypterids
(Clarke and Ruedemann 1912; Wills 1965), and even in a Pennsylvanian centipede
(Levi-Setti, pers. comm.). In trilobites, however, the record is so much more complete
that their evolution can be studied through time, and this to some extent compensates
for the absence of internal parts.

[Palaeontology, Vol. 22, Part 1, 1979, pp. 1-22, pi. 1.]

PALAEONTOLOGY, VOLUME 22

Since the nearest modern analogues to trilobite eyes, in the rather broad and general
sense given by external appearance, are to be found in the eyes of insects and crusta-
ceans, it is appropriate first to discuss the morphology and function of these, and then
to consider what degree of similarity may be found in the eyes of trilobites.

THE COMPOUND EYES OF INSECTS AND CRUSTACEANS

In all modern arthropods possessing compound eyes the over-all structure of the
optical system is relatively constant. Even though compound eyes may have arisen
several times independently in different arthropod stocks they are all remarkably
similar, in that the eyes are constructed of numerous identical units, the ommatidia,
which are usually radially arranged so as to cover a fair angular field of vision (text-
fig. la, b).

In each ommatidium there are three main functional regions :

(a) The dioptric structures, which comprise the corneal lens (a thickened part of
the cuticle) and the crystalline cone below. Light passing through these transparent
modules on its way to the underlying photoreception is refracted to focus on the distal
end of the photoreceptive (retinular) cells which lay just below. The preservable
lenses of trilobite eyes evidently formed part or all of an analogous dioptric system
and are the only part of the eye now available for study.

(b) The photoreceptive part of the ommatidium (retinula), composed of a number
(usually seven or eight) of elongated retinular cells. The inward facing parts of these
cells (rhabdomeres) may or may not be in direct contact, and form a closed or open
cylinder, the rhabdom. These are made up of blocks of stacked microtubules, alter-
nating along the length of the rhabdom, and containing visual pigments. When light
strikes the rhabdom, the pigments are bleached by light and so trigger off nervous
impulses in the optic nerves which connect the photoreceptors to the optic ganglia
below.

(c) Pigment cells, which form an enclosing sheath normally isolating each
ommatidium.

There are two main kinds of ommatidium, generally found in the eyes of diurnal
and nocturnal or crepuscular arthropods respectively. The eyes bearing these two
types were termed apposition and superposition eyes by Exner (1891). In apposition
(daylight) eyes (text-fig. la) the rhabdom extends between the base of the crystalline
cone and the basement membrane, whilst in superposition eyes (text-fig. lb) it forms
only a short spindle located at the base of the ommatidium and connected to the
crystalline cone by a cone stalk. In superposition eyes (adapted to dark conditions)
the pigment shrinks away from the cone stalk region, sometimes both proximally and
distally, to aggregate round the crystalline cone and down near the basement mem-
brane. The ommatidia are thus no longer isolated. This allows greater over-all
sensitivity to dim illumination, for light striking the eye obliquely can then reach the
rhabdom from all angles and not just down the axis of the eye. Resolution, however,
is naturally diminished. When light-adapted, the pigment encloses the cone stalk
entirely, forming a continuous cylinder so that the rhabdom is shielded from oblique
rays, and as in apposition eyes only light travelling parallel with the ommatidial axis
can impinge on the rhabdom.

CLARKSON: TRILOBITE VISUAL SYSTEM

3

In spite of intensive research on compound eye physiology over the last 80 years,
there is still no clear agreement on how the various parts of the compound eye give an
over-all visual sense (Goldsmith and Bernard 1974). Most students, however, accept
some sort of ‘mosaic-vision’ theory, as first put forward by Muller (1829), and
elaborated by Exner (1891). According to this theory each ommatidium, when light-
adapted, should be sensitive only to light coming down the axis or at only a small
angle to it, since the pigment sheath absorbs the oblique light. Relatively little overlap
should therefore be expected between the visual fields of neighbouring ommatidia,
and the over-all ‘image’ formed at the retinular level would be a mosaic pattern of
individual bright dots like a silk-screen photograph. The coarseness of this mosaic
would depend on the number and size of the lenses and the angular separation of the
ommatidia. The dioptric apparatus of each ommatidium focuses light on the tip of
the rhabdom, and a ‘blur-circle’ rather than a sharp point of light is formed because
of diffraction at each small lens. The sensitivity of ommatidia to incident light has
been shown to be greatest along the optic axis, dropping off sharply as the angle of
light to the axis increases.

Whilst the incident light always seems to focus on the distal tip of the rhabdom,
and the rhabdomere ends are contained within the focal plane, the rhabdom itself
acts as a wave guide, and the incident light is conducted down to its base. Besides
this, however, an inverted image is formed in the rhabdom itself at a fixed, distance
below the lens, and further diffraction images may occur at lower levels. Whether any
of these images are actually used by visual system or whether they are merely in-
evitable, but unnecessary by-products of the optical system is still debatable.

text-fig. 1 . (a) structure of apposition-type ommatidium in a longitudinal section and (below) in transverse
section, as compared with ; (b) superposition-type ommatidium ; (c) variation in refractive index within the
lens and cone of the firefly Phausis (redrawn from Seitz 1969); ( d) reconstruction of a single lens of a
holochroal-eyed trilobite, based on Paladin eichwaldi shunnerensis (King).

4

PALAEONTOLOGY, VOLUME 22

In insect and crustacean eyes a very sharp focal point is sometimes assured by
variation in the refractive index of the crystalline cone (text-fig. lc). As within the
spherical lenses of fish eyes (Locket 1977) gradation in refractive index allows light
to come to a perfect focus, and a very clear image results. Elegant maps of several
such dioptric systems in insect eyes have been prepared, e.g. the optical system of the
firefly Phausis described by Seitz (1969). Not all insects have dioptric apparatus of
this kind, however; in the eyes of some fireflies refractive index is constant throughout
the cone.

In all compound eyes the size of the lenses seems to be a compromise between two
conflicting requirements. High sensitivity to light ideally needs a large lens, whilst
high resolution needs large numbers of small lenses in order to pick out fine detail by
making the mosaic finer. How large the lenses are in any given arthropod seems to be
related to the actual use of its eyes, and to the environment it inhabits. The upper size
limit of an ommatidial lens, according to Horridge (1977 a, b ), should theoretically
be no more than 30 ^m, but in some insects and crustaceans there are areas with much
larger lenses; up to 80 or even 120 ^m, concentrated in special foveal areas in which
the angular separation of the ommatidia is very low. In these foveas, high resolution
and high sensitivity are combined. These are the areas of most intense and acute
vision. Horridge has mapped these in several insects and has measured variation in
angular separation (M), and in diameter of the lenses (D) across their eyes. Eyes with
a large eye parameter (DM), i.e. large lenses with a wide angular separation, are
generally adapted for night or deep-sea vision, whilst those with very many small
lenses, of low angular separation (i e. with a low DM), approaching the theoretical
limits set by diffraction), occur in insects living in bright sunlight. The eye parameter
may vary horizontally across the eye. Horridge associates the latter phenomenon
with the fact that in flying insects objects at the side have greater angular velocity than
those at the front, and the size gradation and separation seems to relate to this.

The mosaic theory, in the light of much experimental evidence, is a useful model
or first approximation to understanding the function of the compound eye, but no
more, and whilst many problems have been illuminated there are many others which
for the moment seem intractable.

Some of this information of the function of compound eyes may be useful in inter-
preting the mode of operation of trilobite eyes; indeed, the above discussion has been
largely confined to those aspects of its operation which can be understood from the
dioptric system alone.

In order to form a reasonably clear picture of the visual system in trilobites the
following properties should be established :

The morphology, fine structure, and mineralogy of as many different trilobite
eyes as possible, based on well-preserved material. In particular the various kinds
of eyes must be defined, and primary structures within them must be distinguished
from secondary diagenetic effects.

The evolution and diversification of the various kinds of eye through geological
time, with documentation of any main evolutionary trends.

The time and mode of origin of new eye types.

The form, growth patterns, and geometry of the lentiferous surface, especially
the size, manner of packing, and spacing of the lenses.

CLARKSON: TRILOBITE VISUAL SYSTEM

5

The optics of the lenses, functioning both individually and collectively as a
complete or partial dioptric system.

The angular range of vision of the eye, and if possible the angular bearings of
individual lens-axes within the visual field.

Any striking comparisons or contrasts between trilobite eyes and those of
modern arthropods.

HOLOCHROAL TRILOBITE EYES

Holochroal eyes are by far the commonest type and occur in trilobites of all ages
from Cambrian to Permian. They possess many lenses of relatively small size (30-
100 ix m, rarely larger), closely packed together and in contact. The lenses are covered
by a thin common cornea which is no more than an extension of the outer layer of
the cuticle.

Little is known of the early history of holochroal eyes, and indeed our knowledge of
these for the first 60 million years of trilobite history is very scanty and is based solely
upon a few meraspids of Lower Cambrian age. Adult Lower and Middle Cambrian
trilobites rarely have the visual surface preserved, for an ocular suture was emplaced
below the visual surface in the early adult stage of development, so that the lentiferous
surface was not attached to the librigena and fell out after death. In the later Upper
Cambrian, however, the ocular suture became obsolete in certain groups, and in the
Ordovician most trilobites retained the visual surface for it was welded to the libri-
gena. Only in a few groups, such as the Calymenina, was the ocular suture retained.
Its obsolescence and the consequent retention of the visual surface is probably a
paedomorphic phenomenon, at least in the olenids (Clarkson 1973 b ), and probably
in other trilobites too.

From the early Ordovician to the late Permian the record of holochroal trilobite
eyes and their associated sensory zones is good, and details of structure are known in
olenids, Asaphus, scutelluids, some illaenids, and proetids and cyclopygids (Lind-
strom 1901 ; Clarkson 1975). In some respects there is quite a considerable range of
variation in characters such as, for instance, lens size, form, thickness, and number,
and in the size and shape of the visual surface and the angular range subtended by
the lens array. In other respects, however, the holochroal eye was a very conservative
organ, in which growth of the visual surface as an anteriorly expanding logarithmic
spiral and the emplacement of lenses in a generative zone along the lower rim of the
eye almost always seems to have taken place in the same general way. Indeed it seems
fairly clear that throughout the 350 million year history of the eye in trilobites there
were only three main controls : changes in proportion, size, and surface curvature of
the visual surface due to differential relative growth; paedomorphosis; and the
incorporation of cuticular sensillae into discrete sensory zones marginal to the eye.

There is now a useful body of information on the structure and evolution of the
eye in trilobites, but there are acute problems in trying to interpret this functionally
in terms of modern compound eyes since the soft parts have all gone. There remains
the possibility, however, of disentangling those morphological characters which
result purely from geometrical constraints during growth, from those which may have
been physiologically significant. Certain characters seem to be especially amenable to

6

PALAEONTOLOGY, VOLUME 22

analysis in this way, and of these the form of the visual surface, the geometry of
growth and lens packing, and the form, geometrical optics, and internal structure of
the lenses are here selected for further discussion.

The visual surface

Lens-packing systems. The genetic programmes which governed the growth of many
Palaeozoic organisms were relatively simple, as witness, for instance, the independent
occurrence of dichotomous branching in pteridophytes, graptolites (especially as
seen in anisograptids), and in crinoids. A very straightforward generative instruction
suffices for this: i.e. bifurcate when the branch has reached length X. Length X may
remain constant for successive zones of growth, but usually decreases by an arith-
metic or logarithmic factor. Similarly uncomplicated genetic instructions sufficed to
generate the lentiferous surface of most trilobite eyes, and it is possible to work these
out from the morphology of the eye alone.

It has been established (Beckmann 1951; Clarkson 1971, 1975) that almost all
trilobite eyes, holochroal or schizochroal, begin their growth as a thin strip of exo-
skeletal material (the generative zone) just below the palpebral suture, which is in
this case a logarithmic spiral. The lenses are produced from the generative zone
which moves downwards as the eye enlarges, always forming the base of the visual
surface and adding new lenses at each ecdysis until the eye is fully grown. Thus the
new lenses are tacked on to the ends of dorso-ventral files which form as a result of
this mode of growth. These files form characteristic patterns, sometimes confined to
certain taxonomic groups only; a number of these have been illustrated elsewhere
(Clarkson 1975).

In all cases the size of the lenses is controlled by the spacing of the lens centres,
which may remain constant or may alter as the visual surface grows. The new lenses
can only grow as far as the proximity of neighbouring lenses will permit, and then
their growth is arrested. Thus, for example, the spacing of the lens centres in most
scutelluid eyes decreases by a constant logarithmic factor as the eye grows, and if the
locations of the individual points in the generative zone from which the lenses are
budded off shift laterally during growth, then the eye will come to have large lenses at
the top and small ones towards the base, arranged in curving diagonal dorso-ventral
files (Clarkson 1975, fig. 5k, pi. 1, figs. 1-10).

The important point here is that whilst the change in lens size of the visual surface
may have been useful to the trilobite, it results from nothing more than an answer (one
of many viable ones), to the problem of packing lenses regularly from a single
marginal generative zone, on to a curving surface. Whilst the fact that the lenses were
larger in one region of the visual surface was primarily controlled by geometric neces-
sity alone, some interest attaches to whether trilobites actually did find this useful,
and indeed they may well have done so.

In many modern arthropods variation in the size of the lenses is not uncommon,
and may be more pronounced than in trilobites, reaching an extreme form in the
remarkable bilobed eyes of euphausiids (Chun 1896; Kampa 1965). As mentioned
before, some insects have different parts of the eye specialized for particular functions,
especially where differential surface curvature and change in facet size allows certain
parts of the visual field to be covered by a ‘fovea’, where relatively large lenses with

CLARKSON: TRILOBITE VISUAL SYSTEM

7

very small interommatidial angles between them provide high resolution as well as
good capacity for light gathering.

The examples I have found in trilobites which might suggest a similar function are
somewhat equivocal. As an example, the eyes of Pseudogygites, aff. latimarginatus
(PI. 1, fig. 6) have large biconvex lenses occupying the upper part of the eye and
covering the latitudes between about 45 and 90° above the equator. Below these upper
lenses are smaller ones, decreasing regularly in size downwards, but with small,
laterally directed axial angles. Whilst it is clear enough that this growth pattern, like
the ones previously described, is a geometrical phenomenon, the possible value of
this to the trilobite cannot be established, for the level of functional differentiation in
terms of pronounced bilobation as found in the eyes of mantis shrimps by Horridge,
and in some euphausiids by Chun, is never approached by trilobites.

The curvature of the visual surface, and the extent of the visual field. It might be
expected that the angular range subtended by the eye would of necessity be adaptive,
and indeed this may often be so. Nevertheless, even this frequently seems to be related
to the over-all form of the trilobite as much as to any specific adaptation. Thus inde-
pendently in cheirurids and scutelluids highly vaulted genera such as Crotalocephalus
and Paralejurus have laterally directed eyes with a latitudinal visual range of only
some 30° or so, whilst their more flattened close relatives Cheirurus and Scutellum
have panoramic eyes with a range of more than 90° of latitude. Whilst this might just
be the result of a purely fortuitous convergence, there can be little doubt that in two
closely related species of the French Ordovician phacopid genus Crozonaspis, the
actual form of the schizochroal eye, and hence its visual range, is directly controlled
by the form of the cephalon. Simple Cartesian transformations (text-fig. 2), using
the celebrated methods of D’Arcy Thompson (1961), make this abundantly clear.

text-fig. 2. Cephala of (a) Crozonaspis kerfornei Clarkson and Henry, 1970; and ( b ) Crozonaspis struvei
Henry, 1968; closely related Phacopina from the Ordovician of Brittany, plotted on Cartesian transforma-
tion grids showing that the different shapes of the eye are a function of relative growth alone.

PALAEONTOLOGY, VOLUME 22

The difference in the visual field of the two species is adaptive only in the sense that
the morphology of the whole animal is adaptive, and for the moment one cannot go
beyond this.

Lens structure and function

Radial structure of holochroal lenses. Lindstrom (1901) figured a number of thin
sections of holochroal eyes with their lenses ground parallel with their principal
planes. In many of these there seems to be a pronounced radial symmetry (as shown by
Dysplanus centrotus Dalman (ibid.), p. 55, pi. Ill, figs. 53-54; Illaenus chiron Holm,
p. 58, pi. IV, figs. 22-23; and Symphysurus palpebrosus Dalman, p. 62, pi. IV, figs.
16-17, though in the latter case the radial structure is largely obscured by diagenesis).

The presence of a radial pattern within the lenses has been amply confirmed by
etching the surfaces of various well-preserved holochroal eyes with EDTA. Paladin
eichwaldi shunnerensis (King) is a magnificently preserved Namurian trilobite found
in the North of England (Osmolska 1970), whose eyes have already been studied in
some detail (Clarkson 19696, 1975). Etched specimens (pi. 1, figs. 1-5; text-fig. 1 d),
show that each lens consists of thin lamellae, radially arranged around the c-axis,
and probably originally contiguous. The radial structure is less pronounced in the
central part of the lens, where it is interrupted by concentric rings. This central part
seems as a whole to etch more rapidly, perhaps suggestive of a slightly different
mineralogy and hence refractive index.

Etched sections normal to the principal plane show that the radial lamellae them-
selves consist of slender calcite fibres here termed trabeculae, which turn outwards
in a fan-like manner so that their distal terminations lie near normal to the convex
outer surface of the lens. This is only found in eyes with convex lenses; in Asaphus
raniceps , which has prismatic lenses with a flat outer surface, though the lenses
are made of radial lamellae, the trabeculae of which they consist do not turn
outwards distally. Presumably the outward torsion of the trabeculae in convex
lenses is associated with the minimization of birefringence for oblique incident light
rays, which would thus be conducted down the curving c-axis of each trabecula with-
out being doubly refracted. Although the radial-lamellar and trabecular construc-
tion is made visible by the etching process, it is also thereby modified. The actual
structures which are illustrated are thus to some extent artefacts, and it is not possible
to tell whether the trabeculae and lamellae were originally in contact. Despite the

EXPLANATION OF PLATE 1

Structure of holochroal eyes.

Figs. 1-5. Paladin eichwaldi shunnerensis (King), Namurian E2, Shunner Fell Limestone, near summit of
Great Shunner Fell, North Yorkshire, England. 1, left eye of holotype SM E 10497, x 14. 2, lower part
of visual surface of Gr I 45,668, etched with EDTA to show radial structure of the lenses, x 200. 3, the
same, a single lens enlarged, x 100. 4, oblique lateral view of deeply etched lens, showing radial lamellae,
x 1000. 5, polished and etched surface, ground nearly parallel with the lens axis, showing curving
trabeculae on the right-hand side. Gr I 45,669, x 800 approx.

Fig. 6. Pseudogygites aff. latimarginatus (Hull). Ashgill, Allen Bay, Devon Island, Canada. Right eye on
detached librigena, GSC C-22858, x 45.

PLATE 1

CLARKSON, trilobite eyes

10

PALAEONTOLOGY, VOLUME 22

complex internal structure of the lenses each still consists of a single crystal of calcite
and like the stereom of echinoderms, the whole lens has crystallographic unity. It is
hard to assess the functional significance of this kind of internal organization. It
may be no more than a growth phenomenon, the addition of trabeculae at the ends of
radial lamellae being simply an easy way to grow a calcitic lens. Whether this is so
or not, the system has been found useful, for without it it might have been more
difficult to eliminate or minimize birefringence in this remarkable fashion.

The radial-lamellar and trabecular structure in schizochroal lenses is virtually
identical with that in holochroal eyes, indeed it was first noted in the schizochroal eyes
of Phacops rana miller i (text -fig. 4). The only real difference in the detailed structure
of the lenses is that schizochroal lenses have intralensar bowls whilst, to the best of
our knowledge, bowls are absent in holochroal lenses. But this correspondence in
fine structure suggests a basic unity, and provides a link between two otherwise very
dissimilar-looking kinds of eye. The schizochroal eye was certainly derived from a
holochroal precursor, probably by paedomorphosis (p. 17), and in the process
retained this fundamental radial plan.

Geometrical optics of the lenses. In holochroal-eyed trilobites the thickness of the
lenses is widely variable. As Lindstrom (1901) first showed some trilobites have very
thin biconvex lenses. Those Cambrian trilobites which have been studied in detail
(e.g. the olenids), all have lenses of this kind, but they are found also in the Ordo-
vician Pseudogygites aff. latimarginatus and others. Scutelluids and proetids tend to
have lenses of intermediate type whilst many other kinds of trilobite possess very
elongated and prismatic lenses; of these A. raniceps is well known (Clarkson 1973a).
Are there any significant functional reasons for this diversity? Certain lines of evi-
dence suggest that this too may be a matter of geometric rather than physiological
necessity. Firstly, in the examples mentioned above and in most others there is a
general correlation between the thickness of the lenses and that of the cuticle. Where
the cuticle is thin the associated lenses are thin, whilst prismatic lenses are associated
with a thick cuticle. In general, lens thickness is about two-thirds that of the cuticle,
whatever the form of the lens or prism.

Secondly, the surfaces of these various kinds of lenses are so shaped as to bring
incident light to a focus at approximately the same relative distance below the proxi-
mal surface of each lens. The lenses of the Upper Cambrian olenid Ctenopyge (Clark-
son 19736) are biconvex in form, the lower surface being of slightly greater curvature.
Given the known refractive indices of the external sea water, of calcite along the
c-axis, and that assumed for internal body fluids, as in modern marine arthropods
(Clarkson and Levi-Setti 1975), the focal length is readily calculated using Gaussian
lens-formulae (e.g. Jenkins and White 1976), and ray paths may then be traced
(text-fig. 3).

In the case of Asaphus (Clarkson 1973a), the upper surface of the prism is virtually
flat. The proximal surfaces of the prisms in most of the material which I originally
studied were damaged, probably by solution during ecdysis, for all the specimens
appeared to be exuviae. There are a number of lenses, however, which still retain most
of their original form, showing hemispherical proximal ends of near perfect shape.
The drawing (text-fig. 3a) constructed from a high magnification photograph shows

text-fig. 3. Optics of the lenses of holochroal eyes: (a) Asaphus raniceps Dalman. L. Ord., Oland, Sweden.
Part of horizontal section through Gr I 5512, (figured by Clarkson 1973a, pi. 50, fig. 1); (ft) same, recon-
struction of lenses based on (a), with ray paths traced for incident light normal to surface; (c) Sphaeroph-
thalmus humilis (Phillips), Andrarum, Scania. Reconstruction of lenses in horizontal section (based on
Gr 1 20803, figured by Clarkson 19736, pi. 95, fig. 3) ; (d) same, with ray paths traced for incident light normal
to the surface; ( e ) Bojoscutellum campaniferum L. Dev. Koneprusy, Bohemia. Horizontal section through
lenses in the centre of the eye. Gr I 14202 (figured by Clarkson 1975, fig. 4c); (/) same, thicker lens at
the periphery of the eye; (g) same, lenses illustrated in (e), ray paths traced for incident light parallel with
the axis, using Gaussian thick-lens formulae. The effects of curving trabeculae are minimal for incidence of
light along the axis and are not considered here.

PALAEONTOLOGY, VOLUME 22

how the radius of curvature is readily measured so that the focal length can be cal-
culated. Likewise the focal length of thicker biconvex lenses, e.g. those of Bojoscu-
tellum, are readily calculated using the oblique ray method for thick lenses as
described by Jenkins and White (1976, ch. 5).

Again in Asaphus, the hemispherical proximal surface of the prism compensates
for the virtually flat distal surface so that light is brought to a focus at a similar rela-
tive distance below the proximal surface to that of the biconvex lens of Ctenopyge.
The two lenses have similar / numbers (( f/D ) (focal length / divided by diameter D)
in this case / 1 = 2-5). The actual length of the lens or prism is not important for the
focusing of the light; what is significant is the relative curvature of the two surfaces.
Since both kinds of lenses focus light at a similar relative distance below the proximal
surface, and indeed as do lenses of intermediate length and curvature, differences in
axial length of the lens or prism are clearly not an optical necessity. It is more likely
that this is a simple structural requirement or growth necessity. The relatively simple
genetic instructions of a trilobite could not grow a thick cuticle without also growing
a thick lens, but the terminations of the lenses would need to be of a particular form
if the eyes were to function effectively.

In insects and crustaceans, light is always focused on the distal tips of the rhabdoms
(apposition eyes) or on the tip of the cone stalk (superposition eyes). These lie at a
very similar relative distance below the dioptric apparatus as do the focal planes
calculated for the lenses of various holochroal-eyed trilobites. This tends to support
the assumption that these trilobite eyes contained some kind of ommatidia, one
below each lens. Within such eyes there is plenty of room for ommatidia as well as for
a large central optic ganglion. Despite the obviously greater diameter of the lenses in
trilobites than in modern insects and crustaceans (30-100 fxm as opposed to 8-30 /un)
on average, there seems to be fair grounds for interpreting the holochroal trilobite
eye as having some structural and functional equivalence to the eyes of insects and
crustaceans. And whilst the programmes responsible for its growth seem to have been
generally rather simple, the elegant structures which minimized birefringence do not
suggest that the holochroal trilobite eye was an organ of inferior or inadequate bio-
logical quality.

SCHIZOCHROAL TRILOBITE EYES

Schizochroal eyes appeared quite suddenly in the early Ordovician. They were pre-
sumably derived from a holochroal ancestor. They are confined to the Ordovician to
Devonian suborder Phacopina, though the eyes in some other taxa also have certain
features in common with the schizochroal condition. Schizochroal eyes are usually
large, with thick biconvex lenses, often relatively few in number, and in some cases as
much as 1 mm across. They are arranged on an inclined, curving visual surface, but
their visual range is never more than some 40° above the equator. The lenses are not
in direct contact with one another but are separated by cuticular material or, as it is
known, intralensar sclera. Whilst the gross morphology, and the angular bearings of
the lenses axes within the visual field have been known for many phacopids for some
time, most of the current interest in schizochroal eyes is concerned with detailed
microstructure and optical properties of the lenses, and the significance of this in
understanding the function of the whole eye.

CLARKSON: TRILOBITE VISUAL SYSTEM

Lens structure

A knowledge of the mineralogical composition and detailed microstructure of the
lenses is a necessary prerequisite for understanding their optics. Early work by Lind-
strong 1901) was based on thin sections prepared for optical micrography, and it was
he who first recognized intralensar structures, though he thought that they were
probably secondary rather than primary. Working with Bohemian phacopids I found
(Clarkson 1967a, 1969a) that polished surfaces showed more detail than thin sections.
In these, though it was sometimes hard to distinguish primary from secondary micro-
structure, one consistent element in the lenses of nearly all the specimens was a bowl-
shaped unit in the base of the lens, similar to those which Lindstrom figured. This
intralensar bowl was found to be of very variable shape amongst Phacopina. In some
Ordovician Dalmanitinae, evidence of the three-dimensional shape of the bowls came
from internal and external moulds ( Dalmanitina , Crozonaspis, Zeliszkella), whilst
in Silurian dalmanitids the bowl appeared as a dark, symmetrical, centrally indented
element at the base of the lens (Clarkson 1968; Clarkson and Levi-Setti 1975; Levi-
Setti 1976). Towe (1975), however, has not detected bowls in the material he
studied.

Campbell (1975) likewise used thin sections and polished surfaces and has reported
the existence of intralensar bowls and other internal structures in several Phacopina.
In all cases the bowl was found to be present, though in members of a dimorphic pair
its precise shape varied from one dimorph to another. In addition he gave clear evi-
dence of subconcentric laminae within the upper unit of the lens, similar to those
described by Clarkson (1969a) in Reedops bronni.

Campbell (1975) also described a pyriform central core in a number of phacopid
eyes. It was most clearly identifiable as a primary structure in silicified specimens of
Paciphacops birdsongensis, in which the outer parts of the lenses had been removed by
weathering to expose individually silicified laminae, indented centrally by the distal
tip of the core. Campbell’s photographs (ibid., pi. c, figs. 3-6), leave no doubt that
this is a real structure and not a diagenetic artefact.

Miller and Clarkson (in prep.) have been able to confirm these observations in
Phacops rana miller i from the Devonian Silica Shale of Ohio. We have used thin sec-
tions, polished surfaces, ground surfaces etched with EDTA, and examined with the
scanning electron microscope, and cathodoluminescence micrography, which
enables some of the primary structures to be distinguished from diagenetic artefacts.
Diagenetic effects were common, and unmodified structure is found only rarely. The
etched and scanned material shows that the bowl and the core are of regular form, and
of much denser texture than the upper unit of the lens. Rather curiously, the bowl is
very thin or absent directly below the core, though the lip is quite thick and rounded
(text-fig. 4). The upper part of the lens is traversed by laminae, convex upwards and
more closely spaced towards the top.

In P. rana milleri the highly biconvex upper unit is of quite complex, though regular,
form and it is somewhat difficult to interpret, for the etching process modifies the
original structure whilst accentuating the details. In sections parallel with the principal
plane a radial structure is apparent, though this is less distinct in the intralensar bowl.
The calcite is arranged in thin lamellae radiating out from the central core; these

14

PALAEONTOLOGY, VOLUME 22

lamellae are particularly clear at the top of the lens which is the first part to be re-
formed after ecdysis. Sections cut parallel with the axis, however, show that each of
the radial lamellae, like those of holochroal eyes, are made of a large number of
fibrous calcitic trabeculae, running parallel with the axis in the lower part of the eye,
but diverging outwards towards the outer surface of the lens, each trabecula abutting
this surface near normal to it.

Whilst relatively few of the specimens we have examined are entirely unaffected by
diagenesis, and the above observations are based on a small number of unaltered
eyes, comparative details are present in a number of other schizochroal-eyed genera
and species.

laminae cornea

a b

text-fig. 4. Phacops rana milleri Stumm. Devonian, Silica Shale, Ohio. Reconstruction of lenses in
(a) vertical, and ( b ) horizontal section.

We have also studied the post-ecdysial growth of the eye in P. rana milleri, which
has shown that the post-ecdysial lenses are very thin and biconvex, they thicken as the
cuticle thickens, becoming of Cartesian form and eventually adding the intralensar
bowl last of all. With the thickening of the lens and cuticle, the cylindrical alveolar
cavity below the lens deepens. At about the time when the bowl is added there appears
an annular girdle preserved as grey micrite just below the ambitus of each lens.
Finally, more cuticular material, obliquely laminated, is secreted on the cylindrical
wall of the alveolus, as an alveolar ring lying against the intrascleral membrane which
is continuous with the cornea.

Whilst the ‘phacopiform’ lenses of P. rana milleri are in many ways typical of
advanced Phacopina, other shapes of bowl and lens occur. In the more primitive
‘acastiform’, and ‘dalmanitiform’ eyes the lenses are more numerous, smaller, more
closely packed, and less strongly biconvex. These have flatter bowls, which in Dal-
manitina closely approximate to the ideal aplanatic lens first described by Des Cartes
in 1637. It is not yet known whether these lenses typically contain cores or not.

CLARKSON: TRILOBITE VISUAL SYSTEM

5

Optics of the lenses

Towe (1973) ground phacopid lenses parallel with the principal planes of schizo-
chroal lenses and was able to take quite sharp photographs through the clear calcite
of the remaining half lens. This simple experiment conclusively proved the existence
of orientated calcite, with the c-axis normal to the principal plane, but could not, of
course, establish the optical nature of an intact biconvex lens. Following Levi-Setti’s
discovery of the remarkable correspondence between the shapes of the upper units in
Dalmanitina socialis and Crozonaspis struvei, and the ideal aplanatic lenses of Des
Cartes (1637) and Huygens (1690), optical models were made of the same shape. The
upper unit was constructed of a block of orientated calcite («=l-66), machined to
shape, bowls shaped to fit and made of various plastics were tried in turn. A poly-
sulphone bowl with refractive index (« = 1-63), brought parallel beams of light to a
sharp focus below the lens ; the result of a combination of slightly different refractive
indices, separated by a Cartesian surface (Clarkson and Levi-Setti 1975).

Campbell (1975) pointed out that the core should have some effect on the lens
optics, and indeed this should be the case, so should the fan-like arrangement of the
calcite trabeculae in the upper unit. If each trabecula or lamina did act as a light guide,
then oblique light impinging on the lateral part of the exposed lens surface would, as
in holochroal eyes, then be conducted down the c-axis of each curving trabecula with-
out being broken into two rays.

Thus although the model proposed by Clarkson and Levi-Setti appears to hold
good, the complex upper unit with its core and radial structure appear to be designed
for further optical refinement, particularly in minimizing the acute problems of bire-
fringence caused by the lens being made of calcite.

The nature of the original sublensar structures

Three possibilities have been suggested for the original structures which underlay
each lens of a schizochroal eye : an ommatidium, analogous with those of modern
insects or crustaceans (Clarkson 1967a); a relatively short ocellus with a flat layer of
narrow retinular cells some distance below the lenses (text-fig. 5c) (Campbell 1975;
Clarkson and Levi-Setti 1976 ; Stockton and Cowen 1976) ; a structure with no known
modern analogues, and therefore hard to interpret (Stuermer 1970; Stuermer and
Bergstrom 1973).

In defence of the ocellar theory Campbell points out that ‘the indirect evidence
suggests that the character of the optic units in phacopid eyes are not those associated
with ommatidia’, and notes that even the best ommatidia could not resolve objects
subtending an angle less than the angle of separation of the ommatidia, which may be
very high. He also shows that in most modern eyes, thick biconvex lenses are most
commonly associated with ocelli. The ocellar lens of the larval sawfly Perga even has
an intralensar bowl. Campbell suggests that the sublensar cone I described in Anana-
spis communis ( = Phacops fecundus) and the somewhat vaguer equivalents in Reedops
cephalotes (Clarkson 1967a, 19696), rather than being equivalent to the crystalline
cone, were more likely ocellar capsules at whose base lay the retinular layer (text-
fig. 5c). In all these respects I agree with Campbell that an ocellar hypothesis is more
plausible than the others.

16

PALAEONTOLOGY, VOLUME 22

Perhaps the closest analogue to the schizochroal eye of trilobites is to be found in
the compound eye of Strepsipterida (Insecta), (text-fig. 5a, b), to which Dr. R. A.
Crowson of Glasgow University kindly drew my attention. These eyes, described
most recently by Kinzelbach (1967), are large and hemispherical, with relatively
enormous and separated lenses. In different genera and species the lens size and
number varies considerably, but they usually have Cartesian proximal surfaces. The
size of the lenses varies within a single eye. Below each lens is a short, modified omma-
tidium, which retains the neural structure of normal ommatidia, but in which the
retinulae are more spread out in a concave layer, and hence much more like an ocellus.

text-fig. 5. (a, b). Xenos (Strepsipterida) a ‘schizochroal-eyed’ insect: (a) external view of eye showing
large, separated lenses; ( b ) a single ommatidium in vertical section (redrawn from Kinzelbach 1967);
(c) Campbell’s reconstruction of an optical unit of a phacopid trilobite (redrawn, slightly modified, from

Campbell 1975).

While this would seem a good model for a complete visual unit, the remarkable
structures described by Stuermer (1970), and Stuermer and Bergstrom (1973) from
X-radiographs, must be taken into consideration. These have the form of very
elongated fibres extending from near the visual surface to deep down in the body,
converging on the midline of the trilobite. The specimens in which they occur are
somewhat distorted and the fibres do not connect with the lenses, hence Campbell
(1975) and I (Clarkson 1973a) have suggested that these are not part of the visual
system. Nevertheless, in stereoscopic X-ray pairs which Professor Stuermer kindly
sent me, the fibres do appear to be coming from different levels inside the eye and are
certainly not all in one plane, as would be expected if they were merely gill lamellae.
If these are actually part of the visual system as Stuermer and Bergstrom claim, then
they could actually be very elongated ommatidia or some kind of cone stalks, alterna-
tively they might be nerves connecting the bases of unpreserved, immediately sub-
lensar structures (and there is no reason why these could not have been ocelli), to a
deep ganglion, which for some reason was most curiously placed in the centre of the
body. There are thus conflicting lines of evidence as to what structures lay below the
lenses and until this is resolved, our understanding of the functions of the schizo-
chroal eye must remain to some extent speculative.

core

a

b

c

CLARKSON: TRILOBITE VISUAL SYSTEM

17

Origins, function, and use of the schizochroal eye

The schizochroal eye, as developed in Phacops and related genera, at first sight
seems so different from any holochroal precursor that a search for its origins may
seem futile. Yet there are resemblances both in the logarithmic spiral form of the
visual surface and in the radial pattern of the lenses, which show a clear relationship.
Some meraspid trilobites begin with relatively large lenses, separated by intervening
cuticular material ; and as the eye grows a small patch of lenses which retains to some
extent this morphology, may be left in the centre of the eye, just below the
palpebral suture. The larval eye of Paladin eichwaldi shunnerensis is of this type
(Clarkson 1975, pi. 3, figs. 12, 13) and the small eyes of Pagetia described by
Jell (1975) as ‘abathochroal’ seem to be remarkably similar in external appear-
ance to such meraspid holochroal eyes, though their fine structure is not well
known.

It is quite probable that the earliest schizochroal eyes were derived paedomorphic-
ally from a holochroal ancestor by retaining the relatively large and separated lenses
of the juvenile condition into the adult phase. Further evidence from these rare and
unusual eyes must be sought, however, before this is confirmed.

Assuming such a paedomorphic holochroal precursor what would be the next
stage in development? Examination of the earliest schizochroal-eyed genera shows
that in most respects they are typically schizochroal and of dalmanitiform structure.
Even these, however, come too late in the early evolution of the schizochroal eye for
the processes of their origin to be readily determined, in all respects except one. For
in the Arenig Llanvirnian genus Ormathops the lens packing system is significantly
different and all eyes examined have a less than regular distribution of lenses on the
visual surface (Clarkson 1971, 1975). These lenses, unlike those of other Phacopina,
are all of identical size. Since in Ormathops, the spacing of the lens centres stayed the
same as new lenses were added to the generative zone at the base of the eye, the lenses
were unable to grow beyond a certain size. But as the eye grew downwards the visual
surface expanded, leaving room for the more lenses whenever there was a large
enough space ; these were automatically emplaced by the simple genetic programme
and once formed, acted as a focus for new lens files as the eye grew larger still. In all
the eyes of Ormathops species, there are normally two or three blocks in which the
lenses are regularly arranged and the dorso-ventral files are parallel. These blocks
are separated by discontinuities, which may be sharp and angular (caesurae) or simply
less distinct zones where the lenses are irregularly distributed ; in the latter case the
emplacement of lenses in these zones seems to have been partially controlled from
both sides.

Amongst early Phacopina only Ormathops has an eye of this kind, presumably
retaining the identical size of the lenses from a holochroal ancestor. There must have
been strong selection pressure in favour of regularity of packing at the expense of
identical size, for all other Phacopina have achieved regularity simply by increasing
the distance between lens centres in the generative zone to accommodate for the
increasing girth of the eye as it grows. The lenses therefore become larger towards
the base of the eye. Whilst this might seem too optically disadvantageous, it is possible
that a slight change in biconvexity of the lenses from the top to the bottom of a file

PALAEONTOLOGY, VOLUME 22

could have altered the focal length (/), and since the / number of the lens is f/D,
where D is the diameter, might have allowed a constant / number whatever the lens
diameter. The sensitivity of the lens can be measured as iff2, and hence if the / number
remains the same so does the sensitivity. So far it is not entirely certain whether or
not this theoretical model was adopted by trilobites ; it is merely one of a number of
ways in which the trilobites eye could have come to terms with the problems of change
in lens size.

In nearly all Phacopina, the result of such packing control is the establishment of a
regular system of hexagonal close packing on the eye surface, but an unusual case of
cubic close packing has been described recently in Phacops turco aff. praecedens Haas
by Fortey and Morris (1977), who state that it could be accounted for by a relatively
small initial difference in lens spacing during ontogeny within the dorso-ventral files.
So far, however, this is the only case of such close packing described. It is associated
in this case with lenses of a fairly constant size, but the functional significance of the
system is presently unknown.

Having assessed something of the origin of the schizochroal eye, and knowing that
individual lenses were sensitive (because of their large size) and capable of producing
sharp images, it remains to consider the function of the schizochroal eye as a whole.
One way of approaching this is through the measurement of the visual field, and of
the angular bearing of lens axes within it. In the eyes of all Phacopina the visual field
normally forms a relatively narrow strip, latitudinally aligned, with the upper limit of
vision rarely rising above 40° of latitude, and it is usually below 30°, contrasting with
the frequently panoramic visual field of many holochroal eyes (Clarkson 1966a, b).
Dr. A. W. A. Rushton (pers. comm.) has suggested that since the lenses are very large,
and capable of point focusing, it is possible that they could have overloaded the photo-
receptors had they been directed straight upwards at the source of light, and this may
be one of the reasons why the schizochroal eye never faces direct illumination. On the
other hand, the orientation of the visual field must also be an adaptation to their mode
of life.

One of the most striking characters in the whole visual system is the peculiarly
unhomogeneous distribution of the lenses within the phacopid visual field. The plan
curvature of the visual surface may be much greater than the profile curvature so that
the lens-axes of the dorso-ventral files tend to be clustered together with small latitu-
dinal axial angles, whilst their longitudinal separation is quite wide. This is perhaps
most extreme in Acaste where a narrow visual field, directed 10° above the equator
and covering only 10° of latitude is traversed by distinct ‘visual strips’, within which
the axial angle is only 1-2°, but between which it may be as much as 10-15° longi-
tudinally. Not all schizochroal eyes show this extreme condition, but there is always
some difference between latitudinal and longitudinal axial angles and frequently the
lens axes are clustered towards the base of the visual field. To what extent these
differences and indeed the pattern of lens-axis distribution within the Phacopina as a
whole are actually adaptive, is for the moment hard to determine.

Previously I proposed (Clarkson 1966a) that the schizochroal eyes of trilobites
were adapted for no more than movement perception. This was upon the under-
standing that an approaching object would progressively occlude more lens-axes, as
Professor Rudwick first pointed out, and that a passing object would register as a

CLARKSON: TRILOBITE VISUAL SYSTEM

19

flicker across the visual field. This view, however, was propounded before the remark-
ably elegant structure of the lenses was known. A more embracing theory of the
function of the whole eye, has recently been proposed by Cowen and Kelley (1976),
and elaborated by Stockton and Cowen (1977). This seems to be a good model for
many aspects of schizochroal eye function.

These authors draw attention to the extreme convexity of the lenses and suggest
that adjacent lenses in the one eye, especially those within a dorso-ventral file, could
have been used for stereoscopic vision provided that there were appropriate neural
links and relays connecting the photoreceptive units. These authors adopt the ocellar
theory and assume a flattish retina of individual photoreceptors at the base of a lens
capsule lying some short distance below the lens. The lens capsule is partially con-
tained within the sublensar alveolus. This would accord with the fact that the upper-
most lenses in the dorso-ventral files of many phacopids are set at an angle to the
alveoli. There are many analogues for this system in modern arthropods, the eyes of
spiders, the larval eyes (stemmata) of beetles, and the ocelli of various insects. Indeed,
as Campbell mentions, the larval eye of the sawfly Perga , has a large thick lens, pro-
vided with an intralensar bowl, and a short lens capsule below, with a basal retina
(Meyer-Rochow 1974).

A pair of adjacent lenses covering a particular region of the visual field would both
see the same object but it would appear on opposite sides of their respective retinas.
As it moved towards or away from the lens-pair, it would register as a movement of
the stimulated points on the two retinas— hence ‘the distance of an object would be
inferred by comparison of images in adjacent lenses at one time ; movement of an
object could be detected by comparison at successive times’. Stockton and Cowen
therefore see the schizochroal eye as designed to give a warning of the presence and
movement of near-by objects, and in particular a three-dimensional appreciation of
actual distance. These authors estimate, using simple geometry, that stereoscopic
vision would be effective at up to 25 cm away from the eye, and even up to 2 m if
lenses at opposite ends of the dorso-ventral files were neurally connected, though
they did not especially favour this latter idea. It is clear that such a system would
operate best for adjacent lenses of dorso-ventral files (hence the selection pressure to
dispense with the less regular lens array of Ormathops). There may also have been a
possibility of stereoscopic vision between adjacent files, though this would have
involved a much more complex neural relay system.

This model seems to account for a number of the remarkable features of the
schizochroal eye, especially if the /-number, and hence sensitivity of the lenses could
have been made constant through slight changes in surface curvature, from top to
bottom of a file. As has been shown, the calcitic trabeculae of which the lenses are
constructed radiate outwards so as to abut the lens-surface near normally and could
act as non-birefringent light guides for the conduction of strongly oblique light,
emphasizing the role of the lens in collecting light over a wide angle. This, too, is in
accordance with Stockton and Cowen’s model. It is in fact, quite probable that there
were other features of the schizochroal eye of equal functional importance, and which
have not yet been detected.

The ‘stereoscopic model’ depends, largely, however, upon whether the basic
assumption is justified, i.e. whether there was a short ocellar lens capsule or an

20 PALAEONTOLOGY, VOLUME 22

ommatidium below each lens. If it were the latter the model would need each serious
modification.

Whilst we are still far from a good understanding of the schizochroal eyes of
Phacopina, it is clear that they were in no way primitive or inferior organs, or that
their biological function was very limited. The elegance of the lens design at least and
the various corrections of which the lenses were capable do not imply a low-grade
nervous system associated with them. The not infrequent incidence of blindness in
Phacopina and many holochroal-eyed trilobites may in some cases at least be environ-
mentally related (Clarkson 19676), and does not imply that the eyes were of poor
enough quality to be easily dispensed with.

Finally, what did trilobites, especially those with schizochroal eyes, actually use
them for? Many trilobites seem to have been mud ingesters or filter feeders, though
as Whittington (1975) has shown, the gnathobasic jaws and spiny appendages may
have enabled some species to pick up and triturate small worms from the substrate.
Even if they were predators to this degree, the eyes would not have been much use to
them in their search for worms, since they are located on the dorsal surface of the
cephalon, and the interpretation of hypostomal maculae as ventral eyes by Lind-
strom (1901) is still sub judice. There is, however, a fair general correlation between
the possession of large and well-developed eyes, and the ability to enroll. It is common,
though not invariable to find that those trilobites with large eyes, whether holochroal
or schizochroal, frequently have superior enrollment ability and fine vincular (co-
aptative) structures. The primary function of trilobites eyes as distant early warning
sensors for the detection of approach of predators, seems to be a reasonable deduc-
tion from the evidence, especially since major changes both in eye structure and in
enrollment ability seems to have taken place in many of the early Ordovician groups
at around the same time. The combination of advanced visual and protective systems
may well have been a major factor in prolonging the existence of trilobites until the
end of the Palaeozoic even in the face of fierce competition and predation.

Acknowledgements. I am very grateful to the Council of the Palaeontological Association for the invitation
to deliver the Twenty-first Annual Address.

I thank John Miller for helpful discussion and permission to quote from unpublished joint work, and
Brian S. Norford for sending me the specimens of Pseudogygites from the Geological Survey of Canada
(G.S.C.).

REFERENCES

beckmann, H. 1951. Zur Ontogenie der Sehflache grossaugiger Phacopiden. Palaont. Z. 24, 126-41, pi. 10.
Campbell, K. s. w. 1975. The functional anatomy of phacopid trilobites : musculature and eyes. J. Proc. Roy.
Soc. New South Wales , 108, 168-188, pis. A.-C.

chun, c. 1896. Leuchtorgane und Facettenaugen. En Beitrag zur Theorie des Sehens in grossen Meere-
stiefen. (i) Atlantis. Biol. Stud, iiber pelagische Organismen , 6, 191-262. (ii) Zoologica, 19, 1 -260, pis. 1-6.
clarke, j. m. and ruedemann, R. 1912. The Eurypterida of New York State. Mem. N. Y. St. Mus. Nat. Hist.
14, 439 pp.,88 pis.

clarkson, E. n. K. 1966a. Schizochroal eyes and vision of some Silurian acastid trilobites. Palaeontology,
9, 1-29, pis. 1-3.

— 19666. Schizochroal eyes and vision in some phacopid trilobites. Ibid. 464-487, pis. 73-75.

— 1967a. Fine structure of the eye in two species of Phacops (Trilobita). Ibid. 10, 603-616, pi. 99.

— 1 9676. Environmental significance of eye-reduction in trilobites and Recent arthropods. Mar. Geol.
5, 367-375.

CLARKSON: TRILOBITE VISUAL SYSTEM

clarkson, E. N. K. 1968. Structure of the eye of Crozonaspis struvei (Trilobita, Dalmanitidae, Zeliszkellinae).
Senck. Lethaea, 49, 383-391, pi. 1.

— 1969a. On the schizochroal eyes of three species of Reedops (Trilobita, Phacopidae) from the Lower
Devonian of Bohemia. Trans. R. Soc. Edinburgh, 68, 183-205, pis. 1-3.

— 19696. Dimorphism of the eye in Weberides shunnerensis (King) (Trilobita), 185-195, pi. 13. In Sexual
Dimorphism in Fossil Metazoa and Taxonomic Implications (westermann, g. e. g., ed.), Int. Union.
Geol. Sci., Series A, Stuttgart, Schweizerbart.

— 1971. On the early schizochroal eyes of Ormathops (Trilobita, Zeliszkellinae). Mem. Bur. Rech. Geol.
Mineral. 73, 51-63, pi. 1.

— 1973a. The eyes of Asaphus raniceps Dalman (Trilobita). Palaeontology, 16, 425-444, pis. 48-50.

— 19736. Morphology and evolution of the eye in Upper Cambrian Olenidae (Trilobita). Ibid. 735-763,
pis. 91-95.

— 1975. The evolution of the eye in trilobites. Fossils and Strata, 4, 7-31, pis. 1-6.

— and henry, j. l. 1969. Sur une nouvelle espece du genre Crozonaspis (Trilobite) decouverte dans
TOrdovicien de la Mayenne. Bull. soc. geol. France, 7e serie, 9, 1 16-123, pi. 1 .

— and levi-setti, R. 1975. Trilobite eyes and the Optics of Des Cartes and Huygens. Nature, 254, 5502,
663-667.

cowen, r. and kelley, j. s. 1976. Stereoscopic vision within the schizochroal eye of trilobites. Ibid. 261,
130-131.

d’arcy Thompson, w. 1961. On Growth and Form. (Abridged edition.) (bonner, j. t., ed.), Cambridge,
346 pp.

des cartes, R. 1637. CEuvres de Des Cartes. La Geometrie. Livre 2, Leiden, J. Maire.
eldredge, n. 1974. Revision of the Suborder Synziphosurina (Chelicerata, Merostomata), with remarks on
Merostome Phylogeny. Amer. Mus. Novitates, 2543, 1-41.
exner, s. 1891. Die Physiologie der facettirten Augen von Krebsen und Insekten, Leipzig and Vienna, 206 pp.
fortey, R. a. and morris, f. s. 1977. Variation in lens packing of Phacops. Geol. Mag. 114, 25-32, pis. 1-2.
goldsmith, t. h. and Bernard, G. d. 1974. The Visual System of Insects. Chapter 5 in The Physiology of
Insecta (2nd ed.) ii (rockstein, m., ed.). New York and London, Academic Press. 568 pp.
henry, j. l. 1968. Crozonaspis struvei n.g. n. sp. Zeliszkellinae (Trilobita) de TOrdovicien moyen de la
Bretagne. Senck. Lethaea, 49, 367-380, pis. 1-2.
horridge, G. 1977a. Insects which turn and look. Endeavour (n.s.), 1, 7-16.

19776. The compound eye of insects. Scientific American, 237, 108-120.

huygens, c. 1690. Traite de la Lumiere. Leiden, Pierre van der Aa.

jell, p. a. 1975. The abathochroal eye of Pagetia, a new type of trilobite eye. Fossils and Strata, 4, 33-43,
figs. 1-5.

jenkins, f. a. and white, H. E. 1976. Fundamentals of Optics (4th ed.). McGraw-Hill, Kogakusha, 1-746.
kampa, E. m. 1965. The Euphausiid Eye— a re-evaluation. Vision. Res. 5, 475-81.

kinzelbach, R. 1967. Zur Kopfmorphologie der Facherfliigler (Strepsiptera, Insecta). Zool. Jb. Anat. Bd.
84, 559-684.

levi-setti, r. 1976. Ancient and wonderful eyes. Fossils Magazine, 1, 35-42.

lindstrom, G. 1901. Researches on the visual organs of the Trilobites. K. Svensk. Vetensk. Akad. Handl.
34, 1-86, 6 pis.

locket, N. a. 1977. Adaptations to the deep-sea environment, 67-192. In Handbook of Sensory Physiology,
VII/5. The Visual System in Vertebrates (crescitelli, f., ed.), Springer Verlag, Berlin, Heidelberg.
manton, s. M. 1977. The Arthropods, Habits, Functional Morphology, and Evolution. Clarendon Press,
Oxford, 527 pp.

meyer-rochow, v. b. 1974. Structure and function of the larval eye of the sawfly Perga. J. Insect Physiol.
20, 1565-1591.

miller, j. 1976. The sensory fields and life mode of Phacops rana (Green, 1832). Trans. R. Soc. Edinburgh,
69, 337-367.

— and clarkson, e. n. K. (in prep.). Post-ecdysial development of the cuticle and visual system of the
trilobite Phacops rana milleri (Stewart, 1927).

Muller, J. 1829. Zur Vergleichenden Physiologie des Gesichtsinnes. Leipzig, Cnobloch. Pp. 1-89.
osmolska, h. 1970. Revision of non-crytosymbolinid trilobites from the Tournaisian-Namurian of Eurasia.
Palaeont. Polonica, 23, 1-165, pis. 1-22.

22 PALAEONTOLOGY, VOLUME 22

seitz, G. 1969. Untersuchungen am dioptrischen Apparat des Leuchtkaferauges. Z. vergl. Physiologie, 62,
61-74.

stockton, w. l. and cowen, r. 1976. Stereoscopic vision in one eye: palaeophysiology of the schizochroal
eye of trilobites. Paleobiology , 2, 304-315.

stuermer, w . 1 970. Soft parts of Cephalopods and T rilobites : some surprising results of X-ray examination
of Devonian slates. Science , 170, 1300-1302.

— and bergstrom, j. 1973. New discoveries on trilobites by X-rays. Paldont. Z. 47, 104-141, pis. 16-24.

towe, K. 1973. Trilobite eyes: calcified lenses in vivo. Science, 179, 1007-1009.

wills, J. a. 1965. A supplement to Gerhard Holm’s ‘Uber die Organisation des Eurypterus fischeri Eichw.’,
with special reference to the organs of sight, respiration, and reproduction. Ark. Zool. (ser. 2) 18, 93-142,
pis. 1-8.

Whittington, h. b. 1975. Trilobites with appendages from the Middle Cambrian Burgess Shale, British
Columbia. Fossils and Strata, 4, 97-136, pis. 1-25.

E. N. K. CLARKSON

Grant Institute of Geology
West Mains Rd.
Edinburgh EH9 3JW

Typescript received 11 May 1978

A LATE SILURIAN FLORA FROM THE
LOWER OLD RED SANDSTONE OF
SOUTH-WEST DYFED

by DIANNE EDWARDS

Abstract. This extensive compression flora from the late Silurian (Downtonian) of South Wales contributes to our
understanding of terrestrial vegetation relatively early in the colonization of the land by predominantly vascular
plants. The large number of fertile Cooksonia specimens has resulted in an amplified diagnosis of C. hemisphaerica
Lang for which intraspecific variation is demonstrated ; in the erection of a new species, C. cambrensis, and in records
of plants close to C. caledonica and C. pertonii. Elongate sporangia on unbranched twisted stalks are named Tortili-
caulis transwalliensis gen. et sp. nov. and affinities with bryophytes and vascular plants are discussed. Hostinella is
the commonest stem type, but some smooth axes show departures from dichotomous branching. In addition, axes
with triangular and truncated spines are described indicating, together with the above, a diversity of taxa and morpho-
logical organization. A short progress report on our studies of the early history of vascular plants in the South Wales
area is presented.

W. H. Lang’s classic paper on the Downtonian (approximately equal to Pridoli)
floras of southern Britain was published over forty years ago (1937), but the funda-
mental importance of his meticulous work has been fully appreciated only recently in
the upsurge of interest in the early evolution of land plants (for example, Banks 1975a,
b). His genus Cooksonia is now recognized as one of the most completely known early
vascular plants (see Addendum in Gray and Boucot 1977), although axes with con-
vincing tracheids have never been found attached to sporangia. Cooksonia has since
been recorded from Downtonian localities in Czechoslovakia (Obrhel 1962), Libya
(Daber 1971), U.S.A. (Banks 1973), and Podolia (Ishchenko 1975), and from
younger strata in Scotland (Edwards 1970), Wales (Croft and Lang 1942), and the
U.S.S.R. (Ananiev and Stepanov 1969; Yurina 1969). It is from these Devonian
horizons that new species have been described, workers on Downtonian floras either
having been reluctant to identify often very fragmentary and relatively featureless
fossils to a specific level, or having assigned them to one of Lang’s original species,
C. pertonii or C. hemisphaerica.

Indeed apart from Lang’s work, little is known about land vegetation in the late
Silurian : vascular plant remains usually comprise sterile, rather featureless axes (for
example, Hoeg 1942) often associated with the highly problematical non-vascular
genera Prototaxites Dawson and Pachytheca Hooker. An exception is the Podolian
flora (Ishchenko 1969, 1975) which contains numerous, but unfortunately poorly
illustrated plants, many of which are of uncertain affinity. The flora at Freshwater
East, Dyfed is also an extensive one. It was first recorded by Dixon in 1921, and Lang
himself illustrated a few specimens in his 1937 paper. My own collection, amassed
over several years and now housed at the National Museum of Wales, contains

[Palaeontology, Vol. 22, Part 1, 1979, pp. 23-52, pis. 2-6.]

24

PALAEONTOLOGY, VOLUME 22

abundant plant remains including sterile and fertile axes of rhyniophyte type as well
as Nematothallus Lang, Prototaxites and Pachytheca. The fossils are very frag-
mentary, indeed of the kind commonly cited in geological literature as ‘plant debris’,
but it is important that any plants of this age be carefully and critically described prior
to any speculation on the early history of land vegetation. In this study, the variety of
form in the sterile axes indicates a diversity of taxa hitherto unknown in British
Downtonian floras, while the occurrence of a large number of Cooksonia sporangia
permits, for the first time, a study of inter- and intraspecific variation in the genus.

OCCURRENCE AND PRESERVATION

Patches of drifted plant and animal fragments occur throughout a layer of grey-green
micaceous sandstone some 0-3 m thick within a typical O.R.S. red-bed sequence on
the north side of Freshwater East Bay (Nat. Grid Ref. SS09 0236 9812). Dixon (1921)
had placed these plant-bearing horizons at the top of his Basement Beds of the O.R.S.,
which he regarded as the Thyestes ( Auchenaspis ) stage of the Downtonian, a con-
clusion supported by King (1933) who equated the beds to the Ledbury marls and
sandstones (1.2) of the Welsh Borderland. Further support for a Downtonian age
comes from the spore assemblages of Richardson and Lister (1969). The strati-
graphy and sedimentology of the area are currently being re-investigated by Professor
J. R. L. Allen (Reading University) and Dr. B. P. J. Williams (Bristol University).
The plant-bearing horizon described here is in their third grey sandstone/sandstone
complex, which corresponds to Dixon’s unit 8, his third ‘green and grey band’ (Dixon
1921, p. 41).

The plants are preserved as coalified compressions {sensu Schopf 1975). Streaks of
iron oxide or, more rarely, iron sulphide are seen on some of the axes, but none are
completely petrified. There are two broad categories of axis: narrow parallel-sided
forms which, when they show dichotomous branching, are assignable to Hostinella
Barrande, and wider less regular types often with coarse surface striations, probably
of Prototaxites affinity. In both, a faint surface pattern is sometimes discernible under
low magnification in ordinary light, but the same specimens observed using a stereo-
scan electron microscope or metallurgical incident light microscope show no cellular
detail. The encrusting sheets of coalified material are less easily identified, being of
animal (for example, eurypterid) as well as plant origin. When recovered on film pulls,
the majority are found to be quite featureless, consisting of granular carbonaceous
material, but some show the reticulate patterning characteristic of cuticles of Nemato-
thallus Lang. These, together with the products of bulk maceration, will be described
elsewhere. Film pulls were also taken from Hostinella axes and certain sporangia.
The former were not particularly informative; cuticles were seldom seen although
occasionally narrow strands, longitudinally aligned and possibly representing the
remains of cell walls, were recovered. A few sporangia contained spores.

DIANNE EDWARDS: LATE SILURIAN FLORA

25

SYSTEMATIC PALAEONTOLOGY

Family rhyniaceae Kidston and Lang, 1920
Genus cooksonia Lang, 1937

Type species. Cooksonia pertonii.

Original diagnosis. Dichotomously branched, slender, leafless stems with terminal
sporangia that are short and wide. Epidermis composed of elongated, thick-walled
cells. Central vascular strand consisting of annular tracheids.

Background rationale. Although most of the terminal sporangia in the Freshwater
East assemblage are readily assigned to the genus Cooksonia , these have presented
considerable nomenclatural problems at the specific level. The simplicity of morpho-
logical organization of the fossils, together with their fragmentary nature and lack of
anatomy, results in relatively few characteristics for use in a systematic analysis and
may also result in the over-emphasis of minor morphological differences which them-
selves may not be real. It has been noticed, for example, that removal of a few grains
of rock or organic material can radically alter apparent sporangium shape, while part
and counterpart are occasionally not identical, because some organic material has
been lost on splitting the rock. Thus, before specific diagnoses are presented, I propose
to give a general description of all the Cooksonia sporangia present so that the range
in sporangium form may be fully appreciated.

General description. All the specimens were photographed and then drawn and
measured. In addition to sporangium width and height, the diameter of the axis at its
junction with the sporangium was noted and, in the case of a tapering axis, its
diameter when of constant width. Film pulls were then taken from the less well-
preserved specimens. The data are presented in Table 1. Accuracy of measurement
depended on the mode of preservation as well as on the shape of the sporangium and
its subtending axis. There were no difficulties when the sporangium was clearly
delimited being composed of a greater thickness of carbonaceous material than the
subtending axis nor where, although very little organic material remains on the fossil,
the rock area once covered by the sporangium was stained a different colour. Measure-
ment of sporangium height and axis width was most difficult in the tapering specimens
where the entire coalified compression had a uniform appearance or when little
organic material remained.

As a result of this analysis it was decided to split the eighty-three most complete
fertile specimens into two major groups based on sporangium shape. Thus in the
following description, Group I contains sporangia in which height is roughly equal
to width, and Group II has sporangia which are considerably wider than high. Each
group has been further subdivided into three, using characters of the subtending axis.
Representatives are drawn in text-fig. 1 . It must be emphasized that these categories
have been constructed purely for ease of handling numerous comparatively feature-
less fossils, and they should not at this stage be considered to have any taxonomic
significance. Perhaps more relevant to the latter are a small number of intermediate
morphology which do not fit exactly into any one of the six categories.

26

PALAEONTOLOGY, VOLUME 22

table 1 . Dimensions of fertile specimens assigned to the genus Cooksonia Lang.

Sporangium

Height (in mm)

Width (in mm)

Mean of

sporangium height

No. of

Category

Range

Mean

Range

Mean

sporangium width

specimens

Group I a

0-95-0-5

0-65

0-9 -0-4

0-66

0-99

10

Group lb

1-0 -0-28

0-49

0-96-0-3

0-51

0-98

22

Group Ic

0-95-0-2*

0-58

0-85-0-4

0-57

101*

15

Group II a

1-3 -0-3

0-52

1-77-0-35

0-72

0-73

14

Group II b

0-55-0-3

0-45

0-73-0-4

0-62

0-74

8

Group lie

0-8 -0-27

0-47

1-7 -0-4

0-8

0-63

14

Sporangium width

Axis diameter in mm

Axis width

Immediately below

Axis immediately Axis when

sporangium

When parallel sided

below sporangium parallel sided

No. of

Category

Range

Mean

Range

Mean

Mean Mean

specimens

Group la

0-38-0-1

0-17

-

-

4-08 4-08

10

Group lb

0-7 -0-15

0-27

-

-

1-98

22

Group Ic

0-4 -0-15

6-28*

0-2 -0-07

0-14

2-26* 4-8

12

Group I la

0-5 -0-06

0-19

-

-

4-23 4-23

14

Group lib

0-4 -0-2

0-37

-

-

2-09

8

Group lie

0-9 -0-15

0-33*

0-45-0-1

0-15

2-8* 5-7

11

* Approximate measurements.

Group /, type a (text-fig. 1a). Each sporangium is more or less circular in outline
(ratio of sporangium height to width ranges between 0-86 and 1-25) and is subtended
by a parallel-sided axis whose width is small compared with that of the sporangium
itself. The discrete shape of the two sporangia illustrated in Plate 2, figs. 1, 2, is typical
but some (e.g. PI. 2, fig. 4) have a less regular outline. Four specimens are branched
(e.g. PI. 2, fig. 9) the final dichotomy being some distance below the terminal sporan-
gia. The subtending axes are either quite straight (PI. 2, fig. 1) or gently flexuous.

The most important diagnostic feature of this category is the abrupt junction
between parallel-sided axis and presumably spherical sporangium. There are, how-
ever, a small number of specimens (e.g. PI. 2, fig. 3) where a very slight increase in axis
diameter occurs immediately below the sporangium.

Group II, type a (text-fig. 1e). This parallels Group la in axis characteristics but the
sporangia are more or less oval (elliptical) in outline. Relatively few specimens are
assignable to this category and these show considerable variation in morphology.
For example, the symmetry of one of the largest sporangia in the assemblage (PI. 2,
fig. 8) may be compared with the irregular shape of that illustrated in Plate 2, fig. 7.
The majority do not show branching, with just a short length of subtending axis
preserved. NMW 77.6G.113 (PI. 2, fig. 7) is a branching form with slender axes.

DIANNE EDWARDS: LATE SILURIAN FLORA

27

text-fig. 1. Line drawings of various shapes of
Cooksonia sporangia. All specimens x 14. a. NMW
77.6G.105, B. NMW 77.6G.41, c. NMW 77.6G.23,
d. NMW 77.6G.61, E. NMW 77.6G.13, f. NMW
77.6G.23, G. NMW 77.6G.6, H. NMW 77.6G.52,
i. NMW 77.6G.1 13, J. NMW 77.6G.27, K. NMW
77.6G.33. a is assigned to C. cambrensis sp. nov.
forma a; E to C. cambrensis forma /3; b-d, f-h are
assigned to C. hemisphaerica\ I, K to cf. C. cale-
donica and j is left as Cooksonia sp.

Branching is also seen in NMW 77.6G.21 (PI. 2, fig. 10) where only the sporangia and
distal axes appear to be completely preserved. Note that here the parallel-sided axes
expand very slightly immediately below the sporangia and thus are similar to the
specimens in Group la illustrated in Plate 2, fig. 3.

Group I, type b (text-fig. lc, d). Sporangia have circular outlines, but are sometimes
flattened at the junction with the subtending axis. The ratio of sporangium height to
width ranges between 0-83 and IT 8. The axis is wide at the point of attachment and
then tapers very slightly below. The diameter of the subtending axis when parallel-
sided is large when compared with sporangium width. The specimens are thus readily
distinguished from those in Group la. The sporangium itself is quite distinct, its
carbonaceous residues being much denser than those of the axis (PI. 3, fig. 2). Branch-
ing is common, the final dichotomy occasionally occurring immediately below the
sporangia (PI. 3, fig. 1) but more usually a short distance below (PI. 3, figs. 2, 3, 6).

Group II, type b (text-fig. If, g). Apart from the difference in sporangium shape, my
observations on Group lb are equally relevant here. These sporangia also appear to
be very well preserved, having entire, regular margins. They are not much wider than
the subtending axes so that, as in Group lb, the spore-containing regions are little
more than swollen extensions of the stems (PI. 3, figs. 4, 8, 9).

28

PALAEONTOLOGY, VOLUME 22

Group /, type c (text-fig. 1b). These may be distinguished from the specimens placed
in Group lb by the increased tapering of the subtending axes, which become parallel-
sided some distance below the sporangium junction. A typical example is illustrated
in Plate 4, fig. 2. This is a case where it was impossible to measure sporangium height,
unlike the specimen shown in Plate 4, fig. 4 where the sporangium is clearly delimited
from the tapering stalk. On such specimens the width of the axis at the sporangium
junction is found to be approximately the same as in Group lb. The majority of the
axes are unbranched. Most of the measured sporangia are more or less isodiametric,
but a specimen on NMW 77.6G.21 is extended vertically (PI. 4, fig. 3) while narrower
forms showing comparatively little increase in width are intermediate between
Groups la and Ic (PI. 4, fig. 6).

Group II, type c (text-fig. 1h-k). This is the category which shows greatest variation in
morphology and size. A very small number may be considered the equivalents of
Group Ic, i.e. oval sporangia with strongly tapering subtending axes with occasional
branching. An example with a particularly well-defined sporangium is illustrated in
Plate 4, fig. 8. Specimen NMW 77. 6G. 100a bears a sporangium with slightly irregular
outline (PI. 4, fig. 9), which is preserved as a highly coalifield compression, the inter-
stices between the granular carbonaceous material being filled with an orange sub-
stance, possibly limonite, the whole presenting a reticulate appearance. Some of this
material was scraped off and mounted on a slide. Oval to circular, often irregularly
shaped, light brown crystals were observed, but not spores. Examples of such crystals
are illustrated in Plate 3, fig. 13. In the specimens already described, the junction
between sporangium and stalk is seen as a straight line. There is, however, a small
number of oval sporangia in which the distal part of the stalk extends into the base of
the sporangium cavity so that a convex line marks the junction (PI. 4, figs. 10, 14, 15).
The representative of this type illustrated in Plate 4, fig. 14 is an unusual branching
specimen in that the sporangium is overtopped by the second branch of the ultimate
dichotomy.

Finally there are a few oval sporangia with distinct borders. The example illustrated
in Plate 4, figs. 16, 17 has a short comparatively wide stalk ( ?intermediate with
Group Ic), the junction with the sporangium being almost straight. The central,

EXPLANATION OF PLATE 2

Figs. 1-10. Cooksoniacambrensis sp. nov. 1. NMW 77.6G.105, holotype forma a, x 15. 2, NMW 77.6G.21,
formaa, xl5. 3,NMW77.6G.3,formaa, > 15. 4, NMW 77.6G.10, forma «, x 15. 5, NMW 77.6G.1 13,
branching forma j3, x 15. 6, NMW 77.6G.60, forma j8, 15. 7, NMW 77.6G.32a, cf. C. cambrensis

forma /} with irregularly shaped sporangium, • 15. 8, NMW 77.6G. 13, holotype forma yS, x 15. 9, NMW
77.6G.24, branching forma a, x 13-5. 10, NMW 77.6G.21, branching forma y3, x 13-3.

Figs. 11-14. Terminal sporangia of uncertain affinity. 11, NMW 77.6G.21, elongate sporangium, x 15.
12, NMW 77.6G.108, elongate sporangium with border, x 15. 13, NMW 77.6G.109a, cf. Salopella sp.,
x 20. 14, fragments of smooth-walled spores recovered on film pull from specimen illustrated in fig. 13,
x 340.

PLATE 2

EDWARDS, Late Silurian flora

30

PALAEONTOLOGY, VOLUME 22

presumably spore-containing, region of the sporangium consists of dense carbona-
ceous material which is surrounded distally by a narrow band of less dense material.
This border becomes narrower near the junctions with the axis. A similar structure is
seen partially surrounding a possible sporangium in Plate 4, fig. 18.

Spores. As film pulls result in the almost complete removal of the sporangium from
the rock, they were usually taken from less well-preserved specimens. Small amounts
of coalified material were removed from the more morphologically informative
sporangia and oxidized in Schulze’s solution. In three instances, chips of rock bearing
sporangia were mounted on stubs, coated with gold and examined using a stereoscan
electron microscope. The film-pull method was the only successful one, with spores
recovered from six specimens. In all cases they are fragmentary and have smooth
walls. The diameters of the more complete spores (i.e. those where half or more than
half the spore is present), together with some information on sporangium type, is
given in Table 2. Note that spore size in the globular (Group I) sporangia is more or
less uniform (PI. 3, figs. 10-12) but the spores from the single oval sporangium
(Group II) are markedly larger (PI. 3, figs. 14, 15).

table 2. Spore dimensions from sporangia of Cooksonia type.

Specimen

Range in spore

Mean

No. of spores

Plate 3,

number

Category

size (^m)

(/u.m)

measured

fig-

NMW 77.6G.72

I b (branching)

24-0-30-0

27-0

7

10, 12

NMW 77.6G.17

I b (branching)

22-5, 24

-

2

11

NMW 77.6G.1 1 1

Ic

25-5-37-5

29-5

14

NMW 77.6G.38

I

22-5, 22-5

-

2

NMW 77.6G.112

I

28-5-34-5

30-6

5

NMW 77.6G.53

II (not c)

46-5-54-0

50-4

5

14, 15

Cooksonia hemisphaerica Lang

Plate 3, figs. 1-12, Plate 4, figs. 1-8; text-fig. 1, b-d, f-h.

Amplified diagnosis. Erect part of plant consists of dichotomously branching axes between 0-25 and 1-6 mm
wide. Smooth axes terminate in sporangia of variable size and morphology including hemispherical, spheri-
cal, and elliptical forms. Sporangia are 0-3-2-0 mm wide and 0-2-2-0 mm high. Axes gradually increase in
width below sporangia. Plant homosporous; spores smooth, circular, 22-5 ^m-34-5 /urn in diameter.

EXPLANATION OF PLATE 3

Figs. 1-3, 5-7. Cooksonia hemisphaerica sensu Lang 1937. All specimens x 15. 1, NMW 77.6G.17a.
2, NMW 77.6G.6a. 3, NMW 77.6G.23b. 5, NMW 77.6G.23a. 6, NMW 77.6G.1. 7, NMW 77.6G.94,
unusually large, unbranched form.

Figs. 4, 8, 9. C. hemisphaerica: oval forms. 4, NMW 77.6G.6a, x 15. 8, NMW 77.6G.75b, xl5. 9, NMW
77.6G.27a, x27.

Figs. 10-15. Film pulls from Cooksonia sporangia. 10, NMW 77.6G.72a; (F.P. (i)), fragmentary spores
from isolated circular spore mass, x 320. 11, NMW 77.6G.17b (F.P. (i)), smooth spores from C. hemi-
sphaerica sensu Lang, x 320. 12, NMW 77.6G.72a (F.P. (i)), fragmentary spores from C. hemisphaerica
sensu Lang, * 288. 13, NMW 77. 6G. 109a (F.P. (i)), crystals of various sizes (arrowed) recovered from
Group 1 type Cooksonia sporangium, x 640. 14, 15, NMW 77.6G.53 (F.P. (i)), larger smooth-walled
spores from oval (Group II) Cooksonia sporangium, x 1 32.

PLATE 3

EDWARDS, Cooksonia

32

PALAEONTOLOGY, VOLUME 22

Holotype. V 58012 (Lang No. 181) illustrated in Lang 1937, pi. 9, figs. 31, 32.

Holotype locality. Quarry in Targrove beside drive to Targrove Hall.

Age. Downtonian.

Of the Freshwater East specimens described above, those in Groups I b, Ic, II b, and a small number in
Group lie will be included in C. hemisphaerica. Also included in the amplified species are the plants
illustrated by Obrhel (1962, pi. 1, figs. 1-3, identified as cf. C. hemisphaerica) and certain of Ishchenko’s
specimens (e.g. those illustrated in Ishchenko 1975, pi. 14, figs. 1, 2, and 5). Excluded are the specimens
described as C. hemisphaerica by Ananiev and Stepanov (1969).

Description and discussion. This was the only existing species of Cooksonia that it was
immediately possible to recognize in the assemblage. The sporangia I have described
in Group I b (PI. 3, figs. 1, 2, 5-7) are directly assignable to C. hemisphaerica , which
Lang erected for a small number of specimens from a single locality, a quarry in
Targrove near the top of the Downtonian. He described the terminal sporangium as
hemispherical, almost as high as wide, with thick walls and flat base. He illustrated
three specimens (Lang 1937, pi. 9, figs. 31-36), all of which resemble the Freshwater
East Group lb examples in over-all morphology although the sporangia of the latter
are sometimes isodiametric or slightly taller than wide. The subtending axis in the
Lang specimens is wide at the point of attachment and then tapers. Lang considered
the sporangial width to be approximately three times the diameter of the axis, which
he presumably measured some distance below the junction. This compares favour-
ably with some of the Welsh specimens. All of Lang’s examples are larger, the biggest
sporangium being just under 2 mm high and just over 2 mm wide, while the smallest,
which he considered immature, 1 mm wide and 0-96 mm high. Although he observed
a central strand in one of his fertile specimens, this proved structureless. It is of great
botanical interest that it was in sterile axes associated with these sporangia that he
demonstrated tracheids with annular thickening, which he recognized as ‘the most
ancient piece of vascular tissue as yet demonstrated in position in a fossil plant in
Britain’ (p. 256).

Branching in the Freshwater specimens, like that in Lang’s, is dichotomous.
Neither he nor I have seen a pseudomonopodial branching system with dicho-
tomously branching ‘laterals’ similar to that in Ananiev and Stepanov’s reconstruc-
tion of C. hemisphaerica based on Lower Devonian specimens from Siberia (Ananiev
and Stepanov 1969). I agree with Gensel (1976) that the Russian plants have much
more in common with her Renalia than with the relatively simple dichotomizing
Downtonian species, and perhaps should be excluded from the genus Cooksonia.

One of the most informative specimens in the assemblage is illustrated in Plate 3,
fig. 5. It comprises a branching axis in which the left-hand fork terminates in a globular
sporangium typical of C. hemisphaerica , but the right-hand one bears an elliptical
sporangium of the type assigned to Group II b. (The illustration unfortunately gives
the impression that the right-hand subtending axis is parallel sided, because some
coalified residues have flaked off where it widens below the sporangium). Thus,
assuming that both sporangia are completely preserved, this is a demonstration of
intraspecific variation in sporangia of C. hemisphaerica. There is an alternative inter-
pretation which I find less plausible. It is possible that each sporangium was ellip-
soidal in life so that, depending on its orientation on burial, compression would result

DIANNE EDWARDS: LATE SILURIAN FLORA

33.

in either a circular or an elliptical shape. I consider it more satisfactory to extend the
diagnosis of C. hemisphaerica to include sporangia which are much wider than Lang
originally described (i.e. oblate spheroids in life) as well as the more of less spherical
forms. In both, the subtending axis is wide at the sporangium junction and tapers
only a little below.

As already mentioned, specimens assigned to Group lb (now considered C. hemi-
sphaerica) and Group Ic differ only in the degree of tapering of the fertile axis and in
the distance of the ultimate branching point from the sporangia, and some specimens
may be of intermediate morphology (PI. 4, fig. 3). Indeed it is possible to assemble a
continuous series from sporangia with axes which taper to less than half the sporan-
gial width, to those in which there is little change in axis diameter. Comparing only
sporangial dimensions in Group lb and Ic, there is little difference in the ranges of
height and width (Table 1), although the means are slightly higher in Group Ic. The
diameter of the axis at the sporangium-stalk junction is more or less the same. I
believe, therefore, that specimens assigned to Group Ic should also be included in
C. hemisphaerica. Some support for this comes from Lang’s own specimens, which
also show variation in axis size, although his figured examples are too few to allow
satisfactory comparison.

My observation that sporangia of C. hemisphaerica sensu stricto Lang occur close
to branching points while those of Group Ic terminate narrower unbranched axes or
are some distance away from a fork, leads to the further speculation that the differ-
ence may be a developmental one— i.e. that Group lb specimens are younger than
Group Ic.

Although Cooksonia with its smooth axis is not directly comparable to any extant
pteridophyte, it seems likely that its mode of growth was similar, and that each naked
aerial axis would have possessed an apical cell or group of initials, which at a branch-
ing point would have divided equally to produce two new meristems. The onset of
reproductive activity (sporangia formation) would have resulted in the cessation of
further growth of that particular axis. Thus, considering a fertile specimen such as
NMW 77.6G.17 (PI. 2, fig. 2), dichotomous branching would be followed by a short
period of vegetative activity (involving cell division at the extreme apex and tissue
differentiation and extension behind), after which the vegetative apex would be
converted into sporangial initials and growth would cease. In NMW 77.6G.23 (PI. 2,
fig. 5) the period of vegetative activity would have been much longer, while in
NMW 77.6G.6 (PI. 3, fig. 1) sporangia were produced almost immediately after
branching. On this model, sporangia in Group Ic would develop after a considerable
period of vegetative activity. My hypothesis involves a slight modification of this
determinate growth pattern in that I suggest that the first three examples described
above are young fertile specimens preserved soon after sporangium determination,
while sporangia on tapering narrower axes are mature ones, the thinner axes resulting
from purely extension growth. In support of this speculation are my observations
that sporangia in Group lb almost always appear entirely preserved, while those in
Group Ic tend to be larger on average and less regular in shape, perhaps indicating
some disintegration at maturity. I have found only one completely fertile specimen
(NMW 77.6G.1) which shows unequal development of the products of an ultimate
branch (PI. 3, fig. 6). Here the right-hand branch is considerably longer and thinner

34

PALAEONTOLOGY, VOLUME 22

than the other, possibly indicating that it had begun extension growth. It is, of course,
equally possible that the meristem divided unequally and that the narrower right-
hand branch subsequently grew more rapidly or for a longer time before determina-
tion of the sporangium caused growth to cease. The sporangia themselves are of
different shape, the more globular right-hand one unfortunately being less com-
pletely preserved.

Such delayed extension growth is seen during the maturation of the strobilus axis
in certain lycopods, although not in the sporangial stalks themselves. A far more
striking demonstration is found in the sporophytes of hepatics, where the cells of the
unbranched seta (sporangium stalk) are differentiated in the embryo and further
growth is almost completely by cell elongation. I consider it unlikely that an entire
Cooksonia sporophyte would develop in this way.

Although I have already included Groups lb and c and Group II b in C. hemi-
sphaerica, I am reluctant to extend the specific concept still further to encompass the
more heterogenous Group lie, except for a small number of intact oval sporangia
with tapering axes with sporangium height/width ratio similar to that in Group lib
(PI. 4, fig. 8). I would also include the few specimens of similar size and morphology
which have rather irregular distal margins and which are possibly older, empty
sporangia.

The rather limited information from in situ spores lends some support to my con-
clusions, because spore diameter in sporangia assigned to Groups lb and c (i.e.
C. hemisphaerica ) plus two undetermined globular sporangia is more or less the same
(PI. 3, figs. 10-12); the undetermined, irregular Group II sporangium which I would
not include in C. hemisphaerica on morphological grounds, has markedly larger
spores (PI. 3, figs. 14-15).

I feel less confident in assessing the extent of intraspecific variation in the remaining
specimens, yet I am reluctant to erect numerous species based on very limited numbers

EXPLANATION OF PLATE 4

Figs. 1-7. Tapering, globular forms of Cooksonia hemisphaerica , all 15. 1, NMW 77.6G.41. 2, NMW

77.6G.36. 3, NMW 77.6G.37a. 4, NMW 77.6G.42, intermediate form with C. hemisphaerica sensu
Lang, 1937. 5, NMW 77.6G.68a. 6, NMW 77.6G.63, possible intermediate form with C. cambrensis.
7, NMW 77.6G.21.

Fig. 8. NMW 77.6G.52, C. hemisphaerica (oval form), x 15.

Fig. 9. NMW 77.6G.109, Cooksonia sp. with reticulate patterning, x 10.

Fig. 10. NMW 77.6G.3, ■ 15.

Fig. 11. NMW 77.6G.33, cf. C. caledonica , 15.

Fig. 12. NMW 77.6G.10, cf. C. pertonii, x 15.

Fig. 13. NMW 77.6G. 114, C. pertonii from type locality at Perton Lane, Hereford, x 3-8.

Fig. 14. NMW 77.6G.74, Cooksonia sp., x 15.

Fig. 15. NMW 77.6G.113, cf. C. caledonica (specimen subsequently destroyed), x 15.

Figs. 16, 17. NMW 77.6G.27a and b, Cooksonia with border, x 15.

Fig. 18. NMW 77.6G.35, ? sporangium with border, x 15.

Figs. 19-23. Axes with enations assigned to Psilophytites sp. 19, 20, NMW 77.6G.32a. 19, unbranched
axis, ■ 10-2. 20, single spine from 19 enlarged, x 33-6. 21, NMW 77.6G.69a, unbranched axis with
truncated spines, x 7-2. 22, NMW 77. 6G.69b, single spine from counterpart of fig. 21, x21-4. 23, NMW
77.6G.88, branching axis with crowded spines, x 7.

PLATE 4

EDWARDS, Cooksonia and Psilophytites

afck t ■

36

PALAEONTOLOGY, VOLUME 22

of fossils, often of an exceedingly fragmentary nature. There are, however, a few
sporangia sufficiently distinct to merit standard nomenclatural treatment (e.g.
C. cambrensis sp. nov.). Others have some characteristics in common with, although
not identical to, existing species (e.g. C. pertonii ) while the remainder present new
combinations of characters but occur in insufficient numbers to allow specific
diagnosis.

Cooksonia cambrensis sp. nov.

Plate 2, figs. 1-10; text-fig. 1a, e.

Diagnosis. Aerial part of plant consisting of presumably erect, smooth axes 0-5-0-06 mm wide with occa-
sional dichotomous branching. Terminal sporangia are circular to elliptical in outline, with subtending axes
parallel-sided or increasing slightly in diameter immediately below the sporangium. Forma a includes
sporangia circular in outline, 0-95-0-5 mm high and 0-9 to 0-4 mm wide. Forma |8— sporangia ranging
between elliptical, hemispherical, and irregular in outline, 1 -3—0-3 mm high and 1-77 to 0-35 mm wide.

Holotypes. Forma ct-NMW 77.6G.105 (PI. 2, fig. 1). Forma /3-NMW 77.6G.34 (PI. 2, fig. 8).

Type locality. Foreshore on north side of Freshwater East Bay, near Pembroke, Dyfed. Nat. Grid. Ref.
SS09 0236 9812.

Horizon. Lower Red Marl Group, early Downton (approx, equivalent to Pridoli).

Specific derivation. From Cambria (Wales). Forma a includes specimens described in Group I a above.
Forma /3 includes specimens described in Group II a.

Description and discussion. This will include the specimens described in Group 1
where the sporangia are circular in outline and subtended by straight parallel-sided
axes. The majority of specimens are unbranched but, where branching does occur
(e.g. PI. 2, fig. 9), it is some distance away from the sporangia. As I have no unequivo-
cal evidence that the branched and unbranched specimens belong to the same plant,
this is a case where it could be argued that the specimens showing no branching and
in which vascular tissue has not been demonstrated should not be assigned to the
genus Cooksonia. I believe that there are sufficient similarities in the characteristics
of the sporangium and subtending axis to support the use of the genus, f have
described similar organization in specimens in Group He, although here the sporangia
have more or less elliptical outlines. The outstanding example is on specimen NMW
77.6G.34 (PI. 2, fig. 8), notable for its large size, regular outline, and robust axis. The
remaining specimens are less well preserved. Some have sporangia with flattened
bases and hence hemispherical outlines. Again the majority are unbranched and,
where branching is present, the axis is much narrower. 1 include these Group 1 c
specimens in the new species but, to distinguish the two major sporangium shapes, I
designate the circular ones as forma a and the elliptical, forma /3.

1 also include those specimens where there is a slight increase in axis diameter
immediately beneath the sporangium, as opposed to the gradual increase typical of
C. hemisphaerica. An example of a globular form is illustrated in Plate 2, fig. 3, and
an elliptical one in Plate 2, fig. 10. 1 have considered the possibility that these speci-
mens are intermediate between C. hemisphaerica as defined above and C. cambrensis ,
which would in turn even further extend the specific concept of C. hemisphaerica. The
important diagnostic feature here is the area of contact between sporangium and

DIANNE EDWARDS: LATE SILURIAN FLORA

37

subtending axis, which is more or less the same (and extensive) in all forms of C. hemi-
sphaerica but small in C. cambrensis.

Affinities of the remaining Cooksonia sporangia

1. Comparison with C. pertonii. I have not found any sporangia which are im-
mediately recognizable as C. pertonii Lang, although there is a very small number
included in my Group lie which show some similarity to the less typical examples in
Lang’s collection made from various localities throughout the Welsh Borderland.
Little detail is visible on his plates, but an examination of his figured specimens
housed at the British Museum and my own material collected from the type locality,
Perton Lane, reveals that the majority of sporangia are considerably wider than high,
resulting in a strongly flattened appearance (PL 4, fig. 13). The junction between
sporangium and subtending axis is long and almost straight, so that the tip of the
axis is nearly as wide as the sporangium itself, but it usually tapers rapidly (e.g.
Lang’s plate 8, fig. 7) although occasionally it remains more or less the same width
(Lang’s plate 8, fig. 12). The strongly flattened sporangia from Dyfed (e.g. NMW
77.6G.10 illustrated in PI. 4, fig. 12) are much smaller than any of Lang’s and so
fragmentary that they are better left as cf. C. pertonii.

2. Comparison with C. caledonica. The sporangia in this Scottish Dittonian species
are also variable in size and shape, but most are slightly wider than high. They too have
tapering dichotomously branching axes which extend into the bases of the sporangia
to varying degrees so that in extreme cases sporangium shape is reniform. The
sporangia are further characterized by a narrow distal border (Edwards 1970).

There are no unequivocal representatives of C. caledonica at Freshwater East.
A curved sporangium stalk junction is present in some Group lie specimens (e.g.
PI. 4, figs. 10, 1 1, 14, 15). Of these NMW 77.6G.1 13 (PI. 4, fig. 15), later sacrificed for
spore preparations, is closest to the Scottish species and will be assigned to cf. C. cale-
donica. NMW 77.6G.3 (PI. 4, fig. 10) will be left as Cooksonia sp. as will NMW
77.6G.74 (PI. 4, fig. 14). The latter is unique in that although one of the daughter
branches of the ultimate dichotomy terminates in a sporangium, the other is ap-
parently sterile and considerably longer : an example of overtopping following dicho-
tomous branching.

NMW 77.6G.33 (PI. 4, fig. 1 1) has, in addition to a curved junction, a very narrow
border represented by a strip of easily removed coalified material and will be called
cf. C. caledonica. A much wider, less heavily coalified border is present on NMW
77.6G.27. On the counterpart, the sporangium stalk junction is slightly curved, but
the over-all morphology of the specimen— an oval sporangium borne on a short stout
axis— more closely resembles C. hemisphaerica (Group II b) than C. caledonica (PI. 4,
figs. 16, 17).

It therefore seems likely that there was more than one kind of plant, in which
sporangia had distinctive margins, in the Downtonian. This is also seen in the Lower
Devonian, where sporangia with borders are not unusual, e.g. Gosslingia Heard,
Zosterophyllum Penhallow, Cooksonia crassiparietilis Yurina and Cooksonia sp.
(Croft and Lang 1942). In the first two genera, the rim is more heavily coalified than
the central region and is believed to be involved in sporangial dehiscence (e.g.

38

PALAEONTOLOGY, VOLUME 22

Edwards 1969). A less dense rim, similar to those illustrated in Plate 4, figs. 16-18,
has been recorded for undetermined Cooksonia specimens from the Brecon Beacons
(Croft and Lang 1942) and here it is more likely that the border results from the com-
pression of an almost spherical organ where there was a greater thickness of organic
material in the central region, being composed of sporangial contents plus wall, than
at the periphery where only the wall was compressed.

INCERTAE SEDIS

Genus tortilicaulis gen. nov.

Type species. Tortilicaulis transwalliensis sp. nov.

Diagnosis. Fragments of plants consisting of unbranched presumably upright axes terminating in elongate,
fusiform to oval bodies interpreted as sporangia. Axes show occasional twisting especially immediately
below sporangia.

Tortilicaulis transwalliensis sp. nov.

Plate 5, figs. 1-12; text -fig. 2

Diagnosis. Characters as in generic diagnosis. Axes 0-4-0-1 mm wide and at least 10 mm long. Terminal
sporangia 3-9-1 T mm high and 1-3-0-38 mm wide.

Holotype. NMW 77.6G.2 deposited at National Museum of Wales.

Holotype locality. Foreshore on north side of Freshwater East Bay, near Pembroke, Dyfed. Nat. Grid. Ref.
SS09 0236 9812.

Horizon. Lower Red Marl Group, early Downton (approximately equal to Pridoli).

Derivation of name. Generic name Tortilicaulis is derived from the Latin adjective ‘tortilis’ meaning twisted
and noun ‘caulis’ meaning stem. Specific epithet is from ‘Transwallia’ the Latin name for Pembroke
(= across Wales).

Description. Elongate, fusiform to oval bodies attached to smooth stout axes are
occasionally found. They are two to three times longer than wide and quite variable in
shape. Although all attempts to isolate spores have failed, these elongate structures
are assumed to be terminal sporangia. Branching has not yet been recorded in the
subtending axes, which often appear to be twisted.

Some indication of the variability in sporangium shape and size (particularly the
length/width ratio) and in the morphology of the apex may be seen in text-fig. 2 and
Plate 5. Of the long and comparatively narrow sporangia, some have typically
attenuated tips and may be described as fusiform (text-fig. 2a, o and PI. 5, fig. 7) while
others are more bluntly rounded (text-fig. 2f, PI. 5, fig. 12). This difference is paralleled
in the wider sporangia which may be either ovate (text-fig. 2b, c, PI. 5, fig. 2) or
ellipsoidal (text-fig. 2l, PI. 5, fig. 10) depending on the shape of the distal regions.
Whether or not such distinctions are real is debatable, as the circumscription of the
sporangium apex is often hampered by lack of organic material on the fossil and by
staining of the rock (itself rich in disseminated carbon) in the immediate vicinity of the
fossil. Table 3 shows that sporangia with rounded apices tend on average to be shorter
than those with attenuated tips. Unfortunately the sample is too small for this to be
of any significance. Indeed, evidence for a considerable size range in sporangia of

DIANNE EDWARDS: LATE SILURIAN FLORA

39

table 3. Dimensions of most complete specimens of Tortilicaulis
transwalliensis gen. et sp. nov.

Sporangium Axis

Sporangial shape

Height
in mm

Width
in mm

Width in mm

A. Fusiform

NMW 77.6G.7

2-3

0-67

01

NMW ll.6G.il

1-5

0-45

0-15-^0-18

NMW 77.6G.2

3-9

0-95

0-4

Means for 3 specimens

2-57

0-69

0-22

B. Fusiform with rounded tips

NMW 77.6G.7

1-7

0-5

01

NMW ll.6G.5a

1-3

0-38

0-1

NMW 77.6G.3

1-95

0-7

0-3

NMW 77.6G.47

1-5

0-55

01

Means for 4 specimens

1-6

0-53

015

C. Oval with attenuated tips

NMW 77.6G.75a

3-5

1-3

0-2

NMW 77.6G.23b

2-4

10

0-3

NMW 77.6G.22a

1-9

0-8

0-1

NMW 77.6G.74

2-8 f

11

0-3

NMW 77.6G.49

1-6

0-65

0-2

Means for 5 specimens

2-44

0-97

0-22

D. Oval with rounded tips

NMW 77.6G.81

2-0

0-75

0-2

NMW 77.6G.5a

115

0-55

0-3

E. Oval with mucronate tips

NMW 77.6G.37c

2-0

0-8

018

NMW 77.6G.1

1-45

0-6

0-20

NMW 77.6G.85

2-3

10

-

Means for 3 specimens

1-92

0-8

019

Over-all means (17 specimens)

207

0-80

0-20

identical shape is apparent in text-fig. 2h and n. Here the broad sporangia have
distinctive, almost mucronate, tips (see also PI. 5, figs. 6, 8, 9).

Accurate measurement of sporangial length was most difficult in the fusiform types
where the sporangial bases taper gradually into the subtending axes. In the elliptical
to ovate forms the junction is more obvious. Certain of the sporangia are slightly
asymmetrical at the base, a condition often related to the twisting of the axis im-
mediately below the sporangium (text-fig. 2h, PI. 5, figs. 2, 8, 9, 11). In specimen
NMW 77.6G.85 (PI. 5, fig. 9) a definite constriction of the axis is visible in this region,
but in others the axis is actually broken (text-fig. 2j, n and PI. 5, fig. 7).

Subtending axes are of varying length, the longest being slightly more than a centi-
metre (PI. 5, fig. 1). In contrast to what is seen in Cooksonia, branching has not been
observed in these axes. Although they are more or less parallel-sided, they do not give
an impression of stiffness or rigidity : indeed some are conspicuously curved. A dis-
tinctive feature is the twisting mentioned above. When present, this normally occurs

40

PALAEONTOLOGY, VOLUME 22

just below the sporangium, but some axes are further twisted proximally as in text-
fig. 2b, PI. 5, fig. 5. This specimen shows a definite constriction as opposed to folding,
at each twist.

Very little has been discovered about the anatomy of the specimens. In the sporan-
gia, carbonaceous residues either occur as flat sheets exhibiting clear fracture, or are
granular. Spores were not seen on amyl acetate film pulls nor when small fragments
were oxidized in Schulze’s solution. Some specimens have the reticulate appearance

text-fig. 2. Line drawings of Tortilicaulis transwalliensis gen. et sp. nov. sporangia to show range in shape.
All specimens x 10. Fusiform types: a. NMW 77.6G.2, G. NMW 77.6G.77, o. NMW 77.6G.13. Fusiform
with rounded tips: f. NMW 77.6G.7, i. NMW 77.6G.5a, K. NMW 77.6G.3. Oval with attenuated tips:
b. NMW 77.6G.49, c. NMW 77.6G.75a, D. NMW 77.6G.48, E. NMW 77.6G.23b. Oval with rounded
tips: L. NMW 77.6G.81, p. NMW 77.6G.5a. Oval with mucronate tips: H. NMW 77.6G.85, J. NMW
77.6G.1, N. NMW 77.6G.37c. M is a specimen with tip missing. NMW 77.6G.44. Arrows indicate twisting
immediately below sporangium.

EXPLANATION OF PLATE 5

Figs. 1-13. Sporangia of Tortilicaulis transwalliensis gen. et sp. nov. 1, NMW 77.6G.2, holotype, x 13-5.
2, NMW 77.6G.75a, 15. 3, inset, NMW 77.6G.75a, specimen illustrated in fig. 2 after all organic

material had been removed and photographed using unilateral illumination. Black line indicates approxi-
mate limits of sporangium, x 2-5. 4, NMW 77.6G.23b, robust sporangium with reticulate appearance,
x 15. 5, NMW 77.6G.49, small sporangium with long unbranched twisted stalk, x 15. 6, NMW
77.6G.49, faintly striated sporangium with mucronate tip and constriction immediately below sporan-
gium, xl5. 7, NMW 77. 6G.7, fusiform sporangium, x 15. 8, NMW 77. 6G.1, sporangium with oblique
striations and break at sporangium/axis junction, x 23-25. 9, NMW 77.6G.85, sporangium with distinct
mucronate tip, x 15. 10, NMW 77.6G.81, note reticulate pattern and asymmetric base, x 15. 11, NMW
77.6G.5a, small sporangium with broad twisting stalk, x 15. 12, NMW 77.6G.7, fusiform sporangium
with rounded tip and twist immediately below, x 15. 13, NMW 77.6G.37a, isolated oval mass with
reticulate pattern, x 12-4.

Figs. 14-18. Pellia epiphylla, an extant liverwort. 14, detached mature seta of sporophyte showing twisting,
x2. 15, part of mature seta with twisting, x 8. 16, cellular detail at twist, x 12-5. 17, capsule of sporo-
phyte with twisted seta immediately below, x 12-5. 18, scanning micrograph of seta below capsule, x22.

PLATE 5

EDWARDS, Tortilicaulis and Pellia

42

PALAEONTOLOGY, VOLUME 22

already described for certain Cooksonia sporangia (PI. 5, figs. 4, 10). Here also,
yellow-brown crystals together with some carbonaceous material were recovered from
the film pulls. Also present in the matrix are discrete oval to fusiform bodies similar in
size to the sporangia but lacking subtending axes (e.g. PI. 5, fig. 12). These again failed
to yield spores.

In a few cases, after most of the carbonaceous material had been removed from the
sporangia on film pulls, obliquely running striations were visible on the rock beneath.
This is most clearly seen in the specimen illustrated in Plate 5, fig. 7. Very fragmentary
strands were recovered on the film pulls themselves. If these are the remains of cell
walls, it would suggest that some of the cells of the sporangium wall were spirally
orientated. Specimen NMW 77.6G.75 (PI. 5, fig. 2), when viewed using unilateral
illuminations, has a distinctly corrugated appearance, the interrupted ridges again
being oblique. When all organic material was removed, it was noted that the rock
itself shows a similar patterning (PI. 5, fig. 3).

No structural detail is known for the axes and central strands have not been seen.
It would be inadvisable on this evidence, however, to conclude that the plants were
not vascular, as strands are only rarely present in the associated axes of Cooksonia.

Discussion. My investigations on this group are somewhat disappointing in that the
critical information necessary for a comprehensive and conclusive discussion as to its
affinities has not been discovered. For example, having failed to extract spores, I can
only assume that the elongate bodies are sporangia. I should like to have found
a much larger number of very long twisted axes before concluding that they are
unbranched— absence of tracheids or even a central strand does not permit the con-
clusion that the plants were non-vascular, especially as I failed to demonstrate
tracheids in the Cooksonia and Hostinella axes in the assemblage. Such limitations
should be borne in mind throughout the following discussion.

Elongate sporangia terminating naked axes are characteristic of certain members
of the Rhyniaceae ( sensu Banks 1975). They include Rhynia Kidston and Lang;
Horneophyton Barghoorn and Darrah; Eogaspeseia Daber; Steganotheca Edwards
and Richardson and Eorhynia Ishchenko. Dichotomous branching has been recorded
in all these genera, sometimes very close to the fertile region. Only the first three are
known to have had vascular tissue. The apparent absence of branching in the Fresh-
water East specimens separates them from these rhyniophytes.

Long unbranched naked axes terminating in elongate sporangia typify the Lower
Devonian genus Sporogonites Halle. The type species S. exuberans originally
described from Roragen, Norway (Halle 1916, 1936) and since found in Belgium,
Wales, and France, has large sporangia up to 9 mm long and 2 to 4 mm in diameter
at the widest point. The smooth unbranched parallel-sided axes are approximately
0-5 mm wide and up to 10 cm long. In the original specimens, the sporangium apex
is described as rounded, but the Belgian specimens (Stockmans 1940) have more
pointed tips. The sporangium, of which only the upper half is considered spore-
bearing, tapers gradually into the subtending axis. The surface of the sterile basal
region has heavy longitudinal folding in some specimens, but this feature is lacking in
more compressed fossils. Croft and Lang (1942) described stomata-like structures on
the lower parts of semi-petrified Welsh sporangia. Two further species have since been

DIANNE EDWARDS: LATE SILURIAN FLORA 43

erected : S. chapmanii Lang and Cookson from Australia and S. excellens Frenguelli
from Argentina. Halle tentatively suggested affinities with the psilophytes but this
view was radically altered following Andrew’s reconstruction based on Belgian
material (Andrews 1960). Andrews described several fertile axes aligned in parallel
which appeared to be attached to a carbonaceous film. He interpreted this as a
bryophyte-like thallus to which several sporophytes were attached.

The Freshwater East specimens have much in common with Sporogonites. Al-
though much smaller they are similar in shape, particularly when comparison is made
with the Belgian S. exuberans, where the junction between sporangium and stalk is
more distinct. Neither longitudinal folds nor sterile basal region are evident on the
Welsh specimens, but this could result from preservation differences. The axes of
Sporogonites are narrower (relative to sporangium width), longer, and straighter. It
is interesting to note that one of Stockman’s specimens refigured by Hoeg (Hoeg
1967, fig. 161a, p. 242) has a fold or twist some distance below the sporangium.

On the rather limited morphological data presented above, I conclude that these
Downtonian sporangia have some affinity with Sporogonites, but as they do not
show the well-defined characteristics of that genus, they should be placed in a new
genus.

Systematic position o/Tortilicaulis. As mentioned earlier, the possibility that Tortili-
caulis was a vascular plant cannot be entirely eliminated. Should this have been the
case it would be assignable to the Rhyniaceae in the Rhyniophytina Banks, 1975.
Consideration of possible bryophyte affinities is prompted by its similarities with
Sporogonites, in addition to the twisted nature of its axes. This latter feature is a
characteristic of the mature setae (sporophyte stalks) of certain mosses and liverworts.
Thus in Pellia the young sporophyte has a short stalk which elongates considerably
due to rapid water uptake and at the same time twists (PI. 5, figs. 14-18). The mature
rather flaccid, twisted seta is hollow and when immersed in water (as would occur
during the initial stages of fossilization) it does not straighten out. A compressed
mature sporophyte of Pellia would thus look remarkably similar to Tortilicaulis
except that the capsule in Pellia is spherical. The majority of capsules in the Junger-
manniae are, however, ovoid to cylindrical, but modern forms are rarely indehiscent
and there are no indications of a valvate construction in the fossils. Indeed the chances
of such a delicate organ as a liverwort seta being fossilized must be very slim. Schuster
(1966, p. 583) describes the seta of the Jungermanniae as ‘an exceedingly ephemeral
structure owing to the delicacy of its cells’. He points out that they have no intrinsic
rigidity because all wall thickening is absent and there is no cuticle. The outermost
cells of the mature hollow seta are obviously more robust than the remainder.

The seta of a moss is a much more resilient organ and some, for example Discelium
nudum, are quite regularly twisted as a result of a spiral growth process. Herbarium
material kindly donated by Dr. J. Duckett (Bangor) was immersed in water and the
setae straightened out. Duckett considers this to occur only in young setae and that
mature ones remain untwisted on rewetting.

Tortilicaulis thus has some features in common with bryophyte sporophytes, but
there is little unequivocal evidence to support this grouping. I conclude therefore that
the new genus should be left as Incertae sedis.

44

PALAEONTOLOGY, VOLUME 22

Family rhyniaceae Kidston and Lang, 1920
Incertae sedis

Description and discussion. A small number of terminal sporangia, although elliptical
in shape, are longer than wide. They are not as large as Tortilicaulis specimens, both
sporangium and axis diameters being similar to those in Cooksonia. A typical example
is on NMW 77.6G.39 (PI. 2, fig. 11). The subtending axis (4 mm long) widens at its
base, perhaps indicating a branching point. In the remaining examples, axes are much
shorter and unbranched. Considering sporangium shape, the specimens show some
resemblance to the Cooksonia sp. described by Croft and Lang from the Lower
Devonian of the Brecon Beacons and at present being revised in this laboratory. They
are smaller and all except one lack the border already discussed for the younger speci-
mens. The exception is an elongate sporangium 1-6 mm long and 0-7 mm wide which
has a rounded apex and tapering base (PI. 2, fig. 12). The central region originally
consisted of a flat, smooth sheet of coalified material, which tended to flake off during
the investigation, although some was removed and macerated. The border, a strip of
coalified material, 0T to 0T2 mm wide, but narrowing slightly at the sporangium-
axis junction adhered more closely to the rock surface. This sporangium is obviously
quite different from any of the remaining terminal sporangia in the flora, but more
specimens are required for an adequate circumscription of the new plant.

Finally, Plate 2, fig. 13 shows a small fertile specimen in which two sporangia are
borne immediately above a dichotomy: that on the left is incomplete distally and
0-35 mm wide while the elongate sporangium on the right is at least 0-9 mm long and
03 mm wide. The axis before division is 0-2 mm wide. Smooth- walled spores re-
covered on film pulls are illustrated in Plate 2, fig. 14. This specimen is provisionally
assigned to the form genus Salopella Edwards and Richardson, but until better
material is found will not be given a specific name.

DESCRIPTION OF SMOOTH VEGETATIVE AXES

Hostinella sp. The most common sterile axes in the assemblage are smooth dicho-
tomously branching stems assignable to the form genus Hostinella Barrande. They
are parallel-sided and normally straight or very gently curved. Most specimens
branch just once, the resulting axes being more or less equal in diameter, but very
occasionally an unequal fork is seen (PI. 6, fig. 4). Plate 6, fig. 5 shows a specimen in

EXPLANATION OF PLATE 6

Figs. 1-13. Smooth sterile axes. 1, NMW 77.6G.22a, plant and animal fragments including typical Hosti-
nella. 10-5. 2, NMW 77.6G. 108, block with narrower axes, <17-5. 3, NMW 77.6G.89a, Hostinella
with asymmetric branch pattern, x 7-5. 4, NMW 77.6G. 18b, axis with unequal branching, x6. 5, NMW
77.6G.3, Hostinella with double dichotomy, • 3. 6, NMW 77.6G.52, axis with short ‘lateral’ branches,
8-5. 7, NMW 77.6G. 19, wider axis with anomalous branching, x7. 8, NMW 77. 6G. 40, axis showing
some cellular structure, x 35. 9, 10, NMW 77.6G.37a and 37b, -9, x 10-5. 1 1, NMW 77.6G.27, narrow
axis with short lateral branch below dichotomy, ■ 4-5. 12, NMW 77.6G.52, cluster of branches, x 7.
13, NMW 77.6G.52, Hostinella with presumed vascular strands, 9-35.

Figs. 14, 15. Psilophytites sp. NMW 77.6G.56a and 56b, x 10.

PLATE 6

EDWARDS, Hostinella and Psilophytites

46

PALAEONTOLOGY, VOLUME 22

which more than one branch point is present and Plate 6, fig. 3 an example of a more
profusely branching system with slight asymmetry in the branching pattern itself.
The diameter of the axes varies between 0 05 mm and 2 mm (PI. 6, fig. 1). Some blocks
are covered by very narrow (0 05-0 1 mm) comparatively unbranched axes (PI. 6,
fig. 2). There is little change in diameter along a single specimen even in the more
profusely branched types.

Little has been discovered about the anatomy of the axes. Cuticles have not yet
been isolated from either bulk macerations or film pulls, although longitudinally
aligned irregular strands of organic material, possibly the remains of cortical cell
walls, have been seen on the latter. In the more heavily coalified specimens, longi-
tudinal surface striations can sometimes be seen and also streaks of orange-yellow
material, possibly limonite, perhaps replacing the cell contents (PI. 6, fig. 8).

A few axes have central strands, which are normally narrow when compared with
the total axis diameter and are thus characteristic of the rhyniophytes. Plate 6, fig. 3
shows such a strand having bifurcated some distance below the branch point. The
occurrence of a much wider strand in two specimens suggests that perhaps more than
one major group of vascular plants was present. Tracheids have not been seen.

Anomalous branching forms

(i) Specimen NMW 77.6G.52 (PI. 6, fig. 6). This narrow featureless axis (maximum
width is 0T5 mm) gives off three branches on one side and one on the other. Two of
these lateral projections have wide bases and then taper. The curved longer one is
1 mm long.

(ii) Specimen NMW 77.6G. 19 (PI. 6, fig. 7). The complete specimen is 7-8 mm long
and has a pair of opposite branches at one end and a single branch at the other. The
main axis tapers from 0-6 mm to 0-5 mm in width. One of the pair of branches is
sharply truncated, the other tapers. The solitary branch is rounded distally.

(iii) Specimens NMW 77.6G.27, 37, and 52 (PI. 6, figs. 9-12). These three exhibit a
concentration of branching at one end of a long otherwise unbranched axis. The
orientation of these axes in life is unknown : thus they are all figured with the long
axis horizontal.

Specimen NMW 77.6G.27 (PI. 6, fig. 11) branches almost dichotomously at one
end, but just below the branching point is a lateral projection slightly curved towards
the branched end and ending abruptly. The entire specimen is 13 mm long and the
main axis 0-5 mm wide. The main axis is striated and some of the cell contents have
been replaced by limonite.

Specimen NMW 77.6G.37a (PI. 6, figs. 9, 10). Here a slender axis, 0-4 mm wide and
5-3 mm long terminates in a cluster of four branches. The preservation of the axis is
similar to that in NMW 77.6G.27. Although the ends of the axes appear rounded,
they are actually broken off and end abruptly.

Specimen NMW 77.6G.52 is similar to the last one but has only three short
branches the middle one having a rounded tip (PI. 6, fig. 12).

Discussion. Branching in these three specimens is similar but not identical to the
K-branching characteristic of Zosterophyllum myretonianum Penhallow, where it is
normally confined to the basal region of the plant and is thought to contribute to its

DIANNE EDWARDS: LATE SILURIAN FLORA

47

tufted growth habit (Walton 1964). It is not unlikely, therefore, that these clusters of
branches form the basal holdfast or are part of a more extensive rhizomatous system
of a Hostinella- type plant. It must be emphasized, however, that they have never been
found attached to dichotomizing axes nor has vascular tissue been seen. Thus the tuft
of branches could equally well be part of the erect aerial system of a plant. In 1942
Hoeg described Hostinella with axillary tubercles from the Downtonian of Spits-
bergen. More recently I have shown that in the Lower Devonian Gosslingia brecon-
ensis, the tubercle may be a branch scar or represent the remains of the base of a
branch (Edwards 1970), while Banks and Davis (1968) have described a short branch
in the axillary tubercle position in Crenaticaulis verruculosus. Compression of such a
specimen where overlying branches become fused together could produce a con-
figuration similar to that in NMW 77.6G.27 and NMW 77.6G.52. Such branch
patterns were not recorded by Hoeg, although he did find small fragments of axes
showing the characteristic branching of Zosterophyllum.

DESCRIPTION OF VEGETATIVE AXES WITH ENATIONS

Specimen NMW 77.6G.32a and b (PI. 6, figs. 14, 15). One of the most exciting finds
at this locality was a dichotomously branching axis bearing triangular enations,
interpreted as spines, apparently arranged in two rows, one on either side of the axis.
The over-all height of the specimen is 14 mm. The shorter branch, which is on the left
in Plate 6, fig. 14 and has the best-preserved spines, is just over 4 mm long. The other
which was uncovered extends for approximately one centimetre. It is possible that
this was further branched, but the preservation is not good: little carbonaceous
material remains and some pyrites is present. The axis below the branching point
where spines are few is 0-5 mm wide. Measurement of axis width becomes more diffi-
cult distally because either the spines are crowded and have attenuated bases or the
axis itself is poorly preserved. The axis is sometimes striated but no central strand is
present. The spines appear to be attached to the sides of the axes, but a superficial
attachment for some cannot be ruled out. Their arrangement is random, some
alternate, while others occur in opposite pairs. Their shape is variable. The more com-
plex spines are triangular with length of base roughly equal to height (e.g. 0-4 mm
high and 0-35 mm wide at base). Others are attenuated basally (e.g. 0-47 mm at base
and 0-25 mm high) and some have attenuated apices. An example of this is seen just
below the fork where the needle-like tip is directed forward. In a few the apex is
missing and the tip sharply truncated. The average height of the more complete
spines is 0-29 mm and basal width is 0-34 mm.

Specimen NMW 77.6G.69a and b (PI. 4, figs. 21, 22). Fig. 21 shows the counterpart
of this unbranched axis 11-5 mm long and approximately 1 mm wide. Its margins are
irregular and the surface of this heavily coalified compression is longitudinally
striated. Three prominent projections, slightly asymmetric at the base and sharply
truncated distally (PI. 4, fig. 21), are visible on one edge of the axis. The largest is
0-35 mm high. The other side has the remains of bases of projections only. A pro-
minent depression occurs on the axis surface near one end, indicating that at least one
projection was superficially attached. The axis is far more robust than that in the
preceding specimen. The dimensions of the enations are similar but it is impossible

PALAEONTOLOGY, VOLUME 22

to decide whether the truncated types in this unbranched specimen were actually peg-
like, as for example is seen in Psilophyton princeps s.s. (Hueber 1967), or more spine-
like but with fractured tips.

Specimen NMW 77.6G.32a and b (PL 4, figs. 19, 20). This unbranched flexuous
axis is 6-5 mm long and ranges between 0-2 and 0-3 mm in width. It bears numerous
spines of varying shape. Some are falcate (PI. 4, fig. 20), tapering from a relatively
narrow base to an almost hair-like tip (typical example is 0T5 mm wide at base and
0-4 mm high), while others are more robust with wider bases (c. 0-3 mm) and are less
curved distally or broken off. The bases of some of the spines are overlain by the axis.

Specimen NMW 77.6G.88 (PI. 4, fig. 23). Triangular enations are numerous in the
upper part of this twice branched specimen overlapping both each other and the axis
itself. Some of the spine tips are directed forward. Axis width is approximately
0-5 mm.

Discussion. These very fragmentary axes have some evolutionary importance in that,
as far as I am aware, they are the only pre-Devonian spinous plants morphologically
similar to the later Devonian genera Sawdonia and Psilophyton. There are a number
of plants covered with appendages which are claimed to be lycopods or psilopsids
from older rocks. These include the Cambrian Aldanophyton Kryshtofovich, the
Ordovician Boiophyton Obrhel and Akdalophyton Senkevich, and the Silurian Saxonia
Roselt and Lycopodolica Ishchenko. None have been shown to be vascular. Indeed
some possibly have animal rather than plant affinities. Their numerous needle- or
hair-like emergences do not resemble the non-vascular enations of either Sawdonia
or Psilophyton. Although I have not demonstrated vascular tissue in the Downtonian
axes, which are considerably smaller than later spinous specimens, I am convinced
that should they have been found in Lower Devonian rocks they would have been
assigned to Hoeg’s form genus Psilophytites (Hoeg 1952). He erected this for frag-
mentary sterile axes with spreading undivided spines of psilophytalean affinity intend-
ing it to have a similar usage to Hostinella. This was before the unravelling of the
exceedingly complex taxonomic tangle involving Dawson’s Psilophyton complex.
(For a full account see Banks, Leclercq, and Hueber 1975). It is now accepted that in
the Gaspe flora there are two distinct types of spiny plant, P. princeps (Trimero-
phytina) and Sawdonia ornata (Zosterophyllophytina) (Hueber 1967; Hueber and
Banks 1967). Thus as Hoeg anticipated, the fragmentary sterile spinous stems
described from numerous Lower Devonian localities may well belong to quite
separate taxonomic groups (Banks 1975J) making the usage of a form genus such
as Hoeg’s highly desirable. It is perhaps a little unfortunate that the name may be
taken to imply relationship with the genus Psilophyton itself and furthermore that its
meaning is the exact opposite of what it describes (Gk. psilo- = smooth). But nomen-
clatural considerations and revisions are best centred on Devonian specimens and I
propose provisionally to call these spinous Downtonian axes Psilophytites sp.,
appreciating that they probably belong to at least two taxa of possibly widely separate
affinity.

DIANNE EDWARDS: LATE SILURIAN FLORA

49

ENVIRONMENT OF DEPOSITION

The stratigraphy and sedimentology of the area is currently being revised by Professor
J. R. Allen and Dr. B. P. J. Williams. They consider that Dixon’s Basement Beds were
accumulated on coastal sandflats influenced by the sea (pers. comm.), calcretes and
mudcracks providing evidence for both prolonged and short exposure. It is considered
that the sandstones and conglomerates probably represent channels and sand banks,
perhaps partly intertidal. There is therefore the possibility that the plants grew on a
saltmarsh. The fact that, although fragmentary, the individual fossils are not badly
damaged, suggests limited transport and supports this suggestion. However, it is
also a possibility that the plants lived on river banks or on mudflats surrounding lakes
and were washed into the coastal mudflats where they were fairly rapidly buried.

GENERAL DISCUSSION

The composition of Downtonian floras has been mentioned only briefly here, but is
more adequately surveyed elsewhere (e.g. Banks 1975c). Hostinella and Cooksonia
predominate in the majority with the Podolian, and to a lesser extent, the Dyfed
floras, showing greater diversity. I have actually seen only the Welsh Borderland
fossils described by Lang (1937). My observations on his two Cooksonia species have
already been recorded. In general the most striking difference between Lang’s and the
Dyfed plants is one of size, both axes and sporangia being considerably smaller in the
latter.

My investigation of the Dyfed Downtonian flora is part of a much larger project
involving the collection and description of plant macrofossils from Eltonian (lower-
most Ludlow Series) to Downtonian localities in South Wales, an area where the
stratigraphy is particularly well documented. Preliminary results are summarized in
Table 4. Apart from Freshwater East, our most productive Downtonian locality is
Capel Horeb (Edwards 1970) where in addition to Steganotheca, Cooksonia has now
been recorded (unpublished data). Tracheids have not yet been demonstrated in
Eltonian Bringewoodian Y-axes but it seems not unlikely that vascular plants first
appeared in this region in middle Silurian times or even earlier. The first indication
that such Y-axes had rhyniophyte affinities comes in the Whitcliffian (Edwards and
Davies 1976) where they are found attached to Cooksonia and Steganotheca axes. The
present account produces some evidence for diversification in the Downtonian.
Whether or not a similar evolutionary pattern was repeated throughout the world is
debatable. Records of possible lycopods in the Late Silurian of Podolia (Ishchenko
1975), of lycopods in questionably Silurian rocks of Libya (Klitzsh et al. 1973) and
the prospect of the revision of the age of the lower part of the Baragwanathia flora in
Australia (Gray and Boucot 1977) all indicate that further critical reappraisal is neces-
sary before any generalizations are made.

Hoeg (1952, p. 213) in discussing nomenclatural problems relating to the identifica-
tion of fragmentary axes in the Lower Devonian wrote : Tt may be maintained that if
a plant fossil is so incomplete, it does not merit mention at all.’ To a certain extent this
may be held true for the very fragmentary fossils I have described in this Downtonian

50

PALAEONTOLOGY, VOLUME 22

table 4. Late Silurian stratigraphy in South Wales and Bohemia, incorporating records of macroplant
fossils and environmental interpretations in the South Wales area. (Stratigraphy based on Cocks et al.

1971.)

Series

Stages

Britain

Bohemia

Macroplant remains
in South Wales
(excluding calcified
forms)

Environmental
interpretation in
South Wales

Terrestrial fluviatile

Cooksonia species

Brackish— sub- and

Steganotheca striata
Tortilicaulis
transwalliensis
Psilophytites sp.
IZosterophyllum

intertidal

Y-axes + tracheids

Marine

Cooksonia sp.
Steganotheca sp.

Marine si

allowing

Y-axes + tracheids

Marine inshore

Y-axes

Marine inshore-

shelf

Y-axes

Marine inshore-

shelf

Powysia bassettii

(Edwards 1977)

Marine inshore-

Inopinatella

Y-axes

shelf

Post-Ludlow

pre-Gedinnian

(Downton)

Pridoli

Whitcliffian

Bringewoodian

Kopanina

Leintwardinian

Eltonian

DIANNE EDWARDS: LATE SILURIAN FLORA

5:

flora, but like Hoeg I consider it is important to describe, document, and, where pos-
sible, to identify such plants for future reference. The age of the flora is also respon-
sible for perhaps an over-zealous treatment of minor differences in sporangium
morphology. Certainly such a rigorous approach would not be applied in describing,
for example, a truss of sporangia from the Upper Devonian Rhacophyton Mourlon.
It has resulted, however, in a demonstration of inter- and intra-specific variation in
Cooksonia, while the whole assemblage indicates a diversity of land vegetation much
greater than hitherto known in the uppermost Silurian of Britain.

Acknowledgements. I am grateful to Professor H. P. Banks, Dr. M. G. Bassett, Dr. J. Duckett, Professor
J. R. L. Allen, and Dr. B. P. J. Williams for their advice. I thank Mrs. C. Rogerson who is employed
on an N.E.R.C. research grant for her invaluable assistance with photography. The financial support of the
N.E.R.C. is gratefully acknowledged. I also thank the Keeper of Palaeontology at the British Museum
(Natural History) for permission to borrow specimens from the Lang collection.

REFERENCES

ananiev, a. R. and Stepanov, s. a. 1969. The first finding of the Psilophyton flora in Lower Devonian
Salairsky Ridge (Western Siberia). Tomsk State Univ. Publication , 203, 13-28, pis. 1-2. [In Russian.]
Andrews, H. N. 1960. Notes on Belgian specimens of Sporogonites. Palaeobotanist, 7, 85-89.
banks, H. p. 1973. Occurrence of Cooksonia, the oldest vascular land plant macrofossil, in the Upper
Silurian of New York State. J. Indian bot. Soc. Golden Jubilee Vol. 50A, 227-235, pi. 1.

— 1975a. Early vascular land plants: proof and conjecture. Bioscience, 25, 730-737.

— 19756. The oldest vascular land plants: a note of caution. Rev. Palaeobot. Palynol. 20, 13-25, pis. 1-2.

— 1975c. Palaeogeographic implications of some Silurian-Early Devonian floras. In Campbell, k. s. w.
(ed.). Gondwana Geology. Australian National University Press, Canberra. Pp. 71-97, 4 pis.

— and davis, m. r. 1969. Crenaticaulis, a new genus of Devonian plants allied to Zosterophyllum and its
bearing on the classification of early land plants. Am. J. Bot. 56, 436-449.

— leclercq, s. and hueber, f. M. 1975. Anatomy and morphology of Psilophyton dawsonii sp. n. from the
late Lower Devonian of Quebec (Gaspe) and Ontario, Canada. Palaeontogr. am. 8, 77-137, pis. 17-24.

cocks, l. r. m., Holland, c. h., rickards, R. b. and strachan, i. 1971. A correlation of Silurian rocks in the
British Isles. Jl. geol. Soc. 127, 103-136.

croft, w. n. and lang, w. h. 1942. The Lower Devonian flora of the Senni Beds of Monmouthshire and
Breconshire. Phil. Trans. R. Soc. 231B, 131-163, pis. 9-11.
daber, r. 1971. Cooksonia— one of the most ancient psilophytes— widely distributed, but rare. Botanique,
2, 35-40, pi. 1.

dixon, e. E. l. 1921. The geology of the South Wales Coalfield. Part XIII. The country around Pembroke
and Tenby. Mem. Geol. Surv. U.K.

Edwards, D. 1969. Further observations on Zosterophyllum llanoveranum Croft and Lang, from the Lower
Devonian of South Wales. Am. J. Bot. 56, 201-210.

— 1970a. Fertile Rhyniophytina from the Lower Devonian of Britain. Palaeontology, 13, 451-461,
pis. 84-87.

— 19706. Further observations on the Lower Devonian plant Gosslingia breconensis Heard. Phil. Trans.
R. Soc. 258B, 225-243, pis. 34-38.

— 1977. A new non-calcified alga from the Upper Silurian of Mid Wales. Palaeontology, 20, 823-832,
pis. 110-111.

— and Davies, e. c. w. 1976. Oldest recorded in situ tracheids. Nature, Lond. 263, 494-495.

— and richardson, j. b. 1974. Lower Devonian (Dittonian) plants from the Welsh Borderland. Palae-
ontology, 17, 311-324, pis. 40-41.

gensel, p. g. 1976. Renalia hueberi, a new plant from the Lower Devonian of Gaspe. Rev. Palaeobot.
Palynol. 22, 19-37, pis. 1-5.

gray, j. and boucot, a. J. 1977. Early vascular land plants: proof and conjecture. Lethaia, 10, 145-174.

52 PALAEONTOLOGY, VOLUME 22

halle, t. G. 1916. A fossil sporogonium from the Lower Devonian of Roragen in Norway. Bot. Notiser,
79-81.

— 1936. Notes on the Devonian genus Sporogonites. Svensk bot. Tidskr. 30, 613-623, pis. 3-4.

hoeg, o. a. 1942. The Downtonian and Devonian flora of Spitsbergen. Skr. Svalb. og Ishavet, 83, 1-228,
pis. 1-62.

— 1952. Psilophytites, a new form genus of Devonian plants. Palaeobotanist, 1, 212-214.

— 1967. Psilophyta. In boureau, e. (ed.). Traite de Paleobotanique. Masson et Cie, Paris. Vol. 2, 191-352.
hueber, F. m. 1967. Psilophyton : the genus and the concept. Proc. Int. Symp. Devonian System. D. H.

Oswald (ed.), Calgary, vol. II, 815-822, pis. 1-2.

— and banks, H. p. 1967. Psilophyton princeps : the search for organic connection. Taxon , 16, 81-85.
Ishchenko, T. a. 1969. The Cooksonia palaeoflora in the Skalski Horizon of Podolia and its stratigraphical

significance. Geol. J. Kiev. 29, 101-109, 1 pi. (Transl. Geol. Surv. Canada).

— 1975. The late Silurian flora of Podolia. 80 pp., pis. 1-15. ‘Dymka’ Scientific Publishing House, Kiev.
[In Russian.]

king, w. w. 1934. The Downtonian and Dittonian strata of Great Britain and north-western Europe.
Q. J. geol. Soc. Lond. 90, 526-562.

klitzsh, e., lejal-nicol, a. and massa, d. 1973. Le Siluro-Devonien a psilophytes et lycophytes du bassin de
Mourzouk (Libye), C.R. Acad. Sc. Paris. Series D, 277 , 2465-2467.
lang, w. H. 1937. On the plant remains from the Downtonian of England and Wales. Phil. Trans. R. Soc.
227B, 245-291, pis. 8-14.

obrhel, j. 1962. Die Flora der Pridoli-Schichten (Budnany-Stufe) des mittelbohmischen Silurs. Geologie,
11, 83-97, pis. 1-2.

Richardson, J. B. and lister, t. r. 1969. Upper Silurian and Lower Devonian spore assemblages from the
Welsh Borderland and South Wales. Palaeontology, 12, 201-252, pis. 37-43.
schopf, J. m. 1975. Modes of fossil preservation. Rev. Palaeobot. Palynol. 20, 27-53, pis. 1-3.
schuster, R. m. 1966. The Hepaticae and Anthocerotae of North America. Vol. 1. Columbia University
Press, New York and London.

stockmans, F. 1940. Vegetaux eodevoniens de la Belgique. Mem. Mus. r. Hist. nat. Belg. 93, 1-90, pis. 1-14.
walton, J. 1964. On the morphology of Zosterophyllum and some other early Devonian plants. Phyto-
morphology, 14, 155-160.

yurina, a. 1969. The Devonian flora of Central Kazakhstan. In bogdanov, a. a. (ed.). Materials on the
Geology of Central Kazakhstan. Moscow University Press, Moscow. Vol. 8, 143 pp., pis. 1-30. [In
Russian.]

DIANNE EDWARDS

Typescript received 20 January 1978
Revised typescript received 20 March 1978

Department of Botany
University College
Cardiff CF1 1XL

TRILOBITES FROM THE CONISTON
LIMESTONE GROUP (ASHGILL SERIES)
OF THE LAKE DISTRICT, ENGLAND

by KENNETH J. MCNAMARA

Abstract. Five trilobite species from the Coniston Limestone Group of the southern Lake District are redescribed,
on the basis of type and topotype material, and one new genus, four new species, and one new subspecies erected.
A lectotype of Sphaerexochusl boops Salter is designated and topotype material described; the species is placed
in Pseudosphaerexochus. Encrinurus kingi Dean is redescribed and placed in Enatencrinurus, for which a type
species is designated ; this represents the first record of this genus in Britain. A lectotype of Calymene subdiademata
McCoy is selected; material from northern England previously assigned to C. marginata (Shirley) is referred to
C. subdiademata. Chasmops marri (Reed) is redescribed and placed in a new genus, Toxochasmops', the pygidium and
hypostome are described for the first time. The thorax and hypostome of Acidaspis magno spina Stubblefield are
described for the first time and a lectotype designated. Ascetopeltis apoxys sp. nov., Paraharpes whittingtoni sp. nov.,
Gravicalymene susi sp. nov., Primaspis bucculenta sp. nov., and Tretaspis convergens Dean deliquus subsp. nov. are
described. The evolution of T. convergens in northern England and its distribution and relationships with other
contemporaneous species of Tretaspis are discussed. Trilobites from the Coniston Limestone Group described by
McCoy (1851) are re-evaluated.

Outcrops of the Coniston Limestone Group in the southern part of the Lake
District (Cumbria) occur in a narrow strip, rarely more than one hundred metres
across, which trends north-east to south-west. The most easterly outcrops occur on
the southern side of Shap Fells. Outcrops occur intermittently to the south-west
towards Lake Windermere, principally in Longsleddale, Kentmere, and Troutbeck.
Well-exposed outcrops occur to the south-west in the region of Coniston, and Torver
and Ashgill Becks, thence intermittently to Millom in the extreme south-west of the
Lake District (see McNamara in press, fig. 1).

Considering the relatively large amount of research on this Group that has been
undertaken (particularly in the latter part of the nineteenth century), our knowledge
of the faunas is very limited. These rocks had been studied by some estimable geo-
logists in the nineteenth century, notably Sedgwick, Hughes, Nicholson, Harkness,
and Marr. However, the only purely palaeontological work to result from their
researches were descriptions by McCoy (1851) (which included Calymene brevi-
capitata Portlock, C. subdiademata McCoy, Lichas subpropinqua McCoy, Zethus
atractopyge McCoy, Zethus rugosus Portlock, Cheirurus clavifrons Dalman, and
Illaenus rosenbergi Eichwald); Salter (1864), who described Sphaerexochusl boops,
and Reed (1894), who described Chasmops marri. In the twentieth century only four
workers have dealt specifically with trilobites from the Coniston Limestone Group
in the southern Lake District: Stubblefield (1928) described Acidaspis magnospina\
Whittington (1950) described and figured two harpids which he referred to Paraharpes
cf. hornei (Reed); Temple (1952) described and figured species of Dalmanitina\ and,
most recently, Dean (1963a) has described a meagre trilobite fauna from the Stile End
Formation, in which he identified and described one new species, Encrinurus kingi. In

[Palaeontology, Vol. 22, Part 1, 1979, pp. 53-92, pis. 7-12.]

54

PALAEONTOLOGY, VOLUME 22

addition, Ingham (1968, 1970) has recently illustrated Cybeloides ( Paracybeloides )
girvanensis (Reed) and the types of Illaenus marshalli Salter from the southern Lake
District.

The aim of this paper is threefold : to describe one new genus, four new species, and
one new subspecies; to redescribe Calymene subdiademata, Sphaerexochusl boops,
Encrinurus kingi, Chasmopsmarri, and Acidaspis magno spina ; to re-evaluate McCoy’s
(1851) diagnoses of Coniston Limestone Group trilobites. A list of the trilobites
found in the Coniston Limestone Group is given elsewhere (McNamara, in press).

Terminology. The terminology of Harrington et al. (in Moore 1959) is followed,
except that ’branch’ is preferred to ‘section’, with regard to the facial suture, and
‘hypostome’ is preferred to ‘hypostoma’ or ‘labrum’. Where ambiguity may arise
when terms such as ‘long’, ‘broad’, etc. are used, they are qualified by the use of the
terms ‘sagittal’, ‘exsagittal’, and ‘transverse’ (sag., exsag., and tr.). In describing the
lateral glabellar lobes and furrows the abbreviations lp, 2p, etc. are used, numbering
from the rear. The tubercle notation for Encrinurus devised by Tripp (1957) is
utilized, as is the notation for describing the character of trinucleid pits devised by
Bancroft (1929) and modified by Ingham (1970) and Hughes et al. (1975). One new
term, the ‘orle furrow’ is introduced to describe the furrow near the anterior margin
of the frontal lobe in species of Chasmops, close to, and parallel with, the preglabellar
furrow. The term is derived from the seventeenth-century word ‘orle’, which describes
a border within a heraldic shield a short distance from its edge. The stratigraphical
terminology is that of McNamara (in press).

Temple (1975a, pp. 462-466) has discussed various orientations which may be
employed for the purpose of describing, measuring, and photographing trilobites.
A single orientation encompassing all morphological types is impractical. For this
study cephala were orientated, where possible, with eye lobes lying along a horizontal
plane. If eye lobes were not present, the posterior margin of the occipital ring was set
vertically. Pygidia were orientated with the dorsal surface of the first axial ring lying
along a horizontal plane. The method of orientation of trinucleids is that suggested
by Hughes et al. (1975, p. 546), the anterior and posterior fossulae lying along a
horizontal plane.

Preservation of material. All the specimens have been decalcified and are preserved as
internal and external moulds in mudstone, calcareous mudstone, calcareous siltstone,
and argillaceous limestone. These rocks have suffered unidirectional post-
lithification deformation which has resulted in distortion of the fossils (see PI. 9,
figs. 9, 10 and PI. 11, figs. 1, 3 for typical examples). The deformation is greater in the
mudstone than in the limestone. As a consequence of this distortion of the fossils,
comparisons cannot be made between length and breadth unless the specimen appears
to be relatively undistorted. Consequently comparisons of relative dimensions are
made along single axes. This allows for comparison between forms which are either
longitudinally, laterally, or obliquely compressed. This method can be employed only
so long as the rock has suffered only a single period of deformation which is unidirec-
tional. Such is the case with the material from the southern Lake District. The degree
of deformation of the rocks, and consequently the fossils, increases in a south-
westwards direction across the southern Lake District.

McNAMARA: CONISTON TRILOBITES

55

The material described is housed in: the Sedgwick Museum, Cambridge (prefix
SM), the British Museum (Natural History) (BM), the Institute of Geological
Sciences (GSM), and the Hunterian Museum, Glasgow (HM). The locality numbers
for the topotype material of existing species and for the material of new species are
explained elsewhere (McNamara in press , figs. 2-6, Appendix 1).

SYSTEMATIC PALAEONTOLOGY

Family proetidae Salter, 1864
Subfamily proetinae Salter, 1864
Genus ascetopeltis Owens, 1973a

Type species. By original designation, Ascetopeltis bockeliei Owens, 1973a, p. 125, figs. 1m, n and 2a-k, from
the Harjuan, Tretaspis Series, Stage 5a, Holmen, Oslo-Asker district, Norway.

Discussion. The similarity of species of Ascetopeltis to some species which have been
included in Proetus (s.l.) has been discussed by Owens (1973a, p. 126). The main
differences he noted between the two genera are in pygidial features and surface
sculpture. In addition the glabella of Ascetopeltis appears to be shorter in relation to
its width, with a more broadly rounded anterior. Ascetopeltis is restricted to the
Ashgill Series (Owens 19736, p. 27).

Ascetopeltis apoxys sp. nov.

Plate 7, figs. 1-9; text-fig. 1

Holotype. SM A98177, internal mould of cranidium (PI. 7, fig. 1) from the Stile End Formation (Cautleyan
Stage, Zone 2) 450 m north-north-west of Stockdale Farm, Longsleddale (locality 8b).

Paratypes. HM A15430, internal mould of incomplete, articulated specimen (PI. 7, fig. 4) from Kentmere
(locality 41a); SM A98157, external mould of cranidium (PI. 7, fig. 2) from locality 8b; SM A98159,
internal mould of librigena (PI. 7, fig. 7) from locality 7b; SM A98085, internal mould of pygidium (PI. 7,
fig. 3) from locality 7b.

Material, localities, and horizon. Forty-seven exuviae have been collected from the quarries in the Stile
End Formation north-north-west of Stockdale Farm, Longsleddale (localities 7a, 7b, 8a, 8b). The paratype
articulated specimen (HM A15430, PI. 7, fig. 4) and a cranidium (HM A15431, PI. 7, fig. 5) were collected
by Dr. J. K. Ingham from the same horizon at Kentmere (locality 41a).

Diagnosis. Preglabellar furrow slightly under half length (sag.) of anterior rim; convex anterior rim as
wide (tr.) as base of glabella, tapers abruptly laterally ; anterior branches of facial suture weakly divergent to
border furrow then strongly convergent anteriorly; eye lobe short; pleural and interpleural furrows of
pygidium well-incised anteriorly ; pygidial axis broad and long.

Description. Cranidium longer than broad; moderately convex. Large, tumid, gently tapering glabella
occupies seven-tenths cranidial length ; widest posteriorly; slightly constricted at mid-glabellar length (PI. 7,
fig. 6); broadly rounded anteriorly. Bears three pairs of shallow glabellar furrows: lp most prominent,
extending transversely for a short distance in from axial furrow at one-third glabellar length from occipital
furrow, then recurves strongly to run almost exsagittally towards occipital furrow which it fails to meet
(PI. 7, fig. 2); 2p furrow short, directed more transversely than lp; 3p furrow set a little closer to 2p furrow
than 2p is to lp; fainter and shorter than 2p furrow. Axial furrow well incised. Occipital furrow narrow
(exsag. and sag.) and deep; transverse medially, directed forwards abaxially. Occipital ring more strongly
vaulted than glabella; narrower (tr.) than posterior of glabella; weak occipital lobe present laterally (PI. 7,
fig. 4); bears small occipital node which is set closer to posterior border (PI. 7, fig. 1). Eye lobe short, one-
quarter cranidial length (PI. 7, fig. 2) and set far back opposite posterior part of glabella, extending from

56

PALAEONTOLOGY, VOLUME 22

opposite 2p lobe to close to occipital furrow. Anterior branches of facial sutures run parallel from eye lobe
then are weakly divergent anteriorly to border furrow from where they converge strongly. Their posterior
course is unknown, but posterior position of eye (PI. 7, fig. 8) suggests that angles e and £ (see Owens 19736,
p. 4) are not independent angles. Anterior part of fixigena quite strongly convex; rapidly narrows adaxially
(PI. 7, fig. 2). Preglabellar furrow narrow (sag.), less than half length of anterior rim. Whole preglabellar
area only occupies one-sixth of cranidial length. Anterior rim convex sagittally; strongly acuminate
abaxially (PI. 7, figs. 1, 2); as wide (tr.) as posterior of glabella. Librigena (PI. 7, figs. 7, 8) triangular;

Figs. 1-9. Ascetopeltis apoxys sp. nov. Stile End Formation (Cautleyan Stage, Zone 2). 1, holotype,
SM A98177, dorsal view of internal mould of cranidium, from Stockdale, Longsleddale (locality 8b),
x 8. 2, paratype, SM A98157, dorsal view of cast of external mould of cranidium, from same locality
as 1 , 8.3, paratype, SM A98085, dorsal view of internal mould of pygidium, from Stockdale, Longsled-
dale (locality 7b), x 8. 4, paratype, HM A 15430, dorsal view of incomplete, articulated specimen, from
Kentmere (locality 41a), x 6. 5, HM A15431, dorsal view of internal mould of cranidium, from same
locality as 4, x 7. 6, SM A98087, dorsal view of internal mould of cranidium, distortion having enhanced
glabellar constriction; from same locality as 3, x8. 7, paratype, SM A98159, dorso-lateral view of
internal mould of librigena, from same locality as 3, x 6. 8, SM A98088, dorso-lateral view of internal
mould of librigena, from same locality as 3, x 7. 9, SM A98204, dorsal view of internal mould of pygi-
dium, from Stockdale, Longsleddale (locality 8a), x 6.

Figs. 10, 11. Paraharpes whittingtoni sp. nov. Applethwaite Formation (Cautleyan Stage, Zone 2), north of
Waterhead House, Coniston. 10, holotype, GSM 74456a, dorsal view of cast of external mould of near-
complete individual, x 2; also figured by Whittington (1950, pi. 5, fig. 1). 11, paratype, GSM 74456b,
dorsal view of cast of external mould of incomplete cephalon, x2; also figured by Whittington (1950,
pl. 5, fig. 2).

PLATE 7

McNAMARA, Coniston trilobites

58

PALAEONTOLOGY, VOLUME 22

bounded by strongly convex rim and deep furrow; lateral border thickens anteriorly; posterior border
thickens and flattens abaxially; short genal spine developed. Prominent eye socle reaches back to posterior
furrow. Cheek below eye convex and strongly declined.

Thorax only known from incomplete, distorted specimen (PI. 7, fig. 4) which bears eight segments. Axis
occupies half thoracic width anteriorly, but narrows posteriorly; it is strongly vaulted; each axial ring
separated from articulating half-ring by deep furrow. Pleurae bear prominent furrows; strongly declined
distally and acuminate.

Pygidium (PI. 7, figs. 3, 9) subtriangular, nearly twice as wide as long. Large axis occupies two-fifths
pygidial width anteriorly and reaches almost to posterior border; gently tapered posteriorly; broadly
rounded at posterior; strongly vaulted and gently convex longitudinally. Bears six axial rings; anterior
interannular furrow deep, succeeding one shallow posteriorly, being ill defined sagittally. Axial furrow
broad and shallow. Pleural area gently convex, bearing at least three, possibly four, pleurae. Pleural and
interpleural furrow equally well incised anteriorly, more prominent laterally; become progressively
shallower posteriorly. Anterior pleural band a little broader than posterior band.

Discussion. Ascetopeltis apoxys is similar to the only other described species of
Ascetopeltis from Britain, A. barkingensis Owens, 19736, collected from an erratic
boulder of presumed upper Ordovician in age, from Barking, Dent, Yorkshire (Owens
19736, p. 26). A. apoxys differ from A. barkingensis in a number of features. It has
deeper glabellar furrows, narrower (tr.) occipital ring, and more posteriorly posi-
tioned occipital node. The preglabellar furrow is longer in A. apoxys', it is one-third
the length of the anterior border in A. barkingensis, not one-half as given by Owens
(19736, p. 26) in the species diagnosis. The anterior border of A. apoxys is narrower
(tr.) and tapers more strongly laterally than in A. barkingensis, while the anterior
branch of the facial suture is less divergent initially. The eye is positioned further back
in A. apoxys, the anterior part of the fixigena is more convex, and the pygidial furrows
are more deeply incised. A. apoxys is also similar to the type species A. bockeliei
Owens (1973a, p. 126, fig. 2) from the late Ashgill of the Oslo region; however, this
species possesses a longer eye lobe, shorter (sag.) preglabellar furrow, shallower
lateral border furrow, and longer genal spine. A. lepta Owens (1973a, p. 130, fig.
3e-g, j), from the same horizon as A. bockeliei, has a more cylindrical glabellar and
longer preglabellar furrow than A. apoxys.

Family harpidae Hawle and Corda, 1847 [non Bronn, 1849]

Genus paraharpes Whittington, 1950

Type species. Harpes (Eoharpes) homei Reed, 1914, p. 10, pi. 2, figs. 1,2; from the upper Drummuck Group,
Girvan; by subsequent designation of Whittington (1950, p. 11).

Paraharpes whittingtoni sp. nov.

Plate 7, figs. 10, 11

1892 Harpes Doranni Portlock; Marr, p. 108.

1950 Paraharpes cf. homei (Reed); Whittington, pp. 41-42, pi. 5, figs. 1, 2.

Holotype. GSM 74456a, the external mould of an almost complete articulated individual (PI. 7, fig. 10) from
the Applethwaite Formation (Cautleyan Stage, Zone 2) at ‘Coniston Waterhead’; figured by Whittington
(1950, pi. 5, fig. 1).

Material, locality, and horizon. Whittington (1950, p. 41) was unsure both of the horizon and the locality
from which the holotype and the paratype (GSM 74456b, PI. 7, fig. 11) were collected. It seems likely they
originated from an outcrop north of Waterhead House, 2 km east-north-east of Coniston (SD 321984). The

McNAMARA: CONISTON TRILOBITES

59

specimens of P. whittingtoni are preserved in calcareous mudstone characteristic of the Applethwaite
Formation. The only other harpid material known from the Coniston Limestone Group in the southern
Lake District are fragments of harpid fringe from the High Pike Haw Formation (Cautleyan Stage, Zone 2)
at High Pike Haw. These, however, cannot be specifically determined.

Diagnosis. Brim has about eighteen irregular rows of pits lying in a rough quincuncial pattern ; twelve rows
occur on the cheek roll. Brim of constant anterior and lateral width. Prolongation shorter than length (sag.)
of cephalon; internal rim almost straight. Eye lobe set far forward.

Description. Cephalon two-thirds as long as broad. Bordered by wide brim which occupies a little over one-
quarter cephalic length (sag.) ; of even width anteriorly and laterally to end of cheek roll, then tapers strongly
posteriorly ; almost flat and covered by many pits which are arranged in a rough quincuncial pattern ; pits
close to external rim are larger than inner pits; pits increase in size close to internal border (PI. 7, fig. 1 1).
There are approximately eighteen concentric rows of pits in brim, though they are not always well defined as
many small pits are developed which tend to break up the more regular pattern of larger pits. Prolongation
shorter than length of cephalon, being estimated to end approximately opposite twentieth thoracic segment ;
tapers strongly posteriorly, internal rim being almost straight. Cheek roll horizontal anteriorly, inclining to
about 45° postero-laterally opposite alae ; narrower than brim, bearing about twelve irregular rows of small
pits anteriorly. Cheek roll prolongation wide anteriorly, tapering strongly posteriorly; bears large pits
arranged in a quincuncial pattern. Girder is weakly developed. External rim narrow, convex; internal like-
wise. Glabella occupies a little under one-half cephalic length ; strongly convex ; widest posteriorly where
subtriangular posterior lobe present, this being set much lower than rest of glabella; anteriorly glabella
contracts abruptly, anterior width being two-thirds posterior. Posterior glabellar furrows deeply incised and
straight ; anteriorly divergent at about 40° ; narrow posteriorly, broaden near mid length, then narrow to
axial furrow (PI. 7, fig. 11). Preglabellar furrow short (sag.) and shallow. Axial furrow runs exsagittally
forwards from posterior border to anterior of basal lobe, then runs gently inwards to preglabellar furrow;
shallow posteriorly, but much deeper and broader anteriorly. Ala wider (tr.) than basal lobe, semicircular,
inner half strongly convex, outer half lower and flatter. Raised, pitted, gently convex cheek narrowest (tr.)
outside ala, widens anteriorly, less so posteriorly; central part of wide anterior portion bears conical eye
lobe which is set a little lower than posterior of glabella but higher than basal lobe and ala ; set far forward
opposite anterior of glabella. Prominent eye ridge (PI. 7, fig. 10) runs inwards and forwards to antero-
lateral corner of glabella. Occipital furrow deep and wide (sag.) narrows (exsag.) abaxially. Occipital ring
strongly vaulted and short (sag. and exsag.) ; bears anteriorly positioned median glabellar node (PI. 7, fig. 1 1 ).

Thorax (PI. 7, fig. 10) incomplete, only fifteen segments preserved. Broadens posteriorly to fourth seg-
ment, then of constant width to eleventh, posterior to which it narrows strongly. Axis tapers gently back-
wards and occupies one-third thoracic width anteriorly, one-quarter posteriorly. Articulating half ring long
(sag.) and narrows abaxially to deep axial furrow. Pleurae bear a prominent, oblique furrow, anterior and
posterior pleural bands of equal length (exsag.). Pleurae horizontal to distal fulcrum; then steeply decline
for one-sixth pleural width (tr.).

Discussion. Whittington (1950, p. 41) preferred not to erect a new species for this form,
as the horizon and locality were uncertain. The near-certainty that it is from the
Applethwaite Formation, the good preservation of the material, and the discovery of
further important differences between it and P. hornei (Reed 1914, p. 10, pi. 2, figs. 1,2)
seem to justify the foundation of a new species. P. whittingtoni has a greater number
of more well-ordered pits in the brim than P. hornei (18 as opposed to 12-14) and more
pits in the cheek roll (12 as opposed to 8-10). The prolongation of P. whittingtoni is
shorter and less strongly curved, being less than the length of the cephalon, whereas
it is more than the cephalic length in P. hornei , extending to the twenty-third segment
(Whittington 1950, p. 39). The brim of P. hornei broadens laterally around the
cephalon, but remains of even width in P. whittingtoni. The preglabella area of
P. whittingtoni is relatively longer than that of P. hornei. The glabella of P. hornei
is much longer than broad, whereas it is slightly broader than long in P. whittingtoni.

60

PALAEONTOLOGY, VOLUME 22

The axial furrow is deeper and broader in P. whittingtoni resulting in a well-defined
glabella. The eye lobe is set further forward relative to the glabella in P. whittingtoni.
The eye ridge, prominent in P. whittingtoni , is not apparent in the holotype of
P. hornei. The thorax tapers less strongly in P. whittingtoni than in P. hornei.
P. costata (Angelin), redescribed by Warburg (1925, pp. 214-224, pi. 5, figs. 1-6) from
the Boda Limestone of Sweden, differs from P. whittingtoni in having a much shorter
glabella and more gently tapered, longer prolongation. Dean (1971, pp. 8-10, pi. 2,
figs. 4, 6) has described P. costata from the Chair of Kildare Limestone in Eire on
one incomplete cephalon. This has a glabella which is much narrower posteriorly
than in P. whittingtoni and narrower alae. P. ruddyi Whittington (1950, p. 42, pi. 5,
figs. 6-8, pi. 6, figs. 1-3) from upper Ashgill strata in Wales has a similar brim form to
P. whittingtoni, but has fewer, larger pits.

Family trinucleidae Hawle and Corda, 1847
Subfamily trinucleinae Hawle and Corda, 1847
Genus tretaspis McCoy, 1 849

Type species. Subsequently designated by Bassler 1915, p. 1285, Asaphus seticornis Hisinger, 1840, p. 3,
pi. 37, fig. 2; from the Fjacka Formation, Dalarna, Sweden.

Tretaspis convergens Dean, 1961 deliquus subsp. nov.

Plate 8, figs. 1-8; text-figs. 2, 3

1916 Trinucleus seticornis (Hisinger); Marr, p. 199 (pars.).

1916 Trinucleus bucklandi Barrande; Marr, pp. 199, 202.

1970 Tretaspis hadelandica Stormer brachystichus Ingham; Ingham, p. 49 (pars.).

Holotype of subspecies. SM A99006, external mould of a cephalon and incomplete thorax (PI. 8, figs. 1-4)
from the Torver Formation (Cautleyan Stage, Zone 3) between Torver Beck and its first tributary
(locality 22e).

Material, localities, and horizons. Two complete articulated specimens plus numerous cephalic fragments
are known from the Torver Formation around Torver Beck and its tributaries (localities 21e, 22e), Old Pits
Beck (locality 35b), and Willy Scrow (locality 25a). Two specimens have been collected from the upper part
of the Applethwaite Formation (Cautleyan Stage, Zone 3) at Garbourn Nook (locality 19), whilst three
specimens were collected by Dr. J. K. Ingham from this horizon at Kentmere (locality 41b2).

Diagnosis. Subspecies of Tretaspis convergens in which arc I4 of the fringe is never developed. Arc I3 is
absent frontally, and may be either complete or incomplete laterally; where laterally incomplete 8-10 pits
persist.

Description. Apart from the fringe, the characters of the cephalon and thorax of T. convergens deliquus are
largely like those of T. convergens convergens described by Dean (1961, p. 127) from the Pusgillian of the
Cross Fell Inlier, and by Ingham (1970, p. 45) from the same stage in the Cautley district. Ingham (1970,
p. 46) observed a transverse row of tubercles behind the anterior margin of the axial ring but this does not
appear to be present in T. convergens deliquus.

On the fringe four arcs of pits present anteriorly. E, and I4 pits share sulci to bR9 or bR10, though further
apart after bR5, then diverge, following which E3 is developed from about bR14. Et possesses twenty-one
pits (half fringe) (text -fig. 2). E3 only developed laterally; shares sulci with E, on upper lamella but not on
lower. Laterally pits in E, and E2 are smaller than anteriorly. There are eight to thirteen pits in E2 (half
fringe), ending one pit short of posterior row, which contains seven to nine pits. In front of glabella aR

McNAMARA: CONISTON TRILOBITES

61

radii consist of I2 and In, I3 being absent frontally (PI. 8, figs. 2, 5). Pits in I3 appear at aR3 and persist
to aRn 13 in all specimens (text-fig. 2). In some specimens pits from aR14 to posterior row are present (PI. 8,
fig. 3) whereas in others pits are absent between aR13 and aR16 (PI. 8, figs. 5, 7). On upper lamella In and I3
share sulci laterally but not anteriorly. On lower lamella pits in I2, I3, and In share deep radial sulci ; I2
sometimes has pits in the same sulci but only when aligned with inner pits. I2-In are radially aligned, as are
Ip E1; and E2, the two sets being out of phase with one another (PI. 8, fig. 5).

The pygidium (PI. 8, figs. 6, 8) of T. convergens has not hitherto been described. Its maximum breadth
(tr.) is about three times sagittal length. Anterior border straight, transverse ; postero-lateral border broadly
convex, lateral margin steeply declined. Axis occupies one-quarter anterior pygidial width; gently convex
transversely ; bears five furrows which are broad (sag.) and shallow medially ; opposite these, in axial furrow,
deep apodemal pits present. Axial furrow is broad and shallow. Pleural area bears a prominent anterior
pleural furrow which broadens laterally; second furrow very shallow and narrow close to axial furrow,
broadening and deepening abaxially. A faint third furrow exists as an elongate, broad, shallow depression
a little behind broad part of second furrow (PI. 8, figs. 6, 8). Posteriorly border arches up behind axis (PI. 8,
fig. 8) and bears a deep medial invagination.

pits in E2 pits in posterior row

text -fig. 2. Histograms of selected fringe characters (half fringe data) in Tretaspis
convergens Dean deliquus subsp. nov.

Discussion. T. convergens convergens , which Dean (1961, pp. 127-129, pi. 9, figs. 1-6)
described from the upper Pusgillian Dufton Shales of the Cross Fell Inlier, is charac-
terized by a strongly swollen frontal lobe which overhangs the fringe anteriorly, and
coarsely reticulated frontal lobe and genae. The fringe of the Cross Fell material
contains arcs I2, 13, and In, which are present frontally (text-fig. 3), and I4 which has
eight to ten pits (Ingham 1970, p. 45) but is incomplete both laterally and frontally.
Ingham (1970, p. 45, pi. 6, figs. 1 12) has described T. convergens convergens from the
Pusgillian and early Zone 1 of the Cautleyan Stage of the Cautley district. In contrast
to the Cross Fell specimens, the Cautley form has fewer pits in I4 (0-4) and a wider
range of pits in the posterior row (7-12). In the Lake District T. convergens deliquus
is common in Zone 3 in the Torver Formation and occurs rarely in the upper part of
the Applethwaite Formation, which is also of Zone 3 age (McNamara in press). In

62

PALAEONTOLOGY, VOLUME 22

the lower and middle parts of the Applethwaite Formation (Zone 2) T. corner gens
convergens is locally common. Like the Cautley form, the Lake District form of
T. convergens convergens has fewer pits in I4 than the Cross Fell form, but it continues
a trend towards decreasing the number of pits in this row by generally having I4
absent (although I3 is complete frontally and laterally) (text-fig. 3) though specimens
from one locality (15a) contain two pits in I4.

This trend in reduction of the number of pits is continued in T. convergens deliquus,
I3 becoming discontinuous frontally and, in many specimens, discontinuous laterally.
Ingham (1970, p. 49) thought that specimens of Tretaspis which he collected from the
upper part of the Applethwaite Formation (Zone 3) at Kentmere (locality 41b2)

text-fig. 3. Diagrammatic representation of inner part of fringe (I2-In) in subspecies of
Tretaspis convergens Dean, illustrating their relative stratigraphic positions in northern
England ; data of T. convergens convergens in the Pusgillian and early Zone 1 from Ingham
(1970); filled circles represent arc I3.

McNAMARA: CONISTON TRILOBITES

63

were T. hadelandica St0rmer brachysticus Ingham on account of the presence of a
laterally incomplete row of pits in arc I3. Comparison of these specimens with recent
collections from the Lake District shows Ingham’s specimens to belong to T. con-
vergent deliquus. This subspecies has also been collected from the Applethwaite
Formation at Garbourn Nook (locality 19).

The decrease in number of pits in the fringe in the T. convergens lineage is in direct
contrast to the morphological changes in pit number displayed by the T. hadelandica
lineage in Rawtheyan strata (Ingham 1970, pp. 46-50) and in the early lineages of
the T. moeldenensis group in upper Caradoc and early Ashgill strata at Cautley
(Ingham 1970, pp. 50-55). The disappearance of T. convergens from Cautley during
Zone 1 , but its presence in Zones 2 and 3 in the Lake District might be accounted for
by the increase in dominance of species of the T. moeldenensis group in Cautley during
the early Cautleyan. Whereas T. convergens convergens was able to coexist with an
early member of the T. moeldenensis group, T. colliquia Ingham, T. cf. moeldenensis
Cave became dominant during Zone 1 (Ingham 1970, p. 54) and so may have forced
out T. convergens convergens. The absence of members of the T. moeldenensis group
from the Lake District, the continuance here of T. convergens convergens and the
development of T. convergens deliquus , suggest a barrier to migration of species of
Tretaspis between Cautley and the Lake District during Zones 2 and 3. During the
hiatus in the Lake District between Zones 4 and 6, the T. convergens lineage became
extinct. The barrier between the two regions must have been broken by late Zone 6
as T. aff. latilimbus (Linnarsson) distichus Ingham, a form intermediate between
T. hadelandica brachystichus and T. latilimbus distichus , is present in the White Lime-
stone Formation in the Lake District.

Whereas the T. convergens lineage shows a decrease in number of pits in the fringe,
the T. hadelandica group shows an increase by the completion of I3, then the develop-
ment of arc I4. As a consequence of this, homeomorphy of fringe characters occurs
between the T. convergens and T. hadelandica lineages, which may result in some con-
fusion in the naming of subspecies using fringe characters alone. Consequently,
T. convergens convergens from the Pusgillian and T. latilimbus distichus from the
upper Rawtheyan display a similar fringe morphology. Similar homeomorphy exists
between the Cautleyan T. convergens deliquus and the Rawtheyan T. hadelandica
brachystichus, both species having an incomplete number of pits in I3. The two forms
can be distinguished, however, by the larger frontal lobe of T. convergens deliquus
and its coarser, more complete cephalic reticulation, and by the smaller number of
pits in I3 frontally in older T. hadelandica brachystichus than in T. convergens deliquus.
The Lake District T. convergens convergens from Zones 2 and 3 is like T. hadelandica
hadelandica (which is probably from the Gagnum Shale of Gran, Hadeland (Ingham
1970, p. 49)), I3 being complete laterally.

The pygidium of T. convergens deliquus is similar to that of T. hadelandica brachy-
stichus but differs in bearing five axial furrows, not six; similarly the pleural area
bears three furrows, not four. Species of the T. moeldenensis group generally bear
more than six axial furrows.

64

PALAEONTOLOGY, VOLUME 22

Family cheiruridae Hawle and Corda, 1847
Subfamily eccoptochilinae Lane, 1971
Genus pseudosphaerexochus Schmidt, 1881

Type species. Sphaerexochus hemicranium Kutorga, 1854, p. 112, pi. 1, fig. 2: from the Aseri Stage (Cla),
Estonia; by subsequent designation of Reed (1896).

Discussion. The form described here, Pseudosphaerexochus boops, has never been
formerly assigned to any genus with certainty. The ovate form of the glabella and
form of the glabella lobes is characteristic (Lane 1971, p. 45) of Pseudosphaerexochus,
to which genus P. boops is now referred.

Pseudosphaerexochus boops (Salter, 1864)

Plate 8, figs. 9-14

1851 Ceraurus clavifrons Dalman; McCoy, pp. 154 (pars), 338 (pars), pi. If, fig. 12, 12a.

1864 Sphaerexochus ? boops Salter (pars), p. 79, pi. 6, fig. 28, non 27.

1868 Cheirurus juvenis Salter; Nicholson, p. 54.

1873 Sphaerexochus boops Salter; Salter, p. 50.

1878 Sphaerexochus boops Salter; Marr, p. 873.

1888 Cheirurus bimucronatus Murchison; Aveline et al., p. 55.

1888 Sphaerexochus boops Salter; Aveline et al., p. 55.

1891 Sphaerexochus boops Salter ; W oods, p. 151.

1892 Sphaerexochus boops Salter; Marr, p. 109.

1974 Pseudosphaerexochus ? boops (Salter) (pars); Price, pp. 850-852, pi. 113, fig. 11, non 10.

Lectotype. Herein selected : SM A41905, internal mould of a distorted cranidium (PI. 8, figs. 9, 1 1) from the
Applethwaite Formation (Cautleyan Stage, Zone 2), ‘Applethwaite Common’; figured by McCoy 1851,
pi. If, fig. 12; Salter 1864, pi. 6, fig. 28; Price 1974, pi. 113, fig. 11.

Material, localities, and horizon. Three other specimens of this species are known : an incomplete cranidium
from the same horizon and locality as the lectotype ; another from a similar horizon south of Nettle Crag,
Torver (collected by Mr. T. C. Nicholas) ; a badly deformed cranidium collected by Dr. J. K. Ingham, from
a similar horizon at Stunfel Howe (locality 14) can probably be referred to this species.

EXPLANATION OF PLATE 8

Figs. 1-8. Tretaspis convergens Dean deliquus subsp. nov. Torver Formation (1-4, 6-8) and upper Apple-
thwaite Formation (5) (both Cautleyan Stage, Zone 3). 1-4, holotype of subspecies, SM A99006, from
Torver (locality 22e), cast of external mould of incomplete individual, 1, dorsal, 2, anterior, 3, antero-
lateral, 4, lateral views, all x 6 : pits in row I3 complete laterally. 5, SM A98729, antero-lateral view of
cast of external mould of incomplete cephalon, from Garbourn Nook (locality 19), showing row I3
incomplete laterally, x 6. 6, SM A43178, dorsal view of internal mould of almost complete individual,
from Torver (locality 21e), showing length of genal spines, x 4. 7, 8, SM A43181, antero-lateral, 7, and
dorsal, 8, views of internal mould of incomplete individual, from Torver (locality 23e), x 3}.

Figs. 9-14. Pseudosphaerexochus boops (Salter), Applethwaite Formation (Cautleyan Stage, Zone 2). 9, 1 1,
lectotype, SMA41905, dorsal, 9, and lateral, 1 1 , views of internal mould of distorted cranidium ; 11 shows
greater exsagittal length of 3p lobe over 2p lobe; from ‘Applethwaite Common’, x H; also figured by
McCoy (1851, pi. If, fig. 12, 12a), Salter (1864, pi. 6, fig. 28), and Price (1974, pi. 113, fig. 11). 10, 14,
SM A99016, lateral, 10, and dorsal, 14, views of internal mould of cranidium, from Nettle Crag, Torver
(locality 43), x2. 12, 13, SM A68682, lateral, 12, and palpebral, 13, views of internal mould of incomplete
cranidium, from ‘Applethwaite Common’, x 2.

PLATE 8

McNAMARA, Coniston trilobites

66

PALAEONTOLOGY, VOLUME 22

Emended diagnosis. Glabella ovate and broadly rounded anteriorly ; bears three pairs of glabellar lobes of
which 3p is longer (exsag.) than 2p, whilst lp is longer than 3p. 3p furrow runs almost transversely.
Relatively narrow central glabellar area between sub-quadrate basal lobes. Ornamentation on glabella of
fine granules with fewer coarse tubercles.

Description. Glabella ovate, approximately two-thirds as broad as long, widest at 2p lobes; almost flat
postero-medially, declining anteriorly from opposite 3p lobes at approximately 70° to vertical. Transversely
gently convex medially, steeply declined laterally; broadly rounded anteriorly, lp lobe longest (exsag.),
sub-quadrate, bounded by deep lp furrow which runs inwards and slightly backwards from axial furrow,
then curves back strongly at one-quarter glabellar width, progressively shallowing to meet occipital furrow
(PI. 8, figs. 9, 13). Delineated lp lobe occupies a little under one-third posterior glabellar width; 2p furrow
parallel to abaxial part of lp furrow, but shallower; extends in as far as does outer part of lp furrow;
2p lobe shorter than lp lobe (exsag.), being two-thirds its length; 3p furrow directed more transversely
than 2p or lp furrows; shallower and a little shorter than 2p; 3p lobe a little longer than 2p lobe, but
shorter than lp lobe (PI. 8, figs. 10-12); lengthens adaxially (PI. 8, fig. 10). Frontal lobe short, occupying
one-tenth glabellar length. Preglabellar furrow deep, lying below overhanging frontal lobe (PI. 8, fig. 12).
Anterior border rolled; almost transverse medially, curving back strongly abaxially. Occipital furrow
shallow medially, curving back strongly abaxially. Occipital furrow shallow medially, deeper abaxially;
appears to curve forwards on lectotype, but this is thought to have been caused by distortion ; it probably
runs transversely (PI. 8, fig. 14). Occipital ring imperfectly known, gently convex medially, declining strongly
abaxially. Eye lobe not preserved but thought to be situated opposite 2p lobe (see PI. 8, fig. 12). Posterior
branch of facial suture directed almost transversely from eye lobe, curving back strongly close to genal
angle. Posterior border furrow deep adaxially, shallows abaxially. Posterior border is roll-like. Fixigena
narrows (exsag.) abaxially ; pitted. Glabella covered by an ornamentation of fine granules interspersed with
larger tubercles (PI. 8, figs. 9, 10).

Discussion. Salter (1864) based this species on two specimens: one from ‘Apple-
thwaite Common’, originally described by McCoy (1851, pi. If, fig. 12) as Ceraurus
clavifrons, the other from the Sholeshook Limestone of south Wales, which have since
been refigured by Price (1974, pi. 113, figs. 10, 1 1). The Welsh specimen, which only
has the basal glabellar lobes and posterior part of the glabella preserved, has been
compared with the Lake District form both by Salter and Price due to the apparently
strongly curved occipital furrow. As Price (1974, p. 852) suggests, the Sholeshook
specimen has undergone distortion, as has the Lake District specimen. Consequently
the curved occipital furrow is thought to have been caused by distortion and is not
considered to be diagnostic. The Sholeshook specimen is not considered to be the
same species as the Lake District form as the lp lobe is relatively larger and the
posterior of the central glabelar area is relatively narrower; this form may represent
a distorted specimen of P. juvenis (Salter). Price’s assertion that the 3p lobe of the
Lake District form is shorter than the 2p is erroneous. He based his opinion on
the dorsal view alone, whereas, in fact, the anterior of the glabella has been strongly
bent beneath the cranidium and the 3p lobe is longer (PI. 8, fig. 1 1). Although Lane
(1971) omits to mention P. boops in his account of the British cheirurids, the charac-
teristic form of the glabellar lobes is considered distinctive enough to warrant the
retention of the species.

P. boops is closest to P. juvenis and P. octolobatus (McCoy). P. juvenis (Salter
1848, p. 344, pi. 7, figs. 1, 2, 3, 3a) from the Sholeshook Limestone has an ovate
glabella, but more tapered frontal lobe than P. boops (Price 1974, pi. 113, figs. 5, 6);
the 2p lobe is longer than the 3p. In P. octolobatus (McCoy 1849, p. 407 ; Lane 1971,
p. 48, pi. 8, figs. 1-8) from the Bala Limestone, which also occurs in the Applethwaite
Formation and in the overlying Torver Formation, the ornamentation is like that of

McNAMARA: CONISTON TRILOBITES

67

P. boops, but the glabella is more sphaerical and the 2p and 3p lobes are of equal
length. P. ekphyma Lane (1971, p. 46, pi. 9, figs. 1, 4, 5) from the upper Drummuck
Group at Girvan has an ovoid glabella which like P. boops is broadly rounded
anteriorly, but the 2p lobe is longer than the 3p.

Family encrinuridae Angelin, 1854
Subfamily encrinurinae Angelin, 1854
Genus erratencrinurus Krueger, 1971

Type species. Erratencrinurus nebeni Krueger, 1971, p. 1144, pi. 1, fig. 7; pi. 2, figs. 1, 2; from the Oandu
(D3) and Rakvere (E) Stages of the glacial boulders of Germany ; herein designated.

Discussion. Krueger (1971) erected Erratencrinurus on specimens of middle and upper
Ordovician age from the glacial boulders of north Germany. Typically encrinurids
possess a number of large glabellar tubercles associated with smaller adventitious
ones. Erratencrinurus is typified by having either a pair of very large tubercles or
spines in row II or III, or a single spine or tubercle in row I. The species from the Lake
District which Dean ( 1 963a, p. 53) assigned to Encrinurus based principally on pygidial
characters, is reassigned to Erratencrinurus as the cranidium, which is now known,
shows the inner pair of glabellar tubercles in row III to be much larger than others
on the glabella. This is the first known occurrence of this genus in Britain. Other
penecontemporaneous species from Britain previously referred to Encrinurus have
recently been placed in a new genus, Celtencrinurus, by Evitt and Tripp (1977, p. 1 19).
This genus is typified by the possession of the large tubercle III-O, and a median furrow
on the anterior border of the cranidium.

Erratencrinurus kingi (Dean, 1963a)

Plate 9, figs. 1-7; text -fig. 4

1963a Encrinurus kingi Dean, pp. 53-54, pi. 1, figs. 6, 7, 12; pi. 2, figs. 1, 2, 7.

Holotype. SM A51706, internal mould of a pygidium (PI. 9, fig. 4) figured by Dean 1963a, pi. 2, fig. 1 ; from
the Stile End Formation (Cautleyan Stage, Zone 2) in the quarries above Stockdale Farm, Longsleddale.

Material, localities, and horizons. In addition to the five specimens (one incomplete cranidium, a fragment of
fixigena, and three pygidia) figured by Dean (1963a, pi. 1, figs. 6, 7, 12; pi. 2, figs. 1, 2, 7), there is also
an almost complete cranidium (BM It8328) collected by W. B. R. King (as were Dean’s specimens) from
the same horizon as the holotype and paratypes. In the Sedgwick Museum collection there are two further
pygidia which were collected from the High Pike Haw Formation at 'Millom Park’ in the extreme south-
west of the outcrop of the Coniston Limestone Group. A cranidium of this species has been collected by the
author from 1-5 km north-west of Millom (locality 39). A number of fragmentary specimens have also been
collected from behind Kentmere Hall (locality 33) in the Stile End Formation. Thus this species is restricted
to the Stile End and High Pike Haw Formations (Cautleyan Stage, Zone 2).

Emended diagnosis. Frontal lobe wide, nearly three times posterior glabellar width. Tubercle formula:
1-1, II-3, 2, 1 ; III-3, 2, 1 ; (iv-1, 0); IV-3, 2, 1 ; (v-1, 0); V-2, 1 ; VI-2, 1 (text-fig. 4). Librigena pitted, bearing
only a few tubercles below the eye. Pygidium with twenty-five axial rings and eleven pleurae.

Description. Cranidium subtriangular, convexity unknown due to flattening. Glabella broadest across
frontal lobe ; narrows posteriorly to almost one-third frontal lobe width. Frontal lobe occupies about one-
third glabellar length (sag.) and is well rounded anteriorly. Anterior border broad and convex; separated
from frontal lobe by deep, broad furrow. Glabellar carries three equi-spaced, tuberculate, glabellar lobes
which increase in size anteriorly (PI. 9, figs. 2, 5). These are separated from one another by short, slot-like

68

PALAEONTOLOGY, VOLUME 22

furrows. Posterior to lp lobe a narrow (sag. and exsag.) raised band runs across glabella and widens just
prior to meeting axial furrow. The band is bounded anteriorly by a shallow furrow which narrows adaxially,
and posteriorly by a deeper furrow of more constant depth (PI. 9, figs. 2, 5). Occipital ring long (sag.)
occupying two-sevenths cranidial length. Occipital furrow broad (sag. and exsag.) and shallow. Axial
furrows deep but broad, gently divergent from occipital furrow, more strongly divergent anterior to 3p
lobe. Fixigena strongly convex; narrow (tr.) anteriorly but widens abruptly posterior to eye lobe. Eye lobe
set opposite lp furrow at a distance from axial furrow equal to width of glabella across 2p lobes. Posterior
branch of facial suture meets lateral border a little anterior to genal angle ; anterior branch runs forwards
and inwards at 45°, then runs obliquely across frontal lobe below false preglabellar ridge. Librigena (PI. 9,
fig. 7) known only from an incomplete specimen which lacks anterior portion. It consists of a broad, con-
vex border which is directed horizontally adaxially, then declines outwards, through 90°, and a steeply
declined triangular area adaxially. Glabella is covered by coarse tubercles, the notation for which is given
in the diagnosis. Size of the tubercles is variable : those of row I are small, central pair of row II is larger,
whilst central pair of row III is much larger still (PI. 9, fig. 6); in row IV, central pair is as large as

Figs. 1-7. Erratencrinurus kingi (Dean), Stile End (2-7) and High Pike Haw (1) Formations (Cautleyan
Stage, Zone 2). 1, SM A72668, dorsal view of internal mould of pygidium, from Millom (locality 39),
4. 2, 5, BM It8328, dorsal views of internal mould, 2, and cast of external mould, 5, of cranidium, from
Longsleddale (locality 11), both -4. 4, holotype, SM A5 1 706, dorsal view of internal mould of pygidium,
from Stockdale, Longsleddale (locality 8c), x 3; also figured by Dean (1963a, pi. 2, fig. 1). 3, 6, SM
A98248, antero-lateral, 3, and dorsal, 6, view of cast of external mould of incomplete, distorted cranidium
showing the large pair of glabellar tubercles in row III, from Kentmere (locality 38), both x 4. 7, BM
It8329, lateral view of cast of external mould of librigena, from same locality as 2, x 4.

Figs. 8-16. Calymene (s.l.) subdiademata McCoy, Applethwaite Formation (Cautleyan Stage, Zones 2
and 3). 8, SM A98619, dorsal view of cast of external mould of incomplete individual, from Moor Head,
Troutbeck (locality 17a), x2. 9, lectotype, SM A6806, dorsal view of external mould of distorted
individual, from ‘Coniston Water’, x 3; also figured by McCoy (1851, pi. If, fig. 10). 10, SM A43551,
dorsal view of internal mould of complete individual, from ‘Applethwaite Common’, x2. 11, 12, 13,
SM A98910; lateral, 1 1, x 2}, anterior, 12, x 4, and dorsal, 13, x 2\ , views of cast of external mould of
cranidium, from Garbourn Nook (locality 19). 14, SM A989 1 1, dorsal view of internal mould of cephalon,
from Garbourn Nook (locality 19), x2. 15, SM A98558, dorsal view of cast of external mould of
pygidium, from Garbourn Road, Troutbeck (locality 1 5b), x 1|. 16, SM A98900, ventral view of internal
mould of hypostome, from same locality as 1 1, x 4J.

text-fig. 4. Glabellar tubercle arrangement in
Erratencrinurus kingi (Dean) ; tubercle notation
follows that of Tripp (1957).

EXPLANATION OF PLATE 9

PLATE 9

McNAMARA, Coniston trilobites

70

PALAEONTOLOGY, VOLUME 22

inner pair in row II. Only small tubercles are developed on the frontal lobe. Anterior border bears unknown
number of large tubercles. Fixigena is covered by many large tubercles, whereas steeply declined librigena
is pitted, bearing only a single row of tubercles below elevated eye. Border of librigena bears three rows
of tubercles in a quincuncial pattern (PI. 9, fig. 7); rows decrease in size outwards. Posterior border and
genal spine are smooth.

Hypostome and thorax are unknown. Pygidium has been adequately described by Dean (1963a,
pp. 53-54); however, it is not, as Dean states, ‘twice as long as broad’ but only slightly longer than broad
(PI. 9, figs. 1, 4; Dean 1963a, pi. 2, fig. 1).

Discussion. Erratencrinurus kingi appears to be most closely related to the type
species, E. nebeni Krueger, which is late Caradoc in age. Although both have a
similar tubercle arrangement, E. kingi bears an extra pair of tubercles in row III and
no median tubercle in row II. Although the development of an enlarged pair of
glabella tubercles is the main feature which relates E. kingi to the German species
of Erratencrinurus, it also has a similar pygidium. The German forms bear eight to
eleven pleurae; E. kingi has eleven. Whilst most German species possess more than
thirty axial rings, E. kauschi Krueger, like E. kingi, has only twenty-five. The pygidium
of E. kauschi is also similar to E. kingi in over-all shape, but has one fewer pair of
pleurae. E. kingi was considered by Dean (1963a, p. 53) to bear five tubercles on the
occipital ring; in fact only four tubercles are present on the specimen (SM A51705;
Dean 1963a, pi. 1, fig. 7) and they lie on row III of the glabella, not the occipital ring.
Celtencrinurus multiplicatus Reed (1901, p. 107, pi. 7, fig. 3) from the ‘Middle Bala
at Barking, Dent’ is known from a pygidium which bears a similar number of pleurae
to E. kingi. However, the posterior pleura meets the axis much farther forwards in
C. multiplicatus. It also seems to possess two more axial rings, but the poor preserva-
tion of the holotype makes this difficult to ascertain with certainty.

Family calymenidae Milne Edwards, 1840
Subfamily calymeninae Milne Edwards, 1840
Genus calymene Brongniart, 1822

Type species. Calymene Blumenbachi Brongniart 1822, p. 11, pi. 1, fig. la, b\ from the Wenlock Limestone,
Wren’s Nest, Dudley, Worcestershire.

Discussion. Temple (19756, pp. 147-149) has questioned the emplacement of a number
of late Ordovician and early Llandovery species in Diacalymene as the upturned
anterior part of the preglabellar area lacks a transverse dorsal keel typical of D. dia-
demata (Barrande), from the Silurian of Bohemia. Instead, these forms possess an
anterior border which bears an anterior keel and a posterior ridge, the intervening
area varying in orientation between horizontal, as in ‘Z). marginata' Shirley, gently
inclined, as in ‘ZX’ crassa Shirley, and steeply inclined, as in ‘ZX’ drummuckensis
(Reed). These species show features more in common with Calymene, but until the
type species, C. blumenbachi Brongniart, is redescribed fully, it is thought advisable,
as suggested by Temple (19756, p. 149), to refer the species to Calymene ( sensu lato).

Calymene (s.l.) subdiademata McCoy, 1851

Plate 9, figs. 8-16

71845 Calymene blumenbachi Brongniart; Sedgwick, p. 445.

1851 Calymene subdiademata McCoy, pp. 166-167, pi. If, fig. 10, non 9.

McNAMARA: CONISTON TRILOBITES

71

1868 Calymene subdiademata McCoy; Nicholson, p. 53.

1878 Calymene senaria (Conrad); Marr, p. 873.

1888 Calymene subdiademata McCoy; Aveline et al., p. 55.

71892 Calymene blumenbachi var. caractaci Salter; Marr, p. 108.

1913 Calymene planimarginata Reed (pars) ; Marr, pp. 2, 7.

1916 Calymene planimarginata Reed; Marr, pp. 191, 199, 200.

1933 Calymene subdiademata McCoy; Shirley, p. 65.

1934 Calymene cf. planimarginata Reed; King and Wilcockson, p. 10.

1936 Diacalymene marginata Shirley (pars), p. 416 (pars), pi. 29, fig. 20.

1948 Calymene (Diacalymene) marginata (Shirley); King and Williams, pp. 206, 210, pi. 16, fig. 2.

1959 Diacalymene cf. marginata Shirley; Dean, pp. 204, 208.

1962 Diacalymene cf. marginata Shirley; Dean, p. 1 16, pi. 13, fig. 13; pi. 14, fig. 11.

1966 Diacalymene marginata Shirley; Ingham, pp. 465-468, 484, 486, 489, 494, 497-498.

1977 Calymene (s.l.) cf. marginata (Shirley); Ingham, pp. 98-100, pi. 21, figs. 9-24.

Lectotype. Herein selected; SM A6806, internal mould of complete individual from the Applethwaite
Formation (Cautleyan Stage, Zone 2), ‘Coniston Water’ (PI. 9, fig. 9); figured by McCoy 1851, pi. If,
fig. 10.

Material, localities, and horizon. This species has been collected from localities 9, 14-19, 22f, 41b in the
middle and upper parts of the Applethwaite Formation (Cautleyan Stage, Zones 2 and 3) in the Lake
District. It is particularly common on Moor Head, north of Applethwaite Common (localities 16-19),
where it comprises almost 40 per cent of the trilobite fauna.

Emended diagnosis. A species of Calymene (s.l.) characterized by a strong posterior ridge on the preglabellar
area which is separated from the anterior ridge by a gentle, posteriorly inclined flat surface ; glabella bell-
shaped, projecting only slightly in front of fixigena and bearing a small 2p lobe, half the diameter of the
lp lobe.

Discussion. McCoy (1851, pp. 166-167) based C. subdiademata on a number of
specimens, but only figured two syntypes ; a cranidium from the ‘limestone of Leint-
wardine’ and the complete, though distorted, specimen (SM A6806) from the Lake
District. In his discussion of this species, Shirley (1933, p. 65) recommended that the
name ‘ subdiademata' be dropped, as one of the syntypes the cranidium from Leint-
wardinae, appears to be lost (this is confirmed by Dr. D. Price, Assistant Curator at
the Sedgwick Museum) and the other specimen is distorted. As can be seen from
Plate 9, fig. 9 the distortion has had little effect in diminishing any of the distinguish-
ing characters of this species (as described by Ingham (1977, pp. 98-100) on con-
specific material from the Cautley district). As further conspecific material has
subsequently been collected from the same formation as the lectotype it would
appear to be justifiable to retain McCoy’s specific name.

Ingham (1977, p. 98) has called the northern England form C. (s.l.) cf. marginata.
The holotype of C. (s.l.) marginata from the lower Drummuck Group (Cautleyan),
Craighead Inlier, near Girvan is lost (Ingham 1977, pp. 98-99), but the illustrations
by Shirley (1936, pi. 29, fig. 19) and of topotype material by Ingham (1977, pi. 21,
figs. 7, 8) shows C. (s.l.) marginata to have a glabella which is relatively narrower
posteriorly, a frontal lobe which has a more transverse anterior margin and which
projects forward of the fixigena less than in C. (s.l.) subdiademata, and a 2p lobe
larger relative to the lp lobe. Shirley (1936, pi. 29, fig. 20) illustrated a specimen
which he referred to ‘ Diacalymene marginata' from the ‘Calymene Beds (? Lower
Bala), opposite Taythes Farm, near Cautley, Yorkshire’. Ingham (1977) has not
discussed the taxonomic position of this specimen, but places it, along with the

72

PALAEONTOLOGY, VOLUME 22

holotype of ‘D.’ marginata, in the synonomy of C. (s.l.) cf. marginata. This Cautley
specimen appears to belong to C. (s.l.) subdiademata.

Shirley (1933, p. 65) compared the lectotype of C. (s.l.) subdiademata to ‘C.’
quadrata King as he thought the Lake District form like ‘C.’ quadrata, possesses
only twelve thoracic segments. One segment, however, is largely covered by the
pygidium. Another complete specimen (PI. 9, fig. 10) shows the presence of thirteen
thoracic segments of typical calymemid form. Recently, Siveter (1977, p. 386) erected
the genus Sthenarocalymene and included within it King’s C. quadrata. S. quadrata
belongs within the Flexicalymeninae and not the Calymeninae as it lacks fixigenal
buttressing to the 2p lobe. The number of thoracic segments possessed by Sthenaro-
calymene is variable: the type species S. lirella Siveter (1976, p. 388) has thirteen, one
more than S. quadrata.

C. (s.l.) emicata Ingham (1977, pp. 100-101, pi. 22, figs. 1-6) from the Rawtheyan
Stage, Zone 5 of the Cautley district, has a longer, more inflated glabella than
C. (s.l.) subdiademata and a more arched pregabellar area. C. (s.l.) prolata Ingham
(1977, pp. 102-103, pi. 22, figs. 11-17) from the Cautleyan Stage, Zone 3 in the Cautley
district, and C. (s.l.) prolata from the Torver and upper Applethwaite Formations
(also Zone 3) of the Lake District can be distinguished from C. (s.l.) subdiademata by
its more roll-like preglabellar area and glabella which projects strongly in front of
thefixigenae. C. (s.l.) drummuckensis Reed (1906, p. 135, pi. 17, fig. 14; pi. 18, figs. 1-4)
from the upper Drummuck Group, Girvan, has relatively larger 2p and 3p lobes than
C. (s.l.) subdiademata, anterior keel of the anterior border set relatively higher than
the posterior ridge, more transverse anterior margin of the frontal lobe and glabella
which is narrower posteriorly. C. (s.l.) crassa (Shirley 1936, p. 416, pi. 29, figs. 21-23)
from the lower Llandovery Gasworks Mudstone near Haverfordwest, has a longer,
more convex glabella, which is relatively narrower posteriorly, larger 2p lobes and
longer preglabellar area which has a less strongly developed posterior ridge than in
C. (s.l.) subdiademata (Temple 1975, pi. 25, fig. 4).

Subfamily flexicalymeninae Siveter, 1976
Genus gravicalymene Shirley, 1936

Type species. Gravicalymene convolva Shirley, 1936, p. 409, pi. 29, figs. 16-18; from the Birdshill Limestone
(Ashgill), Birdshill Quarry, near Llandeilo, south Wales.

Gravicalymene susi sp. nov.

Plate 10, figs. 1-13

71845 Calymene n. sp. ; Sedgwick, p. 455.

1851 Calymene brevicapitata Portlock; McCoy, p. 166, pi. If, fig. 6.

1865 Calymene senaria Conrad; Salter, p. 98.

1868 Calymene brevi-capitata Portlock; Nicholson, p. 53.

1873 Calymene senaria Conrad; Salter, p. 53.

1888 Calymene brevicapitata Portlock; Aveline, Hughes, and Strahan, p. 55.

1888 Calymene senaria Conrad; Aveline, Hughes, and Strahan, p. 55.

71888 Calymene n. sp. ; Aveline, Hughes, and Strahan, p. 55.

1892 Calymene senaria Conrad; Marr, p. 108.

1931 non Calymene brevicapitata Portlock; Shirley, p. 4.

1963a Gravicalymene cf. praecox (Bancroft); Dean, p. 55, pi. 1, fig. 11.

1977 Gravicalymene deani Ingham, p. 97 (pars).

McNAMARA: CONISTON TRILOBITES

73

Holotype. BM It8689, cranidium (PI. 10, figs. 1-4), collected by W. T. Dean from the south-east bank
of Stockdale Beck, just below the junction with Brow Gill (locality 4a) ; 25 m above the base of the Apple-
thwaite Formation (Cauleyan Stage, Zone 2).

Material, localities, and horizons. This species commonly occurs (132 exuviae) in the lower-middle part of
the Applethwaite Formation (Cautleyan Stage, Zone 2) in Longsleddale (localities 3, 4a, 4c, 4e, 4f, 9, 9a, 9b),
Kentmere (localities 12a, 12b, 13, 14a, 14c, 14d), Pull Beck (locality 26a), and above Hussey Well Beck
(locality 36). It occurs less commonly in the upper part of the Applethwaite Formation (Cautleyan Stage,
Zone 3) (forty-nine exuviae from localities 16, 16a, 18a, 19 north of Applethwaite Common). The specimen
figured by Dean (1963a, pi. 1, fig. 11) from the Stile End Formation (Cautleyan Stage, Zone 2) above
Stockdale, Longsleddale (locality 8c), is placed in this species. A single cranidium is known from the High
Pike Haw Formation (Cautleyan Stage, Zone 2) at High Pike Haw (locality 32).

Diagnosis. Glabella bearing relatively large lp lobe and small 2p lobe and projecting beyond fixigena.
Preglabellar area long (sag.), gently inclined posteriorly into deep furrow and rolled anteriorly. In plan,
preglabellar area well rounded, laterally acuminate. Axial furrow broad and sinuous.

Description. Cephalon subtraphezoidal, broad. Glabella moderately vaulted; with low longitudinal con-
vexity; strongly ‘bell-shaped’, lp lobes being widely spaced, distance between their abaxial extremities
being about T5 times that across 2p lobes and T9 times that across 3p lobes, lp lobe bluntly subquadrate,
lp furrow narrow; deep laterally, wider and shallow adaxially where bifurcates, anterior branch faintly
encircling 2p lobe, posterior branch curving back and shallowing as runs almost exsagittally to occipital
furrow. 2p lobe small, subcircular, less than half diameter of lp lobe. 2p furrow short, shallow, directed at
15° back from transverse line. 3p lobe is very small and subcircular, less than half diameter of 2p lobe.
Frontal lobe short, occupying one-fifth glabellar length ; gently rounded anteriorly. Occipital ring is widest
medially, narrows distally ; more strongly vaulted than glabella. Occipital furrow is narrow (sag.) medially,
but broadens and deepens laterally. Preglabellar area is twice frontal lobe width (tr.) ; it is long, occupying
almost one-quarter cranidial length (sag.); from deep, broad (sag.) furrow it slopes upwards at 45°, then
curves over in broad roll (PI. 10, figs. 3, 9, 10) and becomes directed postero-ventrally to rostral suture
which is situated below level of furrow. In dorsal aspect border is gently convex forward ; laterally acuminate
towards anterior of facial suture. Axial furrow sinuous and broad (PI. 10, fig. 2), abaxial wall gently rounded
anteriorly; highest at eye lobe, which is prominently raised (PI. 10, fig. 11) and situated opposite 2p lobe.
Fixigena steeply declined to posterior border and flattens postero-laterally to bluntly rounded genal angle.
Anterior branches of facial suture moderately convergent forward ; posterior branch directed transversely
initially from eye lobe, then recurved to bisect genal angle. Librigena triangular ; bears prominent lateral
border which narrows posteriorly; border furrow shallow.

Hypostome (PI. 10, fig. 6) longer than broad; bears a moderately convex median body; median furrow
faint, demarkating long (sag.) anterior lobe and short posterior lobe which is crescentic in outline. Posterior
furrow bears a medially positioned depression. Lateral border furrow bears depression slightly posterior
of lateral part of posterior lobe. Lateral border narrow, rolled, deeply embayed posterior to anterior wings.
Posterior wings situated opposite posterior lobe. Posterior border is bilobed.

Rostral plate and thorax unknown. Pygidial axis bears five rings (PI. 10, fig. 13) ; axis tapers posteriorly to
broad, bluntly rounded tip ; end piece occupies almost one-third length of axis ; axial furrows become fainter
posteriorly. Pleural area gently sloping except near border where it strongly declines ; bears five pleurae and
deep pleural furrows which terminate just short of smooth border. First four pleurae bear faint interpleural
furrows which, close to axial furrow, are situated closer to posterior of pleura, but become medially
positioned distally.

Discussion. The range of Gravicalymene susi (Cautleyan Stage, Zones 2 and 3) lies
between that of G.jugifera Dean (1962, p. 1 16, pi. 13, figs. 9, 1 1 ; pi. 14, figs. 3, 4, 8, 9),
which is restricted to Pusgillian strata in the Cross Fell Inlier (Dean 1962) and the
Cautley district (Ingham 1977, p. 94), and G. deani Ingham (1977, p. 96, pi. 20, figs. 16,
17; pi. 21, figs. 1-6) which, though occurring rarely in Zone 2, is most common in
Zone 3 of the Cautleyan Stage in the Cautley district. G. susi is also intermediate in
morphology between G.jugifera and G. deani. It compares with the older G.jugifera

74

PALAEONTOLOGY, VOLUME 22

in possessing a similarly inclined preglabellar area (though it is relatively longer in
G. susi and has developed a thicker, rolled anterior part) and glabella with well-
rounded frontal lobe which projects beyond the fixigena. The glabella of G. susi,
however, is much more strongly ‘bell-shaped’, the 2p lobe being much smaller relative
to the lp lobe (compare PI. 10, fig. 2 with Ingham 1977, pi. 20, fig. 16 and pi. 21, fig. 2).

Both G. deani and G. susi possess this strongly ‘bell-shaped’ glabella, though it is
even more accentuated in G. susi due to the very small 2p lobes. However, the pre-
glabellar area of G. deani is shorter, occupying only one-sixth of the cranidial length,
more transverse anteriorly and is more strongly upturned and more broadly rolled.
The glabella of G. deani does not project as far forward anteriorly as in G. susi and
has a more transverse frontal lobe. G. susi can be particularly distinguished from
G. jugifera and G. deani by its broad axial furrows.

It would seem likely that these three geographically related species reflect parts
of a phylogenetic lineage. The preglabellar areas of both calymenid species which
occurred during Zone 2 and in early Zone 3 in the southern Lake District, that is
Calymene (s.l.) subdiademata and G. susi, are morphologically similar, their develop-
ment probably reflecting the occupation of similar ecological niches. Consequently,
the form of the preglabellar area must be afforded low taxonomic status in these
calymenids (as suggested by Whittington 1971, p. 459) as it may be a homeomorphic
feature. It would seem unwise to use it as the basis on which to postulate phylogenetic
relationships as Ingham (1977, p. 97) considered may have been possible between
G. deani and G. hagani Ross.

The type species of Gravicalymene, G. convolva Shirley (1936, p. 409, pi. 29,
figs. 16-18) from the Birdshill Limestone (early Ashgill) of south Wales, has a much
larger 2p lobe than G. susi and a more strongly upturned preglabellar area, rather
like that of G. deani.

A specimen (PL 10, fig. 7) from ‘Applethwaite Common’ (probably SM A9569)
referred to C. brevicapitata Portlock by McCoy (1851, p. 165, pi. If, fig. 6), sub-
sequently referred to C. senaria Conrad by Salter (1865, p. 97), mentioned by Shirley
(1931, p. 4) and then placed in G. deani by Ingham (1977, p. 97), is herein placed in
G. susi.

EXPLANATION OF PLATE 10

Figs. 1-13. Gravicalymene susi sp. nov. Applethwaite Formation (Cautleyan Stage, Zones 2 and 3).
1-4, holotype, BM It8689, dorsal, 1, view of internal mould, and dorsal, 2, lateral, 3, and anterior, 4, views
of cast of external mould of cranidium, from Brow Gill, Longsleddale (locality 4a), all x2. 5, 10, 11,
SM A98864, dorsal, 5, lateral, 10, and anterior, 11, views of cast of external mould of cranidium, from
Garboum Nook (locality 19), all x 2. 6, SM A98262, ventral view of internal mould of hypostome from
same locality as 1, x 3. 7, SM A9569, dorsal view of internal mould of cranidium, from ‘Applethwaite
Common’, x2; also figured by McCoy (1851, pi. If, fig. 6). 8, 9, SM A98865, dorsal, 8, and lateral,
9, views of internal mould of cranidium, from same locality as 5, both x 2. 12, SM A98430, dorsal view
of internal mould of cranidium, from Kentmere (locality 12a), x If. 13, SM A98261, dorsal view of
internal mould of pygidium, from same locality as 1, x 2.

Figs. 14-16. Toxochasmops marri (Reed), Applethwaite Formation (Cautleyan Stage, Zone 2). 14, GSM
Zs821, dorsal view of internal mould of pygidium, from ‘Marshall’s Park’, Coniston, x If. 15, 16, holo-
type, SM A40168, dorsal views of cast of the external mould, 15, and internal mould, 16, from ‘Apple-
thwaite Common’, x If; also figured by Reed (1894, pi. 7, figs. 1-3).

PLATE 10

McNAMARA, Coniston trilobites

76

PALAEONTOLOGY, VOLUME 22

Family pterygometopidae Reed, 1905
Subfamily chasmopinae Pillet, 1954
Genus toxochasmops gen. nov.

Type species. Trilobites extensus Boeck, 1838, p. 139, redescribed by Stormer (1941, p. 138, pi. 3, figs. 7-11);
from 4by-S, Gasokalven, Baerum, Norway.

Assigned species. T. extensus (Boeck 1838) [ Chasmops macrourus (Sjogren 1851)], T. wesenbergensis
(Schmidt 1881), T. eichwaldi (Schmidt 1881), T. marri (Reed 1894), T. amphorus (Salter 1864), T. bisseti
(Reed 1906).

Diagnosis. Cephalic length half to almost two-thirds cephalic width. Anterior border moderately to strongly
arched. Glabella gently convex; frontal lobe broad, and of low convexity, height/width ratio varying from
1 : 3-3 to 1 : 5. 3p glabellar lobe long, one-half glabellar length; 3p furrows sinuous, anteriorly divergent at
110°; 2p lobe extremely small. Central glabellar area narrow. Anterior branches of facial suture parallel to
anteriorly divergent. Eye lobe moderate sized, set close to glabella. Hypostome long, median body almost
twice as long as broad ; posterior tongue long, one-quarter hypostome length. Pygidium long and narrow
containing twelve to eighteen segments.

Discussion. Since its original conception, McCoy’s (1849) genus Chasmops has been
assigned twenty-nine species (Haller 1973, p. 729), although five of these have sub-
sequently been assigned to other genera. The forms encompassed by the generic
concept of Chasmops show a wide range of morphologies. Cephalic features which
vary are cephalic shape, length, and convexity of frontal lobe, relative sizes of 2p
and 3p glabellar lobes, size, elevation, and position of eye and orientation of anterior
branch of the facial suture. The hypostome shows a large variation in length. The
pygidium similarly shows a wide variation in length and number of segments. Future
work will show the phylogenetic relationships between the various chasmopines. For
the purpose of this work only the group containing ‘C.’ marri has been separated into
a new chasmopine genus.

The type species of Chasmops , C. odini (Eichwald), can be distinguished from all
species of Toxochasmops by its smaller 3p lobe, which is less than half the glabellar
length ; more prominent 2p lobe ; broader, shorter cephalon ; more strongly convex
frontal lobe ; eye more laterally disposed, such that the anterior branches of the facial
suture are anteriorly convergent ; shorter hypostome ; shorter, broader pygidium with
fewer than ten segments.

Toxochasmops marri (Reed 1894)

Plate 10, figs. 1-8; Plate 11, figs. 15-17; text-fig. 5

71845 Asaphus Powisi (Murchison); Sedgwick, p. 445.

71846 Phacops ( Asaphus ) Powisii Murchison; Sedgwick, p. 238.

1851 Chasmops odini 7 (Eichwald); McCoy in Sedgwick and McCoy, p. 164.
1864 Phacops ( Chasmops ) conophthalmus Boeck?; Salter, p. 41.

1867 Phacops conophthalmus Boeck; Hughes, p. 354.

1868 Phacops conophthalmus Boeck; Nicholson, p. 54.

71872 Phacops sp. ; Aveline and Hughes, p. 5.

1873 Phacops ( Chasmops ) macroura (Sjogren); Salter, p. 52.

1877 Phacops ( Chasmops ) macroura (Sjogren); Harkness and Nicholson, p. 468.

1878 Phacops ( Chasmops ) macroura (Sjogren); Marr, pp. 871, 873.

1878 Phacops ( Chasmops ) conophthalmus Boeck; Marr, pp. 871, 873.

1888 Chasmops odini 7 (Eichwald); Aveline, Hughes, and Strahan, p. 55.

McNAMARA: CONISTON TRILOBITES

77

1888 Phacops conophthalmus Boeck; Aveline, Hughes, and Strahan, p. 55.

1888 Phacops macroura (Sjogren); Aveline, Hughes, and Strahan, p. 55.

1892 Phacops cf. eichwaldi Schmidt; Marr, pp. 99, 109.

1892 Phacops ( Chasmops ) sp. ; Marr, p. 103.

1892 Phacops cf. brevispina Schmidt; Marr, p. 109.

1894 Phacops ( Chasmops ) marri Reed, pp. 241-246, pi. 7, figs. 1-3.

1906 Phacops marri Reed; Reed, p. 159.

1913 Phacops ( Chasmops ) sp.; Marr, p. 9.

1916 Phacops ( Chasmops ) marri Reed; Marr, pp. 191, 195, 197, 199, 200.

71934 Chasmops sp. ; King and Wilcockson, p. 1 1 .

1959 Chasmops cf. marri (Reed); Dean, p. 218, Table 3.

1963a Phacops ( Chasmops ) marri Reed; Dean, p. 50.

1963a Chasmops cf. extensa (Boeck); Dean, p. 54, pi. 1, figs. 1, 2, 5, 10.

1966 Chasmops marri (Reed) ; Ingham, pp. 466-468, 484-486, 489, 498-499.

1968 Chasmops sp. ; Whittington, p. 123.

1970 Chasmops ; Ingham and Wright, p. 234.

1973 Chasmops marry (sic) (Reed); Haller, p. 729.

1973 Chasmops sp. ; Price, pp. 233-234, Tables 2 and 3.

Holotype. SM A40168, an articulated specimen (PI. 10, figs. 15, 16) which lacks the posterior three thoracic
segments and the pygidium; from the Applethwaite Formation of ‘Applethwaite Common’, figured by
Reed 1894, pi. 7, figs. 1-3.

Material, localities , and horizons. This species occurs very commonly in the Stile End Formation (Cautleyan
Stage, Zone 2) comprising 40-7 per cent of the trilobite fauna ; disarticulated specimens have been collected
from the quarries above Stockdale Farm (localities 7a, 7b, 8a, 8b, 8c), and behind Kentmere Hall, Kentmere
(locality 38). It comprises 28 per cent of the trilobite fauna of the High Pike Haw Formation (Cautleyan
Stage, Zone 2), specimens principally coming from locality 32 at High Pike Haw. It is less common in the
calcareous mudstones of the Applethwaite Formation (Cautleyan Stage, Zones 2 and 3), thirty-five speci-
mens from Longsleddale (localities 3, 4c, 4e, 4f, 9b), Kentmere (locality 14a), Moor Head (locality 16b),
Garbourn Nook (locality 19), Torver (localities 23f, 43), comprising only 5-2 per cent of the trilobite fauna.

Emended diagnosis. Anterior border of cephalon with median downward flexure. Frontal lobe very broad
and with low convexity. Orle furrow set close to anterior border furrow. Eye lobe set close to glabella such
that exsagittal line passing lateral extremity of frontal lobe passes through mid-point of eye. Pygidium with
sixteen axial rings and fourteen pleurae.

Description. Cephalon sub-semicircular; the holotype shows an exaggerated length/breadth ration caused
by distortion. Less distorted material suggests the cephalon may have been three-fifths as long as broad.
Glabella is almost flat posteriorly to anterior tips of 3p lobe, then it declines anteriorly at about 45°. Frontal
lobe slopes laterally at about 30° below the horizontal. Glabella broadest anteriorly across frontal lobe,
maximum width being two-thirds cephalic width ; narrowest across 2p lobes where it is one-quarter maxi-
mum glabellar width. Frontal lobe occupies two-thirds length of glabella ; it is broad, ratio to maximum
across 3p lobes being 1 : 0-6 ; it has low convexity with a height to breadth ratio of about 1:4-5. An exsagittal
line passing lateral extremity of frontal lobe would pass through centre of eye lobe (PI. 10, fig. 3). 3p furrow
curves sinuously around 3p lobe, abaxially being directed inwards transversely before curving backwards
at 45° to a transverse line. Neither of the anterior two pairs of glabellar furrows is transglabellar, but stop
such that a central glabellar area remains between 3p lobes. This area is depressed below level of rest of
glabella and is one-fifth width of frontal lobe. 3p lobe is large and triangular ; its maximum length (exsag.)
is two-fifths length of glabella. Posteriorly 3p lobe extends past small lp and 2p lobes; it reaches a little
farther forward than anterior of eye lobe. 2p lobe is very small (PI. 10, figs. 3, 5). 2p furrow exists only
as inner part of deep furrow which bounds posterior of 3p lobe. This furrow bifurcates adaxially, one branch,
the 2p furrow, extending around anterior of 2p lobe, the other, the lp furrow running around anterior of
lp lobe. Deep coexistent lp and 2p furrow is directed forwards adaxially at 45°. lp lobe a little larger than
2p lobe. Occipital ring occupies two-sevenths width of cephalon, and is of constant length (sag. and exsag.),
occupying one-seventh length of cranidium. Occipital furrow narrow (sag. and exsag.), transverse and deep.

78

PALAEONTOLOGY, VOLUME 22

text-fig. 5. Reconstructions of Toxochasmops marri (Reed);
(a) dorsal view; ( b ) anterior view; (c) lateral view of cephalon,
approximately x 1 .

Axial furrows diverge forward at 70° and follow a sinuous course ; rise anteriorly from posterior border,
then narrow and shallow opposite 3p lobes before plunging steeply to anterior border. Preglabellar furrow
very shallow and bounded anteriorly by a thick convex border. Anterior view of cephalon shows that border
rises gradually until it is level with frontal lobe. It then becomes horizontal but dips down medially. Orle
furrow (PI. 1 1, fig. 7) runs close to and parallel with preglabellar furrow within frontal lobe. It is shallowest
antero-laterally and deepest medially, where it forms two pairs of prominent pits from which an oval of
small auxiliary impressions extends back over centre of frontal lobe. Eye lobe strongly crescentic, set close
to glabella and occupies almost one-quarter cranidial length ; anterior margin in line with anterior part of
3p lobe, posterior in line with posterior of 3p lobe. Eye surface bears thirty dorso-ventral files of lenses with
a maximum of sixteen and a minimum of four lenses in each file. A deep, broad furrow runs around base
of eye. Posterior branch of facial suture is directed outwards and slightly forwards then runs in a broad,
shallow furrow in a gentle, convex-forward curve to lateral border, which it intersects in line with posterior
of 3p lobe. Anterior branches of facial sutures diverge anteriorly at 60°; they do not intersect anterior
border but run transversely in preglabellar furrow. Fixigena is narrow (tr.) anteriorly, widens past eye
lobe, then narrows at first posteriorly before greatly widening towards posterior border ; steeply declines
laterally. Posterior border is a narrow (exsag.), raised rim adaxially, which widens and flattens towards genal
angle. As it joins lateral border it becomes vertically orientated and sweeps back as long, gradually narrow-
ing genal spine which reaches to posterior of thorax. Posterior border furrow is narrow and deep adaxially ;
shallows and broadens abaxially before degenerating as it approaches faint lateral border furrow which
similarly shallows towards genal angle. Librigena is gently convex; lateral border is broad, raised, rolled;
bounded by wide, shallow furrow. Glabella is covered by many large tubercles with intervening smaller,
adventious ones (PI. 11, fig. 5). There is a general diminution in tubercle size laterally. Tubercles are also
present on fixigena inside eye lobe.

Middle body of hypostome (PI. 1 1 , fig. 2) is widest opposite wide, triangular wings which extend hori-
zontally antero-laterally, but curve downwards through almost a right angle posteriorly. Wing process is
very prominent. Shoulder behind wing extends posteriorly to two-thirds hypostomal length. A deep middle
furrow begins opposite shoulder and extends inwards and backwards; behind this a crescentic lobe is
defined which occupies one-quarter hypostomal length.

McNAMARA: CONISTON TRILOBITES

79

Thorax (PI. 10, figs. 15, 16; PI. 11, fig. 1) consists of eleven segments. Axis occupies nearly one-third of
thoracic width and is strongly vaulted. Axial furrow is well defined, deep, and narrow. Pleurae almost flat
adaxially, but at fulcrum, which is set at three-quarters pleural width, they sharply decline ; bluntly and
abruptly terminated; bear deep, diagonal pleural furrow. Anterior pleural band narrow (exsag.) adaxially,
broadens abaxially at first, then narrows close to fulcrum.

Pygidium (PI. 10, fig. 14; PI. 11, figs. 1, 8) bears up to sixteen axial rings and fourteen pleurae, the last
two axial furrows being very faint. Axis tapers gently posteriorly, its maximum width occupying nearly
one-third anterior width of pygidium; it occupies four-fifths length of pygidium, the post-axial region
sloping downwards at first and bearing faint longitudinal pleural furrows before becoming smooth and
upturned. Axial furrow is deep and narrow, becoming shallower at tip of axis. Interpleural furrows are
deep and narrow adaxially becoming quite faint close to border ; posterior five or six pairs, in fact, fail to
reach border and become increasingly sinuous. Anteriorly interpleural furrows are straight and directed
almost transversely, but farther back become increasingly posteriorly directed, recurving close to lateral
border, last few pairs running almost parallel to axis. Anterior six pleurae bear faint pleural furrows which
become more poorly developed posteriorly.

Discussion. Dean (1963a, p. 54) referred the species of Toxochasmops from the Stile
End Formation to ‘ Chasmops cf. extensa Further collecting by the author has
revealed that this form is the same as that from the overlying Applethwaite Formation,
T. marri. Dean regarded the Stile End Formation as being Actonian in age, but it
has recently been shown (McNamara in press ) to be early Cautleyan (Zone 2) in age,
like the lower and middle parts of Applethwaite Formation. T. extensus can be
distinguished from T. marri in having the eye lobe positioned farther from the
glabella, a narrower frontal lobe and a pygidium with a greater number of segments.
In addition the glabellar convexity is greater than in T. marri.

The two British species most similar to T. marri are T. amphorus (Salter 1864, p. 42,
pi. 4, fig. 16) from the Crug Fimestone, south Wales, and T. bisseti (Reed 1906, p. 157,
pi. 20, fig. 1-3) from the upper Drummuck Group, Thraive Glen, Girvan. T. ampho-
rus was erected by Salter on a pygidium. Although having a similar number of seg-
ments as T. marri this specimen has an axis which is particularly broad anteriorly and
tapers strongly posteriorly. It is very stongly convex, the distal parts of the pleurae
being vertically inclined. A cranidium (SM A42810) from the same horizon and
locality as the holotype has been discovered in the Sedgwick Museum, Cambridge.
It is similar to the cephalon of T. marri, but lacks the pronounced median flexure of
the anterior border. In addition the orle furrow is not set as close to the anterior
border furrow and the 3p lobe is relatively longer. T. bisseti also lacks the median
flexure of the anterior border, has the eye lobe set closer to the glabella, and has a
narrower frontal lobe and a shorter pygidium with thirteen axial rings and fourteen
pleurae.

The two other species which show a close resemblance to T. marri are T. wesen-
bergensis (Schmidt 1881, pi. 4, figs. 10-12; pi. 5, figs. 1-7) from the Rakvere Stage (E)
and T. eichwaldi (Schmidt 1881, pi. 5, figs. 8-10) from the Nabala (Fxa), Vormsi (Fxb),
Pirgu (FjC), and rarely the Porkuni (F2) stages (RSomusoks 1953, p. 412) of Estonia.
The eye lobe of T. wesenbergensis is set farther from the glabella than in T. marri,
whilst in T. eichwaldi the eye lobe is closer to the glabella and positioned more
posteriorly. Neither of these species has the median flexure of the anterior border,
whilst T. wesenbergensis has two fewer pygidial axial rings than T. marri.

In addition to its occurrence in the Fake District, T. marri also occurs in the
Cautley district where it is restricted to the Cautleyan Stages, Zones 1-3 (Ingham

80

PALAEONTOLOGY, VOLUME 22

1966, p. 486). Material from the Dholhir Beds of the Berwyns, north Wales (Whit-
tington 1968, p. 121) has been examined and can be referred to T. marri as can
‘ Chasmops sp.’ from the Sholeshook Limestone, south Wales, listed by Price (1973,
tables 2 and 3).

Family odontopleuridae Burmeister, 1843
Subfamily odontopleurinae Burmeister, 1 843
Genus acidaspis Murchison, 1839

Type species. By monotypy; Acidaspis brightii Murchison, 1839, p. 658, pi. 14, fig. 15; from the Wenlock
Limestone.

Acidaspis magnospina Stubblefield, 1928

Plate 11, figs. 9-14; Plate 12, figs. 1-9; text-fig. 6

1888 Acidaspis Brightii Murchison; Aveline, Hughes, and Strahan, p. 55.

1891 Acidaspis n. sp. ; Nicholson and Marr, p. 511.

1892 Acidaspis n. sp.; Marr, p. 108.

1928 Acidaspis magnospina Stubblefield, pp. 428-433, p. 14, figs. 1-6.

1933 Acidaspis magnospina Stubblefield; Warburg, pp. 15-16.

1937 Acidaspis magnospina Stubblefield; Opik, p. 43.

1949 Acidaspis magnispina (sic) Stubblefield; Prantl and Pribyl, pp. 26, 141.

1956 Acidaspis magnospina Stubblefield; Whittington, p. 513.

1962 Acidaspis magnospina Stubblefield; Dean, p. 50.

1963a Acidaspis magnospina Stubblefield; Dean, p. 123.

1966 Acidaspis magnospina Stubblefield; Ingham, pp. 465, 467-468, 486.

1967 Acidaspis magnospina Stubblefield; Ross, p. B17.

1970 Acidaspis sp. ; Ingham and Wright, p. 237.

1973 Acidaspis cf. magnospina Stubblefield; Price, p. 244.

EXPLANATION OF PLATE 11

Figs. 1-8. Toxochasmops marri (Reed), Applethwaite Formation (Cautleyan Stage, Zones 2 (4, 7) and 3
(1, 2, 5)), Stile End Formation (Cautleyan Stage, Zone 2) (3, 8), High Pike Haw Formation (Cautleyan
Stage, Zone 2) (6). 1 , 2, SM A98867, dorsal view of internal mould of articulated specimen which lacks
only its librigenae, 1, and ventral view showing the hypostome in position, 2, from Garboum Nook
(locality 19), both x 1. 3, SM A98111, dorsal view of internal mould of cephalon, from Stockdale,
Longsleddale (locality 7b), x 1^. 4, 7, BM It8678, dorsal, 4, and anterior, 7, views of internal mould of
cephalon, from Brow Gill, Longsleddale (locality 4a), both x 2. 5, SM A98857, dorsal view of cast of
external mould of incomplete cranidium, from same locality as 1, x2. 6, SM A98245, dorsal view
of internal mould, from High Pike Haw (locality 32), x If. 8, SM A98 114, dorsal view of internal mould
of pygidium, from same locality as 3, x If.

Figs. 9-14. Acidaspis magnospina Stubblefield, Applethwaite Formation (Cautleyan Stage, Zone 3); all
from Garbourn Nook (locality 19). 9, SM A98783, cast of external mould of ventral surface of occipital
spine showing the median groove and distribution of tubercles, x 2\. 10, SM A98786, dorsal view of
internal mould of pygidium, x3f. 11, SM A98839, dorsal view of internal mould of pygidium, x7.
12, 13, BM In27066, fig. 12 showing Stubblefield’s (1928, pi. 14, figs. 1, 2) holotype and a paratype; the
specimen lower right is the holotype ; the specimens are resting against a cranidium of Calymene sub-
diademata; fig. 12 is the cast of the external mould; fig. 13 shows a dorsal view of the internal mould of
the holotype ; both figs, x 2\. 14, SM A98772, internal view of the upper part of the librigena showing the
eye surface, x 12.

PLATE 11

McNAMARA, Coniston trilobites

82

PALAEONTOLOGY, VOLUME 22

Holotype. BM In27066, incomplete cephalon (PI. 11, figs. 12, 13) from the Applethwaite Formation
(Cautleyan Stage, Zone 3), Garbourn Nook (locality 19); figured by Stubblefield 1928, pi. 14, figs. 1, 2.

Material, localities, and horizon. Twenty-five disarticulated specimens are known from the same area as
the type specimens (Applethwaite Common, localities 16a, 19), those from locality 19 being topotype
material; the localities are in the upper part of the Applethwaite Formation (Cautleyan Stage, Zone 3). One
specimen has been collected from Stunfel Howe (locality 14c).

Emended diagnosis. Glabella with ill-defined 3p lobe and narrow, well-defined central glabellar area.
Librigenal and occipital spines very long, twice cephalic length. Hypostome with poorly developed posterior
lobe and well-developed anterior wings. Pygidium bearing six pairs of border spines, three pairs being
anterior to major pair.

text-fig. 6. Reconstruction of Acidaspis magnospina
Stubblefield; oblique antero-lateral view, approximately
x 4.

Description. Cephalon sub-oval, bearing three posteriorly directed spines, two librigenal, one occipital, of
equal length ; length of spines nearly three times glabellar length (PI. 12, figs. 1 , 2) extending back to posterior
of pygidium. Cranidium semicircular and strongly convex. Glabella widest posteriorly; five-eighths maxi-
mum cranidial width ; tapers gently forward from posterior then, anterior to 2p lobe, tapers more strongly
forward. Transverse width, posteriorly, is a little greater than sagittal length. Glabellar furrows delineate
rectangular central glabellar area whose sagittal length is almost three times transverse width; strongly
convex transversely, whilst longitudinally it curves forwards and downwards through an arc of 75° to
terminate against short (sag.) shallow preglabellar furrow which is bounded anteriorly by gently inclined
border. Ip lobe oval, with long axis set at 20° to sagittal line; occupies one-half length of glabella; lp
furrow deep; from axial furrow directed postero-medially from axial furrow at 45°, then recurves to
run back exsagittally to occipital furrow, thus isolating glabellar lobe from central glabellar area. 2p lobe
sub-oval and occupies just less than half area of lp lobe. 2p furrow runs postero-medially from axial furrow
at 20° to sagittal line, then recurves to run exsagittally to join lp furrow genuflection. A very faint swelling
on side of glabella a little in front of 2p lobe is indicative of a 3p lobe (PI. 12, fig. 3). Occipital furrow
broad (sag.) and shallow; declines steeply laterally, curving backwards, as it is met by lp furrow, to meet
posterior end of axial furrow in a depression behind lp lobe. Within this broad depression is a raised occi-
pital lobe. Occipital ring developed posteriorly into a long, stout spine (PI. 1 1, fig. 12). Anteriorly occipital
area is broad, extending laterally almost to behind eye lobe; rapidly narrows posteriorly to half this width,
then continues as long, stout, hollow spine which gradually tapers posteriorly and occupies three times
length of glabella, extending almost to posterior of pygidium. Initially it rises steeply from occipital furrow,

McNAMARA: CONISTON TRILOBITES

83

at 60° to horizontal, to twice cranidial height ; then, at one-third its length from occipital furrow it becomes
horizontally orientated (PI. 12, fig. 1). Ventral surface of spine rises more shallowly than dorsal surface
initially; it then steepens before arching back posteriorly. This results in the development on internal
moulds, where the spine is missing, of Stubblefield’s (1928, p. 429) transverse ‘post-occipital line’ (PI. 12,
fig. 9). In transverse cross-section, spine is ‘kidney-shaped’, dorsal surface being strongly convex, whilst
ventral surface carries a medially positioned, dorsally directed groove (PI. 11, fig. 9). Ventro-laterally
twelve prominent, large tubercles are present (PI. 11, fig. 9; PI. 12, fig. 1) which are arranged such that
anterior four pairs are equally spaced and next two pairs set close together, this sequence being repeated.

Axial furrow deep posteriorly but shallows as it rises out of occipital furrow, becoming very shallow and
narrow opposite 2p lobe; deepend anterior to this. Fixigena comprises an inner ridge (which runs parallel
to axial furrow and degenerates opposite 2p furrow) and an outer, narrower ridge (the eye ridge) which
parallels inner ridge, but is longer, curving inwards anteriorly to abut against glabella in front of 2p lobe
(PI. 12, fig. 3). Inner ridge carries three rows of tubercles posteriorly which merge to form a single row
anteriorly; eye ridge bears a single, sinuous row of about eight tubercles (PI. 12, fig. 3). Eye lobe set far
back opposite posterior of lp lobe (PI. 12, fig. 2). Anterior branches of facial suture almost parallel close to
eye, but become increasingly more convergent anteriorly; meet anterior border in front of 2p lobe. Area
of fixigena between facial suture and eye ridge develops as an anteriorly widening depression which bears
an apodemal pit antero-lateral to 2p lobe ; it then shallows into anterior border. Posterior branch of facial
suture runs almost transversely for a short distance from eye lobe before curving postero-laterally to meet
posterior border close to inner part of librigenal spine (PI. 12, fig. 6). Librigena bordered by raised, narrow,
convex rim that bears eleven border denticles which lengthen and become more horizontally orientated
posteriorly. Lateral border furrow is broad and shallow. Librigena narrows upwards towards eye; holo-
chroal eye surface is semi-cylindrical, and domed dorsally (PI. 11, fig. 13). Librigena bears a long, curved
slender spine (PI. 1 1, fig. 13; PI. 12, figs. 2, 8, 9) which reaches as far back as occipital spine. In addition to
occurrence of tubercles on fixigena, tubercles are also present on central glabellar area and glabellar lobes,
dorsal surface of occipital spine and dorsal surface of librigenal spine ; these tubercles are arranged irregu-
larly. A further cluster of large tubercles is set on librigena below eye. A prominent swelling is present on
occipital ring anterior to development of occipital spine.

Hypostome (PI. 12, fig. 4) slightly wider (tr.) than long (sag.). Middle body moderately convex, sub-
circular in outline. Posterior lobe crescentic; ill-defined anteriorly by faint furrow. Lateral and posterior
furrows deep, wide. Posterior border a little inflated laterally. Lateral border narrows anteriorly before
widening into short anterior wings ; acuminate postero-laterally. Anterior border is straight.

Total number of thoracic segments unknown, though it may, like all other species of Acidaspis, have
possessed ten. One incomplete specimen (SM A43418; PI. 12, fig. 5) bearing seven displaced thoracic
segments, shows the axial rings to curve forwards laterally as they decline and thicken into axial furrow.
Anterior two segments bear short (tr.) pleurae which only extend as far as inner part of librigenal spine.
Width (tr.) of pleurae from axial furrow to fulcrum in succeeding pleurae is far greater than in anterior two
pairs. Pleurae bear short lateral spines. Broad (exsag.) pleural furrow separates short convex anterior band
from longer convex posterior band.

Pygidium (PI. 11, figs. 10, 11) bears a triangular axis which carries a large, convex anterior ring and a
smaller, lower, posterior ring which tapers back to a small, flat posterior plate that reaches almost to
posterior border. Anterior ring extends laterally into a pleural ridge which curves posterially to run at 25°
from the sagittal line. It continues across border, extending back as a long, stout spine which is three times
as long as the pygidium. Semicircular border bears three spines antero-lateral to macrospine, and two
postero-laterally ; they all extend at about 25° from the sagittal line. Thus border bears a total of twelve
spines. Macrospine carries a single, small tubercle close to its junction with axial ring; otherwise pygidial
surface is smooth.

Discussion. A. magnospina Stubblefield is the oldest-known representative of this
genus in Britain . The only other British Ordovician species is A . aster oidea Reed (1914,
p. 31, pi. 5, figs. 3-7) from the upper Drummuck Group, Girvan. This species differs
from A. magnospina in having a shorter occipital spine and bearing one less pair of
pygidial spines. Two species of Acidaspis of Ordovician age occur in Estonia:
A. viruana (Opik 1937, p. 43, pi. 5, fig. 1; Bruton 1968, p. 296, figs. 8, 9?) and

84

PALAEONTOLOGY, VOLUME 22

A. aviensis Bruton (1968a, p. 299, figs. 11, 12). A. viruana is probably a little older
than A. magnospina, occurring in the Rakvere Stage (E); it differs in possessing a
shorter occipital spine, more prominent 3p lobe, and broader central glabellar area.
The pygidium which Bruton (1968a, figs. 9, 10) referred to this species has one fewer
pair of border spines anterior to the major pair. A. aviensis from the Porkuni Stage
(F2), is a little younger than A. magnospina ; it lacks an occipital spine (Bruton 1968a,
fig. 11), has a prominent 3p lobe, this being absent or minute in the British species,
and a ‘bicomposite’ lp lobe.

A. cincinnatiensis Meek (1873), redescribed by Whittington (1956, p. 512, pi. 58,
figs. 9-11, 13, 14; pi. 59, figs. 1-3, 6, 9-11), is a little older than A. magnospina ,
occurring in upper Trentonian to Maysville strata. It has shorter cephalic spines than
A. magnospina , prominent 3p lobe, and greater number of pygidial spines. The
hypostome of A. cincinnatiensis has a more well-defined posterior lobe, shorter
anterior wings, and lateral projection of the lateral border set more anteriorly. Ross
(1967, p. B17, pi. 5, figs. 19, 20) mentions a species of A cidaspis from Kentucky which
was approximately contemporaneous with A. magnospina. This species also appears
to lack a 3p lobe, but, as Ross states, it differs from A. magnospina in having a shorter
occipital spine.

A. magnospina also occurs in the Cautley district (Ingham 1966), where it ranges
from the Pusgillian Stage to Zone 2 of the Cautleyan Stage, or even possibly as high
as Zone 4. Price (1973) records A. cf. magnospina from the Birdshill Limestone in
south Wales. These specimens have been seen and are similar to the type material.
A. magnospina has also been recorded from the Dufton Shales (Dean 1962) of higher
Pusgillian age in the Cross Fell Inlier. It appears to be restricted to early Zone 3 in
the Lake District.

EXPLANATION OF PLATE 12

Figs. 1-9. Acidaspis magnospina Stubblefield, Applethwaite Formation (Cautleyan Stage, Zone 3) ; all from
Garboum Nook (locality 19). 1, 2, 3, SM A98768, lateral, 1, dorsal, 2, and lateral, 3, views of internal
mould of incomplete cephalon, all x 3. 4, SM A98843, ventral view of internal mould of hypostome,
4. 5, SM A43418, dorso-lateral view of cast of external mould of incomplete, articulated individual,
< 2. 6, 7, SM A98878, dorsal, 6, and anterior, 7, views of internal mould of cephalon, x 3. 8, SM A98787,
lateral view of cast of external mould of librigena, x 3}. 9, BM In27067, dorsal view of internal mould of
incomplete cephalon, x 2.

Figs. 10-19. Primaspis bucculenta sp. nov. Torver Formation (10-14, 17-19) and upper Applethwaite
Formation (16), both Cautleyan Stage, Zone 3; Staurocephalus Limestone (15). 10, SM A43127, dorsal
view of internal mould of cranidium, from Old Pits Beck (locality 35b), > 4. 11, SM A43126, dorsal
view of internal mould of incomplete cranidium, from same locality as 12, x8. 12, 13, 14, holotype,
SM A98931, lateral, 12, anterior, 13, and dorsal, 14, views of cast of external mould of cranidium, from
Willy Scrow (locality 25a), x6. 15, SM A32845b, lateral view of cast of external mould of librigena,
from Swindale, Knock, x 6. 16, SM A98713, dorsal view of cast of external mould of pygidium, from
Moor Head, Troutbeck (locality 18a), x 6. 17, SM A43168, lateral view of internal mould of librigena,
from same locality as 12, x 6. 18, SM A43120, dorsal view of internal mould of pygidium, from same
locality as 12, x 6. 19, SM A43227, dorsal view of cast of external mould of pygidium, from same locality
as 12, x6.

PLATE 12

McNAMARA, Coniston trilobites

PALAEONTOLOGY, VOLUME 22

Genus primaspis R. and E. Richter, 1917

Type species. Odontopleura primordialis Barrande, 1846, from the Liben Formation, Dedu Berouna,
Bohemia, figured by Barrande 1852, pi. 37, fig. 14 and refigured by Bruton 19686, pi. 1, fig. 9.

Primaspis bucculenta sp. nov.

Plate 12, figs. 10-19

1891 Staurocephalus globiceps Portlock; Nicholson and Marr, p. 505.

1913 Acidaspis cf. dalecarlica Tornquist; Marr, p. 7.

1916 Acidaspis asteroidea Reed?; Marr, p. 199.

1916 Acidaspis sp. nov.; Marr, p. 199.

1916 Staurocephalus murchisoni Barrande?; Marr, p. 199.

1948 Acidaspis cf. asteroidea Reed; King and Williams, p. 210, pi. 16, fig. 4.

?1966 Primaspis cf. dalecarlica (Tornquist); Ingham, pp. 466-467, 470, 486, 497-502.

1967 ‘ Acidaspis cf. dalecarlica' (Tornquist); Bruton, p. 7.

1967 Primaspis evoluta (Tornquist); Bruton, p. 7 (pars).

Holotype. SM A98931 (PI. 12, fig. 1), cranidium from the Torver Formation (Cautleyan Stage, Zone 3),
the base of Willy Scrow (locality 25a).

Material , localities, and horizons. Twenty-four disarticulated cranidia, librigenae, and pygidia are known;
all, except for three pygidia and one cranidium, are from the Torver Formation (Cautleyan Stage, Zone 3)
from the region of Torver Beck (localities 21e, 23e, 35b) and from the base of Willy Scrow (locality 25a).
Of the other three specimens, two were found north of Applethwaite Common (locality 18a), near the top
of the Applethwaite Formation (Cautleyan Stage, Zone 3) whilst a cranidium was found by Dr. J. K.
Ingham at the same horizon at Kentmere (locality 41b). The species is thus restricted to Zone 3 in the Lake
District.

Diagnosis. Broad fixigena and basal glabellar lobe, each as wide as central glabellar area posteriorly.
2p lobe two-thirds size of lp lobe. Librigenal spine strongly swollen where joins border. Pygidium bears ten
spines ; only one pair antero-lateral to macrospine ; pygidium three times as wide as long.

Description. Cranidium wider than long; strongly convex. Glabella probably slightly wider than long, being
widest across lp lobes, then gently narrowing to 2p lobes, anterior to which it rapidly narrows, lp lobe oval,
with long axis diverging forwards from exsagittal line at 20° (PI. 12, figs. 10, 1 1) ; occupies one-third posterior
glabellar width and a little under one-half glabellar length ; occupies same width as central glabellar area,
lp furrow directed initially postero-medially at 30° from transverse line, then curves back and runs exsagit-
tally to occipital furrow, thus isolating lp lobe from central glabellar area. 2p lobe subcircular, occupying
about two-thirds area occupied by lp lobe. 2p furrow directed postero-medially more acutely than lp
furrow, at about 45°, then, adaxially, runs exsagittally to meet inner part of lp furrow. Isolated central
glabellar area approximates to a rectangle but narrows opposite glabellar lobes (PI. 12, fig. 1 1). At its widest
it is only just as wide as lp lobe, at its narrowest as wide as 2p lobe. 3p lobe very small (PI. 12, fig. 14),
elongate, with its long axis set close to exsagittal line. Frontal lobe very short, narrow and vertically declined
(Plate 12, fig. 12). Occipital furrow broad (sag.) and shallow. Occipital ring one-quarter as long as glabella;
strongly vaulted; laterally it narrows at first, then widens into occipital lobe, then degenerates a little as
abuts against posterior border behind lp lobe. Small tubercle positioned towards posterior end of occipital
ring (PI. 12, fig. 14). Axial furrows diverge from posterior border, but swing round to converge anteriorly.
Fixigena broad posteriorly, being as broad as lp lobe; gradually narrows forwards as it swings inwards,
degenerating outside 2p lobe. Eye ridge separated from genal ridge by a deep furrow ; it commences at base
of eye lobe, opposite posterior of lp lobe, then runs parallel to genal ridge, merging and degenerating with
it anteriorly. Preglabellar field (PI. 12, fig. 10) gently upturned medially, more strongly upturned laterally.
Posterior border rim very narrow adaxially; broadens (exsag.) laterally, being very swollen where crossed
by posterior branch of facial suture. Posterior border furrow concomitantly deepens laterally. Posterior
branch of facial suture runs at 45° towards posterior border. Anterior branch of facial suture gradually
diverges from eye ridge anteriorly, meeting anterior border in line with posterior part of fixigena. Librigena

McNAMARA: CONISTON TRILOBITES

87

(PI. 12, figs. 15, 17) steeply declined below eye lobe; rapidly widens as declines into broad, shallow border
furrow ; bounded laterally by rounded narrow border which rapidly widens towards genal angle ; at genal
angle border extremely swollen where met by short, stout librigenal spine (PI. 12, fig. 15). Lateral border
bears fourteen denticles (PI. 12, fig. 17). Eye surface sub-spherical, composed of many gently curving
rows of holochroal lenses, eye surface bounded ventrally by strong furrow, below which cheek is
tuberculated. Borders are tuberculated, those on swollen part being larger than those on narrower
borders.

Thorax and hypostome are unknown. Pygidium (PI. 12, figs. 16, 18, 19) three times as wide as long
(excluding spines) ; bears a triangular axis which comprises a high, convex, wide (tr.) anterior ring, a lower,
narrower posterior ring, and a flat triangular plate which joins raised, convex posterior border. Anterior
ring develops laterally into a raised pleural ridge which curves postero-laterally at 30° to transverse line,
gradually thickening as it approaches border, whereupon it swells appreciably and extends beyond border as
major pygidial spine for twice length of minor spines. Semicircular border bears three pairs of minor spines,
central two pairs being positioned equidistantly about the sagittal line, whilst third is smaller and develops
out of base of major spine. Only one spine present antero-lateral to macrospine; making a total of ten
pygidial spines. Axial rings carry small irregularly positioned tubercles. Similar sized tubercles occur on
spines and posterior border; become larger and more concentrated at junction of spines with border.
Tubercles are also scattered on the flat, depressed pygidial border.

Discussion. Bruton (1967, p. 7) assigned these specimens from Torver to Primaspis
evoluta (Tornquist). It can be shown that a number of important differences occur
which separate this form into a different species. The posterior of the fixigena, the
lp lobe and the central glabellar area are all approximately equal in width in
P. bucculenta, whereas the central glabellar area is wider than the lp lobe, which in
turn is wider than the fixigena in P. evoluta (Bruton 1967, pi. 1, figs. 4, 9). Relative
to the lp lobe, the 2p lobe is larger in P. bucculenta. The characteristic indentation of
the adaxial part of the lp lobe in P. evoluta is absent in the Lake District species.
The occipital ring is shorter (sag.) in P. bucculenta. The pygidium of P. bucculenta
differs from P. evoluta in being three times, not four times, as wide as long, and it
carries one less spine antero-lateral to the macrospine.

P. semievoluta , Reed 1910, p. 214, pi. 17, figs. 1-3, from the Longvillian of the
Cross Fell Inlier has a narrower fixigena than P. bucculenta , lacks the minor spine
fused to the macrospine on the pygidium and has one extra spine antero-lateral to it.
P. caractaci (Salter) from the late Caradoc of south Shropshire differs in possessing
a greater number of pygidial spines and narrower fixigena (Dean 19636, p. 239, pi. 44,
figs. 3, 9). P. girvanensis (Reed 1914, p. 33, pi. 5, figs. 8-10; pi. 6, figs. 1-3) from the
upper Drummuck Group, Girvan, has a wider central glabellar area, larger lp lobe,
lacks the fused spine on the pygidium and bears two spines antero-lateral to the
macrospine.

Coniston Limestone trilobites described and illustrated by McCoy

In his description of Palaeozoic fossils, McCoy (1851) discussed and mentioned
eight species of trilobites from the Coniston Limestone of the Lake District, seven
of which he figured. The present generic and specific assignments of some of these
have been discussed earlier: SM A9569 (McCoy, pi. If, fig. 6; PI. 10, fig. 7 herein)
designated by McCoy as Calymene brevicapitata Portlock is now believed to be
Gravicalymene susi sp. nov. ; SM A6806 (McCoy, pi. If, fig. 10; PI. 9, fig. 9 herein),
one of two illustrated syntypes of C. subdiademata McCoy, has been herein selected
as the lectotype of that species; SM A41905 (McCoy, pi. If, fig. 12; PI. 8, figs. 9, 11

PALAEONTOLOGY, VOLUME 22

herein), which McCoy called Cheirurus clavifrons Dalman, has been redesignated
herein as the lectotype of Pseudo sphaerexochus boops (Salter).

A pygidium (SM A41909 ; text-fig. 7 a), probably from the Applethwaite Formation
of ‘Coniston’, was the basis of the new species Lichas subpropinqua proposed by
McCoy (1854, pi. If, fig. 17) in a footnote to the plate explanation. He had earlier
(McCoy 1851, p. 150) questioningly placed it in L. propinqua Barrande. Salter (1873,
p. 50) later referred it to L. laciniatus (Wahlenberg) without comment. Although the
specimen has been transversely shortened by lateral compression, it shows the
characteristic features of L. laciniatus , redescribed by Warburg (1925, p. 295),
notably the shape of the pleurae, pleural and interpleural furrows, and the gently
rounded postero-lateral borders which meet to form a pointed posterior termination.
Reed (1896, p. 427, pi. 31, fig. 10) described L. conformis keisleyensis on the basis of
a pygidium from the Keisley Limestone. This form appears to be identical with that
from the Lake District. Warburg (1925, p. 300) referred this subspecies to
L. laciniatus, as did Dean (1974, p. 79).

text-fig. 7. Three pygidia from the Applethwaite Formation figured by McCoy (1854, pi. If, fig. 17, a;
pi. Ig, figs. 2, 3, b; pi. 1g, fig. 8, c): a, Lichas laciniatus (Wahlenberg), SM A41909, internal mould from
‘Coniston’; b, Atractopyge verrucosa (Dalman), SM A41904, cast of external mould from Applethwaite
Common; c, Cybeloides ( Paracybeloides ) girvanensis (Reed), GSM Z1 9108, internal mould from ‘Coniston

Water’.

McCoy (1851, p. 157) based the species Zethus atractopyge on two pygidia, one
SM A41904 (McCoy, pi. Ig, figs. 2, 3; text-fig. lb herein) from the Applethwaite
Formation of ‘Applethwaite Common’, the other, SM A41875 (McCoy, pi. Ig,
figs. 5, 6) from the mid-Caradoc near Meifod, and an incomplete partly disarticu-
lated specimen (SM A41903) from ‘Ravenstonedale’ (McCoy, pi. Ig, fig. 1). Dean
(1961, p. 319) subsequently chose the Meifod specimen as the lectotype of Atracto-
pyge atractopyge. This pygidium differs from the Lake District form in possessing a
pygidial axis which tapers less rapidly posteriorly. The ‘Ravenstonedale’ specimen (a
disarticulated cephalon and thorax, not a pygidium as stated by Ingham (1975,
p. 81) can, on account of the almost equi-sized 2p and 3p lobes, be assigned to
A. cf. verrucosa (Dalman), as can the Lake District form. This species is based on a

McNAMARA: CONISTON TRILOBITES

recently rediscovered cranidium (see Dean 1974, p. 97, text-fig. 4), probably from the
Crug Limestone, south Wales. Cranidia collected from the same area in the Lake
District as the pygidium SM A41904 are like the type specimen of A. verrucosa.

McCoy (pi. 1g, fig. 8) illustrated, though did not discuss, a pygidium (GSM
Z1 9108) from ‘Coniston Water’ (text-fig. 7c) which he called Z. rugosus (Portlock).
This pygidium is reassigned to Cybeloides {Paracybeloid.es) girvanensis (Reed).

The three specimens (SM A4 1906-4 1908) from Sunny Brow, which McCoy (p. 172,
pi. 1g, figs. 33-35) assigned to Illaenus rosenbergi Eichwald, formed the basis of a
new species, I. marshalli, erected by Salter (1867, p. 200). Later authors, including
Eichwald (1860, p. 1483), retained this species in I. rosenbergi. The cranidium (SM
A41906), however, has a much shorter eye lobe than that of I. rosenbergi (Eichwald
1825, pi. 3, fig. 3). As Ingham (1970, p. 25) suggested, these poorly preserved speci-
mens appear to be identical with Stenopareia bowmanni (Salter). The alleged distinc-
tive feature of 7. marshalli' was the surface pattern of irregular rugae. These, however,
parallel the cleavage direction and are consequently considered to be tectonic
features associated with transverse shortening. Salter, in erecting I. marshalli,
further considered that the eye lobe was positioned closer to the glabella, yet further
forward than in S. bowmanni. In comparison with topotype material of S. bowmanni
figured by Price (1974, pi. 1 12, figs. 1-8), there appears to be no appreciable difference
in eye lobe position between the two forms. Salter’s (1867, pi. 29, fig. 3) figure of the
Lake District cephalon erroneously shows the right eye lobe closer to the glabella
than it is on the actual specimen.

Lastly, the specimens from ‘Applethwaite Common’ and ‘Coniston Water Head’
which McCoy (1851, p. 164) referred to Chasmops odini (Eichwald) belong to Toxo-
chasmops marri (Reed).

Acknowledgements. Professor H. B. Whittington, and Drs. J. K. Ingham and C. P. Hughes have proffered
useful suggestions for the improvement of this work, for which I am grateful. Drs. D. Price (SM), A. Rush-
ton (GSM), R. Fortey, S. Morris (BM), and J. K. Ingham (HM) have kindly lent material in their care.
Assistance in the field was given by Professor and Mrs. H. B. Whittington, S. Radford, and S. Conway
Morris, to all of whom I offer my thanks. Much of this work was carried out whilst in receipt of a
N.E.R.C. Studentship at the Sedgwick Museum.

REFERENCES

angelin, n. p. 1854. Palaeontologia Scandinavia, I: Crustacea formationis transitionis. Fasc. 2, 21-92, Lund.
aveline, w. t. and hughes, t. mck. 1872. The geology of the country around Kendal, Sedbergh, Bowness
and Tebay. Mem. geol. Surv. U.K. 20 pp.

— and strahan, A. 1888. The geology of the country around Kendal, Sedbergh and Tebay. Mem.
geol. Surv. U.K. 94 pp.

Bancroft, B. B. 1929. Some new species of Cryptolithus (s.l.) from the Upper Ordovician. Mem. Proc.
Manchr. lit. phil. Soc. 73, 67-98.

barrande, j. 1846. Notice preliminaire sur le Systeme Silurien et les Trilobites de Boheme. i vi, 97 pp.

— 1852. Systeme silurien du centre de la Boheme. Iere Partie. Recherches paleontologiques. i-xxx, 935 pp.
bassler, r. s. 1915. Bibliographic index of American Ordovician and Silurian fossils. Bull. U.S. natn. Mus.
92, 1521 pp.

boeck, c. 1838. Ubersicht der bisher in Norwegen gefundenen Formen der Trilobiten-Familie. In keilhau,
b. m. Gaea norwegica, Teil 1, 138-145. Christiania.

brongniart, a. 1822. In brongniart, a. and desmarest, a. G. Histoire naturelle des Crustaces fossiles;
les trilobites. 65 pp. Paris.

90 PALAEONTOLOGY, VOLUME 22

bruton, D. L. 1967. A revision of the Swedish Ordovician Odontopleuridae (Trilobita). Bull. Geol. Instn.
Univ. Uppsala, no. 43, 1-40.

— 1968a. Ordovician odontopleurid trilobites from Estonia and Latvia. Lethaia, 1, 288-302.

— 19686. A revision of the odontopleuridae (Trilobita) from the Palaeozoic of Bohemia. Skr. Norske
Vid.-Akad. Oslo, 1. Mat. -Nat. Kl. 25, 1-73.

dean, w. t. 1959. Stratigraphy of the Cross Fell Inlier. Proc. Yorks, geol. Soc. 32, 185-228.

— 1961. Trinucleid trilobites from the Higher Shales of the Caradoc Series in the Cross Fell Inlier. Ibid.
33, 119-134.

— 1962. The trilobites of the Caradoc Series in the Cross Fell Inlier of northern England. Bull. Brit. Mus.
Nat. Hist. (Geol.) 7, 65-134.

— 1963a. The Stile End Beds and Drygill Shales in the east and north of the English Lake District. Ibid.
9, 257-296.

— 19636. The Ordovician trilobite faunas of south Shropshire. III. Ibid. 7, 215-254.

— 1971. The trilobites of the Chair of Kildare Limestone (Upper Ordovician) in eastern Ireland.
Palaeontogr. Soc. ( Monogr .), 1, 1-60.

— 1974. Idem. Ibid. 2, 61-98.

eichwald, E. D. 1825. Geognostico-Zoologicae, per Ingriam Marisque Baltici Provincias, nee non de Tri-
lobitis observationes. 58 pp. Casani.

— 1860. Lethaea rossica un paleontologie de la Russie. lere vol. Ancienne periode. 1633 pp. Stuttgart.
haller, j. 1973. Die Ordovizische Trilobitengattung Chasmops aus baltoskandischen Gescheiben.

Palaeozoologie. A/IV/4, 723-768.

harkness, R. and nicholson, H. 1877. On the strata and their fossil contents between the Borrowdale
Series of the north of England and the Coniston Flags. Q. Jl geol. Soc. Lond. 33, 461-484.
hawle, i. and corda, a. j. c. 1847. Prodrom einer Monographic der bohmischen Trilobiten. 176 pp.
Prague.

hisinger, w. 1840. Lethaea Svecica seu Petrificata Sveciae, iconibus et characteribus illustrata. Suppl.
secundum. Holmiae.

hughes, c. p., ingham, J. K. and addison, r. 1975. The morphology, classification and evolution of the
Trinucleidae (Trilobita). Phil. Trans. R. Soc. B. 272, 537-604.
hughes, T. mck. 1867. On the break between the Upper and Lower Silurian rocks of the Lake District, as
seen between Kirkby Lonsdale and Malham, near Settle. Geol. Mag. 4, 346-356.
ingham, j. k. 1966. The Ordovician rocks in the Cautley and Dent districts of Westmorland and Yorkshire.
Proc. Yorks, geol. Soc. 35, 455-505.

— 1968. British and Swedish Ordovician species of Cybeloides (Trilobita). Scott. J. Geol. 4, 300-316.

— 1970. A monograph of the upper Ordovician trilobites from the Cautley and Dent districts of West-
morland and Yorkshire. Palaeontogr. Soc. (Monogr.), 1, 1-58.

— 1974. Idem. Ibid. 2, 59-87.

— 1977. Idem. Ibid. 3, 89-121.

— and wright, a. d. 1970. A revised classification of the Ashgill Series. Lethaia, 3, 233-242.

king, w. B. r. and wilcockson, w. H. 1934. The Lower Palaeozoic rocks of Austwick and Horton-in-
Ribbleside. Q. Jl geol. Soc. Lond. 90, 7-31.

— and williams, a. 1948. On the lower part of the Ashgillian Series in the north of England. Geol. Mag.
85, 205-212.

krueger, h. 1971. Encrinuriden aus ordovizischen Gescheiben (Teil 1). Geologie, Jg. 20, 10,1132-
1169.

kutorga, s. 1854. Einige Sphaerexochus und Cheirurus aus dem silurischen Kalksteinschichten des
Gouvernements von St. Petersbourg. Zap. imp. miner. Obschch. 13, 105-126.
lane, p. d. 1971. British Cheiruridae (Trilobita). Palaeontogr. Soc. (Monogr.), 95 pp.
marr, j. e. 1878. On some well defined life-zones in the lower part of the Silurian (Sedgwick) of the Lake
District. Q. Jl geol. Soc. Lond. 34, 871-885.

— 1892. The Coniston Limestone Series. Geol. Mag. (dec. 3), 9, 97-110.

— 1913. The Lower Palaeozoic rocks of the Cautley district. Q. Jl geol. Soc. Lond. 69, 1-17.

— 1916. The Ashgillian succession to the west Coniston Lake. Ibid. 71, 189-204.

mccoy, f. 1849. On the classification of some British fossil Crustacea, with notices of some new forms in
the University Collection at Cambridge. Ann. Mag. nat. Hist. (2), 4, 161-179, 330-335, 392-414.

McNAMARA: CONISTON TRILOBITES

91

mccoy, f. 1851. In sedgwick, a. and mccoy, f. A synopsis of the classification of the British Palaeozoic rocks,
with a systematic description of the British Palaeozoic fossils in the geological museum of the University of
Cambridge. Fasc. I, 1-184. London and Cambridge.

mcnamara, K. J. (in press). The age, stratigraphy and genesis of the Coniston Limestone Group in the
Southern Lake District. Geol. J. 14.

meek, f. b. 1873. Descriptions of invertebrate fossils of the Silurian and Devonian systems. Rep. geol. Surv.
Ohio, l,pt. 2, 1-243.

moore, R. c. (ed.). 1959. Treatise on Invertebrate Paleontology, Part O, Arthropoda 1 , xix. 560 pp. Geol.

Soc. Amer. and Univ. Kansas Press.
murchison, R. I. 1839. The Silurian System. London.

nicholson, h. a. 1868. An essay on the Geology of Cumberland and Westmorland. 93 pp.

— and marr, j. e. 1891. The Cross Fell Inlier. Q. Jl geol. Soc. Lond. 47, 500-529.
opik, A. 1937. Trilobiten aus Estland. Acta Comm. Univ. Tartuensis, A, 32, (3), 1-163.

Owens, R. M. 1973a. Ordovician Proetidae (Trilobita) from Scandinavia. Norsk, geol. tidsskr. 53, 117-181.

— 1973 b. British Ordovician and Silurian Proetidae (Trilobita). Palaeontogr. Soc. ( Monogr .), 98 pp.
pillet, J. 1954. La classification des Phacopacea (Trilobites). Bull. Soc. geol. Fr. (Ser. 6), 3, 817-839.
prantl, f. and pribyl, a. 1949. A study of the superfamily Odontopleuracea nov. superfam. (Trilobites).

Rozpr. ustr. Ust. geol Praha 12, 1-221.

price, d. 1973. The age and stratigraphy of the Sholeshook Limestone of south-west Wales. Geol. J. 8,
225-246.

— 1974. Trilobites from the Sholeshook Limestone (Ashgill) of South Wales. Palaeontology, 17, 841-868.
reed, f. r. c. 1894. Woodwardian Museum notes on Phacops ( Chasmops ) marri. Geol. Mag. (dec. 4), 1,

241-246.

1896. The fauna of the Keisley Limestone— Part 1. Q. Jl geol. Soc. Lond. 52, 407-437.

— 1901. Woodwardian Museum notes. Salter’s undescribed species III. Geol. Mag. (dec. 4), 8, 106-110.
1905. The classification of the Phacopidae. Ibid. 52, 1-11.

— 1906. The Lower Palaeozoic trilobites of the Girvan district, Ayrshire. Palaeontogr. Soc. (Monogr.),
3, 97-186.

1910. New fossils from the Dufton Shale. Geol. Mag. (dec. 5), 7, 211-220.

— 1914. The Lower Palaeozoic trilobites of Girvan. Supplement. Palaeontogr. Soc. (Monogr.), 56 pp.
richter, r. and richter, E. 1917. Palaeontologische Beobachtungen im Rheinischen Devon, I. Ueber

einzehne Arten von Acidaspis, Lichas, Cheirurus, etc. aus der Eifel. Jahrb. Nass. Ver.f. Naturkunde, 70,
143-161.

roomusoks, a. 1953. New data on the trilobite genus Chasmops from the Ordovician of the Estonian S.S.R.

Loodusuurijate Seltsi Juubelikoguteos, 396-413, Tallin. [In Russian.]
ross, R. 1967. Calymenid and other Ordovician trilobites from Kentucky and Ohio. U.S. geol. Surv. Prof.
Paper, 583B, 19 pp.

salter, j. w. 1848. In Phillips, J. and salter, j. w. Palaeontological Appendix to Professor John Phillips’
Memoir on the Malvern Hills, compared with the Palaeozoic districts of Abberley etc. Mem. Geol. Surv.
U.K. 2, (1), 331-386.

— 1864. A monograph of the British trilobites from the Cambrian, Silurian and Devonian formations.
Palaeontogr. Soc. (Monogr.), 1, 1-80.

— 1865. Idem. Ibid. 2, 81-128.

— 1867. Idem. Ibid. 4, 177-214.

— 1873. A catalogue of the collection of Cambrian and Silurian fossils contained in the Geological Museum
of the University of Cambridge. 204 pp. Cambridge.

schmidt, f. 1881. Revision der ostbaltischen silurischen Trilobiten. Abt. 1. Phacopiden, Cheiruriden und
Encrinuriden. Mem. I’Acad. Imp. Sci. St. Petersbourg (7), 30, 1-237.
sedgwick, a. 1845. On the comparative classification of the fossiliferous strata of North Wales, with the
corresponding deposits of Cumberland, Westmorland, and Lancashire. Q. Jl geol. Soc. Lond. 1, 442-450.

— 1846. On the classification of the fossiliferous slates of Cumberland, Westmorland and Lancashire.
Ibid. 2, 106-132.

Shirley, j. 1931. A redescription of the known British Ordovician species of Calymene (s.l.). Mem. Proc.
Manchester lit. Phil. Soc. 75, 1-33.

— 1933. A redescription of the known British Silurian species of Calymene (s.l.). Ibid. 77, 51-67.

92

PALAEONTOLOGY, VOLUME 22

Shirley, J. 1936. Some British trilobites of the Calymenidae. Q. Jl geol. Soc. Lond. 92, 384-422.
siveter, d. j. 1977. The middle Ordovician of the Oslo region, Norway, 27. Trilobites of the family Caly-
menidae. Norsk, geol. tiddskr. 56, 335-396.

sthrmer, l. 1941. Early descriptions of Norwegian trilobites. The type specimens of C. Boeck, M. Sars
and M. Esmark. Ibid. 20, 113-151.

Stubblefield, c. J. 1928. A new trilobite Acidaspis magnospina from the Coniston Limestone. Ann. Mag.
Nat. Hist. Lond. (10), 1, 427-433.

temple, j. t. 1952. A revision of the trilobite Dalmanitina mucronata (Brongniart) and related species.
Acta Univ. Lund. (N.F., 2), 48, 1-33.

— 1975a. Standardisation of trilobite orientation and measurement. Fossils and Strata, 4, 461-467.

— 19756. Early Llandovery trilobites from Wales with notes on British Llandovery Calymenids. Palae-
ontology, 18, 137-159.

tripp, r. 1957. The trilobite Encrinurus multisegmentatus (Portlock) and allied middle and upper Ordovician
species. Ibid. 1, 60-72.

warburg, e. 1925. The trilobites of the Leptaena Limestone in Dalarne. Bull. Geol. Instit. Uppsala, 17,
1-162.

— 1933. On the structure of the occipital ring of the Odontopleuridae. Arkiv.fdr Zool. 25A, no. 9, 1-19.
Whittington, H. B. 1941. The Trinucleidae— with special reference to North American genera and species.

J. Paleont. 15,21-41.

— 1950. British trilobites of the family Harpidae. Palaeontogr. Soc. ( Monogr .), 1-55.

— - 1956. Type and other species of Odontopleuridae (Trilobita). J. Paleont. 30, 504-520.

— 1968. A monograph of the Ordovician trilobites of the Bala area, Merioneth. Palaeontogr. Soc.
(Monogr.), 4, 93-138.

1971. Silurian calymenid trilobites. Palaeontology, 14, 455-477.

woods, H. 1891. Catalogue of fossils in the student’s stratigraphical series, Woodwardian Museum. 23 pp.

Typescript received 22 February 1978
Revised typescript received 5 June 1978

K. J. MCNAMARA

Department of Geology and Geophysics
University of Sydney
N.S.W. 2006
Australia

NEW ACROTRETACEAN BRACHIOPODS
FROM THE PALAEOZOIC OF
BRITAIN AND AUSTRIA

by L. R. M. COCKS

Abstract. New acrotretacean brachiopods are described; Caenotreta aldridgei gen. et sp. nov. from the Llandovery of
the Welsh Borderland, England, C. celloni sp. nov. from the Llandovery of the Carnic Alps, Austria and Caenotreta sp.
from the late Silurian of the Carnic Alps. There are also new records from the Devonian of Germany. The possible
ecology of these micromorphic conical forms is discussed.

During a study of Silurian conodonts (Aldridge 1972), in particular from the Welsh
Borderland, Dr. R. J. Aldridge recovered some small conical shells from his acid-
treated residues at several different localities and stratigraphical horizons. Subsequent
collections made by him from the Silurian of the Carnic Alps, Austria, and also the
Devonian of Germany, yielded similar shells which, upon examination, were found to
be inarticulate brachiopods of the superfamily Acrotretacea. Comparable small shells
have been described from the Ordovician, especially by Cooper (1956) from America
and Biernat (1973) from eastern Europe, and from the Devonian by Ludvigsen (1974);
but the only description of representatives of this group from the Silurian is in a short
paper by Ireland (1961). Ireland described two new monospecific genera, Artiotreta
and Acrotretella from the middle part of the Chimneyhill Formation of Oklahoma,
U.S.A. The Chimneyhill is of Llandovery age.

Since Ireland’s paper was written, the scanning electron microscope has become a
standard tool in the study of small fossils, and enables adequate pictures of the Silurian
members of the Acrotretacea to be published for the first time. In addition, in the
present paper a specimen is illustrated with both valves together, an unusual
occurrence at any age in this group.

SYSTEMATIC PALAEONTOLOGY

Order acrotretida Kuhn, 1949
Superfamily acrotretacea Schuchert, 1893

Discussion. Following the appearance of the Acrotretacea in the early Cambrian, the
superfamily diversified until its acme in the late Cambrian and early Ordovician.
Numbers became noticeably fewer by Caradoc time, and by the Silurian the
Acrotretacea appears to have been represented only by the minute forms described by
Ireland (1961), Ludvigsen (1974), and in this paper. In the Treatise (Rowell 1965,
p. H 276) the upper range of the superfamily was in doubt, the only information then
available being from the lower Silurian material in Ireland’s paper, and also some
material which Rowell himself had seen, from float ascribed to the lower Devonian
Kalkburg Limestone of New York State, in which only pedicle valves were found.

[Palaeontology, Vol. 22, Part 1, 1979, pp. 93-100, pis. 13-14.]

94

PALAEONTOLOGY, VOLUME 22

Later Ludvigsen (1974) described a new scaphelasmatine, Opsiconidion from the lower
Emsian of arctic Canada, and discussed other Devonian records, which go up to the
Frasnian. Some rather fragmentary specimens, not illustrated here but probably
belonging to the new genus Caenotreta described below, have been found by Dr.
Aldridge in conodont residues from the Ballersbacher Limestone (late Emsian or early
Eifelian) at its type locality near Ballersbach, and also from the Giinteroder Limestone
(Eifelian) at Blauer Bruch, Bad Wildungen, both localities in the Rhenish Slate
Mountains in Germany.

Family acrotretidae Schuchert, 1893
Subfamily torynelasmatinae Rowell, 1965
Genus caenotreta gen. nov.

Type species. Caenotreta aldridgei sp. nov. (see below).

Age range. Silurian (Llandovery Series) to Devonian (Eifelian Series).

Diagnosis. Acrotretid with simple conical pedicle valve, like Torynelasma, but whose brachial valve has
rods instead of a plate on its median septum.

Discussion. The only other genera attributed to this subfamily (Rowell 1965, p. H 279)
are Torynelasma Cooper, 1956 and doubtfully Acrotetrella Ireland, 1961. Torynelasma
is now widely known from various localities and horizons in the early and Middle
Ordovician; from the Pratt Ferry Formation, Alabama, U.S.A. (Cooper 1956, p. 257),
the Antelope Valley Limestone, Nevada, U.S.A. (Krause and Rowell 1975), the
Leningrad region, U.S.S.R. (Gorjansky 1969), Kazakhstan, U.S.S.R. (Popov 1975),
Estonia and Poland (Biernat 1973) and from unpublished material from the lower
Ordovician of Australia. As may be seen from the photographs of Cooper and the
superb drawings of Biernat, Torynelasma has a very similar external shape and
ornament to Caenotreta. The new genus also shows the distinctive protegular pattern
in both brachial (PI. 13, fig. 1) and pedicle (PI. 14, fig. 1) valves, consisting of the
honeycomb of pits described by Biernat and Williams (1970). However, Caenotreta
differs from Torynelasma in possessing a pair of rods on the brachial valve median
septum, instead of a plate. These rods are not unlike the early growth stages illustrated
for Myotreta by Biernat (1973, p. 44, fig. 17), though the mature Myotreta possesses a

EXPLANATION OF PLATE 13

All photographs taken with the Scanning Electron Microscope.

Figs. 1-8. Caenotreta aldridgei gen. et sp. nov. Figs. 1, 2. BB 75901, conjoined valves, from Minsterley
Formation (Llandovery, probably Fronian), Hope Brook, Minsterley, Salop. Grid Ref. SJ 360 022.
Fig. 1, enlargement of brachial valve protegulum and adjoining edge of pedicle valve, x 500. Fig. 2,
oblique general view, x 100. Figs. 3-5 from Wych Beds (Llandovery, Telychian), lane to Birches Farm,
Cowleigh Park, Malvern Hills, Hereford and Worcester. Grid Ref. SO 760 468. Fig. 3, BB 75911,
interior of brachial valve showing form and growth lines of the pseudointerarea, x 200. Figs. 4, 5, BB
75912, holotype, oblique, and ventral views of brachial valve, x 100 and x50 respectively. Figs. 6-8
from Purple Shale (Llandovery Telychian), stream bank south-west of Ticklerton, Salop. Grid Ref. SO
481 901. Fig. 6, BB 75936, oblique view of pedicle valve, x 60. Fig. 7, BB 75922, lateral view of pedicle
valve, x 50. Fig. 8, BB 75920, exterior of smaller brachial valve, x 200.

PLATE 13

COCKS, acrotretacean brachiopods

96

PALAEONTOLOGY, VOLUME 22

folded median septum (Biernat 1973, p. 84, fig. 30) quite different from Caenotreta or
Torynelasma. In addition Torynelasma possesses a pair of muscle scars each
surrounded by a prominent rim of shell material just anterior to either end of the
pseudointerarea in the brachial valve interior; these rims are absent from Caenotreta.
The only trace of muscle attachment areas on the floor of the brachial valve of
Caenotreta is an ill-defined area of very weak flabellate scars to be seen only in one or
two of the presumably more gerontic specimens (e.g. PL 13, fig. 4).

Acrotretella differs from Caenotreta in its median septum, which starts posteriorly as
a pair of prongs, which merge anteriorly to form a single blade without any kind of
plate, the whole structure resembling a tuning fork when viewed from above (Ireland
1961). In addition, the pedicle valve of Acrotretella is less sharply conical and the
growth lines more prominent, but these last two features by themselves would be
regarded more as specific than generic characters. Details of the pseudointerarea in
Acrotretella are not known.

Hansotreta also has a tall conical pedicle valve which externally resembles
Caenotreta and Torynelasma , but Hansotreta has a prominent apical process, leading
to its placement in the Acrotretinae (Krause and Rowell 1975).

Caenotreta aldridgei sp. nov.

Plate 13, figs. 1-7; Plate 14, figs. 1-4

Description

Pedicle valve. Acutely conical in shape, with posterior part of cone more rectilinear in profile than the more
sloping anterior part. The pseudointerarea is indicated only by a slight asymmetry in cross-section— there
are no well-defined edges to the pseudointerarea and no sign of any intertrough or interridge. Ornament
absent from main part of shell, apart from fine growth lines which continue evenly round the whole anti-
apical margin in mature specimens. Prominent protegulum with honeycomb of pits (PI. 14, fig. 1), some
overlapping in the same way as Opsiconidion (Ludvigsen 1974, fig. 4) ; open foramen at very apex of valve
(PI. 14, fig. 2); further enlargements of the foramen show the protegular pits of the protegular exterior
continue over the edge of the foramen into valve interior; pedicle tube absent. Valve interior smooth,
without visible structures of any kind (PI. 14, fig. 4).

Brachial valve. Subcircular in outline, flat to very slightly convex, apart from a slight suggestion of a very
shallow, even median sulcus (PI. 1 3, fig. 2). Exterior smooth, apart from fine growth lines. Protegular area
with similar ornament to pedicle valve, consisting of pits in a honeycomb arrangement (PI. 13, fig. 1).

EXPLANATION OF PLATE 14

All photographs taken with the Scanning Electron Microscope.

Figs. 1-4. Caenotreta aldridgei gen. et sp. nov. Figs. 1, 3, 4, from Purple Shale (Llandovery, Telychian),
streambank south-west of Ticklerton, Salop. Grid Ref. SO 481 901. Fig. 1, BB 75925, apex of pedicle
valve, showing protegulum, x 500 approx. Fig. 3, BB 75918, oblique view of brachial valve, x 150. Fig.
4, BB 75923, oblique view looking up into the interior of a pedicle valve, x 100. Fig. 2, BB 75913,
oblique view of the apex of a pedicle valve, showing the pedicle opening, x 500, from Venusbank
Formation (Llandovery, Idwian), Hope Quarry, near Minsterley, Salop. Grid Ref. SJ 355 021.

Fig. 5. Caenotreta sp., BB 75941, lateral, but slightly oblique, view of a pedicle valve, x 75, from Horizon
1 8 [ploeckensis Zone, Ludlow), Cellon section, Carnic Alps, Austria.

Figs. 6-8. Caenotreta celloni gen. et sp. nov. Figs. 6-8, BB 75942, holotype, a brachial valve, Fig. 6, ventral
view, x 50. Fig. 7, oblique view from anterior, showing shape of septum and its rods, x 75. Fig. 8, lateral
oblique view, x 60, from Horizon 10J ( celloni Zone, Llandovery, Telychian), Cellon section, Carnic
Alps, Austria.

PLATE 14

COCKS, acrotretacean brachiopods

PALAEONTOLOGY, VOLUME 22

Interior with prominent anacline pseudointerarea (PI. 13, fig. 3), with entire growth lines and no median
structures apart from a broad shallow groove. No trace of muscle scars, apart from a very slightly
flabellate area, with no well-defined boundaries, anterior to the pseudointerarea in just one or two
specimens (e.g. the holotype). Prominent median septum arising from valve floor just anterior to pseudo-
interarea, and incorporating two rods projecting antero-ventrally from the floor of the valve; an upper
rod running the whole length of the top of the septum, in the position corresponding to the plate of
Torynelasma; the lower rod diverging from the upper rod at about one-third valve length and continuing
anteriorly beneath the upper rod in the septum. Both rods project further ventro-anteriorly than the septum,
and end in rounded knobs (PI. 13, fig. 4). The edge of the valve is angled round the entire margin anterior
to the pseudointerarea, so as to fit snugly into the pedicle valve when the valves were together
(PI. 13, fig. 2).

Localities and material. Holotype BB 75912, a brachial valve (PI. 13, figs. 4, 5), one of six brachial valves
and two pedicle valves (including BB 75907 to BB 75912), from Wych Beds (Llandovery, Telychian), lane
leading to Birches Farm, Cowleigh Park, Malvern Hills, Hereford and Worcester. Grid Reference SO 760
468. Locality C.P. 1 of Aldridge (1972). Twenty-three pedicle valves and fourteen brachial valves
(including BB 75914 to BB 75940) from very top of Purple Shale (Llandovery, Telychian), stream bank
south-west of Ticklerton, Salop. Grid Ref. SO 481 901. Localities Ticklerton 1 and 3 of Aldridge (1972),
age later modified by Aldridge 1975, p. 612. Conjoined valves, three further pedicle valves and one further
brachial valve (including BB 75901 to BB 75906) from Minsterley Formation (Llandovery, probably
Fronian), bank of Hope Brook, Minsterley, Salop. Grid Ref. SJ 360 022. Locality H.V. 4 of Aldridge
(1972). Probably referred to the species is a single pedicle valve (BB 75913) from Venusbank Formation
(Idwian), Hope Quarry, south of Minsterley, Salop. Grid Ref. SJ 355 021. Locality H.Q. 3 of Aldridge
(1972). Thus the confirmed range of the species is in the Fronian and Telychian Stages of the Llandovery
Series, and probably extends from the underlying Idwian Stage.

Dimensions. Like the conodonts with which they are associated, measurement of these shells with a
conventional microscope and graticules is difficult, and the range of accuracy has been only an estimated
10%. However, measurement of topotype specimens are as follows: BB 75912, holotype, a brachial valve,
length 0-8 mm, width 0-9 mm; BB 75908, a pedicle valve, height 1-0 mm, width 0-8 mm; BB 75910, a
brachial valve, height of median septum and rods 0-3 mm. The whole series of populations studied have
comparable dimensions.

Caenotreta celloni sp. nov.

Plate 14, figs. 6-8

Description

Pedicle valve. There are no discernible differences between the pedicle valves of C. celloni and C. aldridgei
(for description see above).

Brachial valve. Although the valve outline is very similar to C. aldridgei, and may still be termed
subcircular, the outline of C. celloni is slightly more quadrangular, with the change in angle between lateral
margins and pseudointerarea more marked than in C. aldridgei. Radial ornament absent ; growth lines
slightly coarser than in C. aldridgei, but protegulum of similar size and ornament. Internally the
pseudointerarea is comparable with C. aldridgei, but the septal structure is different. In C. celloni the
median septum rises from the valve floor just anterior to the pseudointerarea, but there is a pair of upper
rods on the top of the septum (PI. 14, fig. 7), as opposed to the single rod of C. aldridgei. Although this pair
of rods coalesces, there is a clear median furrow between the rods, which precludes confusion with the
entire plate of Torynelasma. There is also a lower rod, in a similar style and position to that of C. aldridgei.
There is no trace of any muscle scars on the valve floor.

Localities and material. Holotype BB 75942, a brachial valve (PI. 14, figs. 6-8), one of nine brachial valves
and five pedicle valves (including BB 75942 to BB 75946), from Horizon 10J of Walliser (1964) ( celloni
Zone, Llandovery, Telychian), Cellon section, Carnic Alps, Austria. Eleven brachial valves and six pedicle
valves (including BB 75947 to BB 75951) from Horizon 11D of Walliser (1964) (lower part of
amorphognathoides Zone, Llandovery, Telychian), Cellon section, Carnic Alps, Austria.

COCKS: ACROTRETACEAN BRACHIOPODS

99

Discussion. Caenotreta aldridgei and C. celloni are contemporary species of the late
Llandovery. The chief substantial difference between the two species is in the single rod
at the top of the median septum in C. aldridgei as opposed to the fused pair of rods seen
in C. celloni , and this difference was observed in all of the brachial valves of the
populations concerned.

Caenotreta sp.

Plate 14, fig. 5

Localities and material. From above Horizons 10J and 11D, which yielded C. celloni, Dr. Aldridge has
collected material of Caenotreta from the following horizons in the Cellon section of the Carnic Alps,
Austria. The terminology of the horizons and their conodont zones is from Walliser (1964), and their
correlation from Aldridge (1975). Horizons 12 and 12A, upper part of amorphognathoides Zone
(Wenlock, Sheinwoodian) ; Horizon 14D, upper sagitta Zone (Wenlock, Homerian); Horizon 16, upper
crassa Zone (Ludlow, Eltonian); Horizons 16A, 17, 18 (including BB 75941, PI. 14, fig. 5), ploeckensis
Zone (Ludlow, Eltonian or slightly later); Horizon 27, latialatus Zone (Ludlow); Horizon 31 A, crispus
Zone (Ludlow) ; Horizon 39, eosteinhornensis Zone, just below the base of the Megaera-Schichten (late
Ludlow or early Pridoli).

Discussion. Although there are not enough specimens from any one of these horizons
for it to be safe to found a new species, the material listed above is sufficiently
diagnostic both to extend the range of the new genus without doubt into the late
Silurian of the Carnic Alps, and also to suggest that there exists in this later period a
third and different species of Caenotreta , presumably descended from C. aldridgei or C.
celloni. This species of Wenlock and Ludlow age has a pedicle valve similar to that of
the two described species, but the diagnostic rods on the brachial valve median septum
are relatively smaller, reduced in some cases to a mere swelling at the top of the septum.
In some specimens the rods even appear to be absent, but this may be due to
mechanical abrasion. However, in the youngest representatives of Caenotreta , from
the Giinteroder Limestone (Eifelian) of Blauer Bruch, Bad Wildungen, Germany, a
distinct upper rod is still present, although the lower rod is absent in both of the only
well-preserved brachial valves from the locality.

ECOLOGY

There are few known brachiopods whose adult form is as small as that of Caenotreta.
At an adult size of around 1 mm, a single sand grain would pose a serious threat in even
a medium-strength current. Modern bivalves of comparable size are known which live
in a variety of depths (Moore 1977) but these are all infaunal or semi-infaunal and thus
their ecology is not directly comparable with the epifaunal acrotretaceans. In view of
the shape of Caenotreta (PI. 13, fig. 2), the animal must have lived with the apex of its
pedicle valve down if it lived on the substrate. How deeply an adult would have settled
into the substrate is unknown, but probably at least half of the length of the pedicle
valve would need to be buried to afford reasonable vertical stability to the shell as a
whole. However, the water intake area between the two valves had to be kept as high
off the sea floor as possible. The matrices in which Caenotreta has been found vary in
size from mud (including lime-mud) to silt. The pedicle foramen remained open, and
thus one may presume a pedicle functional throughout life, but the relatively small size
of the opening suggests that the pedicle may have been used as a stabilizing device,

100

PALAEONTOLOGY, VOLUME 22

rather than as a permanent attachment mechanism. Such a minute shell living on the
sea floor would have been very threatened by even weak sedimentation. Caenotreta
may have thrived in areas kept relatively clean by weak currents, or it may have been
capable of using its small pedicle in a more subtle way so as to be able to push itself
upward if the sediment rain was too great, in a manner similar to that described on a
larger scale for some modern terebratulides by Richardson and Watson (1975). Rowell
and Krause (1973) have suggested that some other acrotretaceans may have lived as
epifauna, attached either to floating plankton or to fixed algae, and hanging
downwards from them. Attractive though this theory is, it seems unlikely that
Caenotreta lived in this attitude, since the pedicle opening is so small.

Acknowledgements. I am most grateful to Dr. R. J. Aldridge for passing these specimens on to me, to Dr.
M. G. Bassett and Dr. C. H. C. Brunton for reading the manuscript, and to P. C. Ensom for technical
assistance.

aldridge, R. J. 1972. Llandovery conodonts from the Welsh Borderland. Bull. Br. Mus. nat. Hist. (Geol.)
22, 125-231, pis. 1-9.

— 1975. The stratigraphic distribution of conodonts in the British Silurian. Jlgeol. Soc. Lond. 131, 607-
618, pis. 1-3.

biernat, g. 1973. Ordovician inarticulate brachiopods from Poland and Estonia. Pal. polon. 28, 1-120,
pis. 1-40.

— and williams, a. 1970. Ultrastructure of the protegulum of some acrotretide brachiopods.
Palaeontology, 13, 491-502, pis. 98-101.

cooper, G. a. 1956. Chazyan and related brachiopods. Smithson, misc. Colins. 127, 1-1245, pis. 1-269.
gorjansky, v. u. 1969. Inarticulate brachiopods of the Cambrian and Ordovician deposits of the north-
western Russian Platform. Min. Geol. RSFSR, Sev-Zap. Terr. Geol. Upravl. 6, 1-173 [in Russian].
Ireland, H. a. 1961. New phosphatic brachiopods from the Silurian of Oklahoma. J. Paleont. 35, 1 1 37—
42, pi. 137.

krause, F. F. and rowell, a. J. 1975. Distribution and systematics of the inarticulate brachiopods of the
Ordovician carbonate mud mound of Meiklejohn Peak, Nevada. Pal. Contr. Univ. Kansas, 61, 1-74, pis.
1-12.

ludvigsen, R. 1974. A new Devonian acrotretid (Brachiopoda, Inarticulata) with unique pro tegular
ultrastructure. N. Jb. Geol. Palaont. Monatsh, 1974, 133-148.
moore, D. R. 1977. Small species of Nuculidae (Bivalvia) from the tropical western Atlantic. Nautilus, 91,
119-128.

popov, L. y. 1975. Eocardine brachiopods from themiddle Ordovician of the Chingiz Ridge (Kazakhstan).
Pal. Zh. 32-41, pi. 5 [in Russian],

Richardson, j. r. and watson, J. e. 1975. Form and function in a Recent free living brachiopod.
Paleobiology, 1, 379-387.

rowell, a. J. 1965. Inarticulata. In williams, a. et al. Treatise on Invertebrate Paleontology . H.—
Brachiopoda. Geol. Soc. Amer. and Univ. Kansas Press, 927 pp.

— and krause, f. f. 1973. Habitat diversity in the Acrotretacea (Brachiopoda, Inarticulata). J. Paleont.
47, 791-800, pi. 1.

walliser, o. h. 1964. Conodonten des Silurs. Abh. hess. Landesamt. Bodenforsch. 41, 1-106, pis. 1.32.

REFERENCES

L. R. M. COCKS

Typescript received 17 February 1978
Revised typescript received 21 March 1978

Department of Palaeontology
British Museum (Natural History)
Cromwell Road
London SW7 5BD

THE PHYSIOLOGICAL DIFFERENCES BETWEEN
ARTICULATE BRACHIOPODS AND
FILTER-FEEDING BIVALVES AS A FACTOR IN
THE EVOLUTION OF
MARINE LEVEL-BOTTOM COMMUNITIES

by H. MIRIAM STEELE-PETROVIC

Abstract. Relevant physiological differences are discussed to determine their potential influence on the change in
dominance within marine level-bottom communities from articulate brachiopods in the Palaeozoic to filter-feeding
bivalves in the Mesozoic and Cenozoic. It is shown that (1) the two groups of animals have overlapped greatly in
requirements for food and space during the Phanerozoic, (2) the bivalves have considerably more energy per unit
biomass to spend on the production of offspring, (3) filter-feeding bivalves are potentially able to exploit a larger
number of habits and to cope more effectively with most environmental factors than are the articulates, and (4) that
these bivalves have considerably greater abilities to colonize and to expand their distribution than have the articulates.

Empirical evidence is cited for the importance of competition in Recent marine level-bottom communities in general,
and amongst Recent filter-feeding bivalves in these communities in particular. Although competition may occur only
occasionally, it seems to be important in shaping long-term structures of level-bottom communities. Competition on
the level bottom appears to have been considerably intensified and its effects greatly enhanced at certain times during
the Phanerozoic. It is suggested that partitioning of space contributed greatly to the ability of articulate brachiopods
and filter-feeding bivalves to share the near-shore region during the Palaeozoic. The fact that filter-feeding bivalves
became established close to shore early in the Palaeozoic is attributed to frequent unpredictable physical disruptions of
shallow- water communities. A lack of severe physical disturbances offshore probably enabled community structures to
be maintained there for considerable periods of geological time; when severe physical changes disrupted these
structures at the end of the Permian, bivalves invaded and replaced the articulates as the dominant off-shore
invertebrates. The decline of articulate brachiopods and increase in importance of filter-feeding bivalves occurred in
a series of steps. It is suggested that because of physiological differences between the two groups, filter-feeding
bivalves suffered less than articulates at both the Permian-Triassic and Triassic-Jurassic extinctions; that after each
extinction filter-feeding bivalves were able to invade numerous vacant or partially vacant habitats earlier and faster
than the articulates ; and that competition, particularly from the established bivalves, then prevented articulates from
reoccupying many of the habitats that they had previously held. The Cenozoic decline of articulate brachiopods
resulted from a loss of their preferred habitats (Ager in litt. 1977).

Articulate brachiopods were the dominant fossilizable invertebrates in most level-
bottom communities in the Palaeozoic (Bretsky 1969a), and filter-feeding bivalves
replaced the articulates and dominated these communities in the Mesozoic and
Cenozoic (e.g. Rudwick 1970, pp. 182-184). I became interested in the reasons for this
change in dominance when working on the problem of brachiopod feeding (Steele-
Petrovic 1976) and recognized that the two groups have utilized essentially the same
food during the Phanerozoic, but that compared with filter-feeding bivalves bra-
chiopods waste considerable energy in almost every physiological aspect of feeding.
These facts led me to consider the physiology of other processes, and the ways in which
the physiological differences between the two groups could have influenced their fossil
record since the Middle Ordovician.

In this paper I compare resource requirements of articulate brachiopods and filter-
feeding bivalves, and the ecological advantages of the two groups both in the adult and

[Palaeontology, Vol. 22, Part 1, 1979, pp. 101-134.]

102

PALAEONTOLOGY, VOLUME 22

larval stages. This comparison shows considerable overlap with bivalves superior,
suggesting that competition may have played a role in the changes in dominance. Since
competition in level-bottom communities has been disputed frequently (e.g. Johnson
1964; Stanley 1974a; Connell 1975), I discuss evidence on competition in these
communities, both in the Recent and in the past. I then consider in some detail the
changes in relative abundances of filter-feeding bivalves and articulate brachiopods
during the Phanerozoic, and how these changes, and therefore the major structural
changes in level-bottom communities, can be explained in terms of the physiological
differences that are discussed here.

The brachiopods that I consider are primarily the articulates, as the inarticulates
differ from the articulates in many ways, and have played a significant role only in a
relatively restricted range of marine communities since the early Ordovician. However,
in the first section, on feeding, I commonly discuss brachiopods in general, since all
brachiopods feed in essentially the same way (Steele-Petrovic 1976).

RESOURCE REQUIREMENTS OF BRACHIOPODS AND
FILTER-FEEDING BIVALVES

Space. Brachiopods and filter-feeding bivalves have numerous morphological and
physiological similarities : the shapes of their shells are commonly similar; both groups
have two valves secreted by a mantle that encloses the soft tissue ; all brachiopods and
most filter-feeding bivalves are suspension-feeders (for differences between ‘filter-
feeding’ and ‘suspension-feeding’, see Steele-Petrovic 1975); they obtain food from
water currents brought into the mantle cavity by beating of lateral cilia located on the
feeding organ, and in each case the water passes through the feeding organ, where
particles are trapped, and then out as an exhalant current ; digestion is very similar in
both groups (Steele-Petrovic 1976). It is understandable therefore that both groups
have had considerable overlap in habits and in habitat requirements during the
Phanerozoic. Both groups have had free-living, permanently attached, burrowing, and
swimming forms ; and both have occupied most marine habitats including intertidal,
level-bottom (both shallow and deep water), and reefal.

Food. The available evidence for brachiopods on stomach contents, and on the
anatomy, histology, physiology, and biochemistry of their feeding system indicates
that their food consists, in varying proportions, of dissolved substances, bacteria,
organic colloids, organic detritus, and algae (Steele-Petrovic 1976).

Table 1 summarizes information on the food of filter-feeding bivalves. Many
authors have claimed that bivalves must assimilate organic detritus (for reviews see
Yerwey 1952; Jorgensen 1966, pp. 260-263), but the value of such detritus has not been
sufficiently examined experimentally; as Jorgensen (1966, pp. 257-258) has indicated,
it is very difficult to separate dead organic material from living micro-organisms in the
water, and in most experiments that have been performed to show the importance of
organic detritus as food, the bacterial content has not been considered, so that positive
results could have been due to bacteria rather than to dead organic matter. However,
as in the case of brachiopods, there are several lines of indirect evidence that strongly
suggest that organic detritus can be utilized by filter-feeding bivalves. First, the
dominant enzymes of intracellular and extracellular digestion in bivalves are

STEELE-PETROVIC: BRACHIOPODS AND BIVALVES

103

table 1. Direct and indirect evidence of foods of filter-feeding bivalves.

Food

References

Comments

Dissolved
organic matter

Efford and Tsumura 1973; Khailov et
al. 1973; Bamford and Gingles 1974;
Bamford and McCrea 1975

Uptake may be through body surface.
For same reasons as with brachiopods
(Steele-Petrovic 1976), probably not
important source of food

Bacteria

Zobell and Feltham 1938; Newell 1965;
Jorgensen 1966, p. 259; Fenchel 1971 ;
Bernard 1974; Hyllebergand Gallucci
1975

Colloids

Verwey 1952

Evidence of removal from water. May
obtain from surface of clay particles as
McCammon (1969) suggested for
brachiopods

Organic detritus
Algae

This paper

Yonge 1926; Ballantine and Morton
1956; Dean 1958; Allen 1962; Goreau
et al. 1970; Fenchel 1971; Mathers
1972; Hylleberg and Gallucci 1975

See text

Minute animals

Nelson 1933; Mansour 1946; Hylleberg
and Gallucci 1975

Appears to be relatively unimportant in
most cases

carbohydrates (Sova et al. 1970; Kristensen 1972a; Wojtowicz 1972; Mathers 1973);
and organic detritus is relatively high in carbohydrates and low in proteins and lipids
(Agatova and Bogdanov 1972), while living material is relatively high in proteins and
low in carbohydrates and lipids (Parsons et al. 1961 ; Agatova and Bogdanov 1972).
Secondly, methylchitinase in the absence of chitinase, and methylcellulase in the
absence of cellulase are usually found in bivalve guts; but methylchitin and
methylcellulose only occur as products of the breakdown of chitin and cellulose, and
there is usually evidence to indicate that bivalve guts lack large quantities of bacteria
capable of hydrolysing chitin and cellulose (Kristensen 1972a). These facts suggest that
chitin and cellulose are hydrolysed by bacteria in detritus and that the broken-down
detritus is subsequently ingested, and at least part of it digested by filter-feeding
bivalves. Although suspension-feeders in general may ingest large numbers of larvae
(Woodin 1976), such larvae are probably not assimilated by most invertebrates (e.g.
Jorgensen 1966, p. 149). The apparent absence of extracellular proteases in brachio-
pods (Steele-Petrovic 1976) suggests an inability, by comparison with that of filter-
feeding bivalves, to assimilate minute animals. However, this form of food appears in
general to be relatively unimportant for these bivalves. In all other respects the food of
the two groups is essentially the same, and I see no reason to suggest that this has not
always been the case. Undoubtedly for both groups, as noted by Cowen (1971) for
brachiopods, the food of any particular individual depends on habitat, latitude, depth,
and season.

Discussion. Since articulate brachiopods and filter-feeding bivalves have utilized
essentially the same resources during the Phanerozoic, competition may have played a
significant role in the evolution of the two groups.

104

PALAEONTOLOGY, VOLUME 22

COMPARATIVE ECOLOGICAL ADVANTAGES OF THE ADULTS OF
ARTICULATE BRACHIOPODS AND FILTER-FEEDING BIVALVES

Feeding

When determining whether one group has had a competitive advantage over the
other, one factor to consider is the energy intake and the way in which this energy is
utilized. In evaluating potential competitive advantages one needs to consider the
following three energy factors, each of which is referred to unit biomass and unit time :

1 . Gross energy gain defined as energy obtained from ingested food. This quantity is a measure of feeding
effectiveness, i.e. success in obtaining energy by feeding.

2. Net energy gain defined as gross energy gain minus energy expended in feeding. This quantity is a
measure of the advantages gained from feeding.

3. Excess energy gain defined as net energy gain minus basal metabolic energy (the amount of energy
needed to maintain an animal in a state of rest), or gross energy gain minus unavoidable energy costs.
The group that obtains greatest excess energy gain per unit biomass and unit time must have a feeding
advantage. The excess energy can be channelled into body reserves, fast growth, physiological
processes requiring a high expenditure of energy (e.g. stronger feeding, burrowing, swimming), and
most significantly into the production of large numbers of offspring. Accordingly, excess energy gain is
the measure of competitive ability in feeding.

Feeding efficiency, defined as net energy gain over gross energy gain, is not as
important a factor in competitive ability as is net energy gain. In fact, net energy gain is
frequently increased at the expense of efficiency. A comparison can be drawn with
selling cars, where success depends on maximizing total net profit rather than
percentage profit.

Although no data are available on energy intake and utilization, similar steps in the
feeding processes of the two groups can be compared, and a qualitative estimate of the
relative energy gains at each stage can be made. Providing one group has a feeding
advantage at all stages, one can ascertain which animal has the greater total net energy
gain. The animal with greater excess energy gain can then be determined, and accord-
ingly the one with competitive superiority, provided that the basal metabolic needs
of the two groups are comparable or favour the group with greater net energy gain.

Pumping. The bivalve gill and brachiopod lophophore both act as pumps, and are
fundamentally the same morphologically and physiologically in that both are
composed of filaments with lateral cilia which beat to produce an inhalant current, and
with frontal cilia which transport the trapped material to a groove leading to the mouth
(see Rudwick 1970, pp. 117 et seq. and Steele-Petrovic 1975, 1976, for discussions of
lophophore; and Atkins 1936a, b, c, for discussions of gills). Gill filaments are strongly
attached, laterally to each other, and distally to the mantle or foot ; these attachments
are formed either by ciliary junctions or organic fusion. In contrast, lophophoral
filaments are never attached to each other, nor are their distal ends attached to any
anatomical structure and energy must be expended holding them in place against the
mantle surface or body wall, in order to separate inhalant and exhalant chambers
(Rudwick 1970, p. 118). Therefore, a bivalve gill can produce a greater pressure
difference and hence greater current velocity (Rudwick 1970, pp. 118-120), and can
have a greater pumping capacity, than can a brachiopod lophophore with similar
cumulative filamental length and similar energy consumption. Accordingly, where

STEELE-PETROVIC: BRACHIOPODS AND BIVALVES

105

trapping organs are of comparable size, the net energy gained in pumping is potentially
greater for filter-feeding bivalves than for articulate brachiopods. In addition, the gill
is a more compact structure than the lophophore, because of folding and fusion of
the gill sheets back to back, and in some cases plication. This structural modification
enables a gill that is considerably shorter than a lophophore to have a comparable
pumping capacity.

Trapping of particles. Because of open spaces between the filaments of the lophophore,
much of the material suspended in the feeding current of brachiopods is carried
between the filaments and into the exhalant chamber (Rudwick 19626; Steele-Petrovic

1975) . Only those particles are trapped that make contact with the frontal surfaces of
the filaments (Rudwick 1962 6; Bullivant 1968) or with their short lateral cilia
(Strathmann 1973). In comparison, latero-frontal cirri on adjacent filaments of the gill
of filter-feeding bivalves intermesh across the interfilamentary spaces (Dral 1967;
Moore 1971; Owen 1 974 ; Owen and McCrae 1 976) and trap all particles in the inhalant
current that are larger than the size of the mesh ; mesh size in Mytilus edulis is 0-6 pm by
either 2-4 pm or 4-8 pm, the latter depending on the relative positions of adjacent cirri
(Owen 1974). Although energy is consumed by the latero-frontal cirri of bivalves, the
amount must be small compared with that which is effectively wasted by the inability of
articulates to trap much of the incoming material. Therefore, where pumping organs
are of comparable size, the net energy gained at the trapping stage is considerably
greater for filter-feeding bivalves than for articulate brachiopods.

Transporting particles to the mouth. Particles that have been accepted as potential food
by filter-feeding bivalves are bound in mucus on the frontal surfaces of the filaments
(Atkins 1936a; Jorgensen 1966, p. 83). In contrast, brachiopods appear to locally
reverse the beat of the lateral cilia to prevent particles from escaping while being
carried along the frontal cilia (Strathmann 1973), and trapped particles are bound in
mucus only after reaching the food groove (e.g. Steele-Petrovic 1976). The brachiopod
reversal of ciliary beat must consume very much more energy than the bivalve
production of mucus, and since reversal decreases the pumping capacity, it effectively
wastes considerable energy; in addition, this reversal in brachiopods is unlikely to
prevent the escape of motile protistans (cf. Bullivant 1968), or to be as successful as the
mucus of filter-feeding bivalves in retaining non-swimming forms. Therefore, when
trapping organs of comparable sizes are considered, the net energy gain with respect to
the ability to retain trapped particles is considerably greater for filter-feeding bivalves
than for articulate brachiopods.

The brachiopod lophophore lacks a sorting mechanism (e.g. Steele-Petrovic 1975,

1976) and handles all particles indiscriminately, regardless of potential food value. In
contrast, both the gills (e.g. Atkins 1936a, b , c ) and labial palps (e.g. Yonge 1926;
Purchon 1955; Jorgensen 1966, pp. 77-82; Hughes 1975) of filter-feeding bivalves
effectively, although imperfectly (Atkins 19366; Hughes 1975), sort the trapped
material according to specific gravity (e.g. Atkins 1936a), size (e.g. Verwey 1952;
Hughes 1975; Hylleberg and Gallucci 1975), and possibly to some extent according to
food value (Hughes 1975); the denser and larger particles and particle masses are
rejected as pseudofaeces (Atkins 1936a; Verwey 1952; Jorgensen 1966, pp. 75-81;
Hughes 1975 ; Hylleberg and Gallucci 1975), and small low density particles, which are

106

PALAEONTOLOGY, VOLUME 22

likely to have a greater food value, and can be more readily processed for intracellular
digestion (Jorgensen 1966, p. 85), eventually reach the mouth. When suspended
material is scarce, sorting by the gills (e.g. Atkins 1936a, b) and labial palps (e.g. Ansell
1961; cf. Jorgensen 1966, p. 79) greatly decreases and probably frequently ceases.
However, Atkins (1936c) found that when small quantities of fine carborundum were
experimentally dropped on to a gill, only some of the particles were accepted and
others were rejected as pseudofaeces. The fact that sorting may occur when only small
amounts of material touch the gill suggests that this process is not just a ‘costly’ method
of getting rid of excess trapped particles, but rather that it is advantageous even when
potential food and therefore potential energy are scarce. Accordingly, it can be
concluded that sorting must consume less energy than it ultimately saves.

Digestion. Morphological and histological evidence indicates that assimilation of food
by the digestive diverticula is the same in brachiopods and filter-feeding bivalves
(Steele-Petrovic 1976); therefore, in order to compare the relative advantages of the
methods of digestion in the two groups, it is necessary to consider only those processes
that precede assimilation. Although the digestive tracts of the two kinds of animals
differ in many ways, the effectiveness and energetics of parallel as well as similar
processes can be compared (see Owen 1953, 1955 and Reid 1965, for discussions of
digestion in filter-feeding bivalves; and Steele-Petrovic 1976, for discussion of
brachiopods).

Ingested particles are released from their binding mucus in filter-feeding bivalves by
the combined action of stomach pH (Yonge 1935) and rotation of the crystalline style
against the gastric shield (Reid 1965); in brachiopods this release must be due to pH
alone (Steele-Petrovic 1976), and must be a slower and less ‘costly’ process than in
bivalves. However, rotation of the crystalline style in filter-feeding bivalves (Morton
1952; Reid 1965; Kristensen 19726) has a counterpart in rotation of the pyloric
protostyle in brachiopods (e.g. see Steele-Petrovic 1976), and the energy consumed in
these processes is probably similar in both groups of animals.

Since different regions in the stomach of filter-feeding bivalves have different
functions (Reid 1965), several steps of the digestive process can occur simultaneously:
i.e. sorting of ingested material in the stomach, rejection of dense particles into the
intestinal groove (e.g. Reid 1965), and transport of very small particles to the digestive
diverticula by the beating of cilia in the diverticular ducts (Owen 1955) all occur at the
same time. Therefore, although filter-feeding bivalves do not feed continuously
(Morton 1973), within each feeding cycle food is processed in a continuous-flow system
where the regime can be adjusted according to the circumstances, and handling of
particles of low nutritional value is minimized. In contrast, the digestive tract of
brachiopods is morphologically simpler and comparatively undifferentiated, so that
no sorting takes place, only one phase of digestion occurs at a time, and ingested
particles are handled in batches (e.g. Steele-Petrovic 1976). Accordingly, ingestion
must cease, and then muscular contractions of the digestive diverticula force ingested
particles (regardless of size or potential food value) back and forth between the
stomach and digestive diverticula (Steele-Petrovic 1976); and the fraction of usable
particles in each batch of the handled material must decrease with time. Therefore,
under most conditions the digestive system of a feeding bivalve can process

STEELE-PETROVIC: BRACHIOPODS AND BIVALVES 107

considerably more food in a given time compared with that of a feeding brachiopod of
comparable size. The comparative effect of cyclical feeding (Morton 1973) is unknown
for there are no data on possible cyclicity in brachiopods (Steele-Petrovic 1976). Since
neither the digestive cells of brachiopods nor those of filter-feeding bivalves can
phagocytose particles that are more than a few microns in diameter (Owen 1955;
Steele-Petrovic 1976), the bivalve method of transporting only very small particles into
the digestive gland, must bring a comparatively larger percentage of potentially usable
particles into contact with the digestive cells. The apparent absence of extracellular
proteases and lipases in brachiopods (Steele-Petrovic 1976) and their presence,
although in small concentrations, in filter-feeding bivalves (Mansour-Bek 1945;
George 1952; Reid 1966, 1968; Reid and Dunnill 1969; Reid and Rauchert 1970, 1972;
Vaskovsky and Suppes 1972) is another indication of the greater effectiveness of the
digestive system in these bivalves.

Digestion in filter-feeding bivalves is not only considerably more effective than in
articulate brachiopods, but the net energy gained at this stage must also be very much
greater in bivalves for the following reasons. In most cases cilia perform a function far
more efficiently than can any conceivable muscular mechanism (Prosser and Brown
1961, p. 476); therefore, the ciliary movement of only very small particles to the
digestive cells of filter-feeding bivalves must consume just a small fraction of the energy
used by brachiopods in the muscular pumping of large quantities of material back and
forth between the stomach and digestive diverticula. As with sorting prior to ingestion,
it can be argued that sorting in the stomach of these bivalves consumes less energy than
it ultimately saves; in fact, considerable amounts of energy must be expended by
brachiopods in processing large quantities of ingested material that cannot be
digested; whereas the bivalves get rid of much of the unwanted matter by spending
comparatively less energy in sorting, both before and after ingestion. In addition, it is
hard to imagine that the production of extracellular proteases and lipases is a net
energy drain on these bivalves.

Excess energy gain. The above discussion shows that although both filter-feeding
bivalves and articulate brachiopods expend considerable amounts of energy in feeding,
the energy is in general profitably used by bivalves, whereas much of it is effectively
wasted by brachiopods. As a result, filter-feeding bivalves have a considerably greater
net energy gain in feeding. It follows that the basic metabolic energy requirements per
unit biomass and unit time would have to be much greater for the bivalves than for the
articulates in order for the excess energy gain per unit biomass and unit time to be
similar in the two groups.

The scanty evidence available suggests that the normal rate of consumption of
energy per unit biomass is greater in filter-feeding bivalves than in articulate
brachiopods (Hammen 1977); this is a predictable situation, considering the fact that
bivalves have more ‘costly’ methods of reproduction (discussed below) and are
generally more active than the articulates. There are no comparative data on basal
metabolic rates per unit biomass. However, much of the energy consumed in a resting
state must be for oxygen consumption ; oxygen is obtained from the inhalant water
current which is produced in the same manner by both groups (see above), both types
of animals are relatively simple, and many organs and tissues in the two groups are

PALAEONTOLOGY, VOLUME 22

similar (Steele-Petrovic 1976) and probably have comparable oxygen requirements.
Therefore, it can be argued that as a first approximation the basic metabolic energy re-
quirements per unit biomass and unit time should not differ greatly in the two groups.

These arguments indicate that the excess energy gain per unit biomass and unit time
should be considerably greater for filter-feeding bivalves than for articulate brachio-
pods. Therefore, in a competitive situation the bivalves should have an advantage
over the articulates.

Coping with different environmental factors

Another factor to consider when determining if either filter-feeding bivalves or
articulate brachiopods have a potential competitive advantage over the other is their
relative abilities to cope with different environmental factors. If one group exploits
more effectively a large number of environments, one would expect that group to have
a competitive advantage in many situations. If the same group of animals is superior in
both feeding and in coping with different environmental factors, that group should be
considerably superior to the other in most competitive circumstances.

Morphological plasticity and exploitation of different habits. Articulates have adapted
to different conditions only by changes in shape of the shell, size and form of the
pedicle, and configuration of the lophophore. The morphology of their ‘soft parts’ and
their physiological processes have probably remained the same, in essential features,
throughout the Phanerozoic (Steele-Petrovic 1976). Although some articulates have
attached to soft organic materials such as sponges, tunicates, algae etc. (Rudwick 1961,
1965, 1970, p. 77; Ager 1967a; Foster 1974, p. 23) and floating seaweed (Ager 1962),
have been cemented (Rudwick 1965, 1970, p. 85; Ager 1967a), have lived free on the
bottom (Rudwick 1965, 1970, pp. 87-90; Ager 1967a; Bowen 1968) particularly during
the Upper Palaeozoic, or have lived partially buried in mobile sands (Richardson and
Watson 1975a, b ), these modes of life are not typical, and articulates appear to be best
suited for pedical attachment to hard substrates. In contrast, filter-feeding bivalves
have a greater potential range of typical adaptations, and they have differed greatly,
both in size and shape, and in ‘soft-part’ morphology and physiological processes.
They have developed numerous burrowing, byssally attached, free-living, boring, and
cemented forms which have successfully adopted infaunal, semi-infaunal, or epifaunal
habits in or on soft and hard bottoms (e.g. Stanley 1968, 1970, 1972). Unlike articulates,
some of these bivalves are deposit-feeders (Yonge 1949; Pohlo 1969).

The slight structural differences between the bivalve gill and brachiopod lophophore
have contributed to the abilities of filter-feeding bivalves to successfully exploit
different habits more fully than articulate brachiopods could. If the shape of an animal
does not change with size, surface area of a trapping organ is proportional to
(volume)2^3 of an animal. Hence there is a maximum length of trapping organ beyond
which it is too large to be contained within the mantle of the animal. Since the gill is
more compact than the lophophore, and for a given length has a greater pumping
capacity, it can support a greater biomass than can a lophophore of comparable
length. Therefore, filter-feeding bivalves have the potential for attaining larger sizes
than do brachiopods ; and the greater excess energy gain of the gill over the lophophore
further enhances this potential. The fact that filter-feeding bivalves have the potential

STEELE-PETROVIC: BRACHIOPODS AND BIVALVES

109

to grow to be larger than brachiopods, and therefore can span a greater size range, has
almost certainly contributed to the abilities of these bivalves to successfully exploit
different habits and habitats more fully than articulate brachiopods.

In addition, a compact gill leaves room inside the valves for an even larger gill in
relation to the biomass, or for other anatomical features such as the foot. The
brachiopod lophophore may occupy seven-eighths of the mantle cavity (Reynolds and
McCammon 1977) and appears to leave little space for other structures. Since the
bivalve foot varies greatly from a muscular burrowing organ to a small reduced
structure that produces byssal threads, it has enabled bivalves to assume a number of
adaptive roles that cannot be assumed by brachiopods. The greater excess energy gain
of filter-feeding bivalves over brachiopods must give these bivalves relatively more
energy to expend on such functions as burrowing or swimming, reproduction, and
stronger pumping. The development of the eulamellibranch gill with its large pumping
capacity contributed to the Mesozoic exploitation of the deep infaunal siphonate habit
by filter-feeding bivalves (Stanley 1968). In contrast, a brachiopod lophophore,
because of its basic design, is limited to changes in length, and could never be modified
to have sufficient pumping capacity for siphons to be functional; as a result,
brachiopods have remained epifaunal except for the lingulids, whose valves open at the
surface of the sediment during feeding.

The importance of the eulamellibranch gill in deep burrowing forms has generally
not been emphasized along with that of mantle fusion (Stanley 1968, 1972). A shallow
pallial sinus was present in the Ordovician genus Lyrodesma (Newell and LaRocque
1969); it also occurred in several genera of anomalodesmatans up to 100 million years
before relatively deep-burrowing forms of that Subclass appeared in the
Carboniferous (Runnegar 1974). Therefore, the exploitation of deep infaunal habits
should not have been prevented by the absence of mantle fusion and siphons. Another
consideration is that deep-burrowing siphonate bivalves require eulamellibranch gills
to produce a current with sufficient velocity to overcome the friction of a long siphon. If
the classification of Lyrodesma as a trigonid is correct, that genus must have had a
filibranch gill for one would not expect it to have had a more advanced gill than all
Recent representatives of the group. Palaeontological evidence does not permit one to
determine whether the early anomalodesmatans already had eulamellibranch gills
(Runnegar 1974). Therefore, it is conceivable that it was the lack of a eulamellibranch
gill that prevented early Palaeozoic filter-feeding bivalves from becoming deep
burrowers.

The above discussion indicates that the basic structure of filter-feeding bivalves
lends itself to greater morphological and physiological variability than does that of
brachiopods. (Schopf et al. (1975) argued that articulates are morphologically more
complex than bivalves, but these authors considered only shell morphology.) In the
course of the Phanerozoic, the greater inherent potential of filter-feeding bivalves for
evolutionary change has enabled them to become more specialized and to exploit a
greater number of habits, compared with brachiopods.

Eurytopy versus Stenotopy. Many filter-feeding bivalves are eurytopic as exemplified
by the fact that as a group they have successfully exploited the intertidal zone since the
Middle Ordovician (my own unpublished information). In contrast, articulate

10

PALAEONTOLOGY, VOLUME 22

brachiopods are comparatively stenotopic and only a relatively few species have ever
lived intertidally; most of these intertidal forms such as the Ordovician genus
Zygospira (Walker and Laporte 1970; my own unpublished information from the
Ottawa Valley, Canada), Devonian Howellella (Walker and Laporte 1970), and the
Recent species Waltonia inconspicua, Tegulorhynchia nigrans, and Pumilus antiquatus
(Percival 1944, 1960; Rudwick 1962a, b; Rickwood 1968) occupy (ed) tidal channels or
pools; Percival (1944) and Bowen (1968) reported W. inconspicua and Thayer (1975,
1977) reported Terebratalia transversa from above low tide mark, but such occurrences
are not nearly as common for articulate brachiopods as for filter-feeding bivalves.
Since the salinity tolerance of articulates is greater than was previously thought, it is
uncertain what physiological characteristics prevent(ed) this group from generally
living intertidally (Thayer 1974), although an inability to withstand dessication may be
a factor (cf. Thayer 1975).

Turbulence. Since the attachment strengths of the pedicle and byssus appear to be
comparable (Thayer 1975), both groups of animals are able to tolerate similar energy
regimes, providing a firm area of attachment is available. However, shifting sediment
frequently occurs in turbulent water, and very few articulates can tolerate this
disturbance (but see Richardson and Watson 1975a, b)\ in contrast, infaunal filter-
feeding bivalves, particularly those that can burrow rapidly, are able to readily adjust
to shifts in the substrate.

Sedimentation rate. Articulate brachiopods are generally unable to tolerate a high rate
of sedimentation (Rudwick 1970, p. 159). On the other hand, many filter-feeding
bivalves can move as the position of the sediment-water interface changes.

Turbidity. As mentioned above, brachiopods trap only a fraction of material
suspended in the inhalant current, whereas filter-feeding bivalves trap virtually all
particles. Therefore, in turbid conditions it would be necessary for most filter-feeding
bivalves, but not brachiopods, to spend a great deal of energy sorting and transporting
the trapped material, and beyond a certain turbidity the sorting and transport
mechanisms would probably become clogged and inoperable.

The few bivalve species that live where the water is turbid are secondarily adapted
for handling large amounts of suspended material. For example, in siphonate
suspension-feeders, straining tentacles surround the aperture of the inhalant siphon
(e.g. Yonge 1949; Ansell 1961;Pohlo 1972;Narchi 1972, 1975) and interdigitate across
the opening to keep out suspended material; in at least one family, the Veneracea, the
tentacles are particularly well developed only in those species living in turbid
environments (Pohlo 1 972). Various modifications in the margin of the mantle (Nelson
1938), and different kinds of membranes at the base of the siphons (Dodgson 1928;
Yonge 1949; Ansell 1961) are found in many bivalves living in turbid environments.
These features enable such bivalves to control the flow of water through the mantle
cavity and to direct much of the suspended material away from the gills. Deposit-
feeding tellinaceans, which are modified to suck in large amounts of bottom material,
lack straining tentacles on the end of the siphon (Pohlo 1972); but a pair of mantle
folds at the base of the siphon collects pseudofaeces and prevents them from being
washed forward by the inhalant current (Yonge 1949; Pohlo 1972); these deposit-
feeders also have smaller gills and larger palps than the suspension-feeders, so that

STEELE-PETROVIC: BRACHIOPODS AND BIVALVES 111

most of the trapped material is passed to the palps (Atkins 19366, c), which appear to
reject excess mucus-bound material rather than sorting individual particles according
to size and weight as they do in other filter-feeding bivalves (Reid and Reid 1969).
Therefore, certain filter-feeding bivalves have greatly overcome the basic difficulty that
the bivalve gill has in handling large amounts of suspended material ; and some of these
bivalves may be able to cope with turbid environments as successfully as brachiopods;
however, this tolerance for suspended sediment has been achieved only through
secondary adaptations.

Food requirements. Compared with filter-feeding bivalves, articulate brachiopods
appear to have lower rates of oxygen consumption (Hammen 1977), cannot grow as
large (discussed above), and more commonly brood their larvae (see below). A large
part of energy output in marine invertebrates goes into producing offspring (Vance
1973), and production of brooded larvae requires considerably less energy than
production of planktotrophic larvae (Mileikovsky 1971; Menge 1975). Therefore,
these observations indicate that articulate brachiopods should need less food than
filter-feeding bivalves.

Indirect evidence of the need for less food by articulates is that they frequently live in
deep water (e.g. McCammon 1969, 1973; Foster 1974) where food is relatively scarce
(Jorgensen 1966, pp. 273 et seq. ; Raymont 1971) and where filter-feeding bivalves are
often small or rare (e.g. Foster 1974, p. 23). There is even evidence to suggest that the
Mesozoic articulate Pygope lived in the absence of other suspension-feeders where
food was scarce (Ager 1965, 19676; Vogel 1966). The present occurrence of
brachiopods and absence of filter-feeding bivalves in cryptic habitats in reefs (Jackson
et al. 1971; Logan 1975) may also be related to a shortage of food.

Coping with predation

Stanley (19746) argued that smaller size, thinner shells, epifaunal habit, and lack of
mobility are features of brachiopods that in general make them more susceptible to
predation than bivalves. However, the fact that an animal can be more easily killed
does not ensure that it is preferentially attacked. In fact R. T. Paine (pers. comm.) has a
small amount of evidence suggesting that Recent predators generally take bivalves
before brachiopods. This fact is in accordance with the relative scarcity of brachiopods
today and their comparatively small biomass. However, brachiopods may have been
more susceptible to predation during the Palaeozoic when they were a common part of
the benthos and therefore more easily found.

COMPARATIVE ECOLOGICAL ADVANTAGES OF THE LARVAE OF
ARTICULATE BRACHIOPODS AND FILTER-FEEDING BIVALVES

Larvae of most articulate brachiopods have a relatively short free-swimming period of
usually a few hours or at most a few days (Ager 1967a; Rudwick 1970, p. 155), and in
many species the larvae are brooded within the shell of the mother (e.g. Percival 1944,
1960; Atkins 1960; Rickwood 1968; Rudwick 1970, p. 153; Webb et al. 1976). On the
other hand, most filter-feeding bivalves, particularly those that live in tropical and
temperate shelf zones, have planktotrophic larvae (Mileikovsky 1971), although

112

PALAEONTOLOGY, VOLUME 22

lecithotrophic larvae appear to be important in bivalves inhabiting high latitudes and
great depths (Mileikovsky 1971). This free-swimming larval stage appears to last for up
to five weeks in most bivalves (Muus 1973).

A short planktonic larval stage and brooding within the parent favour a clumped
distribution of a species; clumping can lead to local extinction in the presence of a
minor adversity, or to total extinction at the disappearance of a major habitat.
Conversely, a relatively long planktonic period favours a wider distribution of a
species, permits quicker recovery of populations that have been locally damaged, and
under favourable conditions enables quicker expansion, including colonization of new
habitats (Mileikovsky 1971).

Planktotrophic larvae must have functional digestive tracts long before settling; and
gill-palp feeding organs of bivalves are usually functioning two or three days after
settling has occurred (Bayne 1971). In contrast, new articulate spat have only
incomplete rudimentary guts (Percival 1944, 1960; Rickwood 1968), and the
lophophore does not begin to develop in Waltonia inconspicua until after the gut has
opened through the mouth which is not ‘for some time’ after settling (Percival 1944).
Also, bivalves, including bivalves that do not have planktotrophic larval stages, have
grown a shell by the time settling has occurred (Cox 1969). Studies by Percival (1944)
and Rickwood (1968) indicate that articulates do not begin to grow a shell until after
they have settled; Rickwood reported that the shell of Pumilus antiquatus did not
appear until three or four days after settling and that the spat were particularly
susceptible to predation by ciliates and polychaetes before that time. Rickwood’s
report implies that resistance to predation increases measurably once articulates
produce a shell. Therefore, although newly settled spat of bivalves are also exceedingly
vulnerable (Muus 1973), their weak shells may be protection from the tiniest predators
(e.g. ciliates and very small polychaetes) that effectively attack the newly settled
articulate spat of the only species for which evidence of this type is available. Since
vulnerability to predation generally decreases with size it is advantageous for an animal
to grow as quickly as possible. However, there are insufficient comparative data on the
early growth of articulates and filter-feeding bivalves to determine if either group has
an advantage at this stage of development.

The comparative ecological advantages of articulate brachiopods and filter-feeding
bivalves are summarized in Table 2.

EVIDENCE OF COMPETITION IN RECENT MARINE
LEVEL-BOTTOM COMMUNITIES

During the Phanerozoic, filter-feeding bivalves replaced articulate brachiopods as the
dominant fossilizable invertebrates in most marine level-bottom communities (see
below). I have already shown that these two groups overlapped considerably in
requirements for food and space throughout this time, and that in general the bivalves
have a potential for gaining more energy in feeding, exploiting a larger number of
habits, and coping more effectively with environmental factors than the brachiopods.
Such facts raise the possibility that competition between articulate brachiopods and
filter-feeding bivalves was a significant factor in the evolution of marine level-bottom
communities. However, several authors have recently concluded that competition is

STEELE-PETROVIC: BRACHIOPODS AND BIVALVES

113

table 2. Summary of comparative ecological advantages of articulate brachiopods and filter-feeding

bivalves.

Advantages

Feeding

Coping with
environmental
factors

Larvae

Articulate brachiopods

Filter-feeding bivalves

Greater energy gain in :

Pumping

Trapping

Transporting particles to mouth
Digestion

Greater morphological and physiologi-
cal variability
Greater eurytopy

Tolerate higher rates of sedimentation
Tolerate greater turbulence where sedi-
ments unconsolidated

Generally tolerate greater turbidity (Some species secondarily adapted for
Probably require less food high turbidity)

??Susceptibility to predation??

Long planktonic stage

?Spat possibly more resistant to

predation

unimportant in modern communities of this type (e.g. Johnson 1964), and within
subsets of marine invertebrates (particularly bivalves) in such communities (e.g.
Stanley 1974a). This assessment of competition amongst bivalves as unimportant is
particularly relevant to the present study, for by similar reasoning it can be argued that
interspecific competition must be unimportant amongst co-existing brachiopods, and
between co-occurring species of animals that are as similar as articulate brachiopods
and filter-feeding bivalves. Also, by extrapolation one might contend that competition
was unimportant in level-bottom communities and amongst articulate brachiopods
and filter-feeding bivalves at all times in the past. Before trying to assess whether
competition between these bivalves and articulates, and between these two groups of
animals and other associated invertebrates, has been an important factor in the
evolution of marine level-bottom communities, one should be aware of the cases for
and against competition on the level bottom today.

‘Competition’ is used here in the sense of MacArthur (1972, p. 21): i.e. ‘two species
are competing if an increase in either one harms the other . . . provided the effect is
reciprocal’. Therefore, to obtain proof of competition one has to follow changes in
numbers of individuals of co-existing species, and to show that these changes are the
result of an injurious effect of one species on another, a procedure that is particularly
difficult on the level bottom where many animals live infaunally. Nevertheless, there is
a small amount of direct and indirect evidence of competition, both in level-bottom
communities in general, and amongst filter-feeding bivalves of such communities in
particular.

Empirical evidence of competition

Competition in Recent level-bottom communities. Changes in fauna and often in bottom
sediment accompanied the disappearance of eel grass along the Atlantic coast of North
America in the 1930s. Johnson (1964) concluded, on the basis of evidence from

114

PALAEONTOLOGY, VOLUME 22

localities where the substrate had altered, that except for the disappearance of those
few species that lived directly upon or amongst the eel grass, most faunal changes
resulted indirectly from changes in bottom deposits. On the basis of this study, he
argued that benthic communities are associations of largely independent species ; this
conclusion appears to be widely accepted by palaeontologists. However, Stauffer
(1937) documented changes in a muddy lagoon on the Massachusetts coast, where on a
first approximation no gross change in sediment appears to have occurred. He found
that about 40% of the original fifty-fo.ur common species were no longer present at this
locality after the loss of eel grass (Table 3). In addition, changes appear to have
occurred in the relative abundances of some species that did not disappear. Although
Stauffer’s study was not sufficiently detailed to show such changes except where they
were markedly pronounced, it is evident that at least two of the four burrowing species
that were most common following the disappearance of eel grass had considerably
increased their original numbers. This change in relative abundances appears to have
greatly altered the dominances of the infauna, and therefore changed the character of
the community. A plausible explanation for these results is that prior to the
disappearance of eel grass, competitive interactions restricted the spread of certain
species, but that these species subsequently expanded into the vacancies left either by
the eel grass or by stronger animal competitors that had disappeared or decreased in
numbers with disappearance of the eel grass. Therefore, Stauffer’s study invalidates
Johnson’s conclusions, and appears to supply direct (albeit skimpy) evidence of the
importance of competition within a marine level-bottom community.

The importance of competition can also be inferred indirectly from certain
community structures. It is generally accepted that with time, community diversity
increases by invasion of species from outside; and that where resources are limited,
competition between successive invaders and established species forces all species to
become increasingly more specialized (e.g. Hutchinson 1959). The result is a
community in which the resources are partitioned amongst the different species.
Therefore niche partitioning in a community is indirect evidence that competition has
shaped the community. There are studies that show this type of resource sharing in
level-bottom communities, both in the Temperate Zone (Sanders 1960) and in the
Boreal Zone (Turpaeva 1948, 1949, 1953, 1954, 1957).

Sanders (1960) examined the Nepthys incisa-Nucula proximo community from soft-
bottom sediments at 19 m depths in Buzzards Bay, Massachusetts, and found that
only a few species out of a total of seventy-nine constantly dominate the community
when either numbers of specimens or biomass are considered. In order to establish how

table 3. Common species of eel grass community before
and after destruction of eel grass.

Habitats occupied by species Number of species

Live on eel grass

Before

7

After

1

Swim among eel grass

6

3

Live on mud surface

16

12

Burrow into mud

25

20

STEELE-PETROVIC: BRACHIOPODS AND BIVALVES

15

resources are utilized by the nine most abundant species (93% by number), he
determined the spatial distribution of each, and from their methods of feeding and gut
contents he determined the material that is ingested and from where on the bottom it is
obtained. He found that each species utilizes a different range of the available
resources. The same result is obtained for the eight most abundant species by biomass.

Turpaeva (1948, 1949, 1953, 1954, 1957) examined the feeding relationships
amongst the dominant benthic invertebrates from four distinct regions in the Barents
Sea. The animals were collected from depths of up to 300 m during five expeditions.
The methods of feeding of all species were determined. In order to establish where the
animals collected their food, composition of gut contents was analysed for each species
and was compared with the substrate where the species lived. This information was
supported or augmented by data from a large number of published papers on gut
contents and/or the anatomy and physiology of feeding and digestion in the same or
closely related species. On the basis of these studies Turpaeva recognized four feeding
zones and five feeding groups for benthic invertebrates : within the bottom sediment
(swallowers), on the surface of the bottom (collectors), the extreme bottom layer of
water (filterers A), and higher bottom layer of water (filterers B, and waiters, i.e.
animals that do not create their own currents). She noted that in general each of the
more abundant species by biomass in a community feeds in a different trophic zone,
and that where more than one species feeds in a single zone the biomass of the
dominant species there greatly exceeds the biomass of all the other species. In the
exceptional cases where two species in a single feeding zone have subequal biomasses,
Turpaeva noted that either food or space is partitioned between them. She reported
that other Soviet biologists working in the Barents Sea, White Sea, Sea of Azov, and
Caspian Sea have found these same feeding relationships amongst the benthic
invertebrates.

I have heard it argued that since food is usually abundant, Turpaeva’s feeding zones
have little ecological significance. Nevertheless, partitioning of the feeding space
appears to be a reality in the Boreal communities; and as Turpaeva (1948) noted, food
is strongly cyclical in these far northern waters. I have also heard it argued that
Turpaeva’s A and B filtering groups are ecologically unimportant since suspension-
feeders actively create their own currents and may therefore obtain particles from
different water levels. However, data of Reid and Reid (1969) show that sympatric
suspension-feeding species of Macoma have predominantly different kinds of diatoms
in their guts depending on whether their siphons project just above the bottom or a
little higher into the water column.

Competition among Recent filter-feeding bivalves. Bradley and Cook (1959) noted that
Mya arenaria usually lives abundantly in muddy areas, and Gemma gemma where it is
sandy. However, when the two species co-exist, they found that an average of 25%
fewer specimens of the small species of Gemma occur near the relatively larger Mya in
the direction of current flow, than in other directions. They concluded that Mya has a
deleterious effect on Gemma.

Eight species of Macoma may occur sympatrically along the coast of British
Columbia. Reid and Reid (1969) attempted to determine niche overlap amongst the
species by establishing for each species : orientation and activity of the siphon during

16

PALAEONTOLOGY, VOLUME 22

feeding, acceptance and transport of different kinds of particles to the mouth, and gut
contents at the time of collection. They found the animals to consist of three deposit-
feeders, four suspension-feeders, and one species that could feed in either way. Of the
three deposit-feeders it appears that M. secta lives primarily on bacteria that coat
ingested sand grains; gut contents of both M. calcar ea and M. lipara consist mainly of
small diatoms and flagellates, but M. calcarea can accept finer particles. Each of the
suspension-feeders extends its siphon to a different maximum height ; those that feed
closest to the bottom ( M . elimata at 1-5 cm, M. incongrua at 2 cm) have a large
percentage of diatom chains in their guts, in contrast to those that feed higher in the
water (M. inquinata at 2-5 cm, and M. nasuta at nearly 3 cm), which have mainly large
solitary diatoms ; in addition M. nasuta can accept and transport larger particles than
can M. inquinata. Although further work is needed on the problem, the available
information suggests that there is notable niche partitioning among the different
species of Macoma.

Theoretical considerations on competition

Competition in level-bottom communities. A contradiction appears to exist between
direct and indirect evidence on competition in Recent marine communities. On the one
hand, controlled studies usually fail to demonstrate that competition actively occurs
amongst animals in these communities (evidence summarized by Connell 1975) ; on the
other hand, niche partitioning with respect to both food and space, which is indirect
evidence of competition, appears to occur commonly (cf. Turpaeva 1948, 1949, 1953,
1954, 1957; Sanders 1960; Reid and Reid 1969).

Several authors have shown that intense predation in Recent invertebrate com-
munities prevents competition (e.g. Paine 1966, 1971, 1974; Connell 1975; Menge and
Sutherland 1976). Connell (1975) reviewed published evidence from controlled
experiments illustrating that grazers and predators on the middle and lower levels of
rocky shores in temperate zones usually eliminate their prey before these mature, with
the result that these intertidal communities are normally undersaturated, and
competition does not occur; and that harsh physical conditions in the upper part of the
intertidal zone frequently kill young and small individuals so that likewise, com-
petition is usually prevented. Nevertheless, he also showed that despite this predation,
large numbers of prey in certain widely spaced year-classes survive to maturity and
then persist for many years. As an explanation, Connell proposed that occasionally
(i.e. every few years) natural enemies are reduced or harsh physical conditions are
temporarily ameliorated so that the young of the dominant species survive to a stage
that is invulnerable either to predation or to the severe physical conditions; and he
reasoned that once the prey reaches this invulnerable size it competitively suppresses,
displaces, or excludes other colonists. Therefore, from evidence and arguments
presented by Connell, it appears that although competition can be seen to occur only
rarely in rocky-shore communities, its effects are long-lasting, and it is very important
in shaping the long-term structures of these communities.

Connell (1975) also argued that predation seems to be more intense where physical
conditions are less severe (see also Jackson 1972), and that as a result, competition
should occur in moderate environments even less frequently than on the rocky shore.
Certainly, there is more direct evidence for competition in intertidal and reefal

STEELE-PETROVIC: BRACHIOPODS AND BIVALVES 117

environments than on the level bottom; but this evidence may be a function of the
greater comparative difficulty in studying the ecology of level-bottom communities.
Nevertheless, the argument that predation is more intense subtidally than intertidally
is not the only factor to be considered in attempting to assess the prevalence of
competition in level-bottom communities. For instance, many more species of
invertebrates in general, not just of predators, inhabit the level bottom compared with
the rocky shore, and all of these species (including epifauna, infauna, deposit-feeders,
suspension-feeders, etc.) must be kept low in order for competition, either for food or
space, to be prevented. Furthermore, as predation reduces existing prey there may be
room for other species to move into the community (MacArthur 1972, p. 32).
Interestingly, Woodin (1974) argued that the competitive interactions she studied
amongst polychaetes in an intertidal mud flat should be more important subtidally,
where the same animals also occur, but where disturbances by physical factors are
considerably reduced. At present almost nothing is known about the complex
biological interactions or the relative abundance of resources and general degree of
saturation in level-bottom communities. Therefore, we do not know how often
competition occurs in marine level-bottom communities, although we can argue on the
basis of niche partitioning that it occurs sufficiently frequently to play a significant
structural role, and presumably a significant evolutionary role. In order to study
directly the effects of intermittent competition, it is necessary to monitor communities
for extensive periods of time; the conclusion that competition is unimportant is drawn
usually from studies that were conducted for only a couple of years.

Competition among filter-feeding bivalves. On the basis of considerable indirect
evidence Stanley (1974a) argued that competition among subsets of suspension-
feeding bivalves in marine level-bottom communities is generally weak and therefore
relatively unimportant, and notably less important than that among mammals. Van
Valen (1976) discussed extensively and disputed each of Stanley’s arguments, primarily
from the point of view of the theory of trophic energy in evolution. Van Valen showed
that bivalves compete as much as mammals, but that competition in bivalves is low
pressure, giving the appearance of no interaction at all ; and he noted that the sum of
many weak interactions can equal the sum of a few strong ones. Therefore, one cannot
conclude by extrapolation from the Recent into the past that competition has been
unimportant in the evolution of filter-feeding bivalves or (by further extrapolation) in
the change in dominance from articulate brachiopods to filter-feeding bivalves in
marine level-bottom communities.

Competition on the geological time-scale

Although intense competition may occur only rarely (on the human time-scale) in
marine level-bottom communities, it appears to occur sufficiently frequently to
strongly influence community structures. Therefore, it can be argued that competition
affects changes in these structures, and that its cumulative effects over geological time
are very important in community evolution.

Another point to consider in evaluating the importance of competition in the past is
that ecological conditions have fluctuated greatly during the Phanerozoic, and have
not always been the same as they are today. With every drop in sea-level, available

18

PALAEONTOLOGY, VOLUME 22

space on the shelf was reduced, resulting in smaller ecological niches and increased
species packing (Schopf 1974; Simberloff 1974), or in fewer species at the same level of
packing. Also, it has been argued on the basis of micropalaeontological evidence that
there have been great shortages of food in the past, particularly during periods of
massive extinctions (Tappan 1971 ; Tappan and Loeblich 1973). Therefore, one would
expect competition to have been considerably intensified, and its effects to have been
greatly enhanced, at certain times during the Phanerozoic. Furthermore, if predation
has increased in intensity since the Palaeozoic (Stanley 19746), competition may be
less important in general in shaping communities today than it was in the past.

The above discussion indicates how risky it may be to extrapolate into geological
time on the basis of what is observed in the present. Although some action may appear
to be insignificant when viewed at any particular instant, its cumulative effects over
geological time may be very important. Also, although natural laws remain the same,
conditions under which these laws operate may change so greatly that different times in
the past may bear little resemblance to the present.

RELATIVE ABUNDANCES OF ARTICULATE BRACHIOPODS AND
FILTER-FEEDING BIVALVES THROUGH THE PHANEROZOIC

Brachiopods were the dominant suspension-feeding animals before the Middle
Ordovician, when filter-feeding bivalves suddenly became important components of
near-shore faunas. Although filter-feeding bivalves at this time moved into certain
shallow-water environments that had been previously unexploited by articulate
brachiopods (unpublished data from the Ottawa Valley) they also formed large diverse
populations with these brachiopods, particularly in shallow subtidal habitats. Filter-
feeding bivalves continued to evolve at a steady rate and continually invaded new
adaptive zones (Stanley 1972). However, they failed to become established in deeper
water, and articulates continued to dominate the offshore environments (Bretsky
1969a). These relative abundances were retained in marine benthic communities until
close to the end of the Palaeozoic (Bretsky 1969a).

As late as the early part of the Upper Permian, off-shore faunas were still typically
dominated by articulate brachiopods (Grant and Cooper 1973; Pattison et al. 1973),
and local near-shore areas were dominated by filter-feeding bivalves (Pattison et al.
1973) or molluscs in general (Grant and Cooper 1973). The extensive marine regression
that occurred near the end of the Permian resulted in considerable reduction or even
elimination of the epicontinental seas, and a world-wide emergence of land (see papers
in Logan and Hills 1973). During the early stages of retreat in the Upper Permian the
relative abundance of molluscs, and of brachiopods that inhabited more shallow
environments, increased (Grant and Cooper 1973). In fact even prior to this time the
highly specialized articulates of the West Texas reefs were slowly being replaced by
more broadly adapted pedunculate forms (Grant 1971). As regression continued the
variety and numbers of articulates decreased and the abundance of filter-feeding
bivalves increased (Dagis and Ustritsky 1973 ; Grant and Cooper 1973 ; Nakazawa and
Runnegar 1973; Pattison et al. 1973); this increase in bivalve abundance was in some
cases accompanied by a similar increase in gastropods (Nakazawa and Runnegar
1973) or ammonites (Grant and Cooper 1973). As bivalves increased in abundance.

STEELE-PETROVIC: BRACHIOPODS AND BIVALVES 119

their diversity increased in some places (Kanmera and Nakazawa 1973) and decreased
in others (Pattison et al. 1973). Many families of articulate brachiopods became extinct
in the later Permian (Waterhouse and Bonham-Carter 1976), and many of those that
survived into the Triassic did so with only one genus (Dagis and Ustritsky 1973).
Certain Palaeozoic families of filter-feeding bivalves slowly disappeared during the
Upper Permian (Nakazawa and Runnegar 1973), and this disappearance was
accompanied by a correspondingly slow appearance of the ancestors of Mesozoic
bivalves (Stanley 1972; Kanmera and Nakazawa 1973; Nakazawa and Runnegar
1973). Of particular importance at this time was the expansion of the rapidly
burrowing anomalodesmatids and trigoniaceans (Stanley 1972). The turnover of
bivalve families at the Permian-Triassic boundary was very low, and except for the
Pterioida and Veneroida, widely spread orders of filter-feeding bivalves suffered
essentially no decline in their number of genera (Nakazawa and Runnegar 1973).

Ammonites and to a lesser extent bivalves (many of them pseudoplanktonic filter-
feeders) are the only invertebrates that are known to have occurred in any abundance
in early Triassic seas; lingulids are usually the only brachiopods recorded and they are
comparatively much less abundant than the ammonites and bivalves (Kummel 1973a,
b\ Newell 1973; Rudwick 1970, p. 182). Since the lowermost Triassic has a complete
spectrum of facies except for reefs, the absence of many invertebrate groups cannot be
due to a lack of suitable substrate (Kummel 1973a, b). Articulate brachiopods began to
increase in diversity and numbers in the Middle Triassic (Dagis and Ustritsky 1973),
but never regained their former dominance (Rudwick 1970, p. 183). Typical Mesozoic
genera of filter-feeding bivalves started to appear in the early Triassic and became
common and widely spread by the late Triassic (Nakazawa and Runnegar 1973).

Articulate brachiopods were decimated again by extinctions at the Triassic- Jurassic
boundary (Ager 1971), and by the Middle Jurassic the Atrypida, Spiriferida, and two
of the remaining three aberrant Strophomenida groups had become extinct, so that the
Terebratulida and Rhynchonellida were almost the only remaining articulates
(Rudwick 1970, p. 173). From the Jurassic onwards filter-feeding bivalves dominated
most macro-invertebrate assemblages in both numbers and diversity (Kauffman 1973).
In fact brachiopods were insignificant in North American Jurassic (Hallam 1975, p.
131) and Cretaceous (Reeside 1957) faunas. In the Jurassic of Europe, dense low-
diversity faunas, mainly of filter-feeding bivalves, dominated the littoral and very
shallow subtidal or lagoonal environments, essentially in the absence of brachiopods
(Ager 1965; Hallam 1975, pp. 75, 92; 1976). Although articulates were frequently a
significant component of shallow, normal marine environments in the European
Jurassic, Hallam (1975, pp. 46, 72, 73; 1976) considered that they were in general not
nearly as important as filter-feeding bivalves. However, Ager has found (in lift. 1977)
that throughout the Mesozoic, articulate brachiopods were dominant on rapidly
lithified shallow-water carbonate sea floors, while burrowing suspension-feeding
bivalves prospered on soft bottoms. These soft, shallow-water sediments of Mesozoic
age also supported the occasional oyster (Ager 1965, 1976). Hallam (1976) reported
that shallow marine basins that were poorly oxygenated in the Jurassic of Europe in
some cases supported a fauna consisting mainly of deposit-feeding nuculoids, while
basins with a higher oxygen content contained a more varied fauna in which
brachiopods were subordinate to the relatively more important filter- feeding bivalves.

.20

PALAEONTOLOGY, VOLUME 22

Ager (1965) noted that brachiopods were essentially absent from Mesozoic coral reefs
but that they often thrived in association with these reefs; he noted that the best
development of both Jurassic and Cretaceous brachiopods was in detrital fore-reef
sediments, in contrast to their absence in back-reef lagoonal sediments in the Mesozoic
in general. On the other hand, a diverse and often abundant fauna of filter-feeding
bivalves occurred in Jurassic reefs (Hallam 1976). Deep-water environments generally
lacked benthic invertebrates that left a fossil record, except for the deposit-feeding
forms that left trace fossils (Hallam 1971, 1975, p. 97), and for the articulate Pygope
which lived during the late Jurassic and early Cretaceous in fine-grained sediments
where food was probably scarce (Ager 1965, 1967 b).

Extinctions at the end of the Mesozoic only affected articulate brachiopods at the
generic level, so that the character of the group has changed little since the middle of the
Jurassic (Rudwick 1970, p. 173). However, since the end of the Mesozoic, articulates
have again decreased in importance, and in shallow-water environments have been
largely replaced by bivalves (Ager 1967a). This articulate decline coincided with a
marked Cenozoic decrease in rapidly lithified carbonates, and articulate brachiopods
are found in the Tertiary of Europe where there are limestones (Ager in lift. 1977).
Today articulates are a very insignificant part of the marine fauna (Ager 1967a).

An idealized summary of the relative dominances of the two groups is given in text-
fig. 1.

SHARING THE NEAR-SHORE REGION DURING THE PALAEOZOIC

Although filter-feeding bivalves and articulate brachiopods have overlapped con-
siderably in utilization of resources, the two groups continued to share the near-shore
region from the Middle Ordovician until close to the end of the Permian without the
physiologically superior bivalves ousting the brachiopods. An explanation is that
although the overlap is considerable when the whole of the Phanerozoic is considered,
partitioning generally occurred between these animals in the Palaeozoic. Professor
Ager wrote (in litt. 1977) that in his experience brachiopods and bivalves rarely
occurred together in abundance either in the Palaeozoic or in the Mesozoic. Since
brachiopods are basically considerably more tolerant of high turbidity than are filter-
feeding bivalves, shallow soft muddy bottoms were usually inhabited by free-living
articulates in the absence of filter-feeding bivalves (Steele-Petrovic 1975). In less turbid
Palaeozoic environments where both groups co-existed, although apparently rarely,
articulates typically lived epifaunally and attached by the pedicle, in most cases
probably to firm surfaces such as hard substrates, or on muddier bottoms to exposed
shells and other fragments (discussed above). In contrast, Palaeozoic filter-feeding
bivalves were endobyssate, epibyssate, free-burrowing, or free-living epifaunal forms
(Stanley 1972), and therefore in most cases must have inhabited slightly different
micro-habitats from the articulates. Division of food amongst the different sympatric
species might have occurred (e.g. Walker 1972), but supporting evidence is scarce.
Nevertheless, partitioning of space must have contributed to the ability of the two
groups of animals to share the near-shore region for so long a time.

STEELE-PETROVIC: BRACHIOPODS AND BIVALVES

12

text-fig. 1 . Idealized diagram of dominances of articulate brachiopods and
filter-feeding bivalves in level-bottom communities.

THE RESTRICTION OF FILTER-FEEDING BIVALVES TO NEAR-SHORE
COMMUNITIES UNTIL THE LATTER PART OF THE PERMIAN

Stanley (1972) suggested, but no longer believes (Stanley in litt. 1976), that because of
the stenotopic nature of established off-shore articulates, filter-feeding bivalves were
confined to near-shore regions in the Palaeozoic. If that had been the case these
bivalves would probably have been restricted to the intertidal zone rather than also
occurring abundantly in shallow, normal marine environments. Another explanation
is that deeper environments of the Palaeozoic inland seas were generally inhospitable
for filter-feeding bivalves, or at least more favourable for articulates than for these
bivalves. As mentioned above, brachiopods would probably have had an advantage
over primitive filter-feeding bivalves where turbidity was high. However, since off-
shore sediments during the Palaeozoic appear to have ranged from sand to mud
(Bretsky 1969a), turbid conditions could not have been the only factor that kept filter-
feeding bivalves close to shore. One can also discount the possibility that these bivalves

122

PALAEONTOLOGY, VOLUME 22

were physiologically unable to live in deeper water, for bivalves (such as the Ordovician
genus Lyrodesma, in Bretsky 19696) appeared off-shore at various times during the
Phanerozoic, although they failed to spread and become important numerically. There
is also the consideration that brachiopods compared with filter-feeding bivalves can
probably live where food is less plentiful (discussed above). However, benthic fauna
was abundant during most of the Palaeozoic, suggesting that scarcity of food was not a
problem then.

A more satisfactory explanation can be found in the Stability-Time hypothesis of
Sanders (1968). Ecologists generally agree that an outside species can invade an
undersaturated community more easily than it can a community at carrying capacity.
Hutchinson (1959) argued that invasion probably only succeeds where one or more
species are fluctuating and are under-represented at a given time. Since near-shore
communities are in a region of high environmental stress and are subject to continual
disturbance, their diversity generally remains low (Sanders 1968, 1969; Slobodkin and
Sanders 1969), and their species are subject to large fluctuations. Shallow subtidal
environments must also be subject to frequent (geologically speaking) unpredictable
physical disturbances that are severe enough to disrupt community equilibria and
produce faunal fluctuations. Therefore, once bivalves become successful marine
benthic invertebrates, possibly due to an adaptive breakthrough of the byssus as a
post-larval organ in the Ordovician (Stanley 1972, 1975), they were able to invade
successfully and to become prominent in fluctuating intertidal and shallow subtidal
communities. In contrast, physical disturbances, unless very severe, are considerably
dampened in deeper environments. As a result of long-term predictability, off-shore
communities evolve relatively high diversities (Sanders 1968; Dayton and Hessler
1972), which the Stability-Time hypothesis attributes to increased niche specialization
(Sanders 1968, 1969). Resulting community equilibrium in these environments is
probably rarely disturbed more than slightly by physical forces, and community
structures should be maintained for considerable periods of geological time.

Several authors (Dayton and Hessler 1972; Menge and Sutherland 1976) have
argued that high diversity in the deep sea today is explained better by predation than by
increased niche specialization as suggested by Sanders. The predation theory proposes
that prey populations are maintained at sufficiently low densities so that resources are
rarely limiting, thus permitting great overlaps in resource utilization (Dayton and
Hessler 1972). Although this theory may explain Recent deep-sea diversity (but see
Grassle and Sanders 1973), it does not seem to be an appropriate explanation for
conditions that were present in marine level-bottom communities during the
Palaeozoic; it can be argued that if off-shore Palaeozoic communities had been
undersaturated as predicted by the predation theory, filter-feeding bivalves almost
certainly would have invaded. In contrast, increased niche partitioning which gives rise
to saturated communities adequately explains the restriction of these bivalves close to
shore.

A single species of filter-feeding bivalve attempting to invade an off-shore
community in the Palaeozoic undoubtedly had essential requirements in common both
with a number of established species of articulate brachiopods and with other less
similar animals such as bryozoans, crinoids, annelids, and soft-bodied organisms that
have not been preserved. Although an attempted filter-feeding bivalve invader must

STEELE-PETROVIC: BRACHIOPODS AND BIVALVES

123

have faced competition from several different kinds of established animals, com-
petition was probably greatest from articulates because of their dominance and their
greater ecological overlap with these bivalves. Nevertheless, since the needs and living
processes of animals inhabiting a diverse community are highly inter-related,
fluctuations not only in articulate brachiopods but in a large number of diverse
established species, amounting to disruption of the community structure, might have
been necessary before filter-feeding bivalves were able to invade these ancient
communities. The palaeontological evidence appears to support this suggestion, for it
was not until after the Permo-Triassic crisis, when articulate brachiopods and many
other groups were decimated, that filter-feeding bivalves became established off-shore.

Although faunal changes occurred in off-shore communities throughout the
Palaeozoic, these changes, as recorded in the fossil record, amounted primarily to
replacement of one articulate brachiopod by another, and articulates continuously
dominated the off-shore scene (Bretsky 1969a). As mentioned above, brachiopods as a
group have a relatively low potential for morphological and physiological versatility,
and orders of articulates have differed mainly in shape of the shell and configuration of
the lophophore; they are also ecologically conservative, and have always been
epifaunal suspension-feeders and generally pedically attached, although free-living
forms were common during the Palaeozoic. Therefore, undersaturation, or local
extinction of a single articulate species might lead to its replacement in almost exactly
the same niche, by another, possibly phylogenetically distant, articulate. Such
fluctuations in brachiopods could possibly result from slight environmental changes,
which must have affected even off-shore communities many times during the
Palaeozoic.

Even when major changes occurred off-shore in the articulate faunas at the end of
the Devonian, filter-feeding bivalves did not move into this region ; this fact implies
that community structures were not intensely disrupted at that time, and that much of
the faunal change was the result of substitution of one species for another, rather than
wholesale extinctions and subsequent colonization of open habitats. On the basis of
this reasoning, it can be argued that off-shore communities, at least as indicated in the
fossil record, retained essentially the same ecological structures throughout the
Palaeozoic.

The above arguments may have general implications, particularly for the
Palaeozoic; i.e. new higher taxa of marine invertebrates which differed significantly
from pre-existing forms, and which became ecologically successful during periods of
relative geological stability when off-shore communities were well established, may in
general have achieved initial prominence in shallow environments where physical
disturbances are more intense; these new taxa, even if they could physiologically
tolerate off-shore conditions, may have been restricted to shallow waters until a time
when community structures off-shore were disrupted. However, there is evidence to
suggest that predation has increased in intensity since the Palaeozoic (Stanley 19746)
possibly causing a decrease in the density of prey populations off-shore (Dayton and
Hessler 1972), and hence increasing the ease with which invasions may occur.
Therefore, the above implications may not apply in most post-Palaeozoic situations.

124

PALAEONTOLOGY, VOLUME 22

INCREASE IN IMPORTANCE OF FILTER-FEEDING BIVALVES AND
DECLINE OF ARTICULATE BRACHIOPODS

Stanley (19746) attributed the decline of articulate brachiopods and the change in
dominance of articulates and filter-feeding bivalves to the inability of articulates to
cope with advanced Mesozoic predators, namely teleost fishes, crabs, and drilling
gastropods. However, the first appearances of fossilized members of these groups
occur after the Permo-Triassic and Triassic- Jurassic declines in brachiopod domin-
ance; teleost fishes first appear in the Upper Jurassic (Andrews et al. 1967), drilling
gastropods in the Upper Cretaceous (Sohl 1969), and crabs in level-bottom com-
munities in the Cretaceous (Glaessner 1 969). Therefore, one cannot convincingly argue
that predation by these animals caused either the extinctions or lack of re-expansions
of articulates at the Permian-Triassic and Triassic-Jurassic boundaries; it could have
been a factor only in the final decline of articulate brachiopods at the end of the
Mesozoic and during the Cenozoic. In addition, both articulates and filter-feeding
bivalves probably were subject to considerable predation during the Palaeozoic,
particularly from starfish and possibly also from nautiloids. There is fossil evidence to
show that predatory starfish have been present at least from the Upper Ordovician,
and that they have used the very effective method of external digestion at least from the
Middle Devonian (Spencer and Wright 1966). Although filter-feeding bivalves may
have been better protected from these predators than were articulates (Stanley 19746),
large changes in the relative dominances of the two groups of animals did not occur
between the Middle Ordovician and about the end of the Palaeozoic, thus giving no
evidence that efficient Palaeozoic predators reduced the abundance ratio of articulates
to filter-feeding bivalves. In fact, MacArthur (1972, p. 94) argued that, except on an
island, predators other than man are usually incapable of causing complete
extinctions.

The known relative abundances of articulate brachiopods and filter-feeding bivalves
in Mesozoic level-bottom communities can be explained in terms of physiological
differences between the two groups of animals as follows. There is evidence from the
fossil record to suggest that the Permo-Triassic extinctions were caused or contributed
to greatly by: (1) intolerable physical conditions which resulted as the sea withdrew
(see papers in Logan and Hills 1973); (2) decrease in shelf area due to regressing seas
(Schopf 1974; Simberloff 1974); (3) low productivity of primary producers (Tappan
and Loeblich 1973). As the sea retreated, many species disappeared. Those that
disappeared first and suffered most were highly specialized and/or stenotopic forms
(cf. previous section and section on relative advantages), which strongly suggests that
changing physical conditions contributed significantly to many extinctions. The effect
played by competition, either for food or space, is more difficult to assess. As shelf area
decreased and food became scarce competition may have ensued, unless intolerable
physical conditions had resulted in extinctions that more than compensated for the
decreased availability of the necessary resources. Eventual extinction of many species
may have been accelerated by predation and/or competition once populations dropped
below a critical value (MacArthur 1972, pp. 92-97). Therefore, it can be argued that
filter-feeding bivalves suffered relatively fewer extinctions at the end of the Palaeozoic
than did the articulates because of the greater eurytopy of these bivalves and possibly

STEELE-PETROVIC: BRACHIOPODS AND BIVALVES 125

also to some extent because of greater competitive ability. The Triassic- Jurassic
extinctions also affected articulates considerably more than filter-feeding bivalves,
possibly for the same reasons, as suggested by the fact that I could find no evidence for
biased loss of habitats that were particularly favourable for articulates ; however, these
possible causes for Triassic-Jurassic extinctions cannot be tested at present.

The dearth of benthic faunas in the Lower Triassic contrasts with the more common
occurrence of nektonic forms (see above), and suggests that bottom living conditions
were highly unfavourable at that time. By the Middle Triassic these conditions had
begun to improve, as indicated by a more diverse and abundant benthos. It can be
argued, on the basis of comparative physiology of articulates and filter-feeding
bivalves, that once benthic conditions started to improve, these bivalves should have
had a colonizing advantage over the articulates for the following reasons: (1) because
of the greater eurytopy of filter-feeding bivalves, physical conditions that were
tolerable for them would have been widespread before suitable conditions for the more
stenotopic articulates had developed, and filter-feeding bivalves generally should have
emerged from the environments where they sought refuge during the Lower Triassic
before articulate brachiopods did ; (2) since bivalve larvae generally have a planktonic
stage of several weeks, in contrast to a very short or absent planktonic stage of
articulates, filter-feeding bivalves should have spread into unoccupied areas relatively
more quickly; (3) since these bivalves can effectively cope with a range of en-
vironmental conditions they should have become widespread in a variety of open
habitats ; (4) if food were in short supply, the bivalves, because of their more effective
feeding methods, usually should have been favoured.

This ability of filter-feeding bivalves to extensively colonize level-bottom environ-
ments ahead of articulate brachiopods in the Middle Triassic, adequately explains the
subsequent dominance of these bivalves over the articulates. In addition, after
articulates declined in the Triassic-Jurassic extinctions, filter-feeding bivalves, which
were relatively little affected, must have moved into many of the habitats that were
previously occupied by the articulates, thus increasing the dominance. In agreement
with modern ecological theory, one would expect that as environmental conditions
ameliorated in the Middle Triassic, and invasions (not only of filter-feeding bivalves
but also of other relatively eurytopic benthic invertebrates) increased, the pioneer
level-bottom communities would have developed greater diversity and greater
resistance to outside invasion. When conditions became favourable for most
articulates and they attempted to spread, they probably encountered resistance of
varying intensities from interacting species in different established communities. Many
environments (e.g. shallow subtidal, shallow lagoonal, reefal) that were occupied by
articulate brachiopods during the Palaeozoic, contained filter-feeding bivalves but no
articulates in the Mesozoic (see above). There is no reason to suggest that Mesozoic
articulates were unable to tolerate these environments. Rather, their exclusion can be
explained by an early establishment of other invertebrates. An invading articulate
species would probably have competed mainly with established filter-feeding bivalves,
but also with other less similar animals that utilized some of the same resources. The
success of an articulate attempting to invade a low-diversity community would
probably have depended on the degree of overlap, particularly with the established
bivalves. Its chances of success in a relatively diverse community consisting of highly

126

PALAEONTOLOGY, VOLUME 22

specialized and interacting species would probably have been low, particularly if some
of the established species had been filter-feeding bivalves with considerable overlap in
requirements with the invader. Because of a scarcity of information on Triassic
communities, and on both succession in Mesozoic communities in general and life
histories of the particular Mesozoic species involved, much of the present discussion is
necessarily generalized. There are sufficient data to permit only those communities that
inhabited soft muddy bottoms to be discussed in slightly greater detail.

As noted above, filter-feeding bivalves moved into and occupied soft muddy
sediments during the Mesozoic ; except for the occasional oyster, these bivalves were
predominately infaunal siphonate forms. This situation contrasts greatly with that of
the Palaeozoic, when soft muddy environments were inhabited by abundant free-living
epifaunal articulates. Since siphonate bivalves are much better suited than are most
other forms of filter-feeding bivalves for handling suspended sediment, and as a result
for occupying soft muddy environments, and since eulamellibranch gills are impe-
rative for deep burrowers, the presence of the siphon and almost certainly the
eulamellibranch gill permitted filter-feeding bivalves, early in the Mesozoic, to move in
abundance into these open muddy environments (cf. Stanley 1968, 1972). Also,
certain oysters, both extant and fossil, are secondarily adapted physiologically for
handling large quantities of mud (Nelson 1938). One cannot argue convincingly that
unsuitable physical conditions prevented articulates from re-occupying these environ-
ments during the Mesozoic; after all, these animals had the potential to evolve
numerous highly specialized morphologies which could effectively cope with soft
muddy substrates, as illustrated in the Palaeozoic; also, although most Palaeozoic
mud-dwelling articulates became extinct at the end of the Permian, there were a few
Mesozoic articulates such as Terebratulina and related forms which were specially
adapted for life on a muddy bottom (Ager in lift. 1977), and these genera must have had
a potential to radiate. In contrast, the absence of Mesozoic articulates in muds can be
explained adequately by the presence of filter-feeding bivalves. In general, when
Mesozoic articulates suited for life on muddy substrates began to spread, as conditions
became favourable for them, they probably encountered competition, particularly
from oysters and siphonate bivalves, but also from other invertebrates that had
previously become established in the mud; and any new mutant that could have
exploited muddy bottoms would have faced similar competition. Although articulates
and siphonate bivalves lived at different horizons, siphons of bivalves reached to the
surface or a little above, and bivalves would have fed where articulates lived and fed.
Since both groups utilized essentially the same food and feeding space, intense
competition could have resulted. Therefore, it can be argued that the presence of early
colonizers, and in particular filter-feeding bivalves, prevented the Mesozoic articulates
from realizing their potential for life on soft muddy substrates. If filter-feeding bivalves
had been unable to exploit muddy environments, Mesozoic articulates probably would
have become re-established on these muds.

The ability of filter-feeding bivalves to colonize earlier and faster than articulates
undoubtedly affected the character of radiations of the two groups. Bivalves had more
opportunity to evolve new forms as they expanded into a greater number of different
environments, and articulates remained relatively conservative, even for them, because
they were unable to recolonize many of the environments that they had occupied

STEELE-PETROVIC: BRACHIOPODS AND BIVALVES 127

during the Palaeozoic. The fact that articulates were excluded from Mesozoic soft
muds and reefs, two habitats in which they had been highly specialized in the
Palaeozoic, probably contributed greatly to the Mesozoic-Cenozoic conservativism
of the group.

This situation with articulate brachiopods, where they suffered several severe
extinctions but failed to subsequently regain their previous importance, contrasts with
that of ammonoids (Rudwick 1970, p. 183), which came close to extinction at the end
of both the Permian and Triassic, but unlike brachiopods re-radiated significantly
(Arkell 1957; Teichert 1967) and re-expanded in importance to again become a
dominant part of the fauna. A plausible explanation for this difference is that no other
animals had moved in to fill the roles held by the ammonites before these extinctions.

IMPLICATIONS FOR FURTHER RESEARCH

Changes in dominance of articulate brachiopods and filter-feeding bivalves in marine
level-bottom communities are interpreted in this paper in terms of physiological
differences between the two groups. Although the proposed explanations fit the
available evidence, additional information is essential to verify the ideas presented here
and to permit solving of numerous problems that cannot be answered at present. In
particular, more field and/or laboratory data are needed on the biology of the two
groups (particularly articulates) and on the structure of marine level-bottom
communities through time. The collection of such data requires experts in several
normally unrelated disciplines. I hope that this paper draws the attention of specialists
in the relevant fields to the outstanding problems and to the importance of these
problems for understanding the evolution of marine level-bottom communities.

SUMMARY AND CONCLUSIONS

1 . During the Phanerozoic, articulate brachiopods and filter-feeding bivalves have
overlapped considerably in habits and habitat requirements.

2. The food of brachiopods and filter-feeding bivalves is essentially the same.

3. Compared with brachiopods, filter-feeding bivalves gain more energy in pumping,
trapping, transporting particles to the mouth, and digestion.

4. Filter-feeding bivalves generally cope better than articulate brachiopods with
different environmental conditions.

5. At present there is insufficient information to determine relative susceptibility of
articulates and filter-feeding bivalves to predation.

6. The relatively long planktonic stage of filter-feeding bivalve larvae permits
quicker colonization and expansion than by articulate brachiopods.

7. The importance of competition cannot be dismissed either in Recent level-bottom
communities, in general, or within subsets of marine invertebrates (especially bivalves)
of such communities, in particular.

8. At certain times during the Phanerozoic, competition may have been considerably
more intense than at present.

9. Between the Middle Ordovician and about the end of the Permian, articulate
brachiopods and filter-feeding bivalves shared the near-shore environments, while

128

PALAEONTOLOGY, VOLUME 22

articulates dominated off-shore. After the Permo-Triassic extinctions, filter-feeding
bivalves generally dominated over articulates. This dominance increased again at the
Triassic-Jurassic boundary and again during the Cenozoic.

10. There is evidence that partitioning of space contributed to the co-existence of
articulate brachiopods and filter-feeding bivalves close to shore during the Palaeozoic.

1 1 . Physical disturbances close to shore permitted filter-feeding bivalves to become
established in near-shore communities early in the Palaeozoic. The dampening of
disturbances away from shore enabled off-shore communities to become highly
diverse, and to maintain essentially the same ecological structures throughout the
Palaeozoic. Filter-feeding bivalves could not fit into these established community
structures, and it was not until these communities were disrupted at the end of the
Permian that the bivalves moved off-shore.

12. The decline of articulate brachiopods and change in dominance of articulates
and filter-feeding bivalves cannot be attributed to predation.

13. This decline and change in dominance can be attributed to physiological
differences between the two groups of animals. Filter-feeding bivalves suffered less
than articulates at each period of extinction. Following the extinctions, these bivalves
were able to re-colonize a large variety of widely spread open habitats earlier and faster
than the brachiopods. Competition then probably prevented articulates from invading
many regions. With each wave of extinction, filter-feeding bivalves gained prominence
at the expense of articulate brachiopods.

Acknowledgements. I thank D. V. Ager, P. W. Bretsky, P. A. Heithaus, J. M. Hurst, M. Mancenido, S.
Mancenido, R. T. Paine, R. Petrovic, and S. M. Stanley for comments on parts or all of the manuscript at
various stages in its development. I am particularly grateful to Professor Ager for communicating to me
certain information, mainly on brachiopods and communities of Mesozoic age, which I was unable to
obtain from the literature; and to Dr. E. R. Heithaus and Dr. Petrovic for numerous helpful discussions.
I appreciate the use of facilities at the Department of Geological Sciences, Northwestern University.

REFERENCES

agatova, a. i. and bogdanov, yu. A. 1972. Biochemical composition of suspended organic matter from
the tropical Pacific. Oceanology, 12, 227-235.

ager, D. v. 1962. The occurrence of pedunculate brachiopods in soft sediments. Geol. Mag. 99, 184-186.
— 1965. The adaptation of Mesozoic brachiopods to different environments. Palaeogeogr.
Palaeoclimat. Palaeoecol. 1, 143-172.

1967u. Brachiopod palaeoecology. Earth Sci. Rev. 3, 157-179.

— 1967ft. Some Mesozoic brachiopods in the Tethys region. In adams, c. g. and ager, d. v. (eds.).
Aspects of Tethyan biogeography. Pubis Syst. Ass. 7, 135-151.

— 1971. Space and time in brachiopod history. In middlemiss, f. a., rawson, p. f. and newall, g. (eds.).
Faunal provinces in space and time. Geol. J. Spec. Issue, 4, pp. 95-110.

1976. The nature of the fossil record. Proc. Geol. Ass. 87, 131-160.

allen, J. a. 1962. Preliminary experiments on the feeding and excretion of bivalves using Phaeodactylum
labelled with 32P. J. mar. biol. Ass. U.K. 42, 609-623.

ANDREWS, S. M., GARDINER, B. G., MILES, R. S., and PATTERSON, C. 1967. Pisces. Pp. 637-683. In HARLAND,
w. B. et al. (eds.). The fossil record, xi+827 pp. Geological Society of London.
ansell, a. d. 1961. The functional morphology of the British species of Veneracea (Eulamellibranchia). J.
mar. biol. Ass. U.K. 41, 489-515.

arkell, w. j. 1957. Introduction to Mesozoic Ammonoidea. In moore, r. c. (ed.). Treatise on invertebrate
paleontology. Part L, Mollusca 4. Pp. L80-L129. Geological Society of America and University of
Kansas Press.

STEELE-PETROVIC: BRACHIOPODS AND BIVALVES

129

atkins, D. 1936a. On the ciliary mechanisms and interrelationships of lamellibranchs. Part I. New
observations on sorting mechanisms. Q. Jl microsc. Sci. 79, 181-308.

— 1936 ft. On the ciliary mechanisms and interrelationships of lamellibranchs. Part II. Sorting devices
on the gills. Ibid. 79, 339-373.

— 1936c. On the ciliary mechanisms and interrelationships of lamellibranchs. Part III. Types of
lamellibranch gills and their food currents. Ibid. 79, 375-421.

— 1960. The ciliary feeding mechanism of the Megathyridae (brachiopods), and the growth stages of
the lophophore. J. mar. biol. Ass. U.K. 39, 459-479.
ballantine, d. and morton, J. E. 1956. Filtering, feeding and digestion in the lamellibranch Lasaea rubra.
Ibid. 35, 241-274.

bamford, D. R. and gingles, R. 1974. Absorption of sugars in the gill of the Japanese oyster Crassostrea
gigas. Comp. Biochem. Physiol. 49A, 637-646.

— and mccrea, r. 1975. Active absorption of neutral and basic amino acids by the gill of the common
cockle, Cerastoderma edule. Ibid. 50A, 811-817.

bayne, b. l. 1971. Some morphological changes that occur at the metamorphosis of the larvae of Mytilus
edulis. In crisp, d. j. (ed.). European marine biology symposium. Vol. 4. Pp. 259-280. Cambridge
University Press, New York.

Bernard, F. R. 1974. Particle sorting and labial palp function in the Pacific oyster Crassostrea gigas
(Thunberg 1895). Biol. Bull. 146, 1-10.

bowen, z. p. 1968. A guide to New Zealand Recent brachiopods. Tuatara, 16, 127-150.

Bradley, w. H. and cooke, p. 1959. Living and ancient populations of the clam Gemma gemma in a Maine
coast tidal flat. Fishery Bull. Fish. Wild l. Serv. U.S. 137, 305-334.
bretsky, p. w. 1969a. Evolution of Paleozoic benthic marine invertebrate communities. Palaeogeogr.
Palaeoclimat. Palaeoecol. 6, 45-59.

— 1969ft. Central Appalachian late Ordovician communities. Bull. geol. Soc. Am. 80, 193-212.
bullivant, J. s. 1968. The method of feeding of lophophorates (Bryozoa, Phoronida, Brachiopoda). N.Z.

Jl mar. freshw. Res. 2, 135-146.

CONNELL, J. h. 1975. Some mechanisms producing structure in natural communities : a model and evidence
from field experiments. In cody, m. l. and diamond, j. m. (eds.). Ecology and evolution of communities.
Pp. 460-490. Belknap Press of Harvard University Press, Cambridge, Mass.
cowen, R. 1971. The food of articulate brachiopods— a discussion. J. Paleont. 45, 137-139.
cox, L. r. 1969. General features of Bivalvia. In moore, r. c. (ed.). Treatise on invertebrate paleontology,
Part N, Mollusca 6 Bivalvia. Pp. N2-N128. Geological Society of America and University of Kansas
Press.

dagis, a. s. and ustritsky, v. i. 1973. The main relationships between the changes in marine fauna at the
close of the Permian and the beginning of the Triassic. In logan, a. and hills, l. v. (eds.). The Permian
and Triassic systems and their mutual boundary. Mem. Can. Soc. Petrol. Geol. 2, 647-654.
dayton, p. K. and hessler, r. r. 1972. Role of biological disturbance in maintaining diversity in the deep
sea. Deep-Sea Res. 19, 199-208.

dean, D. 1958. New property of the crystalline style of Crassostrea virginica. Science, N.Y. 128, 837.
dodgson, r. w. 1928. Report on mussel purification. Fishery Invest., Lond. Ser. II, 10, 1-498.
dral, a. d. G. 1967. The movements of the lateral-frontal cilia and the mechanism of particle retention in
the mussel. Neth. J. Sea Res. 3, 391-422.

efford, i. E. and tsumura, K. 1973. Uptake of dissolved glucose and glycine by Pisidium, a freshwater
bivalve. Can. J. Zool. 51, 825-832.

fenchel, t. 1971. Aspects of decomposer food chains in marine benthos. Verh. dt. zool. Ges. 65,
14-23.

foster, m. w. 1974. Recent Antarctic and Subantarctic brachiopods. Antarctic Res. Ser. Am. geophys.
Union, Washington, 21, i-xi, 1-189.

George, w. c. 1952. The digestion and absorption of fat in lamellibranchs. Biol. Bull. 102, 118-127.
glaessner, M. f. 1969. Decapoda. In moore, r. c. (ed.). Treatise on invertebrate paleontology. Part R,
Arthropoda 4(2). Pp. R400-R566. Geological Society of America and University of Kansas Press.
GOREAU, t. F., goreau, n. I., yonge, c. m., and neuman, Y. 1970. On feeding and nutrition in Fungiacava
eilatensis (Bivalvia, Mytilidae), a commensal living in fungiid corals. J. Zool., Lond. 160, 159-
172.

130

PALAEONTOLOGY, VOLUME 22

grant, R. E. 1971. Brachiopods in the Permian reef environment of West Texas. Pp. 1444-1481. In
yochelson, e. L. (ed.). Proceedings of the North American paleontological convention. Vol. 2.
[viii] + 703- 1674 pp. Allen Press, Lawrence, Kansas.

— and cooper, G. a. 1973. Brachiopods and Permian correlations. In logan, a. and hills, l. v. (eds.).
The Permian and Triassic systems and their mutual boundary, Mem. Can. Soc. Petrol. Geol. 2, 572-595.

grassle, J. f. and Sanders, H. L. 1973. Life histories and the role of disturbance. Deep-Sea Res. 20, 643-
659.

hallam, A. 1971. Provinciality in Jurassic faunas in relation to facies and palaeogeography. In
middlemiss, F. A., rawson, p. F., and newall, G. (eds.). Faunal provinces in space and time. Geol. J.
Spec. Issue 4, pp. 129-152.

— 1975. Jurassic environments. 269 pp., Cambridge University Press, New York.

— 1976. Stratigraphic distribution and ecology of European Jurassic bivalves. Lethaia , 9, 245-259.
hammen, c. s. 1977. Brachiopod metabolism and enzymes. Am. Zool. 17, 141-147.
hughes, T. G. 1975. The sorting of food particles by Abra sp. (Bivalvia: Tellinacea). J. exp. mar. Biol. Ecol.
20, 137-156.

hutchinson, G. e. 1959. Homage to Santa Rosalia or Why are there so many kinds of animals? Am. Nat.
193, 145-159.

hylleberg, J. and gallucci, v. F. 1975. Selectivity in feeding by the deposit-feeding bivalve Macoma
nasuta. Mar. Biol. 32, 167-178.

jackson, J. b. c. 1972. The ecology of the molluscs of Thalassia communities, Jamaica, West Indies. II.
Molluscan population variability along an environmental stress gradient. Ibid. 14, 304-337.

— goreau, T., and hartman, w. d. 1971. Recent brachiopod-coralline sponge communities and their
palaeoecological significance. Science, N.Y. 173, 623-625.

Johnson, r. G. 1964. The community approach to paleoecology. Pp. 107-134. In imbrie, j. and newell,
n. D. (eds.). Approaches to paleoecology, viii+432 pp. Wiley, New York, London, and Sydney.
Jorgensen, c. b. 1966. Biology of suspension feeding. 357 pp., Pergamon Press, London.
kanmera, k. and nakazawa, k. 1973. Permian-Triassic relationships and faunal changes in the eastern
Tethys. In logan, a. and hills, l. v. (eds.). The Permian and Triassic systems and their mutual
boundary. Mem. Can. Soc. Petrol. Geol. 2, 100-119.
kauffman, E. G. 1973. Cretaceous Bivalvia. Pp. 353-383. In hallam, a. (ed.). Atlas of palaeobiogeography.

xii + 531 pp. Elsevier, Amsterdam, London, and New York.

KHAiLOV, k. m., filenko, G. a., burlakova, z. p., and smirnov, v. a. 1973. Relation of stationary
concentrations of basic forms of organic matter in sea water and the specific rate of their trophic
utilization of organisms in littoral communities. Hydrobiology, 209, 132-134.
kristensen, j. h. 1972a. Carbohydrases of some marine invertebrates with notes on their food and on the
natural occurrence of the carbohydrates studied. Mar. Biol. 14, 130-142.

— 19726. Structure and function of crystalline styles of bivalves. Ophelia, 10, 91-108.
kummel, B. 1973a. Aspects of the Lower Triassic (Scythian) Stage. In logan, a. and hills, l. v. (eds.). The
Permian and Triassic systems and their mutual boundary. Mem. Can. Soc. Petrol. Geol. 2, 557-571.

— 19736. Lower Triassic (Scythian) molluscs. Pp. 225-233. In hallam, a. (ed.). Atlas of palaeobiogeog-
raphy, xii + 531 pp. Elsevier, Amsterdam, London, and New York.

logan, a. 1975. Ecological observations on the Recent articulate brachiopod Argyrotheca bermudana
Dali from the Bermuda Platform. Bull. mar. Sci. 25, 186-204.

— and hills, l. v. (eds.). 1973. The Permian and Triassic systems and their mutual boundary. Mem.
Can. Soc. Petrol. Geol. 2, 1-766.

macarthur, r. h. 1972. Geographical ecology. Patterns in the distribution of species. 269 pp., Harper and
Row, New York.

mccammon, H. M. 1969. The food of articulate brachiopods. J. Paleont. 43, 976-985.

— 1973. The ecology of Magellania venosa, an articulate brachiopod. Ibid. 47, 266-278.
mansour, K. 1946. Food and digestive processes of the lamellibranchs. Nature, Lond. 157, 482.
mansour-bek, J. j. 1945. The digestive enzymes of Tridacna elongata Lamk. and Pinctada vulgaris L. (A
preliminary communication.) Proc. Egypt. Acad. Sci. 1, 13-20.
mathers, N. F. 1972. The tracing of a natural algal food labelled with Carbon 14 isotope through the
digestive tract of Ostrea edulis L. Proc. malac. Soc. Lond. 40, 115-124.

— 1973. Carbohydrate digestion in Ostrea edulis L. Ibid. 40, 359-367.

STEELE-PETROVIC: BRACHIOPODS AND BIVALVES

131

menge, B. A. 1975. Brood or broadcast? The adaptive significance of different reproductive strategies in the
two intertidal sea stars Leptasterias hexactis and Pisaster orchraceus. Mar. Biol. 31, 87-100.

— and Sutherland, j. p. 1976. Species diversity gradients: synthesis of the roles of predation,
competition, and temporal heterogeneity. Am. Nat. 110, 351-369.

mileikovsky, s. A. 1971. Types of larval development in marine bottom invertebrates, their distribution
and ecological significance: a re-evaluation. Mar. Biol. 10, 193-213.
moore, H. J. 1971. The structure of the latero-frontal cirri on the gills of certain lamellibranch molluscs and
their role in suspension feeding. Ibid. 11, 23-27.

morton, b. 1973. A new theory of feeding and digestion in the filter-feeding Lamellibranchia.
Malacologia, 14, 63-79.

morton, J. E. 1952. The role of the crystalline style. Proc. malac. Soc. Lond. 29, 85-92.
muus, k. 1973. Settling, growth and mortality of young bivalves in the 0resund. Ophelia , 12, 79-116.
nakazawa, K. and runnegar, b. 1973. The Permian-Triassic boundary : a crisis for bivalves? In logan,
a. and hills, L. v. (eds.). The Permian and Triassic systems and their mutual boundary. Mem. Can. Soc.
Petrol. Geol. 2, 608-621.

narchi, w. 1972. Comparative study of the functional morphology of Anomalocardia brasiliana (Gmelin
1791) and Tivela mactroides (Born 1778) (Bivalvia, Veneridae). Bull. mar. Sci. 22, 643-670.

1975. Functional morphology of a new Petricola (Mollusca Bivalvia) from the littoral of Sao Paulo,

Brazil. Proc. malac. Soc. Lond. 41, 451-465.

nelson, T. c. 1933. On the digestion of animal forms by the oyster. Proc. Soc. exp. Biol. Med. 30, 1287-
1289.

— 1938. The feeding mechanism of the oyster. Part I. J. Morph. 63, 1-61.

Newell, N. d. 1973. The very last moment of the Paleozoic Era. In logan, a. and hills, l. v. (eds.). The
Permian and Triassic systems and their mutual boundary. Mem. Can. Soc. Petrol. Geol. 2, 1-10.

— and larocque, A. 1969. ?Family Lyrodesmatidae, Ulrich 1894. In moore, r. c. (ed.). Treatise on
invertebrate paleontology. Part N, Mollusca 6, Bivalvia. P. N471. Geological Society of America and
University of Kansas Press.

Newell, R. 1965. The role of detritus in the nutrition of two marine deposit feeders, the prosobranch
Hydrobia ulvae, and the bivalve Macoma balthica. Proc. zool. Soc. Lond. 144, 25-45.
owen, G. 1953. On the biology of Glossus humanus (L.) ( Isocardia cor Lam.). J. mar. biol. Ass. U.K. 32, 85-
106.

— 1955. Observations on the stomach and digestive diverticula of the Lamellibranchia. I. The
Anisomyaria and Eulamellibranchia. Q. Jl microsc. Sci. 96, 517-537.

— 1974. Feeding and digestion in the Bivalvia. Adv. comp. Physiol. Biochem. 5, 1-35.

— and mccrae, J. m. 1976. Further studies on the latero-frontal tracts of bivalves. Proc. R. Soc. B194,
527-544.

paine, R. T. 1966. Food web complexity and species diversity. Am. Nat. 100, 65-75.

— 1971. A short-term experimental investigation of resource partitioning in a New Zealand rocky
intertidal habitat. Ecology, 52, 1096-1106.

— 1974. Intertidal community structure. Experimental studies on the relationship between a dominant
competitor and its principal predator. Oecologia, 15, 93-120.
parsons, t. r., Stephens, k. and Strickland, j. d. h. 1961. On chemical composition of eleven species of
marine phytoplankters. J. Fish. Res. Bd Can. 18, 1000-1016.
pattison, j., smith, d. b. and Warrington, G. 1973. A review of late Permian and early Triassic
biostratigraphy in the British Isles. In logan, a. and hills, l. v. (eds.). The Permian and Triassic
systems and their mutual boundary. Mem. Can. Soc. Petrol. Geol. 2, 220-260.
percival, e. 1944. A contribution to the life-history of the brachiopod Terebratella inconspicua Sowerby.
Trans. R. Soc. N.Z. 74, 1-23.

— 1960. A contribution to the life-history of the brachiopod Tegulorhynchia nigricans. Q. Jl microsc.
Sci. 101, 439-457.

pohlo, R. 1969. Confusion concerning deposit feeding in the Tellinacea. Proc. malac. Soc. Lond. 38, 361—
364.

— 1972. Feeding and associated morphology in Sanguinolaria nuttallii. Veliger, 14, 298-301.
prosser, c. L. and brown, f. a. 1961. Comparative animal physiology. 2nd edn. 688 pp., W. B. Saunders
Co., Philadelphia.

132

PALAEONTOLOGY, VOLUME 22

purchon, r. d. 1955. The structure and function of the British Pholadidae (rock-boring Lamellibranchia).
Proc. zool. Soc. Lond. 124, 859-911.

raymont, J. E. G. 1971. Alternative sources of food in the sea. In costlow, j. d. (ed.). Fertility of the sea,
Vol. II, pp. 383-399. Gordon and Breach Sci. Publ., New York.
reeside, j. b. 1957. Paleoecology of the Cretaceous seas of the Western Interior of the United States. In
ladd, H. s. (ed.). Treatise on marine ecology and paleoecology. Vol. 2. Mem. geol. Soc. Am. 67(2), 505-
542.

reid, R. G. b. 1965. The structure and function of the stomach in bivalve molluscs. J. Zool., Lond. 147, 1 56—
184.

— 1966. Digestive tract enzymes in the bivalves Lima hians Gmelin and My a arenaria L. Comp.
Biochem. Physiol. 17, 417-433.

— 1968. The distribution of digestive tract enzymes in lamellibranchiate bivalves. Ibid. 24, 727-744.

— and dunnill, R. M. 1969. Specific and individual differences in the esterases of members of the genus
Macoma (Mollusca: Bivalvia). Ibid. 29, 601-610.

— and rauchert, k. 1970. Proteolytic enzymes in the bivalve mollusc Chlamys hericius Gould. Ibid. 35,
689-695.

— 1972. Protein digestion in members of the genus Macoma (Mollusca: Bivalvia). Ibid. 41 A, 887-

895.

— and reid, a. 1969. Feeding processes of members of the genus Macoma (Mollusca : Bivalvia). Can. J.
Zool. 47, 649-657.

Reynolds, w. a. and mccammon, h. m. 1977. Aspects of the functional morphology of the lophophore in
articulate brachiopods. Am. Zool. 17, 121-132.

richardson, j. r. and watson, j. e. 1975a. Locomotory adaptations in a free-lying brachiopod. Science,
N.Y. 189, 381, 382.

19756. Form and function in a Recent free-living brachiopod Magadina cumingi. Paleobiology,

1, 379-387.

rickwood, a. E. 1968. A contribution to the life history and biology of the brachiopod Pumilus antiquatus
Atkins. Trans. R. Soc. N.Z. 10, 163-182.

rudwick, m. j. s. 1961. The anchorage of articulate brachiopods on soft substrata. Palaeontology, 4, 475,
476.

— 1962a. Notes on the ecology of brachiopods in New Zealand. Trans. R. Soc. N.Z. 25, 327-335.

— 19626. Filter-feeding mechanisms in some brachiopods from New Zealand. J. Linn. Soc. 44, 592-
615.

— 1965. Ecology and paleoecology. In moore, r. c. (ed.). Treatise on invertebrate paleontology. Part H,
Brachiopoda. Pp. H199-H214. Geological Society of America and University of Kansas Press.

— 1970. Living and fossil brachiopods. 199 pp., Hutchinson University Library, London.
runnegar, b. 1974. Evolutionary history of the bivalve subclass Anomalodesmata. J. Paleont. 48, 904-

940.

Sanders, H. l. 1960. Benthic studies in Buzzards Bay. III. The structure of the soft-bottom community.
Limnol. Oceanogr. 5, 138-153.

— 1968. Marine Benthic diversity: a comparative study. Am. Nat. 102, 243-282.

— 1969. Benthic marine diversity and the stability-time hypothesis. Brookhaven Symp. Biol. 22, 71-80.
schopf, T. J. M. 1974. Permo-Triassic extinctions: relation to sea-floor spreading. J. Geol. 82, 129-143.

— raup, d. m., gould, s. J., and simberloff, d. s. 1975. Genomic versus morphologic rates of evolution.
Paleobiology, 1, 63-70.

simberloff, D. 1974. Permo-Triassic extinctions: effects of area on biotic equilibrium. J. Geol. 82, 267-
274.

slobodkin, l. b. and Sanders, h. L. 1969. On the contribution of environmental predictability to species
diversity. Brookhaven Symp. Biol. 22, 82-93.
sohl, N. F. 1969. The fossil record of shell boring by snails. Am. Zool. 9, 725-734.

SOVA, v. v., elyakova, L. a., and vaskovsky, v. E. 1970. The distribution of laminarinases in marine
invertebrates. Comp. Biochem. Physiol. 32, 459-464.
spencer, w. k. and wright, c. w. 1966. Asterozoans. Pp. U4-U107. In moore, r. c. (ed.). Treatise on
invertebrate paleontology. Part U, Echinodermata 3(1). Geological Society of America and University of
Kansas Press.

STEELE-PETROVIC: BRACHIOPODS AND BIVALVES

133

Stanley, s. M. 1968. Post-Paleozoic adaptive radiation of infaunal bivalve molluscs— a consequence of
mantle fusion and siphon formation. J. Paleont. 42, 214-229.

— 1970. Relation of shell form to life habits of the Bivalvia. Mem. geol. Soc. Am. 125, 1-296.

— 1972. Functional morphology and evolution of byssally attached bivalve molluscs. J. Paleont. 46,
165-212.

— 1974a. Effects of competition on rates of evolution, with special reference to bivalve mollusks and
mammals. Syst. Zool. 22, 486-506.

— 19746. What has happened to the articulate brachiopods? Abstr. Progm geol. Soc. Am. 966, 967.

— 1975. Adaptive themes in the evolution of the Bivalvia (Mollusca). A. Rev. Earth planet. Sci. 3, 361 -
385.

Stauffer, r. c. 1937. Changes in the invertebrate community of a lagoon after disappearance of the eel
grass. Ecology, 18, 427-431.

steele-petrovic, H. M. 1975. An explanation for the tolerance of brachiopods and relative intolerance of
filter-feeding bivalves for soft muddy bottoms. J. Paleont. 49, 552-556.

— 1976. Brachiopod food and feeding processes. Palaeontology , 19, 417-436.

STRATHMANN, r. 1973. Function of lateral cilia in suspension feeding of Lophophorates (Brachiopoda,
Phoronida, Ectoprocta). Mar. Biol. 23, 129-136.

tappan, H. 1971. Microplankton, ecological succession and evolution. Pp. 1058-1103. In yochelson, e. l.
(ed.). Proceedings of the North American paleontological convention , vol. 2, [viii] + 703- 1674 pp. Allen
Press, Lawrence, Kansas.

and loeblich, a. r. 1973. Smaller protistan evidence and explanation of the Permian-Triassic crisis.

In logan, a. and hills, l. v. (eds.). The Permian and Triassic systems and their mutual boundary. Mem.
Can. Soc. Petrol. Geol. 2, 465-480.

teichert, c. 1967. Major features of cephalopod evolution. Pp. 162-210. In teichert, c. and yochelson,
e. L. (eds.). Essays in paleontology and stratigraphy, R. C. Moore Commemorative Volume, [vi] + 626 pp.
Department of Geology, University of Kansas Special Publication 2. University of Kansas Press,
Lawrence.

thayer, c. w. 1974. Salinity tolerances of articulate brachiopods. Abstr. geol. Soc. Am. Northeastern
Section, 80-81.

— 1975. Strength of pedicle attachment in articulate brachiopods: ecologic and paleoecologic
significance. Paleobiology, 1, 388-399.

1977. Recruitment, growth, and mortality of a living articulate brachiopod, with implications for the

interpretation of survivorship curves. Ibid. 3, 98-109.
turpaeva, E. p. 1948. The feeding of some benthic invertebrates of the Barents Sea. Zool. Zh. 27, 503-512.
[In Russian.]

— 1949. The significance of food relations in the structure of marine benthic Biocoenoses. Dokl. Akad.
Nauk SSSR, 65, 93-96. [In Russian.]

1953. Feeding and food groups of marine benthic invertebrates. Trudy Inst. Okeanol. 7, 259-299. [In

Russian.]

— 1954. Types of marine benthic biocoenoses and the dependence of their distribution on abiotic
environmental factors. Ibid. 11, 36-55. [In Russian.]

— 1957. Food interrelationships of dominant species in marine benthic biocoenoses. In nikitin, b. n.
(ed.). Marine biology [translated from Trudy Inst. Okeanol. 20], Am. Inst. biol. Sci., 137-148.

vance, r. 1973. On reproductive strategies in marine benthic invertebrates. Am. Nat. 107, 339-352.
van valen, L. 1976. Energy and evolution. Evolutionary Theory, 1, 179-229.

vaskovsky, v. e. and suppes, z. s. 1972. Phospholipases of marine invertebrates — I. Distribution of
Phospholipase A. Comp. Biochem. Physiol. 43B, 601-609.
verwey, J. 1952. On the ecology of distribution of cockle and mussel in the Dutch Wadden Sea. Archs
need. Zool. 10, 171-239.

vogel, k. 1966. Eine Funktionsmorphologische Studie an der Brachiopodengattung Pygope (Malm bis
Unterkreide). Neus Jb. Geol. Paldont. 125, 423-442.

walker, k. r. 1972. Trophic analysis: a method for studying the function of ancient communities. J.
Paleont. 46, 82-93.

— and laporte, l. f. 1970. Congruent fossil communities from Ordovician and Devonian carbonates of
New York. Ibid. 44, 928-944.

134

PALAEONTOLOGY, VOLUME 22

Waterhouse, J. b. and bonham-carter, G. 1976. Range, proportionate representation, and demise of
brachiopod families through Permian Period. Geol. Mag. 113, 401-428.
webb, G. R., logan, a., and noble, J. p. a. 1976. Occurrence and significance of brooded larva in a Recent
brachiopod, Bay of Fundy, Canada. J. Paleont. 50, 869-871.
wojtowicz, m. b. 1972. Carbohydrases of the digestive gland and the crystalline style of the Atlanticdeep-
sea scallop ( Placopecten magellanicus, Gmelin). Comp. Biochem. Physiol. 43A, 131-141.
woodin, s. a. 1974. Polychaete abundance patterns in a marine soft-sediment environment: the
importance of biological interactions. Ecol. Monogr. 44, 171-187.

— 1976. Adult-larval interactions in dense infaunal assemblages: patterns of abundance. J. mar. Res.
34, 25-41.

yonge, c. m. 1926. Structure and physiology of the organs of feeding and digestion in Ostrea edulis. J. mar.
biol. Ass. U.K. 14, 295-386.

— 1935. On some aspects of digestion in ciliary feeding animals. Ibid. 20, 341-346.

— 1949. On the structure and adaptations of the Tellinacea, deposit-feeding Eulamellibranchia. Phil.
Trans. R. Soc. B234, 29-76.

zobell, c. E. and feltham, c. b. 1938. Bacteria as food for certain marine invertebrates. J. mar. Res. 1,
312-327.

Manuscript received 7 December 1977
Revised manuscript received 17 April 1978

H. MIRIAM STEELE-PETROVIC

4904 S.E. Princeton Drive
Bartlesville, Oklahoma 74003
U.S.A.

A MIDDLE JURASSIC MAMMAL BED
FROM OXFORDSHIRE

by ERIC F. FREEMAN

Abstract. The geology, palaeoecology, and vertebrate fauna of a mammaliferous sediment in the Forest Marble of
Oxfordshire are described, as are the techniques used in its processing. The mammalian fossils appear to have been
derived from the faeces of predatory animals, probably small theropods, the occurrence of theropod teeth possibly
having value in the search for new Mesozoic mammal localities. In the mammal (s.l.) fauna from the new site there
occur late representatives of the Morganucodontidae ( Wareolestes rex gen. et sp. nov.), Kuehneotheriidae ( Cyrtla -
therium canei gen. et sp. nov.), and Tritylodontidae, as well as the earliest known members of the families Dryolestidae
and (?)Peramuridae ( Palaeoxonodon ooliticus Freeman, 1976). The holotype of the ( ?)peramurid appears to show an
early stage in the development of the talonid basin. The first known upper molars of Middle Jurassic representatives
of the suborders Amphitheria and Docodonta are described and illustrated, as are the positions of the wear facets
thereon. The mammals now appear to have diversified at the level of the family long before the Bathonian, and
probably during the early Jurassic.

In recent years there has been a revival of interest in the mammalian faunas of the
Mesozoic formations, an area of study largely dormant since the publication of
the classic monographs of Owen and Simpson in 1871 and 1928/9 respectively. The
recent large increase in both the quantity of fossil material and in the number of pro-
ductive localities has been due principally to the introduction of various techniques
for the processing of sediments on a large scale.

In spite of this increased level of activity, the known mammalian fauna of the
Middle Jurassic has not been substantially increased since the nineteenth century,
when the first Mesozoic mammals known to science were found by the miners work-
ing the Stonesfield Slate. The only major addition published since then has been the
fauna from the Great Estuarine Series of Skye, (Waldman and Savage 1972).

Nevertheless the British Middle Jurassic seems to abound in fossil mammalia and
I have managed to add another three mammaliferous localities to the two previously
known, one in Dorset (Freeman 1976a) and two in Oxfordshire (Freeman 19766).
Undoubtedly, many more such localities must await discovery.

The following account deals with the most significant of the three sites, the one in
the Forest Marble of the Old Cement Works Quarry, Kirtlington, Oxfordshire
(SP 494200).

STRATIGRAPHY

The Kirtlington Old Cement Works Quarry has been disused since 1929 (Dodsworth
1972), but still provides a highly fossiliferous exposure of the White Limestone,
Forest Marble, and Cornbrash. Together with similar quarries in the neighbourhood,
it has been the subject of numerous papers on Middle Jurassic geology and palaeonto-
logy. Notable among these are the works of Arkell (1931) and McKerrow, Johnson,
and Jakobson (1969). The better exposure available to Arkell permitted a section to

[Palaeontology, Vol. 22, Part 1, 1979, pp. 135-166, pis. 15-21.]

136

PALAEONTOLOGY, VOLUME 22

be measured which facilitated correlation with other exposures in Oxfordshire and
Gloucestershire. Since then, the main face of the quarry (the east face) has deteriorated
to the point where the more detailed stratigraphical and palaeoecological study of
McKerrow et al. (1969) had to be confined to scattered exposures away from Arkell’s
area of study. The following remarks are primarily concerned with establishing the
stratigraphic position of the mammal bed, in both the immediate context of the
quarry, and more generally in the regional context ; they do not extend to such prob-
lems as the position of the White Limestone/Forest Marble junction, nor the sub-
division and long-range correlation of the Forest Marble (see Odling 1913;
Richardson, Arkell, and Dines 1946; and Palmer 1973).

The bed of unconsolidated brown marl which has yielded the mammalian fossils
is without doubt that labelled 3p in McKerrow et al. (1969); in the present account it
is named the Kirtlington Mammal Bed. It is impersistent in its occurrence, forming
a lenticle in the north-eastern corner of the quarry with a total outcrop of 21 - 5 metres
(see PI. 15, fig. 1). The bed is of variable thickness (c. 4 to 25 cm); at its extremities
it rapidly thins away to nothing, being replaced from below by an increasingly
greater thickness of a deep-yellow friable oolitic limestone which here, and elsewhere
along the outcrop, grades downwards into the massive ‘Coral Epithyris Limestone’
(3o) of McKerrow et al. (1969). A similar friable yellow limestone occurs immediately
above the mammal bed. The contact of the mammal bed with the soft limestones
above and below it are extremely abrupt, which suggests that the contacts are
erosional features.

From its position overlying the ‘Coral Epithyris Limestone’ (=Arkell’s ‘Fossili-
ferous Cream Cheese Bed’), the Kirtlington Mammal Bed would seem almost
certainly to be laterally equivalent to part of the ‘Kemble Beds’ shown in Arkell’s
1931 section. More specifically, as can be seen from text-fig. 1, it seems likely that it
is at the same stratigraphic level as a thin bed of friable oolitic limestone sandwiched
between the ‘Fossiliferous’ and ‘Unfossiliferous Cream Cheese’ beds at the northern
end (i.e. left-hand side) of Arkell’s 1931 section (q.v.).

In the broader context, as the ascending sequence of the Great Oolite Group in the
Oxford area is Chipping Norton Formation-Sharp’s Hill Formation (including the
Stonesfield Slate)-Taynton Limestone Formation-Hampen (Marly) Formation-
White Limestone-Forest Marble-Cornbrash (McKerrow and Kennedy 1973), the
new mammal fauna is clearly slightly younger than that from Stonesfield. As
ammonites have not been reported from the Forest Marble of the Kirtlington
area, the mammal bed cannot be ascribed to any of the Bathonian ammonite zones
directly. However, using indirect evidence, the occurrence of the ostracod Glypto-
cy there penni in the fimbriatus-waltoni Clay at Kirtlington (see text-fig. 1) has led
Bate and Mayes (1977) to suggest that this formation belongs in the Clydoniceras
discus Zone. If so, as the Lower Cornbrash contains this zonal ammonite, the beds
between the fimbriatus-waltoni Clay and the Lower Cornbrash, among them the
Kirtlington Mammal Bed, must also be assigned to the C. discus Zone. The contrary
opinion of Torrens (1969) places the Kemble Beds (and therefore presumably the
mammal bed) in the underlying Oxycerites aspidoides Zone (Upper Bathonian).

Further afield, the age relationships of the Kirtlington mammals to those described
elsewhere from the Middle Jurassic are at present known only approximately. As

Obscured

Distance from c.140m c.21-5m

Section A

Massive Limestone

Flaggy Limestone

R R

Rubbly Limestone

50cm

pTinfj with bored top
O with ooliths L with lignite
with cross bedding
^c^with corals & brachiopods
r — ) with limestone pebbles

? -

D

Obscured

: ; ? —

— G
O —

— o

7"

7"

WYCHW00D /\^
BEDS 10ft. 3 in

KEMBLE BEDS

BED 12 3ft.

BED 11

7ft. 3in.

Unfossiliferous
Cream Cheese
Bed

Fossiliferous Cream
Cheese Bed

BED 10 3 ft. 6in.

FIMBRIATA-
WALTONI CLAY

c.60m Equivalent of
Section D in
Arkell 1931

text-fig. 1. The short-range stratigraphy of the Kirtlington Mammal Bed. Section B corresponds in part
to the upper part of Profile 3 of McKerrow et al. 1969, as does the numbering of beds 3o to 3v. The mammal
bed is bed 3p. Sections A and C show the north-west and south-east extremities of the bed, while Section D
corresponds to part of the northern extremity of the section in Arkell 1931.

138

PALAEONTOLOGY, VOLUME 22

Waldman and Savage (1972) consider that their mammal site in the Isle of Skye is of
Middle Bathonian age, it would appear to be slightly older than Kirtlington, which
presumably is roughly contemporary with the other known Forest Marble site, at
Watton Cliff in Dorset (Freeman 1976a).

SEDIMENT PROCESSING

Individual batches of the mammal bed were slurried with water, either by boiling
for a short time or by soaking in cold water overnight, and were then sieved through
a 0-35 mm mesh. After cleaning the shelly residues by washing with aqueous detergent
solutions, they were dried and resieved into narrow size-range fractions of which the
coarser (>2-l mm) were hand sorted without further treatment. The remaining
fractions down to 0-5 mm were digested with aqueous acetic acid (15% by volume),
to leave residues consisting almost entirely of particles of limonite and vertebrate
fossils.

These residues were so concentrated (0-10% of the original sediment weight) that
further concentration was not necessary, the vertebrate fossils being easily isolated by
simple naked-eye searching under natural daylight.

However, purely as an experiment, two different techniques were applied to
separate batches of one of these acetic acid-insoluble residues in order to remove the
limonite, and thus concentrate the vertebrate fossils still further. The first batch was
density separated using bromoform (density 2-63-2-69 g ml-1 at 20 °C). The floating
fraction constituted 51% by weight of the original and still consisted of a mixture of
limonite and fossil bone, albeit enriched in the latter. Moreover, the movement of
the bone into the floating fraction was not complete, some remaining in the fraction
denser than bromoform. The second batch of acetic acid-insoluble residue was
treated with an aqueous solution of thioglycollic acid (5% by volume), which had
been pre-treated with powdered modern bone (see Rixon 1976, p. 1 12). The limonite
dissolved almost totally, the undissolved bony residue constituting 19% by weight
of the original. Thus treatment with thioglycollic acid was considerably more effec-
tive (enrichment factor 5-3) than was the bromoform flotation (enrichment factor
2-0) in the further enrichment of the vertebrate-containing residue.

INVERTEBRATE PALAEONTOLOGY, LITHOLOGY, AND
PALAEOENVIRONMENT

The Kirtlington Mammal Bed appears to contain assemblages of both indigenous
and derived fossils. The latter, characterized by their abraded condition and the
frequent presence of patches of adhering oolitic limestone matrix, comprise fragments
of oysters and such typically marine forms as compound corals, polyzoa, brachio-
pods, echinoids, and crinoids. It is evident that these derived fossils originated from
one or more of the marine limestones lower in the Great Oolite sequence, very
probably the White Limestone.

In contrast, the indigenous biota appears to be of non-marine origin. Apart from
the bulk of the vertebrate fauna (q.v.), the indigenous fossils are largely micro-
scopic, comprising abundant charophyte gyrogonites and ostracods. This ostracod

FREEMAN: JURASSIC MAMMALS

139

fauna has been examined in detail by Dr. Martin Ware (Ware 1978), who reports
that it contains, amongst others, the species Timiriasevia mackerrowi (abundant) and
Theriosynoecum kirtlingtonense, both of which were considered by Bate (1965) to
be freshwater forms ; as both species show the characteristic population age structure
expected of autochthonous species, they provide good evidence for the freshwater
origin of the mammal bed.

Small indeterminate plant fragments occur abundantly as sooty or limonitic
impressions in the mammal bed. A casual search for more complete plant remains,
and for insect fossils, has so far been unsuccessful.

The mammal bed contains an apparently unsorted mixture of clastic materials,
ranging in size from subangular pebbles of oolitic limestones down to abundant
comminuted shell debris, individual ooliths, and rare silica sand grains. Particle-size
analysis of the sediment shows it to be remarkably uniform in texture, both vertically
and horizontally, over the whole outcrop (see Table 1), and thus to be free of large-
scale current sorting. Apart from localized and minor concentrations of comminuted
shell and ooliths along individual bedding planes, no sedimentary structures have
been seen; in particular, there is no sign of cross-bedding.

The mammal bed contains iron in the + 3 oxidation state (as limonite), rather than
in the +2 oxidation state (as in iron pyrites), suggesting that deposition occurred
under oxidizing rather than reducing conditions.

Mr. Noel Shelton of GR-Stein Refractories Ltd. has investigated the mineralogy
of the <0-35 mm fraction of the sediment by X-ray diffraction (X.R.D.). He reports
that it consists largely of calcite, with a subordinate amount of quartz, findings which
were supported by both the loss-on-ignition results (34-23%), and qualitatively, by
acid digestion. No clay minerals were detected by X.R.D. My own analysis of the
<0-35 mm fraction by acid-base titrimetry gave a CaC03 content of about 66-68%
(by weight). Sugden and McKerrow (1962) discussed the composition of marls within

table 1. Particle-size analysis of the residues from the Kirtlington Mammal Bed, showing its uniformity
of texture both horizontally and vertically.

Source of sediment (
distance from Section A
(metres)

Weight of
sediment
(kg.)

>0-5in

Mesh
perct
0-5-
0-25 in.

analysi
sntages
0-25in-
2-1 mm.

s of wc
of initic
2-1-
1-7mm.

ishing
nl sedir
1-7-
1-4-mm.

residue
nent w<
1-4--
1-0mm.

>s as
sight
1-0-
0-5mm.

l |

in ro
o o

0 - 0-85

7-4

0 19

0-64

0-28

0-25

0-82

1-22

2-35

c.4-3

165-7

0-03

0-09

0-42

0-21

0-16

0-58

1-01

2-34

10-4-17-2

207-5

0-21

0-21

0-52

0-25

0-19

0-66

1 20

5-60

17-2-20-2

401-6

0-15

0-13

0-51

0-23

0-17

0-63

0-97

2-09

c.18, top half of bed

2-819

0-50

0-18

0-12

0-39

0-88

1-71

c.18, bottom half of bed

2-462

0-4 8

0-18

0-13

0-47

0-98

1-88

140

PALAEONTOLOGY, VOLUME 22

the Great Oolite Series of Oxfordshire, concluding that the substantial proportion of
clay in such sediments, by forming films around the crystals of calcium carbonate,
prevents its recrystallization and thus inhibits the induration of the sediment. The
non-detection of clay minerals in the Kirtlington Mammal Bed by X.R.D. is therefore
surprising, as the sediment displays all the properties mentioned by Sugden and
McKerrow as typical of marls. Thus it is friable when dry, becoming plastic when
wetted with water. In other words, in the apparent absence of clay minerals, it is not
clear what has prevented the induration of the sediment.

It seems probable that during a temporary marine regression, a shallow, non-
stagnant body of freshwater received periodic influxes of poorly sorted sediment
which consisted largely of calcite mud, and which was obtained locally by the erosion
of earlier Middle Jurassic oolitic limestones.

THE VERTEBRATE FAUNA OF THE MAMMAL BED

In order to prepare a vertebrate faunal list a sample of the mammal bed weighing
141-9 kg was processed by the standard method described earlier, and the residues
hand picked as quantitatively as possible down to a particle size of 0-5 mm. The
vertebrate fossils were sorted into their respective categories, and both complete
and incomplete specimens counted and weighed. In the context of this faunal list
(Table 2), and Tables 3 and 4, an ‘incomplete’ specimen is regarded as one whose
missing portion could reasonably be expected to be isolatable and identifiable.
Certain items in Table 2 do not appear to be even approximately representative of the
sediment as a whole ; such unrepresentative occurrences are indicated where appro-
priate. The remarks following apply to the vertebrate fauna in toto, and not just to
that in Table 2.

Virtually all of the vertebrate fauna appears to be truly indigenous to the mammal
bed, rather than being derived from the oolitic limestone clasts that occur so
abundantly in the deposit, a 78-5 g hand-picked sample of these clasts yielding only
thirteen minute fish scales by acid-digestion. Similarly, as care was taken whenever
possible to remove the soft limestones that tended to adhere to the top and bottom
surfaces of the mammal bed, contamination from these sources should also be
minimal.

Nearly all of the abundant crocodilian teeth are shed crowns, only three of the
468 complete teeth in the faunal list having roots. This, coupled with the absence of
wear other than that produced by use, suggests that the accumulation of the croco-
dilian teeth was largely a biocoenosis. They vary in shape from sub-conical to elon-
gated and strongly recurved, and range in crown height from 0-8 to 9-5 mm. A pair of
carinae usually extends to the apex of each tooth, the enamel of which is ornamented
with vertical striae ; this ornamentation varies markedly in its degree of prominence,
suggesting that more than one species is represented. The teeth bear little resemblance
to those of the Teleosaurus spp., marine crocodilia which have been found in the
Bathonian of Oxfordshire (Phillips 1871), but instead are similar to the teeth of the
small Late Jurassic freshwater goniopholid Nannosuchus. The status of this latter
animal is uncertain, being regarded by its original describer, Owen (1879), as a new
genus with adults dwarfed to match the size of the mammals upon which Owen

FREEMAN: JURASSIC MAMMALS

141

table 2. Vertebrate fossils from 141-9 kg of the Kirtlington Mammal Bed, by weight and number: ^0-5 mm.

TEETH

Number

Weight

%

by wt.

OTHER REMAINS

Number

Weight

MAMMALIA (s.U

5+141

REPTILIA

REPTILIA

Cetiosaurus (?) bone

14]

1 535g

Crocodilia

468+1208]

2-6046g

75-17

Crocodilian scutes

145]

1-40 61 g

Ornithischia.typeA

19+110]

0 2251 g

6-50

Crocodilian (?) vertebrae

19+119]

0-9392 g

Ornithischia.typeB

4+11]

0-0627g

1-81

Crocodilian (?) claws

11+13]

0-21 1 3 g

*

Theropoda

2 + 13]

0 • 01 92 g

0-55

Chelonian plates

2+189]

10-99g

Pterosauria

8

0 0145g

0-42

Lepidosaurian jaws

14 81

0-0804g

Incertae Sedis

5+122]

0-2536g

7-32

AMPHIBIA

ACTINOPTERYGH

4k

Anuran limb bones

13]

0-03 50 g

Lepidotidae +
Pycnodontidae

51 + I 8]

0-1 953g

5-64

ACTINOPTERYGH

?Caturus sd*

6

0-0034g

0-10

Scales

>1 97+[>59]

>0-71 09g

ELASMOBRANCHII

Vertebrae

6+11]

0-0259 g

Asteracanthus sd.

1

0-0824g

2-38

CLASS INCERTAE SEDIS

Hvbodus sd.

2

0-0042 g

0-12

Vertebrae

9

0-0729g

„ „
Lamnid

1

Long Bones

19+ [>228]

>0-711 7 g

TOTAL

3-4650g

100-01

Miscellaneous

>10-0g

Square brackets enclose the numbers of ‘incomplete’ specimens as defined in the text. An asterisk indicates
that the specimens concerned were present in an atypically high concentration in the particular batch of
sediment analysed. The fossils identified as Lepidosaurian jaws may also include those of small fish.

considered they preyed, while a later worker, Joffe (1967), suggested that Nanno-
suchus was merely an assemblage of juvenile specimens of Goniopholis simus. In an
earlier paper (Freeman 1975), I adopted a non-commital attitude towards this prob-
lem, but pointed out the seeming paradox of the remains of crocodilia in a Lower
Cretaceous lignite bed greatly outweighing those of their likely prey ; this imbalance
is again to be seen in the fauna of the Kirtlington Mammal Bed. The supposed
crocodilian vertebrae are amphicoelous, and are of the right order of size to be from
the same animals as the teeth, as are the dermal scutes and claws.

142

PALAEONTOLOGY, VOLUME 22

Of the ornithischian teeth, type A closely resembles those of the Early Liassic
?ankylosaur Scelidosaurus (Swinton 1973) and the Late Jurassic hypsilophodont
Echinodon (Owen 1861); they range in crown height from 0-8 to 5-2 mm. Type B are
of the same general size and shape as type A, but do not have cutting edges which are
crenulated, nor crowns which are vertically ribbed. It is not clear whether these
differences are of taxonomic significance or are merely the result of wear upon type A
teeth, although the former is suspected.

The theropod teeth, with only one exception, are smaller than those of the Middle
Jurassic carnosaur Megalosaurus bucklandi ranging in height from T6 to 7-4 mm.
They may be the teeth of either juvenile M. bucklandi or of coelurosaurs. The pos-
sible ecological significance of these theropod teeth is discussed later.

Three incomplete long bones are identified as the fused tibia/fibula or radius/ulna
of anurans. This skeletal modification is known in the Early Jurassic anuran Neo-
batrachus (see Romer 1966, fig. 145), and so the Kirtlington specimens add nothing
to the evolutionary history of the order. Their only significance is to suggest that
deposition of the mammal bed took place under non-marine conditions, as does the
greater part of the vertebrate fauna in general.

THE ORIGIN OF THE MAMMALIAN FOSSILS— 7COPROCOENOSIS

Mellett (1974) noted the similarity in preservation between the contents of recent
carnivore ‘scat’, and terrestrial microvertebrate fossil accumulations, and advanced
the hypothesis that the latter were composed predominately of the contents of the
faeces or regurgitata of predators. He proposed a term, ‘coprocoenosis’, to describe
such an accumulation. The observations described below support Mellett’s ideas and
suggest that they may be of value in the search for new Mesozoic mammal sites.

The Kirtlington Mammal Bed was of necessity processed in individual batches,
ranging in weight from 36T to 141-9 kg, each batch being collected from a relatively
small area of accessible outcrop. Unexpectedly, it was found that the similarity
between the mammal teeth within an individual batch tended to be greater than
between batches (see Table 3). This phenomenon, only weakly displayed at
Kirtlington, has been seen by the author in a more striking form elsewhere. Further-
more, the isolated mammal teeth frequently appear to have been hollowed out
from within, their pulp cavities being greatly enlarged. In some cases this process
has resulted in the destruction of the roots of the teeth, leaving only the hollowed-out
husks of the highly durable, enamel-coated, crowns. Most of the mammal teeth from
Kirtlington have incomplete crowns, the fractures being always sharp and fresh
looking. The absence of matching fragments of these broken teeth and the generally
good condition of the non-mammalian fossils suggest that most of the damage suf-
fered by the mammal teeth was not caused by the isolation procedures, but instead
pre-dates fossilization. Finally, one edentulous fragment of a mammalian dentary has
been found which bears on its external surface two circular depressed fractures, of
approximate diameters 0-9 and 0-4 mm, which appear to be the marks of a predator’s
teeth.

It therefore seems likely that the mammals of the Kirtlington Mammal Bed were
largely the victims of predators. The action of digestive juices in the predator’s

FREEMAN: JURASSIC MAMMALS

143

stomach would usually destroy all the non-compacted bone in the prey, leaving only
the relatively resistant teeth to be excreted. Even these would usually be broken and
partially digested. However, the enclosure of the teeth within a faecal pellet would
protect them from further damage, and provide them with a buoyant vehicle for
water transportation to an environment suitable for their fossilization. Burial and
subsequent oxidative decay of the faecal pellet would then liberate the teeth as
independent but spatially associated entities.

The similarity of the teeth within a given batch of sediment may therefore be the
result of original association within the same faecal pellet, and working backwards,
even within the jaws of the same individual animal. This suggests that careful batch-
wise processing of mammaliferous sediments might occasionally yield worthwhile
evidence of association between various parts of mammalian dentitions.

As to the identity of the predators responsible, I suggest that they were primarily
small theropods, either coelurosaurs or juvenile carnosaurs. At Kirtlington, teeth of
small theropods are rarer even than those of mammals and occur in the same sporadic
way in the sediment (see Table 3), suggesting that the same erratic mechanism
(?coprocoenosis) was responsible for the fossilization of both groups of animals.

table 3. Distribution of teeth of mammals, theropods, and ornithischians between individual batches of
the Kirtlington Mammal Bed, suggesting a possible coprolitic origin for the mammal teeth.

WEIGHT OF

44-6 kg.

141 -9 kg.

39-5 kg.

49-9 kg.

137-3 kg.

76-6 kg.

3 6-1 kg.

Palaeoxonodon ootiticus

Q°7

?Amphitherium sp.

Dryolestidae

Cvrtlatherium canei

Oh

Eupantotheria incert. sed.

^7 ^H7?

Wareolestes rex

Morganucodontidae inc.sed.

<$> ^7

©

Docodontidae

00

©

@@?

Triconodonta incert. sed.

©?

© © ©

Tritylodontidae

©

©

© ©

Mammalia (s.l.) incert. sed.

© m

A

©A

A A

cutAA

[771?

MAMMALIA, no. of teeth

[two]

f i ve + [four]

four

+[one]

one
♦ [four]

seven + [nine]

three + [six]

[four]

THEROPOOA , no. of teeth

one
+ [one]

two + [three]

none

two

none

two + [one]

none

ORNITHISCHIA, no. of teeth

ten
♦ [two]

twenty three + [eleven]

nine

+(one]

twel ve
+ [one]

twenty nin e + [six]

sixteen
+ [three]

nine

.[two]

Note the irregular distribution of the mammal and theropod teeth compared with that of the ornithischian
teeth. Note also the greater degree of fragmentation of the mammal teeth (square brackets enclose the
numbers of ‘incomplete’ specimens as defined in the text).

Symbols: upright semicircles = lower molars; inverted semicircles = upper molars; diamonds = pre-
molars; circles = cheek teeth of uncertain position; rectangles = incisors; triangles = tooth fragments of
uncertain position. The numerals enclosed within these symbols are the ‘FM/K’ numbers of the specimens.

144

PALAEONTOLOGY, VOLUME 22

This contrasts with the occurrence of the other important group of terrestrial verte-
brates, the ornithischians, whose teeth are present in a high and relatively constant
concentration (one tooth per 3-3 to 4-2 kg). Further afield, mammal fossils occur
with those of theropods not only at Kirtlington, but also in the Cliff End Bone Bed
(author), and in sediments from Swanage (Simpson 1928, p. 190), Watton Cliff,
Dorset (personal communication from David Ward), Woodeaton Quarry, Oxford-
shire (observations of author and David Ward), Stonesfield (Buckland 1824), and
Hanover Point in the Isle of Wight (personal communication from Richard Ford;
see Butler and Ford 1977). No other terrestrial groups are consistently associated
with one another in this way. With the exception of the Stonesfield Megalosaurus , the
teeth of the theropods are small and are thus compatible with a diet of small mammals.

Even irrespective of any possible predator/prey relationship, considered empiri-
cally, the presence of theropod teeth within a sediment would suggest that mammal
fossils are also present. Indeed, it was only the finding of a small theropod tooth by
Dr. Ware in November 1974 that prompted my search for mammal fossils at
Kirtlington.

Finally, as Mellett pointed out, his hypothesis causes problems in certain aspects of
vertebrate palaeoecology. In particular, coprocoenosis would be expected to produce
a heavy preservational bias in favour of small-size animals, which are more likely
to be eaten whole. Thus while the observed small size of Mesozoic mammals may well
be a genuine reflection of the ecological constraints placed upon them by the dino-
saurs (see, for example, Desmond 1975), it should be recognized that the actual data
of the fossil record may be unreliable in this regard. Secondly, even within the size
range of the animals whose preservation would be favoured by coprocoenosis, there
would be subtle factors at work that would cause certain species to be selectively
preserved at the expense of the others, for example, the taste preferences of the pre-
dator and the evasiveness of its prey. It is therefore a highly dangerous practice to
assume that a given Mesozoic mammal fossil assemblage even approximately repre-
sents the live fauna from which it was derived. In this regard the qualitative and
quantitative differences between the assemblages of mammal fossils from Stonesfield,
the Isle of Skye, Watton Cliff, Dorset, and Kirtlington are especially noteworthy,
as in these cases stratigraphic and palaeogeographic differences are minimal (see
Table 4).

THE MAMMALIAN FAUNA

Except for FM/K 43, all the specimens from Kirtlington described in this section are
the more important finds from the first 525-9 kg of the mammal bed to be processed,
the collection being summarized in Table 3. The specimen numbers which are pre-
fixed by ‘FM/K’ (Forest Marble/Alrtlington) are those used in the author’s collection ;
the corresponding numbers of the British Museum (Natural History) are given in
the Appendix. Other specimens with numbers prefixed by ‘M’, ‘R’, or ’B.M.(N.H.)’
are also in the national collection.

The Scanning Electron Micrographs used to illustrate this account were taken
after the specimens had been coated with aluminium, the use of which allows the
specimens to be cleaned afterwards by immersion in a dilute aqueous solution of a
weak alkali, such as ‘Decon’.

FREEMAN: JURASSIC MAMMALS

145

table 4. Assemblages of mammal fossils from four Middle Jurassic localities, showing the disparity in

faunal elements.

Stonesfield

Isle of Skye

Watton

Cliff

Kirtlington

CHEEK TEETH OF
Amphitheria

31 in 4 jaws

2

mu

7 + [21

Dryolestidae

1

2

Kuehneotheriidae

2 ♦ (11?

Amphilestidae

42 in 7 jaws

1

Morganucodontidae

1 +

[4]

3 + [31*111?

Docodontidae

9 in 2 jaws

1

2 + [5]

Triconodonta incert.sed.

(41

Multituberculata

5

Tritylodontidae

7 in 2jaws

4

[4]

CANINES, INCISORS and
INDETERMINATE

9 in 3 jaws

1 +

[6]

4 + |9]

Sources of data : Stonesfield, Simpson 1 928 ; Isle of Skye, Waldman and Savage 1 972 ; Watton Cliff, author’s
examination of David Ward’s collection from 90-5 kg of sediment (specimen numbers M 34984-M 35007
inclusive); Kirtlington, author’s collection from 525-9 kg of sediment. For convenience FM/K 31, an
incomplete possible eupantothere upper molar, has been omitted from the table, as has FM/K 43, an
incisor of either a multituberculate or a tritylodontid. Square brackets enclose the numbers of ‘incomplete’
specimens as defined in the text.

Terminology relating to dental morphology is generally as explained in Patterson
1956, except where otherwise indicated. The term ‘buccal’ is synonymous with
‘external’, ‘lingual’ is synonymous with ‘internal’, while for the cheek teeth ‘mesial’
approximates to ‘anterior’ and ‘distal’ to ‘posterior’.

The classification of Mesozoic mammals is at present an area of debate, and even
controversy. In this account, unless otherwise stated, the formal taxonomy at the
subclass level and lower follows the schemes outlined in Kermack, Kermack, and
Mussett (1968) and in Kermack, Mussett, and Rigney (1973). This is done partly for
reasons of internal consistency, and partly to facilitate comparison of the present
account with the particularly relevant work of Clemens and Mills (1971). No value
judgement is intended, and it may well be that the forms described herein will finally
be classified in a manner other than that currently employed.

For the taxonomy at the level of the class, I favour the opinion of Bakker (1975),
who groups together the Therapsida and the Mammalia (as understood to consist of
the subclasses Theria and Atheria). However, unlike Bakker, I consider that the
simple enlargement of the pre-existing class Mammalia achieves this purpose quite
satisfactorily, and that a new class name, Bakker’s Theropsida, is unnecessary. There

146

PALAEONTOLOGY, VOLUME 22

is an increasing body of evidence to suggest that at least certain members of the
Therapsida possessed what in a living mammal would be considered highly diagnostic
characteristics, namely an endothermic metabolism and an insulating layer of hair
(see Ager 1977). Inclusion of the Therapsida in the Mammalia would also collect
together what is essentially a unified evolutionary complex, instead of dividing it
between the classes Reptilia and Mammalia, which is inevitable if an artificially
rigid definition of the Mammalia is attempted by means of osteological (and especially
dental) characters alone.

SYSTEMATIC DESCRIPTIONS

Class mammalia Linnaeus
Subclass theria Parker and Haswell
Infraclass pantotheria Simpson
Order eupantotheria Kermack and Mussett
Suborder amphitheria Kermack, Kermack, and Mussett
Family kuehneotheriidae Kermack, Kermack, and Mussett
Genus cyrtlatherium gen. nov.

Type and only known species. Cyrtlatherium canei sp. nov.

Derivation of name. From Cyrtla, an Anglo-Saxon proper name and the root of Kirtlington, and therios,
Greek for ‘wild beast’.

Diagnosis. Lower molars with a recurved protoconid, more gracile than in Kuehneo-
therium praecursoris Kermack, Kermack, and Mussett. Lingual cingulum strongly
curved upwards underneath the protoconid.

Cyrtlatherium canei sp. nov.

Plate 15, figs. 2-5

1976 FM/K 38; Freeman, pp. 1053-1054, fig. 2c.

Derivation of trivial name. To honour Derek J. Cane, who contributed so much to this exercise, and who
found the holotype.

Holotype. FM/K 11, Kirtlington Mammal Bed, Upper Bathonian, Kirtlington.

Diagnosis. As for genus.

EXPLANATION OF PLATE 15

Fig. 1 . Newly excavated section in the Old Cement Works Quarry, Kirtlington, showing part of the outcrop
of the Kirtlington Mammal Bed (Bed 3p), June 1976. The photograph is centred on Section B (see
text-fig. 1), and corresponds roughly to the upper part of plate 9 in McKerrow et al. 1969. The horizontal
distance across the photograph is c. 7-5 m, the mammal bed extending for another c. 10 m to the left
(NW.) and for another c. 4 m to the right (SE.).

Figs. 2-5. Cyrtlatherium canei gen. et sp. nov. 2, 3, 4, holotype (FM/K 11), lower molar from buccal ( x 44),
occlusal ( • 57), and lingual ( x 52) aspects. 5, lower molar FM/K 38 from lingual aspect ( x 40). Figs. 2-4
stereophotographs.

PLATE 15

FREEMAN, Jurassic mammals

48

PALAEONTOLOGY, VOLUME 22

Description. This taxon is based on two right lower molars. The holotype (FM/K 1 1) is complete except for
the lower parts of its roots. Its protoconid is much higher than the metaconid and paraconid, and is recurved
near its tip. In contrast to the specimens illustrated for K. praecursoris (Kermack et al. 1968), the paraconid
is in line with the protoconid, and is not displaced lingually to any extent. A marked crest runs along the
mesiolingual face of the protoconid from the tip of the protoconid to the tip of the paraconid ; this does not
appear to be an artefact of wear, but rather is an original feature of the tooth. In general, the tooth appears
not to be worn to any significant degree. The metaconid is displaced to the lingual side of the protoconid,
indeed to quite a considerable extent; as a consequence the angle of the trigonid is approximately 140°.
A cingulum extends along the whole length of the lingual face, curving upwards in a manner reminiscent of
the amphilestids. It extends around the ends of the tooth, to fade away at the mesial and distal ends of the
buccal face. An accessory cusp occurs on the cingulum at its distal extremity (the ‘hypoconulid’ of Kermack
et al. 1968), but not at its mesial extremity.

The other tooth (FM/K 38) lacks its paraconid and mesial root through breakage. As in the holotype, a
prominent crest runs from the tip of the protoconid down its mesiolingual face. In general, the cingulum
resembles that of the holotype, but the distal accessory cusp is a more prominent feature. The distal root is
complete, tapers gently towards its apex, and curves in a mesial direction.

Finally, a badly worn and incomplete tooth (FM/K 27) is assigned with reservations to C. canei. It
adds nothing significant to the knowledge of the taxon.

Dimensions (in mm) FM/K 1 1

Length of crown 0-83

Maximum width of crown 0-37

Height of protoconid (from gum line) 0-57

FM/K 38
>0-73
0-37
0-56

FM/K 27
>0-83
0-34
>0-54

Comments. While, for reasons previously explained, I have formally classified
Cyrtlatherium and the other kuehneotheriids within Kermack et al.' s Amphitheria,
and not within the Symmetrodonta, I do in fact favour the latter placement. This
reflects my opinion that Kuehneotherium is not part of a group that is ancestral to
both of the orders Symmetrodonta and Pantotheria (in the sense of Patterson 1956).
In the original description of Kuehneotherium, Kermack et al. (1968) attached great
importance to the lingual cingulum of the upper molars, which they considered to be
the precursor of the protocone; as thus interpreted, Kuehneotherium shows that the
evolution of the protocone was underway in Rhaeto-Liassic times. However, this
conclusion is not supported by the later therians Palaeoxonodon and Peramus,
neither of which have even rudimentary protocones in their supposed upper molars.
Kuehneotherium is also disqualified as the ancestor of the order Pantotheria (sensu
Patterson) by the pronounced lingual cingula of its lower molars, the like of which
are not seen in the early therians outside of the Symmetrodonta.

Cyrtlatherium is not the first symmetrodont to be described from the Middle
Jurassic, being anticipated for over a century by the amphilestids from Stonesfield.
However, the latter are not conventionally recognized as such, and are usually
regarded as early triconodonts. Only recently has this assignment been questioned

EXPLANATION OF PLATE 16

Figs. 1-2. Amphilestes broderipii (Owen), lower molar from the Forest Marble of Watton Cliff, Dorset
(M 35000), from lingual ( x 26) and occlusal ( x 30) aspects.

Figs. 3-7. Palaeoxonodon ooliticus Freeman. Holotype (FM/K 8), lower molar. 3, mesial aspect ( x 37).
4, lingual aspect ( x 44). 5, oblique lingual view of the talonid ( x c. 72). 6, occlusal aspect ( x 56).
7, apical aspect ( x 56).

Figs. 1-7 stereophotographs.

PLATE 16

FREEMAN, Jurassic mammals

150

PALAEONTOLOGY, VOLUME 22

and an excellent set of arguments put forward for placing the amphilestids with
the symmetrodonts (Mills 1971). Through the kindness of David Ward, I have
had the opportunity to examine a lower molar of Amphilestes broderipii from the
Forest Marble of Watton Cliff, Dorset. The specimen (M 35000) is the first material
of Amphilestes to be seen free of matrix and as thus seen the resemblance to the
lower molars of the typical symmetrodonts is compelling (see PI. 16, figs. 1, 2). In
particular, the concave lingual face and the lingual displacement of the accessory
cusps (now the paraconid and metaconid) are well shown.

After revision to accommodate the above changes, the classification of the infra-
class Pantotheria becomes :

Infraclass Pantotheria

Order Eupantotheria Kermack and Mussett
Suborder Amphitheria Kermack, Kermack, and Mussett
Family Amphitheriidae Owen
Family Paurodontidae Marsh
Family Peramuridae Kretzoi
Suborder Dryolestoidea Butler
Family Dryolestidae Marsh
Suborder Symmetrodonta Simpson
Family Amphidontidae Simpson
Family Amphilestidae Kiihne

Family Kuehneotheriidae Kermack, Kermack, and Mussett
Family Spalacotheriidae Marsh

The removal of the Amphilestidae from the subclass Atheria leaves the suborder
Eutriconodonta Kermack, Mussett, and Rigney with only one family, the Tricono-
dontidae Marsh.

Family (?)peramuridae Kretzoi
Genus palaeoxonodon Freeman, 1976

Type and only known species. Palaeoxonodon ooliticus Freeman, 1976.

Derivation of name. From palaeos, Greek for ‘ancient’, oxonia, the Latinized name of Oxfordshire, and
odons, Greek for ‘tooth’.

Palaeoxonodon ooliticus Freeman, 1976

Plate 16, figs. 3-7, Plate 17, figs. 1-4, 8, Plate 18
1976 Palaeoxonodon ooliticus Freeman, pp. 1053-5, figs. la-c.
Derivation of trivial name. From the Great Oolite Series, Middle Jurassic.

EXPLANATION OF PLATE 17

Figs. 1-4. Palaeoxonodon ooliticus Freeman. Lower molar FM/K 7. 1, lingual aspect (x49). 2, oblique
distal aspect ( x 51). 3, occlusal aspect ( x 52). 4, mesial aspect ( x 48).

Figs. 5-7. lAmphitherium sp. Incomplete lower molar FM/K 16. 5, lingual aspect (x25). 6, oblique
distal aspect ( x 25). 7, oblique mesial aspect ( x c. 25).

Fig. 8. ? Palaeoxonodon ooliticus Freeman. Lower premolar FM/K 9, from buccal aspect ( x 35).

Figs. 2-4, 7, stereophotographs.

PLATE 17

FREEMAN, Jurassic mammals

152

PALAEONTOLOGY, VOLUME 22

Description. The holotype (FM/K 8) is a right lower molar, complete except for its paraconid and the lower
extremities of both roots, and the buccal side of the mesial root. The protoconid is high, sharply acute, and
has a recurved mesial edge. Its lingual face is roughly planar, giving the cusp an approximately semicircular
cross-section. The metaconid is roughly two-thirds the height of the protoconid, and appears to have been
quite sharply separated from the now-missing paraconid ; no cingulum joins the bases of these two cusps.
The mesial face of the protoconid bears only a poorly developed anterobuccal cingulum (see PI. 16, fig. 3),
which contrasts with the condition seen in Amphitherium and Peramus, and in the dryolestid teeth described
later in this account.

Most of the interest of the holotype lies in its talonid, which appears to be roughly intermediate in form
between those seen in Amphitherium prevostii (Middle Jurassic) and Peramus tenuirostris (Late Jurassic),
and foreshadows those of tribosphenic dentitions. As in Amphitherium, the talonid of FM/K 8 consists
primarily of a single cusp (the hypoconid?) situated on the distobuccal end of a crista obliqua that extends
from the distal edge of the metaconid and which forms a continuation of the crest on the metaconid. How-
ever, unlike Amphitherium, the crista obliqua is raised into a small cusp at its approximate median point,
in which feature it resembles certain specimens of Peramus, ‘Trinity molar type 6’ (Slaughter 1971), and
certain Tertiary Insectivores (see Scraeva and Arvaldus in Cray 1973). Following Mills (1964), this median
cusp is called the ‘posterior accessory cusp’ in this account. From the ?hypoconid a low ridge runs in an arc
mesiolingually towards the base of the metaconid. This low ridge and the crista obliqua together encompass
an area which is concave and includes the lingual face of the ?hypoconid ; this would seem to be an initial
stage in the development of the talonid basin which was to become so important in the later stages of
mammalian dental evolution. It is possible that the concavity of the ?hypoconid in Palaeoxonodon assisted
a piercing function of the cusp, after the fashion of a fuller on a bayonet or stabbing knife. The low ridge
that forms the lingual rim of the incipient talonid basin of FM/K 8 is expanded in two places to form what
from their positions appear to be incipient developments of an entoconid and a hypoconulid. Although these
incipient cusps are barely perceptible, they are sufficiently elevated above the ridge from which they arise
to have been preferentially abraded (see PI. 16, fig. 6). Their positions seem to indicate that the first of the
talonid cusps to evolve, i.e. the one seen in Amphitherium, was the hypoconid (for a summary of the argu-
ments on this topic, see Slaughter 1971, pp. 138-140).

Another right lower molar (FM/K 7) is similar to the holotype but has lower trigonid and talonid cusps.
In the case of FM/K 7 there is no anterobuccal cingulum at all, and the talonid comprises only one cusp
(the ?hypoconid), the other two talonid cusps not being seen even under the Scanning Electron Microscope
(see PI. 17, figs. 2 and 3). Also, there is no development at all of a talonid basin, the lingual face of the
?hypoconid being entirely convex. As in the holotype, there is a ‘posterior accessory cusp’ upon the crista
obliqua.

A premolar (FM/K 9) has a single recurved cusp with a buccal cingulum that ends distally in a small heel.
It has two roots which are separated only near their lower extremities (see PI. 17, fig. 8). The specimen is
doubtfully identified as a left lower premolar of P. ooliticus, primarily because of its size and its general
resemblance to the lower premolars of Amphitherium and Peramus.

In the light of the remarks made earlier regarding the origins of the mammal fossils, it may be significant
that FM/K 7, FM/K 9, and the holotype all came from the same 39-5 kg batch of sediment.

Dimensions (in mm)

FM/K 8

FM/K 7

FM/K 9

Length of crown

>0-87

>0-90

0-51

Length of trigonid

>0-59

>0-63

—

Maximum width of crown

0-49

0-54

Height of protoconid (from gum line)

0-89

0-80

0 61

EXPLANATION OF PLATE 18

Figs. 1-5. Palaeoxonodon ooliticus Freeman. Referred upper molar FM/K 12. 1, occlusal aspect ( x c. 52).
2, mesial aspect ( x 28). 3, distal aspect ( x 38). Referred hindmost upper molar FM/K 4. 4, oblique
distal aspect ( x 50). 5, occlusal aspect ( x 50).

Figs. 1-3, 5, stereophotographs.

PLATE 18

FREEMAN, Jurassic mammals

154

PALAEONTOLOGY, VOLUME 22

Three eupantothere upper molars are also referred to Palaeoxonodon ooliticus, largely on account of
their size, and at the risk of engaging in a circular argument, because they are generally similar to those
presumed, on better evidence, to be of Peramus tenuirostris (see Clemens and Mills 1971).

Specimen FM/K 12 is a superbly preserved left upper molar, complete except for its two buccal roots.
The paracone is the highest cusp of the crown, and forms its most lingual feature. Consider firstly the distal
edge of the tooth, apart from the paracone it bears in all four cusps. The metacone is immediately disto-
lingual to the paracone, from which it is only incompletely separated. From the gum-line the heights of the
paracone and metacone are in the approximate ratio 2 : 1 . A metacrista runs distobuccally from the meta-
cone, and bears three small, incompletely separated cusps, of which the one adjacent to the metacone is
the largest ; in position it corresponds with the virtually obsolete cusp labelled ‘cusp c’ in the figure of the
Peramus upper molar in Crompton (1971). As to the mesial edge of the tooth, the paracrista runs mesio-
buccally from the paracone to end at the buccal edge of the tooth in a prominent stylocone. In FM/K 12
the stylocone is distinctly mesial to the position of the stylocone in the dryolestids, where it forms a pro-
minent feature near the middle of the buccal edge of the tooth (the centroexternal cusp of Simpson 1929).
As in the case of the ‘cusp c’, the stylocone of FM/K 12 is more developed than in Peramus , and is com-
parable to the corresponding feature in Pappotherium pattersoni (Middle Cretaceous) (see Slaughter 1971).
However, unlike Pappotherium, there is no protocone on the lingual side of the paracone ; indeed, there is
not even a lingual cingulum that could be considered to be a forerunner of the protocone.

FM/K 30 is a much less complete left upper molar, as it lacks the buccal side of the tooth, i.e. the stylo-
cone, parastyle, and the end of the metacrista. What remains of the distal edge of the tooth bears two cusps
apart from the paracone ; as in FM/K 12 the metacone is distobuccally situated with respect to the paracone,
but is more distinctly separated from it. The ‘cusp c’ is also a more sharply defined feature than in FM/K 12,
and in general what remains of the tooth resembles the unidentified specimen FM/K 32 (see PI. 19, figs. 1-3).

The third of the upper molars referred to Palaeoxonodon ooliticus (FM/K 4) differs from the other two in
its distal edge, which is markedly shorter than the mesial and bears only two cusps, the paracone and the
metacone. In spite of the damage (apparently post mortem) that it has suffered, the short metacrista does not
appear to have borne a ‘cusp c’ or other cusps. As in FM/K 12, the paracone and metacone are only par-
tially separated from one another. The form of the tooth suggests that it is a hindmost left upper molar.

Dimensions (in mm)

Length of crown
Width of crown

Height of paracone (from gum line)

FM/K 12
0-95
0-96
0-80

FM/K 30
>0-75
>0-85
0-82

FM/K 4
>0-84
0-99
c. 0-6

Family ?amphitheriidae Owen
1 Amphitherium sp.

Plate 17, figs. 5-7

Description. A eupantothere left lower molar (FM/K 16), minus its talonid, is distinguished from P. ooliticus
by its larger size, the slightly convex lingual face of the protoconid, and by its prominent anterobuccal cusp.
From the size of the scar left by the removal of the talonid, it must have been a sizeable structure, the
tooth thus differing from the teeth of the dryolestids described later. In general what remains of the tooth
resembles the molars of A. prevostii, but in the absence of the talonid a positive identification, even at the
level of the family, is not possible.

EXPLANATION OF PLATE 19

Figs. 1-6. Eupantothere upper molars. Unworn specimen FM/K 32. 1, oblique occlusal aspect (x34).
2, oblique mesial aspect ( < c. 38). 3, distal aspect ( x 37). Worn specimen FM/K 26. 4, mesiobuccal
aspect (x 47). 5, distal aspect ( x 47). 6, oblique occlusal aspect ( x 38).

Figs. 1, 3, 6, stereophotographs.

PLATE 19

FREEMAN, Jurassic mammals

156

PALAEONTOLOGY, VOLUME 22

Dimensions :

Length of trigonid 0-79 mm.

Maximum width of trigonid 0-75 mm.

Height of protoconid (from gum line) 1 -02 mm.

Compare with A. prevostii B.M.(N.H.) 36822, from the Stonesfield Slate (dimensions in mm).

m5

m4

m3

m2

M,

P4

Length of crown

M0

1 69

106

Ml

104

0-85

Length of trigonid

0-81

0-89

0-86

0-81

0-73

—

Maximum width of trigonid

0-73

0-69

0-65

0 61

0-65

—

Height of protoconid (from gum line)

102

108

1-08

1-08

0-98

—

Suborder amphitheria Kermack, Kermack, and Mussett
Family incertae sedis

Plate 19, figs. 1-6

1976 FM/K 32; Freeman, p. 1054, figs. 2a, b.

Description. Two right upper molars have been found which are significantly larger than the three assigned
to P. ooliticus, although they are otherwise generally similar to them. They are roughly compatible in size
with the lower molar described above as ‘lAmphitherium sp., although any such identification in this case
must be little more than a guess.

Specimen FM/K 32 is a complete, virtually unworn, crown supported on the remnants of three roots. As
in the upper molars referred to P. ooliticus, the stylocone is a prominent cusp, situated on the buccal edge
of the tooth in a position distinctly mesial to the line that bisects the tooth through the paracone. Once
again the paracone is the largest of the cusps, and forms the most lingual part of the tooth. The metacone
is quite distinct from the paracone, and is situated distobuccally with respect to it, more so than in the case
of FM/K 12. In addition, the metacrista that runs distobuccally from the apex of the metacone bears only
two cusps, both of which are quite separate from one another, (see PI. 19, figs. 1-3 and text-fig. 2d).

In contrast, specimen FM/K 26 has been subjected to extensive wear, apparently before death (see
PI. 19, figs. 4-6, and text-fig. 2b). This has resulted in the truncation of the paracone (Facet 4), and the
removal of the metacone and the other cusps on the distal edge of the crown to leave an elongated concavity
in their stead (Facet 6+7). On the mesial side of the crown the stylocone has largely been removed by an
extensive wear facet (2) that runs obliquely to the mesiobuccal extremity of the parastyle, little of which
now remains. The mesial crest of the paracone has also been worn away to leave an elongated concavity
(Facet 2'). The ornament and numbering of the wear facets in text-fig. 2b follows that used in the figure of
the Eury lambda upper molar in Crompton and Jenkins 1967, where the pattern of wear is markedly similar.
Crompton and Jenkins assign their facets 2 to wear against the apex of a metaconid of an opposing lower
molar, while facets 4, 6, and 7 were produced by occlusion against, respectively, the mesial cingulum cusp,
the paraconid, and the protoconid of the lower molar situated behind that responsible for facets 2. How-
ever, facets 3 and 5 on the mesiolingual and distolingual faces of the paracone of Eurylambda are not
represented at all in FM/K 26; from Crompton and Jenkins’s figure these two facets were produced by,
respectively, the distobuccal and mesiobuccal surfaces of the two adjacent lower molars that abutted the
upper molar.

Nor is FM/K 26 alone in failing to display wear facets on the lingual side of the paracone. The minimal
amount of wear on FM/K 32 is fully in accord with that seen on FM/K 26, and if extrapolated would
produce a tooth of identical appearance. As other workers have reported or predicted wear facets on the
lingual side of the paracone, and in particular between the metacone and paracone (see, for example,
Mills 1964, Clemens and Mills 1971, p. 105, and Crompton 1971), their manifest absence in FM/K 26
should be noted.

Dimensions (in mm) FM/K 32 FM/K 26

Length of crown Ml 1-02

Width of crown 1-20 M0

Height of paracone (from gum line) 106 >0-83

FREEMAN: JURASSIC MAMMALS

157

text-fig. 2. Eupantothere upper molars from occlusal aspect ; unworn (fig. 2a,
FM/K 32) and worn (fig. 2b, FM/K 26). Abbreviations: pa., paracone; me.,
metacone; c, cusp ‘c’; sty., stylocone. The scale bar represents 0-5 mm.

Suborder dryolestoidea Butler
Family dryolestidae Marsh

Plate 20, figs. 1-8

1976 FM/K 29; Freeman, p. 1053, fig. 2d-f.

Description. This family is represented in the collection from Kirtlington by two incomplete lower molars.

Specimen FM/K 29 is a left lower molar which lacks its talonid and the greater part of its roots. As is
typical of the dryolestids, its trigonid is mesiodistally compressed compared to Palaeoxonodon, the ratio
H/L' of the height of the protoconid (H) to the length of the trigonid (L') being T 70-2-38 for FM/K 29
(the uncertainty is caused by doubts as to the exact position of the base of the protoconid), but less than
T51 and T27 for the two lower molars of Palaeoxonodon. Although the talonid has been removed by
breakage, the scar that it has left indicates that it must have been a small structure situated low on the
distal flank of the metaconid. Again in contrast to Palaeoxonodon, the crista obliqua does not extend to the
tip of the metaconid. A prominent and distinctly pointed cuspule occurs on the mesial face of the proto-
conid. The paraconid and metaconid are of comparable size, and are distinctly smaller than the protoconid.
The roots appear to have been much closer together than in Palaeoxonodon , the mesial root having an
elliptical cross-section, not circular as in the holotype of P. ooliticus.

Specimen FM/K 14, a right lower molar, is even less complete than FM/K 29, as it lacks its talonid and
roots, and much of its metaconid. In general it is similar to FM/K 29 in the proportions and dispositions
of its cusps, but differs in the joining of the bases of the paraconid and metaconid by a distinct cingulum.
As before, the ratio H/L' is greater (1-93) than in the material of Palaeoxonodon.

Dimensions (in mm)

Length of trigonid

Maximum width of trigonid

Height of protoconid (from gum line)

FM/K 29
0-56
0-58

095-T33

FM/K 14
0-61
0-55
118

Order (?) eupantotheria Kermack and Mussett

Specimen FM/K 31 is incomplete, consisting of two well-separated and acute cusps, supported on a
single stout root. The fossil may be a eupantothere left upper molar of a type unlike those previously
described, i.e. with a metacone lying directly distal to the paracone, instead of mesiodistally to it. Once
again, there is no lingual cingulum.

Dimensions : Height of ?paracone from gum line 0-67 mm.

Height of ?metacone from gum line 0-58 mm.

158

PALAEONTOLOGY, VOLUME 22

Subclass atheria Kermack, Mussett, and Rigney
Order triconodonta Osborn

Suborder morganucodonta Kermack, Mussett, and Rigney
Family morganucodontidae Kuhne
Genus wareolestes gen. nov.

Type and only known species. Wareolestes rex sp. nov.

Derivation of name. After Dr. Martin Ware, in recognition of his major contribution to the success of my
work at Kirtlington ; and lestes, Greek for ‘brigand’, alluding to the presumed carnivorous nature of the
animal.

Diagnosis. Lower molars with kiihnecone directly lingual to the main cusp, and much
smaller than the main cusp. Lower molars with a poorly defined buccal cingulum.

Wareolestes rex sp. nov.

Plate 21, figs. 1, 2

1976 FM/K 25; Freeman, pp. 1053-4, fig. 2g.

Derivation of trivial name. From the Latin for ‘king’, an allusion to the relatively large size of the animal
and also a pun on the name of Mr. E. J. King.

Holotype. FM/K 25, Kirtlington Mammal Bed, Upper Bathonian, Kirtlington.

Diagnosis. As for genus. Upper Bathonian.

Description. The holotype and only known specimen is a well-preserved nearly complete lower molar. It
is substantially larger and more inflated in shape than the lower molars of the triconodont from the Welsh
Rhaeto-Liassic fissures (Kuhne 1949; Parrington 1967). However, in most other respects, especially in the
distribution and relative proportions of its principal cusps, the holotype of the new species is similar to the
Welsh form (see text-fig. 3). In particular, both have a high central cusp (a, using the nomenclature of

a

text-fig. 3. Morganucodontid lower molars from lingual aspect. Fig. 3a,
holotype of Wareolestes rex, FM/K 25. Fig. 3b, Welsh Rhaeto-Lias tricono-
dont, M 16536. The scale bar represents 0-5 mm.

EXPLANATION OF PLATE 20

Figs. 1-8. Dryolestid lower molars. Specimen FM/K 29. 1 , lingual aspect ( x 50). 2, distal aspect ( x 45).

3, mesial aspect (x45). 4, apical aspect (xc. 75). 5, occlusal aspect (xc. 75). Specimen FM/K 14.
6, lingual aspect ( x 48). 7, distal aspect ( x 45). 8, mesial aspect ( x 45).

Figs. 1-3, 6, stereophotographs.

PLATE 20

6

FREEMAN, Jurassic mammals

160

PALAEONTOLOGY, VOLUME 22

Crompton and Jenkins 1968), in line mesially and distally with two smaller cusps (b and c respectively),
and an expanded lingual cingulum on which are developed a series of small cusps. In FM/K 25 there are
three of these cingulum cusps preserved ; they have the same relative sizes as in the Welsh triconodont,
with the present central cusp being the highest. Wear or damage may have removed additional cusps origin-
ally present on the lingual cingulum mesiolingual to the main cusp. In contrast to the Welsh triconodont,
the highest cingulum cusp (the kiihnecone or cusp g) of FM/K 25 is displaced mesially to a position directly
lingual to the main cusp (a), instead of being lingual to the valley between cusp a and the distal cusp c.
The lingual cingulum curves around the distal edge of the tooth where a substantial cusp (d) arises from it,
directly distal to the distal cusp c. The cingulum then continues in a subdued, non-cuspidate, form around
the buccal face of the crown, in which feature it contrasts with the lower molars of the Welsh triconodont.

The lingual face of the main cusp (a) is divided into four shallow embayments, giving the lingual face a
scalloped outline from above. The most distal of these embayments is more concave than the others, and
results in the distal edge of cusp a having a sharp concavo-convex cross-section.

The crown is supported by the remains of two stout roots, both sharply broken across at their point of
separation.

Dimensions: Length of tooth 2-31 mm.

Width of crown 1 -24 mm.

Height of principal cusp from lingual cingulum c. 1 -2 mm.

Comments. The occurrence of a typical morganucodontid in the Upper Bathonian is a
substantial and unexpected upwards extension of the known stratigraphic range of
the family, which hitherto has only been known from the Rhaeto-Lias. It indicates
that the morganucodontids were a stable and distinctive group for at least thirty
million years.

Wareolestes shows no features intermediate between the Welsh triconodont and
the triconodontids of the Late Jurassic. In particular, FM/K 25 shows no sign of the
equalization of the principal cusps and the elimination of the cingulum cusps seen
in the lower molars of the triconodontids.

Suborder docodonta Kretzoi
Family docodontidae Simpson
1 Borealestes sp.

Plate 21, figs. 3-5

1976 FM/K 24; Freeman, p. 1054, fig. 2h.

Description. Only one of the docodont teeth from Kirtlington is sufficiently complete to merit description
in this account. It is with great diffidence ascribed to Borealestes, which is the only known Middle Jurassic
docodont (Waldman and Savage 1972). Its size is of the same order as the lower molars of Borealestes
serendipitus.

EXPLANATION OF PLATE 21

Figs. 1, 2. Wareolestes rex gen. et sp. nov. Holotype (FM/K 25), lower molar. 1, lingual aspect (x 19).
2, buccal aspect ( x 19).

Figs. 3-5. Docodontid upper molar, FM/K 24. 3, occlusal aspect ( x 29). 4, linguo-occlusal aspect ( X 28).
5, mesial aspect ( x 26).

Figs. 6-8. Multituberculate or tritylodontid incisor, FM/K 43. 6, occlusal aspect ( x 21). 7, buccal aspect
(x 11). 8, apical aspect (x 21).

Fig. 9. Cusp of a tritylodontid cheek tooth, FM/K 3; oblique lateral aspect ( x 12).

Figs. 1, 3, 4, 9, stereophotographs.

PLATE 21

FREEMAN, Jurassic mammals

162

PALAEONTOLOGY, VOLUME 22

By analogy with Haldanodon from the Kimmeridgian of Portugal (Kiihne 1968), the specimen FM/K 24
is a left upper molar. It bears four well-marked cusps on a roughly trapezoidal crown.

The mesiobuccal cusp is the highest, has a curved buccal face, and a lingual face that is divided into two,
distinct, essentially flat ‘facets’. The upper molars of Haldanodon also has two facets similarly positioned,
which Hopson and Crompton (1969) attributed to wear against the lower molars. However, in FM/K 24
at least, these ‘facets’ do not appear to be artefacts of wear. If they had been, almost certainly the dentine
would have been exposed at the centre of each ‘facet’, assuming, of course, that the cusp was initially conical.
Furthermore, the three crests that separate the two ‘facets’ from one another and from the buccal face run
from the apex of the cusp to well-defined features of the tooth which could not have been produced by
wear, namely two of the other cusps and a salient on the mesial cingulum.

The distobuccal cusp also has a curved buccal face and a lingual face sharply divided into essentially flat
facets, again two in number. It is connected to the mesiobuccal cusp by a sharply angulated crest which
bears a pair of quite unmistakable wear facets ; these do expose the dentine, one facet being on the distal
flank of the mesiobuccal cusp, the other on the mesial flank of the distobuccal cusp.

The crest that separates the two lingual ‘facets’ of the mesiobuccal cusp from one another also serves to
join this cusp to the mesiolingual cusp. The latter is a relatively small feature, which only displays the
‘facetted’ appearance of the buccal cusps in a subdued fashion.

Finally, the distolingual cusp is the smallest of the four, and has only rounded surfaces.

A pronounced cingulum occurs along the buccal edge of the tooth, and extends as a distinct structure
around the mesial and distal edges of the tooth ; thereafter it merges into the complex of the lingual cusps.
As would be expected if the docodonts arose from morganucodontid ancestors, the lingual cusps appear to
be greatly expanded cingulum cusps (see Mills 1971, p. 56).

Of the three roots originally present, only that on the lingual side of the tooth has survived. It is roughly
cylindrical and of constant diameter except near its apex, where it flares outwards slightly and irregularly.

In contrast to the upper molars of the Late Jurassic docodonts Haldanodon and Docodon, the mesial and
distal edges of FM/K 24 are not deeply indented, although they do curve inwards to some extent. According
to Mills (1971, p. 56), the upper molars of Docodon (Latest Jurassic) evolved from the type seen in Haldano-
don (Kimmeridgian) by the addition of a fourth, distolingual cusp to the tooth. However, the presence of a
small, but distinct distolingual cusp in FM/K 24 shows that this feature was already in existence by the
Bathonian, and suggests that it appeared in the earliest phases of docodont evolution.

Dimensions: Maximum length of crown (across buccal cusps) 1-53 mm.

Length of crown across lingual cusps 1 -05 mm.

Maximum width of crown (across mesial cusps) 1-58 mm.

Height of mesiobuccal cusp (from top of cingulum) 0-62 mm.

Height of distobuccal cusp (from top of cingulum) 0-31 mm.

Class mammalia Linnaeus (or mammalia s.l.)

Order multituberculata Cope (or therapsida Broom)

Family ?paulchoffatiidae Hahn (or ?tritylodontidae Cope)

Plate 21, figs. 6-8

Description. An incomplete incisor (FM/K 43), found by Dr. Martin Ware, is somewhat similar to those
figured for the multituberculate family Paulchoffatiidae (Hahn 1969, fig. 44) and for the tritylodontid
Oligokyphus (Kiihne 1956, text-figs. 33, 34), except that its crown has a roughly triangular cross-section,
not an elliptical one. It is of the right order of size to have belonged to the same type of animal as the
supposed multituberculate molar from Watton Cliff, Dorset (Freeman 1976a). A similar (though unrolled)
multituberculate molar has also been found at Kirtlington, although it is outside the scope of the present
account.

Dimensions of FM/K 43 : Length of tooth >5-6 mm.

Length of crown > 1-54 mm.

Maximum width of tooth 1T6 mm.

FREEMAN: JURASSIC MAMMALS

163

Class mammalia Linnaeus (sensu lato)

Order therapsida Broom
Family tritylodontidae Cope

Plate 21, fig. 9

Description. Four cusps or parts of cusps which appear to be from tritylodontid cheek teeth have been found
at Kirtlington. They add nothing to the knowledge of the family, except to slightly extend its known range
upwards from the Middle Bathonian to the Late Bathonian. A complete incisor (FM/K 6) is tentatively
referred to the Tritylodontidae, because of its strong similarity to the Oligokyphus incisor R7304 (see text-
fig. 4 and compare with Kuhne 1956, text-fig. 38f).

Dimensions of FM/K 6 : Total height of tooth 2-36 mm.

Height of crown from gum line 0-89 mm.

Length of crown 1-07 mm.

Width of crown 0-52 mm.

TEXT-fiG. 4. Possible tritylodontid incisor FM/K 6, from
a, lingual aspect; and b, buccal aspect. The scale bar
represents 0-5 mm.

CONCLUSIONS

With the benefit of hindsight, the mammalian fauna of the Middle Jurassic is now
seen to be much as would be expected from its intermediate stratigraphic position
between the Rhaeto-Liassic and the Late Jurassic faunas. Thus, on the one hand,
there are the relict representatives of the Rhaeto-Lias families Morganucodontidae,
Kuehneotheriidae, and Tritylodontidae, and, on the other, early members of the
Late Jurassic families Dryolestidae, Docodontidae, and, possibly, Peramuridae. If
the Amphilestidae are placed in the Symmetrodonta, and if the specimen described
from Watton Cliff, Dorset, really is a Multituberculate (Freeman 1976a), then all of
the major groupings of Mesozoic mammals now appear to have existed in the Middle
Jurassic, with the apparent exceptions of the Eutriconodonta and the higher Theria.
Even these may be awaiting discovery in some as yet unknown Middle Jurassic
locality. The marked faunal differences between the known Middle Jurassic mammal
sites do not inspire confidence that even now we have a fair and balanced view of the
mammalian life of the Middle Jurassic.

The newly revealed diversity of the Bathonian mammalia at the level of the family,
and the absence of intermediate forms suggest that the higher mammalian taxa
became differentiated from one another long before the Bathonian. If the apparent
faunal poverty of the Rhaeto-Lias is genuine, it would seem that the Early Jurassic

164

PALAEONTOLOGY, VOLUME 22

was a time of rapid mammalian diversification. On the other hand, perhaps the
Rhaeto-Lias assemblages known at present are as biased and unrepresentative as
the Stonesfield fauna is now shown to be. Only time will tell.

Acknowledgements. The active help and encouragement of Dr. Martin Ware at the inception of this work
were critical to its success. I also thank the following for their help and/or interest : Dr. W. A. Clemens and
Messrs. C. T. Bilby, A. R. Elmes, E. J. King, D. K. Mugridge, F. J. Pinchin, D. J. Ward, and R. H. Whit-
field. I thank Dr. H. W. Ball and the staff of the Dept, of Palaeontology of the British Museum (Natural
History) for access to material in the national collection. I am most grateful to Mr. Noel F. C. Shelton
of GR-Stein Refractories Ltd. for the X.R.D. work, and to Mr. R. J. Gale of British Industrial Sand Ltd.
for arranging it. The Scanning Electron Photomicrographs were taken by Messrs. G. McTurk and
D. Bagley. This work is dedicated to the memory of my parents.

APPENDIX

The specimens described in this paper have been deposited in the Department of Palaeontology of the
British Museum (Natural History). The list below correlates the author’s collection numbers (prefaced with
'FM/K’) with the corresponding B.M.(N.H.) numbers (prefaced with ‘M’).

FM/K 1=M 36501
FM/K 2 = M 36502
FM/K 3=M 36503
FM/K 4=M 36504
FM/K 5=M 36505
FM/K 6=M 36506
FM/K 7= M 36507
FM/K 8=M 36508
FM/K 9=M 36509
FM/K 10=M 36510
FM/K 11 = M 36511
FM/K 12=M 36512
FM/K 13 = M 36513
FM/K 14=M 36514
FM/K 15=M 36515
FM/K 16=M 36516
FM/K 17=M 36517

FM/K 18 = M 36518
FM/K 19=M 36519
FM/K 20 =M 36520
FM/K 21 = M 36521
FM/K 22 =M 36522
FM/K 23 = M 36523
FM/K 24=M 36524
FM/K 25=M 36525
FM/K 26= M 36526
FM/K 27=M 36527
FM/K 28 = M 36528
FM/K 29= M 36529
FM/K 30=M 36530
FM/K 31 =M 36531
FM/K 32= M 36532
FM/K 33= M 36533
FM/K 34= M 36534

FM/K 35 = M 36535
FM/K 36=M 36536
FM/K 37= M 36537
FM/K 38 = M 36538
FM/K 39=M 36539
FM/K 40 =M 36540
FM/K 41 = M 36541
FM/K 42 =M 36542
FM/K 43 = M 36543
FM/K 52 =M 36552
FM/K 53=M 36553
FM/K 54=M 36554
FM/K 58 = M 36558
FM/K 60 =M 36560
FM/K 61 =M 36561
FM/K 62= M 36562
FM/K 63 = M 36563

REFERENCES

ager, D. v. 1977. On hairy reptiles. Proc. Geol. Ass. 88, 127-128.

arkell, w. j. 1931. The Upper Great Oolite, Bradford Beds, and Forest Marble of south Oxfordshire, and
the succession of gastropod faunas in the Great Oolite. Q. Jl geol. Soc. Lond. 87, 563-629.

barker, R. T. 1975. Dinosaur renaissance. Scientific American, 232, 58-78.

bate, R. J. 1965. Freshwater ostracods from the Bathonian of Oxfordshire. Palaeontology , 8, 749-759.

— and mayes, c. 1977. On Glyptocythere penni Bate and Mayes sp. nov. Stereo-Atlas of Ostracod Shells,
4(6).

buckland, w. 1824. Notice on Megalosaurus. Trans. Geol. Soc. London, (2), 1, 390-396.

butler, p. m. and ford, R. 1977. Discovery of Cretaceous mammals on the Isle of Wight. Proc. Isle Wight
nat. Hist, archaeol. Soc. 6, 662-663.

Clemens, w. a. and mills, J. R. E. 1971. Review of Peramus tenuirostris Owen (Eupantotheria, Mammalia).
Bull. Br. Mus. nat. Hist. {Geol.), 20 (3), 87-113.

gray, p. e. 1973. Marsupialia, insectivora, primates, creodonta and carnivora from the Headon Beds
(Upper Eocene) of southern England. Bull. Br. Mus. nat. Hist. {Geol.), 23 (1), 1-102.

FREEMAN: JURASSIC MAMMALS

165

Crompton, a. w. 1971. The origin of the tribosphenic molar. In kermack, d. m. and kermack, k. a. (eds.),
Early Mammals. Zool. J. Linn. Soc. 50, Suppl. 1, 65-87.

— and jenkins, f. a. 1967. American Jurassic Symmetrodonts and Rhaetic ‘Pantotheres’. Science, 155,
1006-1009.

— 1968. Molar occlusion in late Triassic mammals. Biol. Rev. 43, 427-458.
desmond, a. j. 1975. The Hot-blooded Dinosaurs — A Revolution in Palaeontology. Blond and Briggs Ltd.,
London, 238 pp.

dodsworth, c. 1972. The early years of the Oxford cement industry. Industrial Archaeology, 9, 285-295.
freeman, E. F. 1975. The isolation and ecological implications of the microvertebrate fauna of a Lower
Cretaceous lignite bed. Proc. Geol. Ass. 86, 307-312.

— 1976a. A mammalian fossil from the Forest Marble (Middle Jurassic) of Dorset. Proc. Geol. Ass. 87,
231-235.

— 19766. Mammal teeth from the Forest Marble (Middle Jurassic) of Oxfordshire, England. Science,
194, 1053-1055.

hahn, G. 1969. Beitrage zur Fauna der Grube Guimarota Nr. 3. Die Multituberculata. Palaeontographica
(ser. A), 133, 1-100.

hopson, J. A. and crompton, a. w. 1969. Origin of mammals. Evolutionary Biology, 3, 15-72.
joffe, J. 1967. The ‘dwarf’ crocodiles of the Purbeck Formation, Dorset: A reappraisal. Palaeontology, 10,
629-639.

kermack, d. m., kermack, K. A., and mussett, F. 1968. The Welsh pantothere Kuehneotherium praecursoris.
J. Linn. Soc. (Zool.), 47, 407-423.

kermack, K. a., mussett, f. and rigney, H. w. 1973. The lower jaw of Morganucodon. Ibid. 53, 87-175.
kuhne, w. G. 1949. On a Triconodont tooth of a new pattern from a fissure-filling in south Glamorgan.
Proc. zool. Soc. Lond. 119, 345-350.

— 1956. The Liassic therapsid Oligokyphus. Br. Mus. (Nat. Hist.), London, 145 pp.

— 1968. Kimeridge mammals and their bearing on the phylogeny of the Mammalia. In drake, e. j. (ed.).
Evolution and environment. Yale University Press, New Haven, 109-123.

mckerrow, w. s., Johnson, R. T. and jakobson, m. E. 1969. Palaeoecological studies in the Great Oolite at
Kirtlington, Oxfordshire. Palaeontology , 12, 56-83.

— and Kennedy, w. j. 1973. The Oxford District. Geol. Ass. Guide , no. 3, 46 pp.

mellett, J. s. 1974. Scatological origin of microvertebrate fossil accumulations. Science, 185, 349-350.
mills, J. r. E. 1964. The dentitions of Peramus and Amphitherium. Proc. Linn. Soc. Lond. 175, 1 17-133.

— 1971. The dentition of Morganucodon. In kermack, d. m. and kermack, k. a. (eds.). Early Mammals.
Zool. J. Linn. Soc. 50, Suppl. 1, 29-63.

odling, M. 1913. The Bathonian rocks of the Oxford District. Q. Jl geol. Soc. Lond. 69, 484-513.
owen, R. 1861. Monograph on the fossil reptilia of the Wealden and Purbeck formations. Part V. Lacer-
tilia ( Nuthetes , etc.). Palaeontogr. Soc. ( Monogr .), pp. 35-39.

— 1879. Monograph on the fossil reptilia of the Wealden and Purbeck formations. Supplement IX.
Crocodilia (Goniopholis, Brachidectes, Nannosuchus, Theriosuchus, and Nuthetes). Palaeontogr. Soc.
(Monogr.), pp. 6-10, 19.

palmer, t. j. 1973. Field meeting in the Great Oolite of Oxfordshire. Proc. Geol. Ass. 84, 53-64.
parrington, f. r. 1967. The origin of mammals. Advmt. Sci., Lond. 24, 165-173.

patterson, B. 1956. Early Cretaceous mammals and the evolution of mammalian molar teeth. Fieldiana
Geol. 13, 1-105.

Phillips, J. 1871. Geology of Oxford and the Valley of the Thames. Macmillan and Co., Oxford, 523 pp.
Richardson, l., arkell, w. J. and dines, h. G. 1946. Geology of the country around Witney. Mem. geol.
Surv. U.K., pp. 36-39, 69-71.

rixon, a. e. 1976. Fossil Animal Remains: their preparation and conservation. Athlone Press, London,
304 pp.

romer, a. s. 1966. Vertebrate Paleontology (3rd edition), University of Chicago Press, Chicago, 468 pp.
Simpson, G. G. 1928. A Catalogue of the Mesozoic Mammalia in the Geological Department of the British
Museum. Br. Mus. (Nat. Hist.), London, 215 pp.

— 1929. American Mesozoic Mammalia. Mem. Peabody Mus. Yale, 3, 1-235.

slaughter, b. h. 1971 . Mid-Cretaceous (Albian) therians of the Butler Farm local fauna, Texas. In kermack,
d. m. and kermack, K. a. (eds.). Early Mammals. Zool. J. Linn. Soc. 50, Suppl. 1, 131-143.

66

PALAEONTOLOGY, VOLUME 22

sugden, w. and mckerrow, w. s. 1962. The composition of marls and limestones in the Great Oolite
Series of Oxfordshire. Geol. Mag. 99, 363-368.

swinton, w. E. 1973. Fossil Amphibians and Reptiles (5th edition), Br. Mus. (Nat. Hist.), London, 133 pp.
torrens, H. s. 1969. The stratigraphical distribution of Bathonian ammonites in central England. Geol.
Mag. 106, 63-76.

waldman, M. and savage, R. J. G. 1972. The first Jurassic mammal from Scotland. ,// geol. Soc. Lond.,
128, 119-125.

ware, m. 1978. Palaeoecology and Ostracoda of a Bathonian mammal bed in Oxfordshire. Unpublished
M.Sc. thesis, Aberystwyth.

ERIC F. FREEMAN

Typescript received 7 April 1978
Revised typescript received 21 June 1978

146 Haydons Road
Wimbledon
London SW19 1AE

APPENDAGES OF THE ARTHROPOD
AGLASPIS SPINIFER (UPPER CAMBRIAN,
WISCONSIN) AND THEIR SIGNIFICANCE

by D. E. G. BRIGGS, D. L. BRUTON, and H. B. WHITTINGTON

Abstract. A single specimen of Aglaspis spinifer is unique in having the appendages preserved approximately in
place. Re-examination has shown that the cephalic region bore four (perhaps five) pairs of appendages, and not six
as Raasch (1939) claimed. The first pair was uniramous, apparently cylindrical and jointed, but cannot be confirmed
as being chelate. The remaining pairs on the cephalic region were like those on the anterior half of the trunk, uniramous
walking legs composed of five podomeres. The aglaspidids are therefore not chelicerates, and we do not consider them
to be closely related to trilobites. The order is not assigned to any higher taxon. The record of Chelicerata in the
Cambrian is dramatically reduced.

T he majority of specimens used by Raasch (1939) to describe the arthropod Aglaspis
spinifer came from a single locality in the Lodi Member, St. Lawrence Formation,
Trempealeau Stage, Upper Cambrian, Wisconsin. This account deals with only one
of Raasch’s specimens, which is unique in that appendages are preserved approxi-
mately in place beneath the dorsal exoskeleton. Raasch (1939, pp. 12-13, pi. 1;
pi. 2, fig. 11; pi. 4) stated that six pairs of appendages were present on the carapace
(cephalic region), the first being chelate, with those posterior to it and those on the
trunk being simple, unspecialized ‘walking legs’. In the light of these observations
Raasch (1939, pp. 69-84) reviewed the systematics of aglaspidids and concluded that
they were merostomes, a conclusion widely accepted by subsequent workers (Stormer
1944, pp. 74-77, fig. 14, 16 a,b\ 1955, pp. P10-P12, fig. 7, 4a, b; in Grasse 1949, p. 217;
Waterlot in Piveteau 1953, p. 546; Novojilov in Orlov 1962, p. 389; Bergstrom 1968,
p. 501; Eldredge 1974, p. 38). The accepted classification of Aglaspis and its allies
thus hinges on this single specimen (Stormer’s 1944, fig. 14, 16 b, is misleading in that
it portrays appendages taken from this specimen and transferred to a different
species, Aglaspella eatoni ), and our aim was to re-examine the basis for Raasch’s
statements. We are most grateful to Dr. Robert M. West, Milwaukee Public Museum
(abbreviated as MPM), Milwaukee, Wisconsin, for the opportunity to examine and
prepare the specimen. The ‘part’ (PI. 22, fig. 1) is an internal mould of the dorsal
exoskeleton, which shows also the external mould of parts of the ventral exoskeleton,
and portions of appendages beneath the internal mould; the ‘counterpart’ (PI. 23,
fig. 1) is an external mould of the dorsal exoskeleton. We do not follow Raasch in
dividing the body into ‘cephalothorax’, and ‘abdomen’, preferring to use the terms
‘cephalic region’ and ‘trunk’, as carrying fewer implications of supposed affinities.
The use of other terms is indicated in the explanatory text-figs. 1 and 2.

[Palaeontology, Vol. 22, Part 1, 1979, pp. 167-180, pis. 22-25.]

168

PALAEONTOLOGY, VOLUME 22

SYSTEMATIC PALAEONTOLOGY

Family aglaspididae Miller, 1877
Genus aglaspis Hall, 1862

Aglaspis spinifer Raasch, 1939

Plates 22-25 ; text-figs. 1, 2a, 2b

The original description and discussion by Raasch (1939, pp. 10-14, 62, 63, 65, 66,
pis. 1-4, pi. 7, figs. 1-4, pi. 9, figs. 9, 10, pi. 10, figs. 8-10) is based on sixty specimens,
most of which came from what he termed the merostome parting, in the Lodi Member,
St. Lawrence Formation, at Point Jude, three miles east of Gotham, Richland County,
Wisconsin. We have studied only the unique specimen from this locality which has
the appendages preserved approximately in place (MPM 11154, 11155, part and
counterpart respectively). Raasch (1939, fig. 6) gave a composite section at the
locality, and subsequently (1951, p. 141) explained that the trilobite Dikelocephalus
gracilis ovatus, which occurs at this horizon, is a synonym of D. oweni. Other trilo-
bites, a lingulid brachiopod, and species of four other genera of aglaspidids occur
with A. spinifer. Current stratigraphical terminology is given by Ostrom (1970, fig. 4).
Our comments on the unique specimen amplify or emend Raasch’s original
description.

Dorsal exoskeleton. The internal mould (PI. 22, fig. 1) is of the size given by Raasch
(1939, p. 10; his plate 1 is approximately xO-75 natural size). The counterpart
(PI. 23, fig. 1) shows the division of the dorsal exoskeleton into cephalic shield, eleven
trunk tergites, and the twelfth portion a long terminal spine (‘telson segment’ of
Raasch, 1939, p. 12). The spine was presumably circular or oval in cross-section
before compression, and there is no evidence that it was other than horizontal in
life. A relatively short anterior portion of each trunk tergite is set off by the articulating
ridge as an articulating flange ; the flange is smooth, the main portion of the tergite
faintly tuberculate (PI. 23, fig. 1). The posterior margin of the tergite bears a con-
spicuous narrow band of close-packed, coarser tubercles, similar to those on the
posterior margin of the cephalic shield. It appears that tergites 1-11 freely articulated
with each other, the cephalic region, and the terminal spine (PI. 25, fig. 1 ; text-fig. 2a).
Raasch (1939, p. 12) considered that tergite 1 1 was considerably longer than the first,
but much narrower than those preceding it, and so regarded somite 11 and the
terminal spine as comprising the ‘postabdomen’. Tergites 1 11 appear to change
progressively in form, the pleural regions becoming narrower backwards and more

EXPLANATION OF PLATE 22

Figs. 1-3. Aglaspis spinifer Raasch, 1939. Holotype, MPM 11154, part, internal mould, Lodi Member,
St. Lawrence Formation, Trempealeau Stage, Upper Cambrian, Point Jude, 3 miles E of Gotham,
Richland County, Wisconsin, U.S.A. 1, entire part, x 1-05. 2, posterior portion of part, showing post-
ventral plates and ventral view of proximal portion of terminal spine, x 2. 3, latex cast, anterior portion
of part, showing appendages in ventral view, x 1-5.

PLATE 22

BRIGGS, BRUTON, and WHITTINGTON, Aglaspis spinifer

170

PALAEONTOLOGY, VOLUME 22

strongly curved, and tergite 1 1 (the pleural regions of which (PI. 25, fig. 1 ; text -fig. 2a)
appear to have been crumpled by flattening of original convexity) seems to belong in
this gradational series. There is thus little evidence for a subdivision of the trunk.
Some separation of trunk tergites has occurred, so that part of the articulating flange
is exposed axially between them (PI. 23, fig. 1). Laterally the posterior margin of an
individual tergite diverges from the articulating ridge on the succeeding tergite
(against which it would have abutted), and the entire articulating flange of the
pleural region may be exposed (PI. 23, fig. 1 ; text-fig. 1). This is presumably the
result of flattening of the original convexity, combined with the evident separation
of the tergites in the axial region.

Appendages. The anteriormost appendage (PI. 22, fig. 3; PI. 23, fig. 2; PI. 24, fig. 1 ;
text-fig. 1) runs forward and outward from beneath the eye-lobe on each side of the
cephalic region to the margin. Raasch (1939, pi. 4, fig. 1) outlined four podomeres of
the left anterior appendage (the distal two forming the chela) in thick, black lines.
The lines of fracture or change in level traced by Raasch may be identified in PI. 22,
fig. 3 and PI. 23, fig. 2. The margins of a parallel-sided structure are preserved,
traversed by changes in level which curve from longitudinal to transverse in direction.
These impressions may represent a probably cylindrical, possibly jointed appendage
which has been flattened during preservation, but are not sufficient to conclude that
the preserved portion of the appendage was chelate. A featureless strip outlines the
position of the right anterior appendage (PI. 22, fig. 3 ; PI. 24, fig. 1), the margins not
clearly outlined and transverse lines lacking.

Behind the first appendage (numbered 1 in text-fig. 1) a series of nine similar pairs
(numbered 2 to 10) are exposed. The distal podomeres are best preserved, all append-
ages except the last on the right showing two or three of them. The fourth podomere
from the distal end is evident in right appendage 2, in pair 5, and in left 6 and those
posterior to it. A fifth podomere (the coxa) is apparent only in appendages right 5
and left 8. Small spines are visible in places along the margins of some of the append-
ages (right 2, left 5, 6, right 8) but the limbs do not appear to have borne an armature
of heavy spines. The proximal ends of the appendages are poorly preserved, presum-
ably because they lay close to the dorsal exoskeleton and were pressed against it
during preservation. As a result the two layers are difficult or impossible to separate
by preparation. In contrast, the thicker layer of matrix which intervenes distally
between appendages and exoskeleton makes preparation easier. The proximal
podomeres of appendages 2 to 10 are aligned more or less normal to the trunk axis, the
majority of the limbs flexed so that the distal two podomeres are preserved directed
backward or forward. The appendages were probably flattened antero-posteriorly

EXPLANATION OF PLATE 23

Figs. 1, 2. Aglaspis spinifer Raasch, 1939. 1, Holotype, MPM 11155, counterpart, showing posterior
portion of cephalic shield and portions of trunk tergites 1-7, x 2. 2, holotype, MPM 11154, part, median
and left anterolateral portion of cephalic region, for comparison with Raasch 1939, pi. 4, figs. 1, 2.
Symbols as text -fig. 1, those for appendages placed immediately to left of particular appendage, x 5.
Horizon and locality as PI. 22.

PLATE 23

BRIGGS, BRUTON, and WHITTINGTON, Aflaspis spinifer

text-fig. 1. Aglaspis spinifer Raasch, 1939. Camera-lucida drawing of MPM 11154, holotype, part,
internal mould; ar— articulating ridge, shown in fine stipple; arf— articulating flange; c— appendage of
cephalic region; cbj— coxa-body junction; e— area of eye-lobe outlined by dashed line; fix— flexure;
fr— fracture; la— labrum; lvp— left ventral plate; rvp— right ventral plate; t— appendage of trunk;
te— tergite of trunk ; tsp— terminal spine. Appendages are numbered 1-10, with suffix ; tergites are numbered
1-11, with suffix. Coarser stipple indicates matrix along fractures, around appendages, and between post-
ventral plates, hachures run down-slope from line indicating break in slope. Query indicates areas in which
interpretation is uncertain.

EXPLANATION OF PLATE 24

Fig. 1. Aglaspis spinifer Raasch, 1939. Holotype, MPM 11154, antero-median portion of part to show
appendages, x 2-5. For interpretation see text-fig. 1. Horizon and locality as PI. 22.

PLATE 24

BRIGGS, BRUTON, and WHITTINGTON, Aglaspis spinifer

174

PALAEONTOLOGY, VOLUME 22

in life, oval in cross-section, tapering distally so that the terminal podomere was
elongate-conical in form. This podomere shows one or two longitudinal grooves or
ridges, which may have served to strengthen it. In the course of burial the flexed limbs
have been rotated about the coxa-body junction (text-fig. 1) into the plane of bedding,
so that either the anterior (pairs 2 to 4) or posterior (pairs 8 to 10) surface is uppermost.
In pair 5 the left has been swung forward, while the right is unusual in being extended
straight, and may be compressed with the dorsal surface uppermost. In pairs 6 and 7
the left has been rotated forward, the right backward. The outline of the appendages
suggests that when extended the dorsal margin was straight. A narrow triangular
area separating some of the podomeres (e.g. the two most proximal in right 7 and 9)
presumably represents the less sclerotized arthrodial membrane of a hinge joint,
articulating dorsally. The terminal podomere is never sufficiently well preserved
distally to confirm that the extremity of the appendages was a blunt point and did not
bear a spine or spines.

Raasch (1939, p. 123) considered that the mid-portion of the cephalic region,
adjacent to the prominent fracture, shows the crushed remains of the basal joints of
appendages. This region lies below the level of the dorsal exoskeleton, and lacks the
characteristic external sculpture (PI. 22, fig. 3; PI. 23, fig. 2; PI. 24, fig. 1). Some
poorly defined ventral structures appear to be preserved, and are also evident in
the mid-region of tergites 1 to 4 (indicated by ? in text-fig. 1). They may represent
proximal parts of appendages, but the number of podomeres in each limb does not
appear to have exceeded five (it is considered unlikely that further podomeres are
concealed as a result of folding or overlap during burial). Raasch (1939, pi. 4, fig. 2)
outlined a possible ventral plate (‘epistoma’) in the cephalic region between the eye-
lobes. The margins of this supposed labrum are evident (PI. 23, fig. 2; PI. 24, fig. 1 ;
text-fig. 1) posteriorly and posterolaterally, but are not as clearly defined anteriorly
as suggested by Raasch.

The configuration of the ten paired appendages, including the similarity of spacing
between them, leads us to consider that, despite the portions missing and uncertain-
ties of outline, no further limbs remain to be exposed within the series. However, if
the original relation between dorsal exoskeleton and appendages is to be assessed, a
second assumption must be made, that each appendage is now in the same, or little
modified, position relative to the dorsal exoskeleton as it was in life. In other words,
it is assumed that decay of soft parts, burial, and compaction of the sediment did not
lead to displacement of the ventral cuticle and attached appendages relative to the
dorsal exoskeleton, but merely to the rotation of each appendage 2-10 about the
coxa-body junction so that it came to lie anterior or posterior face upward. This is a
large assumption for two reasons. Firstly, in the only other specimen of Aglaspis
showing appendages, A. barrandei (Raasch 1939, pi. 5, figs. 1-4), they are detached
and displaced. Secondly, in a study of a trilobite with appendages, Olenoides serratus
(Whittington 1975, pp. 102-104), it was shown that in all specimens displacement
relative to the dorsal exoskeleton occurs, despite the evidence that ventral cuticle and
appendages were not broken up but remained a unit. In the present specimen of
A. spinifer, the configuration of the appendages in relation to the dorsal exoskeleton
does, however, provide evidence which appears to favour the above assumptions.
Text-fig. 1 shows that the proximal ends of lc, and the coxa-body junction (dorsal

BRIGGS, BRUTON, AND WHITTINGTON: AGLASPIS SPINIFER 175

margin of proximal podomere) in 5t, 6t left, 7- lOt, lie at approximately equal distances
in a transverse line from the sagittal line. Further, the coxa-body junctions of 5t, 6t
left, and 7-10t lie beneath the axial region of the trunk, and successively beneath
tergites 1 to 6. This strikingly regular and symmetrical arrangement can only mean that
ventral cuticle and attached appendages remained as a unit and were not markedly
displaced in relation to the dorsal exoskeleton. The qualification ‘not markedly’ is
intentionally imprecise because the exact relationship must remain uncertain. In this
specimen the exact position of the coxa-body junction is not defined, and further
uncertainty as to the original relationship arises from the separation now evident
between dorsal tergites. We suggest that appendages lc to 4c belonged to the cephalic
region, and that probably the fifth pair (5t) belonged to the first trunk somite. That
the right fifth appendage belonged to this somite seems a reasonable assumption,
because the three most proximal podomeres lie beneath the first trunk tergite.
Interpretation of exactly where the dorsal margin of the equivalent podomeres of the
fifth left appendage lie (compare text-fig. 1 and PI. 24, fig. 1) is less certain, as the
dashed line in text-fig. 1 shows. On balance we conclude that the coxa-body junction
appears to lie in a position symmetrical to that of the right fifth limb. Raasch (1939,
pp. 12, 13) stated that six paired appendages were present in the cephalic region, the
first a chela, but our restudy suggests that there were only four pairs, and that
the first was not chelate. The uncertainties in our interpretations are evident, and the
possibility that the cephalic region bore five pairs of appendages cannot be excluded.

Postventral plate. This plate was defined by Raasch (1939, p. 12; pi. 1 ; pi. 2, fig. 12;
pi. 9, figs. 9, 10) and illustrated in position in this specimen, in one other, and as an
isolated plate. The cast from the counterpart (PI. 25, fig. 1 ; text-fig. 2a) shows the
posterior margins of tergites 10 and 11, straight medially, coarsely tuberculate, and
the articulating flanges of tergite 1 1 and the terminal spine, evidence of free articula-
tion between somites 10 and 1 1 and the spine. The cast from the part (PI. 25, fig. 2;
text-fig. 2b) appears quite different in the axial region. A pair of plates, subsemi-
circular in outline, occupy much of the axial region of somite 1 1 and the base of the
terminal spine, and project beneath a portion of somite 10. The adaxial margin of
each plate is bent dorsally, the outline gently curved convexly, and medially they are
in contact. The external surface is tuberculate. The abaxial margin is ill defined,
because of flattening and because the split between part and counterpart has not
followed the plate to this margin. Posteriorly each plate may have abutted against
the curved inner margin of the doublure of the terminal spine (imd in text-fig. 2b).
Because of compaction and flattening of the specimen, the curled (and therefore
more resistant to flattening) edge of the exoskeleton tends to be impressed into the
exoskeleton of the opposite side. Thus in the counterpart (PI. 25, fig. 1) the impression
made by the adaxial margin of each plate is marked, as is the impression in the part
(PI. 22, fig. 2; PI. 25, fig. 2) of the posterior margin of tergites 10 and 11, crossing the
plates. This specimen thus provides the type of evidence on which Raasch (1939,
p. 65) based his postventral plate, which he regarded as ‘divided into two longitudinal
halves presumably united anteriorly by a connecting membrane’. In A. spinifer
Raasch (1939, p. 12) described the plate as ‘almost completely bisected longitu-
dinally’, considering that the posterior cleft was open, the anterior possibly closed

lcm

text-fig. 2. Aglaspis spinifer Raasch, 1939. Camera-lucida drawings. A, MPM 11155, latex cast of posterior
portion of counterpart. B, MPM 11154, latex cast of posterior portion of part, ar— articulating ridge;
arf— articulating flange; fix— flexure; fr — fracture; imd— internal margin of doublure; lvp — left ventral
plate; pm— posterior margin; rvp— right ventral plate; te— tergite; tsp— terminal spine. Tergites numbered
7-11, 12 is terminal spine. Tubercles indicated by open circles. Lines with hachures indicate break in slope,
hachures are directed down-slope.

EXPLANATION OF PLATE 25

Figs. 1, 2. Aglaspis spinifer Raasch, 1939. 1, holotype, MPM 11155, portion of latex cast of counterpart,
x 3 ; for interpretation see text-fig. 2a. 2, holotype, MPM 11154, portion of latex cast of part, x 3 ; for
interpretation see text-fig. 2b. Horizon and locality as PI. 22.

PLATE 25

BRIGGS, BRUTON, and WHITTINGTON, Aglaspis spinifer

178

PALAEONTOLOGY, VOLUME 22

by a membrane. It is thus not clear whether Raasch thought the plate was divided
into two separate portions (as Stormer 1955, p. P10 implies). If it were, the occur-
rence of isolated examples implies that the two halves were held together by a less
sclerotized membrane which was not preserved. In the specimen studied here the
inner edges of the two halves, which are dorsally upturned, appear to be separated by
matrix along almost the entire length (PI. 25, fig. 2; text-fig. 1), the two edges in
contact medially for a very short distance. It does not appear that the two halves
were fused medially. The anus may have opened in the posterior portion of the cleft
in the postventral plate (Raasch 1939, p. 65).

DISCUSSION

Our objective was limited to the re-examination of the single specimen on which
depended the view that Aglaspis and its allies were to be assigned to Class Mero-
stomata; we have not studied other specimens of A. spinifer , nor searched for any
which also may have appendages preserved approximately in place. Without such
further work, and a wider study of aglaspidids, we would not attempt a restoration
of the animal. What we consider may be inferred about appendages of A. spinifer is
summarized in text-fig. 1 . This knowledge is both equivocal and incomplete, reflect-
ing the preservation of the specimen. We contend that the left anterior appendage
(lc in text -fig. 1 ; compare PI. 23, fig. 2) is too poorly preserved to be interpreted as
a chela, and that the cephalic region bore fewer than six pairs of appendages. If
we are correct, then Aglaspis and presumably other genera constituting the family
Aglaspididae are not Merostomata nor Chelicerata. Stormer (1955) included three
other families with the Aglaspididae in the Order Aglaspidida, and placed the order in
Subclass Xiphosura of the Merostomata. Various modifications to this taxonomy
have been proposed since 1955. Genera have been added to the family Aglaspididae
or the order by Chlupac (1965), Chlupac and Havlicek (1965), Novojilov (in Orlov
1962), and Repina and Okuneva (1969), a new family added by Flower (1969), while
Bergstrom (1968, 1971) has suggested subtractions from this grouping. Our work
bears only marginally on these matters. Now that we know the number of pairs of
appendages in the cephalic region of Aglaspis is 4 or 5, not 6, the force of Bergstrom’s
(1971) arguments for removing the families Strabopidae and Paleomeridae from the
Aglaspidida are diminished, and we doubt their validity. Further, the work in pro-
gress by Bruton does not support the attribution by Bergstrom of these two families
to the Subclass Merostomoidea Stormer, 1959, resemblances between them and
genera placed in this Subclass by Stormer being only superficial. The Cambrian
specimens of Khankaspis bzahnovi described as an aglaspidid by Repina and Okuneva
(1969) show structures beneath the exoskeleton which they have interpreted as
lamellate gill branches, but show no traces of the segmented limbs of Raasch’s
specimen. It appears to us an unwarranted inference (Bergstrom 1975, p. 291) to
state that aglaspidid appendages are biramous, for no trace of a branch is preserved
in the Wisconsin specimen, and the limb appears uniramous. More information is
needed on the nature of aglaspidid appendages, and the content of the order remains
uncertain.

BRIGGS, BRUTON, AND WHITTINGTON: AGLASPIS SPINIFER 179

If our restudy had unequivocally shown only four pairs of appendages in the
cephalic region of Aglaspis, the same number as in the cephalon of certain trilobites
(Cisne 1975; Whittington 1975, 1977), we might have argued for some relationship
between them. Such a relationship has been considered (Raasch 1939, pp. 69, 70;
Stormer 1944, pp. 76, 115, 116), but cannot be argued for on the similarity of the
body regions or of the appendages. The resemblance between olenellid trilobites
(in which only antennae are known) and aglaspidids appears to us superficial. For
example, the furrowed olenellid glabella, the long, curved eye-lobe, the opisthothorax
and pygidium, are exoskeletal features having no parallel in aglaspidids, and the
postventral plate is unique to the latter. Whatever genera and families may be
grouped with Aglaspis into the Order Aglaspidida, present knowledge excludes it
from merostomes and we prefer not to assign it to any higher taxon. Bergstrom (1968 ;
1975, pi. 1, fig. 1) described a fragmentary early Cambrian xiphosuran, and gave a
diagram (1968, fig. 8) of main evolutionary lines among early merostomes. The
removal of aglaspidids from the Merostomata widens the gap between these lines,
and dramatically reduces the Cambrian record of this class.

Acknowledgements. Natural Environment Research Council grant GR3/285 supported laboratory work in
the National Museum of Natural History, Washington D.C., U.S.A., by Briggs and Whittington, and
Bruton’s studies were supported by the University of Oslo. We are indebted to Drs. Porter M. Kier and
Richard E. Grant for facilities provided in the National Museum, and Bruton thanks Professor L. Stormer,
University of Oslo, for advice.

REFERENCES

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

1971. Paleomerus— merostome or merostomoid. Lethaia, 4, 393-401.

— 1975. Functional morphology and evolution of xiphosurids. Fossils Strata , 4, 291-305.
chlupac, i. 1965. Xiphosuran merostomes from the Bohemian Ordovician. Shorn, geol. Ved. Praha, P5,

7-38.

— and havlicek, v. 1965. Kodymirus n.g., a new aglaspid merostome of the Cambrian of Bohemia.
Ibid. P6, 7-20.

cisne, J. L. 1975. Anatomy of Triarthrus and the relationships of the Trilobita. Fossils Strata, 4, 45-63.
eldredge, N. 1974. Revision of the Suborder Synziphosurina (Chelicerata, Merostomata), with remarks on
merostome phylogeny. Novitates, 2543, 1-41.

flower, r. h. 1969. Merostomes from a Cotter horizon of the El Paso Group. Mem. Inst. Min. Technol.
New Mex. 22 (4), 35-44.

grasse, p. p. 1949. Traite de Zoologie. Vol. 6. Masson, Paris.

orlov, yu A. (ed.) 1962. Oznovy Paleontologii : Arthropoda tracheate and cheliferous. Academy of Sciences,
USSR, Moscow. [In Russian.]

ostrom, M. e. 1970. Field trip guidebook for Cambrian-Ordovician geology of western Wisconsin. Inf.
Circ. Wis. geol. nat. Hist. Surv. 11, 1-131.

piveteau, J. (ed.) 1953. Traite de Paleontologie . Vol. 3. [viii]+ 1063 pp., Masson, Paris.
raasch, G. o. 1939. Cambrian Merostomata. Sp. Pap. geol. Soc. Am. 19, i-ix, 1-146.

— 1951. Revision of Croixan Dikelocephalidae. Trans. III. St. Acad. Sci. 44, 137-151 (also Circ. III. St.
geol. Surv. 179, 1952, 137-151).

Repina, l. n. and okuneva, o. G. 1969. Cambrian arthropods of the Maritime Territory. Paleont. J. 3,
95-103. [Translation of Paleont. Zh. by Am. geol. Inst.]
stormer, l. 1944. On the relationships and phylogeny of fossil and Recent Arachnomorpha. Skr. norske
Vidensk-Akad. mat. -nat. Kl. 5, 1-158.

180

PALAEONTOLOGY, VOLUME 22

ST0RMER, L. 1955. Chelicerata, Merostomata. pp. 1-41. In moore, r. c. (ed.). Treatise on invertebrate
paleontology, Part P, Arthropoda 2. Geological Society of America and University of Kansas Press.

— 1959. Trilobitomorpha, Trilobitoidea. pp. 22-37. In moore, r. c. (ed.). Treatise on invertebrate
paleontology. Part O, Arthropoda 1. Geological Society of America and University of Kansas Press.

Whittington, H. b. 1975. Trilobites with appendages from the Middle Cambrian, Burgess Shale, British
Columbia. Fossils Strata, 4, 97-136.

— 1977. The Middle Cambrian trilobite Naraoia, Burgess Shale, British Columbia. Phil. Trans. R. Soc.
B280, 409-433.

D. E. G. BRIGGS
Department of Geology
Goldsmith’s College
New Cross
London SE14 6NW

D. L. BRUTON

Paleontologisk Museum
Sars Gate 1
Oslo 5
Norway

Manuscript received 20 April 1978
Revised manuscript received 27 June 1978

H. B. WHITTINGTON
Department of Geology
Sedgwick Museum
Downing Street
Cambridge CB2 3EQ

A NEW FOR AMINIFER FROM THE
MIDDFE EOCENE OF PAPUA NEW GUINEA

by c. g. adams and d. j. belford

Abstract. Reticulogyra mirata, a miliolacean with some unusual morphological characters, is described from the
Chimbu Limestone of Papua New Guinea.

The stratigraphy and foraminifera of the Eocene/Oligocene Chimbu Limestone,
Papua New Guinea, were described by Bain and Binnekamp in 1973. While their
work was in press we noticed that a rather unusual foraminifer, not mentioned by
Binnekamp in his faunal description, occurred through some 12 m of the Middle
Eocene part of the 300 m thick sequence in the Chimbu Gorge. Since few limestones
of Middle Eocene age have as yet been described from the Malay Archipelago and
the western Pacific, the discovery of a new species is not particularly surprising. It
is, however, unusually interesting because its short range and striking appearance
could make it a valuable marker fossil in this part of the Indo-Pacific region.

The Chimbu Limestone forms a prominent scarp on the western limb of the
Yaveufa Syncline and crops out over a distance of about 1 km along the Chimbu
River near Kundiawa (text-fig. 1). As the succession was described in some detail by
Bain and Binnekamp (1973) only the lower part is figured here (text-fig. 2). R. mirata
occurs in samples 20NG 0094-0099, in a hard, grey, dense limestone, unsuited to the
extraction of foraminifera which have, therefore, to be studied by means of random
thin sections. Other foraminifera present in these samples were identified by Binne-
kamp as Fasciolites cf. elongata d’Orbigny, Nummulites javanus Verbeek, and
Dictyoconus chimbuensis Binnekamp, an assemblage clearly indicative of the Middle
Eocene. Numerous small miliolids are also present, as is a new flabelliform larger
foraminifer referred to later. The general aspect of the assemblages in the six samples
studied suggests deposition in fairly shallow water under low energy conditions.

The holotype and figured paratypes are deposited in the Commonwealth Palaeontological Collections,
Bureau of Mineral Resources, Canberra, under numbers CPC 18101 to CPC 18117. Unfigured paratypes
are deposited in the British Museum (Natural History).

SYSTEMATIC PALAEONTOLOGY
Family miliolidae

Subfamily fabularinae Ehrenberg, 1839
Genus reticulogyra nov.

Diagnosis. A planispirally coiled miliolacean possessing short transverse and parallel
(i.e. normal and parallel to the septa respectively) subepidermal partitions throughout
most of the test.

Derivation of name. From the Latin, meaning netted spire.
[Palaeontology, Vol. 22, Part 1, 1979, pp. 181-187, pi. 26.]

182

PALAEONTOLOGY, VOLUME 22

Chimbu

Section

Limestone

measured and sampled

B55/A/33

text-fig. 1 . Locality maps showing position and area of outcrop of the Chimbu Limestone.

ADAMS AND BELFORD: EOCENE FORAMINIFER 183

text-fig. 2. Stratigraphical section through the
lower part of the Chimbu Limestone in the
measured section showing sample positions and
principal elements of the foraminiferal fauna
in each sample. D =Dictyoconus chimbuensis;
F - Fasciolites cf. elongatus\ N =Nummulites
javanus ; R = Reticulogyra mirata', m=miliolids.

m

ES

i T

HII

0099

0098

0097

0096

0095

0094

0091

0090

D.F.m

D,F R,m
D,F,R,m
D,F,R,m

F.R.m

DfF,N,R,m
D,F R,m

m

m

0

ft

0

10

5-

20

10

-30

III Massive
Mil limestone
Calcareous
I'v.MFv.I sandstone

F,m

AH samples have the
F,m prefix 20 NG

B55/A/34

Reticulogyra mirata sp. nov.

Plate 26, figs. 1-10; text-figs. 3, 4

Diagnosis. As for the genus.

Derivation of name. From the Latin, meaning to be wondered at.

Material. Hundreds of specimens in random thin sections of limestone.

Description. Test porcellaneous, laterally compressed or subspherical chambers
arranged in a planispiral coil of two to three whorls with five and a half to seven
chambers in the last whorl. In some forms the final chamber shows a tendency to
flare or to become uncoiled (PI. 26, fig. 5; text-fig. 3). Short transverse and parallel
subepidermal partitions extend into each chamber lumen from the third onwards,
the former usually being both thicker and slightly deeper than those parallel to the

text-fig. 3. Outline drawing of an off-centre axial
section of Reticulogyra mirata showing a strongly
compressed and flared terminal chamber. Sample
NG 0095.

184

PALAEONTOLOGY, VOLUME 22

septa ; they produce a pitted effect when seen in tangential sections just cutting the
test’s surface (PI. 26, fig. 10; text-fig. 4a). These partitions may be visible externally
as a reticulum in matrix-free individuals since the outer wall is very thin. All but the
first few chambers possess a basal wall which is usually thickest in the median plane
and occasionally gives the chambers an angular appearance when seen in thin section
(PI. 26, fig. 1). The proloculus is spherical, subspherical, or irregular, and is followed

c

text-fig. 4. Photographs of specimens on surface of polished blocks using reflected light.
All from sample NG 0095. a. Reticulogyra mirata. Tangential section through coil showing
longitudinal and parallel subepidermal partitions forming a reticulum, - 33. CPC 18101.
b. Gen. et sp. indet. Sagittal/oblique section through flared portion of test showing septa
and subepidermal partitions, 22. CPC 18102. c. Gen. et sp. indet. Semi-equatorial section
through flared portion of test showing septa and the reticulum formed by the subepidermal
partitions, x5. CPC 18103.

EXPLANATION OF PLATE 26

All figs, x 20 approx, unless otherwise stated.

Figs. 1-10. Reticulogyra mirata gen. et sp. nov. 1-4, transverse equatorial sections showing planispiral
coil, septa, and variation in size and shape of proloculus. In 1 the basal layer gives the chambers an angular
appearance; 3, holotype (CPC 18106); 4b ( x 32), enlargement showing parallel subepidermal partitions.
5-9, axial or near axial sections showing variation in shape; lb and 8 b ( x 42 and x 30), enlargements
showing basal wall (lb) and transverse partitions more clearly; 10 ( < 32), tangential section cutting
surface of an inner whorl and showing reticulum produced by intersection of parallel and transverse
subepidermal partitions.

Figs. 11-14. Either the microspheric form of R. mirata or an undescribed and indeterminable mean-
dropsinid. 11 ( ■ 41), sagittal section through flared portion of test showing septa and parallel sub-
epidermal partitions. 12-14 ( x 25, <41, x 20), subequatorial sections through flaring individuals. All
three specimens show septa, transverse and parallel subepidermal partitions as does R. mirata.

Figs. 1 - 1 4 registered as CPC 18104-18117. Figs. 1-4, 8, 9, 1 1 from sample 20NG 0098; figs. 5-7, 10, 12, 14
from sample 20NG 0095; fig. 13 from sample 20NG 0097.

PLATE 26

ADAMS and BELFORD, Reticulogyra mirata

186

PALAEONTOLOGY, VOLUME 22

by a short tube a quarter to half a turn in length. The diameter of the proloculus
ranges from 0-20 to 0-55 mm; the larger proloculi are sometimes irregular. The
aperture has not been seen clearly and could be either single or multiple.

Holotype. CPC 18106; Plate 26, fig. 3, from sample 20NG 0098.

Measurements. Max. diameter 0T8 mm, min. diameter 0-17 mm. The microspheric generation has not
been seen.

Remarks. Reticulogyra is difficult to place in any miliolid family as currently defined
since no other genus has both transverse and parallel subepidermal partitions.
Indeed, in this respect its structure resembles that of some members of the subfamily
Dicyclininae (Lituolacea) from which group it is, however, excluded by the possession
of a basal wall, a feature believed to be confined to porcellaneous foraminifera. The
only similar genus within the subfamily Fabularinae is Raadshoovenia van den Bold,
but this has a milioline coil and lacks parallel partitions. Cuvillierinella Papetti and
Tedeschi (1965), a Cretaceous genus, although described as having a planispiral
initial coil, clearly begins with a milioline coil which quickly becomes planispiral
(Papetti and Tedeschi 1965, figs. 2a, b, 4c, d ) ; it also lacks parallel partitions. Taberina
Keijer, a soritid, is planispiral then uncoiling and has incomplete transverse inter-
septal partitions and interseptal pillars.

Although Bain and Binnekamp (1973) mention only Dictyoconus, Fasciolites, and
Nummulites from the samples in which Reticulogyra occurs, another important genus
is also present (PI. 26, figs. 11-14; text-fig. 4b, c ). Like Reticulogyra, it possesses both
longitudinal and transverse subepidermal partitions and appears to be porcellaneous.
The adult test is flabelliform and comprises some nine to sixteen chambers. The most
complete individuals so far obtained range from 8 to 11 mm in length and 0-35-
0-50 mm in thickness. One individual (PI. 26, fig. 13) is at least 10 mm wide. The trans-
verse partitions are thicker and more widely spaced than those parallel with the septa.
Unfortunately, we have not seen the initial stage of this taxon and cannot therefore
assign it to an existing genus or describe it as new. It may, just possibly, be the micro-
spheric form of Reticulogyra, although its flabelliform habit suggests that it is more
probably a meandropsinid, possibly related to Saudia.

Wall structure. When viewed in thin section (e.g. PI. 26, figs, lb, 8b), the walls of the
foraminifera described here bear a striking resemblance to that of Austrotrillina
Parr. One of us (C. G. A.) has long been puzzled by the development of the alveolar
wall in this particular genus since it appeared to serve no structural purpose or to
offer any selective advantage. It could hardly have been primarily intended to confer
additional structural rigidity since species of Triloculina (normal milioline wall) of
similar shape and size inhabited the same environments successfully. Physiologically,
an alveolar wall seems actually to be disadvantageous since it could only impede the
free internal streaming of cytoplasm. The porcellaneous wall of R. mirata, while
structurally slightly different from that of Austrotrillina, appears to offer the same
physiological disadvantage while similarly lacking any primary structural advantage ;
other calcareous foraminifera of similar shape and size were perfectly successful
without this structural modification. A. howchini, the last member of the Austro-
trillina lineage, when viewed in reflected light, is seen to have an outer wall which is
so thin that the internal structures can be seen through it (Adams 1968, pi. 2, figs. 1,2).

ADAMS AND BELFORD: EOCENE FORAMINIFER

187

We think that the same is true of Reticulogyra. Because neither genus possesses open-
ings (other than the main aperture) to the exterior, the thinning of the wall cannot
have been intended to permit cytoplasm to leave the test. If the modification was
neither to strengthen the wall nor to let anything escape through it, then its purpose
was presumably to allow something to enter while maintaining the original rigidity :
we suggest that this something may have been light, and that the ridges formed
between the alveolae maintained the original strength of the test. In this connection,
it is worth noting that Lee and Zucker (1969, p. 75) observed that in living Archaias
the algal symbionts appeared to be concentrated in window-like areas in the test
wall, while Ross (1972) noted that the outer walls of the lateral chamberlets in
Marginopora vertebralis served as calcite windows for the symbionts.

Austrotrillina modified its shell continuously during the 18 million years or so of
its existence (Early Oligocene to early Middle Miocene), and then became extinct
for no apparent reason. So far as we know Reticulogyra enjoyed only a very brief
existence during the Middle Eocene. They both occupied tropical shallow-water
carbonate environments which are not known to have been undergoing any profound
changes when these genera became extinct. We therefore suggest that Austro-
trillina and Reticulogyra may have modified the original miliolacean wall in order to
provide better illumination for symbionts. If, after modifying the wall, they became
dependent upon particular symbionts for the maintenance of their metabolism, then
the extinction of these symbionts would necessarily have encompassed the extinction
of the host species without any change being visible in the sedimentary environment.

Acknowledgements. One of us (D. J. B.) has received permission to publish from the Director, Bureau of
Mineral Resources, Geology and Geophysics; we are both grateful to Miss C. A. Harrison and Mr. R. L.
Hodgkinson (B.M. N.H.) for technical assistance.

REFERENCES

adams, c. G. 1968. A revision of the foraminiferal genus Austrotrillina Parr. Bull. Brit. Mus. Nat. Hist.
(Geol.), 16, (2), 73-97.

bain, J. h. c. and binnekamp, J. G. 1973. The foraminifera and stratigraphy of the Chimbu Limestone, New
Guinea. Bull. Bur. miner. Resour. Geol. Geophys. Aust. 139, 1-12.
lee, J. j. and zucker, w. 1969. Algal flagellate symbiosis in the foraminifer Archaias. J. Protozool. 16 (1),
71-81.

papetti, I. and tedeschi, d. 1965. Nuovo genere di foraminifero del Santoniano superiore. Geologica
romana, Rome, 4, 119-128, figs. 4-8.

ROSS, c. A. 1972. Biology and ecology of Marginopora vertebralis (Foraminiferida), Great Barrier Reef.
J. Protozool. 19(1), 181-192.

C. G. ADAMS

Department of Palaeontology
British Museum (Natural History)
London SW7 5BD

D. J. BELFORD

Typescript received 17 February 1978
Revised typescript received 15 May 1978

Bureau of Mineral Resources
Geology and Geophysics
PO Box 378

Canberra City, ACT 2601

THE HAMPEN MARLY AND WHITE
LIMESTONE FORMATIONS: FLORIDA-TYPE
CARBONATE LAGOONS IN THE JURASSIC
OF CENTRAL ENGLAND

by T. J. PALMER

Abstract. The Hampen Marly Formation of central England was laid down on a shallow shelf fringing the shore of
the London Landmass to the east. Three principal environments can be recognized : (i) a nearshore lagoonal region,
dominated by clastic sediments, with faunal and floral evidence of periodic brackish and freshwater influence;

(ii) a shallow marine, more offshore region dominated by oyster reefs, where clastic and carbonate sediments intermix ;

(iii) a deeper, more fully marine area further offshore, where the carbonate sediments and their marine fauna are
identical to those of the overlying White Limestone Formation. The White Limestone Formation of the area studied
can be divided lithostratigraphically into three members. These are, in ascending order, the Shipton Member, the
Ardley Member, and the Bladon Member. Each member represents a shallowing of the sea by sediment accumulation
after an initial deepening that was probably under tectonic control. Eight recurrent assemblages of invertebrates, which
are strongly correlated with the nature of the substrate, can be recognized. Close similarities exist between these
assemblages, and those found today in comparable habitats, such as those in Florida. A picture of the region’s
palaeogeography and the distribution of the main habitats is built up from the facies analysis. Some ideas on the
tectonic control of the sedimentation are put forward.

This paper describes and discusses the depositional environments and the ecological
control on the faunas of two stratigraphic units in the Great Oolite Group (Bathonian)
over part of central England. It contains a summary of a broader study over a wider
area (Palmer 1974, D.Phil. thesis), where the evidence on which the conclusions
herein are based is more fully presented. The principal geographic region under
consideration runs along the outcrop from Burford (SP 2512) to Buckingham
(SP 6933). South-west of this area there is a marked drop in the number of extant
exposures in both the Hampen Marly and the White Limestone Formations. There is
also a facies change in the White Limestone representing the passage from pre-
dominantly marine lagoonal micrites to predominantly well-sorted current-swept
lime sands associated with a shelf break at the western limit of the Oxfordshire
shallows (Worssam and Bisson 1961). The region chosen for study therefore has both
artificial and natural limits.

The study involved the investigation of over thirty extant exposures, some of
considerable size. In addition, some information has been taken from the observa-
tions and accounts of previous workers particularly in the now poorly exposed
Hampen Marly Formation in the region around and to the south-west of Burford.
The sections studied are listed in Table 1.

THE HAMPEN MARLY FORMATION

Definition and stratigraphy

The Hampen Marly Formation ( = Hampen Marly Beds of Arkell, Richardson, and
Pringle 1933) consists of a variable series of clays, limestones, marls, and sands, which

[Palaeontology, Vol. 22, Part 1, 1979, pp. 189-228.]

table 1. Map references of the sections in the Hampen Marly Formation and the White
Limestone Formation discussed in this account. The principal previous descriptions are
also given.

PRINCIPAL

LOCALITY MAP REFERENCE PREVIOUS ACCOUNT

Boreholes

Stony Stratford

SP 785405

Woodward 1894

Calvert

SP 689247

Davies and Pringle 1913

Oxford

SP 505063

Pringle 1926

Witney

SP 344147

Poole 1969

Stowell Park

SP 088128

Green and Melville 1956

Latton

SU 093957

Arkell 1933c

Charlton

ST 963888

Geological Survey field notes

Ready Token

SP 105044

Richardson 1932

Quarries and Cuttings

Deanshanger

SP 753395

Woodward 1894

Brackley

SP 757377

Walford 1912

Aynhoe

SP 516335

Woodward 1894

Croughton*

SP 563336

Palmer 1974

Stratton Audley*

SP 6025

Palmer 1973

Blackthorn Hill*

SP 616211

Arkell 1931

Ardley*

SP 543264

Palmer 1973

Ardley Railway Cutting*

SP 518287-
531276

Arkell et al. 1933

Northbrook Farm*

SP 497222

Palmer 1974

Kirtlington*

SP 494199

McKerrow et al. 1969

Shipton*

SP 4717

Arkell 1931

Whitehill Quarry, Gibralter*

SP 497187

Palmer 1974

Greenhill*

SP 486178

Arkell 1931

Wood Eaton*

SP 534123

Palmer 1973

Tolley’s New Quarry, Bladon*

SP 450150

Arkell 1933ft

Old White House Quarry, Bladon*

SP 448150

Arkell 1933ft

Long Hanborough Station*

SP 436142

Arkell 1931

Layshill Wood*

SP 418147

Richardson 1946

Whitehill Wood*

SP 388153

Richardson 1946

Whitehill Wood Railway Cutting

SP 395157

Arkell 1931

Fisher’s Gate, North Leigh*

SP 391142

Richardson 1946

Stonesfield

SP 389169

Walford 1896

Minster Lovell War Memorial*

SP 316109

Arkell 1931

Astall*

SP 300118

Worssam and Bisson 1961

Eton College*

SP 297102

Worssam and Bisson 1961

Whitehill (Sturt Farm)*

SP 271109

Palmer 1974

Swinbrook

SP 278124

Woodwood 1894

Stonelands*

SP 278098

Palmer 1974

Milton*

SP 257157

Richardson 1910

Taynton

SP 236155

Richardson 1933

Little Barrington

SP 205121

Richardson 1933

Windrush

SP 190125

Richardson 1933

Slape Hill*

SP 423196

Palmer 1973

Whiteways, Middle Barton*

SP 420246

Palmer 1974

Great Rollright*

SP 322304

Palmer 1974

Temple Mills*

SP 345361

Whitehead and Arkell 1946

Hampen*

SP 062205

Richardson 1929

Pinswell

SP 035126

Richardson 1933

North Cerney

SP 023080

Richardson 1933

Chedworth Cutting

SP 061019

Richardson 1933

Foss Cross*

SP 056091

Torrens 1969

Dagham Downs*

SP 003060

Torrens 1967

An asterisk indicates that the section is extant, and has been examined by the author.

PALMER: HAMPEN MARLY AND WHITE LIMESTONE FORMATIONS

91

J

This Account

McKerrow & others 1969

Arkell 1947

Arkell 1931

CORNBRASH

Opper Cornbrash
Member

FORMATION

Lower Cornbrash
Member

LOWER CORNBRASH

LOWER CORNBRASH

LOWER CORNBRASH

FOREST

Wvchwood Beds

WYCHW00D BEDS

a.

3

FOREST MARBLE FORMATION

FOREST

MARBLE'

_

MARBLE

Bradford Beds

BRADFORD BEDS

KEMBLE BEDS (cr Cheese&U Epithyris Beds

O

CD

WHITE

LIMESTONE

FORMATION

Bladon Member

Fimbriatus-Waltoni Clay
Oyster Epithyris Marl

Kemble Beds
Bladon Beds.incl
Fimbriata-Waltoni Beds

Fimbriata-Waltoni Beds
Middle Epithyris Bed

a

CO

Ardley Member

WHITE

LIMESTONE

GREAT

OOLITE

Lower Epithyris Bed
Ardleyensis Bed

'E

©

“5

o

Shipton Member

WHITE LIMESTONE

Ardley Beds

BLOCKS fl&in
(White Lst.)

Nerineaeudesii Bed
Roach Bed

GO

re

E

CD

HAMPEN MARLY FORMATION

HAMPEN MARLY BEDS

TAYNTON LIMESTONE FORMATION

NOT DISCUSSED

TAYNTON STONE

NOT DISCUSSED

text-fig. 1 . Scheme of stratigraphic subdivision of the Upper Great Oolite Group adopted in this account,
compared with selected previous divisions.

occur between the Taynton Limestone Formation (where present), and the overlying
White Limestone Formation (text-fig. 1). The 8-7-m-thick type section is at Hampen
Railway Cutting, west of Salperton, Gloucestershire (Table 1), and has been described
by Woodward (1894), and Richardson (1929). Within the main study area the
Hampen Marly Formation is only well exposed at Wood Eaton Quarry (Palmer 1973).
The locations of former exposures are shown in text-fig. 2, and listed in Table 1.

To the south-west of the type section the Hampen Marly Formation seems to
disappear, and is replaced laterally by beds of the White Limestone Formation. In
the railway cutting at Chedworth (Richardson 191 1) it is barely distinguishable from
the latter, being represented only by thin marl beds in a predominantly limestone
sequence. Similarly in the Stowell Park borehole, 2-5 km north-east of the Chedworth
cutting (Green and Melville 1956), there is doubt about the development of the forma-
tion, which is probably represented only by an uncored 2-6-m interval immediately
overlying the Taynton Limestone Formation. This was inferred to be marly from the
gamma-ray log. At Latton, 10 km south-west of Cirencester, the formation appears
to be entirely absent (Arkell 1933c; Richardson 1933), having passed laterally into
expanded White Limestone (Arkell 1947; Arkell and Donovan 1952). At Charlton,
15 km south-south-west of Cirencester, 3 m of doubtful Hampen Marly Formation
were recorded in the Geological Survey borehole. All these beds, however, were
limestones, and their inclusion in the Hampen Marly Formation seems to have been
based on the presence of local wisps of mudstone. Otherwise they seem to have been
identical to what was referred to as White Limestone in the Latton section.

To the north-east of the study area the formation shows another change in facies
into the clays of the Upper Estuarine Formation (Arkell 1951-1958, p. 14), although
the precise relationship between the two formations is not entirely clear. Only one
ammonite has ever been recorded from the Hampen Marly Formation— Procerites
imitator (S. Buckman) from the base in the Ardley-Fritwell railway cutting— and
none have been found in the Upper Estuarine Formation. However, the base of the
overlying Great Oolite Limestone in Northamptonshire is considered by Torrens
(1968), on the basis of ammonite evidence, to be younger than the base of the laterally

192

PALAEONTOLOGY, VOLUME 22

equivalent White Limestone Formation in the Oxfordshire/Gloucestershire region.
Similar reasoning indicates that the Upper Estuarine Formation is at least partly
younger than the Hampen Marly Formation. The section at Croughton (text-fig. 3)
supplies supporting evidence. Here the bottom beds of the White Limestone (the
Shipton Member ; see below) contain considerably more clay than the corresponding
beds in Oxfordshire and Gloucestershire; they also contain rootlets, which are a
feature typical of Upper Estuarine and Hampen Marly Formation conditions of
deposition, rather than those of the White Limestone. Thus these lower beds at
Croughton may represent the lateral transition from beds of White Limestone to the
south-west, into beds of the Upper Estuarine Formation to the north and east.

At the base of the formation in the region around the type locality there are some-
times alternations of marls and cross-bedded calcarenites of typical Taynton Lime-
stone facies, and it is not easy to decide on the precise junction between the two
lithostratigraphic units (Richardson 1933). Similarly, the junction between the
Hampen Marly Formation and the overlying White Limestone is seldom clear-cut,
and the one formation passes into the other through an alternating sequence of marls
and marly limestones. The nature of Hampen Marly Formation sediments is further
affected by local variations between different depositional regimes within the forma-
tion; thus clays with freshwater fossils may pass vertically and laterally into lime-
stones with marine fossils, and then into marls with rootlets (as at Wood Eaton),
suggesting small-scale variations in relative sea-level of purely local distribution;
perhaps associated with local changes in the conformation of emergent and sub-
merged areas.

However, in spite of these complications, there are noticeable general trends
within the formation. Field studies and previous accounts of the Hampen Marly
Formation indicate that there is a marked change in their dominant lithology as one
passes from north-east to south-west in the area over which they outcrop. To the
north and east the formation is dominated by clays (sometimes sandy) and thin
sands. Towards the south-east the proportion of marly clay and marl increases;
further in this direction the marls start to become associated with small calcium
carbonate peloids (usually referred to as ooliths by earlier authors) and beds of marly
limestone. Eventually limestones completely replace the marls and the formation
passes laterally into the base of the White Limestone Formation in the region around
Cirencester.

These changes are shown on text-fig. 2. All available sectional details (see Table 1
for references) have been analysed, and the thicknesses of seven different lithologies
(sand; sandy clay; clay; clayey marl and marly clay; marl; marly limestone; lime-
stone) have been expressed graphically as a percentage of the total thickness of the
Hampen Marley Formation seen at that particular locality. The variety of lithologies
present in the Hampen Marly Formation at any one exposure is immediately obvious ;
in addition the three broad zones characterized by different sediment types can be
distinguished, (i) The north-eastern and eastern area where clay and sand pre-
dominate. (ii) The central region where marls, marly clays, and marly limestones
predominate, (iii) The south-western region where marly limestones and limestones
predominate.

An exercise like this is not necessarily reduced in value by uncertainties in correla-

PALMER: HAMPEN MARLY AND WHITE LIMESTONE FORMATIONS

193

text-fig. 2. Relative proportions of different sediment types in exposures of the Hampen Marly Formation ; see Table 1 for sources of information and
map references of localities. The inset shows a putative environmental reconstruction.

194

PALAEONTOLOGY, VOLUME 22

tion. In the eastern part of the study area the Taynton Limestone Formation is absent
beneath the Hampen Marly Formation, and beds of similar clayey facies extend to
within a few metres of the Lias (see Davies and Pringle 1913 on the Calvert Borehole
section). Possibly some of this thickening represents beds which are the lateral
equivalent of the Taynton Limestone Formation, and perhaps part of the lower
Great Oolite as well, although Sellwood and McKerrow (1974) suggest that this is
absent east of Ardley. Similarly, it is likely that the bottom member of the White
Limestone Formation (the Shipton Member) passes into a nearshore clayey and
sandy faciei in the region around Brackley and Buckingham. However, these con-
siderations enhance the opinion that the sediments of both the Hampen Marly Forma-
tion and other formations of the Great Oolite Group, show greater terrigenous
clastic influence to the north-east and east. Although the stratigraphic limits of the
Hampen Marly Formation may become less distinct in this direction, the sediments
are still predominantly clastic, as opposed to autochthonous.

Fauna and flora

At Wood Eaton the base of Hampen Marly Formation bed 5 (the Monster Bed of
Palmer 1973) contains a particularly well-preserved fauna and flora. I am very grate-
ful to Professor T. M. Harris and Dr. R. H. Bate for their comments on the macro-
flora and the ostracods respectively. Both these categories are richly represented
and would fully repay closer study.

Flora. Professor Harris has kindly pointed out several features of the flora: Equise-
tites is represented by in situ rootlets; otherwise this genus is unrepresented, being
totally absent in the well-preserved fragmentary plant material. Equisetites was
common in low-lying, marshy deposits of the Jurassic, while it is likely that the
fragmentary flora comes from some way inland, being deposited by a river in the
lowland area fringing the sea. The upland flora found in the Middle Jurassic Deltaic
Series of Yorkshire consists largely of conifers, as does the fragmentary Wood Eaton
plant material. However, here the resemblance stops, and Harris (pers. comm.)
states: T was slightly astonished that none of the cuticles was of a species I knew
from the Yorkshire Bajocian and Bathonian.’ Clearly, a considerable separation
between the Mid North Sea High and the London Landmass is indicated.

The cuticles of the Wood Eaton species are stated by Harris to be ‘neither specially
thick, nor specially thin in relation to other members of their classes, and do not
suggest an arid climate’. This supports the idea of river drainage having been the
transporting and depositional agent. The abundance of charophytes also suggests
the influence of fresh water. Only the robust female receptacles (oogonia) are found,
and fragments of other parts of the thallus are absent. This implies that they were
derived from a source some way removed from the deposits where they are found,
probably in lakes upstream.

Microfauna. Of the nine species of ostracod found in the bed, only two are marine,
and these are both rare. The ostracod fauna has much in common with that of the
Upper Estuarine Series (Aslin 1965), the following forms being common to both:
Lophocy there scabra , Micropneumatocy there postrotunda , Bisulcocypris sp., Faba-

PALMER: HAMPEN MARLY AND WHITE LIMESTONE FORMATIONS 195

nella bathonica, Macrodentina ( Mediodentina ) bathonica, and Kliena levis. In terms of
both species represented, and total numbers of specimens, the ostracod fauna is
predominantly brackish (R. H. Bate, pers. comm.). There are no wholly freshwater
forms, and it seems likely that the ostracods lived in brackish lagoons fringing the
shore, rather than being derived from up-river.

Macrofauna. The presence of several bones of Cetiosaurus suggests shallow swampy
conditions with plenty of vegetation. Otherwise, the fauna of this bed is entirely
molluscan (Table 2). The presence of Viviparus and Valvata indicates freshwater
conditions. Usually when these forms occur in the Great Oolite they do so in great
abundance, frequently to the exclusion of nearly everything else (Hudleston 1896;
Palmer 1973). In this case, their rarity suggests that they have been derived from
elsewhere, possibly the same freshwater source as the charophytes.

Several further species in the bed have been implicated as brackish or euryhaline
forms, and there is a strong similarity between this fauna and the Bathonian faunas
of the Hebrides (Hudson 1963a, b). Cuspidaria ibbetsoni and Placunopsis socialis
are both suggested as being tolerant of slightly reduced salinities 25-30°/oo (Hudson
1963a), and Neomiodon is thought by Hudson to be the dominant form at brackish
water salinities of about 9°/00, though ranging up into brackish marine and down into
brackish/freshwater salinities. Tancredia and Protocardia have not before been impli-
cated as brackish water forms, and probably had a fairly wide salinity tolerance.

It is well known that brackish-water faunas show a reduction in the number of
species with large numbers of individuals, when compared with fully marine faunas
(e.g. Hedgpeth 1957), and that at a higher taxonomic level certain groups (e.g.
echinoderms, sponges, ectoprocts, and articulate brachiopods) are generally
intolerant of salinities much below 35°/00. This reduction in faunal diversity is true for
the whole of the Hampen Marly Formation at Wood Eaton (Palmer 1973). Bed 2 of
Palmer (1973), for example, also contains (except for a single echinoid spine) only
molluscs, with Eomiodon and two species of Corbula accounting for 50% of the
fauna (Table 2). On the other hand, the invertebrate macrofauna from these two beds
still each contain fifteen different species (Table 1)— which is very different from the
monotypic shell-beds typical of the Great Estuarine Series and other brackish-water
deposits. There are two possible explanations for this apparent high diversity. First
the species minimum is found not at salinities midway between fully marine (35%0)
and freshwater (0-0-5°/oo), but shifted towards the freshwater end of the spectrum, at
about 9°/00 (Remane 1934). To the marine side of this value the fauna is of a restricted
marine character, but not at a diversity minimum. Such a situation could explain the
fauna in question if all the molluscs were euryhaline. The second, and more likely,
explanation for their occurrence together is that some degree of mixing has occurred.
This may have been brought about by current activity, but the relatively uniform
nature of the sediment suggests that this has played only a minor part. What is more
likely is that environmental conditions in the region of deposition fluctuated quite
rapidly, as features such as rainfall, run-off, and conformation of the land varied, and
the salinity of the water frequently changed. Under these changing conditions the
nature of the benthos would fluctuate also, and the resulting death assemblage would
be a mixture of species which lived in different environments. These fluctuations are

196

PALAEONTOLOGY, VOLUME 22

table 2. Quantitative details of the fauna of selected beds from the Hampen Marly and White Limestone
Formations of the study area.

TABLE 2

KEY TO BIVALVE LIFE HABITS
(All are suspension feeders)

A ADPRESSED BYSSATE
A' FREE-SWINGING BYSSATE
A RECLINING
A CEMENTED

O SEMI-INFAUNAL BYSSATE
W SLUGGISH. SHALLOW BURROWING
ACTIVE. SHALLOW BURROWING
• DEEP BURROWING
-w BORING

Taxon

Strat.Unit

Facies

Bed

Life

Habit

HAMPEN

MARLY

FORMATION

SHIPTON MEMBER

ARDLEY MEMBER

BLADON MEMBER

llelodon hirsonensis (d' Archaic)

llelodon bynei Cox & Arkell

Cucullaea sp

Eonavicula minuta (J. de C. Sowerby)

Barbatia pratti (Morris & Lycett)

Modiolus imbricatus J. Sowerby

Modiolus (Inoperna) plicatus J. Sowerby

Lithophaga fabella (J.A. Eudes-Deslongchamps) —
Myoconcha acteon Morris & Lycett ex d'Orbigny —
Pteroperna costatula (J.A. Eudes-Deslongchamps)-

Bakevellia waltoni (Lycett)

Costigervi l lia crassicosta (Morris & Lycett)

Gervillella ovata (J. de C. Sowerby)

Isognomon isognomonoides (Stahl)

Isognomon (Myti loperna) bathonicus (Morris & Lyc

Isognomon (Myti loperna) murchisonii (Forbes)

Pinna odlingi Arkell

Chlamys (Radulopecten) vagans (J. de C. Sowerby)

Camptonectes annulatus (J. de C. Sowerby)

Camptonectes rigidus (J. Sowerby)

Hypotrema sp

Placunopsis fibrosa Laube

Placunopsis socialis Morris & Lycett

i (Plagiostoma) subcardi iformis Greppin

i (Plagiostoma) bynei Cox & Arkell

i (Plagiostoma) sp

Pseudolimea duplicata (J. de C. Sowerby)

Limatula gibbosa (J. Sowerby)

Ctenostreon rugosum (W. Smith)

Praeexogyra hebridica (Forbes)

^ Lopha costata (J. de C. Sowerby)

jonia pullus J. de C. Sowerby

^ Trigonia (Vaugonia) moretoni Morris S Lycett

Neomiodon brycei (Tate)

locardia loweana (Morris & Lycett) :

iocardia islipensis (Lycett)

locardia minima (J. Sowerby)

Pseudotrapezium cordiforme (Deshayes)

illista antiopa (Thevenin ex d'Orbigny)

na bellona d’Orbigny

Corbis lajoyi (d‘ Archaic)

Sphaeriola oolithica (Rollier)

Tancredia truncata Morris & lycett

Tancredia extensa Lycett

Tancredia planata Morris & Lycett

Quenstedtia bathonica (Morris & lycett)

Quenstedtia morrisi Cossmann

Protocardia lycetti (Rollier)

Protocardia buckmani (Morris & Lycett)

Protocardia stricklandi (Morris & Lycett)

Protocardia sp

Pleuromya uniformis (J. Sowerby)

Gresslya sp

Corbula hul liana Morris 1

Corbula sp —

Myopholas acuti costa (J. de C. Sowerby)

Pholadomya deltoidea( J. Sowerby)

>mya gibbosa (J. Sowerby)

Arcomya sp —

Cercomya undulata (J. de C. Sowerby)

Gastrochaenopsis sp

Cuspidaria ibbetsoni (Morris)

-0.7 0.8

-51.9 9.2 16.2

-3.3 2.3 5.0

-2.7 7.2 3.8

0.8

3.8 2.5

PALMER: HAMPEN MARLY AND WHITE LIMESTONE FORMATIONS

197

TABLE 2 (cont.)

Strat.Umt

KEY TO MAIN GROUPS

I = SPONGES

ii = corals Facies

III = WORMS
IV = CRUSTACEAN
V = GASTROPODS
VI = BRYOZOANS

VII = BRACHIOPODS E^d

VIII = ECHINOIDS

Life

Taxon Habit

HAMPEN

MARLY

FORMATION

WHITE LIMESTONE FORMATION

Ardley WL 18 coral bed =2

Croughton WL 16 \ z

1 non

Croughton WL 19 \ coraliferous m

Ardley WL 19 )

ARDLEY MEMBER

Shipton WL 1 coral bed 2

1

Shipton WL 2 fimb - wait clay 3

I

B

in

w

Y

m

m

11

Limnoria sp

CD

CD

FB

FB

AR

ES

CD

ES

TS

VD

?BD

?VS
?VG
VG
VG
?VG
VG
VG
?VG
VG
VG
VG
?VG
? IS
IS
IS
? IS
?VG
IC

?vc

-1.7

Enaulofungia sp

15.8

0.5

0.2

0.7

Cyathophora pratt i Edwards & Haime

-0.2

Choinatoseris sp

0.5

Serpula (Cyc loserpula) sp

-1.5

Serpula (Tetraserpula) sp

terebel loid

0.5

4.5 7.9

0.2

0.5

Helminthopsis sp

0.8

Trochotoma obtusa (Morn's & Lycett)

Scurria bathiensis (Rollier)

Chartrionella bunburii (Morris & Lycett)

-4.4

1.8

Amber leya bathonica Cox & Arkell

Notodelphinula depauperata (Morris & Lycett)

-0.2 3.0

1.8

0.6

3.7

2.4

40.0

49.6

Proconulus sp

Pseudomelania sp

-1.0 2.2 4.6

- 0.3

1.3 0.8

0.5 0.4 1.6

Neridomus cooksoni (Deslongchamps)

Brachytrema sp

—6.0
1.3

3.0 3.7

Ceritella sp

Fibula phasianoides (Morris & Lycett)

5.0

Cossmannea bathonica (Rigaux & Sauvage)

1.2

-48.8

Endiaplocus roissyi (d'Archaic)

Dicroloma sp

-0.5

—1.5

0.3

0.3

—0.3

Globularia sp

Pa laeohydatina sp

-2.5 6.1 3.8 2.5

Cylindrites sp

Viviparus sp

?vc

vs

vs

ES

CD

ES

ES

ES

ES

ES

CD

VG

VG

DP

1.3

*—1.2

11.0 12.6
6.8

Valvata sp

1.3

Stomatopora dichotoma (Lamouroux)

3.2 3.0

? ctenostome

Epithyris oxonica Arkell

-12.1 9.4

2.1

22.2

0.7

Epithyris sp

Digonel la digonoides (S.S. Buckman)

-35.8 5.3 3.7

1.5 0.8

Moorellina sp

Acrosalenia hemicidaroides Wright

—0.5

1.5

0.5 1.1 1.5

Pseudodiadema sp

Clypeus muelleri Wright

27.5

Figures indicate percentages of a bulk-sampled collection of the bed in question. Key to life habits is as
follows: cd— cavity dwelling encruster (suspension feeder); fb— frame builder (suspension feeder /carni-
vore); ar— active recliner (suspension feeder); es— exposed surface suspension feeder; ts— tube dwelling
suspension feeder ; vd— vagile deposit feeder ; bd— burrowing deposit feeder ; vg— vagile grazer ; vs— vagile
scavenger; vc— vagile carnivore; is— infaunal suspension feeder; ic— infaunal carnivore; dp— deposit
feeding plougher.

198

PALAEONTOLOGY, VOLUME 22

characteristic of lagoonal environments today (Shepard and Moore 1960), and are of
sufficiently short period for the different shell assemblages not to be separated into
different sedimentary strata.

That the restricted nature of the fauna is a response only to the adverse substrate
is unlikely. Approximately contemporaneous clays, usually considered to have a
fully marine origin, contain a very different bivalve fauna, e.g. Nucula, Grammatodon,
Oxytoma , Trigonia, Lucina , and Goniomeris in the Fuller’s Earth Clays of Dorset
(Arkell 1933a, p. 253).

There is evidence in previously published literature that the sands and clays of the
Hampen Marly Formation in the eastern part of the study area (see text-fig. 2) may
also have been laid down under conditions of reduced salinity. In the Oxford City
Brewery borehole (Pocock 1908), 8-60 m of ’Upper Estuarine Series’, which may be
referred to the Hampen Marly Formation, are recorded immediately beneath the
White Limestone. Predominantly, they consist of sandy clays and black shales with
abundant plant debris, some of it well-preserved. In two of these beds, is the only
recorded species ‘ Cyrena\ a name often applied by early workers. In the sands and
clays of the Ardley/Fritwell railway cutting (see Arkell et al. 1933), Odling (1913)
records ‘Car diurn stricklandi ( = Protocardia stricklandi ), ‘ Cardium ’ incertum ( = 1 Pro-
to car dia morrisi\ see Cox and Arkell 1948) and ‘ Astarte ’ angulata ( = Eomiodon
angulata ). Both these genera include species found at Wood Eaton in the Hampen
Marly Formation which may have been euryhaline. In their bed 17 at Ardley, Arkell
et al. (1933) recorded rootlets of Equisetum and abundant ostracods ‘of the Darwinula
style’. Darwinula is a freshwater form (Van Morkhoven 1962-1963).

It would seem, then, that reduced salinities were of more than local influence in
the control of facies and faunas over the eastern region.

Associated with the changing lithofacies across the region of Hampen Marly
Formation outcrop are faunal changes. The brackish- and fresh-water indicators
disappear; at Milton (text -fig. 2), for example, Aslin (1965) identified ten ostracod
species, of which nine are considered marine and one brackish-marine, and there is a
similar loss of beds with rootlets. Marls with abundant Praeexogyra hebridica
become common, some reaching over 2 m in thickness (Richardson 1933). Kalli-
rhynchia concinna is recorded occurring in some of these oyster beds, and, with the
exception of occasional encrusters on the oysters, other species are not common.
Where they are recorded, they are usually confined to limestones. This is probably due
to preservational effects, the marls are soft and have undergone compaction. As a
result of this all the originally aragonitic fossils are preserved as rather indeterminate
composite moulds. These are difficult to identify even when sections are fresh, but as
the beds weather the fossils are destroyed completely. Poorly preserved and unidenti-
fiable bivalves can be found by digging back into the marls at old exposures, such as
Milton Quarry. In contrast to this, forms which were originally calcite (e.g. K. con-
cinna and P. hebridica '), have not dissolved away during diagenesis, and weather out
of the marls in good preservation. This gives the false impression that these forms
dominated the assemblages when the rocks were laid down.

Nevertheless, K. concinna and P. hebridica do occur in enormous numbers in the
Hampen Marly Formation. Richardson (1929) records a bed at Lower Swell which
was over 2 m thick, and composed entirely of valves of P. hebridica. Similar beds, up

PALMER: HAMPEN MARLY AND WHITE LIMESTONE FORMATIONS 199

to 2 m thick, are met frequently, and are reminiscent of the reefs formed by other
oyster species ranging from the Jurassic (e.g. Liostrea distorta in the Purbeck) to the
present day (e.g. Crassostrea virginica in the lagoons of the Texas Gulf coast).
Hudson (19636) and Hudson and Palmer (1976) have argued the similarities of
Praeexogyra to Crassostrea, and it seems reasonable to look at the Hampen Marly
Formation oyster reefs in the light of our knowledge of the present-day ones.

Most oysters require a hard surface on which to settle (Stenzel 1971, p. N1012).
In gregarious oysters, larvae prefer to settle close to conspecific adults, which is
obviously important in the maintenance of reefs. Presumably the first generation
founding a new reef requires a substrate such as a shelly or coarse-grained bottom,
rather than a soupy mud. The only way of determining whether the Hampen Marly
Formation oysters obey this rule was to survey the previous literature (see Palmer
1974 for details) for all mention of what appear to have been oyster reefs (either
oyster beds over 15 cm thick, or beds in which oysters are stated to have been very
abundant), and take note of the underlying bed. Of the twenty-one Hampen Marly
Formation oyster reefs considered in this way, all nineteen for which details of the
underlying bed are known are supported by limestones. Although there is no evidence
that these limestones were cemented at the time of oyster settlement, they no doubt
formed a substrate that was firm enough and contained enough hard particles for the
oysters to get established.

Once an oyster reef has become established, it greatly influences the nature of
sedimentation in the vicinity (Lund 1957). Inorganic material in suspension is caught
up in the feeding currents, and compacted into either pseudofaeces or faeces. These
compacted pellets then behave as coarse sediment particles and sink to the bottom in
the neighbourhood of the reef. Oyster beds are thus effective sediment accumulators,
taking fine material in suspension, and depositing it as coarse grains, so that the beds
are centres of rapid sedimentation. The fine pelleted nature of such deposits in the
fossil record is not immediately obvious. However, where the sediment has been pro-
tected from agitation and compaction (e.g. inside a shell), it may sometimes still be
seen to have a pelleted form.

It seems likely that the development of the predominantly marly facies of the
Hampen Marly Formation, between the muddy brackish lagoons adjacent to the
London Landmass and the offshore marine region, was closely associated with
the development of the oyster reefs. The environment in which these reefs flourished
was quiet, and not greatly current influenced. Hence, unpelleted fine material may
also have been able to settle out. This situation would have been enhanced by the
baffling effect of the reefs themselves.

On this model it is possible to explain the lateral transition of Hampen Marly
Formation into expanded White Limestone in the Cirencester region: conditions
became fully marine in the offshore, south-westerly direction, such that the con-
ditions for the oyster-reef growth became unsuitable. At the present day, Crassostrea
reefs cease normal growth during the high salinities above 36°/00, and are replaced by
Ostrea equestris under permanent fully marine conditions (Stenzel 1971). However,
the encrusting fauna, typically of ectoprocts and foraminiferans, which is sometimes
associated with the Hampen Marly Formation P. hebridica, suggests tolerance of the
fully marine salinities. Therefore depth and increasing substrate instability seem the

200

PALAEONTOLOGY, VOLUME 22

most likely reasons for the sudden loss of oyster reefs and the associated marls.
Similarly, recent oyster reefs thrive best in shallow water (1-10 m deep— see Stenzel
1971).

In the Texas Gulf coastal lagoons at the present day the reefs themselves and the
inter-reef areas support different animal communities (Parker 1960). In the Hampen
Marly Formation the main inhabitant of the reef other than P. hebridica itself seems
to have been K. concinna. This is in contrast to the more clayey environments in the
Upper Estuarine Series and the Great Estuarine Series, where the main assiciate of
Praeexogyra is Modiolus (Hudson 1963a) and the Recent situation where Crassostrea
is associated with Brachidontes (Parker 1960). In addition to this, the Hampen Marly
reefs are often encrusted by ectoprocts (‘ Berenicea' and Stomatopora ), serpulids
(Dorsoserpula sp.) and foraminiferans ( Nubeculinella sp.), and occasionally bored by
worms and ?phoronids. These associations are not found in the Upper Estuarine
Series, and seem to suggest a much higher degree of marine influence than in those
formations. The inter-reef assemblage of the Hampen Marly Formation, if present,
is likely to have consisted of aragonitic infaunal bivalves and new fresh sections are
needed before any attempt can be made to elucidate it.

Summary

The available evidence, therefore, is consistent with the following model for the
environment of deposition of the Hampen Marly Formation. In the east and north-
east the London Landmass supported a fully terrestrial flora, and was drained by
rivers. These rivers discharged into a system of lagoons around the edge of the land,
whose salinities showed fluctuations. Further south-westwards these brackish lagoons
came under a more marine influence. Occasionally, fully marine deposits were laid
down. These sediments subsequently provided stable substrates for reefs of P. hebri-
dica and an associated fauna, which greatly influenced the nature of the sediments in
their vicinity by the self-sedimentation process and baffle action. These sediments
were predominantly marls. Still further offshore (further to the south-west) deeper
open marine conditions were unsuitable for the maintenance of oyster reefs, and the
formation passes laterally into White Limestone facies. The approximate limits of
these three broad regions are shown in text-fig. 2.

THE WHITE LIMESTONE FORMATION

The White Limestone Formation is the best exposed and most fossiliferous sub-
division of the Great Oolite Group between Burford and Buckingham. It varies in
thickness along the outcrop between about 20 m in the Cherwell valley, and about
6 m in the most north-easterly part of the region. It is predominantly composed of
well-cemented limestones. Locally, clay horizons are developed, but these nowhere
attain the thickness and lateral extent which would justify their being classed as
separate members within the formation.

The limestones which compose the bulk of the formation vary between sparites,
originally deposited as grain-supported lime sands, and micrites, originally deposited
as fine-grained lime muds. The dominant particles which make up the lime sands,
and which are usually also common in the lime muds, are peloids and shell fragments.

PALMER: HAMPEN MARLY AND WHITE LIMESTONE FORMATIONS 201

The term ‘peloid’ implies doubtful origin (McKee and Gutschick 1969) but the
vast majority of peloids in the Great Oolite Group were probably produced as the
pseudofaeces or faeces of suspension and deposit-feeding animals. In Recent car-
bonate environments, such pellets start as soft accretions of sediment, but are known
sometimes to harden slowly by a poorly understood process of intra-cementation,
to form resistant particles of micritized rock (Illing 1954; Bathurst 1971), which
behave as sand-sized grains. This process is likely to be the single most important
factor controlling deposition of sediment in the White Limestone Formation. Other
peloids may represent micritized shell fragments, and ooliths which have lost their
concentric structure during diagenesis according to the sort of process described by
Shearman, Twyman, and Karimi (1970).

Of the shell fragments, those derived from species whose shells were composed of
calcite (at least in part) are invariably preserved in their original mineralogy, whilst
fragments that were originally aragonite have always been subsequently dissolved,
and are usually replaced by void-filling sparry calcite. They frequently retain their
original outline in the form of a micrite envelope (Bathurst 1966).

Ooliths are locally common and sometimes abundant. However, many of the beds
described as oolitic by earlier authors who saw small spherical grains in them are
likely to have been peloidal rather than oolitic; it is often difficult to recognize the
true oolitic structure in a hand specimen, even with a lens. Even if fractured pellets
are seen, they frequently appear to be concentrically banded, because a micrite
envelope or a ferruginous pellicle penetrates to some distance beneath the surface,
and can give a false impression of accretionary banding. Although this distinction
was recognized as important in the Great Oolite Group by Woodward (1894),
subsequent workers tended to over-generalize, and the distinction between ‘ooliths’
and ‘pseudo-ooliths’ never became in England the commonplace that it did, for
example, in France. Locally, compound intraclasts and larger limestone pebbles are
also found, indicating nearby cementation and erosion.

The limestones at the top of the White Limestone Formation are almost invariably
micrites, with or without scattered peloids and shell fragments. This fact makes the
formation easily distinguishable from the overlying Forest Marble Formation in
which micrite is usually absent. The two formations are further distinguished by the
abundance of bioturbation in the White Limestone, which is probably associated
with a high micrite content giving the sediment stability. The Forest Marble, in
contrast, was a loose grain-supported lime sand and contains very few burrows.
Consequently the original bedding is usually preserved.

The origin of the micrite in the White Limestone Formation is not obvious. That
it originated as lime mud, and is not a diagenetic fabric, is indicated by the frequency
with which it forms a matrix within which other particles are embedded. Under the
scanning electron microscope it appears as minute granular calcite crystals, which
are presumably recrystallized aragonite. Therefore neither morphology nor isotope
chemistry of the mud particles can be expected to show whether the original sediment
had an inorganic or an organic origin. It is assumed that breakdown of skeletal
material, particularly of calcareous algae, was the principal source of the lime mud
as it is today.

202

PALAEONTOLOGY, VOLUME 22

Litho stratigraphy of the White Limestone

The White Limestone Formation falls within the ammonite zones of Tulites sub-
contracts to Prohecticoceras retrocostatum (Torrens 1968). Schemes of further
subdivision have been based on specific faunal marker horizons, some of which appear
to extend over most of the region between Burford and Buckingham. Others, however
(e.g. the Upper Epithyris Bed of the Cherwell valley), are of more limited geographic
extent. In both cases these horizons mark constant lithofacies, supporting a charac-
teristic fauna. They do not necessarily represent time lines.

Arkell (1931) introduced a scheme of subdivision based on nerineid gastropods,
distinguishing the upper third of the formation ( Aphanoptyxis bladonensis Beds).
Recently, Barker (1976) has extended this scheme by recognition of a third species of
Aphanoptyxis , A. excavata , which occupies a zone approximately equivalent to the
lower half of Arkell’s ardleyensis Beds.

These high-spired gastropod horizons, which also include the Nerinea eudesii bed
(actually a species of Nerinella) which Arkell (1931) first noted at a level about 4 m
below the top of his ardleyensis zone, have considerable value in correlation over
short distances and in field recognition of sections. Nevertheless, the control of all
these species by lithofacies is strong, and their use as stratigraphic markers must be
seen in the light of considerations of changing palaeogeography. The scheme of
subdivision employed here is based on lithology, and it has proved possible to recog-
nize three cycles of sedimentation within the White Limestone Formation. These
cycles are interpreted as representing relatively rapid episodes of deepening, followed
by a filling-in of the basins thus created. The resultant sequence of facies from deep to
shallow is not everywhere identical, but correlation of slightly different sequences
which can nevertheless both be interpreted as shallowing upwards, seems empirically
more useful than correlation by facies dependent species in nearshore heterogeneous
environments.

The three shallowing upwards cycles here recognized comprise, in ascending
order, the Shipton Member, the Ardley Member, and the Bladon Member (text-
fig. 1).

The Shipton Member. The type section of the lowest member of the White Limestone
Formation is at Shipton quarry (see text-fig. 3). This member has previously been
called the Croughton Member (Palmer 1974). However, these beds can no longer be
seen at Croughton, whereas a new section throughout the whole of the member can
now be seen at Shipton Cement Works.

The Shipton Member is characterized by both sparitic, micritic, and muddy lime-
stones, alternating with thin clays and marls. In the south-eastern part of the study
area (Wood Eaton and Shipton) there is a tendency for the proportion of fines to
increase upwards. The over-all proportion of clastic material, both as individual
beds and as disseminated fine sand grade quartz and clay in the limestones, is greater
than in the overlying Ardley Member. The limestones are furthermore rather more
poorly cemented than those in the Ardley Member. This is clearly seen in thin section,
where the inter-granular pores are often filled with clay, unfilled, or collapsed, whereas
a well-developed calcite cement is the rule in the Ardley Member. This poorer
cementation is also obvious in the field : the limestones are softer, and allow the frost

PALMER: HAMPEN MARLY AND WHITE LIMESTONE FORMATIONS 203

to penetrate their surfaces and cause a rough exfoliation of the superficial layers.
Furthermore, after the solution of the original aragonite of the gastropods and many
of the bivalves, the surrounding matrix was often not sufficiently cemented to hold
open the resulting voids, so there is usually no calcite replacement. Thus fossils that
were originally aragonite are frequently preserved in the Shipton Member as com-
posite moulds.

SHIPTON MEMBER

text-fig. 3. Thickness and sediment types of the Shipton Member (White Limestone Formation) in the

main study area.

The junction between these two members is easy to recognize at all the localities
shown in text -fig. 3. The bottom bed of the overlying Ardley Member throughout the
study area is invariably a thick, massive, well-cemented biopelsparite, which may
become dearagonitized at its base (e.g. the Roach Bed at Ardley). This bed contains
a rich fauna of burrowing bivalves as well as large nerineid gastropods (Table 2),
and the solitary coral Chomatoseris. Beneath this the top bed of the Shipton Member
is a clay in the north-eastern region (Croughton, Ardley) and passes south-westwards
into a sandy clay or marl which is often laminated. The soft nature of this bed allows

204

PALAEONTOLOGY, VOLUME 22

it to weather back rapidly beneath the hard overlying basal bed of the Ardley Member,
which then sticks out and forms a prominent overhang which is a characteristic of
all weathered vertical faces in this part of the White Limestone Formation (e.g.
Slape Hill; Fisher’s Gate, North Leigh).

As discussed above, the junction between the Hampen Marly Formation and the
base of the Shipton Member is gradational. At some localities (e.g. Shipton) the basal
Shipton Member Limestones are sparites, in contrast to the marly limestones in the
upper parts of the Hampen Marly Formation. However, the easiest way of defining
the base of the latter seems to be on faunal and floral criteria ; where good exposures
of the beds either side of the contact are seen, the fauna of Kallirhynchia concinna,
Praeexogyra hebridica, and shallow burrowing bivalves which characterizes the
Hampen Marly Formation, is replaced upwards by beds containing a more diverse
fauna often including corals, Stiphrothyris, Epithyris, Lopha costatum, Clypeus
muelleri , nerineid gastropods, and deep burrowing bivalves such as Pholadomya and
Homomya. This transition was clearly seen in the railway cutting at Ardley by Odling
(1913) and Pringle (in Arkell et al. 1933); the junction between the two units here lies
at the base of bed 20 in the latter account. Rootlets, on the other hand, are typical
Hampen Marly Formation features and are absent from the Shipton Member except
in the north-east part of the study area. Here, as discussed above, there is reason to
suspect that the Shipton Member is passing laterally into beds of Upper Estuarine
facies. At Stratton Audley, on the eastern edge of the main study area, test borings
have shown the member to contain too much clay and marl to be worked for road
metal.

There are, furthermore, several faunal criteria, the consequences of original
environmental differences, which can be used with considerable confidence to
distinguish Shipton Member from the overlying Ardley Member in the field. Coral
shoals and their associated fauna (discussed below) are typical of the former. The
presence of rhynchonellids, a small Epithyris (here called Epithyris sp.), Clypeus
muelleri , Lopha costatum, and Solenopora jurassica all indicate Shipton Member
(Table 2; text-fig. 3). Barker (1976) has recognized a new species of Aphanoptyxis,
A. excavata which is also typical of the Shipton Member. None of these species have I
found in the Ardley Member.

West of Burford these criteria cannot be used for distinguishing the Shipton Mem-
ber, and the unit probably has no value in that region. The limestones become more
pure, and were predominantly well-sorted oolitic and peloidal lime sands. At the
same time the coral shoals and their associated faunas die out. Presumably the sea
bottom was sufficiently agitated and unstable to be unsuitable for predominantly
sedentary epifaunal faunas of this sort. As Worssam and Bisson (1960) point out,
this region is also marked by a general thickening of the White Limestone as a whole,
and represents the boundary between the London Platform and the Severn Basin.

The Ardley Member. The type section of the Ardley Member is taken at Ardley
Fields Farm Quarry (text-fig. 4) on the north-eastern boundary of Oxfordshire.
The member includes all the beds given in the account of Palmer (1973) between the
hardground at the top of bed WL 5, and the base of bed WL 1 5. The top of the member
coincides with the top of the Ardley Beds of Arkell (1947), as defined in the railway

PALMER: HAMPEN MARLY AND WHITE LIMESTONE FORMATIONS 205

cutting adjacent to the quarry at Ardley. The lower part of the massive basal bed,
which immediately overlies the top clay of the Shipton Member, contains a fauna
exquisitely preserved as moulds. This is the Roach Bed of Arkell et al. (1933). The
basal bed is a well-cemented biopelsparite with a characteristic molluscan fauna
which is discussed more fully below. As stated above, it is easily distinguishable
throughout the study area standing out above the softer Shipton Member. In the
Burford region this bed has been burrowed, and takes on the appearance of Dagham
Stone (text-fig. 4). This name is given to a particular texture of hard limestone with
open crustacean burrow systems preserved (Fiirsich and Palmer 1975).

text-fig. 4. Thickness and sediment types of the Ardley Member (White Limestone Formation) in the main

study area.

The top of the member represents a considerable period of time during which no,
or minimal, deposition of sediment took place. This depositional break is frequently
marked by a hardground (text-fig. 4), which is bored and encrusted, and often has a
knobby ferruginous surface suggestive of burrowing activity before cementation
occurred. Locally, the borers and encrusters are absent, but the same knobby ferru-
ginous surface is recognizable.

206

PALAEONTOLOGY, VOLUME 22

In the north and eastern part of the study area, the depositional break is marked
by a different fabric. This is a thin bed (up to 10 cm) of finely laminated limestone.
In thin section the laminae are seen to be composed of thin fining-upwards units
(1-4 mm) entirely made up of small peloids and occasional quartz grains. The top
surfaces of the laminae sometimes support a very thin ferruginous layer, which may
be an indication of a former algal coating. The bed contains crustacean burrows
filled with faecal pellets, vertical sparry calcite filled worm burrows, and the gastropod
Valvata. This represents a similar fauna to that locally seen in the bed which caps the
Bladon Member in the north-east region, and which is supratidal in origin (Palmer
and Jenkyns 1975). However, the laminated bed at the top of the Ardley Member is
less micritic and does not show desiccation shrinkage. The similar vertical burrows,
and the presence of laminations which are possibly algal, suggest that it may have
formed in a zone subject to periodic emergence. The local presence of Valvata
towards the top suggests periodic freshwater saturation which may have been supplied
by rainfall. This laminated bed is itself capped by a hardground, and encrusted by
large flat oysters ( Liostrea wiltonensis) at Great Rollright, and by small Nanogyra at
Stratton Audley. The bed is also well developed at Temple Mills, Whiteways, and
Shipton (text-fig. 4). Whereas both the hardground and the laminated bed are repre-
sentative of the sedimentation break, the latter is also suggestive of emergence and
is evidence for a shallowing, in the north and east of the study area, towards the
London Landmass.

A shallowing in this direction is suggested by other features as well. There is a
greater proportion of clastic material in the limestones of the north-eastern region
(Odling 1913; Martin 1958). At Stratton Audley the top bed of the member is locally
cut by small channels, up to 4 m across, which contain a basal lag of large pieces of
driftwood. Channels of this sort are not seen in the Ardley Member to the south-west,
and further indicate nearby land.

In the south-west region of the study area the hardground which marks the top of
the member at Stonelands and Whitehill quarries lies only just over 2 m above the
base of the massive bottom bed of the member (text-fig. 4). At Eton College quarry
the same massive bottom bed occurs, and the top of the member, although not now
seen, is probably represented by the ‘marked plane of erosion’ 3-2 m above the base
of the member, and commented on by both Arkell (1931) and Worssam and Bisson
(1961). These thicknesses compare with 4-55 m at Asthall, 4-8 m at Fisher’s Gate,
5-8 m at Layshill Wood, and 4-55 m at Ardley (text-fig. 4). It seems, therefore, that
the Ardley Member shows a marked thinning at Whitehill and Stonelands which may
be due to erosion and reworking of the sediment. At Stonelands the hardground at
the top of the member truncates the fossils embedded in the matrix, clearly indicating
some compaction or early cementation followed by an erosive period. Similarly,
there is clear evidence of reworking at Whitehill, where the cross-bedded biopel-
sparite immediately below the hardground contains large angular clasts of compacted
marly limestone which were presumably torn up from nearby during some disturb-
ances such as a storm. Probably these trends are again associated with the passage
westwards from the London Platform into deeper water.

Over the study area as a whole, examination of the rocks both in the field and in
thin section allows three principal lithofacies to be recognized.

PALMER: HAMPEN MARLY AND WHITE LIMESTONE FORMATIONS 207

1. A well-sorted pelletal and skeletal lime sand, locally with ooliths, and contain-
ing virtually no lime mud. Except in the case of roach beds, this facies is invariably
cemented into hard massive units by a coarse sparry calcite cement which also replaces
the predominantly aragonitic fauna. The peloids and shell fragments often show a
well-developed micrite envelope or a ferruginous pellicle which suggests that they
were rigid discrete units, rather than soft compacted ones, and the original soft sedi-
ment is interpreted as being loose. This facies predominates in the south-west of the
region, where the member shows marked thinning; in this region, cross-bedding is
common. Over the rest of the study area this facies is characteristic of the lower beds
of the Ardley Member, and cross-bedding is generally absent. Locally beds of this
facies are made up of thin (5-mm) fining-upwards units which may be storm settle-
outs (Klein 1965).

2. A poorly sorted pelleted lime mud, which locally consists of scattered peloids
and shell fragments in a matrix of lime mud, and elsewhere shows distinct pelleting
of the lime mud itself. These peloids are often less distinct than those in facies 1, and
do not always have the thin ferruginous pellicles. Shells and shell fragments replaced
by sparry calcite are common. The matrix of these beds varies between micrite,
sparite, and microspar which may represent recrystallization of original micrite. The
original texture of the soft sediment is interpreted as having been more sticky than
the lime sand considered above, and many of the discrete micrite peloids seen in
sections of this facies were probably more or less soft, rather than rigid particles.
The top surface of the soft sediment may have been somewhat soupy. This facies is
more common towards the top of the member (text-fig. 4).

3. A shelly micrite, with associated small shell fragments and peloids. In thin
section this facies is always composed of a micritic groundmass, and is locally poorly
cemented. In places the micrite is compacted into pellets, but these are always the
discrete microcoprolites ( Favreina decemlunulatus) of crustaceans, and have a charac-
teristic shape and structure (Kennedy, Jakobson, and Johnson 1969) not seen in the
micrite peloids of the pelleted lime mud facies. They are also frequently confined to
burrow systems. The soft sediment of this facies is interpreted as having been firm and
fairly stiff. It supported a rich epifauna (see below), which could not have thrived
if the lime mud was of a soupy texture. Locally, as at Kirtlington and Wood Eaton,
this facies contains cross-bedded units in which the millimetre scale laminations are
picked out as alternations of shelly debris and fine lime mud. The shelly micrite
facies is only seen at Wood Eaton, in the Cherwell valley (Shipton and Kirtlington),
and at Northbrook Farm. It occurs at the top of the member, forming what is
probably a continuous unit over this region (text-fig. 4). It sometimes occurs above
beds of the pelleted lime mud facies (as at Kirtlington), and sometimes replaces them
(as at Wood Eaton).

The Bladon Member. The Bladon Member is the topmost subdivision of the White
Limestone Formation. Arkell (1931) did not appreciate the autonomy of this unit,
and ascribed the beds here included in it to the sublithographic facies of the Kemble
Beds, the fimbriata-waltoni Beds (in the region of the Cherwell valley), and the
Middle Epithyris Bed. Later, Arkell (1947) revised the classification and regarded the
latter two units as within his Bladon Beds (subzone of Aphanoptyxis bladonensis).

208

PALAEONTOLOGY, VOLUME 22

which also included locally beds here taken as the top of the Ardley Member (e.g.
Arkell 1947, p. 57).

It is because of Arkell’s use of the term Bladon Beds, that the name is retained and
given member status here. The type section is therefore taken to be at Old White
House Quarry, Bladon, and includes beds 2 to 7 inclusive of Arkell (19336). In the
Cherwell Valley (e.g. McKerrow, Johnson, and Jakobson 1969) this includes the beds
between the top of the Upper Epithyris Bed and the base of the Middle Epithyris Bed.
Principal variations in facies and thicknesses are shown in text -fig. 5.

[5H Algal laminae and birdseyes
I I Unfossiliferous micrite
SS§] Fossiliferous micrite
E44I Lime sands
HH Marl
EM Clay
tWd Cross-bedding
V& Colonial corals
Hardground

-L^L Top I of member
■cm Bottom ) not seen

Top [ of member
Bottom ) seen burford

BLADON MEMBER

Great

Rollright

N Bladon Member probably
' developed as clays and
lagoonal micrites
' indistinguishable

from Forest Marble

Croughton ^For™.a,ion in
\,

. Stratton
Ardley q-FAudley \

fe°?¥| \

Whitehill

Stonelands

Eton College

text-fig. 5. Thickness and sediment types of the Bladon Member (White Limestone Formation) in the main

study area.

The stratigraphic limits of the Bladon Member are easily recognizable. At the
bottom it is nearly always seen resting on top of the hardground or non-deposition
surface which caps the underlying Ardley Member. At the top the sediments of the
overlying Forest Marble Formation usually consist of thick pure clays or cross-
bedded shelly oolitic limestones with thin clay partings, which stand out in contrast
to the underlying micrites of the Bladon Member. Furthermore, in the north-east
part of the study area the top bed of the Member is developed as an algally laminated
micrite with birdseyes and desiccation cracks (Palmer and Jenkyns 1975) which is
quite unmistakable and totally unlike any other bed in the Great Oolite. In the
Burford region, where the Bladon Member becomes cross-bedded and oolitic, the

PALMER: HAMPEN MARLY AND WHITE LIMESTONE FORMATIONS 209

top is usually marked by a hardground. Above this the oobiosparites of the Forest
Marble contain more oyster debris than the underlying beds. Abundance of oysters,
however, cannot be used to distinguish Forest Marble from Bladon Member over the
whole of the study area.

The lithologies represented in the Bladon Member are more varied than those seen
in the lower members of the White Limestone ; nevertheless their geographic distribu-
tion fits in to form a concise and simple pattern of marginal marine sedimentation
and palaeogeography. The laminated birdseye bed mentioned above indicates
supratidal deposition. The genesis, distribution, and palaeogeographic implications
of this bed (called ‘unfossiliferous Cream Cheese Beds’ by earlier workers) have been
fully discussed by Palmer and Jenkyns (1975) and do not need stating again. The
distribution of this formerly emergent island barrier, however, is the key to the
processes controlling sedimentation at the top of the White Limestone and through-
out the Forest Marble Formation.

Wherever the laminated micrites of the emersion horizon are seen, the laminations
diminish downwards, and the bed becomes fossiliferous. At the same time, the sedi-
ment contains more shell fragments, and may locally become pelleted so that it
resembles the pelleted lime mud facies of the Ardley Member. It is in this facies,
immediately underlying an horizon which was clearly deposited in supratidal condi-
tions, that one would expect to find indications of intertidal conditions if such existed.
I have found none in spite of careful searching. The fossiliferous micrites at this level
do show bioturbation mottling, but the absence anywhere of signs of tidal activity or
intertidal exposure are taken as strong evidence against the Great Oolite sea having
exhibited significant diurnal tidality. This is discussed more fully below.

The unlaminated fossiliferous micrite was probably deposited as a somewhat soupy
lime mud in less than a metre of water. Locally (as at Bladon, Long Hanborough
Station, Temple Mills), this facies forms the top of the Bladon Member, and the
emersion horizon is absent (text-fig. 5).

In addition to the tendency for coarser-grained sediments to predominate in the
lower part of the Bladon Member, there is also a tendency for the sediments to become
coarser passing south-westwards. Massive biopelsparites with subsidiary ooliths
occur in the lower Bladon Member at Layshill Wood, Fisher’s Gate, and Minster
Lovell. Further towards Burford the proportion of ooliths increases, and cross-
bedded calcarenites of varying composition make up most of the member. These
cross-bedded calcarenites are well developed at Whitehill and Stonelands quarries
where they have been called Kemble Beds (e.g. Martin 1958). In thin section, well-
developed oolitic structures are not usually seen clearly. On close scrutiny, however,
some of the grains show a poorly developed system of concentric bands, sometimes
round a nucleus. They are interpreted as ooliths which have undergone diagenetic
micritization, with almost total loss of their original concentric laminations, in a
process similar to that described by Shearman et al. (1970). They appear to have been
deposited on submarine shoals deposited by the agency of periodic currents sweeping
from the north-west (Palmer and Jenkyns 1974). Locally these shoals emerged, and
subaerial cementation with early dissolution of the aragonite fauna occurred. Such a
hardground is seen capping the Bladon Member at Stonelands quarry.

Beds containing abundant branching corals occur in the central part of the study

210

PALAEONTOLOGY, VOLUME 22

area (see text-fig. 5). One of these beds, the Upper Epithyris Bed of Arkell (1931),
forms a constant horizon at the top of the Bladon Member, in the region of the Cher-
well Valley. This is the same region as that to which the fimbriata-waltoni Beds are
confined (see below). This coral bed is the lateral equivalent of the emersion horizon.

The corals appear to have colonized firm substrates containing shell material, and
then to have grown up in arborescent growths above the surface of the substrate. As
in the coral beds of the Shipton Member, regions of still water persisted down amongst
the stems of the corals, and dead shells and micrite accumulated there.

McKerrow et al. (1969) assumed that beds of this facies at Kirtlington represented
a transported assemblage. This, as Allen and Kaye (1973) point out, is contrary to
the evidence of (i) abundant micrite ; (ii) the geopetal fills in most of the articulated
brachiopods; (iii) the concave-up position of most of the single Epithyris valves.
Allen and Kaye’s analogue of Florida back-reef corals is equally mistaken : the quiet
conditions amongst the fronds are much more analogous to conditions in Florida
Porites shoals where mud and shells similarly accumulate due to baffle action.

Well developed in the quarries of the Cherwell valley, and best seen at Shipton
quarry, is a thin sequence of fine-grained elastics and associated marls called the
fimbriata-waltoni Beds by Arkell (1931), and the fimbriatus-waltoni clays by
McKerrow et al. (1969).

At Shipton, this horizon can be followed along the whole of the south-west face,
and can be seen to vary between thicknesses of 0-8 and 1-9 m. McKerrow et al. (1969)
noted that it appeared similarly continuous at Kirtlington. On close inspection,
however, this horizon at Shipton is seen to be made up of complex interdigitation
of laterally impersistent clay beds, some of which are lens-shaped when seen on a
vertical face. Disturbed contacts between these beds occur locally and suggest that
deposition of some of the clays was also local and catastrophic. The coarser silty
beds are confined to channels within the clays, and are frequently lined by a drift-
wood lag. Some of the clays contain evidence of erosion in the form of stringers of
carbonate pellets or larger pebbles of micritic limestone of a similar lithology to the
‘Cream-Cheese’ limestone mentioned above. Klein (1965) considered these to have
been caliche nodules formed by in situ growth of calcium carbonate, forming irregular
masses. Many of these nodules, however, contain shells and microfossils, and are
locally clearly peloidal in texture rather than uniformly crystalline or banded as is
real caliche (Bernard, Le Blanc, and Major 1962). The rounded, knobbly shape of the
pebbles is atypical of eroded clasts, and this has probably contributed to their not
being recognized in the past. This shape may be due to partial nodular cementation
within the bed from which they were originally derived, so that only the nodules
withstood reworking.

The fimbriata-waltoni clays are interpreted as having been deposited under pre-
dominantly quiet-water lagoonal conditions, subject to periodic current activity and
influx of new sediment, perhaps during storms. The impersistent nature of the clays,
and the local development of shelly marls suggests that the environment was hetero-
genous, probably a complex system of marsh and creeks, with local carbonate mud
mounds, the precise conformation of which was always changing. This picture is
supported by faunal evidence of brackish water (see below) and driftwood. Orienta-
tion of this driftwood indicates origin of the clays from the north-east (Allen and

PALMER: HAMPEN MARLY AND WHITE LIMESTONE FORMATIONS 211

Kaye 1973); the direction of the London Landmass, from which one would expect
freshwater influence to come.

Fauna of the White Limestone Formation

The nature of the fauna in the White Limestone was naturally strongly influenced by
the nature of the soft substrate. This is true both for those assemblages dominated
by infaunal species, of which the vast majority are bivalves and gastropods, and also
for the sessile epifaunal forms. The latter are predominantly found in those micrites
or muddy limestones which are interpreted as having been coherent, pasty, and firm
prior to lithification. Loose, unstable lime sands (now forming sparites), and certain
clays and lime muds which were probably soupy when forming the sea floor, contain
few sessile epifaunal species. Their faunas are dominated by burrowers, the morpholo-
gies of many of which can be interpreted as being directly adapted to coping with the
exigencies of the particular enclosing sediment.

The following account of the principal faunal communities recognizable in the
White Limestone is based on field observation over several years of hundreds of beds
of the principal lithological types discussed in the previous section (see Palmer 1974).
Where they have been readily available, bulk samples of the faunas typical of the
various lithologies have been made. These were prepared in the laboratory and
analysed quantitatively. The details of some of these bulk collections are shown in
Table 2, and serve as examples, rather than as rigid descriptions, of the communities
herein recognized.

Coral Beds. The most conspicuous of the White Limestone communities is provided
by the beds with abundant colonial corals seen in the Shipton and Bladon Members
(see text-figs. 3 and 5). This community is absent from the Ardley Member in the
study area. These coral beds are soft or well-cemented limestones, and the corals
themselves are invariably replaced by sparry calcite, so that only their superficial
features are observed. In previous accounts they have usually been referred to species
in either of the two genera Thamnasteria and Isastrea, according to whether they
have, respectively, confluent septa with beaded margins, or distinct corallite walls
and entire septa. McKerrow et al. (1969) recognized that intermediate forms occurred
in the White Limestone, and that corallites of apparently different genera sometimes
appear on the same specimen. Since the fine details of septal structure on which these
two genera can best be distinguished (Wells 1956) are never preserved, it is best to
consider the distinction between them as not proven. I have therefore not separated
them in this account.

The total fauna of the coral beds may be subdivided into five different ecological
groups :

1. Frame builders: Thamnasteria and/or Isastrea predominate in this role. Other
corals occasionally occur, belonging to two genera : Cyathophora, a massive form with
a well-developed coenosteum, is found in the Shipton Member at Ardley, Wood
Eaton, and Shipton, and the solitary Montlivaltia is found in the Shipton Member
only, at Milton and Shipton.

PALAEONTOLOGY, VOLUME 22

The colonial corals usually show a branching, dendritic growth form suggestive of
quiet-water conditions, but massive knobbly heads also occur. The corals in the base
of the Bladon Member at Shipton, however, grow as thin platy laminae over the
surface of shell debris. This form is suggestive of the platy growth form of Recent
Montastrea annularis , in which it is a light gathering response restricted to darker
environments (Graus and McIntyre 1976). It is not clear whether a similar control in
the base of the Bladon Member might have been a response to greater depth or
greater turbidity than higher up in the unit.

2. An organohesive fauna: this term is coined to cover sessile animals which are
attached to the substrate by organic tissue, which decays on death, so that the shell
may not finally be buried in life position. The main groups falling into this category
are pedically attached brachiopods, such as a small Epithyris sp. which occurs in all
the Shipton Member coral beds, and byssally attached nestling bivalves, of which
Eonavicula minuta and Plagiostoma subcar diiformis are the most common. In the
Bladon Member, Epithyris sp. is absent, and the same niche is filled by the larger,
more plicate E. oxonica. It occurs in vast numbers in all growth stages. Individuals
often show interference in their growth caused by crowding, and large specimens
may have up to thirty sites of attachment of younger individuals ( Podichnus ) on their
surfaces. They presumably lived in clusters amongst the coral fronds. Rhyncho-
nellids also occur, but are much more common in the Shipton than the Bladon
Member.

3. A boring and encrusting fauna: predominantly of small and inconspicuous
species which are obligate hard substrate dwellers, relying on pieces of coral rubble
and shell material for attachment ; the main encrusting species are the thecideacean
brachiopod Moorellina, the ectoprocts ‘ Berenicea ’ and Stomatopora dichotoma,
Serpula (Cy closer pula) sp. and small calcisponges. These forms probably lived well
down in sheltered microenvironments amongst the coral rubble (see Palmer and
Fursich 1974). Praeexogyra hebridica is also common.

Coral rubble, when found in the White Limestone Formation, is often heavily
bored by bivalves, whose mud- and limestone-filled crypts can easily be collected
complete from the surrounding mass of sparry calcite which replaces the coral. The
principal borer of coral debris in the Shipton Member is not Lithophaga , as generally
supposed, but a hitherto unrecognized species of the gastrochaenacean Gastro-
chaenopsis. Members of this superfamily of the Myoida have not been widely recorded
below the Upper Jurassic, although they also occur in the Sharp’s Hill Formation
(Lower Bathonian), and the Lincolnshire Limestone of Bajocian age.

4. A vagile epifauna: this consists mainly of Acrosalenia and ‘ Pleurotomaria"
(usually too poorly preserved for a closer identification), together with other gastro-
pods. These were the grazing and predatory species, which had to search for their
food, rather than let it come to them as did the predominantly filter-feeding groups
(2 and 3). Other likely predators and scavengers are fish and decapod crustaceans,
of which small unidentifiable fragments are often found.

5. A burrowing and ploughing fauna, which occupied patches of soft sediment
beneath and between the patches of coral and coral rubble, but to which the corals
and other members of the coral community were not a life requirement. This group
includes Pholadomya, Cuspidaria, Protocardia , Clypeus, Globularia, and the occa-

PALMER: HAMPEN MARLY AND WHITE LIMESTONE FORMATIONS 213

sional nerineid gastropod; all of them are also frequently encountered in non-
coralliferous beds of the Shipton Member.

A complete faunal analysis of representative Shipton and Bladon Member coral
beds is given in Table 2. In addition to the Epithyris mentioned above, the main
differences between the communities in the two members lie in the presence of
Isognomon isognomonoides and Lopha costata, both of which are common in the
Shipton Member (also in non-coralliferous beds) but absent at the higher level. On
the other hand, the Epithyris in the Bladon Member often provided attachment sites
for the Anomia- like pulvinitid bivalve Hypotrema which does not occur in the Shipton
Member and which has never before been recognized in the Great Oolite.

The coral beds of the White Limestone Formation are not reefs in the modern sense
of the word : they do not have the relief of either the fringing reefs or the patch reefs
of modern tropical seas. Their closest equivalent is the inshore coral shoals which
are common in water less than 3 m deep along the seaward edge of the Florida Keys
(Ginsburg 1964). The principal corals of these shoals are Porites divaricata and the
rose coral Manicina areolata on sandy bottoms, and Siderastrea radians and Solen-
astrea hyades on rocky bottoms (Ginsburg 1964). Solenastrea is remarkably similar
in appearance to the Bathonian Cyathophora , both having raised corallites separated
by a vesicular coenosteum (Wells 1956). Similarly, Porites and Manicina of the
Recent, bear a superficial resemblance to Thamnasteria/ Isastrea and Montlivaltia
respectively. Porites is either massive or branching, with small, closely united coral-
lites characterized by more or less weakly defined synapticular rings (Wells 1956),
all of which features it has in common with Thamnasteria; Montlivaltia resembles
Manicina in that both are large solitary corals of the superfamily Faviicae. It may be,
therefore, that members of this particular lineage have occupied similar niches since
at least the Bathonian. Manicina can clear loose sediment off its surface (Ginsburg
1964), which is an obvious adaptation to life on a sandy substrate in shallow turbulent
seas. Such an adaptation would also have benefited Montlivaltia, which was likely
to have been subjected to the same problem.

Neither the colonial corals (Isastrea/ Thamnasteria) nor the solitary Montlivaltia
seem to have needed an extensive hard substrate for initial attachment, and they are
found in the White Limestone Formation entirely surrounded by sediment in which
the most extensive hard surfaces are shells and shell fragments. Probably initial larval
settlement took place on a shell fragment. The shell fragment became superfluous as
the coral grew and became stable due to its own weight, or to support from adjacent
colonies and rubble. Cores through Recent but smothered Porites shoals from
Florida, show a very similar array of coral rubble and associated fauna (see below)
entirely surrounded by lime sand, lime mud, and shell fragments (pers. obs.).

Observations on the accessory faunas associated with Porites shoals show that
these faunas are very similar in life habits to those found in White Limestone coral
beds. Barbatia cancellaria (Lamarck) is often abundant amongst the terminal
branches of Porites colonies, attached by a weak byssus which may be released if
disturbed : B. domingensis (Lamarck) is gregarious, but prefers the lower region of the
Porites colonies, amongst the dead branches ; Isognomon radiata (Anton) also prefers
the region of dead branches; Pinctada radiata (Leach) occurs rarely between
branches; and Lima scabra (Born) is loosely attached either within or on top of

214

PALAEONTOLOGY, VOLUME 22

Porites colonies, and may detach itself and swim away if disturbed (Stanley 1970).
The Bathonian equivalents to these species, based on morphology and inferred life
habits are Eonavicula minuta ( =Barbatia ), I. isognomonoides (=/. radiata), I. ooliticus
(=P. radiata ), and Plagiostoma subcar diiformis (=L. scabra ). In addition to Stanley’s
information, Coogan (1971) records Recent Lithophaga boring in dead coral (=Litho-
phaga and Gastrochaenopsis in the White Limestone). It is locally abundant amongst
the dead rubble of Porites shoals, together with at least three species of serpulid,
several ectoprocts, Homotrema , hydrozoans, Clypeaster rosaceus, and shells of
several burrowing bivalves (pers. obs.).

These Recent Porites shoals and their associated fauna occur in water down to
about 3 m deep in Florida (Ginsburg 1964). To the south of Rodriguez Bank where
they are particularly well developed, the lower limit of the most abundant growth
is about 1-3 m in depth. The upward limit of growth is low- water spring-tide level
(Turmel and Swanson 1964). A similar depth range is envisaged for the Bathonian
coral shoals.

The one great difference between Recent and Bathonian shoals is the apparent lack
of algae in the latter. Today, Porites shoals are often accompanied by abundant
green and red algae, such as Halimeda opuntia and Goniolithon, which incidentally
are both major producers of medium and coarse sediment in the shoals. In none of
the thin sections made for this study is there any obvious trace of any codiacean or
dasycladacean, even though both groups go back to the Cambrian (Wray 1971) and
they have a characteristic texture in thin section, even after replacement of the original
aragonite by calcite (Stanley 1966). The only alga found in the White Limestone is
Solenopora jurassica, the long-ranging rhodophyte, thought to have been the ancestor
of some of the Recent corallines. In the study area, it is confined to the Shipton Mem-
ber (text-fig. 3). Red algae can utilize the blue end of the light spectrum whereas green
algae need the red end, which penetrates less far into sea water; therefore the red
algae can tolerate deeper and more turbid water. Since we are postulating a depth of
only a few metres for the coral shoals of the Shipton Member, however, depth alone is
unlikely to have been the controlling factor. It may be that a high proportion of fine
clastic material derived from the land to the north-east tended to make the shallow
water murky.

Muddy lime sands. Where corals are absent from the muddy limestones of the Shipton
Member, there is a reduction or absence of several of the encrusting, boring, and
nestling forms (e.g. calcisponges, Cycloserpula, Moorellina , Kallirhynchia concinna,
E. minuta , Epithyris sp., Gastrochaenopsis ). Nevertheless, the sedentary epifauna is
still quite varied and again indicates a stable sea bottom, probably littered with dead
shells and soft-bodied organisms which provided attachment sites for a few cementers
and a range of byssally attached Pteriomorphia. The principal species found are
Praeexogyra hebridica (cemented, but never forming reefs as in the Hampen Marly
Formation), and I. isognomonoides (byssally attached). Lopha costatum (cemented),
Girvillella ovata, Camptonectes annulatus, C. rigidus (all byssally attached) and
Epithyris sp. are also common. Streamlined, alate forms, such as Pteroperna costatula
occur only occasionally. From analogy with the Recent Pteria, whose shape is very
similar, it might be concluded that this form lived in exposed water perhaps amongst

PALMER: HAMPEN MARLY AND WHITE LIMESTONE FORMATIONS 215

the fronds of soft-bodied animals and plants which are not preserved. The rarity of
P. costatula may reflect rarity of this habitat. The inequivalve condition of G. ovata
and Costigervillia crassicosta (which occurs only rarely in the Shipton Member), and
the unstreamlined shape of the other byssate forms mentioned above, suggests that
most of the community lived close to the substrate, or in cracks and fissures (cf.
Kauffman in Moore (ed.) 1969; Stanley 1970).

The other main faunal component of the non-coralliferous Shipton Member muddy
limestones consists of semi-infaunal and infaunal species. Some of these species also
burrowed amongst the rubble in the coral beds, but always in much fewer numbers.

The semi-infauna is principally represented by Modiolus imbricatus and
M. ( Inoperna ) plicatus (as judged on the criteria discussed by Stanley 1970). The
main infaunal species is the cyprinid Anisocardia loweana, a smooth, streamlined form
which was probably shallow burrowing and active. Other common shallow burrowers
include Trigonia pullus, Pseudotr ape sium cor diforme, Quenstedtia sp.. Protocar dia
stricklandi, and P. lycetti. The relative abundances of these species vary considerably,
not only from bed to bed, but also in different parts of the same bed, in what appears to
be a uniform sediment. This means either that control of distribution was to some
extent fortuitous (spatfall just happened to take place in a particular region at a
particular time), or that it was controlled by micro-environmental variations of the
sort that are difficult to observe in living systems, and quite impossible in ancient
situations; for example, minor differences in chemical composition of the seawater
(see Wilson 1951; Cooper 1951). It may be for these local reasons that relative
abundances of species in these Bathonian shallow marine communities appear so
variable: nevertheless, certain species can be recognized as characterizing certain
habitats and substrates.

Deep-burrowing Pholadomyoida also occur abundantly in the Shipton Member,
whereas they are rare except at certain restricted horizons in the higher members.
They include Pholadomya, Goniomya, Homomya, Gresslya, and Pleuromya, all of
which have large pallial sinuses, and are almost invariably found in life positions. The
first three of the above-mentioned genera are similar in form to Recent Mya arenaria,
which shows a preference for muddy sand substrates, and is intolerant of strong
currents and waves (Stanley 1970 ; Swan 1952) ; such an environment would seem from
the sediments to fit much of the Croughton Member. M. arenaria is a slow burrower,
and is subject to mass-mortality during storms, since the adult form can neither
reburrow easily if exposed nor clear material which plugs its siphons (Weymouth
1920). It is likely that plugging of siphons was a major cause of mortality of those
forms found in life position in White Limestone sediments. Pleuromya and Gresslya,
being smaller and more streamlined, are more likely to have been able to excavate
themselves if smothered in adulthood.

The non-bivalve fauna is represented by high-spired gastropods, Clypeus muelleri,
and open crustacean dwelling-burrows of the Spongeliomorpha- type which frequently
contain the faecal pellets of the species which lived in them. Such open burrows in the
White Limestone are again limited to beds with a high enough proportion of fines to
have rendered the substrate pasty and not liable to collapse. Clypeus also appears to
have demanded such substrates. It is usually found within or on top of beds of muddy
lime sand, and which probably ploughed through the soft sediment, either just below

216

PALAEONTOLOGY, VOLUME 22

or at the sediment-water interface, detritus feeding in the way Clypeaster rosaceus
does today. It is absent from sediments which are not muddy, and is consequently
highly characteristic of the Shipton Member in the study area.

Clean washed lime sands. The well-developed Roach Bed at Ardley yields an abundant
fauna which is typical of the well-sorted biopelsparites which are particularly
characteristic of the lower part of the Ardley Member, and which also occur in the
Shipton Member. This is shown in Table 2. The only typical form absent from the
Roach Bed, but common in most other beds of this facies, is the button-coral Choma-
toseris. Apart from this coral, the fauna of this facies is dominated by molluscs. The
most abundant single species is the high-spired Nerinella scalaris (d’Orbigny), but
other high-spired nerineids also occur. These are interpreted as having been sedentary
burrowing filter feeders, similar to Recent Turritella. The most extreme of these forms
is Bactroptyxis bacillus which starts to occur in the lime sand facies of the Ardley
Member in the region west of Witney and becomes common in the White Limestone
around Cirencester. B. bacillus is the ultimate in slender high-spired gastropods, with
a spiral angle of only 4-5°, and a length of up to 15 cm. As in other nerineids, the
shape is extremely cumbersome for a mobile form and it seems most likely that the
shell acted as an anchor. That this extreme form becomes more common away from
the Oxfordshire shallows, into the more open-sea current-affected lime sands of
the Cotswolds, supports this hypothesis. The longer shell was a more efficient
anchoring system, more appropriate to these energy conditions than the somewhat
shorter and fatter shells of the nerineids more common in Oxfordshire. These forms,
in contrast, become rarer south-westwards as B. bacillus becomes more common.

The forms associated with the nerineids are primarily bivalves. The principal
species are Eocallista antiopa and Protocardia stricklandi. Also common are Vaugonia
moretoni, Corbis lajoyi, Lucina bellona, Pseudo trapesium cordiforme, Protocardium
buckmani, and Pleuromya uniformis. Less common, but also occurring are Trigonia
pullus, Quenstedtia bathonica, Parallelodon hirsonensis, and Modiolus imbricatus.

The most striking feature of this fauna is the rarity of epifaunal forms ; the pectina-
ceans which are relatively common in the Shipton Member and Bladon Member, and
also in the micritic facies at the top of the Ardley Member, are very rare. If M. imbri-
catus was semi-infaunal as discussed above, then the only epifaunal species which
characterizes this facies is the occasional large P. hirsonensis. This, again, is probably
a reflection of the loose nature of the substrate, which was susceptible to current scour.

The burrowing bivalves considered above exhibit a variety of morphologies from
smooth streamlined species like the very common E. antiopa , to inflated forms like
C. lajoyi and heavily ornamented ones like V. moretoni. The more streamlined species
were probably rapid burrowers, perhaps moving actively within the sediment,
whereas the more inflated or ornamented forms were probably slow, inactive
burrowers. Today, slow burrowers are not found in lime sands which are readily
affected by currents, and only fast burrowers which can reburrow quickly if exhumed
by current scour, can thrive (e.g. the Tivela biofacies of Newell, Imbrie, Purdy, and
Thurber 1959). This suggests that the lime sand facies of the Ardley Member was not
subject to frequent high current activity. On the other hand, there is no stabilizing
micrite or fine mud in the sediment, and the deep-burrowing anomalodesmatans and

PALMER: HAMPEN MARLY AND WHITE LIMESTONE FORMATIONS

217

burrow-dwelling crustaceans appear to have found conditions unsuitable. Possibly
gentle winnowing and occasional more severe storm currents prevented accumulation
of fines and kept the sediment loose.

It was beds of this facies at the base of the Ardley Member which Arkell (1931)
recognized as his Nerinea eudesii beds. They are by no means confined to this level,
however, and also occur widely in the Shipton Member of the study region.

Cross-bedded lime sands. In the west of the study region (around Burford) some units
within the Ardley Member, and most of the Bladon Member, become cross-bedded.
This is seen as being the result of a slight deepening and increase of current activity
towards the edge of the Oxfordshire shallows. Further into the shallows the develop-
ment of banks and shoals dampened current activity. The cross-bedded units are
often oolitic, and massive oolites also occur in the lower part of the Bladon Member
at Bladon.

Today, oolites and other cross-bedded lime sands contain a conspicuously reduced
fauna because so few species can cope with the exigencies of a constantly shifting
substrate. This appears also to have been so in the White Limestone. The cross-
bedded units locally contain lag deposits of bivalves and high-spired gastropods
which are obviously not in situ , but which probably constituted the more stable lime
sand community discussed above during occasional periods when current activity
was reduced. The indigenous fauna consists of occasional large Purpuroidea and a
species of pectinid with an umbonal angle of 110° and symmetrical auricles. This
reduced fauna is reminiscent of that of shifting oolite dunes in the Bahama region
today where the large herbivorous gastropod Strombus samba (morphologically
similar to Purpuroidea) is the principal species (Newell et al. 1959). The pectinid is
clearly a swimming form (Stanley 1970, fig. 11), and could presumably have moved
away from encroaching sediment.

Where the tops of the cross-bedded units remained undisturbed for a period, an
infauna became established, and a stabilizing fine-grained component was intro-
duced into the otherwise loose lime sand. An example of this, with a rich and well-
preserved fauna, is seen at the top of the Bladon Member at Stonelands Quarry
(Table 2). The lithology of this bed resembles the micritic and slightly clayey lime
sands of the Shipton Member at Croughton and Ardley and it contains some of the
same burrowing fauna and semi-infauna as these beds ( M . imbricatus, Protocardia
lycetti, Anisocardia loweana, Pseudotrapesium cor diforme, Pleuromya sp., Phola-
domya sp. and crustacean burrows). Also present are some of the species found in
the looser lime sands of the Ardley Member, including V. moretoni and L. bellona.

Pelleted lime mud. The pelleted lime mud lithofacies is regularly developed in the
Ardley Member above the lime sands at the base, and also in the fossiliferous micrites
in the upper parts of the Bladon Member. The pellets (or peloids) are clearly seen
where protected from compaction within shells, but are usually squashed together in
the matrix of the rock to give a more or less uniform micritic appearance. As a soft
sediment, its mechanical properties are interpreted as having been intermediate
between those of the loose, clean washed lime sand, and the stable pasty sands with
abundant fines which are common in the Shipton Member.

218

PALAEONTOLOGY, VOLUME 22

The fauna of this lithofacies is dominated by high-spired gastropods of the genus
Aphanoptyxis. A. ardleyensis is the dominant species in the Ardley Member, but this
is replaced by the slightly blunter A. bladonensis in the Bladon Member. These are
again thought to have been sedentary infaunal filter feeders.

Unfortunately, there is no single locality within the Ardley Member at which good
preservation allows a close analysis of all those species which occur with A. ardley-
ensis. However, a picture of the main components of the community can be built up
from observations over the whole study area.

The associated fauna includes some of the species present in the lime sand facies.
Trigonia pullus and L. bellona are occasionally found. More common, but never
occurring in the vast abundance of A. ardleyensis , is a species of Cossmannea or
Nerinella, easily recognizable in cross-section from its internal ribs. In addition,
however, there is a fauna which is absent from the lower beds : this consists of Corbula
hulliana, Eomiodon fimbriata, Bakevellia waltoni, and small vertical worm burrows.
Species from the lime sands which do not persist into these higher beds are Parallelo-
don hirsonensis, Corbis lajoyi, Protocardia stricklandi, V. moretoni, and Chomato-
seris sp.

In the pelleted lime mud facies it is the infauna which again predominates;
B. waltoni is the only common epifaunal form. The burrowing bivalves are principally
ornamented forms, and were probably shallow burrowing and rather inactive. The
two principal incomers Corbula hulliana and E. fimbriata, both become abundant in
the micrites and clays of the overlying Bladon Member, and seem to show a marked
preference for fine-grained sediments. The asymmetry of the valves in Recent
Corbula has been interpreted by Yonge (1946) as an adaptation to this sort of environ-
ment. When the valves close, it results in complete collapse of the inhalent chamber,
and hence complete voiding of pseudofaeces. Such behaviour may be contributory to
the pelleted nature of the facies under discussion.

Deep-burrowing bivalves are again rare in the micritic facies, although they may
occur abundantly in interbedded thin sandy or marly beds. Bed WL7 at Ardley
Fields Farm quarry (see Palmer 1973) is a good example of this: it contains a large
number of small Pholadomya ? lirata in life position, and occasional Globularia
which probably preyed on them.

This facies of the White Limestone Formation contains few body fossils other than
molluscs, and some of the mollusc species may also be implicated as euryhaline in
other Great Oolite deposits (e.g. C. hulliana in the Hampen Marly and Forest Marble
Formations). This may suggest slight brackish water influence in the pelleted lime mud
facies, but it is equally likely that the apparent low diversity was the result of un-
favourable substrate, rather than unfavourable salinity. The loose and sometimes
soupy texture of the sediment would have been best suited to infauna rather than
epifauna, and the apparent dominance of molluscs reflects the success of this group in
infaunal niches.

The fauna of the Bladon Member pelleted lime muds does not include any of the
common lime sand species ( T . pullus, L. bellona, Cossmannea/ Nerinella). On the other
hand, the newcomers in the Ardley Member pelleted lime muds ( Bakevellia waltoni,
Corbula hulliana, E. fimbriata ) become more common. This implies a continuum of
decreasing sediment looseness, perhaps with increasing soupiness which was also

PALMER: HAMPEN MARLY AND WHITE LIMESTONE FORMATIONS 219

unfavourable to epifauna, passing from lime sand to Ardley Member pelleted lime
muds to Bladon Member pelleted lime muds. Cuspidaria ibbetsoni (in fact probably a
corbulid) and Amberleya nodosa also occur in the Bladon Member.

Shelly micrite. The most conspicuous fossil beds in the White Limestone occur in
this facies which is confined to the upper part of the Ardley Member (text-fig. 4). The
dense white lime mud contains abundant shells in all stages of breakdown, and is
frequently pelleted into the large microcoprolite Favreina decemlunulatus (see
Kennedy et al. 1969). Favreina is often packed into large burrow systems of the
Spongeliomorpha suevica- type. The abundance of such dwelling burrow systems and
the many common epifaunal species in this facies again indicates a firm stable bottom,
similar to those in the Shipton Member discussed above.

The epifauna is dominated by the brachiopod Epithyris oxonica ; bivalves such as
Gervillella ovata, Pseudolimea duplicata , Costigervillia crassicosta , Hypotrema sp. and
Praeexogyra hebridica are also common. Semi-infaunal species are represented by
Modiolus imbricatus and M. ( Inoperna ) plicatus, and infauna particularly by Aniso-
cardia islipensis and Sphaeriola oolithica. The fauna other than bivalves includes
many ectoprocts and serpulids encrusting shell material (see Table 2). There are
also abundant terebelloid worms (McKerrow et al. 1969), represented by tubes
composed of peloids and shell fragments originally embedded in an organic matrix.
These tubes represent a large constituent of the sediment, and were presumably of
major importance in removing sand-sized particles from circulation as sediment
grains.

McKerrow et al. (1969) studied size/frequency distribution of Epithyris and
Modiolus from beds of the micritic facies at Kirtlington (their beds le, 2j, 2k, and 3k).
They concluded that articulated specimens represented an in situ fauna, but that
single valves had been transported. However, the large amounts of micrite, the
general rarity of disarticulated brachiopods, and the undamaged condition of many of
the fossils in general, suggests little transportation in beds of this facies, and for small
distances when it did occur.

In all the beds of this facies studied, the epifauna and semi-infauna forms the
dominant part of the total community, both in numbers of species and numbers of
individuals (Table 2). Some of these species (e.g. M. imbricatus ) are forms which
occur throughout the Great Oolite, wherever loose shell debris lay on or within a
firm substrate; whereas others (e.g. Hypotrema , C. crassicosta , Digonella digonoides,
terebelloid worms) seem to be either restricted to, or else show a marked preference
for, this particular micritic facies. A third group (e.g. G. ovata , E. oxonica, Campto-
nectes annulatus, M. ( Inoperna ) plicatus ) occurs in the White Limestone micrites and
more clayey beds, but is absent from the lime sands and the pelleted lime muds of the
lower part of the Ardley Member.

It is impossible to envisage all the different niches which were occupied by this
variety of epifaunal filter-feeding species. The more inequivalve bivalves are likely
to have lived close to the substrate with a horizontal plane of commissure, but it is
tempting to think that the more equivalve and streamlined forms (e.g. G. ovata ,
Pseudolimea duplicata) may have been attached to the fronds of sponges, sea-whips,
and other octocorals, which leave no trace in the fossil record, in the same way that

220

PALAEONTOLOGY, VOLUME 22

Recent Pteria colymbus (Roding) in Florida shows a marked preference for the
upper portions of alcyonarians (Stanley 1970).

The bivalve infauna is dominated by shallow burrowers. A. loweana, common
throughout the White Limestone is a streamlined form which was probably a fairly
active burrower. A. islipensis and S. oolithica are inflated forms with thick shells, and
were probably more sluggish.

The crustacean infauna, although packing its burrows with faecal pellets, does not
appear to have covered the walls of these burrows with a supporting pelleted lining.
This implies that the original micrite substrate was firm enough to support open
burrow systems, rather than of a soupy fluid consistency. This observation is sup-
ported by the greater variety of molluscan epifauna, which would easily have become
choked on a soupy substrate. Epifauna today are absent on soupy mud-mounds in
Florida Bay, except where they grow on sea-grass, above the surface of the sediment
(pers. obs.).

In the micrite facies, some of the species reach a size larger than that reached by the
same species in any other facies of the Great Oolite. M. imbricatus, for example,
reaches a length of over 70 mm, whereas 50 mm is large for other beds. Similarly,
E. oxonica grows to a length of over 50 mm, longer even than the large individuals
which occur in the coral beds of the Bladon Member. The environment, therefore,
seems to have been very equable. The geographical distribution of the micritic facies
(see text-fig. 4) the inferred nature of the substrate, and the apparent equability of
the environment support the model of a quiet, shallow, lagoonal environment, pro-
tected from open marine currents, but not so near the shore that there was any
apparent reduction of salinity. Presumably this setting was highly productive in
energetic terms, providing the algae on which the apparently large and abundant
fauna fed. This high productivity may also have resulted in high algal production
of the micrite.

At certain localities in the south-east of the study area (e.g. Kirtlington; Wood
Eaton; Whitehill Quarry, Gibralter), this facies passes laterally into cross-bedded
limestones with much broken shell debris which appears on close examination to be
derived from the shelly micrite community. The shelly micrite is envisaged as having
formed extensive stable muddy mounds on the sea floor, perhaps less than a metre in
depth. The cross-bedded units are interpreted as having filled in extensive deeper
areas (perhaps up to 5 m deep) in between the mounds. These deeper areas were
wider and more irregularly distributed than channels, but they were probably main-
tained by storm- and wind-generated currents which periodically eroded material
from the sides of the mounds, and deposited it as current-bedded units in the deeper
water.

Nearshore lagoonal clays. The fimbriata-waltoni beds of the Bladon Member in the
Cherwell Valley (text -fig. 5) contain a fauna which has much in common with that
of the pelleted micrites beneath the laminated emersion horizon. As Arkell (1931)
first pointed out, the most abundant species are Bakevellia waltoni, Eomiodon fim-
briata and Cuspidaria ibbetsoni, with occasional Protocardia lycetti. The fauna from a
locally developed bed within the fimbriata-waltoni clay at Shipton is detailed in
Table 2. However, these clays contain an additional fauna and flora which is absent

PALMER: HAMPEN MARLY AND WHITE LIMESTONE FORMATIONS 221

from the limestones. This difference lies foremost in the enormous volume of coalified
wood which is found in the clays. Most is derived from conifers which must have
grown on land. These pieces of wood reach a length of 4 m, and may be up to 12 cm in
diameter ; locally they are concentrated into layers up to 4 cm thick which constitute
impure coals. Also occurring commonly in these clays is the freshwater gastropod
Valvata.

The bivalve fauna of the fimbriata-waltoni clays does not specifically suggest fresh-
water of fresh/brackish conditions, although some species (e.g. Corbula hulliana,
Eomiodon fimbriata) have already, been implicated as euryhaline in the Hampen
Marly Formation. The association of wood and Valvata, however, strongly suggests a
freshwater influence not far away. This conclusion was also reached by Bate (1965)
who recorded the freshwater ostracods Bisulcocypris sp. A, Theriosynoecum kirtling-
tonense Bate, and Timiriasevia mackerrowi Bate from the fimbriata-waltoni clay at
Kirtlington, in association with charophytes.

The bivalves preserved in the clays consist largely of disarticulated and broken
specimens arranged on bedding planes. In contrast, material collected from the
marly limestones seen interdigitating with the clays at Kirtlington by Arkell, and
now in the Oxford University Museum, contains many articulated specimens. The
similar fauna of the fossiliferous micrite facies also includes articulated specimens,
and it seems likely that some at least of the common bivalves recorded from the clay
actually represent the fauna which lived associated with the more calcareous sedi-
ments. These marly sediments probably formed the banks and the surrounding areas
of the extensive system of creeks through which water from the land to the north-east
drained towards the sea. A north-easterly derivation is supported by the orientation
of the driftwood in the clays (Allen and Kaye 1973), and the system probably drained
through a series of slow-moving freshwater lakes and gullies in which the freshwater
fauna flourished. The general sedimentary picture of horizons of derived shells,
pellets, and micrite pebbles against a background of clay suggests some variation in
the strength of the discharge; perhaps again associated with storms, or a wet season.
Locally, silts occur in lensoid beds within the clays; the molluscan fauna of these beds
is confined to broken and disarticulated specimens of Praeexogyra hebridica, and
suggests origin from a slightly different source.

At Shipton, horizons of rootlets have been seen in the fimbriata-waltoni clays. At
Enslow Bridge (Phillips 1871) and Kirtlington (Arkell 1931), Cetiosaurus oxoniensis
has also been found. The similarity between this fauna and flora and that of both the
Hampen Marly Formation at Wood Eaton, and the Forest Marble Formation around
Stratton Audley is very striking, and further indicates a very shallow, swampy region,
channelled by creeks. The environment was probably somewhat similar to the
Everglades of Florida today.

Hardgrounds. Hardgrounds occur at the top of the Ardley and Bladon Members at
the localities shown in text -figs. 4 and 5 respectively. The faunas associated with them
are Exogyra crassa, Liostrea wiltonensis, rare Serpula (Cy closer pula) sp., and the
crypts of boring worms and bivalves. The species diversity on these surfaces is thus
much lower than that on contemporaneous hardgrounds in the region to the south-
west, e.g. that at the base of the Bradford Clay (Palmer and Fursich 1974) which sup-
ports seventeen different species of borers and encrusters. This is a reflection of a

222

PALAEONTOLOGY, VOLUME 22

general trend within the Bathonian; a marked diversity increase towards the more
stable open marine conditions to the south-west, with the incoming of stenotopic
organisms that found the exigencies of the nearshore Oxfordshire shallows unsuitable.

DISCUSSION

The interplay of two trends is responsible for the lateral and stratigraphic variation
which characterizes the White Limestone of the study area. The first is the shallowing
in the direction of the shore line of the London Landmass in the north-east part of the
region. The second is the cyclicity which gives rise to the three members.

We have already seen how the deposits and fauna of the Hampen Marly Formation
became influenced by river discharge at Wood Eaton and Ardley. Similarly, at the
south-western edge of the Oxfordshire shallows, they pass into a deeper water, more
fully marine, facies. The same is true for the whole of the White Limestone Forma-
tion. Evidence for shallowing towards the London Landmass north-eastwards
includes the following observations (member and locality in brackets where appro-
priate). There is an increase in the insoluble component of the limestones and in the
amount of clastic material generally (particularly clear in the Shipton Member at
Stratton Audley). In the Bladon Member the fimbriata-waltoni clays have been
interpreted as representing a south-westwards extension of the clastic trap contained
between the land and the emergent lime mud barrier formed by the laminated
micrite (Palmer and Jenkyns 1974). There is also an increase in the number of
channels at the top of the Ardley and Bladon Members (Stratton Audley), which are
clay filled and contain a basal driftwood lag. Rootlets also occur in the north-eastern
region only (Shipton Member, Croughton). The trends of more elastics, more wood,
more channelling, more rootlets, and increasing numbers of brackish and freshwater
fossils is also seen very clearly in the overlying Forest Marble Formation as one passes
north-eastwards across the region where the top of the underlying Bladon Member is
developed as the emergent horizon (Palmer and Jenkyns 1974). These trends con-
tinue as far as the most easterly locality for which published information exists, the
Calvert Borehole (Davies and Pringle 1913). Here, the White Limestone is reduced to
less than 4 m of impure, marly limestones. Conversely, the Hampen Marly Forma-
tion is 13 m thick, and the Forest Marble (here developed as lagoonal clays with
rootlets) 12 m. This expansion of the non-marine units above and below the White
Limestone almost certainly involves the development of both the Bladon and the
Shipton Members in a non-marine facies. The Great Oolite feathers out completely
about 20 km east of Calvert.

The evidence for deepening and the attainment of more open marine conditions
at the south-western edge of the Oxfordshire shallows in the Burford region has been
mentioned several times in the text above. Primarily, this lies in the changing nature
of the lithologies of all three members in this region. Micrites, marls, muddy lime-
stones, and subordinate clays are common on the shallows. They represent very
shallow water (perhaps in the order of 1-5 m), where sediment banks and shoals
developed. Wave fetch was necessarily low, and currents were dampened by the
shallowness of the water and the baffle action of the sediment shoals. South-
westwards, at and off the edge of the shelf, larger wind-generated waves and currents

PALMER: HAMPEN MARLY AND WHITE LIMESTONE FORMATIONS

223

could develop in somewhat deeper water (perhaps about 5- 1 5 m), causing winnowing
of the substrate and forming loose lime sands. Locally, these became oolitic, and
periodically they were swept by storm currents and became cross-bedded. The faunal
changes discussed above are directly or indirectly (via the changing nature of the
sediment) the result of more fully marine conditions and better circulation of sea-
water south-west of the Burford region, than in the Oxfordshire shallows.

Superimposed on this palaeogeographic trend are the three shallowing upward
cycles of the three members. Much of the evidence for shallowing north-eastwards
discussed above comes from observations made in the top beds of the members.
There are rootlets at the top of the Shipton Member at Croughton (Palmer 1974),
and there is a putative dinosaur trackway at the same level at Ardley. There is the
laminated bed with Valvata and the wood-lined channels at the top of the Ardley
Member (text-fig. 4). There is the emergent birdseye limestone at the top of the
Bladon Member over much of the north-east of the region. Similarly, it is the lowest
beds of the members which appear to be best sorted by current winnowing ; and the
proportion of fine-grained material increases upwards until some sort of stand-still
is represented at the top. This is most clearly seen in the Ardley Member which passes
from lime sand to pelleted lime mud to shelly micrite, representing a drop in current
activity as the region of deposition becomes silted up.

The increase in micrite upwards is seen partly as a reduction in current winnowing,
and partly as an increase in the abundance of aragonite needle-producing green algae
as conditions became more protected and lagoonal.

The shallowing of each member is thus seen as being primarily dependent on
sedimentation. The deepening which heralded the onset of each member, however,
is seen as a relatively rapid event probably under tectonic control. The recent dis-
coveries of Bathonian vulcanicity in the northern North Sea place Britain firmly in a
tectonic province at this time, and indeed contemporaneous faulting of the Jurassic
strata or within the underlying basement has recently been postulated as the dominant
mechanism affecting the facies distribution of Jurassic sediments over the whole of
southern Britain (e.g. Sellwood and Jenkyns 1975; Jenkyns and Senior 1977). On the
stable platform of the Oxfordshire shallows vertical displacement at any one time
probably never exceeded a few metres, but in such shallow water this greatly affected
the nature of sedimentation before and after deepening. It is no coincidence that the
change of facies along the edge of the Oxfordshire shallows runs along the line of the
Moreton Axis which greatly affected sediment distribution and facies in the Lower
Jurassic and Inferior Oolite. Recently, Sellwood and McKerrow (1974) have discussed
facies change in the lower Great Oolite over this area and suggested a fault zone in the
underlying Palaeozoics. It is quite likely that the relatively sudden thickening and
change of facies over the same area in the upper Great Oolite was under the same
control.

Previous interpretations of the environmental picture in the White Limestone
(Klein 1965; McKerrow et al. 1969) have concluded deposition in intertidal condi-
tions. Klein (1965) extended this interpretation into Gloucestershire, and based his
conclusions on the evidence of thick laterally persistent, graded beds, becoming finer
upwards, which he states could have been formed by regression of high tidal flat
sediments (finer grained) over low tidal flat sediments (coarser grained).

224

PALAEONTOLOGY, VOLUME 22

It has been shown how the members as a whole fine upwards, and this has indeed
been interpreted as a shallowing. The graded units recognized by Klein, however, are
in the order of y-lf m, and are burrowed at the top and capped by hardgrounds
(Klein 1965, p. 187). It seems that a simpler explanation of the grading is that after
deposition of the coarse lime sands, there was a period of non-deposition and reduced
current activity. During this period the burrowing fauna colonized the bed from the
top surface, and the upper part of the sediment was bioturbated. During this period
fine sediment was produced, principally by the breakdown of animal skeletons, but
perhaps also by Codiacea similar to Penicillus which were able to establish themselves
on the relatively stable bottom sediment. The fine lime mud so produced, together
with fines removed from suspension and pelleted by filter-feeders, became incor-
porated into the sediment and mixed up by the bioturbation in the upper part of the
bed, giving the whole thickness a fining-upward appearance. Support for this hypo-
thesis comes from one of the fining-upward sequences at Foss Cross Quarry (see
Table 1) where the top of the bed contains a rich molluscan infauna, and a high
proportion of fines are seen in thin section. The same bed seen at Dagham Downs
Quarry (bed 6 of Torrens 1967) contains no molluscan infauna, and does not show an
upward increase in the proportion of fines.

Even without bioturbation, beds which are capped by hardgrounds may also show
a greater proportion of fines at their tops. Hardground formation may have started
by the growth of thin veneer of cement around each constituent particle, which holds
adjacent particles rigidly together; the inter-granular spaces may then have acted
as sediment traps for suspended fines, or mud-size particles produced by boring
organisms. I have noted this phenomenon frequently in Bathonian hardgrounds
in Normandy.

There are other, more empirical, objections to Klein’s intertidal hypothesis. First,
it may be surmised that such a variety of lithofacies and faunas over such a wide area
cannot all represent the same depositional environment. Second, there is a complete
absence of depositional structures indicating tidal activity. These including herring-
bone cross-bedding, flaser bedding, and tidal bedding which he later (Klein 1972)
considered to be so important in the recognition of tidalites. Third, none of the
fabrics and desiccation structures which have been found associated with Recent
intertidal carbonate sequences have been seen in Klein’s study area. These include
algal stromatolites; wide, shallow mud cracks; deep prism mud cracks (Ginsburg,
Bricker, Wanless, and Garrett 1970); and birdseye structures (Shinn 1968). Although
Klein (1965, p. 178) states that mud cracks are common to many limestones in the
Great Oolite, in the present study they have only been noted in one bed within the
Bladon Member, in a facies not seen by Klein (pers. comm. 1972). It seems more likely
that Klein mistook burrows for mud cracks.

In the total absence of any evidence for intertidal conditions, other than that
graded bedding is known to have been produced by prograding of high tidal flats
over low tidal flat (e.g. Evans 1965), and in the light of a reasonable alternative
explanation for the same phenomenon in the subtidal zone, it seems most unlikely
that Klein’s hypothesis is correct. Where indications of emergence are seen at the
top of the White Limestone to the north-east of Klein’s study area (around Ardley
and Stratton Audley), the fabrics are quite distinctive. Millimetre laminations, curled

PALMER: HAMPEN MARLY AND WHITE LIMESTONE FORMATIONS 225

algal mats, mud flake conglomerates, and birdseyes, associated with shrinkage
cracks, give a rock of totally different appearance to Klein’s White Limestone. It
seems likely that this was a very limited area of emergence and that most of the White
Limestone was deposited in shallow, subtidal conditions.

The presence of both supratidal and subtidal deposits, and the absence of un-
equivocal intertidal structures strongly suggests very reduced diurnal tidality in this
part of the Bathonian sea. Such a situation is to be expected over wide shallow shelves,
where tidal currents coming in from the open sea are dampened by friction and the
baffling effect of sediment shoals. In Florida Bay today the tidal range is reduced to
13 cm in the eastern part, and there is no exchange of water with the open sea.
Similarly, tidal range across the Great Bahama Bank falls from 0-78 m at the edge to
virtually nothing in the middle (Bathurst 1971). The situation is likely to have been
the same crossing on to the Oxfordshire shallows, and to have become accentuated in
the shallowest upper parts of the members. Only during storms was current activity
increased, forming and maintaining channels, causing deposition of lime mud layers
on the algally covered supratidal flats, and causing fluctuations in salinity. Under
these periodically harsh and variable conditions, only the most tolerant of eurytopic
organisms were able to exist. Some open marine groups, such as cephalopods, were
absent. Other predominantly stenotopic groups are represented only by one or two
species which were particularly tolerant of the adverse conditions. These include
sponges, corals, foraminifera, brachiopods, bryozoans, and echinoderms. The group
that thrived best was the bivalves, perhaps partially because of their ability to seal
themselves from the outside environment and thus at least ameliorate the adverse
effects of short-term fluctuations in the physical environment.

South-west of the Oxfordshire shallows, on the deeper more open marine carbonate
shelf of Bath and the Cotswolds, water circulation was much better, and fluctuations
in the physical environment were dampened. The same is true in the contemporaneous
shallow-water carbonates around the Armorican Massif in Normandy. As a result,
a much more diverse stenotopic fauna became established, with groups such as
sponges, corals, foraminifera, brachiopods, bryozoans, and echinoderms much more
fully represented. Over-all diversity of invertebrate species was about four times as
high as in Oxfordshire. The marine fauna of the Great Oolite discussed above, there-
fore, consists principally of those forms which, by physiology or ecological strategy,
could withstand the exigencies of an environment in which considerable fluctuations
in various physical parameters were to be expected. The more diverse stenotopic
counterpart to this fauna is more rarely seen in southern England. Consequently,
where it does occur, at least some of its elements have been given their own name
(Bradfordian fauna) and invested with a stratigraphic significance that they do not
have. This problem will be dealt with separately.

REFERENCES

allen, J. R. L. and kaye, p. 1973. Sedimentary facies of the Forest Marble (Bathonian), Shipton-on-Cherwell
quarry, Oxfordshire. Geol. Mag. 110, 153-163.

arkell, w. J. 1931. The Upper Great Oolite, Bradford Beds, and Forest Marble of Southern England, and
the succession of gastropod faunas in the Great Oolite. Q. Jl geol. Soc. Lond. 87, 563-629.

— 1933a. The Jurassic system in Great Britain. Oxford, 681 pp.

226

PALAEONTOLOGY, VOLUME 22

arkell, w. j. 1933 b. New evidence on the Great Oolite succession at Bladon, near Woodstock, Oxford-
shire. Proc. Geol. Ass. 44, 177-183.

— 1933c. Boring at Latton, near Cricklade. Proc. Cottes. Nat. Field Club , 24, 181-183.

1947. The geology of Oxford. Oxford, 267 pp.

— 1951-1958. A monograph of English Bathonian Ammonites. Palaeontogr. Soc. ( Monogr .), 1-264.

— and donovan, d. t. 1952. The Fuller’s Earth of the Cotswolds, and its relation to the Great Oolite.
Q. Jl geol. Soc. Lond. 107, 227-253.

— Richardson, l. and pringle, J. 1933. The Lower Oolites exposed in the Ardley and Fritwell railway
cuttings, between Bicester and Banbury, Oxon. Proc. Geol. Ass. 44, 340-354.

aslin, c. J. 1965. The Upper Estuarine Series of the East Midlands. Unpublished Ph.D. thesis, Univ. of
Leicester.

barker, m. j. 1976. Gastropods of the Great Oolite. Unpublished Ph.D. thesis, Univ. of Keele.
bate, R. H. 1965. Freshwater ostracods from the Bathonian of Oxfordshire. Palaeontology, 8, 749-759.
Bathurst, r. g. c. 1966. Boring algae, micrite envelopes, and lithification of mollusc biosparites. Geol. Jl,
5, 15-32.

— 1971. Carbonate sediments and their diagenesis. Developments in Sedimentology, 12, 1-620.
Bernard, H. a., le blanc, R. J. and major, c. J. 1962. Recent and Pleistocene geology of southeast Texas.

In rainwater, e. h. and zingula, R. p. (eds.). Geology of the Gulf coast and central Texas. Geol. Soc.
Amer. Guidebook for 1962 Ann. Mtg. 175-224.

coogan, A. H. 1971. Bahaman and Floridian biofacies. In multer, h. g. (ed.). Field guide to some carbonate
rock environments, Florida Keys and Western Bahamas. Madison, N.J., 159 pp.
cooper, l. h. N. 1951. Chemical properties of the sea water in the neighbourhood of the Labadie Bank.
Jl Mar. Biol. Assoc. 30, 21-26.

cox, l. R. and arkell, w. j. 1948. A survey of the Mollusca of the British Great Oolite Series. Palaeontogr.
Soc. {Monogr.), 1-105.

davies, a. m. and pringle, J. 1913. On two deep borings at Calvert station. Jlgeol. Soc. Lond. 69, 308-342.
evans, G. 1965. Intertidal flat sediments and their environments of deposition in the Wash. Ibid. 121,
209-245.

fursich, F. T. and palmer, t. j. 1975. Open crustacean burrows associated with hardgrounds in the Jurassic
of the Cotswolds. Proc. Geol. Ass. 86, 171-181.

ginsburg, r. n. 1964. South Florida Carbonate Sediments. Guide book for field-trip no. 1, Geol. Soc. Amer.
Convention 1964. New York, 72 pp.

— bricker, o. p., wanless, H. r. and Garrett, p. 1970. Exposure Index and sedimentary structures of a
Bahama Tidal flat. Geol. Soc. Amer., Abstracts, 2, 744.

graus, R. R. and macintyre, i. G. 1976. Light control of growth form in colonial coral reefs. Science, 193,
895-897.

green, G. w. and melville, R. v. 1956. The stratigraphy of the Stowell Park borehole. Bull. geol. Surv. G.B.
11, 1-34.

hedgpeth, J. w. 1957. Estuaries and lagoons II: biological aspects. In hedgpeth, j. w. (ed.). Treatise on
Marine Ecology and Paleoecology, Vol. 1, Ecology. Mem. Geol. Soc. Amer. 67, 693-729.
hudleston, w. H. 1896. British Jurassic Gasteropoda; Part 1 : Gasteropoda of the Inferior Oolite. Palae-
ontogr. Soc. {Monogr.), 1-514.

Hudson, J. d. 1963a. The recognition of salinity-controlled mollusc assemblages in the Great Estuarine
Series (Middle Jurassic) of the Inner Hebrides. Palaeontology, 6, 318-326.

— 19636. The ecology and stratigraphical distribution of the invertebrate fauna of the Great Estuarine
Series. Ibid. 327-348.

— and palmer, t. j. 1976. A euryhaline oyster from the Middle Jurassic and the origin of the true oysters.
Ibid. 19, 79-93.

illing, l. v. 1954. Bahaman calcareous sands. Bull. Am. Ass. Petrol. Geologists, 38, 1-95.
jenkyns, h. c. and senior, j. r. 1977. A Liassic palaeofault from Dorset. Geol. Mag. 114, 57-62.

Kennedy, w. j., jakobson, M. E. and Johnson, R. t. 1969. A Favreina-Thalassinoides association from the
Great Oolite of Oxfordshire. Palaeontology, 12, 549-554.
klein, G. de v. 1965. Dynamic significance of primary structures in the Middle Jurassic Great Oolite Series,
Southern England. In Middleton, g. v. (ed.). Primary sedimentary structures and their hydrodynamic
interpretation. Soc. econ. Paleont. Mineral., Spec. Publ. 12, 173-191.

PALMER: HAMPEN MARLY AND WHITE LIMESTONE FORMATIONS

227

klein, G. de v. 1972. Sedimentary model for determining paleotidal range: reply. Bull. Geol. Soc. Amer.

83, 539-546.

lund, e. j. 1957. Self-silting, survival of the oyster as a closed system, and reducing tendencies of the environ-
ment of the oyster. Univ. Texas Inst. Marine Sci. Publ. 4, 313-319.
martin, A. J. 1958. The sedimentary petrology of the Bathonian rocks of the North Cotswolds. Unpublished
Ph.D. thesis, Univ. of London.

mckee, E. d. and gutschick, r. c. 1969. Sedimentary belts in lagoon of Kapingamarangi Atoll. Bull. Am.
ylsj. Petrol. Geologists. 43, 501-562.

mckerrow, w. s., Johnson, r. t. and jakobson, M. E. 1969. Palaeoecological studies in the Great Oolite at
Kirtlington, Oxfordshire. Palaeontology, 12, 56-83.

moore, R. c. (ed.). 1969. Treatise on invertebrate Paleontology ; part N: Mollusca6; Bivalvia. University of
Kansas, 952 pp.

Newell, n. d., imbrie, J., purdie, e. g. and thurber, D. L. 1959. Organism communities and bottom facies,
Great Bahama Bank. Bull. Amer. Mus. Nat. Hist. 117, 181-228.
odling, m. 1913. The bathonian rocks of the Oxford district. Jl geol. Soc. Lond. 69, 484-513.
palmer, t. J. 1973. Field meeting in the Great Oolite of Oxfordshire : report by the director. Proc. Geol. Ass.

84, 53-64.

— 1974. Some palaeoecological studies in the Middle and Upper Bathonian of central England and northern
France. Unpublished D.Phil. thesis, Univ. of Oxford.

— and fursich, f. t. 1974. The ecology of a Middle Jurassic hardground and crevice fauna. Palaeontology,
17, 507-523.

— and jenkyns, H. c. 1975. A carbonate island barrier from the Great Oolite (Middle Jurassic) of Central
England. Sedimentology, 22, 125-135.

Parker, r. h. 1960. Ecology and distributional patterns of marine macroinvertebrates, northern Gulf of
Mexico. In shepard, f. p., phleger, f. b. and andel, t. h. van (eds.). Recent sediments, north-west Gulf of
Mexico. Am. Ass. Petrol. Geologists, Spec. Publ. 302-337.

Phillips, J. 1971. The geology of Oxford and the valley of the Thames. Oxford, 523 pp.
pocock, T. i. 1908. The geology of the country round Oxford. Mem. Geol. Surv. England and Wales,
1-142.

poole, e. g. 1969. The stratigraphy of the Geological Survey Apley Barn Borehole, Witney, Oxfordshire.
Bull. Geol. Surv. G.B. 29, 1-104.

pringle, j. 1926. Geology of the country round Oxford. Mem. Geol. Surv. England and Wales, 1-191.
remane, a. 1934. Die brackwasserfauna. Zool. Anz., Supplementband, 7, 34-74.

Richardson, L. 1910. The Great Oolite section at Groves Quarry, Milton-under-Wychwood, Oxfordshire.
Geol. Mag., Dec. 5, 7, 537-542.

— 1911. On the sections of Forest Marble and Great Oolite on the railway between Cirencester and
Chedworth, Gloucestershire. Proc. Geol. Ass. 22, 95-115.

— 1929. The country around Moreton-in-Marsh. Mem. Geol. Surv. England and Wales, sheet 217,
1-162.

— 1932. A boring at Ready Token, near Cirencester. Proc. Cottes. Nat. Field Club, 24, 185-188.

— 1933. The country around Cirencester. Mem. Geol. Surv. England and Wales, sheet 235, 1-119.

— 1946. The country around Witney. Ibid., sheet 236, 1-150.

sellwood, B. w. and jenkyns, h. c. 1975. Basins and swells and the evolution of an epeiric sea
(Pliensbachian-Bajocian of Great Britain). Jl geol. Soc. Lond. 131, 373-388.

— and mckerrow, w. s. 1974. Depositional environments in the lower part of the Great Oolite Group of
Oxfordshire and North Gloucestershire. Proc. Geol. Ass. 85, 189-210.

shearman, d. j., twyman, J. and karim, m. z. 1970. The genesis and diagenesis of oolites. Ibid. 81, 561-575.
shepard, f. p. and moore, d. g. 1960. Bays of Central Texas coast. In shepard, f. p., phleger, f. b. and
andel, t. H. van (eds.). Recent sediments, north-west Gulf of Mexico. Am. Ass. Petrol. Geologists,
Spec. Publ. 302-337.

shinn, E. a. 1968. Practical significance of bird’s-eye structures in carbonate rocks. Jl Sediment. Petrol. 38,
215-223.

Stanley, s. M. 1966. Palaeoecology and diagenesis of Key Largo Limestone. Bull. Am. Ass. Petrol.
Geologists, 50, 1927-1947.

— 1970. Relation of shell-form to life habits in the Bivalvia. Mem. Geol. Soc. Amer. 125, 1-296.

228

PALAEONTOLOGY, VOLUME 22

stenzel, h. b. 1971. Mollusca 6, Bivalvia part 3 {Oysters). In moore, r. c. (ed.). Treatise on invertebrate
Paleontology. University of Kansas, N953-N1124.

swan, E. F. 1952. Growth of the clam My a arenaria, as affected by the substratum. Ecology, 33, 530-534.
torrens, H. s. 1967. The Great Oolite Limestone of the Midlands. Trans. Leicester Lit. and Phil. Soc. 61,
65-90.

— 1968. The Great Oolite Series. In sylvester-bradley, p. c. and ford, t. d. (eds.). The Geology of the
East Midlands. Leicester, 400 pp.

— 1969. Guidebook of the International Field Symposium on the British Jurassic. Keele, 38 pp.
turmel, r. and swanson, r. 1964. Rodriguez Bank. In ginsburg, r. n. (ed.). South Florida carbonate sedi-
ments, Guidebook for field-trip no. 1, Geol. Soc. Amer. Convention 1964. New York, 72 pp.

van morkhoven, f. p. c. m. 1962-1963. Post-Palaeozoic Ostracoda, their morphology , taxonomy, and
economic use. Amsterdam, 2 vols., 682 pp.

walford, E. a. 1896. Stonesfield Slate. Third and final report of the committee. Rep. Brit. Ass., Liverpool.
356 pp.

— 1912. An account of the geology of the Brackley town well. Buckingham, 14 pp.

wells, J. w. 1956. Scleractinia. In moore, r. c. (ed.). Treatise on invertebrate Paleontology, part F, Coelent-
erata. University of Kansas, 498 pp.

weymouth, F. w. 1920. The edible clams mussels and scallops of California. California Fish and Game
Comm. Fish Bull. 4, 1-74.

whitehead, t. h. and arkell, w. J. 1946. Field meeting at Hook Norton and Sibford, Oxfordshire: report
by the directors. Proc. Geol. Ass. 57, 16-18.

wilson, d. p. 1951. A biological difference between natural sea-waters. Jl Mar. Biol. Assoc. 30, 1-20.
woodward, h. b. 1894. The Jurassic rocks of Britain, Vol. 4: The Lower Oolitic rocks of England {Yorkshire
excepted). Mem. Geol. Surv. U.K., 628 pp.

worssam, B. c. and bisson, g. 1961. The geology of the country between Sherborne, Gloucestershire, and
Burford, Oxfordshire. Bull. Geol. Surv. G.B. 17, 75-115.
wray, J. l. 1971. Ecology and geologic distribution. In ginsburg, r. n., rezak, r. and wray, j. l. Geology
of Calcareous Algae. Sedimenta, 1. (Publication of the Comparative Sedimentology Laboratory, Univer-
sity of Miami.)

yonge, c. m. 1946. On the habits of Turritella communis. Jl Mar. Biol. Assoc., U.K. 26, 377-380.

Typescript received 3 February 1978
Revised typescript received 17 May 1978

T. J. PALMER

Department of Geology and Mineralogy
Parks Road
Oxford OX1 3PR

CARADOC MARINE BENTHIC
COMMUNITIES OF THE SOUTH BERWYN
HILLS, NORTH WALES

by R. k. pickerill and p. j. brenchley

Abstract. The Upper Ordovician (mid Caradoc: Soudleyan-Longvillian) clastic rocks of the south Berwyn Hills,
North Wales, contain an abundant and diverse macrobenthic fauna dominated by epifaunal brachiopods. Based on
studies in the south Berwyns, Shropshire, and Snowdonia, four communities are recognized and examined in terms of
composition and related palaeoenvironmental parameters. It is concluded that the communities are intergrading and
exhibit a close correlation with substrate and a broad correlation with depth, distance from shore, turbulence, and
rates of sedimentation. The low diversity Howellites community was best developed on muddy silt and silty mud
substrates in low energy turbid conditions and in water depths of less than 25 m. The Dinorthis community exhibits a
low to moderate diversity and based on the balance of constituent genera and relationship to substrate is divisible into
two sub-communities : the Dinorthis sub-community was best developed on shifting coarse sand substrates in high
energy, non-turbid, well-oxygenated environments of water depths of less than approximately 10 m. The Macrocoelia
sub-community was developed on finer sand substrates deposited in lower energy conditions and slightly more off-
shore in deeper water (25 m). The Dalmanella community developed on non-turbid, well-oxygenated, mobile fine sand
substrates in water depths of 25 m or less and in areas of reduced sedimentation. The Nicolella community inhabited a
variety of substrates but developed best on calcareous silt and fine sands. Energy conditions were variable at any one
time but in general low-energy conditions prevailed, sedimentation rates were low and water depth was in the order of
approximately 30 m. The communities are examined in terms of their stratigraphical distribution within the Berwyn
succession and are discussed in relation to previously described Lower Palaeozoic communities. It is suggested that
benthic faunas progressively migrated into deeper waters throughout the Lower Palaeozoic.

Since the pioneering work of Petersen (1911, 1913) marine benthic communities
have generally been considered to be real phenomena and only on occasion has this
reality been contested (e.g. Lindroth 1935; MacGinitie 1939; Muller 1958). One of
the major goals of marine palaeoecology is the description of community structure
and evolution over long periods of geological time and the ultimate development of
general models relating them to controlling environmental parameters. Attempts
have already been made to trace the evolution of certain communities, for example,
Bretsky (1969a), Anderson (1971), Watkins and Boucot (1975), and Boucot (1975),
though as Thayer (1974) points out, such efforts must be regarded with some caution
until a sufficient number of detailed palaeoecological investigations have been com-
pleted. Unfortunately, relatively few detailed studies of Ordovician community
palaeoecology have been undertaken, notable exceptions being those of Walker and
Laporte (1970) and Walker and Alberstadt (1975) in carbonate sequences, and
Bretsky (19696, 1970a, b ) and Bretsky and Bretsky (1975) in clastic sequences. This
paper is therefore intended to document just such an investigation which was under-
taken in the Caradoc rocks (Ordovician) of the south Berwyn Hills, North Wales.

Brachiopoda were the numerically dominant taxonomic group of sedentary
marine macrobenthos during Ordovician times (Williams 1976) and predominate in
the communities described here. Additional taxa comprising these communities,

[Palaeontology, Vol. 22, Part 1, 1979, pp. 229-264.]

230

PALAEONTOLOGY, VOLUME 22

though of lesser abundance, include bivalves, gastropods, ostracodes, trilobites, and
crinoids, and these elements are also included. The work arose from the observations
of Williams (1963, 1973), who in his treatment of Caradoc brachiopods of the Bala
district, North Wales, noted briefly that the faunas occurred in four ‘associations’,
the Nicolella, Dinorthis, Onniella, and Howellites associations, which were sub-
sequently adopted by Pickerill (1975, 1976, 1977). Williams (1963), however, stated
‘whether these associations represent remains of biotic communities that existed in a
north Welsh province during Caradocian times is a matter for further exploration’.
Research was therefore undertaken in the Berwyn Hills to investigate this statement
in some detail. In addition to detailed sampling and data collection in the Berwyn
Hills, further material was obtained from south Shropshire (now south Salop) and
east Snowdonia, where lower Caradoc lithofacies are more clearly differentiated.

text-fig. 1 . Generalized map of outcrop of Ordovician rocks in North Wales, and more detailed map of the
south Berwyn Hills indicating localities referred to in the text.

PICKERILL AND BRENCHLEY: ORDOVICIAN MARINE COMMUNITIES

Stratigraphical setting

The Berwyn Hills (text-fig. 1) are formed by a dome-like upfold of Ordovician rocks
flanked on three sides by Silurian strata, and on the eastern edge by sediments of
Permo-Carboniferous age. Stratigraphically the oldest material exposed is Llandeilo
in age (MacGregor 1961) and this is overlain by several hundred metres of unfossili-
ferous sediments with thin tuff's and lavas which have not been precisely dated but
are probably of lower Caradoc age. These are followed by a succession of fossili-
ferous sediments and volcanics of proved Caradoc (Soudleyan and Longvillian) age
(King 1923; Brenchley 1969; Pickerill 1977). These fossiliferous sediments, with
which this study is concerned, are composed of a thick sequence ( c . 900 m) of thinly
interbedded muds and silts and/or sands. The stratigraphical sequence in the south
Berwyns is illustrated in text-fig. 2. The present study does not include in detail the
graptolitic Pen-y-Garnedd Shales (=Nod Glas) as these are generally poorly
exposed and have a limited shelly fauna.

Samples collected in Snowdonia were obtained from the Glanrafon Beds and
Snowdon Volcanic Group (Soudleyan and Longvillian), which are composed essen-
tially of interbedded mudstones and siltstones with thick intercalations of volcani-
clastic sandstones. In south Shropshire samples were taken mainly from coarse
sandstone facies within the Hoar Edge Grit (Costonian), which lies unconformably
on rocks of Precambrian age and represents a transgressive fining-upward sequence,
but some collections were made for comparative purposes from rocks of Harnagian
and Soudleyan age (text-fig. 2).

text-fig. 2. Stratigraphical columns illustrating lithological sequences and Stage boundaries in south
Shropshire, the south Berwyns, and eastern Snowdonia. Thicknesses are approximately to scale.

232

PALAEONTOLOGY, VOLUME 22

Environmental setting

The environmental setting briefly outlined here is based on a combination of know-
ledge of the general position of the Berwyn Hills in relation to the well-established
regional environmental framework of the Welsh Basin (Brenchley 1969) and on more
detailed examination of lithofacies, associated bedforms, and ichnofacies within the
region itself. Faunal information is used only in a very general sense in environmental
reconstruction as the objective is to interpret the fauna in terms of its environmental
distribution (cf. Thayer 1974).

Geographically the Berwyn Hills now lie in what was formerly part of the central
region of the NE-SW trending Welsh Basin during Caradoc times. The Welsh Basin
itself was a fault bounded tectonically active graben during the Ordovician, approxi-
mately 120 km in width and bounded to the north-west by the Anglesey-Rosslare
Horst and to the south-west by the Church Stretton or Pontesford Linley Fault. A land
area composed of Borrowdale Volcanics lay to the north until at least Longvillian
times (Brenchley 1969). Within the graben a series of volcanic islands extended east-
west across W ales and erosion of many of these islands provided a maj or source of intra-
basinal sediment (Bassett 1963). It has long been recognized that sedimentation in the
northern part of the Welsh Basin was shallow- water in origin (Brenchley 1969) though
accurate delineation of specific sub-environments has not really been attempted.

For the purposes of this paper we present the pertinent observations and conclu-
sions on the sedimentary environments of the Caradocian rocks in the south Berwyns
in Table 1 . Basically, the whole succession was deposited in a generally quiet subtidal
to shallow marine shelf environment (inner infralittoral, 0-30 m). In the Soudleyan
the environmental conditions were essentially homogeneous, apart from episodic
volcanicity, and the relatively thick succession was deposited rapidly. The much
thinner Longvillian succession was deposited in more varied environments charac-
terized by slower rates of sedimentation and representing a greater bathymetric
range, though still within the inner infralittoral spectrum. Sediment was supplied
from intrabasinal volcanic sources and was redistributed and deposited by tidal
currents and occasionally modified by storm and wave activity (Brenchley, in press).

Material

Four hundred and three samples were obtained from the south Berwyn Hills, includ-
ing in total some 72,000 brachiopods and some 6,000 additional elements (bivalves,
gastropods, trilobites, ostracodes, etc.). Additional sampling in Shropshire and
Snowdonia was undertaken because sediment types were more clearly differentiated
than in the south Berwyns and it was hoped and eventually realized that a clearer
picture of faunal distribution would be obtained. Sample sizes of all collections
ranged from between 60 to 500 individuals.

Stanton and Evans (1972) have pointed out that the ability to define or recognize
communities is determined, in addition to the structural characteristics of the com-
munity being investigated, by the number and size of available samples. A large
number of inter- and intralocality samples were taken to define the faunal patterns.
Some assemblages were collected and recorded in the laboratory, others were
recorded directly in the field. Full locality, stratigraphical, and faunal details of each

PICKERILL AND BRENCHLEY: ORDOVICIAN MARINE COMMUNITIES

233

table 1. Summary of environmental analysis of the Caradoc succession in the south Berwyn Hills.

Description

Interpretation

VOLCANICS

Volcanics, including welded tuffs occur
associated with the mudstone sequence
in the north and west Berwyns. The tuffs
form widespread sheets of relatively
constant thickness.

The normally marine area was subject to
periodic emergence. Deposition of sub-
aerial tuffs occurred on a relatively flat,
undissected surface (Brenchley 1969).

Facies

A vertically and laterally variable associa-
tion of mudstones with thin, parallel
bedded, lenticular bedded, and biotur-
bated siltstones and fine sandstones.

A facies association commonly found in
shallow subtidal environments (depths
10-30 m).

Sedimentary

Very common small scale cross lamination

Evidence of both current and wave ripples.

Structures

and common ripple marks, both asym-
metric and symmetric. Interference
ripple patterns frequent. Small (<30 cm
deep) sharply incised scour and fill struc-
tures, and some broader channels, mud
clasts in sandstones.

Locally, stream-like current channelled
muds.

Ichnology

A variable ichnofaunal suite including
Cruziana spp., Rusophycus spp., Tricho-
phycus, Skolithos, Planolites, ? Gyro-
chorte, Diplocraterion, Arenicolites,
? Palaeophycus, Vermiforichnus, Teichi-
chnus, and ? Imponoglyphus.

Coexistence of the Skolithos and Cruziana
ichnofacies in a shallow subtidal inner
infralittoral shelf environment (Pickerill
1976, 1977).

Body Fossils

Normal ‘shallow water’ marine benthic
assemblages.

Marine shelf.

sample are given in Pickerill (1974). A locality list, maps, and sample details have
been deposited with the British Library, Boston Spa, Wetherby, Yorkshire LS23 7BQ
U.K. as Supplementary Publication No. SUP 14011 (175 pages). A selection of the
fauna has been deposited in the British Museum (Natural History).

Most of the fauna in the south Berwyns is well known taxonomically through the
comprehensive description of the brachiopods of the near-by Bala area by Williams
(1963) and through the monographs of the trilobites of the Bala area by Whittington
(1962, 1965, 1966, 1968) and of the south Shropshire area by Dean (1960 b, 1961,
1963a, b). The crinoids have been monographed by Ramsbottom (1961) but the
remainder of the fauna is still imperfectly described. Quantification of faunal data was
facilitated by the good preservation of most taxa and each sample was counted with a
view to assessing its composition in terms of relative abundance of individual genera.
For brachiopods and bivalves, unless the specimen was articulated, the number of
individuals was taken as the number of the most abundant valve. For trilobites the
number of the most abundant part (cephalon or pygidium) was divided by a factor
of ten to allow for ecdysis (see Harrington 1959, p. Ol 1 1). Crinoids, bryozoans, and
ostracodes proved more difficult to handle quantitatively. These elements usually
occurred as only a very minor part of the collections and were for ease of data hand-
ling referred to as single individuals.

234

PALAEONTOLOGY, VOLUME 22

ASSEMBLAGES AS LIFE ASSEMBLAGES

Johnson (1960) and Lawrence (1968) have discussed the problem of deriving from
fossil assemblages an understanding of the original community. Unfortunately
many of these problems, such as non-preservation as a result of chemical dissolution
and diagenetic activity and selective winnowing of extremely small forms, etc., must
remain obscure as there is no positive evidence by which to assess them. However, in
spite of such problems we consider that the assemblages are representative of the
original communities based on our following observations :

1 . Life clusters of completely articulated shells have been observed in most sedi-
ment types, and these assemblages are of similar composition to assemblages in which
some degree of disturbance is clearly demonstrable. Such life clusters are most com-
mon in muds and silty muds which have not been subjected to higher energy current
regimes or intense biogenic reworking.

Commonly observed are monospecific clusters of three to seven individuals of the
dalmanellid Howellites, which lived with the umbo pointing down and the commissure
vertical or nearly vertical (Richards 1972). Monospecific clusters of the plectambonitid
Sowerbyella are also common and this genus appears to have lived on the substratum
supported by the gently convex ventral valve. Less frequently observed are life
clusters of associated Sowerbyella and Howellites and monospecific clusters of the
craniopsid Paracraniops, which appear to have been free-living craniids like Pseudo-
crania (Williams 1963). Rare examples of the orthid Platystrophia sublimis and the
triplesid Bicuspina spiriferoides were also observed and these appear to have lived
umbo down with the plane of the commissure nearly vertical. Finally the lingulid
Lingulasma tenuigranulatum was also observed in vertical burrowing position, details
of which are given by Pickerill (1973).

2. Detailed analysis of the various taxa has indicated that post-mortem reworking,
transportation, and hydraulic mixing of the assemblages was limited. This is par-
ticularly true for those assemblages scattered within mudstones and muddy silt-
stones, which accumulated under low-energy conditions. Of these assemblages
60-7% contain some articulated valves and these articulated valves commonly form a
high percentage of all valves present (median 40-50%, text-fig. 3), but the remaining
39-3% of the assemblages contain no articulated valves. Therefore there are apparently
two types of assemblages present ; those which have a percentage of articulation and
which have suffered little or no hydraulic disturbance, and those assemblages which
have been sufficiently disturbed to cause complete disarticulation.

Assemblages occurring on bedding planes or in calcareous lenses less commonly
have articulated valves (text-fig. 3), and amongst the assemblages which have articu-
lated valves the percentage articulation is generally low (text-fig. 3). This indicates
that nearly all these assemblages have been disturbed but that the degree of disturb-
ance in some cases is limited. The wide size distribution and lack of strongly preferred
orientation of the valves (text-fig. 4) and a ratio of pedicle to brachial valves approach-
ing unity (text-fig. 3) in all these disturbed assemblages indicates that they have not
been subjected to prolonged winnowing and sorting by currents. Poorly sorted,
poorly oriented assemblages could have been moved and dumped by violent wave and
current activity during storms, particularly in near-shore and littoral environments.

PICKERILL AND BRENCHLEY: ORDOVICIAN MARINE COMMUNITIES

235

« «

c/5 co

u> o

S. o

Q co

T3 -S
<D X!

■g £

g a>

(0

1 £

£ S2

Howellites

% Articulation

text-fig. 3. (A) Histograms showing ratio of pedicle to brachial valves in n collections of Howellites.
(B) Histograms showing percentage of valves which are articulated in collections containing Howellites,
Sowerbyella, or Heterorthis. Tn=total number of collections. An = collections with articulated specimens.
An

crowded on bedding planes; d= assemblages in calcareous lenses.

236

PALAEONTOLOGY, VOLUME 22

text-fig. 4. Histograms showing size-frequency distribution of brachial valves in assemblages containing
Dalmanella (a, bx ct dx b2) or Howellites (a2 c2 d2). The rose diagram shows the orientation of the hinge lines
of the brachial valves in the same assemblages. n= number of specimens, a = assemblage widely dispersed
on a bedding plane; b= medium dispersed; c= crowded on a bedding plane; d= assemblages in calcareous

lenses.

In more off-shore subtidal situations, such as occur in the Berwyns, the ability
of waves and currents to move and dump large grains is much more limited and a
significant degree of transport takes a considerably longer period. Because the
assemblages described here are generally unfragmented and lack abrasion it is unlikely
that transport has been either violent or prolonged. Instead it is probable that the
assemblages are nearly in situ but have been disturbed by brief episodes of turbulence
such as might be expected in a wave-influenced subtidal environment.

3. The similarity of taxonomic composition between those assemblages which are
in life position, those scattered assemblages which have a high percentage of articu-
lated valves, and those clearly transported assemblages which occur on bedding
planes, in lenses, or in cross-stratified beds, suggests that in all cases there has been
little mixing of faunal elements.

PICKERILL AND BRENCHLEY: ORDOVICIAN MARINE COMMUNITIES 237

Thus, we agree with Johnson (1965) and Walker and Bambach (1971), who point
out that most benthic assemblages on clastic substrates represent the remains of
organisms that lived nearly in place, even though the resultant assemblages may be
‘time averaged’.

In summary, though it is clear that some transport and winnowing of many of the
assemblages has occurred, we conclude that this was not sufficiently extensive to
modify the original associations and therefore regard the assemblages as representa-
tive of the original communities.

TERMINOLOGY

The classification of ecological and palaeoecological units is still beset by numerous
differences of approach and terminology (Kauffman and Scott 1976). Therefore to
avoid confusion we have defined below certain terms used in this account.

(i) An assemblage refers to a single collected sample.

(ii) An association refers to the recurrent association of taxa in a group of assemblages.

(iii) A community refers to a spatially repeated and temporally recurring group of organisms usually
related to specific environmental parameters. Palaeontologically a community is usually inferred from the
occurrence of one or more associations. We have used sub-community to distinguish associations within a
community which have a different abundance of the constituent genera. Generally, communities have been
named either after their environment of occurrence or, as in the case here, after a genus which, in the sense of
Johnson (1972), is dominant or characteristic and hence is one of the most abundant. More recently Hurst
(1975) proposed that communities should be named after one of the most abundant and ecologically
restricted invertebrate species. This avoids confusion as far as extra-regional applicability of localized com-
munity types are concerned, and, in addition, it attempts to avoid the problem that different species of the
same genus can possibly have different ecological preferences (e.g. Hurst 1975). However, this procedure
can lead to a proliferation of community names and we have preferred a generic designation.

Johnson (1972) has pointed out that a community is composed of three kinds of species : (a) characteristic
species, which occur more frequently in the community than in any other and are therefore characteristic of
the particular community. It is not necessary for a species to be abundant for it to be characteristic and in
this account we have referred to genera as being characteristic even if they are uncommon ; (b) intergrading
species, which are characteristic of another usually adjacent community ; and (c) ubiquitous species, which
occur in several communities but are not characteristic of any one. In the description below we adopt this
general scheme and therefore describe what we regard as the elements characteristic of a particular com-
munity and also indicate those elements of adjacent co-existing communities which may sometimes form
part of the particular association.

(iv) Diversity refers to the number of genera in an assemblage. We have calculated mean diversity for
each community by totalling the number of genera and dividing this total by the number of assemblages in
the community.

THE CARADOC COMMUNITIES

The nature of shallow-water benthic communities on clastic substrates has been
succinctly summarized by Johnson (1972, p. 152), who states that such communities
. . exhibit low diversity, recur in variable combinations of species and are often
revised by fluctuations in the physical environment. The environmental gradients
tend to be low. As a consequence benthic communities on clastic substrates tend to be
continuous and intergrading.’ Thus, the actual delineation of distinct communities is
often quite arbitrary and, indeed, this has been our experience in examining the
Caradoc rocks of the Berwyn Hills where the mudstone and siltstone lithofacies are
poorly differentiated and the communities exhibit nearly continuous intergradation.

238

PALAEONTOLOGY, VOLUME 22

table 2. Faunal list and stratigraphical distribution of taxa from the south Berwyn Hills.

Brachiopoda

Bicuspina spiriferoides (M’Coy) 1, 2, 3, 5

Cremnorthis parva (Williams) 5

Dalmanella horderleyensis (Whitting-
ton) 3

Dalmanella indica (Whittington) 3

Dalmanella cf. modica (Williams) 4, 5

Dalmanella sp. 1,2

Dinorthis berwynensis (Whittington) 1, 2

Dinorthis berwynensis angusta
(Williams) 2, 3

Dinorthis cf. flabellulum (M’Coy) 3

Dinorthis sp. 5

Dolerorthis duftonensis prolixa
(Williams) 3, 4, 5

Eoplectodonta rhombica (M’Coy) 4, 5

Eoplectodonta sp. 2

Howellites antiquior (M’Coy) 3, 4

Howellites spp. _ 1,2,5

Kiaeromena kjerulfi (Holtedahl) 2, 3, 4, 5

Kiaeromena sp. 5

Kjaerina hedstroemi (Bancroft) 3

Kjaerina jonesi (Bancroft) 3, 4

Kjaerina sp. 3

Kjerulfina sp. 4

Leptaena ventricosa (Williams) 2, 3

Leptaena sp. 3, 5

Leptestiina oepiki (Whittington) 2, 3

Lingulasma tenuigranulatum (M’Coy) 5
Lingulasma sp. 5

Trilobita

Deacybele pauca (Whittington) 3, 4, 5

Broeggerolithus broeggeri (Bancroft) 1 , 2, 3

Broeggerolithus nicholsoni (Reed) 3

Broeggerolithus soudleyensis (Bancroft) 2, 3
Broeggerolithus sp. 4, 5

Brongniartella cf. ascripta (Reed) 2, 3

Brongniartella minor (Salter) 1 , 2, 3

Brongniartella sp. 3, 4, 5

Calyptaulax sp. 4

Chasmops cambrensis (Whittington) 3, 4, 5

Conolichas sp. 3, 4

Estoniops alifrons (M’Coy) 5

Gastropoda

Bellerophontid gen. et sp. indet. 2, 3

Bucania sp. 3

Bucaniopsia sp. 1,2

Clathrospira sp. 1 , 2, 3

Cyclonema crebristria (M’Coy) 1 , 2

Cyrtolites sp. 1,2,3

Lingulella cf. ovata (M’Coy) 1, 2,

Lingulella sp. 3

Linguloid gen. et sp. indet. 1,2,

Macrocoelia expansa (Sowerby) 3

Macrocoelia prolata (Williams) 2, 3

Nicolella actoniae (Sowerby) 4, 5

O bolus sp. 5

Onniella cf. soudleyensis (Bancroft) 2
Orbiculoidea sp. 3

Oxoplecia sp. 1 , 2,

Paracraniops sp. 1,2

Platystrophia sublimis (Opik) 3, 5

Reuschella horderleyensis (Bancroft) 1 , 2

Reuschella horderleyensis undulata
(Williams) 3, 4

Reuschella oblonga (Whittington) 1

Rhactorthis cf. crassa (Williams) 4, 5

Rhynchotrema sp. 4

Rostricellula sparsa (Williams) 1, 2,

Sericoidea sp. 2

Skenidioides costatus (Cooper) 3, 4,

Sowerby ella spp. 1, 2,

Strophomena sp. 4

Trematis sp. 2

Vellamo sp. 5

Strophomenid gen. et sp. indet. 5

Plectambonitid gen. et sp. indet. 5

Clitambonitid gen. et sp. indet. 5

Flexicalymene caractaci (Salter) 3, 4

Flexicalymene ( Reacalymene ) limb a
(Shirley) 1

Flexicalymene planimarginata (Reed) 3, 4

Flexicalymene sp. 2

Illaenus sp. 2

Kloucekia apiculata (M’Coy) 3, 4,

Otarion sp. 3

Parabasilicus powisi (Salter) 1 , 2,

Pharostoma sp. 5

Proetidella sp. 2, 3

Kokenospira sp. 1,2

Lophospira spp. 1, 2,

Murchisonia sp. 2

? Seely a sp. 2

Sinuites soudleyensis (Reed) 1, 2

Sinuites spp. 1,2,

5

3, 4,5

3,5

4

5

3,4

5

3

3, 4, 5
3, 4,5

PICKERILL AND BRENCHLEY: ORDOVICIAN MARINE COMMUNITIES

239

Table 2 ( cont.) \

Bivalvia

Ambonychia sp. 2

? Area sp. 3

Colpomya sp. 2

Ctenodonta sp. 3

Cyrtodontid gen. et sp. indet. 1 , 2

Goniophorina sp. 2

Others

Cystoid plates 1 , 2

Crinoid ossicles 1, 2, 3, 4, 5

Favosites sp. 3

Monticulipora sp. 1,2, 3, 4, 5

Orthoceras sp. 1, 2, 3

1 = Lower Soudleyan

2= Middle and upper Soudleyan

3= Lower Longvillian Cwm Rhiwarth Siltstones

Modiolopsis modiolaris (Conrad) 1 , 2

Modiomorphid gen. et sp. indet. 1, 2

? Psilonychia sp. 2

Pterineid gen. et sp. indet. 2

Vlasta sp. 2

Pyritonema sp. 3

Stenopora sp. 1,2, 3, 4, 5

Stick bryozoa indet. 1, 2, 3, 4, 5

Tallinnella scripta (Harper) 1 , 2, 3

Tentaculites sp. 3, 4, 5

4= Lower Longvillian Pen-y-Garnedd Limestone
5= Upper Longvillian Pen-y-Garnedd Limestone

Our experience from a study of the 403 samples from the south Berwyns was that
certain genera commonly occurred together forming recurrent associations and also
that particular genera numerically dominated certain associations. Consequently our
definition of the communities has attempted to take account of both the association
of taxa and their abundance. Q-mode and R-mode cluster analysis could have pro-
duced results which were reproducible but could not handle taxa association and
abundance at the same time (MacDonald 1975) and we are not convinced that the
statistically defined clusters would have been more meaningful than our intuitively
clustered groups. We were influenced in our decisions as to which taxa should be
included in a particular community in the south Berwyns by our parallel studies in
Shropshire and Snowdonia, where coarse sandstone lithofacies are better differen-
tiated and contain more discrete brachiopod associations.

We have defined each community by the presence of certain characteristic genera
which must form a greater percentage of an assemblage than elements from any single
other community. Most assemblages in the south Berwyns could be assigned to a com-
munity on this basis though there are a small number of assemblages composed of
ubiquitous genera which cannot be rigorously assigned.

Wherever possible the material was identified to species level and a faunal list is
given in Table 2. We failed to recognize any species of the same genus which showed
different ecological preferences and therefore believe that there is little loss of informa-
tion in describing the communities at generic rather than specific level. This procedure
has avoided the cumbersome formulation of many communities which have essen-
tially the same structure and composition.

The communities which we recognize commonly show a moderately good correla-
tion with a particular substrate. This is hardly surprising as the majority of taxa in
this study are benthic and therefore the substrate is a variable with potentially power-
ful ecological effects. Not only is substrate readily observable, but it also reflects

240

PALAEONTOLOGY, VOLUME 22

table 3. Composition of the Howellites community in the south Berwyn Hills

Included here are all assemblages in which Howellites, Sowerbyella, and Paracraniops form a greater
percentage of the assemblage than elements from any single other community. The column headed ‘charac-
teristic community’ adopts the general model proposed by Johnson (1972) to indicate the types of genera
present in the particular community. Thus C=a characteristic genus, I=an intergrading genus, and U=an
ubiquitous genus. For the intergrading genera we have also indicated their characteristic community or
sub-community where H = Howellites community, Da = Dalmanella community, N = Nicolella community,
O = Onniella community, DD = Dinorthis community, Dinorthis sub-community, and DM = Dinorthis com-
munity, Macrocoelia sub-community. Thus, for example, IDa and IN we regard as intergrading genera from
the Dalmanella and Nicolella communities respectively. Column A represents a presence percentage where
the number of collections in which a genus occurs is divided by the total number of collections of a particular
community ( x 100) and is therefore a measure of how widespread a particular genus is. Thus, for example,
Howellites occurs in 95-8% of all collections assigned to this community (229 out of 239). Column B repre-
sents the average percentage abundance of an individual genus in those collections of a community where
the particular genus is present, thus indicating its average percentage of occurrence. Therefore, employing
the same example, Howellites occurs in 95-8% of all collections and in these collections occurs with an
average abundance of 47-2%. Column C represents the average percentage occurrence of each genus within
all the collections allocated to the community. Thus, out of a total of 239 samples, Howellites has an average
abundance of 45-2%.

Genera

Group
(Superfamily
or order)

Characteristic

Community

A

Presence

%

B

%

Abundance

C

Average

%

Brachiopods

1. Howellites

Enteletacea

C

95-8

47-19

45-21

2. Sowerbyella

Plectambonitacea

C-U

74-9

43-18

32-34

3. Paracraniops

Lingulacea

c

42-3

11-95

5-05

4. Macrocoelia

Strophomenacea

IDM

38-9

9 10

3-54

5. Dinorthis

Orthacea

IDD

36-4

9-86

3-59

6. Reuschella

Enteletacea

IDD

28-0

6-37

1-79

7. Bicuspina

Triplesiacea

IDa

8-8

3-68

0-32

8. Onniella

Enteletacea

IO

50

6-59

0-33

9. Leptaena

Strophomenacea

IDa

4-6

1-21

0-06

10. Rostricellula

Rhynchonellacea

IDM

3-8

3-44

0-13

11. Dalmanella

Enteletacea

IDa

2-9

13-35

0-39

12. Heterorthis

Enteletacea

IDD

2-5

13-98

0-35

13. Lingula

Lingulacea

U

2-1

1-55

0-01

14. Kjaerina

Strophomenacea

IDa

1-7

3-77

0-01

15. Kiaeromena

Strophomenacea

IDa

1-3

2-73

0-03

16. Eoplectodonta

Plectambonitacea

IN

0-8

1-77

0-02

17. Dolerorthis

Orthacea

IN

0-8

7-34

0-06

18. Oxoplecia

Triplesiacea

IDa

0-4

1-21

0-01

19. Sericoidea

Plectambonitacea

IO

0-4

0-72

0-01

Trilobites

20. Broeggerolithus

Trinucleina

U

64-7

1-02

0-66

21. Brongniartella

Calymenina

C

38-9

0-65

0-25

22. Parabasilicus

Asaphacea

C

230

0-70

0 16

23. Flexicalymene

Calymenina

IDa

6-3

0-31

0-02

24. Illaenus

Illaenina

?

0-4

0-21

0-01

PICKERILL AND BRENCHLEY: ORDOVICIAN MARINE COMMUNITIES

241

Genera

Group
(Superfamily
or order)

Characteristic

Community

A

Presence

%

B

%

Abundance

C

Average

%

Gastropods

25. Cyclonema

Platyceratacea

C

13-0

4-26

0-58

26. Cyrtolites

Bellerophontacea

IDa

2-1

304

0-24

27. ? Seelya

Murchisoniacea

C

2-1

1-65

0-08

28. Sinuites

Bellerophontacea

C?

0-4

0-24

0-04

29. Bucaniopsis

Bellerophontacea

C?

0-4

0-22

004

Bivalves

30. Modiolopsis

Modiomorphacea

c

13-8

2-56

0-45

31. Goniophorina

Modiomorphacea

c

2-9

0-21

0-14

32. Area

Arcacea

c

2-1

0-23

0-14

33. Psilonychia

Ambonychacea

c

2-1

0-18

0-10

34. Vlasta

Praecardiacea

c?

0-8

0-14

009

Others

35. Tallinnella

Ostracode

u

30-5

7-46

2-28

36. Crinoids

—

u

45-6

4-02

1-83

37. Bryozoa

—

u

13-4

4-60

0-62

38. Tentaculites

Cricoconarid

IDa

1-7

2-15

004

39. Orthoceras

Nautiloid

U

1-7

2-57

004

Ichnofauna includes Skolithos, Planolites, Teichichnus, ? Gyrochorte, ? Palaeophycus, Monomorphichnus,
Dimorphichnus, and Cruziana.

Number of collections— 239 (Cwm Rhiwarth Siltstones, locality details in SUP 14011).

Relationship to substrate— mudstone 21, silty mudstone 27, muddy siltstone 69, siltstone 72, fine sandstone 40.
Brachiopod diversity — 3-6.

Total diversity— 6-1.

other factors such as aeration, stability, degree of consolidation, and organic matter
(Fiirsich 1976). Indeed, the importance of substrate in influencing the distribution of
organisms or communities has been demonstrated by workers in both Recent (e.g.
Petersen 1911, 1913; Craig and Jones 1966; Driscoll 1967; Johnson 1971) and ancient
(e.g. Wobber 1968; Fiirsich 1976) examples. However, substrate does not appear to
have been the only factor in the community distribution and we therefore discuss other
likely limiting environmental parameters for each community.

The Howellites community

The Howellites community is characterized by the brachiopod genera Howellites,
Sowerbyella, and Paracraniops, which all appear to have been ecologically tolerant
genera as they also occur within all the other communities described here but with a
lower abundance. Commonly occurring trilobites include the ubiquitous Broeg-
gerolithus, Brongniartella , and Parabasilicus. Other characteristic taxa include the
gastropods Cyclonema, Sinuites, Bucaniopsis, and ? Seelya, and the bivalves Modio-
lopsis, Goniophorina, Vlasta, lArca, and Psilonychia. Apart from Cyclonema and
Modiolopsis, however, these taxa always occur in a low number of collections and

242

PALAEONTOLOGY, VOLUME 22

with low abundance. The community must also have possessed an important poly-
chaete and oligochaete annelid element as ichnofaunas produced by these taxa,
particularly Skolithos, Planolites, and Teichichnus, are also characteristic and occur
with moderate to high frequency of abundance (Pickerill 1977). Other elements found
in associations referred to the Howellites community are considered to be taxa from
adjacent coexisting communities. Thus, Macrocoelia, Reuschella, Dinorthis, Bicus-
pina, Dalmanella, Leptaena, Rostricellula, Heterorthis, Kjaerina, Kiaeromena, and
Flexicalymene intergrade from the more sandy substrates associated with the Dinor-
this and Dalmanella communities, Eoplectodonta and Nicolella intergrade from the
more silty substrates associated with the Nicolella community, and Onniella and
Sericoidea intergrade from more muddy substrates. Such intergrading taxa, apart
from Macrocoelia , Reuschella , and Dinorthis, are found in only occasional assem-
blages and usually with a low abundance (Table 3).

The community was widespread on silty mud and muddy silt substrates in the
Soudleyan throughout both the north and south Berwyn Hills, and was also common
in the lower Longvillian when similar substrates and environmental conditions pre-
vailed. The community was also present in the Bala district and east Snowdonia in
rocks of similar age and facies, and appears to have been the most widespread com-
munity in the Anglo-Welsh area (Williams 1963). In Snowdonia, for example, it can
be observed in muddy siltstones and similarly in the Arenig-Bala district, at Dolwyd-
delan and Betwys-y-Coed (e.g. Bryn Eithen SH 810518). In the south Berwyns good
localities exist throughout the region and the community can be particularly well
observed in rocks of Soudleyan age in Cwm Llech (SJ 016248) and lower Longvillian
age in the Main Quarry on the southern slopes of Allt y Main (SJ 178157).

The Howellites community has a notably low diversity of characteristic shelly
benthos and, in fact, most assemblages contain only three or four brachiopod genera
and occasional additional taxa such as bivalves, gastropods, and trilobites. Apart
from Howellites, Sowerbyella, and Paracraniops only the soft-bodied infaunal
benthos appears to have been reasonably common, as both ichnofauna and general
bioturbation are frequent. The sediments in which the Howellites community is most
frequently found are bioturbated silty mudstones and muddy siltstones with thin,
interbedded, parallel, cross-laminated, or rippled siltstones. The generally muddy
silt or silty mud substrate appears to have been relatively soft with moderate cohesion
which enabled the preservation of a varied ichnofauna and widespread bioturbation.
Muddy substrates with moderate cohesion are usually related to relatively high sedi-
mentation rates and such conditions are indicated for the Soudleyan which is repre-
sented by a large thickness (300-400 m) of muddy sediments. The facies association,
sedimentary structures, and ichnofaunas all suggest a generally low energy, shallow
subtidal environment of less than 25 m water depth (Pickerill 1977; Brenchley, in
press). A rather quiet environment is also suggested by the presence of numerous
articulated brachiopod shells and the preservation of many life clusters of Howellites,
in particular, Sowerbyella and Paracraniops.

The Dinorthis community

The characteristic brachiopods of the Dinorthis community in the south Berwyns are
Dinorthis, Heterorthis, Reuschella, Macrocoelia, and Rostricellula. Intergrading

PICKERILL AND BRENCHLEY: ORDOVICIAN MARINE COMMUNITIES 243

brachiopod genera include Howellites, Sowerbyella , and Paracraniops from the
muddy silt related Howellites community, and Bicuspina , Kiaeromena, Leptaena , and
Dalmanella from the Dalmanella community (Table 4). The associated non-
brachiopod fauna is typically quite sparse but does include occasional gastropods,
such as Cyrtolites, Lophospira , and Sinuites ; occasional bivalves, such as Byssodesma
and Psilonychia and low numbers of trilobites which include the ubiquitous Broeg-
gerolithus, in particular, and Brongniartella, Parabasilicus, and Flexicalymene. The
associated ichnofauna is also quite rare apart from Vermiforichnus, which occurs in
association with both Macrocoelia and Heterorthis (Pickerill 1976), and occasional
intergrading Skolithos and Planolites.

Within the Dinorthis community we recognize two sub-communities (‘populations’
in the sense of Bretsky (19706)), which are characterized by a different balance in the
abundance of the constituent genera and a different distribution in relation to
substrate (Table 4). The Dinorthis sub-community is characterized particularly by
Dinorthis together with variable proportions of Heterorthis and the common inter-
grading elements Reuschella and Macrocoelia and is particularly associated with
coarse silt and fine to coarse sand substrates. The Macrocoelia sub-community is
dominated by Macrocoelia , which often forms monospecific assemblages, but also
typically contains Reuschella and Rostricellula and intergrading Dinorthis and
Heterorthis. This latter sub-community is particularly associated with laminated fine
sandstones. The following discussion will therefore be directed towards each of the
sub-communities in turn.

1 . Dinorthis sub- community. In the south Berwyns the Dinorthis sub-community
is clearly defined only in the coarse volcaniclastic sandstones of the Swch Gorge Tuff.
Elsewhere, on finer substrates, Dinorthis and Heterorthis are commonly associated,
but often with intergrading elements from other communities, such as Howellites ,
Reuschella, and Sowerbyella (Table 4). Because of the limited development of a
coarse sandstone facies in the south Berwyns it was not clear whether this was
generally the preferred facies of the Dinorthis sub-community. We therefore examined
comparable lithofacies in Snowdonia and south Shropshire. In Snowdonia at Llyn
Cowlyd (SH 719615), for example, the sub-community contains abundant Dinorthis
and subordinate Macrocoelia and occurs in coarse sandstones (the Soudleyan
Multiplicata Sandstone of Diggens and Romano 1968). At Capel Curig (SH 709578)
a similar association occurs in massively bedded tuffaceous sandstones. In the
Costonian of Shropshire, Dinorthis- rich assemblages also containing Heterorthis,
harknesselids, and Salopia as major elements occur in fine conglomerates and coarse
sandstone facies of the Hoar Edge Grit (Table 5). The presence of a Dinorthis associa-
tion in coarse lithofacies at all the above localities suggests to us a preference for a
coarse substrate, but the presence of Dinorthis and Heterorthis, in particular, on
finer substrates in the south Berwyns suggests that these genera were sufficiently
eurytopic to colonize other environments.

In all those localities where coarse sandstones are present the assemblages occur as
disarticulated valve op bedding planes or as dispersed and sometimes fragmented
valves within massive coarse sandstones, which indicates that the assemblages were
reworked. However, the similarity of taxonomic composition of the assemblages

244

PALAEONTOLOGY, VOLUME 22

table 4. Composition of the Dinorthis community in the south Berwyn Hills. Included here are all assem-
blages in which Dinorthis, Macrocoelia, Reuschella, Heterorthis, and Rostricellula form a greater percentage
of the assemblage than elements from any single other community. Legend as in Table 3.

Genera Group Characteristic ABC

(Superfamily Community Presence % Average

or order) % Abundance %

A. Dinorthis Sub-community

Brachiopods

1. Heterorthis

Enteletacea

C

80-9

65-47

52-96

2. Howellites

Enteletacea

IH

69-7

16-16

11-26

3. Reuschella

Enteletacea

IDM

561

7-99

4-40

4. Dinorthis

Orthacea

C

51-7

15-02

7-76

5. Sowerbyella

Plectambonitacea

IH-U

49-4

26-26

12-98

6. Paracraniops

Lingulacea

IH

34-8

11-60

4-04

7. Macrocoelia

Strophomenacea

IDM

10-1

10-73

1-08

8. Lingula

Lingulacea

U

5-6

0-89

0-05

9. Bicuspina

Triplesiacea

IDa

4-5

1-77

0-08

10. Kiaeromena

Strophomenacea

IDa

2-3

6-71

016

11. Leptaena

Strophomenacea

IDa

2-3

2-22

0-05

12. Onniella

Enteletacea

IO

2-3

2-04

0-06

13. Dalmanella

Enteletacea

IDa

11

9-00

0-10

14. Oxoplecia

Triplesiacea

IDa

11

1-96

0-02

Trilobites

15. Broeggerolithus

Trinucleina

U

49.4

1-37

0-68

16. Brongniartella

Calymenina

IH

38-1

1-33

0-49

17. Parabasilicus

Asaphacea

IH

11-2

0-99

0-11

18. Flexicalymene

Calymenina

IDa

4-5

3-86

0-18

19. Proetus

Proetacea

?

1-1

0-66

0-01

Gastropods

20. Cyrtolites

Bellerophontacea

IDa

3-4

4-06

0-14

21. Lophospira

Pleurotomaracea

IDa

3-4

4-06

0-14

Bivalves

22. Byssodesma

Modiomorphacea

IDa

11

1-04

0-04

23. Pteriaceid indet.

?

1-1

0-90

0-01

Others

24. Crinoids

—

U

66-3

1-46

0-97

25. Tallinella

Ostracode

U

27-0

7-15

1-93

26. Bryozoa

—

u

101

2-82

0-29

27. Orthoceras

Orthocone nautiloid

u

3-4

1-01

0-03

Ichnofauna includes Vermiforichnus and occasional intergrading Skolithos and Planolites.

Number of collections— 89 (Cwm Rhiwarth Siltstones, locality details in SUP 14011).
Relationship to substrate— muddy siltstone 10, coarse siltstone 52, fine sandstone 27.
Brachiopod diversity— 3-0.

Total diversity— 5-0.

PICKERILL AND BRENCHLEY: ORDOVICIAN MARINE COMMUNITIES

245

Genera

Group

Characteristic

ABC

(Superfamily

Community

Presence % Average

or order)

% Abundance %

B. Macrocoelia Sub-community

Brachiopods

1. Macrocoelia

Strophomenacea

C

1000

39-52

39-52

2. Howellites

Enteletacea

IH

63-2

26-62

16-85

3. Reuschella

Enteletacea

C

57-9

7-52

4-35

4. Bicuspina

Triplesiacea

IDa

57-9

5-49

3-18

5. Paracraniops

Lingulacea

IH

47-4

11-75

5-57

6. Sowerbyella

Plectambonitacea

IH-U

26-3

36-73

9-67

7. Strophomena

Strophomenacea

C?

26-3

2-48

0-66

8. Leptaena

Strophomenacea

IDa

15-8

4-83

0-76

9. Dalmanella

Enteletacea

IDa

10-5

29-32

3-09

10. Rostricellula

Rhynchonellacea

C

10-5

10-85

1-14

11. Onniella

Enteletacea

IO

10-5

3-70

0-39

12. Dinorthis

Orthacea

IDD

10-5

3-45

0-36

13. Kiaeromena

Strophomenacea

IDa

10-5

2-02

0-21

14. Eoplectodonta

Plectambonitacea

IN

5-3

9-69

0-51

15. Kjerulfina

Strophomenacea

IDa

5-3

1-32

0-07

Trilobites

16. Broeggerolithus

Trinucleina

U

68-4

1-55

1-06

17. Brongniartella

Calymenina

IH

42-1

0-76

0-35

18. Parabasilicus

Asaphacea

IH

26-3

0-47

0-12

19. Flexicalymene

Calymenina

IDa

5-3

0-35

0-02

Gastropods

20. Cyrtolites

Bellerophontacea

IDa

10-5

6-25

0-66

21. Lophospira

Pleurotomaracea

IDa

10-5

6-25

0-66

22. Sinuites

Bellerophontacea

IDa

5-3

5-39

0-29

Bivalves

23. Psilonychia

Ambonychacea

IH

15-8

13-78

2-18

24. Pteraceid indet.

?

5-3

10-00

0-53

Others

25. Crinoids

_

U

89-5

12-01

10-75

26. Bryozoa

—

U

211

4-16

0-88

27. Tallinnella

Ostracode

u

15 8

5-26

0-83

28. Orthoceras

Orthocone nautiloid

u

5-3

1-48

0-08

Ichnofauna includes Vermiforichnus and occasional intergrading Planolites.

Number of collections— 19 (Cwm Rhiwarth Siltstones, locality details in SUP 14011).
Relationship to substrate— mudstone 0, coarse siltstone 10, laminated fine sandstone 9.
Brachiopod diversity— 3-9.

Total diversity — 7-5.

246

PALAEONTOLOGY, VOLUME 22

table 5. Composition of the Dinorthis association in south Shropshire. Included here are all assemblages
in which Dinorthis, Harknessella, Heterorthis, and Salopia form a greater percentage of the assemblage than
elements from any single other community. Legend as in Table 3.

Genera

Group
(Superfamily
or order)

Characteristic

Community

A

Presence

%

B

%

Abundance

C

Average

%

Brachiopods

1 . Dinorthis flabellulum

Orthacea

C

1000

28-8

28-8

2. Heterorthis patera

Enteletacea

C

1000

25-3

25-3

3. Harknessella

vespertilio

Enteletacea

?c

1000

19 0

19 0

4. Salopia salteri

Enteletacea

c

500

8-7

4-5

5. Dolerorthis sp.

Orthacea

IN

33-3

8-8

2-9

6. Leptaena sp.

Strophomenacea

IDa

33-3

3-3

1-6

7. Dalmanella sp.

Enteletacea

IDa

16-7

27-0

4-5

8. Rafinesquina cf.

complanata

Strophomenacea

IDM

16-7

180

3-3

9. Oxoplecia sp.

Triplesiacea

IDa

16-7

40

0-7

10. Platystrophia sp.

Orthacea

IN

16-7

2-5

0-4

11. Howellites sp.

Enteletacea

IN

16-7

1-3

0-2

Trilobites

12. Flexicalymene cf.

acantha

Calymenina

IDa

16-7

2-5

0-4

13. Costonia ultima

Trinucleina

U

16-7

1-2

0-2

Others

14. Bryozoa

—

U

500

10-7

5-8

15. Solenopora

16. Crinoids

Solenoporaceae

?

U

16-7

150

indet.

2-5

Ichnofauna includes rare intergrading Skolithos and occasional Planolites.

Number of collections— 6 (Hoar Edge Grit).

Relationship to substrate — coarse sandstone 6.

Brachiopod diversity— 4-7.

Total diversity— 5-8.

suggests that in spite of some transport the assemblages reflect original benthic
associations.

The epiclastic volcaniclastic sandstones of the Swch Gorge Tuff were deposited
in extremely shallow-water, sublittoral, well-oxygenated conditions, as laterally
they pass rapidly northwards into a subaerially deposited ignimbritic facies. In the
Costonian of Shropshire the coarse sediments are commonly massive or have large-
scale cross-stratification and lie immediately above the unconformity between the
basal Costonian and the Precambrian (see Greig et al. 1968). They are frequently
poorly sorted and appear to represent sand sheets deposited rather rapidly during
the basal Caradoc transgression across the irregular Precambrian surface. We regard
these sediments as an extremely shallow sublittoral facies and probably formed in
less than approximately 10 m water depth.

PICKERILL AND BRENCHLEY: ORDOVICIAN MARINE COMMUNITIES 247

The nature of the coarse substrates indicates that there must have been high-energy
conditions. Sedimentation rates must have been high for any single unit, as indicated
by the internal bedforms and the general absence of bioturbation and ichnofauna.
Turbidity must have been negligible, as indicated by the ‘clean’ nature of individual
sandstone units and we therefore infer that the Dinorthis sub-community was best
developed on shifting, coarse sand substrates in high-energy, non-turbid, well-
oxygenated environments of water depths of less than approximately 10 m.

2. Macrocoelia sub-community : In the south Berwyns the Macrocoelia sub-
community is best developed in rocks of Lower Longvillian age, particularly on the
slopes of Gallt yr Ancr (SJ 143125) and in Bryngwyn Quarry (SJ 182175), where it
also includes several intergrading elements from adjacent communities. At these
localities the coarse siltstones and laminated fine sandstones containing the Macro-
coelia sub-community are interbedded with muddy siltstones and silty mudstones
containing an associated Howellites community and with massively bedded fine
sandstones containing an associated Dalmanella community. Typically, the sub-
community has abundant Macrocoelia , common Reuschella, and relatively rare
Rostricellula as characteristic elements but also contains a variable proportion of
intergrading elements (Table 4).

In the south Berwyns the preferred substrate for the Macrocoelia sub-community
was coarse silt and laminated fine sand. This is also the case where the sub-community
is found in Snowdonia, Shropshire, and the Breidden Hills. For example, in
Snowdonia at Capel Curig (SH 709578) Macrocoelia- dominated assemblages with
occasional intergrading elements from the Dalmanella community are present in
parallel laminated sandstones of lower Longvillian age. In the Breidden Hills virtually
monospecific Macrocoelia assemblages are present in rocks of questionable age
(? upper Costonian-lower Soudleyan) but of similar lithology, and in Shropshire the
sub-community is also present in similar lithologies of Costonian age.

The preferred environment of the Macrocoelia sub-community is difficult to assess
from sedimentary evidence, but its intimate association with the Howellites com-
munity suggests that it occupied a similar range of 25 m or less. Furthermore, the
presence of the fossil-boring Vermiforichnus in association with Macrocoelia has been
taken by Pickerill (1976) to indicate water depths of 25 m or less. Sedimentation
rates for any single unit must have been at least moderately high, as indicated by the
parallel laminated nature of the sandstones, the absence of ichnofaunas and general
bioturbation, and the presence of the sub-community in reworked coquinite assem-
blages. As with the Dinorthis sub-community and for similar reasons turbidity must
have been low. We therefore interpret the preferred environment of the Macrocoelia
sub-community as having been reasonably similar to that of the Dinorthis sub-
community, but regard it as having been slightly more off-shore, in deeper waters,
and in slightly lower energy situations perhaps associated with more stable substrates.

The Dalmanella community

The characteristic elements of the Dalmanella community are the brachiopod genera
Dalmanella , Kjaerina , Bicuspina , Leptaena, and to a lesser extent Kiaeromena; the
trilobites Flexicalymene, Kloucekia, and possibly Otarion; the gastropods Lopho-
spira, Sinuites, and Murchisonia, and the bivalves Byssodesma and Colpomya.

248

PALAEONTOLOGY, VOLUME 22

Though the majority of these genera occur in many of the collections assigned to the
community they usually occur with a low or moderate frequency of abundance, and
of the characteristic genera only Dalmanella appears to have been abundant. Ichno-
faunas are also relatively common, though the majority appear to have been produced
by trilobites, for example, Rusophycus, Cruziana , and Trichophycus, or bivalves, for
example, Lockeia, and ichnofaunas produced by soft-bodied organisms are infrequent
though they do include Arenicolites and intergrading Skolithos and Planolites. Other
elements found in associations assigned to the Dalmanella community are considered
to be intergrading elements from adjacent coexisting communities. Of these inter-
grading elements the brachiopod genera Howellites and Sowerbyella from the
Hoxvellites community are major components, occurring in the majority of associa-
tions and with a moderate frequency of abundance (Table 6). Similarly the ubiquitous

table 6. Composition of the Dalmanella community in the south Berwyn Hills. Included here are all
assemblages in which Dalmanella , Bicuspina , Kjaerina, Kiaeromena, Kjerulfina, or Leptaena form a greater
percentage of the assemblage than elements from any single other community. Legend as in Table 3.

Genera

Group
(Superfamily
or order)

Characteristic

Community

A

Presence

%

B

%

Abundance

C

Average

%

Brachiopods

1. Dalmanella

Enteletacea

C

90-9

34-36

31-23

2. Howellites

Enteletacea

IH

88-6

30-06

26-64

3. Sowerbyella

Plectambonitacea

IH-U

79-6

20-27

16-13

4. Kjaerina

Strophomenacea

C

500

5-21

2-60

5. Bicuspina

Triplesiacea

C

45-5

7-04

3-20

6. Leptaena

Strophomenacea

C

40-9

1-71

0-70

7. Kiaeromena

Strophomenacea

C

36-4

2-42

0-88

8. Macrocoelia

Strophomenacea

IDM

29-6

8-99

2-65

9. Dinorthis

Orthacea

IDD

11-4

7-41

0-84

10. Nicolella

Orthacea

IN

11-4

6-90

0-78

11. Dolerorthis

Orthacea

IN

91

2-93

0-27

12. Paracraniops

Lingulacea

IH

6-8

1-69

0-11

13. Strophomena

Strophomenacea

C

6-8

0-43

0-03

14. Oxoplecia

Triplesiacea

C

4-6

2-12

010

15. Skenidioides

Orthacea

IN

4-6

2-30

0-09

16. Kjerulfina

Strophomenacea

C

4-6

1-89

0-09

17. Lingula

Lingulacea

u

4-6

0 61

0-01

18. Rostricellula

Rhynchonellacea

IDM

2-3

2-95

0-07

19. Strophomenid indet.

?

2-3

0-37

0-01

Trilobites

20. Broeggerolithus

Trinucleina

E

93-2

0-85

0-79

21. Flexicalymene

Calymenina

C

70-5

0-56

0-40

22. Kloucekia

Dalmanitacea

C

500

0-46

0-23

23. Parabasilicus

Asaphacea

IH

50-0

0-33

0-16

24. Brongniartella

Calymenina

IH

29-6

0-15

0-05

25. Conolichas

Lichidacea

IN

2-3

0-05

0 01

26. Chasmops

Dalmanitacea

IN

2-3

0-05

0-01

27. Otarion

Proetacea

C?

2-3

0-05

0-01

PICKERILL AND BRENCHLEY: ORDOVICIAN MARINE COMMUNITIES 249

Genera

Group
(Superfamily
or order)

Characteristic

Community

A

Presence

B

%

Abundance

C

Average

Gastropods

28. Lophospira

Pleurotomaracea

C

40-9

5-48

2-24

29. Sinuites

Bellerophontacea

C

36-4

4-03

1-46

30. Murchisonia

Murchisoniacea

c

29-6

2-66

0-78

31. Cyrtolites

Bellerophontacea

c

29-6

1-85

0-55

32. Cyclonema

Platyceratacea

IH

2-3

6-85

016

33. Bellerophontacean
indet.

?

2-3

30

0-07

Bivalves

34. Byssodesma

Modiomorphacea

c

91

1-89

0-17

35. Colpomya

Modiomorphacea

c

2-3

3-42

008

36. Pteriacean indet.

?

2-3

2-10

0-05

37. Modiomorphacean
indet.

?

2-3

1-30

0-02

Others

38. Bryozoa

—

u

56-8

10-03

5-70

39. Tallinnella

Ostracode

u

47-7

2-59

1-24

40. Crinoids

—

u

36-4

0-63

0-23

41. Tentaculites

Cricoconarid

c

22-8

1-40

0-32

42. Orthoceras

Orthocone nautiloid

u

15 9

0-60

0-10

43. Favosites

Tabulate coral

?

2-3

0-43

0-01

Trace fossils include Arenicolites , Lockeia, Trichophycus, Rusophycus, Cruziana and intergrading Skolithos
and Planolites.

Number of collections— 44 (Cwm Rhiwarth Siltstones, locality details in SUP 14011).
Relationship to substrate— mudstone 0, coarse siltstone 5, fine sandstone 39.
Brachiopod diversity— 51.

Total diversity — 10-7.

trilobite Broeggerolithus occurs in most assemblages but usually with a low frequency
of abundance. Other intergrading elements characteristically occur in only a few
collections and with a low frequency of abundance. These include the brachiopods
Macrocoelia, Dinorthis, Strophomena, and Rostricellula from the Dinorthis com-
munity, and Nicolella and Dolerorthis from the Nicolella community, and the trilo-
bite genera Parabasilicus and Brongniartella from the Howellites community and
Deacybele, Platylichas, and Chasmops from the Nicolella community. Finally, the
Dalmanella community frequently contains bryozoa which appear to have been
moderately abundant, ostracodes and crinoids (Table 6).

In the south Berwyns the preferred substrate for the Dalmanella community was
fine sandstones or sometimes coarse siltstones (Table 6). A similar substrate pre-
ference can be observed in other areas. For example, in Snowdonia at Capel Curig
(SH 709578) and Betwys-y-Coed (SH 810518), it can be observed in similar litho-
logies of lower Longvillian age. In the south Berwyns the fine sandstones may be
massive, cross-stratified, or more occasionally exhibit small scale cross-lamination.
The community here is developed particularly in rocks of lower Longvillian age and

250

PALAEONTOLOGY, VOLUME 22

good localities exist in the south-east Berwyns where such sandstones have been
extensively quarried. Thus, on the south-eastern facing slopes of Allt y Main quarries
such as the Main Quarry (SJ 178157), Glascoed Quarry (SJ 142122), and the Bron-y-
main Quarries (SJ 167145) all illustrate the typical Dalmanella community. At these
localities, as elsewhere in the south Berwyns and Snowdonia, the sandstone facies are
invariably interbedded with muddy siltstones and silty mudstones, which charac-
teristically contain an associated Howellites community and also are commonly
associated with thick laminated fine sandstones containing the Macrocoelia sub-
community of the Dinorthis community.

Unfortunately there are no diagnostic environmental indicators of the precise
environmental position of the Dalmanella community. However, as it is frequently
associated with both the Howellites and Dinorthis ( Macrocoelia sub-community)
communities it would be imprudent to infer a substantial difference in depth. Con-
sequently, we believe that the community lived in water depths of 25 m or less. Where
the community is best developed in the south Berwyns the lower Longvillian sedi-
ments are at maximum just over 20 m in thickness, and we therefore infer that accumu-
lation rates were relatively slow, probably being related to reduced rates of
intrabasinal subsidence (Brenchley 1969). However, the presence of cross-
stratification and parallel lamination indicates that sedimentation rates for individual
sandstone units must have been relatively high and we therefore infer that sedimenta-
tion rate and substrate mobility may have been significantly related to this
community.

The Nicolella community

The characteristic brachiopod genera of the Nicolella community are Nicolella ,
Dolerorthis, Eoplectodonta , Platystrophia, Skenidioides, and Leptestiina. These genera
are present in the majority of collections and commonly occur with moderate to high
frequency of abundance (Table 7). Additional characteristic brachiopod genera
forming part of the community but only occurring in a few assemblages and usually
with a low or moderate abundance are Cremnorthis, Rhactorthis, Vellamo, Obolus,
and Lingulasma. It is notable that these latter genera are entirely restricted to the
Nicolella community. Characteristic trilobites include Conolichas, Deacybele,
Calyptaulax, Chasmops , and Estoniops, which though present in several assemblages
are not necessarily abundant. Other taxa found in assemblages designated to the
Nicolella community are considered to be intergrading elements from adjacent
coexisting communities. These elements include the brachiopod genera Sowerbyella,
Bicuspina, Dalmanella, Kiaeromena, Reuschella, Rostricellula, Howellites, Para-
craniops, Lingulella, Kjaerina, Strophomena, and Onniella; the trilobites Flexicaly-
mene, Broeggerolithus, Brongniartella, Parabasilicus, and Kloucekia; the gastropods
Sinuites and Lophospira, and the ichnofauna Skolithos and Planolites (Table 7). These
intergrading taxa are normally found in only a few assemblages and characteristically
exhibit a low frequency of abundance. Finally, the community frequently contains
indeterminate bryozoan and crinoidal debris.

Williams (1963) noted that the Nicolella ‘association’ was of lower Longvillian age
in the Bala district and Marshbrookian-Actonian in Shropshire. In the south Berwyns
there is evidence to suggest that the Pen-y-Garnedd Formation and its associated

PICKERILL AND BRENCHLEY: ORDOVICIAN MARINE COMMUNITIES

251

Bala

text-fig. 5. Diachronism of the Pen-y-Garnedd Formation from the north-west to the south-east Berwyns
based on the first recorded occurrence of listed trilobite species. Rocks similar to the Pen-y-Garnedd Forma-
tion in the Bala area are known as the Gelli Grin Calcareous Ashes, and the distribution of the trilobites is
recorded in Bassett et al. (1966).

Nicolella community is diachronous within the Longvillian (text-fig. 5). Thus, early
in the Longvillian in the north-western part of the south Berwyns, for example at Nant
Achlas (SJ 014267), the Pen-y-Garnedd Formation and its associated Nicolella
community passes laterally south-eastwards into fine sandstones of the Upper Cwm
Rhiwarth Formation and its associated Dalmanella community, with the result that
the Nicolella community was ‘diluted’ by intergrading taxa of this adjacent com-
munity (text -fig. 6). Thus, intergrading elements from the Dalmanella community are
typically found in association with the Nicolella community in the lower horizons of
the Pen-y-Garnedd Formation and include Dalmanella , Leptaena, Kjaerina, Kiaero-
mena, and Bicuspina. Progressively during the Longvillian, calcareous sediments of
the Pen-y-Garnedd Formation became more widespread and the Dalmanella com-
munity was excluded from the area, with the result that the Nicolella community
sensu stricto became firmly established and intergrading elements decreased in both

252

PALAEONTOLOGY, VOLUME 22

frequency of occurrence and abundance vertically in any one section. A similar
pattern exists with components of the coexisting Howellites community, particularly
Howellites and Paracraniops, which initially intergrade from adjacent muddy silt and
silty mud substrates, but as the Nicolella community became more firmly established
this intergradation became less pronounced. Conversely, components of the Onniella
association of Williams (1973), particularly Onniella itself, is found to increase in
frequency of occurrence and abundance vertically in any one section, this presumably
reflecting or ‘anticipating’ in the sense of Worsley (1971) the incoming of muds (Pen-y-
Garnedd Shale) with a related Onniella association containing Onniella and
Sericoidea.

The Nicolella community does not appear to be specifically related to substrate as
it may be found in a variety of lithofacies. For example, in Nant Achlas (SJ 014267)

Nant

Cwm

Achlas

Llech

Llanfyllin Meifod

i r

DALMANELLA COMMUNITY
with Da/manella, Bicuspina. Leptaena etc. sometimes;
intergrading with:- D/NORTHIS COMMUNITY with AV
Macrocoelia and Dinorthis ■' ■ .... ..

text-fig. 6. Vertical and lateral community and facies relationships at the top of the Cwm Rhiwarth
Siltstones and base of the Pen-y-Garnedd Formation.

PICKERILL AND BRENCHLEY: ORDOVICIAN MARINE COMMUNITIES 253

it is associated both with sandy wackite conglomerates and bedded calcareous silt-
stones; in the south-eastern Berwyns, on the southern slopes of Allt Fawr (SJ 147138),
it is associated with metre-thick massively bedded siltstones, and in the Main Quarry
(SJ 178157) it is associated with bedded muddy siltstones and fine sandstones. With
the exception of the latter locality the sediments are generally calcareous owing to a
variable bioclastic component.

The general absence of sedimentary structures, the virtually complete bioturbation
of the sediment, and the increased accumulation of fine bioclastic debris suggests that
sedimentation rates were relatively slow. Energy conditions were likely to have been
variable from time to time, as reflected by the variation in lithology. However, though
high-energy conditions briefly moved coarse sediment, the rocks usually consist of
coarse silt mixed with poorly sorted bioclastic debris and a low-energy situation must
generally have prevailed. The most common substrates associated with the Nicolella
community appear to have been silt and fine sand (Table 7) and this suggests that the
community colonized an environment similar to the Dalmanella community. How-
ever, the lower rates of sedimentation and lower-energy conditions suggest that the
community occupied a slightly more olf-shore position. The depth parameter cannot
be accurately assessed but it is unlikely to have greatly exceeded the 25 m suggested
previously for the Dalmanella community.

Diversity of the Nicolella community is moderate to high, which presumably
reflects low environmental stress typically associated with more off-shore environ-
ments (Bretsky and Lorenz 1970) and the increasing stability and predictability of
the environment (Slobodkin and Sanders 1969).

Summary of communities. Though the four communities which we recognize in the
Caradoc of the south Berwyn Hills all show some correlation with substrate (Tables
3-7) we do not consider this to have been the only limiting environmental factor.
Despite the fact that it is difficult to assign an absolute depth of habitation to any
of the defined communities, all four probably existed within a relatively narrow
depth range (0-730 m) and we do not recognize, for example, the distinctive depth
zonation commonly described in the Silurian (Ziegler 1965; Ziegler et al. 1968; Calef
and Hancock 1974). There does, however, appear to be a broad correlation with depth
and, perhaps more importantly, distance from shore. Thus, with increasing distance
from shore we find the Dinorthis community, the Howellites and Dalmanella com-
munities, and the Nicolella community. The more in-shore communities were affected
by a relatively higher sedimentation rate which is reflected in their lower diversities.
The two most off-shore communities, the Dalmanella and Nicolella communities,
occupied more stable environments and exhibit slightly higher diversities (cf. Bretsky
and Lorenz 1970).

There is an approximate correlation between distance from shore and turbulence
and we find the Dinorthis community related to higher-energy environments, and the
Howellites and Nicolella communities to quieter regimes, with the Dalmanella com-
munity in an intermediate situation. It should be noted, however, that both the
higher diversity Nicolella community and the lower diversity Howellites community
occupied environments of similar low energy so that at least in these cases turbulence
does not appear to have been a limiting factor in community distribution.

table 7. Composition of the Nicolella community in the south Berwyn Hills. Included here are all assem-
blages in which Nicolella , Eoplectodonta, Platystrophia, Skenidioides, Leptestiina, and Dolerorthis form
a greater percentage of the assemblage than elements from any other community. Legend as in Table 3.

Genera Group Characteristic ABC

(Superfamily Community Presence % Average

or order) % Abundance %

Brachiopods

1.

Nicolella

Orthacea

2.

Dolerorthis

Orthacea

3.

Platystrophia

Orthacea

4.

Skenidioides

Orthacea

5.

Leptestiina

Plectambonitacea

6.

Eoplectodonta

Plectambonitacea

7.

Howellites

Enteletacea

8.

Sowerbyella

Plectambonitacea

9.

Cremnorthis

Orthacea

10.

Vellamo

Clitambonitacea

11.

Leptaena

Strophomenacea

12.

Bicuspina

Triplesiacea

13.

Rhactorthis

Orthacea

14.

Kiaeromena

Strophomenacea

15.

Strophomena

Strophomenacea

16.

Lingulasma

Lingulacea

17.

Dinorthis

Orthacea

18.

Reuschella

Enteletacea

19.

Rostricellula

Rhynchonellacea

20.

Obolus

Lingulacea

21.

Kjaerina

Strophomenacea

22.

Onniella

Enteletacea

23.

Paracraniops

Lingulacea

24. Lingulella

25. Strophomenid indet.

26. Plectambonitid indet.

27. Clitambonitid indet.

Trilobites

Lingulacea

28.

Chasmops

Dalmanitacea

29.

Estoniops

Dalmanitacea

30.

Conolichas

Lichidacea

31.

Deacybele

Cheirurina

32.

Flexicalymene

Calymenina

33.

Broeggerolithus

Trinucleina

34.

Kloucekia

Dalmanitacea

35.

Calyptaulax

Dalmanitacea

36.

Brongniartella

Calymenina

37. Parabasilicus

Gastropods

Asaphacea

38.

Sinuites

Bellerophontacea

39. Lophospira

Others

Pleurotomaracea

40. Bryozoa

—

41.

Crinoids

—

42.

Pyritonema

Sponge spicules

C

950

3605

34-25

C

85-0

24-24

20-60

c

80-0

19-43

15-54

c

600

15-06

9-04

c

55-0

10-39

5-71

c

500

8-82

4-41

IH

45-0

6-43

2-89

IH-U

40-0

8-46

3-38

C

25-0

2-65

0-66

C

200

1-67

0-33

IDa

200

2-20

0-44

IDa

20-0

2-86

0-57

C

15-0

2-49

0-37

IDa

15-0

1-87

0-28

IDM

15-0

1-24

0-19

C

15-0

4-33

0-65

IDD

100

2-01

0-20

IDM

100

0-94

0-09

IDM

100

1-13

0-11

C?

100

1-07

0-10

IDa

100

1-64

0-17

IO

100

3-43

0-34

IH

50

2-62

0-13

U

5-0

1-83

0-09

50

0-89

0-04

50

0-86

0-04

50

0-88

0-04

C

55-0

1-82

1-00

C

50-0

1 01

0-50

C

45-0

1-05

0-47

C

35-0

0-48

0-19

IDa

35-0

1-26

0-50

U

25-0

0-69

0-17

IDa

15-0

0-38

0-06

C

10-0

0-16

0-02

IH

100

0-25

0-03

IH

50

0-23

0-01

IDa

15-0

1-62

0-24

IDa

150

1 68

0-25

U

600

5-49

3-29

U

50-0

3-46

1-73

?

5-0

0-20

0-01

Ichnofauna includes rare intergrading Skolithos and Planolites.

Number of collections— 20 (Pen y Garnedd Formation, locality details in SUP 14011).

Relationship to substrate— wackite conglomerates 2, sandstone 1, fine sandstone 8, coarse siltstones 8, mudstones 1.
Brachiopod diversity— 14-0.

Total diversity — 19-0.

PICKERILL AND BRENCHLEY: ORDOVICIAN MARINE COMMUNITIES 255

Other environmental factors which are commonly related to the distribution of
marine organisms or communities, such as salinity and water temperature (Jones
1950; Berry and Boucot 1967) are considered to have been unimportant in determin-
ing the distribution of the Caradoc communities described here. There is no evidence
of a near-by major land area, which implies that the marine region was not subject to
fluviatile influences and was not one of restricted near-shore marine circulation, thus
removing the possibility of local brackish or hypersaline conditions. In addition, the
area is of such a restricted geographical extent that geographical variations in
temperature are unlikely. Availability of food as a controlling environmental para-
meter in the distribution of benthic communities has been stressed by Marshall
(1954), Calef and Hancock (1974), and Fursich and Hurst (1974). Generally, increase
in depth is accompanied by a decrease in organism density which reflects a decreased
food supply. Unfortunately there is no positive evidence by which to assess food
supply in the rocks described here, but we feel that because the communities all
existed within a relatively narrow depth range, variation in food supply was unlikely
to have been pronounced, though we assume that it did decrease with increased depth
and distance off-shore. Other depth-related factors, such as oxygen content of the
water, pressure and light, etc., are considered to have been unimportant. The salient
environmental parameters for each of the communities are summarized in Table 8
and represented schematically in text-fig. 7.

Within the Nant Hir Group (Costonian-Harnagian) of the Bala district, Williams
(1973) recognized, but unfortunately without giving details, a further community,
the Onniella community, which he correlated with a mud substrate. In the Berwyn
Hills there is a possibility that this community is present within the lower horizons
of the Pen-y-Garnedd Shales (text-fig. 2), where Onniella is associated with Sericoidea ,
Paterula, and graptolites, but outcrop is limited and therefore sample numbers for this
horizon are small and the significance of this association is therefore difficult to assess.

table 8. Relationship between the Caradoc communities and interpreted environmental parameters.

Sedimentation Distance

Community

Diversity

Substrate

DINORTHIS

low

medium to

( Dinorthis sub-

3-0 brachiopods

coarse sand

community)

5-0 total

DINORTHIS

low-medium

fine sand

(Macrocoelia

3-9 brachiopods

sub-community)

7-5 total

DALMANELLA

medium

fine sand to

5 1 brachiopods
10-7 total

coarse silt

HOWELLITES

low

3-6 brachiopods
6T total

muddy silt

NICOLELLA

high

variable

14-0 brachiopods

calcareous

19-0 total

silt or sand

Rate

Energy

from Shore

Depth

high

high

nearshore

70-10 m

high

medium

nearshore

moderate

25 m

low but
variable

low but
variable

moderate

25 m

high

low

moderate

25 m

low

low

offshore

730 m

256

PALAEONTOLOGY, VOLUME 22

text-fig. 7. Schematic illustration of the relationship of the Caradoc communities and environmental

parameters.

If, however, the associations observed in the Pen-y-Garnedd Shales are representative
of the Onniella community, it is likely that the community was associated with a mud
substrate accumulating under off-shore conditions in a very low-energy environment
and with low rates of sedimentation. Lower in the Berwyn succession species of
Onniella are found occasionally within the Howellites community (Table 3), and pro-
visionally might be regarded as intergrading elements from the ‘ Onniella community’.

STRATIGRAPHICAL DISTRIBUTION OF COMMUNITIES IN THE
SOUTH BERWYNS

The above discussion of Caradoc communities is based essentially on Soudleyan to
upper Longvillian assemblages from the south Berwyns. However, similar associations
are found in the north Berwyns (Brenchley, in press) and we have found similar
communities in Soudleyan and Longvillian rocks of Snowdonia. Communities with
the same general composition occur in the Breidden Hills and Shropshire, in rocks
ranging in age from Costonian to Longvillian but additional communities are almost
certainly present, as for example in the Shelve area of Shropshire where Williams
(1974) has described other faunal associations which he assigned (Williams 1976) to
a broadly delineated Bicuspina Set. We conclude that some of the communities had a
relatively stable composition at generic level throughout the lower part of the Caradoc
of North Wales and Shropshire, though there were many examples of species replace-
ment and some changes in generic composition (see Williams 1963).

PICKERILL AND BRENCHLEY: ORDOVICIAN MARINE COMMUNITIES

257

In the south Berwyns the rocks of Soudleyan age characteristically contain the
Howellites community, which was the indigenous community on muddy-silty sub-
strates throughout this stage. Intergradation from adjacent co-existing communities
occurred throughout the succession, particularly when there was intergradation of
lithofacies. Thus, at the coarse end of the silt-grade spectrum intergradation took
place with genera from the sand-related Dinorthis community or Dalmanella com-
munity. Occasionally, when environmental parameters were suitable, the indigenous
Howellites community was replaced by the Dinorthis community (text-fig. 8). This

U.Longvillian

L.Longvillian

U + M. Soudleyan

L. Soudleyan

NICOLELLA COMMUNITY

DALMANELLA COMMUNITY (with elements of

Dinorthis and Macrocoelia sub- community)

-Enrichment from Dinorthis community

HOWELLITES COMMUNITY plus intergrading elements
from Dinorthis and Onniella community

Enrichment and replacement from Dinorthis community
( Dinorthis in particular, and Macrocoelia sub -community)

DINORTHIS COMMUNITY (with Dinorthis sub- community)

HOWELLITES COMMUNITY

Calcareous siltstones and pebbly calcareous siltstone
Sandstone and siltstone

m Volcaniclastic sandstones of the Swch Gorge Tuff
jjgllll Mudstone and silty mudstone

Intergradation between adjacent communities

text-fig. 8. Generalized stratigraphical section for the south Berwyns showing the predominance of the
Howellites community and the periodical invasions of elements from other communities, and finally the
establishment of the more off-shore Dalmanella and Nicolella communities.

258

PALAEONTOLOGY, VOLUME 22

occurred when the near-shore volcaniclastic sandstones of the Swch Gorge Tuff were
developed in the lower Soudleyan, when virtually monospecific Dinorthis assemblages
predominated, and again when fine sandstone deposition prevailed throughout the
area a short distance (40-50 m) above the tuff, when assemblages dominated by
Heterorthis together with Dinorthis and Reuschella prevailed. Near the top of the
upper Soudleyan the Howellites community was enriched by intergrading elements
from both the Dinorthis community, such as Dinorthis, Reuschella, Macrocoelia, and
Rostricellula, and from the Dalmanella community, such as Leptaena and Kiaeromena.

In the lower Longvillian the widespread development of fine sandstones was accom-
panied by an associated Dalmanella community. Periodically, when environmental
conditions were amenable, the Dalmanella community was replaced by the Dinorthis
community, particularly by the Macrocoelia sub-community. Irrespective of which-
ever community was prevalent, intergradation took place between these communities
themselves and from the Howellites community. The Dalmanella community was
replaced upwards during the lower Longvillian by the Nicolella community where
there was reduced sedimentation but elsewhere in the substage the Dalmanella com-
munity continued to prevail. At this time intergradation between the Nicolella and
Dalmanella communities was prominent. By upper Longvillian times the Nicolella
community had become more widespread and firmly established throughout the
whole of the south Berwyns so that only ubiquitous elements such as Sowerbyella are
found in association with the characteristic Nicolella community. In the upper part of
the upper Longvillian, Onniella, which occurs in the overlying Pen-y-Garnedd Shales,
is found, possibly representing an intergrading element from adjacent muddy
substrates.

DISCUSSION

Although the faunal assemblages in the Berwyn area are similar to those around Bala
and show comparable recurrent associations, we differ in some respects from
Williams (1963, 1973) as to how the series of intergrading associations might best be
grouped into communities. Nevertheless, we confirm the prevalence of a Howellites
community in muddy-silty sediments and the presence of a Nicolella community in
the relatively calcareous sediments of central north Wales. The Dinorthis association
of Williams (1963) we regard as being divisible into two communities, viz. the Dinor-
this ahd Dalmanella communities, and the former can be usefully split into two sub-
communities. Where sandy and silty facies are poorly differentiated, as in the Berwyn
and Bala areas, there is intergradation of the Dalmanella and Dinorthis communities,
but where substrates are well differentiated, as in the Costonian of Shropshire or the
Soudleyan of Snowdonia, the communities are clearly defined. We consider that this
use of end members in a chain of intergrading benthic associations is helpful in the
identification and naming of communities. It is our experience that throughout the
whole Berwyn area and at many other localities in North Wales and Shropshire, most
faunal assemblages of Soudleyan or Longvillian age can be reasonably assigned to
one of the five communities discussed in this paper. We therefore conclude that in
spite of the extensive intergradation of the benthic faunas they can be usefully
partitioned into communities representing environmentally controlled, re-occurring
natural associations.

PICKERILL AND BRENCHLEY: ORDOVICIAN MARINE COMMUNITIES 259

The degree to which communities intergrade appears to be related to the environ-
mental gradients within a region. Where, as in North Wales, the basin succession is
thick and environmental gradients are low, both laterally and vertically, there is
considerable intergradation between communities, and mixed assemblages are found.
The variable composition of these assemblages implies a high rate of local immigra-
tion and extinction, and consequently a high equilibrium number for each community
(e.g. Bretsky and Bretsky 1976). In contrast, where a shelf succession is relatively
thin, and lateral and vertical environmental gradients are sharp (e.g. east Shropshire)
the communities are more clearly partitioned into distinct lithotopes, and the com-
positions of the communities are less variable. Such communities tend to show a lower
equilibrium number and therefore resemble those small, isolated geographical areas
such as oceanic islands (MacArthur 1972). Clearly a recognition of the local tectonic
and sedimentological framework in which particular communities are found is essen-
tial before more generalized models of communities can be framed.

Communities show many temporal changes in their composition as a result
of species evolution, the establishment of new niches within the community,
the extinction of some species, and the immigration of elements from other areas.
Changes of these kinds are found in the communities of the Soudleyan-Long-
villian of the Anglo-Welsh area. For example, Howellites and Sowerbyella are
represented by five stratigraphically arranged species (Williams 1963), Reuschella by
two species and a sub-species, and Dalmanella by four species. Most of these new
species appear to have arisen indigenously within the Anglo-Welsh area, and indeed
most of the Soudleyan and Longvillian brachiopod genera are either represented by
earlier species within the area or appear for the first time within this region. In
contrast, the appearance of some trilobites, notably Chasmops and Estoniops,
apparently results from immigration from the Baltic region (Dean 1960a).

Some of the mid-Caradoc communities, as defined at the generic level, occur more
widely within the Ordovician. The Dinorthis and Macrocoelia sub-communities are
well established by Costonian times in south Shropshire and existed until at least the
top Soudleyan or lower Longvillian in Wales. Assemblages which are possibly related
to the Dinorthis community, with Horderleyella and Rafinesquina, together with
coarsely ribbed orthids, such as Hesperorthis and Orthis , are found as early as the
upper Llanvirn of the Llandeilo District (Williams 1953), but these early assemblages
also include Dalmanella and Sowerbyella as major elements. Dalmanella is again
associated with Rafinesquina and Macrocoelia in the upper Llandeilo Calcareous
Ashes of the Berwyn Hills (MacGregor 1961), which suggests that the Dalmanella
community might not have become clearly differentiated from the Dinorthis com-
munity until the Caradoc. Faunas from the Spy Wood Grit (Costonian) of west
Shropshire which contain Dalmanella , Bicuspina , Kjaerina, and Rostricellula
(Williams 1974) are more reminiscent of a mid-Caradoc Dalmanella community.
Associations apparently similar to the Dalmanella community are also present in the
Marshbrookian of Shropshire (Dean 1958).

We do not find records of low diversity assemblages dominated by Howellites and
Sowerbyella outside the Soudleyan-Longvillian. In contrast, the Nicolella com-
munity appears to have a range extending from the Costonian at least into the Ashgill
where in zones 1-3 (Cautleyan) of the Cautley area, Ingham (1966) records faunas

260

PALAEONTOLOGY, VOLUME 22

including Nicolella, Dolerorthis, Glyptorthis, Platystrophia, Sampo , and Skenidioides.
Similar faunal associations, with the addition of Christiania , are known elsewhere in
the Ashgill, e.g. from flank beds of the Boda Limestone of Sweden (personal observa-
tion), from Belgium (Sheehan 1975), the Portrane Limestone of east Ireland (Wright
1964), and from the Drummuck Group of Girvan (Lamont 1935). It is possible that
this off-shore Ordovician community subsequently developed into the Dicoelosia-
Skenidioides community of the Lower and Middle Llandovery (Boucot 1975) with
the addition of new Silurian elements.

Attempts have been made to generalize Lower Palaeozoic communities into a few
major depth related types (Bretsky 1969a; Anderson 1971), and in these broad terms
the North Wales communities could all be referred to as orthid-strophomenid-
trilobite communities. More recently, Boucot (1975) has advocated the use of benthic
assemblages which comprise a group of communities that occur repeatedly in
different parts of a region in the same position relative to a shoreline. Following this
scheme the Caradoc communities could empirically be allocated to the Benthic
Assemblages of Boucot, as illustrated in column A below.

Benthic Assemblage 1 is typically represented elsewhere by a linguloid-bivalve
dominated community which we have not observed in the south Berwyns. The five
communities which we have discussed have been assigned to benthic assemblages
according to their relative depth.

Benthic Assemblage 1
Benthic Assemblage 2
Benthic Assemblage 3

Benthic Assemblage 4
Benthic Assemblage 5

A

Not recognized
Dinorthis community
Dalmanella and Howellites
communities
Nicolella community
Onniella community

B

Not recognized
Dinorthis community
Dalmanella , Howellites, and
Nicolella communities
? Onniella community

On this interpretation the Dinorthis community would be the Anglo- Welsh Caradoc
equivalent of the Upper Llandovery Eocoelia community, and the Nicolella com-
munity would equate with the Costistricklandia community. However, there are
alternative interpretations of the distribution of Caradoc communities relative to
benthic assemblages, and we prefer the distribution shown in Column B because,
as previously suggested, the depth ranges of the Dinorthis to Nicolella communities
could be as low as 0-30 m. The Onniella community in the Pen-y-Garnedd Shales
probably occupied a more off-shore position, though this does not necessarily imply
substantially greater depths. For example, Cave (1965) interpreted the black grapto-
litic shales of the Pen-y-Garnedd Shales and its lateral equivalents as occupying a
positive area of no great depth within the Welsh Basin.

Sparse faunas comparable to the Onniella community, comprising small shells, in
particular Sericoidea, and found in graptolitic shales, have been interpreted by
Sheehan ( 1977) as being benthic organisms attached to seaweed fronds or other firm
areas of the sea floor. Such associations appear to have been some of the deepest
in the Ordovician but may, nevertheless, have occurred within normal shelf depths.
Sheehan (1977) suggests that the limited biomass and diversity of the fauna might be

PICKERILL AND BRENCHLEY: ORDOVICIAN MARINE COMMUNITIES 261

the result of a deficient nutrient supply deriving from a lower level of marine pro-
ductivity in the Lower Palaeozoic (Tappan and Loeblich 1973).

The restricted depth range of at least four of the Caradoc communities contrasts
with the suggestion of Boucot (1975, p. 50) that Benthic Assemblages 1 to 5 occupy a
depth range of 0 to 1 50-200 m in the Silurian-Devonian, and contrasts even more
strongly with the suggested depth range of 0 to 300-500 m for the Salopina - Visbyella
communities in the Silurian (Hancock et al. 1974). It is possible that we have recog-
nized only shallow near-shore faunas, and that other communities existed elsewhere.
However, within the Welsh Basin where deeper water facies exist, benthic faunas are
sparse or absent. It follows that either we have underestimated the depth range of
Caradoc communities or the Silurian depth range has been overestimated (see Hurst
1976), or alternatively there has been a migration of benthic communities into pro-
gressively greater water depths during the Lower Palaeozoic and the number of
benthic assemblages has increased. We regard it as likely that the comparison of
Silurian and modern diversity distributions has led to an overestimate of the absolute
depth at which Silurian communities lived. In addition, we also believe that there is
evidence for migration of benthic faunas into deeper water during the Lower Palaeo-
zoic. For example. Crimes (1974) has noted a progressive increase in the diversity
of ichnofossils in deep-water environments during the Palaeozoic and has commented
on the appearance at the beginning of the Ordovician of Zoophycos in intermediate
depths, and the first significant colonization of the deeper ocean floor. The recorded
distribution of sessile epifaunal benthic faunas in the Cambrian also suggests that
most such filter-feeding associations were found only in in-shore situations, i.e. the
Skolithos facies in clastic rocks or the archaeocyathid reefs in carbonate rocks
(Copper 1974). Subsequently, the development during the Ordovician of diverse
filter-feeding communities composed of brachiopods, bryozoa, echinoderms, and
corals changed the ecology of Palaeozoic shelves. The early evolutionary history of
these complex communities is still to be determined, but it seems probable that they
initially colonized the trophically rich in-shore environments and later migrated into
more off-shore situations. The development of an advanced lophophore in the
spiriferids and pentamerids may well have been an adaptation to deeper-water
environments (see Fiirsich and Hurst 1974), and the diversification of these groups in
the late Ordovician and early Silurian could be related to the colonization of more
oxygen-deficient, trophically poor, off-shore shelf situations.

Acknowledgements. This work was undertaken during the tenure of a N.E.R.C. grant by one of us (R. K. P.)
which is gratefully acknowledged. We thank Drs. J. M. Hurst and J. P. A. Noble for critically reading pre-
liminary versions of the manuscript. Joe Lynch prepared the diagrams and Sherri T ownsend the manuscript.

REFERENCES

anderson, e. J. 1971. Environmental models for Palaeozoic communities. Lethaia, 4, 287-302.
bassett, D. a. 1963. The Welsh Palaeozoic geosyncline: a review of recent work on stratigraphy and sedi-
mentation. Pp. 35-69. In Johnson, M. r. w. and stewart, f. h. (eds.). The British Caledonides, ix+280 pp.
Oliver and Boyd, Edinburgh and London.

— Whittington, h. b. and williams, A. 1966. The stratigraphy of the Bala district, Merionethshire.
Q. Jlgeol. Soc. Lond. 122, 219-271.

berry, w. b. n. and boucot, a. j. 1967. Pelecypod-graptolite association in the Old World Silurian. Bull,
geol. Soc. Am. 78, 1515-1522.

262

PALAEONTOLOGY, VOLUME 22

boucot, a. J. 1975. Evolution and extinction rate controls. 427 pp. Elsevier, Amsterdam.
brenchley, p. j. 1969. The relationship between Caradocian volcanicity and sedimentation in North Wales.
Pp. 181-202. In wood, a. (ed.). The Pre-Cambrian and Lower Palaeozoic rocks of Wales. Report of a
symposium. . . . x+461 pp. University of Wales Press, Cardiff.

— (in press). The Caradoc rocks of the north and west Berwyn Hills, North Wales. Geol. J.
bretsky, p. w. 1969a. Evolution of Palaeozoic benthic marine invertebrate communities. Palaeogeogr.,

Palaeoclimat. Palaeoecol. 6, 45-59.

— 1969 b. Central Appalachian late Ordovician communities. Bull. geol. Soc. Am. 80, 193-212.

1970a. Late Ordovician benthic marine communities in north central New York. Bull. N.Y. St. Mus.

Sci. Serv. 44, 1-34.

1970ft. Upper Ordovician ecology of the Central Appalachians. Bull. Peabody Mus. nat. Hist. 34,

1-150.

— and bretsky, s. s. 1975. Succession and repetition of late Ordovician fossil assemblages from the
Nicolet River Valley, Quebec. Paleobiology, 1, 225-237.

— 1976. The maintenance of evolutionary equilibrium in late Ordovician benthic marine inverte-
brate faunas. Lethaia, 9, 223-233.

— and lorenz, d. m. 1970. Adaptive response to environmental stability: a unifying concept in paleo-
ecology. Pp. 522-550. In yochelson, e. l. (ed.). Proceedings of the North American paleontological con-
vention. . . . Chicago, September 5-7 , 1969. Vol. 1. [xiv]+703 pp. Allen Press, Lawrence, Kansas.

calef, c. E. and Hancock, n. j. 1974. Wenlock and Ludlow marine communities in Wales and the Welsh
Borderlands. Palaeontology, 17, 779-810.

cave, R. 1965. The Nod Glas sediments of Caradoc age in North Wales. Geol. J. 4, 279-298.
copper, p. 1974. Structure and development of early Paleozoic reefs. Proc. Second Int. Coral Reef Symp. 1,
365-386.

craig, G. Y. and jones, N. s. 1966. Marine benthos, substrate and palaeoecology. Palaeontology, 9, 30-38.
crimes, T. p. 1974. Colonization of the early ocean floor. Nature, Lond. 248, 328-330.
dean, w. T. 1958. The faunal succession in the Caradoc Series of south Shropshire. Bull. Br. Mus. nat. Hist.
(Geol.), 3, 191-231, pis. 24-26.

— 1960a. The use of shelly faunas in a comparison of the Caradoc Series in England, Wales and parts of
Scandinavia. Rep. 21st session Int. geol. Congr. Norden, 7, 82-87.

1960ft. The Ordovician trilobite faunas of south Shropshire, I. Bull. Br. Mus. nat. Hist. (Geol.). 4,

73-143, pis. 11-19.

1961. The Ordovician trilobite faunas of south Shropshire, II. Ibid. 5, 313-357, pis. 45-55.

— 1963a. The Ordovician trilobite faunas of south Shropshire, III. Ibid. 7, 215-254, pis. 37-46.

1963ft. The Ordovician trilobite faunas of south Shropshire, IV. Ibid. 9, 3-18, pis. 1, 2.

diggins, J. n. and romano, M. 1968. The Caradoc rocks around Llyn Cowlyd, North Wales. Geol. J. 6,
31-48.

driscoll, E. G. 1967. Attached epifauna— substrate relations. Limnol. Oceanogr. 12, 633-641.
fursich, f. t. 1976. Fauna-substrate relationships in the Corallian of England and Normandy. Lethaia, 9,
343-356.

— and hurst, j. m. 1974. Environmental factors determining the distribution of brachiopods. Palae-
ontology, 17, 879-900.

greig, d. c., wright, J. e., hains, B. a. and mitchell, G. h. 1968. Geology of the country around Church
Stretton, Craven Arms, Wenlock Edge and Brown Clee (Explanation of One-inch Geological Sheet 166).
Mem. geol. Surv. U.K. i-xiv, 1-380.

HANCOCK, n. j., hurst, j. m. and FURSICH, f. t. 1974. The depth inhabited by Silurian brachiopod communi-
ties. Jl geol. Soc. Lond. 130, 151-156.

Harrington, H. j. 1959. General description of trilobita. Pp. 038-0117. In moore, r. c. (ed.). Treatise on
invertebrate paleontology. Part O, Arthropoda. Geological Society of America and University of Kansas
Press.

hurst, J. M. 1975. Wenlock carbonate, level bottom, brachiopod-dominated communities from Wales and
the Welsh Borderland. Palaeogeogr. Palaeoclimat. Palaeoecol. 17, 227-255.

— 1976. The depths inhabited by Silurian brachiopod communities: comment. Geology, 4, 709, 710.
ingham, J. k. 1966. The Ordovician rocks in the Cautley and Dent districts of Westmorland and Yorkshire.

Proc. Yorks, geol. Soc. 35, 455-505, pis. 25-28.

PICKERILL AND BRENCHLEY: ORDOVICIAN MARINE COMMUNITIES

263

Johnson, R. G. 1960. Models and methods for the mode and formation of fossil assemblages. Bull. geol. Soc.
Am. 71, 1075-1086.

— 1965. Pelecypod death assemblages in Tomales Bay, California. J. Paleont. 39, 80-85.

— 1971. Animal-sediment relations in shallow water benthic communities. Mar. Geol. 11, 93-104.

— 1972. Conceptual models of benthic marine communities. Pp. 148-159. In schopf, t. j. m. (ed.).
Models in paleobiology. Freeman-Cooper, San Francisco.

jones, N. s. 1950. Marine bottom communities. Biol. Rev. 25, 283-313.

kauffman, E. G. and SCOTT, R. w. 1976 Basic concepts of community ecology and paleoecology. Pp. 1-28.
In scott, R. w. and west, r. r. (eds.). Structure and classification of paleocommunities. Dowden, Hutchin-
son and Ross.

king, w. b. r. 1923. The Upper Ordovician rocks of the south-west Berwyns. Q. Jl geol. Soc. Lond. 84,
487-507.

lamont, A. 1935. The Drummuck Group, Girvan; a stratigraphical revision with descriptions of fossils
from the lower part of the Group. Trans, geol. Soc. Glasg. 19, 288-332, pis. 7-9.

Lawrence, D. R. 1968. Taphonomy and information losses in fossil communities. Bull. geol. Soc. Am. 79,
1315-1330.

lindroth, a. 1935. Die associationen der marinen Weichboden. Zool. Bidr. Upps. 15, 331-366.
macarthur, R. H. 1972. Geographical ecology: patterns in the distribution of species. 269 pp. Harper and
Row, New York.

Macdonald, K. B. 1975. Quantitative community analysis: recurrent group and cluster techniques applied
to the fauna of the Upper Devonian Sonyea Group, New York. J. Geol. 82, 473-499.
macginitie, G. e. 1939. Littoral marine communities. Am. Midi. Nat. 21, 28-55

macgregor, a. r. 1961. Upper Llandeilo brachiopods from the Berwyn Hills, North Wales. Palaeontology,
4, 177-209, pis. 19-23.

marshall, N. B. 1954. Aspects of deep sea biology. 380 pp. Hutchinson, London.
muller, c. H. 1958. Science and philosophy of the community concept. Am. Scient. 46, 294-308.
petersen, c. G. J. 191 1. Valuation of the sea. I. Animal life of the sea bottom, its food and quantity. Rep.
Dan. biol. Stn, 20, 1-81.

— 1913. Valuation of the sea. II. The animal communities of the sea bottom and their importance for
marine zoogeography. Ibid. 21, 1-44.

pickerill, r. k. 1973. Lingulasma tenuigranulata — palaeoecology of a large Ordovician linguloid that lived
within a strophomenid-trilobite community. Palaeogeogr. Palaeoclimat. Palaeoecol. 13, 143-156.

— 1974. Geology of the south Berwyn Hills, North Wales, with particular reference to Upper Ordovician
marine benthic communities. Ph.D. thesis, Univ. of Liverpool.

— 1975. Application of ichnology to the study of ancient marine benthic community palaeoecology.
A discussion and case example. Marit. Sediments, 1 1, 49-52.

— 1976. Vermiforichnus borings from the Ordovician of central Wales. Geol. Mag. 113, 159-164.

— 1977. Trace fossils from the Upper Ordovician (Caradoc) of the Berwyn Hills, central Wales. Geol. J.
12, 1-16.

ramsbottom, w. H. c. 1961. The British Ordovician Crinoidea. Palaeontogr. Soc. [Monogr.], 1-37, pis. 1-8.
Richards, R. p. 1972. Autecology of Richmondian brachiopods (late Ordovician of Indiana and Ohio).
J. Paleont. 46, 386-405.

sheehan, p. m. 1975. Late Ordovician brachiopods from Belgium. Abstr. Progm geol. Soc. Am. 7,
1267.

— 1977. Ordovician and Silurian brachiopods from graptolitic shales and related deep-water argillaceous
rocks. Lethaia, 10, 201-203.

slobodkin, l. B. and Sanders, h. l. 1969. On the contribution of environmental predictability to species
diversity. In woodwell, g. m. and smith, h. h. (eds.). Diversity and stability in ecological systems.
Brookhaven Symp. Biol. 22, 82-95.

Stanton, r. j., jr. and evans, I. 1972. Community structure and sampling requirements in paleoecology.
J. Paleont. 46, 845-858.

tappan, H. and loeblich, A. R. 1973. Smaller protistan evidence and explanation of the Permian-Triassic
crisis. In LOGAN, A. and hills, l. v. (eds.). The Permian and Triassic systems and their mutual boundary.
Mem. Can. Soc. Petrol. Geol. 2, 465-480.

thayer, c. w. 1974. Marine paleoecology in the Upper Devonian of New York. Lethaia, 7, 121-155.

264

PALAEONTOLOGY, VOLUME 22

walker, K. r. and alberstadt, L. p. 1975. Ecological succession as an aspect of structure in fossil com-
munities. Paleobiology, 1, 238-257.

— and bambach, r. k. 1971. The significance of fossil assemblages from fine grained sediments— time
averaged communities. Abstr. Progm geol. Soc. Am. 3, 783, 784.

— and laporte, L. F. 1970. Congruent fossil communities from the Ordovician and Devonian of New
York. J. Paleont. 44, 928-944.

watkins, r. and boucot, a. j. 1975. Evolution of Silurian brachiopod communities along the southeastern
coast of Acadia. Bull. geol. Soc. Am. 86, 243-254.

Whittington, H. b. 1962-1968. The Ordovician trilobites of the Bala area, Merioneth, Part I. Palaeontogr.
Soc. [. Monogr .] 116(1962), 1-32, pis. 1-8; Part II, 118(1965), 33-62, pis. 9-18; Part III, 120(1966), 63-92,
pis. 19-28; Part IV, 122 (1968), 93-138, pis. 29-32.

williams, a. 1953. The geology of the Llandeilo area Carmarthenshire. Q. Jlgeol. Soc. Lond. 108, Ml-201.

— 1963. The Caradocian brachiopod faunas of the Bala district, Merionethshire. Bull. Br. Mus. nat.
Hist. (Geol.), 8, 327-471, pis. 1-16.

— 1973. Distribution of brachiopod assemblages in relation to Ordovician palaeogeography. In hughes,
n. F. (ed.). Organisms and continents through time. Spec. Pap. Palaeont. 12, 241-269.

1974. Ordovician brachiopoda from the Shelve district, Shropshire. Bull. Br. Mus. nat. Hist. (Geol.),

Supp. 11, 1-163, pis. 1-28.

— 1976. Plate tectonics and biofacies evolution as factors in Ordovician correlation. Pp. 29-65. In
bassett, M. G. (ed.). The Ordovician System: proceedings of a Palaeontological Association symposium,
Birmingham, September 1974. 696 pp. University of Wales Press and National Museum of Wales,
Cardiff.

wobber, f. J. 1968. A faunal analysis of the Lias (Lower Jurassic) of South Wales (Great Britain). Palaeo-
geogr. Palaeoclimat. Palaeoecol. 5, 269-308.

worsley, d. 1971. Faunal anticipation in the lower Llandovery of the Oslo region, Norway. Norsk, geol.
Tidsskr. 51, 161-167.

wright, a. D. 1964. The fauna of the Portrane Limestone, II. Bull. Br. Mus. nat. Hist. (Geol.), 9, 157-256,
pis. 1-11.

ziegler, a. m. 1965. Silurian marine communities and their environmental significance. Nature, Lond. 207,
270-272.

— cocks, L. R. M. and bambach, r. k. 1968. The composition and structure of Lower Silurian marine
communities. Lethaia, 1, 1-27.

R. K. PICKERILL
Department of Geology
University of New Brunswick
P.O. Box 4400
Fredericton, N.B.

E3B 5A3
Canada

Manuscript received 16 May 1977
Revised manuscript received 8 February 1978

p. J. BRENCHLEY
Department of Geology
University of Liverpool
Liverpool L69 3BX

THE JAWS AND RADULA OF THE JURASSIC
AMMONITE DACTYLIOCERAS

by ULRICH LEHMANN

Abstract. Anaptychi ( = lower jaws) and upper jaws of two specimens of Dactylioceras from the Yorkshire coast,
found within their living chambers, are described and analysed. The corresponding lower and upper jaws are about
equal in length and much shorter than the height of the living chambers. One of the ammonites contained numerous
radular teeth between the lower and upper jaw. In addition, isolated lower and upper jaws attributed to Dactylioceras
are reported from Reichenschwand (near Niirnberg, Federal Republic of Germany).

When investigating specimens of Dactylioceras from the Yorkshire coast, several
anaptychi were discovered within their living chambers. Two of them were sawed
out of the surrounding matrix and subsequently ground down. Both were found to be
associated with upper jaws. The sectioned anaptychi are here treated as Nos. 1 and 2.

DESCRIPTION OF MATERIAL

Anaptychus No. 1 was located within the living chamber of an adult specimen of
Dactylioceras ( Orthodactylites ) tenuicostatum (Young and Bird) from bed 22
(Howarth 1973) at Port Mulgrave, lying about 90° behind the aperture. The diameter
of the ammonite conch is 82 mm, and the apertural rim of the shell is thickened. The
innermost whorls are pyritized ; the septa of the outer part of the phragmocone are
destroyed. The living chamber seems to be complete, but there may have been water
movement within it prior to entombment, caused possibly by implosion of the septa
and resulting in the anaptychus being oriented with its longitudinal axis at an angle
of about 60-70° to the living chamber.

The anaptychus was found after the ammonite had been cut along its median plane.
It was then ground and photographs and drawings made at intervals of about
0-05 mm, resulting in seventy-seven sections. The drawings were transferred to thick
transparent plastic plates, and from them text-fig. 1 was drawn. This shows the rela-
tive position of the anaptychus with the upper jaw between its wings. The upper jaw
is pressed rather closely against the right flank of the anaptychus, causing some diffi-
culty with the reconstruction. There is no indication of a calcareous coating of the
flanks of the anaptychus. The front part is quite straight; an inner lamella bends
backwards, in the median part its length attains one-quarter of that of the outer
lamella. In most previous illustrations of anaptychi the inner lamella is not
represented.

The upper jaw is similar in shape to upper jaws of recent cephalopods ; it consists of
two wings joining anteriorly to form a wedge and then bending outwards and back-
wards as an outer lamella, the hood. There is no real beak as in recent forms. Between
the anaptychus and the upper jaw remains of the radula were found.

[Palaeontology, Vol. 22, Part 1, pp. 265-271, pi. 27.]

266

PALAEONTOLOGY, VOLUME 22

text-fig. 1 . Reconstruction of the lower jaw (= anaptychus) and of the upper jaw of a specimen of Dactylio-
ceras ( Orthodactylites ) tenuicostatum (Young and Bird) (= specimen No. 1) as found within the living
chamber. The thick lines indicate the position of the first ‘section’, a small part of the left side of the structure
beyond the lines being lost.

Anaptychus No. 2 was found in the living chamber of an adult specimen of Dactylio-
ceras ( Orthodactylites ) semicelatum (Simpson) from bed 28 probably at Port Mulgrave
(obliteration of locality sign during preparation leaves some doubt about locality).
It lay about 90° behind the aperture of the ammonite.

The diameter measures 82 mm, and the apertural rim of the shell is thickened. The
state of preservation is very good, all septa being preserved, and the last ones are
more closely spaced than the preceding ones. The inner half of the living chamber is
filled with clear calcite crystals, the outer half contains sediment and the anaptychus
embedded in it. Its orientation is ‘normal’ : front towards the aperture, base towards
the exterior (ventrally) (Trauth 1927; Morton 1975).

As with No. 1, anaptychus No. 2 was found after the ammonite had been cut in the
median plane. It was subsequently treated the same way as No. 1, but at intervals of
about 0T mm, resulting in forty-two sections from which text-fig. 2 was drawn.
Nothing is now left of the two anaptychi; the photographs and drawings are kept in
the Hamburg collections.

The position of the upper jaw within the anaptychus was easier to analyse because
they were not as closely pressed against each other. No radular teeth were found
between the jaws.

The two text-figures are not identical, but they give the exact form and relative

LEHMANN: JAWS OF DACTYLIOCERAS

267

E

E

text-fig. 2. Reconstruction of the lower jaw (= anaptychus) and of the upper jaw of a specimen of Dactylio-
ceras ( Orthodactylites ) semicelatum (Simpson) (= specimen No. 2) as found within the living chamber.

position of each pair of structures as they were found. Differences between the figures
are due to the kind of preservation and to different angles of view, so they should not
be overemphasized. In text -fig. 1 the thick lines indicate the first section; the missing
parts of the left side of anaptychus and upper jaw fell victims to the saw before they
were detected.

Measurements

Diameter of

Length of

Approximate

Length of

Height of

ammonite

anaptychus

height of
anaptychus

upper jaw

upper jaw

Number 1 82 mm

8-2

5-5

81

3-7

Number 2 82 mm

8-6

5-9

8-7

3-8

Radula. In sections 38 to 62 of anaptychus No. 1, a considerable number of radular
teeth were seen to be preserved between anaptychus and upper jaw. They are clearly
visible on the photographs (Plate 27, fig. 1, 5). All are simple, thin-walled hollow
cones. Although the individual teeth are well preserved, so many of them are detached
that it is not possible to reconstruct the complete radular ribbon.

Three types of teeth are distinguishable (text-fig. 3). On each side of the radula one
long marginal tooth (1-2 mm in length); the next towards the centre are half as long;

268

PALAEONTOLOGY, VOLUME 22

three short teeth, 0-2-0- 3 mm in length, of which the central or rhachidian tooth is the
largest, seem to form the centre of each transverse row of teeth. However, the short
central teeth are so much disarranged that their exact number per transverse row
cannot be given with certainty, whereas the arrangement of the lateral teeth is certain.
Text-fig. 3, therefore, only indicates the most probable arrangement. If it is correct it
corresponds to the arrangement of the transverse rows in the other known ammonite
radulae, all of which have seven teeth per transverse row (Lehmann 1967, 1971, 1976).

text-fig. 3. Reconstruction of a transverse row of the
radula as preserved between anaptychus No. 1 and the
upper jaw. The number of the three innermost teeth is
most probable, but not certain, owing to advanced
disintegration.

DISCUSSION

Previous studies by the author (Lehmann 1970, 1971, 1976) have established that
anaptychi and aptychi of ammonites are their lower jaws, in contrast to their former
interpretation as opercula (Trauth 1927, 1935; Arkell 1957; Schindewolf 1958).

Up to now no aptychi or anaptychi of Dactylioceras have been known (Schmidt-
Effing 1973, p. 29). This had led to the assumption that Dactylioceras may possibly not
have possessed any jaws at all, and therefore may have been a plankton feeder. How-
ever, jaws of Dactylioceras are not as rare as it may seem. For example, a limestone
slab from Ziegelei Reichenschwand, east of Niirnberg, W. Germany (shown on
Plate 27, figs. 2-4) has the surface covered with several upper and lower jaws of the
same shape as the ones described from the coast of Yorkshire. The strata exposed at
Reichenschwand have so far produced ammonites of the genera Harpoceras and
Dactylioceras, but only jaws of the type shown on Plate 27, figs. 2-4. The jaws of
Harpoceras being well known it is highly probable that the jaws belong to Dactylio-
ceras, particularly as they resemble the jaws from Yorkshire. Upper and lower jaws

EXPLANATION OF PLATE 27

Fig. 1. Part of the living chamber of Dactylioceras ( Orthodactylites ) tenuicostatum (Young and Bird) from
bed 22, Port Mulgrave, Yorkshire, with the jaw apparatus and sections of radula teeth (white arrow) as
seen in section No. 41, approx, x 15.

Fig. 2. Surface of a slab of limestone from Ziegelei Reichenschwand, near Niirnberg, Lias epsilon (lower
Toarcian), with isolated upper and lower jaws presumably of Dactylioceras sp., x 3.

Fig. 3. Lower jaw (=anaptychus) presumably of Dactylioceras sp., x 14. Same locality as Fig. 2.

Fig. 4. Upper jaw presumably of Dactylioceras sp., x 17. Same locality as Fig. 2.

Fig. 5. Marginal radular tooth of same specimen as Fig. 1. Section 56, x 25.

PLATE 27

LEHMANN, Jaws and radula of Dactylioceras

270

PALAEONTOLOGY, VOLUME 22

are detached but they have not been transported very far. Their size is less than that of
the jaws from Yorkshire. Due to compaction they are pressed into a plane and this has
caused rupture of the outer margins of the lower jaw.

The jaws of recent cephalopods and of ammonoids are very similar : both have no
articulation, their relative movements being effected only by muscles and both owe
their cutting ability to a shearing effect (Kaiser and Lehmann 1971). This has been
shown to apply for Nautilus jaws as well, in spite of the calcification of their mandibles
which even produces prominent denticles (Saunders et al. 1978). Even thin and seem-
ingly weak jaws of recent cephalopods may prove to be surprisingly powerful biting
organs. ‘ Amphitretus , an octopod, has flat, delicate jaws, which G. C. Robson con-
sidered incapable of biting, but a specimen was collected taking the bait off fish hooks’
(A. Bidder, in litteris, Nov. 1977).

The fact that the radulae of recent cephalopods and of ammonoids are rather
narrow and uniform, in contrast to radulae of prosobranch gastropods, seems to
indicate rather similar use for all of them, which means more or less omnivorous or
carnivorous diet. Plankton feeders would not need radulae either to seize or to
swallow their food. That does not mean that ammonoids did not occasionally eat
plankton, but that they did not depend on it.

Plankton and other soft-bodied organisms are not found in the stomachs of
ammonoids, although they may very well have composed part of their diet. The
presence of ostracods, foraminifers, crinoids, and other ammonites, etc. (Lehmann
1971, 1975, 1976) points to the diet of a scavenger and carnivore much like that of
many recent octopods. ‘However, ammonites are a diverse and variable group, and a
single feeding strategy seems unlikely’ (Kennedy and Cobban 1976).

Acknowledgements. Thanks are due to Eva Mehrling for preparing and sectioning the jaws and for docu-
mentation of the sections, to Doris Lewandowski for drawing text-figs. 1-2, and H.-J. Lierl for assistance in
analysing the radula and for providing the slab from Reichenschwand from his private collection.

REFERENCES

arkell, w. j. 1957. Aptychi. In moore, r. c. (ed.). Treatise on invertebrate paleontology. Part L, Mollusca 4,
L437-L441.

howarth, M. K. 1973. The stratigraphy and ammonite fauna of the upper Liassic Grey Shales of the York-
shire coast. Bull. Brit. Mus. (Nat. Hist.), Geology, 24, (4), 237-277.
kaiser, p. and lehmann, u. 1971. Vergleichende Studien zur Evolution des Kieferapparates rezenter und
fossiler Cephalopoden. Paldont. Z. 45, 18-32.

Kennedy, w. j. and cobban, w. a. 1976. Aspects of ammonite biology, biogeography, and biostratigraphy.
Spec. Pap. Palaeont. 17, 1-94.

lehmann, u. 1967. Ammoniten mit Kieferapparat und Radula aus Lias-Geschieben. Paldont. Z. 41, 38-45.

— 1970. Lias-Anaptychen als Kieferelemente (Ammonoidea). Paldont. Z. 44, 25-31.

— 1971. Jaws, radula, and crop of Arnioceras (Ammonoidea). Palaeontology, 14, 238-341.

— 1976. Ammoniten, ihr Leben und ihre Umwelt. Enke-Verlag, Stuttgart 1976, 171 pp.

morton, N. 1975. The position of the aptychus in some Jurassic ammonites. N. Jb. Geo'l. Paldont. Mh.
Jg. 1975,409-411.

saunders, w. B., spinosa, c., teichert, c. and banks, R. c. 1978. The jaw apparatus of recent Nautilus and
its palaeontological implications. Palaeontology, 21, pp. 129-141.
schindewolf, o. H. 1958. Uber Aptychen (Ammonoidea). Palaeontogr . A, 111, 1-46.

LEHMANN: JAWS OF DACTYLIOCERAS

271

schmidt-effing, r. 1972. Die Dactylioceratidae, eine Ammoniten-Familie des unteren Jura. Munster.
Forsch. Geol. Palaont. 25/26, 255 pp.

trauth, F. 1927. Aptychenstudien. I. liber die Aptychen im Allgemeinen. Ann. naturhist. Mus. Wien, 41,
171-259.

1935. Anaptychi und anaptychusahnliche Aptychi der Kreide. N. Jb. Miner., Beil.-Bd. (B) 74,

448-468.

U. LEHMANN

Typescript received 23 January 1978
Revised typescript received 17 April 1978

Geolog.-Palaontologisches Institut
Universitat Hamburg
Bundesstrasse 55
2 Hamburg 13
W. Germany

THE PALAEONTOLOGICAL ASSOCIATION

The Association was founded in 1957 to further the study of palaeontology. It holds meetings
and demonstrations as well as publishing Palaeontology and Special Papers in Palaeontology.
Membership is open to individuals and to institutions on payment of the appropriate annual
subscription. Rates for 1978 are:

Institutional membership .... £25-00 (U.S. $50.00)

Ordinary membership .... £12-00 (U.S. $24.00)

Student membership . . . £7-50 (U.S. $12.00)

There is no admission fee. Correspondence concerned with Institutional Membership should be
addressed to Dr. C. H. C. Brunton, Department of Palaeontology, British Museum (Natural
History), Cromwell Road, London SW7 5BD, England. 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. Sub-
scriptions cover one calendar year and are due each January ; they should be sent to the Membership
Treasurer, Dr. J. C. W. Cope, Department of Geology, University College, Swansea SA2 8PP,
Wales, U.K.

PALAEONTOLOGY

All members who join for 1978 will receive Volume 21, Parts 1-4. All back numbers are still in
print and may be ordered from B. H. Blackwell, Broad Street, Oxford OX1 3BQ, England, at
£12-00 per part (post free). A complete set, Volumes 1-20, consists of 79 parts and costs £948.

SPECIAL PAPERS IN PALAEONTOLOGY

The subscription rate for 1978 is £15 (U.S. $30.00) for Institutional Members and £7-50 (U.S.
$1 5.00) for Ordinary and Student Members. Individual subscriptions should be placed through the
Membership Treasurer, Dr. J. C. W. Cope, Department of Geology, University College, Swansea
SA2 8PP, Wales, U.K. Ordinary and Student members only may obtain individual Special Papers
from Dr. Cope at reduced rates. Non-members may obtain them at the stated prices from B. H.
Blackwell, Broad Street, Oxford OX1 3BQ, England.

COUNCIL 1978-1979

President : Professor H. B. Whittington, Department of Geology, Sedgwick Museum, Cambridge CB2 3EQ
Vice-Presidents : Dr. L. R. M. Cocks, Department of Palaeontology, British Museum (Natural History),
Cromwell Road, London SW7 5BD

Dr. E. P. F. Rose, Department of Geology, Bedford College, Regent’s Park, London
NW1 4NS

Treasurer: Mr. R. P. Tripp, High Wood, West Kingsdown, Sevenoaks, Kent TNI 5 6BN
Membership Treasurer : Dr. J. C. W. Cope, Department of Geology, University College, Swansea SA2 8PP
Secretary : Dr. C. T. Scrutton, Department of Geology, The University, Newcastle upon Tyne NE1 7RU

Editors

Professor C. B. Cox, Department of Zoology, King’s College, Strand, London WC2R 2LS
Dr. M. G. Bassett, Department of Geology, National Museum of Wales, Cardiff CF1 3NP
Dr. K. C. Allen, Department of Botany, Bristol University, Bristol BS8 1UG
Dr. R. A. Fortey, Department of Palaeontology, British Museum (Natural History),
Cromwell Road, London SW7 5BD

Other Members of Council

Dr. R. J. Aldridge, Nottingham
Dr. M. C. Boulter, London
Dr. M. D. Brasier, Hull
Dr. P. J. Brenchley, Liverpool
Dr. C. H. C. Brunton, London
Dr. S. Conway Morris, Cambridge
Dr. G. E. Farrow, Glasgow

Dr. R. A. Fortey, London
Dr. M. B. Hart, Plymouth
Dr. P. M. Kier. Washington
Dr. S. C. Matthews, Bristol
Dr. I. E. Penn, London
Dr. G. D. Sevastopulo, Dublin

Overseas Representatives

Australia: Professor B. D. Webby, Department of Geology, Sydney University, Sydney, N.S.W., 2006
Canada: Dr. B. S. Norford, Institute of Sedimentary and Petroleum Geology, 3303-33rd Street NW., Calgary, Alberta
India : Professor M. R. Sahni, 98 Mahatma Gandhi Marg, Lucknow (U.P.), India
New Zealand: Dr. G. R. Stevens, New Zealand Geological Survey, P.O. Box 30368, Lower Hutt
West Indies and Central America : Mr. John B. Saunders, Geological Laboratory, Texaco Trinidad, Inc., Pointe-a-Pierre,
Trinidad, West Indies

Western U.S. A.: Professor J. Wyatt Durham,- Department of Paleontology, University of California, Berkeley 4,
California

Eastern U.S. A. : Professor J. W. Wells. Department of Geology, Cornell University, Ithaca, New York
South America: Dr. O. A. Reig, Departmento de Ecologia, Universidad Simon Bolivar, Caracas 108, Venezuela

Palaeontology

VOLUME 22 ■ PART 1

CONTENTS

The visual system of trilobites

E. N. K. CLARKSON 1

A late Silurian flora from the Lower Old Red Sandstone of
south-west Dyfed

D. EDWARDS 23

Trilobites from the Coniston Limestone Group (Ashgill
Series) of the Lake District, England

K. J. MCNAMARA 53

New acrotretacean brachiopods from the Palaeozoic of
Britain and Austria

L. R. M. cocks 93

The physiological differences between articulate brachiopods
and filter-feeding bivalves as a factor in the evolution of
marine level-bottom communities

H. M. STEELE-PETROVIC 101

A Middle Jurassic mammal bed from Oxfordshire

E. F. FREEMAN 135

Appendages of the arthropod Aglaspis spinifer (Upper
Cambrian, Wisconsin) and their significance

D. E. G. BRIGGS, D. L. BRUTON and H. B. WHITTINGTON 167

A new foraminifer from the Middle Eocene of Papua
New Guinea

C. G. ADAMS and D. J. BELFORD 181

The Hampen Marly and White Limestone formations:
Florida-type carbonate lagoons in the Jurassic of central
England

T. J. PALMER 189

Caradoc marine benthic communities of the south Berwyn
Hills, North Wales

R. K. PICKERILL and P. J. BRENCHLEY 229

The jaws and radula of the Jurassic ammonite Dactylioceras

U. LEHMANN 265

Printed in Great Britain at the University Press, Oxford
by Eric Buckley, Printer to the University

Palaeontology

VOLUME 22 • PART 2 MAY 1979

Published by

The Palaeontological Association London
Price £1 2

THE PALAEONTOLOGICAL ASSOCIATION

The Association publishes Palaeontology and Special Papers in Palaeontology. Details of membership and
subscription rates may be found inside the back cover.

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 on all aspects of palaeontology and stratigraphical palaeontology are invited.
They should conform in style to those already published in this journal, and should be sent to Dr. R. A.
Fortey, Palaeontological Association, Department of Palaeontology, British Museum (Natural History),
Cromwell Road, London SW7 5BD, England, who will supply detailed instructions for authors on request
(these were published in Palaeontology 1970, 20, pp. 921-929).

Special Papers in Palaeontology is a series of substantial separate works; the following are available (post
free).

1. (for 1967): Miospores in the Coal Seams of the Carboniferous of Great Britain, by A. H. v. smith and
M. A. butterworth. 324 pp., 72 text-figs., 27 plates. Price £8 (U.S. $16).

2. (for 1968) : Evolution of the Shell Structure of Articulate Brachiopods, by A. williams. 55 pp., 27 text-
figs., 24 plates. Price £5 (U.S. $10).

3. (for 1968) : Upper Maestrichtian Radiolaria of California, by Helen p. foreman. 82 pp., 8 plates.
Price £3 (U.S. $6).

4. (for 1969): Lower Turonian Ammonites from Israel, by r. freund and m. raab. 83 pp., 15 text-figs.,
10 plates. Price £3 (U.S. $6).

5. (for 1969): Chitinozoa from the Ordovician Viola and Fernvale Limestones of the Arbuckle Moun-
tains, Oklahoma, by w. a. m. jenkins. 44 pp., 10 text-figs., 9 plates. Price £2 (U.S. $4).

6. (for 1969): Ammonoidea from the Mata Series (Santonian-Maastrichtian) of New Zealand, by
R. a. Henderson. 82 pp., 13 text-figs., 15 plates. Price £3 (U.S. $6).

7. (for 1970): Shell Structure of the Craniacea and other Calcareous Inarticulate Brachiopoda, by
a. williams and A. D. wright. 51 pp., 17 text-figs., 15 plates. Price £1-50 (U.S. $3).

8. (for 1970): Cenomanian Ammonites from Southern England, by w. j. Kennedy. 272 pp., 5 tables,
64 plates. Price £8 (U.S. $16).

9. (for 1971): Fish from the Freshwater Lower Cretaceous of Victoria, Australia, with Comments on
the Palaeo-environment, by m. waldman. 130 pp., 37 text-figs., 18 plates. Price £5 (U.S. $10).

10. (for 1971): Upper Cretaceous Ostracoda from the Carnarvon Basin, Western Australia, by R. H. bate.
148 pp., 43 text-figs., 27 plates. Price £5 (U.S. $10).

11. (for 1972): Stromatolites and the Biostratigraphy of the Australian Precambrian and Cambrian, by
m. R. Walter. 268 pp., 55 text-figs., 34 plates. Price £10 (U.S. $20).

12. (for 1973): Organisms and Continents through Time. A Symposium of 23 papers edited by n. f. hughes.
340 pp., 132 text-figs. Price £10 (U.S. $20) (published with the Systematics Association).

13. (for 1974) : Graptolite studies in honour of O. M. B. Bulman. Edited by R. b. rickards, d. e. jackson,
and c. p. hughes. 261 pp., 26 plates. Price £10 (U.S. $20).

14. (for 1974): Palaeogene Foraminiferida and Palaeoecology, Hampshire and Paris Basins and the
English Channel, by j. w. Murray and c. a. wright. 171 pp., 45 text-figs., 20 plates. Price £8
(U.S. $16).

15. (for 1975): Lower and Middle Devonian Conodonts from the Broken River Embayment, North
Queensland, Australia, by p. g. telford. 100 pp., 9 text-figs., 16 plates. Price £5-50 (U.S. $11).

16. (for 1975): The Ostracod Fauna from the Santonian Chalk (Upper Cretaceous) of Gingin, Western
Australia, by J. w. neale. 131 pp., 40 text-figs., 22 plates. Price £6-50 (U.S. $13).

17. (for 1976): Aspects of Ammonite Biology, Biogeography, and Biostratigraphy, by w. J. Kennedy
and w. a. cobban. 94 pp., 24 text-figs., 11 plates. Price £6 (U.S. $12).

18. (for 1976): Ostracoderm Faunas of the Delorme and Associated Siluro-Devonian Formations,
North West Territories, Canada, by D. L. dineley and E. J. loeffler. 218 pp., 78 text-figs., 33 plates.
Price £20 (U.S. $40).

19. (for 1977): The Palynology of Early Tertiary Sediments, Ninetyeast Ridge, Indian Ocean, by E. m. kemp
and w. K. Harris. 74 pp., 2 text-figs., 8 plates. Price £7 (U.S. $14).

20. (for 1977): Fossil Priapulid Worms, by s. c. morris. 159 pp., 99 text-figs., 30 plates. Price £16
(U.S. $32).

21. (for 1978): Devonian Ammonoids from the Appalachians and their bearing on International Zonation
and Correlation, by M. R. house. Price £12 (U.S. $24).

© The Palaeontological Association, 1979

Cover. The lobopod animal Aysheaia pedunculata Walcott, 1911, from the Burgess Shale, Middle
Cambrian, British Columbia, x4. Specimen in National Museum of Natural History, Smithsonian
Institution, Washington D.C., U.S.A. Photograph by H. B. Whittington, in ultraviolet radiation reflected
from surface (see Phil. Trans. R. Soc., London).

A GIANT MYRIAPOD TRAIL FROM THE
NAMURIAN OF ARRAN, SCOTLAND

by D. E. G. BRIGGS, W. D. I. ROLFE, and J. BRANNAN

Abstract. A large trace fossil in the Limestone Coal Group of Arran is preserved in a deltaic, channel-fill sandstone
from a cyclic sequence including coals. The trail, Diplichnites cuithensis ichnosp. nov., consists of two parallel series
of closely spaced imprints, and is attributed to the giant Carboniferous myriapod Arthropleura, making this the earliest
evidence for the genus. Analysis of the trail suggests the individual responsible was c. 1 m long, and had twenty-three
pairs of appendages. Knowledge of modern myriapod gaits has been used to extrapolate a theoretical trail for Arthro-
pleura, which compares well with the fossil trail. The trail suggests that a gait pattern of forestroke : backstroke
of 5-5: 4-5 was used in walking across the sand substrate. This contrasts with the previous estimated gait of
3:7 that Arthropleura might have used in pushing through coal-forest litter. Such a range of gaits is well within
that recorded for individual Recent species of millipede. Diplichnites is emended to exclude most trilobite loco-
motion trails.

‘As Professor Phillips has remarked, every geologist who visits Arran is tempted to
write about it, and finds something new to add to what has been already put on
record’ (Bryce 1859, p. 61). It is therefore remarkable that the present large trail
(PI. 28), 6-25 m long and 0-36 m wide, has escaped previous attention, lying as it
does on one of the popular geological excursion routes (south-east of locality 12 of
Macgregor 1965, pp. 128-129) beside the footpath from Laggan to the Cock of Arran.
The oversight is due to the trail’s near invisibility on the dipping bedding plane,
except under the raking light of early morning and evening. The trail was drawn to
the attention of one author (D. E. G. B.) on a Cambridge University excursion
in 1975. It will form a useful new locality on future itineraries (Macgregor 1965,
p. 126, fig. 13), the usual conservation courtesies being observed. An exhibit of
a replica of the trail is mounted in the Arran Nature Centre, near Brodick. Speci-
mens referred to are held by the Hunterian Museum, University of Glasgow (pre-
fixed HM).

GEOLOGICAL SETTING

Occurrence

The trail runs (vector 087°) across the south face of a small sandstone quarry
(NR 9722 5112) which opens off the south shore of the salt pans harbour between
Cock Farm and Laggan, 1-75 km south-east of the Cock of Arran. The bedding plane
dips 32° at 358°. The quarry, which is flooded by the sea at high tide, was opened no
later than the eighteenth century : it is quite likely that quarried blocks of the sand-
stone bearing the missing part of the trail are built into the walls of the ruined malt
kiln, above the south-east corner of the quarry (at NR 9720 5111). A second trail
(PI. 29, fig. 3) 1-9 m long (lineation 167°-347°) is located on the slab, 2-32 m at 181°
from a point on the main trail 5-5 m from the west face of the quarry.

[Palaeontology, Vol. 22, Part 2, 1979, pp. 273-291, pis. 28-30.]

274

PALAEONTOLOGY, VOLUME 22

Stratigraphy

The trail occurs on the bedding plane forming the south face of the quarry, c. 60 mm
below the top of a 6 m thick white sandstone (text-fig. 1). This is sandstone unit 136
on the large-scale unpublished maps of the area by Dr. Grace Page, Bedford College,
London University. The unit formed the roof of the coal seams (Ramsay 1841, p. 31)
which ceased to be worked in the eighteenth century (Gunn et al. 1903, p. 146; Gunn
in Tyrrell 1928, p. 268). The top of this sandstone is 42-6 m (dyke omitted) below the
base of the limestone taken as the Index Limestone (Gunn et al. 1903, p. 49 ; Gunn and
Lee in Tyrrell 1928, p. 60; Macgregor 1965, p. 32) and 92-7 m above the top of the
supposed Hosie Limestone equivalent (Gunn et al. 1903, p. 48; Gunn and Lee in
Tyrrell, 1928, p. 58). This places the trail in the Limestone Coal Group, of Namurian
Series, Pendleian Stage (Ex— Ramsbottom 1978). The Limestone Coal Group section
around Laggan was measured (text-fig. 1) and found to total 135-3 m. This is thicker
than the whole of the Carboniferous Limestone Series of the Corrie section ( c . 119 m
according to Macgregor 1965, pp. 104-105) and emphasizes the southerly attenua-
tion (George 1960, pp. 74-76) of the Limestone Coal Group (estimated to be 83 m
thick at Corrie from George et al. 1976, p. 48, fig. 13 : 2).

Sedimentary environment

The Limestone Coal Group (text-fig. 1) is predominantly composed of clastic
sediments and is bounded above and below by limestones (Index and Hosie respec-
tively) representing marine transgressions. The succession appears to have been
deposited in a proximal deltaic environment (Frazier and Ozanik 1969), similar to
that of the Midland Valley Limestone Coal Group (Moore 1959; Francis 1965,
p. 352; Read and Dean 1976 and references therein). The trail is preserved on the
surface of a bedding plane of heterogeneous sandstone. At the eastern end the sand-
stone is thicker, coarser, and purer (HM TS. 17841— quartz arenite, McBride 1963)
than elsewhere and it grades westwards into a finer sandstone (HM TS. 17843 —
lithic arkose) containing discontinuous layers of shale.

Sections of the western slab (PI. 28 ; PI. 29, fig. 2) reveal small-scale ripples with silt
drapes, a structure analogous to flaser bedding and indicating a variable low-velocity
flow regime. Some of the ripples appear to be flat topped, although the evidence is
equivocal, which may indicate planing in very shallow water. The coarser cross-
bedding lower down in the sandstone unit indicates a higher current velocity and the
accumulated evidence suggests that the unit may represent a channel which was
gradually silted up. The surface of the slab preserves traces of roots to which biogenic
disruption of the bedding structures may also be attributed. The material immediately
overlying the trail lacks evidence of penetration by roots which suggests that the
vegetation was penecontemporaneous with the formation of the trail and was not
established subsequently. This is supported by the occurrence of a root apparently
pulled into alignment by the arthropod (PI. 29, fig. 1). The clarity of the imprints in
such a coarse lithology (PI. 30, figs. 5, 6) renders it unlikely that they were subaqueous
in origin.

Roots occur at many horizons in the succession (text-fig. 1), as do Lepidodendron
sp. (Gunn et al. 1903, pp. 49, 156, 166; Gunn in Tyrrell 1928, p. 60) and Stigmaria sp.

BRIGGS ET AL.\ MYRIAPOD TRAIL

275

Abundant plants occur in a dark-grey micaceous siltstone 1-4 m above the trail
horizon (text-fig. 1). These consist (Dr. A. C. Scott, pers. comm.) of fragmentary
‘carbonized’ compressions, stem and root fragments of which about half are unidentifi-
able, and include one piece of unidentified lycopod (HM Pb.4577). The most common
plant is a pteridosperm, Sphenopteris elegans (Brongniart) Sternberg = Heterangium
grievii (Williamson), represented mainly by ribbed stems (an up-to-date synonymy
and description is given by Van Amerom 1975, pp. 9-14, pis. 1-3). Stratigraphically
S. elegans is most widely recorded (Kidston 1923, pp. 242-248; Walton 1940, p. 124
as Diplotmema adiantoides) in the Scottish Upper Limestone and Limestone Coal
Groups. Maceration of the siltstone yielded one unidentified Lagenicula- like
megaspore.

KEY

hVi LIMESTONE
Wl SHALE

EE3 SILTSTONE

1 I SANDSTONE

^1 FOUL COAL
lllllllllll IRONSTONE
IH FIRECLAY
1**1 GANISTER

CROSS-BEDDING

kSD DYKE

SCALE

text-fig. 1. Measured section through the Limestone Coal Group north-west of Laggan, showing the posi-
tion of the trace fossil and the overlying plant-rich horizon.

276

PALAEONTOLOGY, VOLUME 22

THE TRACE FOSSIL

Description

Both trails are somewhat indistinct from over two hundred years of weathering. The
longer one (PI. 28) is 36 cm wide, and extends some 6-25 m across the bedding plane.
It may be considered, for descriptive purposes, to run east-west, although it curves
gently southward at the western and northward at the eastern end of the exposure.
A large slab removed from the central part of the trail (PI. 28), presumably by the
quarriers, divides it into two sections. The western end is thinner bedded and separated
readily from the substrate, whereas the unit thickens to the east and becomes fused
to the subjacent sandstone. The western section was collected by us (HM X.1041,
PI. 29, fig. 2), and reinforced latex peels taken of the eastern section (and shorter trail)
were used to prepare fibreglass casts (HM X.1042).

The imprints of individual appendages are usually represented by a single depres-
sion oriented roughly normal to the axis of the trail (PI. 29, fig. 1 ; PI. 30, fig. 1). Each
depression is generally deeper and wider at the end nearest the trail margin and
shallows and tapers towards the axis. The appendages also pushed up mounds of
sediment between the imprints which may be preserved raised slightly above the
general level of the bed surface. The tracks become progressively more distinctly
preserved towards the eastern end of the exposure, where sharper impressions were
revealed by removal of overburden from the northern set in 1977, the width of the
paired series of imprints increasing from about 5 cm to a maximum of 9 cm (PI. 30,
figs. 1-4). The footfalls of the shorter appendages, which occur inside the majority,
are only evident at this end. This is thought to be due to a variation in the water
content of the substrate when the track was formed; the sediment probably became
wetter (or perhaps drier) and less cohesive, accounting for the poorer preservation
westward (PI. 29, fig. 2). The poorly defined prints appear to be different in character
rather than the product of weathering; the entire slab is thought to have been exposed
for the same length of time (c. 200 years). Only one series of imprints (PI. 29, fig. 3)
is preserved in the shorter trail, but the reason for this is not clear. During removal of
overburden from the eastern section of the longer trail, a careful search was made to
see if the previously exposed trail was only an undertrack (Goldring and Seilacher
1971, p. 424), the original being within the removed material. No tracks were found
in the overlying sandstone even though it split along planes that might be expected
to show them had they been present. The trail therefore appears to be a true surface
feature, albeit modified by erosion ; no loss by omission of track elements (‘fallout’
of Goldring and Seilacher 1971) has occurred. This is borne out by the imprints’
general lack of sharpness (Goldring and Seilacher 1971, p. 428).

EXPLANATION OF PLATE 28

Diplichnites cuithensis sp. nov. View westward across the bedding plane bearing the trail as it was in 1975.
The slab just in view at the western end, beyond the missing portion, has since been removed to con-
stitute a holotype (HM X.1041, PI. 29, fig. 2). The scale bar represents 10 cm.

PLATE 28

BRIGGS, ROLFE and BRANNAN, Myriapod trail

278

PALAEONTOLOGY, VOLUME 22

Individual imprints are spaced closely and fairly regularly at a linear density of
about one per cm (PI. 29, fig. 1 ; PI. 30, fig. 1). A certain amount of interference occurs
due to footfalls almost coinciding, but this can usually be identified except where the
preservation is poor or the impressions have merged due to water in the substrate
(PI. 29, fig. 2). At the newly exposed eastern extremity the well-preserved imprint of
a shorter appendage has survived apparently undistorted (PI. 30, figs. 5, 6). It shows
the distinct impression of three spines, the median one deepest and most distal with
respect to the appendage which made the imprint, the other two shallower and flank-
ing it symmetrically separated by a distance of about 14 mm. The imprints of the three
spines are elongate and they shallow towards the axis of the trail. The individual
traces produced by these three projections usually coalesce to form a single depression.
The groove made by the largest median spine can sometimes be distinguished as an
adaxially tapering extension of the imprint. The imprints of some of the shorter
appendages which fall adaxial of the rest, are relatively narrow and shallow and may
represent the median spine alone; these appendages may have supported less weight.
Evidence for the direction of locomotion based on the orientation of the tracks to the
axis of the trail and the mounds of sediment pushed up by the appendages is equivocal
(cf. Hanken and Stormer 1975, pp. 262, 263; Goldring and Seilacher 1971, p. 426).

Interpretation

A number of factors indicate that the trail is that of Arthropleura, although body
fossils of this giant myriapod have yet to be reported from Arran. The large number
of regularly spaced imprints eliminates any vertebrate of equivalent size, and all but
multipedal arthropods (on grounds partly discussed by Gevers et al. 1971, pp. 87-91).
Eurypterids are heteropodous, and produce a series of distinctive tracks repeated en
echelon (Hanken and Stormer 1975; Goldring and Seilacher 1971, p. 429; Waterston
pers. comm, although cf. Merostomichnites trails), quite unlike the present trail.
Eurypterid trails may also show a median longitudinal groove, made by the genital
appendage, and such a feature is lacking. Scorpions, such as Gigantoscorpio , pre-
sumably left distinctive Paleohelcura trails, and individual tracks should show a
characteristic trifid (or even pentadactyle) imprint from the plantigrade foot with its
terminal claw (Stormer 1963, fig. 37). Such imprints are found in Paleohelcura
( = Beaconichnus ) antarcticum (Gevers), which might have been left by a large
amphibious scorpion like Gigantoscorpio, rather than by the eurypterid suggested by
Gevers et al. (1971, p. 90). A number of large, unusual, partially known chelicerates
existed in the Carboniferous such as Cyrtoctenus Stormer and Waterston, 1968, but
these would presumably also have produced trails with widely separated en echelon

EXPLANATION OF PLATE 29

Figs. 1-3. Diplichnites cuithensis sp. nov. 1, HM X. 1042/2, fibreglass replica of central portion of trail,
the arrow indicating a root aligned by the arthropod (west, xO-15). 2, HM X.1041, holotype, slab
originally at western extremity of exposure showing gradual reduction in sharpness and finally non-
preservation of tracks (east, xO-15). 3, HM X. 1 043, fibreglass replica of best preserved series of imprints
of additional trail which runs approximately normal to the main example (west, x 0-15).

PLATE 29

BRIGGS, ROLFE and BRANNAN, Myriapod trail

280

PALAEONTOLOGY, VOLUME 22

groups of heteropodous tracks. The large size of the Arran trail is consistent with
Arthropleura , one of the largest arthropods known, which reaches lengths of up to
1 -8 m. The individual making the trail would have measured only about 1 m. The sedi-
mentary environment is similar to that in which the body fossils of Arthropleura occur,
a cyclic sequence of sediments including coal, representing flood plain or deltaic
swamp. In addition, the isolated imprint described above (PI. 30, figs. 5, 6) conforms
exactly to the distal extremity of the walking limbs which terminate in a single spine
flanked by the paired spines of the preceding segment (Rolfe and Ingham 1967;
Rolfe 1969, fig. 390).

Analysis

The trail produced by an arthropod may be envisaged as a ‘set’ of footprints (the
imprints of three paired appendages in the case of a hexapod) repeated at an interval
determined by the distance between successive footfalls of the same appendage,
i.e. the stride length. The stride is made up of two elements, a propulsive backstroke
when the appendage is in contact with the ground, and a recovery forward stroke.
The distance travelled during the backstroke depends on the angle of swing of the
limb and the dimensions of the arthropod. The distance travelled during the forward
stroke depends on the relative durations of forward and backward strokes.

The size of the trail indicates an Arthropleura about 1 m long, assuming that Rolfe
and Ingham (1967, fig. 1) have reconstructed the attitude of the appendages to the
trunk correctly, and a study of the relative dimensions of their reconstruction shows
that the appendages would have been about 10 cm long. Angle of swing varies with the
rate of progress of the arthropod (Manton 1954, p. 335) but the large size of Arthro-
pleura and simple appendage structure suggest a maximum value of about 60°.
The base of an appendage 10 cm long swinging through an angle of 60° will travel

Notes on the plates

Plate 29 and Plate 30, figs. 1, 5, and 6 were taken in low-angle incident light. Plate 30, figs. 2-4 were
taken with the fibreglass replica illuminated from behind (in transmitted light). The deepest part of individual
imprints is replicated in slightly thinner fibreglass than elsewhere, and therefore appears as lighter areas
using this technique ; the tracks are more easily identified as they are unobscured by shadows. The figures
are arranged on the plates so that the trail is orientated as it is viewed in situ— the western end to the right
or top of the plate. Directions of illumination are given in the explanations with reference to the margins of
the plate and are bracketed with the magnification.

EXPLANATION OF PLATE 30

Figs. 1-6. Diplichnites cuithensis sp. nov. 1-4, HM X. 1042/3, fibreglass replica of best-preserved portion
of the trail exposed at the eastern end. 1, (north, x 0T5). 2, pair of closely spaced repeated imprints
indicating stride length ringed, scale bar 10 cm (transmitted, x 0T5). 3, 4, northern set of imprints (on
the right side in fig. 2) largely exposed by excavation in 1977 (transmitted, x 0-35); individual imprints
outlined, closely spaced repeated imprints ringed, those flanking them marked by dots in fig. 4 to show
evidence of stride length and twenty-three pairs of walking appendages. 5, 6, HM X.1045, latex of
counterpart (X.1044) removed in 1977 showing detail of well-preserved imprint indicated x in fig. 4
(north-east, x 0-75).

PLATE 30

BRIGGS, ROLFE and BRANNAN, Myriapod trail

282

PALAEONTOLOGY, VOLUME 22

10 cm during the execution of the backstroke; this provides an estimate for this
component of the stride length.

If the length of stride can be deduced from the trail, the pattern of gait (the ratio
of time spent in the forward stroke to that in the backstroke; Manton 1950, p. 534;
1977, p. 298) can then be calculated using the distance travelled during the back-
stroke estimated above. The speed and power achieved by an arthropod depend on
the relative duration of the forward and backstroke. More powerful gaits (such as
those employed by polydesmoid millipedes pushing through leaf litter) at relatively
slower speeds are the result of a longer backstroke and hence a larger number of
appendages in contact with the ground at any given moment. Stride lengths are
notoriously difficult to determine even in Recent millipedes (Manton 1954, p. 337).
A large number of uniform appendages produces a dense trail of evenly spaced foot-
falls and Manton (1954; 1977, p. 504) often found it necessary to place a ‘boot’ on
one limb to determine the stride length. The smaller number of appendages may make
the stride length easier to identify in the tracks of chilopods. Variations in limb
length may ensure that the footfalls of some appendages occupy characteristic
positions.

A rhythm or repetition in the track will not necessarily conform to stride length
(cf. Manton 1952, p. 153, fig. 11; 1977, fig. 7.6c, d ) but may represent imprints of
intermediate appendages. This is the most likely explanation for the repetition of
groups of footfalls at intervals of 6 cm, apparent in the most easterly part of the trail
exposed (top section of series on the right of PI. 30, fig. 1). 6 cm is an unlikely stride
length as the distance travelled by the leg bases during the backstroke alone with an
angle of 60° is 10 cm. More satisfactory evidence of the stride is provided by the
individual imprints of the shorter appendages which plot adaxial of the majority,
where they are not obscured by subsequent footfalls. At the eastern end of the
northern series of imprints (PI. 30, figs. 2-4) two closely spaced footfalls and one more
widely spaced flanking them on either side are repeated at an interval of about 22 cm.
The repeated pair is also evident, although poorly preserved, in equivalent positions
in the southern series (PI. 30, fig. 2) indicating that the left and right appendages were
in phase. No satisfactory evidence for any other stride length has been observed.

The stride length of 22 cm evident in the trail combined with a backstroke estimated
as 10 cm (assuming an angle of swing of 60°) gives a ratio of the duration of forward
to backstroke of 12:10, i.e. a pattern of gait of 5-5 : 4-5 (diagrammatically represented
in text-fig. 2). Thus approximately 45% of the walking appendages were in contact
with the ground at any given time. The trail provides no satisfactory basis for deducing
the angle of swing; it cannot be assumed, for example, that a relationship exists
between this angle and that made by the elongate imprints with the axis of the trail.
In places the latter approaches 90° suggesting an angle of swing of zero ! It is quite
possible, however, that the angle was less than 60° implying an even shorter back-
stroke and correspondingly faster gait, assuming that the evidence for a stride length
of 22 cm is reliable. The number of walking appendages is indicated by the number
of footprints between two successive imprints of the same limb, i.e. within a stride
length; there appears to have been about twenty-three pairs (PI. 30, fig. 4). The
number of appendages employed in a single metachronal cycle (executing various
stages of the same stride) depends on the phase difference between them, i.e. the

BRIGGS ET AL.-. MYRIAPOD TRAIL

283

TIME

text-fig. 2. Diagrammatic representation of the gait of Arthropleura based on the trail.

The movements of three successive appendages are shown, the forward swing by a thin
upward-sloping line, the propulsive backstroke by a thicker downward-sloping line.

The relative durations of the forward and backward strokes are 5 5: 4-5 and the phase
difference ( p d) between successive appendages is 0T4. t is the time separating the foot-
falls of the two successive appendages. (Diagram constructed after Manton.)

fraction of the duration of a pace separating two successive limbs (the smaller the
phase difference below 0-5 the greater the number per cycle). The phase difference is
expressed as that proportion of a pace by which appendage n-\- 1 is in advance of
appendages (text-fig. 2; cf. Manton 1950, fig. 1; 1977, fig. 7.1). Changes in the pattern
of gait in Recent arthropods are always accompanied by alterations in the phase
difference, but this relationship cannot be expressed as a formula which would allow
phase difference to be calculated from the gait deduced above. However, Recent
arthropods tend to space the propulsive limbs so that the distance between the foot-
fall of the last of one metachronal wave and the first of the following approximately
equals that between propulsive limbs within a wave (Manton 1954, pp. 329, 330,
text-fig. 5; 1977, fig. 7.4). This factor can be used to assess the number of appendages
in a wave, and hence the phase difference (text-fig. 3), assuming a pattern of 5-5: 4-5
and an angle of swing of 60° (a value of less than 0-5 is indicated by the even spacing
of the footprints in the trail and the fact that the left and right appendages were in
phase). The most regular spacing is achieved with seven appendages in a wave, suggest-
ing a phase difference of 0T4. Thus three metachronal waves would have been
evident travelling forward along the length of the arthropod (a gait similar in most
respects to that employed by Polydesmus running freely; Manton 1954, pi. 55,
fig. 38; 1977, pi. 5, fig./).

The data deduced above can be used to plot a theoretical trail (cf. Manton 1950).
The stride length is known ; the relative spacing of the imprints in a single set, how-
ever, must be determined. Distance from the axis of the trail depends mainly on the
length of the appendages. Rolfe and Ingham’s (1967, fig. 1) reconstructed individual
has twenty-eight pairs, but the anamorphic ontogeny of Arthropleura suggests that
the relative dimensions of an individual with only twenty-three may be assessed by
removing five median somites and closing the gap. The footprints of shorter limbs
at the anterior and posterior end will plot inside the rest, thus accounting for a maximum
width of each series of imprints of about 9 cm (PI. 30, figs. 1-4). The linear distance
separating a footprint of limb n from the first subsequent footprint of limb w + 1 (the

284

PALAEONTOLOGY, VOLUME 22
0.2

0.14

text-fig. 3. Diagram showing the effect of altering the phase difference between
successive appendages on the spacing of the propulsive limbs. A relative duration
of forward to backstroke of 5-5 : 4-5 (i.e. approximately 45% of the appendages
in contact with the ground) and an angle of swing of 60° are assumed. Propulsive
limbs (in contact with the substrate) are represented by thick lines, the dots
emphasizing the spacing between them; legs performing the forward recovery
stroke are shown as thin lines. (After Manton 1954.)

appendage behind n) is given as d (the distance travelled by the arthropod during the
time between the footfalls of the two successive appendages) minus the exsagittal
distance s separating the middle of the bases of the two limbs (Manton 1950, p. 537).
d is the product of the time t (text-fig. 2) separating the footfalls of two successive
appendages, and the speed y. t can be calculated from the duration of pace (stride
length/speed, i.e. 22 cm /y unknown), if the phase difference (0-14) is known: thus
t = (22/y) — 0T4(22/y). Hence d = y( 0-86 x 22 )/y =18-9; the speed of progression y is
eliminated and need not be known. The distance 5 between the bases of the appendages
varies along the length of the arthropod and was estimated on the basis of Rolfe and
Ingham’s (1967, fig. 1) reconstruction. The resultant of ( d—s ) is positive and therefore
the footprint of «+ 1 is plotted that distance anterior of the footprint of appendage n
(Manton 1950, p. 537). The theoretical trail (text-fig. 4) is similar to the best-preserved
section of the fossil trail. Individual footprints in the latter could even be attributed
to particular appendages, but this is considered unwarranted on the available evidence.
It is likely, however, that the two closely spaced imprints upon which the deduced
stride length is largely based (PI. 30, fig. 4) represent the shortest appendages, i.e. the
first and last, which plot together in the theoretical trail (text-fig. 4). A theoretical

BRIGGS ET AL.\ MYRIAPOD TRAIL

285

o

n

2

DIRECTION OF PROGRESS

text-fig. 4. Theoretical trail of Arthropleura based on Rolfe and Ingham’s (1967) reconstruction of the
body, and the gait deduced from the trail (text-fig. 2). The arthropod is assumed to have had twenty-three
pairs of walking appendages. Each dot represents the imprint of a single appendage. The right side of the
trail represents a single set of imprints made by the right appendages (numbered in Arabic). Part of the next
set is also shown (numbered in Roman) offset but separated by the stride length of 22 cm. The left side shows
the trail generated when sets produced by the left appendages are superimposed overlapping as they would
in life. Individual imprints tend to merge giving the impression of a dashed line along the outer margin.
The imprints of appendages 1 and 23, which plot close together, are ringed at the beginning of the trail to
show how they indicate the stride length.

trail could also be generated by constructing a series of ‘gait diagrams’ showing the
positions which the appendages would occupy at equally spaced intervals of time.
Such diagrams have been used by Manton (1977, pp. 503, 504) in the analysis of gaits
of Recent uniramians, and have the advantage that the relationship between the
gait and the imprints produced is instantly apparent; the approach cannot, however,
be applied so readily to the trace fossil in question.

DISCUSSION

Rolfe and Ingham (1967, fig. 1) reconstructed Arthropleura employing a lower-
geared gait than that deduced from the trail, in which nineteen of the twenty-eight
pairs of walking appendages borne by the individual were included in a metachronal
wave (i.e. a phase difference of c. 0-05). This corresponds to a pattern of about 3:7
(text-fig. 5) in which 70% of the appendages are in contact with the ground at any
given time. If it is assumed that the distance travelled during the backstroke was not
greater than the maximum 10 cm postulated for the gait which produced the trail
(and it would likely have been reduced in a lo3ver-geared pattern), the resultant stride
would not have exceeded 14-3 cm. It might therefore be argued that Rolfe and Ingham’s
interpretation of the mode of life of Arthropleura— pushing through the debris on
the swamp floor— is incorrect in the light of the trace fossil attributed to it, which
shows evidence of a much higher geared gait. It is perfectly feasible, however, that
both gaits were performed by the same arthropod. Manton (1954, pi. 55, figs. 39, 40)
figured a female Polydesmus angustus running freely with a pattern of 5-9:41, and
also harnessed to a ‘sledge’ (to simulate the effect of pushing through leaf litter)
employing a pattern of 2-2 : 7-8. P. angustus thus displays a wider range of gaits than
that separating Rolfe and Ingham’s reconstruction of Arthropleura from the pattern
that apparently produced the trail. The low-gear gait would only have been employed

286

PALAEONTOLOGY, VOLUME 22

TIME

text-fig. 5 . Diagrammatic representation of the gait of Arthropleura reconstructed
by Rolfe and Ingham (1967, fig. 1). The relative durations of the forward and
backward strokes are 3 : 7 and the phase difference between successive appendages
is 0-05. (Symbols as on text-fig. 2.)

by Arthropleura when forcing its way through vegetation, an environment un-
conducive to the preservation of tracks. The arthropod would have quickened its
pace unimpeded by plant debris on more open ground, reducing the time spent on
the propulsive backstroke. The nature and interpretation of the trail imply a greater
flexibility of the appendages than that assumed by Manton (1977, pp. 234, 235).
Arthropleura apparently walked on the three distalmost spines of the limb (PI. 30,
figs. 5, 6), suggesting that the paired spines on the more proximal podomeres only
supported the appendage when walking over loose vegetation, for example. The
simple, undifferentiated nature of the podomeres would have imposed limitations on
flexibility, but this is compensated to an extent by the large number of podomeres
(Manton 1973, p. 273). The poorer preservation of a triangular area of cuticle at the
junction between the podomeres of some specimens (Rolfe and Ingham 1967, pi. 1,
fig. 11; Rolfe 1969, fig. 389) may represent the less sclerotized arthrodial membrane
of a simple hinge joint (cf. Anomalocaris , Briggs, in preparation).

The trail provides little basis for revising Rolfe’s (1969) interpretation of the mode
of life of Arthropleura , which was based on the occurrence of the body fossils, although
it does indicate that the arthropod ventured out from the coal forest to cross abandoned
distributary channel sands. The observation that one root seems to have been aligned
(PI. 29, fig. 1, top right) by a limb dactyl, if this is not merely coincidental, suggests
that some vegetation had become established on the old channel fill.

A similar trace fossil to that on Arran, from the celebrated Joggins section
(Westphalian B) in Nova Scotia, was reported and figured by Ferguson (1965, 1966,
1975). It occurs in a sheet sand thickening into a channel sand (bed 39/S2 of Ferguson
1975, p. 74; Duff and Walton 1973, p. 370) and the sedimentary sequence, albeit
younger, represents a similar deltaic environment (Duff and Walton 1973; Way
1968). One of the three trails on the slab recovered (Ferguson 1975, fig. 4) was origin-
ally attributed to an amphibian (Ferguson 1965, p. 13), but all are now considered
to have been made by Arthropleura (Ferguson 1966, p. 128; 1975, p. 74, cf. fig. 4;
Baird in Carroll et al. 1972, p. 75). When the Nova Scotia trail maker was identified,
body fossils of the arthropod were not known from the Joggins section. Since then the

BRIGGS ET A L. : MYRIAPOD TRAIL

287

myriapod described as Amynilyspes springhillensis by Copeland (1957) from the same
Cumberland Group facies B at Springhill, has been recognized as a juvenile Arthro-
pleura (Rolfe 1969, p. R617). In addition, it is possible that the telson of the supposed
Hastimimal sp. (Copeland and Bolton 1960, p. 43) from Joggins also belongs to
Arthropleura. Similarly Eurypterusl pulicaris Salter from the Upper Carboniferous
of New Brunswick and the spined ‘supposed limbs of myriapods’ from Joggins
(Copeland 1957, p. 59, pi. 15, fig. 3) could be arthropleurid limbs: reinvestigation is
needed. Large, poorly preserved trails from the Westphalian D north of Florence,
Cape Breton County, Nova Scotia, have also been attributed to Arthropleura (Baird
in Carroll et al. 1972, p. 54). Some of the large cuticle fragments from the Joggins
hollow tree stumps previously thought to be tetrapod skin, then eurypterid, were
referred by Carroll (1972, p. 71) to Arthropleura. However, Dr. C. D. Waterston
(pers. comm.) has pointed out that all the material figured by Dawson in 1863, and
most of that in 1882, is comparable with large eurypterids such as Vernonopterus and
Dunsopterus (Waterston 1957, 1968). Hibbertopterus (= Campylocephalus) cf.
scouleri, a similarly ornamented form, has been recorded from the Upper Carboni-
ferous of Port Hood, Nova Scotia (Copeland and Bolton 1960). The Joggins trails
are smaller than the Arran example, ranging in total width from 20 to 26 cm. The
largest consists of a paired series of regularly spaced oval depressions, elongate
normal to the axis, presumably representing groups of near coincident footfalls which
cannot be distinguished. The smaller trails, which are better preserved (Ferguson
1966, fig. 2; 1975, fig. 4) show individual imprints apparently arranged in closely
spaced diagonals. Ferguson (pers. comm.) ascribes the difference in preservation to
a decrease in water content of the sediment, and presumably increased cohesiveness,
when the smaller trails were formed. The trails are clearly those of a multipedal
arthropod, and the paired limbs apparently moved in phase, indicating a phase
difference between successive appendages of less than 0-5. Although Scudder (1891,
pp. 10-18; 1895) reported eight millipedes from Joggins, one scorpion (?) (Scudder
1 895, cf. Petrunkevitch 1913), and the crustacean Pygocephalus , to which may be added
the eurypterid mentioned above, the trace fossil is attributed to Arthropleura as the
only myriapod of sufficient size. It proved impossible to interpret the best-preserved
example with confidence, although Dr. Ferguson kindly provided a cast of a 0-25 m
length of the clearest imprints. The gait employed was probably similar to that which
produced the Arran trail (a detailed study of the entire length of the smallest trail
using the approach described above might confirm a stride of about 1 3 cm and approxi-
mately thirty pairs of limbs). The paired series of imprints, however, occupy a
maximum of about 60% of the total width as compared to 50% in the Arran trail.
Further they are not concentrated along the abaxial margins to the same extent as in
the Arran example, but appear to show a more even density throughout the width of
the paired series. This suggests a greater variation in appendage length and flexibility
in the smaller Joggins arthropleurids.

The body fossil of Arthropleura has been recorded in sediments from Westphalian
A to Stephanian C in age (Rolfe 1969). Other trace fossils attributed to this arthropod
(Ferguson 1966, 1975; Baird in Carroll et al. 1972, p. 54) fall within this range and
occur in a similar environment. The trail in the Namurian (EJ Limestone Coal Group
of Arran therefore represents the earliest evidence of Arthropleura ; it probably

288

PALAEONTOLOGY, VOLUME 22

simply reflects the establishment of coal swamp conditions in Scotland sooner than
elsewhere.

Stormer’s (1976) description of a genus from Aiken an der Mosel, Germany,
extended the range of the arthropleurids to the Lower Devonian. He reconstructed
Eoarthropleura (1976, figs. 45, 46) with the opposing limbs out of phase, the body
undulating laterally as the propulsive limbs converged. Stormer justified this rela-
tively rapid gait, leading to horizontal undulations, by the wide posterior doublure
of the tergites which allowed considerable movement between them, and the over-all
morphology which indicates that ‘the Devonian form was more agile than the
Carboniferous one’ (1976, p. 43). It seems unlikely, however, that the gait recon-
structed would have been suitable for the habitat envisaged by Stormer (p. 113),
which is essentially similar to that of the Carboniferous genus. Body undulations
occur in some myriapods when the two limbs of a pair are used in opposite phase.
They are a hindrance to locomotion and are controlled as far as possible by a variety
of adaptations (tergite heteronomy, segmental musculature), generally only appear-
ing during fast running (in epimorphic chilopods). Stormer (1976, p. 95) draws an
analogy between the trunk flexibility in Eoarthropleura and in the Recent Symphyla,
which are adapted to follow the tortuous passages between soil particles. These
myriapods, however, do not employ a gait with the opposing limbs out of phase
(Manton 1977, fig. 7.8) and there is no tendency to produce body undulations; phase
differences are less than 0-5. The relatively small Eoarthropleura ( c . 1 1 cm), although
lacking the extra tergites of the Symphyla (Manton 1977, pp. 373-375) may have
twisted and turned its way through swamp vegetation rather than forcing a passage
by pushing, as Arthropleura did. It is unlikely to have walked on open ground with
the opposing appendages out of phase, because this, combined with the flexibility
of the articulations, would have thrown the body into severe undulations. It pre-
sumably employed a similar range of gaits to that deduced for Arthropleura.

TAXONOMY

Many ichnogenera of myriapods have been described: Acanthichnus, Acripes, Arthro-
podichnus, Beaconichnus, Diplichnites, Diplop odichnus, Diplopodomorpha, Hamipes,
Harpepus , Merostomichnites, Myriapodites, Pterichnus, Tasmanadia, Umfolozia
are among those listed by Hantzschel (1975). To these may be added Dunstairia and
Stiaria from the overlooked work by Smith (1909). The type specimens of Diplichnites
and Myriapodites came from the Joggins section that has yielded the trail described
by Ferguson. Indeed, the two distinct types of trails on Ferguson’s (1975, fig. 4) slab
could well be referred to those two genera. Unfortunately, Diplichnites has become
firmly entrenched in recent literature as a trilobite locomotion trail, despite the fact
that Dawson’s holotype of his type species came from deltaic Westphalian, and com-
prised two rows of tracks six inches apart. This usage stems from Seilacher’s (1955,
pp. 342-343) suggestion that when a trilobite moves straight forward or backward,
it will leave a trail that will be difficult to differentiate from trails of other arthropods.
He ‘provisionally’ used Dawson’s name Diplichnites for such trails of trilobites, as
well as for those of trilobites moving obliquely forward. Since such trails were of
rather generalized type, however, the actual omnibus name applied to them was felt

BRIGGS ET AL.: MYRIAPOD TRAIL

289

by Seilacher to be of secondary importance. Granted that trace fossil genera are
liable to expansion with use (Hantzschel 1975, p. W35), this nevertheless seems an
undue extension in meaning of the term, and it would be better to revive one of the
undoubted trilobite ichnogenera (as by Osgood 1970; Anderson 1975) at present
regarded as junior synonyms of Diplichnites. Most workers since Dawson have
ignored the fact that he originally correctly deduced that his large trails ‘were prob-
ably produced by a land or freshwater animal— possibly a large crustacean or gigantic
annelid or myriapod ’ (1862, p. 7— our italics). Dawson (1891, p. 389) also noted that
‘the space between the rows of marks is slightly depressed and smoothed, as if with
a heavy body’, a feature more likely in terrestrial myriapods than trilobite trails.

The holotype of the type species of Diplichnites has not yet been located. Baird
(in Carroll et al. 1972, p. 54) has used the name Duovestigia Butts, 1891 for the arthro-
pleurid trail from Florence, Nova Scotia. This genus (cf. Kuhn in Hantzschel 1975,
p. W184) is based on small trails from the Upper Carboniferous of Missouri, and the
name could equally be used for the Arran trail. Baird (letter to Ferguson, 1966) has
also suggested that the Westphalian (sic) Acripes is a junior synonym, and that similar
trails are figured by Abel (1935, figs. 222?, 238). The figured, but undescribed,
Beaconichnus giganteum Gevers in Hantzschel, 1975, p. W45 should also be assigned
to Duovestigia, thus extending the record to the Devonian of Antarctica. Clearly,
much revision is required of myriapod ichnogenera, in the light of work on modern
myriapod gaits, and with study of living myriapod trails. In the interim, Duovestigia
may be regarded as a junior synonym of Diplichnites as emended below (cf. Hantzschel
1975, p. W61).

Ichnogenus diplichnites Dawson, 1873 (emend.)

Type ichnospecies. D. aenigma Dawson, 1873, by original monotypy.

Emended diagnosis. Morphologically simple trail, up to 36 cm wide, consisting of two
parallel series of tracks (each up to 9 cm wide); individual tracks elongate roughly
normal to trail axis, spaced closely and regularly at up to about one per cm.

Diplichnites cuithensis ichnosp. nov.

Plates 28-30

Derivation of name. From Cuithe (Gaelic: cattle-fold) — the name of the cleared clachan near the locality.
Holotype. Slab Hunterian Museum X.1041 (PI. 29, fig. 2).

Paratypes. Four fibreglass replicas of trail, HM X. 1042/1-3, X.1043; and small fragments of counterpart
X.1044/1-3.

Type locality. Salt pans harbour quarry, Laggan, Arran, Scotland.

Horizon. Limestone Coal Group, Pendleian Stage, Namurian Series, Carboniferous.

Diagnosis. Very large Diplichnites, with rare trifid tracks shallowing towards axis of trail.

Description. See pp. 276-278 above.

Interpretation. Locomotion trail of large myriapod Arthropleura.

290

PALAEONTOLOGY, VOLUME 22

Acknowledgements. We thank Drs. A. G. Smith and R. Johns for drawing D. E. G. B.’s attention to this
trail; Mr. C. Fforde for permitting its collection; the Manpower Services Commission for a grant to
collect and prepare the specimen; and to the following for guidance and assistance in this project: Dr.
R. Anderton, Dr. D. Baird, Dr. B. J. Bluck, Dr. R. L. Carroll, Professor P. McL. D. Duff, Mrs. V. Edwards,
Dr. L. Ferguson, Dr. R. Goldring, Mrs. M. Leeper, Mr. W. Lindsay, Mr. D. McLean, Mr. F. Murchie,
Dr. G. Page, Dr. W. A. Read, Professor W. A. S. Sarjeant, Dr. A. C. Scott and Dr. C. D. Waterston.

REFERENCES

abel, o. 1935. Vorzeitliche Lebensspuren. Jena, xv+644 pp.

anderson, a. M. 1975. The ‘trilobite’ trackways in the Table Mountain Group (Ordovician) of South Africa.
Palaeont. afr. 18, 35-45.

bryce, J. 1859. The geology of Clydesdale and Arran. Glasgow, 197 pp.

butts, E. 1891. Recently discovered foot-prints of the Amphibian Age, in the Upper Coal Measure Group
of Kansas City, Missouri. Kansas City Scientist 5 (2), 17-19.

CARROLL, R. L., belt, E. s., dineley, D. l., baird, d. and mcgregor, D. c. 1972. Vertebrate paleontology of
Eastern Canada. Guidebook, Excursion A59, 24th Internat. geol. Congr., Montreal, pp. 1-113.
copeland, M. J. 1957. The arthropod fauna of the Upper Carboniferous rocks of the Maritime Provinces.
Mem. geol. Surv. Canada , 286, 1-64, 21 pis.

— and bolton, t. e. 1960. The Eurypterida of Canada. Bull. geol. Surv. Canada, 60, 13-47.
dawson, J. w. 1862. Notice of the discovery of additional remains of land animals in the Coal-Measures of

the South Joggins, Nova Scotia. Q. Jl geol. Soc. Lond. 18, 5-7.

— 1863. The air-breathers of the Coal Period in Nova Scotia. Canad. Nat. Geol. also Proc. Nat. Hist.
Soc. Montreal. 8, 1-12, 81-92, 159-175, 268-295, pis. 1-6.

— 1873. Impressions and footprints of aquatic animals and imitative markings, on Carboniferous rocks.
Amer. J. Sci. 105, 16-24.

— 1882. On the results of recent explorations of erect trees containing animal remains in the Coal
Formation of Nova Scotia. Phil. Trans. R. Soc. Lond. 173, 621-659, pis. 39-47.

1891. Acadian geology, 4th edn. London, xxvi+833 pp.

duff, p. mcl. D. and walton, E. k. 1973. Carboniferous sediments at Joggins, Nova Scotia. Septieme
Congr. internat. Strat. geol. Carb., Krefeld 1971, C.R. 2, 365-379.
ferguson, l. 1965. [Survey of research topics.] Maritime Sediments, 1, 13.

— 1966. The recovery of some large track-bearing slabs from Joggins, Nova Scotia. Ibid. 2, 128-130.

— 1975. The Joggins section. Ibid. 11, 69-76 [= 1976. In Ancient sediments of Nova Scotia. Fieldtrip
guidebook. Eastern Section, Soc. Econ. Paleont. Mineral. 111-118]

francis, E. H. 1965. Carboniferous. In craig, g. y. (ed.), The geology of Scotland. Edinburgh and London,
pp. 309-357.

frazier, D. E. and ozanik, a. 1969. Recent peat deposits— Louisiana coastal plain. Spec. Pap. geol. Soc.
Am. 114,63-85.

GEORGE, T. N. 1960. The stratigraphical evolution of the Midland Valley. Trans, geol. Soc. Glasg. 24, 32-107.

JOHNSON, G. A. L., MITCHELL, M., PRENTICE, J. E., RAMSBOTTOM, W. H. C., SEVASTOPULO, G. D. and WILSON,

R. b. 1976. A correlation of Dinantian rocks in the British Isles. Geol. Soc. Lond., Spec. Rep. 7, 87 pp.
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.
goldring, R. and seilacher, a. 1971. Limulid undertracks and their sedimentological implications.
N. Jb. Geol. Paldont.-Abh. 137, 422-442.

gunn, w., geikie, a., peach, b. n. and harker, a. 1903. The geology of north Arran, south Bute and the
Cumbraes, with parts of Ayrshire and Kintyre (Sheet 21, Scotland). Mem. geol. Surv. U.K. vii+200 pp.
hantzschel, w. 1975. Miscellanea (Supplement I), trace fossils and problematica. In teichert, c. (ed.),
Treatise on invertebrate paleontology, part W. Boulder, Colorado and Lawrence, Kansas, xxi + 269 pp.
hanken, n. m. and stormer, L. 1975. The trail of a large Silurian eurypterid. Fossils and strata, 4, 255-270.
kidston, R. 1923. Fossil plants of the Carboniferous rocks of Great Britain, 2 (3). Mem. geol. Surv. U.K.—
Palaeont., pp. 199-274, pis. 48-68.

mcbride, E. F. 1963. Classification of common sandstones. J. sedim. Petrol. 33, 664-669.

BRIGGS ET AL.: MYRIAPOD TRAIL

291

macgregor, m. 1965. Excursion guide to the geology of Arran. Glasgow, 192 pp.

manton, s. M. 1950. The evolution of arthropodan locomotory mechanisms. Part 1. The locomotion of
Peripatus. J. Linn. Soc. ( Zool .), 41, 529-570.

— 1952. Id. Part 3. The locomotion of the Chilopoda and Pauropoda. Ibid. 42, 118-166.

— 1954. Id. Part 4. The structure, habits and evolution of the Diplopoda. Ibid. 42, 299-368.

— 1973. Id. Part 11. Habits, morphology and evolution of the Uniramia (Onychophora, Myriapoda,
Hexapoda) and comparisons with the Arachnida, together with a functional review of Uniramian
musculature. Ibid. 53, 257-375.

— 1977. The Arthropoda — habits, functional morphology, and evolution. Oxford, xxii + 527 pp.
moore, d. 1959. Role of deltas in the formation of some British Lower Carboniferous cyclothems. J. Geol.

67, 522-539.

osgood, R. G. 1970. Trace fossils of the Cincinnati area. Palaeontogr. Amer. 6 (41), 277-444.
petrunkevitch, A. 1913. A monograph on the terrestrial Palaeozoic Arachnida of North America. Trans.
Conn. Acad. Arts. Sci. 18, 1-137.

ramsay, A. c. 1841. The geology of the Island of Arran from original survey. Glasgow, 78 pp.
ramsbottom, w. h. c. 1978. Correlation of the Scottish Upper Limestone Group (Namurian) with that of
the North of England. Scott. J. Geol. 13, 327-330.

read, w. a. and dean, J. M. 1976. Cycles and subsidence: their relationship in different sedimentary and
tectonic environments in the Scottish Carboniferous. Sedimentology, 23, 107-120.
rolfe, w. d. i. 1969. Arthropleurida. In moore, r. c. (ed.). Treatise on invertebrate paleontology, part R,
pp. R. 607-620. Boulder, Colorado and Lawrence, Kansas.

— and ingham, j. K. 1967. Limb structure, affinity and diet of the Carboniferous ‘centipede’ Arthro-
pleura. Scott. J. Geol. 3, 118-124, 1 pi.

scudder, s. H. 1891. Index to described fossil insects. Bull. U.S. geol. Surv. 71, 744 pp.

— 1895. Notes on myriapods and arachnids found in sigillarian stumps in the Nova Scotia coal field.
Geol. Surv. Canada, Contrib. Canad. Palaeont. 2, 57-66, pis. 4, 5.

seilacher, a. 1955. Spuren und Lebensweise der Trilobiten. In schindewolf, o. h. and seilacher, a.
Beitrage zur Kenntnis des Kambriums in der Salt Range (Pakistan). Abh. Akad. Wiss. Lit. Mainz, math.-
nat. Kl. 10, 86-143, pis. 16-27.

smith, J. 1909. Upland fauna of the Old Red Sandstone formation of Carrick, Ayrshire. Kilwinning, 41 pp.
stormer, l. 1963. Gigantoscorpio willsi, a new scorpion from the Lower Carboniferous of Scotland, and its
associated preying microorganisms. Skr. Norske Vidensk.-Akad. Oslo 1. Mat.-Nat. Kl. N.S. 8, 1-171.

— 1976. Arthropods from the Lower Devonian of Aiken an der Mosel, Germany, part 5. Senck. leth.
57, 87-183, 10 pis.

— and waterston, c. d. 1968. Cyrtoctenus gen. nov., a large late Palaeozoic arthropod with pectinate
appendages. Trans. R. Soc. Edinb. 68, 63-104, 6 pis.

Tyrrell, G. w. 1928. The geology of Arran. Mem. geol. Surv. U.K. viii+292 pp.

van amerom, H. w. J. 1975. Die Eusphenopteridischen Pteridophyllen. Med. Ryks. Geol. Dienst. Ser.
C. 111-1-No. 7, 101 pp.

walton, i. 1940. An introduction to the study of fossil plants. London, x+188 pp.

waterston, c. d. 1957. The Scottish Carboniferous Eurypterida. Trans. R. Soc. Edinb. 63, 265-288.

— 1968. Further observations on the Scottish Carboniferous eurypterids. Ibid. 68, 1-20.

way, j. h. 1968. Bed thickness analysis of some Carboniferous fluvial sedimentary rocks near Joggins,
Nova Scotia. J. sedim. Petrol. 38, 424-433.

D. E. G. BRIGGS

Typescript received 28 April 1978
Revised typescript received 20 August 1978

Geology Department
Goldsmiths’ College
University of London
New Cross
London SE146NW

W. D. I. ROLFE
J. BRANNAN

Hunterian Museum
University of Glasgow
Glasgow G12 8QQ

THE MIDDLE PLEISTOCENE OSTRACOD
FAUNA OF THE WEST RUNTON
FRESHWATER BED, NORFOLK

by P. DE DECKKER

Abstract. Five freshwater ostracod assemblages, ecologically bound to environmental changes reflected by the
sediments, and corresponding to the late Beestonian and early Cromerian pollen assemblage zones, are described.
The ecological information recorded from the ostracods conforms to the climatic data offered by the palaeobotanical
record.

The ostracod fauna was analysed from sixty-two samples collected, in a strati-
graphical sequence, from a section of Goss’s Gap, West Runton, the same locality
from which West collected material for his palynological study of the type Cromer
Forest-bed Formation (West, in press), and which Stuart (1975) studied for his
investigation of the vertebrate fauna of the type Cromerian. For further details on
the locality, refer to Stuart’s diagram of the schematic section at West Runton (Stuart
1975, p. 65). The position of Stuart’s section labelled AJS corresponds to the one
studied here.

The sixty-two samples were taken from a 1 -625 m section dug along the beach,
15m north of the gap, in a channel cut in ferruginous gravel and capped with a leached
peaty layer, reddish black in colour. All the samples were collected at intervals of
2-5 cm, except for the bottom four samples which were at intervals of 5 cm.

In the laboratory, 40 gr. of each sample were treated in a sodium hexametaphosphate
solution for a period of 24-48 hours, and later washed through two sieves of SOOOjum
and 210/um each. The small fraction (<210|U,m) was not studied, as juveniles smaller
than 210/xm cannot usually be identified at specific level. Every ostracod and ostracod
fragment was later picked from the residues under a binocular microscope.

DESCRIPTION OF THE SECTION STUDIED AT WEST RUNTON

A detailed sketch of the profile at Goss’s Gap, West Runton, is presented in text-fig. 1 .
However, the lithology of the section is recorded below :

157-5- 162-5 cm: reddish brown to black leached peat with minor sand.

140-157-5 cm: almost entirely plant debris with some dark brown sand.

122-5-140 cm: interlayering of black silt, rich in plant debris, and beige to brown
fossiliferous sand (mainly molluscs remain).

100-122-5 cm: brown to beige sandy layers, highly fossiliferous (containing mainly
a macroscopic molluscan fauna) with few mud fragments interlayered with some
black silty ones with abundant plant debris.

87-5-100 cm: zone of brecciation where beige sand with few black pebbles is mixed in
black silt rich in shells and shell fragments, with small lenses of beige to dark grey
sandy material.

[Palaeontology, Vol. 22, Part 2, 1979, pp. 293-316, pis. 31-35.]

294 PALAEONTOLOGY, VOLUME 22

ds recorded at West Runton. Note that the ostracod assemblages correspond to the pollen assemblage zones
by West (in press). Sample numbers are to be preceded with the letters CR.

DE DECKKER: FRESHWATER OSTRACODS

295

55-87-5 cm: plant debris. Leaves present in the bottom 10 cm, then small fragments
of grey clay common in the next 25 cm. Shells and shell fragments frequent.

25-55 cm: brown to beige silt at the bottom and sand in the top half; few pebbles at
the bottom. Plant debris increasing from bottom to top.

0-25 cm: dark grey to black sand becoming silty towards the top with layers of black
cherty pebbles (often rounded) and black fragments.

Bottom: ferruginous gravel layer consisting mainly of quartz with fragments of
marine shells. This layer is at least 30 cm thick.

West (in press) studied a similar section (labelled WRAQ) at the same locality. Minor
differences in thickness of the various lithologies were noticed in comparing this
section with the one presented here. The reason for this is that the position of both
sections lies within a large channel (see Stuart 1975, fig. 2) accessible at Goss’s Gap.
Consequently, such small discrepancies were considered as insignificant and correla-
tion between various levels were made, even though small differences between the
two sections exist.

TAXONOMY

All the specimens collected at West Runton for this study are deposited in the British
Museum (collection OS9230-OS9302). The ostracod species are described in the same
taxonomic order as recorded by Hartmann and Puri (1974) in their palaeontological
classification of Ostracoda. Much of the data on ecology, stratigraphical range and
geographical distribution for most of the species is taken from publications by Diebel
and Pietrzeniuk (1969, 1975a, 19756, 1977) and Diebel and Wolfschlager (1975).

The following abbreviations are used: L = left valve, R = right valve, C = carapace.

Genus darwinula Brady and Norman, 1889
Darwinula sp.

Plate 33, fig. 1 1

Material. 1 R, 1 L.

Measurements. Length, 0-70 mm, height: 0-25 mm.

Genus paralimnocythere Carbonnel, 1965
Paralimnocythere compressa (Brady and Norman, 1889)

Plate 33, figs. 1-5, 7-10, 13

1889 Limnocythere inopinata var. compressa Brady and Norman, p. 170.

Material. 21 L, 16 R.

Measurements. Females, length: 0-51-0-60 mm, height: 0-30-0-32 mm; males, length: 0-50-0-56 mm,
height: 0-26 mm. Greatest height: females, \ from anterior; males, \ from anterior.

296

PALAEONTOLOGY, VOLUME 22

Description. Strongly sexually dimorphic ; shell very thin and extremely fragile. General outline rectangular,
but narrower in males. Externally, various protuberances present on valves (see PI. 33, figs. 1-3, 7-9, 13);
mediodorsally, anterior protuberance slightly smaller in females and posterior one very large in males and
pointed towards posterior. Dorsum inclined in adult females and almost horizontal in adult males. Ventrum
strongly concave in both sexes, sometimes more incurved in adult males. Shell reticulate all over, following
curvature of valves in posterior area. Internally, central muscle scars arranged in vertical row of four.
Numerous radial pore canals curved with some branched, as in all Paralimnocythere species.

Ecology. At West Runton, P. compressa is restricted to cold climates (late Beestonian A and late Beestonian
B Zones), accompanied by abundant aquatic plants typical of muddy substrates, and shallow rather still
bodies of water. Females are more common. A similar environment was postulated for this species by Diebel
and Pietrzeniuk (1969) at Sussenborn.

Geographical distribution. Recent: Great Britain. Fossil: Sussenborn near Weimar, Paludinenbank, near
Oranienburg, and Kumro near Neuzelle. Negadaev-Nikonov’s (1971) description and illustration of P.l
cf. compressa from fluviatile sediments at Tiraspol (Elster in age) is identical to P. compressa found at
West Runton. The Tiraspol specimens are larger (length: 0-75 mm, height: 0-4-0-35 mm).

Stratigraphical range. Pleistocene to Recent (Diebel and Pietrzeniuk 1969). The Paralimnocythere species
need revision before their stratigraphical importance in the Pleistocene can be recognized. The sediments
at Sussenborn, in which P. compressa is found, are Middle Pleistocene (Elster I) in age.

Genus ilyocypris Brady and Norman, 1889
Ilyocypris bradyi Sars, 1890

Plate 33, fig. 14

1890 Ilyocypris bradyi Sars, p. 59.

Material. 4 L, 2 R, 1 C.

Measurements. Length: 0-812 mm, height: 0-45 mm.

Ecology. Found in cold waters, sometimes slightly saline, in temporary ponds or springs; can withstand
only slight changes in water temperature.

Geographical distribution. Recent : holarctic regions.

Stratigraphical range. At least Pleistocene to Recent.

EXPLANATION OF PLATE 31

Figs. 1-14. Scottia browniana. 1, internal lateral view of juvenile LV, OS 9230, sample CR 39, L: 0-725 mm.
2, internal lateral view of juvenile LV, OS 9231, sample CR 39, L: 0-625 mm. 3, internal lateral view of
juvenile RV, OS 9232, sample CR 35, L: 0-70 mm. 4, internal lateral view of juvenile RV, OS 9233,
sample CR 35, L: 0-70 mm. 5, internal lateral view of adult LV, OS 9234, sample CR 44, L: 0-885 mm.
6, external lateral view of juvenile RV, OS 9235, sample CR 45, L: 0-725 mm. 7, external lateral view
of juvenile RV, OS 9236, sample CR 39, L: 0-725 mm. 8, external lateral view of adult RV, OS 9237,
sample CR 39, L: 0-825 mm. 9, external lateral view of adult RV, OS 9238, sample CR 39, L: 0-80 mm
10, enlargement of fig. 5 to show central muscle field. 11, external lateral view of juvenile RV, OS 9239,
sample CR 45, L : 0-50 mm. 12, external lateral view of juvenile RV, OS 9240, sample CR 15, L : 0-70 mm.
13, external lateral view of juvenile LV, OS 9241, sample CR 39, L: 0-55 mm. 14, external lateral view
of juvenile LV, OS 9242, sample CR 39, L: 0-625 mm.

Fig. 15. Scottia tumida External lateral view of adult RV, OS 9243, sample CR 33, L: 0-675 mm.

RV : right valve, LV : left valve, L : length.

PLATE 31

DE DECKKER, Pleistocene ostracods

298

PALAEONTOLOGY, VOLUME 22

Ilyocypris gibba (Ramdohr, 1 808)

Plate 33, fig. 15

1808 Cypris gibba Ramdohr, p. 91, pi. 3, figs. 13-17.

Material 2 L, 1 R, 1 C.

Measurements. Length: 0-975 mm, height: 0-525 mm.

Ecology. Found in small bodies of water, not subject to drying up, with clayey or muddy substrates (Klie
1938); also in quiet flowing water with temperatures ranging between 4 and 19-5 °C (Aim 1916).

Geographical distribution. Recent : Europe, North Africa, and North America.

Stratigraphical range. Middle Oligocene, Pleistocene to Recent.

Genus candona Baird, 1845
Candona angulata G. W. Muller, 1900

Plate 32, fig. 1 1

1900 Candona angulata G. W. Muller, p. 18, pi. 1, figs. 1-5.

Material. 1 L, 3 R.

Measurements. Length: 1-13 mm, height: 0-625 mm.

Ecology. Spring-form found in slightly saline waters (ranging between 0-4 and 13-4°/00)-
Geographical distribution. Recent : Europe, North Africa.

Stratigraphical range. At least Middle Pleistocene to Recent.

Remarks. Characterized by typically pointed posterior end to right valve in both
sexes and faint reticulation covering external posterior part of both valves in both
sexes at the adult stage. Greatest height at f from the anterior ; angle of posterior area
of valve steeply inclined.

EXPLANATION OF PLATE 32

Figs. 1-5. Scottia browniana. 1, external lateral view of juvenile RV, OS 9244, sample CR 45, L: 0-675 mm.
2, external lateral view of juvenile RV, OS 9245, sample CR 45, L: 0-687 mm. 3, external lateral view
of juvenile LV, OS 9246, sample CR 15, L: 0-70 mm. 4, dorsal view of carapace of adult, OS 9247,
sample CR 39, L: 0-825 mm. 5, ventral view of carapace of adult, OS 9248, sample CR 39, L: 0-75 mm.

Fig. 6. Candona Candida. Internal lateral view of adult RV, OS 9249, sample CR 3, L: 0-95 mm.

Figs. 7-9. C.parallela. 7, internal lateral view of adult LV, OS 9250, sample CR 13, L : 0-825 mm. 8, external
lateral view of LV of adult carapace, OS 9251, sample CR 29, L: 0-825 mm. 9, external lateral view of
adult RV, OS 9252, sample CR 13, L: 0-80 mm.

Fig. 10. C.fabaeformis. External lateral view of adult carapace, OS 9253, sample CR 18, L: 0-90 mm.

Fig. 11. C. angulata. External lateral view of adult RV, OS 9254, sample CR 45, L: 1-137 mm.

Fig. 12. C. compressa. External lateral view of adult RV, OS 9255, sample CR 36, L: 0-925 mm.

Figs. 13-14. C. neglecta. 13, internal lateral view of adult RV, OS 9256, sample CR 13, L: 1-00 mm. 14,
external lateral view of adult LV, OS 9257, sample CR 13, L: 1-10 mm.

Figs. 15-18. C. levanderi. 15, internal lateral view of adult RV, OS 9258, sample CR 33, L: 0-91 mm.
16, internal lateral view of adult RV, OS 9259, sample CR 6, L: 0-78 mm. 17, internal lateral view of
adult LV, OS 9260, sample CR 6, L: 0-90 mm. 18, external lateral view of adult LV, OS 9261, sample
CR 33, L: 1-05 mm.

RV : right valve; LV : left valve; L: length.

PLATE 32

DE DECKKER, Pleistocene ostracods

300

PALAEONTOLOGY, VOLUME 22

Candona Candida (O. F. Muller, 1776)

Plate 32, fig. 6

1776 Cypris Candida O. F. Muller, 1776, p. 198.

Material. 1 L, 1 R.

Measurements. Length: 0-95 mm, height: 0-55 mm.

Ecology. Typical cold water form, resistant to slight changes of temperature ( = stenothermal) ; occurs in
most types of water-bodies and springs, even in saline waters. Often found in marshy vegetation.

Geographical distribution. Recent : holarctic regions.

Stratigraphical range. At least Middle Pleistocene to Recent.

Remarks. Length- width ratio of this species smaller than for Candona angulata. Right
valve in adults, without pointed posterior end.

Candona compressa (Koch, 1837)

Plate 32, fig. 12

1837 Cypris compressa Koch, no. 16.

Material. 2 L, 2 R.

Measurements. Length: 0-925 mm, height: 0-55 mm.

Ecology. Found at present on sandy edges of lakes. Commonly found in Pleistocene travertine sediments,
and sometimes in slightly saline lakes.

Geographical distribution. Recent: North and East Europe, Siberia.

Stratigraphical range. Pleistocene to Recent.

EXPLANATION OF PLATE 33

Figs. 1-5, 7-10, 13. Paralimnocythere compressa. 1, external lateral view of adult female RV, OS 9262,
sample CR 10, L: 0-55 mm. 2, external lateral view of adult female RV, OS 9263, sample CR 9, L:
0-537 mm. 3, external lateral view of adult female RV, OS 9264, sample CR 13, L: 0-525 mm. 4, internal
lateral view of adult female RV, OS 9265, sample CR 11, L: 0-55 mm. 5, external lateral view of adult
female LV, OS 9266, sample CR 9, L: 0-512 mm. 7, external lateral view of adult male RV, OS 9267,
sample CR 9, L : 0-50 mm. 8, external lateral view of adult male LV, OS 9268, sample CR 9, L : 0-525 mm.
9, external lateral view of adult male LV, OS 9269, sample CR 13, L: 0-55 mm. 10, internal lateral view
of adult male RV, OS 9270, sample CR 12, L: 0-525 mm. 13, dorsal view of adult RV, OS 9271, sample
CR 13, L: 0-562 mm.

Fig. 6. Potamocypris sp. External lateral view of carapace showing RV, OS 9272, sample CR 1 7, L : 0-60 mm.

Fig. 11. Darwinula sp. External lateral view of LV, OS 9273, sample CR 16, L: 0-70 mm.

Fig. 12. P. wolfi. External lateral view of carapace showing LV, OS 9274, sample CR 4, L: 0-725 mm.

Fig. 14. Ilyocypris bradyi. Dorsal view of carapace of adult, OS 9275, sample CR 13, L: 0-812 mm.

Fig. 15. I. gibba. Dorsal view of carapace of adult, OS 9276, sample CR 40, L: 0-975 mm.

RV : right valve ; LV : left valve ; L : length.

PLATE 33

DE DECKKER, Pleistocene ostracods

302

PALAEONTOLOGY, VOLUME 22

Remarks. Characterized by flat dorsum inclined anteriorly, and by concave impres-
sion anterodorsally. When viewed dorsally, right valve slightly pinched at anterior.
Length-height ratio 1-8 with greatest height at § from anterior.

Candona fabaeformis (Fisher, 1851)

Plate 32, fig. 10

1851 Cypris fabaeformis Fisher, p. 146, table 3, figs. 6, 7, 9, 10.

Material. 1 C.

Measurements. Length: 0-90 mm, height: 0-425 mm.

Ecology. Usually found at present, in ponds with muddy bottoms.

Stratigraphical range. At least Middle Pleistocene to Recent.

Geographical distribution. Recent: holarctic regions (found at Prezletice (Czechoslovakia) in sediments of
Cromerian age, sensu Alsolon 1974).

Remarks. Characterized by the strong overlapping of left valve over right in flattened
posterodorsal area of shell.

Candona levanderi Hirshman, 1912

Plate 32, figs. 15-18

1912 Candona levanderi Hirshman, p. 13, table 1, figs. 1-15.

Material. 2 L, 2 R.

Measurements. Length: 0-78-1-05 mm, height: 0-475-0-50 mm.

Ecology. Spring-form ostracod.

Geographical distribution. At present in Finland, Bulgaria, north-east Germany, and the Alps (from Diebel
and Pietrzeniuk 1969).

Stratigraphical range. Pleistocene to Recent.

Remarks. Shape of shell highly variable (Diebel and Pietrzeniuk 1969). Greatest height
at | from anterior, with posterior part tapering at steep angle from dorsum. Inner
lamella broad anteriorly and posteriorly, very narrow posterodorsally.

Candona neglecta Sars, 1887

Plate 32, figs. 13-14

1887 Candona neglecta Sars, p. 107, table 15, figs. 5-7, table 19.

Material. 1 L, 1 R.

Measurements. Length: 1-00-1-10 mm, height: 0-567-0-575 mm.

Ecology. Found in all kinds of water-bodies today (often in marshy vegetation) including saline waters.
Recorded in water with a temperature range between 5 and 8 °C (Aim 1916).

Geographical distribution. Recent : Europe, central Asia, North Africa (= Palaeoarctic).

Stratigraphical range. Danian (Hanganu 1977), Pleistocene to Recent.

DE DECKKER: FRESHWATER OSTRACODS

303

Remarks. Adult males much larger than females ; in lateral view shape of each valve
characterized by almost circular posterior area. Females flatter, with inner lamella
broader anteriorly; length-height ratio of 2. Dorsum slightly concave.

Candona parallela G. W. Muller, 1900

Plate 32, figs. 7-9; Plate 34, figs. 5-10, 13
1900 Candona parallela G. W. Muller, p. 25, table 5, figs. 5, 6, 23-25.

Material. Adults: 3 L, 5 R, 2 C; juveniles: 61 L, 76 R, 19 C.

Measurements. Adults: length: 0-80-0-825 mm, height: 0-467-0-475 mm.

Description. Juveniles characterized by flattened trapezoidal shape in lateral view, with valves nearly always
reticulate. Left valve overlaps right, and is taller in dorsal area. Externally, reticulation almost absent in
central muscle scars area. Adult forms similar in lateral outline to C. compressa but with slightly convex
dorsum. Reticulation absent in adults.

Ecology. Spring and early summer form inhabiting high range of temperatures from 3-23-5 °C (Aim 1916).
Commonly found in small water-bodies (sometimes subject to drying up) rich in plants; also in small
streams close to springs filled with vegetal matter (Diebel and Pietrzeniuk 1975a). At West Runton, C. paral-
lela lived in abundant fen and reed swamp ; an increase in percentage of these aquatic plants is accompanied
by an increase in numbers of C. parallela. This species occurs in climates ranging from Open ( Betula ) Forest
to Boreal Forest.

Geographical distribution. Recent : Europe, North America, and Eurasia.

Stratigraphical range. At least Middle Pleistocene to Recent.

Remarks. Juveniles of C. parallela are almost identical in shape and size to juvenile
specimens of C. fer tills fer tills illustrated by Triebel (1963, pi. 28), which are also faintly
reticulate. Triebel’s adult specimens have an anteriorly inclined dorsum.

Candona sp. 1

Plate 34, figs. 1,2, 4, 11, 12

Material. 22 L, 20 R, 22 C.

Measurements. Sizes range for these juveniles from: length: 0-38-0-74 mm, height: 0-26-0-39 mm.

Description. All specimens of Candona sp. 1, collected at West Runton are juveniles. In lateral view, shape
of a flat ellipsoid, except in the slightly concave ventral area. Anterior rounded, posterior more pointed,
especially in youngest juvenile stages. Left valve overlaps right by about 0-03 mm. In dorsal view, carapace
narrow with both sides slightly curved except for pointed anterior and posterior ends. Greatest height at
about middle of carapace. External surface of shell smooth. Central muscle scars consisting of four,
occasionally five scars in front and two in back. Two mandibular scars, well separated from one another,
with posterior one situated almost below the four vertical ones.

Ecology. At West Runton, Candona sp. 1 is found mostly with C. parallela , suggesting similar ecological
requirements for both species. At the transition, Candona sp. 1, is the most abundant species where fen and
reed swamp vegetation was very rich.

Candona sp. juveniles

Many juvenile valves and fragments were found in samples from West Runton, but
these could not be identified at species level. They were particularly abundant in four

304

PALAEONTOLOGY, VOLUME 22

of the six samples from the late Beestonian A pollen assemblage zone, typical of a
cold climate.

cyclocypris Brady and Norman, 1889
Cyclocypris laevis (O. F. Muller, 1776)

Plate 35, figs. 1-4

1776 Cypris laevis O. F. Muller, p. 198.

Material. 31 L, 27 R, 27 C.

Measurements. Length: 0-475-0-525 mm, height: 0-325-0-375 mm.

Ecology. Adaptable to any type of aquatic environment, but, usually found in small water-bodies with
marshy vegetation, and also in water springs. Found at most times of the year.

Geographical distribution. Recent : holarctic regions.

Stratigraphical range. Danian (Flanganu 1977), Pleistocene to Recent.

Remarks. Almost egg-shaped in dorsal view with right valve slightly smaller anteriorly
and posteriorly. In lateral view, dorsum semicircular. Externally, smooth valves
except for a few pore canals.

Cyclocypris ovum (Jurine, 1820)

Plate 35, figs. 5-8

1820 Monoculus ovum Jurine, p. 179, table 19, figs. 18, 19.

Material. 33 L, 26 R, 52 C.

Measurements. Length: 0-462-0-482 mm, height: 0-30-0-367 mm.

Ecology. Today found in water-bodies of all kinds ; able to withstand many environmental changes (tempera-
ture, salinity). Found at most times of the year. C. ovum is the most common Cyclocypris species at West
Runton ; found where aquatics were abundant and where water-body is shallow.

Geographical distribution. Recent: holarctic regions.

Stratigraphical range. Miocene to Recent.

EXPLANATION OF PLATE 34

Figs. 1, 2, 4, 11, 12. Candona sp. 1. 1, external lateral view of juvenile RV, OS 9277, sample CR 27, L:
0-575 mm. 2, external lateral view of juvenile LV, OS 9278, sample CR 27, L: 0-567 mm. 4, internal
lateral view of juvenile RV, OS 9279, sample CR 27, L : 0-55 mm. 1 1, dorsal view of carapace of juvenile,
OS 9280, sample CR 27, L: 0-515 mm. 12, ventral view of carapace of juvenile (anterior facing down-
ward), OS 9281, sample CR 27, L: 0-60 mm.

Fig. 3. Candona sp. 1 ? External lateral view of carapace of juvenile to show RV, OS 9282, sample CR 27,
L: 0-725 mm.

Figs. 5-10, 13. C. parallela. 5, external lateral view of juvenile RV, OS 9283, sample CR 12, L: 0-60 mm.
6, external lateral view of juvenile carapace to show RV, OS 9284, sample CR 29, L : 0-537 mm. 7, external
lateral view of juvenile LV, OS 9285, sample CR 14, L: 0-687 mm. 8, external lateral view of juvenile
LV, OS 9286, sample CR 29, L : 0-65 mm. 9, internal lateral view of juvenile RV, OS 9287, sample CR 29,
L: 0-525 mm. 10, internal lateral view of juvenile RV, OS 9288, sample CR 14, L: 0-625 mm. 13, dorsal
view of juvenile carapace, OS 9289, sample CR 29, L: 0-537 mm.

RV : right valve ; LV : left valve ; L : length.

PLATE 34

DE DECKKER, Pleistocene ostracods

306

PALAEONTOLOGY, VOLUME 22

Remarks. Differentiated from C. laevis by being much narrower in dorsal view;
greatest width at about middle of carapace ; right valve slightly larger, overlapping
left one ventrally. In lateral view, slightly less circular than C. laevis with both antero-
ventral and posteroventral areas flatter. Carapace slightly pitted externally.

Cyclocypris serena (Koch, 1837)

Plate 35, figs. 12-13

1837 Cypris serena Koch, no. 22.

Material. 4 L, 2 R, 3 C.

Measurements. Length: 0-475-0-50 mm, height: 0-30-0-337 mm.

Ecology. At present found in still waters and temporary ponds ; perhaps cold water form able to withstand
slight changes of temperature. Found at most times of the year.

Geographical distribution. Recent: North and Central Europe, North America.

Stratigraphical range. At least Middle Pleistocene to Recent. Absolon (1976) stated that this species is
typically Holocene but had already appeared during the Pleistocene.

Remarks. Larger than C. ovum and C. laevis ; flatter when viewed laterally.

cypria Zenker, 1854
Cypria sp.

Remarks. Only one specimen found at West Runton (CR 46).

scottia Brady and Norman, 1889
Scottia browniana (Jones, 1850)

Plate 31, figs. 1-14; Plate 32, figs. 1-5
1850 Cypris browniana Jones, p. 25, table 3, fig. 1 a-d.

Material. 305 L, 299 R, 26 C.

Measurements. See text-fig. 2.

Description. Characterized by trapezoidal form, almost horizontal dorsum and thick shell ; two bosses on
right valve dorsally in juvenile stages, especially in stages A-2 and A-3 (see PI. 3 1 , figs. 3-4 ; PL 32, figs. 1 -3).
Usually anterodorsal boss present on right valves in stage A-l. In dorsal view carapace almost ellipsoid in
shape. Left valve slightly longer in posterior area, overlapping ventrally and anterodorsally ; right valve
slightly longer anteriorly, overlapping posterodorsally. In juvenile stages, dorsal overlapping of right valve
over left accentuated by two bosses. External surface of shell smooth in adults; reticulate in most juvenile
specimens. Internally, many scattered pore canals ; inner lamella thick and broader anteroventrally. Ventral
area slightly concave. Central muscle scars (see PI. 31, fig. 10).

Geographical distribution and stratigraphical range. Lower Pliocene (Romania (Hanganu 1977) see below),
and Lower and Middle Pleistocene (see Kempf 1971, fig. 5). For size distributions of S', browniana at West
Runton and at other localities see text-fig. 2. S. browniana appears identical to Hanganu’s description of
Cypria bonei from the Upper Danian of Romania (paratypes examined) except the latter is of smaller size
(see text-fig. 2). The stratigraphical range of this species is therefore extended from Lower Pliocene to
Middle Pleistocene. Hanganu’s specimens are found in sandy sediments, a similar environment to that of
S. browniana at West Runton, Clacton-on-Sea, and Sugworth.

DE DECKKER: FRESHWATER OSTRACODS

307

um >||XH9I3H

3

text-fig. 2. Length-width measurements made on valves of Scottia browniana and S. tumida recovered from the West Runton section, and

compared with Scottia specimens from other localities.

308

PALAEONTOLOGY, VOLUME 22

Ecology. In most localities from which S. browniana is known, the predominant sediment is fine sand
(calcareous fine sand and gitja in Kempf 1971). From the present study, it appears that this species could
live in varying types of climate, from those associated with Open ( Betula ) Forest to a warm Temperate
Forest. It was most abundant under the Temperate Forest climate (where sandy substrate occurred) and
numbers decreased drastically in the warm Temperate climate where vegetal matter proliferated (=
? eutrophication of the water body). When S. browniana was extremely abundant, very few specimens of
other ostracod species were present. Because of its thick shell and tendency to be found within sandy sedi-
ment, it is suggested that S. browniana was either a benthic or an epibenthic species.

Remarks. Cypria candonaeformis (Schweyer, 1949) described and illustrated by
Negadaev-Nikonov (1971, tables 9 (10) and 11 (4)) is identical to S. browniana. This
was confirmed by examining topotype material held in the Senckenberg Museum.
The Moldavian specimens are larger than those found at West Runton, and have
a greater length-width ratio of the carapace. Note that at Tiraspol C. candonaeformis
occurs within sandy sediment.

Scottia tumida (Jones, 1850)

Plate 31, fig. 15

1850 Cypris tumida Jones, p. 26, pi. 3, fig. 2 a-c.

Material. 7 valves.

Measurements. See text-fig. 2.

Description. Shape almost globular with curved dorsum; ventral area nearly flat. Length-height ratio
smaller than for S', browniana (see text-fig. 2), and the adults of S. tumida are also smaller in size.

Ecology. As for S. browniana.

Geographical distribution and stratigraphical distribution. See Kempf 1971, fig. 5.

EXPLANATION OF PLATE 35

Figs. 1-4. Cyclocypris laevis. 1, dorsal view of carapace of adult, OS 9290, sample CR 14, L: 0-525 mm.
2, ventral view of carapace of adult, OS 929 1 , sample CR 14, L : 0-50 mm. 3, external lateral view of adult
LV, OS 9292, sample CR 14, L : 0-475 mm. 4, internal lateral view of adult RV, OS 9293, sample CR 14,
L: 0-50 mm.

Figs. 5-8. C. ovum. 5, dorsal view of carapace of adult, OS 9294, sample CR 9, L : 0-475 mm. 6, ventral
view of carapace of adult, OS 9295, sample CR 9, L : 0-487 mm. 7, external lateral view of adult LV,
OS 9296, sample CR 9, L: 0-475 mm. 8, internal lateral view of adult LV, OS 9297, sample CR 9,
L: 0-462 mm.

Fig. 9. Eucypris dulcifons. External lateral view of adult RV, OS 9298, sample CR 11, L: 0-90 mm.

Figs. 10, 11. Cypridopsis vidua. 10, dorsal view of carapace of adult, OS 9299, sample CR 10, L: 0-65 mm.

11, internal lateral view of adult LV, OS 9300, sample CR 14, L: 0-725 mm.

Figs. 12, 13. Cyclocypris serena. 12, external lateral view of carapace of adult to show LV, OS 9301, sample
CR 36, L: 0-475 mm. 12, dorsal view of carapace of adult, OS 9302, sample CR 36, L: 0-50 mm.

RV : right valve ; LV : left valve ; L : length.

PLATE 35

DE DECKKER, Pleistocene ostracods

310

PALAEONTOLOGY, VOLUME 22

eucypris Vavra, 1891

Eucypris dulcifons Diebel and Pietrzeniuk, 1969

Plate 35, fig. 9

1969 Eucypris dulcifons Diebel and Pietrzeniuk, p. 479, fig. 9, table 9, figs. 5-8.

Material. 1 R in sample CR 1 1 .

Measurements. Length: 0-90 mm, height: 0-55 mm.

Ecology. Described by Diebel and Pietrzeniuk (1975ft) from lake sediments and also small water bodies,
and by Robinson (in press) from sediments deposited at Sugworth in sluggishly flowing waters, rich in
growing vegetation or accumulations of plant debris.

Geographical distribution. Middle Pleistocene: Germany and England; Upper Pleistocene: Germany.
Stratigraphical range. At least Middle to Upper Pleistocene.

Remarks. Characterized by finger print reticulation on outside of the carapace
(Robinson, in press). E. dulcifons in dorsal view is much narrower than E. pigra
(Fisher, 1851), but otherwise very similar in outline.

herpetocypris Brady and Norman, 1889
Herpetocypris sp.

Remarks. Only one juvenile specimen was found (sample CR 6).

cypridopsis Brady, 1 868
Cypridopsis vidua (O. F. Muller, 1776)

Plate 35, figs. 10-11

1776 Cypris vidua O. F. Muller, p. 198.

Material. 4 L, 1 R, 4 C.

Measurements. Length: 0-65-0-725 mm, height: 0-425-0-45 mm.

Ecology. Found in small and large water-bodies rich in aquatic vegetation; common in entrophic and
usually warm waters.

Geographical distribution. Pleistocene: Kashmir, Germany, USA, England (Swain 1976). Recent: cos-
mopolitan.

Stratigraphical range. Late Pliocene (Swain, 1976) to Recent.

Remarks. Characterized by its large size compared with the three Cyclocypris species
found at West Runton; strongly pitted surface, pointed dorsum and sinuous hinge in
dorsal view ; left valve overlaps dorsally at about j from anterior.

DE DECKKER: FRESHWATER OSTRACODS

311

potamocypris Brady, 1870
Potamocypris wolfi Brehm, 1920

Plate 33, fig. 12

1920 Potamocypris wolfi Brehm, p. 6.

Material. 1 C.

Measurements. Length: 0-725 mm, height: 0-36 mm.

Ecology. Typical of cold water springs.

Geographical distribution. Recent : Europe.

Stratigraphical range. At least Middle Pleistocene to Recent.

Remarks. Characterized by length-height ratio 1-95-2-10. Both valves reticulate;
right valve overlaps left, especially in dorsal area.

Potamocypris sp.

Plate 33, fig. 6

Material. 1 C (CR 17).

Measurements. Length: 0-60 mm, height: 0-375 mm.

Ecology. Found in sandy sediment rich in vegetal debris; pollen assemblage indicates Open ( Betula ) Forest
type of climate and shallow water body.

Stratigraphical range. At least Middle Pleistocene.

Remarks. Characterized by a length-height ratio of 1 -6, with dense and deep reticula-
tion on the outside of the carapace; left valve overlaps the right, especially in the
dorsal area.

Note. At West Runton, fragments of large cypridid ostracods were found, but were unidentifiable even at
generic level. At the same locality, Dr. J. E. Robinson found three additional species of ostracods, namely
Candona tricicatricosa, Limnocythere falcata, and Sclerocypris ? clavata prisca. These were recovered in the
silty beds below the detritus peat, and are believed to correspond approximately with samples CR 8-10
in my section (Robinson, per litt. 28.10.77).

OSTRACOD ASSEMBLAGES

The distribution of the ostracods found in each of the sixty-two samples is represented
in text-fig. 1. This chart shows the five main zones based on ostracod assemblages,
which correspond almost exactly with West’s pollen assemblage zones (West, in
press and Funnell and West 1977) and with the lithological changes. These five
assemblages are recorded below, in a stratigraphical sequence.

For the ecological notes on the ostracods, refer to the taxonomy section of this
paper.

Ostracod assemblage 1 : late Beestonian A pollen assemblage zone ( 0-26 cm)

Pollen assemblage and local macroscopic vegetation. Gramineae-Cyperaceae-
Artemisia with plants typical of shallow water and muddy substrates (93%). Climate

312

PALAEONTOLOGY, VOLUME 22

postulated: cold. Herbaceous vegetation existed and charophyte gyrogonites were
abundant (from 30 to 60 per 40 gr. of sediment) inferring a still and shallow water-body.
Environment postulated. Small pond(s) and/or intermittent meandering streams con-
sisting of cold water under a cold climate ; plants of shallow muddy substrates present.
Intermittent occurrence of ostracod species which are few in number, indicating the
water body was shallow and often subject to drying up, or was covered by coarser
river sediments as shown by occasional black cherty pebbles. The water could have
been slightly brackish or stagnant at stages (shown by Candona Candida CR 3,
C. compressa CR 6, Ilyocypris bradyi CR 3-6). Note the absence of Scottia browniana.

Ostracod assemblage 2: late Beestonian B pollen assemblage zone ( 26-55 cm)

Pollen assemblage and local macroscopic vegetation. Gramineae-Cyperaceae-item/a
as well as plants typical of muddy substrates and shallow water, decreasing rapidly
towards top (between CR 9-14 fen and reed swamp with approximately 75% halo-
phytes). The climate was typical of an open ( Betula ) forest. Charophyte gyrogonites
present, but less abundant than before.

Environment postulated. Small lake, pond, or swamp, becoming deeper but not con-
nected to a river (no pebbles brought in). Possible episode of brackish water shown by
ostracods (C. angulata CR 9-10, C. compressa CR 11, C. neglecta CR 11, I. bradyi
CR 13-14), between samples CR 9-14 (? proximity to Sea or high rate of evaporation),
when the halophytes were also abundant. Ostracods seem to indicate the presence of
marshy vegetation in places, as well as possible eutrophication of the water (shown by
Cypridopsis vidua ) at some levels. Note that charophytes increase in number when
water was probably slightly saline. Water temperature uncertain.

Ostracod assemblage 3: Cromerian I A and IB pollen assemblage zones ( 55-87-5 cm)

Pollen assemblage and local macroscopic vegetation. Cromerian IA: Betula- Pinus \
plants of shallow water and muddy substrates almost absent; plants typical of fen
and reed swamp nearly 30%. Climate postulated is a boreal forest type. Cromerian
IB: Pinus- Ulmus, few plants of shallow water and muddy substrates (about 20% at
one stage) but aquatics abundant with high percentage of fen and reed swamp halo-
phytes. Climate postulated, is a transition between boreal and temperate forest types.

Environment postulated. Climate warming up towards the top as shown by palyno-
logical data. Fen and reed swamp type of vegetation abundant (most abundant at
transition Cromerian I A IB, the time during which Candona sp. 1 is the most abundant
in the section studied, and there is a high number of C. parallela). Ostracods give no
indication of the depth of the water-body.

Note the two pollen assemblages (Cromerian IA and IB) are here included in the
one ostracod assemblage.

Ostracod assemblage 4 : Cromerian II A pollen assemblage zone ( 102-122-5 cm)

Pollen assemblage and local macroscopic vegetation. Pinus- Quercus-Ulmus with
aquatics fairly abundant (45%) and fen and reed swamp plants up to 30%. Charo-
phyte gyrogonites present in all samples (except CR 38) indicating a probable decrease
in water level.

DE DECKKER: FRESHWATER OSTRACODS

313

Other data. Vertebrate fragments occasional, fish scales present in most samples.

Environment postulated. Temperate forest type of climate. Small shallow lake with
fairly abundant aquatics, and a fen and reed swamp vegetation. At level CR 45, the
water could have been slightly saline as there is an increase in charophyte gyrogonites
in that sample. Ostracods offer no data on water temperature except for Ilyocypris
gibba (CR 40, 42, 43, 46) which can be found in waters up to 19 °C. Supply of sand and
vegetal debris very high.

Note that Scottia browniana is extremely abundant.

Ostracod assemblage 5 : Cromerian IIB pollen assemblage zone ( 122-5-140 cm)

Pollen assemblage. Quercus-Ulmus-Tilia ; no data on macroscopic plants, except
that some charophyte gyrogonites are present.

Other data. Vertebrate fragments and fish scales common.

Environment postulated. Warm temperate forest type of climate. Sandy substrate,
covered with vegetal debris, possibly preventing S. browniana occurring in large
numbers. The presence of charophytes suggest that the water was shallow. Vertebrates
must have frequented the surroundings of the water body (see Stuart 1975).

Above assemblage 5

140-15.7-5 cm: plant debris in dark brown sand. 157-5-161-5 cm: reddish brown to
black leached vegetal debris with less sand.

Apart from a few specimens of S. browniana recovered from samples CR 54-56,
and two specimens of S. tumida (CR 54), ostracods are absent from this uppermost
horizon. This absence is probably associated with leaching of the ostracod valves.
Molluscan shells are also absent.

COMPARISON WITH OTHER MIDDLE PLEISTOCENE DEPOSITS IN EUROPE
YIELDING OSTRACODS

The only deposit known in the British Isles, apart from the sites around West Runton,
yielding ostracods of a similar stratigraphical horizon is at Sugworth, near Oxford.
The fauna was described by Robinson (in press) from a deposit formed by a sluggish
stream with a strong tendency to stagnation and vegetation overgrowth, broken
by intermittent flooding. This description corresponds well with the deposit at
West Runton. The Sugworth ostracod fauna is reminiscent of that at West Runton
with abundant S. browniana. However, at Sugworth Metacypris cordata (a species
typical of lakes choked with vegetation) is present. At Clacton-on-Sea, an ostracod
fauna of Hoxnian age yielding abundant S. browniana was found in sandy sediment.

In Europe and the USSR, six Middle Pleistocene deposits appear to be analogous
to the one studied here. At Siissenborn (Elster in age) Diebel and Pietrzeniuk (1969)
described a fauna with Paralimnocy there compressa, similar to that at West Runton.
However S. browniana is absent, probably because the substrate was not sandy. At
Prezletice, north of Prague, a Cromerian deposit (94 cm thick), was described by
Absolon (1974), who postulated that the sediment was probably deposited on a

314

PALAEONTOLOGY, VOLUME 22

flood-plain and that the water had a salinity level of about 3°/00- The ostracod fauna
was part of a different biotope from that at West Runton even though I. gibba,
Cyclocypris laevis, and Candona compressa were present throughout the entire deposit.
At Tiraspol in Moldavia, Negadaev-Nikonov (1971) described an ostracod fauna of
Cromerian age from fluviatile sediments. Twenty-five ostracod species were recorded
from a sequence 14-6 m thick. Most of the species found at West Runton are also
represented there including S. browniana ( = Cypria candonaeformis ) and P. compressa
(= P. ?cf. compressa ). Kazmina (1975) described freshwater ostracods from the West
Siberian Platform and showed illustrations of Cypria candonaeformis, which from
descriptions and measurements, appears to be synonymous with S. browniana.
Kazmina’s specimens are middle to late Pliocene in age. At Voigstedt, Diebel (1965)
recovered only five ostracod species from four different horizons, the reason being that
they form part of a reworked fauna transported in running water. The ostracods give
no indication of the age of the deposit known to be Cromerian. Two other deposits,
both Holstein in age, yield ostracod faunas strongly resembling that at West Runton.
One occurs at Syrniki in Poland, where Diebel (1961) described fifteen ostracods,
S. tumida being the most abundant species (forming about 50% of the fauna). One major
difference from the West Runton fauna is the presence of Metacypris cordata. The
other deposit, described by Kempf (1967), occurs at Tonisberg, in the lower Rhine
district in West Germany where the ostracod fauna is similar to that at West Runton.
The presence of Metacypris cordata, Limnocy there ' sanctipatricii, and Cytherissa
lacustris, though small in number, suggests that the body of water could have been
more extensive than that at West Runton.

REWORKED FAUNA

The only foraminifera found were derived from Upper Cretaceous and Lower
Pleistocene sediments. These were probably contained within the chalk fragments
which are present in parts of the section. These fragments, which were brought into
the area at West Runton, tend to confirm the idea that intermittent streams were
occasionally connected to the water-body where the ostracods were living. The
presence of a number of broken ostracod valves (especially candonid ones) at the
bottom of the section, as well as occasional fragments of large cypridids, suggests
that some ostracods could have been transported. This cannot be substantiated,
especially as the type of water body postulated above, would favour partial breakage
of ostracod valves by incoming intermittent streams. These were not continuous, as
some very fragile ostracod valves (e.g. P. compressa) were found intact, as were some
ostracod carapaces with both valves still holding together. Partial reworking would
have either destroyed or separated the valves. The large cypridid ostracods after
death, were often attacked by micro-organisms which render the shell very thin and
extremely fragile in places. Slight undercurrents, as for example those created by wind
over the water body, could cause the breakage of large fragile ostracod valves.

DE DECKKER: FRESHWATER OSTRACODS

315

AGE OF THE DEPOSIT

Unlike the pollen, the ostracod assemblages cannot provide us with accurate dating.
The only valuable information is provided by S. browniana and S. tumida which are
absent from the fossil record in late Pleistocene and Holocene sediments. All the other
ostracods found at West Runt on are still living today.

CONCLUSIONS

The pollen assemblage data infers an improvement in the climate from the bottom
of the section towards the top (from cold to warm temperate). Similarly, the ostracod
faunal assemblages partly show a change from cold to warm water, but they are also
more controlled by local factors. Sedimentological changes, perhaps associated with
climatic changes, seem to directly affect the composition of the ostracod assemblages.
At West Runton, the deposit studied, represents above the gravel layer at the bottom,
a sedimentary sequence deposited in a large channel. With the information provided
by the palaeobotanical and ostracod data, as well as a study of the sediments, the
following environment is inferred. Initially small ponds and/or intermittent streams
meandering under a cold climate occurred at West Runton, where aquatic plants
were abundant and ostracods very few but diversified. Water could have been slightly
saline at the time. Later, with improvement of the climate (open ( Betula ) forest)
ostracods became more abundant amongst typical fen and reed swamp vegetation.
The water body could have been enlarged and deepened and the water could have
become saline at some stage. For most of the time, fen and reed swamp vegetation
prevailed during the boreal forest climate and the transition to a temperate forest
type of climate. At that time ostracod species were less abundant. The environment
was still that of a swamp with sluggishly moving water bringing in sandy sediments
and perhaps fragments of clay. A zone of mixing of sediments and faunas, probably
signifying a period of drying up of the water body and perhaps accompanied by
erosion, preceded a period typical of a temperate climate during which the shallowing
swamp was covered by vegetation, and sandy material was brought into the swamp
by flowing water. One species of ostracod S. browniana , was extremely abundant.
Vertebrates (fish, rodents, and small mammals (Stuart 1975)) frequented the swamp
and its surroundings. A small change in salinity of the water could have occurred.
Finally the ostracod fauna decreased rather rapidly under the warm temperate forest
type of climate where vegetal debris was profuse and covered the still shallowing
swamp, in which the remains of many molluscs are found.

Acknowledgements. I am much indebted to Professor R. G. West who suggested this project and who read
and offered comments on the manuscript. My understanding of the ostracod fauna at West Runton gained
much from discussions with Dr. J. E. Robinson. I also wish to thank Dr. A. R. Lord for identifying the
foraminifera which I recovered in my samples. My gratitude goes to Ms L. M. Sheppard who kindly took
the SEM photographs on the Cambridge Instruments Stereoscan II at the British Museum.

316

PALAEONTOLOGY, VOLUME 22
REFERENCES

absolon, a. 1974. Cromerian ostracode fauna from Prezletice and its palaeoecological interpretation. Vest.
Ustred. Ust. geol. 49, 41-47.

— 1976. Neue Daten zur Evolution der Siisswasserostracoden im Pleistozan. Abh. Verh. naturw. Ver.
Hamburg. N.F. 18/19 (suppl.), 229-237.

alm, G. 1916. Monographie der schwedischen Stisswasser-Ostracoden. Zool. Bidr. Upps. 4, 1-248.
diebel, K. 1961. Ostracoden des Paludinenbank-Interglazials von Syrniki am Wieprz (Polen). Geologie 10,
533-545.

— 1965. Bemerkungen zur Ostracoden fauna von Voigstedt in Thiiringen. Palaont. Abh. 2, 293-297.

— and pietrzeniuk, E. 1969. Ostracoden aus dem Mittelpleistozan von Sussenborn bei Weimar. Ibid.
A, 3, 463-488.

— 1975a. Ostracoden aus dem holozanen Travertin von Bad Langensalza. Quartarpalaontologie,
1, 27-55.

— 1975 b. Mittel- und jungpleistozane Ostracodenfaunen des Raums Postdam-Brandenburg-
Rathenow in Stratigraphisher und Okologischer Sicht. Z. Geol. Wiss. Themenheft. 3, A, 1197-1233.

— 1977. Ostracoden aus dem Travertin von Taubach bei Weimar. Quartarpalaontologie, 2, 1 19-137.

— and wolfschlager, h. 1975. Ostracoden aus dem jungpleistozanen Travertin von Ehringsdorf bei
Weimar. Abh. zent. geol. Inst. (Berk), 23, 91-136.

funnell, b. m. and west, r. G. 1977. Preglacial Pleistocene deposits of East Anglia. In shotton, f. w. (ed.).

British Quaternary Studies— Recent Advances, 247-265, Oxford University Press.
hanganu, E. 1976. Nouvelles especes de Cyprididae dans le Dacien Superieur de la Muntenne Orientale
(Roumanie). Bull. Soc. beige Geol. Paleont. Hydrol. 85, 51-61.
hartmann, G. and puri, H. s. 1974. Summary of Neontological and Paleontological classification of
Ostracoda. Mitt. hamb. zool. Mus. Inst. 70, 7-73.

kazmina, t. a. 1975. Stratigrafia i Ostrakody Pleistosena i Rannego Pleistosena Ionga Zapadnago-Sikirskoi
Ravniny. Trudy Inst. Geol. Geofis Sib. Otd. 264a, 1-108. [In Russian.]
kempf, E. K. 1967. Ostracoden aus dem Holstein-Interglazial von Tonisberg (Niederrheingebiet), Mber.
Dt. Akad. Wiss. Bert. 9, 119-139.

— 1971. Okologie, Taxonomie und Verbreitung der nichtmarinen Ostrakoden-Gattung Scottia im
Quartar von Europa. Eiszeitalter Gegenw. 22, 46-63.

klie, w. 1938. Krebstiere oder Crustacea III: Ostracoda, Muschelkrebse. Tierwelt. Dtl. 34, 1-230.
negadaev-nikonov, K. n. 1971. Crustacea-Rakoobraznye. In Plejstocen Tiraspolia, 55-71. Kishinev. [In
Russian.]

robinson, J. E. in press. The ostracods from Sugworth.

Stuart, a. j. 1975. The vertebrate fauna of the type Cromerian. Boreas, 4, 63-76.
swain, f. m. 1976. Evolutionary Development of Cypridopsid Ostracoda. Abh. Verh. naturw. Ver. Hamburg.
N.F. 18/19, (suppl.), 103-118.

triebel, E. 1963. Ostracoden aus dem Sannois und jungeren Schichten des Mainzer Beckens : 1 . Cyprididae.
Senckenberg. leth. 44, 157-207.

west, r. G. in press. The preglacial Pleistocene of the Norfolk and Suffolk coasts. Cambridge University
Press.

P. DE deckker
Zoology Department
University of Adelaide

Typescript received 14 December 1977 G.P.O. Box 498, Adelaide

Revised typescript received 10 June 1978 South Australia 5001

FUNCTIONAL MORPHOLOGY AND
ONTOGENETIC VARIATION IN THE
CALLOVIAN BRACHIOPOD SEPTIRHYNCHIA
FROM TUNISIA

by MIGUEL O. MANCENIDO and CHRISTOPHER D. WALLEY

Abstract. The brachiopod Septirhynchia numidiensis sp. nov. is described from the Callovian (Jurassic) of southern
Tunisia. Studies of the internal characters and their development reveal that only the adult forms possess the cardinal
process and ventral median septum regarded as diagnostic of the genus. Inferred functional morphology suggests
that these adult features represent direct or indirect adaptations to living partially buried in sediment, and their
gradual development reflects a progressive change from an epifaunal to a semi-infaunal mode of life during ontogeny.
Reports in the literature of synchronous homeomorphs with other species of Septirhynchia seemingly refer to juvenile
forms. The implications for brachiopod taxonomy of the observed ontogenetic variation in Septirhynchia are discussed.
The possible affinities and origins of the genus are re-examined and it is proposed that the present monotypic family
based on Septirhynchia should be lowered to the level of a subfamily, the Septirhynchiinae within the Rhynchonellidae.
The known stratigraphical and geographical distribution of the genus is reviewed, showing an apparent restriction
to the Callovian of the south Tethyan region.

The Jurassic rhynchonellid brachiopod genus Septirhynchia (Muir-Wood 1935) has
been considered sufficiently distinct from other Mesozoic forms to be placed in
a separate family, the Septirhynchiidae (Muir-Wood and Cooper 1951). The presence
of a ventral median septum and a cardinal process in the brachial valve, in conjunction
with large size and pentameroid appearance, were features that suggested a possible
complete separation from the mainstream of Mesozoic brachiopod evolution (e.g.
Ager 1959, p. 327; Ager et al. 1972, p. 221).

Members of this genus have been reported from Callovian and Kimmeridgian
sediments of East Africa, the Middle East, and North Africa. The internal structures
of the group have been little studied, mainly because of limited material, so that
previous knowledge of internal structures rested almost entirely on a few silicified
specimens described by Muir- Wood and Cooper (1951). In that paper, however, the
occurrence in the same sediments as Septirhynchia of forms externally similar but
lacking one or both diagnostic internal structures was noted. Subsequently, Dubar
(1967) and Rousselle (1970) have reported other forms apparently homeomorphic
with Septirhynchia from the Jurassic of North Africa. The consequent uncertainty
over the validity and constancy of the internal morphology of Septirhynchia has
compounded the problem of relating this genus to other rhynchonellid groups.

In the course of fieldwork in southern Tunisia, one of us (C. D. W.) collected
nearly 100 specimens of a new species of Septirhynchia from Callovian sediments.
Detailed studies of the internal morphology of this species at different ontogenetic
stages have revealed information on the characteristics, status, and probable affinities

[Palaeontology, Vol. 22, Part 2, 1979, pp. 317-337, pi. 36.]

318

PALAEONTOLOGY, VOLUME 22

of the genus. An interpretation of these internal and external features in terms of
functional morphology has produced conclusions which have wider ranging implica-
tions for brachiopod taxonomy.

SYSTEMATIC PALAEONTOLOGY

Abbreviations. L, length; W. width; T, thickness (all measurements to the nearest mm). NVC, no. of ventral valve
costae; NDC, no. of dorsal valve costae (as counted at commissure).

Family rhynchonellidae Gray, 1848

Subfamily septirhynchiinae (Muir-Wood and Cooper, 1951 ; nom. transl. herein)
Genus septirhynchia Muir-Wood, 1935

Septirhynchia numidiensis sp. nov.

Plate 36, figs. 1-14; text-figs. 1-4

Derivation of name. After Numidia, the ancient kingdom and Roman province of northern Africa.

Type locality. 60 m below the summit of Djebel Brourmett marked by the ‘Poste Optique’, approximately
1-5 km north-west of Tatahouine, near Medenine, southern Tunisia.

Age and distribution. Confined to the lower portion of the Foum Tatahouine Limestone and Marls Forma-
tion (Busson 1967) in the area around Tatahouine. All material was collected from a single horizon, probably
of early Callovian age (Busson 1967). The presence of Erymnoceras sp. some 20 m above this bed would
tend to confirm this age.

Material. Approximately 100 specimens, some very fragmentary. Holotype and figured paratypes deposited
at the Department of Palaeontology, British Museum (Natural History), London, catalogue numbers
BB 76530-76539. Other paratypes will be housed in the Department of Geology, University College,
Swansea, and Service geologique de Tunisie.

Diagnosis. Septirhynchia with subtriangular outline, adult size large for genus.
Typically with a strong median sinus on the ventral valve. The number of costae on
each valve decreases from 13 in the juvenile to 8 or 9 in the adult, the average being 10,
of which 2-3 lie in the ventral sinus. The umbo of the ventral valve is sharply pointed
and strongly incurved in the adult form.

Dimensions. The material may be divided conveniently into two groups, the smaller specimens, relatively
much thinner, with width often greater than length, and the larger, more globose forms which have thickness

EXPLANATION OF PLATE 36

Septirhynchia numidiensis sp. nov., Callovian, Djebel Brourmett (southern Tunisia). All specimens are
oriented with the pedicle valve uppermost. All x 1 .

Figs. 1-5. Holotype BB 76530 (adult). 1, posterior view; 2, dorsal view; 3, lateral view; 4, anterior view;
5, ventral view.

Figs. 6, 10. Paratype BB 76534 (narrow adult). 6, posterior view; 10, lateral view.

Figs. 7, 8, 11. Paratype BB 76535 (small adult). 7, lateral view; 8, posterior view; 1 1, dorsal view.

Fig. 9. Anterior view of juvenile specimen BB 76539a.

Figs. 12-14. Paratype BB 76532 (juvenile). 12, dorsal view; 13, posterior view; 14, lateral view.

PLATE 36

MANCENIDO and WALLEY, Septirhynchia numidiensis

320

PALAEONTOLOGY, VOLUME 22

as their greatest measurement with length being greater than width. There is, however, a continuous grada-
tion between these two groups which are interpreted as being the juvenile and adult individuals respectively.
The following are typical values. Specimens BB 76532-76533 are the extremes of the size range collected.

Holotype

(BB 76530)

L = 45

W = 44;

T = 45

NVC= 11;

NDC = 10

Paratype

(BB 76531)

L = 49

W — 41 ;

T = 46

NVC= 10;

NDC = 9

(BB 76532)

L = 25

W = 26;

T= 19

NVC = 12;

NDC = 13

(BB 76533)

L = 57

W = 62;

T = 58

NVC = 9;

NDC = 8

(BB 76534)

L = 54

W = 43;

T = 57

NVC = 9;

NDC = 8

(BB 76535)

L = 36

W = 41 ;

T = 31

NVC = 10;

NDC = 9

(BB 76536)

L = 47

W = 40* ;

T = 42

NVC = 12*;

NDC = 1 1

(BB 76537)

L = 34

W = 34;

T = 26

NVC = 13;

NDC = 14

(BB 76538)

L = 36

W = 33;

T = 29

NVC = 12;

NDC = 1 1

(BB 76539a)

L = 29

W = 18*;

T = 21

NVC = 10*;

NDC = 10*

(BB 16539b)

L = 50

W = 46;

T = 49

NVC = 10*;

NDC = 9*

* Approximate, due to crushing of specimen.

Variation

in the number of ribs of the fifty-nine measurable specimens according to size classes (based

on averaging length and width)

7 8 9

1 2
1 1 2
1 1

Number of ribs (ventral valve)

10 11 12

1 1

7 7 6

6 10 4

1 2 1

L + W

13 14 2

50-59
40-49
1 30-39

2 1 20-29

Description

Exterior. Strong differences are present in external morphology between adult and juvenile forms.
There are, however, many transitional forms, and the shape of the juveniles can be recognized in the growth
lines of the adults, suggesting a simple ontogenetic series.

In juveniles the valves are moderately convex, with width approximately equal to or greater than length.
The ventral umbo is gently curved and the foramen faces dorsally. There is only a slight trace of a ventral
sinus and the anterior commissure is a simple rectimarginate type with a zig-zag serial deflexion super-
imposed. Lateral commissure nearly straight with only the slightest dorsal projection of the posterior of
the ventral valve. Juveniles have a maximum of thirteen or fourteen low, angular costae (as measured at
the commissure) and the bare planareas of the valves are small and only slightly concave.

Adult forms tend to be gibbous, with both valves strongly incurved. In particular, the umbo of the
ventral valve almost touches the dorsal valve in some adults and the minute foramen faces ventrally (PI. 36,
fig. 10). The ventral valve completely encloses the umbo of the incurved dorsal valve and the deltidial plates
are concealed. The conspicuous ventral sinus originates some distance from the apex of the pedicle valve
and deepens gradually towards the anterior commissure where it is marked by a simple uniplication com-
bined with a zig-zag serial deflexion. The lateral commissure is deflected dorsally along its posterior portion
by a dorsal protrusion of the ventral valve, which in fact projects some way inside the brachial valve form-
ing a feature comparable to the ‘squama and glotta structure’ of uncinulid brachiopods discussed by
Westbroek et al. (1975). The valves have from eight to ten coarse angular costae which may be up to 10 mm
wide, the sinus generally containing two or three costae. The decrease in number of costae from juvenile to
adult stage is due to the relative expansion of the planareas at the expense of the small lateral costae, in the
same fashion as described by previous workers in other species of Septirhynchia. The planareas of both
valves are smooth and deeply concave, giving the posterior part of the shell a characteristic ‘hour-glass’
outline in transverse section (text-fig. 2).

There is a good deal of variation in the morphology of the adult form and a few large specimens with
a wide anterior commissure and less pronounced gibbosity are known.

MANCENlDO AND WALLEY: SEPTIRHYNCHIA FROM TUNISIA

321

text-fig. 1. Enlarged transverse serial sections drawn from acetate peels of juvenile specimens of Septi-
rhynchia numidiensis. a-g, individual crushed anteriorly; original length 29 mm; this sequence extends
to 3-5 mm from the tip of the beak. 1-11, individual with a broken beak (so distance from ventral umbo not
shown); original length of specimen approx. 25 mm; this sequence 1-11 covers an interval of 5 mm from
the brachial umbo. Key a = deltidial plates, b = incipient cardinal process, c = normal septalium.

Interior. Remarkable differences have been observed between the internal structures of the sectioned
juvenile and adult individuals, although the examination of a number of other specimens has revealed
a continuous series of intermediate forms. Among the juveniles the smallest specimen examined was
posteriorly broken but originally was probably only a little over 25 mm long. The internal structure of this
type, as revealed by serial grinding, is shown in text-fig. 1 (1-11).

The delthyrial chamber is bounded by two dental lamellae which lie parallel to the plane of symmetry
in section, diverging from each other anteriorly. The dental lamellae sharply decrease in height anteriorly
and project straight into the dorsal valve forming massive, smooth teeth which are completely enclosed by
the sockets of the dorsal valve. The deltidial plates are disjunct.

322

PALAEONTOLOGY, VOLUME 22

A ventral median septum is absent in this specimen. Careful sectioning of the umbonal region of a slightly
larger (29 mm long) specimen (text-fig. 1a-g) and examination of other fragmentary forms has shown that
there is no trace of a ventral septum in juveniles of this species. In a number of what might be termed
younger adult forms the ventral septum is represented only by a low ridge. Unbroken weathered adult
specimens (and also the silicified S. pulchra and S. azaisi figured by Muir-Wood and Cooper, 1951, pi. 1,
fig. 8 ; pi. 2, figs. 2, 6) show that the ventral median septum disappears towards the umbo, confirming the
absence of this feature in the juvenile.

text-fig. 2. Transverse serial sections drawn from acetate peels of an adult specimen of Septirhynchia
numidiensis. Distance from ventral umbo given in mm. Length of specimen 39 mm. Key : a = fused deltidial
plates, b = cardinal process, c = overturned septalium.

text-fig. 3. Longitudinal serial sections drawn from acetate peels of an adult specimen of Septirhynchia
numidiensis. Measurements given are from lateral edge of specimen in mm. Length of specimen 38 mm.
Key : a = fused deltidial plates, b = recurved cardinal process, c = overturned septalium, d = raduliform

crus.

324

PALAEONTOLOGY, VOLUME 22

The brachial valve has a simple septalium [used here in the restricted sense of Childs (1969), which is
equivalent to Pearson’s (1977) ‘muscle trough’] and delicate raduliform crura. The dorsal median septum
is present in the very smallest forms examined. A pedicle collar has been observed in one or two juveniles.

Two adult forms have been transversely serial-sectioned, of which one is figured (text-fig. 2). A third
adult has been sectioned longitudinally (text-fig. 3). All the sectioned forms and some fragmentary specimens
show consistently similar internal features. The ventral valve contains a large blade-like median septum,
up to 8 mm in depth in some specimens, which extends from near the valve apex towards the anterior where
it ends suddenly, generally about midway along the valve. In specimens which have been developed out of
the matrix the septum surface can be seen to be smooth and to lack any signs of muscle attachment areas.
The dental lamellae are thin and decrease in height rapidly anteriorly, and project straight dorsally into the
brachial valve where they form massive teeth. The sockets in the dorsal valve do not enclose the teeth
antero-dorsally (text-fig. 2, 10.2-11.8). The deltidial plates are fused anteriorly to form a henidium and
a small pedicle collar can sometimes be recognized.

The dorsal valve is so incurved that the septalial cavity faces dorsally. The umbo is covered and partially
enveloped by a series of thin calcite sheets which form a complex structure on top of, and to some extent
enfolding, the septalium. This is the so-called cardinal process of Muir-Wood and Cooper (1951) and can
be clearly seen in text-fig. 3, 10.8-16.0. The main feature of this structure is the thin sheet extending back
posteriorly over the brachial umbo (see also text-fig. 2). The simple raduliform crura, triangular in cross-
section, are strongly curved towards the ventral valve and lack terminal processes. The high, thin, blade-
like dorsal median septum ends abruptly in the anterior portion of the valve. In both adults and juveniles
neither Buckman’s calcination technique nor examination of natural internal casts has yielded any reliable
evidence of muscle scars.

text-fig. 4. Selected set of acetate peels to illustrate details of internal structure, a-d, Septirhynchia
numidiensis sp. nov. from the Callovian of Djebel Brourmett (southern Tunisia); a, section of juvenile
specimen showing typical septalium and incipient cardinal process (equivalent to section 6 of text-fig. 1),
x 7-5; B, posterior portion of cardinal process showing overturned septalium (equivalent to section 12-2
of text-fig. 2), -4; c, apex of brachial valve enveloped by skeletal material from the cardinal process
(equivalent to intermediate condition between sections 12.8 and 13.7 of text-fig. 2), x 4; d, enlarged portion
of saggittal section of paratype BB 76536 (adult) showing nature of cardinal process and crura base (equi-
valent to sections 12.3-15.6 of text-fig. 3), x 3 ; e, enlarged portion of transverse section across the umbo of
a left valve of the Recent pholadacean bivalve Barnea Candida (Linnaeus) from Swansea Bay (South Wales),
showing the nature of the umbonal reflection for comparison, x 4.

MANCENIDO AND WALLEY: SEPTIRHYNCHIA FROM TUNISIA 325

Taxonomic remarks. Septirhynchia numidiensis is distinguished from previously
described species by its subtrigonal outline, the presence of a prominent ventral sinus,
and the nine to twelve, wide, angular costae. Only two other species (each described
from single specimens) are sufficiently similar to S. numidiensis to deserve more
detailed comparison. S. madashonensis Muir-Wood (1935) is based on a fragmentary
specimen with fifteen broad costae. No specimen of S. numidiensis of a similar size
is either as spherical or possesses such a deep sinus. S.pulchra Muir-Wood and Cooper
(1951) has the closest resemblance to the new species but has more costae (13), is less
trigonal in outline, and seems to have a less well-defined sinus. The only specimen
described appears, on the basis of internal and external evidence, to have attained an
ontogenetic stage equivalent to much larger forms of S. numidiensis. The material
described from other Tunisian localities by Dubar (1967) as S. pulchra more closely
resembles the new species and might be conspecific. In addition the specimen figured
by Stefanini (1932, pi. 4, fig. 9) as Rhynchonella azaisi, from what was then Italian
Somaliland does not appear to belong to Cottreau’s species but seems more closely
related to either S. numidiensis or even S. pulchra.

Palaeoecology and functional morphology

S. numidiensis occurs in a pale-grey micrite, in which it is locally abundant. There
is little associated fauna other than rare, poorly preserved regular echinoids and
nerineid gastropods. Some slight transportation seems to have taken place in that
a number of the larger specimens are disarticulated and there is an absence of forms
below 20 mm in length. A number of the larger specimens are found buried with their
umbones downwards.

Almost all the adult specimens studied have their anterior commissures crushed.
This, together with their strongly incurved gibbous profile and the tiny, obscured
foramen implying extreme reduction (or loss) of a functional pedicle, suggests that
these adult forms lived unattached, their umbones buried in soft sediment with the
lateral commissure approximately vertical. In fact such an orientation is the only
stable one which would have kept the gaping portion of the commissure clear of the
muddy substrate. That specimens have actually been found in such a position would
strongly support this hypothesis as it is most improbable that any transported shells
would have come to rest in such an orientation.

Nearly all the features of the adult morphology can be interpreted as adaptations
to this mode of life. The projection of the ventral planareas into the dorsal valve
appears to have been a device to allow the gaping of the valves without mud seeping
through the buried posterior portion of the commissure. Westbroek et al. (1975)
proposed a similar function for an analogous structure in Devonian uncinulids.
A consideration of how these overlapping commissural portions worked in Septi-
rhynchia as the valves were opened reveals that this structure was a very efficient yet
economic way of sealing the lateral commissure using the minimum of skeletal
material (text-fig. 5). The anterior limits of this ‘double sealing’ of the shell may be
assumed to mark the maximum permissible burial depth of the organism.

The fused deltidial plates (the henidium) of the adult appear to have served a similar
purpose in preventing mud from entering over the enclosed dorsal umbo. This seal-
ing of the umbonal part of the shell is so complete that when serial grinding it is easy

326

PALAEONTOLOGY, VOLUME 22

text-fig. 5. Idealized internal reconstruction of S. numidiensis at two ontogenetic stages: a, iate’
juvenile, and b, adult. These approximately correspond to the growth stages of the sectioned
specimens shown in text-figs. 1 and 2. Both are oriented in the same way as in text-fig. 6, but the
juvenile has been enlarged twice relative to the adult to facilitate comparison. The specimens are
drawn as though the shell were transparent but with the septalium and cardinal process shown
in sagittal section, and with crura and dental lamellae omitted for clarity. Key : cross-hatching =9
dorsal and ventral median septa; continuous lines = diductor muscles (contracted); broken
lines = adductor muscles (relaxed). Circle marks hinge axis. Musculature and pedicle morphology
necessarily somewhat speculative.

to overlook the abrupt passage from the henidium to the thin recurved sheet of the
cardinal process (text-fig. 3, 1 1 -8- 15-6).

The gibbous shape of the adult and the strongly incurved pedicle beak would have
allowed both umbones to lie at approximately the same level within the mud. This
and the symmetrical balance of the approximately equal valves (when the com-
missure is vertical) would have given the animal a certain stability and helped prevent
toppling over. The development of a fold and sulcus almost certainly aided the
separation of the inhalant from the exhalant water currents (cf. Rudwick 1970). The
associated ventral sinus would have helped to channel the exhalent currents away.

The juveniles of this species, lacking any of these features and with a less incurved
beak and proportionately larger foramen, can be assumed to have had a functional
pedicle by which they were attached to the substrate either by a stalk-like pedicle or
else by a ramifying, byssus-like structure (cf. Rudwick 1961). The differences in
external morphology between the juveniles and adults reflect these very different
modes of life and for the most part the variation in internal structure can be similarly
explained. These differences are summarized in text-fig. 6.

A comparison of text-fig. 5 a and b suggests that the fractional volume of the
organism occupied by skeleton was not constant but diminished throughout ontogeny,
thus decreasing the bulk density of the animal. Thayer (1975, p. 182) has shown how
this may be advantageous for organisms living on soft sediment and also (p. 186)
that a certain degree of sinking would have provided a greater load bearing area for
the adult shell. The increase in gibbosity of the valves during ontogeny appears to
have been largely due to an allometric mode of growth, clearly evident in the gradual
increase of the spiral angle of growth (cf. text-figs. 5-6 and Thompson 1942, fig. 352).

MANCEftlDO AND WALLEY: SEPTIRHYNCHIA FROM TUNISIA

327

MODEL OF
ORGANISM /
SUBSTRATE
RELATIONSHIP:

1 cm

attached, epifaunal > unattached, semi-infaunal

Overall Shape:

Length /Width/
Thickness Ratio:

Ventral Umbo:

Number
of Costae:

Anterior

Commissure:

Lateral
Com missure.

Planareas :

Deltidial Plates:

D iduct or
Muscles dorsally
attached to:

Ventral Median
Septum:

Dental Sockets:

increasing gibbosity and valve curvature

L ^ W > T -D> L«W^T -> L^W^T

sub-erect, exposing
foramen

highly incurved,
concealing foramen

12-14

-t> 10-12

H> 8-10

rectimarginate

almost

straight

. gently
^ uniplicate

gently arched
dorsally

strongly
^ uniplicate

strongly arched
dorsally

increasing relative area and concavity

disjunct 1> conjunct [> fused (henidium )

s e p t al i u m

<1 >

<

cardinal process

absent -O low ridge — 1> blade-like, high and long
^ do not enclose teeth

enclose teeth fully

antero-dorsa lly

text-fig. 6. General morphological trends observed during the ontogeny of S. numidiensis, showing
relationship to inferred mode of life. Though size, age, and growth stage are undoubtedly strongly corre-
lated parameters, some variation in the degree of development of any feature at any given size (between
the ontogenetic extremes) can occur, but without invalidating the basic sequence outlined.

328

PALAEONTOLOGY, VOLUME 22

In the dorsal valve, such increasing curvature would have presented problems for the
attachment of the diductor muscles in that the orientation of the septalium would
have been shifted until it faced dorsally in a useless position. This problem was solved
by the development of a thin enveloping calcite sheet which grew back over the
umbo, keeping pace with the incurvature and forming the cardinal process (text-
figs. 3 ; 4b, c, d; and 5). The calcite sheet also wrapped around the septalium forming
an extremely strong, wide base for muscle attachment. In the mode-of-life model
proposed, the pressure of mud against the posterior portion of the shell would have
been such as to require powerful diductor muscles (and hence a wide, firm muscle
platform). The approximate arrangement of these muscles, inferred from the relation-
ships of the hinge mechanism (cf. Jaanusson and Neuhaus 1965; Rudwick 1970), are
indicated in text-fig. 5.

The differences in dentition between the juveniles and adults are also significant.
In the adult the dental sockets are open antero-dorsally, a feature that is only explicable
if the anterior commissure was orientated upwards so that the teeth were held in by
gravity. Disarticulation has not been noted in the juvenile specimens whilst dis-
articulated (but largely unbroken) adult valves are not uncommon.

This saving of skeletal material in the formation of the adult sockets, the thinness
of the valves, and the minimal nature of the overlap of the lateral commissure suggest
that skeletal economies were important in S. numidiensis. It is possible that this may
have been due to the need to keep body weight as low as possible in order to prevent
excessive sinking in the soft muddy substrate. In view of this it seems unlikely that the
ventral median septum was non-functional, although the precise function it performed
cannot be determined with certainty. It was most probably involved in strengthening
the shell, although this is difficult to prove. A paradigmatic approach to interpreting
this structure, in the fashion of Rudwick (1964, 1968), is difficult because any such
structure must be a compromise between architectural optimization and biological
constraints, with major uncertainties attached to both. It would seem likely, how-
ever, that the main stresses would have been vertical (due to shell weight) and per-
pendicular to the commissural plane (due to sediment pressure against the opening
valves). A thin blade-like structure oriented in the plane of both stresses might have
been the best compromise, especially in view of the apparent restrictions on building
more massive skeletal features.

Evidence for such a function comes from the abrupt termination of the septum,
coinciding with the level of the anterior end of the overlapping portion of the lateral
commissure, a point indicative of the maximum tolerable depth of burial (see above).
The absence of the septum in other, smaller Jurassic rhynchonellids which are heavily
thickened and in the juveniles of the species under discussion, further suggests
a function associated with one of the unusual features of Septirhynchia, such as its
mode of life or large adult size. Whilst proving a strengthening function for the
ventral septum is difficult, the fact remains that alternative functions that would
have been of value to the adult Septirhynchia are difficult to imagine. It has been
suggested to the authors that the lack of a ventral median septum in the juvenile
might have been due to the need for a large pedicle capsule, which subsequently
atrophied allowing the growth of a median septum to help support the body wall
anteriorly.

MANCENIDO AND WALLEY: SEPTIRHYNCHIA FROM TUNISIA

329

There are other features which suggest possible behavioural adaptations to the semi-
infaunal mode of life. It seems likely that S. numidiensis could have maintained a con-
stant depth of burial during periods of sedimentation as the very process of opening
the valves even slightly would have tended to lever the organism upwards in the sedi-
ment. In this case the loss of a functional pedicle would have been a distinct advantage
in allowing free movement upwards.

The gregarious association of S. numidiensis would have been of advantage in
breaking up slow water currents into more turbulent eddying flow, which would have
been better for the feeding current system of the individuals.

A broadly similar mode of life has been invoked for a number of Palaeozoic groups.
Ivanova et al. 1964 (re-figured in Ager 1967a) and Ziegler et al. (1966) have suggested
on the basis of field evidence that a number of Silurian and Devonian pentameroid
brachiopod species lived with their umbones downwards. Westbroek et al. (1975)
made a similar suggestion for Devonian uncinulids, as did Grant (1971) for the
Permian tetracamerid Septacamera. Many of these forms share common morpho-
logical features with Septirhynchia, probably representing convergence due to similar
life habits.

THE GENUS SEPTIRHYNCHIA— A REAPPRAISAL
In this section an attempt is made to reinterpret all the available information on the
genus Septirhynchia in the light of the results of the studies of S. numidiensis. Parti-
cular problems considered are those of the supposed homeomorphs, the generic
characteristics, the origin of the cardinal process, and the taxonomic affinities of the
genus.

The variation between juvenile and adult forms of S. numidiensis has been discussed
above and is summarized in text-fig. 6. There is considerable evidence that such varia-
tion is also present in the other species of Septirhynchia. For instance, Dubar (1967)
described a new species of rhynchonellid, Rhynchonella pseudoazaisi for a form which,
except for its lack of a ventral median septum, seemed to him to have the appearance
of being a juvenile of S. azaisi (Cottreau) with which it was found associated. The
serial sections that he figured (p. 51, fig. 3) are very similar to those from the young
specimens of S. numidiensis (text-fig. 1) with a septalium and a cardinal process in
an early stage of formation. It would seem therefore that R. pseudoazaisi is best
regarded as a juvenile form (and consequently a junior synonym) of S. azaisi
(Cottreau).

Moreover, the small specimens that Dubar described as R. cf. budulcaensis on the
grounds that they resembled the somewhat larger S. ? budulcaensis apart from the
absence of a ventral median septum, are probably juveniles of the latter species.

In addition to these examples from Tunisia, a few other cases of alleged homeo-
morphy with Septirhynchia have to be considered before reassessing the nature and
status of this genus.

Two specimens were mentioned briefly by Muir-Wood and Cooper (1951, p. 4)
as lacking in the one case a ventral median septum, and in the other, both septum and
cardinal process. Yet, from what is now known about the ontogeny of Septirhynchia
and the limited information given about these specimens it would seem that they
might well have been juveniles of S. azaisi with which they occurred.

330

PALAEONTOLOGY, VOLUME 22

By contrast, the specimens of Rousselle (1970) from the Oxfordian/Kimmeridgian
of southern Algeria which she called ? S. cf. budulcaensis are almost certainly not
members of this genus. Their outline, ‘normal’ rhynchonellid lateral commissure,
large number of ribs, and only moderately large size distinguishes them from Septi-
rhynchia. Internally there are only a few points of resemblance and the extensive
secondary thickening is quite uncharacteristic of Septirhynchia. Very similar forms
to this species have been collected from the Oxfordian of Morocco by Dr. A. E. Adams
and on the basis of internal and external evidence are attributed to the genus Somali-
rhynchia. Also closely comparable are the specimens of S. cf. africana (Weir) from
eastern Spain, serial sections of which are shown by Ager and Walley (1977, fig. 6).
The ‘ventral septum’ found in one of Rousselle’s specimens appears to be little more
than a myophragm (euseptoidum) which probably has little generic significance
(Cooper 1970, p. 223) and is also known in some species of Somalirhynchia. In
addition, as discussed below, there is very little evidence for the much cited
Kimmeridgian age of S.? budulcaensis , and an early Callovian age is more likely. It
would seem therefore that the only records of Septirhynchia lacking either the ventral
median septum and/or cardinal process that can be accepted almost certainly refer
to juveniles.

Hence, taking together this re-interpreted evidence with the known ontogeny of
S. numidiensis, it seems unavoidable to draw the conclusion that both the ventral
median septum and the cardinal process are developed rather late in the ontogeny
throughout the genus, with a tentative suggestion that the process was formed slightly
earlier than the septum (the fact that forms with a ventral median septum but without
cardinal process have not been described may be considered significant).

The development of both features only in the adult forms of Septirhynchia involves
the variation of internal structures through ontogeny to an extent previously un-
recognized in this group, where the ‘classical’ viewpoint, currently epitomized by
Cooper (1970, pp. 220 ff.) confers great weight on the existence of reputedly stable
internal features (such as cardinal process, septa, etc.) for supraspecific taxonomy.

Rudwick (1970, p. 32) perceptively pointed out that failure to recognize the change
in external morphology undergone by a brachiopod during ontogeny ‘may even
result in an assignment of juvenile and adult individuals to different species or even
different genera!’ In the case of Septirhynchia , with its additional internal modifica-
tions, a traditional typological approach to taxonomy would have almost inevitably
resulted in juveniles and adults being separated at family level.

This demonstrates the manifest inadequacy of a classification based on the mere
presence or absence of a character at an adult stage. Far more desirable is the erection
of taxa on the basis of both juvenile and adult morphology and the inclusion of this
in the diagnosis. In the absence of juveniles some idea of ontogenetic variation can
generally be obtained by a more interpretative approach to adult characters. The
development of external features can frequently be traced by examination of the
growth lines, and the origins of internal features can often be similarly inferred from
details of shell structure (see Ager 1965a, pp. 212-213; Pearson 1977, p. 13). For
example, despite the remarkable variation in S. numidiensis, there is in fact little
about the external and internal morphology of the juvenile that cannot be deduced
from the adult form.

MANCENIDO AND WALLEY: SEPTIRHYNCHIA FROM TUNISIA

331

Whether the degree of ontogenetic variation seen in Septirhynchia is peculiar to that
genus or else is common in other brachiopods is unknown and can be ascertained
only by much careful work. Such cases might be expected when adults and juveniles of
a species are clearly adapted to markedly different modes of life.

Revised diagnosis of Septirhynchia ( Muir-Wood , 1935 )

On the basis of this new information it is necessary to alter greatly the original
diagnosis of this genus, proposed by Muir-Wood (1935, p. 106), and only slightly
modified in Muir-Wood and Cooper (1951, p. 2). Confirmation of this revision awaits
the detailed study of the type species S. azaisi, the internal structure of which (especi-
ally the brachial valve) has never been properly investigated. However, it has been
possible to examine a good photograph (kindly provided by Mme J. Drot) of the
sagittally sectioned specimen which was rather curiously chosen as the lectotype of
Cottreau’s species by Muir-Wood. This clearly shows a cardinal process similar to
S. numidiensis in that it is built up of a number of recurving sheets. The assumption
throughout this emended diagnosis is therefore that, as discussed above, the onto-
genetic variation observed in specimens of S. numidiensis is typical of the genus as
a whole :

Adult size large, pentameroid in shape, dorsal fold low, median sinus of variable depth. Ventral umbo
long and acute, greatly incurved in the adult, concealing minute foramen and deltidial plates which are
disjunct in the juvenile but fused later in ontogeny. Dental lamellae strong, uniting with the base of the
ventral median septum when present. Ventral septum occurs only in the adult when it may extend to up
to two-thirds of the pedicle valve length ; in the juvenile it is absent or represented by a low ridge. Dorsal
median septum well developed throughout ontogeny. Septalium present in juvenile forms and overturned
and concealed by the cardinal process in the adult. The cardinal process is formed by the partial envelop-
ment of the dorsal umbo and septalium by recurved thin calcite sheets. Crura raduliform, simple, and
uniformly curved. A pedicle collar may be present. Lateral planareas smooth, well developed, increasing in
importance in the adult at the expense of lateral ribs. Posterior lateral commissure arched dorsally for a short
distance where the dorsal valve overlaps the extension of the ventral valve. Shell typically ornamented by
coarse unbranching costae which may vary greatly in number depending on species and ontogenetic stage.

Distribution of species of Septirhynchia

Partly for the sake of completeness and partly because of the interest in Septirhynchia
as a typical faunal element of the Jurassic Ethiopian Province (e.g. see Ager 19676,
1973; Hallam 1975) a brief review of the records of the component species and their
stratigraphical and geographical distribution is given below (see also text-fig. 7).

S. azaisi (Cottreau) (1924, p. 581, pi. 17, figs. 1-4) was made the type species of the genus Septirhynchia
by Muir-Wood (1935, p. 107, pi. 9, figs. 3, 4) who figured another specimen also from the Somalia/Ethiopia
region and gave a short list of early synonyms. Muir-Wood and Cooper (1951, p. 4, pi. 1, figs. 6-10) and
Jaboli (1959, p. 26, pi. 4, fig. 1) also described specimens of this species from the same area. Dubar (1967,
p. 48, pi. 1, figs. 4, 7) described S. azaisi from southern Tunisia as well as Rhynchonetla pseudoazaisi which
we believe to be its juvenile form as discussed above. In the East African region the age of S. azaisi is not
really known with precision but is generally given as Callovian. In Tunisia, the specimens of Dubar come
from horizons of early to possibly mid Callovian age. S. madashonensis Muir-Wood (1935, p. 109, pi. 10,
fig. 11) was described from a single incomplete specimen from the Callovian of eastern British Somaliland.
Another single specimen has been identified by us from Tunisia, where it is found slightly below the
S. numidiensis horizon and is therefore probably of early Callovian age (see also the list of unconfirmed
records of Septirhynchia below). S. mogharaensis Muir-Wood (1935, p. 110) was based on the specimen
figured by Douville and Cossmann (1926, p. 326, pi. 7, fig. 9) as R. decorata (Schlotheim) from the Callovian
(possibly Bathonian) of Sinai.

332

PALAEONTOLOGY, VOLUME 22

5.? budulcaensis (Stefanini) (1932, p. 109, pi. 5, figs. 1, 2) was originally referred to Rhynchonella
(Stolmorhynchia!), and included a form figured by Weir (1929, p. 36, pi. 4, fig. 5) as Stolmorhynchia azaisi
(Cottr.) var. from the ‘Kimmeridgian’ of Somalia, but Muir-Wood (1935, p. 106) transferred it to her new
genus Septirhynchia. Possibly conspecific are Rhynchonella inflata Jaboli (1959, pp. 17-18, pi. 2, fig. 1)
from approximately the same area, and R. ( Stolmorhynchia! ) afifii Farag and Gatinaud (1962, p. 85,
pis. 1-1 1, fig. 5) from the Callovian of Maghara, Sinai, which were both included by Dubar in the synonymy
of the forms he described as R. cf. budulcaensis Dubar (1967, p. 52, pi. 1, figs. 8, 9). As discussed previously
these were almost certainly juveniles of Septirhynchia ? budulcaensis'. we have identified further specimens
of this species in material from approximately the same horizon and locality in southern Tunisia. Although
frequently quoted as being a Kimmeridgian species, there is very little real evidence for this and the age of
the Tunisian specimens is either earliest Callovian or latest Bathonian. S. pulchra Muir-Wood and Cooper
(1951, p. 2, pi. 1, figs. 1-5, 11, 12, pi. 2, figs. 1-6) was described from a single silicified specimen from the
Callovian of Ethiopia. Whether the specimens attributed to this species by Dubar ( 1 967, p. 46, pi. 1 , figs. 1 , 2)
belong here with S. pulchra or with S. numidiensis is not completely clear.

S. numidiensis is currently known with certainty only from the Lower or Middle Callovian sediments of
southern Tunisia.

A survey of the literature and an examination of the brachiopod collection of the British Museum
(Natural History) has provided a great deal of information on the geographical range of Septirhynchia.

We have examined a plaster cast of a single specimen identified by Doudet (1959, p. 36, pi. 4, fig. 2) as
? Stolmorhynchia sp. from the Callovian of Andranomantsy, northern Madagascar (kindly provided by
Dr. J. H. Delance). It has external features very similar to Septirhynchia madashonensis Muir-Wood but
unfortunately information on internal structures is lacking. ,

Specimens from the Callovian Muddo Erri Limestone of north-east Kenya (Muddo River locality)
include forms identified as S. azaisi by Muir-Wood (MS.) and others attributable to S. madashonensis and
S. ? cf. budulcaensis.

A number of poorly preserved specimens from Djebel Hagab on the Oman Peninsula are definitely
Septirhynchia but cannot be satisfactorily identified below generic level. This suggests that the unfigured
record of S. azaisi from the Middle-Upper Jurassic Surmeh Formation, at Surmeh, in Iran by James and
Wynd (1965, p. 2198) may be genuine but so far this has not been confirmed.

Similarly, two references to S. azaisi are made in the Lexique Stratigraphique for Lebanon and Syria
(Dubertret 1963). In both cases the ages are imprecisely known but possibly Callovian. In one case (p. 75)
Dubertret records ‘ Rhynchonella azaisi Stef, (non Cott.)’ from northern Lebanon and in another (p. 77)
R. azaisi (Cott.) is reported from Djebel Ansariyeh, north-western Syria. Furthermore, there are a few
fragmentary specimens from the latter area in the British Museum which can be confidently assigned to
Septirhynchia.

Museum material from the Callovian strata of the Zerka River area of western Jordan contains a single
specimen which is very similar in morphology to S'. ? budulcaensis.

Other forms which resemble 5.? cf. budulcaensis have been described and illustrated, under a so far un-
published (and hence unavilable) generic and specific name, from the Middle-Upper Dhruma Formation
(Bathonian-Callovian) of Djebel Tuwaiq, central Saudi Arabia, by Nazer (1970, p. 40, pi. 3, figs. 2-5,
pi. 4, fig. 1).

Other very large Callovian rhynchonellids in the British Museum from Hadramaut and Kutch (India)
have also been examined but they appear to have only the most superficial similarities to Septirhynchia.
A decision on the affinities of such forms must await further work on these poorly known faunas.

The geographical distribution of Septirhynchia is illustrated in text-fig. 7. The distribution around the
palaeoequator is particularly noteworthy. Summarizing the stratigraphical distribution, Septirhynchia
appears to be confined to Lower and Middle Callovian sediments, with the possibility of some species
being late Bathonian in age.

There is unfortunately little stratigraphical data on the provenance of the earliest Septirhynchia faunas.
Despite the genus being commonly considered a typical faunal representative of the Ethiopian province,
there is in fact no unequivocal evidence that Septirhynchia originated in the East African area.

There is a little more evidence for the stratigraphical distribution of the species within the genus, if the
sequence in Tunisia is representative. Dubar (1967) found 5.? budulcaensis as the earliest form, well below
strata containing S. azaisi and S. ''pulchra'. Work by one of us (C. D. W.) has in part confirmed this, having
found S.! budulcaensis at the same level as Dubar (late Bathonian-early Callovian), and some distance

MANCENIDO AND WALLEY: SEPTIRHYNCHIA FROM TUNISIA

333

text-fig. 7. Geographical distribution of Septirhynchia. Map based on the Callovian palaeocontinental
reconstructions of Smith and Briden (1977, p. 66) using an equal-area projection. Approximate position
of shoreline derived from various sources.

above this, a horizon with S. madashonensis a few metres below the horizon with S. numidiensis (to which
Dubar’s S. pulchra may belong). S. numidiensis appears to be the youngest species, at least in Tunisia.
Whether this sequence has any more than local significance is not known.

Ancestry and higher taxonomy of Septirhynchia

The results of the studies of S. numidiensis have cast new light on the possible
ancestry and taxonomic position of the genus Septirhynchia. In particular the
emphasis placed on the presence of the cardinal process needs to be re-examined.

The cardinal process of Septirhynchia seems to differ fundamentally from all other
brachiopod cardinal processes so far described. No structure comparable in mode of
formation to this recurving and growing back of skeletal sheets over the brachial
umbo is known to us amongst the brachiopods. There is a very close analogue in
Recent pholadacean bivalves where, in such species as Barnea Candida (Linnaeus)
(the white piddock), the anterior adductor muscle has migrated from just inside the
valve over the umbo. This process results in the muscles, with an associated skeletal
accessory plate, becoming attached to the so-called umbonal reflection, thus over-
lying what was originally the external surface of the umbo (see text-fig. 4d, e).

The cardinal process of Septirhynchia is well seen in the single silicified specimen of

334

PALAEONTOLOGY, VOLUME 22

S. pulchra figured by Muir-Wood and Cooper (1951, pi. 1, figs. 11, 12, pi. 2, figs. 3-6)
showing many of the features described in S. numidiensis. Although Muir- Wood and
Cooper noted that a septalium was present and ‘partially concealed by the incurva-
ture of the umbo’ (p. 3) no attempt was made to explain the presence of two structures
for the same function.

The uniqueness, and yet at the same time, the fundamental simplicity of the
cardinal process in Septirhynchia, coupled with the presence of a septalium in the
juvenile, suggests that it has developed secondarily. There is no evidence for this
feature having been inherited from a Palaeozoic ancestor. It will be very interesting
to see if similar cardinal processes, probably also associated with enclosed, highly
incurved brachial umbones can be found in other rhynchonellid groups.

Some answers to the questions surrounding the ancestry and affinities of Septi-
rhynchia can now be proposed in view of our new knowledge of the genus. In parti-
cular it would seem that if the characteristic adult features of the genus are directly
or indirectly adaptations for a certain mode of life and they are absent in the youngest
forms, then it can be assumed that even an immediate ancestor need not have pos-
sessed these features. In such a case a suitable ancestral stock would have probably
been characterized by a sharp beak, a septalium, raduliform crura, and a dorsal
septum, all features that are common in the sitbfamily Tetrarhynchiinae of the
Rhynchonellidae.

The most immediately obvious candidate for such an ancestral form amongst the
Bathonian Tetrarhynchiinae is Isjuminella decorata (Schlotheim). This large, coarsely
costate form, known from France and Portugal, has recently been reviewed by
Almeras (1966) and Drot and Fischer (1966) and has a number of features in common
with Septirhynchia. Externally it is a fairly globose form, with strongly incurved
umbones, a tiny foramen, and well-developed planareas. It is also extremely variable
(Drot and Fischer 1966, figs. 10-21) with some forms looking very much like Septi-
rhynchia (in many early records, species of Septirhynchia were either compared with
or identified as Rhynchonella decorata Schlotheim). Internally Isjuminella has a
simple septalium but no cardinal process, raduliform crura, and shows no ventral
septum, but is extremely thickened instead. The evidence necessary to decide whether
these similarities are due to a very close phylogenetic relationship or merely to con-
vergent evolution is not at present available. However, when attention is turned from
obviously similar forms to what little is known about trends within the genus Septi-
rhynchia other possibilities suggest themselves.

As discussed above, the earliest known species of Septirhynchia, at least in Tunisia,
is S.l budulcaensis. This is a form which looks far more like a ‘normal’ rhynchonellid
than any of the other species of the genus, having a smaller size, more numerous, much
finer costae (26-28), and few pentameroid characters. If Septirhynchia did develop
from a S.l budulcaensis-type form then it is easy to imagine the derivation of the
remaining species through a tendency towards larger size accompanied by decreasing
numbers of costae. Such a trend is also seen in ontogeny (see above, and text-fig. 6).

Curiously enough it would seem that in the Tunisian sedimentary column the
vertical disposition of the species agrees broadly with this trend, with S. numidiensis
(with eight to ten costae in the adult) being the youngest species known, and the more
costate S. madashonensis, S. ‘ pulchra ’, and S. azaisi being found immediately below.

MANCEftlDO AND WALLEY: SEPTIRHYNCHIA FROM TUNISIA 335

In view of the fact that the Tunisian sequences are the only ones that have been
examined in detail it would be premature to draw any wider conclusions from what
may be a purely local phenomenon.

If S. ? budulcaensis is the species closest to the immediate ancestors of the genus
Septirhynchia then it follows that such forms may have been so morphologically
unremarkable that it is difficult to know how they might be recognized. However, at
a broader level, it can be reasonably assumed that such forms would probably have
sprung from that plexus of the Tetrarhynchiinae which gave rise to such genera as
Kutchirhynchia, Burmirhynchia, Daghanirhynchia, and Somalirhynchia (see Ager
et al. 1972, pp. 95 ff., fig. 6).

Were the evidence of S.l budulcaensis ignored and origins from something like
a shared immediate ancestor with Isjuminella proposed, the roots of the genus
Septirhynchia would still lie in the mid (or early) Jurassic Tetrarhynchiinae. For the
purposes of taxonomic classification, despite the affinities with the Tetrarhynchiinae,
it cannot be denied that there are major differences. The presence of a cardinal process
and ventral median septum represent structural innovations apparently without
parallel in the history of that subfamily.

It is here proposed, therefore, that Septirhynchia is more appropriately classified
in a subfamily (the Septirhynchiinae) rather than in a family of its own, as originally
envisaged by Muir-Wood and Cooper (1951, p. 5). This subfamily (being monotypic
and apparently restricted stratigraphically to a single stage or so) is to be regarded
as a minor taxon by comparison with the other established subfamilies of the
Rhynchonellidae to which it is assigned. If allowance is made for ontogenetic varia-
tion, the diagnosis given by Muir-Wood and Cooper (1951, p. 5) and in the Treatise
(Ager 19656, p. H619) for the family is satisfactory for the subfamily. This proposal
does not preclude the possibility that future research may reduce even further the
taxonomic category to which Septirhynchia is assigned.

In summary the genus Septirhynchia presents something of an enigma, a highly
specialized form with a number of anomalous and apparently aberrant features, yet
seemingly very closely related to ordinary rhynchonellids. The problem is further
deepened by the striking contrast between the extensive geographical success of the
group and its brief stratigraphical range. Can these curious features be combined
into any sort of satisfactory picture?

A tentative solution can be proposed if it is remembered that in the Silurian and
Devonian many brachiopods lived semi-infaunally in soft muds, but gradually the
bivalves, better adapted to such a substrate, came increasingly to dominate this
environment in post-Palaeozoic times (Stanley 1968). It would seem that in Septi-
rhynchia the brachiopods, however briefly, regained this niche.

What appears to have happened is that in Bathonian-early Callovian times a wide-
spread marine transgression created large areas of shallow marine carbonate mud
across the southern portion of the Tethys. With what was probably the minimum of
competition, Septirhynchia seems to have evolved rapidly to fill the niche, in the
process developing again many features possessed by the Palaeozoic occupants of
this habitat. The apparently abrupt demise of this genus was probably due to the loss
of these environments by regression and also to increasing competition from the
more tolerant and flexible bivalves.

336

PALAEONTOLOGY, VOLUME 22

Acknowledgements. The material was collected during the course of a more extensive study on the palaeoeco-
logy of the Jurassic of southern Tunisia by C. D. W. The fieldwork for this was funded by the University of
South Carolina (N.S.F. grant G.F. 39074X1). The assistance of M. Ahmed Azzouz, director of the Tunisian
Geological Survey is gratefully acknowledged. M. O. M. carried out this work at University College of
Swansea during the tenure of a British Council overseas fellowship. Finally, special thanks are also due to
Professor D. V. Ager for his continuous encouragement and assistance, and for critically reading the
manuscript.

REFERENCES

ager, D. v. 1959. The classification of the Mesozoic Rhynchonelloidea. J. Paleont. 33, 324-332, pi. 49.

— 1965a. Serial grinding techniques. Pp. 212-224. In kummel, b. and raup, d. (eds.). Handbook of
paleontological techniques, xiv+852 pp. Freeman, San Francisco and London.

— 1965 b. Mesozoic and Cenozoic Rhynchonellacea. Pp. H597-H625. In moore, r. c. (ed.). Treatise on
invertebrate paleontology. Part H, Brachiopoda, vol. 2. Geological Society of America and University
of Kansas Press.

1967a. Brachiopod palaeoecology. Earth-Sci. Rev. 3, 157-179.

— 19676. Some Mesozoic brachiopods in the Tethys region. In adams, c. g. and ager, d. v. (eds.).
Aspects ofTethyan biogeography. Pubis. Syst. Ass. 7, 135-151.

— 1973. Mesozoic Brachiopoda. Pp. 431-436. In hallam, a. (ed.). Atlas of palaeobiogeography. xii +
531 pp. Elsevier, Amsterdam, London, and New York.

— childs, a. and pearson, d. a. 1972. The evolution of the Mesozoic Rhynchonellida. Geobios, 5 (2-3),
157-235.

— and walley, c. d. 1977. Mesozoic brachiopod migrations and the opening of the North Atlantic.
Palaeogeogr. Palaeoclimat. Palaeoecol. 21, 85-99.

almeras, y. 1966. Types de la collection Schlotheim (Brachiopodes) : figuration et remarques. Trav. Lab.
Geol. Univ. Lyon, N.s. 13, 277-287, pis. 17-19.

busson, G. 1967. Le Mesozoique saharien, l6re partie, Extreme Sud-tunisien. Centre Natl. Rech. Sci., ser.
geol. 8, 1-194.

childs, a. 1969. Upper Jurassic rhynchonellid brachiopods from northwestern Europe. Bull. Br. Mus. nat.
Hist. {Geol.), 6, 1-119, pis. 1-12.

cooper, G. a. 1970. Generic characters of brachiopods. Pp. 194-263, pis. 1-5. In yochelson, e. l. (ed.).
Proceedings of the North American Paleontological Convention . . ., vol. 1, part C. xiv+703 pp. Allen
Press, Lawrence, Kansas.

cottreau, J. 1924. Invertebres jurassiques de la region de Harar (Abyssinie). Bull. Soc. geol. Pr., Ser. 4,
24 (7-8), 579-591, pis. 17-18.

doudet, m. 1959. Contribution a l’etude de quelques brachiopodes du Jurassique moyen de Madagascar.
Thesis (unpubl.), Univ. Grenoble.

douville, h. and cossmann, m. 1926. Le Callovien dans le massif du Moghara, avec description des fossiles.

Bull. Soc. geol. Pr., Ser. 4, 25 (4-5), 303-328, pis. 5-8.
drot, J. and fischer, J. c. 1966. Nouvelles observations sur ‘ Rhynchonella' decorata (Schlotheim), Brachio-
pode bathonien. Annls Soc. geol. N. 86, 53-63.

dubar, G. 1967. Brachiopodes Jurassiques du Sahara Tunisien. Annls Paleont. (Inv.) 53 (1), 33-101,
pis. 1-4.

dubertret, L. 1963. Liban, Syrie : chaine des grands massifs cotiers et confins a l’Est. In dubertret, l. (ed.).
Lexique Strat. Int. Vol. III. Asie. Ease. lOcl, 7-155.

farag, i. A. M. and gatinaud, w. 1962. Six especes nouvelles du genre Rhynchonella dans le roches jurassiques
d’Egypte. J. Geol. un. Arab. Repub. 1960, 4 (1), 81-88, pis. 1-2.
grant, R. e. 1971. Taxonomy and autecology of two Arctic Permian rhynchonellid brachiopods. In dutro,
j. T. (ed.). Paleozoic perspectives: a paleontological tribute to G. Arthur Cooper. Smithson. Contr.
Paleobiol. 3, 313-335.

hallam, a. 1975. Jurassic environments. 269 pp. Cambridge University Press, Cambridge. (Earth-Science
Series.)

jaanusson, v. and neuhaus, h. 1965. Mechanism of the diductor muscles in articulate brachiopods. Stockh.
Contr. Geol. 13, 1-8.

MANCEfJIDO AND WALLEY: SEPTIRHYNCHIA FROM TUNISIA

337

jaboli, D. 1959. Fossili giurassici dell’Harar (Africa Orientale). Brachiopodi, Lamellibranchi e Gasteropodi.

Atti Accad. naz. Lincei Memorie, 4 (1), 3-100, pis. 1-11.
james, G. A. and wynd, j. G. 1965. Stratigraphic nomenclature of Iranian Oil Consortium Agreement Area.
Bull. Am. Ass. Petrol. Geol. 49, 2182-2245.

muir-wood, h. m. 1935. Jurassic Brachiopoda. In macfadyen, w. a., et al. The geology and palaeontology
of British Somaliland. II. The Mesozoic palaeontology of British Somaliland, 7, 75-147, pis. 8-13. Publ.
Govt. Somal. Protect. London.

— and cooper, G. a. 1951. A new species of the Jurassic brachiopod genus Septirhynchia. Smithson,
misc. Colins, 116, 1-6, pis. 1-2.

nazer, N. m. 1 970. Rhynchonellide brachiopods of the Middle and Upper Jurassic of the Tuwaiq Mountains,
Saudi Arabia. M.Sc. thesis (unpubl.), Univ. Kansas.

pearson, D. a. b. 1977. Rhaetian brachiopods of Europe. Neue Denkschr. Naturhist. Mus. Wien, 1, 1-85.
rousselle, L. 1970. Sur une ‘faunule’ de Rhynchonelles du Jurassique superieur de la region de Laghouat
(Sud algerien). Bull. Soc. geol. Fr., Ser. 7, 12 (3), 573-578.
rudwick, M. J. s. 1961. The anchorage of articulate brachiopods in soft substrata. Palaeontology, 4, 475,
476.

— 1964. The inference of function from structure in fossils. Br. J. Phil. Sci. 15, 27-40.

— 1968. Some analytic methods in the study of ontogeny in fossils with accretionary skeletons. In
macurda, D. B. jun. (ed.). Paleobiological aspects of growth and development. A symposium. J. Paleont.
42, Pt. 2 of 2, Suppl. No. 5, 35-49.

1970. Living and fossil brachiopods. 199 pp. Hutchinson, London.

smith, A. G. and briden, j. c. 1977. Mesozoic and Cenozoic paleocontinental maps. 63 pp. Cambridge Uni-
versity Press, Cambridge.

Stanley, s. m. 1968. Post-Paleozoic adaptive radiation of infaunal bivalve molluscs— a consequence of
mantle fusion and siphon formation. J. Paleont. 42, 214-229.
stefanini, G. 1932. Echinodermi, Yermi, Briozoi e Brachiopodi del Giura-Lias della Somalia. Palaeontogr.
ital. 32 (1931), n.s. 2, 81-130, pis. 4-8.

thayer, c. w. 1975. Morphologic adaptations of benthic invertebrates to soft substrata. J. mar. Res. 33,

Thompson, d. w. 1942. On growth and form. 1 1 16 pp. Cambridge University Press, Cambridge.
weir, J. 1929. Jurassic fossils from Jubaland, East Africa, collected by V. G. Glenday, and the Jurassic
geology of Somaliland. Monogr. geol. Dept. Hunter. Mus. 3, 1-63, pis. 1-5.
westbroek, p., NEIJNDORFF, F. and stel, J. 1975. Ecology and functional morphology of an uncinulid
brachiopod from the Devonian of Spain. Palaeontology, 18, 367-375.
ziegler, A. m., boucot, a. J. and sheldon, r. p. 1966. Silurian pentameroid brachiopods preserved in position
of growth. J. Paleont. 40, 1032-1036, pis. 121, 122.

177-189.

Manuscript received 23 January 1978
Revised manuscript received 2 May 1978

M. O. MANCENIDO
C. D. WALLEY

Department of Geology
University College of Swansea
Swansea SA2 8PP

TRILOBITES FROM THE ORDOVICIAN
AUCHENSOUL AND STINCHAR LIMESTONES
OF THE GIRVAN DISTRICT, STRATHCLYDE

by R. p. tripp

Abstract. Four new trilobite faunules from the Barr Group of the Girvan District are recorded ; three new species
are described —Remopleurides ateuchetos, Raymondaspis brocklochensis, and Hemiarges inghami. A new specific
name, Xylabion kirkdandiensis, is proposed for a form from the Confinis Flags. The trilobite assemblages indicate
inshore conditions of deposition ; the Stinchar Limestone fauna at Minuntion Quarry is a good example of the shallow-
water illaenid-cheirurid community. The close resemblance to certain lower Esbataottine Formation species indicates
equivalence in age, probably upper Chazyan. The upper part of the Stinchar Limestone is Llandeilo in terms of the
British succession ; it is not certain whether the lower horizons are Llanvirn or not, but there is no sign of a break in
the trilobite faunas of the Barr Group, apart from that attributable to transgression.

The Auchensoul Limestone yields the earliest trilobite fauna in the Barr Group.
The lower/middle Stinchar Limestone faunas bridge the gap between the Confinis
Flags (Tripp 1962) and the platy upper Stinchar Limestone (Tripp 1967). Brockloch
Quarry exposes the youngest Stinchar Limestone. The trilobites are mainly from the
following localities.

Auchensoul Limestone, Auchensoul Bridge. Nat. Grid Ref. NX 261929 (Williams 1962, p. 11,

p. 255 (al); Tripp 1962, p. 34). Hunterian Museum Collection.

Lower/middle Stinchar Limestone, Water of Gregg, near Barr, (a) East of Barr. Nat. Grid Ref.

NX 282941 (Williams 1962, p. 255 (bl)). Hunterian Museum Collection. ( b ) East of Barr. Nat.

Grid Ref. NX 279940. I.G.S. Edinburgh Collection.

Lower/middle Stinchar Limestone, Minuntion Quarry. Nat. Grid Ref. NX 220911; Text-fig. 1.

Massive Limestone (Williams 1962, p. 259). Mrs. Robert Gray’s and Mr. John Smith’s collections.

Decalcified Limestone, uppermost beds, discovered by Dr. J. K. Ingham in 1972. Hunterian Museum

Collection.

Top Stinchar Limestone, Brockloch Quarry, 0-8 kilometre north of east of Brockloch Farmhouse.

Nat. Grid Ref. NX 256951 (Williams 1962, p. 13, p. 255 (b3)). Hunterian Museum Collection.

The terminology is essentially that adopted in the Treatise on Invertebrate Paleonto-
logy, Part O. Most rare taxa are illustrated, but not described. Specimens collected
by the author have been purchased by the Hunterian Museum, Glasgow University,
with the assistance of a Treasury Grant-in-aid. Records of the numbers of trilobite
parts summarized in the Table 1 have been deposited with the British Library, Boston
Spa, Wetherby, Yorkshire, LS23 7BQ, U.K. as Supplementary Publication No. SUP
14015 (4 pages).

[Palaeontology, Vol. 22, Part 2, 1979, pp. 339-361, pis. 37-40.]

table 1. List of species recorded, and of their relative frequency, vc indicates very common, over 100 specimens; c indicates common, 11-100 specimens;

r indicates rare, 2-10 specimens; vr indicates very rare, 1 specimen

Pliomerella spp.

Encrinuroides stincharensis (Reed)
Quinquecosta aff. stincharensis Tripp
Thulincola barbarus Tripp
Calyptaulax georgei Tripp

amilchell Quarry (1), i

i Quarry (2), Benan Burn (3), I

342

PALAEONTOLOGY, VOLUME 22

SYSTEMATIC PALAEONTOLOGY

Family remopleurididae Hawle and Corda, 1847
Genus remopleurides Portlock, 1843
Remopleurides ateuchetos sp. nov.

Plate 37, figs. 1-6, 8

Diagnosis. Anterior tongue more than 50% maximum width of glabella, projecting.
Genal and dorsal spines absent. Glabella smooth.

Holotype. A.13990a, b (cranidium). Plate 37, figs. 1-3. Auchensoul Limestone, Auchensoul Bridge.

Other material. 13 cranidia, 7 free cheeks, 1 thorax (lacking first segment) with pygidium attached, 8
incomplete thoraces or single segments; Auchensoul Limestone, Auchensoul Bridge.

Material from other horizon. 14 cranidia, 4 free cheeks, 1 hypostome, 1 thorax (lacking first segment) with
pygidium attached; Stinchar Limestone, Water of Gregg, east of Barr.

Dimensions of holotype.

Length of cranidium (sag.)

8-1

mm

Width of cranidium

100

mm

Width of glabella

9-3

mm

Width of anterior tongue

4-6

mm

Distance between posterior extremities of eyes

5-6

mm

Description. Length of cranidium 80% width, moderately convex in both directions, greatest width opposite
midlength. Lateral glabellar furrows absent. Anterior tongue short, 50% maximum width of glabella,
convex transversely and projecting, curving downwards. Occipital ring long (sag.); occipital furrow
shallow. Palpebral lobe narrow, not extending forwards beyond base of tongue. Palpebral furrow shallow.
Eye of moderate size, almost vertical ; furrow above socle shallower than furrow below. Outer area of cheek
narrows rapidly forwards, narrowing out anteriorly, gently convex. Genal angle acute ; genal spine absent.
Posterior border furrow dies out half-way across cheek. Doublure of free cheek extends to eye, flattened
posteriorly ; vincular ledge curves forwards and outwards from midwidth.

Hypostome moderately convex transversely, much wider than long. Middle body elliptical, not bilobate ;
oval maculae faintly swollen. Median boss absent (on internal mould). Anterior margin weakly convex.
Anterior border short (sag.), expanding abaxially. Anteriorly wing projects laterally. Lateral and posterior
borders narrow, convex; posterolateral fork short, broad-based, tapering rapidly, directed straight back-
wards. Posterior margin transverse.

Thorax narrows steadily backwards. Rachial rings weakly convex longitudinally and transversely, no
spine on eighth ring. Rachial furrow deep. Pleurae approximately 30% width of ring, directed outwards
anteriorly, free points curve successively more strongly backwards. Seventh pleurae slightly enlarged,
extending backwards as far as eighth; posterior four pleurae reduced. Eleventh pleurae do not stretch
beyond pygidium. Articulating bosses and sockets small but tall. Articulating half ring short; articulating
furrow deep. Doublure extends to rachial furrow.

Pygidium more than twice as wide as long. Rachis short. First ring represented by a pair of transversely
elongated lobes. Second ring weakly developed. Pleurae fused, free points hardly developed. Pleural
furrows absent. Median embayment broad. Doublure convex, extending to rachis.

Glabella smooth. Occipital and thoracic rings smooth or with transverse raised lines ; a row of incon-
spicuous tubercles at posterior margin of ring. Free cheek smooth or with faint raised lines which die out
posteriorly. Doublure with terrace lines well developed.

Remarks. This species differs from described species in the absence of a spine on the eighth thoracic ring.
The free cheek resembles that of R. vulgaris Tripp (1967, p. 46, pi. 1, figs. 5-19) but the vincular ledge is
more oblique. The cranidium has a characteristic outline, and is usually smooth.

TRIPP: ORDOVICIAN TRILOBITES

343

Remopleurides sp. ( vulgaris species group)

Plate 37, fig. 11

1903 1 Remopleurides girvanensis Reed; Reed, p. 41.

Material. 3 cranidia, 40 incomplete thoraces (17 with pygidia attached); massive Stinchar Limestone,
Minuntion Quarry.

Remarks. Thoraces and pygidia correspond with R. vulgaris Tripp in gross morphology; the spine on the
eighth ring, which extends backwards no further than half-way across the ninth ring in R. vulgaris, is even
shorter in the above material, and pygidia are similar in outline. Cranidia differ in their stronger convexity
and broader tongue, over 50% maximum width of glabella (compared with 40% in R. vulgaris). The occipital
ring (and the margin of the glabella) are transversely striate, as is occasionally the case in R. vulgaris but
not in R. girvanensis.

? Remopleurides sp.

Plate 37, fig. 7

Material. 1 incomplete thorax (see explanation of plate for registered numbers) ; Auchensoul Limestone,
Auchensoul Bridge.

Remarks. One specimen consisting of four articulated thoracic segments, probably the seventh to the
eleventh, has such long pleurae (exceeding the width of the rachis), that reference to the genus Remo-
pleurides, despite the characteristic articulating system, must be treated as questionable.

text-fig. 1. Diagrammatic section in lower /middle Stinchar
Limestone at Minuntion Quarry, near Girvan. Prepared by
Dr. J. K. Ingham.

344

PALAEONTOLOGY, VOLUME 22

Family asaphidae Burmeister, 1843
Genus isotelus Dekay, 1824
Isotelus sp.

Plate 37, fig. 15

Material. 1 cranidium, 2 free cheeks, 2 hypostomes, 6 thoracic segments, 1 pygidium; Auchensoul Lime-
stone, Auchensoul Bridge.

Remarks. This form bears a general resemblance to I. stincharensis (Begg 1950, p. 288, pi. 14, figs. 4-5;
Tripp 1962, p. 6, pi. 1, figs. 20-25) from the Confinis Flags. The following are the main points of difference.
The branches of the facial suture are more divergent, the hypostome is narrower anteriorly, the rachis of the
pygidium is considerably shorter.

EXPLANATION OF PLATE 37

The specimens are testiferous unless otherwise stated; all were lightly coated with ammonium chloride

before being photographed.

Figs. 1-6, 8. Remopleurides ateuchetos sp. nov. 1. Cranidium (holotype, A.13990a). Dorsal view, x3-4.
2-3. The same, frontal and right lateral views, x 3. Auchensoul Limestone, Auchensoul Bridge. 4. Crani-
dium (A. 13991), x4. Lower Stinchar Limestone, Water of Gregg, east of Barr. 5. Right free cheek
(A. 13992a), x 6. Auchensoul Limestone, Auchensoul Bridge. 6. Ten thoracic segments and pygidium
(A. 13993a). Internal mould, x 4. Auchensoul Limestone, Auchensoul Bridge. 8. Hypostome (A. 13995),
x 4-5. Stinchar Limestone, Water of Gregg, east of Barr.

Fig. 7. 7 Remopleurides sp. Four thoracic segments (A.13994a), x 3-7. Auchensoul Limestone, Auchensoul
Bridge.

Fig. 9. Remopleurides sp. Hypostome (A. 13996a), x4. Top Stinchar Limestone, Brockloch Quarry.

Fig. 10. Hypodicranotus sp. Hypostome (IGSE 13327), x4-5. Stinchar Limestone, Water of Gregg, east
of Barr.

Fig. 11. Remopleurides sp. ( vulgaris species group). Cranidium (BM. In.21136). Massive Stinchar Lime-
stone, Minuntion Quarry, x 2.

Fig. 12. Remopleurides sp. Left free cheek (A. 13997) with subgenal spine, x6. Top Stinchar Limestone,
Brockloch Quarry.

Fig. 13. Remopleurides sp. Doublure of left free cheek (A. 13998), x 2-8. Auchensoul Limestone, Auchensoul
Bridge.

Fig. 14. Robergia sp. Left free cheek (A. 13999a), x4. Auchensoul Limestone, Auchensoul Bridge.

Fig. 15. Isotelus sp. Hypostome (A. 14000a), x2-5. Auchensoul Limestone, Auchensoul Bridge.

Figs. 16-23. Raymondaspis brocklochensis sp. nov. Top Stinchar Limestone, Brockloch Quarry.
16-17. Cranidium (holotype, A. 14001), dorsal and frontal views, • 4. 18. Hypostome (A. 14002), show-
ing small macula at end of lateral furrow, x 10. 19. Cephalon (A. 14003). Oblique left view, x 3-7.
20-21. Pygidium (A. 14004a). Posterior and dorsal views, x3. 22. Hypostome (A. 14005), x9. 23. Right
free cheek (A. 14006), -4.

Fig. 24. Eobronteus tousTripp. Cranidium (A. 14064a). Internal mould, x 2. Decalcified Stinchar Limestone,
Minuntion Quarry.

Fig. 25. Illaenid thorax and pygidium (A. 14073a), x 3. Auchensoul Limestone, Auchensoul Bridge.

Figs. 26-27. Bumastoides sp. Five thoracic segments and pygidium (BM In.21786). 26. Internal mould,
showing bicuspid anterior margin of doublure, x 2-5. 27. Oblique posterior view of latex cast from
external mould, x 3. Massive Stinchar Limestone, Minuntion Quarry.

PLATE 37

TRIPP, Ordovician trilobites

346

PALAEONTOLOGY, VOLUME 22

Family scutelluidae Richter and Richter, 1955
Genus Raymondaspis Pribyl, 1949
Raymondaspis brocklochensis sp. nov.

Plate 37, figs. 16-23

Diagnosis. Glabella gently swollen, basal width about 60% maximum width, IS
distinct. Anterior border clearly defined. Fine raised lines on glabella. Posterior
margin of pygidium transverse; postrachial extension short. Pleural ribs absent.

Holotype. A. 14001 (cranidium). Plate 37, figs. 16-17. Top Stinchar Limestone, Brockloch Quarry.

Other material. 6 cephala, 8 cranidia, 23 free cheeks, 21 hypostomes, 1 incomplete thorax with pygidium
attached, 25 pygidia ; top Stinchar Limestone, Brockloch Quarry.

Dimensions of holotype.

Length of cranidium (est.)

6-7 mm

Width of cranidium

8-3 mm

Length of glabella (sag.)

5-2 mm

Width of glabella (max.)

51 mm

Basal width of glabella

2-9 mm

Description. Differs from R. reedi Tripp (1976, p. 382, pi. 3, figs. 19-28) from the basal Superstes Mud-
stones in the following features: glabella much wider posteriorly, about 60% anterior width (compared
with 50%), convexity weaker. Glabella more strongly rounded in outline anteriorly. Preglabellar furrow
deeper. Rachial furrow shallower. Anterior border longer (exs.). Eye smaller. Intramarginal furrow extends
for full length of free cheek. Middle body of hypostome more convex. Pygidium concave abaxially, posterior
margin transverse. Rachis with one ring well defined ; postrachial ridge short. Faint terrace lines on pygidium,
convex forwards on rachis, slanting backwards and inwards adaxially on pleural lobe.

Remarks. R. brocklochensis resembles R. reedi and R. brumleyi (Cooper 1953, p. 25, pi. 9, figs. 8-10) in the
presence of IS and absence of pygidial pleurae, but differs markedly in the greater posterior width of the
glabella, and in other features mentioned above. Raymondaspis is the most abundant trilobite at Brockloch.

Family illaenidae Hawle and Corda, 1847
Unassigned illaenid parts

Plate 37, fig. 25, text-fig. 2

Remarks. A multiplicity of taxa is represented in the Auchensoul and Stinchar Limestones, but evidence for
the association of parts is lacking. Text-fig. 2 illustrates the diversity of free cheeks, rostral plates, and
hypostomes.

Family dimeropygidae Hupe, 1953
Genus dimeropyge Opik, 1937
Dimer opyge aff. labrosa Tripp, 1967

Plate 38, fig. 1

Material. 5 cranidia, 1 free cheek; decalcified Stinchar Limestone, Minuntion Quarry.

Remarks. Differs from D. labrosa Tripp (1967, p. 54, pi. 2, figs. 21-32) from the platy upper Stinchar Lime-
stone, in the following features : the three aciculate tubercles on the occipital ring are absent, the adaxial
and abaxial tubercles on the posterior border of the cranidium are not so greatly enlarged, the field of the
free cheek is broader and more tuberculate. D. labrosa was compared with D. spinifera Whittington and
Evitt (1954, p. 42, pi. 22, 23; text-figs. 9-10) from the Lincolnshire Limestone. Chatterton and Ludvigsen
(1976, p. 51) have drawn attention to the close similarity between D. labrosa and D. clintonensis Shaw

TRIPP: ORDOVICIAN TRILOBITES

347

text-fig. 2. Illaenid free cheeks, rostral plates, and hypostomes. a. Left free cheek (A. 1 4065a, b),
x3. Auchensoul Limestone, Auchensoul Bridge. B. Left free cheek (A. 14066a, b), x3.
Decalcified Stinchar Limestone, Minuntion Quarry, c. Left free cheek (A. 14067). Ventral
view showing vincular furrow and terrace lines, x 2. Top Stinchar Limestone, Brockloch
Quarry, d. Rostral plate (A. 14068a, b). Dotted lines indicate shallow terrace lines between
deeper ones; broken line indicates anterior margin of flange, x 3-5. Stinchar Limestone, Water
of Gregg, east of Barr. e. Rostral plate (A. 14069a, b), x 3-5. Decalcified Stinchar Limestone,
Minuntion Quarry. F. Rostral plate (A. 14070), x 6. Decalcified Stinchar Limestone, Minuntion
Quarry. G. Hypostome (BM In. 152709), x 3. Massive Stinchar Limestone, Minuntion Quarry.
H. Hypostome (A. 14071), x6. Decalcified Stinchar Limestone, Minuntion Quarry, i. Hypo-
stome (A. 14072a, b), x 12. Decalcified Stinchar Limestone, Minuntion Quarry.

(1968, p. 40, pi. 11, figs. 12-38) from the Chazy Group. The Minuntion form of D. labrosa is closer both to
D. clintonensis and to D. spinifera than is the type form, but both differ from the North American species
in the circumflex outline of the anterior margin of the cranidium, and the straight genal spine.

Family dionididae Giirich, 1907
Genus dionide Barrande, 1 847
IDionide sp.

Plate 38, fig. 3

Material. 1 cranidium (internal and external moulds of upper lamella) ; top Stinchar Limestone, Brockloch
Quarry.

Description. Cephalon semicircular in outline. Glabella longer than wide, moderately convex in both
directions. Glabella widens abruptly to twice basal width at 25% length from back, narrowing steadily
anteriorly. IS short, running inwards and forwards at anterior extremity of neck. 2S short, transverse,
placed near midlength of glabella, terminating in a large, shallow, rounded pit. Occipital ring conforms in
width and convexity with neck of glabella. Occipital furrow shallow. Cheek gently convex adaxially, more

348

PALAEONTOLOGY, VOLUME 22

strongly convex abaxially. A shallow furrow commences opposite IS, curving forwards and outwards,
then backwards and outwards and becoming shallower : a faint, transverse furrow runs from the rachial
furrow to the apex of this genal furrow, tangentially. Posterior border furrow possibly represented by a short
(tr.) depression almost opposite occipital furrow : posterior border otherwise obsolete. Articulating ledge
adaxially at back of cephalon. Lateral border weakly developed: genal angle not preserved. Glabella and
occipital ring smooth, except for a low tubercle opposite 2S, with two smaller tubercles in line behind.
Cheek anterior to genal furrow bears a network of low, anastomosing raised lines, with pits between;
independent caecal ridge not developed. Inner limit of lower lamella probably indicated by change of
convexity on right side. Cheek posterior to genal furrow pebbled and faintly pitted, without raised lines.
Lateral border smooth.

Remarks. This cephalon combines characters of Dionide turnbulli Whittington (1952, p. 8, pi. 2, figs. 1-6,
10-11) and Dionidella incisa Prantl and Pribyl (1949, p. 6, pi. 1, fig. 1, text-fig. 4; Whittard 1958, p. 96,
text-figs. 5a, b). Features of the former are the elongate, moderately convex glabella, with narrow neck,
and narrow occipital ring; features of the latter are the presence of 2S terminating in a pit, and the uniform
sculpture of the inner and outer parts of the cheek.

The course of the genal furrow recalls that of the posterior border furrow in a number of species, parti-
cularly that referred to Dionide formosa (Barrande) by Curtis (1961, p. 14, pi. 7, figs. 1-2, pi. 8, fig. 1) but
in the Brockloch cephalon the furrow curves much more strongly forwards, and commences opposite IS.
A cephalon referred by Butts (1941, pi. 82, fig. 26) to D. holdeni Raymond, has a strong genal furrow

EXPLANATION OF PLATE 38

The specimens are internal moulds unless otherwise stated.

Fig. 1. Dimeropyge aff. labrosa Tripp. Broad cranidium (A. 14007a), x 16. Decalcified Stinchar Limestone,
Minuntion Quarry.

Fig. 2. ? Paraharpes sp. Testiferous brim of cephalon (A. 14008), x 3. Top Stinchar Limestone, Brockloch
Quarry.

Fig. 3. ? Dionide sp. Internal mould of upper lamella of cephalon (A. 14009a), > 4. Top Stinchar Limestone,
Brockloch Quarry.

Fig. 4. Lonchodomas sp. Testiferous cranidium (A. 14010), x2-5. Stinchar Limestone, Water of Gregg,
east of Barr.

Figs. 5-11. Ceraurinella aff. magnilobata Tripp. Decalcified Stinchar Limestone, Minuntion Quarry.
5. Cranidium (A. 1401 la), x3. 6. Small cranidium (A. 14012a), x 16. 7. Hypostome (A. 14063b). Latex
cast from external mould, x 12. 8. Left free cheek (A. 14013a), x6. 9. Pygidium (A. 14014). Oblique
left posterolateral view, x 5. 10. Doublure of pygidium (A.14015b). Latex cast from external mould,
x4-5. 11. Pygidium (A. 14016). Latex cast from external mould, illuminated to show downsloping

anterior flange, x 4-5.

Figs. 12-14. Sphaerocoryphe sp. Top Stinchar Limestone, Brockloch Quarry. 12-13. Testiferous cephalon
(A. 14017). Dorsal and frontal views. Note absence of lateral spines, x5. 14. Testiferous hypostome
(A. 14018), x 6.

Figs. 15-16. Sphaerocoryphe saba Tripp. Decalcified Stinchar Limestone, Minuntion Quarry. 15. Crani-
dium (A. 14019), x 6. Left free cheek (A. 14020), x 15.

Figs. 17-19. Sphaerexochus filius Tripp. Stinchar Limestone, Minuntion Quarry. 17. Small holaspis with
external mould of hypostome exposed (A. 14021), x8. Decalcified Stinchar Limestone. 18. Cephalon
(A. 14022a). Right lateral view, x 3. Decalcified Stinchar Limestone. 19. Testiferous cephalon (BM
In.23522). Right lateral view, x 4. Massive Stinchar Limestone.

Fig. 20. Acanthoparypha or Pandaspinapyga sp. Testiferous cranidium (A. 14023) closely associated with
a left free cheek of Quinquecosta aff. stinchar ensis, x 2. Auchensoul Limestone, Auchensoul Bridge.

Fig. 21. Pliomerella sp. Testiferous left free cheek (A. 14024), x 3-5. Top Stinchar Limestone, Brockloch
Quarry.

Figs. 22-23. Quinquecosta aff. stincharensis Tripp. 22. Left free cheek (A. 14025), ■ 4. Stinchar Limestone,
Water of Gregg, east of Barr. 23. Testiferous pygidium (A. 14026), x5. Auchensoul Limestone,
Auchensoul Bridge.

PLATE 38

TRIPP, Ordovician trilobites

350

PALAEONTOLOGY, VOLUME 22

approximately in the position of the anterior branch of the specimen under description, but the sculpture
anterior and posterior to the furrow is not modified. It seems likely that genal and posterior border furrows
were homologous in function ; the changes in morphology may prove to be of taxonomic significance, and
for this reason the specimen is referred to Dionide with reserve.

Family cheiruridae Salter, 1864
Genus ceraurinella Cooper, 1953
Ceraurinella aff. magnilobata Tripp, 1967
Plate 38, figs. 5-11

Material. 2 cranidia from the massive Stinchar Limestone, Minuntion Quarry. 2 cephala, 137 cranidia,
35 free cheeks, 49 hypostomes, 77 thoracic segments, 25 pygidia ; decalcified Stinchar Limestone, Minuntion
Quarry.

Remarks. Some specimens are indistinguishable from C. magnilobata Tripp (1967, p. 61, pi. 3, figs. 14-28)
from the platy upper Stinchar Limestone, but most show the following differences : the convexity of the
glabella is stronger, 1L is shorter compared with 2L and 3L, the eye is larger compared with the width
of the cheek, the middle body of the hypostome is more convex, the great spine of the pygidium is set at
a steeper angle to the slope of the rachis, and the spine curves upwards near the tip, the anterior flange of
the pygidium slopes downwards.

New Specific Name. Lane (1971, p. 24) considered the holotype of Bartoninus dispersus Tripp (1962, p. 18,
pi. 2, fig. 22 a, b) from the Confinis Flags, to be attributable to Ceraurinella, and generically distinct from
the remainder of the material referred to that species. I am in agreement with that conclusion and hereby
propose the specific name kirkdandiensis for the other specimens referred to B. dispersus. I select as holotype
the pygidium HM A. 5312 from the Confinis Flags, Kirkdominae (Tripp 1962, pi. 2, fig. 31). The species is
attributable to the genus Xylabion Lane (1971, p. 40). The weak definition of the anterior border of the
cranidium, and the short, outwardly directed spine on the anterior flange of the pygidium distinguish
X. kirkdandiensis from other species. C. magnilobata and C. aff. magnilobata differ from C. dispersa in
their stronger glabellar furrows and larger eye.

In describing C. kingstoni Chatterton and Ludvigsen (1976, p. 52, pi. 8) compared their species most
closely with C. magnilobata, and the resemblance to the affiliated form from Minuntion is even closer. The
Scottish populations can be distinguished from the Canadian by the larger 1L of the cranidium and larger
anterior flange of the pygidium.

The resemblance of C. magnilobata to C. chondra Whittington and Evitt, and to A. typa Whittington
and Evitt has been discussed by Lane (1971, p. 19). Again, the Minuntion form stands closer to these two
species than does the type material of C. magnilobata, particularly in the steeper slope of the great spine.

The degree of resemblance between the species discussed above is sufficiently great to indicate close
relationship and therefore equivalence to the lower Esbataottine in age (see p. 359).

Genus sphaerocoryphe Angelin, 1854
Sphaerocoryphe sp.

Plate 38, figs. 12-14

Material. 1 cephalon, 44 cranidia (mainly swollen anterior glabellar lobes), 7 hypostomes, 2 pygidia ; top
Stinchar Limestone, Brockloch Quarry.

Description. Neck of glabella short. 1L large. Lateral spines absent. Fixigenal spine stout, short, directed
strongly outwards. Hypostome with elongate middle body lacking posterior lobe and middle furrow;
borders narrow. Pygidial spines not greatly divergent. Anterior lobe of glabella and middle body of hypo-
stome granulate.

Remarks. The above material resembles S.pemphis Lane (1971, p. 62, pi. 14, figs. 1-18) from the Balclatchie
Group, in its short neck, but differs in its larger 1 L, absence of lateral cranidial spines, and stouter genal
spine.

TRIPP: ORDOVICIAN TRILOBITES

351

Family encrinuridae Angelin, 1854
Genus encrinuroides Reed, 1931
Encrinuroides stincharensis (Reed, 1928)

Plate 39, figs. 1-17

1906 Encrinurus punctatus (Briinnich), var. arenaceus, Salter; Reed, p. 120, pi. 16, fig. 8.

1928 Encrinus stincharensis Reed, p. 64.

1931 Encrinurus stincharensis Reed; Reed p. 19.

Holotype (by monotypy). BM In.23157 (testiferous pygidium, massive Stinchar Limestone, Minuntion
Quarry). Reed 1906, pi. 16, fig. 8. This paper, Plate 39, figs. 16-17.

Other material. 3 pygidia from the massive Stinchar Limestone, Minuntion Quarry. 1 cephalon, 9 cranidia,
4 free cheeks, 1 hypostome, 7 thoracic segments, 21 pygidia; decalcified Stinchar Limestone, Minuntion
Quarry.

Description. Holotype pygidium triangular in outline, strongly vaulted, approximately as wide as long;
21 rachial rings and 9 pleurae, ninth pair not reaching posterior margin. Rachis almost 30% maximum
width of pygidium, narrowing slowly and steadily to an ill-defined apex at 20% length from back, strongly
convex longitudinally and transversely. Ring furrows continuous, increasingly shallow mesially on suc-
cessive rings. Rachial furrow deep and narrow anteriorly, becoming shallow posteriorly. Pleural lobe
curves strongly downwards near midwidth. Pleural ribs prominent ; first four ribs end in blunt free tips,
posterior ribs merge with lateral border. Anterior pleurae gently convex forwards, posterior pleurae
straight. First four rib furrows deep and narrow throughout, fifth furrow shallow abaxially, subsequent
furrows die out near margin. Lateral border of moderate and uniform width, extending horizontally inwards
anteriorly, sloping slightly upwards posteriorly; inner margin almost straight. Articulating half-ring and
facet short (sag.) ; articulating furrow shallow. Surface granulate ; pairs of irregularly spaced, small tubercles
or large granules on rachis. Pygidia from the decalcified limestone are identical with specimens from the
massive limestone, and associated parts are attributed to E. stincharensis accordingly.

Cephalon elliptical in outline, sagittal length slightly less than 50% width at base of genal spines, which
are small. Glabella longer than wide, width across 1L about 70% width across frontal lobe, strongly convex
longitudinally and transversely. Frontal lobe 45% length of glabella, broadly rounded in outline anteriorly.
Longitudinal median furrow broad, extending backwards from preglabellar furrow for 20% length of
glabella. Lateral lobes represented by gentle swellings on steep lateral slope of glabella, 2L slightly longer
than 3L, 1L shortest (exsag.), ridge-like and connected to neck of glabella. IS, 2S, and 3S short and broad,
successively shallower. Occipital ring short, of uniform length, wider (tr.) than base of glabella. Occipital
furrow well defined, transverse. Preglabellar furrow uniformly deep and broad, undercutting glabella.
Rachial furrow deep and narrow, straight ; deep fossula at anterior extremity ; stout apodemes adaxially
at junctions with 2S, 1 S, and occipital furrow. Anterior border of craniduum short (sag.), lengthening slightly
abaxially. Fixed cheek convex, sloping steeply towards sides. Palpebral lobe small, elevated, midlength
opposite 2S, anterior extremities 175% anterior width of glabella apart. Eye ridge absent. Posterior border
short, approximately equal in length (exs.) to occipital ring, not widening abaxially, curving forwards
abaxially. Genal spine short and slender, directed outwards, and only slightly backwards, forwardly placed.
Anterior branch of facial suture runs obliquely inwards and downwards to fossula, and then curves forwards
and inwards, defining anterior border of cranidium. Posterior branch of facial suture curves outwards and
backwards cutting border opposite IS.

Eye lobe small, elevated, separated from cheek by deep, broad furrow. Lens surface convex, expanded,
occupying more than 50% height of lobe. Cheek slopes steeply outwards. Field flattened, considerably
wider than border. Precranidial lobe short (exs.), indistinctly marked off from anterior border. Lateral
border narrow, continuous with anterior border in outline and convexity, not markedly incurved posteriorly.
Lateral border furrow deep and broad, U-shaped in cross section.

Hypostome subtriangular, anterior outline strongly rounded. Middle body oval, 75% length of hypo-
stome, very strongly convex. Rhynchos projects anteriorly but not overhanging border, ill defined posteriorly.
Anterior border strongly developed. Anterior wing large, placed anteriorly to mid-length of hypostome
rounded wing process near extremity. Lateral border narrow, depressed. Posterior tongue moderately long,
flattened, pointed. Doublure unknown.

352

PALAEONTOLOGY, VOLUME 22

Cephalon irregularly tuberculate, lateral lobes sparsely tuberculate; about 90 tubercles on glabella,
tubercle size index (Temple and Tripp in press, i.e. width of largest glabella tubercles as % of width of
frontal lobe) 7%. An irregular row of about eleven small tubercles on anterior border of cranidium. Cheeks
tuberculate, field of cheeks closely pitted between tubercles. External surface of middle body of hypostome
granulate.

Attribute coding (Temple and Tripp in press). Pygidium: 1-0; 2-21; 3-1; 4-0; 5-0; 6-9; 7-0; 8-0; 9-1.
Cranidium: 10-1; 11-1; 12-15-1100; 16-1; 17-0; 18-0; 19-0; 20-0; 21-0; 22-0; 23-0; 24-0; 25-0; 26-0;
27-90; 28-11; 29-0; 30-2; 31-0; 32-2; 33-7; 34-0.

Development. A meraspis cranidium (A. 14078) 1-2 mm in sagittal length has a narrow cranidium, IS and
2S connected across the glabella, and a small torular tubercle (Evitt and Tripp 1977, p. 114) situated
adaxially on the fixed cheek opposite 2S. A small free cheek 1-9 mm in lateral length has eye lobe larger
compared with the adult. Small pygidia lack full number of segments ring, furrows are deeper, and first
four pairs of pleurae end in out-turned free points.

Remarks. E. stincharensis, E. autochthon Tripp (1962, p. 22, pi. 3, figs. 18-25) from the Confinis Flags,
and E. polypleura Tripp (1967, p. 70, pi. 5, figs. 1-8) from the platy upper Stinchar Limestone, constitute
a species group unified by the following features: glabella widens steadily forwards, glabellar furrows
short, genal spine short, anterior border of free cheek not set at an angle to lateral border, 9 pairs of pleurae
in pygidium, posterior pleurae fused abaxially, large paired granules: on rachis of pygidium.

Genus quinquecosta Tripp, 1965
Quinquecosta aff. stincharensis Tripp, 1967

Plate 38, figs. 22-23

Material. 4 free cheeks, 1 thoracic segment, 1 pygidium; Auchensoul Limestone, Auchensoul Bridge.
2 cranidia, 1 cranidium with 12 thoracic segments attached, 9 free cheeks, 2 hypostomes, 6 pygidia ; Stinchar
Limestone, Water of Gregg, east of Barr.

Remarks. The above material agrees closely with topotypes from Auchensoul Quarry (Tripp 1967,
p. 74, pi. 5, figs. 21-30) except that the eye is smaller, the field of the free cheek is larger and bears a greater
number of pits, and the pygidial pleurae are longer. The thorax provides the first evidence that there were
at least twelve segments present in Quinquecosta , none of which are spined or macropleural. Q. stincharensis
is the only species to range (under open nomenclature) from the Auchensoul Limestone to the basal Superstes

EXPLANATION OF PLATE 39

The specimens are internal moulds from the decalcified Stinchar Limestone, unless otherwise stated ; all

are from Minuntion Quarry.

Figs. 1-17. Encrinuroides stincharensis (Reed). 1-3. Cranidium (A. 14027). Dorsal, oblique right lateral,
and frontal views, x 3. 4. Hypostome (A. 14028), x 5. 5. Two incomplete cranidia (A.14029a) showing
left fixigenal spine, x7. 6. Small pygidium (A. 14030a) showing free points on first four ribs, x7-5.
7. Cranidium (A. 14031a) showing longitudinal anterior median furrow and coarse tuberculation.
Oblique left lateral view, x 6. 8. Cranidium (A. 14032a) with glabella widening strongly forwards, x 5.
9. Incomplete thoracic segment (A.14033a), x8. 10. Right free cheek (A. 14034), x8. 11. Pygidium
(A. 14035). Latex cast from external mould, x 8. 12-13. Pygidium (A. 14036a). Left lateral and dorsal
views, x4-5. 14-15. Testiferous pygidium (BM In. 52699). Right lateral and dorsal views. Massive
Stinchar Limestone, x2. 16-17. Pygidium (holotype, BM In.23157). Ventral view showing lateral

border and dorsal view. Massive Stinchar Limestone, x 3.

Figs. 18-22. Calyptaulax georgeiThpp. 18. Cranidium (A. 14037a), x 6. 19. Incomplete thoracic segment
(A. 14038a), x 9. 20. Enlargement oflens surface ofeye (A. 14039b). External mould, - 12. 21. Pygidium
(A. 14040b). Latex cast from external mould, x 8. 22. Testiferous pygidium (BM In. 2^ 564). Massive
Stinchar Limestone, x 4.

Figs. 23-24. Thulincola barbarus Tripp. Cranidium (A. 14041a). Right lateral and dorsal views, x9.

Fig. 25. Amphilichas sp. Hypostome (A. 14042a), x2-2.

PLATE 39

TRIPP, Ordovician trilobites

354

PALAEONTOLOGY, VOLUME 22

Mudstones ; the genus is not known outside the Barr Group of the Girvan District. The closest relationship
is with a new genus described in manuscript by DeMott (1963) from the Platteville (equivalent in age to
upper Edinburg) Formation of Beloit, Wisconsin. The pygidium and free cheek closely resemble Quinque-
costa, but the glabella, in which the basal lobe (1L) is obsolete, has a strong pliomerid appearance; the
thorax, consisting of thirteen or more segments, is not unlike that recorded above from the Water of Gregg.

Family lichidae Hawle and Corda, 1847
Genus amphilichas Raymond, 1905
Amphilichas spp.

Plate 39, fig. 25

Remarks. A pygidium from the Auchensoul Limestone (A. 14079) closely resembles that figured from the
Confinis Flags, Minuntion (Tripp 1962, p. 31, fig. 28) but the rachis is shorter. An exceptionally large
hypostome from the decalcified Stinchar Limestone, Minuntion Quarry, is illustrated.

Genus hemiarges Giirich, 1901
Hemiarges inghami sp. nov.

Plate 40, figs. 1-17; text-fig. 3
1967 Hemiarges sp. Tripp, p. 81, pi. 6, fig. 26.

Diagnosis. Central lobe of glabella narrow. Bullar lobe confluent with basal lobe on
external surface. Large paired tubercles consistently arranged on glabella and fixed
cheek. A pair of short spines on occipital ring.

Holotype. A. 14063a, b (cranidium). Plate 40, figs. 1-3. Decalcified Stinchar Limestone, Minuntion Quarry.

Other material. 105 cranidia, 4 free cheeks, 35 hypostomes, 1 thoracic segment, 23 pygidia. Decalcified
Stinchar Limestone, Minuntion Quarry.

Material from other horizon. 1 cranidium from the platy upper Stinchar Limestone, Aldons Quarry.

EXPLANATION OF PLATE 40

The specimens are internal moulds from the decalcified Stinchar Limestone, unless otherwise stated ; all
are from Minuntion Quarry, except fig. 26.

Figs. 1-17. Hemiarges inghami sp. nov. 1-3. Cranidium (holotype, A. 14043a). Dorsal, frontal, and
oblique right lateral views, x 9. 4. Cranidium (A. 14044a). Oblique frontal view showing left palpebral
lobe, x 12. 5. Cranidium (A. 14045a) ; tubercles at back of central lobe far apart, x 8. 6. Left free cheek
(A. 14046). External mould, x 9. 7. Left free cheek (A. 14047a), doublure partly exposed, x 9. 8. Crani-
dium (A. 14048). External mould, showing tubercles with apical pillars, indicating setiforous pores,
x 18. 9-10. Cranidium(A. 14049a). Oblique left lateral and oblique posterior views, x 8. 11. Hypostome
(A. 14050a), x9. 12. Small cranidium (A. 14051), x 16. 13. Cranidium (A. 14052). Frontal view of
latex mould,. showing occipital spines, x 10. 14. Pygidium (A. 14053b). Latex cast from external mould,
x 12. 15. Broad pygidium (A. 14054a), x8. 16. Incomplete thoracic segment (A. 14055a), doublure
partly exposed, x 8. 17. Cranidium (A. 14049b). External mould, x 8.

Figs. 18-24. Ceratocephala relativa Tripp. 18-19. Cranidium (A.14056a). Dorsal and frontal views, x 6.

20. The same (A. 14056b). Latex cast from external mould; note problematical alga top right, x6.

21. Small cranidium (A. 14057a), showing glabella with narrow, convex central lobe, x9. 22. Dorsal
shield (A. 14058a), x4-5. 23. Left free cheek (A. 14059b). External mould, x9. 24. Hypostome (A. 14060).
External mould, x 9.

Fig. 25. Unassigned metaprotaspis (or degree zero meraspis) (A.14062), x 18.

Fig. 26. Dasycladacean alga, Vermiporella sp. (BM V.59776). Thin section, x 30. Top Stinchar Limestone,
Brockloch Quarry.

PLATE 40

TRIPP, Ordovician trilobites and alga

356

PALAEONTOLOGY, VOLUME 22

Dimensions of holotype.

Length of cranidium (sag.)

Width of glabella (maximum)
Width of central lobe (minimum)
Length of bullar lobe (exs.)

3-3 mm
2-4 mm

0- 8 mm

1- 5 mm

Description. Cranidium moderately convex longitudinally, weakly so transversely. Central lobe weakly
expanded anteriorly, extending half-way across bullar lobe, narrowing slowly backwards, not marked off
from posterior lobe. Bullar lobe equal to, or slightly wider than, central lobe opposite eye, with strong
independent convexity. Basal lobe slightly depressed, indistinctly demarcated on all sides. Posterolateral
swelling subelliptical, steeply inclined, confluent with basal lobe, strongly marked off from occipital ring.
Occipital ring slopes forwards, tapering to a point abaxially, moderately arched transversely. Longitudinal
furrow deep and narrow, curving outwards for a short distance posterior to bullar lobe before dying out
on external surface. Occipital furrow broad and shallow mesially, deep and narrow behind posterolateral
swelling. Preglabellar furrow deep and well defined, continuous with rachial furrow, which curves inwards
towards back of bullar lobe and dies out on external surface.

Anterior border long (sag.), expanding abaxially. Palpebral lobe slopes steeply upwards (Plate 40,
fig. 4); posterior extremity posterior to bullar lobe. Palpebral furrow distinct. Eye ridge low, running back-
wards and slightly outwards from anterolateral angle of glabella to palpebral lobe. Fixed cheek convex,
sloping steeply outwards and backwards. Posterior border short, narrowing steadily abaxially, sloping
forwards, set low compared with occipital ring and fixed cheek. Posterior border furrow deep, widening
strongly abaxially. Anterior branch of facial suture runs forwards and cuts anterior border at a steep
angle ; posterior branch curves outwards and backwards, crossing posterior border abaxially to midwidth.

Eye lobe elevated, rounded, moderately large; lens surface occupies upper half of lobe. Free cheek
falcate, gently sigmoidal in outline. Field subtriangular in shape, convex. Lateral and posterior borders
moderately wide, weakly convex, merging to form backwardly curving librigenal spine, which narrows
steadily. Border furrows broad and shallow, faintly continued as a single furrow on spine. Doublure wide,
strongly convex abaxially, bearing faint terrace lines parallel to margin.

Cephalon tuberculate, with large paired tubercles consistently arranged. A pair of short spines at back of
occipital ring, directed straight backwards. Five pairs of large tubercles: three pairs on central lobe along-
side bullar lobe ; one pair towards back of bullar lobe ; one pair adaxially on fixed cheek, alongside basal
lobe. Tubercles of intermediate size placed posteriorly on central lobe, anteriorly on bullar lobe, alongside
palpebral lobe, and on field of free cheek. Positions of smaller tubercles is indicated on text-fig. 3. Well-
preserved external moulds (Plate 40, fig. 8) show a pillar in the centre of the tubercle, possibly indicating
presence of an open pore, presumably setiferous.

Hypostome 70% as long as wide. Central body broadly rounded in front, though narrower and more
strongly rounded on some specimens than on others, 65% length of hypostome, weakly convex. Posterior
lobe small. Lateral and posterior furrows deep and broad on internal surface, shallow on external. Lateral
margin curves outwards and backwards at front, forming obtusely pointed posterior wing, converging
posteriorly. Posterior border broad, slightly tumid mesially; posterior margin straight. Doublure of
posterior margin extends as far as posterior furrow, convex except for an elliptical median boss which is
separated by a narrow furrow from anterior margin of doublure. Surface smooth.

Thoracic segment with pleura curving strongly backwards. Pleural furrow sharp, mesial for most of its
length, becoming shallow beyond midwidth. Doublure broad, with faint, closely spaced terrace lines.
Surface coarsely granulate.

Pygidium much wider than long. Rachis 40% maximum width, strongly arched transversely, narrowing
strongly at 65% pygidial length from front ; postrachial ridge narrow, strongly developed, reaching to border.
One ring well defined by a strong ring furrow ; second ring marked off abaxially only. Rachial furrow deep
and narrow. Pleural lobe weakly convex. First pleura short : posterior band swollen, extended into free
point; pleural furrow dies out near margin. Second pleura larger than first, expanding abaxially. Anterior
band moderately swollen, separated by border furrow from independently convex lateral border. Posterior
band swollen, extended into long free point which reaches well beyond third pleura. Second pleural furrow
dies out abruptly where it joins border furrow. Third pleura swollen, unfurrowed, circumscribed by second
interpleural furrow and border furrow. Border well defined. Terminal free points tiny, placed close together
behind postrachial ridge. A slightly larger free point at anterior extremity of third pleurae separated by

TRIPP: ORDOVICIAN TRILOBITES

357

narrow, rounded notch from second free point. Doublure broad, convex; terrace lines faint. A tubercle on
posterior bands of first and second pleurae, second adaxially placed ; pairs of small tubercles on rachis and
third pleurae (text-fig. 2).

Development. Smallest cranidia (Plate 40, fig. 12) 1 0 mm in sagittal length differ from full-grown specimens
in that bullar lobe is shorter and does not extend so far forwards. Eye ridge more strongly developed, and
separated by broad depressed area from anterior border. Longitudinal furrow extends beyond bullar lobe.
Large paired tubercles arranged as in adult.

Remarks. The pair of occipital spines distinguishes H. inghami from all other species. The closest com-
parison is with H. turneri Chatterton and Ludvigsen (1976, p. 85, pi. 19, figs. 1-41) from the lower
Esbataottine Formation. Both species have a lateral border to the free cheek bearing a few larger tubercles;
the pygidia are remarkably similar except that in H. inghami the adaxial pair of free points on the third
pleurae are smaller than the abaxial pair. The well-defined border distinguishes the pygidium from that of
other Barr Group species— H. insolitus Tripp (1967, p. 80, pi. 6, figs. 20-25) from the platy upper Stinchar
Limestone, and H. sp. (Tripp 1962, p. 32, pi. 4, figs. 29-32) from the Confinis Flags. The single cranidium
from the upper Stinchar Limestone, Aldons, quoted in the synonymy possesses the diagnostic pair of occipital
spines and large paired tubercles, but those on the central glabellar lobe are less regularly arranged.

Family odontopleuridae Burmeister, 1843
Genus ceratocephala Warder, 1838
Ceratocephala relativa Tripp, 1967

Plate 40, figs. 18-24; text-fig. 4

1967 Ceratocephala relativa Tripp, p. 81, pi. 6, figs. 27-34.

Material. 1 dorsal shield, 19 cranidia, 10 free cheeks, 1 hypostome, 1 thoracic segment, 2 pygidia ; decalcified
Stinchar Limestone, Minuntion Quarry.

Remarks. The articulated dorsal shield (Plate 40, fig. 22 ; text-fig. 3) resembles C. laciniata Whittington and

text-fig. 3. Hemiarges inghami sp.
nov. Reconstruction of dorsal shield.
Decalcified Stinchar Limestone,
Minuntion Quarry.

text-fig. 4. Ceratocephala relativa
Tripp. Outline drawing of incomplete
articulated dorsal shield (Plate 40,
fig. 22), x 10. Decalcified Stinchar
Limestone, Minuntion Quarry.

358

PALAEONTOLOGY, VOLUME 22

Evitt (1954, text-figs. 1,13-14) from the Edinburg Limestone in general construction, but differs as follows :
fourth to fifth pleural spines are the longest, whereas in C. laciniata the length increases on successive
segments, in C. relativa spines curve slightly backwards near tips, in C. relativa spines of ninth segment are
convergent, not divergent, and run parallel to first pygidial spines, surface is granulate and does not bear
small spines as in C. laciniata.

Unassigned metaprotaspis (or degree zero meraspis)

Plate 40, fig. 25 ; text-fig. 5

Material. 1 metaprotaspis (or degree zero meraspis) 0-7 mm in sagittal length; decalcified Stinchar Lime-
stone, Minuntion Quarry.

Description. Dorsal shield as long as wide, gently convex; width of rachis 20% maximum width; cranidium
occupies 55% length. Glabella parallel sided; moderately arched transversely. Preglabellar and rachial
furrows well defined, continuous. Faint transverse glabellar furrows at 30% and 60% length from occipital
furrow. Occipital ring short, well defined. Occipital furrow deep. Fixed cheek large, gently convex, joining
in a narrow band anterior to glabella without forming a distinct anterior border. Palpebral lobe and eye
ridge not preserved. Facial suture apparently indents anterolateral margins, suggesting presence of a narrow
free cheek. Posterior border of cranidium short (exs.), widening slightly abaxially. Posterior border furrow
firmly impressed, transverse, subcontinuous with occipital furrow. Anterior outline of pygidium sub-
transverse. Rachis of pygidium narrows backwards to a point 40% length of pygidium from posterior margin ;
two segments and a small, rounded terminal piece, which is strongly marked off. Rachial furrows almost
straight, broad anteriorly, narrowing backwards. Pleural lobes gently convex ; two very faint furrows are
directed outwards and slightly backwards across lobe. Border absent.

Remarks. The closest comparison is with the asaphid degree zero meraspis figured by Hughes (in press,
pi. 1, fig. 10) from the Lower Llandeilo. Our specimen is smaller, with pygidium less developed, but is
possibly attributable to the family Asaphidae.

text-fig. 5. Unassigned meta-
protaspis (or degree zero
meraspis). Reconstruction
(Plate 40, fig. 26), x 55.
Decalcified Stinchar Lime-
stone, Minuntion Quarry.

REMARKS AND CONCLUSIONS

The Auchensoul and Stinchar Limestones are sparingly fossiliferous in the main,
but yield mixed shelly faunas locally. Except for Remopleurides, nearly all the trilo-
bites consist of isolated parts, frequently broken. Most of the specimens appear to
have been transported a limited distance from the original habitat; the occasional
occurrence of delicate articulated exoskeletons indicates temporarily quiet, un-
disturbed conditions.

The decalcified lower/middle Stinchar Limestone at Minuntion Quarry provides
the shallowest-water fauna, the best example of the illaenid-cheirurid community
(Fortey 1975, p. 340) in the Girvan District; Ceraurinella is the most common genus,

TRIPP: ORDOVICIAN TRILOBITES

359

illaenids are represented by a diversity of forms although not numerically dominant,
Dimeropyge is rare, and raphiophorids and other deeper- water elements are absent.
Diversity is low. The frequency of a problematical spherical alga (Plate 40, fig. 20)
and gastropods support the suggestion of shallow-water conditions. The trilobites
are most closely related to those of the platy upper Stinchar Limestone, but that fauna
includes deeper water genera and is more diverse; algae are absent, and gastropods
rare, as the following comparison shows (specimens are lodged in the Hunterian
Museum) :

DECALCIFIED STINCHAR LIMESTONE

Algae?

Ostracodes

Bryozoa

Gastropods

Trilobite cranidia

Brachiopods

Pelmatozoan cup plates

Bivalves

Conodonts

Machaerids

Other

MINUNTION QUARRY

very common
very common
very common
158
156
95
52
20
6
9
4

UPPER STINCHAR LIMESTONE
AUCHENSOUL QUARRY
absent

very common
33
3

189

184

16

37

very common
33
5

The remaining lower/middle Stinchar Limestone assemblages indicate slightly
deeper conditions of deposition compared with the Minuntion, but shallower than
the platy upper Stinchar Limestone. The top Stinchar Limestone of Brockloch
Quarry is exceptional in the poverty of the trilobite fauna compared with the wealth
of brachiopods. No species of trilobite present is certainly known to occur at any
other locality. The occurrence of a dasycladacean alga (Plate 40, fig. 26) indicates
warm, shallow conditions. The Stinchar Limestone is succeeded by the Superstes
Mudstones, the basal part of which has yielded the offshore nileid community of
Fortey.

A study of the conodonts led Bergstrom ( 1 97 1 , p. 1 1 4) to conclude that the boundary
between the Llanvirn and the Llandeilo falls within the Stinchar Limestone. Williams
et al. (1972, text-fig. 9) placed the base of the Barr Group within the Llandeilo Series.
There is no indication of a break in the Stinchar Limestone trilobite faunas.

No trilobite species occurring in the Barr Group is known outside the Girvan
District. The closest similarity is to certain species of upper Chazyan age from North
America, in particular the lower Esbataottine Formation of the southern Mackenzie
Mountains, Canada (see pp. 346, 350, 357).

Chatterton and Ludvigsen (1976) recognized four biofacies, dominated by different
trilobite genera, in the lower Esbataottine. The Barr Group faunules do not compare
closely with any of these, though the Minuntion illaenid-cheirurid community would
fit best in the Calyptaulax-Ceraurinella biofacies. The platy upper Stinchar Lime-
stone fauna might correspond to the deeper Dimeropyge biofacies ( Dimeropyge is
the most abundant trilobite at Auchensoul Quarry) though diversity is less than at
some Calyptaulax-Ceraurinella localities. Chatterton and Ludvigsen (1976, p. 16)

360

PALAEONTOLOGY, VOLUME 22

table a list showing 74% of genera common to the lower Esbataottine Formation
and to the Girvan District (possibly all Stinchar Limestone occurrences). The
reciprocal figure, the proportion of genera common to the Stinchar Limestone and
the lower Esbataottine Formation, is less than 40%, a reflection of the exclusive
character of the Barr Group faunas.

Acknowledgements. I thank Dr. J. T. Temple and Dr. J. K. Ingham for suggestions regarding the manu-
script, Dr. G. F. Eliott for advice regarding the algae, and M. L. Holloway for drawing Text-fig. 3. I also
thank those responsible for the following collections for the loan of specimens : British Museum (Natural
History) (BM) ; I.G.S., Edinburgh (IGSE) ; Royal Scottish Museum, Edinburgh (RSM) ; Sedgwick Museum,
Cambridge.

REFERENCES

angelin, n. p. 1854. Palaeontologica Scandinavica. Part 1. Crustacea formationis transitionis, Fasc. 2,
Lund, i-ix, 21-92, pis. 25-41.

barrande, j. 1852. Systeme Silurien du centre de la Boheme, lere Partie: Recherches paleontologiques.

Vol. 1. Crustaces : Trilobites, Prague and Paris, i-xxx, 1-935, pis. 1-50.
begg, J. l. 1950. New Trilobites from Girvan. Geol. Mag. 87, 285-291, pi. 14.

bergstrom, s. m. 1971. Conodont biostratigraphy of the Middle and Upper Ordovician of Europe and
eastern North America. In sweet, w. c. and bergstrom, s. m. (eds.). Symposium on conodont biostrati-
graphy. Mem. Geol. Soc. Amer. 127, 83-161.
burmeister, H. 1843. Die Organisation der Trilobiten, Berlin, i-viii, 1-147, pis. 1-6.
butts, c. 1941. Geology of the Appalachian Valley in Virginia. Va. Geol. Survey, Bull. 52, pt. 2, 1-271,
pis. 73-95.

chatterton, b. d. e. and ludvigsen, r. 1976. Silicified middle Ordovician trilobites from the South Nahanni
River area, District of Mackenzie, Canada. Palaeontographica (A), 154, 1-106, pis. 1-22.
cooper, b. n. 1953. Trilobites from the lower Champlainian formations of the Appalachian Valley. Mem.
Geol. Soc. Amer. 55, 1-69, pis. 1-19.

curtis, M. L. k. 1961. Ordovician trilobites from the Valongo area, Portugal. Cheiruridae, Pliomeridae and
Dionididae. Bol. Soc. Geol. Portugal, 14, 1-18, pis. 1-8.
dekay, J. E. 1824. Observations on the structure of trilobites, and description of an apparently new genus,
with notes on the geology of Trenton Falls, by J. Renwick. Ann. Lyceum nat. Hist. 1, 174-189, pis. 12-13.
demott, l. l. 1963. Middle Ordovician trilobites of the upper Mississippi Valley. Unpublished Ph.D. thesis,
Harvard Univ.

evitt, w. R. and tripp, r. p. 1977. Silicified Middle Ordovician trilobites from the families Encrinuridae
and Staurocephalidae. Palaeontographica (A), 157, 109-174, pis. 1-24.
fortey, R. a. 1975. Early Ordovician trilobite communities. Fossils and Strata, 4, 331-352.
gurich, G. 1901. Ueber eine neue Lichas-Art aus dem Devon von Neu-Siid-Wales und fiber die Gattung
Lichas fiberhaupt. Neues Jahrb. Beil.-Bd. 14, 519-539, pi. 18.
hawle, i. and corda, A. j. c. 1847. Prodrom einer Monographic der bohmischen Trilobiten, Prague, 1-176,
pis. 1-6.

hughes, c. p. The Ordovician trilobite faunas of the Builth-Llandrindod inlier, central Wales, pt. 3. Bull.
Brit. Mus. nat. hist. Geology (in press).

hup£, p. 1953. Classe des Trilobites. In Traite de Paleontologie, 3, 44-246, ed. J. Piveteau, Paris.
lane, p. D. 1971. British Cheiruridae (Trilobita). Palaeontogr. Soc. ( Monogr .), 125, 1-95, pis. 1-16.
moore, R.c.(ed.). 1959. Treatise on Invertebrate Paleontology. Part O, Arthropoda l.i-xix, 1-560, Lawrence
and Meriden.

opik, a. 1937. Trilobiten aus Estland. Acta Comment. Univ. Tartu. 32, 1-163, pis. 1-26.
portlock, J. E. 1843. Report on the geology of the County of Londonderry, and of parts of Tyrone
and Fermanagh, Geol. Survey, Dublin and London, i-xxxi, 1-784, pis. 1-38.
prantl, f. and pribyl, a. 1949. O novych nebo malo znamych trilobitech ceskeno ordoviku. Rozpr. ceske
Akad. ved Uneni, 48, pt. 8, 1-22, pis. 1-2.

Raymond, p. E. 1905. The trilobites of the Chazy Limestone. Ann. Carneg. Mus. 3, 328-386, pis. 10-14.

TRIPP: ORDOVICIAN TRILOBITES

361

reed, f. r. c. 1903-1935. The Lower Palaeozoic trilobites of the Girvan district, Ayrshire. Palaeontogr.

Soc. ( Monogr .), 1903, Pt. 1, 1-48; 1906, Pt. 3, 97-186; 1931, Suppl. 2, 1-30.

1928. Notes on the family Encrinuridae. Geol. Mag. 65, 51-77.

richter, R. and richter, e. 1955. Scutelluidae n. n. (Tril.) durch ‘kleine Anderung’ eines Familien-Namens
wegen Homonymie. Senck. leth. 36, 291-293.

Salter, J. w. 1864. A monograph of the British trilobites. Palaeontogr. Soc. {Monogr.), 1-80, pis. 1-30.
shaw, f. c. 1968. Early middle Ordovician Chazy trilobites of New York. Mem. N.Y. St. Mus. Sci. Serv.
17, 1-163, pis. 1-24.

temple, j. t. and tripp, r. p. An investigation of the Encrinurinae (Trilobita) by numerical taxonomic
methods. (In press.)

tripp, R. p. 1962. Trilobites from the confinis Flags (Ordovician) of the district, Ayrshire. Trans. Roy. Soc.
Edinb. 65, 1-40, pis. 1-4.

— 1 965. Trilobites from the Albany division (Ordovician) of the Girvan district, Ayrshire. Palaeontology,
8, 577-603, pis. 80-83.

1967. Trilobites from the upper Stinchar Limestone (Ordovician) of the Girvan district, Ayrshire.

Trans. Roy. Soc. Edinb. 67, 43-93, pis. 1-6.

— 1976. Trilobites from the basal superstes Mudstones (Ordovician) at Aldons Quarry, near Girvan,
Ayrshire. Ibid. 69, 369-423, pis. 1-7.

warder, J. A. 1838. New trilobites. Am. Journ. Sci. 34, 377-380.

whittard, w. f. 1958. The Ordovician trilobites of the Shelve inlier. West Shropshire, pt. 3. Palaeontogr.
Soc. {Monogr.), Ill, 71-116, pis. 10-15.

Whittington, H. b. 1952. The trilobite family Dionididae. J. Paleont. 26, 111, pis. 1-2.

— and evitt, w. r. 1954. Silicified middle Ordovician trilobites. Mem. Geol. Soc. Amer. 59, 1-137,
pis. 1-33.

williams, a. 1962. The Barr and lower Ardmillan Series (Caradoc) of the Girvan district, south-west
Ayrshire, with descriptions of the Brachiopoda. Mem. geol. Soc. Lond. 3, 1-267, pis. 1-25.

WILLIAMS, A., STRACHAN, I., 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. Spec. Rep. Geol. Soc. Lond. 33, 1-74.

R. P. TRIPP

British Museum (Natural History)
Cromwell Road
London SW7 5BD

Typescript received 25 January 1978
Revised typescript received 28 July 1978

TAXONOMY, FUNCTIONAL MORPHOLOGY
AND PALAEOECOLOGY OF THE ORDOVICIAN
CYSTOID FAMILY HEMICOSMITID AE

by J. FREDRIK BOCKELIE

Abstract. Morphological aspects of the Middle-Upper Ordovician genus Hemicosmites are discussed. Most of
the important features are related to respiratory functions, and the distribution and functions of the rhombs are
evaluated. Five different patterns of rhomb distribution are outlined. Phyletic changes occur throughout the Ordovician
in: (1) respiratory structures, (2) expansion of oral area, including addition of new plates, and (3) increased size of
the arms. It is suggested that gonads were located in specialized appendages connected to the ambulacral furrows as
previously suggested by Jaekel (1899), and that the genus, normally found associated with reefs and carbonate mud
mounds, gave rise to Caryocrinites (Caryocrinitidae). Four new species are described, three from Norway and one
from Sweden : Hemicosmites papaveris, H. sculptus, H. sphaericus, and H. variabilis. The type species of the genus,
H. pyriformis , which is known from one specimen only, differs in many respects from all other known Hemicosmites
species. It is either an aberrant specimen or the species is generically distinct from all others. This taxonomic question
is left open due to lack of additional material.

Hemicosmites (Hemicosmitidae) is a rhombiferan cystoid consisting of a globular
to poppy-head shaped theca with three armlike appendages, a stem, and probably
a root structure (text-fig. 3). Hemicosmites belongs to a distinct group of cystoids, the
Hemicosmitida, with an endothecal respiratory pore system (see Paul 1968). The
inhalant pores consist of numerous holes forming a sieve (sieve-pores) ; the exhalant
pores are simple (pore holes; text-figs. 3, 5). Representatives of the family Hemi-
cosmitidae were confined to shallow-water environments in the Ordovician tropical
seas of the Baltic Basin. Few people have studied Hemicosmites in detail, mostly
because of the lack of well-preserved material and the fact that at first inspection the
pore system does not seem to exhibit any distinct pattern of distribution. Hemi-
cosmites pyriformis from Russia was first described by von Buch in 1840. In 1865
Kjerulf identified a loose plate from Norway as ? Hemicosmites pyriformis. During
the latter half of that century several new species were described from different
localities and strata (text -fig. 1). Jaekel (1899) gave a thorough description of Hemi-
cosmites and other cystoids, and detected important morphological features of
Hemicosmites. The first Hemicosmites recognized from Sweden (Thorslund 1936)
was identified as H. extraneus Eichwald by Regnell (1945) and described together
with a new species, H. oelandicus. New material, particularly from Norway, but also
involving a closer examination of Swedish and Estonian material, has made it possible
to arrive at a more detailed concept of the genus. Modifications of the pore systems
of Hemicosmites seem as important as those of diploporite cystoids. The respiratory
system apparently developed more rapidly in tropical areas than in other climatic
zones during the Ordovician. It is possible to discuss relationships between the two
families Hemicosmitidae and Caryocrinitidae on the basis of the configurations in

[Palaeontology, Vol. 22, Part 2, 1979, pp. 363-406, pis. 41-44.]

364

PALAEONTOLOGY, VOLUME 22

Middle (squares), and Upper (circles) Ordovician times. 1, Hemicosmites'! sp. A. ; 2, H. papaveris
sp. nov. ; 3, H. variabilis sp. nov.; 6, H. sculptus sp. nov.; 7, H. oelandicus ; 8, H. sphaericus sp.
nov. ; 9, H. rudis ; 10, H. extraneus; 11, H. pocillunr, 12. H. pulcherrimus ; 13, H. verrucosus',

14, H. grandis; 15, H. malum', 16, H. levior ; 17, H. oblongus', 18, H. pyriformis.

the material studied. Not all the known Hemicosmites species have been investigated
because the majority of the Estonian and Russian species are being revised by
Professor R. F. Hecker (pers. comm.).

GEOGRAPHICAL AND STRATIGRAPHICAL DISTRIBUTION

From Norway most of the species are new, including Hemicosmites papaveris from ‘Stage’ 4b§ (upper
Caradoc), H. variabilis from ‘Stage’ 5a (upper Ashgill), and H. sculptus from ‘Stage’ 5b (uppermost Ash-
gillian). In addition several loose thecal plates and parts of thecae probably belonging to Hemicosmites
have been found in ‘Stage’ 3c/3-y (Kjerulf 1865) and ‘Stage’ 5a. The Norwegian faunas thus contain species
throughout the whole time range of the genus.

From Sweden, Hemicosmites is known in the Lower and Middle Ordovician only. In Lower Ordovician
strata (upper Arenig) loose thecal plates are found on Oland, contemporary with the Estonian species
H. malum (Pander), but they do not seem to be conspecific; indeed, they may belong to the Caryocrinitidae.
Two species of Hemicosmites can be identified in Sweden, H. oelandicus Regnell from the Dalby Limestone
(Jaanusson 1960), equivalent to the Johvi-Keila Stage boundary of Estonia, and H. sphaericus sp. nov.
from the Kullsberg Limestone of Kullsberg, Dalarna, equivalent to the Idavere-Johvi Stage of Estonia
(text-fig. 2). Norwegian and Swedish faunas contain fewer species than those of Estonia.

There are several uncertainties as to the precise stratigraphical positions of the different species. H. malum
(Stage CJ, H. pyriformis (Stage unknown), H. pulcherrimus (Stage Dj), H. grandis (Stage F2), and H. verru-
cosus (Stage F,) occur in Estonia, the former three being Middle, the latter two Upper Ordovician in

BOCKELIE: ORDOVICIAN CYSTOIDS

365

age. All known species from Ingermanland (Leningrad district) are of Middle Ordovician age (Stage Q);
here are included H. malum, H. oblongus, and H. pyriformis (Stage unknown).

It is interesting to note that individual species within the Baltic Basin have a restricted geographical as
well as stratigraphical distribution. In fact Ingermanland and Estonia have only one species in common.
It seems somewhat strange that the Hemicosmites associated with the Kullsberg carbonate mud mounds
should be replaced by other cystoids in the late Ordovician of Sweden, whereas they were still flourishing
in similar environments in the Oslo Region. The reason may be the high degree of specialization that can
be seen in the respiratory structures and possibly also in other structures.

The distributional pattern of Hemicosmites has been somewhat uncertain since many loose thecal plates
were formerly all identified within the genus. The same even applied to complete specimens of other genera,
and Hemicosmites has been used by many authors as a collective name for different cystoids, playing a role
similar to the name Echinosphaerites. * Hemicosmites ' was described from the U.S.A. by Hall (1864), and
from Britain by Forbes (1848) and Salter (1866). By the end of the last century many of the problems were
clarified. Hemicosmites subglobosus Hall is now known as Coelocystis, and H. rugatus (Salter) and a specimen
identified by Forbes as H. oblongus belong to the family Caryocrinitidae. Thoral (1935, p. 159) mentioned
Hemicosmites from the Upper Ordovician of the south of France, but no species name was given. Regnell
(1945, p. 98 ; 1948, p. 17) also referred to this record. However, as far as can be seen from available sources,
all southern European genera appear to be caryocrinitids. Hemicosmites jaekeli Sun (1936) from Yunnan
may be a Hemicosmites, judging from the plate pattern. If it is, it is the first representative known from out-
side the Baltic Basin. Hemicosmites species are known from Norway (Kjerulf 1865; Brogger 1882; new
data here), Sweden (Thorslund 1936; Regnell 1945, 19486; new here), Estonia and Russia (Pander 1830;
von Buch 1840, 1845; Verneuil and Keyserling 1845; Jaekel 1899). Most of the described specimens were
collected during the last century, and details concerning localities and stratigraphical position are often
unknown. The oldest known species was H. malum (Pander 1830) according to Jaekel (1899, p. 308; see also
Regnell 19486, p. 16, footnote), from the ‘ Echinosphaerites Limestone’, Cx, which is probably of Aseri
age, CXa (text-fig. 2), and thus a correlative of the Didymograptus murchisoni Zone, but a loose thecal plate
of an even older possible Hemicosmites has been found at the Arenig-Llanvirn boundary of Norway. The
youngest species of the genus are found in Norway ( H . sculptus sp. nov.) and Estonia (H. verrucosus Jaekel),
both from the uppermost Ordovician (Hirnantian).

At present sixteen species are considered to belong to Hemicosmites, four of which are new. A list of the
species and present locations of types is given below. For those types that I have seen the species is marked
with an asterisk.

* 1 . Hemicosmites pyriformis von Buch, 1 840 (p. 20, pi. 1 , figs. 11,12); type species of genus. Palaontolo-
gisches Museum, Museum der Naturkunde der Humboldt Univ., D.D.R. Unnumbered specimen. Type
locality, Pulkowa, Ingermanland, U.S.S.R. Type stratum not given. Figured here PI. 43, figs. 3-5.

2. H. extraneus Eichwald, 1840 (figured specimen from pi. 11, fig. 5 of Eichwald 1860). A specimen con-
sidered to be the type is located at Leningrad University, Dept. Hist. Geol., Catalogue no. 1/3207. Type
locality, Spitham, Estonia (see Regnell 1945, p. 101). Type stratum, D3 (Vasalemma). Figured here PI. 44,
figs. 1-3.

3. H. grandis Jaekel, 1899 (p. 310, text-fig. 73). According to Jaekel (1899) preserved in Leningrad, but
now at Paleontological Institute in Moscow. Type locality, Haapsalu, Estonia. Type stratum, F2 (upper-
most Ashgill).

4. H. jaekeli Sun 1936 (p. 481, pi. 2, fig. 3). Catalogue no. S. 1281 in Geol. Surv. China, Nanking. Type
locality, Kweichou, China. Type stratum, Llandeilo age beds. The true nature of this species is difficult to
settle, but it may belong to Corylocrinus (Caryocrinitidae).

5. H. levior Jaekel, 1899 (p. 309). No figure given and description inadequate. Type not located. Accord-
ing to Bassler and Moodey (1943) the type locality is Zarskoje Selo, near Leningrad, U.S.S.R. Type stratum,
Cj. This species is coeval with H. malum Pander and it may be a junior synonym.

6. H. malum (Pander, 1830), pi. 29, fig. 1. Leningrad University, Dept. Hist. Geol., Catalogue no. 1/324.
Type locality, Zarskoje Selo, near Leningrad, U.S.S.R. Type stratum, Cx.

7. H. oblongus (Pander, 1830) (p. 146, pi. 2, figs. 22, 23). No information given as to where the type is
located. According to Bassler and Moodey (1943) the type locality is near Leningrad, U.S.S.R., and the
type stratum is Cx. Jaekel (1899) did not mention H. oblongus and may have rejected it as a separate species,
because of its deformed theca.

366

PALAEONTOLOGY, VOLUME 22

*8. H. oelandicus Regnell, 1945 (p. 99, pi. 3, fig. 12). Naturhistoriska Riksmuseet, Stockholm, Sweden.
Catalogue no. Ec4362. Type locality, Boda, Oland, Sweden. Type stratum, Lower Chasmops Limestone
(= Dalby Limestone), upper Llandeilo.

*9. H. papaveris sp. nov. Paleontologisk Museum, Oslo, Norway. Catalogue no. 94644. Type locality,
Steinvika, Langesund, Oslo Region, Norway. Type stratum, Encrinite Limestone (upper Caradoc).

10. H.pocillum Jaekel 1899 (p. 310). No figure given. According to Jaekel (1899) the material is preserved
at the University of Tartu (formerly Jurjew) and in Leningrad. No type was chosen. According to Hecker
(written comm.) this material or parts of it is at the Paleontological Institute in Moscow. Type locality,
unknown. Type stratum, Dt.

1 1 . H.pulcherrimus Jaekel, 1 899 (p. 310, pi. 1 8, fig. 6). According to Jaekel (1 899) the type was in Leningrad,
but it is now at the Paleontological Institute in Moscow (Hecker, pers. comm.). Catalogue no. 15355
(257/657). Type locality, Sack, near Tallinn (formerly Reval), Estonia. Type stratum, D,. Figured here
PI. 44, fig. 9.

12. H. rudis Jaekel, 1899 (p. 310, fig. 74). According to Jaekel (1899) one specimen is located at the
University of Tartu and one in Leningrad. No type was chosen. One or both of these specimens are now at
the Paleontological Institute, Moscow, according to Hecker (pers. comm.). Type locality, Vassalemma,
Estonia. Type stratum, D,.

*13. H. sphaericus sp. nov. = H. extraneus Regnell, non Eichwald. Paleontologisk Institute, Uppsala,
Sweden. Catalogue no. D 325 (=UM ec95 as used by Regnell 1945, p. 101, pi. 3, fig. 13). Type locality,
Kullsberg, Dalarna, Sweden. Type stratum, Kullsberg Limestone, lowermost part (lower Caradoc).

*14. H. sculptus sp. nov. Paleontologisk Museum, Oslo, Norway. Catalogue no. 97066. Type locality,
Svartoy, Ringerike, Oslo Region, Norway. Type stratum, ‘Stage’ 5b (uppermost Ashgill — Hirnantian).

*15. H. variabilis sp. nov. Paleontologisk Museum, Oslo, Norway. Catalogue no. 97079. Type locality,
Kalvsjo Quarry, Hadeland, Oslo Region, Norway. Type stratum, ‘Stage’ 5a (uppermost Ashgill).

16. H. verrucosus Eichwald 1856 (p. 124). Leningrad University Dept. Hist. Geol. (figured specimen from
pi. 1 1, fig. 3a, b, of Eichwald 1861). Cat. no. 1/327. Type locality, Soutlep, Nouck Peninsula, Estonia. Type
stratum, Fj (upper Ashgill).

The following additional species names were included in the list given by Bassler and Moodey (1943):
Hemicosmites altus Jaekel (1918, p. 97, fig. 88), based upon a single plate only; this is hardly a recognizable
feature. H. elongatus Pander (1867, p. 124) is a manuscript name ‘that got into literature by accident’
(Regnell 19486, p. 17, footnote). H. porosus Eichwald 1860 is, according to Jaekel (1899, p. 310), probably
a synonym of H. verrucosus. Eichwald (1860, p. 636) figured only two infra-laterals of H. porosus.

PALAEOECOLOGY OF HEMICOSMITES

Within the Baltic Basin Hemicosmites species have been found only in beds indicative
of regressive conditions, whether locally or regionally, and are mostly associated
with bioherms or carbonate mud mounds. Representatives of the genus may have
been present within the Basin throughout the Ordovician, but shallow-water areas
are now seldom preserved in the area. The fact that individual species have a limited
stratigraphical as well as geographical distribution may indicate a high degree of
specialization. Within some populations, however, there is considerable structural
variation even within the same age groups of individuals. Such variation has also been
found amongst other cystoids (Bockelie 1978), and may either be due to natural
variation within age groups, or indicative of dimorphism.

A study of Hemicosmites papaveris sp. nov. from the Oslo Region gives some possible
indication of adaptation to narrow ecological niches. Three bioherms in the Encrinite
Limestone (upper Caradoc) are found along a line perpendicular to the palaeoshore.
Each of these bioherms has a different echinoderm fauna. H. papaveris lived on the
bioherm furthest away from the palaeoshore at Steinvika, Langesund, and on death

BOCKELIE: ORDOVICIAN CYSTOIDS

367

Series

Graptolite

zones

East Baltic
stages

Oslo

Region

stages

NORWAY

SWEDEN

ESTONIA

INGER-

MANLAND

Dicetlograptus
anceps

Porkuni Fj

5b

Ashgi

D.complanatus

P'rgu FIc

6 1

m

Pleurograptus

linearis

Vormsi Fj^

Nabala Fja

4c

[?]

Dicranograptus.
clingani
-[?]—

Rakvere E

Lb6

Oandu Dqj

Caradoc

Keila

DH

Diplograptus

multidens

Johvi

Dl

4b<x

Idavere Cjjj

Nemagraptus

gracilis

Kukruse

4a/9

Llandeilo

Glyptograptus

teretiusculus

Uhaku

C

Ic

4a«c

Llanvirn

Didymograptus

murchisoni

Lasnamagi Cjb

Aseri

Cla

4a«i

Arenig

Didymograptus
bifidus
Didymograptus
hirundo
Didymograptus
extensus

Kunda B

m

Volkhov Bjj

Latorp Bj

3b

9I1J12I

U-? 118?|
0

,1

181

J 1 6|l7 J

text-fig. 2. Stratigraphical distribution of Hemicosmites species. The numbers refer to the same species as
in text-fig. 1. In addition: 4, H.l sp. B; 5, 77. ? sp. C; Stratigraphical correlation based mainly on data
from Mannil (1966). Upper and lower boundaries of D. clingani Zone, however, are somewhat uncertain.

fell into hollows in the bioherm. Other specimens were washed off the bioherm and
broken, and disarticulated plates were distributed in the ‘back reef’ zone. A second
bioherm about 250 m further towards the palaeoshore was also inhabited by Hemi-
cosmites, but in much smaller numbers. On this bioherm, crinoids dominated together
with some Heliocrinites. Roots of crinoids are very common on the flanks of this
bioherm. A third bioherm, at Asstranda, Porsgrunn, of approximately the same size
as the other two, and located about 10 km further towards the shore, contains no
Hemicosmites. Instead, numerous Heliocrinites seem to have replaced Hemicosmites.
Even in other areas with presumed shallow water deposits, Heliocrinites seem to have
replaced Hemicosmites, both in the Oslo Region and in Dalarna, Sweden.

‘Stages’ 5a and 5b (uppermost Ashgill) of the Oslo Region contain Hemicosmites
species associated with carbonate mud mounds (5a) and with ‘reefs’ (5b), respectively.
The two environments, each with a different species, may have differed in current
velocities, substrate, suspended nutrition, and sedimentation rates.

In Dalarna several carbonate mud mounds of Middle and Upper Ordovician ages

368

PALAEONTOLOGY, VOLUME 22

occur. Only in the Middle Ordovician of the Kullsberg carbonate mound is a Hemi-
cosmites species found ( H . sphaericus sp. nov.). Its relation to the mound is not quite
clear because of faults within the mound itself, but it may have lived on the flanks or
have been an element in a pioneer community found immediately below the mound.
A few loose thecal plates possibly belonging to the same species have been found
immediately below the Kullsberg Limestone at Amtjarn. No remains of Hemi-
cosmites are known from the Upper Ordovician of Sweden. However, a species of the
closely related Caryocrinites (Caryocrinitidae) has been described from Upper
Ordovician beds at Skalberget in Dalarna (Regnell 1945). Several loose thecal plates
from other localities in Dalarna may belong to this genus.

In Estonia Hemicosmites is also known to be associated with bioherms, both in the
Middle and Upper Ordovician (Mannil 1966). Some are known from outside bioherms
and carbonate mud mounds, but such occurrences are rare. The oldest Hemicosmitesl
sp. A, from the upper Arenig-lower Llanvirn of the Oslo Region, Norway, is from
detrital shallow-water limestones. Only a loose thecal plate has been found (Kjerulf
1865). H. oelandicus Regnell, represented by a single specimen only, also occurs in
shallow-water limestones of clastic type, in the Dalby Limestone (lower Caradoc) of
Oland, Sweden.

Hemicosmites lived attached to the sea bottom and may have had a root structure
of the same type as the closely related Caryocrinites. Hemicosmites may have been
a passive rheophile type filter feeder in the same way as known in several stalked
crinoids (Breimer 1969). This type of feeding position has also been suggested as one
of the possibilities for Caryocrinites (Sprinkle 1975). Hemicosmites, living in a shallow
water environment could, as inferred from the arm construction, also direct its arms
into the currents (text-fig. 3).

RELATIONSHIP TO OTHER FAUNAS

During the Ordovician Period most cystoids occurred within particular provinces.
The superfamily Hemicosmitida was dispersed throughout the northern Hemisphere.
Of the Hemicosmitidae, Hemicosmites was restricted to the Baltic Basin, possibly
with the exception of one Chinese species (Sun 1936). The oldest unquestionable
species is found in Estonia, probably indicating that the genus arose in the Baltic
Basin. The other genus, Tricosmites, is known exclusively from the lower Silurian of
Estonia. The family Caryocrinitidae may have had a wider distribution, both strati-
graphically and geographically. The oldest representatives of the Hemicosmitida are
loose thecal plates recently collected from Oland, Sweden, but their generic position
is uncertain at present. From the Middle Ordovician of Burma, Bather (1906)
described Caryocrinites aurorus, C. avellanus, and C. turbo, the oldest known more or
less complete specimens. Caryocrinites (under the name Stribalocystites, now con-
sidered to be a synonym by Frest 1975) is common in the Silurian of North America,
and has also been reported from the Silurian of Estonia (Schmidt 1858).

Various authors have commented on the distribution of the Hemicosmitida in
southern Europe, but the various genera are difficult to determine, because of the lack
of oral areas. The family Heterocystitidae is based on the Silurian Heterocystis from
North America, and may be a crinoid (type lost or mislaid; Paul 1969). Representa-

BOCKELIE: ORDOVICIAN CYSTOIDS

369

tives of the family Thomacystidae ( Thomacystis ) are found only in the Upper Ordo-
vician of Britain.

Within the Hemicosmitida, the genera Hemicosmites, Tricosmites , and Caryo-
crinites occur together in the Baltic Basin. Similarities in construction of Hemi-
cosmites and Caryocrinites, and the fact that Caryocrinites later replaces Hemicosmites
in its environment strengthen the possibility that these two genera are closely related.

MATERIAL AND METHODS

For this study museum collections have been used only to a limited extent. In museum collections in Oslo
only a few thecal plates were present. Some specimens of H. extraneus and H. sphaericus sp. nov. were
borrowed from Naturhistoriska Riksmuseet, Stockholm (RM) and from the Paleontologiska Institutionen
(UM), Uppsala, both Sweden. From the Humboldt Museum (HM), Berlin, D.D.R., the type specimen of
H. pyriformis was borrowed, and some working material was also borrowed from Estonia. Photographs
of types and figured specimens in Leningrad and Moscow have been placed at my disposal by Professor
Hecker and Dr. Tanja Koren.

The greater part of the material has been collected personally in Norway and Sweden from 1971 onwards.
Field work was intended to assemble as much new information as possible on morphology, stratigraphy,
and palaeoecology of cystoids. The precise stratigraphical and geographical positions of several other
cystoid species and genera were ascertained and several cystoids new to Norway were obtained. It has
proved valuable to collect large blocks containing cystoids, often completely surrounded by matrix, to
ensure that fragile and delicate structures are not lost, and also to collect associated fossils for stratigraphical
purposes. Orientated samples were taken, as well as samples for thin section analyses to facilitate palaeo-
ecological interpretations.

Hemicosmitids are generally found in marlstones and limestones, and the thecae are often more or less
complete, whereas arms, stems, and roots are mostly lost. Disarticulation of cystoids may be due to dif-
ferences in the construction of the sutural area of the thecal plates. Infilling or replacement of the fossils
are common, secondary calcite or fine-grained sediment being most usual. Cystoids other than hemico-
smitids often have pyrite infilling skeletal cavities or lining the thecae, occasionally totally replacing the
skeleton. Pyrite commonly fills the pore canals in all cystoids.

The majority of the samples have been prepared chemically because of the toughness of the sediments.
Cystoids have been completely dissolved to reveal both a mould of the external side of the thecae and the
internal steinkern. When decalcified the material was cast in silicone rubber. This preparation has allowed
study of structures which could not have been detected by mechanical preparation.

FUNCTIONAL MORPHOLOGY

A cystoid can be divided conveniently into three separate parts, the theca, the sub-
vective system, and the stem (text-fig. 3), and separate elements of each can be
described individually.

Thecal plates. Representatives of the Hemicosmitida have a restricted number of
thecal plates arranged in definite circlets. The exact number of plates in each of these
circlets may vary, but three or four circlets are always present (text -fig. 4a). Members
of the superfamily Glyptocystitida on the other hand, have a theca with five circlets
of plates termed basals (B, plural BB), infra laterals (IL, ILL), laterals (L, LL),
radials (R, RR), and orals (O, OO). Jaekel (1899) distinguished the Glyptocystitida
(referred to by him as Regularia) from the other rhombiferan cystoids (referred to as
Irregularia) by the number and the arrangement of thecal plates. He considered the
regularity in the Glyptocystitida to indicate a primitive stock, and believed a homology
to exist between plates of members of this group and the others. Jaekel (1899, p. 308,

370

PALAEONTOLOGY, VOLUME 22

text-fig. 3. Reconstruction of Hemicosmites based on
material of H. papaveris sp. nov. (theca) and similarities in
construction of arms of Caryocrinites (Sprinkle 1975). Stem
reconstruction based on H. porosus (Hecker, 1964), and
roots based on structures associated with H. papaveris sp.
nov.; ap— additional appendages (gonads?); B— basal plate;

H— hydropore; IL— infra-lateral plate; L— lateral plate;

Pe— periproct (anal opening); R— radial plate.

text-fig. 72) showed an interpretation of the plate arrangement in Hemicosmites
extraneus Eichwald, with four basals (B1-B4), six infra laterals (ILl-IL5 + ILa),
nine laterals (L1-L5+ three unnumbered LL), and nine radial plates, unnumbered.
He considered the plates given numbers (e.g. Bl, IL2 etc.) to be homologous with
corresponding plates of the Glyptocystitida. It is doubtful if this is true. Not only are
the plates of Hemicosmitida less constant than Jaekel thought, but even in Glypto-
cystitida some variation occurs.

A general formula for the Hemicosmitida may be given as 3-4 BB, 6 ?ILL, 8LL,
9 RR, and 700. This formula may not always be constant even within a species,
though it usually is. Several modifications occur and are most clearly expressed in the
family Caryocrinitidae, where tegminal plates replace the radials and orals. Even in
the Hemicosmitidae new plates may be intercalated between the radials (see p. 375),
and the number of lateral plates can vary within populations. Jaekel (1899, p. 294)
noted that one specimen of H. malum had seven ILL plates instead of the usual six.
For descriptive purposes it may thus be useful to regard the numbering of plates not

BOCKEL1E: ORDOVICIAN CYSTOIDS

371

as indicative of homology, but only as a descriptive convenience. I prefer to number
the plates consecutively in each of the plate circlets clockwise in oral aspect (as sug-
gested by Paul 1969, p. 194), commencing just left of the hydropore, which later seems
to be a fixed point in the Hemicosmitida (text-fig. 4b).

text-fig. 4. Plate diagram (a) and pore distribution (b) on Hemicosmites. B1-B4— basal plates; IL1
IL6— infra-laterals; L1-L9— lateral plates; R1-R9— radial plates. All plates are numbered in clockwise
manner; using R9 as a starting point. This plate contains the hydropore. H— hydropore; F— arm facet;
N — nodes on lateral plates ; Pe— periproct. Cluster of pores in fig. b : sieve-pores are inhalant ; larger simple

pores are exhalant.

Thecal pores. Pore rhombs in all rhombiferan cystoids are composed of sets of thecal
canals which open at the external surface of the thecal plates. Jaekel (1899) used the
term dichopore for such canals. However, the canals can open either in slits (in
Glyptocystitida) or in a pair of pores (Hemicosmitida). Paul (1972) suggested the
term cryptorhomb to denote rhombs in which the dichopores open in pores, as in
Hemicosmites. Each such dichopore opens in a pair of pores, one large simple pore
at one end and a sieve-pore at the other (text-fig. 5). In a few cases, as noted in Caryo-
crinites ornatus by Jaekel (1899), sieve-pores can be present at both ends of the
dichopore and still be functional. This situation, however, is unusual in the genus,
but typical of C. roemeri Jaekel.

The simple pores generally have an internal diameter of 0-20-0-22 mm and the
canals leading from them seem to have the same diameter throughout. The sieve-
pores, as seen on the outer surface, occur as a variable number of holes, often radially
arranged. If the plate surface is slightly abraded the appearance of the sieve-pores
may alter (text-figs. 5b, 6e). With increased abrasion it becomes evident that all the
sieve-pores are connected to a single canal leading from the plate interior towards the
external surface. As the canal approaches the external surface it shows a fourfold
petaloid cross section. Further outwards the four lobes separate and are themselves
divided into four; thus the sieve-pores have a basic fourfold pattern.

In thecal plates in which numerous sieve-pores occur, notably in the lower half of
most LL-plates, it may not be easy to decide which of the pore-holes belong to
individual sieve-pores. This is most difficult in some of the later species ( H . papaveris
sp. nov. and H. sculptus sp. nov.). For convenience I have chosen to call such pores

372

PALAEONTOLOGY, VOLUME 22

text-fig. 5. Section through a theca of Hemicosmites. A, longitudinal section through a dichopore showing
water inlet through a sieve-pore and the outlet through the simple pore ; immediately above the longitudinal
section is the appearance of sieve-pores and simple pores on the outer surface, with the dichopore canal
dashed, b, vertical section through a sieve-pore and corresponding horizontal sections at different levels
(and growth stages) to show increased complexity of sieve-pores through the plate (1-4); highly schematic.

aggregated pores (text-fig. 6e). Aggregates of pore-holes are related to current system 2
(text-fig. 7d and text below) and show a radiating pattern as if the pores belong to one
sieve-pore (text-fig. 6e). The tendency to aggregation of pores takes place in strati-
graphically older species such as H. extraneus , but is not as marked as in later species.
Aggregated pores form by the amalgamation of adjacent sieve-pore clusters during
growth and are most obvious near the centre of the LL-plates.

Distribution of pore-rhombs. The rhombs on the theca of Hemicosmites are grouped
to form five distinct flow systems (text-fig. 7b). Three of these inhalant-exhalant
systems are vertical systems and exhibit a flow direction more or less parallel to the

% M % »*

text-fig. 6. Stratigraphical reduction of sieve-pores of Hemicosmites species (a-d). a, H. extraneus (RM
Ec5280, lower Caradoc, Cm); b, H. sphaericus sp. nov. (PMO A35441, Kullsberg Limestone, lower
Caradoc) ; c, H. papaveris sp. nov. (PMO 94650, Encrinite Limestone, upper Caradoc— abraded specimen) ;
d, H. sculptus sp. nov. (PMO 97066, ‘Stage’ 5b, upper Ashgill) ; e, aggregated sieve-pores (ag) of the earliest-
formed pores (close to the plate centre) as seen in H. papaveris (PMO 94627). Scale bar is 1 mm (a-d at

same scale).

BOCKELIE: ORDOVICIAN CYSTOIDS

373

text-fig. 7. Rhomb distribution and current patterns on a theca of Hemicosmites. A, Distribution of
symmetrical and asymmetrical rhombs over the theca, mainly based on data from H. extraneus (RM
Ec27604), but found in all species, b, Current patterns over the theca inferred from the distribution of
inhalant and exhalant pores. Arrows indicate ex-current flow directions. Black arrows— exhalant vertical
systems (1-3); white arrows— exhalant horizontal systems (4-5). c, symmetrical rhomb; d, asymmetrical
rhomb. Plate terminology as in text-fig. 4a.

long axis of the theca (text-fig. 7b). Of these the lower system (1) involves rhombs
across the basal circlet and the lower part of the infra-lateral circlet, the exhalant
current being directed from the base upwards. The second system (2) involves pores
of the upper half of the ILL plates and the lower part of the LL plates, with the
exhalant currents directed from LL towards ILL plates (text-fig. 7b). The third and
upper system (3) is formed by rhombs of the lower half of the RR plates and the upper
half of the LL plates. The exhalant current was directed from the RR plates to the
LL plates (text-fig. 7b). Sea water would pass around the theca with different speed at
different levels. Knobs in the upper half of the LL plates probably caused strong water
currents around the theca at this level and reduced the possibility of recirculation in
the pore-system. Exhalant water from flow system (3) would enter the sea-water
which was passing quickly around the theca, and the possibility of recirculation would
be reduced. A strong water current may also have been present at about mid-height
of the theca just below the anus. Exhalant water of systems (1) and (2) would emerge at
this zone. A resultant respiratory flow pattern would be directed from the base
upwards and from the adoral area downwards. This type of flow pattern around the
theca agrees well with the ideas expressed by Paul (1972, text-fig. 12) for the diplo-
porite cystoid Haplosphaeronis, and may be typical for cystoids in general.

Two other systems of pore-rhomb distribution are also present, with flow directions
perpendicular to the long axis of the theca ; these systems (4 and 5) may be termed

374

PALAEONTOLOGY, VOLUME 22

horizontal systems (text-fig. 7b). They are located at the two zones of maximum water
current flow; system 4 is on the ILL plates, and system 5 on the LL plates. The hori-
zontal systems may be partly or completely missing (see below).

Symmetrical and asymmetrical pore-rhombs. Some of the rhombs show symmetrical
arrangements of the pore-pairs, whereas others do not. In symmetrical rhombs the
number of pore-pairs in each of the half rhombs is equal (text-fig. 7c), whereas in
asymmetrical rhombs it is not (text-fig. 7d). In extreme cases there may be twice as
many canals in one half rhomb as in the other. Analysing twenty-five specimens of
H. extraneus, asymmetrical rhombs are found to be arranged as in text-fig. 7a. This
pattern occurs without exceptions. In all cases where rhombs can be seen in other
species they conform to this pattern. Asymmetrical rhombs of the lateral series are
most typical in the lower part of LI, L3, L5, L7, and L8. L9 has both a symmetrical
rhomb and an asymmetrical one. This may have something to do with the position
of the anal opening. Plates L2, L4, and L6 all have asymmetrical rhombs. The pore
distribution of the lateral series would of course be reflected in the upper half of the
infra-lateral plates. In the basal series the two pentagonal plates have asymmetrical
rhombs, whereas in B1 two asymmetrical and one symmetrical rhomb are present.
In B4 all three rhombs are symmetrical. The distribution of the type of rhombs in
the lower part of the ILL plates is naturally dependent upon the pattern of the basal
plates (text-fig. 7a). The pores of the radial series are more complex because of the
small space available for their development, but even here the limited available
information suggests that a pattern following the preceding one exists in most species.
The distribution of the symmetrical and asymmetrical rhombs seems to be constant
within all species of Hemicosmites, probably including the type species H. pyriformis.
In the holotype of this last species, however, there is some doubt because no pores
have been traced in the radial plates. Whether the absence of such pores is typical of
H. pyriformis cannot be ascertained at present, because no more material of the species
has been found.

The pattern of rhomb distribution may be important for respiration, and seems to
indicate that the highest respiratory potential of the animal was on the side opposite
the anus, and involving the infra-lateral and lateral plates (respiratory systems 1
and 2). In general, asymmetrical rhombs have the higher number of pores along the
longer sides of the inhalant half-rhombs. In the laterals this side is always along the
vertical line of the plates (text-fig. 7a). The presence of asymmetrical rhombs would
have led to different amounts of oxygen/carbon dioxide exchange in the two sides of
the rhombs, causing respiratory gradients within the theca.

Most of the specimens studied have a fixed pattern of rhombs, with sieve-pores
and simple pores arranged as in text-fig. 4b. In two cases, however, anomalies have
been found, where the pores have different positions. In H. extraneus (RM Ec27604)
the rhomb connecting IL5:IL6 (consisting of one canal only) has a sieve-pore
where the simple pore should be present, and vice versa. In H. sphaericus sp. nov.
(PMO A35441) the abnormal condition is found in rhomb L9:R9, where one
canal has the pores reversed. This situation took place at some early ontogenetic
stage, because pores formed later in that same rhomb continued to grow in a
normal manner. These pathological conditions do not seem to have affected the

BOCKELIE: ORDOVICIAN CYSTOIDS

375

animal in any way, and these canals may have functioned like all other pores of
that rhomb.

Plates of the oral area. The triangular mouth is situated at the summit of the theca
and is bordered by nine radials (R1-R9). Between these, wedge plates are often inter-
calated, but their number varies (text-figs. 8, 9). The relationship between wedge
plates of Hemicosmites, adoral accessory plates of Thomacystis (Paul 1969, p. 192),
and the various accessory plates present in the Caryocrinitidae (Frest 1975) is diffi-
cult to evaluate at present. However, the wedge plates are formed mostly by initial
growth under the radial plates as wedges and they appear on the surface in (?)adult
specimens. Wedge plates are found between radials 3 and 4 and between radials 6
and 7 in H. papaveris sp. nov.

text-fig. 8. Oral area of Hemicosmites. a, positions of arm facets (F) in relation to lateral circlet (lower
circlet) and radial circlet (containing arm facets) of H. extraneus ; oral plates omitted (black, centre).
H— hydropore; N— nodes on lateral plates; B— oral area of H. pyriformis. Mouth (black centre) roughly
triangular. Ambulacral furrows leading out to arm facets (F) ending on lateral circlet. Three wedge plates
(dotted) in radial circlet. Specimen figured by Jaekel (1899, fig. 69). c, H. pyriformis (Holotype). Relation-
ship between different plates of the oral area. Arm facets reaching down to lateral circlet (white). Radial
plates overgrown or incorporated in arm plates (R5 and R2 not exposed). Rl, R8, R7, R6, R4 partly over-
grown or incorporated in arm plates. Oral cover plates are biserially arranged, and this arrangement pre-
sumably continues on to the arms. Three accessory facets present adjacent to the arm facets, and arranged
in an anti-clockwise manner. They may represent reproductive structures. I, III, V— radii; ap— arm plates;
af — additional facet; H— hydropore; Fl-9 — radial plates; wp — wedge plate.

376

PALAEONTOLOGY, VOLUME 22

In some specimens of H. papaveris sp. nov. a wedge plate may be present between
R9 and Rl. Some or all the plates bordering the upper portion of R9 may also be
wedge plates. In H. papaveris (PMO 94635) three plates occur along the upper
border of R9, and they are all formed in the same way as the wedge plates, and may
thus be comparable to these (text-fig. 9c). In this species important modifications of
many of the plates of the oral area have taken place. Species stratigraphically older
than H. papaveris, such as H. extraneus from the lower Caradoc, have only occa-
sional wedge plates, which, when present, are very small. The slightly younger
H. sphaericus sp. nov. (Caradoc) from Sweden has wedge plates clearly developed
along the same sutures as H. papaveris. Thus wedge plates were present by Caradoc
times. Hemicosmites with large wedge plates normally have larger oral areas than
species with smaller such plates. In consequence, wedge plates may be important for
increasing the size of the oral area relative to the theca. Many species show this,
including the type species, H. pyriformis (text-fig. 8b, c). Jaekal (1899, fig. 69) figured
these wedge plates (text-fig. 8b), but did not comment on them. The mouth of Hemi-
cosmites was covered by a number of small, irregular plates, for which the name oral
cover plates is appropriate. Extensive material of H. papaveris has shown that these
cover plates are formed by growth at the upper surface of the radial plates or the wedge
plates (text-fig. 9c). Cover plates are most probably homologous with the ambulacral
covering plates. If homologous with the oral plates of other rhombiferan cystoids
as suggested by Jaekel (1899), they were not homologous with the orals of glypto-
cystitids. The number of cover plates varies from one to ten, possibly more, and the
sequence in which they are formed does not appear to be systematic. The correlation
of such plates even in different individuals of the same population is difficult, let alone
between species. In young specimens of H. papaveris only one plate may be present,
covering the mouth; this may be equivalent to the central plate of Caryocrinites. In
adult specimens of H. papaveris and H. extraneous alternating plates continue from
the central plate towards the arm facets. They may well continue on to the arms as
cover plates, possibly biserial, but in most specimens of different species the con-
figuration is not simple. In the holotype of H. pyriformis (text-fig. 8c), the oral cover
plates are developed in the same manner as ambulacral cover plates in rhombiferan
cystoids such as the glyptocystitids, with a biserial arrangement continuing on to
the arms. Even here, a few regularities occur (text-fig. 8c, PI. 43, fig. 3). The variability
of such plates among the species of Hemicosmites may indicate that from this genus
may have developed the tegminal plates of Caryocrinites. To trace these evolutionary
trends in detail would require well-preserved material, which is not currently available.

Hydropore. At the upper border of R9 a pore or a slit is always present. This pore
is here interpreted as a hydropore, which is thus located on a straight line between the
mouth and the anus, as in most other cystoids. The gonopore in various cystoids is
generally found slightly offset to the left of this line.

Gonopore. No pore is present that could be identified as a gonopore, and for this
reason the pore here identified as a hydropore is taken as a gonopore by some authors.

Just before the ambulacral furrows reach on to the large arm facets, anti-clockwise
directed side branches occur ; one branch to each of the three ambulacral furrows is
present. The best examples are seen in the holotype of H. pyriformis (text-fig. 8c,

BOCKELIE: ORDOVICIAN CYSTOIDS

377

text-fig. 9. Oral areas of Hemicosmites. A, H. extraneus (Alliku, Estonia, Roomusoks Collection, 1975, no
number (TE)). Note that no wedge plates are present and number and distribution of oral cover plates
(1-6) is simple. Accessory facets (af) are also present; F— arm facets; h— hydropore, b, H. papaveris sp.
nov. (PMO 94635) containing wedge plates (dotted) along sutural area of radial plates (1-9); ap— arm
plate; H— hydropore; n — nerve canal, c, Growth of wedge plates from underneath the radials in contrast
to oral cover plates with lateral growth (PMO 94635); H— hydropore at upper part of R9.

PI. 43, fig. 3) and in one specimen of H. extraneus from Alliku, in Estonia (text-fig. 9,
PL 43, fig. 12). The side branches are about 1 mm long and are slightly irregular. The
branches end in a deep pit surrounded by an elevated facet-like structure, but no
facets of the types generally associated with appendages occur. If appendages were
directed from these facets, they would probably not consist of calcareous plate
elements, but rather appear as soft papillate structures (text-fig. 3). Such accessory
facets occur in specimens of all species that I have studied and they are all built in the
same manner. It is difficult to decide, because of different states of preservation,
whether such structures did occur on every single specimen. The fact that no gonopore
has been found might indicate the nature of the accessory facets. Jaekel (1899, p. 301)
suggested that these facets had ‘Abhangsorgane’ which might have served a different
function from that of the arms, and he suggested that these ‘arm rudiments’ served
some function in reproduction. I am inclined to agree with Jaekel in this. Accessory
facets have been found in many specimens of different species. They may have been
present early in ontogeny, because at least one specimen of H. papaveris sp. nov.,
with a thecal diameter of 6 mm, had these facets. Whether or not accessory facets are
found in both sexes cannot be ascertained at present.

In recent crinoids it is well known that the gonads are located in the arms. In
Caryocrinites (Caryocrinitidae) Frest (1975) described holes associated with arm
facets and he pointed out that these might have borne small brachioles or pinnules

378

PALAEONTOLOGY, VOLUME 22

having a sensory function. It may be possible that the ‘brachiole rudiments’ served
two functions, both sexual and sensory. In other cystoids, like Eucystis (Diploporita),
respiratory pores are found arranged in a particular way around the mouth and basal
portion of the ambulacra (Bockelie 1972) and in these animals the papillae covering
the pores could perform a sensory function in addition to the primarily respiratory
function. In a blastoid ( Pentremites rusticus) Katz and Sprinkle (1976) described some
of the respiratory structures as specialized to serve as brood pouches, besides function-
ing in normal respiration.

In most of the specimens studied the ‘arm rudiments’ are associated with the addi-
tional wedge plates of the oral area. In the holotype of H. pyriformis one of the
accessory facets is located on one of the arm plates (text -fig. 8c).

All recent echinoids and crinoids have separate sexes as far as is known, but it is
often difficult to distinguish them morphologically, at least externally (Hyman 1955,
p. 6). Sexual dimorphism may have been present in cystoids. In populations of
Hemicosmites it is not unusual to find different types of thecal shape: flat-topped
forms and forms with a dome-shaped oral area. This has been observed both in
H. papaveris and in H. variabilis, and may be a sexual character.

Periproct. The periproctal opening is located on the side of the theca, about mid-
height, and is surrounded by three plates, IL1, IL6, and L9 (text-fig. 4b). Jaekel
(1899, p. 295), however, noted some anomalies in the position of the anal opening.
In H. malum this opening was located along the IL1 :IL6 suture, and thus slightly
below the normal position. Jaekel did not mention whether this was found in one
specimen only, or if it really is typical for primitive species as he suggested. One
additional specimen of H. malum (B.M. (N.H.) 29094) has also been found with
have the anal opening in this position. In one specimen of H. pyriformis (see Jaekel
1899, pi. 19, fig. 4) the anal opening occurs in the ‘ Caryocrinites ’ position, i.e. at the
upper border of L9 (Jaekel 1899, p. 296). To my knowledge, in all other specimens of
known species the anal opening occurs where IL1, IL6, and L9 meet. This anomaly
may well indicate the close relationship between Hemicosmitidae and Caryocrinitidae,
and the trends of evolution.

The anal opening was covered by a pyramid of five, or occasionally four, triangular
plates (text -fig. 21c, d). These plates were moveable, and could open to let out faecal
matter, probably in the form of pellets.

Subvective system. Most cystoids lose their subvective appendages very soon after
death. However, in a few specimens of Hemicosmites extraneus and H. papaveris, as
well as in the holotype of H. pyriformis, remains of the basal portions of such
appendages are preserved. Hemicosmites possessed biserial appendages (arms) with
alternating sizes of brachiolar plates (text-figs. 3, 8c). In cross section a U-shaped
food groove is present, similar to that seen in Caryocrinites (Sprinkle 1975, fig. Id).
Immediately below the food groove an additional canal, probably representing one
of the aboral radial nerve canals(?) is present (text-fig. 10c, d). Because only the basal
portions of the arms are preserved, it is difficult to compare these plates of Hemi-
cosmites with the arms of Caryocrinites as described by Sprinkle (1975). One specimen
of Caryocrinites (text-fig. 10e; PI. 43, fig. 2) has a basal portion of the arms that

BOCKELIE: ORDOVICIAN CYSTOIDS

379

resembles Hemicosmites (text-fig. 10c; PI. 43, fig. 1), but both genera seem to be
variable (text-fig. 8c). The arms were attached to the facets formed at the sutures of
two adjacent radial plates ; occasionally an adjacent lateral plate may be incorporated,
in which case three plates meet in the facetal area. In the holotype of H. pyriformis
the major portion of the facet is on the lateral plates and either one or two laterals
are incorporated in the facetal area in addition to the distal portion of the adjacent
radial plates (text-fig. 8c). This type of facet is considered to be a highly evolved stage
for Hemicosmites.

text-fig. 10. Arm facets and arm plates of Hemicosmites (a-d) and Caryocrinites (e).
a, B, view of both sides of an arm of H. extraneus (RM Ec5283); c, frontal view of arm of
H. extraneus (RM Ec5283); d, frontal view of arm of H. papaveris sp. nov. (PMO 94635);

E, arm of Caryocrinites ornatus (PMO A21227). a— brachial element; amb— ambulacral
furrow; ap— arm cover plate; n— nerve canal ; wp— wedge plate. Scale bars indicate 1 mm.

The outer limits of the arm facets are somewhat irregular and several arm plates
form the proximal portions. In the oldest species, like H. extraneus, a fairly clear
distinction can be made between arm plates and oral cover plates (text-fig. 10a, b).
Five arm plates are preserved on either side of the arm in RM Ec5283 (text-fig. 10a, b ;
PI. 43, fig. 1). The plates on either side may vary considerably in shape and size. The
oral cover plates continue on to the arms where they are gradually reduced in size
as they change into arm cover plates. However, no clear distinction can be made as

380

PALAEONTOLOGY, VOLUME 22

to where the oral cover plates are transformed into arm cover plates. Arm cover
plates may be biserial, but certain irregularities have been found both in H. extraneus
and H. papaveris. Problems in detecting sutures between such cover plates because
of their small size and abrasion may account for these irregularities. In H. pyriformis
the distinction between arm plates, radial plates, and wedge plates may be difficult
to establish. In the holotype some arm plates are incorporated in the oral portion and
cover some of the radial plates (i.e. R5 and the greater part of R2). The shapes of the
arm plates may vary considerably from one arm to the other in the basal portions
(text-fig. 8c). The arm cover plates, however, are arranged biserially outwards on to
the arms (text-fig. 8c). In this respect the arms of this species resemble those of other
cystoids, e.g. Echinosphaerites (work in preparation). Whereas the oral cover plates
were more irregularly distributed in the other Hemicosmites species, they show a dis-
tinctive orderly distribution in H. pyriformis. When present as oral cover plates, they
have an irregular ornament, but become smoother when entering the arms. Arm
types found in Caryocrinites (Sprinkle 1975) may be similar to those of Hemicosmites,
since these two genera have many other features in common. One specimen of
H. extraneus (RM Ec5283) with the basal portion of the arm preserved may have
a facet for a pinnule preserved just where the arm plates become free from the facet
(text -fig. 10c; PI. 43, fig. 1). The structure is not well preserved, but if this really is
a facet for a pinnule it may further strengthen the hypothesis of a close relationship
between Hemicosmites and Caryocrinites, even with regard to the arm structures.

Nerve system. In contrast to crinoids, cystoids show no traces of nerve canals on the
inner side of the thecal plates. For this reason very little is known about the nerve
system of cystoids. In Hemicosmites and a few other cystoids nerve canals have been
found leading out into the arms. Only three radial nerves off a presumed oral nerve
ring were present in Hemicosmites leading to the arms (text-fig. 11a). A basal nerve

text-fig. 1 1 . Nerve system of Hemicosmites. a, nerve canals (n) leading off the oral nerve ring into the arms
in H. papaveris sp. nov. (PMO 94644). b, interior of the basal portion of the theca of H. variabilis sp. nov.
(PMO 97096). d— depressions in sutural area, three of which continue somewhat upwards, one being closed ;
s— hollows for presumed nerve canal leading into stem interior; t— thickened portion of basal plate, also
in the other plates; tr— triangular space where basal nerve ring may have been situated.

BOCKELIE: ORDOVICIAN CYSTOIDS

381

ring giving off branches to the stem was probably also present in Hemicosmites. One
specimen of H. variabilis (text-fig. 11b; PL 42, fig. 3) shows the interior of the basal
portion. The basal plates are thickened internally; and these areas (t) approach each
other at the plate sutures, leaving three open and one closed depression along the
plate sutures (d). In the base the depressions form triangular spaces (tr). In addition
other minor depressions occur in the basal portion of the basal plates. Consequently
several depressions exist along the inner side of the basal portion of the theca, where
an aboral nerve ring could have been present. A hollow centre of the basal portion
could have contained nerve canals leading down into the stem(s). According to
Sprinkle (1975) no traces of the aboral nerve system seem to continue into the arms of
Caryocrinites. On the basis of current information it cannot be decided if this holds
for Hemicosmites. An aboral nerve ring also seems to be present in the sessile diplo-
poritic cystoid Haplosphaeronis (Sphaeronitidae) from Norway (work in preparation).

Stem and attachment. In some specimens of H. papaveris parts of the proximal
portion of the stem are found. It is set in a socket, reaching about 1 mm into the base
of the theca. The lumen of the columnals (Moore et al. 1968) is circular, its radius
about one-third that of the columnal. The circular individual columnal has
numerous, simple crenulae and a faint crenulate suture.

In H. papaveris no differentiation of the nodals has been seen, whereas in H. extraneus
(RM Ec5280) the noditaxis (series of one nodal and all the corresponding internodals ;
see Moore et al. 1968) may contain three internodals. In a specimen of Caryocrinites
sp. (PMO A21227) the noditaxis may be similar to that of H. extraneus. H. porosus
(Hecker 1964, pi. 2, fig. 9) may have a xenomorphic stem (e.g. the proximal and distal
columnals are quite different). So far no other species with such a long stem has been
found, for which reason it can only be assumed that there is a possible xenomorphic
structure of the stem in other Hemicosmites species (text-fig. 3).

As yet neither the root structures nor attachment plates are known, and there are no
indications of what the distal portion of the stem looked like. Since Hemicosmites
always seems to be present within or around carbonate mud mounds and bioherms,
where other echinoderms have been found attached, it can be assumed that Hemi-
cosmites species were provided with root structures similar to that of Caryocrinites.
At the locality at which H. papaveris is most common, root structures are very common,
but it is not known if these roots are related to Hemicosmites or to crinoids. No
attachment discs are known from localities in the Oslo Region or Sweden where
Hemicosmites occur, whereas root structures are present. It is thus assumed that root
structures may be present in Hemicosmites (text-fig. 3).

Growth. The life-span of Hemicosmites is unknown, but it has been suggested that
other cystoids could live for 3-4 years (Paul 1967). During growth certain changes
and modifications of the theca and the thecal structures take place. Some of these
changes are of minor importance, whereas others may be important or even critical
for the animal. The most important features are related to the respiratory pores and
the oral area. The distinction of which observed features are due to ontogenetic
changes and which merely fall into the natural variation of the population at different
stages of development, can only be made on the basis of large collections illustrating
different growth stages in the same population (Bockelie 1978). As the animal grew

382

PALAEONTOLOGY, VOLUME 22

some new features were added (e.g. pores) and some could become modified by
resorption. Several growth stages, however, are perceivable on individual specimens ,
because growth stages can be incorporated in the structures of the adult. Such
features may pertain to growth lines of thecal plates (Breimer and Macurda 1972)
and to the pore system.

The ratio of thecal height to width of H. papaveris changes very little ; thus the shape
is almost constant, irrespective of the animal’s age (text-fig 12b). In H. variabilis on
the other hand, the large spines of the lateral plates are most obvious in young
individuals (PI. 41, fig. 1 1), but less so in adults.

©

text-fig. 12. a, phylogenetic trends in the oral area of Hemicosmites. Changes of shape and size of arm
facets with time; (a) H. extraneus (RM Ec2760); (b) H. papaveris sp. nov. ; and (c) H. variabilis sp. nov.
Upper row in oral view, lower row in lateral view. Note the change from spindle shape to irregular poppy-
head shape (a-c). H. papaveris (holotype, PMO 94644) is shown in different lateral position than the other
two, because the periproctal area of this species is not preserved. B, thecal height/width ratios in different
Hemicosmites species shows almost no change with time (species 2-6). 1 —H. pyrijormis (type of genus);
2 —H. extraneus ; 3 —H. oelandicus\ 4 — H. sphaericus sp. nov.; 5 —H. papaveris sp. nov.; 6 —H. sculptus

sp. nov.

Growth of rhombs. The rhombs expand in size during life by addition of new dicho-
pores. As a result the record of ontogenetic development may be retained in the
skeleton, even though parts of individual dichopores may change somewhat by
resorption to fulfil functional requirements as the animal grew. The simple pores of
a pore-rhomb seem to have the same diameter and general appearance whether
formed at an early or late stage. The number of holes in individual sieve-pores on the
other hand may show significant differences. There is a tendency towards fewer
numbers of such holes in sieve-pores close to the plate sutures (text-fig. 13b). The
greatest differences relate to sieve-pores of the lateral plates. In a specimen of Hemi-
cosmites rudis the sieve-pores contain symmetrically arranged, slightly elongated
pores. In the first formed sieve-pores seven to eight holes are present, whereas in the
last formed, closest to the plate suture, only two holes were observed (text-fig. 13c).

BOCKELIE: ORDOVICIAN CYSTOIDS

383

Similarly 9-16 pores occur in the first formed sieve-pores of H. pyriformis and only
4-8 in the last formed (text-fig. 13c). The number of holes in sieve-pores of the basal
plates is not well known because of the poor preservation of many specimens, but
in general fewer pores are present than elsewhere in the theca. In most cases, smaller
examples of H. extraneus have fewer holes per sieve-pore than larger specimens. This
implies that the number of holes in a sieve-pore increases with growth in this species.
This would explain why the largest number of pores per sieve -pore occurs in the
central, oldest parts of the LL plates, and also why aggregated pores are found in the
oldest-formed dichopores of the LL plates.

Pore-rhombs of the thecal plates do not normally reach the plate centre, which
means that in the earliest ontogenetic stages the cystoids may have had direct
02/C02 exchange through the thin skeletal plates. Infra-lateral and lateral plates of
Hemicosmites have rhombs, some of which reach closer to the plate centre than others
of the same plate. This means that some rhombs were formed at an early ontogenetic
stage, and some later during ontogeny. The major rhomb system comprises the
vertical systems 1, 2, and 3. The horizontal system 4 and 5 came into operation later
in ontogeny. It would be advantageous for a cystoid that relied on diffusion to dis-
tribute oxygen internally to have as many rhombs as possible equally distributed
over the thecal surface. All points inside the theca would then be equidistant from the
oxygen source. With growth of the theca larger areas without rhombs may have
developed adjacent to the intercirclet sutures, and rhombs of systems 4 and 5 were
then developed here.

text-fig. 13. Pores of Hemicosmites species, a, trend towards reduction in the number of holes in
sieve-pores throughout the Ordovician. 1—//. extraneus ; 2—77. sphaericus sp. nov. ; 3—77. pyri-
formis; 4—77. papaveris sp. nov.; 5—77. variabilis sp. nov.; 6 — 77. sculptus sp. nov. Mean value
(horizontal line) and total range of observations, n— number of holes in sieve-pores. B, number of
holes in individual sieve-pores in L9 of five specimens of 77. extraneus. n— number of pore holes;
s — order of sieve-pore formation, c, same in laterals of 77. pyriformis (solid lines), and 77. rudis
(various types of broken lines). Note that all the last formed sieve-pores (end of connected lines)
always have fewer pores than earlier formed sieve-pores.

384

PALAEONTOLOGY, VOLUME 22

Plates of the oral area. As mentioned above there is an increase in the number of oral
cover plates with increasing size of the animal, but the sequence in which these plates
were formed is not known in detail. Wedge plates may occur early in ontogeny, and
seem to be associated with the facets of the sexual structures, and this may indicate
that additional facets were formed after the major parts of the theca were in operation.

PHYLOGENY OF HEM/COSMITES AND RELATION TO OTHER GENERA

Several features of Hemicosmites show systematic changes with time, such as reduction
in the number of pore holes in sieve-pores (text-fig. 13a). The size of the individual
holes increased during the Ordovician, and seems to correspond with an increase of
mean size of sediment particles in the Hemicosmites- dominated environments.
Better material is needed for detailed measurements to calculate the dependencies.
A reduction of the pore systems during the Ordovician can also be observed. Not only
does the number of dichopore canals become reduced in number, but some of the
rhomb systems (text-fig. 7) show marked reductions. However, more material is
needed to distinguish ontogenetic changes from the phyletic ones, particularly
because most of the Upper Ordovician species are smaller than those in the Middle
Ordovician. In Hemicosmites it is possible to distinguish between ‘early’ and ‘late’
forms. H. extraneus appearing in the lower Caradoc has arm facets concentrated in
a small area at the thecal summit (text-fig. 12a). During the Middle and Upper
Ordovician the facets migrated outwards from the summit and the species changed
their shapes from spindle to poppy-head types (text-fig. 12a). The changes can best
be studied in the radial plates which alter from being long and narrow to short and
broad. In addition, wedge plates expand in the same manner. The expansion of the
wedge plates is probably a consequence of the enlargement of the size of the oral
surface.

With changes of thecal shape the arm facets gradually increased in size relative
to the thecal diameter. In later species like H. papaveris and H. sculptus the facets may
be in contact with, and even incorporate, the upper part of the laterals. The most
extreme situation is reached in H. pyriformis (age unknown) where the arm facets
occupy the greater part of the lateral plates (text-fig. 8c). The radial plates are here
reduced even more than in other Hemicosmites species.

These changes as outlined above are stages in a phyletic trend that, if continued,
would lead to a Caryocrinites organization. A further reduction of the radial plates
involves the facets migrating further outwards from the mouth. Also, an increase
occurs in the number of oral cover plates organized in the manner found in H. papa-
veris, among others, with a central plate (text-fig. 20a). The addition of wedge plates
is also an indication of the same phyletic trend. The only other difference between
representatives of the two families involves the position of the periproct, which in
Caryocrinites is located at the upper border of a lateral plate. Phylogenetically this
could be achieved by an orally directed migration of the periproct, as suggested by
Jaekel (1899). One specimen of H. pyriformis, characterized as ‘kummerliches
Exemplar’, was recognized by Jaekel to have the periproct in a ‘ Caryocrinites position’
(Jaekel 1899, pi. 18, fig. 4). This highly abnormal situation shows that these two
representatives of two different families may be closely related. Most known species

BOCKEL1E: ORDOVICIAN CYSTOIDS

385

of Caryocrinites have numerous arm facets, but several species of this genus are
known with only three facets, as in Hemicosmites , e.g. Caryocrinites sphaeroidalis
(Miller and Gurley), C. stellatus, and C. tribrachiatus Frest (see Frest 1975). The
only constant difference between representatives of the two families is the number of
thecal plates. Both families are characterized by an equal number of basals and infra-
laterals of the same type. In Hemicosmites species nine laterals are normal, but up
to eleven such plates are present in anomalous specimens. Jaekel (1899, p. 295) noted
that the IL plates around the periproct, i.e. IL1, IL6, and IL5, are smaller than IL2,
IL3, and IL4. Jaekel stated that the width of the IL plates increased in Hemicosmites
during Ordovician time. He further suggested that this would lead to a gradual
reduction in size of L6, and this can be seen on available material. He then predicted
that a total reduction of L6 might have taken place. Such a reduction would lead to
a Caryocrinites organization.

I regard the phyletic changes in Hemicosmites as indicative of the gradual trans-
formation of morphological structures from those found in Hemicosmitidae to those
of the Caryocrinitidae. This and the fact that Upper Ordovician species of Caryo-
crinites occur in Sweden (C. septentrionalis Regnell 1945), and in Estonia during the
lower Silurian (C. ornatus ; F. Schmidt 1858, p. 221 ; Regnell 19486, p. 35), I take as
reasonable grounds to suggest that Caryocrinites is a descendent of Hemicosmites.
Other representatives of the two families are not known in sufficient detail to indicate
phyletic trends, but the variability in morphology of some Hemicosmites species
indicates that this genus could have given rise to other genera of Caryocrinitidae. The
latter family had a greater number of successful representatives than the Hemi-
comitidae, and this may be related to the construction and number of arms that occur
in some of the Caryocrinitid genera.

text-fig. 14. H. pyriformis von Buch. Plate arrangement and thecal pores as observed on
the holotype. Note that the arm facets reach well down on to the LL-plates. No pores are
found on the radials (cf. text-fig. 8c). F— arm facets; ruled lines indicate parts not exposed.
Distribution of pores gives minimum information. More pores may be present but have not
been observed. Thick dots in some laterals indicate strong abrasion of that part of the thecal
surface. All pores in lateral series are sieve-pores.

386

PALAEONTOLOGY, VOLUME 22

SUMMARY AND CONCLUSIONS

This study covers only a limited portion of the existing material of Hemicosmites
species. Altogether about 100 specimens have been studied, belonging to seven or
eight species, of which by far the most belong to H. extraneus and H. papaveris.
Many of the observations may be relevant for all species of the genus, but when
considering species variation and the diversity of recent faunas it is not possible to
judge whether all the observations are typical for all species. The remaining species
are being studied in Moscow by Professor Hecker, and this will add valuable informa-
tion to complete the picture of this genus. It seems likely, however, that the genus as
a whole was limited to the Baltic Basin, that it occurred exclusively in Middle and
Upper Ordovician shallow-water environments, and that no more than about twenty
species existed. In fact there is a possibility that some of the existing species may be
regarded as synonyms when studied in detail.

The degree of variability within the described species has shown the importance of
using as large collections as possible for taxonomic discrimination. This applies
particularly to H. variabilis which, if single specimens had been found at different
localities, would easily have ranked as at least two, possibly three species. Other
species, such as H. extraneus, seem remarkably constant, though some variation does
exist. The variation in number of thecal plates in each of the plate series has made it
necessary to regard the plate numbering as only a convenient way of grouping them,
and suggests that any attempt to compare individual plates of Hemicosmites with
those of other genera of other families may, at present, be pure speculation.

The distribution and function of the thecal pores seem to be important for the
animal, as in other cystoids. In fact, all cystoids studied in detail have specialized
thecal pores that underwent modifications during their evolutionary history (text-
fig. 13a). Hemicosmites, with its typical endothecal pore system, is only one lineage
of experiments in respiratory specialization. From the available material the forma-
tion and distribution of the pore system over the theca of various species can be
recognized. Five different patterns of pore-rhomb distributions exist, and it is possible
to see these structures in terms of function and the utilization of current patterns
around the theca to carry waste water away (text-fig. 7). Because of poor preserva-
tion, this study has not incorporated specimens with abnormal plate distribution.
Such studies, however, may be important for understanding details of pore function.

Important ontogenetic and phyletic changes are found in the formation of inter-
calated plates, such as wedge plates, and the number and distribution of oral cover
plates. The sequence of formation of these plates, and their correlation with tegminal
plates of Caryocrinites or even other genera of other families in the same super-
family, are obscure. Nevertheless, during the Ordovician an increase in both size and
number of such plates enlarged the oral area. This may have been effected by the
increased size of the arm facets. Larger facets may indicate longer arms. The facets,
and therefore the anus, also came to be situated further apart and together all these
changes may have increased the efficiency of food collecting. Specialization of the
thecal pores and the evolution of the arms (of which we have only indirect evidence)
seem to be the most important evolutionary changes.

The locations of gonopore and hydropore have often been disputed, since in fact

BOCKEL1E: ORDOVICIAN CYSTOIDS

387

no typical gonopore is present in Hemicosmites. There were accessory furrows
adjacent to, and connected with, the ambulacral furrows. As Jaekel suggested, these
may have borne papillate structures that could have contained gonads and them-
selves have been sexual structures. It may be argued that such structures could even
serve more than one function.

The subvective system appears to be very similar to that of Caryocrinites, and this,
together with other morphological features that are manifest both ontogenetically
and phylogenetically, has led to the suggestion that Caryocrinites is a descendent of
Hemicosmites, even though they are placed in two different families (Caryocrinitidae/
Hemicosmitidae).

SYSTEMATIC PALAEONTOLOGY
Superfamily hemicosmitida Jaekel, 1918

Diagnosis. Theca composed of circlets of plates with three or four BB, six to ten
ILL, eight or more LL and a variable number of RR plates. Rhombs with inner side
of strongly folded delicate laminae perpendicular to suture, each pore usually
terminating in a simple pore at one end and a sieve-pore at the other. Lower
Ordovician-Lower Devonian.

Remarks. The diagnosis is emended to take account of Thomacystis Paul (1969) in
the Thomacystidae. At present four families are assigned to the Hemicosmitida,
i.e. Hemicosmitidae, Caryocrinitidae, Thomacystidae, and Heterocystitidae. How-
ever, the latter may be a crinoid (Paul 1969, p. 191). Similarities between Caryo-
crinites and Hemicosmites may be due to a closer relationship than recognized by
some authors, but differences are sufficiently clear to separate them into two families.

Family hemicosmitidae Jaekel, 1918

Diagnosis. Family of the Hemicosmitida with four BB, six ILL, eight or nine LL, and
a circlet of nine RR plates. Variable number of oral cover plates and inserted plates
in the oral area. Facets few, one in each of the three radii, or occasionally two in each
radius. Periproct lateral, never above LL.

Remarks. At present three genera can be included within the family : Hemicosmites
von Buch, 1840, Tricosmites Jaekel, 1918, and Oocystis Dreyfuss, 1939. Oocystis has
a plate arrangement similar to that of Hemicosmites, containing 4BB, 6ILL, 8LL,
and ‘a set of small plates present above LL’ (Kesling 1967, p. S225; = radials). The
oral surface of Oocystis is not known in detail. Its systematic position should thus
be left open at present. Tricosmites has a periproct in the same position as in Hemi-
cosmites', other features also suggest closeness to Hemicosmites.

Corylocrinus von Koenen, 1886, formerly assigned to the Hemicosmitidae, cannot
now be placed there because it lacks radial plates. The variable number of plates in
the oral area puts this genus closer to Caryocrinites than to Hemicosmites, though it
is very much in an intermediate position. Unfortunately, no material of Oocystis,
Tricosmites, and Corylocrinus has been available for this study. For Corylocrinus,
however, important diagrams have been presented by Renard (1968).

388

PALAEONTOLOGY, VOLUME 22

Genus hemicosmites von Buch, 1840

Type species. Hemicosmites pyriformis von Buch, 1840.

Synonyms. Hexalacystis Haeckel, 1896 (Type: Hemicosmites verrucosus Eichwald, 1856). See also Jaekel
1899, p. 307.

Diagnosis. A genus of Hemicosmitidae with four BB, six ILL, nine LL, and nine RR.
Oral area with variable number of cover plates and inserted plates. Periproct situated
at the border of IL1 :IL6 and L9, except in one species, where it is along the suture
IL1 :IL6. Hydropore present at upper part of R9. Three short ambulacral grooves,
ending in large facets at approximately 120° intervals, each facet generally shared by
two RR and two RR and one L plate. One R without facets between each pair of
facet-bearing RR. With or without wedge plates.

Remarks. A gonopore has not been found on the theca. Jaekel (1899, p. 308) noted the
presence of ‘additional brachiole facets’. Such structures have been found in some
specimens of different species, anticlockwise to the larger facets, and with ambulacral
furrows branching and leading to them. The minor facets are small and have no real
facetal area. They may have been supports for structures containing gonads (see p. 377).
The hydropore is located in the upper part of R9 and may be an irregular hole or slit,
surrounded by complex knobs, folds, etc. These give a rugose pattern to that part of
the plate, and often obscure the pore hole. The basal portion of the holotype of
H. extraneus has an additional plate between the BB and ILL series (bordering
B1 :B4:IL6:IL5). This must be regarded as abnormal and is not apparent in any
other material available to me. One specimen of H. papaveris has ten LL plates, and
this is also abnormal.

Jaekel (1899) distinguished three groups of Hemicosmites species. Group A, which
includes the type species, has a globular theca with nodes and facets reaching down
on to three laterals. Group B, including H. extraneus and H. sphaericus, contains
species with oval to pyriform thecae with facets adjacent to the mouth, not reaching
to the laterals, and nodes on three or all laterals. The third group has cup-shaped
thecae, flattened above the strong, but not always equally developed nodes on the
laterals (Jaekel 1899, p. 310). H . papaveris can be assigned to the latter group, although
some specimens also seem to belong to group A in that the facets extend far down on
to the laterals. Whether Jaekel’s grouping reflects differences sufficiently great to
erect new taxa cannot be established at present.

Details of the type specimens of the species described below are given on p. 365.

Hemicosmites pyriformis von Buch, 1840

Plate 43, figs. 3-5; text-figs. 8c, 14, 15

1840 Hemicosmites pyriformis von Buch, p. 59, pi. 1, figs. 1-3, 6-8, 11, 13.
1844 Hemicosmites pyriformis von Buch; von Buch, p. 108, pi. 1, figs. 11, 12.
1899 Hemicosmites pyriformis von Buch; Jaekel, p. 97, pi. 17, fig. 6.

1943 Hemicosmites pyriformis von Buch; Bassler and Moodey, p. 166.

1945 Hemicosmites pyriformis von Buch; Regnell, pp. 98, 100.

1967 Hemicosmites pyriformis von Buch; Kesling, p. S219, fig. 120, la-b.

BOCKELIE: ORDOVICIAN CYSTOIDS

389

Diagnosis. Hemicosmites with smooth globular theca, at least one wedge plate, facets
below thecal summit, system 4 somewhat reduced, systems 3 and 5 not present. Holes
in sieve-pores radially arranged, numerous ; all thecal plates short, no knobs present.

Material. One specimen, the holotype, has been seen for this study. Jaekel ( 1 899, fig. 69) also figured another
specimen and there is one specimen in the B.M.(N.H.) London labelled H. pyriformis, but it is probably
misidentified.

Description. Thecal outline globular. Mouth situated on the summit of the theca and present above three
facets of unequal construction. The hydropore is present at the upper left margin of R9 (text-figs. 8c, 15).
The circular anal opening, lacking a surrounding rim, has five movable triangular plates covering the anus.
A stem was present in life. Thecal height/width ratio is 1-09 (text-fig. 12b). One wedge plate is present
between R6 and R7. Plate thickness about 1 mm. Basals of moderate size (height/width ratio 0-9), infra-
laterals of moderate length (h/w ratio 1-2) and laterals long (h/w ratio 1-5). Three laterals have a rect-
angular shape. The plate surface is almost smooth. Simple pores have a tube, less prominent than in H.
extraneus. No umbones are present on the laterals. Rhombs of system 4 are almost completely reduced,
and of systems 3 and 5 completely reduced (text-fig. 14). Complete rhombs occur across all the intercirclet
sutures. In the basals sieve-pores have mostly four, but occasionally up to eight, pore holes. Due to abrasion
of the surface these pores are not well preserved (text-fig. 1 3c). Sieve-pores are present in small depressions
which rarely coalesce, and individual pores lack rims. The pores are mostly elongated, often radially
arranged. Simple pores are not connected by rims, but occur as single cylinders. Pore diameter is about
0-20-0-2 1 mm. No LL:RR rhombs are present.

The mouth, situated at the thecal summit, was completely covered in life with plates associated with the
ambulacra. These strongly ornamented cover plates continue on to the arms (text-fig. 8c ; PI. 43, fig. 3) in
a similar fashion to that of Lovenicystis Regnell, 1945 and Echinosphaerites (Bockelie, work in preparation).
A series of smooth biserially arranged arm plates (text-fig. 8c) grew from the contact between the radials
and the oral cover plates. Their shapes and sizes are variable. On the holotype the plates forming the arm in
radius I have covered R2 almost completely on their left side, thus being in contact with L2 (text-fig. 8c).
Arm plates of radius III cover R5 completely. Arm plates of radius V have an arrangement more typical of
other Hemicosmites species.

Arm facets are tongue-shaped and reach well down on the laterals. In radius I four plates are incorporated
in the facet: Rl, R2, LI, and L2; in radius III R4 (part of this plate is reduced or not exposed), parts of
R6 (remaining covered by arm plates), L4 and L5 (R5 is not exposed and may be reduced). In radius V
parts of R7, R8, and L7 are incorporated in the facet (text-figs. 8c, 15).

text-fig. 15. H. pyriformis von Buch. Holotype in oral and lateral view.
Terminology as in text-fig. 3. Note that arm facets reach on to laterals. Radials
reduced or assimilated. Oral cover plates numerous. Scale bar = 1 mm.

390

PALAEONTOLOGY, VOLUME 22

Remarks. H. pyriformis differs from all other known Hemicosmites species in the
construction of the oral area. No other species has such strongly developed arm plates,
nor shows such a degree of reduction of radial plates. There are no LL : RR rhombs
in the type. Whether this is constant in the species is not known at present. Although
known only from one specimen, H. pyriformis may differ sufficiently from all other
species generally referred to Hemicosmites to necessitate the eventual erection of
a new genus for these species. However, I consider that more material of H. pyriformis
is needed to verify whether the differences as outlined above are constant before
making such a split at the generic level. At present this cannot be safely established;
in fact, the type specimen of H. pyriformis may be simply an aberrant specimen.
A specimen referred to as H. pyriformis in the collections at the British Museum
(Natural History; E7592) does not have the same features as the type specimen, but
instead the facets lie closer to the mouth and do not appear below the radials, which
also have pores. Of other specimens also referred to H. pyriformis in the same col-
lection, E29094 may belong to H. malum. A further specimen, (E29093), strongly
abraded, bears holes or depressions caused by an organism that at some stage sat on
the thecal surface in the basal portion. These depressions are similar to those found
in H. malum in other collections.

Hemicosmites sphaericus sp. nov.

Plate 41, figs. 2, 4, 6; text-figs. 6b, 15, 16

1936 Hemicosmites sp. ; Thorslund, p. 26.

1945 Hemicosmites extraneus\ Regnell {non Eichwald 1860), p. 100, pi. 3, figs. 13-14.

1948 Hemicosmites extraneus\ Regnell {non Eichwald 1860), pp. 30-31.

1968 Hemicosmites sp. nov. ; Paul, p. 729, pi. 138, fig. 1.

Diagnosis. Hemicosmites with variable thecal shape; mostly spherical; smooth
surface; wedge plates developed; facets just below thecal summit; slight reduction of
horizontal system 4 ; holes in sieve-pores radially arranged ; all plates relatively short ;
knobs not normally present.

Material. Fourteen more or less complete specimens and several isolated plates in RM, UM, and PMO
collections, from Kullsberg Limestone, flank facies (upper Llandeilo-lower Caradoc), Kullsberg Quarry,
Kullsberg, Dalarna, Sweden.

EXPLANATION OF PLATE 41

Figs. 1, 3, 5. Stereophotograph of Hemicosmites extraneus Eichwald, RM Ec27604, lateral, oral, and basal
views, > 1 ; Odinsholm, Estonia.

Figs. 2, 4, 6. Hemicosmites sphaericus sp. nov., Holotype (UM D325 = Ec95 of Regnell 1945), lateral, oral,
and basal views, x 1 ; Kullsberg Quarry, Sweden.

Figs. 7, 8. H. variabilis sp. nov., cast of PMO 79081, lateral and oral views, x 1-5. Kalvsjo, Hadeland, Oslo
Region, Norway.

Fig. 9. H. variabilis sp. nov., cast of PMO 97079, lateral view; locality as for fig. 7; x 1-5.

Fig. 10. H. sculptus sp. nov., cast of Holotype (PMO 97066), oral area, note strongly granulated surface,
x 1 ; 0. Svartoy, Ringerike, Oslo Region, Norway.

Figs. 11, 12. H. variabilis sp. nov., cast of PMO 97075, oral area and lateral view, x4; Note three spiny
laterals. Locality as for fig. 7.

PLATE 41

BOCKELIE, Hemicosmites

392

PALAEONTOLOGY, VOLUME 22

text-fig. 16. H. sphaericus sp. nov. Holotype (UM D325 = Ec95 as used by
Regnell 1945). Terminology as in text-fig. 3. Note that arm facets do not reach
on to the laterals. One wedge plate is present (dotted). Scale bar = 1 cm.

Description. Thecal outline generally oval to globular, tapering slightly towards the base. Mouth situated on
the thecal summit and usually above or level with the three facets (text-fig. 16). The hydropore is present on
the raised portion of R9. The circular anal opening may have a rim, and may also project slightly outwards
(text-fig. 15). Five movable triangular plates cover the anus. A stem was present in life. Thecal height/
width ratio varies between IT and 1-4, but 1-2 is the average (text-fig. 12). Wedge plates present between
R3 : R4 and R6 : R7, but may occasionally also be present along other sutures. Plate thickness about 1 mm.
Basals short (height/width ratio 0-9), infra-laterals (h/w ratio 1-2) and the laterals and radials are shorter
than in H. extraneus (text-fig. 17). Plate surface smooth, but uneven in well-preserved specimens. The simple
pores have a thin tube, less prominent than that of H. extraneus. The umbones of the laterals below the
facets may have weakly developed nodes.

EXPLANATION OF PLATE 42

Fig. 1. Hemicosmites variabilis sp. nov., cast of Holotype (PMO 97079), lateral view, xl-5; Kalvsjo,
Hadeland, Oslo Region, Norway.

Fig. 2. H. variabilis sp. nov., cast of PMO 97077, basal view, xl-5; locality as for fig. 1.

Fig. 3. H. variabilis sp. nov., cast of PMO 97096, internal view of base, xl-5; locality as for fig. 1.

Fig. 4. H. variabilis sp. nov., cast of holotype (PMO 97079), oral view, xl-5; locality as for fig. 1.

Figs. 5, 6. H. papaveris sp. nov., cast of holotype (PMO 94644), oral view, xl-5; Encrinite limestone
(Caradoc), Steinvika, Langesund, Oslo Region, Norway.

Figs. 7, 8. H. variabilis sp. nov. ; 7, cast of IL-plate (PMO 97097) ; 8, cast of L-plate (PMO 97098) ; locality
as for fig. 1 . x 4.

Fig. 9. H. papaveris sp. nov., cast of base of PMO 94628; locality as for fig. 5. x 1-5.

Fig. 10. H. papaveris sp. nov., cast of oral area of PMO 94635, showing wedge plates at the boundary
between the lower radials (R9) and the triangular mouth; locality as for fig. 5. x 2.

Fig. 1 1 . H. ? sp. A, PMO 20147, presumably an IL-plate ; note the strongly developed half-rhombs of vertical
system, x 2; ‘Stage’ 3c/3 -y (Arenig-Llanvirn), Vaekkero, Oslo.

Fig. 12. H.l sp. C, PMO 64616, lateral plate, larger than others found in the Oslo Region, x 1 ; ‘Stage’ 5a
(Ashgill), Gran, Hadeland, Oslo Region, Norway.

Fig. 13. H. papaveris sp. nov., cast of PMO 94606 to show presence of stem adjoint to the base, xl-5;
locality as for fig. 5.

PLATE 42

BOCKELIE, Hemicosmite ;

394

PALAEONTOLOGY, VOLUME 22

Complete rhombs are developed across all intercirclet sutures. Only between the laterals and the radials
do some irregularities occur, with half rhombs in a few places. Occasionally some rhombs are also reduced.
Between sutures of infra-laterals weakly developed rhombs are present, but are not seen in IL4 : IL5. In the
basals a reduction occurs in some specimens between B 1 : IL2. Incomplete rhombs are reduced between most
lateral plates of system 5, but present between L8 :L9. In the basals each sieve-pore has four holes; sieve-
pores of the laterals may have varying numbers of such holes (maximum 20) depending on the age of the
animal, and whether or not it is the last-formed pore (text-fig. 22). PMO A35440 (12 mm thecal diameter)
has few holes in sieve-pores of laterals, and always four holes in the last-formed sieve-pore (closest to the
plate suture). Sieve-pores of radial plates generally have four holes only. The sieve-pores are always set in
shallow depressions which rarely coalesce, and individual pores lack rims. The pores may be circular
(0-07-0T0 mm diameter) or elongated, and are often arranged radially. Simple pores are also set in slight
depressions and have weak rims. The internal diameter of simple pores is generally about 0-20-0-22 m.
PMO A35424 was cut and polished and shows calcified dichopores, similar to those of H. extraneus (PI. 44,
figs. 5, 7, 8).

The mouth, situated on the thecal summit, was completely covered in life with plates associated with the
ambulacra. The facets are located immediately below the thecal summit. Plates have been found leading
down on to the facets, including one or two biserially arranged arm plates, probably similar to those of
Caryocrinites (Sprinkle 1975).

text-fig. 17. H. extraneus Eichwald. RM Ec27604. Terminology as in text-
fig. 3. Note the short ambulacral furrows, and arm facets not reaching on to
laterals. Spindle shape common. Ruled lines = orals not exposed. Scale
bar = 1 cm.

Remarks. Of known Hemicosmites species, H. sphaericus is close to H. extraneus,
but differs in having a smooth external surface, whereas the surface of H. extraneus
is rugose. Further, the thecal outline of H. sphaericus is more globular, and may
occasionally also be pear-shaped. All thecal plates tend to be shorter than those of
H. extraneus (text-fig. 17) and the thecal plates of H. sphaericus do not have strongly
impressed sutures. In adult specimens H. extraneus has more holes in individual sieve-
pores and strongly calcified dichopores within the theca. H. sphaericus differs from

BOCKELIE: ORDOVICIAN CYSTOIDS

395

H. oelandicus in that the latter is more spherical and has shorter IL plates (text-
fig. 18). Wedge plates occur in H. sphaericus, but have not been seen in the unique
type of H. oelandicus , which is not sufficiently well preserved to allow comparisons
of plates of the oral area or the pore systems.

Hemicosmites papaveris sp. nov.

Plate 42, figs. 2-5; text-figs. 19, 20a, 21a-d

Diagnosis. Poppy-head shaped species with slightly granulated surface; wedge plates
developed; facets level with thecal summit; reductions of most horizontal pore
systems ; holes in sieve-pores usually radially arranged ; basals and infra-laterals long ;
laterals and radials short ; knobs developed on three or all laterals.

Material. Several artificially produced moulds, of which fifty more or less complete specimens and numerous
plates were cast. All specimens in PMO collection, from inside bioherm of Encrinite Limestone (upper
Caradoc), Steinvika, Langesund, Oslo Region, Norway.

Description. Theca poppy-head shaped with slightly convex or flattened oral area, almost straight sides,
and concave basal plates. Height/width ratio of theca varies between T3 and 1-6 (average T4). Mouth
situated on thecal summit and usually level with the three facets. A possible hydropore is present at the upper
raised portion of R9. The circular anal opening is covered with five triangular plates ; in PMO 94629 the
anal opening is quadrangular, with four cover plates only (text-fig. 21c, d). Wedge plates are present in
most specimens, but best seen in adults. Thin lath-like additional plates are present in PMO 94619 (text-
fig. 21b) along the upper portion of sutures R3:R4 and R6:R7. Plate thickness is about 1 mm. Height/
width ratio of basals T05 (text-fig. 23), of infra-laterals T3, of laterals 1 -4, and of radials T5. Most specimens
have nine laterals, but PMO 94619 and 94633 each have ten lateral plates.

Due to severe pressure solution the exact configuration of pore rhombs in individual specimens is difficult
to establish. A composite diagram shows only slight reduction of most pore systems ; system 5 is almost
completely reduced. Traces of calcified dichopores were found, but severe recrystallization obliterates most
of the thecal canals. The peristome is covered by a number of oral cover plates, one central plate in young
specimens (text-fig. 21a) and several plates in adults (text-fig. 19). Arm facets are often 2-5 mm in diameter.
The size and width of corresponding ambulacral furrows are dependent on the size of the individual. In
PMO 94635 parts of the arm plates are present. When the oral cover plates are lost, a triangular mouth can
be seen. A hydropore is located in the upper portion of R9 and in contact with the oral cover plates or
wedge plates; the slit opens within a tumid, rugose area. In one specimen a stem fragment, 2 mm diameter
and 3 mm long, is present (PI. 43, fig. 13). Numerous root structures occur in the bioherm with this species,
and the lack of discs and other attachment structures may indicate that a root structure was present in
H. papaveris.

Remarks. H. papaveris belongs to group C (Jaekel 1899, p. 310) comprising poppy-
head shaped Hemicosmites species, often with strongly developed nodes on the upper
half of the laterals. It differs from the somewhat older H. rudis in having more faintly
developed ridges with simple pores. It also differs from H. pocillum in the lack of
knobs on the umbones of the infra-laterals.

Hemicosmites variabilis sp. nov.

Plate 41, figs. 11, 12; Plate 42, figs. 1-4; text-fig. 22

Diagnosis. Hemicosmites with variable thecal shape; smooth or rugose surface orna-
ment ; wedge plates present, facets just below or level with thecal summit ; reduction
of horizontal systems 4 and 5; sieve-pores with few holes; ILL plates and LL plates
with strongly developed, often projecting, umbones; basals, laterals and radials
short; infra-laterals long.

396

PALAEONTOLOGY, VOLUME 22

text-fig. 18. H. oelandicus Regnell. Holotype (RM Ec4362). Terminology as
in text-fig. 3. Note that arm facets do not reach on to laterals. No proper
wedge plates are present, but one additional radial plate (dotted) is present.

Scale bar = 1 cm.

Material. Eight more or less complete specimens and several isolated thecal plates, all in PMO collections,
from flank facies of ‘Stage’ 5a lime mudbank (upper Ashgill), Kalvsjo Quarry, eastern side, Lunner, Hade-
land, Oslo Region, Norway.

Description. Thecal outline variable, with two different kinds occurring within the same collections: one
has very strongly developed projections on the umbones of the laterals, and weak projections on the infra-
laterals (PI. 41, figs. 7, 12), while the other has equally strong development of projections on the umbones
of both plate series (PI. 41, fig. 9). All specimens come from the same locality and strata. The mouth is
situated on the thecal summit and is usually level with the three facets (text-fig. 22). A hydropore is present
in the upper raised portion of R9. The anal opening is circular and may have a rim; it never projects out-
wards. No cover plates have been found on the anal opening. One specimen (PMO 97081) has a quad-
rangular anal opening. A stem was present in life. Wedge plates are present in most specimens, but are not
always easy to detect because the sutures are indistinct. Plate thickness is about 1-5 mm, in a few cases as
much as T9 mm. Basals are short (height/width ratio 0-9), infra-laterals long (ratio T3), laterals and radials
short (ratios 1-3 and 1-3 respectively), as seen in text-fig. 23.

The plate surface is smooth in some specimens and strongly rugose in others, but never has spiny orna-
ment. Simple pores are raised and relatively prominent, but very few in number. Rhomb systems 4 and 5

EXPLANATION OF PLATE 43

Fig. 1 . Hemicosmitesextraneus Eichwald, RM Ec5283, showing base of arm ; Odinsholm, Estonia (Llandeilo-
Caradoc). x T7.

Fig. 2. Caryocrinites ornatus Say, PMO A21227, showing base of arm; Silurian. Lockport, New York,
U.S.A. x 1-7.

Figs. 3, 4, 5. H. pyriformis von Buch, Holotype, Humboldt Museum, D.D.R., oral area, lateral view, and
basal view, ■ 1-3; Pulkowa, Leningrad district, strata unknown.

Figs. 6, 7. H. extraneus Eichwald, oral area and lateral view, X 1*5; Stage (Caradoc), Alliku, Estonia.
(A. Roomusoks Collection, no number).

Figs. 8-11.//. verrncoinjJaekel, Holotype, Leningrad Geol. Inst. 1/327. 8, lateral view; 9, IL-plate; 10, basal
view; 11, oral view, x 3.

Fig. 12. H. extraneus Eichwald, details of oral area of fig. 6; scar in anticlockwise position for arm facets:
additional facets, x 3.

PLATE 43

BOCKELIE, Hemicosmites, Caryocrinites

398

PALAEONTOLOGY, VOLUME 22

text-fig. 19. H. papaveris sp. nov. Holotype (PMO 94644). Terminology
as in text-fig. 3. Note poppy-head shape and arm facets which do not reach on
to the laterals in this specimen. The specimen is arranged differently in lateral
view from previous figures, because periproctal area is not exposed. Scale
bar = 0-5 cm.

text-fig. 20. Oral area of Hemicosmites species. A, H. papaveris sp. nov. (PMO 94638). b, H. sculptus
sp. nov. 1-9— radials; H— hydropore; af— accessory facet; ap— arm plate; F— arm facet;
n— nerve canal ; wp— wedge plate. In fig. B inserted square indicates details of shape of hydropore.
Scale bar = 1 mm.

BOCKELIE: ORDOVICIAN CYSTOIDS

399

seem to be completely absent and the remaining systems are weakly developed. Only traces of vertical
system 1 have been seen on the basals, and most sieve-pores of the other systems seem to have from two to
four relatively large holes.

Calcified dichopores were found in some specimens, whereas the internal surfaces of isolated plates
show only faint traces of dichopore walls. The pore system is set in deeply sunken areas of the plate interior,
particularly in basal and radial plates, but also on the inside of laterals.

In life the mouth was completely covered with plates associated with the ambulacra ; the facets are large
and level with the thecal summit. Oral cover plates are spinose and form a strongly irregular area. Arm
facets grew on to the lateral plates which were then incorporated in the facets (text-fig. 22). No arm plates
have been found.

A young specimen, only 5 mm in thecal diameter, has three very large, flaring laterals, giving it a stellate
appearance (PI. 41, figs. 11, 12). In young stages the oral area thus apparently grew faster relative to other
parts of the body.

text-fig. 21. a and B, oral area of Hemicosmites papaveris sp. nov. a, young specimen
with a central plate only, ruled (PMO 94642) ; b, specimen with two oral laths in radial
circlet, dotted (PMO 94619) ; c and d, anal pyramid of H. papaveris sp. nov. c, specimen
with four plates (PMO 94629) ; D, pyramid with the normal five plates (PMO 94605) ;
E, IL-plate of H. ? sp. A (PMO 20147). Note the asymmetrical distribution of pore holes
in a rhomb, typical of Hemicosmites species. Well-developed rhombs of series four
present ; somewhat unusual in Hemicosmites, but present in early species. Scale bar =
0-5 cm.

Remarks. H. variabilis shows resemblances to the possibly contemporaneous
H. verrucosus Eichwald (cf. Jaekel 1899, pi. 18, fig. 5). Photographs of the latter
species (PI. 43, figs. 8-11) indicate that it has more numerous pore rhombs, and less
strongly developed ornament which is often of a different type; in addition, the
umbones of the infra-laterals of H. variabilis are more strongly developed, com-
prising almost the entire plate. In extreme cases the infra-laterals of H. variabilis
may have a blunt spine-like shape. As mentioned above (p. 396), the variation in shape

400

PALAEONTOLOGY, VOLUME 22

of H. variabilis is so great that if the specimens had not been found at the same
locality and from the same strata, they would have been regarded as belonging to at
least two different species.

Hemicosmites sculptus sp. nov.

Plate 41, fig. 10; text-fig. 20b

1897 Caryocrinus sp. ; Kiaer, pp. 17, 75.

1945 ‘ Caryocrinus ’ sp. ; Regnell, p. 105.

19486 Caryocrinus sp.; Regnell, p. 37.

1967 Caryocrinites sp. ; Kesling, p. S141, fig. 64.

Diagnosis. Hemicosmites with strongly granulated surface; raised arm facets; limited
number of oral cover plates; sieve-pores few, normally 2-4 irregular pores in each
sieve-pore.

Material. The holotype is from ‘Stage’ 5b ‘reef’ limestone (uppermost Ashgill-Hirnantian), East Svartoy,
Ringerike, Oslo Region, Norway. In addition two worn specimens (PMO 7288 and PMO 97047) and
a thecal plate (PMO 97067) are thought to belong to this species. All the additional specimens are from
‘Stage’ 5b at Stavnestangen, Ringerike.

Description. Mouth covered by six spiny, densely packed cover plates and two wedge plates. Oral surface
strongly irregular with rough sculpture and deep incisions between three arm facets. The facets are raised
well above the oral cover plates. The hydropore-bearing R9 is pointed and reaches above the level of the
arm facets. Sieve-pores were detected on R1 (3 pores, each with 3 pore holes), R3 (4 pores with 2, 3, 4,
and 5 pore holes), R4 (4 pores with 3, 2, 1 ?, and 2 pore holes) ; the average number of pore holes per sieve-
pore is 3 (text-fig. 13a). At least one wedge plate and six oral cover plates are present on the holotype
(text-fig. 20b).

Remarks. The strongly spinose surface and raised arm facets in H. sculptus differ
from those in other known Hemicosmites species. The proximal portion of the facets
has a gentle slope in an anti-clockwise direction (text-fig. 20b, PI. 41, fig. 10) in the
holotype, in the position where the additional facets are present in other species.
These slopes may be traces of similar structures.

Hemicosmitesl sp. A

Plate 42, fig. 11; text-fig. 21e

1865 ( 1)Hemicosmites pyriformis von Buch; Kjerulf, p. 4.

1882 Hemicosmites sp. ; Brogger, p. 42.

1945 Hemicosmites sp.; Regnell, p. 100.

1948a Hemicosmites ? sp.; Regnell, p. 33, fig. 5.

Material. One single plate of the lateral series (PMO 20147). This is the specimen mentioned by Kjerulf,
Brogger, and Regnell and subsequently described by Regnell (1948a) ; it is from the Orthoceratid Limestone,
‘Stage’ 30(8-7 (upper Arenig-lower Llanvirn), Vaekkero, Oslo, Norway.

Description. The plate is an IL, possibly IL3 (text-fig. 21e). The height of the plate is 12 mm, the width
9 mm. The plate thickness (1-5 mm) is like that of most Hemicosmites species. The surface is corroded, and
the pore holes of the simple pores are clearly visible ; the number of the pore holes depends on the size of
the seven half rhombs. The configuration of the pores indicates a strongly developed current system 4, more
strongly developed than in other known Hemicosmites species.

BOCKELIE: ORDOVICIAN CYSTOIDS

401

Remarks. The plate is clearly similar to other Hemicosmites species in some respects,
but the fact that only a single example is known makes it impossible at present to put
a firm generic name to it. Recent finds of plates of a possible caryocrinitid in con-
temporaneous deposits on Oland, Sweden, may indicate that the Hemicosmites! sp.
A plate could also be a caryocrinitid, but more and better-preserved material is
needed to decide this. However, the plate described here is more than twice the size
of those from Oland and is certainly not conspecific. Some uncertainties as to the
precise stratigraphical position of the plate exist. The lithologies in which it was
found may indicate the upper half of ‘Stage’ 3c/3 or all of ‘Stage’ 3cy; in either case
it is very close to the Arenig-Llanvirn boundary.

text-fig. 22. Oral and lateral view of H. variabilis sp. nov. Holotype (PMO
97079). Terminology as in text-fig. 3. Note the strongly developed nodes on
all laterals, and that at least two facets reach well down on to the laterals.

Several specimens have all three facets reaching this far down. Note also the
large periproct and irregular shape of theca. Scale bar = 0-5 cm.

Hemicosmites! sp. B

Text-fig. 23

Material. One specimen only, a basal portion of a theca (PMO 97095), from ‘Stage’ 5a (upper Ashgill),
Kalvsjo Quarry, Lunner, Hadeland, Oslo Region, Norway.

Description. Two basal plates, B2 and B3, and three infra-laterals are preserved. The theca might have been
about 40-45 mm high. Apparently four basals were present. The surface ornament of the plate is strongly
irregular and deep incisions exist between contiguous plates. Because of partial pressure solution, details
of the plate surface are missing, and no sieve-pores can be seen. The open pores on the other hand are
fairly well preserved and their distribution differs somewhat from that normally found in Hemicosmites
species in that they all occur on the crests of strong ridges. Intra-circlet rhombs are present on at least one
suture, indicating that pore system 4 was present.

402

PALAEONTOLOGY, VOLUME 22

Remarks. Since the distribution of pores differs from other species referred to
Hemicosmites, and also from patterns seen in Tricosmites and Caryocrinites
(Caryocrinitidae), the generic affinity of this specimen is left open at present.

text-fig. 23. A, average plate height/plate width of basals (B),
infra-laterals (IL), laterals (L), and radials (R) of various
Hemicosmites species. Note that most plates tend to become
shorter with time (1-7). 1 — H. extraneus; 2 —H. pyriformis
(stratigraphical position unknown); 3 —H. oelandicus; 4—
H. sphaericus sp. nov. ; 5 — H. papaveris sp. nov. ; 6 — H. variabilis
sp. nov.; 1—H. sculptus sp. nov. b, plate diagram of Hemi-
cosmitesl sp. B (PMO 97095). Pores few, size of specimen large
(scale bar = 1 cm), c, section through a plate of H. sphaericus
sp. nov. (PMO A 35424/2) showing size and distribution of
dichopores inside the theca. Note that two groups of dichopores
converge and that they do not reach equally deeply into the
theca. Scale bar = 0-5 cm.

EXPLANATION OF PLATE 44

Figs. 1-3. Hemicosmites extraneus Eichwald, Holotype, oral and lateral views, and details of lateral side;

Leningrad University Geol. Dept. 1/3207; D3 (Caradoc), Spitham, Estonia. l-2x 1 ; 3 x 3.

Fig. 4. H. variabilis sp. nov., cast of inner side of IL-plate (PMO 79099), x 2 ; locality as for PI. 42, fig. 1 .
Fig. 5. H. extraneus Eichwald, sagittal section of RM Ec5282, showing dichopore canals inside theca, x 1 ;
Odensholm, Estonia.

Fig. 6. H. variabilis sp. nov., cast of inner side of L-plate (PMO 97100), x 2; locality as for PI. 42, fig. 2.
Figs. 7, 8. H. sphaericus sp. nov., saggital section of PMO A35424 to show dichopores inside theca ; Kullsberg
Limestone. Kullsberg, Sweden. 7, x 1 ; 8 x 8.

Fig. 9. H. pulcherrimus Jaekel, Holotype. Palaeontol. Inst. Moscow 15355 (257/657), lateral view, x2;
D3 (Caradoc), Sack, near Tallinn, Estonia.

PLATE 44

BOCKELIE, Hemicosmites

404

PALAEONTOLOGY, VOLUME 22

Hemicosmites! sp. C

Plate 42, fig. 12

Material. Several loose blocks with plates (PMO 64616, 68422, 68423), from ‘Stage’ 5a (upper Ashgill),
Gjovik farm, Gran, Hadeland, Oslo Region, Norway.

Description. Large loose thecal plates of a Hemicosmitesl species. Lateral plates measuring 20 by 12 mm
in maximum dimensions. Plates are often strongly abraded, showing no trace of surface ornament. Traces
of pores are present.

Remarks. The large size of the plates indicates that the complete specimen must have
been about 10 cm high. The large size of the theca might be an indication of close
relationship to the contemporary H. grandis Jaekel (1899).

Acknowledgements. I am grateful for the help given by a number of colleagues, and particularly Dr. C. R. C.
Paul (University of Liverpool, England) for the use of unpublished data and information, for valuable
discussions, and for critically reading the manuscript. Dr. Tanja Koren (VSEGEI, Leningrad) has given
valuable information concerning the whereabouts of Hemicosmites species, provided photographs of type
material, and also made contact with people who could add even more information. Professor R. F. Hecker
(Moscow) has provided both photographs and information on the location of some of the species. Professor
A. Roomusoks (Tartu University, Estonia (TE)) kindly loaned me some of his material for comparative
work. Material has also been put at my disposal by Dr. Valdar Jaanusson (Naturhistoriska Riksmuseet,
Stockholm, Sweden (RM)) and by Professor R. A. Reyment (Paleontologiska Institutionen, Uppsala,
Sweden (UM). Professor H. Jaeger (Humboldt Museum (HM), Berlin, D.D.R.) kindly loaned me the type
of H.pyriformis. Material has also been borrowed from the British Museum (Natural History— B.M.(N.H.),
England). Financial support has been provided by Nansenfondet (Oslo) and Gustav Lindstrom’s
Minnesfond (Stockholm) and is gratefully acknowledged. Professor G. Henningsmoen, Oslo, arranged
working facilities at Paleontologisk Museum (PMO) and kindly read and commented on my manuscript.
Thanks are also due to the technicians at the latter museum, and in particular to Mr. Aa. Jensen for modelling
a reconstruction of Hemicosmites.

REFERENCES

bassler, R. s. and moodey, m. w. 1943. Bibliographic and faunal index of Paleozoic pelmatozoan echino-
derms. Sp. Pap. geol. Soc. Am. 45, i-vi, 1-743.

bather, f. a. 1906. Ordovician cystidea from Burma. In reed, f. r. c. The Lower Palaeozoic fossils from the
northern Shan States, Burma. Mem. geol. Surv. India Palaeont. indica, N.s. 2, 3, 1-154, pis. 1-8.
bockelie, J. F. 1972. Diploporitic cystoids ( Echinodermata ) from the Ordovician of Norway with remarks on
some Swedish forms. Thesis (unpubl.), Oslo Univ.

— 1978. Variability of ambulacral structures in some diploporite cystoids. Thalassia jugosl. 12 (1), 31-39.
breimer, a. 1969. A contribution to the paleoecology of Palaeozoic stalked crinoids. Konkl. Nederl. Akad.

Wetensch.-Amsterd. Proc., Ser. B, 72 (2), 139-150.

— and macurda, d. b. jun. 1972. The phylogeny of the fissiculate blastoids. Konkl. Nederl. Akad.
Wetensch. Naturkl. 26 (3), 1-390, pis. 1-34. Amsterdam.

brogger, w. c. 1882. Die Silurischen Etagen 2 und 3 im Kristianiagebiet und auf Eker. Kgl. Frederiks
Univ., Chra. Univ. Program, 2. sem., 1882.

buch, L. von. 1840. Uber Sphaeroniten und einige andere Geschlechter, aus welchen Crinoideen enstehen.
K. Akad. Wiss. Berlin, Verhandl. 1840, 56-60.

— 1845. Uber Cystoideen eingeleitet durch die Entwicklung der Eigenthumlichkeiten von Caryocrinus
ornatus Say. K. Akad. Wiss. Berlin, Abhandl. 1844, 89-116.

dreyfuss, m. 1939. Les cystoides de l’Ordovicien superieur de Languedoc. Bull. Soc. geol. Fr., Ser. 5, 9,
117-134, pis. 1-2.

eichwald, E. von. 1840. Sur le systeme silurien de l’Esthonie. Med. Hist. Nat. Acad. Med. St. Petersburg,
Jour. 1, 1-222.

BOCKELIE: ORDOVICIAN CYSTOIDS

405

eichwald, e. von. 1856. Beitrag zur geographischen Verbreitung der fossilien Thiere Russlands. Alte
Periode. Bull. Soc. Nat. Moscou, 29 (1), 88-127.

— 1860. Lethaea Rossica ou Paleontologie de la Russie. 1. Premiere section de Vancienne periode. v-xix+
681+8 pp., atlas, 59 pis., Schweizerbart, Stuttgart.

forbes, E. 1848. On the Cystidea of the Silurian rocks of the British Islands. Mem. geol. Surv. U.K. 2,
483-534.

frest, t. 1975. Caryocrinitidae (Echinodermata : Rhombifera) of the Laurel Limestone of southeastern
Indiana. Fieldiana, Geol. 30 (4), 81-106.

haeckel, e. 1896. Die Amphorideen und Cystoideen. Beitrdge zur Morphologie und Phylogenie der Echino-
dermen: Festschrift zum siebenzigsten Geburtstage von Carl Gegenbauer am 21 August 1896. Vol. 1.
179 pp., 5 pis., Engelmann, Leipzig.

hall, J. 1864. Account of some new or little known species of fossils from rocks of the age of the Niagara
Group. N.Y. State Cab. Nat. Hist. 20th Ann. Report, 305-401 (1867); originally printed in advance for
18th report, 1864 (Albany).

hecker, r. F. 1964. Class Cystoidea. In hecker, r. f. (ed.). Osnovy Paleontologii, 10, 35-45, pis. 2-5. Moscow.
[In Russian.]

hyman, l. h. 1955. The invertebrates. 4: Echinodermata. The coelomate Bilateralia. New York.
jaanusson, v. 1960. On the series of the Ordovician System. Rep. Int. geol. Congr., 21st Session, Norden, 7,
70-81.

jaekel, o. 1899. Stammegeschichte der Pelmatozoen 1, Thecoidea und Cystoidea. x+442 pp., 18 pis.,
Springer, Berlin.

1918. Phylogenie und System der Pelmatozoen. Palaont. Z. 3 (1), 1-128.

katz, s. G. and sprinkle, j. 1976. Fossilized eggs in a Pennsylvanian Blastoid. Science, N.Y. 1137-1139.
kesling, r. v. 1967. Cystoids. Pp. S268-S288. In moore, r. c. (ed.). Treatise on invertebrate paleontology.

Part S, Echinodermata. Vol. 1, part 1, Geological Society of America and University of Kansas Press.
ki/er, j. 1897. Faunistische Uebersicht der Etage 5 des norwegischen Silursystems. Vid. Selsk. Skr. 1.
Math. Naturv. kl. 1897, 3, 1-76. (Kristiania.)

kjerulf, th. 1865. Veiviser ved Geologiske Excursioner i Christiania omegn. Univ. Program. Christiania
(Oslo).

koenen, E. f. r. k. von. 1886. Ueber neue Cystideen aus den Caradoc-Schichten der Gegend von Mont-
pellier. Neues Jb. Miner. Geol. and Palaont. 1886, 2, 1-848. (Leipzig.)
mannil, R. m. 1966. Istoria razvitia Baltiiskogo Basseina v. Ordovike. [Evolution of the Baltic Basin during
the Ordovician.] Inst. geol. Akad. Nauk Estonoskoi SSR. Tallinn. [In Russian with English summary.]
moore, R. c. and jeffords, r. m. 1968. Classification and nomenclature of fossil crinoids based on studies
of dissociated parts of their columnals. Paleont. Contr. Univ. Kans., Echinodermata , Art. 9, 1 -86.
pander, c. h. 1830. Beitrdge zur Geognostisie des Russischen Reiches, xx+165 pp., 31 pis. -f 4B, 4C, 16B,
St. Petersburg.

PAUL, c. R. c. 1967. The British Silurian cystoids. Bull. Br. Mus. nat. Hist. (Geol.), 13, 299-355, pis. 1-10.

— 1968. The morphology and function of dichoporite pore-structures in cystoids. Palaeontology, 11,
697-730, pis. 134-140.

— 1969. Thomacystis, a unique new hemicosmitid cystoid from Wales. Geol. Mag. 106, 190-196, pi. 11.

— 1972. Morphology and function of exothecal pore-structures in cystoids. Palaeontology, 15, 1-28,
pis. 1-7.

regnell, G. 1945. Non-crinoid Pelmatozoa from the Palaeozoic of Sweden. A taxonomic study. Meddn
Lunds geol-min. Instn. 108, 1-225, pis. 1-15.

— 1948a. Echinoderms (Hydrophoridea, Ophiocista) from the Ordovician (Upper Skiddavian, 3c/3)
of the Oslo Region. Norsk geol. Tidsskr. 27, 14-57.

— 19486. An outline of the succession and migration of non-crinoid pelmatozoan faunas in the Lower
Paleozoic of Scandinavia. Ark. Kemi Miner. Geol. 26A, 1-55.

— 1960. The Lower Palaeozoic echinoderm faunas of the British Isles and Balto-Scandia. Palaeontology,
2, 161-179.

renard, H. 1968. Contribution a la revision des cysto'ides de TAshgill de la Montagne Noire. 124 pp., 25 pis.,
Montpellier.

salter, J. w. 1866. On the fossils of North Wales. Appendix to ramsay, a. c. The geology of North Wales.
Mem. geol. Surv. U.K. 3, 239-381, pis. 1-26.

406

PALAEONTOLOGY, VOLUME 22

schmidt, f. von. 1858. Untersuchungen iiber die Silurischen Formationen von Esthland, Nord-Livland
und Oesel. Arch. Naturk. Liv.-Est.-u. Kur lands, Ser. 1, 2.
sprinkle, J. 1973. Morphology and evolution of blastozoan echinoderms. Harvard Univ. Mus. Comp. Zool.
Spec. Publ. 1-283.

— 1975. The ‘arms’ of Caryocrinites, a rhombiferan cystoid convergent on crinoids. J. Paleont. 49, 1062-
1073, pi. 1.

sun, Y. c. 1936. On the early occurrence of some Ordovician and Silurian cystoids from Western Yunnan
and its significance. Palaeont. Novit. 1, 1-9.

thorslund, p. 1936. Siljansomradets brannkalkstenar och kalk-industri. Sver. geol. Unders. Afh., Ser. C,
398, 1-64, pis. 1-3.

verneuil, E. de and keyserling, a. de. 1845. In Murchison, r. i. et al. Geologie de la Russie d' Europe et des
montages de POural. Vol. II, Troisieme partie, Paleontologie. xxii,+ 511 pp., 43 pis., Londres et Paris.
Yakovlev, n. n. 1918. Some new data on Cryptocrinites and the connection between the Crinoidea and
Cystoidea. Ann. Soc. Palaeont. Russie, 2, 7-26. Petrograd, 1917. [In Russian with English summary.]

Manuscript received 22 March 1978
Revised manuscript received 18 September 1978

JOHAN FREDRIK BOCKELIE

Paleontologisk Museum
Sarsgate 1, Oslo 5
Norway

TAXONOMY AND OPERCULAR FUNCTION OF
THE JURASSIC ALGA STIC HO PO RELLA

by GRAHAM F. ELLIOTT

Abstract. Stichoporella stutterdi (Carruthers) Edwards is shown to differ from S. cylindrica (Lignier) Pia in the form
of the distinctive opercula which terminated the side-branches in the reproductive stage of development ; previous
accounts of the difference were based on damaged examples. The opercula of S. stutterdi support the suggestion of
Fily and Rioult (1976) based on S. cylindrica , as to their being formed immediately prior to cyst formation, and being
shed to allow cyst-dispersal. The microstructure of the opercula in S. stutterdi is described. S. cylindrica is recorded
only from the Middle Jurassic of northern France; S. stutterdi has a more northerly area of occurrence, overlapping
with S. cylindrica in France but also occurring in the English midlands.

Stichoporella is a large distinctive dasycladacean alga, occurring uncommonly
in the Middle Jurassic of northern France and rarely in England. The genus was
created for Goniolina cylindrica Lignier from the French Bathonian by J. Pia, who
showed that this species was different in structural plan from a true Goniolina (Lignier
1913; Pia 1923, p. 68). Later Edwards (1928) recognized that a problematic English
fossil earlier regarded as coral or crinoid, and described as a higher plant ( Aroides
stutterdi Carruthers), is in fact a second species of Stichoporella.

More recently Fily and Rioult (1976) have made a careful and very detailed
revision of Stichoporella cylindrica, based on a new specimen of this rare dasycla-
dacean. Rioult’s emended generic diagnosis (op. cit. , p. 40) may be translated as ‘thallus
cylindrical, not articulated, of large size, 15-16-mm diameter, with large axial cavity
without constrictions or swellings. Alternate euspondyl verticils of phloiophore-type
primary branches. At the exterior of the calcified cylinder, each pore (branch-
termination) is closed by a laminated cupuliform operculum, free of the (calcified)
cylinder and presumed deciduous, and closely inserted in the hexagonal areas (pores)
which are delimited by the (honeycomb) ridges of the cylinder. All the branches could
have been “fertile ampoules” (cladospore reproductive structures)’ (see text-fig. 1).

Apart from its large size and very simple structure (many other Jurassic dasycla-
daceans are much more elaborate), Stichoporella is unusual in its distinctive branch-
opercula, to which Fily and Rioult rightly draw attention, and largely because of
which they place Stichoporella in a new diplopore tribe, the Stichoporellinae (Fily
and Rioult 1976, p. 40).

Both Edwards and Fily and Rioult indicate that the only significant difference
between S. cylindrica and S. stutterdi lies in the form of these opercula, the two species
otherwise being closely similar in structure, proportions, and size. Edwards (1928,
p. 80) wrote, ‘It [S. stutterdi ] differs from S. cylindrica in the dentate and dovetailing
margins of the calcified membranes.’ Fily and Rioult (1976, p. 40) wrote that S. cylind-
rica apparently differs only from S. stutterdi by the absence of denticules at the inter-

[Palaeontologj, Vol. 22, Pari 2, 1979, pp. 407-412, pi. 45.]

PALAEONTOLOGY, VOLUME 22

text-fig. 1. Stichoporella\ diagram of structure. Left: surface pattern of opercula. Centre: radial swollen
branches, terminating in opercula, cysts in outer parts of branches, calcification shaded. Right : axial or
stem-cell cavity, to show initiation of branch- verticils.

section of the surface ridges delimiting the opercula. Their figure (op. cit., pi. 1, fig. 2)
clearly shows the smoothly rounded hexagonal outlines of these opercula.

An attempt by me to evaluate this character in S. stutterdi and so determine the
validity of the two species, led to the results set out below.

THE FORM OF THE OPERCULUM IN S. STUTTERDI

The English material is very limited, and comes from two localities only : Stones-
field in Oxfordshire (Bathonian) and Wittering, formerly in Huntingdonshire but now
Cambridgeshire (Bajocian). Both are localities long since disused and overgrown; the
fossils are from old collections. Stonesfield is the type-locality for the species, and
the material from there is best preserved.

The cylindrical algae occur more or less compressed, to show a pavement of the
opercula (PL 45, fig. 3). Most of these are represented by the underlying mould only,

EXPLANATION OF PLATE 45

Figs. 1-4. Stichoporella stutterdi (Carruthers) Edwards; Stonesfield Beds, Middle Jurassic (Lower
Bathonian); Stonesfield, Oxfordshire, England. 1. Single operculum, showing hexagonal-stellate out-
line, and gentle convexities and concavities on lower smooth well-preserved portion, x 20. Reg. no.
V. 10968. 2. Group of four opercula, showing characteristic flaked weathering which influenced earlier
descriptions, x 20. Reg. no. V.5585. 3. Surface view of part of curved ‘pavement’ of opercula of this
cylindrical fossil, x8. Reg. no. V. 10968. 4. Near-vertical thin-section of operculum, x80. Left: edge
of section ; right : matrix (black and white). The section between shows the finely layered nature of the
component calcareous lamellae : between these are developments of small calcite crystals, extensive on
the left, minor on the right. Reg. no. V. 5585a. Registration numbers above are those of the British Museum
(Natural History), Department of Palaeontology.

PLATE 45

ELLIOTT, Stichoporella

410

PALAEONTOLOGY, VOLUME 22

which can be recrystallized to a varying degree. Where the actual laminated oper-
culum is preserved, it has usually flaked or scaled away to a varying extent. Careful
search reveals the occasional near-perfect example. It is then seen that these are
rounded and hexagonal to stellate in outline, with a porcellanous or enamelled surface
appearance. From the flattened very slightly convex centre of each, alternate con-
vexities and concavities in the outer slopes cause the stellate outline (PI. 45, fig. 1),
much as in certain brachiopods, e.g. Zeilleria (Z.) quadrifida (Lmk). There is a
tendency during early weathering for the lamellar operculum to flake away dif-
ferentially on the slopes of the convexities, so giving a spiky stellate outline (PI. 45,
fig. 2), and this is the origin of earlier statements on the ‘dentate’ outline and its inter-
locking properties. This is not so; the un weathered opercula are separated by the
sinuous course of the terminations of the underlying calcareous coatings of the
branches (the skeletal honeycomb structure). Whilst normally concavity is opposed
to convexity in adjacent opercula, two opposed concavities sometimes occasion
widening of the interopercular skeleton. Each rounded hexagonal-stellate operculum
fits its own branch-termination (‘honeycomb cell’) and seems to have rested in the
aperture.

S. stutterdi then, does differ consistently from S. cylindrica in operculum-outline,
if not quite as previously described, and the character is accentuated by weathering.
The Bajocian example available is very coarsely recrystallized, but the stellate outline
is still recognizable. How significant is this one difference between the two described
species? This is now considered in the light of the assumed function of the operculum.

STRUCTURE AND SIGNIFICANCE OF THE OPERCULUM IN
STICHOPORELLA

Fily and Rioult (1976) have drawn attention to the markedly lamellar nature of
the operculum in S. cylindrica, like a laminated watch-glass, and to the fact that this
dense white structure has been preserved differently from the main skeletal calcium
carbonate. They suppose the opercular flaking to be assisted by original interlamellar
films of organic matter within the operculum (and record a negative test for aragonite,
and the absence of silicification in this white calcareous structure). The opercula are
said to rest in the rounded-hexagonal branch apertures on narrow rims. They are
considered by these authors to be a special device behind which the reproductive
cysts developed, and which were shed when the latter were released.

In thin-section the calcareous material of the operculum of S. stutterdi (PI. 45,
fig. 4) is seen to be finely layered and transparent within the different component
lamellae. These are separated by planes of parting along which strings of small calcite
crystals have developed during diagenesis. In contrast, the wall-material of the
cylindrical skeleton is of uniformly coarse calcite crystals, representing the results of
diagenesis on the original aragonitic skeleton, assumed by analogy both with living
and with little-altered fossil dasycladaceans. This confirms an original difference in
the composition of the operculum.

ELLIOTT: JURASSIC STIC HO PO RELLA

LIFE HISTORY

In living dasycladaceans the early growth-stages are not calcified. Only during
later stages of growth does calcification develop to the usual extent for the particular
taxon; it is especially liable to develop around reproductive structures when these
individualize. This stage follows in the mature plant, on the break-up of the large
nucleus in the rhizoid and the swarming of daughter nuclei into the stem-cell and
branches, to originate the reproductive elements. These are contained in ‘fertile
ampoules’ of varying form, usually filled with cysts but sometimes shedding free
gametes direct, as in Dasycladus itself. Valet (1969) gives a detailed account of the
various sexual and reproductive mechanisms in living Dasycladales. Fily and Rioult
(1976) suggest that the outer swollen ends of the primary branches in S. cylindrica
functioned as fertile ampoules, with the opercula as special containing devices.

If this was so, I would reconstruct the life-history of S. stutterdi as follows. It was
an inhabitant of sheltered shallow coastal marine waters, as evidenced by the other
fossils at Stonesfield and, when mature, was a thick-stemmed, thick-branched green
alga, which eventually laid down aragonite in the mucilage between and coating the
branches, to give the outer cylindrical honeycomb skeleton. At the end of its short
life-span, probably one or at most two years, swarming of small nuclei up the stem-
cell and into the terminal portions of the side-branches initiated copious cyst-
formation. Hitherto, the rounded branch-tips had protruded beyond the skeletal
mesh, with an assimilatory function. On the cessation of vegetative growth, they
shrank. Against them, inside, from the biochemically changed cytoplasm of the
reproductive period, there was laid down intermittently lamellar calcareous matter.
I suggest this could have been calcium carbonate mixed with calcium oxalate, as is
known from the reproductive discs of living Acetabularia. Whatever it was, it behaved
differently during subsequent diagenesis to the granular skeletal aragonitic calcium
carbonate. If this is what happened, the outer lamella of each operculum was the
earliest formed. Subsequently, with cyst-ripening, the opercula were shed, the cysts
dispersed, and the individual alga died or was perhaps regenerated from the shrunken
rhizoid. Those now preserved as fossils were a minority, for some reason buried just
before completion of the cycle.

If this hypothesis is true, why did the branches of S. cylindrica and S. stutterdi shrink
differently, so occasioning the specific difference? Both species are recorded as
occurring in France, around the Paris Basin (Dutertre 1926a, b, Gardet 1952, Fischer
1969, Fily and Rioult 1976), but S. stutterdi also occurs further to the north, in the
English midlands. Were the two species perhaps ecophenes responding to differences
in the marine climate? I have suggested (Elliott 1977) that Stichoporella was
characteristic of a narrow climatic belt north of a Jurassic isocryme of major algal
significance; this could perhaps be significant. However, as fossils they are morpho-
logically distinct, and I confirm them as separate palaeontological species.

Acknowledgements. My thanks are due to staff at the British Museum (Natural History): to Messrs. J. V.
Brown and P. V. York for photography, to Mr. M. Crawley for preparation of the text-figure, and to
Mr. R. L. Hodgkinson for his skilful preparation of a thin-section from a tiny piece of specimen.

412

PALAEONTOLOGY, VOLUME 22

REFERENCES

dutertre, a. p. 1926a. Decouverte d’un Aroides dans l’etage bathonien du Boulonnais. C.r. somm. Seanc.
Soc. Geo I. Fr. 1926, 32-33.

— 1926 b. Decouverte d’un ‘Aroides’ dans le Bathonien des Ardennes. Annls Soc. geol. N., 51, 211-212.
Edwards, w. N. 1928. On the algal nature of Aroides stutterdi Carruthers. Ann. Mag. nat. Hist. (10) 1, 79-81.
elliott, G. f. 1977. Inferred isocrymal distribution of Jurassic dasycladacean algae in Europe, north
Africa and southwestern Asia. J. geol. Soc. Lond. 133, 363-373.
fily, G. and rioult, M. 1976. Stichoporella cylindrica (Lignier), dasycladacee dans le Bathonien de Cour-
tomer, Orne (Normandie). Position stratigraphique et precisions systematiques. Bull. Soc. geol. Nor-
mandie, 63, 33-44.

fischer, J. c. 1969. Geologie, paleontologie et paleoecologie du Bathonien au sud-ouest massif ardennais.
Mem. Mus. natn. Hist. nat. Paris, C 20, 320 pp.

gardet, g. 1952. Sur deux fossiles interessants du Jurassique haut-marnais. C.r. somm. Seanc. Soc. geol.
Fr. 1952, 324-325.

lignier, o. 1913. Vegetaux fossiles de Normandie. VII. Contribution a la flore jurassique. Mem. Soc. linn.
Normandie, 24, 67-105.

pia, j. 1923. Einige Ergebnisse neuerer Untersuchungen liber die Geschichte der Siphoneae verticillatae.
Z. indukt. Abstamm. u. Vererblehre, 30, 63-98.

valet, G. 1969. Contribution a l’etude des Dasycladales. 2. Cytologie et reproduction. 3. Revision syste-
matique. Nova Hedwigia, 17, 551-644.

Typescript received 5 April 1978
Revised typescript received 1 June 1978

G. F. ELLIOTT

Department of Palaeontology
British Museum (Natural History)
Cromwell Road
London SW7 5BD

MICROPALAEONTOLOGICAL STUDIES OF
THE UPPER JURASSIC AND LOWER
CRETACEOUS OF AND0YA,
NORTHERN NORWAY

by MAGNE L0FALDLI and BINDRA THUSU

Abstract. Seventy species of foraminifera are recognized and grouped into three assemblages. Assemblage 1, entirely
dominated by Haplophragmoides represents a restricted, marginal marine environment and is confined to the lower
part of Ratjonna Member of the Middle Volgian. Assemblage 2, dominated by Haplophragmoides in association
with Lenticulina, represents a shallow, open marine environment and is confined to the upper part of Ratjonna
Member of the Ryazanian. Assemblage 3, dominated by Nodosariidae and Glomospira , represents an open marine,
neritic environment and is associated with the Nybrua Formation of Valanginian-Hauterivian age.

Species of Haplophragmoides in assemblages 1 and 2 of the Volgian-Ryazanian are poorly preserved and left under
open nomenclature. However, these species are broadly comparable with forms reported from the Upper Jurassic-
Lower Cretaceous of north-west Europe and the Arctic areas. Assemblage 3 shows close similarity to the Valanginian-
Hauterivian microfaunas from north-west Europe. However, the dominance of calcareous species in assemblage 3
in Andoya and coeval beds in north-west Europe is in marked contrast to neritic faunas reported from Agardhfjellet,
Spitsbergen, where Early Cretaceous microfaunas are dominated by simple arenaceous forms. These faunal differences
are probably the result of substrate, latitudinal, or climatic factors.

The present investigation reports micropalaeontological analysis of surface and
mechanically excavated sections of the Upper Jurassic (Volgian) to Lower Cretaceous
(Hauterivian) sequence from Andoya, an island in the Vesteralen archipelago, in
northern Norway (text-fig. 1). The Jurassic and Cretaceous sequence (text-fig. 2)
consists predominantly of sandstones and shales resting unconformably on granitic
Pre-Cambrian basement. For detailed account on sedimentology and biostratigraphy
the reader is referred to Dalland (1975), Thusu and Vigran (1975), and Birkelund
et al. (1978). In the present paper a complete list of foraminifera is given and an
assessment of their biostratigraphic, palaeoecologic, and palaeobiogeographic
significance is attempted.

Foraminifera were recovered by boiling the samples in a weak solution of NaOH
for a short time before sieving in the usual manner.

STRATIGRAPHY

The Jurassic and Cretaceous sequence of Andoya is about 650 m thick (text-fig. 2) and
consists predominantly of sandstones and shales deposited in two small troughs. The
age of the sediments is Middle to Late Jurassic in the southern, and Early Cretaceous
in the northern trough. The Jurassic includes the Ramsa Formation and the bulk of
the Dragneset Formation (text-fig. 2). The Ramsa Formation consists of sandstones,
shales, and coal layers that did not yield any microfauna. The overlying Dragneset

[Paiaiontolog), Vol. 22, Part 2, 1979, pp. 413-425, pis. 46-47.]

414

PALAEONTOLOGY, VOLUME 22

Formation consists of three members, the youngest of which the Ratjonna Member,
a predominantly silty shale unit of the Middle Volgian-Ryazanian age, contains
agglutinated foraminifera. The Lower Cretaceous Nybrua Formation consists of
calcareous sandstone, siltstone, and marl of Valanginian-Hauterivian age. The Leira
and Skjaermyrbekken members of the Nybrua Formation contain abundant cal-
careous and agglutinated foraminifera. The two remaining units, Nordelva and
Helinesset formations, consist of sandstone and shale that contain sparse microfaunas.

PALAEOENVIRONMENT

Three foraminiferal assemblages are recognized in the Dragneset and Nybrua forma-
tions (text-fig. 2). These assemblages and their age and environmental interpretations
are summarized below.

Assemblage 1. This assemblage is restricted to the lower part of the Ratjonna Member
(text-fig. 2), correlated with the Pavlovia rotunda- Progalbanites zones of the Middle

L0FALDLI AND THUSU: NORWEGIAN MESOZOIC MICROFAUNAS 415

text-fig. 2. Selected foraminifera and dinocysts close to the Jurassic/Cretaceous boundary in Andoya, Norway.

416

PALAEONTOLOGY, VOLUME 22

Volgian. The assemblage is made up exclusively of poorly preserved arenaceous
specimens of Haplophragmoides. Nearly 100% of the assemblage consists of H. aff.
neocomiana (Chapman). Some of the individuals seem to be intermediate forms
between H. neocomiana (Chapman) and H. concava (Chapman). Some individuals
of H. aff. volgensis Myatliuk are also recorded in this assemblage. Although H. neo-
comiana is known to occur in the Lower Cretaceous of western Europe (Chapman
1894; Ten Dam 1948 ; Fletcher 1972) and H. volgensis in the Upper Jurassic of Poland
and U.S.S.R. (Bielecka 1975), all the forms recorded here are left under open nomen-
clature because of the poor preservation.

Species of Haplophragmoides are known to have wide environmental tolerances.
Chamney (1977) considered this genus to reach its optimum support in the normal
marine environment of the shelf-slope contact, at an approximate depth of 130 m.
The incoming of dinocysts Gonyaulacysta cladophora-perforans group in abundance
suggests a shallow marine environment. However, this poorly preserved, restricted
arenaceous fauna also could be the result of reduced oxygen supply. This is supported
by the presence of dark, laminated siltstones composing most of the Ratjonna

EXPLANATION OF PLATE 46

Taxa from assemblage 1, figures 1, 6; Volgian, Dragneset Formation, lower part of Ratjonna Member;
assemblage 2, figures 5, 8; Ryazanian, Dragneset Formation, upper part of Ratjonna Member. All other
taxa from assemblage 3, Valanginian-Hauterivian, Nybrua Formation, Leira Member, Andoya. All
figures are side views.

Fig. 1. Haplophragmoides aff. volgensis Myatliuk, x 105.

Fig. 2. Ammodiscus tenuissima (Gumbel), x 140.

Fig. 3. Glomospira gordialis (Jones and Parker), x 130.

Fig. 4. Glomospirella gaultina (Berthelin), x 140.

Fig. 5. Haplophragmoides cf. excavata Cushman and Waters, x 65.

Fig. 6. Haplophragmoides aff. neocomiana (Chapman), x 65.

Fig. 7. Verneuilinoides inaequalis Bartenstein and Brand, x 80.

Fig. 8. Haplophragmoides aff. goodenoughensis Chamney, x 40.

Fig. 9. Glomospira cf. charoides (Jones and Parker), x 80.

Fig. 10. Nodosaria loeblichae Ten Dam, x 125.

Fig. 11. Haplophragmium aequale (Roemer), x 45.

Fig. 12. Ammobaculites cf. subcretacea Cushman and Alexander, x70.

Fig. 13. Uvigerinammina sp., x80.

Fig. 14. Textularia foeda Reuss, x 105.

Fig. 15. Verneuilinoides cf. neocomiensis (Myatliuk), x50.

Fig. 16. Nodosaria cf. regularis Terquem, x75.

Fig. 17. Astacolus cf. cephalotes (Reuss), x85.

Fig. 18. Lenticulina gaultina (Berthelin), x35.

Fig. 19. Lagena sulcata (Walker and Jacob), x 155.

Fig. 20. Dentalina inepta Reuss, x 1 10.

Fig. 21. Astacolus cf. gratus (Reuss), x 105.

Fig. 22. Lenticulina aff. ovalis (Reuss), x 80.

Fig. 23. Dentalina linearis (Roemer), x 85.

Fig. 24. Dorothia cf. hechti Dieni and Massari, x95.

Fig. 25. Dentalina cf. communis d’Orbigny, x 35.

Fig. 26. Lenticulina miinsteri (Roemer), x 75.

Fig. 27. Dentalina cylindroides Reuss, x 50.

PLATE 46

L0FALDLI and THUSU, Norwegian Mesozoic Foraminifera

418

PALAEONTOLOGY, VOLUME 22

Member, which according to Dalland (1975) may have been deposited in somewhat
deeper water with a deficiency of oxygen. The presence of coaly matter and also the
absence of calcareous foraminifera, in contrast to assemblages 2 and 3 also might
indicate somewhat restricted marginal marine environment for this assemblage.

Assemblage 2. This assemblage occurs in the upper part of the Ratjonna Member,
correlated with the ammonite zone Surites ( Bojarkia ) mesezhnikovi, of the Ryazanian.
The fauna consists of poorly preserved arenaceous forms and a few calcareous
species. H. aff. goodenoughensis Chamney, H. cf. excavata Cushman and Waters
appear together with species of Bathysiphon , Reophax, and Lenticulina. H. aff.
goodenoughensis is the most common form in this assemblage and seems most similar
to individuals recorded from the Lower Cretaceous of Arctic Canada (Chamney,
1969; Souaya, 1976) and Spitsbergen (Lofaldli, unpublished data). However, the
forms recorded here are left under open nomenclature because of the poor preserva-
tion of the fauna.

The appearance of calcareous forms together with rich invertebrate faunas (Dalland
1975) and dinocysts (Birkelund et al. 1978) indicates a shallow, open marine environ-
ment for the assemblage.

EXPLANATION OF PLATE 47

All taxa from assemblage 3, Valanginian-Hauterivian, Nybrua Formation, Skjaermyrbekken Member,
Andoya. Except where otherwise stated figures are side views.

Fig. 1. Planularia cf. bradyana (Chapman), X 85.

Fig. 2. Dentalina nana Reuss, x 35.

Fig. 3. Pseudonodosaria mutabilis { Reuss), x75.

Fig. 4. Pseudonodosaria humilis (Roemer), x 135.

Fig. 5. Lenticulina perobliqua (Reuss), x 60.

Fig. 6. Pseudonodosaria tenuis (Bornemann), x 105.

Fig. 7. Saracenaria frankei Ten Dam, x 75.

Fig. 8. Lenticulina aff. sigali Bartenstein, Bettenstaedt and Bolli, x 70.

Fig. 9. Marginulinopsis comma (Roemer), x 55.

Fig. 10. Marginulinopsis gracillissima (Reuss), xllO.

Fig. 11. Lingulina loryi (Berthelin), 1 10.

Fig. 12. Globulina prisca Reuss, x 115.

Fig. 13. Quadrulina brunsviga Zedler, x 140.

Fig. 14. Oolina globosa (Montagu), x225.

Fig. 15. Bullopora tuberculata (Sollas), 180.

Fig. 16. Marginulina robusta Reuss, x 120.

Fig. 17. Sigmomorphina aff. neocomiensis Sztejn, x 160.

Fig. 18. Vaginulinopsis humilis praecursoria Bartenstein and Brand, x45.

Fig. 19. Lamarckina lamplughi (Sherlock), ventral view, x 185.

Fig. 20. Lingulina semiornata Reuss, x 115.

Fig. 21. Rosalina nitens Reuss, ventral view, x 145.

Fig. 22. Ramulina aculeata Wright, x 50.

Fig. 23. Patellina subcretacea Cushman and Alexander, dorsal view, x 145.

Fig. 24. Lamarckina lamplughi (Sherlock), dorsal view, x 185.

Fig. 25. Conorboides valendisensis (Bartenstein and Brand), dorsal view, x 140.

Fig. 26. Rosalina nitens Reuss, dorsal view, x 145.

Fig. 27. Conorboides valendisensis (Bartenstein and Brand), ventral view, x 155.

PLATE 47

L0FALDLI and THUSU, Norwegian Mesozoic Foraminifera

420

PALAEONTOLOGY, VOLUME 22

Assemblage 3. The microfaunas in samples from the Nybrua Formation alternate
from rich to very sparse, and are partly well-preserved. A mixed arenaceous-
calcareous assemblage is recorded both in Leira and Skjaermyrbekken members.
The characteristic of the calcareous fauna is displayed by the diversity of the Nodo-
sariidae, of which Lenticulina, Dentalina, Astacolus, Lagena, Marginulina, and
Marginulinopsis are best represented. The calcareous forms are represented by
numerous species, although most of them are represented by fewer specimens. The
most important index-fossils seem to be the following :

Conorboides valendisensis, recorded from Valanginian of western Germany
(Kemper 1961) and in Berriasian-Valanginian of Yorkshire (Fletcher 1972).

Lamarckina lamplughi, recorded in Hauterivian to Albian of Europe (Bartenstein
etal. 1971).

Vaginulina recta, known from Valanginian to Albian of Europe (Sztejn 1957).

Lenticulina wisselmanni, ranges from Hauterivian to Aptian in Germany and
England (Khan 1962).

Lagena hauteriviana hauteriviana, known from Berriasian to Barremian in Germany
(Michael 1967) and England (Fletcher 1972).

Marginulinopsis comma, previously recorded from Valanginian to Albian of the
Netherlands and Germany (Ten Dam 1948).

Patellina subcretacea, recorded from Valanginian to Albian of Europe (Sztejn
1957).

Arenaceous foraminifera include Glomospira, Glomospirella, Bathysiphon, Hap-
lophragmium, Bigenerina, Uvigerinammina, Textularia, Ammobaculites, Ammodiscus,
Verneuilinoides, and Reophax. Of note is the presence of Lenticulina and Glomospira
in large numbers. The commonest species are Glomospira cf. charoides, G. gordialis,
Haplophragmium aequale, Lenticulina aff. ovalis, and L. munsteri. Such an assemblage
is characteristic of open marine, neritic environment. This assemblage is most similar
to those reported from the Valanginian-Hauterivian sediments of north-western
Europe and Poland (Hecht 1938; Ten Dam 1946, 1948; Bartenstein 1956; Sztejn
1957; Bartenstein and Kaever 1973; Fletcher 1972).

Discussion. The great distances that isolate Late Jurassic and Early Cretaceous
foraminiferal assemblages in Andoya from other reported assemblages of similar age
render the island an important link for palaeobiogeographic reconstruction. In the
Middle Volgian-Ryazanian times Andoya lay near the marginal areas of the northern-
most Atlantic epicontinental sea which was connected to the north with the boreal
sea of the Arctic region and to the boreal Atlantic Sea in the south-west. Species of
Haplophragmoides in assemblages 1 and 2 of the Volgian-Ryazanian are poorly pre-
served and left under open nomenclature. However, these species are broadly com-
parable to forms reported from the Upper Jurassic-Lower Cretaceous of north-west
Europe and the Arctic areas. In Valanginian-Hauterivian times a major transgression
began as a result of Late Kimmerian phase of faulting. The influx of calcareous forms
in assemblage 3 of Valanginian-Hauterivian age in the Nybrua Formation is probably
the result of this transgression. Well over 90% of calcareous forms recorded, in
assemblage 3 are known to occur in coeval beds in north-west Europe. However, the
dominance of calcareous forms in assemblage 3 in Andoya is in marked contrast to

L0FALDLI AND THUSU: NORWEGIAN MESOZOIC MICROFAUNAS 421

neritic faunas reported from Agardhfjellet, Spitsbergen where Early Cretaceous
assemblages are dominated by simple arenaceous forms in which one or several
species make up the bulk of the assemblage. Lofaldli and Thusu (1976, p. 76) conclude
that these faunal differences are probably the result of sedimentary substrate, lati-
tudinal, or climatic factors in Spitsbergen.

SYSTEMATIC ACCOUNT OF FORAMINIFERA

In this paper some seventy species of foraminifera are recorded. All of these are
benthonic. More than 70% of the recorded species are calcareous, and the remaining
arenaceous. Some of the difficulties concerns the poor state of preservation of the
tests and the rare occurrences of most of the species. This explains the use of open
nomenclature in many identifications. The literature that has mainly been used for
the identifications are Hecht (1938), Ten Dam (1946, 1948), Bartenstein and Brand
(1951), Bartenstein, Bettenstaedt, and Bolli (1957), Sztejn (1957), Bartenstein and
Bettenstaedt (1962), Bartenstein and Kaever (1973), and Dailey (1973).

The families and genera of the foraminifera are arranged according to the classifica-
tion of Loeblich and Tappan (1964).

Astrorhizidae
Bathysiphon spp.

Ammodiscidae

Ammodiscus tenuissima (Giimbel) = Spirillina tenuissima Giimbel, 1862. Plate 46, fig. 2
Glomospira cf. charoides (Jones and Parker) = Trochammina squamata var. charoides Jones and Parker,
1860. Plate 46, fig. 9

Glomospira gordialis (Jones and Parker) = Trochammina squamata gordialis Jones and Parker, 1860.
Plate 46, fig. 3

Glomospirella gaultina (Berthelin) = Ammodiscus gaultinus Berthelin, 1880. Plate 46, fig. 4

Hormosinidae

Reophax spp.

Lituolidae

Haplophragmoides cf. excavata Cushman and Waters, 1927. Plate 46, fig. 5
Haplophragmoides aff. goodenoughensis Chamney, 1969. Plate 46, fig. 8

Haplophragmoides aff. neocomiana (Chapman) = Haplophragmium neocomianum Chapman, 1894. Plate 46,
fig. 6

Haplophragmoides aff. volgensis Myatliuk, 1939. Plate 46, fig. 1
Ammobaculites cf. subcretacea Cushman and Alexander, 1930. Plate 46, fig. 12
Haplophragmium aequale (Roemer) = Spirolina aequalis Roemer, 1841. Plate 46, fig. 1 1
Textulariidae

Textularia foeda Reuss, 1846. Plate 46, fig. 14
Bigenerina sp.

Ataxophragmiidae
Uvigerinammina sp. Plate 46, fig. 13

Verneuilinoides inaequalis Bartenstein and Brand, 1951. Plate 46, fig. 7

Verneuilinoides cf. neocomiensis (Myatliuk) = Verneuilina neocomiensis Myatliuk, 1939. Plate 46, fig. 15

Dorothia cf. hechti Dieni and Massari, 1966. Plate 46, fig. 24

Nodosariidae

Nodosaria loeblichae Ten Dam, 1948. Plate 46, fig. 10
Nodosaria cf. regularis Terquem, 1862. Plate 46, fig. 16

Astacolus cf. cephalotes (Reuss) = Cristellaria ( Cristellaria ) cephalotes Reuss, 1863. Plate 46, fig. 17
Astacolus cf. gratus (Reuss) = Cristellaria ( Cristellaria ) grata Reuss, 1863. Plate 46, fig. 21
Astacolus scitula (Berthelin) = Cristellaria scitula Berthelin, 1880
Astacolus schlonbachi (Reuss) = Cristellaria ( Cristellaria ) schlonbachi Reuss, 1 863

422

PALAEONTOLOGY, VOLUME 22

Dentalina cf. communis d’Orbigny, 1826. Plate 46, fig. 25
Dentalina cylindroides Reuss, 1 860. Plate 46, fig. 27
Dentalina inepta Reuss, 1863. Plate 46, fig. 20

Dentalina linearis (Roemer) = Nodosaria linearis Roemer, 1841. Plate 46, fig. 23
Dentalina nana Reuss, 1863. Plate 47, fig. 2
Dentalina oligostegia (Reuss) = Nodosaria oligostegia Reuss, 1845
Frondicularia hastata Roemer, 1842

Lagena hauteriviana hauteriviana Bartenstein and Brand, 1951

Lagena sulcata Walker and Jacob = Serpula ( Lagena ) sulcata Walker and Jacob, 1798. Plate 46, fig. 19

Lenticulina gaultina (Berthelin) = Cristellaria gaultina Berthelin, 1880. Plate 46, fig. 18

Lenticulina incurvata (Reuss) = Cristellaria incurvata Reuss, 1863

Lenticulina miinsteri (Roemer) = Robulina munsteri Roemer, 1839. Plate 46, fig. 26

Lenticulina aff. ovalis (Reuss) = Cristellaria ovalis Reuss, 1845. Plate 46, fig. 22

Lenticulina perobliqua (Reuss) = Cristellaria ( Cristellaria ) perobliqua Reuss, 1863. Plate 47, fig. 5

Lenticulina saxonica Bartenstein and Brand, 1951

Lenticulina aff. sigali Bartenstein, Bettenstaedt and Bolli = Lenticulina ( Marginulinopsis ) sigali Bartenstein,
Bettenstaedt and Bolli, 1957. Plate 47, fig. 8
Lenticulina wisselmanni Bettenstaedt, 1952

Marginulina robusta Reuss = Cristellaria ( Marginulina ) robusta Reuss, 1862. Plate 47, fig. 16
Marginulinopsis comma (Roemer) = Marginulina comma Roemer, 1841. Plate 47, fig. 9
Marginulinopsis gracillissima (Reuss) = Cristellaria gracillissima Reuss, 1862. Plate 47, fig. 10
Planularia cf. bradyana (Chapman) = Cristellaria bradyana Chapman, 1 894. Plate 47, fig. 1
Pseudonodosaria humilis (Roemer) = Nodosaria humilis Roemer, 1841. Plate 47, fig. 4
Pseudonodosaria mutabilis (Reuss) = Glandulina mutabilis Reuss, 1863. Plate 47, fig. 3
Pseudonodosaria tenuis (Bornemann) = Glandulina tenuis Bornemann, 1854. Plate 47, fig. 6
Saracenaria frankei Ten Dam, 1946. Plate 47, fig. 7
Saracenaria triangularis d’Orbigny, 1840
Vaginulina recta Reuss, 1 863

Vaginulinopsis humilis praecursoria Bartenstein and Brand = Lenticulina ( Vaginulinopsis ) humilis prae-
cursoria Bartenstein and Brand, 1951. Plate 47, fig. 18
Lingulina lame data Tappan, 1940

Lingulina loryi (Berthelin) = Frondicularia loryi Berthelin, 1880. Plate 47, fig. 1 1

Lingulina semiornata Reuss, 1863. Plate 47, fig. 20

Polymorphinidae

Globulina prisca Reuss, 1863. Plate 47, fig. 12

Pyrulina infracretacea Bartenstein, 1952

Sigmomorphina aff. neocomiensis Sztejn, 1957. Plate 47, fig. 17

Quadrulina brunsviga Zedler, 1961. Plate 47, fig. 13

Bullopora tuberculata (Sollas) = Webbina tuber culata Sollas, 1877. Plate 47, fig. 15

Ramulina aculeata Wright, 1886. Plate 47, fig. 22

Glandulinidae

Tristix acutangulum (Reuss) = Rhabdogonium acutangulum Reuss, 1862

Oolina globosa (Montagu) = Vermiculum globosum Montagu, 1803. Plate 47, fig. 14

Discorbidae

Rosalina nitens Reuss, 1863. Plate 47, figs. 21, 26
Spirillinidae

Spirillina minima Schacko, 1892

Patellina subcretacea Cushman and Alexander, 1930. Plate 47, fig. 23
Involutinidae

Trocholina infragranulata Noth, 1951
Anomalinidae

Gavelinella sigmoicosta (Ten Dam) = Anomalina sigmoicosta Ten Dam, 1948
Ceratobuliminidae

Conorboides valendisensis (Bartenstein and Brand) = Conorbis valendisensis Bartenstein and Brand, 1951.
Plate 47, figs. 25, 27

Lamarckina lamplughi (Sherlock) = Pulvinulina lamplughi Sherlock, 1914. Plate 47, figs. 19, 24

L0FALDLI AND THUSU: NORWEGIAN MESOZOIC MICROFAUNAS

423

table 1 . Occurrence of foraminifera
(R = rare, C = common, A = abundant)

Ammobaculites cf. subcretacea (PI. 46, fig. 12)
Ammodiscus tenuissima (PI. 46, fig. 2)
Astacolus cf. cephalotes (PI. 46, fig. 17)

A. cf. gratus (PI. 46, fig. 21)

A. schlonbachi
A. scitula
Bathy siphon spp.

Bigenerina spp.

Bullopora tuberculata (PI. 47, fig. 15)
Conorboides valendisensis (PI. 47, figs. 25, 27)
Dentalina cf. communis (PI. 46, fig. 25)

D. cylindroides (PI. 46, fig. 27)

D. inepta (PI. 46, fig. 20)

D. linearis (PI. 46, fig. 23)

D. nana (PI. 47, fig. 2)

D. cf. oligostegia

Dorothia cf. hechti (PI. 46, fig. 24)
Frondicularia hastata
Gavelinella sigmoicosta
Globulina prisca (PI. 47, fig. 12)

Glomospira cf. charoides (PI. 46, fig. 9)

G. gordialis (PI. 46, fig. 3)

Glomospirella gaultina (PI. 46, fig. 4)
Haplophragmium aequale (PI. 46, fig. 1 1)
Haplophragmoides cf. excavata (PI. 46, fig. 5)

H. aff. goodenoughensis (PI. 46, fig. 8)

H. aff. neocomiana (PI. 46, fig. 6)

H. aff. volgensis (PI. 46, fig. 1)

Lagena hauteriviana hauteriviana
L. sulcata (PI. 46, fig. 19)

Lamar ckina lamplughi (PI. 47, figs. 19, 24)
Lenticulina gaultina (PI. 46, fig. 18)

L. incurvata

L. munsteri (PI. 46, fig. 26)

L. aff. ovalis (PI. 46, fig. 22)

L. perobliqua (PI. 47, fig. 5)

L. saxonica

L. aff. sigali (PI. 47, fig. 8)

L. wisselmanni
Lingulina lamellata
L.loryi (PL 47, fig. 11)

L. semiornata (PI. 47, fig. 20)

Marginulina robusta (PI. 47, fig. 16)
Marginulinopsis comma (PI. 47, fig. 9)

M. gracillissima (PI. 47, fig. 10)

Nodosaria loeblichae (PI. 46, fig. 10)

N. cf. regularis (PI. 46, fig. 16)

Oolina globosa (PI. 47, fig. 14)

Patellina subcretacea (PI. 47, fig. 23)

Member

Ratjonna Leira Skjaermyrbekken

C C

R R

C
C
C
C

R C C

R R

R

R

C

C

R

C

C

R

R

R

R

C

A A

A A

C C

C C

R
C
C
R

R

R R

R
R

R R

R A A

A A

R R

R

R

R

R

R

R

R

R

R

R

R

R

R

424

PALAEONTOLOGY, VOLUME 22

Member

Ratjonna Leira Skjaermyrbekken

Planularia cf. bradyana (PI. 47, fig. 1)
Pseudonodosaria humilis (PI. 47, fig. 4)

P. mutabilis (PI. 47, fig. 3)

P. tenuis (PI. 47, fig. 6)

Pyrulina infracretacea
Quadrulina brunsviga (PI. 47, fig. 13)

Ramulina aculeata (PI. 47, fig. 22)

Reophax spp.

Rosalina nitens (PI. 47, figs. 21, 26)

Saracenaria frankei (PI. 47, fig. 7)

S. triangularis

Sigmomorphina aff. neocomiensis (PI. 47, fig. 17)

Spirillina minima

Textularia foeda (PI. 46, fig. 14)

Tristix acutangulum
Trocholina infragranulata
Uvigerinammina sp. (PI. 46, fig. 13)

Vaginulina recta

Vaginulinopsis humilis praecursoria (PI. 47, fig. 18)
Verneuilinoides inaequalis (PI. 46, fig. 7)

V. cf. neocomiensis (PI. 46, fig. 15)

R R R

R

R

R

R

R

R

C

R

R

R

C

R

R

R

R

R

R

R

R

R

R

R

Acknowledgements. We wish to thank cand. real. Arne Dalland (University of Bergen) for providing samples
and pertinent stratigraphic information, cand. real. Jeno Nagy (University of Oslo) for comments on the
manuscript, engineer Kari Baardseth (Central Institute for Industrial Research, Oslo) for taking scanning
micrographs, and Mr. Robin J. Godwin (Bristol University, U.K.) for photography.

bartenstein, H. 1956. Zur Mikrofauna des Englischen Hauterive. Senckenbergiana leth. 37, 509-533.

— and bettenstaedt, f. 1962. Marine Unterkreide (Boreal and Tehtys). In Leitfossilien der Mikropaldon-
tologie B7 (pp. 225-297). Arbeitskreis dtsch. Mikropalaont. Borntraeger, Berlin.

— and brand, E. 1951. Micropalaontologische Untersuchungen zur Stratigraphie des norddeutschen
Valendis. Abh. senckenb. naturf. Ges. 485, 239-336.

— and kaever, m. 1973. Die Unterkreide von Helgoland und ihre mikropalaontologische Gliederung.
Senckenbergiana leth. 54, 207-264.

— bettenstaedt, f. and bolli, H. m. 1957. Die Foraminiferen der Unterkreide von Trinidad. B.W.I.
Erster Teil: Cuche- und Toco-Formation. Eclogae geol. Helv. 50, 5-67.

— and kovatcheva, T. 1971. Foraminiferen des bulgarischen Barreme. Neues Jahrb. f. Geol. u.
Palaontologie, Abh. 139, 125-162.

bielecka, w. 1975. Foraminifera and brackish Ostracoda from the Portlandian of Polish Lowlands. Acta
Palaeont. Pol. 20, 295-393.

birkelund, T., thusu, b. and viGRAN, J. 1978. Jurassic and Cretaceous biostratigraphy of North Norway
with comments on British ammonite Rasenia evoluta. Palaeontology, 21, 31-63.
chamney, t. p. 1969. Barremian Textulariina, Foraminiferida, from Lower Cretaceous beds, Mount
Goodenough section, Aklavik Range, District of Mackenzie. Bull. Canada Geol. Survey, 185, 1-41.

— 1977. Foraminiferal morphogroup symbol for paleoenvironmental interpretation of drill cutting
samples: Arctic America. 1st. Int. Symp. on Benthonic Foraminifera of Continental Margins. Part B.
Paleoecology and Biostratigraphy. Maritime Sediments Spec. Pub. 1, 585-624.

chapman, f. 1894. Bargate beds of Surrey and their microscopic contents. Quart. Jour. Geol. Soc. 50,
677-730.

REFERENCES

L0FALDLI AND THUSU: NORWEGIAN MESOZOIC MICROFAUNAS

425

dailey, d. H. 1973. Early Cretaceous foraminifera from the Budden Canyon Formation, northwestern
Sacramento Valley, California. Univ. Calif. Publ. Geol. Sci. 106, 111 pp.
dalland, a. 1975. The Mesozoic rocks of Andoy, Northern Norway. Norges Geol. Unders. 316, 271-287.
fletcher, b. n. 1972. The distribution of Lower Cretaceous (Berriasian-Barremian) foraminifera in the
Speeton Clay of Yorkshire, England. In casey, r. and rawson, p. f. (eds.). The Boreal Lower Cretaceous,
161-168. Seal House Press, Liverpool.

hecht, f. E. 1938. Standard-Gliederung der Nordwest-deutschen Unterkreide nach Foraminiferen. Ahh.
senckenb. naturf. Ges. 443, 42 pp.

kemper, e. 1961. Mikrofauna und Faziesfossilien im unteren Mittelvalendis Nordwestdeutschlands. Neues
Jahrb. f. Geol. u. Palaontologie, Monatshefte, 87-94.

khan, M. H. 1962. Lower Cretaceous index foraminifera from northwestern Germany and England. Micro-
paleontology, 8, 385-390.

loeblich, a. r. and tappan, H. 1964. Sarcodina, chiefly ‘Thecamoebians’ and Foraminiferida. In moore,
r. c. (ed.). Treatise on Invertebrate Paleontology, Part C, Protista 2. I— II. Geol. Soc. Amer. and Univ.
Kansas Press, 900 pp.

lofaldli, m. and thusu, B. 1976. Microfossils from the Janusfjellet Subgroup (Jurassic-Lower Cretaceous)
at Agardhfjellet and Keilhaufjellet, Spitsbergen. Norsk Polarinstitutt Arbok 1975, 69-77.

Michael, e. 1967. Die Mikrofauna des nw-deutschen Barreme. Teil I : Die Foraminiferen des nw-deutschen
Barreme. Palaeontographica Suppl. 12, 176 pp. Stuttgart.
souaya, f. j. 1976. Foraminifera of Sun-Gulf-Global Linckens Island Well P-46, Arctic Archipelago,
Canada. Micropaleontology, 22, 249-306.

sztejn, J. 1957. Micropaleontological stratigraphy of the Lower Cretaceous in central Poland. Inst. Geol.
Prace, 22, 185-263.

ten dam, a. 1946. Arenaceous Foraminifera and Lagenidae from the Neocomian (Lower Cretaceous) of the
Netherlands. J. Paleont. 20, 570-577.

— 1948. Foraminifera from the Middle Neocomian of the Netherlands. Ibid. 22, 175-192.
thusu, b. and vigran, j. 1975. A review of the palynostratigraphy of the Jurassic system in Norway. 16 pp.
In NPF— Jurassic Northern North Sea Symposium. Stavanger, 1975.

M. L0FALDLI

Continental Shelf Institute
Hakon Magussons gt. 18
P.O. Box 1883
7001 Trondheim
Norway

B. THUSU

Typescript received 24 April 1978

Revised typescript received 11 September 1978

Arabian Gulf Exploration
P.O. Box 263
Benghazi
Libya

TWO NEW EARLY CRETACEOUS DINOCYST
SPECIES FROM THE NORTHERN NORTH SEA

by ROGER J. DAVEY

Abstract. Two new species of dinocyst, Oligosphaeridium abaculum and Systematophora silyba, are described from
a Barremian assemblage obtained from the northern North Sea north-east of the Shetlands. O. abaculum is the first-
known hystrichosphere with plate-centred tubular processes on which a clearly defined paratabulation is present and
this is described in detail. The paratabulation is of the Gonyaulax- type, and it is inferred that chorate cysts with similar
morphology are also of this type.

During the routine dating of core samples from the Institute of Geological Sciences
Offshore boreholes, a well-preserved and unusual dinocyst assemblage was recovered
from the terminal depth sample (depth 78-95 m) of borehole 77/80B. Two new species
from this assemblage are described below. The borehole was drilled during June 1977
and was located 30 km north-east of Unst, Shetlands in the northern North Sea
licence block 1 /4 (unallocated) at latitude 60° 56-7' N., longitude 0° 15-7' W. The water
depth at this drilling site was 141 m. A sedimentary thickness of 78-95 m was penetrated
which consisted of 33-60 m of Quaternary clay overlying 45-35 m of dark greenish-
grey mudstone. The upper part of this mudstone is micropalaeontologically dated as
being of Barremian-Aptian age and the lower part is of Barremian age.

All type and figured specimens having MPK numbers and the SEM stubs are
housed in the collections of the Institute of Geological Sciences, Leeds.

ASSEMBLAGE DETAILS

After standard palynologic preparation, the core sample at 78-95 m yielded an
assemblage composed almost entirely of dinocysts. Spores, pollen grains, and
terrestrially derived plant debris are virtually absent. The dinocyst assemblage is very
well preserved and dominated (97%) by Oligosphaeridium abaculum sp. nov. It, how-
ever, contains many other species including Achomosphaera neptuni (Eisenack 1958)
Davey and Williams 1966a, Avellodinium falsificum Duxbury 1977, Cassiculo-
sphaeridia magna Davey 1974, Cyclonephelium hystrix (Eisenack 1958) Sarjeant and
Stover 1978, Hystrichodinium voigtii (Alberti 1961) Davey 1974, Kleithriaspheridium
corrugatum Davey 1974, Odontochitina operculata (Wetzel 1933) Deflandre and
Cookson 1955, Phoberocysta neocomica (G ocht 1957) Millioud 1969, Pseudoceratium
pelliferum Gocht 1957, and Systematophora silyba sp. nov. This assemblage, but
for the new species, closely resembles those obtained from the Barremian part of the
Speeton Clay in north-east England (Davey 1974; Duxbury 1977). The presence of
O. operculata together with A. falsificum and K. corrugatum indicates that this core
sample is of early Barremian age and assignable to the Cassiculosphaeridia magna
Subzone, O. operculata Zone of Davey in press1.

[Palaeontology, Vol. 22, Part 2, 1979, pp. 427-437, pis. 48-50.)

428

PALAEONTOLOGY, VOLUME 22

The relatively rich dinocyst assemblage, associated with the paucity of land-
derived plant remains, strongly suggests that deposition at this borehole site during
the early Barremian took place during a marine transgressive phase.

SYSTEMATIC DESCRIPTIONS

Class dinophyceae Fritsch
Order peridiniales Haeckel
Genus oligosphaeridium Davey and Williams 19666

Remarks. O. abaculum sp. nov. is a tabulate dinocyst species, and by its inclusion
in the previously non-tabulate genus Oligosphaeridium the concept of this genus, as
envisaged by Davey and Williams 19666, is widened. However, except for the para-
tabulation, O. abaculum is so similar to the other members of this genus that it is
undoubtedly closely related to them and the paratabulation is considered to fall into
the category of specific variation. The Gonyaulax- type paratabulation and process
formula determined for this species can undoubtedly be applied to the other members
of Oligosphaeridium. In addition, it is probable that morphologically similar genera
such as Hystrichosphaeridium Deflandre 1937 emend. Davey and Williams 19666,
Kleithriasphaeridium Davey 1974, and Systematophora Klement 1960 also have the
same paratabulation.

Oligosphaeridium abaculum sp. nov.

Plate 48, figs. 1-6; Plate 49, figs. 1-7 ; Plate 50, figs. 1,4, 10, 11; text-figs. 1, 2

Derivation of name. Latin, abaculus, small tile for mosaic work, with reference to the tabulate nature of the
cyst wall.

Diagnosis. Shape : the pericyst, excluding processes, was originally spheroidal with
only minor dorso-ventral flattening.

Wall : the cyst wall is apparently two-layered, the two layers being closely appressed
except where the periphragm alone forms the processes. The wall is lightly to densely
intraperforate, the surface is smooth to scabrate and sparsely pitted. The processes
are smooth.

EXPLANATION OF PLATE 48

Figs. 1-6. Oligosphaeridium abaculum sp. nov. Note variation in cyst wall texture from smooth to scabrate.
1 , offset ventral view with process arising from paraplate 6". The wall is sparsely pitted and the parasutures
very faintly indicated, x 1500 (SEM). 2, lateral view; the precingular process has a strongly perforate
base and on the adjacent paraplate (upper left) the precingular process has become completely detached
at this point. Parasutures defined by very low ridges, x 1400 (SEM). 3, view of archaeopyle margin and
base of precingular process, x 5000 (SEM). 4, lateral view illustrating strong parasutural ridges and
the distal spines of a precingular process, x 1000 (SEM). 5, view of base of a partially detached process,
x 10,000 (SEM). 6, cyst fragment illustrating smooth internal wall, x 2000 (SEM).

Figs. 7, 8. Systematophora silyba sp. nov. 7, lateral view of specimen with operculum remaining partially
attached. Note the alignment of cingular processes dividing the cyst into approximately equal halves.
The pre- and postcingular annulate complexes are particularly noticeable because of the smooth peri-
phragm surrounding, and within, each complex, x 1500 (SEM). 8, apical-lateral view of specimen with
archaeopyle to illustrate the structure of a precingular annulate complex, x 2000 (SEM).

PLATE 48

DAVEY, Cretaceous Dinocysts

430

PALAEONTOLOGY, VOLUME 22

Paratabulation : the parasutures may be defined by very low ridges or by an
apparent change in the internal wall structure ; the clarity of the parasutures varies
from poor to good with those in the parasulcal region being the most poorly defined.
The paratabulation formula is pr, 4', 6", 6c, 6"', lp, 5s, l"".

Processes: the processes are of tubiform shape, and flare distally typically giving
rise to six slight flexuous spines; neighbouring spines may be joined medially to each
other. Process stem width varies according to position on the cyst, with the apical,
posterior, intercalary, and parasulcal processes being the most narrow. These pro-
cesses are also the shortest but variation in length is not pronounced ; process length
is approximately equal to half the endocyst diameter. The process formula is 4', 6",
5"', lp. Is, \"" ; the first postcingular paraplate (T") does not bear a process and the
parasulcal process occurs on the posterior sulcal paraplate (ps).

Archaeopyle : a apical archaeopyle always appears to be present, and is developed
by the detachment of the apical paraplates as a unit (type A). It has a strongly zigzag
margin with a deep parasulcal notch.

Holotype. MPK 2145, slide CSC 1824/4, IGS borehole 77/80B, depth 78-95 m, northern North Sea,
block 1/4. Barremian.

Paratype. MPK 2146, slide CSC 1824/4. IGS borehole 77/80B, depth 78-95 m, northern North Sea,
block 1/4. Barremian.

Holotype

Paratype

Range

Pericyst diameter (excluding processes)

(jum)

(fim)

w

length (archaeopyle developed)

54

—

51(56)62

width

57

52

52(57)64

Process stem length

28-36

22-38

16-38

(av. max. 31)

Number of specimens measured 15.

Description. Wall : the pericyst wall is approximately 0-5 ^m thick and is not obviously
two-layered. The density of the intraperforation varies only slightly with the individual
and the consequent spongy appearance is obvious under the light microscope (PI. 50,
figs. 10, 1 1); under the SEM the wall surface is smooth to very lightly pitted (PI. 48,
figs. 1-6). The parasutures are immediately obvious under the light microscope but
under the SEM are difficult (PI. 48, figs. 1, 2), and sometimes impossible, to discern.
Rarely do they take the form of low ridges (PI. 48, fig. 4). This lack or paucity of
surface expression suggests that the parasutural markings may be mainly features
formed by a change in wall structure such as a marked decrease in wall intraperforation.

EXPLANATION OF PLATE 49

Figs. 1-7. Oligosphaeridium abaculum sp. nov. All figures x 300. 1, ventral view of holotype. 2, view of
posterior part of the parasulcus. Note that processes have been detached from paraplates ps and lp.
Slide CSC 1824/4, MPK 2160, phase contrast. 3, dorsal view of the holotype. 4, detached operculum;
note centrally situated preapical paraplate. Slide CSC 1824/4, MPK 2161. 5, lateral view of specimen
illustrating the shape of the cingular paraplates. Slide CSC 1 824/4, MPK 2162, phase contrast. 6, antapical
view of paratype. 7, lateral view of specimen illustrating the spinose distal extremities of the processes.
Slide CSC 1824/4, MPK 2163, interference contrast.

PLATE 49

DAVEY, Cretaceous Dinocysts

432

PALAEONTOLOGY, VOLUME 22

Processes : the processes are smooth, without internal wall structure and the wall
thickness is less than 0-5 ^m. Proximally, the bases of the processes typically have
small to relatively large perforations (PL 48, figs. 2, 3). When these are well developed
a definite weakness of the process wall is developed and the process easily becomes
broken at this point (PI. 48, fig. 5; PI. 49, figs. 2, 5).

Paratabulation : the epicyst and hypocyst are of approximate equal size and are
separated by a narrow paracingulum which on the ventral surface, is displaced by its
own width (text-fig. 1a). The ends of the paracingulum do not overlap. The parasulcus
is not sinusoidal but is rather displaced approximately half its width to the left at the
paracingulum. The sulcal paraplates as, rs, Is, and ps are well defined ; as is large and
elongate, rs and Is are small and rectangular, and ps is relatively large and rect-
angular. The latter has a semi-circular boundary with paraplate l"". Paraplate ra
lies above paraplates rs and Is and to the right of Y"; its boundary with cingular
paraplate 6c is poorly defined. At the posterior end of paraplate as, abutting against
paraplates 6c, ra and V", are two depressions which correspond to the thecal flagellar
pores. The right depression is the smaller and more deeply indented.

text-fig. 1 . Drawings of the holotype of Oligosphaeridium abaculum sp. nov. to illustrate the paratabulation
and the position of the processes (process terminations omitted), a. Ventral surface, as, anterior sulcal;

ra, right accessory; rs, right sulcal; Is, left sulcal; ps, posterior sulcal paraplates. B, dorsal surface.

A small ovoidal preapical paraplate (pr) occupies the centre of the apical series
(text-fig. 2b). The apical paraplates T and 4' are essentially four-sided with T having
an additional small side that abutts against paraplate as. Paraplates 2' and 3' are
basically five-sided. The precingular paraplates are five-sided with 6" being the
smallest. The elongate cingular paraplates have arcuate long margins and are
narrowest at the cingular paraplate boundaries and widest, particularly on the dorsal
surface (text-fig. 1b), where the precingular parasutures abutt against them. The
postcingular paraplates are large and basically rectangular except for Y" which is
small, rectangular, and occurs beneath as and above lp. The latter is five-sided. The

DAVEY: CRETACEOUS DINOCYSTS

433

antapical paraplate (l"") is basically five-sided with paraplate ps indenting its
ventral margin (text-fig. 2a).

Intraspecific variation: this is slight and involves details such as process length,
width, and clarity of paratabulation. However, one rare distinctive variant exists in
which the stems of the processes are very short or apparently absent (PI. 50, figs. 1, 4).
In the latter case, the stellate distal process terminations lie directly on the endocyst.

text-fig. 2. Drawing of Oligosphaeridium abaculum sp. nov. to illustrate the para-
tabulation and the position of the processes (process terminations omitted).

A, antapical view of paratype. B, detached operculum (MPK 2161). Note small
preapical paraplate occupying the centre of the apical series.

Remarks. The more or less clear paratabulation of O. abaculum sp. nov., which is
defined by low ridges or, more usually, by internal wall structuring, differentiates this
species from all previously described species. Davey, in press2, erected a granular
species of Oligosphaeridium, O. verrucosum, from the Aptian of the northern Bay of
Biscay which sometimes possesses smooth parasutural areas. However, the granular
nature of the periphragm and the form of the paratabulation differentiates O. verru-
cosum from O. abaculum. Evitt et al. (1977, p. 4) record that Wiggins and Engelhardt
have found tabulate specimens of Oligosphaeridium in the Lower Cretaceous of
Alaska; it may be that these are comparable to the present species.

Genus systematophora Klement, 1960
Systematophora silyba sp. nov.

Plate 48, figs. 7, 8; Plate 50, figs. 2, 3, 5, 6, 7-9

Derivation of name. Latin, silybum, a kind of thistle — with reference to the spiny appearance of the cyst.

Diagnosis. Shape: the pericyst, excluding processes, is subspherical to ovoidal, the
long axis being in the apical-antapical plane. Dorso-ventral flattening is minor.

Wall : the cyst wall is apparently two-layered, the two layers being closely appressed
except where the periphragm alone forms the processes. The pericyst surface is rarely

434

PALAEONTOLOGY, VOLUME 22

smooth and is typically densely granular ; the granules are concentrated in the para-
cingular and pandasutural areas.

Processes : the processes are solid and smooth walled. Distally they terminate with
an irregular bifurcation ; proximally they may divide several times before joining the
endocyst. Rarely neighbouring processes may be linked medially or distally. They
vary in width but are approximately of equal length— typically between one quarter
and one third the endocyst diameter. The processes are aligned along the paracingulum,
and in the pre- and postcingular and antapical regions tend to form annulate
complexes.

Archaeopyle: an apical archaeopyle (type A) always appears to be developed
although the operculum often remains attached. The archaeopyle has a strongly
zigzag margin with a moderately deep parasulcal notch.

Holotype. MPK 2147, slide CSC 1824/3, IGS borehole 77/80B, depth 78-95 m, northern North Sea, block
1/4. Barremian.

Paratype. MPK 2148, slide CSC 1824/4, IGS borehole 77/80B, depth 78-95 m, northern North Sea,
block 1/4. Barremian.

Holotype

Paratype

Range

Pericyst diameter (excluding processes)

ip m)

(am)

Oxm)

length (complete)

36

36-38

length (archaeopyle developed)

33

—

30(32)36

width

38

33

29(34)38

Process length

8-14

7-10

7-14

(av. max. 13)

Number of specimens measured 15.

Description. Wall: the pericyst wall is approximately 0-5 ,um in thickness. Smooth
specimens are apparently present and although the majority of specimens are strongly
granular (PI. 48, figs. 7, 8), the exact nature of the surface ornamentation is sometimes
difficult to discern under the light microscope. The granules, which measure less than
0-5 pm in diameter, vary somewhat in size and appear to be restricted to the portions
of the cyst surface outside the annulate complexes. This unusual distribution results

EXPLANATION OF PLATE 50

Figs. 1, 4, 10, 11. Oligosphaeridium abaculum sp. nov. 1, ventral view to specimen with very faint para-
sutures and reduced process stems, x 2000 (SEM). 4, dorsal view of similar aberrant specimen with
more clearly defined paracingulum, x 1500 (SEM). 10, 11, Slide CSC 1824/1, MPK 2164. Ventral views
of specimen to illustrate wall texture and structure, 10, x 800 (interference contrast); 11, x 800 (plase
contrast).

Figs. 2, 3, 5, 6, 7-9. Systematophora silyba sp. nov. 2, dorsal view of holotype illustrating pre- and post-
cingular annulate complexes and aligned cingular processes, x 640. 3, ventral view. Slide CSC 1824/4,
MPK 2165, '640. 5, medial view of paratype, x640. 6, dorsal view of paratype, x 640. 7, detail of
cyst surface to illustrate smooth processes and granular surface periphragm, x 5000 (SEM). 8, lateral
view of a particular granular specimen; operculum partially attached. Slide CSC 1824/4, MPK 2166,
x 640. 9, lateral view of specimen illustrating process alignment and archaeopyle development. Slide
CSC 1824/1, MPK 2167, ■ 640.

PLATE 50

DAVEY, Cretaceous Dinocysts

11

436

PALAEONTOLOGY, VOLUME 22

because the smooth walled processes tend to branch proximally to give root-like
extensions over the cyst surface within the annulate complexes (PI. 48, figs. 7, 8).
A less well defined, non-granulate area sometimes surrounds the complexes and
similarly results from the proximal branching of the processes. The processes vary
in width from under 0-5 to 2 ^m, are mainly either bifid distally or are broadly
but irregularly capitate.

Process distribution : the epicyst and hypocyst are of approximate equal size and
are separated by a single alignment of processes that mark the position of the para-
cingulum. They are here distributed in pairs and triplets, each linear group being
indicative of a single cingular paraplate. Annulate complexes occupy the pre- and
postcingular and antapical regions and each indicates a single paraplate. There appear
to be six precingular, five postcingular, and one antapical complex. Obvious annulate
complexes or process alignments are absent in the apical and parasulcal regions.

Remarks. The combination of apical archaeopyle and annulate process complexes
clearly indicate that S. silyba sp. nov. belongs to the genus Systematophora. The
simple form of the processes, however, differentiates this species from most other
members of the genus, which bear complexly branching and anastomozing processes.
S. areolata Klement 1960, from the Upper Jurassic, is the most similar species but
differs significantly in having fewer processes which are more orderly arranged; in
particular, the annulate complexes are very obvious.

The specimen illustrated by Duxbury 1977 (pi. 11, fig. 3) as Cleistosphaeridium
polypes (Cookson and Eisenack 1962), from the Hauterivian and Barremian of
north-east England, is similar to S. silyba and may be conspecific.

Acknowledgements. I wish to thank Drs. R. Harland and B. Owens for their critical review of the manu-
script. The facilities of the Institute of Geological Sciences, Leeds, and the permission of the Director,
Institute of Geological Sciences, London, to publish, are gratefully acknowledged.

REFERENCES

alberti, G. 1961. Zur Kenntnis mesozoischer und alttertiarer Dinoflagellaten und Hystrichosphaerideen
von Nord- und Mitteldeutschland sowie einigen anderen Europaischen Gebieten. Palaeontographica, A,
116, 1-58, pis. 1-12.

cookson, i. c. and eisenack, a. 1962. Additional microplankton from Australian Cretaceous sediments.
Micropaleontology , 8, 485-507, pis. 1-7.

davey, r. j. 1974. Dinoflagellate cysts from the Barremian of the Speeton Clay, England. In Symposium on
Stratigraphic Palynology, Birbal Sahni Institute of Palaeobotany, Special Pub. No. 3, 41-75, pis. 1-9.

— in press1. The stratigraphic distribution of dinocysts in the Portlandian (latest Jurassic) to Barremian
(Early Cretaceous) of north-west Europe. Amer. Assoc. Strat. Palynol. Contrib. Ser., pis. 1-4.

— in press2. Marine Apto-Albian palynomorphs from Sites 400A and 402A, IPOD Leg 48, northern Bay
of Biscay. Initial Reports of the Deep Sea Drilling Project , 48, pis. 1-8.

— and williams, G. l. 1966a. The genera Hystrichosphaera and Achomosphaera. In Studies on Mesozoic
and Cainozoic dinoflagellate cysts. Bull. Br. Mus. nat. Hist. (Geol.), suppl. 3, 28-52.

— 19666. The genus Hystrichosphaeridium and its allies. In Studies on Mesozoic and Cainozoic
dinoflagellate cysts. Ibid. 53-106.

deflandre, g. 1937. Microfossiles des silex cretaces. Deuxieme partie. Flagelles incertae sedis Hystricho-
sphaerides. Sarcodones. Organisms divers. Ann. Paleont. 26, 51-103, pis. 11-18.

DAVEY: CRETACEOUS DINOCYSTS

437

deflandre, G. and cookson, i. E. 1955. Fossil microplankton from Australian late Mesozoic and Tertiary
sediments. Aust. J. Mar. Freshwat. Res. 6, 242-313, pis. 1-9.

duxbury, s. 1977. A palynostratigraphy of the Berriasian to Barremian of the Speeton Clay of Speeton,
England. Palaeontographia, B, 160, 17-67, pis. 1-15.

eisenack, A. 1958. Mikroplankton aus dem norddeutschen Apt. Neues Jb. Geol. Palaont., Abh ., 106, 383—
422, pis. 21-27.

evitt, w. r., lentin, J. K., millioud, m. e., stover, l. e. and williams, G. L. 1977. Dinoflagellate cyst termino-
logy. Pap. geol. Surv. Can. 76-24, 1-9.

gocht, H. 1957. Mikroplankton aus dem nordwestdeutschen Neokom (Teil 1). Palaont. Z. 31, 163-185,
pis. 18-20.

klement, k. w. 1960. Dinoflagellaten und Hystrichosphaerideen aus dem unteren und mittleren Malm
Siidwestdeutschlands. Palaeontographica, A, 114, 1-104, pis. 1-10.

millioud, m. e. 1969. Dinoflagellates and acritarchs from some western European Lower Cretaceous type
localities. In bronnimann, p. and renz, h. h. (eds.). Proceedings First International Conference Planktonic
Microfossils, Geneva, 1967, 2, 420-434, pis. 1-3, E. J. Brill, Leiden.

sarjeant, w. a. s. and stover, L. e. 1978. Cyclonephelium and Tenua: A problem in dinoflagellate cyst
taxonomy. Grana, 17, 47-54.

wetzel, o. 1933. Die in organischer Substanz erhaltenen Mikrofossilien des baltischen Kreide-Feuersteins
mit einem sediment-petrographischen und stratigraphischen Anhang. Palaeontographica, A, 77, 1 4 1 - 1 88.

R. J. DAVEY

Robertson Research International Ltd.

‘Ty’n-y-Coed’

Llanrhos

Typescript received 15 April 1978
Revised typescript received 30 July 1978

Llandudno LL30 ISA
Gwynedd
Wales

)•

THE HORSE C O RM 0 H I P PA RIO N THEOBALDI
FROM THE NEOGENE OF PAKISTAN, WITH
COMMENTS ON SIWALIK HIPPARIONS

by BRUCE J. M ACFADDEN and ABU BAKR

Abstract. A well-preserved skull of a three-toed horse, which is one of only a few known from the Neogene Middle
Siwaliks of the Pot war Plateau, Pakistan, is referred to Cormohipparion theobaldi. This conclusion is based on cranial,
dental, and size criteria that are diagnostic of this genus and species and differentiate it from the small Siwalik
hipparions. The Siwalik hipparions appear to represent a polyphyletic assemblage of at least two, and probably three,
forms. As a result of recent field work it is suggested, in contrast to previous hypotheses, that the ‘Hipparion Datum’
involves more than one form. The Siwalik hipparions have phylogenetic affinities with two or more groups in Holarctica.
Cormohipparion theobaldi appears most closely related to other species of this genus found in Clarendonian (late
Miocene) deposits of North America.

The Siwalik Hills of Pakistan and India have been one of the most important sources
of Eurasian Neogene mammals for more than a century. This rich faunal sequence
has attracted special attention because of the presence of early hominoid fossils,
including Ramapithecus and Sivapithecus (Pilbeam et al. 19776). During the last
century, large collections of Siwalik mammals were made for the Geological Survey
of India, Calcutta, and the British Museum (Natural History), London. These col-
lections have been monographed by workers such as Falconer and Cautley (e.g.
1846-1849), Lydekker (e.g. 1880-1884), and Pilgrim (e.g. 1911, 1915, 1926, 1932).
During the 1920s and 1930s, significant collections were made for the Yale Peabody
Museum by G. Edward Lewis, and for the American Museum of Natural History by
Barnum Brown (see Matthew 1929; Colbert 1935). Later collections were made by
other institutions, including the Bavarian State Museum, Munich, and the Geo-
logical Institute, University of Utrecht. During this decade, extensive collections have
been made by several groups including the Yale University-Geological Survey of
Pakistan expedition (Pilbeam et al. 1977a). The Equidae, represented by hipparions
and Equus, are of great abundance in the Siwaliks ; they are second only to the Bovidae
in number of specimens collected.

As is the case elsewhere with fossil horses, because of their cosmopolitan nature, the
Siwalik hipparions are of great importance in establishing a framework for Holarctic
biochronology and palaeozoogeography. Siwalik hipparions are mostly known from
isolated teeth and postcranial elements. Therefore, it is not surprising that the
taxonomy of these horses has traditionally been based on those elements. Skinner
and MacFadden (1977, see discussion below) have recently shown that certain cranial
characters can be used to sort out hipparions into several morphologically and
phylogenetically distinctive groups. However, skulls of Siwalik hipparions are very
rare and they have only been described in a few cases (e.g. Colbert 1935). The purpose
of this report is to describe a well-preserved skull of Cormohipparion theobaldi from

[Palaeontology, Vol. 22, Part 2, 1979, pp. 439-447.]

440

PALAEONTOLOGY, VOLUME 22

the Neogene Potwar Plateau of Pakistan and to interpret the phylogeny, biostrati-
graphy, and palaeozoogeography of Siwalik hipparions in the light of cranial and
other characters.

The dental nomenclature used in this paper follows Stirton (1941), Skinner and
Taylor (1967), and Skinner and MacFadden (1977).

The following specimen abbreviations are used in the text: AMNH: Department
of Vertebrate Paleontology, the American Museum of Natural History, New York;
GSI: Division of Palaeontology, Geological Survey of India, Calcutta; PUZ:
University of the Punjab, Department of Zoology Collection, Lahore.

SYSTEMATICS OF SIWALIK HIPPARIONS

Only a synopsis of Siwalik hipparion systematics will be presented here, because this
subject was recently reviewed elsewhere (Hussain 1971). Falconer and Cautley (1849)
described the first species of Siwalik hipparion as Hippotherium antilopinum. Lydekker
(1877) described a large species of Hipparion as Sivalhippus theobaldi. Subsequently
three small species were described: H. punjabiense (Lydekker 1886), H. perimense
(Pilgrim 1940), and H. chisholmi (Pilgrim 1910). Although in his taxonomy of
Hipparion Colbert (1935) listed all five previously named species, he appears to support
Matthew’s (1929) idea of only two species, the large H. theobaldi and small H. anti-
lopinum (including, in synonomy, the other three small species). Hussain (1971)
follows the idea of two species in the Dhok Pathan Formation and, in addition,
erects H. nagriensis for an intermediate-sized species that is restricted to the under-
lying Nagri Formation. These systematic discussions were largely based on analysis
of isolated teeth and postcranial elements.

As a result of this traditional taxonomy of Hipparion, a total of several hundred
species have to date been named from Holarctica and Africa. In some cases a new
species is created out of despair because it is virtually impossible to fit one’s local
sample into a pre-existing species. Forsten (1968), in her revision of Old World
hipparions, synonomizes the few hundred previously recognised species to thirteen,
which are defined on dental and postcranial characters. This solution might appear to
have alleviated the problems of unwieldy taxonomy; however, use of other charac-
ters such as cranial morphology suggests that Forsten’s species may include poly-
phyletic or horizontal ‘species’ assemblages.

Skinner and MacFadden (1977) present a discussion of primarily North American
hipparions in which discrete morphological groups at the generic level are based on
cranial characters. Of particular relevance to the present discussion, one hipparion
genus, Cormohipparion, is diagnosed by the presence of a preorbital facial fossa that
is deeply pocketed posteriorly and has a relatively well-developed and usually
continuous anterior rim (as well as a continuous posterior rim, but other hipparions
also exhibit this character). It was also demonstrated that, based on analysis of three
relatively large quarry samples (‘populations’) of Cormohipparion, the development of
the preorbital facial fossa is of taxonomic value at the generic level, despite claims
to the contrary. Besides Cormohipparion, several other discrete hipparion groups can
be recognised by characters of cranial morphology. It would be satisfying from
a practical point of view if cranial, dental, and postcranial characters could all be

M ACFADDEN AND BAKR: HORSE CORMOHIPPARION

441

used together to elucidate both generic and specific relationships. The present dis-
cussion is directed towards that goal.

SYSTEMATIC PALAEONTOLOGY

Order perissodactyla Owen, 1 848
Family equidae Gray, 1821

Genus cormohipparion Skinner and MacFadden, 1977
Cormohipparion theobaldi (Lydekker, 1877)

Text-fig. 1, Table 1

1877 Sivalhippus theobaldi Lydekker, pp. 31-32 {nomen oblitum).

1882 Hippotherium theobaldi Lydekker, pp. 81-87, pi. 1 1, figs. 3, 4; pi. 12, figs. 2, 4; pi. 13, figs. 1-3.
1885 Hipparion theobaldi Lydekker, pp. 58-60.

1929 Hipparion theobaldi Matthew, pp. 461, 524-526.

1935 Hipparion theobaldi Colbert, pp. 133-160, figs. 60-62, 64-70.

1968 ‘ Hipparion theobaldi ’ Forsten, pp. 83, 87, 88, fig. 34.

1971 Hipparion theobaldi Hussain, pp. 37-47, 51-63, pi. 1, figs. 7-10; pi. 2, figs. 7, 8; pi. 3, figs. 1-3;
pi. 4, figs. 1,2; text-fig. 18a, b.

1977 Cormohipparion sp. Skinner and MacFadden, pp. 923, 924; text-fig. 7. (Selected synonomy,
see Hussain 1971, for other references.)

Holotype. GS1 C153, left maxillary fragment with dP2-dP4, from the Middle Siwaliks of the Potwar
Plateau, Pakistan, see Lydekker (1882), pi. 11, fig. 4.

Revised diagnosis. Large and robust hipparion. Well-developed preorbital facial fossa that usually has
a continuous anterior rim and is pocketed posteriorly. Upper cheek teeth large with thick cement, rela-
tively complex enamel plications (especially on the posterior border of the prefossette and anterior border
of the postfossette), plicaballin usually consists of two loops, and protocone oval to bean-shaped. Lower
cheek teeth with deep ectoflexids in both the premolars and molars, and widely separated metaconids and
metastylids. Robust tridactyl metapodials.

Referred specimen. PUZ 69/371, skull with facial, palatal, and dorsal occipital regions, and left and right
P2-M3.

Occurrence. Collected by Professor A. Bakr and field party in 1969 from PUZ locality 45 near the village
of Lehri (Middle Siwaliks, Dhok Pathan ‘faunal stage’), Potwar Plateau, Pakistan (additional data are on
file in the Department of Zoology, University of the Punjab, Lahore).

Description. PUZ 69/371 is a relatively well-preserved skull of a mature individual (text-fig. 1). The skull is
not significantly crushed and the diagnostic generic and specific characters are preserved.

The buccinator fossa is anterior to P2. The premaxillary-maxillary suture extends posteriorly to above P2.
The infraorbital foramen lies above P3 anteroventral to the preorbital facial fossa. The preorbital facial
fossa, which is developed on the nasal and maxillary bones, is relatively large (right antero-posterior
length = 62 mm; right dorso- ventral width = 50 mm; left antero-posterior length = 64 mm; left dorso-
ventral width = 46 mm ; accuracy ± 5 mm) and lies above P3-M 1 . This fossa, with well-developed anterior
and posterior rims and deep posterior pocket, is the principal diagnostic character of Cormohipparion.
Although the sutures are difficult to distinguish, it appears that the lacrimal is triangular and lies postero-
ventral to the fossa. The anterior (maxillary) portion of the malar crest is moderately developed and the
posterior (jugal) portion is not preserved. On the dorsal occipital region there are bilaterally symmetrical
parietal crests that converge posteriorly. The posterior portion of the occipital region is not preserved and
an endocast of sediment is exposed.

A complete set of the cheek tooth dentition (right and left P2-M3) is preserved. These teeth are large with
relatively thick cement (Table 1). There is a progressive reduction posteriorly of the occlusal area of each
tooth, i.e. P2 is largest and M3 is smallest.

1

text-fig. 1. Cormohipparion theobaldi, PUZ 69/371, from the Neogene Middle Siwaliks of Pakistan.
1 , dorsal view of skull, < i ; 2, right lateral view of skull, x ^ ; 3, ventral view of skull, x i ; 4, occlusal view
of left upper cheek teeth, x

M ACFADDEN AND BAKR: HORSE CORMOH1PP ARION

443

table 1. Dental measurements (mm) of Cormohipparion theobaldi, PUZ 69/371.

Left side:

M3

M2

M1

P4

P3

P2

Greatest antero-posterior length

26-5

23-4

23-4

26-1

27-7

34-4

Greatest transverse width

22-2

24-8

26-5

29-6

29-5

27-8

Right Side :

Greatest antero-posterior length

26-2

23-9

23-1

27-3

27-8

35-1

Greatest transverse width

22-1

260

270

29-1

290

290

The P2 is triangular with a characteristically well-developed anterior extension of the parastyle. P3-M3 are
roughly square. The parastyles and mesostyles are well-developed ; the metastyles are weak. The prefossettes
and postfossettes are richly plicated, with the maximum enamel folding on the posterior border of the pre-
fossette and anterior border of the postfossette. The protolophs and metalophs are crescent shaped. The
plicaballin consists of two principal folds on each tooth. The protocone is oval or bean shaped and it is
characteristically isolated from the protoloph. The hypocone is not separated anteriorly from the metaloph
(as is the case in some other equids where a postprotoconal groove separates the metaloph and hypocone).
Posteriorly, the hypocone is separated from the metaloph by a moderately developed hypoconal groove.

Discussion. In his review of Siwalik hipparions, Hussain (1971) recognized three
species. Hipparion nagriensis was characterized as being restricted to the Nagri
Formation and of intermediate size relative to the two other species. H. theobaldi,
found in the Dhok Pathan and Tatrot (Upper Siwaliks) Formations, was charac-
terized by its large size and relatively robust metapodials. H. antilopinum, found con-
temporaneously with H. theobaldi, was characterized by its small size and relatively
gracile metapodials.

It is evident from the present study that, based on the configuration of the preorbital
facial fossa, the large species theobaldi can be assigned to the genus Cormohipparion.
There are few other skulls of Cormohipparion from the Siwaliks known or described ;
as well as PUZ 69/371, similar facial configurations are illustrated for AMNH 19466
(Colbert 1935, p. 143, fig. 64) and AMNH 98728 (Skinner and MacFadden 1977, p. 923,
fig. la). There are also dental characters that can be used to diagnose C. theobaldi.
Certainly its large size has been noted by workers for more than a century. Present
work in progress (MacFadden, in preparation) suggests that, besides size, there are
distinctive differences in dental pattern. For example, in the upper cheek teeth, the
plications on the posterior border of the prefossette and anterior border of the post-
fossette are more complex than in other Siwalik hipparions. Also, the plicaballin is
more complex. In the lower cheek teeth the ectoflexids (external valleys) are very deep
in both the premolars and molars. Correlated with the deep ectoflexids is the forma-
tion of a divided isthmus (including antero- and post-isthmus) and plicaballinid that
is poorly developed or absent. In the smaller Siwalik hipparions, which we will refer
to as the ‘Small Hipparion Complex’ (including H. antilopinum s.s. and H. nagriensis),
the deep ectoflexids and correlated characters are also found in the molars. However,
in contrast to C. theobaldi, there are usually shallow ectoflexids and prominent
plicaballinids in the premolars. In summary, C. theobaldi can be set apart from the
Small Hipparion Complex by characters including configuration of preorbital facial
fossa, size, and dental pattern.

The Small Hipparion Complex is at present more difficult to diagnose on a com-
bination of cranial, size, and dental characters. Within this complex there are at least

444

PALAEONTOLOGY, VOLUME 22

two cranial forms (phena), or species, belonging to two genera. The first group includes
the skull described by Colbert (1935, p. 142, fig. 63), AMNH 19761, incorrectly
referred to Cormohipparion by Skinner and MacFadden (1977, p. 923, figs, lb and 7c).
It now appears that AMNH 19761 should be considered a different form because of
its significantly smaller preorbital facial fossa (even taking into consideration the
small size of this skull). The taxonomic assignment of skulls exemplified by AMNH
19761 (and other specimens referred to H. antilopinum in Hussain 1971) is at present
left as ‘Small Hipparion Complex, incertae sedis ’. It would be satisfying if all the small
Siwalik hipparions had a facial morphology similar to AMNH 19761 ; if this were so,
we could conclude that two genera and species are represented in the Siwaliks,
C. theobaldi and "HI antilopinum. However, it does not appear that all skulls of the
Small Hipparion Complex are of a similar morphology. One specimen, GSI C349,
from the Middle Siwaliks of Perim Island, India, is morphologically different from the
Small Hipparion Complex, incertae sedis. In GSI C349, the preorbital facial fossa
has a complete and well-defined posterior rim and pocket, but anteriorly the fossa is
poorly defined and blends into the facial region. This configuration is most similar to
that of forms such as Hipparion s.s. from the type locality of this genus, Mt. Leberon,
France. Unfortunately, there are no other characters, such as size or pattern of the
dentitions, that can also serve to separate the forms within the Small Hipparion
Complex. This is not surprising, as a similar problem exists for western Eurasian
hipparions, where Forsten (1968) has included in her earliest (Vallesian) species,

H. primigenius, an assemblage that cannot be distinguished solely upon dental or size
criteria, but is certainly polyphyletic if viewed in light of differences in cranial
morphology.

In summary, the Siwalik hipparions are provisionally allocated to the following
categories, based on cranial morphology, size differences, and dental pattern.

I. Cormohipparion theobaldi ; well-developed preorbital facial fossa with complete
anterior rim, deep pocket posteriorly, large size, complex plications and plicaballin
in upper cheek teeth, deep ectoflexids in both the lower premolars and molars, robust
metapodials (exemplified by PUZ 69/371).

2. Small Hipparion Complex ( antilopinum-\ike ) ; at least two skull forms, small size,
plications less complex than C. theobaldi , lower molars with deep ectoflexids but lower
premolars with shallow ectoflexids and well-developed plicaballinids. Metapodials
relatively gracile.

a. Small Hipparion Complex, incertae sedis; significantly smaller preorbital facial
fossa relative to C. theobaldi (exemplified by AMNH 19761).

b. Hipparion , sensu stricto , preorbital facial fossa rimmed and pocketed posteriorly
but anteriorly poorly defined and blending into the facial region (exemplified by
GSI C349).

The biostratigraphy of Siwalik hipparions is important from the aspects of Holarctic
correlation and palaeozoogeography. Earlier workers stated that hipparions first
occurred in the Chinji Formation (Lower Siwaliks), became abundant in the Nagri
and Dhok Pathan Formations (Middle Siwaliks), and persisted into the Tatrot
Formation (Upper Siwaliks). Recent workers have rejected the idea that hipparions

MACFADDEN AND BAKR: HORSE CORMOHJPPARION 445

occur in the Chinji Formation, and it now appears that hipparions first appear in the
lower part of the Nagri Formation (Hussain 1971 ; Simons et al. 1971 ; Pilbeam et al.
1977a). This first occurrence, or ‘datum plane’ apparently represents dispersal of
hipparions into the Indo-Pakistan subcontinent. The value of the Hipparion Datum
Plane was recognized by earlier workers (e.g. Colbert 1935) to be an important event
in Neogene Holarctic biochronology, assuming that dispersal occurred rapidly. This
assumption of rapid dispersal could not be tested until the recent advent of absolute
dating techniques. Berggren and Van Couvering (1974), using a combination of
radiometric dates and marine-terrestrial correlations, stated that the Hipparion
Datum in the western Old World occurred at about 12-5 mya. Redating of the critical
Howenegg site in the Hegau region at the fossiliferous horizon suggests that the
Hipparion Datum in the western Old World is about 2 million years younger than
the Berggren and Van Couvering date of 12-5 mya (Becker-Platen et al. 1977). In
the Potwar Plateau of Pakistan there are as yet no radiometric dates from the Middle
Siwaliks. Recent chronological investigations in this region have concentrated
on independently dating the faunas by magnetic polarity stratigraphy. In the
Siwaliks, the Hipparion Datum occurs about 100 m below a distinctive long zone
of normal polarity interpreted to be magnetic Epoch 9 (Barndt et al. 1978). There-
fore, the age limits for the Hipparion Datum range from a maximum of 1 1 -47 mya
(Epoch 10-Epoch 11 boundary) to a minimum of 10-21 mya (Epoch 10-Epoch 9
boundary). If an extrapolation is performed within Epoch 10, then the Hipparion
Datum, because it is near the upper boundary limit, is probably closer to the 10-21
age (radiometric data from Watkins and Walker 1977).

These data may be interpreted in two different ways. If the date of about 12-5 mya
is accepted for the Hipparion Datum in the western Old World, then it appears that
hipparions were palaeogeographically isolated from the Indo-Pakistan subcontinent
for about two million years. On the other hand, if the dates of about 10-5 mya are
accepted for the Hipparion Datum in the western Old World, then the dispersal of
hipparions throughout this region, the Siwaliks, and probably the rest of Eurasia
was a geologically instantaneous event. This latter hypothesis seems preferable,
pending further absolute dates in critical late Miocene Old World sequences.

In his phylogeny of Siwalik hipparions, Hussain (1971) states that one intermediate-
sized species, H. nagriensis, was involved in the Hipparion Datum. He also hypo-
thesized that both the smaller H. antilopinum and larger H. theobaldi were descended
from H. nagriensis. Therefore, Hussain implies speciation of Siwalik hipparions in
situ. Based on recent field work by the Yale University-Geological Survey of Pakistan
expeditions, Hussain’s phylogenetic and biostratigraphic hypotheses should be
modified. Numerous Yale-GSP localities in the Nagri Formation yield two distinct
size classes (small and large). This distribution appears to continue through the Dhok
Pathan Formation and possibly into the Tatrot. As a result of this study, these two
forms are assigned to Cormohipp avion theobaldi and the Small Hipparion Complex.
Furthermore, it appears that at least two hipparion genera are involved in the
Hipparion Datum and that these forms diversified prior to dispersal into the Indo-
Pakistan subcontinent. C. theobaldi appears closely related to forms such as C. occi-
dentale in the Clarendonian (late Miocene) of North America (see Skinner and
MacFadden 1977). This genus is last found in latest Clarendonian— early Hemphillian

446

PALAEONTOLOGY, VOLUME 22

sediments in North America, but with its occurrence throughout the Dhok Pathan
Formation and possibly the Tatrot Formation, which are roughly equivalent to the
Hemphillian and younger Blancan North American Land Mammal Ages, Cormo-
hipparion appears to have persisted later in the Siwaliks than in North America.
Within the Small Hipparion Complex in the Siwaliks, the affinities of one form,
‘ incertae sedis ’ is at present not certain, whereas the other form appears closely
related to Hipparion sensu stricto from the Turolian of Mt. Leberon, France.

Acknowledgements. Financial support for this study was provided by National Science Foundation —
Smithsonian Foreign Currency grants. Special thanks are extended to the Principal Investigator of these
grants, David R. Pilbeam, Chairman of the Department of Anthropology, Yale University. The manu-
script was critically read by Jon A. Baskin, S. David Webb, and Ronald G. Wolff. I also thank Ronald G.
Wolff for his help with the preparation of the photographs. Mrs. Silvie Sidaway typed the manuscript. The
University of Florida contributed funds to defray part of the cost of publication.

REFERENCES

barndt, J. et al. 1978. The magnetic polarity stratigraphy and age of the type locality of the Dhok Pathan
village, Potwar Plateau, Pakistan. Earth Planet. Sci. Letters , 41, 355-364.
becker-platen, j. d., benda, l. and steffens, p. 1977. Litho- und biostratigraphische Deutung radio-
metrischer Altersbestimmungen aus dem Jungtertiar der Tiirkei (Kainozoikum und Braunkohlen der
Turkei, 18). Geol. Jb. B, 25, 139-167.

berggren, w. a. and van coiJVERiNG, j. a. 1974. The late Neogene: Biostratigraphy, geochronology, and
paleoclimatology of the last 15 million years in marine and continental sequences. Palaeogeogr., Palaeo-
climat ., Palaeoecol. 16, 1-216.

Colbert, e. h. 1935. Siwalik mammals in the American Museum of Natural History. Trans. Amer. Phil.
Soc. n.s. 26, 1-401.

falconer, H. and cautley, p. t. 1846-1849. Fauna antiqua sivalensis, being the fossil zoology of the Sewalik
Hills, in the north of India. London, Pts. 1-9.

forsten, a.-m. 1968. Revision of the Palearctic Hipparion. Acta Zool. Fennica. 119, 1-134.
hussain, s. T. 1971. Revision of Hipparion (Equidae, Mammalia) from the Siwalik Hills of Pakistan and
India. Verlag. bayer. Akad. fViss. N.s. 147, 1-68.

lydekker, r. 1877. Notices of new and other Vertebrata from Indian Tertiary and Secondary rocks. Pec.
geol. Surv. India. 10, 30-43.

1880-1884. Indian Tertiary and Post-Tertiary Vertebrata. Pal. Indica. 10, 1-355.

— 1885. Catalogue of the remains of Siwalik Vertebrata contained in the geological department of the
Indian Museum , Calcutta. Pt. 1, Mammalia. Calcutta, 1-116.

— 1886. Catalogue of the fossil Mammalia in the British Museum. Pt. 11. Containing the order Ungulata,
suborders Perissodactyla, Toxodontia, Condylarthra, and Amblypoda. London, 1-186.

Matthew, w. d. 1929. Critical observations upon Siwalik mammals. Bull. Am. Mus. nat. Hist. 56, 437-560.
pilbeam, d. R. et al. 1977 a. Geology and palaeontology of Neogene strata of Pakistan. Nature, Lond. 270,
684-689.

— et al. 1911b. New hominoid primates from the Siwaliks of Pakistan and their bearing on hominoid
evolution. Ibid. 270, 689-695.

pilgrim, G. e. 1910. Notices of new mammalian genera and species from the tertiaries of India. Rec. geol.
Surv. India. 40, 185-205.

1911. The fossil Giraffidae of India. Pal. Indica. 4, 1-29.

— 1915. New Siwalik primates and their bearing on the questions of evolution of man and the Anthro-
poidea. Rec. geol. Surv. India. 45, 1-74.

1926. The fossil Suidae of India. Pal. Indica. N.s. 8, 1-65.

— 1932. The fossil Carnivora of India. Ibid. 18, 1-232.

Simons, E. l., pilbeam, D. R. and boyer, s. J. 1971. Appearance of Hipparion in the Tertiary of the Siwalik
Hills of north India, Kashmir and West Pakistan. Nature, Lond. 229, 408-409.

M ACFADDEN AND BAKR: HORSE CO RM 0 H I P PARIO N 447

skinner, M. f. and taylor, b. e. 1967. A revision of the geology and paleontology of the Bijou Hills, South
Dakota. Novitates, Amer. Mus. nat. Hist. 2300, 1-53.

— SKINNER, M. F. and macfadden, b. j. 1977. Cormohipparion. N. Gen. (Mammalia, Equidae) from the
North American Miocene (Barstovian-Clarendonian). J. Paleont. 51, 912-926.
stirton, R. a. 1941 . Development of characters in horse teeth and the dental nomenclature. Jour. Mammal.
22, 434-446.

watkins, N. D. and walker, g. p. l. 1977. Magnetostratigraphy of eastern Iceland. Am. Jour. Sci. 277,
513-584.

BRUCE J. MACFADDEN
Florida State Museum
University of Florida
Gainesville, Florida 32611
U.S.A.

ABU BAKR

Department of Zoology
University of the Punjab
Fahore

Typescript received 25 April 1978 Pakistan

LIVE AND DEAD FAUNAS FROM CORALLINE
ALGAL GRAVELS, CO. GALWAY

by DANIEL W. J. BOSENCE

Abstract. A visual assessment, supported by multivariate statistical analysis, of the fauna from seventy-one benthic
samples from Mannin Bay shows the existence of five communities, which are coincident with the sedimentary facies
of the area. The Bank Community is found in algal banks constructed by the free-living corallines Lithothamnium
corallioides and Phymatolithon calcareum. This community is characterized by a varied and abundant epifauna living
on the algal thalli. The Muddy Algal Gravel Community also has a rich epifauna but has a more diverse burrowing
infauna. The Clean Algal Gravel Community is found in high-energy areas and has a poorer epifauna but a specialized
burrowing infauna. The Fine Sand Community is distinct from the algal gravel faunas, being composed of sand-living
bivalves, echinoderms, and gastropods. The Mud Community is very poor in both species and numbers. The fauna
from Mannin Bay is similar to communities described from coralline algal sediments of Ireland, Brittany, and the
English Channel.

The post-mortem history of abrasion, encrustation, and boring of the skeletal material is described. The dead
fauna is analysed to ascertain whether the previously defined communities are recognizable from the skeletal debris.
Most of the dead molluscan species in a facies are exotic but the bulk of the individuals come from that facies.
Exotic species are most dominant in the Clean Algal Gravel Facies and the Fine Sand Facies. Gastropods
from the rocky substrates form the most abundant group of exotic species. The trophic nuclei of the benthic
communities cannot be reconstructed from the dead fauna even though the live faunas are dominated by skeletal
organisms. Multivariate statistical analysis shows that the living communities cannot be reconstructed from the
dead fauna.

Community analysis of macrobenthic organisms is now popular among palaeonto-
logists as it provides a greater understanding of fossil assemblages, ancient environ-
ments, biostratigraphy, and community evolution. However, the recognition of
original life communities from fossil assemblages is difficult, and frequently has to
be based on slender evidence. One of the main reasons for the difficulties lies in the
paucity of studies on recent death assemblages and their relationships to the com-
munities from which they are derived. Detailed studies of this problem in sublittoral
environments comes from shallow coastal lagoons in California (Johnson 1965;
Warme 1969; Petersen 1976), a Spanish estuary (Cadee 1968), and the Yucatan coast
(Warme et al. 1976). The results from these studies show that death assemblages
reflect, with varying degrees of accuracy, the composition of the living communities.
Warme et al. (1976) go further and suggest that the death assemblages supply more
information on the benthic communities than the results from one sampling pro-
gramme. This is because of the temporal fluctuations in living populations and the
time averaging effect of death assemblages. These results contrast markedly with
studies from littoral environments (e.g. Wilson 1967), where major differences are
found between live and dead populations.

The main aim of this paper is to describe the macrobenthic communities and death
assemblages from a shallow marine environment in Co. Galway, Eire, where the
death assemblages do not permit the reconstruction of the original trophic nuclei or

[Palaeontology, Vol. 22, Part 2, 1979, pp. 449-478, pis. 51-52.]

450

PALAEONTOLOGY, VOLUME 22

text-fig. 1. Location, sedimentary facies, and sample station numbers of Mannin Bay, Co. Galway, Eire.

communities. Mannin Bay (text-fig. 1) is moderately exposed to wave and tidal
currents and therefore provides a comparison with the previous studies, many of
which are from low energy environments.

The environment of the study area has been well documented by Lees et al. (1969),
Buller (1969), Bosence (1976a, b, 1978) and Gunatilaka (1977). Most of the sediments
of the inner part of Mannin Bay are coralline algal gravels (maerl) together with
molluscan, echinoderm, foraminiferal, bryozoan, sponge, and ostracode debris.
Five sedimentary facies have been described from the results of a detailed programme
of scuba-collected samples and observations (Bosence 1976 b). A Bank Facies is
found in shallow water (1-8 m), reasonably sheltered areas and comprises an auto-
chthonous build-up of the branching coralline algae Lithothamnium corallioides
Crouan and Phymatolithon calcareum (Pallas) Adey and McKibbin (PI. 51, fig. 1).
The banks have a relief of up to 30 cm and cover areas up to half a kilometre square.
A Clean Algal Gravel Facies is found in exposed areas and is formed of algal and

BOSENCE: CORALLINE ALGAL GRAVEL FAUNAS

45:

molluscan debris swept into wave ripples. A Muddy Algal Gravel Facies is found
intermediate between the two previous facies and comprises algal and molluscan
material with a sandy mud matrix. A Fine Sand Facies of mixed biogenic origin is
found offshore and overlapping the algal gravels. A Mud Facies of both terrigenous
and carbonate origin is found in the very sheltered creeks around the bay.

The biological production of calcium carbonate within each of the facies is to be
described in a future publication.

METHODS

The area was sampled by scuba-diving at seventy-seven stations during the summer of 1972 (text-fig. 1).
The fauna was sampled by excavating 50 x 50 x 30 cm deep samples into strong polythene bags, sealed at
the top, with a hole in the side and a 2 mm mesh net over the base. After sampling, the bags were floated
to the surface with air introduced to the top of the bag and the sample was washed to remove the sand and
mud-sized material. The small fauna (less than 2 mm) was collected in samples 12 x 20 x 10 cm deep (i.e.
one-tenth the area of the large sample). The sample area was photographed in colour and notes were taken
on the ecology and taphonomy of the fauna and flora, together with substrate structures and textures. The
samples were sorted by hand (a x 10 lens was used for the small fauna). The live material was stored in
formaldehyde solution neutralized in Hexamine. All dead material recognizable to specific and generic
level was washed and dried. Because the study was mainly concerned with the carbonate-producing
organisms the soft bodied macrofauna was not identified below family level. Specimens for SEM work
were washed, dried, fractured, and coated with gold.

Duplicate sampling of a visually uniform substrate (stations at 1a, b, c, and d) showed that 60% of the
live and 68% of the dead taxa found in one square metre were being sampled in the 50 x 50 cm samples.
However, two species were making up 75% of the live individuals and four species account for 78% of
the dead individuals, and these common species were being adequately sampled. Similarly, air-lift suction
sampling (Keegan 1974) to depths of 100 cm in the substrate showed that all species were being sampled in
the top 20 cm. Complete faunal lists from the samples are stored at the British Library, Boston Spa, Wetherby ,
Yorkshire, LS23 7BQ, U.K. as Supplementary Publication No. SUP 14012. 124 taxa were identified,
comprising 6,687 live specimens and 100,680 dead specimens.

COMMUNITY ANALYSIS

Community descriptions

Observations made whilst diving, together with examination of species/stations data
suggested that similar groupings of species were occurring in the same sedimentary
facies. These qualitative observations were tested with multivariate statistical analyses
(see below) which confirm that species can be grouped into recurring associations or
communities. These communities correspond to previously defined sedimentary
facies and are, therefore, named after the facies (cf. Jones 1950): Bank Community,
Clean Algal Gravel Community, Muddy Algal Gravel Community, Find Sand
Community, and Mud Community. The composition and synecology of these com-
munities are described below.

Bank Community. The Bank Community (Table 1) contains the most abundant
vagile epifauna of all the facies. The unattached corallines provide a sheltered three-
dimensional structure in which food is supplied by the algae and by the debris trapped
within the algal branches (PI. 51, figs. 1, 2). This microenvironment allows the small-
sized epifauna to live throughout the bank structure so that they are in effect infaunal
(cf. ‘subsurface epifauna’ of Keegan 1974). Gastropods are common in this algal

table 1. Composition and abundance of fauna in Bank Community (18 stations).

Taxa

Trophic

Abundance (No. per 0-25 sq.

m)

group

Mean

No. stations Maximum

VAGILE EPIFAUNA

present

Bittium reticulatum

H

31-60

7

270

Xantho sp.

C/S

6-61

16

24

Porcellana longicornis

C/S

616

7

74

Gibbula cinerarea

H

600

3

93

Rissoa parva

H

3-88

2

40

Idotea sp.

S

2-22

2

30

Tricolia pullus

H

1-66

1

30

Portunus sp.

C/S

111

9

4

Gibbula magus

H

0-88

7

8

Chitonida

H

0-61

2

10

Nassarius incrassatus

C

0-44

3

6

Eupagurus sp.

C/S

Oil

2

1

Nassarius reticulatus

c

005

1

1

Galathea squamifera

c/s

005

1

1

Asterias rubens
Marthasterias glacialis

§

observed present, not sampled

SESSILE EPIFAUNA

Musculus discors

SF

5-55

2

80

Anemonia sulcata

C/S

0-44

3

5

Halichondria sp.

SF

Oil

2

1

Chlamys varia

SF

005

1

1

Spirorbidae

SF]

Serpulidae
Tubulipora phalangea

SF

SF

observed present, not sampled

Other bryozoa

sf)

BURROWING INFAUNA

Golfingia sp.
Mysella bidentata

DF ) 2-94

SF Jcomm- 2-77

11

2

16

40

Lucinoma borealis

SF

2-00

9

11

Nucula turgida

DF

0-94

5

9

Nereidae

C/S

0-88

7

6

Glyceridae

c

0-64

6

4

Venus verrucosa

SF

0-44

6

2

Amphitritidae

DF

0-38

4

4

Eunicidae

C/S

0-33

6

1

Parvicardium ovale

SF

0-27

1

5

Venerupis rhomboides

SF

0-22

2

3

Venerupis aurea

SF

0-22

3

2

Parvicardium exiguum

SF

016

2

2

Thyasira flexuosa

SF

Oil

1

2

Terribellidae

SF

Oil

1

2

Gouldia minimum

SF

005

1

1

Abra nitida

SF

005

1

1

Dosinia exoleta

SF

005

1

1

Nemertini

C

0-05

1

1

BORING INFAUNA

Hiatella arctica

SF

0-55

1

10

Gastrochaena dubia

SF

Poly dor a sp.
Cliona sp.

DF/SF)

DF/SFJ

observed present, not sampled

ey : H = herbivore ; C/S =

carnivore/scavenger ; DF =

deposit feeder, SF = suspension feeder; comm.

= commensals.

BOSENCE: CORALLINE ALGAL GRAVEL FAUNAS

453

table 2. Composition and abundance of fauna in Muddy Algal Gravel Community (1 1 stations).
Key as for Table 1.

Taxa Trophic

group

VAGILE EPIFAUNA

Bittium reticulatum

H

Rissoa parva

H

Gibbula cinerarea

H

Mangelia sp.

C

Cingula semicostata

H

Xantho sp.

C/S

Eupagurus sp.

C/S

Idotea sp.

s

Diadora apertura

H

Portunus sp.

C/S

Patina pellucida

H

Chitonida

H

Natica

C

Nassarius incrassatus

C

Nassarius reticulatus

C

Maidae

C/S

Asterias rubens

C)

Marthasterias glacialis

C)

SESSILE EPIFAUNA

Anemonia sulcata

C/SF

Anomia ephippium

SF

Halichondria sp.

SF

Spirorbidae

SF)

Serpulidae

SF

Bryozoans

sfJ

BURROWING INFAUNA

Mysella bidentata

SF

Nucula turgida

DF

Lucinoma borealis

SF

Astarte triangularis

SF

Thyasira flexuosa

SF

Golfingia sp.

DF

Turritella communis

SF

Parvicardium exiguum

SF

Nereidae

C/S

Parvicardium ovale

SF

Venerupis rhomboides

SF

Venerupis aurea

SF

Tellina tenuis

SF

Glyceridae

C/S

Nemertini

c

Haminoe navicula

c

Venus pullastra

SF

Sabella pavonina

SF

Abundance (No. per 0-25 sq. m)

Mean

No. stations
present

Maximum

53-64

5

230

4-54

1

50

3-45

8

20

2-72

1

30

1-81

1

20

1 09

5

5

0-81

6

4

0-36

3

2

0-27

1

3

0-27

2

2

0-09

1

1

0-09

1

1

009

1

1

0-09

1

1

0-09

1

1

0-09

1

1

observed present, not sampled

1-00 5 5

009 1 1

009 1 1

observed present, not sampled

7-27

2

50

6-72

4

69

6-36

10

28

3-63

2

10

2-18

2

23

1-82

5

6

1-45

2

15

1-36

4

10

1 36

4

43

1-27

4

10

0-91

2

6

0-91

5

5

0-54

1

6

0-54

4

3

0-54

5

4

0-36

1

4

0-27

1

3

0-27

2

2

[ Table continued overleaf.]

454 PALAEONTOLOGY, VOLUME 22

[Table 2 continued .]

Taxa

Trophic

group

Mean

Abundance (No. per 0-25 sq. m)

No. stations Maximum

present

Dosinia exoleta

SF

018

2 1

Venus verrucosa

SF

018

2 1

Dentalium sp.

DF

018

2 1

Venus fasciata

SF

009

1 1

Gouldia minimum

SF

009

1 1

Lutraria lutraria

SF

009

1 1

Abra alba

SF

009

1 1

Branchiomma vesiculosum

SF

009

1 1

Echinocyamus pusillus

DF

009

1 1

Leptosynapta inhaerens

DF

009

1 1

BORING INFAUNA

Gastrochaena dubia

SF

009

1 1

Hiatella arctica

SF

009

1 1

Polydora sp.
Cliona sp.

SF j

SF/DFJ

observed present, not sampled

lattice and may reach abundances of up to 93 Gibbula cinerarea and 270 Bittium
reticulatum per 0-25 sq. m. Another common group are decapod Crustacea which live
on and within the bank. The small forms, exemplified by Porcellana longicornis and
Galathea squamifera, can move within the lattice of the corallines but the larger
xanthids and portunids either swim over the surface, nestle into crevices, or occupy
rough burrows within the bank (PI. 51, fig. 3).

The sessile epifauna comprises two crevice-dwelling byssate bivalves, Musculus
discors and Chlamys varia, and various encrusting organisms. When found on the

EXPLANATION OF PLATE 51

Fig. 1. Underwater photograph of the free-living coralline algae Phymatolithon calcareum (left) and
Lithothamnium corallioides (right). Note epiphytes and Halichondria (upper right and lower left on
P. calcareum ) on corallines, ■ 1 .

Fig. 2. SEM micrograph of surface of Lithothamnium corallioides illustrating rich diatom population, x 200.

Fig. 3. Underwater photograph of portunid crab burrow in margin of bank of Lithothamnium corallioides.
Brown macrophyte Dictyota dichotoma growing on bank (upper left), x T

Fig. 4. Foraminifer encrusting surface of coralline, SEM, x 225.

Fig. 5. Foraminifer overgrown and buried within algal thallus, SEM, x 230.

Fig. 6. Dried specimen of Halichondria encrusting coralline. The sponge does not visibly alter the surface
of the alga, SEM, x 200.

Fig. 7. Pomatoceras triqueter encrusting Phymatolithon calcareum, x 1-5.

Fig. 8. Underwater photograph of Pomatoceras triqueter encrusting live Lithothamnium corallioides, X 1-5.

Fig. 9. SEM micrograph of Phymatolithon calcareum bored by Polydora, x 15.

Fig. 10. SEM micrograph showing spreite in central partition of U-shaped boring of Polydora, x 400.

Fig. 1 1 . SEM micrograph of opening of Polydora boring. The alga has reacted to the worm by building
a callus around the boring, x 90.

PM

PLATE 51

BOSENCE, algal gravel fauna

456

PALAEONTOLOGY, VOLUME 22

live corallines, the various encrusting forms show growth competition between the
algal host and the encrusters. Bryozoans and foraminifers are seen to be overgrown
by the corallines (PL 51, figs. 4, 5) whereas Halichondria (PI. 51, figs. 1, 6), Anemonia
sulcata , serpulids (PI. 51, figs. 7, 8), and epiphytic algae (e.g. Dictyota dichotoma )
overgrow the living corallines (PI. 51, fig. 3).

The burrowing infauna lives below the level of matrix sedimentation (between
10 and 30 cm below bank surface) and is dominated by polychaetes, bivalves and the
sipunculid Golfingia. Small-sized bivalves (e.g. Nucula turgida, Lucinoma borealis ,
and Parvicardium spp.) and polychaetes are able to burrow between the buried
branches of the corallines but the large venerupid bivalves are thought to be stationary
for most of their lives. There would be little predation or physical disturbance at this
depth in the bank.

The infaunal boring organisms utilize the algal skeletons, and bivalve and gastropod
shells as hard substrates. The most conspicuous borer into live algae is Polydora
(PL 51, figs. 9-11). A growth reaction rim (PL 51, fig. 11) in fossil examples would
indicate boring into the live as opposed to dead algal thalli. The U-shaped bores of
Polydora are thought to be the result of both chemical and mechanical activity
(Haigler 1969). Mechanical boring is suggested in this case because of the spreite
present in the central portion of the boring (PL 51, fig. 10). Several species of un-
identified filamentous red and green algae are common borers in the corallines (cf.
Cabioch 1969) forming inward branching cylindrical borings 0-015-0 089 mm in
diameter. The boring filaments support the large growths of filamentous algae found
on the living corallines in the quieter areas of the bay (Bosence 1976a). The sponge
Cliona is not found in the algal thalli but is abundant in large mollusc shells (PL 52,
fig. 9). The bivalves Hiatella arctica and Gastrochaena dubia bore, or nestle very
closely, as is more often the case with Hiatella , into the large coralline thalli and thick
mollusc shells.

Muddy Algal Gravel Community (Table 2). The surface of the Muddy Algal Gravel
Facies is a sandy mud on which lie live and dead corallines together with shell debris
(PL 52, fig. 2). The vagile epifauna is very similar to that of the Bank Community
except in the lower abundance of species present. Included are occasional patellids,
Diadora apertura and Patina pellucida. These gastropods are capable of deeply scrap-
ing the substrate for food whereas the more common trochiid gastropods have
a radula which is only capable of brushing the surface (Fretter and Graham 1962).
The sessile epifauna of the Muddy Algal Gravel Community is similar to the bank
fauna except for the occasional presence of two byssate species of bivalve, Anomia
ephippium and Crenella sp.

The substrate of this facies is similar to that of the lower, dead level of the algal
banks except that the algal grains are not branching and interlocking. This renders
the substrate easier for burrowing and may account for the greater diversity of
burrowing infauna in this community in comparison with that of the Bank Com-
munity (cf. Tables 1 and 2). The boring infauna is identical to that of the Bank
Community.

Clean Algal Gravel Community (Table 3). The vagile epifauna of this community is
not as abundant as that of the Bank or Muddy Algal Gravel communities because of

BOSENCE: CORALLINE ALGAL GRAVEL FAUNAS

457

table 3. Composition and abundance of fauna in Clean Algal Gravel Community (23 stations).
Key as for Table 1.

Taxa

Trophic

Abundance (No. per 0-25 sq

. m)

group

Mean

No. stations

Maximum

VAGILE EPIFAUNA

present

Bittium reticulatum

H

20-00

7

120

Idotea sp.

S

3-08

10

30

Cingula semicostata

H

2-61

1

60

Chitonida

H

2-30

6

20

Xantho sp.

C/S

1-30

12

7

Gibbula magus

H

0-83

4

10

Gibbula cinerarea

H

0-54

3

4

Nassarius reticulatus

C

0-48

2

10

Acmaea sp.

H

0-43

1

10

Tricolia pullus

H

0-43

1

10

Alvania punctura

H

0-43

1

10

Portunus sp.

C/S

0-35

6

2

Eupagurus sp.

C/S

0-17

3

2

Asterias rubens

c

0-04

1

1

Marthasterias glacialis

c

0-04

1

1

Porcellana longicornis

c/s

0-04

1

1

Trivia sp.

H

0-04

1

1

SESSILE EPIFAUNA

Mytilus edulis

SF

0-43

1

10

Anemonia sulcata

C/SF

0-30

5

1

Spirorbidae

SF)

Serpulidae

SF

observed present, not sampled

Bryozoans

sfJ

BURROWING INFAUNA

Astarte triangularis

SF

79-10

14

490

Lucinoma borealis

SF

1-99

8

26

Nucula turgida

DF

1 90

1

44

Golfingia sp.

DF

1-26

8

12

Venerupis aurea

SF

0-61

4

10

Echinocyamus pusillus

DF

0-52

2

10

Nereidae

C/S

0-52

9

3

Tellina tenuis

SF

0-48

2

10

Caecum sp.

SF

0-43

1

10

Mysella bidentata

SF

0-43

1

10

Crenella sp.

SF

0-43

1

10

Venerupis rhomboides

SF

0-39

6

4

Glyceridae

C/S

0-39

8

2

Dosinia exoleta

SF

0-35

7

2

Parvicardium ovale

SF

0-30

5

2

Eunicidae

C/S

0-26

5

2

Natica sp.

c

0-26

2

3

Gari tellinella

DF

0-22

2

3

Parvicardium exiguum

SF

0-17

2

3

[Table continued overleaf.]

458

PALAEONTOLOGY, VOLUME 22

[Table 3 continued .]

Taxa

Trophic

group

Venerupis pullastra

SF

Nemertini

C

Aphroditidae

C/DF

Amphitritidae

DF

Leptosynapta inhaerens

DF

Venus fasciata

SF

Thyasira flexuosa

SF

Dentalium sp.

DF

Branchiomma vesiculosum

SF

Venus verrucosa

SF

Lanice conchilega

SF

Pseudococcumis mixta

SF

BORING INFAUNA

Poly dor a sp.

SF

Cliona sp.

SF/DF

Abundance (No. per 0-25 sq. m)

Mean

No. stations
present

Maximum

0-17

1

3

017

2

3

0-17

1

4

017

1

4

0-17

3

2

013

3

1

0-08

2

1

008

2

1

008

2

1

0-04

1

1

0-04

1

1

observed present not sampled

the unstable nature of the substrate. The gravel is commonly swept into ripples with
ripple crests which are inhospitable to small members of the epifauna. Most of the
epifauna is, therefore, found in the ripple troughs associated with the live coralline
algae, coarse shell debris and clusters of grains held together by weed. Similarly, the
sessile epifauna is found mainly in the ripple troughs attached to the shells and
corallines. The burrowing infauna is dominated by bivalves and polychaetes. Because
the grain size is smaller, and the algal branches are not interlocking, it is probably
easier for the large bivalves to burrow in this sediment than in the Bank Facies. The
bivalve fauna differs from that of the Bank and Muddy Algal Gravel communities in
the lower abundances of the mud-inhabiting Nucula turgida and the thin-shelled
lucinoids Thyasira flexuosa and Lucinoma borealis. These are replaced by the small
Astarte triangularis which may reach numbers up to 490 per 0-25 sq. m. When com-
pared to the Bank Community there are more thick-shelled veneracean bivalves and
three species of the Tellinacea appear. The polychaete fauna includes the tube
builders Lanice conchilega (PI. 52, fig. 3) and Branchiomma vesiculosum , which use the
algal debris for tube construction. There is also an amphitritid polychaete which
forms a loose tube with algae, foraminifers, and shells.

The boring infauna is confined to the large grains found in the ripple troughs, in
particular large bivalve shells and rhodoliths.

Fine Sand Community (Table 4). The mobile surface of this substrate does not
normally provide niches for a diverse epifauna. However, an epifauna is found on
local accumulations of shell debris and local cover of Zostera marina and sea weeds.
No species are unique to this community and the diversity and abundances of the
vagile epifauna is less than the coralline algal faunas. The sessile epifauna is, as with

table 4. Composition and abundance of fauna in Fine Sand Community (12 stations).
Key as for Table 1.

Taxa Trophic

group

VAGILE EPIFAUNA

Bittium reticulatum

H

Gibbula cinerarea

H

Ido tea sp.

S

Nassarius incrassatus

c

Eupagurus sp.

c/s

Xantho sp.

c/s

Nassarius reticulatus

c

Buccinum undatum

c

Natica sp.

c

Mange lia sp.

c

Gibbula magus

H

SESSILE EPIFAUNA

Musculus discors

SF

Anemonia sulcata

C/SF

Spirorbidae

SF)

Serpulidae

SF

Bryozoans

SF j

BURROWING INFAUNA

Thyasira flexuosa

SF

Mysella bidentata

SF

Acrocnida brachiata

SF

Lucinoma borealis

SF

Abra alba

DF

Turritella communis

SF

Sabella pavonina

SF

Montacuta ferruginosa

SF

Abra prismatica

DF

Nereidae

C/S

Tellina tenuis

DF

Dentalium sp.

DF

Venerupis aurea

SF

Maldanidae

DF

Echinocardium caudatum

DF

Parvicardium ovale

SF

Owenia sp.

DF

Golfingia sp.

DF

Parvicardium exiguum

SF

Tellina fabula

DF

Leptosynapta inhaerens

DF

Glyceridae

C/S

Ensis sp.

SF

Venerupis rhomboides

SF

Venerupis pullastra

SF

Venus verrucosa

SF

Gari tellinella

DF

Branchiomma vesiculosum

SF

BORING INFAUNA

Poly dor a sp.

SF)

Cliona sp.

SF J

Abundance (No. per 0-25 sq. m)

Mean

No. stations
present

Maximum

17-50

5

180

9-25

5

100

5-00

1

60

1-83

3

20

1-33

4

8

0-58

4

3

0-25

2

2

0-08

1

1

0-08

1

1

0-08

1

1

008

1

1

1-67

1

20

0-92

3

5

observed present, not sampled

14-50

6

10

6-66

2

70

2-58

2

30

2-08

6

10

1-83

3

15

1-75

4

14

1-50

3

15

0-83

1

10

0-83

1

10

0-67

6

2

0-5

3

3

0-5

2

4

0-42

3

2

0-42

1

5

0-42

1

3

0-33

2

2

0-33

2

3

0-33

1

4

0-17

2

1

0-17

2

1

0-17

2

1

0-08

1

1

0-08

1

1

0-08

1

1

0-08

1

1

0-08

1

1

0-08

1

1

0-08

1

1

observed present, not sampled

460

PALAEONTOLOGY, VOLUME 22

the vagile epifauna, only found on weed or encrusting shell debris on the sediment
surface. The sediment is too fine-grained, mobile, and does not provide cover to
protect encrusting or vagile organisms. The burrowing infauna of this substrate
contains diverse and locally very abundant species of suspension feeders, deposit
feeders, and carnivores. The fauna is clearly distinct from that of the algal gravels and
contains species of bivalves, gastropods, polychaetes, and echinoderms commonly
associated with sandy substrates (see below). The sand substrate does not support
a boring macrofauna but the common carbonate grain borers are found in the shell
debris on the sediment surface.

Mud Community (Table 5). The Mud Community was only sampled at four stations
and the numbers of ecological groups are low, as are the species and their abundances.
There are no species, apart from Priapulus, which are specially adapted for life in
muddy substrates. The mud appears to be inhabited by species from neighbouring
facies which can tolerate fine sediment deposition.

table 5. Composition and abundance of fauna in Mud Community (4 stations).
Key as for Table 1.

Taxa

Trophic

Abundance (No. per 0-25 sq

. m)

group

Mean

No. stations

Maximum

VAGILE EPIFAUNA

present

Bittium reticulatum

H

1000

1

40

Littorina littorea

H

2-50

1

10

Gibbula magus

H

0-75

1

3

Xantho sp.

C/S

0-75

1

3

BURROWING INFAUNA

Parvicardium exiguum

SF

2-75

2

10

Mysella bidentata

SF

2-50

1

10

Lucinoma borealis

SF

200

2

7

Verier upis aurea

SF

1-50

1

6

Lanice conchilega

SF/DF

100

2

3

Nereidae

C/S

100

3

2

Priapulus sp.

DF

0-25

1

1

Eunicidae

C/S

0-25

1

1

Glyceridae

c/s

0-25

1

1

Quantitative community analysis

The results from regional benthic surveys produce very large matrices of data
from which species may or may not be grouped into recurring groups or communities.
The data may be assessed qualitatively, as for example by Keegan (1973) for 2000
samples containing about 200 species, and Dorjes (1972) for 103 samples contain-

BOSENCE: CORALLINE ALGAL GRAVEL FAUNAS

461

ing 268 species. In attempting to establish discrete faunal groupings from data of this
size, clearly decisions are subjective and errors may be made. If objective grouping of
the data is made then the communities have more value for comparative purposes
and for investigating the relationships between faunal groups and sedimentary facies.
Two multivariate statistical techniques are now widely used for grouping samples
on the basis of their fauna. They are both Q-mode analyses which group objects
(stations) on the basis of their variables (fauna), as opposed to R-mode analyses
which group related variables. The techniques are cluster analysis which classifies
groups of similar stations, and principal components analysis or ordination, which
can be used to show the differences and similarities between stations. Both these
methods are used here for the live and dead faunas and the results are compared. Full
descriptions of these methods can be obtained from Sneath and Sokal (1973, p. 214)
and Davis (1973, p. 456).

Methods. A non-parametric cluster analysis was selected as it analyses data on a presence-absence basis
(Bonham-Carter 1967). This is particularly useful for the analysis of fossil assemblages where species
abundances are difficult to obtain (Buzas 1972). The Jaccard coefficient of association was used to measure
the similarity of the samples as it ignores the frequency of mismatches. This is important when comparing
samples with large numbers of absences of species as samples may be considered similar because they
have the same species absences. In the clustering procedure the groups were ‘unweighted’ (Sneath and Sokal
1973, p. 228). The analyses are presented as dendrograms with samples scaled against their coefficients of
association (text-figs. 2, 4). The clusters of the live fauna are chosen at level of association of about 0-2
which is considerably higher than would be obtained from a clustering of the same data if it was randomly
arranged (randomness level = 0-04). The clusters of the dead fauna are chosen at a level of association of
0-4 and have a randomness level of 0T (text-fig. 4). The principal-components analysis was modified from
Wahlstedt and Davis (1968). Species which occurred in less than two stations were removed because of the
large number of zeros in the matrix. For the live fauna this left a matrix of 51 taxa occurring at 71 stations,
and for the dead fauna 59 species of molluses were analysed from 7 1 stations. The abundance data (x) were
transformed to log10(x+ 1) in accordance with other workers (e.g. Hughes and Thomas 1971) to make the
data less skewed. Originally the first five eigenvectors were plotted for the live and dead faunas as they
account for 60% and just over 50% of the total variation of the data respectively (Bosence 19766). The
similarities between the samples can be seen from their proximity when plotted with respect to the eigen-
vectors. Samples which consistently plot close to one another have similar faunas and samples which are
always separate have dissimilar faunas. The eigenvectors one, three, and four provide the best separation
of the stations on the basis of their live fauna. The analysis of the dead fauna shows the best separations of
the stations when plotted with respect to eigenvalues one, two, and three (text-figs. 3, 5). The eigenvalues
and their contributions are shown in Tables 6 and 8.

Cluster analysis. The cluster analysis, which is based on species presence or absence, shows a general low
level of association between the samples. At the 0-2 level of association about one half of the stations are
clustered in groups (text-fig. 2) which in the main correspond to the previously defined sedimentary facies.
Stations from the Bank Facies are clustered in Group 4. Group 2, although containing the majority of
samples from the Muddy Algal Gravel Facies contains a number of samples from neighbouring facies.
Most of these stations are located near the boundary of the Muddy Algal Gravel Facies and this suggests
a mixing of faunas at facies boundaries. The Clean Algal Gravel Facies stations are separated into three
groups. Samples taken on Mannin Platform comprise Groups 3 and 5 whilst the more offshore stations
make up Group 6. The Fine Sand stations are similarly divided into the offshore stations (Group 8) and
the stations from Mannin Platform which are classified with the Clean and Muddy Algal Gravel Facies
stations. Group 1 contains stations from terrigenous and intertidal stations. Stations from the Mud Facies
do not form a cluster.

From this classification of stations based on species presence or absence there is some correspondence
between fauna and sedimentary facies. However, the classification does not show an exact grouping of
samples which can be equated with the previously defined sedimentary facies.

462

PALAEONTOLOGY, VOLUME 22

text-fig. 2. Dendrogram illustrating clustering of stations by cluster analysis of live fauna.

Principal-components analysis. This analysis is more sophisticated than the previous analysis in that the
stations are compared by species abundances. The eigenvalues obtained from an analysis of the live fauna
are shown in Table 6. Text-fig. 3 shows the original samples plotted with respect to the eigenvectors one,
three, and four.

Stations from the Clean Algal Gravel Facies plot as a separate group due to variation in eigenvector one.
The species contributing most of the loading to this vector is Astarte triangularis (Table 3), a species parti-
cularly common in this sedimentary facies. Another distinct group of stations are those from the offshore
area of the Fine Sand Facies. The inshore Fine Sand Facies stations show more similarity with the fauna
of the adjacent algal gravel stations. The Fine Sand Facies stations are mainly separated by eigenvector
two, the main variation of which is controlled by the occurrences of the sand-living species Echinocardium
cordatum, Acrocnida brachiata, and tellinid bivalves. Stations from the Bank Facies, in the main, plot in
a separate area and the stations are separated by eigenvectors one and three. Eigenvector three is mainly
defined by nereid polychaetes and xanthid and portunid crabs. Stations from the Muddy Algal Gravel

BOSENCE: CORALLINE ALGAL GRAVEL FAUNAS

463

table 6. First five eigenvalues and eigenvectors from principal-components analysis of live fauna.

one

two

three

four

five

Eigenvalues
Percentage of total

0-99

0-59

0-31

0-24

0-23

contribution per eigenvalue

25 01

15-01

7-80

6 10

5-91

text-fig. 3. Live fauna from stations 1-73 plotted with respect to eigenvectors one, three, and four. Key.
• Bank Facies, A Muddy Algal Gravel Facies, O Clean Algal Gravel Facies, □ Fine Sand Facies, ▲ Mud

Facies.

Facies and inshore Fine Sand Facies generally have a scattered distribution in the centre of the field. Stations
from the Mud Facies consistently plot together.

The principal-components analysis therefore supports the results suggested by the cluster analysis and
shows a reasonably good correlation of sedimentary facies and the distribution of macrobenthic animals.
The most distinct sedimentary facies and communities are the Bank, Clean Algal Gravel, and the offshore
Fine Sand. The Muddy Algal Gravel, some Clean Algal Gravel, and the inshore Fine Sand facies have more
similar faunas.

From this analysis of species abundances the fauna can be shown to be grouped into recurring associa-
tions or communities and this confirms the results of the initial qualitative work.

Discussion

The fauna of each of the algal communities is similar, but in detail differences can
be seen in the abundances of certain groups and the presence or absence of species
which appear particularly sensitive. The Bank Community has the most diverse and

464

PALAEONTOLOGY, VOLUME 22

abundant epifauna and boring infauna. The Muddy Algal Gravel Community has
a similar epifauna but the species abundances are lower. However, the burrowing
infauna is richer in species and their abundances: in particular, the nuculacean,
lucinacean, veneracean, and tellinacean bivalves and polychaetes. The Clean Algal
Gravel Community is the most distinct algal gravel community but is most similar to
the Muddy Algal Gravel Community with the exception of the nuculacean and
lucinacean bivalves. These are replaced by the abundant Astarte triangularis. Most
other species are present, but with lower abundances, in this exposed environment.
The Fine Sand Community is clearly distinct, particularly offshore, both in respect
of its poor epifauna and the composition and abundance of the infauna.

Keegan (1974) has described similar faunas from the coralline algal sediments in
near-by Kilkerrin and Galway Bays. The Bank Community is closely comparable with
the fauna of Keegan’s ‘sublittoral living maerl’. In particular the varied epifauna
and low numbers of infaunal burrowers are common to the two regions. The Clean
Algal Gravel Community is similar to, but less diverse than, Keegan’s ‘maerl debris
unstable sediment’ fauna. Keegan does not describe facies similar to the Muddy
Algal Gravel or the Fine Sand. In general the area studied by Keegan yielded much
higher species abundances than did Mannin Bay. This is unlikely to be the result of
the different sampling procedures as these were compared in the two areas (see
Methods, above). In some cases the high abundances in Galway and Kilkerrin Bays
may be related to larval retention in these enclosed areas (Keegan 1974). The difference
in abundance of fauna between these two areas is not thought to be permanent but
probably reflects a short-term population fluctuation in Mannin Bay. Evidence for
this comes, for example, from the presence of abundant large dead shells of Venerupis
rhomboides in the Mannin Bay gravels. In the live populations this species is only
represented by occasional juveniles.

The results from Mannin Bay are similar to those of Cabioch (1968) from Brittany ;
in particular the Bank Community which, again, has the characteristic rich epifauna.
Cabioch considers the maerl fauna to be basically a Venus fasciata Community
(Ford 1923) with the addition of a rich epifauna. The similarities of the maerl and
sand faunas to Ford’s communities can also be seen when a comparison is made with
the fauna from the sands and gravels of the English Channel (Jones 1950; Holme
1966; Howell and Shelton 1970).

Coralline algal banks are also known from subtropical areas (e.g. Rodriguez Bank,
Florida Bay), and their sedimentology has been described by Turmel and Swanson
(1971, 1976). However, as far as I am aware, no detailed description of the fauna has
been published. Turmel and Swanson (1971) describe the fauna as being diverse
and abundant, particularly with regard to epifaunal organisms which live within the
‘ Goniolithon forest’ on Rodriguez Bank.

Similarly, no detailed palaeoecological studies have been published on Tertiary
coralline agal limestones. Pedley (1976) lists abundant molluscs, echinoids,
crustaceans, brachipods, calcareous sponges, and foraminifera from the coralline
algal bioherms of Malta. Bualuk and Radwanski (1968) describe the fauna from
rhodoliths in the Lithothamnium Limestone from the southern Holy Cross moun-
tains in Poland. They list diverse gastropods, bivalves, bryozoa, cirrepedes, starfish,
and echinoids from sands with abundant corallines. Shalekova (1964) mentions

BOSENCE: CORALLINE ALGAL GRAVEL FAUNAS

465

diverse faunas associated with lower Tertiary bioherms in western and central
Slovakia.

From this previous work it can be seen that diverse faunas are found in subtropical
and ancient coralline algal banks. However, more work is required before useful
comparisons can be made or community evolution be discussed.

ANALYSIS OF DEATH ASSEMBLAGES

A persistent problem in palaeontology is the difficulty of establishing the relation-
ships of fossil assemblages to living communities. Apart from the occasional organism
preserved in growth position the majority of fossil assemblages contain a biota which
has been transported and abraded to a greater or lesser degree. One approach to this
problem is to analyse the relation between live and dead assemblages in recent
environments. Apart from post-mortem transport there are several ecological factors
causing differences between live and dead faunal assemblages. These causes include
the patchiness in distribution of live benthic faunas which is well exemplified in this
study (Tables 1-5). In addition, several authors have shown that the abundance or
presence of species can change with time even if there are no apparent changes taking
place in the environment (Fischer 1931; Moore 1933; Birkett 1953; Holme 1966).
The importance of these temporal changes in the fauna can be assessed by either long-
term sampling programmes (Cadee 1968) or by comparison with the fauna in similar
neighbouring areas as is done in this study. Furthermore, rare species may be inadequ-
ately sampled. In Mannin Bay only 60% of the species of an area are being recovered
(see Methods, above). Finally, when habitats are changing the dead fauna of an
area will be composed of a mixture of faunas from different environments. On
a large scale the sedimentary facies of Mannin Bay appear to be stable and have been
shown to be related to present-day physical processes except for the onlapping of the
Fine Sand Facies. In this area a mixed live fauna has been described (above) and
a mixed dead fauna would be expected. On a smaller scale habitats may change, for
example during storms the margins of the algal banks are eroded. This would result
in mixing of dead populations at the margins of communities.

Notes on the taphonomy of the major invertebrate groups

Bivalvia. The most easily recognizable mode of death of the bivalves is from preda-
tion by naticid gastropods which are occasionally found living in the area (PL 52,
fig. 4). However, gastropod predation is only rarely seen and is far less important than
predation by the asteroids Marthasterias glacialis and Asterias rubens, which leave
no direct evidence of their activity in the fossil record. Indirect evidence is in the form
of asteroid plates and digging traces (PI. 52, fig. 1); the later would only rarely be
preserved in the coarse open algal sediments. Other predation on bivalves is by skates
(Rajidae) which excavate saucer-like depressions to obtain infaunal bivalves. After
the soft parts are eaten the fractured shells are seen scattered in the depression. In
summary, only occasionally are damaged shells found and the majority of the
specimens which have recently died show no evidence of predation. This is the con-
verse of that found by Wilson (1967) on tidal flats where bird predation is high.

466

PALAEONTOLOGY, VOLUME 22

The individual valves of the dead bivalves are usually disarticulated, but they may
stay attached for a long period. Initially the shells are held together by the ligament,
dentition, and suction in closely fitting valves. In forms with closely fitting dentitions
and margins (e.g. Nucula ) the valves are held together until the shell is considerably
abraded (cf. Boyd and Newell 1972). The joined valves of bivalves are frequently
covered by encrusters which grow over the shell commissure to seal the valves
together (PI. 52, fig. 7). These observations suggest that attached valves should not be
used indiscriminately as indicators of transport in fossil assemblages.

The post-mortem history of bivalve shells depends to a large extent on their size.
Large shells (greater than about 2 cm diameter, e.g. some Mytilacea, Ostracea,
Pectinacea, some Cardiacea, Veneracea, Mactracea, and Solenacea) are frequently
bored, encrusted, abraded, and fractured, whereas small shells (less than about 2 cm
diameter, e.g. some Mytilacea, Astartacea, Lucinacea, Erycinacea, some Cardiacea,
and Tellinacea) are not encrusted and bored at the macroscopic level, but are abraded
and fractured. This is thought to be because the larger shells are less frequently
transported than the small shells and can, therefore, be colonized and bored. In
addition the large shells will be on the sea bed longer than the small shells as they will
take a longer time to be broken down. This size division of the bored shells is suggested
by Boekschoten (1966) for Venus striatula in the North Sea. In shallow waters the
shells are never bored but shells dredged from deeper, quieter waters show evidence
of boring.

Larger shells are found mainly in the hydrodynamically stable convex up-position
in the high energy Clean Algal Gravel Facies judging from the face value of photo-
graphs and underwater observations. This is also the case for the buried shall layer
found 35 cm below the sediment surface at a station just to the south of Ardillaun
(73, text-fig. 1). This convex up-orientation leads to a differentiation of the encrusting
organism (PI. 2, figs. 9, 10). The lower concave surface is commonly encrusted by
serpulids, spirorbids, bryozoa, and barnacles, whereas the upper convex surface
is encrusted and bored by epilithic corallines, filamentous algae, and sponges. Ryland
(1970, p. 79) states that bryozoa are abundant on the concave surfaces of shells and

EXPLANATION OF PLATE 52

Fig. 1 . Underwater photograph of Asterias rubens excavating algal gravel with tube feet for bivalve ( Venerupis
aurea), xl/8.

Fig. 2. Underwater photograph of Venus verrucaria burrowing into muddy algal gravel, x 1.

Fig. 3. Underwater photograph of Lanice conchilega with tube and fan constructed of algal and mollusc
grains in clean algal gravel, x 1 .

Fig. 4. Venerupis rhomboides encrusted by corallines and a serpulid worm growing over the commissure, x 1 .

Fig. 5. Fractured specimens of Littorina littorea. Crab predation is suggested for the three right-hand
specimens but the others may have been fractured during transport, x 1 .

Fig. 6. Nucula turgida (upper left), Gibbula cinerarea (upper right), and Turritella communis bored by
naticid gastropods, x 1.

Fig. 7. Specimens of Bittium reticulatum showing abrasion of outer whorls, x 1.

Fig. 8. Heavily encrusted and bored specimen of Buccinum undatum, x 1.

Fig. 9. Upper convex surfaces of bivalves encrusted mainly by corallines and filamentous algae and bored
by sponges, x 0-5.

Fig. 10. Lower concave surfaces of bivalves in fig. 9 encrusted mainly by serpulids and bryozoa, x0-5.

PLATE 52

BOSENCE, algal gravel fauna

468

PALAEONTOLOGY, VOLUME 22

that in current swept areas the shells are convex down with the bryozoa inside. He
suggests that settlement may occur on the inside of the shells because of the strong
currents on the outer, upper surface. This separation of epifauna on the lower, and
epiflora on the upper, more brightly lit, surfaces has been reported from settlement
plates by Dybern (1967), Sentz-Braconnot (1968), and Bosence (in press). There
are many factors which may cause this separation of encrusting organisms. Positive
geotropism and negative phototropism have been shown for the larvae of rock-
encrusting spirorbids by de Silva (1962). He considers that the larvae are led away
from the bright areas which algae inhabit. The upper surfaces of the shells in Mannin
would have more light and would, therefore, likely to be attractive to the algae.
Once the algae are established the encrusting organisms would prefer the clean
surfaces away from the algae.

This separation of the epiflora and fauna is likely to be preserved and would indicate
small breaks in sedimentation or omission surfaces.

Gastropoda. Only rarely do the shells of gastropods show evidence of the way in
which they died. The sedentary infaunal Turritella communis is sometimes found bored
by naticid gastropods together with some bored Gibbula cinerarea (PI. 52, fig. 4).
Asterias rubens and Marthasterias glacialis have been observed preying on gastropods
and have been seen with up to six snails at a time in their enfolded arms. The recog-
nition and possible extent of crab predation on gastropods is difficult to interpret
from shell fracture. Boekschoten (1967) has shown how crabs break away the outer
whorl to eat gastropods and this is seen in Mannin Bay. However, this outer section
of the shell is the weakest part and would be the first to fracture by physical processes
as is suggested for some of the littorinids in PI. 2, fig. 5. The post-mortem history of
gastropods, as with bivalves, is related to the original shell size. The large shells
(greater than about 1 cm high: e.g. some Patellidae, Trochidae, Littorinidae, Tur-
ritellidae, Buccinidae (PI. 52, fig. 8), and Nassidae) show boring and encrusting in
addition to some physical abrasion and fracturing. Small gastropods (less than about
1 cm high: e.g. Acmaeaidae, Turbinidae, Rissoidae, Cerithiidae, Triphoridae and
Conidae) show only abrasion and fracturing (PI. 52, fig. 6). As argued for the bivalves
this division is considered to be the result of the small shells being transported more
often than the large shells.

Polychaeta. Potential polychaete macrofossils are the agglutinated tubes of the
terribellids, amphitritids, maldanids, Pectinaria and Chaetopterus, the calcified tubes
of serpulids, and the burrows of nereid, glycerid, and eunicid worms. The agglutinated
tubes would normally be preserved in their subvertical life position. But they are
frequently eroded and may be buried with their tubes parallel to the bedding (e.g.
Goldring 1964). The calcareous tubed polychaetes are mainly found encrusting shell
debris and would be buried with the shell, or fractured off and buried as sedimentary
grains.

Crustacea. Decapod Crustacea and, in particular, the more heavily calcified Brachi-
uridae, are abundant in the coralline algal gravels. As the exoskeleton is renewed
many times in their life it would be expected that these would be common in the
sediment. However, this is not found to be the case in Mannin Bay where only the

BOSENCE: CORALLINE ALGAL GRAVEL LAUNAS

469

occasional worn tips of chelae are found. The major cause of this is the extensive
decalcification of the cuticle which occurs prior to ecdysis (Robertson 1941) leaving
soft exoskeleton which soon decays. Major predators of the smaller decapods are
members of the family Rajidae. Stomach contents of Raja clavata have shown up to
fourteen small crabs per fish.

Occurrences and relative abundances of species and individuals of molluscs

Molluscan species and individuals are here classified as : (a) only occurring live in
a facies; ( b ) only occurring dead in a facies; and (c) occurring live and dead (Table 7).
From this classification it can be seen that there are about twice as many dead species
as there are live species in a facies. The dead individuals are up to three orders of
magnitude more abundant than the live individuals. In most facies exotic species are
more abundant than live and dead only species. However, the greatest numbers of
occurrences of dead species and individuals from a facies are from species which live
within that facies. Occurrences of exotic species are the next most abundant group,
followed by very few species and individuals of molluscs which occur living in a facies
but are not found dead. If the non-molluscan members of the fauna had been included
within this study then this last group would have been much more important. In
summary, although there may be large numbers of exotic species in a facies, their
occurrences are lower and the numbers of individuals are lower still. The exotic
species are characterized by low occurrences (often one specimen) of a large number
of species. The smallest numbers of exotic shells are found in the lower energy Bank
and Muddy Algal Gravel facies. The high-energy Clean Algal Gravel Facies has
a large proportion of exotic species (56-4%) and individuals (23-8%). The largest single
identifiable ecological group present in the exotic species are prosobranch gastropods
which usually inhabit weed-covered rocky substrates (Patellidae, Littorinidae,
Rissoidae). These species are thought to be transported in from the adjacent rocky

table 7. Occurrences of mollusc species and individuals in sedimentary facies.

Facies

Bank

Muddy Algal
Gravel

Clean Algal
Gravef

Fine Sand

Mud

No. of stations

18

11

23

12

4

Species

% dead species exotic to facies

46-3

491

56-4

67-6

76-0

Occurrences of live only species

2 (0-8%)

3 (0-9%)

8(1-8%)

7 (3%)

0-

Occurrences of dead only species

73 (27-7%)

97 (28-9%)

145 (32-6%)

119(50-6%)

34 (64-1%)

(exotic)

Occurrences of live and dead species

188 (71-5%)

236 (70-2%)

292 (65-6%)

109 (46-4%)

19(35-9%)

Total species occurrences

263

336

445

235

53

Individuals

Occurrences of live only individuals

3—

11-

35 (0-1%)

20 (0-1%)

0-

Occurrences of dead only individuals

842 (4-9%)

1625 (4-8%)

9039 (23-8%)

5590 (39-5%)

419 (10-9%)

(exotic)

Occurrences of live and dead

16260 (95 1%)

32170 (95-2%)

28845 (76-1%)

8556 (60-4%)

3435 (89-1%)

individuals

Total individuals

17105

33806

37919

14166

3854

470

PALAEONTOLOGY, VOLUME 22

areas or floated in on weed dislodged during storms (Cadee 1968, p. 88). In addition,
the exotic species in the Clean Algal Gravel Facies includes bivalves which are known
to inhabit adjacent facies which may well have been transported in. Some of the
exotic species in the Clean Algal Gravel Facies are known from similar facies in
adjacent bays (Keegan 1974) and it is probable that they once lived in Mannin Bay.
A persistent exotic species in the algal gravels is the common oyster ( Ostrea edulis)
which used to be cultivated in this region. It is only rarely found live today.

The Fine Sand Facies also has a high proportion of exotic species (67-6%) and
individuals (39-4%) which includes the gastropods from rocky areas. Another
identifiable group are six species of tellinid bivalves which are known to inhabit sand
substrates (Keegan 1974; Tebble 1966). It seems likely that these species are either
only occasionally inhabiting this facies or that they are living in the sediment but in
too low numbers to be sampled in the live populations. Other exotic species of the
sand are found living in the adjacent coralline algal gravels and are thought to be
mixed with the sand fauna as the Fine Sand Facies is overlapping the algal facies.

table 8. First five eigenvalues and eigenvectors from principal-components analysis of dead fauna.

Eigenvalues
Percentage of total
contribution per eigenvalue

one

two

three

four

five

1-71

1-17

0-99

0-72

0-59

16-79

11-51

9-70

7-07

5-79

The low diversity fauna found in the Mud Facies is not reflected in the dead faunas.
A high percentage of exotic species (76%) is found which is made up of a small
number of individuals (10-9%). These exotic species are largely accounted for by
specimens from the rocky areas, with the addition of molluscs from the algal gravels
which may have been transported into Mannin Creek.

The numbers of species which occur only living in a facies is very small and the
numbers of individuals of the species is very low. An examination of the species
involved shows that they are nearly all either thin-shelled bivalves (lucinids and
tellinids) or that they are very thin-shelled gastropods ( Haminoea navicula and
Natica ). After death these thin-shelled forms would be expected to break down
quickly and so be unrecognizable in the dead material.

In summary, the dead molluscan fauna bears a complex relationship to the live
fauna. Most of the dead species in a facies are exotic but the bulk of the individuals
come from that facies. The higher-energy facies have the greatest numbers of exotic
shells. The largest single identifiable group of exotic shells are gastropods from
near-by rocky areas ; others are thought to have been transported in or to have formerly
lived in the area.

Reconstruction of trophic nuclei from the dead fauna
This analysis attempts to reconstruct the original trophic nuclei of the benthic
communities using the dead material from each facies (Table 9). The six most abundant
species of a death assemblage are taken to represent the trophic nucleus. If the more

BOSENCE: CORALLINE ALGAL GRAVEL FAUNAS

47:

commonly accepted definition of a trophic nucleus is used (species making up 80%
of the fauna ; Neyman 1967) for the dead material, the nuclei would be made up of the
same three species ( Bittium reticulatum, Rissoa parva, and Gibbula magus ) for the
coralline algal and Fine Sand faunas (Table 9).

table 9. A comparison of actual and reconstructed trophic nuclei.

Actual Reconstructed

Trophic

Cum. %

Trophic

Cum. %

Species

group

abndce.

Species group

abndce.

BANK

COMMUNITY

TROPHIC NUCLEUS

Bittium reticulatum

H

37-5

Bittium reticulatum

H

72-9

Xantho sp.

C/S

45-3

Rissoa parva

H

85-4

Porcellana longicornis

C/S

52-5

Gibbula cinerarea

H

88-8

Gibbula cinerarea

H

59-5

Tricolia pullus

H

90-7

Musculus discors

SF

660

Parvicardium exiguum

SF

92 1

Rissoa parva

H

70-5

(E) Acmaea sp.

H

93-6

CLEAN ALGAL

GRAVEL COMMUNITY TROPHIC NUCLEUS

Astarte triangularis

SF

62-1

Bittium reticulatum

H

55-7

Bittium reticulatum

H

77-8

(E) Rissoa parva

H

77-5

Ido tea sp.

S

80-2

Gibbula cinerarea

H

83-7

Lucinoma borealis

SF

82-4

Astarte triangularis

SF

88-0

Cingula semicostata

H

84-4

Tricolia pullus

H

911

Chitonida

H

86-2

Gibbula magus

H

92-2

MUDDY ALGAL

GRAVEL COMMUNITY TROPHIC NUCLEUS

Bittium reticulatum

H

490

Bittium reticulatum

H

77-8

Mysella bidentata

SF

55 6

Rissoa parva

H

83-2

Nucula turgida

DF

61-7

Gibbula cinerarea

H

86-8

Lucinoma borealis

SF

67-5

Parvicardium exiguum

SF

88-5

Rissoa parva

H

71-6

Acmaea sp.

H

91-2

FINE SAND COMMUNITY TROPHIC NUCLEUS

Bittium reticulatum

H

23-2

Bittium reticulatum

H

43-8

Thyasira flexuosa

SF

42-4

(E) Rissoa parva

H

72-8

Gibbula cinerarea

H

54-7

Gibbula cinerarea

H

78-7

Mysella bidentata

SF

63-5

Turritella communis

SF

82-3

Idotea sp.

s

70-1

(E) Rissoa lilacina

H

84-8

Acrocnida brachiata

SF

73-5

Thyasira flexuosa

SF

86-7

MUD

COMMUNITY

TROPHIC NUCLEUS

Bittium reticulatum

H

39-2

Bittium reticulatum

H

82-5

Parvicardium exiguum

SF

49-9

(E) Littorina littoralis

H

86-9

Mysella bidentata

SF

59-7

Littorina littorea

H

90-3

Littorina littorea

H

69-5

(E) Gibbula cinerarea

H

92-4

Lucinoma borealis

SF

77-3

(E) Rissoa parva

H

94-2

Venus aurea

SF

83-1

Parvicardium exiguum

SF

95-5

Key. H = herbivore; C/S = carnivore/scavenger ; SF = suspension feeder; DF = deposit feeder; (E) = exotic
to community.

472

PALAEONTOLOGY, VOLUME 22

The results from this analysis show that the correct composition and order of
abundance of the species making up the trophic nuclei cannot be reconstructed from
the dead fauna. Apart from the ubiquitous B. reticulatum , which is the commonest
live and dead organism, three of the six most abundant organisms in the dead material
are not present in the six commonest live species. In addition, up to one-third of the
organisms in the ‘dead trophic nuclei’ are exotic. The reconstructed trophic nuclei

text-fig. 4. Dendrogram illustrating clustering of stations by cluster analysis of dead fauna.

BOSENCE: CORALLINE ALGAL GRAVEL FAUNAS

473

are dominated by herbivorous and detritus-feeding gastropods, whereas the actual
nuclei have abundant carnivores/scavengers, suspension feeders, and deposit feeders
in addition to the herbivores. The epifaunal gastropods which are frequently exotic,
are being preferentially preserved to dominate the reconstructed trophic nuclei and
thus forming an anomalous ‘trophic zone’ (Sokolova 1964; Neyman 1967).

Stanton (1976) has recently criticized the applicability of Turpaeva’s (1949, 1957)
work on trophic nuclei to palaeontology by showing that the preservable organisms
in a community do not always reflect the original feeding groups in benthic com-
munities with many soft-bodied organisms. In this study from Mannin Bay the trophic
nuclei are, in each case, dominated by taxa with hard parts but, due to various tapho-
nomic factors, the dead material cannot be used to reconstruct the living trophic
nuclei.

Quantitative analysis of death assemblages

For this study the dead fauna was analysed using the same programmes as were
used for the live fauna (see Methods, above) to see if similar groupings could be recog-
nized. The non-parametric cluster analysis produces the clusters shown in the dendro-
gram in text-fig. 4. Four main clusters are seen at the 04 level, the rest of the stations
showing no grouping. The clusters produced from this analysis show complete mixing
of the stations from the previously defined sedimentary facies and communities.
None of the previous clusters can be recognized and no interpretable pattern is pro-
duced if the clustered stations are plotted on maps. This result is to be expected from
the presence-absence analysis of the fauna because of the high proportion of exotic
species in each facies (Table 7). This mixing of the fauna from each facies is also
reflected in the level of association which is very much higher than for the live fauna.

The results of the principal-components analysis (based on fifty-nine mollusc
species) are shown in Table 8 and text-fig. 5. The first five eigenvalues are used as they
account for just over 50% of the total variation of the data. The plots of the stations
with respect to the first three eigenvectors (text-fig. 5) show that the samples from the
previously defined communities do not plot as separated clusters. However, samples
from similar communities plot as very poorly defined clusters reflecting the Clean
Algal Gravel, Bank, and offshore Fine Sand communities. The Muddy Algal Gravel
and inshore Fine Sand faunas do not plot as recognizable clusters, as was previously
predicted, because of the onlap of the Fine Sand Facies (Bosence 19766). The Clean
Algal Gravel Community stations are separated by eigenvectors one, two, and three
and the Fine Sand Community stations by eigenvector three. The species causing the
highest loading on eigenvector three are the sand living Thyasira flexuosa, Gastrana
fragilis, Turritella communis , and Natica sp., together with the fauna from the rocky
substrates, Calliostoma zizyphinum and Alvania punctura. The Bank Community
stations are defined by variation in eigenvectors one and two.

The principal-components analysis of the dead fauna shows that faunas from
stations are more similar, and the clusters are very much more indistinct than those
of the live fauna (cf. .text-figs. 3 and 5). In conclusion the presence-absence analysis
of the dead material cannot recognize the previously defined communities. From the
principal-components analysis, which is based on species abundances, some of the
stations from the more distinct communities can be recognized but the clustering

474

PALAEONTOLOGY, VOLUME 22

□17 A7
*27 043

*54 49 *23

^ ‘22
,•40

,*55

□ 39

□18

o2CL

EIGENVECTOR 2

10 1-2 1-4 18 18

□53

*70

047

'ib

,„* *11 d51

49o o18 *4

o2?1B

> 050 n2E

,1C24o«45

63«55 *4 a75

•67o29

a27

•1A

EIGENVECTOR 3

48»57

6 8 10 12 14 16 18 20 22 24 2 6

text-fig. 5. Dead fauna from stations 1-73 plotted with respect to eigenvectors one, two, and three. Key.

• Bank Facies, A Muddy Algal Gravel Facies, O Clean Algal Gravel Facies, □ Fine Sand Facies, A Mud

Facies.

based on the live fauna cannot be seen. Therefore it is unlikely that the previously
defined communities in Mannin Bay could be recognized from an analysis of the dead
material.

Comparison with previous work

Perhaps the most detailed study on the relationships between subtidal live and dead
faunas is Cadee’s (1968) study, over a period of three years, of the molluscs from Ria
de Arosa, Spain. His long-term sampling reduced the differences between the live and
dead faunas caused by small-scale time fluctuations and patchy distributions. In
addition, with surface currents having a mean maximum of 20-30 cm/sec and muddy
sediments over most of the bay little physical transport of the shells would be expected.
Cadee finds good correlations of species presences between live and dead faunas, but
their abundances vary considerably. The differences between the live and dead faunas
are thought to be caused by the long-term effects of population fluctuations, patchy
distributions, post-mortem transport of shells (particularly epifauna transported
on weed from rocky areas), and selective removal by predators.

Johnson (1965) working on molluscs from Tomales Bay, and Warme’s (1969)
study of molluscs from Mugu Lagoon, California, show that the greatest numbers of
dead species occurrences are in the dead only (exotic) group and that the occurrences
of the dead only individuals is similar to the occurrences of the live and dead individuals .
The large proportion of exotic species in Tomales Bay and Mugu Lagoon may be

BOSENCE: CORALLINE ALGAL GRAVEL FAUNAS

475

explained by the fact that the exotic material was taken with respect to each sample
and not from the sedimentary facies as is done in this study. This confirms the diffi-
culties of sampling the species of low abundances. Nevertheless, Warme and Johnson
consider that their results show a good correspondence of live and dead molluscan
faunas. The two environments are characterized by areas of both high and low
hydraulic energy.

Petersen’s (1976) study of the mollusc fauna from the very high-energy (250 cm/sec)
sandy-channel habitat in Mugu Lagoon and Tijuana Slough in California produced
results similar to those of Johnson (1965) and Warme (1969) for the presence and
absence of live and dead molluscs and the relative abundances of the live and dead
species. Of particular value in Petersen’s study is his repeated sampling of the two
environments for three times a year over three years. This repeated sampling showed
that the temporal variability of the communities was the major cause of the dis-
crepancies between his live and dead populations.

In a larger-scale study Warme et al. (1976) studied molluscs from both high- and
low-energy environments on the Yucatan shelf. They found that there was a good
correspondence between the live and dead faunas on the basis of species abundances
analysed by cluster analysis. They emphasized the time-average effect of shell
accumulation which results in the shell assemblages closely reflecting community
structure. In addition they suggest that the assemblages provide more information
on the community than one sampling of the live fauna. Post-mortem transport is not
considered important.

The results from Mannin Bay show some similarities with this previous work in
that the dead fauna contains many exotic species. However, this exotic material
consists of large numbers of species with small numbers of individuals so that the bulk
of the dead individuals come from species that live within the facies. It is perhaps
surprising that the results from the sheltered coastal lagoons are similar to those from
the more exposed environment of Mannin Bay. This similarity suggests that the
ecological factors which effect dead-shell distribution are as important as the more
commonly cited physical factors. In this study some of the exotic species can be
identified as coming from the near-by rocky areas. Where one facies is overlapping
another there is considerable mixing of both the live and the dead faunas. If the dead
faunas are examined in more detail to see if the communities and trophic nuclei can
still be recognized, then my results differ from those of previous workers. Whereas
Warme et al. (1976) find good correlation between communities recognized from the
live and the dead material, no such relation is found in Mannin Bay. Statistical
analysis of the dead material shows that the fauna has been extensively mixed and
that the live communities and the trophic nuclei can no longer be recognized. There
are a number of reasons which could explain these differences. Firstly, my community
analysis is based on the entire macrofauna whereas previous workers have just
studied molluscs. Clearly the molluscs have a more equal preservation potential than
the total fauna with both hard- and soft-bodied organisms represented. In fossil
examples from similar habitats it may be possible to predict the molluscan com-
munities from the molluscan death assemblages but these may not be a good repre-
sentation of the total macrofaunal communities or biocoenoses (Stanton 1976).
Secondly, this study is on a more detailed scale than those of Cadee (1968), Johnson

476

PALAEONTOLOGY, VOLUME 22

(1965), Warme (1969), and Warme et al. (1976). The Mannin Bay communities and
sedimentary facies have dimensions measured in terms of hundreds of metres whereas
those of previous workers are on a kilometre scale. If there are areas of mixing of dead
faunas at community boundaries then these will be more important in small-scale
studies than in large-scale studies. In addition, the initial differences between the
communities may not be as great. Thirdly, the hydraulic energy of the environment
will affect the transport and distribution of the dead fauna. Although detailed studies
are lacking, it would appear that Mannin Bay is more exposed than some of the other
areas described and therefore more post-mortem transport would be expected.

CONCLUSIONS

Descriptions of the benthic fauna from Mannin Bay, backed up with multivariate
statistical analysis of species abundances, show the existence of five communities.
The communities are specific to previously defined sedimentary facies. The fauna
from the coralline algal sediments is similar to Ford’s (1923) Venus fasciata com-
munity, but in addition has a diverse epifauna of gastropods and Crustacea together
with a boring infauna. Although no detailed studies are available from the tropical
coralline algal banks the information to date suggests a similarly diverse fauna. The
lack of any palaeosynecological studies from Tertiary coralline algal limestones
precludes comparisons or assessment of community evolution.

A study of the death assemblages produces results which differ from those of
previous workers. In particular, Warme et al. (1976) suggest that molluscan death
assemblages ‘contain more information on the total community living in a parti-
cular habitat than do the live assemblages’. In Mannin Bay the dead material cannot
be used to reconstruct the trophic nuclei or the communities. The good correlation
between the distribution of the live and dead material found by previous workers
could well be the result of the good preservation potential of molluscs. If the whole
fauna is taken into account for community evaluation then the correlation may not
be so good. In addition in this study exotic molluscs are found which are probably
derived from near-by rocky areas. The relationships between live and dead faunas
would be expected to be different in different environments. These differences will
depend particularly on temporal changes in community structure, patchy faunal
distributions, and varying amounts of post-mortem transport. Further work on this
problem is clearly required, particularly on the geologically important open-shelf
areas.

Acknowledgements. Finance and facilities for this project were provided by the Department of Geology,
Reading University. I thank Professor P. Allen for his support and Dr. R. Goldring for his supervision of
this work. The computer studies were carried out under the expert guidance of Dr. R. Till with help from
Mrs. Moss. I wish to thank Drs. G. Warner (Reading University) and B. Keegan (University College,
Galway) for helpful discussion on the faunas and their extraction. An early draft of the manuscript was read
and criticized by Dr. R. Goldring, and I am grateful for his suggestions. The boat and diving work could
not have been carried out without sterling help from Jim Watkins, Charlie Grey from Reading, and Alison
and Oliver Bosence.

BOSENCE: CORALLINE ALGAL GRAVEL FAUNAS

477

REFERENCES

birkett, l. 1953. Change in composition of fauna of Dogger Bank. Nature, Lond. 171, 265.
boekschoten, G. t. 1966. Shell borings of sessile epibiontic organisms as palaeoecological guides (with
examples from the Dutch coast). Palaeogeogr. Palaeoclimat. Palaeoecol. 2, 333-379.

— 1967. Palaeoecology of some mollusca from Tielrode sands Pliocene, Belgium. Ibid. 3, 311-362.
bonham-carter, G. F. 1967. Fortran IV program for Q-mode cluster analysis of non-quantitative data

useing IBM 7090/7094 computers. University of Kansas Contribution, 17, 1-28.
bosence, D. w. J. 1 976a . Ecological studies on two unattached coralline algae from western Ireland. Palaeonto-
logy, 19, 365-395, pis. 52, 53.

— 19766. Ecological and sedimentological studies on some carbonate sediment producing organisms,
Co. Galway, Ireland. Ph.D. thesis (unpubl.). Univ. of Reading.

— 1978. Recent carbonate sedimentation in Connemara, western Eire. A comment by H. A. Gunatilaka.
Estuar. cst. mar. Sci. 7, 303-306.

— In press. The factors leading to reef formation in Serpula vermicularis L. In rosen, b. (ed.). Biology
and systematics of colonial organisms. Systematics Association Special Publication, Academic Press.

boyd, D. w. and Newell, n. d. 1972. Taphonomy and diagenesis of a Permian fossil assemblage from
Wyoming. J. Paleont. 46, 1-14.

bualuk, w. and radwanski, a. 1968. Lower Tortonian sands at Nawodice (southern slopes of the Holy
Cross mountains); their faunal and facial development. Acta geol. Pol. 33, 447-472.
buller, a. t. 1969. Source and distribution of calcareous skeletal components in recent marine carbonate
sediments, Mannin and Clifden Bays area, Connemara, Ireland. Ph.D. thesis (unpubl.). Univ. of Reading.
buzas, m. a. 1972. Biofacies analysis of presence and absence data through canonical variate analysis.
J. Paleont. 46, 55-57.

cabioch, J. 1969. Les fonds de maerl de la Baie de Morlaix et leur peuplement vegetal. Cah. Biol. mar.
10, 139-161.

cabioch, l. 1968. Contributions a la connaissance des peuplements benthiques de la Manche occidentale.
Ibid. 9, 439-611.

cadee, G. c. 1968. Molluscan biocoenoses and thanatocoenoses in the Ria de Arrosa, Galicia, Spain.
Zool. Verh. Leiden, 95, 1-121.

davis, J. c. 1973. Statistics and data analysis in geology. 550 pp., Wiley, New York.
dorjes, J. 1972. Distribution and zonation of macrobenthic animals. In Georgia and coastal region, Sapelo
Island, U.S.A. Sedimentation and biology. Senkenberg. marit. 4, 170-183.
dybern, e. i. 1967. The settlement of sessile animals on eternite slabs in two polls near Bergen. Sarsia,
29, 137-150.

fischer, N. 1931. Local extinction of recently abundant molluscs. J. Conch. Lond. 19, 152-153.
ford, E. J. 1923. Animal communities of the level sea bottom in waters adjacent to Plymouth. J. mar. biol.
Ass. U.K. 13, 164-224.

frettar, v. and graham, A. 1962. British Prosobranch molluscs. 755 pp., Ray Society, London.
goldring, R. 1964. Trace-fossils and the sedimentary surface in shallow-water marine sediments. Pp. 136-
143. In straaten, l. m. J. u. van (ed.). Developments in sedimentology, Vol. I. Deltaic and shallow marine
deposits. Elsevier, Amsterdam.

gunatilaka, h. a. 1977. Recent carbonate sedimentation in Connemara, western Eire. Estuar. cst. mar.
Sci. 5, 609-629.

haigler, s. a. 1969. Boring mechanism of Polydora websteri inhabiting Crassostrea virginica. Am. Zool. 9,
821-828.

holme, n. a. 1966. The bottom fauna of the English Channel. Part II. J. mar. biol. Ass. U.K. 46, 401-493.
howell, B. r. and shelton, R. G. J. 1970. The effect of china clay on the fauna of St. Austell and Mevagissey
Bays. Ibid. 50, 593-607.

hughes, r. n. and thomas, m. l. h. 1971 . The classification and ordination of shallow water benthic samples
from Prince Edward Is. Canada. J. exp. mar. Biol. Ecol. 7, 1-39.

Johnson, r. G. 1965. Pelecypod death assemblages in Tomales Bay, California. J. Paleont. 39, 80-85.
jones, n. s. 1950. Marine bottom communities. Biol. Rev. 25, 283-313.

keegan, b. 1974. The marine fauna of maerl substrates on the west coast of Ireland. Cah. Biol. mar. 15,
513-530.

478

PALAEONTOLOGY, VOLUME 22

lees, a., buller, a. T. and scott, J. 1969, Marine carbonate sedimentation processes, Connemara, Ireland.

64 pp., University of Reading Geology Department Rept. No. 2.
moore, h. B. 1933. A comparison of the sand fauna of Port Erin Bay in 1900 and 1933. Proc. malac. soc.
Lond. 20, 285-294.

neyman, A. a. 1967. Limits to the application of the trophic group concept in benthic studies. Oceanology,
Acad. Sci. USSR, 7, 149-155.

pedley, H. m. 1976. A palaeoecological study of the Upper Coralline Limestone, Terebratula-Aphelesia
bed (Miocene, Malta) based on bryozoan growth form studies and the brachiopod distributions. Palaeo-
geogr. Palaeoclimat. Palaeoecol. 20, 209-234.

petersen, c. H. 1976. Relative abundance of living and dead molluscs in two Californian lagoons. Lethaia,
9, 137-148.

robertson, J. D. 1941. The function and metabolism of calcium in the invertebrates. Biol. Rev. 16, 106-133.
ryland, j. s. 1970. Bryozoans. 175 pp., Hutchinson University Library, London.

sentz-braconnot, e. 1968. Donnes ecologiques et biologiques sur la fixation des serpulidae dans la Rade
de Villefranche-sur-mar. Vie Milieu, 19, Ser.B, 109-132.
shalekova, a. 1964. New information on the calcareous algae in the bioherm limestones of the Palaeocene-
lower Eocene in western and central Slovakia. Czechoslovakia Geol. sbovnik, 15, 57-73.
silva, de p. H. D. H. 1962. Experiments on the choice of substrates by spirorbid larvae. J. exp. Biol. 39,
483-490.

sneath, p. N. A. and sokal, R. R. 1973. Numerical taxonomy. 573 pp., Freeman, San Francisco.
sokolova, M. N. 1964. Some features of the distribution of feeding groups of the deep-water benthos.
Okeanologiya, 4, 17-25.

Stanton, R. J. 1976. Relationship of fossil communities to original communities of living organisms. In
scott, R. w. and west, r. r. (eds.). Structure and classification of palaeocommunities. Dowdon Hutchison
and Ross, Stroudberg.

tebble, n. 1966. British bivalve seashells. 212 pp., British Museum (Natural History), London.
turmel, r. j. and swanson, r. 1971. The development of Rodriguez Bank. A recent carbonate mud mound.
Pp. 82-86. In multer, H. G. (ed.). Field guide to some carbonate rock environments.

— 1976. The development of Rodriguez Bank, a Holocene mudbank in the Florida reef tract.
J. sedim. Petrol. 46, 497-518.

turpayeva, ye. p. 1949. Significance of food interelationships in the structure of marine benthic bio-
coenoses. Dokl. Akad. Nauk SSSR, 15. [In Russian.]

1957. Food interelationships in marine benthic biocoenoses. Tr. In-ta okeanol. Akad. Nauk SSSR, 7.

[In Russian.]

wahlstedt, w. c. and davis, J. c. 1968. Fortran IV program for computation and display of principal
components. University of Kansas Contribution, 21, 1-27
warme, J. 1969. Live and dead molluscs in a coastal lagoon. J. Paleont. 43, 141-150.

— EKDALE, a. a., ekdale, s. F. and petersen, c. h. 1976. Raw material of the fossil record. In scott, r. w.
and west, r. r. (eds.). Structure and classification of palaeocommunities. Dowdon, Hutchison, and Ross,
Stroudberg.

wilson, J. B. 1967. Palaeoecological studies on shell beds and associated sediments in the Solway Firth.
Scott. J. Geol. 3, 329-371.

D. W. J. BOSENCE

Department of Geology

Manuscript received 3 May 1978

Revised manuscript received 29 September 1978

Goldsmiths’ College
University of London

New Cross

London SE146NW

STATOLITHS OF CENOZOIC TEUTHOID
CEPHALOPODS FROM NORTH AMERICA

by MALCOLM R. CLARKE and JOHN E. FITCH

Abstract. Statoliths of fossil teuthoids are described in detail for the first time. Several hundred statoliths collected
at eleven North American sites including Eocene, Oligocene, Miocene, Pliocene, and Pleistocene deposits are from
fourteen kinds of teuthoid. These include six species of the family Loliginidae. Five new species, Loligo applegatei
n. sp., L. mississippiensis n. sp., L. barkeri n. sp., L. valeriae n. sp., and L. stillmani n. sp. are extinct; L. opalescens
is still living off California. The new ommastrephid species Dosidicus lomita n. sp. and Symplectoteuthis pedroensis
n. sp. and the new onychoteuthid species Moroteuthis addicotti n. sp. are extinct species of genera which include
species now living off California. Fossils of Berryteuthis differ from B. magister but have not been given a specific
name. Four other kinds of Loligo statoliths are described but are immature or damaged and are not named. The
genus Loligo is found as early as the Eocene while squids of the genera Dosidicus, Symplectoteuthis, Berryteuthis,
and Moroteuthis were living in the Pliocene. The evolution of Loligo and the ecology of the species are discussed.

Statoliths of cephalopods are small, hard, calcareous stones which lie in fluid-
filled cavities or statocysts within the cartilaginous skull of members of the Octopoda,
Sepioidea, and Teuthoidea. While the internal shapes of the statocysts have been
described for many species (Ishikawa 1924; Dilly et al. 1975), very little attention has
been paid to the form of the statolith (Clarke and Fitch 1975 ; Dilly 1976). As pointed
out previously (Clarke and Fitch 1975) statoliths have shapes which are often charac-
teristic for species and description of statoliths from living cephalopods is currently
under way (Clarke 1978 and in preparation) to form a basis for a detailed study of
fossil ancestors of the living Teuthoidea and Sepioidea. The present work is the first
description of fossil statoliths and includes specimens from ten North American
Tertiary formations of Eocene, Oligocene, Miocene, Pliocene, and Pleistocene age.

All the species represented in the present collection belong to the Teuthoidea and
the majority belong to the Loliginidae. Loliginid squids are neritic, living exclusively
on or near continental shelves in near coastal waters. No loliginid is known to live
exclusively in oceanic waters and, near the edge of the continental shelf, their numbers
may change from extreme abundance to total absence within a few kilometres
although they can extend down the continental slope, near the sea bottom, to about
500-600 m. As all the formations described here contain loliginid squid statoliths
and there is no evidence to suggest that extinct loliginids had different habits from the
living forms, the formations are almost certainly derived from shallow, coastal seas.

The majority of oceanic species of teuthoid, generally grouped in the Oegopsida,
do not enter shallow coastal seas and only a few species are regularly found on the
continental shelf (Clarke 1966); species that do, nearly all belong to the family
Ommastrephidae. In the present collection four species of oegopsid are represented
and of these, two belong to this family. The close relationship of these four oegopsids
is not certainly established and this must await further work on oegopsid statoliths.

[Palaeontology, Vol. 22, Part 2, 1979, pp. 479-511, pis. 53-55.]

PALAEONTOLOGY, VOLUME 22

However, sufficient is known to make it likely that three species are new to science
and they have therefore been given specific names and provisional generic names.

Sepioids are also neritic, shelf-living cephalopods but are not represented in the
present collection. Sepia does not live in waters around North America but the genus
is known in fossil deposits (Jeletzky 1966).

Fourteen distinct types of cephalopod are present in this collection. While there is
individual variation and changes of form due to growth, the samples are large enough
and experience with living species is great enough to be sure that at least ten species
are represented and nine of these are different from living species whose statoliths
have been examined (Clarke 1978 and in preparation). Statoliths of all living species
have not been examined since they dissolve in the most commonly used preservatives
and fresh specimens need to be collected for statoliths to become available. However,
over fifty species including more than forty genera and twenty families have now been
examined by one of the authors (M. R. C.). Several hundred specimens of several
species have been examined to ascertain changes during growth and the extent of
individual variation. This basic work, while not exhaustive, gives sufficient con-
fidence to name eight new fossil species of teuthoid on the basis of the statolith alone.
While the creation of names based upon a small part of the animals is usually inadvis-
able, in the present instance it is justified by the almost total lack of other remains of
teuthoids and Octopods in the fossil record and the not inconsiderable number of
features of the statolith which can be used as criteria for identification (Clarke 1978).
Thus, the description and naming of statoliths will not conflict with names based
upon other parts of the same cephalopods (unless members of the Ammonoidea and
Belemnitida are also found to have possessed statoliths; living Nautilus does not
possess similar calcareous statoliths).

MATERIALS AND METHODS

Since squid statoliths, fish otoliths, and shells of most marine molluscs are composed
of one polymorph of calcium carbonate (i.e. aragonite), one can assume that a fossil
deposit containing an assortment of sea shells will probably contain statoliths and
earstones also. By using this philosophy as a working hypothesis, one of us (J. E. F.)
routinely sampled every shelly fossil exposure that he encountered in North America
if it was friable or could be broken down without destroying or dissolving the aragonitic
components. Not all such beds produced statoliths, but twelve of more than thirty
deposits that were sampled yielded one or more of these tiny artifacts.

When a potentially productive fossiliferous exposure was observed in a road cut, cliff, stream bank, or
at a construction site, a field sample comprising 5 to 50 kg or more of fossiliferous matrix was collected.
At the first opportunity, the field sample was allowed to soak in a tub of water and then screened through
a series of three sieves with mesh sizes of 2, 1, and 0-5 mm (approximately 10, 20, and 30 openings per
inch). To start the process, several handfuls of the saturated ‘dirt’ solution would be placed in the largest-
mesh sieve which was then partially submerged in a second tub of water where the mixture was filtered by
gently rotating and shaking the screen. The retained residue was then dumped on to several layers of news-
paper and allowed to dry in the sun. This process was continued until the entire field sample had been
sieved with the 2 mm screen. At that time, the residue which had passed through the 2 mm mesh, was pro-
cessed with the 1 mm sieve in the same manner, and finally with the 0-5 mm screen.

When the samples were dry, the coarsest fraction was screened through 6-3 mm (Finch) mesh to remove
large shells, rocks, bone fragments, etc. The residue retained by the f-inch mesh would be checked by eye,

CLARKE AND FITCH: CENOZOIC TEUTHOID STATOLITHS

but all other residue was examined by spreading a spoonful at a time in a flat dish with raised edges, and
systematically searching through this residue with a pair of forceps while viewing through a binocular
microscope at six magnifications. By having washed the samples through three sieves, the particles were
graded by size and the task of searching through the material under the microscope was greatly simplified.

If a particular exposure proved highly productive of the types of fossils desired, additional matrix, some-
times in excess of one ton, would be removed and processed. To date, no squid statoliths have been found
in the residue retained by the 2 mm sieve, but both the 1 and 0-5 mm fractions have yielded statoliths.
Unfortunately, only recently did we determine that statoliths of meso- and bathypelagic cephalopods will
pass through the finest-mesh sieves (0-5 mm) that we used.

Field samples which would not break down by soaking in water (e.g. siltstone, heavy clay, etc.), were
successfully processed by first drying the sample thoroughly and then submerging a bucketful at a time in
paraffin (kerosene) for 24 hours. When removed from the kerosene and placed in water, even rock-hard
pieces of siltstone would crumble quite rapidly so that sieving was easily accomplished.

The geographical positions of the eleven locations where samples contained statoliths of teuthoids are
given in text-fig. 1 . Description of the statoliths is based upon criteria for identification and the method
outlined elsewhere (Clarke 1978).

The living species whose statoliths have been examined and are most pertinent to the present work are
Loligo opalescens Berry, 1911, L. pealei Lesueur, 1821, L. vulgaris Lamarck, L.forbesi Steenstrup, 1856,
L. plei Blainville, 1823, Lolliguncula panamensis Berry, L. brevis (Blainville, 1823), Alloteuthis subulata
(Lamarck, 1798), Symplectoteuths oualaniensis (Lesson, 1830). Dosidicus gigas (d'Orbigny, 1835), Todarodes

text-fig. 1. Geographical positions and age of the sites from which deposits have been examined for
cephalopod statoliths. x= Eocene; • - Oligocene; ▼ = Miocene ;■= Pliocene ; A Pleistocene.

482

PALAEONTOLOGY, VOLUME 22

saqittatus (Lamarck, 1799), Berryteuthis magister (Berry, 1913), and Moroteuthis robusta (Verrill, 1876).
For each fossil species, a drawing or, if several specimens are available, a scanning electron micrograph,
measurements, and a verbal description mentioning criteria important for its identification are given. Where
a sufficient number of statoliths represent a species, regressions for measurements have been calculated,
plotted, and compared statistically. SEM photographs of a size range of two species are given to show the
usual variation seen within a species. Terms and dimensions used in the descriptions are shown in text-fig. 2.
Measurements were made with a Wild M5 microscope with an eyepiece micrometer.

Holotypes have been deposited in the Natural History Museum of Los Angeles County (LACM) and,
where numbers of specimens allow, representatives will also be deposited in the U.S. National Museum of
Natural History (USNM) Washington and the British Museum (Natural History) London (BMNH).

DESCRIPTION OF FORMATIONS AND CONTAINED STATOLITHS
WALLMEYER’S BLUFF, VIRGINIA
Age. Middle Eocene (or Lower?).

Location. Wallmeyer’s Bluff, Hanover County, Virginia. South Bank of Pamunkey River on property
owned by Mr. Wallmeyer. 1-6 km (10 miles) north of state road 732 from intersection with state road
629 then 0-3 km (0-2 miles) on private road to River Nanjemoy Formation.

Previous reports and associated fauna. Approximately 90 kg (200 lbs) of fossiliferous matrix were sampled
from this site, but only the residue retained by 1-0 mm and larger mesh was examined very carefully. The
residue retained by the 0-5 mm sieve consisted mostly of rounded, water-worn quartz sand and contained
very few otoliths, so less than a pint (0-47 1) of it was examined under the microscope. Since the single
squid statolith from this site turned up in this small amount of fine residue, a considerably greater number
of statoliths probably could be obtained with additional sampling.

The residue that was examined yielded several hundred otoliths and teeth representing about thirty-five
species of sharks, rays, and bony fishes belonging to at least twenty families. Otoliths ofcusk-eels (Ophidiidae),
eels (Congridae), flatfishes (Bothidae), and croakers and drums (Sciaenidae) were most abundant. Other
shallow-water forms included salt-water catfish (Ariidae), pearlfish (Carapidae), pterothrissids (Ptero-
thrissidae), and herrings (Clupeidae). Extant boarfishes ( Antigonia spp.) and berycids (Berycidae) are
mostly deep-water forms, but their remains were not common in this deposit. Based only upon the fish
remains at this one locality, the Nanjemoy Formation represents deposition at sea depths no greater than
about 100 m.

CEPHALOPODA, TEUTHOIDEA

Family loliginidae d’Orbigny, 1835
Loligo applegatei n. sp.

Text-fig. 3a-e

Material. The single right statolith found in this deposit is clearly distinct from any species of living loliginids
examined previously (Clarke, in preparation) or fossil loliginids described here. Its condition is good except
for a few deeply pitted areas which are probably caused by some organism capable of dissolving aragonite.
Holotype LACMIP 5756.

Diagnosis. This statolith (the holotype) may be distinguished by the following
diagnostic features (text-fig. 3a-e): The lateral dome is broadest at its dorsal end,
i.e. it is pointed in the manner typical of Loligo (Clarke, in preparation) ; it is more
sharply pointed than most other species examined. The lateral dome represents a large
area in anterior and posterior aspects and is both broad and long in comparison with
the rostrum. In outline, the lateral margin of the dome is smoothly curved and there
is no inferior lobe to the dome. There is no distinct posterior dome groove. The dorsal
dome is deflected anteriorly and extends dorsal to the point of the lateral dome.

Posterior

dorsal

ridge

ATTACHMENT

AREA

text-fig. 2. Diagrams of a generalized teuthoid right statolith to show dimensions and terms used in
descriptions. A, view of anterior side; B, view of posterior side; c, view of lateral side; D, view of anterior
side to show the principal measurements used and terms used to describe the form of the indentations.
W = width ; L = length.

PALAEONTOLOGY, VOLUME 22

text-fig. 3. Right statoliths of three species of fossil Loligo. a-e, Loligo applegatei n. sp., holotype (total
L= 1-45 mm); a, anterior view; B, posterior view; c, lateral view; D, medial; E, ventral view; f-j, the
same views of the statolith of L. mississippiensis from Chipola Fm. (total L = 1-32 mm); K-o, the same
views of the statolith of L. valeriae n. sp. Paratype, total L == 119 mm.

The well-defined spur has an eroded ventral margin but was probably fairly wide.
The dorsal and ventral indentations are approximately the same size. There is a very
small dorsal spur and no posterior indentation.

Remarks. Dimensions of this statolith are given in Table 1. As will be seen from
text-figs. 4-6 the only dimension which clearly distinguishes the species from the
other ones plotted is the maximum width (text-fig. 4). This dimension also clearly
distinguishes the specimen from the living loliginids so far examined (Clarke, in
preparation). L. plei is closest to it in this respect.

The very unusual dome of this specimen clearly separates it from other species and
justifies the erection of the species based upon one specimen only.

This is named after Shelton P. Applegate who first collected at this site and gave
us directions for locating it.

table 1. Dimensions of holotypes, some paratypes, and some figured specimens of the species described here.

485

§

SSSSS is

0.5 1.0 1.5

Total Length mm.

text-fig. 4. Statolith dimensions. Maximum width and rostral length plotted
against the total length of the statolith. Regressions for a, Loligo barkeri
n. sp. and b, L. stillmani n. sp. are included (see Table 2). □ L. applegatei
n. sp. ; ▲ Loligo sp. A ; + L. mississippiensis n. sp. from Glendon Limestone;
3 L. mississippiensis n. sp. from Chipola Formation; O • L. barkeri n. sp.
from Barker’s Ranch. Larger symbols for two and three specimens;
T L. barkeri n. sp. from Round Mount silt; A L. valeriae n. sp.; v Loligo
sp. C ; x L. stillmani n. sp.

VACAVILLE SHALE— CALIFORNIA

Age. The Vacaville Shale correlates with the Lutetian Stage of Europe, and thus is younger than the London
Clay and older than the Barton Beds (Bartonian) of England.

Location. 6-4 km (4 miles) north of Vacaville, California, in the vicinity of Dunn’s Peak along Vlatis Creek
(township 6 N., range 1 W., Vaca Valley Quadrangle, U.S. Geological Survey topographic map).

Previous reports and associated fauna. Approximately 18 kg (40 lbs) of fossiliferous matrix were collected.
This material, when screened with 50-mesh sieves, yielded about twenty otoliths and otolith fragments
(four or five species), a few shark teeth (one species), and a single tiny squid statolith (Welton, pers. comm.).

Foraminifers from this locality indicate a deep-water coastal embayment, warm subtropical to tropical
waters, and normal oceanic circulation. Although few in numbers, vertebrate remains tend to substantiate
these conclusions regarding depth and temperature at time of deposition. Previous reports on the site were
published by Palmer (1923) and Mallory (1959).

CLARKE AND FITCH: CENOZOIC TEUTHOID. STATOLITHS

487

l .O

Dorso- 0. 8 _

lateral

Length mm. 0.6 —

0.4

0.2 _

! r__ — |

0.5 1.0 1.5

Ventro- lateral Length mm.

text-fig. 5. Statolith dimensions. Dorso-lateral length plotted against ventro-
lateral length. Symbols as in text-fig. 4.

0.5 1.0 1.5 2.0

Total Length mm.

text-fig. 6. Statolith dimensions. Dorsal tip to spur and spur length plotted
against total length of statolith. Symbols as in text-fig. 4. Regressions for A,
Loligo barkeri, n. sp. and b, L. stillmani n. sp. are included (see Table 2).

PALAEONTOLOGY, VOLUME 22

CEPHALOPODA, TEUTHOIDEA

Family loliginidae d’Orbigny, 1835
Loligo sp. A

Text-fig. 7a-d

Material. One right statolith was collected in good condition but was unfortunately broken into two frag-
ments when transferred between containers.

Remarks. This is probably from a young squid since it is the smallest collected and only
has a total length of 0-84 mm. It has the narrow rounded lateral dome characteristic
of young loliginids (PL 53 and Clarke, in preparation). Because it is juvenile, has
a rather uncharacteristic shape, and there is only one, this specimen has not been
named but is probably distinct from our other Eocene species Loligo applegatei n. sp.
It almost certainly belongs to the genus Loligo since the lateral dome is widest
dorsally, the dorsal dome is small and does not protrude dorsally more than the
lateral dome and the spur and indentations are very distinct.

Distinctive features of this statolith are as follows (text-fig. 7a-d): The lateral
dome is widest near its dorsal end but is not pointed as in L. applegatei n. sp. There
is no clear posterior dome groove. The rostral angle is very obtuse. The dorsal dome
barely protrudes dorsally further than the lateral dome. The distinct spur is longer
than wide.

Dimensions of this statolith are given in Table 1. Comparing these measurements
with the other fossil loliginids described here, the dorso-lateral length (text-fig. 5), the
dorsal tip to spur and the spur length (text-fig. 6) all seem rather lower than expected
in smaller specimens of other available species. The dorso-lateral length is relatively
smaller than in living L. opalescens of the same size from California (Clarke, in
preparation).

This specimen certainly suggests that more samples from this deposit will yield
a west coast Eocene species distinct from the east coast L. applegatei n. sp.

GLENDON LIMESTONE— MISSISSIPPI
Age. Lower-Middle Oligocene.

Location. Road cut on east side of US-61 just north of intersection of US-61 bypass from Interstate-20
(Vicksburg, Mississippi).

Previous reports and associated fauna. Approximately 18 kg (40 lbs) of fossiliferous matrix yielded nearly
4000 otoliths representing at least twenty-five kinds of fish belonging to twenty or more families. Teeth
from sharks, rays, and a triggerfish added a half-dozen other species. Nearly 90% of the otoliths were from
two kinds of flatfish (Bothidae and/or Pleuronectidae), five kinds of eels (mostly Congridae), and one kind
of codiet (Bregmacerotidae). Although some species of Bregmaceros are inhabitants of the pelagic realm,
others are known to enter very shallow water including estuaries. If one assumes the Bregmaceros otoliths
in this deposit were from a shallow-water form, it would not be unreasonable to speculate that the entire
fish fauna represented a shallow, near-shore environment.

Previously Frizzell and Dante (1965) reported upon the otoliths of two species of sciaenids they found in
Glendon Limestone.

CLARKE AND FITCH: CENOZOIC TEUTHOID STATOLITHS

489

text-fig. 7. Statoliths of four species of fossil Loligo. a-d, right statolith of Loligo sp. A, total L = 0-84 mm;
a, anterior view; b, posterior view; c, lateral view; d, ventral view; e-h, the same views of the left statolith
of Loligo sp. B, Length = 1T9 mm; i-l, the same views of the right statolith of Loligo sp. C, Length =
106 mm; m-p, the same views of the left statolith of Loligo sp. D, Length = M3 mm.

490

PALAEONTOLOGY, VOLUME 22

CEPHALOPODA, TEUTHOIDEA

Family loliginidae
Loligo mississippiensis n. sp.

Plate 54, figs. 11-15

Material. Of ten statoliths belonging to this species five are in good condition and are taken as types. All
but one of the rest are almost complete. They all have the shape of adult loliginids except for two which
are smaller than the others. Holotype LACMIP 5757, paratypes LACMIP 5758-5760.

Diagnosis. Diagnostic features of this species based upon the holotype are as follows
(PI. 54, figs. 11-15): The lateral dome is widest and rather pointed at its dorsal end.
The lateral dome is not divided into superior and inferior lobes and it has a distinctly
flattened ventro-lateral side. The lateral dome is thin from anterior to posterior. The
dorsal dome extends dorsally further than the lateral dome. There is a prominent
spur which is square in shape. The ventral indentation is longer than the dorsal
indentation (and often wider in paratypes). There is no distinct antero-inferior lobe
of the lateral dome in the holotype or larger statoliths.

Remarks. Dimensions of the types are given in Table 1, and are plotted in text-figs. 4-6.
Compared with other fossil species of the same length described here, the rostral
length (text-fig. 4) and the dorsal tip to spur (text-fig. 6) tend to be smaller.

CHIPOLA FORMATION— FLORIDA

Age. Lower Miocene. Stratigraphically, the Chipola Formation is the lowest (oldest) of three deposits
which are generally conceded as comprising the Alum Bluff Group (Vernon 1942). At the Ten Mile Creek
locality, the fossiliferous portion of the Chipola Formation is approximately 4 m (12 ft) thick, although
total thickness of the formation ranges from about 6 m to 17 m (20-56 ft) (Cooke 1945).

Location. Chipola Formation, Calhoun Co., Florida. North bank of Ten Mile Creek east of bridge on state
highway 73, about 8 km (5 miles) north of Clarksville. Sites sampled are immediately below bridge and
100 m downstream.

Previous reports and associated fauna. Approximately 300 kg (700 lbs) of fossiliferous matrix from this
deposit yielded several thousand otoliths from at least forty-five species belonging to more than twenty-five
familes. In addition, there were teeth from two kinds of sharks, two rays, and a skate plus two other kinds
of bony fishes. Almost all of the fish remains were from shallow-water forms with those from gobies
(Gobiidae), grunts (Pomadasyidae), flatfish (Bothidae), and sparids (Sparidae) being most abundant. Bone-
fish ( Albula ), croaker (Sciaenidae), cardinalfish (Apogon), herring (Clupeidae), dactyloscopid, and mojarra
(Gerreidae) otoliths were also present in considerable numbers. The outcrop at this site is a fine blue-grey
to yellowish sandy clay which when thoroughly dry, breaks down readily after soaking in paraffin. The
aragonitic components (molluscs, otoliths, statoliths, etc.) are in an excellent state of preservation.

Cushman and Ponton (1932) reported upon the Foraminifera which occur in the Chipola Formation,
while the molluscan fauna has been reported upon by Gardner (1926-1950, 1936) and Vernon (1942).

q

EXPLANATION OF PLATE 53

Figs. 1-16. Eight statoliths of Loligo barkeri n. sp. to show variation in shape. 1-4, 9-12, posterior views
of right statoliths. 5-8, 13-16, anterior views of the same statoliths. Total lengths: 1 and 5 = 1-30 mm;
2 and 6=1 -44 mm; 3 and 7 = 1-50 mm; 4 and 8 = 1-56 mm; 9 and 13 = 1-60 mm; 10 and 14= 1 -64 mm;
1 1 and 15 = 1 -70 mm; 12 and 16 = 1-76 mm.

PLATE 53

CLARKE and FITCH, Loligo barkeri n.sp.

492

PALAEONTOLOGY, VOLUME 22

CEPHALOPODA, TEUTHOIDEA

Family loliginidae d’Orbigny, 1835
Loligo mississippiensis

Text-fig. 3f-j

Material. This species is represented by one right statolith which is in very good condition and has the form
of an adult Loligo.

Remarks. Dimensions of the specimen are given in Table 1, and text-figs. 4-6. They
do not separate this specimen from other species described here.

The specimen cannot be distinguished from L. mississippiensis n. sp. from the
Glendon Limestone. It has a slightly more pointed and thicker lateral dome but
these differences are insufficient to erect a new species.

BARKER’S RANCH— CALIFORNIA

Age. Middle Miocene.

Location. Barkers Ranch. Olcese Sand— Temblor Formation. Numerous localities east of Bakersfield, Kern
Co., California, and north of Kern River. The 221 squid statoliths found in Barker’s Ranch strata came from
nearly every exposure sampled, but most of them came from two outcrops, dubbed ‘Ledge Site’ and ‘Clupeid
Site’, which were sampled more extensively than the others. These two sites and five other beds that were
sampled rather heavily were exposed in several north-south trending canyons in the north-west corner of
section 33, township 28 S., range 29 E., Rio Bravo Ranch quadrangle (U.S. Geological Survey topographic
map).

Previous reports and associated fauna. Addicott (1970) in a monumental report upon the gastropods and
biostratigraphy of the Kern River area of California said that the Barker’s Ranch beds contained ‘a gastropod-
rich sublittoral molluscan fauna that has proved to be the largest, most diverse pre-Pliocene faunal unit
of the Pacific Coast States’. He further stated that ‘This fauna serves as the standard of reference ... for
the “Temblor Stage” of the Pacific coast megafossil chronology, a provincial time-stratigraphic unit
regarded as middle Miocene.’ He noted that the Middle Miocene fauna of the Olcese Sand and Round
Mountain Silt, including the Barker’s Ranch fauna, consisted of 157 species of gastropods.

The upper part of the Olcese Sand consists of fossiliferous fine to very fine sand and siltstone that are
weathered to light grey to tan. In places it is 21 m to 35 m (70-1 16 ft) thick and included in the upper part
of the Olcese Sand is the Barker’s Ranch assemblage. Over a several-year period, Fitch removed and pro-
cessed approximately 2 tons (1800 kg; 4000 lbs) of fossiliferous matrix from perhaps a dozen exposures of
Barker’s Ranch beds. This material has yielded more than 100 000 fish otoliths representing upwards of
sixty-five species belonging to thirty or more families. Additional species are represented among several
thousand shark, skate, and ray teeth, and basking shark ( Cetorhinus ) gill rakers.

Among the fish remains there is nothing to refute Addicott’s (1970) conclusion that it represents sub-
littoral deposition. There are a dozen species of drums and croakers (Sciaenidae), seven kinds of right- and
left-eyed flatfishes (Pleuronectidae and Bothidae), several basses (Serranidae), plus silversides (Atherinidae),
mullets (Mugilidae), gobies (Gobiidae), herrings (Clupeidae), and several other families which are typical
inhabitants of nearshore waters. Deep-water forms (Melamphaidae, Moridae, Myctophidae, Macrouridae,
etc.) are present, but are relatively scarce (Fitch, unpublished data).

Statoliths. These were indistinguishable from those of the Round Mountain Silt and these are treated
together below.

CLARKE AND FITCH: CENOZOIC TEUTHOID STATOLITHS

493

ROUND MOUNTAIN SILT— CALIFORNIA

Age. Middle Miocene.

Location. All of the Round Mountain Silt squid statoliths came from the canyon containing LSJU locality
2121 reported upon by Keen (1943), and from an adjacent north-south trending canyon that drains into the
Kern River. The Round Mountain Silt rests conformably upon the Olcese Sand in most localities and is
overlain by the Miocene Santa Margarita Formation.

Previous reports and associated fauna. Keen (1943) described nineteen new species of molluscs from the
Round Mountain Silt and noted that seventy-seven molluscan species had been found at a single locality
(i.e. LSJU no. 2121) east of Bakersfield, California. Although Addicott (1970) reviewed previous reports
of Round Mountain Silt assemblages and recognized that Round Mountain Silt was distinguishable
stratigraphically from Barker’s Ranch strata, he did not give separate listings for the gastropods
from the two. Addicott (1970) included the bone beds of Sharktooth Hill in the Round Mountain
Silt.

Approximately 320 kg (700 lbs) of fossiliferous matrix were processed by Fitch and examined for identifi-
able fish remains. In all, approximately 3000 otoliths were found which represented more than forty-five
species belonging to at least twenty-five families. Croakers and drums (Sciaenidae) and left- and right-eyed
flatfishes (Bothidae and Pleuronectidae) comprised about 60% of the total otoliths. Lanternfishes (Mycto-
phidae), morids (Moridae), and a few other deep-water forms were more abundant than in Barker’s Ranch
deposits, and probably reflect a slightly deeper environment than the ‘sublittoral’ assignment of Addicott
(1970) for the Barker’s Ranch molluscan assemblage.

CEPHALOPODA, TEUTHOIDEA

Family loliginidae d’Orbigny, 1835
Loligo barkeri n. sp.

Plate 53, figs. 1-16; Plate 54, figs. 1-5

Material. This species is represented by 221 statoliths in good condition and thirty-four fragments pro-
visionally identified as this species. The holotype is a left statolith having the characteristics of a mature
loliginid (LACMIP 5761). Paratypes LACMIP 5762-5812, together with material in the U.S. National
Museum of Natural History and the British Museum (Natural History).

Diagnosis. Distinctive features of this species based on the holotype are as follows
(PI. 54, figs. 1-5): The lateral dome is broadest at the dorsal end which has no sharp
point and is rounded on the dorsal side. There is no indication of an inferior lobe of
the lateral dome. The ventral border of the lateral dome may be rounded as in the
holotype or almost flat as in a few paratypes. Posteriorly, the lateral dome is pro-
minent and there is a distinct posterior dome groove (not present in a few paratypes).
The rostral angle is obtuse but distinct (not distinct in a few paratypes). The dorsal
dome extends dorsally slightly more than the lateral dome and is wider than long. It is
not distinctly separated from the lateral dome except anteriorly. The spur is very
distinct and is either ‘square’ as in the holotype or wider than long. The ventral
indentation is longer and wider than the dorsal indentation.

Remarks. Dimensions of the holotype and the paratype in Plate 54, figs. 1-5 are
given in Table 1 and are plotted for other specimens in text -figs. 4-6. This west coast
species mainly differs from the Florida Miocene Loligo mississippiensis n. sp. by
having a dorsally more rounded and thicker lateral dome. While some dimensions

494

PALAEONTOLOGY, VOLUME 22

table 2. Coefficients in mm and standard errors of regressions plotted in text-figs. 4, 5, 6 of Loligo barkeri
n. sp. and L. stillmani n. sp. and text-fig. 9 of Berryteuthis sp.

n a b sy.x sa sb

Maximum width

Loligo barkeri

50

-009

0-66

004

007

005

L. stillmani

16

015

0-45

0-07

014

010

Berryteuthis sp.

15

0-51

0-32

009

0-28

010

Lateral dome length

Berryteuthis sp.

16

0-80

014

0-06

0-16

006

Thickness

Berryteuthis sp.

16

-0-01

0-29

006

0-18

006

Rostral length

L. barkeri

50

-016

0-50

004

0-07

005

L. stillmani

16

-Oil

0-41

006

012

009

Berryteuthis sp.

15

-0-18

0-49

007

0-21

0-07

Dorsal dome length

Berryteuthis sp.

15

-0-12

0-29

007

0-22

008

Dorsal tip to spur

L. barkeri

48

019

0-31

003

006

0-04

L. stillmani

15

0-21

0-34

004

009

006

Spur length

L. barkeri

49

010

Oil

003

0-05

0-07

L. stillmani

14

0-18

0-07

004

009

006

show a different distribution to those of L. stillmani n. sp. differences between their
regressions are not statistically significant (Table 2).

The species is named after Mr. Lloyd W. Barker, a student helper who showed great
promise as a palaeontologist before his untimely death.

DAY’S POINT— VIRGINIA

Age. Upper Miocene or possibly old Pliocene. McLean (1956) reviewed most of the palaeontologic research
on the formation, and assigned it to the Miocene as had been done historically to that date. Recent research
with planktonic foraminifera (Blow 1969; Akers 1972), echinoids (Kier 1972), ostracods (Hazel 1971,
1977), and fish remains (R. L. Meyer, pers. comm. ; J. Fitch, unpublished data) has shown that its age ranges
upwards, from late Miocene into early Pliocene, at least. Gardner (1943, 1948) included Yorktown Forma-
tion mollusca in her extensive reports on Miocene and Lower Pliocene pelecypods, gastropods, and
scaphopods of Virginia and North Carolina. The beds from which our squid statolith was gleaned fit almost
in the centre of Hazel’s (1971) Orionina vaughani assemblage and thus would be middle early Pliocene in age.

Location. Day’s Point, Yorktown Formation on Edward’s Ranch 1-2 km (0-7 miles) down river from
FFA-FHA camp on south bank of James River which is 4-5 km (2-8 miles) on state highway 673 from its
junction with state highway 674 north of Smithfield Va on highway 10.

Previous reports and associated fauna. The Yorktown Formation and equivalents comprise a widespread
transgressive unit over much of the Atlantic Coastal Plain. In Virginia and North Carolina there are
numerous surface outcrops which are composed of silty sands, clays, shell marl, and coquinas. Greatest
thickness of the Yorktown Formation is reported to exceed 150 m (500 ft) (Gernant et al. 1971).

Approximately 140 kg (300 lbs) of fossiliferous matrix from this deposit yielded more than 1300 otoliths
from more than twenty-five species belonging to sixteen families. In addition, there were teeth from six

CLARKE AND FITCH: CENOZOIC TEUTHOID STATOLITHS

495

sharks, skates and rays, and a single squid statolith. Otoliths of Ammodytes sp., two kinds of bothid flat-
fishes, and three kinds of cusk-eels (Ophidiidae) comprised nearly 85% of the total otolith yield. Similar or
identical fish species live in the western Atlantic at that latitude today, and all can be taken at depths
shallower than 9 1 m (50 fathoms). The fauna identified from bottom sediments off Massachusetts (Wigley and
Stinton 1973), although several hundred miles farther north, had many similar components, especially at
the shallower depths.

CEPHALOPODA, TEUTHOIDEA

Family loliginidae d’Orbigny
Loligo sp. B

Text-fig. 7e-h

Material and remarks. The single statolith from this formation has the tip of its rostrum broken off, is
small, and has the typically rounded lateral dome and indistinct rostral angle of an immature loliginid (text-
fig. 7e-h). It closely resembles Loligo stillmani n. sp. specimens of the same size.

BOWDEN, JAMAICA

Age. Lower Pliocene? Historically, the Bowden Formation has been reported as late Miocene, but recent
investigations of planktonic foraminifera (Blow 1969) and a critical evaluation of the fish fauna (J. Fitch,
unpublished data) indicate that the Bowden ranges upwards from late Miocene into early Pliocene, at least.
The outcrops from which our squid statoliths were gleaned appear to be equivalent to Blow’s (1969) Zones
N 19 and N 20, and thus would be early Pliocene in age.

Location. Road cut on property of Bowden Estates Ltd., approximately 1-3 km (0-8 miles) on Bowden
Parochial road limits junction with main road. Exposures are at fork in road and approximately 1-2 m and
2-5 m (4-8 ft) above the road bed.

Previous reports and associated fauna. Approximately 225 kg (500 lbs) of fossiliferous matrix from this site
yielded some 25 000 otoliths representing more than 1 10 kinds of fish belonging to fifty to sixty families.
In addition, upwards of 100 teeth gleaned from this material are from at least three kinds of sharks and rays
and three families of bony fish. Obviously, the fish assemblage of the Bowden Formation is extremely rich,
a situation previously noted for the invertebrate faunas. Foraminifera, corals, bryozoans, and molluscs
have been reported upon (Woodring 1925, 1928; Blow 1969) and these as well as crustacean, echinoderm,
sponge, barnacle, and gorgonian remains were abundant in the material examined by Fitch. What was
reported as a needlefish jaw fragment from the Bowden Formation by Caldwell (1965) is actually part of
a claw from the shrimp Ctenocheles sp., family Callianassidae (W. C. Blow, pers. comm.).

Caldwell (1965) speculated that the ‘marine species from Bowden seem to represent a drifted (mixed)
assemblage from several ecological environments, as both bottom and pelagic types are represented, includ-
ing forms from the intertidal zone to a maximum depth of some 600 feet’. The fish fauna too is comprised
of both shallow- and deep-water forms, but if drift occurred, it was before the fishes deteriorated and their
otoliths fell out. One suspects that this deposit represents a mass mortality similar to those resulting from
dinoflagellate blooms (red tide organisms) off the Florida coast within the past thirty years.

Among the otoliths from this locality there were such shallow-water forms as bonefish ( Albula ), croakers
(Micropogon, Larimus, and Equetus), soldierfish ( Myripristis and Holocentrus), cardinalfish (Apogon),
jawfish ( Opisthognathus ), pearlfish ( Carapus ), plus herrings, anchovies, gobies, goatfish, gurnards, cusk-
eels, catfish, and a host of others. Fitch (196%) reported upon a laternfish found in this deposit, but deep-
water forms which are unreported include several other kinds of lanternfish, a morid ( Physiculus ), a berycid
{Diretmus), codiets ( Bregmaceros spp.), a lightfish ( Polyipnus ), a pelagic cardinal-fish ( Synagrops ), plus
eels, gonostomatids, brotulids, macrourids, melamphaids, and others. A majority of the otoliths have
crisp, uneroded margins, usually with delicate spinules and other ornamentation still intact.

496

PALAEONTOLOGY, VOLUME 22

CEPHALOPODA, TEUTHOIDEA

loliginidae d’Orbigny, 1835
Loligo valeriae n. sp.

Text-fig. 3k-o

Material. This species is represented by nine statoliths, seven of which are in good condition. Eight have
features such as long domes and short rostra which suggest the squids were not fully mature. Holotype
LACMIP 5813, paratypes LACMIP 5814-5815.

Diagnosis. The holotype is a right statolith having the following diagnostic features
(paratype in text-fig. 3 k-o is slightly less mature). The lateral dome smooth, is
widest and pointed at its dorsal end, the ventro-lateral side rounded. The lateral
dome is oval in lateral aspect. The dorsal dome extends dorsally very little further
than the ‘peak’ of the lateral dome and is not separated from the lateral dome on the
dorsal side. The spur is distinct and is longer than wide. A rather prominent antero-
inferior lobe of the lateral dome extends on to the spur. The ventral indentation is
much wider and longer than the dorsal indentation. If a posterior dome groove is
present it is indistinct.

Remarks. Dimensions of the three types are given in Table 1 and those for all seven
types are plotted in text-figs. 4-6. The species is distinguished from the other fossil
species described here by the shape of the lateral dome and the spur.

The species is named in honour of Mrs. Maurice W. Facey (Valerie), Kingston,
Jamaica without whose help the Bowden beds could not have been sampled.

NEWPORT MESA— FERNANDO FORM ATION— CALIFORNIA
Age. Late Pliocene.

Location. Fernando Formation. 240 m (800 ft) south and 15 m (50 ft) west of north-east corner section 24,
township 65, range 10 W., Newport Mesa south of Upper Newport Bay, Orange Co., California. This
deposit (LACMIP 471) was reported by Mount (1970) as being a ‘6 foot thick lens of sandy cobble con-
glomerate which occurs approximately 425 feet stratigraphically above the base of the formation’. It lies
at lat. 33° 38' 21" N., long. 117° 53' 02" W. and at the present time ‘is located under the residence at 2161
Vista Entrada, Newport Beach, California’.

Previous reports and associated fauna. Mount (1970) noted that he and associates had recovered over 200
species of larger invertebrates from the site, and listed four bivalve molluscs and three gastropods which
were characteristic of the fauna. He described a new species of Neadmete, and reported the age of the deposit
as late Pliocene. He did not offer a suggestion as to depth of deposition.

EXPLANATION OF PLATE 54

Figs. 1-20. Right statoliths of four species of fossil teuthoids. 1-5, Loligo barkerin. sp., total L = l-52mm.
1 , anterior view. 2, posterior view. 3, lateral view. 4medialview. 5, ventral view. 6-10, the same views
of the statolith of L. stillmani n. sp., holotype, total L = 1 -42 mm. 1 1 - 1 5, the same views of the statolith
of L. mississippiensis n. sp., holotype, total L = 1 -48 mm. 16-20, the same views of Moroteuthis addicotti
n. sp., holotype, total L= T48 mm.

PLATE 54

CLARKE and FITCH, Cenozoic teuthoid statoliths

498

PALAEONTOLOGY, VOLUME 22

Zinsmeister (1970) listed thirty-one species of bivalve molluscs, forty-eight of gastropods, two scapho-
pods, one brachiopod, three sharks, and thirteen bony fishes from LACMIP 471. He noted that the fauna
was comprised of forms which typically inhabit shallow as well as moderately deep water, and speculated
that the recovered fossils reflected ‘a depth between 20 and 100 fathoms’.

Fitch (1969a, 19696) reported finding over 5100 fish otoliths and 1200 elasmobranch teeth in approxi-
mately 230 kg (500 lbs) of fossiliferous matrix that he dug from this site (noted as possibly being Pico
Formation). Although the 5100 otoloths represented more than fifty-five species belonging to thirty
families, twenty-five species belonging to six families (i.e. Moridae, Myctophidae, Pleuronectidae, Scor-
paenidae, Cottidae, and Merlucciidae) contributed 90% of these (Fitch, unpublished data). Twenty-two
of these twenty-five species are still living off California today, and most can be found where water depths
range between 1 10 and 183 m (60 and 100 fathoms). Today, two of the more than fifty-five fish species are
extinct, one does not approach within approximately 1600 km of California (offshore), and the southern
limit of range for four others does not approach within 480 to 1600 km the latitude of Newport Beach
(Fitch, unpublished data).

The 230 kg (500 lbs) of matrix sampled by Fitch from this site also yielded the twenty-one squid statoliths
and fragments (three species) being reported upon in this paper.

CEPHALOPODA, TEUTHOIDEA

Family loliginidae d’Orbigny
Loligo sp. C

Text-fig. 7i-l

Material. This is only represented by one left statolith with the tip of the rostrum missing.

Remarks. It has the following diagnostic features (text-fig. 7i-l) : The lateral dome is
widest at its dorsal end but is not pointed and there is an inferior lobe. There is
a distinct posterior dome groove. The rostral angle is obtuse. The spur is longer
than wide.

This species seems close to Loligo valeriae n. sp. from the Pliocene of Jamaica
(p. 496), but is thicker and has a more distinct inferior lobe of the lateral dome.

Family onychoteuthidae

Living species of this family are oceanic but Moroteuthis robusta (Verrill) is some-
times caught in bays on the west coast of North America (Clarke 1966).

Moroteuthis addicotti n. sp.

Plate 54, figs. 16-20

Material. Three complete statoliths and three broken statoliths. Holotype LACMIP 5816, paratypes
LACMIP 5817-5821.

Diagnosis. The most important features (Plate 54, figs. 16-20) are as follows: The
shape of the lateral dome is almost semi-circular in outline in anterior aspect in the
holotype and three largest specimens. In the two smallest specimens it is broadest
at its dorsal end. The dorsal dome is very small, antero-posteriorly flattened, and is
not distinctly separated from the lateral dome except on the anterior surface. There
is no distinct medial fissure. The spur is very ill-defined or absent. The rostrum is
rather thick in comparison with its length and has a well-defined lateral lobe closer
to its ventral tip than to its dorsal end. The lateral dome has a smoother outline in
anterior aspect than in Moroteuthis robusta. The rostrum is relatively longer and
thinner than in M. robusta.

CLARKE AND FITCH: CENOZOIC TEUTHOID STATOLITHS

499

Remarks. The holotype (Plate 54, figs. 16-20) is a right statolith having the size and
appearance of a mature oegopsid and showing greater affinities to the Onychoteuthidae
and to Moroteuthis than to other families and genera (unpublished data of M. R. C.).
The first five features listed in the diagnosis indicate the generic relationship.
Dimensions of the complete types are given in Table 1.

It is named in honour of W. O. Addicott, U.S. Geological Survey, Menlo Park,
California who has been instrumental in bringing numerous fossil sites to our atten-
tion and has been helpful in numerous other ways.

Family gonatidae
Berryteuthis sp.

Remarks. Seven statoliths in good condition and seven broken statoliths represent
this species which is described below from more numerous specimens found in the
Lomita Marl.

FERNANDO FORM ATION— L ACMIP 466— CALIFORNIA
Age. Late Pliocene.

Location. During excavation of a sub-basement for the Crocker Citizens plaza building to the south side
of 6th Street between Hope and Grand, Los Angeles, a lens of the fossiliferous Fernando Formation was
exposed.

Previous reports and associated fauna. From approximately 140 kg (300 lbs) of matrix that was salvaged by
personnel from the Natural History Museum of Los Angeles County, Fitch was able to glean more than
4200 fish otoliths representing at least fifty-six species belonging to twenty-eight families. Teeth from sharks,
skates, and rays represented an additional ten species in eight families.

This deposit (LACMIP 466) was mentioned briefly (as Pico Formation) by Fitch (1969a) who reported
that hake, Merluccius productus, otoliths comprised more than 13% of the total otolith yield. The seven
kinds of lanternfish (Myctophidae) otoliths found in this deposit also were reported upon by Fitch
(19696). Four fish families (fourteen species) contributed 77% of the 4285 otoliths found in this deposit:
Bothidae (three species, 28%), Gobiidae (three species, 23%), Merlucciidae (one species, 13%), and Scor-
paenidae (seven species, 10%). Three of the fish species found in this deposit (a goby, a morid, and a cottid)
are extinct, four others ( Ammodytes , Microgadus , Lyconectes, and Malacocottus) are locally extinct
northern forms, and one (Benthosema) is a locally extinct offshore species (Fitch, unpublished data). Only
a single broken squid statolith was found in LACMIP 466.

LOMITA MARL— CALIFORNIA

Age. Late Pliocene. Based upon the molluscan assemblage, Woodring et al. (1946) assigned the Lomita
Marl to Lower Pleistocene as did both Kennedy (1975) and Langen waiter (1975).

Kanakoff and McLean (1966) described a new species of Neadmete from LACMIP 435, and interpreted
the Lomita Marl as being Late Pliocene in age although ‘previously reported in the literature as Early
Pleistocene’. They also reported that extensive excavations of these outcrops had yielded a large and unique
fauna with new fossil records of molluscs, but did not elaborate. They speculated that LACMIP 425 prob-
ably corresponded to U.S. Geological Survey locality no. 12222 of Woodring et al. (1946).

Zinsmeister (1970) in reporting upon a Pliocene faunal assemblage at Newport Beach stated that he made
comparisons with material from Lomita Marl and found numerous similarities; whereas, Hertlein (1970)
in describing a new species of Kelletia from LACMIP 435 noted that the Lomita Marl was ‘Late Pliocene
or Early Pleistocene’.

Based upon his work with fossil fish otoliths, and intensive sampling of the Lomita Marl by personnel
from the Natural History Museum of Los Angeles County, Fitch (1969a) concluded there was substantial
evidence that the Lomita Marl was in fact the youngest marine Pliocene unit in southern California.

500

PALAEONTOLOGY, VOLUME 22

Location. LACMIP 435. 12 m (40 ft) long exposure on south side of gully north-west of intersection of Park
Western Drive and Host Place, San Pedro, California. Exposure is 10-7 m (35 ft) below Host Place road bed.

Previous reports and associated fauna. Woodring et at. (1946) noted that the Lomita Marl consists of
a variety of calcareous rocks, principally marl and calcareous sand. They used the term ‘calcareous sand’ to
include ‘unconsolidated calcareous material of sand or granule size composed chiefly of calcareous organic
remains— calcareous algae, Foraminifera, Bryozoa, small shells, and shell fragments’. Their investigations
led them to believe that the Lomita Marl included several faunal associations which they interpreted as
being different depth associations ranging from ‘shallow water to about 100 fathoms’. In a listing of ‘Pleisto-
cene mollusks locally extinct in the latitude of San Pedro but now living farther north or south, including
forms that are not known to be living but that are closely related to living forms’, they note four gastropods
and three pelecypods having northern affinities and four gastropods and one pelecypod with southern ties.

Kennedy (1975) gave a relatively complete bibliography for faunal assemblages reported from Lomita
Marl. Included were references to publications on foraminifera, ostracods, molluscs, and vertebrates.
Langenwalter (1975) reported fish species, from a list supplied by Fitch, that had been identified from
Lomita Marl.

Over one tonne (2000 lbs) of fossiliferous matrix from LACMIP 435 yielded nearly 25 000 fish otoliths
representing more than eighty-seven species belonging to thirty families (Fitch 1969a, and unpublished
data). In addition, teeth and other remains from seventeen kinds of sharks, skates, and rays belonging to
twelve families have been found at this site. The fish fauna is composed of many mesopelagics, a few
bathypelagics, two extinct forms, nine locally extinct northern and/or offshore species, and an assortment
of species that are typical inhabitants of 37-90 m depths at the latitude of San Pedro today. The 2424
otoliths from an extinct morid represented the greatest yield for a single species, but four other species
yielded more than 1000 otoliths each; a flatfish, Lyopsetta exilis (2158); a goby Coryphopterus nicholsii
(1842); a cottid, Radulinus asprellus (1373); and a myctophid, Stenobrachius leucopsarus (1119). The
fourteen species of myctophids from this site (over 3100 otoliths) have been reported upon by Fitch (19696),
who speculated that the Lomita Marl was laid down ‘at depths exceeding 600 feet’.

The 185 squid statoliths noted by Clarke and Fitch (1975) from the Lomita Marl were from this site
(LACMIP 435) and represent four or possibly five species as reported herein.

CEPHALOPODA, TEUTHOIDEA

Family loliginidae
Loligo stillmani n. sp.

Plate 54, figs. 6-10

Material. This species is represented by thirteen statoliths in good condition and twenty-five broken or
badly chipped statoliths. The holotype (LACMIP 5822) is a right statolith having the characteristics of
a mature loliginid (PI. 54, figs. 6-10). Paratypes LACMIP 5823-5826.

Diagnosis. Distinctive features of this species based on the holotype are as follows
(PL 54, figs. 6-10): The lateral dome is broadest at the dorsal end which is slightly
more pointed than in Loligo barkeri n. sp. or L. opalescens. There is no indication
of an inferior lobe to the lateral dome. There is no posterior dome groove. The lateral
dome is thinner than in L. barkeri n. sp. The rostral angle is obtuse but distinct.
The dorsal dome extends dorsally much more than the lateral dome and is longer
than wide. It is separated from the lateral dome on the dorsal edge by a shallow but
distinct indentation and forms a rather narrow anteriorly curving ‘blade’. The spur
is distinct but not so prominent as in L. opalescens and is much longer than wide.

Remarks. The larger paratypes agree with the holotype in the above features.
Dimensions of the holotype are given in Table 1 and of paratypes in text-figs. 4-6.
The most obvious differences between the species and L. barkeri n. sp. are the more

CLARKE AND FITCH: CENOZOIC TEUTHOID STATOLITHS 501

dorsally directed dorsal dome, the sharper lateral point and flatter ventro-lateral side
of the lateral dome, the slimmer rostrum and the longer spur of this species.

From text-figs. 4-6 the dimensions of this species and L. barkeri n. sp. would appear
to have different but overlapping distributions. Rostral length is generally shorter in
L. stillmani for statoliths of the same length (text-fig. 4), the maximum width tends to
be smaller in statoliths of the same size (above about 1-3 mm length, text-fig. 4), the
distance from the dorsal tip to the spur is greater in statoliths of the same size and the
spur length tends to be greater (text-fig. 6). However, when the calculated regressions
were tested (Table 2) they were not statistically significant at the 95% level.

The species is named after Dr. Stillman S. Berry who has collected extensively in
the Lomita Marl and has made a great contribution to the taxonomy of living
cephalopods over a period of almost 70 years.

Loligo sp. D

Text-fig. 7m-o

Material. One small statolith having characteristics which suggest immaturity were also present in this
sample. It is possibly an immature Loligo stillmani n. sp. but may represent a distinct species.

Remarks. Characteristic features of this statolith are as follows : The lateral dome is
widest at the dorsal end. The inferior lobe of the lateral dome is relatively large. The
lateral dome is long compared with the rostrum (suggesting immaturity). The rostral
angle is obtuse and indistinct. The dorsal dome does not extend dorsally much more
than the lateral dome from which it is only very indistinctly separated. The medial
fissure is shallow. The spur is longer than wide. The surface is rough and in places
pitted.

Dimensions of the specimen shown in text-fig. 7m-o are given in Table 1.

Family ommastrephidae

Two statoliths, although clearly belonging to different species have some features
which relate them to this family. While the one is probably in the genus Dosidicus
the generic status of the other is not certain but it has close similarities to Symplecto-
teuthis. Characteristics which suggest that these belong to this family are the large
size and rough surface of the dorsal dome, the smooth inferior lobe of the lateral
dome, the distinct medial fissure, the prominent spur and the rostral angle which is
almost a right angle.

Dosidicus lomita n. sp.

Text-fig. 8a-e

Material. A single specimen, holotype LACMIP 5827, is a right statolith having mature features.

Diagnosis. Distinctive features of this statolith are as follows (text-fig. 8a-e): The
lateral dome is divided into a slightly rough superior lobe, a prominent, smooth
postero-inferior lobe, and a smooth antero-lateral lobe. The surface is slightly lumpy
on the antero-ventral side. There is a distinct posterior dome groove. The rostral
angle is close to a right angle. The dorsal dome is very large, rough, and indistinctly

502

PALAEONTOLOGY, VOLUME 22

separated from the lateral dome. A distinct spur is present but is eroded on the dorsal
side. Well-defined indentations are present but the margins are eroded. The medial
fissure is deep and forms a well-defined, triangular posterior indentation. A distinct
posterior ventral ridge is present. The rostrum has a lumpy surface, is approximately
circular in cross section at the base flattening towards the tip, and is fairly thick. The
tip is slightly flexed laterally.

Remarks. Dimensions of this statolith are given in Table 1.

This statolith resembles Dosidicus gigas caught off. the Californian coast by its
general shape, particularly the relative proportions of the dorsal dome, lateral dome,
and rostrum, by the lumpy surface of the dorsal dome and superior lobe of the lateral
dome, by the possession of a distinct superior lobe of the lateral dome, by the very
distinct posterior dome groove, by the lumpy surface of the rostrum, and by having
a posterior ventral ridge. It differs from D. gigas in having a smaller lobe placed
anteriorly instead of laterally on the dorsal end of the lateral dome and in having
a rather thicker and blunter rostrum. The first difference is certainly sufficient to show
the fossil is a distinct species although it seems possible the antero-lateral lobe of the
fossil has evolved into the superior lobe of the living species.

The species name is taken from the name of the fossil deposit in which it was first
discovered.

text-fig. 8. Statoliths of two species of fossil ommastrephid squids, a-e, left statolith of Dosidicus lomita
n. sp. Holotype, total L = 2-52 mm; A, anterior view; b, posterior view; c, lateral view; D, medial view;
e, ventral view ; f-j, the same views of the left statolith of Symplectoteuthis pedroensis n. sp. Holotype, total

L = 2-26 mm.

CLARKE AND FITCH: CENOZOIC TEUTHOID STATOLITHS

503

Symplectoteuthis pedroensis n. sp.

Text-fig. 8f-j

Material. Holotype LACMIP 5828.

Diagnosis. Distinctive features of this species are as follows (text-fig. 8f-j) : The out-
line of the lateral dome and dorsal dome together form an approximate semicircle
in anterior or posterior aspect. The statolith is wide. The lateral dome has a distinct,
shallow groove separating a lobe from a postero-inferior lobe. There is no distinct
posterior dome groove. While the rostral angle is obtuse because of the presence of
a small dorso-lateral rostral lobe, the ventral side of the lateral dome forms a right
angle with the rostrum. The dorsal dome is rough and is not separated from the lateral
dome by a distinct groove. The spur is well-defined and is joined by a prominence to
the superior lobe of the lateral dome. The medial fissure is deep and forms a pointed
posterior indentation with a posterior ventral ridge. There is a dorsal spur, and
a distinct anterior dorsal ridge. The rostrum is relatively small, equalling about one
quarter of the total length of the statolith and being rather narrow.

Remarks. Dimensions of this statolith are given in Table 1. This species differs from
Symplectoteuthis oualaniensis caught off California by having a relatively shorter
rostrum and a lateral lobe near the base of the rostrum which is closer to the rostral
angle than in S. oualaniensis. It is quite possible that S. pedroensis n. sp. is ancestral
to S. oualaniensis. The fossil, however, is much larger than statoliths of the
S. oualaniensis normally collected off California and very large specimens of the
latter might differ less from the fossil. S. oualaniensis grows much larger in some other
localities (e.g. the Indian Ocean).

Family gonatidae
Berryteuthis sp.

Plate 55, figs. 1-20; text-fig. 11, figs. 1-5

Material. This species is represented by twenty-six almost intact statoliths and eighty-five large, identifiable
fragments. LACMIP 5829-5839.

Diagnosis. This species has the following distinctive features (PI. 55, figs. 1-20 and
text-fig. 11, figs. 1-5): The lateral dome looks almost spherical in posterior and, in
some specimens, also in lateral view. In lateral aspect it is usually oval with its longi-
tudinal axis on the same axis as the statolith. The lateral dome has a smooth surface
and has no posterior dome groove. The rostral angle is almost a right angle (in some
specimens the lateral dome bulges ventral to it and so forms an acute angle). The
dorsal dome has a smooth surface and is large; it extends dorsally well beyond
the lateral dome from which it is clearly separated by a shallow groove on the dorso-
lateral surface. The spur is distinct but is much longer than wide. The indentations are
indistinct, narrow and shallow. The medial fissure is shallow. There is no dorsal spur
or anterior dorsal ridge. The rostrum is long, about the same length as the lateral
dome and thick. Its thickest point is not usually at its base. The wing is thick.

Remarks. Dimensions of one specimen are given in Table 1 and of others in text-fig. 9,
and variation in form is shown in Plate 55, figs. 1 -20.

504

PALAEONTOLOGY, VOLUME 22

The shape of the large statoliths is very close to that of the very characteristic
statolith of Berryteuthis magister and there can be little doubt of their generic affinity.
However, statistical comparisons of the measurements show conclusively that the
fossil is a different species from B. magister (in preparation). A second species of the
genus, B. anonychus (Pearcy and Voss, 1963) lives off the western coast of North
America and, as we have not yet had an opportunity to examine statoliths from this
species, we have not given the fossils a specific name in case they should later prove to
belong to the latter species.

The form of the statoliths of Berryteuthis has some superficial similarities to both
the ommastrephid Todarodes and the sepioid Sepia. All three genera have large
statoliths with lateral domes having semicircular outlines in posterior aspect, large
dorsal domes, and thick rostra. As all three species live close to the sea bottom on the
continental shelf and upper slope at some time in their lives, the similarities in
statolith shape could have some relationship to their way of life.

RINCON HILL ROAD

Age. Early Pleistocene.

Location. Rincon Hill Road (Bates Road) Santa Barbara, California. Santa Barbara Formation (this
deposit is equivalent to Timm’s Point Silt to judge from the type of matrix and fish otoliths sorted by Fitch).
The exact spot is on the west bank of Rincon Hill Road about 0-25 km from the intersection of Rincon Hill
Road and California State Highway 1 50.

Previous reports and associated fauna. Approximately half a tonne (1000 lbs) of fossiliferous matrix from
a Pleistocene exposure on Rincon Hill Road contained slightly more than 2000 fish otoliths, three squid
statoliths, and a small assortment of shark, skate, and ray teeth (Fitch and Huddleston MS.). Some of the
fish otoliths found at this site were mentioned under the heading of ‘Bates Road locality’ by Fitch (1969a,
19696). In all, more than fifty species of fish from twenty-one families were represented among the 2000
otoliths. Four kinds of otoliths (two kinds of Gobiidae, a morid, and an embiotocid) were from extinct
species, one was from a locally extinct northern species ( Lyconectes ), and one was from a locally extinct
offshore myctophid ( Benthosema) \ the remaining forty-five or more kinds were from fishes that today
inhabit depths of 20-180 m at the latitude of this site. Fitch (1969a, 19696) speculated that this deposit was
laid down at a depth of 120-180 m (400-600 ft) at a time when ocean temperatures were colder at this
latitude than they are today.

The most conspicuous molluscs at this site were species of Turritella and Dentalium. Other large molluscs
encountered while digging out field samples were species of Astraea, Bursa, Polinices, Nassarius, Panopea,
and Saxidomus. Several large abalone shells, Haliotis rufescens, have been unearthed in a lens of material
adjacent to this site but unquestionably of the same age.

This site has been registered as vertebrate palaeontology collecting locality LACM 3784.

CEPHALOPODA, TEUTHOIDEA

Family loliginidae
Loligo opalescens Berry, 1911

Remarks. One complete right statolith and two damaged statoliths cannot be dis-
tinguished from this species which now lives in inshore waters off the western coast
of the United States of America. The statolith of this species will be described else-
where with other living species of the genus (Clarke, in preparation).

maximum width

H-

lateral dome length

text-fig. 9. Dimensions of statoliths of Berryteuthis sp. plotted against their total length. Coefficients and
standard errors of the regressions plotted are given in Table 2.

506

PALAEONTOLOGY, VOLUME 22

EVOLUTION

While the specimens described are too few to enable general deductions on the
evolution of teuthoids, the fact that at least six species of Loligo are represented,
provides a basis for a discussion of the evolution of this one genus. If we first consider
the Loligo from the Atlantic, L. applegatei n. sp. is distinguished from any other
Loligo examined, fossil or living, by its very large, pointed lateral dome. It is unlikely
to have evolved into the Oligocene L. mississippiensis n. sp. The latter might have
evolved from the Loligo sp. A of the Vacaville which is not likely to be a young stage
of L. applegatei n. sp.

The early Pliocene L. valeriae n. sp., also from the Atlantic side of the continent,
differs from L. mississippiensis n. sp. in having a rounded ventro-lateral side of the
lateral dome, a very long spur, and a relatively smaller and less dorsally directed
dorsal dome. These suggest an independent line of development.

On the Pacific side of the continent the Miocene L. barkeri n. sp. is much closer to
the living L. opalescens than either is to the Pliocene L. stillmani n. sp. which has
a sharper point to the lateral dome, no posterior dome groove, a very long spur, and
a narrower, more pointed, rostrum. L. barkeri sp. quite possibly evolved into the
present-day L. opalescens but L. stillmani n. sp. is likely to be from a different line of
evolution.

The two living ommastrephid genera Dosidicus and Symplectoteuthis are now
known to have lived in the Pliocene although the statoliths are sufficiently different
from the living species to consider them different species; their form does not pre-
clude evolution of the Pliocene forms into the living species.

Similarly, the fossil Berryteuthis sp. is very close to B. magister from the North
Pacific and it is clear that squids of this genus lived in the Pliocene seas of the western
side of North America.

Moroteuthis addicotti n. sp. is close to, but distinct from the living M. robusta
(Verrill, 1876) which is frequently caught off California and may be its ancestor.

ECOLOGY

The living species of Loligo are neritic species living in midwater or just above the
bottom in water less than 200 m for much of their lives. At times, some species extend
into deeper waters to 500-600 m near the bottom on the continental slope but other-
wise they are not found away from the continental shelf or in oceanic water. Only
one species of Loligo , L. opalescens lives off California and two, L. pealei Lesueur,

EXPLANATION OF PLATE 55

Figs. 1-20. Ten statoliths of Berryteuthis sp. to show variation in shape. 1-5, 11-15, posterior views of
right statoliths. 6-10, 16-20, anterior views of the same statoliths. Total lengths 1 and 6 = 2-56 mm;
2 and 7 = 2-68 mm; 3 and 8 = 2-76 mm; 4 and 9 = 2-80 mm; 5 and 10 = 2-80 mm; 11 and 16 = 2-92 mm;
12 and 17 = 3-00 mm; 13 and 18 = 3 00 mm; 14 and 19 = 3-20 mm; 15 and 20 = 3-40 mm.

PLATE 55

CLARKE and FITCH, Statoliteuthis enigmaticus n.sp.

508

PALAEONTOLOGY, VOLUME 22

1821 and L. plei (Blainville, 1823), live off the Atlantic side of the continent. Other
loliginids from the shelf around North America are Sepioteuthis sepioidea (Blainville,
1823), and Lolliguncula brevis (Blainville, 1823) which only live on the Atlantic side
and L. panamensis from the Gulf of California. The statoliths of all these except
S. sepioidea have been examined and compared with the fossil statoliths (Clarke, in
preparation).

The ommastrephids Symplectoteuthis oualaniensis and D. gigas and the onycho-
teuthid M. robusta are all caught close to the Californian coast at times. Berryteuthis
magister lives close to the sea floor on the upper part of the continental slope.

text-fig. 10. Percentage frequency distributions for the
two most numerous fossil statoliths. A, Statoliteuthis
enigmaticus n. sp. based on 117 measurements of lateral
dome length, b, Loligo barkeri n. sp. based on 3 1 1 measure-
ments of total length.

Thus, all the species described here support the view that all sediments examined
in the present work originated from water which was probably shallower than 200 m
on the margin of a continent.

Size frequency histograms of the most numerous species in the collection, Loligo
barkeri n. sp. and Berryteuthis sp., are not distinctly skewed (text-fig. 10). This
suggests that the mesh size used in their collection was sufficiently small to collect the
majority of the statoliths of these species which were present in the ‘dirt’ sample;
otherwise the distributions would be positively skewed. In addition, the lack of such
skewness also shows that the statoliths were not from a population including every
stage of growing squid. Instead, they were from ‘adult’ or near ‘adult’ squid which,
from our knowledge of living species, were probably spawning and dying in the
locality. When such spawning aggregations occur in living squids they are seasonal.
We may therefore be fairly confident that L. barkeri n. sp. and Berryteuthis sp.
aggregated at some season of the year for spawning and death as do living species of
Loligo and Todarodes.

50-

A

20

CLARKE AND FITCH: CENOZOIC TEUTHOID STATOLITHS

509

1 2 3 4 5

text-fig. 11. Figs. 1-5. The right statolith of Berryteuthis sp., total L = 2-77 mm; 1, anterior view;
2, posterior view ; 3, lateral view ; 4, medial view ; 5, ventral view.

DISCUSSION

The present work forms a basic framework for a study of the evolution of shallow-
water teuthoids, particularly Loligo, during the Cenozoic of North America. Because
living teuthoid species are not very numerous and there are only three living species
of Loligo known from North America, an evolutionary tree is not likely to be as
complicated as in many other animal classes. Because there are so few fossil remains of
teuthoids, evolutionary conclusions based upon statoliths are not likely to conflict
seriously with conclusions based upon other remains.

The function of the cephalopod statolith is not known precisely at present but it is
probably concerned with monitoring motion of the animal. It might be expected,
therefore, that the shape relates in some way to the type of motion and hence to the
features of the animal concerned with locomotion such as development of mantle
musculature, fin size, and presence of a means to attain neutral buoyancy. Such
a relationship is not obvious at present but may become more obvious as our know-
ledge of locomotion and the role of the statolith develops.

Acknowledgements. Many individuals contributed towards this work. Some informed us of fossil exposures
or helped excavate and process samples from various sites. Others gave or lent comparative material, and
almost everyone we consulted willingly shared their ideas or special talents. We especially appreciate
assistance received from Warren O. Addicott, J. A. R. Ally, Shelton P. Applegate, Warren C. Blow, Lloyd, W.
Barker, Martin W. Cawthorn, S. Stillman Berry, David K. Caldwell, William L. Craig, Valerie Facey,
Richard A. Fitch, Kurt Geitzenauer, Jack and Mary Hopkins, Richard W. Huddleston, George Kanakoff,
Tsunemi Kubodera, Robert J. Lavenberg, Heinz Lowenstam, Richard McGinnis, Mamoru Murata,
Roger D. Reimer, Mark Roeder, Graham J. B. Ross, Jack W. Schott, Teruo Tobayama, Bruce Welton, and
Charles Wenner. If we have failed to acknowledge assistance, and in a project extending over such a lengthy
period we are certain to have overlooked someone, it has not been intentional. John Fitch received partial
support with a grant from the Penrose Fund of the American Philosophical Society.

510

PALAEONTOLOGY, VOLUME 22

REFERENCES

addicott, w. o. 1970. Miocene gastropods and biostratigraphy of the Kern River area, California. Prof.
Pap. U.S. geol. Surv. 642, 1-174.

akers, w. h. 1972. Planktonic foraminifera and biostratigraphy of some Neogene formations, northern
Florida and Atlantic coastal plain. Tulane Stud. Geol. Paleontology, 9, 1-139.
blow, w. h. 1969. Late middle Eocene to recent planktonic foraminiferal biostratigraphy. Proc. First int.
Conf. plankton. Microfossils. 1, 199-422.

caldwell, D. k. 1965. A Miocene needlefish from Bowden, Jamaica. Q. Jl Fla Acad. Sci. 28, 339-344.
clarke, m. R. 1966. A review of the systematics and ecology of oceanic squids. Adv. mar. Biol. 4,
91-300.

— 1978. The cephalopod statolith: An introduction to its form. J. mar. biol. Ass. U.K. 58, 701-712.

— and fitch, J. e. 1975. First fossil records of cephalopod statoliths. Nature, Lond. 257, 380-381.
cooke, G. w. 1945. Geology of Florida. Geol. Bull. Fla. 29, 1-339.

cushman, J. a., and ponton, G. m. 1932. The foraminifera of the upper, middle and part of the lower
Miocene of Florida. Geol. Bull. Fla. 9, 1-147.

dilly, p. n. 1976. The structure of some cephalopod statoliths. Cell Tissue Res. 175, 147-163.

— Stephens, p. r. and young, J. z. 1975. Receptors in the statocyst of squids. J. Physiol., Lond. 249,
59-61.

fitch, J. e. 1969a. Fossil records of certain schooling fishes of the California Current system. California
Marine Research Committee. CALCOFI Rept. 13, 71-80.

— 1969ft. Fossil lanternfish otoliths of California, with notes on fossil Myctophidae of North America.
Contr. Sci. Los. Ang. City Mus. nat. Hist. 173, 1-20.

— and Huddleston, r. w. Marine fish remains from an early Pleistocene deposit in Ventura County,
California. (Unpubl. MS.)

Frizzell, D. L. and dante, J. h. 1965. Otoliths of some early Cenozoic fishes of the Gulf Coast. J. Paleont.
39, 687-718.

Gardner, J. a. 1926-1950. The molluscan fauna of the Alum Bluff Group of Florida. Prof. Pap. U.S. geol.
Surv. 142, 709 pp.

— 1936. Additions to the fauna of the Alum Bluff group of Florida. Geol. Bull. Fla. 14, 1-82.

— 1943. Mollusca from the Miocene and lower Pliocene of Virginia and North Carolina. Part 1.
Pelecypoda. Prof. Pap. U.S. geol. Surv. 199A, 1-178.

— 1948. Mollusca from the Miocene and lower Pliocene of Virginia and North Carolina. Part 2.
Scaphopoda and Gastropoda. Ibid. 199B, 179-310.

gernant, r. e., gibson, T. G. and Whitmore, f. c. jun. 1971. Environmental history of Maryland Miocene.
Guidebk Md Geol. Surv. 3, 1-58.

hazel, j. e. 1971. Ostracode biostratigraphy of the Yorktown Formation (upper Miocene and lower
Pliocene) of Virginia and North Carolina. Prof. Pap. U.S. geol. Surv. 704, 1-13.

— 1977. Distribution of some biostratigraphically diagnostic ostracodes in the Pliocene and lower
Pleistocene of Virginia and northern North Carolina. J. Res. U.S. geol. Surv. 5, 373-388.

hertlein, l. G. 1970. A new species of fossil Kelletia (Mollusca: Gastropoda) from the Lomita Marl,
Late Cenozoic of San Pedro, California. Contr. Sci. Los Ang. Cty Mus. nat. Hist. 190, 1-8.
ishikawa, m. 1924. On the phylogenetic position of the cephalopod genera of Japan based on the structure
of statocysts. J. Coll. Agric. imp. Univ. Tokyo, 7, 165-210.
jeletzky, j. a. 1966. Comparative morphology, phylogeny and classification of fossil Coleoidea. Paleont.

Contr. Univ. Kansas, 42 (Mollusca, article 7), 162 pp.
kanakoff, G. p. and mclean, J. h. 1966. Recognition of the cancellariid genus Neadmete Habe, 1961, in
the West American fauna, with description of a new species from the Lomita Marl of Los Angeles County,
California. Contr. Sci. Los Ang. Cty Mus. nat. Hist. 116, 1-6.
keen, a. m. 1943. New mollusks from the Round Mountain silt (Temblor) Miocene of California. Trans.
S. Diego, Soc. nat. Hist. 10, 25-60.

Kennedy, G. L. 1975. Paleontologic record of areas adjacent to the Los Angeles and Long Beach Harbors,
Los Angeles County, California. In soule, d. f. and oguri, m. (eds.). Marine studies of San Pedro Bay,
California. Part 9. Paleontology, 1-35. Allan Hancock Foundation and University of Southern California
Sea Grant Programs, Los Angeles.

CLARKE AND FITCH: CENOZOIC TEUTHOID STATOLITHS

511

kier, p. M. 1972. Upper Miocene echinoids from the Yorktown Formation of Virginia and their environ-
mental significance. Smithson. Contr. Paleobiol. 13, 1-41.
langenwalter, p. E., ii. 1975. The fossil vertebrates of the Los Angeles-Long Beach Harbors region. In
soule, d. f. and oguri, m. (eds.). Marine studies of San Pedro Bay, California. Part 9. Paleontology,
36-54. Allan Hancock Foundation and University of Southern California Sea Grant Programs, Los
Angeles.

mclean, J. d., jun. 1956. The foraminifera of the Yorktown Formation in the York-James Peninsula of
Virginia, with notes on the associated mollusks. Bull. Am. Paleont. 36, 260-394.
mallory, v. s. 1959. Lower Tertiary biostratigraphy of the California Coast Ranges. American Association
Petroleum Geologists, Tulsa, Oklahoma, 416 pp.

mount, J. d. 1970. A new species of Neadmete (Neogastropoda) from the Pliocene of California. Contr.
Sci. Los Ang. Cty Mus. nat. Hist. 177, 1-4.

palmer, d. b. k. 1923. A fauna from the Middle Eocene shales near Vacaville, California. Univ. Calif.
Pubis Bull. Dep. Geol. 14, 289-318.

vernon, R. o. 1942. Geology of Holmes and Washington Counties, Florida. Geol. Bull. Fla. 21, 1-161.
wigley, r. l. and stinton, f. c. 1973. Distribution of macroscopic remains of Recent animals from marine
sediments off Massachusetts. Fishery Bull. natn. mar. Fish. Serv. 71, 1-40.
woodring, w. p. 1925. Miocene mollusks from Bowden, Jamaica. I. Pelecypods and scaphopods. Pubis
Carnegie Instn. Wash., no. 366,222 pp.

— 1928. Miocene mollusks from Bowden, Jamaica. II. Gastropods and discussion of results. Ibid.,
no. 385, 564 pp.

— bramlette, m. n. and kew, w. s. w. 1946. Geology and paleontology of Palos Verdes Hills, California.
Prof. Pap. U.S. geol. Surv. 207, 1-145.

zinsmeister, w. J. 1970. A Late Pliocene macrofossil fauna of Newport Beach, Orange County, California.
Bull. sth. Calif. Acad. Sci. 69, 121-125.

M. R. CLARKE

The Laboratory
Citadel Hill
Plymouth PL1 2PB

J. E. FITCH

Typescript received 21 July 1978
Revised typescript received 31 October 1978

Operations Research Branch
Dept, of Fish and Game
350 Golden Shore
Long Beach
California 90802
U.S.A.

THE PALAEONTOLOGICAL ASSOCIATION

The Association was founded in 1957 to further the study of palaeontology. It holds meetings and
demonstrations as well as publishing Palaeontology and Special Papers in Palaeontology. Membership is
open to individuals and to institutions on payment of the appropriate annual subscription. Rates for
1979 are:

Institutional membership ..... £25 (U.S. $50)

Ordinary membership . . . £12 (U.S. $24)

Student membership ..... £7-50 (U.S. $12)

There is no admission fee. Correspondence concerned with Institutional Membership should be addressed to
Dr. C. H. C. Brunton, Department of Palaeontology, British Museum (Natural History), Cromwell Road,
London SW7 5BD, England. 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. Subscriptions cover one calendar year and are due each
January; they should be sent to the Membership Treasurer (for name and address see Council list
below).

PALAEONTOLOGY

All members whojoinfor 1979 will receive Volume 22, Parts 1-4. All back numbers are still in print and may
be ordered from B. H. Blackwell, Broad Street, Oxford 0X1 3BQ, England, at £12 per part (post free).
A complete set, Volumes 1-21, consists of 83 parts and costs £996.

SPECIAL PAPERS IN PALAEONTOLOGY

Members may subscribe to this by writing to the Membership Treasurer (for name and address see Council
list below); the subscription rate for 1979 is £20 (U.S. $40) for Institutional Members, and£10 (U.S. $20) 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 prices shown in the list inside the front
cover. Non-members may obtain copies, at the listed prices, from B. H. Blackwell, Broad Street, Oxford
OX1 3BQ, England.

COUNCIL 1979-1980

President : Professor H. B. Whittington, Department of Geology, Sedgwick Museum, Cambridge CB2 3EQ
Vice-Presidents : Dr. E. P. F. Rose, Department of Geology, Bedford College, Regent’s Park, London NW1 4NS
Dr. C. T. Scrutton, Department of Geology, The University, Newcastle upon Tyne NE1 7RU
Treasurer: Mr. R. P. Tripp, High Wood, West Kingsdown, Sevenoaks, Kent TNI 5 6BN
Membership Treasurer: Dr. J. C. W. Cope, Department of Geology, University College, Swansea SA2 8PP
Secretary: Dr. R. Riding, Department of Geology, University College, Cardiff CF1 1XL
Editors

Professor C. B. Cox, Department of Zoology, King’s College, Strand, London WC2R 2LS
Dr. M. G. Bassett, Department of Geology, National Museum of Wales, Cardiff CF1 3NP
Dr. K. C. Allen, Department of Botany, Bristol University, Bristol BS8 1 UG
Dr. R. A. Fortey, Department of Palaeontology, British Museum (Natural History),

Cromwell Road, London SW7 5BD

Other Members of Council

Dr. R. J. Aldridge, Nottingham
Dr. M. C. Boulter, London
Dr. M. D. Brasier, Hull
Dr. P. J. Brenchley, Liverpool
Dr. D. E. G. Briggs, London
Dr. C. H. C. Brunton, London
Dr. S. Conway Morris, Cambridge

Dr. M. B. Hart, Plymouth
Dr. P. M. Kier, Washington
Dr. S. C. Matthews, Bristol
Dr. I. E. Penn, London
Dr. M. Romano, Sheffield
Dr. D. J. Siveter, Leicester
Dr. J. Watson, Manchester

Overseas Representatives

Australia: Professor B. D. Webby, Department of Geology, Sydney University, Sydney, N.S.W., 2006
Canada : Dr. B. S. Norford, Institute of Sedimentary and Petroleum Geology, 3303-33rd Street NW., Calgary, Alberta
India: Professor M. R. Sahni, 98 Mahatma Gandhi Marg. Lucknow (U.P.), India
New Zealand: Dr. G. R. Stevens, New Zealand Geological Survey, P.O. Box 30368, Lower Hutt
West Indies and Central America: Mr. John B. Saunders, Geological Laboratory, Texaco Trinidad, Inc., Pointe-a-Pierre,
Trinidad, West Indies

Western U.S. A. : Professor J. Wyatt Durham, Department of Paleontology, University of California, Berkeley 4,
California

Eastern U.S.A. : Professor J. W. Wells, Department of Geology, Cornell University, Ithaca, New York
South America: Dr. O. A. Reig, Departmento de Ecologia, Universidad Simon Bolivar, Caracas 108, Venezuela

Palaeontology

VOLUME 22 ■ PART 2

CONTENTS

A gianl myriapod trail front the Namurian of Arran, Scotland

D. E. G. BRIGGS, W D. I. ROLFE and J. BRANNAN 273

The middle Pleistocene ostracod fauna of the West Runton
freshwater bed, Norfolk

P. DE DECKKER 293

Functional morphology and ontogenetic variation in the Callovian
brachiopod Septirhynchia from Tunisia

M. O. MANCENXDO and C. D. WALLEY 317

Trilobites from the Ordovician Auchensoul and Stinchar
Limestones of the Girvan District, Strathclyde

R. P. TRIPP 339

Taxonomy, functional morphology, and palaeoecology of the
Ordovician cystoid family Hemicosmitidae

J. F. BOCKELIE 363

Taxonomy and opercular function of the Jurassic alga Stichoporella
G. F. ELLIOTT 407

Micropalaeontological studies of the Upper Jurassic and Lower
Cretaceous of Andoya, northern Norway

M. L0FALDLI and B. THUSU 413

Two new early Cretaceous dinocyst species from the northern
North Sea

R. J. davey 427

The horse Cormohipparion theobaldi from the Neogene of Pakistan,
with comments on Siwalik hipparions

B. J. MACFADDEN and A. BAKR 439

Live and dead faunas from coralline algal gravels, Co. Galway

D. W. J. BOSENCE 449

Statoliths of Cenozoic teuthoid cephalopods from North America
M. R. CLARKE and J. E. FITCH 479

Printed in Great Britain at the University Press, Oxford
by Eric Buckley, Printer to the University

STITUTION Z NOlinUlSNI^NVINOSHimS^SB lavyail^LIBRARI ES*2 SMITHSONIAN-* INSTITUTION 2 NOli

Z 1 > Z r- z “

O m ^!\ Q XSstT T(y>S. x^3<3 VrX O

1/7 m 52 m 52 * m 52

pi s' w ^ m

to — co

lavaan libraries Smithsonian institution NoumiiSNi nvinoshiiws ssiavaan l i b r

z 52 _ . z co z 52 _ z

> 2 > 2 ^ w

STITUTION NOlinilJLSNI NVINOSHIIWS^SS i a va a 11 UBRAR JES^SMITHSONIAN INSTITUTION N0I1
to _ ^ > to — to

Z X- .V. Z r' z

* LJ® I 2 ®LJE! c liL Jl) 2

ra 5 ~ 5 Xi^ = -

z j z _J z

uavaan libraries smithsonian institution NoumiiSNi nvinoshiiws saiavaan libf

z r- z

m XQyocjz ^ m X(ov^ v, xopTTTyiX m X 'il xgiiisgx m

CO _ to £ </> X “ CO

STITUTION NOlinilJLSNI NVINOSH1MS S3iavaail LIBRARIES SMITHSONIAN INSTITUTION N0I1
... co z to z to z .... co

> ' 5 3» '<«F~ 2: J> NgliilSS>

uavaan libraries^smithsonian institution NoumiiSNi nvinoshiikis^sb lavaan libi

in ~ in — in

O Xgiy pcJX “ o ^ — XPtasgX O

STITUTION2 NOimillSNI^NVINOSHimS SBiavaail^LIBRARI ES2 SMITHSON IAN-* INSTITUTION 2 N0I1
~ z I- Z f- z

to = CO

lavaan libraries smithsonian institution NoumiiSNi nvinqshiiws saiavaan libf

z , to _ z > to z to z

K. 2

I
to

il O

5.

2 2 " > 2 Xjrasv^/ > ' 2

STITUTION NOlinillSNI^NVINOSHlIWS^SBiavaan libraries^smithsonian INSTITUTION NOli
to 5 ' '** - - —

O - W _ o

Z _J z _> z

uavaan libraries smithsonian institution NoumiiSNi nvinoshiiws S3 lavaan libi

Z r- Z r- _ z *“

JVINOSHimS^Sa I HVaa II^LI B RAR I ES^SMITHSONIANJNSTITUTION^NOlinillSNI^NVINOSHlIlNS.^ I

S ><.7^ — r-'i ^ o ><CIstit/7>v \v

5MITHS0NIAN INSTITUTION NOlinillSNI NVINOSHIIINS" S3 I a Va 8 11 LI B R AR I ES SMITHSONIAN INS'
^ , co z w z - — — z

^

w/ — 2 cO

NVINOSHHWS^Sa I a va a n^LI B RAR I Es" SMITHSONIAN INSTITUTION NOIinillSNI_NVINOSHllWS S3 I

</> z

?/ c

SMITHSONIAN INSTITUTION NQIlfUUSNI NVIN0SH1IINS S3iavaai1 LIBRARIES SMITHSONIAN INS
z r* ... z

CD
>

4#^ m n^voc^ w

NVINOSHltWS S3iavaan~ LIBRARIES SMITHSONIAN “INSTITUTION’ NOlinillSNI NVINOSHIHNS S3
co z .... w z .v- w z ^

if) CO ■V'/

> ■?r 5 s ^ 5 -^2^ >

SMITHSONIAN Z I NSTITUTlON^NOlirflllSNl NVINOSHIIWS^SS I a Va a IT LIBRARIES SMITHSONIAN_INS
“ co - ,v . to 5 co z

CO

fE CO — CO

SMITHSONIAN INSTITUTION^NOlinillSNI NVIN0SH1IWS S3 I a Va 311 LI B RAR I ES SMITHSONIAN INS
z * co z w z >

s §j ^

| | ^^3) g I ^4^5- 2

//i •'/*' 2 C/5 2 C/5 “• 2 </)

nvinoshiiws S3 1 a va an libraries Smithsonian institution NoiiniiiSNi_NvjNOSHiMS S3

5 5 o

z _J Z _J z

smc-’-^nian institution NoiimiiSNi NViNOSHiiiAis S3 1 a va an libraries Smithsonian in
z r- z Z ^ . §

^ ~ v O ^TjcvaT^v — y^vASOATrs. ^

r> 1