• - •- - . ■ ■ • :

'UUH:

>gi r kp.£.i

mm

mm

§®§f$

Mi X

O

> 2 Xg^oc^ > " ^F' 2 v. > xguiivjx 2 ’■v >

z in z co <o K- z

avaan libraries Smithsonian institution NoiinnisNi nvinoshiiws S3i avaan li

co ^ co — to

* 1 1
1 K ^

O XI*. D.c>^ ~ XgUiisgX o y" 2 o

titution- NoiiniusNi"NvmosHims S3 1 a va a n li b rar i es^smithsonian”1 institution^ nc

2 2 v z f- 2 2 z

O X^IlTcoX V ° X^vwDX ~ /oo^\ O _ /Too^oX O

3

H*

CO

m v \\v ^ 'x^jvasv^/ m

_ </> = co ~ co

avaan libraries Smithsonian institution Nouninsm nvinoshiiws saiavaan li

2 ■■, to Z t CO 2 CO z

< \, S XCivVoX <£ '\» S S ^5%- . <

CO

O
2 ,

2 •" > x^vas*^ 2 "' 5 xg»os*£^ > ’

T1TUTI0N NOIlfUliSNI NVIN0SHilWSWS3iavaan2LlBRARIES£°SMITHS0NlAN INSTITUTION N(
co 5 \ co — co =: co

u ^ ^37^ UJ

-

a: fit
< ‘

^ x^snrjx co v 2 w w w xxgx;

wOvT?/*/ j&y Jt* i » A- He

<<

££

oa ~ ' / m .

i avaan libraries Smithsonian institution NonniiiSNi nvinoshiiws S3iavaan__L

2 r*

m v^\pc^ z m m

5TITUTI0N W N0linillSNrNVIN0SHllWSWS3 I H VH 8 I1~L I B R A R I ES^SMITHSONIAN INSTITUTION N

JMIUMWI'C w z co 2 CO Z oV. w

X

to

O

z

>

z

2 \rn.i K?/ ^ £ "WXX

z ^ in " ^ 2 w * Z CO

i avaan libraries Smithsonian institution NOiinmsNi nvinoshiiws S3iavaan_L

CO 5 CO = .v co 2

<T^°*7X CU co u c/> 1x1 CO

5 2 | 2 x;^?gp/ |

STITUTION2 N0linillSNl“!NVIN0SHlllMS2S3 IHVaail LIBRARIES SMITHSONIAN INSTITUTION N
2 v Z r- 2 2 S

O XorrX O ^rcTlrO^ O ^ /VOO VqX CD

m c ys^y ; ' x>yAst'>^ m '<^owy ™ ^

_ to ^ _ to — co

i avaan libraries Smithsonian institution Noiiniusm NvmosHims^saiavaan^L

2 , CO 2 r. CO Z

< v s V S S

z \\ ,<k H /$ATP>&\ 2 , A ^ z

X
CO

vX#5"' I

• ,^p > _ ^

stitution w NouniiiSNi^NViNOSHiiws^sa i ava a n^LI B r ar i es^smithsonian institution n

CO X T CO - CO „-7.

)

3

X

~ _

> 'W^ 2 ^ > xwcustjx s -v >

2 w ‘2 ^ OT Z " to

SMITHSONIAN INSTITUTION NOiinillSNS NVIN0SH1IWS S3IHVaS!l LIBRARIES SMITHSONIAN

3 “ _ to = . CO S £0

O _ x^jqsgx o NV „ x^uas^x q

WlNOSHlIlAIS SHiavyan LIBRARIES SMITHSONIAN^INSTITUTION NOlinillSNI^NVINOSHIIiNS
‘ > z r- _ 2 r- 2 r~ >

#% s m

r/S<€4:™ 1

1 C )W ~ X^VASJA^ m Xftosv^ ^ m X^os*^/ £ XJvasv^X rn

) ' ‘ _ co ± to ± to

SMITHSONIAN INSTITUTION NOlinillSNI NVINOSHIISAJS S3tHVHai1 LIBRARIES SMITHSONIAN
l z \ to z tn . z \ co „

; x^vasvj^' | ^ t | t 1

: w ■'••' 2 CO 2 to •■*' 2

NVIN0SH1ISAIS SHiavaan LIBRARIES SMITHSONIAN INSTITUTION NounuiSNi_ nvinoshiiins

CO — CO — CO

to

■£i

.#=

/ v ' H

o

2

SMITHSONIAN INSTITUTION NOlifllllSNI NVINOSH1IWS SSiaVHail LIBRARIES SMITHSONIAN
_ 2 r- x . z r» 2

* — 2 k. O Q i Cv'V-iV'a /’"T^SOiCaX O

m " xouu.sgx id ^ ^ N^aosto/ m

to f: to \ “ co

NVIN0SH1IINS S3 lava an LIBRARIES SMITHSONIAN INSTITUTION NOlinillSNI nvinoshiiins

2 .... CO 2

> V 2 '• . > Xi2lut£X *§ ^ >

0 2 CO *2 </j k. 2 ' CO

SMITHSONIAN INSTITUTION NOlinillSNI NVIN0SH1IWS SHiaVHail LIBRARIES SMITHSONIAN

O — , CO “ CO 2 co

O X<£!_£L2 =_ x^uas^ q " __ x^uiis^x q

"NVINOSHIIWS S3 lava a 11 I! BRAR I ES^SMITHSONIAN^INSTITUTION^NOlinillSNI^NVINOSHimS
- >2 r- 2 r- 2 -

T\ i= AWWA £

d c>x5?r 2 XjvAS>^2 m X^vos m X^c/vos*^ ^ XJvas'^X rn

ASMITHS0NIAN”lNST!TUTI0NWN0!iniIlSN!™NV!N0SHlUAJSWS3 I aVH 8 11 ~LI B RAR i ES SMITHSONIAN

2 ... . Z \ CO 2 to Z

H 2 -4 2 —i 2

_NVIN0SH1IWS S3 I ava a ll_LI BRAR I ES SMITHSONIAN INSTITUTION NOIinillSNI_NVINOSHlltNS

s.

THE PALAEONTOLOGICAL ASSOCIATION

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

COUNCIL 1990-1991

President'. Professor J. W. Murray, Department of Geology, The University, Southampton S09 5NH
Vice-Presidents : Dr M. Romano, Department of Geology, University of Sheffield, Sheffield S3 7HF
Dr P. R. Crowther, City of Bristol Museum and Art Gallery, Queen’s Road, Bristol BS8 1RL
Treasurer : Dr M. E. Collinson, Department of Biology, King’s College, London W8 7AH
Membership Treasurer : Dr H. A. Armstrong, Department of Geology, University of Newcastle,

Newcastle upon Tyne NE1 7RU

Institutional Membership Treasurer : Dr A. W. Owen, Department of Geology and Applied Geology,

University of Glasgow, Glasgow G12 8QQ

Acting Secretary. Dr J. A. Crame, British Antarctic Survey, High Cross, Madingley Road, Cambridge CB3 OET
Circular Reporter : Dr D. Palmer, Department of Geology, Trinity College, Dublin 2
Marketing Manager '. Dr C. R. Hill, Department of Palaeontology, British Museum (Natural History), London SW7 5BD
Public Relations Officer : Dr M. J. Benton, Department of Geology, University of Bristol, Bristol BS8 1RJ

Editors

Dr M. J. Benton, Department of Geology, University of Bristol, Bristol BS8 1RJ
Dr J. E. Dalingwater, Department of Environmental Biology, University of Manchester, Manchester M13 9PL
Dr D. Edwards, Department of Geology, University of Wales College of Cardiff, Cardiff CF1 3YE
Dr P. D. Lane, Department of Geology, University of Keele, Keele, Staffordshire ST5 5BG (co-opted)

Dr P. A. Selden, Department of Extra-Mural Studies, University of Manchester, Manchester M13 9PL
Dr P. D. Taylor, Department of Palaeontology, British Museum (Natural History), London SW7 5BD

Other Members

Dr E. A. Jarzembowski, Brighton Dr E. A. Kennedy, Oxford Dr D. M. Martill, Milton Keynes
Dr A. R. Milner, London Dr R. A. Spicer, Oxford

Overseas Representatives

Argentina: Dr M. O. Mancenido, Division Paleozoologia invertebrados, Facultad de Ciencias Naturales y Museo, Paseo del
Bosque, 1900 La Plata. Australia: Dr K. J. McNamara, Western Australian Museum, Francis Street, Perth, Western
Australia 6000. Canada: Professor S. H. Williams, Department of Earth Sciences, Memorial University, St John’s,
Newfoundland A1B 3X5. China : Dr Chang Mee-mann, Institute of Vertebrate Palaeontology and Paleoanthropology,
Academia Sinica, P.O. Box 643, Beijing. Dr Rong Jia-yu, Nanjing Institute of Geology and Palaeontology, Chi-Ming-Ssu,
Nanjing. France: Dr J.-L. Henry, Institut de Geologie, Universite de Rennes, Campus de Beaulieu, Avenue du General
Leclerc, 35042 Rennes Cedex. Iberia : Prof. F. Alvarez, Departamento de Geologia, Universidad de Oviedo, C/. Jesus Arias
de Velasco, s/n. 33005 Oviedo, Spain. Japan: Dr I. Hayami, University Museum, University of Tokyo, Hongo 7-3-1, Tokyo.
New Zealand: Dr R. A. Cooper, New Zealand Geological Survey, P.O. Box 30368, Lower Hutt. Scandinavia : Dr R.
Bromley, Fredskovvej 4, 2840 Holte, Denmark. U.S.A. : Prof. A. J. Rowell, Department of Geology, University of Kansas,
Lawrence, Kansas 66044. Prof. N. M. Savage, Department of Geology, University of Oregon, Eugene, Oregon 97403. Prof.
M. A. Wilson, Department of Geology, College of Wooster, Wooster, Ohio 44961. W. Germany : Prof. F. T. Fursich,
Institut fur Palaontologie, Universitat, D8700 Wurzburg, Pliecherwall 1

MEMBERSHIP

Membership is open to individuals and institutions on payment of the appropriate annual subscription. Rates for 1990 are:
Institutional membership . . £60 00 (U.S. $108) Student membership . . . . £1 L50 (U.S. $20)

Ordinary membership . £28 00 (U.S. $50) Retired membership .... £14 00 (U.S. $25)

There is no admission fee. Correspondence concerned with Institutional Membership should be addressed to Dr A. W. Owen,
Department of Geology and Applied Geology, The University, Glasgow G12 8QQ. Student members are persons receiving full-
time instruction at educational institutions recognized by the Council. On first applying for membership, an application form
should be obtained from the Membership Treasurer: Dr H. A. Armstrong, Department of Geology, University of Newcastle,
Newcastle upon Tyne NE1 7RU. Subscriptions cover one calendar year and are due each January; they should be sent to the
Membership Treasurer. All members who join for 1990 will receive Palaeontology , Volume 33, Parts 1-4. Enquiries
concerning back numbers should be directed to the Marketing Manager.

Non-members may subscribe, and also obtain back issues up to 3 years old, at cover price through Basil Blackwell Ltd,
Journal Subscription Department, Marston Book Services, P.O. Box 87, Oxford OX2 0DT, England. For issues before Vol.
29 contact the Marketing Manager.

Cover: Framboidal pyrite with lumen of a tracheid of the early land plant, Gosslingia breconensis, from the Lower Old Red
Sandstone of South Wales. x700. See P. Kenrick and D. Edwards. Botanical Journal of the Linnean Society , 97, 95-123.

THE CLASSIFICATION OF THE FORAMINIFERA
A REVIEW OF HISTORICAL AND PHILOSOPHICAL

PERSPECTIVES

by JOHN R. HAYNES

Abstract. A historical review of the classification of Foraminifera reveals that the latest by Loeblich and
Tappan is in the central tradition, leading from the work of the early English School, in which it is attempted
to relate test composition and fine structure to physiology in formulating a sound basis for distinction of
suprageneric rank. Although the early English School first distinguished homologous from analogous
structures, their view that Foraminifera constituted a primitive, plastic group has inhibited biological and
stratigraphical work, and distorted classification down to the present. The search for a single ‘key’ to turn the
problem has also led to swings of fashion and, using Brady’s metaphor, to ‘cutting rather than untying the
Gordian Knot’. Recent classifications are shown to retain considerable artificial elements. Attempts to
substitute numerical taxonomy or cladistics for ‘intuitive’ taxonomy are misguided. Principles to assist the way
forward are given.

The Foraminifera are the most abundant marine invertebrates and play a major role in the
economy of nature. The classification given in the Treatise of invertebrate paleontology (Loeblich
and Tappan (1964) hereafter referred to as Treatise (1964)) has been treated as a standard for a
quarter of a century, so the appearance of a comprehensive revision in book form (Loeblich and
Tappan 1988u) is a considerable scientific event, not only for palaeontology but also for zoology.
Its importance for geology is well expressed by the authors in their introduction : ‘They are the most
widely used fossil organisms for biostratigraphy, age dating and correlation of sediments, and
paleoenvironmental interpretation, both as organisms whose living representatives provide ecologic
data and as mineralized shells that are a geochemical record of paleotemperatures, extent of
glaciation and other paleogeographic features’ (p. 1 ). However, the very success of Foraminifera in
applied palaeontology and palaeoecology has hindered advances in classification because workers
have resisted changes, preferring the ‘stability’ created, at least in the West, by general acceptance
of the Treatise. Indeed, suggested changes may provoke hostility. Symptomatic of the general
attitude is this complaint in Nature about the alternative classifications put forward in Comparative
primate biology (Martin 1988, p. 589): ‘The Editors did not impose a uniform system... use of
alternative schemes seen as a sign of healthy science better seen as an insidious ailment ...a relatively
stable classification is a prerequisite for other work ... it is high time matters were put to rights . . . the
present situation is comparable to a team of computer programmers arguing endlessly about
whether they should use BASIC, FORTRAN or PASCAL, rather than getting on with the actual
programming job in hand’.

Views of this kind mistake the fundamental purpose of classification which is to provide an
orderly hierarchical array of taxa reflecting the genetic lines of descent, not simply to provide a
convenient pigeon-holing, i.e. a natural classification is confused with an artificial identification key
or retrieval system. A taxonomic scheme is useful for ‘other work’ in direct proportion to the extent
it reflects the true lines of descent. An artificial scheme will probably be misleading in biostratigraphy
and palaeoecology and certainly useless for studies of macro-evolution. It is the advances towards
a genetic classification that are a prerequisite of greater precision in other work, especially in study
of rates of evolution and diversity patterns. A good example is provided by the foraminiferal genera

I Palaeontology, Vol. 33, Part 3, 1990, pp. 503-528.|

© The Palaeontological Association

504

PALAEONTOLOGY, VOLUME 33

GJobigerina and Globorotalia , formerly included in one family and both considered to range back
into the early Cretaceous. In Loeblich and Tappan’s book they are now placed in different
superfamilies and both are restricted to the Cenozoic, Globorotalia to the Neogene. The discovery
that the globigerinids are near-surface plankton while the globorotaliids are deep-dwellers is in tune
with this more discriminating approach. Similarly, the large orbitoidal foraminifera were originally
placed in one family that straddled the Cretaceous-Tertiary boundary. Again, these genera are now
recognized as belonging to distinct superfamilies with ranges which reveal and emphasize the
marked break at that horizon. Changes of this kind are clearly a sign of scientific health and
complaints about them reminiscent of the lamentations that arose because Copernicus destroyed the
stability of the Ptolemaic system.

Another reason for the ‘bad press’ given to taxonomy is the common misconception that the
important work on the major groups has been done, with the main goals achieved. Like the
traditional comparative anatomy once taught at Oxford (as described by Sir Peter Medawar in his
engaging autobiography. Memoirs of a thinking radish), it is seen as ‘deadly dull’ and no longer a
‘live’ issue, despite the ‘grandeur’ of the demonstration of evolution. Controversies in taxonomy
are dismissed as mere squabbles about trivial matters, such as priority of names, and descriptive
work on new species as a higher form of stamp collecting. However, there are very few groups in
which the phylogeny is known with any certainty. In the Foraminifera it is a matter of continuing
lively debate and a comparison between the classification set out by Loeblich and Tappan in their
new book and those of earlier workers, including their own in the Treatise (1964), will silence
criticism of this kind.

The general reader may be given a clearer picture of the problems faced in the attempt to produce
a natural classification by a historical summary of past debates about the correct principles to be
followed. The clash between the ‘English School' and the ‘Continental School' in the nineteenth
century and between the English School and the ‘American School' in the early twentieth century
being particularly instructive, because the extreme positions taken up highlight the difficulties
involved as well as differences in philosophy and attitude. This will also help to explain the paradox
that late into this century the Foraminifera could be regarded as a simple, plastic group in which
the shell composition reflected habitat, and differences in shape and ornament the impact of
physico-chemical forces acting on the protoplasm (D’Arcy Thompson 1942), at the same time as
they were being successfully used in biostratigraphy. No attempt is made to summarize all the
classifications that have been proposed (for an exhaustive coverage see Loeblich and Tappan 1964,
pp. 140-63). Instead, attention is directed to the most influential classifications in what may be
called the central tradition; in particular, to the founder members of the English School because
they were the first successfully to distinguish homologous from analogous structures and because
the classification now produced by Loeblich and Tappan is, in many respects, ‘the direct lineal
descendant’ of that of Carpenter (1861, 1862).

ATTEMPTS TO UNTIE THE GORDIAN KNOT
The clash between catastrophism and unif ormitcirianism

The dominant figures in the English School were largely biological in outlook and, appropriately
for a maritime nation, were chiefly concerned with the study of the Recent and the dredge samples
that became available as a result of the strong nineteenth century drive towards exploration of the
oceans. The first work on foraminifera in England was on the Recent species of the south coast
(Walker and Boys 1874). The dominant figures in the Continental School were largely
palaeontological and geological in outlook, the first work in France being that on the Eocene of the
Paris Basin (Lamarck 1804). There was also a deeper dichotomy in that Alcide d’Orbigny, the
founding father of scientific foraminiferal studies and of the Continental School, was also a
catastrophist (appropriate in post-revolutionary France), whereas, in contrast, the English School
was strongly influenced by Lyellian uniformitarianism (see Heron- Allen 1917).

HAYNES: FO R A M I N I F E R AN CLASSIFICATION

505

D'Orbigny (1826) was the first to recognize the Foraminifera as a separate group, considering
them an order of the Cephalopoda distinguished by possession of a series of internal, basal chamber
openings (foramina), rather than possessing a siphuncle. Concordant with his idea that the
Foraminifera were cephalopods he divided them into families entirely on the basis of chamber
arrangement or coiling mode, the simplest (and most severely parsimonious) solution.

After the discovery that Foraminifera were Protozoa by Dujardin (1835). d’Orbigny ( 1839) raised
the group to Class status with six orders based on chamber arrangement, plus an extra one for single
chambered forms. Despite the simplicity of his taxonomic scheme, d'Orbigny recognized a large
number of genera and species. This was because he believed in the complete replacement of faunas
at the end of each successive geological stage by special creation.

Although the discovery that the Foraminifera were Protozoa prompted d’Orbigny to give them
class status, it had little other effect on his view of their classification. In contrast it had a profound
influence on the whole philosophy of the English School, dominated by Williamson and Carpenter
in the mid nineteenth century. Thus, Williamson (1858, p. ix) saw it as demonstrating the Tow
organization’ of ‘diversified inferior creatures’ with an ‘anomalous and obscure' history. The
external features of the shell, such as direction of growth and sculpture, considered important earlier
when foraminifera were thought to be molluscs, were now seen as due to ‘age and local
circumstance’ and possessing Tittle value’. Carpenter (1862, p. vii) in classing them with the
Rhizopoda commented on the soft parts as ‘a little particle of homogeneous jelly ’, the ‘protoplasm
not yet differentiated into cell-wall and cell-contents’ and more primitive than Amoeba , because,
apparently, without nucleus. For this reason he concluded that variation in form of the shell
aperture, which in higher animals with an alimentary canal would warrant generic or family
distinction, was of ‘comparatively little moment’ and perhaps even a matter of ‘individual
difference’.

Both these key figures in the English School readily accepted the new doctrine of Darwinian
evolution: Carpenter was close to the Darwin circle, wrote a supporting review (1860) and took a
prominent part in debates on the subject as, ‘one of the most active and respected interpreters of
evolutionary theory’ (Young 1985). Combined with their view of the low position of Foraminifera
on the tree of life, this led them to reject outright d’Orbigny’s classification based on geometrical
plan. The ‘extraordinary diversity’ of Foraminifera and The gradational nature of these differences’
which indicated ‘community of descent more or less remote ... from a small number of original
types’, also led them to dismiss not only his successive special creations but also his
‘needlessly ... multiplied ... number of species’.

Williamson ( 1 858, p. x) decided that ‘ real specific ’ features would only be shown by the soft parts,
and that The hard shells of the Foraminifera do not constitute a sufficiently constant and important
element in their organization to justify our trusting to them as guides in the discrimination of
species’. While Carpenter (1862, p. vii) was forced to the conclusion that ‘sharply defined
divisions - whether between species, genera, families or orders do not exist’ and that (p. x) the
‘ordinary notion of species ... was inapplicable’, the term becoming one of convenience for inter-
grading assemblages grouped round a small number of family types, showing ‘continuity through
a vast succession of geological epochs’.

It would be anachronistic to complain that Carpenter and Williamson lacked an adequate species
concept. The attempt to apply evolutionary ideas was premature. Adequate knowledge of the
stratigraphical record, of true genetic diversity as distinct from dimorphism and polymorphism
resulting from alternation of generations, and of the complications caused by changing apertural
style during the growth of multiform species, lay in the future. But they were correct in their scathing
attack on members of the Continental School, such as de Montfort (1808) who recognized
individual variations as generic, e.g. nine genera based on ‘ Nautilus calcar ’ - all synonymized with
Lenticulina in Loeblich and Tappan’s book. But in their extreme reaction to d’Orbigny on the side
of Lyellian uniformitarianism they denied (Carpenter 1862, p. xi) there was ‘any fundamental
modification in the foraminiferous type from the Palaeozoic period to the present’, and deduced
(Williamson 1858, p. xii) that foraminifera have Tittle value... in determining the relations between

506

PALAEONTOLOGY, VOLUME 33

zoological provinces or in identifying stratified deposits'. These views cast a long shadow down the
years, inhibiting stratigraphical work in Britain into the present century, despite their countervailing
(and contradictory) remarks on the value of foraminifera as indicators of temperature and depth.
They undoubtedly influenced Darwin, who otherwise might have found a paradigm in these
‘intergrading forms' regarded as the ‘direct lineal descendants’ of remote ancestors. On this subject
the first edition of the ‘Origin' preserves a resounding silence, and, significantly. Carpenter does not
mention foraminifera in his extended review of it ( 1 860). In later editions (from the 4th, 1 866, p. 402,
see 1929 reprint of 6th edition, p. 293) Darwin notes 'Foraminifera have not, as insisted by Dr.
Carpenter, progressed in organisation since even the Laurentian epoch’. Even a hundred years later
it was concluded ‘ the morphological series ... do not always seem to have very strong claims to being
evolutionary series’ (Challinor 1959, p. 79).

The confusion about the value of the foraminifera in stratigraphy is a symptom of the confusion
about species caused by the advent of Darwinism. When Carpenter (1862, p. x) states that the idea
of the species as ‘marked out from each other by definite characters ... from original prototypes
similarly distinguished’, is quite inapplicable to this group, with the suggestion that this is unlike
the case in the more advanced Cephalopoda, we see the persistence of the Linnaean idea of the
species, i.e. separated by morphological gaps from their nearest relatives in the great chain of being.
Underlined by his view (1860, p. 214) that the chief difficulties for Darwin’s theory were those
structures or habits ‘difficult to imagine to have been acquired gradually by any process of
consecutive modification'. Again echoed by Challinor (1959, p. 62), ‘the apparent rarity of closely
graded evolutionary series may express a real rarity and be due... to evolution usually proceeding
in jumps rather than perfectly smooth gradation’. In this way, paradoxically, the profoundly anti-
Darwinian, Linnaean species, the discrete ‘particle’ of Lyell, became incorporated into the
philosophy of the English School as a justification of extreme lumping in the search for the
‘morphological gap’ between ‘true’ species. The pervasive influence of this idea, even today, is seen
in ill-advised attempts to delineate this ‘true’ species by ‘scientifically rigorous’ methods, involving
numerical taxonomy and cladistics (see p. 521).

Although the attempt to apply evolutionary ideas at the specific and generic level was a failure,
the attempt to find a more natural basis for the higher divisions was much more successful. Both
men (Williamson, a medical doctor and Professor of Natural History and Carpenter, one of the
leading physiologists of his time) were highly skilled microscopists and their background and
training prompted them to look for the possible connection between soft parts and fine structure
and wall composition (texture) of the test, as a means of recognizing major groups. Carpenter (1850,
1856, 1862) made particular use of both natural and artificial casts as well as thin-sections in his
studies of foraminifera from off Australia and the Philippines which included a number of larger
calcareous genera. This led him to the discovery of canals and the distinction that could be drawn
on the basis of the presence or absence of pores between annular discoid genera grouped together
by d'Orbigny. By means of thin-sections and transmitted light, Williamson (1852, 1858)
distinguished the three major wall structure groups of the post-Palaeozoic and the lamellar
character of the hyaline group.

These observations led to a comprehensive reclassification by Carpenter (1861, 1862). Two
suborders were recognized: Imperforata, to include membranous, ‘porcellanous’ (porcelaneous)
and ‘arenaceous’ (agglutinated) families and the Perforata, taken to include three calcareous
families - one finely perforate (Lagemda), one coarsely perforate (Globigerinida) and one finely
perforated with canal system and supplementary skeleton (Nummulida). With its emphasis on fine
structure and composition this classification is decidedly modern in outlook and it might be thought
that simple elaboration of its main ideas would have taken place through the late nineteenth
century. For a number of reasons this was not to be the case. In particular, although in outline the
classification ostensibly presumes a necessary connection between wall structure and perforation,
detailed perusal reveals a major contradiction. The tendency of some ‘rough-cast’ porcelaneous
miliolids to acquire an outer agglutinated coat was considered ‘a trifling variation’, not justifying
their separate treatment. Similarly, perforate agglutinated forms with calcareous cement were

HAYNES: FOR A M I N I FER AN CLASSIFICATION

507

united with their calcareous homeomorphs and (1862, p. 48) separated from ‘proper ... purely
arenaceous’ forms with organic cement in which ‘the absence of any pseudopodial pores ... shows
their affinity to be rather with the porcellanous than with the hyaline series’. In this way. Carpenter
tried to rationalize a system based on the idea of the primacy of perforation as an indication of
fundamental physiological distinction, but at the cost of considerably devaluing wall composition
as well as chamber arrangement. It is interesting that in claiming that the porcelaneous and hyaline
series are isomorphic in chamber arrangement he overlooked the distinctive milioline coiling ot a
major proportion of the porcelaneous group, including the ‘rough cast’ forms which could have
assisted his argument. Quite apart from this contradiction at the heart of the classification, the very
success of the rhetoric of the founder members of the English School regarding the impossibility of
recognizing species on the basis of the hard parts (if at all) helped to undermine it from the start.
Thus, ironically, despite its brilliant insights, it came to be seen as over-elaborate, like the parallel
scheme of Reuss (1862), also based on the discoveries of the English School, but with twenty-eight
families recognized.

The triumph of the English School

These changing ideas were well shown by the new classification proposed by Brady (1884), on the
basis of comprehensive collections from all the major oceans, made during the voyage of the
Challenger (1872-74).

Although during the intervening years there had been an attempt to solve the problem of the
relationships of perforation and wall composition by putting composition at the higher hierarchical
level (Rupert Jones 1875), Brady considered this ‘cutting rather than untying the knot’, and
abandoned ‘minute structure’ as a means of subdividing the order as ‘not uniformly applicable'.
Carpenter’s six families were refined and increased to ten (including twenty-nine subfamilies) with
much more emphasis on chamber arrangement, particularly evident at subfamily level. Also, against
the spirit of Carpenter's ‘Principles', the unilocular family Astrorhizidae was set up to cover single
chambered agglutinated genera from the Abyssal Plains and the Globigerinidae was restricted to
planktonic genera. However, the family Textulariidae was still considered to include both
agglutinated and hyaline, biserial and triserial genera. The Lyellian steady-state, uniformitarian
dogma of the English School is also apparent in the idea that Nummulites (‘ Nummulina') ranged
back to the Carboniferous, in the inclusion of Fusulina in the same family, and in the failure to
separate Endothyra from the agglutinated group.

The continuing confusion about species was also still evident. On the one hand Brady quotes
Carpenter's views with approval and states ‘so-called species represent no more than terms of a
series’, but on the other hand he recognizes (p. vi) ‘modifications ... whether species or not which
differ not merely in details of form and structure but in habit... with characters that afford means
of easy identification’. The attempt to group varieties around a few primary types was abandoned
and the treatment of species and genera is nearer to that of Reuss and the Continental School.
Despite the continuing disagreement about stratigraphical ranges, it is possible to see here the
beginning of a rapprochment, although Brady’s species limits are extremely wide and he found the
‘needless multiplication’ of species by European workers ‘a bar to progress’. In the event,
reconciliation between the two schools lay well into the next century.

Brady’s simple classification, backed by the beautiful lithographed plates in the Challenger
Report (unsurpassed until the invention of the scanning electron microscope and still superior in
some respects), was widely influential and remained in use, modified by Lister (1903), almost into
our own day (Jepps 1956). This reflects the long dominance of the English School and the apparent
confirmation of its views : first, by the detection of pores in the proloculus of the porcelaneous genus
Peneroplis (Rhumbler 1895); second, by the discovery of calcareous genera, in supposed Cambrian
limestone from Malvern, some specimens actually being put into three extant species by Chapman
(1900). The issue of the ‘Cambrian’ calcareous foraminifera became the ‘cause celebre’ of
foraminiferal studies, not solved until Wood (1947) demonstrated that the fauna was derived from
an erratic of early Mesozoic age. Meanwhile workers like Galloway (1933) and Chapman and Parr

508

PALAEONTOLOGY. VOLUME 33

(1936) who tried to take the occurrence into account in classification, were almost bound to decide
that calcareous forms sprang from tectinous ancestors and that agglutinating forms arose later and
were more advanced. The germ of this idea may be found in Carpenter (1862). Those that did not,
like Cushman (1927), who on the basis of his stratigraphical work had good reason to find it suspect,
were forced to try and explain it away as a result of post-mortem calcification. Inevitably, wall
structure was further devalued to the point that Douville (1906) decided it was determined entirely
by habitat.

Economic imperatives and the rise of the American School

It is interesting historically, that just as the philosophy of the English School apparently reached its
apotheosis, developments elsewhere in the world, in particular, the successful application of larger
foraminifera in the Far East and the Caribbean, and of smaller foraminifera in the oilfields of Texas,
vindicated their use as biostratigraphical index fossils.

The name chiefly associated with the taxonomic work on the great number of new genera and
species that now came to light was that of Joseph Cushman, the key figure in what came to be
known as the American School (Galloway 1928). He began his career working mainly on Recent
Foraminifera, indeed his interest in them continued until his death (Cushman 1932, 1933, 1942) and
his knowledge was as extensive in this area as that of Brady. But he also became well known as a
consultant for the U.S. Geological Survey and various oil companies. This allowed him to set up
his own laboratory in 1924 and to begin the serial publication of the famous Contributions from the
Cushman Laboratory for Foraminifera/ research which attest to his unrivalled knowledge of
foraminifera from both the subsurface and outcrop. With this background he was well suited to
syncretize the philosophies of the English and Continental Schools and to mitigate their tendency
towards excessive lumping on the one hand and excessive splitting on the other. This was not
achieved immediately because the sheer volume of new material necessitated the creation of many
new taxa and this was fiercely criticized by members of the English School, still labouring under the
misapprehension that the fauna from the Malverns was Cambrian and clinging to Carpenter’s
central dogma: that the Foraminifera were a simple plastic group.

The initial reaction of the English School is well shown by Heron-Alien’s review (1929) of
Cushman's new classification (1927, 1928) which recognized some 600 genera (compared to the 141
of Brady), ‘an overwhelming avalanche’ causing the ‘brain of the beginner to reel’. In particular,
he deplored the splitting of many genera, with the revival of many of d’Orbigny’s names, as well
as previously abandoned names. Although admitting that the time had come for an ‘exhaustive’
reclassification, he was ‘assailed by doubt’ in his response to the argument for splitting based on
stratigraphic utility and as well as pointing out the problem of facies, suggested that unconformities
and faults could ‘seriously embarrass’ the petroleum geologist and that it was ‘self-evident that the
pelagic Foraminifera must be useless as stratigraphical indications’. That the battle was now
between applied and academic workers rather than schools is revealed by later correspondence in
Nature (Heron-Alien 1935) where he quotes Hans Thalmann’s description of the creation of new
genera and species in 1933 and 1934 (more than 300 per annum) as ‘ dementia nomenclatorica
americana ’. It is interesting that, despite quoting Thalmann’s remarks with apparent approval, he
admits the responsibility of Earland (his co-worker on the Discovery material) for seventy-three of
the new species for 1934. This contradiction is reminiscent of Brady who similarly found it necessary
to legitimize his own taxonomic offspring by appealing for general adherence to the standards of
rectitude set by the early Fathers of the English School.

Cushman not only increased the number of genera to some 600 but replaced earlier groupings by
forty-five families (many based on the subfamilies of Brady) which were increased to fifty in the last
edition of his book, Foraminifera (1948). These families were clearly distinguished on grounds of
wall structure as far as it was then known, chamber arrangement and aperture form, with coiled
genera regarded as leading to uncoiled genera. Unilocular agglutinating forms were considered the
most primitive group and ancestral to most other families via coiled, unilocular genera already
present in the Lower Palaeozoic. In his general acceptance of ontogeny and recapitulation as a key

HAYNES: FO R A M I N I FE R AN CLASSIFICATION

509

to unlock the phylogeny of multiform species, and of uncoiling as a dominant trend, he followed
Lister (1903) and Schubert (1908, 1921) rather than Rhumbler (1895, 1911, 1923) and Douville
(1906) who believed in proterogenesis and general coiling-up. There is an interesting correspondence
here with the rejection by Spath (1933) of the idea of a general, coiling-up trend in the cephalopods
as postulated by Hyatt (1900).

Like Brady, Cushman did not group his families in phylogenetic lines (probably because of the
difficulties caused by the confused stratigraphical evidence) but fifteen of the calcareous perforate
families were regarded as having common descent from a rotaliid ancestor. As was shown by Wood
(1949), apart from Peneroplis and its allies which Cushman separates from other porcelaneous
forms on the groups of their postulated perforate ancestry, his families can be grouped into distinct
higher categories based on wall structure. That this was not done reflects the continuing doubts
about the importance of wall structure and the persistence of the idea of the primary importance
of perforation. Cushman also followed Brady in including the Palaeozoic endothyrids with the
agglutinating lituolids.

The families were grouped into higher categories by Glaessner (1945) who recognized seven
superfamilies based on wall structure, chamber arrangement and apertural form. The endothyrids
were placed in a separate superfamily with the fusulines on the grounds of their microgranular wall
structure (Henbest 1937) and the porcelaneous families were also regarded as a natural group
originating from a coiled unilocular ancestor. Glaessner laid particular stress upon the value of
biological studies in interpreting the fossil record and pointed out how the work of Myers (1935,
1936), Le Calvez (1938) and Jepps (1942) not only clarified problems of dimorphism and
reproduction, but proved the stability of species and of wall structure, perforation and chamber
form under different ecological conditions. However, although the superfamilies were considered to
have a common origin in coiled, unilocular, agglutinated ancestors, the continuing existence of the
‘stumbling block’ to confident phylogenetic work, created by the supposed Cambrian fauna of the
Malverns, meant that no attempt was made to recognize suborders on the basis of wall structure.

The return to wall structure as the primary basis of subdivision

Wood (1947, 1949) not only removed the ‘stumbling block’ in the way of recognition of the
stratigraphical succession of the major groups of Foraminifera but greatly clarified knowledge of
wall structure. The optical characteristics of the main types were clearly defined and the hyaline
group was shown to include both radial and oblique (‘optically granular’) genera. It also
encouraged other work on fine structure which with the help of TEM now proceeded apace, leading
to the suggestion there were three types of lamellar wall, monolamellar, monolamellar with septal
flap and bilamellar (Smout 1954; Reiss 1958, 1965). This microscopical work on fine structure is
reminiscent of that of the early English School and in the same way came to underpin a new
classification, that of Loeblich and Tappan ( 1 964), the most important and widely followed of those
proposed in the wake of the great expansion of foraminiferal studies during and immediately after
the Second World War. Compared with Cushman’s scheme the number of genera and families was
practically doubled, to 1194 and 95, grouped in seventeen superfamilies and five suborders. The
outcry that greeted the appearance of Cushman’s classification was not repeated in the case of that
of Loeblich and Tappan. Although Berggren in his review (1965), remarked that few could have
been prepared for the appearance of such a ‘two volume tome’ he went on to say that comparison
of the numbers of taxa recognized with those in other invertebrate groups, revealed it was actually
conservative. It is also significant that the main criticism made by Lee (1966) in his review of the
biological section, is of the assertion that ‘the simpler genera’ with a single-chambered test of
pseudochitin or pseudochitin combined with agglutinated particles also have simpler histories,
remarking that the life cycle in Allogromia is complicated and indicates the opposite. This general
change in attitudes was largely due to the influence of Cushman’s book, which went through four
editions, to the impact of biological work, and to the successful stratigraphical application of larger,
smaller and planktonic foraminifera in the subsurface (in which British micropalaeontologists fully
participated, particularly in the Middle East and Caribbean). Although problems of definition

510

PALAEONTOLOGY, VOLUME 33

remained, the ‘new synthesis’ between population genetics and Darwinism in the 1930s and 1940s
established the modern species concept and thereafter the debate was between individuals rather
than schools. The idea that foraminifera were a simple, plastic group in which true species did not
exist largely disappeared.

Wall composition and fine structure were now regarded as most important and the basis for
recognition of the suborders and superfamilies, with apertural characters and chamber arrangement
second and third. Thus we find that, after just over a hundred years, wall composition and fine
structure have returned to favour, on the lines of Rupert Jones’s attempted revision of Carpenter’s
scheme with perforation made subordinate :

Allogromiina - membraneous
Textulariina - agglutinated
Fusulinina - microgranular

Miliolina - porcelaneous (imperforate in post-embryonic stage)

Rotaliina - calcareous perforate

Apart from the microgranular group, the major subdivisions are based on the wall structure
categories of Williamson, and reliance on wall composition and layering is carried to great lengths
in the recognition of the superfamilies. In particular, aragonitic, calcitic spicular, calcitic single
crystal and , hyaline-oblique (‘optically granular’) structures were given superfamily status.
However, the attempt to use wall structure as a key led to inherent difficulties like those in
Carpenter’s scheme. Although the Rotaliina were defined as calcareous perforate and the Miliolina
as imperforate in the post-embryonic stage, separate status was not given to the perforate families
in the calcareous microgranular group or to those in the agglutinated group (also discounted by
Carpenter). Again, although aperture form is stated to be more important than chamber
arrangement, in practice, the lower divisions were recognized on a combination of characters, and
where wall structure and apertural changes with ontogeny were ill-understood, chamber
arrangement inevitably took precedence.

History also repeated itself in that opinion swiftly swung away from the extremes of this mid-
twentieth century cutting of the Gordian Knot. Towe and Cifelli (1967) showed that although radial
and hyaline-oblique structure are distinct optically they are close crystallographically, and one may
easily give rise to the other. Indeed, intermediate forms have been shown to exist (Bellemo 1974).
This confirmed the variation noted at the specific and generic level in many families by Haynes
(1956) and Wood and Haynes (1957). It thus became clear that the Cassidulinacea was an artificial
grouping and that wall structure shows progressive changes in many groups, e.g. in the Elphidiidae
and Cibicididae (Anomalinidae) where chamber arrangement remains stable, in contrast to the
Nodosariidae in which wall structure is stable while chamber arrangement shows rapid progressive
change along many lines. Similarly, the attempt to subdivide the hyaline group on the basis of
layering ran into problems of conflicting observations until it was decided that all the Rotaliina,
apart from the monolamellar Nodosariacea were bilamellar (Hansen and Reiss 1972; Gronlund and
Hansen 1976). Some of these problems were taken into account by Loeblich and Tappan in various
modifications of their scheme (1974, 1984). In particular, in 1974, they abandoned the attempt to
make hyaline-oblique and bilamellar structure the basis of separate superfamilies, although wall
composition was retained as the basis of the overall grouping.

The Treatise classification became the dominant influence on foraminiferal studies for more than
two decades. Of other classifications proposed, the most interesting was that of Hofker (1951) in
which the hyaline foraminifera, considered to be derived from agglutinated forms with a toothplate,
are divided into three suborders:

Protoforaminata - single aperture, finely perforate (protopores)

Biforaminata - two apertures, finely perforate

Deuteroforaminata - secondary aperture dominant, coarsely perforate (deuteropores)

Unfortunately, although this arrangement emphasizes the aperture and internal characters and

HAYNES: FOR A M I N I FE R AN CLASSIFICATION

511

brings perforation back into consideration as a major factor, it largely ignores wall structure and
stratigraphical relationships and although modified by Reiss (1958) has not been followed.

The classification of Haynes (1981) retained wall structure as the major consideration at the
higher level but with recognition of nine orders and a return to class distinction. The hyaline group
was considered polyphyletic and broken into five orders, e.g. the Buliminida were considered to
have been derived from high trochospiral agglutinated forms (following Hofker in echoing
Carpenter) and the ‘vitreous’ Palaeozoic genera were separated as a suborder within the
microgranular group. The superfamilies were recognized on a combination of coiling mode and
apertural form and emphasis was placed on umbilical characters. A major weakness is that
perforation is ignored and the presence of pores in the megalosphere of Peneroplis is dismissed
(along with the much more doubtful ‘pores’ in the walls of adult peneroplids and miliolids). This
leads to the contradiction that although the hyaline group is regarded as polyphyletic, porcelaneous
genera are considered to constitute a natural group. Further contradictions are that milioline
agglutinating genera are left within the Miliolida (as by Brady) and that the single crystal
Spirillinacea are considered the possible rootstock for the microcrystalline Rotaliida, although the
genome is apparently more complex in genera such as Patellina than in simple discorbaceans.

Summary

This review of the most influential of past classifications reveals: the baneful influence of the idea
that the Foraminifera are a very simple, plastic group; the strong influence of ideas about species
on classification and biostratigraphical use; the uncertainty about the value of wall composition and
fine structure until biological studies showed modern species to be stable and the stratigraphical
relationships of the major groups were clarified. It is also clear that use of a single factor at the
higher hierarchical levels, cuts rather than unties the Gordian Knot, as Brady remarked, i.e. a
solution likely to be as short-lived as Alexander’s empire. In this respect, it is interesting to see how
well the new classification of Loeblich and Tappan avoids the pitfalls of the past.

THE NEW CLASSIFICATION OF LOEBLICH AND TAPPAN ( 1988)

The period since the publication of the classification in the Treatise of 1964 has been a time of
unprecedented growth in foraminiferal studies, especially as applied to subsurface stratigraphy by
the oil companies and to the core material made available since the initiation, in 1968, of the Deep
Sea Drilling Project which has triggered an explosive outburst of activity on planktonic groups. The
providential, commercial appearance of the Scanning Electron Microscope, at almost the same
time, has also allowed more accurate discrimination of taxa on the basis of internal, apertural and
external features, leading to the recognition of large numbers of new genera. In addition, many of
those synonymized in 1964 are now regarded as distinct. This has brought the total number of
genera considered acceptable by Loeblich and Tappan in 1964 (1,192) to more than double that
figure in 1988 (2,446). These have not just been added to existing suprageneric taxa, as might be
expected in a ‘relatively stable’ classification, because these categories have also more than doubled :
suborders (5 to 12); superfamilies (17 to 65); families (95 to 296); subfamilies (129 to 300). This is
not only because of the discovery of new supra-generic taxa but also because of the higher weighting
now given to certain criteria as compared to their approach in 1964. This explains why the number
of families has trebled and the number of superfamilies almost quadrupled, although the number
of genera has little more than doubled. A possible cause for concern is that fifty-six of the
subfamilies, forty-four of the families, three of the superfamilies (two in open nomenclature) and
two of the suborders are represented by one genus only (105 taxa in all). A simple explanation is
that many genera cannot be placed close to existing taxa because of the incomplete (as well as
incompletely known) stratigraphic record. Alternatively, it may be argued that the group is being
oversplit. Probably both tendencies are at work. The large number of families recognized is clearly
important for macro-evolution, presuming that ‘the taxonomic family is the conventional unit for
studies of large scale extinction and rates of change of diversity’ (Newell 1982, p. 260). It may also

512

PALAEONTOLOGY, VOLUME 33

be considered encouraging that the average number of genera per family is 8-26 which falls between
the 5 to 10 considered the general average by Van Valen (1973) presuming these numbers are not
artifacts and are telling us something about niche partitioning. This may also suggest that the
families and subfamilies with very high numbers are probable candidates for further splitting, if we
consider that ideally a subfamily should represent a single evolutionary lineage, e.g. thirty-seven
genera are included in the subfamily Miliolinellinae. However, there will clearly be differences
according to whether evolution in the family is dominantly allopatric and branching, or sympatric
and phyletic.

Criteria employed at generic level

As the title suggests, the work was conceived as a handbook of genera. There is a prodigious number of
excellent plates and most genera considered synonymous are also illustrated. The success of the authors in
carrying out this massively daunting task, is attested by the comments of reviewers: thus Banner (1988)
considered it ‘virtually miraculous’ that not only had they managed to complete the task but had ‘done it very
well'; and Haman (1988), although he regretted that ‘rigorous criteria were not used’, considered the work a
‘scholarly endeavour’, and a ‘useful, well-illustrated catalogue’. The reaction to the decision not to recognize
subgenera was more mixed. Although approved by Haman, Banner considered recognition would have helped
to diminish the rigidity of the generic diagnoses and Jenkins (1989) that they are essential to show iterative,
evolutionary lines. Adams (pers. comm.) has also pointed out that subgenera have been raised to full genera
without discussion, as in the case of the Lepidocyclinidae (see Adams 1987a).

Although it is not made explicit, the genera are recognized on the basis of test architecture and coiling mode,
aperture form and accessory structures and internal structures. The authors also carry through the
presence/absence of keels, umbilical bosses and supplementary apertures as generic criteria much more
consistently than in the Treatise. An important innovation is the use of ornament at generic level, following
its successful use recently in the planktonic forms. The well known Lingulina tenera plexus of the Lias is thus
transferred to Paralingulina. Although this may cause raised eyebrows, the re-instatement of many genera
previously put in synonomy is to be generally welcomed. A good example is that of Ornatanomalina from the
Palaeocene of Pakistan, formerly put in synonomy with Thalmannita , described from the Palaeocene of Cuba.
This led to widespread records of Thalmannita madrugaensis in Africa and Europe which on inspection are
found to refer to species of Ornatanomalina (Haynes and Nwabufo-Ene 1988). This perfectly illustrates the
palaeogeographical importance of refinements in taxonomy. A common fault in palaeontology (which also
tends to bring taxonomy into disrepute) is the confusion of a diagnosis with a full palaeontological description.
As pointed out by Banner (1988). this is not avoided here. In fairness to the authors, they have obviously tried
to include all the features that may be relevant, including in some cases long accounts of reproduction. But this
means that the reader has to search, often vainly, for the distinguishing features of the genus amongst specific
details and repeated suprageneric characters sometimes, but all too rarely, finding what is required only under
the Remarks. A minor point, but a possible source of confusion, is the continued use of spiral and umbilical
as descriptive terms for the dorsal and ventral sides of the test. As both sides in a trochoid test are spiral (and
may be umbilicate) the older terms are much to be preferred.

Criteria employed at the suprageneric level

Predictably, the reaction to the changes at suprageneric level have been more adverse. Thus Haman (1988) who
provides a synopsis of the main taxonomic changes, considered the ‘rationale not easily determined’, and some
of the family distinctions only generic. He clearly felt the new scheme inferior to the ‘incisive and decisive’
classification of the Treatise and that ‘desired stability had not been achieved’. Banner (1988) was also
‘ unhappy about some of the concepts . . . and the suprageneric importance . . . given to some morphocharacters ’.
How far can these criticisms be sustained?

Loeblich and Tappan have had the problem of revising their suprageneric arrangement at the same time as
incorporating a large number of new genera. They have made their task somewhat harder by continuing to
cramp the Foraminifera within one Order, following Levine et al. (1980), a procedure that looks increasingly
outdated in the light of the statement by Sleigh (1989) that the amoeboid protists comprise a number of
evolutionary lines and are ‘almost certainly polyphyletic’. Margulis (1974) has even suggested the Foraminifera
should be considered a phylum. Upgrading them at least to class or subclass status would seem to be warranted
by their position as the largest and most varied group of invertebrates and unique style of reproduction, with,
like plants, a diploid asexual generation.

HAYNES: FOR A M I N I FER A N CLASSIFICATION

513

The five suborders of the Treatise are now increased to twelve:

Allogromiina

Textulariina

Fusulinina

Involutinina

Spirillinina

Carterinina

Miliolina

Silicolocuhnina

Lagenina

Robertinina

Globigerinina

Rotaliina

organic wall, unilocular (may tend to multilocular)
agglutinated

homogeneously microgranular
aragomtic, non-septate

calcitic, single crystal, non-septate (may have few chambers)
spicular, secreted calcite crystals

porcelaneous imperforate, true pores in protoconch of some

opaline silica, imperforate

calcitic, hyaline-radial, monolamellar

aragonitic, hyaline-radial, septate, internal partitions, finely perforate

calcitic, hyaline-radial, perforate, bilamellar, planktonic.

calcitic, hyaline-radial or oblique, perforate, bilamellar, benthonic.

This breakdown reveals the authors continued reliance on wall structure at this level, though there are
inconsistencies. Two, the Globigerinina (planktonic habit) and the Lagenina (fine structure of lamellae), are
recognized on other criteria which leaves the Rotaliina inadequately diagnosed. On this basis the buliminids
also merit equal recognition and separation from the rotaliids. The Fusulinina are defined as microgranular
but include the Archaediscacea and Colaniellacea, described as having radial layers, and included in the
Archaediscina by Haynes (1981). Other difficulties include separation of the non-septate, aragonitic group, as
the Involutinina, from the septate Robertinina, whereas in the Allogromiina, Spirillinina and the Textulariina,
non-septate and septate forms are included together. Agglutinated genera with milioline coiling (Rzehakinacea)
are included in the Textulariina but their relationships with miliolids in the Recent (Haynes 1973) and first
appearance in the Lower Cretaceous associated with acidic and low oxygen environments suggests a separate
origin from within the Miliolida, dentate forms via the Miliolacea, edentate forms via the Ophthalmidiacea.
The single species, Miliammelus legis , included in the supposedly siliceous Silicolocuhnina may represent a
similar response, in this case to life below the Calcite Compensation Depth (CCD) by agglutination of siliceous
rods. The Carterinina is also based on a single genus in which the test is composed of calcareous spicules.
However, these were regarded as agglutinated by Bronniman and Whittaker (1983). Perhaps Carterina should
have been placed under ‘Genera of Uncertain Status’ pending the necessary tank studies.

The features now considered important at superfamily level are the unilocular, bilocular or multilocular
character of the test, presence or absence of perforation, canals and major apertural features. Bilocular is not
defined in the glossary and generally includes non-septate forms in which the initial end is expanded to form
a cavity, considered homologous with a proloculus in septate forms. Non-septate multilocular forms, where
the chambers are connected by stolons, are not separately considered. Family features now include free or
attached habit, presence of internal subdivisions and surface texture as well as coiling mode. Because they
regard wall structure as more fundamental than number and arrangement of chambers, both of which may
change during ontogeny and therefore being more indicative of evolutionary relationships, they (paradoxically)
tend to discount ontogeny as a key to phylogeny at the highest levels so that coiled, uncoiled, uniserial and
unilocular genera tend to be put in separate groups. For instance, fully uniserial agglutinating genera are
separated from those which become uniserial in the adult and regardless of apertural style are included in a
single superfamily, the Hormosinacea.

They may have been unduly influenced here by the current, fashionable revival of the idea of a general coiling
up trend involving proterogenesis, as against uncoiling involving recapitulation ; exemplified by Brasier’s
attempt (1982) to show that the overriding imperative in the evolution of test architecture is towards minimum
lines of communication between proloculus and aperture (‘MINLOC’). This idea is somewhat weakened by
the observation that the soft parts are generally confined, or retract into, the penultimate chamber. Their
approach makes particular coiling modes ‘stable’ by default and produces a horizontal arrangement. On the
whole, forams show new features in the adult and the stratigraphic record provides evidence of repeated
uncoiling trends with connected serial changes in aperture form and position. The ideal shape for rapid
movement through sediment by retraction of the podostyle directed forwards from a terminal aperture is
apparently reached in many unrelated lines of smaller deep-sea forams. Recognition of these lines would have
simplified the classification at family and superfamily level, especially if the difficulties encountered by the
authors at the subordinal level were eased by the recognition of orders (as by Haynes 1981). It would also have
helped to remove some of the inconsistency inevitable in horizontal groupings of a class that exhibits rapid
evolution with serial development of all features of the test at some time, usually in combination. The only

514

PALAEONTOLOGY, VOLUME 33

‘key’ to these developments is the stratigraphic and evolutionary record and it cannot be predetermined. To
be fair to the authors much of the inconsistency arises because of the different levels reached by research into
different families. The Nummulitidae, probably the most famous fossil forarns, are a good case. In the Treatise
(1964), Cole lumped most of the genera previously recognized under Nummulites because they grade into each
other. This extreme lumping, neo-linnaean in its search for morphological gaps, pleased no-one, except
perhaps, punctuationists convinced that evolution proceeds by genuine saltations (‘hopeful monsters’). Most
of the genera remained in use and are reinstated in the present work. However, genera with subdivided
chambers, even the sub-annular complex form Spiroclypeus , are included in the same family, rather than being
separated as Cycloclypeidae (Haynes 1981) though this would be consistent with their own criteria. Already
since publication, the wave of research has undermined this systematic sand castle. Adams (1988) has shown
that the type species of Nummulites has secondary chamberlets and Haynes (1988) has redefined Ranikothalia
with revival of Nummulitoides. This illustrates how the numbers of taxa recognized (diversity) varies not only
with the data available but also according to the current philosophical paradigm and underlines the futility of
pious prayers for ‘stability’. As Knoll and Butterfield (1989, p. 602) wrote recently, ‘Taxonomic changes both
reflect and influence the way we think about evolutionary pattern and process’.

A review of the treatment of major suborders will indicate some of the difficulties to be overcome, especially
necessary considering the use already made of the taxa recognized (in the evolutionary and diversity study
published by Loeblich and Tappan 19886).

Problems in the treatment of the suborders by Loeblich and Tappan (1988a)

Allogromiina. Although the naked Allogromiina are defined as membranous or proteinaceous, the included
families Lagynidae and Allogromiidae are defined as being both, the Allogromiidae as proteinaceous on a
plasma membrane. Within the Allogromiidae the only difference between the subfamilies Shepheardellinae and
Argillotubinae appears to be that the Argillotubinae are wrinkled and may be fixed.

Textulariina. Within the non-septate, agglutinating superfamily Astrorhizacea the only difference between the
family Astrorhizidae and the Rhabdamminidae, as defined, appears to be the greater selectivity shown by the
latter, because Rhabdammina also has branches leading off a ‘globular central area’. The Rhabdamminidae
also includes both free and fixed forms, in contradiction to the stated criteria. In the remarks on the sub-family
Halyphyseminae the comparison is presumably being made with the Dendrophyrinae, not Rhabdammininae.
The family Psammosphaeridae is defined as coarsely agglutinated but the subfamily Psammosphaermae as
finely to coarsely agglutinated. Technitella is figured upside down and on the basis of the spicular wall and the
broken attachment end (Haman 1967) mistaken formerly as the aperture, should be transferred from the family
Saccamminidae to the Rhabdamminidae, subfamily Halyphyseminae. No clear distinction is made between the
family Hemisphaeramminidae and the Saccamminidae and Psammosphaeridae. The Diffusilinidae is defined
as being distinguished by an irregular mass of branching tubes but this does not agree with the diagnosis of
the two genera included. There is no mention of tubes in Diffusilina.

The superfamily Hippocrepinacea is defined as possessing a proloculus and tubular or flaring second
chamber. However, the subfamily Hippocrepininae does not show a prolocular initial end and the
Notodendrodidae (one genus) actually has a bulbous central region leading to tubular arms. Within the
planispiral to uncoiled superfamily Lituolacea, the family Lituolidae is defined by possession of simple interiors
and terminal apertures which is taken to differentiate them from the Discamminidae which is diagnosed as
having internal partitions, i.e. secondary septa. However, the figures given of Discammina and Ammoscalaria
seem to show primary septa (organic without agglutination) only, with no subdivision of the chambers. There
seems little reason to separate them from Eratidus and other genera included in the Lituolidae, subfamily
Ammomarginulininae. Also, it is not clear why a simple wall is considered important in the definition of most
of the families and subfamilies when it is not mentioned in the superfamily definition and no cases of
complicated walls are included. No reason is given for including the Adhaerentiinae which lack an initial coil
and are biserial, in the superfamily.

Although the superfamily Haplophragmiacea is defined as streptospiral to uncoiled, it includes the
streptospiral to planispiral Recurvoidinae (as a subfamily of the streptospiral Ammosphaeroidinidae). It also
includes the calcareous trochospiral and planispiral families Nezzazatidae and Barkerinidae. (Note that
although the Nezzazatidae are defined as trochospiral to planispiral, the subfamily Nezzazatinae is defined as
trochospiral or planispiral.) These families, with their digitate internal partitions and areal apertures, may
represent diagenetically altered robertinids. They are certainly not close to Ammosphaeroidina.

The family relationships within the superfamily Cyclolinacea are not made clear and the diagnosis given for

HAYNES: FO R A M I N I F E R AN CLASSIFICATION

515

the family Orbitopsellidae does not adequately distinguish it from the Cyclolinidae. The terminology used here,
as in the case of the Loftusiacea (inherited from previous workers it must be said) is not helpful and the fog-
index is high. The family Cyclamminidae is defined as involute and rarely uncoiling, yet it includes the
subfamilies Pseudochoffatellinae (evolute) and the Choffatellinae (only one out of eight genera with no
tendency to uncoil). Neither the subfamilies Alveophragmiinae nor the Hemicyclamminae seem to be
adequately diagnosed on the criteria given.

The superfamily Spiroplectamminacea is defined as planispiral to biserial and uniserial development is not
mentioned (nor is streptospiral coiling) although the family Spiroplectamminidae is defined as planispiral or
streptospiral to biserial, rarely becoming uniserial. In fact, two of the eight members of the Spiroplectamminae
become uniserial and both the genera placed in the Vulvulininae. Three of the seven genera placed in the family
Textulariopsidae become uniserial and one of the three placed in the Pseudobolivinidae. It is not possible to
distinguish between these two families on the criteria given.

The subfamily and family distinctions are not clear in the diagnoses given for the superfamily Pavonitinacea.
This group may be of polyphyletic origin.

Within the superfamily Trochamminacea the sub-family Trochammininae are diagnosed as ‘trochospiral or
may tend to uncoil’. None of the ten genera included and illustrated shows uncoiling.

All the canaliculate, high trochospiral agglutinating genera are included, with the biserial Textulariidae, in
the superfamily Textulariacea and the non-canaliculate genera are accommodated in three other superfamilies.
Because this emphasis on perforation tends to cut across groupings made on other features the superfamilies
show a mixture of coiling modes and apertural styles. The Verneuilinacea and Ataxophragmiacea are not
actually distinguishable as defined and it must be pointed out that the family Prolixoplectidae are included in
the Verneuilinacea on the assumption they are non-canaliculate but the fine structure is only known in three
of the nine genera. A number of families in the Ataxophragmiacea are defined as possessing a two-layered wall
but the terminology is confusing. In the family Textulariellidae ‘beams and rafters’ are said to produce an
‘alveolar wall’. In the Dicyclinidae there are ‘transverse or radial partitions’. The Dictyopsellidae have a ‘sub-
epidermal network’, and partitions are not mentioned, but in the Pfenderimdae, ‘vertical or horizontal
partitions’ result in a ‘reticulate subepidermal layer’. However the wall structure is defined, these families do
not appear to be close to the Ataxophragmiidae and the family Cuneolinidae (which includes Sabaudia ,
characterized by a double-walled embyron with microgranular inner and radial outer layer) probably does not
belong here either. The Coskinolmidae appear more closely related to the superfamily Orbitolinacea. They
have been excluded because they lack marginal exoskeleton (marginal subepidermal partitions). They are also
stated to lack the embyron of protoconch and deuteroconch diagnostic for the subfamily Dictyocomnae but
that feature has only been found in eight of the twenty six genera included.

Within the superfamily Textulariacea the subfamily Minouxiinae in the family Eggerellidae is defined as
cribrate which conflicts with the definition of the Dorothiinae as also developing multiple apertures and which
includes Arenodosaria in which the aperture is described as sometimes becoming cribrate. The definition of the
family Pseudogaudryinidae is incomplete, the aperture being described as an interiomarginal arch with the
trend to multiple apertures not included.

Fusulinina. Within the Fusulinina, defined as ‘homogeneously microgranular’, the unilocular Para-
thuramminacea is divided into families mainly on the number of layers and character of the wall. However,
the Parathuramminidae appears to be distinguished from the Chrysothuramminidae by wall thickness alone.

The superfamily Moravamminacea which includes genera with proloculus and tubular second chamber is
distinguished from the Earlandiacea simply on the basis of the appearance of incipient septa and otherwise
shows a very similar range of morphology. The distinctions as defined between the families in the two groups
are not clear. The Earlandiidae are not distinguished as straight but defined as free, although Warnantella , put
in the Pseudoammodiscidae, may be free. Both the Pseudoammodiscidae and Pseudolituotubidae include
streptospiral forms. The Caligellidae, defined as curved to straight, includes genera showing initial coiling and
therefore cannot be distinguished from the Moravamminidae.

As defined, the Tourneyellacea cannot be distinguished from the Moravamminacea and the definition
actually should exclude the subfamily Palaeospiroplectamminae which would be better placed within the
family Palaeotextulariidae.

The definition of the superfamily Archaediscacea does not make a clear distinction from these two
superfamilies either, because the appearance of a ‘radially built’ outer wall layer in this group is only
introduced in the family diagnoses. Inclusion of this superfamily in the Fusulinina clearly offends the stated
criteria as does the inclusion of the Nodosinellacea ‘fibrous inner layer’, the Geinitzinacea ‘radially fibrous'
outer layer and the Colaniellacea outer ‘vitreous’ layer. The imprecision of this terminology and the

516

PALAEONTOLOGY, VOLUME 33

description of the Nodosinellidae as ‘fibrous or perforate', suggests that more detailed research is required into
this whole group and that perpendicularly arranged granules without optical alignment (fibrous structure as
in Palaeotextularia ) is being confused with radial hyaline structure.

The superfamily Endothyracea is not defined as multilocular or septate, although they follow the
Tourneyellacea, and although most of the genera show streptospiral to planispiral coiling with growth the
superfamily and the family Endothyridae and the subfamily Endothyrinae are defined as ‘planispiral to
streptospiral'.

The coiling mode in the superfamily Fusulinacea is not given in the superfamily definition. In the definition
given for the family Ozawainellidae it is stated that geological early taxa are planispiral and evolute but three
of the four Carboniferous genera are involute and only Pseudonovella is evolute with the final whorl
enveloping. This genus should probably be removed to the Loeblichiidae. This weakens the supposed
distinction between the Ozawainellinae and the Pseudostaffcllinae and between the Ozawainellidae and
Staffellidae. The presence of early streptospiral coiling is inadvertently left out of the definition of
Pseudostaffella. Within the Fusulinidae the distinction between the subfamilies Fusulinellinae and
Wedekindellininae is not clear. Within the Schwagerinidae the subfamily Chusenellinae is defined as possessing
plane or weakly folded septa in the early stage but both the genera described have plane septa initially. The
definition of the Verbeekinidae does not include the vital information that the septa are plane (only given in
the case of the subfamily Pseudodoliolininae). The subfamily Kahlerininae, defined as small with single tunnel,
does not fit the family definition.

Miliolina. Within the superfamily Cornuspiracea the distinction between the family Cornuspiridae (planispiral
to streptospiral) and the Hemigordiopsidae (streptospiral to planispiral) is not clear. This is because the
subfamily Meandrospirinae (planispiral or streptospiral) is included in the Cornuspiridae. Also, as defined, the
Meandrospirinae do not appear to be separable from the Calcivertellinae. Genera in which incipient chambers
and definite septation appear in the adult whorls are included in the Baisalinidae and those with a few chambers
throughout, and definite flexostyle, in the Fischerinidae. However, Dolosella in which the holotype has several
undivided whorls, is included in the Fischerinidae. The subfamily Spiriamphorellinae is defined as like
Ophthalmidium in the early stage. If so (the type figures are obscure) it cannot be included in the
Nubeculariidae, as defined.

The definition of the superfamily Miliolacea which rests entirely on coiling mode does not exclude the
Ophthalmidiidae and other forms with chambers of half-coil length included in the Cornuspiracea. This also
applies to that given for the family Hauerinidae because genera with simple apertures without teeth are not
excluded. Because the authors have now defined the family Miliolidae on the basis of its pseudoporous wall,
simple wall structure should have been included in the diagnosis of the Hauerinidae. Quinqueloculine genera
are included in the Miliolinellinae, so the distinction with the Hauerininae is not clear. The Riveroinidae are
defined as planispiral but only Riveroina is planispiral throughout. Pseudohauerinella is quinqueloculine and
may become planispiral in the final whorl.

The definition given for the superfamily Alveolinacea does not make a clear distinction with the Miliolacea.
This is largely because of the inclusion of the Fabulariidae which are milioline. They are defined as, ‘with
milioline early stage’ but the coiling mode in Fabularia and its allies is milioline throughout, the biloculine adult
chambers still being added end to end. Pseudofabularia and Pseudolacazina show a fundamentally different
coiling mode. The Mesozoic and Tertiary alveolines form two stratigraphically distinct groups. This is not
recognized at subfamily level.

The inclusion of the family Milioporidae within the Soritacea distorts the definition of that superfamily so
that it is described as ‘early stage pitted or perforate and less commonly may be perforate throughout growth’
which disagrees with the suborder diagnosis. However, although the Milioporidae are defined as perforate
throughout ontogeny, in three of the six subfamilies it is said to occur in the outer chambers or adult and in
Kamurana, the single genus in the Kamuraninae, it is also described as an adult character. In the other two
superfamily diagnoses it is not mentioned. Although the Milioporidae are defined as being coiled in various
planes (not in the superfamily diagnosis) the subfamilies show a variety of styles, even uncoiled rectilinear.
Kamurana is described as consisting of a globular proloculus followed by an individual tube, so may have
strayed in from the Cornuspiracea. The best place for this ‘family’ would seem to be amongst the genera of
uncertain status.

Lagenina. The Lagenina are defined as radiate (as is the genus Cryptoseptida ) rather than radial, and genera
within the family Syzraniidae are confusingly described as ‘radial fibrous’ and ‘hyaline pseudofibrous’ without
explanation. The family Robuloididae is inadvertently described as uniserial instead of planispiral. The

HAYNES: FO R A M I N I FER AN CLASSIFICATION

517

diagnoses given do not make a clear distinction between the Nodosariidae and Vaginulinidae because both can
include arcuate to uncoiled genera. The subfamily Marginulininae cannot be distinguished from the
Vaginulininae on the diagnoses given. The inclusion of Rimalina with apertural slit in the Glandulininae means
that this subfamily cannot be distinguished from the Entolingulininae on the criteria given.

Robertinina. The Robertinma are defined as planispiral to trochospiral and with internal partitions. However,
genera which becomes uniserial or uncoiled biserial are included in three of the four superfamilies recognized.
Only one of the three families within the Duostominacea is known to have partitions. The family
Epistominidae within the Ceratobuliminacea is defined as having peripheral apertural slits but the subfamily
Garantellinae is defined on the basis of its umbilical, ovate and areal apertures.

Globigerinina. The Globigerinina are defined as ‘radiate’ rather than radial hyaline and within the superfamily
Heterohelicacea, the family Guembelitriidae is defined as having trochospiral, triserial or biserial early stage
although the superfamily is defined as biserial or triserial which would appear to rule this family out. The
family Heterohelicidae is defined entirely on aperture form and the distinction with the Guembelitriidae is
therefore not clear.

The superfamily Rotaliporacea is defined as having extra-umbilical-umbilical aperture but the family
Globuligerinidae have umbilical apertures as well as imperforate pustules and should probably be transferred
to the Guembelitriidae. Within the family Hedbergellidae the subfamily Helvetoglobotruncaninae is defined as
possessing portici extending to the umbilicus. Whiteinella , included in the Hedbergellinae is also described as
having this feature so that distinction between the subfamilies breaks down.

The superfamily Globotruncanacea is defined as having umbilical aperture and tegilla covering the umbilical
area. However, in many genera the aperture is described as extra-umbilical-umbilical, e.g. Marginotruncana ,
and many are described as possessing portici rather than tegilla. The subfamily Abathomphalinae is defined
as having narrow umbilicus with ‘tegilla, that of final chamber covering the umbilical area’. This is inadequate
considering the superfamily definition. What is presumably meant is that the tegillum of the final chamber
becomes a single umbilical cover or ‘techo’. Within the family Globotruncanidae, the subfamily
Globotruncanellinae is defined as single keeled but genera with single keel are included in the
Globotruncaninae, e.g. Globotruncanita, and a number become single keeled with growth.

The superfamily Globorotaliacea cannot be distinguished from the Globigerinacea as defined. The
Globorotaliacea are defined as ‘ non-spinose, but may be pustulose or pitted’, whereas the Globigerinacea
‘May be covered with narrow spines’. The additional character of ‘numerous small pores or fewer large ones’
given for the Globigerinacea does not overcome the lack of a clear distinction. Within the Globorotaliacea
smooth genera are included within the family Eoglobigerinidae, also defined as having extra-umbilical
aperture without lip. However, of the four genera included only Parvularugoglobigerina fits this diagnosis. I
agree with Brinkhuis and Zachariasse (1988) that names of this length should be excluded under the rules of
the 1CZN, and in this case, Planconusa be substituted. The other three genera including Eoglobigerina , should
be transferred to the Guembelitriidae. As both the Globorotaliidae and Truncorotaliidae can be pustulose the
distinction between these family groups is not clear. The family Pulleniatinidae is described as streptospiral.
This character is not included in the superfamily diagnosis.

Within the Hantkeninacea the family Globanomalinidae includes Globanomalina which is low trochospiral
to planispiral. This is not included in the family or superfamily definitions.

The diagnosis of the superfamily Globigerinacea does not include the trochospiral to streptospiral coiling
of the included family Hastigerinidae or the areal apertures of the subfamily Orbulininae. The distinction
between the Porticulasphaerinae and the Orbulininae is not clear because both can have a later enveloping or
enclosing chamber.

Rotaliina. The authors subsume the Buliminida sensu Haynes (1981) within their Rotaliina. This leads to a very
unwieldly subordinal diagnosis and attempted recognition of twenty-four superfamilies but the criteria used are
inconsistent. The Bolivinitacea is reduced to two genera only. These can be distinguished from the others,
removed to the new superfamily Bolivinacea only by possession of truncate and carinate periphery. Genera
without toothplate are removed to the Loxostomatacea. Neither of these superfamilies would appear to
warrant recognition on the authors’ own stated criteria. Within the Bolivinitacea, the family Bolivinoididae
is set up to accommodate one genus, Bolivinoides, characterized by heavy costae, and a well known lineage in
the Upper Cretaceous. However, a number of Tertiary species within the Bolivinidae show a similar type of
ornament. Unless they are all to be given separate names, family distinction is hardly warranted.

The superfamily Cassidulinacea is much reduced, compared with the 1964 Treatise , when it accommodated

518

PALAEONTOLOGY, VOLUME 33

all hyaline-oblique rotaliids. Indeed, both radial and hyaline forms are now included in the family
Cassidulinidae, as are genera with and without toothplates. But although Orthoplecta is given separate
subfamily status because of possession of toothplates which make a ‘spiral column’, Islandiella , with its cornet
shaped toothplate is included with Cassidulina in the Cassidulininae. I consider these genera, both, incidentally,
radial, should be removed to the Bolivinitacea (Islandiellidae). The one trochoid, enrolled biserial genus known
is placed in a family of its own, Cassidulinitidae but the superfamily definition does not accommodate it.

As defined, the Eouvigerinacea cannot be satisfactorily separated from the Bolivinacea. The Lacosteinidae
includes some disparate forms probably of separate origin although initially planispiral. Lacosteina has a
drawn out ‘buliminelline’ adult spiral with loop-shaped aperture, while Elhasaella is twisted biserial with
terminal aperture. Interestingly, the surface textures are quite different.

The diagnoses given do not clearly separate the Turrilinacea and the Buliminacea. Although defined as
possessing simple aperture and toothplate the Turrilinacea include the Stainforthiidae with loop-shaped
aperture and well developed toothplate. A triserial genus without toothplate is also given family status.

Within the Buliminacea the family Reussellidae is not well defined because the angular chambers are not
mentioned. The distinction with the family Trimosinidae is not clear and this group probably deserves sub-
family status only. Despite the difficulties experienced earlier with the Cassidulinacea the authors exclude
hyaline-oblique forms from the Buliminacea and Turrilinacea and place them in the Fursenkoinacea. The
family Fursenkoinidae can only be distinguished from the Stainforthiidae on this basis. I consider that the
superfamilies Fursenkoinacea and Turrilinacea should be reunited in the Buliminacea, which should probably
include the Delosinacea as well which shows development of septal pits and bridges like the Virgulinellidae.
Similarly, I would include the Stilostomellacea (five genera only) in the Bolivinitacea on the grounds that they
represent uniserial end forms. Note that Revets (1989), considers that BuUminella lacks a toothplate and should
be excluded from the Buliminacea.

The superfamily Discorbacea is defined as low trochospiral but the first family within it, the Conorbinidae
is defined as low to high trochospiral and it also includes, within the Eponididae, the subfamily Rectoeponidinae
which becomes uncoiled and uniserial in the adult. The superfamily includes fifteen disparate families. No clear
distinction is made between families such as the Discorbidae and Rosalinidae which have open umbilicus
tending to become secondarily closed by umbilical flaps or bosses and those such as the Eponididae in which
the umbilicus is primarily closed. I consider these features should carry more weight than the presence of the
poreless patch used to bring together Baggina and Cancris which have very different umbilical characters.
Poreless areas are common in genera of diverse groups and like ornament and surface texture should not be
used at family level in isolation from other features. Incidentally, Baggina is defined as possessing closed
umbilicus but on pi. 591, fig. 6, the umbilicus is shown to be open. The Discorbidae are described as having
the chambers subdivided by a ‘paries proximus’ defined (Glossary) as a septal flap. This confuses the
toothplate (umbilical flap) with the septal flap attached to the previous apertural face.

The superfamily Glabratellacea is distinguished from the Discorbacea by the presence of radial striae around
the umbilicus on the ventral side which are presumed to facilitate and indicate plastogamic reproduction. The
mode of reproduction in most foraminifera is unknown and it is difficult to accept that plastogamy and radial
umbilical ornament is necessarily confined to one superfamily (largely to one family) considering the range of
the morphologies (including the high trochoid Buliminoides considered incertae familiae by Revets (1989)) that
are brought together by this approach. A case in point is that of Rosalina parisiensis. Specimens in the British
Museum (Natural History) collections, ascribed to this species on grounds of prominent ventral ornament,
include forms that otherwise clearly belong to Discorbis, Rosalina , Discorbitura , Discorbinella , Neoglabratella ,
Neoconorbina , etc. (Haynes and Whittaker in prep.). This suggests that ornament should be allowed little more
than specific weight in this group.

The superfamily Siphoninacea is not adequately defined. The development of planispiral and streptospiral
growth is only introduced in the definition of the family Siphoninidae. Uniserial should have been added to
the description of the subfamily Siphonminae to distinguish them from the trochospiral to uncoiled, biserial
Siphonidinae. However, there are only five genera in all.

The superfamily Discorbinellacea is distinguished from the Discorbacea on the basis of possession of an
arched or slit-like equatorial, basal aperture but this superfamily definition is inadequate because genera with
areal apertures such as Discorbitura are also included. This approach also brings together the Pseudoparrellidae,
with closed ventral umbilicus and the Discorbinellidae with umbilical open or secondarily closed with umbilical
flaps and/or bosses. It is also inaccurate in that four of the nine genera included in the family Discorbinellidae
do not possess equatorial apertures. One of these genera, Carlfranklinoides is clearly shown to have radiating
striae around the ventral umbilical aperture in pi. 631, fig. 14, although it is stated that ‘none is present on the
holotype’. On the authors’ own criteria this genus should have been transferred to the Glabratellacea. The

HAYNES: FOR A M IN I FER AN CFASSI FICATION

519

genus Cibicidoides is transferred to the family Parrelloididae although the aperture appears to extend around
the dorsal spiral suture as in other members of the Cibicides group. Woodella does not show the differentiation
into a relatively poreless ventral side and coarsely perforate dorsal side given as a family characteristic, so all
three genera placed here probably belong in different groups elsewhere.

As defined, the superfamily Planorbulinacea is stated to be distinguished by extra-umbilical to nearly
equatorial aperture and coarse perforations while plamspiral coiling is not mentioned. This would exclude the
Planulinidae, trochospiral to planispiral and finely perforate with equatorial aperture, and the Cibicididae,
aperture equatorial and possibly extending over the periphery and with a number of finely perforate genera.
The Bisacciidae are irregularly planispiral with sutural canals so probably should, indeed, be excluded. Within
the Cibicididae the genera are finely split but the distinctions are not clear, especially between Cibicides and
Lobatula. Discorbia with umbilical flaps round an open ventral umbilicus would be bi tter placed in the
Discorbinellacea. The Planorbulinidae clearly derive from the Cibicididae but the Cymbaloporidae included
here with them show no such relationship. The presence of umbilical apertures and cover plates, and lack of
dorsal attachment surface indicates a closer relationship with the Discorbacea. Montfortella , included in the
Cibicididae, should probably be transferred to the Victoriellidae which show the development of a
pseudumbilicus with secondary umbilical openings on the spiral, ventral side. The range of the Planorbulinacea
is given as Early Cretaceous to Recent on the basis of one genus only, Epithemella. Its ventral, umbilical
aperture and attachment surface suggest removal to the Discorbacea. This would make the range of the
superfamily, Late Cretaceous to Recent.

Spreading, irregular and branching forms are included in the superfamily Acervulinucea, described as
lacking an aperture apart from mural pores. In Acervulina the aperture is stated to consist only of coarse
perforations. However, the figures of A. inhaerens given on plate 659 clearly show peripheral and dorsal,
sutural apertures resembling those of Planorbulina. This seems to indicate that Acervulina probably represents
an extreme, irregular variant of Planorbulina and cannot be used as the type of a new superfamily.

The definition of the superfamily Asterigerinacea is inadequate in that it does not cover the orbitoidal
Lepidocyclinidae which are included. It is also defective in that it includes the formation of supplementary
chambers as a key feature which excludes the Epistomariidae, subfamily Nuttallidinae, on the authors’ own
criteria. Further, the families, Alfredinidae, Asterigerinatidae, Asterigerinidae (one genus only) and
Amphisteginidae (one genus only) cannot be distinguished on the criteria given. The authors synonymize the
Anomalinacea with the Asterigerinacea following the suggestion of Hansen and Rogl (1980) that Anomalina
(Anomalinidae) is identical with Epistomaroides (Alfredinidae). Here the authors have acted hastily. The type
of Anomalina has certainly been lost but the International Committee on Zoological Nomenclature has not yet
suppressed the name. Although Hansen and Rogl recovered Epistomaroides from Mauritius but failed to
recover the type species of Anomalina (A. punctulata d’Orbigny), in what is its type locality, this does not (and
cannot) prove that d’Orbigny actually described what we now call Epistomaroides as Anomalina. This is
unlikely because although there is a general similarity in coiling mode (shared by many genera) the apertural
and sutural details are quite different. The case of Anomalina should remain open.

The superfamily Nonionacea is defined as ‘planispiral to slightly asymmetrical' but a number of the genera
included are described as low trochoid, such as Spirotectina and in the case of Queraltina , described as
‘distinctly trochoid’. The Pullenia group is included in the family Nonionidae although they have primarily
closed umbilicus and lack the distinct perforation of Nonion and its allies. Melonis , although it is described as
having ten to twelve chambers in the final whorl, is included in the subfamily Pulleniinae, defined as possessing
‘few chambers per whorl'. The Almaenidae are also included in this superfamily although the general form,
peripheral apertures and coarse perforation suggest closer links with the Planorbulinacea.

The definition of the superfamily Chilostomellacea as trochoid and hyaline-oblique (‘granular’) is defective
because the family Chilostomellidae is then defined as trochospiral to planispiral and hyaline-radial genera are
included in the family Trichohyalidae. It also fails to cover the families included with the deep folds in the
apertural face (Alabaminidae), supplementary apertures (Oridorsalidae) and toothplate (Coleitidae). Although
the superfamily is defined as with ventral (‘umbilical’) aperture, the family Heterolepidae in which the aperture
runs on to the dorsal side, as in Cibicides, is also included. This is largely because of wall structure, revealed
by the authors’ statement that ‘optically granular’ species of Cibicidoides should be transferred to Gemellides ,
a hang-over from their attempted splitting of the Anomalinidae on these grounds in 1964, shown to be unsound
(Wood and Haynes 1957; Bellemo 1974; Haynes 1981). This family is better included in the Planorbulinacea.
The family Gavelinellidae differs from the Chilostomellidae in that the ventral aperture runs into the umbilicus
which tends to become closed with umbilical flaps and is nearer to the Discorbacea. Genera with primarily
closed umbilicus such as Hollandina and Paralabamina belong elsewhere. Hanzawaia, with flaps (lappets) on
the dorsal side, is another anomalinid. Discanomalina as illustrated appears to include disparate forms and

520

PALAEONTOLOGY, VOLUME 33

requires restudy; the inclusion of Irene McCullock’s figures of Pseudorosalinoides chatamensis being possibly
a lapsus calami. Unrelated forms also appear to be included in the family Trichohyalidae. Trichohyalus appears
to be a highly ornamented discorbid, near to Neoglabratella , whereas Aubignya lacks radial costae and has
umbilical lobes. The type of Aubignya , A. mariei, is closely similar to Ammonia perlucida and should probably
be transferred to the Rotaliacea.

The treatment of the superfamily Orbitoidacea is confused. The family Orbitoididae is split into two
subfamilies according to whether lateral chamberlets are differentiated or not but Schlumbergeria is
synonymized with Orbitoides , although it is without lateral chamberlets and appears to be its undifferentiated
ancestor. The superfamily is diagnosed as without canals and this is repeated in the definition of the family
Lepidorbitoididae. However, the subfamily Lepidorbitoidinae is taken to include Arnaudiella which has radial
canals. The subfamily is also taken to include the non-orbitoidal, canaliculate Daviesina , Praesiderolites,
Pseudosiderolites and Sulcoperculina. These genera belong elsewhere but this treatment supports the idea of
Van Gorsel that the Lepidorbitoididae originated from Pseudosiderolites (see Haynes 1981) and its inclusion
in the Rotaliacea.

The superfamily Rotaliacea is inadequately defined because it does not cover the orbitoidal families.
Pseudorbitoididae and Miogypsinidae which are included, or the conical Chapmaninidae. The Rotaliidae are
defined as trochoid but include the subfamily Cuvillierininae defined as trochospiral to nearly planispiral. On
examination this subfamily is found to include both trochoid forms, Pseudowoodella and fully planispiral
forms, Fissoelphidium. These should be excluded, with the Cuvillierina group recognized as a full family. It is
not clear what is meant by the distinction drawn between the subfamilies Rotaliinae and the Ammoniinae on
the basis that the umbilical area is 'primarily closed' by an umbilical flap in Rotalia , while it is 'secondarily
closed’ by a foraminal plate in Ammonia. A scroll-like toothplate is attached below the areal foramen in
Ammonia , whereas a partial umbilical partition is formed in Rotalia and the foramen remains a basal opening,
like the primary aperture. Both appear to be built at the same time as the chamber. In both genera
communication into the umbilical area below the umbilical ends of the chambers is initially open. In Ammonia
the umbilical area tends to be filled with a plug or plugs. In Rotalia , the umbilical (astral) lobes coalesce and
fuse as an umbilical coverplate similar to that in Discorbis. Together with the presence of the umbilical
partitions ('toothplate') this prompted the suggestion by Levy et al. (1984, 1986), that the Rotaliidae should
be subsumed within the Discorbidae. However, in the adult this coverplate is broken up by growth of pillars
and development of open sutural canals by resorption (Haynes and Whittaker, in press) which clearly
distinguish the genus from Discorbis. The authors also include the Elphidiidae, a primarily planispiral group
in this superfamily. However, I believe a more natural place for this group is within the Nomonacea, as early
forms grade with the Nonionidae both in wall structure and gross morphology. The appearance of the septal
flap seems to have been an independent development in this superfamily (Haynes 1973).

The Nummulitacea are defined as planispiral but include the family Pellatispiridae described as planispiral
to low trochoid. This family, without marginal cord and showing development of fissures, would more
naturally find a place in the Rotaliacea. Incidentally, although the authors state 'vertical canals or fissures may
be well developed’, in the Pellatispiridae, they are not mentioned in any of the genera described, e.g.
Miscellanea. The orbitoidal nummulitaceans are split into two families according to whether the annular
chambers of the equatorial layer are divided or not. However, Aster ophragmina is included in the
Discocyclinidae, defined as subdivided, although the adult annular rings of this genus are clearly undivided (pi.
819), and in the genera included in the Asterocycylinidae, defined as undivided, the annular rings are clearly
divided (Pis. 824 and 825). The reason for this apparent anomaly, is that the authors regard the annular rings
of the Asterocychmdae as consisting of cycles of small, spatulate to rectangular, primary chambers, whereas
in the Discocyclinidae each adult equatorial chamber is a ring, that may or may not be subdivided into small
secondary chamberlets. This should have been made clear. If it is actually the case, it indicates that these two
families are of quite separate origin and probably belong to different superfamilies, other superfamilies in the
scheme have been separated on features of rather less weight.

DISCUSSION AND CONCLUSIONS

Understandably, considering the complexity of the problem and the present state of knowledge, the
authors have only been partially successful in producing a more natural (genetic) classification.
Despite a brave attempt to 'untie the knot’, they would almost certainly be the first to admit that
large elements remain clearly artificial. Also, as pointed out by the reviewers, in many cases the
diagnoses are vague and even contradictory and they have not been able to push through a

HAYNES: FO R A M I N I FE R AN CLASSIFICATION

521

thorough-going, logical hierarchical scheme. As might be expected, the large amount of new
material and information that has become available (and is still in full flood) has produced more
problems than it has solved. It has forced them to abandon their 1964 attempt to make a clearcut
division on the basis of wall structure, aperture form and coiling mode and to raise other characters
to suborder and superfamily rank. In so far as this is a measure of the real complexity of
relationships it is to be welcomed, because there can be no return to past simplicities, however
beguiling.

A positive attempt is made to solve the problem of Carpenter’s perforate ‘arenaceous’ group by
recognizing them as a separate superfamily. Although it is still not clear whether the diverse
canaliculi, passages, pseudopores and tubuli in other agglutinated genera all represent separate
lines, this move is supported by work on wall structure by Bronniman and Whittaker (1988) and
by Bender (1989). Other changes, such as the use of surface texture at superfamily rank, are likely
to be more controversial. Certain weaknesses also persist from their former classification. Despite
the demonstration that hyaline-radial and hyaline-oblique (‘granular’) wall structure are close to
each other, so that one is easily derived from the other, and the consequent abandonment of
hyaline-oblique structure as a basis for recognition of the Cassidulinacea, this distinction is still a
major factor in regrouping of genera in other superfamilies. The tendency to discount ontogeny as
an indication of evolution is also inherited from the Treatise and also leads, I believe, to horizontal
groupings, particularly in the case of the Textulariina and the Lagenina.

It is evident from this historical account, down to the latest and most ambitious attempt, that a
stable, i.e. natural, classification of the Foraminifera is still a distant goal. The lesson for
stratigraphers and, particularly, macro-evolutionists, is that they must treat suprageneric taxa and
even genera with great caution. In the circumstances, workers not primarily interested in
classification and evolution can be forgiven for thinking they should be abandoned. For instance.
Culver et al. (1987, p. 1 69) suggest that, ideally, species-level data should be employed, because they
doubt ‘the significance of evolutionary generalisations based on the entirely human constructs of
higher taxonomic categories’. But the horizontal limits we draw between species are also artificial,
and as viewed in time, species are human artifacts, whether recognized as part of a phyletic line or
as an allopatric branch (in both space and time) where the connecting cline is known. The objective
biological reality of living species viewed as interbreeding population groups, separated from their
closest relatives by morphological discontinuities, must not be confused with the subjective limits
we are bound to draw between these species and their ancestors. For instance, even where we know
the ancestry in a clear cut case of allopatric speciation, we still have the problem of deciding the
limits between parent species (type subspecies), the peripheral subspecies (possible founder
populations) and the new species (type subspecies). Most species are cryptogenes (the ancestry is not
known) and we can only estimate the possible number of ancestral taxa and intermediate forms lost
in the stratigraphical gaps or not preserved. We have to be careful not to confuse the species, as
preserved, with the ‘natural species’ and appearance in the stratigraphical record with ‘origination’.

It is precisely these confusions which vitiate recent attempts to replace ‘traditional intuitive’
taxonomy, by the methods of numerical taxonomy, or those of cladistics, with which it is hoped to
draw hard and fast lines between species. In the case of numerical taxonomy, there will be no quarrel
with the introduction of more refined measurements. It is the attempt to place reliance entirely upon
numerical values, given equal weight, that cannot be sustained. Of course, if weighting is applied,
the supposed ‘objective’ nature of the enterprise is destroyed. Numerical methodology cannot be
more than the handmaid of taxonomy and in this role provides a very useful adjunct, as is shown
by the work of Barnett (1974), Malmgren (1974) and Scott (1974).

In the case of cladistics we have to separate the extraordinary claims made for it as ‘scientifically
rigorous and operational’ compared with traditional taxonomy, and its results in practice. Like
numerical taxonomy, it pretends to objectivity in replying on morphological characters given equal
weight to establish the degree of genetic difference. It is supposed that ‘ real ’ species, as distinct from
‘hypothetical’ species, can be defined on the basis of ‘splitting events’, i.e. ‘discontinuities’, and
further, that these ‘speciation events’ are the only objective (‘ non-intuitive ') boundaries, apart from

522

PALAEONTOLOGY, VOLUME 33

terminal extinction, available for definition of species taxa. Genera are taken to include daughter
species and ancestral species back to the next available name.

The most ambitious attempt to apply cladistics to foraminifera is that by Fordham (1986), on the
Cenozoic planktonic fauna of the Pacific, reviewed by Banner (1987), Chaproniere (1987) and
Adams (1987). These authors are as one in dismissing it as neither objective, nor operational or
practical. The discontinuities between species-clusters are distinguished by the first and last
appearances of constituent subspecies (phena), some by only one, although it is admitted they are
arbitrary. The fossil record is presumed to be complete and no account is taken of discontinuities
caused by variations in sedimentation rate, sedimentary breaks, ocean'ographical changes and
inclusion in the death assemblage of species and subspecies adapted to different levels of the
stratified water column. A further serious objection is that the method relies on a severely restricted
model of allopatric speciation in which the dichotomous appearance of two daughter species leads
to elimination of the ancestral species. As Adams points out, most biostratigraphical work proceeds
on the ‘ well-justified ' assumption that ancestral species do not necessarily die out as a consequence
of speciation. Not only are geological events confused with originations of palaeontological species
but also with biological ‘speciation events’. Clearly, cladistics is a metaphysical system imposed
upon the facts rather than arising from them. It is unlikely to lead to more soundly based species
or supraspecific categories. The bold claim that ‘Histories, terminations and the reproductive nexus
are easy’ (Kitcher 1989) will, therefore, ring hollowly in the ear of the practising taxonomist.

The central place occupied in cladistics by an extreme model of allopatric speciation reflects
recent fashion, particularly in America (one does not have to be a marxist to see the connection).
However, although Haldane ( 1 959) prophesied that allopatric speciation would come to be accepted
as the norm, and Ruse (1982) decided that ‘totally sympatric speciation is rare’, the stratigraphical
record of the Foraminifera suggests that both rapid allopatric branching and relatively slower,
sympatric transformation take place. Well known evolutionary trends in larger foraminifera
interpreted as examples of phyletic evolution are shown by Lepidolina (Ozawa 1975), the Orbitoides
lineage (Gorsel 1978), Cycloclypeus (MacGillavry 1978) and in planktonics by the Globorotalia
menardii / tumida lineage (Malmgren et al. 1983). The gradual trend in Lepidolina was dismissed as
‘pseudogradualism' by Gould and Eldredge (1977) and a possible result of ‘clone selection’.
However, these larger foraminifera show both sexual and asexual generations.

One of the reasons for the popularity of exclusively, allopatric solutions may be that formerly,
average selection pressure was thought to be quite low; Fisher (1930) considered I % selection for
advantageous qualities was a reasonable average throughout geological time. As a consequence,
emphasis was put on the isolation of small populations (even reduction to a biblical pair) with only
a fraction of the parental gene-pool, to account for rapid change. However, it is now known (Ford
in Huxley 1974) that selective forces are up to thirty times greater in nature than had been realized,
and therefore more effective on larger populations than previously admitted. After all, although the
Galapogos finches are a remarkable example of rapid differentiation from small, immigrant,
founder populations, the case of the medium ground finch (Geospiza fords), which was permanently
reduced to the large beaked form by a major drought (Boag and Grant 1981), represents rapid
sympatric evolution through operation of an environmental ‘bottleneck’ on a relatively large
population.

The many convergent lines of evolution towards complex orbitoidal structure in the large
rotaliids appear to be related to the development of low fertility, reefal environments, perhaps with
initial ‘bottleneck’ reduction to the most effective symbiotic species, amongst small rotaliids,
followed by gradual acquisition of a more efficient algal-greenhouse structure. All the stages can be
seen particularly well in the Miogypsinidae, Orbitoididae and Lepidocyclinidae. There is, of course,
no occasion for the use of the ‘morphological gap', but genera can be distinguished on the basis of
important structural changes and innovations (‘novelties’ of Adams 1983), e.g. the appearance of
lateral chamberlets which distinguish Miogypsina from Miogypsinoides, and Orbitoides from
Schlumbergeria.

Generic and suprageneric categories are natural as far as they represent clusters of species that

HAYNES: FORAMINIFERAN CLASSIFICATION

523

have evolved from ancestral groups by changes at the species (subspecies) level. Again, as far as they
are natural, they will be 'real entities’ that reflect major evolutionary responses to the challenge of
global physiographic change and the opportunities provided by extinction of groups that formerly
dominated a particular environmental range. Loeblich and Tappan’s new classification, despite its
imperfections, reveals even more dramatically than formerly how replacement has been a constant
theme in the evolution of the Foraminifera (for ranges of the suprageneric groups see Ross and
Hainan 1989). Generic ranges are given by Danielle Decrouez (1989). However, it is sobering to
realize that in no case is there direct proof of the evolutionary origin of any of the twelve suborders
recognized and the ancestral groups remain speculative.

PRINCIPLES OF CLASSIFICATION

After the historical study it will be helpful to provide a brief statement of what appear to be the
prerequisites of a more natural classification.

1 . Foraminifera are the largest and most varied invertebrate group, have the most complex shells
and include some of the largest protozoans. Although acellular, life histories are complex, there may
be multinucleate phases and chromosome numbers are high. Even Class status may provide too
narrow a framework for the complexity of relationships to be expressed.

2. A simple ordinal/subordinal scheme based on wall structure is not possible. There are at least
two calcitic Palaeozoic groups, one microgranular, one largely 'vitreous'; both perforate and
imperforate groups of agglutinated forms, plus a group secondarily derived from the miliolids; two
groups of porcelaneous forms, one with perforate ancestry; four groups of microcrystalline calcitic,
hyaline forms, the buliminids possibly derived from perforate agglutinated forms, the nodosariids
from 'vitreous' ancestors, and the globigerinids and rotaliids of unknown origin. The relationship
of the calcitic groups to the poorly known aragonitic group is obscure.

3. Coiling modes are not ' isomorphous’ in the different wall structure groups but adaptive
radiation at different times into broadly similar environments and ecological niches has led to
heterochronous homoeomorphy with the constant repetition of certain styles. There are trends in
test architecture towards both increased complexity and simplification and both coiling-up and
uncoiling (even recoiling and uncoiling) occur. In general, groups adapted to vagrant surface
feeding in high energy environments remain close-coiled; groups secondarily adapted to infaunal
environments often show uncoiling and simplification with growth and attendant serial, apertural
modifications; groups adapted to symbiotic life show increased size and complication of structure.
Contrasting modes within the larger Foraminifera illustrate the 'developmental constraints’ of the
non-lamellar test, i.e. only the Somalina line in the porcelaneous group achieves 'orbitoidal'
structure. New features tend to occur in the adult and proterogenesis is less common, but especially
in symbiotic forms there is often striking differentiation of the juvenile (possibly connected with
dispersal). Multiform genera probably adopt different life habits at different stages. Neotony may
be a factor in the rise of certain groups, such as the planktonics. Ontogeny, therefore, has to be used
with great care as a guide to phylogeny.

4. Only detailed stratigraphical and evolutionary studies can solve the problems of iterative
evolution, parallel development and convergence peculiar to Foraminifera. In these circumstances,
numerical taxonomy and cladistics (which in its 'transformed' manifestation excludes these
problems from consideration) cannot be substituted for traditional taxonomy, i.e. the careful
comparison of test architecture and wall structure, apertural and internal features, taking gross
similarities and differences into account, with the evidence of ontogeny and the stratigraphical
record, to establish phylogeny.

5. Although it is not true that traditional taxonomy relies exclusively on unique rather than
shared characters, as supposed by Fortey ( 1989), care must be taken not to employ any one feature
as a single 'key'. As Carpenter put it, ‘a natural system depends on the whole aggregate of
ascertainable characters’. In this aggregate we must include coiling-mode; banished by Carpenter
but continuing to haunt classification like Banquo's ghost. The supposition that features established

524

PALAEONTOLOGY, VOLUME 33

at one level in one order can be applied with equal weight in another order, with a different
evolutionary history, must also be avoided.

6. 'Species are made by isolation,’ as Darwin put it, but populations can be ‘islanded’ in time
as well as space and evolutionary patterns are complex in Foraminifera, suggesting that both
allopatric branching and sympatric phylogenesis have taken place. As is to be expected, rates are
variable, though sympatric phylogenesis appears generally to be relatively slow, perhaps because
selection is working on larger populations. Attempts to arrive at an ‘objective’, supraspecific
classification by application of a stereotyped model of dichotomous allopatric speciation should be
avoided. Darwin’s famous model of the irregularly branching tree of life (the only diagram in the
Origin) is more appropriate. Also, as pointed out by Adams (1983) the relatively rapid appearance
of ‘novelties’ which then become stable, together with the slower modification of existing characters
can give the appearance of punctuated equilibrium and gradualistic evolution in the same family.

7. The contribution of biological studies, although important, is so far limited. Paradoxically,
despite their stratigraphical importance, less is known about living Foraminifera than about the
living representatives of any other major group. Further studies of life history, functional
morphology, genetics and cytology are a sine qua non of an improved classification. The study of
living, naked forms (Allogromiida) is also a particular need.

Acknowledgements . I am indebted to the two referees who commented on the original draft of this paper,
intended as an essay review of Loeblich and Tappan’s book, and who encouraged me to examine it in historical
context, especially in regard to problems that can be traced back to the work of the English School. In
particular. Dr C. G. Adams and Dr J. E. Whittaker provided references and copies of reviews of earlier
classifications, including the attempt to use cladistics by Fordham. Unfortunately, I was unable to borrow a
copy of the doctoral thesis on foraminiferal classification by Martin Janal for copyright reasons. The paper was
typed by Mrs Janice Evans.

REFERENCES

adams, c. G. 1983. Speciation, phylogenesis, tectonism, climate and eustacy: Factors in the evolution of
Cenozoic larger foraminiferal bioprovinces. 255-289. In sims, r. w., price, j. h. and whalley, p. e. s. (eds.).
Evolution , time and space. The emergence of the biosphere. Systematics Association Special Volume 23.

- 1987u. On the classification of the Lepidocyclinidae (Foraminiferida) with redescriptions of the unrelated
Paleocene genera Actinosiphon and Orbitosiphon. Micropaleontology , 33, 289-317.

1987A Review: Miocene- Pleistocene planktonic foraminifers from DSDP Sites 208 & 77, and phylogeny
and classification of Cenozoic species. Fordham, B.G., 1986. Journal of Natural History , 21, 1078-79.

- 1988. Septa, septal traces and septal filaments in the foraminiferal genus Numnudites Lamarck. Journal
of Micropalaeontology, 7, 89-102.

banner, f. t. 1987. Review: Miocene-Pleistocene planktonic foraminifers from DSDP sites 208 & 77, and
phylogeny and classification of Cenozoic species, Fordham, B.G., 1986. British Micropalaeontological
Society , 31, 26-28.

- 1988. Review: Foraminiferal genera and their classification, by Alfred R. Loeblich and Helen Tappan.
British Micropalaeontological Society Newsletter , 36, 13-15.

barnett, r. s. 1974. An application of numerical taxonomy to the classification of the Nummulilidae
(Foraminiferida). Journal of Paleontology, 48, 1249-1263.

bellemo, s. 1974. The compound and intermediate wall structures in Cibicidinae (Foraminifera) with remarks
on the radial and granular wall structures. Bulletin of the Geological Institutions of the University of Uppsala,
New Series, 6, 1-1 1 .

bender, h. 1989. Gehauseaufbau, Gehausegense und Biologie agglutinieter Foraminiferen. Jalirbuch
Geologischen Budesamstalt, 132, 259-347.

berggren, w. a. 1965. Review: Treatise on Invertebrate Paleontology, Part C: Protista 2 - Sarcodina, by
Alfred R. Loeblich and Helen Tappan. Micropaleontology, 11, 122-124.

boag, p. t. and grant, p. r. 1981. Intense natural selection in a population of Darwins Finches (Geospizinae)
in the Galapogos. Science, 214, 82-85.

brady, h. b. 1884. Report on the Foraminifera dredged by H.M.S. Challenger during the years 1873-1876.
Scientific results of the voyage of H.M.S. Challenger during the years 1873-1876. Zoology, 9, 1-814.

HAYNES: FO R A M I N I FE R AN CFASSIFICATION

525

brasier, M. d. 1982. Architecture and evolution of the foraminiferid test - a theoretical approach. 1-41. In
banner, F. T. and lord, A. R. (eds. ). Aspects of Micropalaeontology, Allen and Unwin, London.
brinkhuis, h. and zachariasse, w. J. 1988. Dinoflagellate cysts, sea level changes and planktonic foraminifera
across the Cretaceous-Tertiary boundary at El Haria, Northwest Tunisia. Marine Micropalaeontology, 13,
153-191

bronniman, p. and Whittaker, j. E. 1983. Zaninettia n.gen., a spicular-walled remaneicid (Foraminiferida,
Trochamminacea) from the Indian and South Atlantic Oceans with remarks on the origin of the spicules.
Revue de Paleobiologie, 2, 13-33.

— 1988. An agglutinated wall structure and the new foraminiferal suborder Trochamminina. Revue de
Paleobiologie , 7, 109-119.

carpenter, w. B. 1850. On the microscopic structure of Nummulina , Orbitolites and Orbitoides, Quarterly
Journal of the Geological Society of London (1849), 6, 21-39.

— 1856. Researches in Foraminifera. Part 2. On the genera Orbiculina, Alveolina, Cycloclypeus and
Heterostegina. Philosophical Transactions of the Royal Society of London, 146, 547-569.

1860. Darwin on the origin of species. National Review, 10, 188-214.

— 1861. On the systematic arrangement of the Rhizopods. Natural History Review, 1, 456-472.

— assisted by Parker, w. k. and jones, t. r. 1862. Introduction to the study of foraminifera. Ray Society,
Hardwicke, London, xxii + 319 pp.

challinor, j. 1959. Palaeontology and evolution. 50-100. In bell, p. r. (ed.). Darwin’s biological work.

Cambridge University Press, Cambridge, xiii + 345 pp.
chapman, f. 1900. Foraminifera from an Upper Cambrian horizon in the Malverns: together with a note on
some of the earliest known foraminifera. Quarterly Journal of the Geological Society of London, 56, 257-263.

— and parr, w. j. 1936. A classification of the Foraminifera. Proceedings of the Royal Society of Victoria,
49, 139-151

chaproniere, c. h. 1987. Review: Miocene-Pleistocene planktonic foraminifers from DSDP Sites 208 & 77,
and phylogeny and classification of Cenozoic species. Fordham, B.C. 1986. Nomen Nudum, 16, 25-28.
culver, s. j., buzas, m. a. and collins, l. s. 1987. On the value of taxonomic standardization in evolutionary
studies. Paleobiology, 13, 169-176.

Cushman, j. a. 1927. An outline of a reclassification of the foraminifera. Contributions from the Cushman
Laboratory for Foraminiferal Research , 3, 1-105.

— 1928. Foraminifera their classification and economic use. 1st Edit. Special Publications of the Cushman
Laboratory for Foraminiferal Research, 1, 1-401.

— 1948. Foraminifera their classification and economic use. 4th edn. Harvard University Press, Cambridge,
Massachusetts, ix + 605 pp.

— 1932. The Foraminifera of the Tropical Pacific Collections of the ‘Albatross’, 1899 1900. Pt. 1. Bulletin
of the United States National Museum , 161, iv+ 1-88.

1933. The Foraminifera of the Tropical Pacific Collections of the ‘Albatross’, 1899-1900. Pt. 2. Bulletin

of the United States National Museum, 161, vi+ 1-79.

1942. The Foraminifera of the Tropical Pacific Collections of the ‘Albatross’, 1899-1900. Pt. 3. Bulletin
of the United States National Museum, 161, 1-67.

darwin, c. 1866. On the origin of species by means of natural selection, or the preservation of favoured races in
the struggle for life. Forth edition, with additions and corrections. John Murray, London, 593 pp.

— 1945. The origin of species by means of natural selection. Reprint of 6th edn. Thinker’s Library no. 8,
Rationalist Press, Watts and Co., London, vii + 434 pp.

decrouez, danielle, 1989. Generic ranges of Foraminifera. Revue de Paleobiologie, 8, 263-321
defrance, m. j. l. 1822. Dictionnaire des Sciences Naturelles vol. 25. F.G. Levrault, Paris, 485 pp.
douville, h. 1906. Evolution et enchainement des Foraminiferes. Bulletin de la Societe Geologique de France ,
series 4, 6, 588-602.

dujardin, f. 1835. Observations nouvelles sur les pretendu Cephalopodes microscopiques. Annales des
Sciences Naturelles, series 2, 3, 312-314.

earland, a. 1934. Foraminifera, Part 3. The Falklands Sector of the Antarctic (excluding South Georgia).
Discovery Reports, 10, 1-208.

fisher, r. a. 1930. The genetical theory of natural selection. Clarendon Press, Oxford, xvi + 272 pp.
fordham, b. g. 1986. Miocene- Pliocene planktic foraminifers from DSDP Sites 208 and 77, and phylogeny and
classification of Cenozoic species. Evolutionary Monographs , 6, Chicago, 200 pp.
fortey, r. a. 1989. The collection connection. Review: Wonderful Life. The Burgess Shale and the nature of
history, by S. J. Gould. Nature, London , 342, 303.

526

PALAEONTOLOGY, VOLUME 33

galloway, J. j. 1928. The change in ideas about foraminifera. Journal of Paleontology , 2, 216-228.

— 1933. A manual of Foraminifera. Principia Press, Bloomington, Indiana, xii + 483 pp.

GLAESSNER, m. f. 1945. Principles of micropalaeontology . Melbourne University Press and Oxford University
Press, Oxford, xvi + 296 pp.

gorsel, j. t. van 1978. Late Cretaceous orbitoidal foraminifera, 1-120. In hedley, r. j. and adams, c. g. (eds.).

Foraminifera , 3. Academic Press, London and New York.
gould, j. s. and eldredge, n. 1977. Punctuated equilibria: the tempo and mode of evolution. Paleobiology ,
3, 115-151.

gronlund, h. and hansen, h. j. 1976. Scanning electron microscopy of some Recent and fossil nodosariid
foraminifera. Bulletin of the Geological Society of Denmark , 25, 121-134.
haldane, j. B. s. 1959. Natural selection, 101-149. In bell, p. r. (ed.). Darwin's biological work. Cambridge
University Press, Cambridge, xiii + 343 pp.

haman, d. 1967. A taxonomic re-interpretation and emendation of the genus Technitella Norman, 1878.
Contributions from the Cushman Foundation for Foraminifera I Research , 18, 27-29, pi. 6.

1988. Review: Foraminiferal genera and their classification, by Alfred R. Loeblich and Helen Tappan.
Journal of Foraminiferal Research , 18, 271-274.

hansen, H. j. and rogl, F. 1980. On Anomalina puctulata d’Orbigny 1826. Journal of Foraminiferal Research,
10, 153-155.

— and reiss, z. 1972. Scanning electron microscopy of wall structures in some benthonic and planktonic
Foraminiferida. Revista Espanola de Micropaleontologia, 4, 169-179.

haynes, j. r. 1956. Certain smaller British Paleocene Foraminifera. Part 1. Contributions from the Cushman
Foundation for Foraminiferal Research , 7, 79-101.

1981. Foraminifera , Macmillan, London and Basingstoke, 433 pp.

1988. Massively cordate nummulilids (foraminifera) at the Palaeocene/Eocene boundary in Tethys. Sixth
Abanathi Chaudhuri Memorial Lecture, University of Calcutta. Indian Journal of Geology , 60, 215-230.

— and nwabufo-ene, K. 1988. Palaeocene nummulitids from Libya and West Africa. Indian Journal of
Geology, 60, 230-247.

— et al. 1973. Cardigan Bay Recent Foraminifera (cruises of the R.V. Antur, 1962-64). Bulletin of the British
Museum ( Natural History ) (Zoology), Supplement 4, 1-245.

henbest, l. G. 1937. Keriothecal wall structure in Fusulina and its influence on fusuline classification. Journal
of Paleontology, 11, 212-230.

heron-allen, e. 1917. ‘ Alcide d’Orbigny, his life and work’. Journal of the Royal Microscopical Society, 1917,
1-105.

— 1929. Foraminifera. Review of Foraminifera : their classification and economic use. J. A. Cushman,
Nature, London, November 2, 680-682.

— 1935. ‘ Dementia nomenclatorica americana" . Correspondence, Nature, London, November 16, 797.
hofk.er, j. 1951. The foraminifera of the Siboga Expedition. Part III. Siboga-Expeditie, Monographic IVa,

1-513. E. J. Brill, Leiden.

huxley, j. s. 1974. Evolution - the modern synthesis. Third Edition, baker, j. r. (ed.). Introduction. Allen and
Unwin, London, ixxvii+ 13-705 pp.

hyatt, a. 1900. Cephalopoda. 502-604. In zittell, k. a.. Textbook of palaeontology, 1, London, 706 pp.
jenkins, D. G 1989. Subgenera in foraminiferal classification. British Micropalaeontological Society Newsletter,

38, 18.

jepps, M. w. 1942. Studies on Polystomella Lamarck. Journal of the Marine Biological Association of the United
Kingdom , 25, 607-666.

1956. The Protozoa, Sarcodina. Oliver and Boyd, Edinburgh and London, 183 pp.
jones, T. R. in Griffiths, J. w. and henfrey, A. 1875. The micrographic dictionary 1, 316-320. Third edition, van
Voorst, London.

kitcher, p. 1989. Bewitchment of the biologist. Review of Unfinished synthesis, N. Eldredge. Nature, London,
320, 649-650.

knoll, a. h. and Butterfield, n. j. 1989. New window on Proterozoic life. Nature, London, 337, 602-603.
lamarck, j. p. b. a. de, 1804. Suite des memoires sur les fossiles des environs de Paris. Annales Museum
National d’Histoire Naturelle, 5, 179-188.

lecalvez, j. 1938. Recherche sur les foraminiferes. Archives de Zoologie Experimental et Generate , 80,
163-333.

lee, j. j. 1966. Review: Treatise on Invertebrate Paleontology, Part C: Protista 2- Sarcodina, by Alfred R.
Loeblich and Helen Tappan. Micropaleontology, 12, 246-247.

HAYNES: FO R A M I N I FER AN CLASSIFICATION

527

levine, n. d., corliss, J. o., cox, F. E. G. et at. 1980. A newly revised classification of the Protozoa. Journal of
Protozoology , 27, 37-59.

levy, a., mathieu, R., poignant, a. and rosset-moulinier, m. 1984. Une nouvelle conception des families
Discorbidae et Rotaliidae. Benthos ‘83 ; 2nd International Symposium on Benthic Foraminifera. Pau, April
1983, 381-387.

1986. Discorbidae and Rotaliidae; a classification to be revised. Journal of Foraminiferal

Research , 16, 63-70.

lister, j. j. 1903. The Protozoa; Section I - The Foraminifera, 147-148. In lankester, e. r. (ed. ). A treatise
on zoology , 1, fascicle 2, Black, London.

loeblich, a. and tappan, h. 1964. Protista 2. Sarcodina chiefly ‘Thecamoebians’ and Foraminiferida. In
MOORE, R. c. (ed.). Treatise on invertebrate paleontology , Part C, 2 vols. Geological Society of America and
University of Kansas Press, Boulder, Colorado and Lawrence, Kansas, 900 pp.

— 1974. Recent advances in the classification of the Foraminiferida, 1—53. In hedley, r. h. and adams,
c. G. (eds.). Foraminifera 1. Academic Press, London.

1984. Suprageneric classification of the Foraminiferida (Protozoa). Micropaleontology , 30, 1-70.

— 1988a. Foraminiferal genera and their classification. Van Nostrand Reinhold Company, New York,
viii + 970 pp. 847 plates in separate volume. Library of Congress Card No. 87-26960.

— 1988A Foraminiferal evolution, diversification and extinction. Journal of Paleontology , 62, 695-714.
macgillavry, h. j. 1978. Foraminifera and parallel evolution - how or why? Geologie en Mijnbouw , 57,

385-394.

malmgren, b. a. 1974. Morphometric studies of planktonic foraminifers from the type Danian ol southern
Scandinavia. Stockholm Contributions in Geology , 29, 1 126.

berggren, w. a. and lohman, g. p. 1983. Evidence for punctuated gradualism in the late Neogene

Globorotalia tumida lineage of planktonic foraminifera. Paleobiology , 9, 377-389.
margulis, l. 1974. Five kingdom classification and the origin and evolution of cells, 45-78. In dobzhansky,
t., hecht, m. k. and steer, w. c. (eds.). Evolutionary Biology, 7, Plenum Press, New York.
martin, r. d. 1988. Life at the top of the tree. Nature, London , 335, 589.

montfort, p. denys DE 1808. Conchy liologie systematique et classification methodique des coquilles 1, F.
Schoell, Paris, 410 pp.

myers, E. H. 1935. The life history of Patellina corrugata. Williamson, Bulletin of Scripps Institute of
Oceanography, Technical Series, 3, 355-392.

- 1936. The life-cycle of Spirillina vivipara Ehrenberg with notes on morphogenesis, systematics and
distribution of the foraminifera. Journal of the Royal Microscopical Society, 56, 120-146.

NEWELL, N. D. 1982. Mass extinctions - illusion or realities. Geological Society of America Special Publications,
190, 257-263.

orbigny, a. d. d’ 1826. Tableau methodique de la classe des Cephalopodes, 111 order Foraminiferes. Annals
des Sciences Naturelles , 7, 245-314.

— 1839. Voyage dans I'Amerique Meridional - Foraminiferes, 5, 1-86. Pitois-Levrault, Paris.

ozawa, T. 1975. Evolution of Lepidolina multiseptata (Permian Foraminifer) in East Asia. Memoirs of the
Faculty of Science, Kyushu University, 23, 117-164.
reiss, z. 1958. Classification of lamellar foraminifera. Micropaleontology, 4, 51-70.

- 1963. Reclassification of perforate Foraminifera. Bulletin of the Geological Survey of Israel. 35, 1-111.
reuss, a. E. 1862. Entwurf emer systematischen Zusammenstellung der Foraminiferen. Sitzungsberichte der

Kaiserlichen Akademie der wissenschaften in Wien, Mathematisch-Naturwissenschaftliche Classe, 44 (1861), 1,
355-396.

revets, s. a. 1989. Structure and comparative anatomy of the toothplate in the Buliminacea. Journal of
Micropalaeontology, 8, 23-36.

rhumbler, L. 1895. Entwurf eines naturlichen Systems dem Thalamophoren. Nachrichten von der Gesellschaft
der Wissenschaften zu Gottingen, Math-Physik. Klasse 1895, 1, 51-98.

— 1911. Die foraminiferen (Thalamophoren) der Plankton-Expedition; Erster Teil: Die allgemeinen
Organizations verhaltnisse der Foraminiferen. Ergebnisse der Plankton - Expedition der Humboldt- Stif lung.
Kiel und Leipzig, 3 (1C), 1-331.

1923. Foraminiferen (Thalmophoren). 50-89. In kukenthal, w. and krumbach, t. (eds.). Handbuch der
Zoologie. 1, Berlin, Leipzig.

ross, c. a. and haman, d. Suprageneric ranges of Foraminiferida. Journal of Foraminiferal Research, 19, 72-83.
ruse, M. 1982. Darwinism defended - a guide to the evolution controversies. Benjamin/Cummings, Menlo Park,
California, xvii + 356 pp.

528

PALAEONTOLOGY, VOLUME 33

Schubert, r. J. 1908. Beitrage zu einer natiirlichen Systematik der Foraminiferen. Neues Jahrbuch fiir Geologie
und Paldontologie , Beilage-Band, 25, 232-260.

— 1921. Palaontologische Daten zur Stammesgeschichte der Protozoen. Palaontologische Zeitschrift, 3,
129-188.

scott, G. h. 1974. Biometry of the foraminiferal shell. 55-151. In hedley, r. h. and adams, c. g. (eds.).

Foraminifera 1. Academic Press, London and New York.
sleigh, M. 1989. Protozoa and other Protists. Edward Arnold, London, New York, x + 342 pp.

SMOUT, a. h. 1954. Lower Tertiary Foraminifera of the Qatar Peninsular. Monograph of the British Museum
( Natural History ), London, viii + 96 pp.

spath, l. f. 1933. Evolution of the Cephalopoda. Biological Reviews , 8, 418-462.

Thompson, d’arcy w. 1942. On growth and form. Second Edition. Cambridge University Press, Cambridge,

1116 pp.

towe, k. m. and cifelli, r. 1967. Wall ultrastructure in the calcareous foraminifera: crystallographic aspects
and a model for calcification. Journal of Paleontology , 43, 742-762.
valen, l, van. 1973. Are categories in different phyla comparable? Taxon , 22, 333-373.
walker, G. and boys, w. 1774. Testacea minuta rariora. G. Walker, London, 25 pp.

Williamson, w. c. 1852. On the minute structure of the calcareous shells of some Recent species of
foraminifera. Transactions of the Microscopical Society of London , 3, 105-128.

1858. On the Recent Foraminifera of Great Britain. Ray Society of London Publications, xx, London,
107 pp.

wood, a. 1947. The supposed Cambrian Foraminifera from the Malverns. Quarterly Journal of the Geological
Society of London, 102, 447 460.

1949. The structure of the wall of the test in the Foraminifera; its value in classification. Quarterly Journal
of the Geological Society of London, 104, 229-255.

and haynes, J. r. 1957. Certain smaller British Paleocenc Foraminifera. Part 2, Cibicides and its allies.
Contributions from the Cushman Foundation for Foraminiferal Research , 8, 45-53.
young, r. M. 1985. Darwin’s metaphor. Nature's place in Victorian culture. Cambridge University Press,
Cambridge, xvii + 341 pp.

JOHN R. HAYNES

Department of Geology, Institute of Earth Studies
Typescript received 5 April 1989 University College of Wales, Llandinam Building

Revised typescript received 4 December 1989 Penglais, Aberystwyth Dyfed, SY23 3DB

NOTE ADDED IN PROOF

Since the publication of Loeblich and Tappan’s book in 1988 the following modifications have been made to
their subordinal scheme. First, Loeblich and Tappan (1989) have recognized the new suborder Trochamminina
of Bronniman and Whittaker (1988) and in addition they have reinstated the suborders Haplophragmiina and
Astrorhizina. This gives the following four-fold breakdown for the agglutinated group:

Astrorhizina - unilocular with organic cement
Trochamminina - multilocular, cement organic, wall simple
Flaplophragmiina - multilocular, cement organic wall simple to alveolar
Textulariina - multilocular, cement calcareous, typically canaliculate
Secondly, Revets (1989) has erected the new suborder Delosinina to accommodate the genus Delosina ,
considered to have a unique monolamellar, irregular microgranular, calcitic wall. The addition of this last
would bring the number of suborders in Loeblich and Tappan’s scheme to sixteen.

Further to the above it can now also be stated that the International Committee on Zoological
Nomenclature has decided to take no action in the case of Anomalina and the Anomalinidae and their status
therefore remains unchanged (Whittaker, pers. comm.).

REFERENCES

loeblich, a. r. and tappan, h. 1989. Implications of wall composition and structure in agglutinated
foraminifers. Journal of Paleontology, 63, 769-777.

revets, S. A. 1989. Structure and taxonomy of the genus Delosina Wiesner, 1931 (Protozoa: Foraminiferida).
Bulletin of the British Museum (Natural History) (Zoology), 55, 1-9.

ONTOGENY, HYPOSTOME ATTACHMENT AND
TRILOBSTE CLASSIFICATION

by R. A. FORTEY

Abstract. The high level classification of trilobites has proved particularly difficult. This paper discusses the
classification of those trilobites which have been placed in the Order Ptychopariida, together with other
relevant groups, including Agnostida. Although often considered 'generalized’, the ptychoparioids have a
distinctive derived character: the hypostome is not exoskeletally connected to the cephalic doublure (natant
hypostomal condition). The polarity of the natant hypostome as a derived character is supported both by
known ontogenies, and by comparison with other trilobites, and is unusual in primitive arthropods as a whole.
The wide distribution within the Trilobita of the natant hypostomal condition is established. The primitive
state, found in redlichiids, is conterminant, in which (he hypostome is attached to the cephalic doublure and
closely corresponds ventrally with the frontal glabellar lobe. Several trilobite groups which primitively had
natant hypostomal condition became secondarily conterminant, e.g. among Asaphida and Proetida. All
trilobites having natant hypostomal condition, together with those which were primitively natant and only
secondarily conterminant or impendent, are believed to constitute a new monophyletic group termed the
Subclass Libristoma. This group includes the majority of the trilobites. A number of major monophyletic
groups within the Libristoma are recognized, and one or more autapomorphy for each group is described. The
Ptychopariida can be reduced in scope to a paraphyletic group of primitive libristomates which cannot yet be
assigned to a clade. The other major groups of Libristoma include: Asaphida, Proetida, Olenina, Harpina.
There is slim evidence that the Phacopida also had libristomate ancestry. Some groups included in
Ptychopariida in the 1959 Treatise can now be assigned to different higher taxa. Illaenacea (= Scutelluina) and
Leiostegiacea probably belong within a large corynexochoid clade; Damesellacea are regarded as the primitive
sister group of Odontopleurida. Agnostida are not libristomates, but they are trilobites; they are more
advanced than olenellids, and comprise the sister taxon of Redlichiida + all other trilobites. An important
autapomorphy defining Agnostida is the loss of calcification of an olenelloid-like rostral plate.

Natant hypostomes are conservative in morphology, and it is often difficult to assign the hypostome among
natant families. Specialized hypostomes are associated with the conterminant or impendent hypostomal
condition. The natant condition is lost polyphyletically, and there is some evidence that secondarily
conterminant trilobites arose when primarily conterminant groups became extinct. The relationships of
subgroups within the Libristoma, and the relationships of other major conterminant taxa (corynexochoids,
Odontopleurida, Lichida, Phacopida) to one another and to the Libristoma remain to be explored before the
taxonomic status of major groups can be decided. The Redlichiida is probably another paraphyletic group
having the same relationship to all higher trilobites as does the Ptychopariida within the Libristoma.

Trilobites are one of the most morphologically diverse invertebrate groups with a good fossil
record. Their long stratigraphical history and abundance has meant that students of the group have
specialized in faunas of a particular age, or upon particular groups. It is possible to spend a
productive working life studying one or two families through one or two systems, and no single
worker has a detailed knowledge of the whole Class. Most specialists are concerned with
classification and relationships of trilobites at subfamily, genus or species level, for which there is
now an overwhelming literature. There is no uniformity of taxonomic usage throughout the group,
and it is probably true that taxonomic practice tends to change, with the specialists, at system
boundaries. Cambrian workers tend to be much more influenced by palaeogeography and
stratigraphy than are Ordovician specialists; Devonian and Carboniferous workers tend to employ
more taxonomic categories (subgenera and subspecies) than do workers lower in the column.
Generally these disparities in approach cause little difficulty. However, a more important

| Palaeontology, Vol. 33, Part 3, 1990, pp. 529-576, 1 pl.|

© The Palaeontological Association

530

PALAEONTOLOGY, VOLUME 33

consequence of specialization concerns classification at the taxonomic level above the family, and
the relationships between families. Changeover between families often coincides with stratigraphical
boundaries, and falls between the provinces of different specialists. The latter may take the higher
categories as 'given' and concentrate instead upon the kind of within-group evolution that can be
revealed by detailed stratigraphic studies. The problem is compounded by the fact that families
often appear to be good, clear categories, with distinctive autapomorphies characterizing them,
but the synapomorphies which can relate one family to another are far less clear. Nor have
stratigraphical 'transformation series’ of species connecting one higher taxonomic category with
another been discovered in many cases - the families appear, as it were, ready made. Many of these
natural groups were recognized already by Salter (1864, p. 2). No trilobite taxon is more intractable
than the Order Ptychopariida with regard to the groups that should, or should not be included
within it. Nor is it clear from the literature whether the Ptychopariida is a natural group. This paper
is an attempt to approach some of these problems by discussing characters which are of service in
characterizing ptychoparioid trilobites. In this endeavour I am acutely aware of the limits of my
own detailed knowledge, particularly of Cambrian trilobites. But new insights into high-level
classification are needed before the publication of the revision of the volume on trilobites of the
Treatise on invertebrate paleontology.

HIGH LEVEL TRILOBITE CLASSIFICATION

The few trilobite workers who have attempted to discuss higher level trilobite classification, for
example, Stubblefield (1936), Henningsmoen (1951), Harrington (in Moore 1959), Bergstrom (1973)
and Fortey and Owens (1975), have stressed the need for a natural classification, by which is meant
a classification that reflects phylogeny. Even so, it is apparent that higher level taxa claimed
'natural' by one worker may be disputed by another - see, for example, Bergstrom’s (1977)
discussion of Fortey and Owens’ (1975) Order Proetida, and Fortey and Owens’ (1979) reply. In
dispute in such discussions is the meaning of particular characters: the claimed synapomorphies of
one authority are disputed as homoplasies by another. Trilobites are not simple subjects for
phylogenetic analysis. Cladistic analyses of character distributions offer a new approach to the
problems of classification. A pioneering attempt in this direction was made by Eldredge (1977). But
the cladogram he produced was based on few characters, was replete with unresolved polychotomies,
and many of the groups indicated would now be in dispute. Eldredge admitted that his analysis was
not satisfactory, referring to 'the chaotic and rudimentary state of our understanding of trilobite
relationships’ (p. 326). A very detailed account of one group, the Suborder Asaphina, which
included a cladistic analysis based on a large matrix of characters, was published by Fortey and
Chatterton (1988). This analysis depended on the recognition of at least one character (the ventral
median suture) which served to define the whole asaphine group. Although reasons were advanced
why this character was likely to be a good synapomorphy there may be workers who disagree (i.e.
who claim that the median suture was polyphyletically derived). Higher Asaphina also have a
distinctive kind of protaspis, and this, too, was regarded as important in recognizing the
relationships of some very disparate adults. The foundation of Fortey and Chatterton’s analysis was
the recognition of at least one apomorphic character uniting the Asaphina.

Cladistic classifications also attempt to resolve taxonomic units into monophyletic groups (Wiley
1981). In this respect they may part company with more traditional approaches used by trilobite
taxonomists in which the status of a higher taxonomic unit may be determined by some assessment
of its morphological difference compared with other trilobites, rather than its common ancestry.
The formal recognition of paraphyletic groups (that is, groups descended from a common ancestor,
but not including all its descendants) is also common practice in existing classifications. The writer
regards the definition of groups by common ancestry as a desirable goal in producing a classification
which is truly 'natural’. This is not the place to develop the theoretical basis for this approach, but
I broadly follow the reasoning set out previously (Fortey and Chatterton 1988) and by such authors
as Jefferies (1986). However, for reasons which will become clear below, it is not always possible.

FORTEY: TRILOBITE CLASSIFICATION

531

even if it is desirable, to eliminate paraphyletic taxa completely from groups with a good fossil
record.

When Fortey and Chatterton (1988) treated the Asaphina, they left aside the question of the
relationships of that suborder within the more inclusive Order Ptychopariida. In the Treatise
(Moore 1959) the Ptychopariida occupies well over half the systematic section - as conceived there
it is a vast assemblage of trilobites. Bergstrom (1973) reduced its scope, largely by elevating previous
subgroups of Ptychopariida into separate orders. Of the relationships between these orders there
was little discussion. Moreover, even the constituent families of Asaphina sensu Fortey and
Chatterton (1988) were distributed among three different orders and several different suborders in
Bergstrom’s tabular classification. Nor were the characters on which the classification was based
discussed in a way which is helpful for a phylogenetic classification. For these reasons a more
inclusive view of Ptychopariida is the one adopted here.

The purpose of this paper is to investigate the problems in a phylogenetic definition of
‘ptychoparioid ’ trilobites: do they constitute a monophyletic group, and if so which higher taxa
may be eligible for inclusion within it? I employ the informal term ‘ptychoparioid’ for all the
trilobites that have been considered in this context. My original intention was to tackle the problem
using a computer-based cladistic analysis of probable constituent taxa, based on as many characters
as possible. However, the characters that appear to be important as potential synapomorphies have
either not been recognized, or not discussed in a phylogenetic context, and it is necessary to make
a start by examining these characters. It is the characters which define the scope of the problem by
indicating which taxa should be included in any more detailed review in the future. Although this
paper concerns trilobites which have been regarded, or might prove to be Ptychopariida, it is hardly
possible to avoid mentioning some of the other trilobite orders (Redlichiida, Phacopida,
Corynexochida, etc.). There are problems in these groups also, but to keep the discussion within
sensible bounds the discussion of such problems is briefer than it should be.

A sceptic might contend that, given the difficulties in producing a satisfactory high level trilobite
classification, it might be better to adopt the 1959 Treatise classification (Harrington in Moore 1959)
as a convention and forego attempts to produce a phylogenetic classification, or, like Zhang and Jell
(1987), abandon all categories above family. One could at least ‘pigeon hole' trilobites in this case.
Such a view would quickly become self-defeating because there would be no criteria to decide the
status (family? genus? superfamily?) of the various groups, and would result in a proliferation of
names as more trilobites with intermediate combinations of characters were discovered. Also,
broader palaeobiological theories which may be of interest to the non-specialist require a taxonomic
base of ‘natural’ units. To cite one example, extinction patterns are a current concern, but if the
‘extinction’ is merely one of the disappearance of an arbitrary unit at a boundary between separate
specialists it becomes meaningless (a ‘taxonomic pseudoextinction’ in the usage of Briggs et al.
1988). Conversely, it is of some interest that the extinction event at the Ordovician-Silurian
boundary affects the suborder Asaphina more than any other group - a fact that was not known
until the Asaphina were phylogenetically defined. I firmly believe that the recognition of natural
groups is a priority in trilobite classification, even if the process is likely to prove contentious.

HYPOSTOME ATTACHMENT CONDITIONS

Fortey and Chatterton (1988, text-fig. 5) briefly introduced three terms to describe the relationship
of the hypostome to the cephalic doublure on the venter, and the glabella on the dorsal surface (text-
fig. 1). This character is of importance to the arguments in this paper. Whittington (1988u, h) has
provided timely reviews of hypostomal attachment in the trilobites. He claimed that the anterior
wings of the hypostome correspond with and were connected to the anterior pits in the glabellar
axial furrows (or, if the pits are not clearly developed, to the appropriate point in the axial furrows);
these structures are homologous in all trilobites. There are some differences in terminology between
Whittington (1988r/, h) and Fortey and Chatterton (1988) but only minor discrepancies in

532

PALAEONTOLOGY. VOLUME 33

text-fig. I . The three conditions of hypostome attachment illustrated by cephalic shields viewed from the
underside. The explanatory cartoons below show cephalic doublure as coarse stipple, the outline of the glabella
in fine stipple, and the hypostomal outline transparent to make clear the relationship between the hypostomal
anterior margin and the preglabellar furrow, a, natant hypostomal condition, with ‘detached’ hypostome,
illustrated by Elrathia kingi , Middle Cambrian, x5; b , conterminant hypostomal condition, hypostome
attached and closely correspondent with glabella outline, illustrated by Gog catillus , early Ordovician (after
Fortey 1975), x 1 ; c, impendent hypostomal condition, with no close relationship between anterior margin of
glabella and that of the hypostome (see also text-fig. 5 a-c), Raymondaspis reticulata , mid Ordovician (after

Whittington 1965), x 3.

interpretation. However, a greater emphasis is placed here on the relationships of the hypostome
to other dorsal structures.

Natant hypostomal condition

This term (text-fig. la) was applied by Fortey and Chatterton ( 1988) to those trilobites in which the
hypostome was not attached to the cephalic doublure, but lay beneath the front of the glabella,
separated from the doublure by a gap. Whittington (1988a) used the term ‘detached’ to describe the
same condition. Because of this absence of firm connection it is unusual to find trilobites with natant
hypostomal condition with the hypostome in place. One relies upon individuals that died on their
backs, the hypostome falling down into the forward part of the glabella (Rasetti 1952a) upon the
decay of the soft tissue (text-fig. 2). Despite this, hypostomes of a wide range of families are known
or inferred to have had natant hypostomal condition, as listed below. The existence of the same
hypostomal condition can be inferred for many more trilobites of which complete remains are
unknown. This is when the genal doublure (or by extension the rostral plate) corresponded in width

text-fig. 2. Natant hypostome in life position (a), and as it falls down into the glabella on death in those
trilobites which died on their backs ( b ) illustrated by generalized ptychoparioid. Compare with Plate 1. Note
how glabellar furrows occupy area to each side of the hypostome.

FORTEY: TRILOBITE CLASSIFICATION

533

with the marginal rim (PI. 1, fig. 2). If the marginal rim was separated from the front of the glabella
by a preglabellar field, and given that the hypostome lay beneath the frontal part of the glabella (in
the position consistent with the anterior wing/anterior pit correspondance), then the hypostomal
condition was natant.

Those hypostomes which were not natant were rigidly attached to the cephalic doublure
(Whittington 1988m b). It is more difficult to assess the degree to which the natant hypostome was
fixed in place by the anterior hypostomal wings. The fact that the hypostomes in life position are
so readily displaced implies that the connection was not likely to have been particularly rigid. Also
I have seen numerous examples of entire cephalic shields which should retain a natant hypostome.
but of which there is now no trace (PI. 1, fig. 2), implying that it was easily separated. Dr R. M.
Owens informs me (pers. comm. 1987) that of the many proetide exoskeletons he has studied with
natant hypostomal condition when the hypostome is found at all it is invariably slumped to one side
in the forward part of the glabella, in the manner of text-fig. 2b, implying that its support was not
rigid (a similar conclusion was reached by Whittington and Campbell 1967). For these reasons it
is likely that the connection between anterior wings and anterior fossulae was ligamental. and that,
in life, there was probably a degree of freedom for movement of the hypostome. For example, up
and down movement of the adoral margin might have been possible by means of muscles attached
to the middle furrows on the hypostomal middle body, using the ligamental attachment of the
anterior wings as ‘rockers'.

Natant (swimming) is an appropriate term to describe this hypostomal condition, even though the
labral plate was probably tethered or anchored in at least some cases rather than truly free-lying and
supported only by the ventral membrane. But some ptychoparioid hypostomes have such reduced
anterior wings (e.g. Bienvillia tetragonalis broeggeri , see Henningsmoen 1957, pi. 1 1, fig. 7) that it
is possible to be sceptical whether they were attached at all. The term natant is preferred to
Whittington’s ‘detached' for two reasons: (a) ‘detached’ has a vernacular use which can apply to
fortuitously separated hypostomes with other kinds of attachment modes; (b) if the natant
hypostome is homologous in various trilobites, and is an unusual character in arthropods as a whole
(below) then it requires a diagnostic epithet to describe it. Note that there is no direct connection
between a trilobite having natant hypostomal condition and the configuration of the ventral sutures,
so that natant hypostomes can be found with rostral plate (Ptychopariacea), median suture (e.g.
Pterocephaliidae) or fused cheeks (most Olenidae).

Orientation. The connection between hypostomal anterior wings and anterior fossulae or pits in the
axial furrows (Whittington 1988«) positions the hypostome in a way which corresponds with the
anterior part of the glabella on the dorsal surface. This conforms with the position in which ‘in
place' natant hypostomes are found (see text-fig. 2 and PI. 1). The anterior pits are in the axial
furrows in front of the eye ridges, which in ptychoparioids are directed towards the anterolateral
corners of the glabella. Because the anterior outline of the hypostome curves forwards, in the life
orientation this profile approximates closely to the front glabellar margin, i.e. the outline of the
preglabellar furrow. The outline of the anterior margin of the hypostome closely corresponds with that
of the glabella, and lies a short distance ventrally below it. If Whittington’s view of the upward
projection of the anterior wings is correct (e.g. PI. 1, fig. 5) there are some forms in which the front
margin of the hypostome cannot have been far removed from the preglabellar furrow. The close
correspondence dorso-ventrally between hypostome and glabella is an important point because it
is retained in all trilobites except those in which the hypostomal condition is impendent.

With the hypostome restored in this position, in the typical ptychoparioid ( Ptychoparia , Elrathia
and the like) it is probable that the distribution of the glabellar furrows is related to the position
of the hypostome (text-fig. 2a). The longest, posterior IS furrows are posterior to the hypostome;
the progressively shorter three anterior pairs of furrows occupy the gap between the hypostome and
the edge of the glabella.

534

PALAEONTOLOGY, VOLUME 33

Conterminant hypostomal condition

In the conterminant hypostomal condition (text-figs. 1 b, 3) the hypostome is in the same relative
position with regard to the glabella as in the natant condition, i.e. the front margin of the hypostome
and that of the glabella closely correspond. However, the hypostome is suturally joined to the inner
edge of the doublure, and in all the described cases (Fortey and Chatterton 1988; Whittington
1988a, b ) this attachment was rigid. Even in the absence of articulated material the conterminant

text-fig. 3. Conterminant hypostomal condition of various types, illustrated by ventral views of cephalic
shields with hypostomes in place, a, the primitive calymenid Neseuretus (early Ordovician), in which
conterminant condition is maintained by a broad anterior flange on the hypostome (mostly after Henry 1980),
x 2; b, secondary conterminant hypostomal condition with backward curvature of doublure in Proceratopyge
(Asaphida; Ceratopygidae : late Cambrian; after Jago, 1987), x 3; c, primary conterminant hypostomal
condition in which rostral plate and hypostome are fused together as a single unit in Fieldaspis (Corynexochida :
mid Cambrian; mostly after Whittington 1988a) x3; rf, primary conterminant hypostomal condition with
narrow, transverse rostral plate, in the odontopleurid Leonaspis (Silurian; after Chatterton and Perry 1983),
x8; e, secondary conterminant hypostomal condition with hypostome recessed in doublure and occupying
large part of venter, Isotelus (Asaphidae: middle Ordovician) x 1.

EXPLANATION OF PLATE 1

Natant hypostomal condition, including entire exoskeletons with hypostome more-or-less in life position (see
text-fig. 2).

Figs. 1, 5, Olenidae: Hypermecaspis. 1, dorsal exoskeleton lying on its back within nodule, early Ordovician,
Bolivia, USNM 380856, x 1 -5. 5, hypostome from lateral view to show upward-directed anterior wings, early
Ordovician, Spitsbergen, Sedgwick Museum (SM) A84079 (see also Fortey 1974, pi. 14, fig. 10), x4.

Figs. 2, 8. Asaphiscidae : Blainia , two exoskeletons from same block, mid Cambrian, Alabama, USNM 62801,
one showing narrow (sag.) rostral plate but lacking hypostome, the other showing hypostome in usual
position in natant forms, x4.

Fig. 3, Cedariinae: Cedaria, showing rostral plate and position of hypostome, U. Cambrian, Utah, USNM 300,
x 4.

Fig. 4, late olenid hypostome, Balnibarbi, early Ordovician, Spitsbergen, SM A84021, x 5.

Fig. 6, Alokistocaridae : Elrathia, mid Cambrian, Utah, preserved dorsal side up but clearly showing rostral
plate and position of hypostome, BM (NH) It 1734, x 2.

Fig. 7, Proetidae; Proetus , Silurian, Wenlock, SM A28263 (Owens 1973, pi. 3, fig. 9), x 5.

PLATE 1

FORTEY, natant hypostomes

536

PALAEONTOLOGY, VOLUME 33

hypostomal condition can be inferred if the inner margin of the doublure extends as far as the front
of the glabella, but no further (text-fig. 1 b). Conterminant hypostomal condition can be found in
trilobites with rostral plate, median suture, or with the cheeks conjoined ventrally. The hypostomal
suture may become obsolete in trilobites with conterminant hypostomal condition, so that the
hypostome and rostral plate form a single unit, as in Corynexochida (text-fig. 3c) and some
Paradoxididae. Conterminant hypostomal condition is found in numerous trilobite families, of
which several examples were described by Whittington (1988a, b) or by Fortey and Chatterton
(1988): Dalmanitidae, Asaphidae, Remopleurididae, Ceratopygidae, Cheiruridae, Calymenidae,
Corynexochidae, Lichidae, Odontopleuridae, Encrinuridae.

1 . Secondary conterminant hypostomal condition. Conterminant hypostomal condition may develop
secondarily from the natant condition. This is shown by examples within accepted monophyletic
groups: Anomocaridae-Asaphacea; Remopleuridacea and Proetida. This secondary development is
proved from stratigraphic evidence which shows that early and primitive members of the group are
natant, and that later, more derived members are conterminant; and from comparison with the
respective out-groups, in which the natant condition also pertains. These examples are important
because they show: (a) that the conterminant condition is certainly not primitive for some Trilobita,
and (b) that the conterminant condition can be attained from the natant by two different
mechanisms (text-fig. 4). There is no known case where the conterminant condition returns to the
natant condition - and hence it becomes a good derived character in classification.

Anomocaridae-Asaphacea. The relationships between the families included in the Asaphacea were
reviewed by Fortey and Chatterton (1988). Most Asaphacea, in which Fortey and Chatterton
included Ceratopygidae as well as Asaphidae, had a conterminant hypostomal condition. The
primitive sister group, which probably included the common ancestor of both families, was the
paraphyletic group Anomocaridae. Anomocarids certainly included some genera, such as
Anomocarioides , which had natant hypostomal condition, the inner margin of the cephalic doublure
falling well short of the front of the glabella. In fact, Fortey and Chatterton could find no member
of the Anomocaracea which was convincingly conterminant. However, some species placed within
this group had relatively wide cephalic doublure, so that the 'gap' between its anteromedian margin
and the front of the hypostome was reduced. In early Ceratopygidae (Proceratopyge spp.) it was
shown that the hypostome was attached to the doublure, and this was achieved by a median
widening of the doublure (Jago 1987). This widening was indicated on the dorsal surface by adaxial
backward curvature of the paradoublural line(s). In Asaphidae the doublure is wide both beneath
free cheeks, and under the (usually wide) preglabellar area, before terminating on a line coincident
with the preglabellar furrow (text-fig. 1 b). The relatively great width of the cephalic doublure
possible in this group suggests that the preglabellar field was ‘lost’ by backward extension of the
doublure beyond the marginal rim at which it terminated in the ptychoparioids. In some
conterminant asaphids doublure and frontal area are again narrow, but stratigraphic evidence
suggests that this narrowness may be secondary. In any case the anomocarid-asaphacean transition
(text-fig. 4a) involves secondary docking of the hypostome with the doublure - which may become
very wide in the process - and attendant loss of the natant hypostomal condition. A few advanced
asaphaceans attain the impendent hypostomal condition.

Remopleuridacea and Dikelocephalacea. These two superfamilies were regarded as closely related
within the Asaphina by Fortey and Chatterton (1988), and Dikelocephalacea is used in the sense
of Fudvigsen and Westrop (1983) to include Ptychaspidacea of authors. Early Remopleuridacea,
such as kainellids and Elkanaspis Fudvigsen, 1982 have a narrow marginal cephalic rim, and the
doublure ventrally corresponds with the rim. Species having a wide preglabellar field are therefore
likely to have had natant hypostomal condition (Fortey and Chatterton 1988, text-fig. 14). In later
and more advanced Remopleuridacea, especially Apatokephalus and its allies, the glabella extends
forwards as far as the border, and the width of the cephalic doublure medially (e.g. Ross 1951, pi.

FORTEY: TRILOBITE CLASSIFICATION

537

backward extension
of doublure

CONTERMINANT

loss of preglabellar
field

text-fig. 4. Two possible mechanisms (above) for attaining secondary conterminant hypostomal condition
from natant hypostomal condition illustrated by generalized ptychoparioid of elrathiid type. Two examples
(below) of secondary conterminant condition, primitive natant morphology to left and position of hypostome
shown as dotted outline: a, anomocaracean Litocephalus (after Palmer 1960) (left) and primitive asaphid
Promegalaspides (right); b, remopleuridaceans Elkanaspis (left) and Menoparia (right) (after Fortey and

Chatterton 1988).

20, fig. 30) indicates that the hypostome had docked at this stage to become conterminant (text-fig.
4b). In the most advanced members of the family Remopleurididae the glabella has encroached
forwards still further so that the hypostomal condition becomes impendent, and the rigid
hypostomal attachment of such forms has been familiar for some time through the work of
Whittington (1959). Remopleuridacea thus afford a second example of secondary conterminant
condition having been attained from a natant condition. Unlike at least some of the Asaphacea, the
achievement of conterminant hypostomal condition was made by elimination of the preglabellar
field by forward encroachment of the glabella the cephalic doublure remains coincident with the

538

PALAEONTOLOGY. VOLUME 33

cephalic rim as it is in ptychoparioids. Even though the conterminant hypostomal condition is
functionally the same as it is in asaphaceans it seems possible that the means of achieving it may
be different.

Less is known of ventral structures in Dikelocephalacea. Conterminant hypostomal condition
certainly pertained in saukiids such as Saukia , and also in ptychaspids with a narrow tube-like
cephalic doublure, and in dikelocephalids, such as Dikelocephalus , with a wide and flat doublure.
Idahoiidae may prove to be the sister group of Dikelocephalacea, and idahoiid genera such as
Wilbernia and Idahoia show wide borders, but which do not reach the front of the glabella; these
were in all probability still natant.

Proetido. The Order Proetida Fortey and Owens, 1975 includes numerous post-Cambrian families,
and was the only trilobite group to survive the Devonian. I shall return to the characters defining
it below. Early Proetida, and perhaps the majority of the Order, have natant hypostomal condition.
The cephalic doublure closely corresponds with the cephalic rim, the two together forming a tube-
like structure in many proetides. Natant hypostomal condition is accordingly obvious for Proetacea
having narrow borders and wide preglabellar fields, such as the primitive genus Decoroproetus
(Owens 1973, pi. 7, fig. 15), and the majority of Aulacopleuridae. On more advanced genera with
reduced preglabellar field, such as Proetus itself, it is not immediately clear whether the hypostome
is or is not attached to the doublure, but specimens preserved ‘on their backs’ (e.g. Proetus
concinnus (Dalman), see Owens 1973, pi. 3, fig. 9; PI. 1, fig. 6 herein) show the hypostome in the
fallen position typical of other taxa with natant hypostomal condition. The same lack of attachment
of the hypostome probably persisted into the Carboniferous in such genera as Namuropyge , which
has a preglabellar field, and certain Cummingella species (e.g. Woodward 1883, pi. 1, fig. 4a).
Elowever, some Proetida did attain secondary hypostomal attachment, as shown in text-figure 5.
For example, the Carboniferous genus Paladin (Whittington 1 988/?, text-fig. 12) has the glabella
encroaching on the cephalic border, and the hypostome is attached to the doublure in the typical
conterminant fashion (also text-fig. 5 <7); the same probably applied to many other Carboniferous
phillipsiids. Campbell (1977; also Whittington 1988/?) showed that the cephalic doublure on the
Devonian brachymetopid genus Cordania was relatively wide - wider than the ‘border’ -and
terminated precisely on a line around the front of the preglabellar furrow. If it was in its usual
position co-incident with the glabellar frontal lobe it is not unreasonable to assume that the
hypostome was conterminant in this case also, although there are no examples known to me of
brachymetopids with hypostome in situ. Campbell (1977, p. 20) thought that the ‘border’ furrow
in this case was not homologous with that in other Proetida (the contrary opinion was held, for
example, by Owens and Thomas 1975), which he regarded as really coincident with the inner
doublural margin. If this view is correct then loss of the natant hypostomal condition would be the
result of encroachment of the glabella on the border as in Paladin and remopleuridids. However,
if the other view is held - that the doublure has ‘grown beyond' the true border - then the loss of
the natant condition was the result of the doublure growing out to meet the hypostomal margin,
a case like the Asaphacea described above (and text-fig. 3, top). Whatever the correct interpretation
of border homology, these proetaceans evidently show the secondary acquisition of conterminant
hypostomal condition by analogous mechanisms to those operating in Asaphina.

Non-proetacean Proetida include several families in Fortey and Owens’ (1975) concept. Their
original scenario was that these were ultimately derived from the subfamily Hystricurinae, a
typically ‘ptychoparioid ’ group with narrow cephalic rims and doublures (e.g. Ross 1951, pi. 8, figs.
7 and 8) and wide enough preglabellar fields to be likely to have had natant hypostomal condition.
Early Bathyuridae ( Peltabellia , see Ross 1951) were generally like hystricurids in these features and
there is no reason to suppose that they were other than natant. However, some advanced forms with
reduced preglabellar fields, such as Bathyurus itself, have been claimed as having rostrum and
hypostome in contact (Ludvigsen 1979); Whittington (1988/?) also considered that the hypostome
was firmly braced in Bathyurus. Whittington (1963, pi. 11, fig. 6) figured the cephalic doublure of
the bathyurelline Punka nitida (Billings) which is wider than in hystricurids but does not extend

FORTEY: TRILOBITE CLASSIFICATION

539

c d

text-fig. 5. Secondarily attached hypostonres in Proetida. a , b , d, impendent hypostomal condition in
Griffithides acanlhiceps Woodward; Dinantian, Treak Cliff, Castleton, Derbyshire: a , dorsal, and b , ventral
views, x 6; d, oblique ventral view, x 5. Note anterior pits in axial furrows well posterior to usual position in
natant forms. National Museum of Wales (NMW) 86.25G.2332. d , Secondary conterminant hypostomal
condition in Archegone ( Phillibole ) aff. aprathensis Richter and Richter, Dinantian, slopes of Pendle Hill,
Lancashire, x 3, NMW 88.36G.10. Photographs kindly supplied by R. M. Owens.

across the preglabellar field. Ross ( 1953) described the silicified hypostome of another bathyurelline,
Licnocephala cavigladia Hintze, which shows remarkably long prong-like anterior wings and a poor
match between anterior profile of hypostome and posterior margin of cephalic doublure. It seems
then, that in Bathyuridae secondary conterminant condition occurred only rarely, and only in the
stratigraphically youngest members of the family. Nor are there any described examples of
Dimeropygidae or Glaphuridae in which conterminant hypostomal condition is suspected.

To summarize, Proetida primitively have natant hypostomal condition, and some families remain
so. Several other proetide families include advanced members in which the hypostome is secondarily
conterminant; this happens by the loss of the preglabellar field by encroachment of the glabella on
the border, or, possibly, by widening of the doublure to cover the preglabellar field. The former
happens also in the Remopleuridacea; the latter in the anomocarid asaphid - ceratopygid
transition. In no case is there a reversal from conterminant back to natant.

2. Primary conterminant hypostomal condition. In the cases just described the conterminant
hypostomal condition is derived, within accepted monophyletic groups, from the natant condition.
Although these trilobites provide the theme of this paper, it is important to recognize that there are

540

PALAEONTOLOGY, VOLUME 33

other trilobite groups with conterminant hypostomal condition in which it is not derived from the
natant condition; indeed, it is considered to be the primitive mode of hypostomal attachment for
the Trilobita as a whole.

The primitive state: Redlichiida. The Redlichiida are regarded as the sister taxon of all ‘higher’
trilobites, which excludes only the Agnostida and olenelloids (see text-fig. 14), and includes all
ptychoparioids. Redlichiids have functional dorsal ecdysial sutures and have a rostral plate
bounded by connective sutures. Following Bergstrom (1973) and Fortey and Whittington (1989),
olenelloids are regarded as primitive trilobites lacking dorsal ecdysial sutures, a view different from
that of Lauterbach (1980), who regarded certain olenelloids as more closely related to chelicerates.
Ax (1987) summarized Lauterbach's views in English. Here it may be stated that the greater number
of derived characters favour the trilobite hypothesis of olenelloid relationships, if the characters
identified as synapomorphies by Fortey and Whittington ( 1989) are correct. The systematic position
of the Agnostida is discussed below. If the Redlichiida is correctly identified as the sister group of
higher trilobites the redlichiid mode of hypostomal attachment (text-fig. 6) is the primitive one for
such trilobites.

text-fig. 6. Conterminant condition as the primitive hypostomal attachment mode as exemplified in
Redlichiida, the sister taxon of1 higher’ trilobites. Redlichia longtangensis Zhang and Lin (see Zhang et at. 1980,
text-fig. 53), late Lower Cambrian, Longtang, E. Guizhou, China, showing extension of cephalic doublure and
position of hypostome, x 15. Nanjing Institute of Geology and Palaeontology 7057. Photograph kindly

supplied by W. T. Chang (Zhang Wentang).

Lauterbach (1980) regarded the Emuellidae as the primitive sister group of redlichiids. Pocock
(1970, text-fig. 4) clearly described the nature of the attachment of the hypostome in Emuella. It is
conterminant, the anterior edge corresponding both with the front of the glabella dorsally, and the
back end of the relatively narrow (tr.) rostral plate. In Redlichia (Opik 1958; Zhang et al. 1980;
Whittington 1988c/) the hypostomal attachment is similar, that is, conterminant (also Resserops , see
Hupe 1953, fig. 39). Opik (1958) showed that the rostral plate and hypostome are fused together in
R. idoneci Whitehouse, whereas in Emuella ‘the hypostomal suture is functional’ (Pocock 1970, p.
528). Of numerous Redlichia species figured by Zhang et al. (1980) some (e.g. pi. 20, fig. 5) have

FORTEY: TRILOBITE CLASSIFICATION

541

hypostomes without attached rostrum, as does Sardoredlichia (Rasetti 1972, pi. 9, fig. 19), and it
seems likely that the fusion of rostrum and hypostome was an advanced, not a primitive character.
Additional evidence for this can be cited in that such secondary fusion happens again, later, in some
species only of Paradoxides (Whittington 1988a), and in the Corynexochida. The important point
is that the primitive mode of attachment in the group Redlichiida + all other non-olenellid trilobites
is conterminant. This can be referred to as primary conterminant hypostomal condition. This is not
homologous with the secondary conterminant hypostomal condition, described above, in which it
is derived from the natant hypostomal condition.

Certain redlichiids develop a short preglabellar field: see, for example, Redlichia yichangensis
Zhang and Lin (in Zhang et cd. 1980) and R. ( Pteroredlichia ) chinensis Walcott (Zhang et al. 1980,
pi. 25, fig. 5). This does not mean however, that the hypostome became detached from the rostrum
and natant. Zhang et a/.'s pi. 20, fig. 9 and pi. 25, fig. 25 clearly show the hypostome in place beneath
the frontal glabellar lobe, and the cephalon in the latter also shows that the posterior part of the
rostral plate extends backwards (see text-figure 6) across the ‘gap’ between marginal rim and
preglabellar furrow. Conterminant hypostomal condition is retained. Many redlichiids for which
ventral structures are not known show a median plectrum in the preglabellar field (e.g. Zhang et cd.
1980, pi. 21, fig. 7), and it is my contention that this is the dorsal expression of such a backward
extension from the rostral plate. A plectrum is a median backward projection of the border, and
usually distinctly defined; however, other redlichiids show a vaguer median depression in the same
region, which may still represent the same ventral structure.

Incidentally, whether or not olenellids were trilobites, they do appear to show some comparable
ventral structures to redlichiines. Some, such as Holmia , were clearly conterminant, with close
connection between doublure and hypostome (Whittington 1988a). Others, such as Olenellus , show
narrow spine-like connections across the preglabellar field, and median plectra dorsally.

Other probable examples of primary conterminant hypostomal condition. If the conterminant
hypostomal condition is primitive for the higher trilobites as a whole it is not surprising that it is
retained in many families. Some of these were reviewed by Whittington (1988a, b). In some groups
the conterminant condition intergrades with the impendent condition discussed in the following
section. In many trilobite families the conterminant condition is constant, and retained during
modifications to the border. For example, in the calymenid Neseuretus (text-fig. 3a) the cephalic
border is exceptionally extended for the family and the conterminant hypostomal condition is
retained by virtue of an anterior plate-like extension of the hypostomal margin which extends
beneath the area between the front of the glabella and the anterior cephalic rim. There is no a priori
reason why Neseuretus should not have become natant - i.e. the hypostome simply ‘drifted away’
from the rostrum as the border extended - but the fact that this does not happen proves the
conservatism of the hypostomal attachment mode.

A list of non-redlichiid higher taxa having primary conterminant hypostomal condition includes
the following, most of which I have confirmed through examination of specimens in the British
Museum (Natural History) collections. This list is not hierarchically arranged, and those taxa
discussed above under secondary conterminant hypostomal attachment are not included.

Order Corynexochida: Most of this group are conterminant (text-fig. 3c); a few advanced forms may
possibly have been impendent. Loss of hypostomal suture by ankylosis may make the anterior limits of the
hypostome difficult to see in some cases.

Superfamily Leiostegiacea : Apparently all are conterminant, but there are no described examples with
hypostome in place. The doublure-hypostome relationship is fairly obvious in those genera, like Leiostegium
itself (text-fig. 17a, p. 564), with distinct anterior pits far forwards in the axial furrows with which the anterior
wings of the hypostome would have engaged. Broad-bordered genera like Palocorona Shergold, 1980 develop
a ‘false border furrow’ but the course of the inner margin of the doublure is still revealed by the paradoublural
line extending to the front of the glabella.

Superfamily Damesellacea : These have not been fully described from the ventral side, but examination of
Stephanocare and Blackwelderia specimens in the B.M. (N.H.) shows that the cephalic doublure is closely

542

PALAEONTOLOGY, VOLUME 33

coincident with the cephalic border, which is narrow and ledge-like in front of the glabella, a structure very
like that of the Leiostegiacea. The hypostomes of several genera are known, and assuming the correspondence
of anterior wings with the anteriorly placed pits in the axial furrows, it seems likely that the Damesellacea were
also conterminant.

Superfamily Cheiruracea: Many entire specimens of cheiruraceans have been described (Lane 1971;
Whittington 1988/?; Jell 1985, pi. 13, fig. 10); all Pliomeridae and Cheiruridae are conterminant, except where
the frontal lobe of the glabella has expanded forwards (especially Encrinuridae; some Cheiruridae) and
incorporated the border within it, and these may perhaps be better described as impendent.

Superfamily Calymenacea : Advanced calymenids are consistently conterminant, even when, as in Neseuretus
(above), the preglabellar area is long. The extended preglabellar area in this case is ‘covered’ by an extension
of the hypostome, but this is unusual. Homalonotids with extended (sag.) borders, and such calymenids as
Calymenesun (Derek J. Siveter pers. comm. 1986) cover the extended area by growth of the rostral plate and
the adjacent genal doublure. The hypostome remains in its usual position. So far as can be ascertained primitive
calymenaceans with preglabellar field were also conterminant. For example Prionocheilus (= Pharostoma)
appears to show a backward extension of the rostral plate sufficient to bridge the distance between anterior
border and the front of glabella and hypostome (Whittard 1960, pi. 18, fig. 4).

Suborder Phacopina: Jaanusson (1975) discussed the morphological steps leading to the suborder
Phacopina, and described Gyrometopus as a primitive representative. This genus had a rather narrow cephalic
doublure, but the hypostome is not known and it is difficult to say whether it was or was not conterminant.
The more primitive Phacopina are classified within the Dalmanitidae. Many of the stratigraphically early
dalmanitids described by Henry (1980) show details of the ventral cephalic surface, including hypostomes in
place (Henry 1980, pi. 28, fig. 4; pi. 29, fig. 4). These were probably conterminant. Early Pterygometopidae,
such as Calyptaulax (e.g. Chatterton and Ludvigsen 1976, pi. 16) are apparently slightly impendent.
Phacopidae are impendent, even early genera such as Morgatia (Henry 1980, pi. 43, fig. 6).

Order Odontopleurida : Odontopleurids (text-fig. 3d) are known from numerous silicified species; they are
all conterminant (see, for example, cephalic shields of several species figured by Chatterton and Perry 1983).
Both cephalic border and doublure are narrow, and the hypostome is rigidly attached (Whittington 1988/?).

Order Lichida. Thomas and Holloway’s (1988) recent review of this group shows conterminant condition
to be the rule.

This brief review shows that, although the conterminant condition is primitive for the higher
Trilobita, it is present, and consistently so, within a variety of higher taxa extending to the later
history of the group. While it is likely that in mid-Cambrian groups this is a retained plesiomorphic
character, there is at least the possibility among later groups such as the Calymenacea that it is
secondarily derived from the natant condition, as in Asaphina and Proetida, and if so this would
have considerable taxonomic implications. I return to this topic below.

Impendent hypostomal condition

In the impendent hypostomal condition (text-figs, lc, 5 a-c) the close relationship between the
position of the hypostome and the front of the glabella which pertains in natant or conterminant
hypostomal condition is lost. The term was coined by Fortey and Chatterton (1988) with reference
to Cyclopygacea, but applies to many other groups. Its taxonomic importance is proved by the fact
that there is no known example where impendent hypostomal condition reverts to conterminant or
natant. The impendent condition results from the relative forward growth and often inflation of the
glabella to engulf the cranidial border; in the process the anterior glabellar margin more nearly
coincides with the cephalic margin, and the hypostomal suture occupies a relatively posterior
position. The relationship between the anterior wing of the hypostome and its connection in the
axial furrow is not lost. This is shown by the hypostomes of Nileus and Remopleurides , for example,
in which the anterior hypostomal wings become elongated in the dorsal direction (e.g. Whittington
1965, pi. 31, figs. 2 and 3; 1988/?, text-figs. 5 and 6). Impendent hypostomal condition can
characterize higher level taxa (Cyclopygacea, Illaenina, Phacopidae), and is independent of the
development of the ventral sutures, occurring with rostral plate ( IUaenus ), median suture
( Amphytrion ) or fused cheeks ( Symphysurus , Phacops).

In those groups which attained secondary conterminant hypostomal attachment a number of
subgroups went further to become impendent (Cyclopygacea, some Asaphinae, some phillipsiids

FORTEY: TRILOBITE CLASSIFICATION

543

such as Griffithides and Hentigia, see Whittington 1988/)). It is clear in these that there is a
morphological and phylogenetic progression running natant secondary conterminant impend-
ent. Impendent hypostomal condition may have been capable of developing directly from primary
conterminant condition. But in, for example, Corynexochida ( Corynexochus ) having the glabella
extending far forwards it is likely that the rostral plate became extremely narrow (sag.) and
impendent attachment was not attained.

A detailed discussion of the impendent condition is not as important to the phylogenetics of
Ptychopariida as is the conterminant condition, but it is important in the wider context of trilobite
classification. In addition to those Asaphina mentioned already several large groups are typified by
having impendent hypostomal condition :

Superfamily Illaenacea: as restricted to Illaenidae plus Scutelluidae, and excluding groups placed in Proetida
by Fortey and Owens (1975). Even early members of this group such as Raymondaspis (Whittington 1965, pi.
57, fig. 9) appear to have the glabella extending well beyond the inner doublural margin.

Advanced Phacopida: The Phacopidae, Calmoniidae and Monorakidae are all impendent. The first two of
these groups had their origins in the Ordovician, and presumably had sister groups within what would at
present be classified as dalmanitaceans having conterminant hypostomal condition.

Late Proetida: A number of stratigraphically late genera of Proetida have glabellas that expand forwards
to the cephalic margin - much in the manner of Nileidae in the Asaphina. Examples would include the Permian
genus Paraphillipsia (see Owens 1983), and Griffithides (text-fig. 5 a-c). Because such forms had cephalic
doublures of normal width they would have had impendent hypostomal condition also. They must have been
derived from one of the secondarily conterminant proetides.

One example will suffice to show that the discrimination of the different modes of hypostomal
attachment can be of immediate taxonomic use (text-fig. 7). Bornemann (1883) described the
trilobite Illaenus meneghinii from Sardinia, which subsecpiently became the type species of

text-fig. 7. Hypostomal attachment condition as a key to revealing homoeomorphy. a and b, dorsal cephalic
and pygidial morphology of the early Ordovician genus Platypeltoides , and the early Cambrian genus
Giordanella (after Rasetti 1972) showing general similarity, c. x 3. Below, inferred ventral views of the same.
c, cartoons illustrating hypostomal attachment mode using same conventions as in text-figure 1, Giordanella
above, probably conterminant, Platypeltoides below, impendent.

544

PALAEONTOLOGY, VOLUME 33

Giordanella. Because of the general resemblance of Giordanella to early Ordovician Nileidae, the
beds yielding it came to be regarded as transitional between Cambrian and Ordovician. Later it
became apparent that the horizon was, in fact. Lower Cambrian, and the time gap between this
horizon and the early Ordovician was so considerable that Nicosia and Rasetti (1971) rejected nileid
affities in favour of incertae familiae . Rasetti (1972, p. 44) went further in using Giordanella as a
good example of how homoeomorphy confuses and, by implication, as part of his argument on how
misleading the morphology of Cambrian trilobites can be ‘without taking into consideration the age
and geographic province’. In the present context, it is clear that, effaced dorsally though it is, the
cranidium of Giordanella can show the anterior outline of the glabella (e.g. Rasetti 1972, pi. 10, figs.
3, 9) well inside the cranidial margin. The Family Nileidae, and the superfamily Cyclopygacea to
which it belongs, are characterized as having impendent hypostomal condition, and as a
consequence the glabella dorsally extends to the cephalic margin. There is thus an easy way to show
that Giordanella and, for example, the nileid Platypeltoides , a genus it generally resembles dorsally
(text-fig. 7), are not related. If one adds to this the fact that Giordanella has a wide rostral plate
(Rasetti 1972, pi. 10, fig. 22) whereas the Nileidae either have a median suture or fused cheeks, then
the likelihood of a close taxonomic relationship between nileids and Giordanella becomes remote.
Giordanella probably had primary conterminant hypostomal condition, if we are to judge from the
width of the doublure and the indication of a paradoublural line running to the front of the glabella.
This suggests a comparison with the family Ellipsocephalidae, which is widespread and familiar in
the Lower Cambrian. Giordanella would then become just another example of a relatively
macropygous and effaced trilobite, which is known to be one of the most iterative (and therefore
of least taxonomic moment) trends in trilobite evolution (Lane and Thomas 1983). One might
perhaps turn Rasetti’s argument on its head and use this example to demonstrate that it is necessary
to look at characters critically first before jumping to stratigraphical conclusions based on general
resemblance.

NATANT HYPOSTOMAL CONDITION AS A SHARED DERIVED CHARACTER

It was concluded above that the primitive hypostomal attachment mode is conterminant. If so,
the natant condition is a derived character and hence a synapomorphy of potential use in defining
a monophyletic group. Opik (1963, p. 77) has already stated that the natant hypostomal condition
is of importance in higher level trilobite classification. There are a few criteria which can be used
to establish the polarity of a character shift (Schoch 1986, pp. 1 34-142) - in this case, that the natant
condition is a derived character. The comparison with Redlichia described above exemplifies the
criterion of comparison with the out-group. There is also the criterion of ontogeny: during
morphogenesis the developmental sequence should parallel the inferred polarity (conterminant to
natant in this case). The stratigraphic criterion suggests that the primitive condition will also be
stratigraphically earlier. Whittington ( 1 988<r/) has shown that the natant condition (his ‘detached’)
pertained in the type species of Ptychoparia - and so it will be, by definition, a ptychoparioid
character. In order to establish how unusual the natant hypostome is as a derived character it is also
worth looking at a wider range of arthropod out groups. If unusual, then it is reasonable to attribute
to it some ‘weight’ in classification.

How unusual a character is the natant hypostome?

The relationship of the trilobites to the other Arthropoda is not resolved (Cisne 1974; Hessler and
Newman 1975; Lauterbach 1980; Schram 1986). Most authors would agree that they are part of
a clade that includes both crustaceans and arachnates. A preliminary cladistic analysis by Briggs
and Fortey (1989) indicated that the trilobites were advanced relative to crustacean groups.
Antennae and biramous trunk limbs were primitive attributes of the crustacean/arachnate clade.
There is no occasion here to discuss the different theories of trilobite relationships but it is worth
considering the Crustacea as an out-group for hypostomal attachment. The homologue of the
hypostome may be the labrum (labral plate if sclerotized), and, if so, ‘labral plate’ is probably the

FORTEY: TRILOBITE CLASSIFICATION

545

correct morphological term for hypostome, although it has been used by few authors (e.g.
Jaanusson 1975). In crustaceans the labrum is attached to and continuous with the dorsal cuticle.
There is no consensus about what constitutes the most primitive living crustacean, but the attached
labrum applies whether this is taken as a remipede (Schram 1986), or a cephalocarid (Sanders 1963)
or a notostracan (text-fig. 8). The same attachment is observed even in highly specialized crustacean
groups such as the copepods (Boxshall 1982, text-fig. 46), and is apparently invariable. In the
nauplius larva, which is generally regarded as primitively present in Crustacea, the labrum is
prominent (text-fig. 8), and a comparison may be drawn with the prominent hypostome of trilobite
protaspides.

Hence, if the attached labral plate is so widespread among these crustacean groups, then the
detached, natant condition becomes the more unusual as an advanced, apomorphic character.

text-fig. 8. Labral plate attached to carapace in primitive crustaceans and their early growth stages, a ,
carapace of notostracan Triops ; b, immature cephalocarid Hutchinsoniella (after Sanders 1963); c, anostracan

Artemia nauplius (after Sanders 1963).

Ontogenetic evidence for character polarity

Trilobites are unusual among fossil groups in having the exoskeletal ontogeny nearly completely
preserved, and described from numerous taxa. It is feasible to see how the natant hypostomal
condition arises during ontogeny. The earliest larval stage is the protaspis. Protaspides of some
species are known with the hypostome in place, but many more are known from the dorsal shield
minus cheeks. The hypostome is invariably prominent in complete protaspides, occupying a
relatively large area of the venter by comparison with meraspides or adults. In all protaspides from
which the hypostome is described it was attached to the doublure (see, for example, Whittington in
Moore 1959; Fortey and Chatterton 1988, fig. 10; text-fig. 10o-g herein). The protaspis did not have
natant hypostomal condition. Even in those protaspides for which the hypostome is not known it
can be inferred that it was attached. This is because the anterior pits are often prominent near the
anterior margin of the glabella, and the glabella extends far forwards (see text-fig. 10 h-m), so that
the correspondence of anterior pits with hypostomal attachment means that the hypostome had to
be adjacent to the cephalic doublure in the usual position. The prominence of the anterior pits in
many protaspides is presumably a reflection of the large relative size of the hypostome. Such
prominent pits are even visible on smoothed out, planktic, asaphoid protaspides such as those of
asaphids (Tripp and Evitt 1986), nileids (Fortey and Chatterton 1988) and remopleuridids
(Whittington 1959), where they may afford the only evidence of the extent of the cephalic axis (text-
bg. 10m). The only possible exceptions are discussed below.

546

PALAEONTOLOGY, VOLUME 33

CONTERMINANT

STAGE

PLECTRUM

STAGE

STAGE

text-fig. 9. Natant hypostomal condition as a derived character. Ontogenies of typical ptychoparioids,
showing change from conterminant to natant, by way of a plectrate stage, a-e , Ehmaniella waptaensis Rasetti
(after Hu 1986, text-fig. 11), Middle Cambrian, Canada, dorsal views; a , small protaspis, x20; b , large
protaspis, x 16; c, meraspis cranidium c. xl5; d, late meraspis, x5; e, holaspis, x5. f—m, ontogeny of
Crassifimbra (after Palmer 1958), Late Lower Cambrian, Nevada, dorsal (f-i) and ventral (j-rri) views of
comparable stages. The rostral form at the ‘plectrum stage’ is not yet described for this species, but is inferred
from such specimens as that depicted by Zhang and Jell (1987, pi. 80, fig. I n-p)\ the ontogenetic series is
matched by adult cranidial morphologies among Lower Cambrian trilobites, which may record the
evolutionary steps involved in the origin of natant groups; n, Redlichia (after Zhang et al. 1980); o, Dolerolenus

(after Rasetti, 1972); p, Hamatolenus.

Hence during ontogeny the hypostome of trilobites with adults having the natant hypostomal
condition must have become separated from the doublure. How this may have happened in
Ptychopariidae has already been well described in Crassifimbra (Palmer 1958) and can be inferred
from the ontogeny of other ‘generalized’ ptychoparioids such as Ehmaniella (Hu 1984, 1986).
Protaspides of such ptychoparioids are all very similar (Palmer 1962; Hu 1971 ; text-figs. 9, 10 h-m
herein) having a narrow axis with transverse annulations of varying inflation, the anterior of the axis
expanding in width near the front margin of the protaspis. During the meraspis stage (text-fig. 9),
a distinct frontal cranidial border appeared (e.g. Palmer 1958, pi. 26, fig. 4), with which the doublure

FORTEY: TRILOBITE CLASSIFICATION

547

corresponds ventrally, but at this stage the glabella still reached the border, and it is a reasonable
interpretation that the hypostome was attached and conterminant. Subsequent growth allows for
the appearance of the preglabellar field, at first narrow, and widening during subsequent instars.
This marks the separation of border from glabella, and presumably the detachment of hypostome
from doublure, with the onset of the natant hypostomal condition. In both Crassifimbra and
Ehmaniella this happens during the later meraspis. The conterminant stage is separated from the
natant by a phase in which the border furrow curves towards the glabella as a plectrum (plectrum
stage in text-fig. 9). In the ontogeny of Sao (Whittington in Moore 1959) the degree I meraspis was
conterminant, the degree 6 meraspis natant. Thus the ontogeny (conterminant to natant) parallels
the inferred phylogeny. Again this is consistent with the recognition of the natant condition as a
synapomorphy. Adult trilobites representing the plectrum stage are described from several families
(e.g. text-fig. 9). In those Ordovician forms known to be secondarily conterminant from
phylogenetic analysis, there is no evidence to suggest that they went through a natant growth stage
during ontogeny (e.g. Asaphidae; Tripp and Evitt 1986) and so the mature conterminant con-
dition was attained by its retention and displacement into the adult, essentially a process of
paedomorphosis. However in certain Cambrian genera there is evidence to suggest that there may
still be an intervening natant phase of ontogenetic development (i.e. the sequence may run attached
hypostome - free - secondarily attached). In Auritama trihmata , illustrated by Opik (1967, pi. 15,
figs. 3-6), a series of cranidia ranging through probable meraspis to large holaspis show a
progressively widening border and reduced preglabellar field, the border being known to correspond
with the cephalic doublure. It is not known whether the hypostomal attachment became secondarily
conterminant in this species, although the close coincidence of border furrow and preglabellar
furrow in the largest cranidium suggest that it may well have been so.

text-fig. 10. a-g, attachment of hypostome to doublure as the general condition in larval trilobites, illustrated
by protaspides of major trilobite groups shown from the ventral side, a, Ptychopariina, Aphelaspis (after
Palmer 1962); 6, Lichida, Acanthopyge (after Chatterton 1971); c, Scutelluina, Dentaloscutellum (after
Chatterton 1971); <7, Asaphida, Isotelus (after Evitt, 1961); <?, Phacopida (Calymenina), Flexicalymene (after
Fortey and Chatterton 1988);/, Proetida, Proetus (after Chatterton 1971); g, generalized Ptychopariina,
Spencella (after Fortey and Chatterton 1988). Not to scale, h-m , dorsal views of Cambrian libristomate
protaspides of various families showing general (plesiomorphic) similarity, and frontal position of anterior pits
indicating that the hypostome was attached to the doublure on the ventral surface. These protaspides are
arranged in order of greater effacement from left to right, h, late protaspis of Leptoplastus (Olenidae) (after
Whitworth 1970), x 60; /, late protaspis of Welleraspis (Catillicephalidae) (after Rasetti 1954), x 50 ; /, late
protaspis of Dytremacephalus (Pterocephaliidae) (after Hu 1971, pi. 16, fig. 4), x50; k, late protaspis of
Hardyoides (Menomoniidae) (after Palmer 1962), x60; /, late protaspis of Dunderbergia (Alokistocaridae)
(after Hu 1971, pi. 15, fig. 8), x 50; in, early protaspis of Glyphaspis (Anomocaridae) (after Hu 1971, pi. 11,

figs. I and 2), x 50.

548

PALAEONTOLOGY, VOLUME 33

Relatively little is known about early Cambrian redlichiid ontogeny. Kobayashi and Kato (1951)
described the ontogeny of Relichia chinensis Walcott, and figured an early conterminant stage. This
has been augmented by Zhang et al. (1980) who have described ontogenetic series of Redlichia
tcikooensis Lu and R. (Pteroredliehia) murakamii Resser and Endo, including a protaspis stage. By
early meraspis stage (Zhang et al. 1980, pi. 23, figs. 5-7) there was a distinct, ledge-like anterior
border, and it is likely that the hypostome was conterminant. As in olenel 1 ids Redlichia larvae show
well-developed palpebro-ocular ridges which curve strongly backwards. Palmer ( 1957) described the
ontogeny of Oienelius , which apparently lacked a protaspis stage; by the early meraspis hypostome
and doublure were connected only by a median prong, which appears to be the same condition as
pertained in adult olenellids having a preglabellar field. Palmer and Halley (1979) described the
ontogeny of another olenellid, Bristolia , in which it seems likely that the conterminant condition
pertained throughout ontogeny, because the border and preglabellar furrow remain in contact.
Because the redlichiids rather than the olenellids are likely to be the primitive sister group of the
ptychoparioids the ontogeny of olenellids is less relevant to trilobites having natant hypostomal
condition. The familiar mid-Cambrian redlichiine Paradoxides is presumably conterminant
throughout ontogeny. The ontogenies of early representatives of trilobites with natant hypostomal
condition will prove relevant to the morphogenesis of the structures involved, but little is known at
present. A critical group may prove to be the Protolenidae. Protolemis itself was probably natant
in the adult, with the glabella falling well short of the border; closely related genera such as
Mvopsolenus certainly were natant, as shown by specimens showing the inner margin of the cephalic
doublure (e.g. Bassett et al. 1976, pi. 1, fig. I) which is well-removed from the preglabellar furrow.
Romanenko (1977, pi. 13, figs. 3-5) described Protolenoides fasciferrus, and included a figure of a
meraspide cranidium which appears to be of the usual ptychoparioid type, and hence presumably
had an attached hypostome. Small, submature cranidia of the same species show a strong plectrum
extending to the glabellar front. If this represents the dorsal expression of a backward extension of
the rostrum then it might be suggested that the hypostome was still in primary conterminant
attachment mode. The plectrum is less distinct in fragmentary larger cranidia. Plectra are also
present in some other protolenids ( Bergeroniaspis , Olekmaspis ), and in Dolerolenidae (Dolerolenus
- see Rasetti 1972, pi. 15, fig. 4). 1 would speculate that this plectrate condition might represent the
transition between primary conterminant and natant hypostomal condition as shown in text-figure
9. Interestingly, it parallels the situation in Proceratopyge (e.g. Rushton 1983, text-fig. 6 b\ Jago
1987; text-fig. 3 b herein) in which the reverse was happening: as the hypostome achieved secondary
conterminant hypostomal condition a plectrum is developed dorsally corresponding with a
backward extension of the cephalic doublure (in this case including a median suture). The ontogeny
of the Corynexochida, in which the hypostome remained attached to the doublure throughout
growth, was described by Robison (1967).

Stratigraphic evidence for natant condition as a derived character

International correlation of early Cambrian rocks is still controversial. However, in no section with
a relatively complete sequence does a ptychoparioid appear first below a redlichiid (Brasier 1989).
Hence so far as it goes the stratigraphic record is in accord with the proposed character
polarity.

To summarize, ontogeny and stratigraphy confirms the conclusion drawn from comparison with
other arthropods, and from the phylogenetic analysis of trilobite relationships used by previous
authors, that an attached hypostome is primitive and the natant hypostomal condition a derived
character. It is concluded that those trilobites having natant hypostomal condition can be regarded as
constituting a monophyletic group. Those trilobites which were primitively natant, but secondarily
conterminant, such as were discussed above, belong within the same group.

Distribution of natant hypostomal condition in the various trilobite families

There now follows an inventory of trilobite families in which the natant hypostomal condition
occurs, or can be reasonably inferred to occur, with documentation. Some examples are illustrated

FORTEY: TRILOBITE CLASSIFICATION

549

on Plate I ; for most others a literature reference is given, either showing the hypostome as it fell
into the forward part of the glabella upon decay of the soft parts, separated from the cephalic
doublure, or showing that the extent of the cephalic doublure fell short of the preglabellar furrow.
Some modes of hypostomal attachment were described in detail by Whittington (1988a, b). The list
is not exhaustive, in that one or two examples are sufficient for each family, and there may be
additional families for which I have not found evidence. For convenience families are arranged
alphabetically. Families are accepted uncritically, although many are not satisfactory, monophyletic
taxa. Even so I cannot accept Raymondinidae, which is an absurdly heterogeneous collection of
genera in the Treatise.

Agraulidae. Agraulos was discussed in detail by Whittington 1988a.

Alokistocaridae. See Elrathia (PI. 1, fig. 6).

Anomocaridae. See discussion in Fortey and Chatterton (1988); Anomocarioides limbatus (Egorova et at.

1982, pi. 44, fig. 2).

Andrarinidae. A somewhat doubtful taxon. Dean ( 1 972) illustrated narrow cephalic doublure of Holasaphus ,
which shows likelihood of natant hypostomal condition. Although Holasaphus was placed in this family by
Howell (in Moore 1959), it was assigned with question to Ptychopariidae by Dean.

Asaphiscidae. Blainia , see PI. 1, figs. 2, 8.

Aulacopleuridae. Natant condition can be inferred on Aulacopleura with narrow and tubular cephalic
doublure and wide preglabellar field.

Auritamiidae. See ontogenetic series of Auritama trilunata (Opik 1967, pi. 1 5, figs. 3-6.) and discussion above
(p. 547).

Bathyuridae. See discussion above (p. 538); most bathyurids are regarded as likely to have been natant
except Bathyurus (Whittington, 1988 b).

Cedariidae. Cedaria (PI. 1, fig. 3).

Conocoryphidae. Probably a polyphyletic family, united only by secondary blindness. Conocoryphe itself
discussed in detail by Whittington (1988a).

Dimeropygidae. This family is problematic, in that Dimeropyge has a curious rostral plate (Tripp and Evitt

1983, pi. 33, fig. II; Chatterton and Ludvigsen 1976, fig. 8) with a prong-like extension which may have
engaged with the unusually small hypostome (which has reduced anterior wings). If this is so it qualifies as a
unique form of (derived) hypostome attachment, and is neither natant nor conventionally secondarily
conterminant. On the other hand I have searched inside entire cephalic shields of early dimeropygids such as
Ischyrotoma anataphra Fortey, 1979 without finding any sign of an attached hypostome, and I think it likely
that these early forms were natant.

Dokimocephalidae. Little known of ventral surface. Narrow doublure of free cheeks shown in \ Vuhuia by
Zhang and Jell (1987, pi. 1 19, fig. 5) and natant condition likely in this member of the family.

Ellipsocephalidae. Snajdr (1958) has illustrated the cephalic doublure of the type species of Ellipsoceplialus ,
which demonstrates the likelihood of a gap between hypostome and doublure.

Hapalopleuridae. Skjarella sp., a new discovery in the Tremadoc of Shropshire by myself and R. M. Owens
(in prep.) shows natant condition in entire specimen with hypostome in place.

Idahoiidae. Saratogia , narrow width of genal doublure illustrated by Ludvigsen and Westrop (1983, pi. 8,
fig. 10); if this followed border furrow on cranidium then natant condition pertained.

Loganopeltidae and Harpididae. Lower lamella in Harpides falls well short of glabellar front (holotype of
H. grimmi Barrande, 1872 shows hypostome in place). These two families are united by some authors.

Lonchocephalidae. Rasetti (1954, pi. 62, fig. 13) illustrated narrow genal doublure, and rostral suture of
Welleraspis swartzi. It is possible that some lonchocephalids lacking preglabellar field were secondarily
conterminant (cf. Whittington 1988a).

Marjumiidae. See Robison (1964, pi. 10, figs. 10-15) for silicified specimens of Modocia showing how
doublure falls short of preglabellar furrow.

Menomoniidae. Bolaspidella discussed by Whittington 1988a.

Nepeiidae. Penarosa netenta , see Jell (1977, especially pi. 21, figs. 5, 7).

Olenidae. Diagnosis in Nikolaisen and Henningsmoen 1985; Peltura see Henningsmoen (1957, pi. 26, fig. 2);
Porterfieldia (Fortey and Owens 1978, pi. 1, fig. 5); Parabolinella (Whittington 1988a); Hypermecaspis (PI. 1,
fig. 1). Some of the in situ hypostomes seem to be too small to have been attached by the anterior hypostomal
wings as suggested by Whittington 1988a. Some derived forms which have lost the preglabellar field ( Triarthrus
see Whittington and Almond 1987) may have become secondarily conterminant.

Papyriaspididae. Doublure of free cheek of Papyriaspis lanceola figured by Opik (1961, pi. 17, fig. 3).

550

PALAEONTOLOGY, VOLUME 33

Parabolinoididae. See, for example, Westrop (1986, pi. 18, fig. 18) for narrow genal doublure coinciding with
border in Orygmaspis ( Parabolinoides ) species with wide preglabcllar field.

Proetidae. Proetus (PI. 1, fig. 7) and Decoroproetus , see Owens (1973) for specimens preserved from ventral
side with hypostomes nearly in situ.

Protolenidae. Hamatolenus see Bassett et al. 1976, and comments above.

Pterocephaliidae. Strigitambus? blepharina (Palmer 1968, pi. 16, fig. 15), Litocephalus (Palmer 1960, pi. 8,
figs. 13, 16). See also discussion in Fortey and Chatterton 1988, which excludes Aphelaspis (also natant) from
this family.

Ptychopariidae. Ptychoparia striata (see Whittington 1988a); Asthenopsis (see Jell 1978a). Snajdr (1958, pi.
40, fig. 3) showed Ptychoparia hypostome in near-life position.

Remopleuridacea. Richardsonella (see Palmer 1968, pi. 14, fig. 8); Elkanaspis (see Ludvigsen 1982 for
silicified specimens showing narrow cephalic doublure combined with wide preglabcllar field).

Solenopleuridae. Sao , discussed in Whittington 1988a.

Trinucleidae. Statement in Hughes et al. (1975, p. 541); Whittington (19886, p. 328).

This long list proves two important points:

1. If the natant hypostomal condition defines a monophyletic group as argued above, it is
retained in a variety of families within that group - it is generally a stable character.

2. Within this group it is also a primitive character, being shared with the common ancestor of
the group. This means that it is of no consequence in defining phylogenetically based subgroups. This
is the usual situation where a synapomorphy becomes symplesiomorphic at the next level in the
taxonomic hierarchy (Eldredge and Cracraft 1982). Hence within the group defined by having
natant hypostomal condition, any subgroups have to be defined on other, derived characters.

This means that natant hypostomal condition is of use in suggesting common ancestry, but we
must look to other characters to define natural groups within this great assemblage of trilobites.
Those groups (currently Asaphina, Proetida) with secondarily conterminant or impendent
hypostomal condition belong within the overall natant group, and have to be defined on other
derived characters - for example, the Asaphina are defined by having a median suture - and the fact
that early members of the group have natant hypostomal condition while later ones generally have
attached hypostomes (Fortey and Chatterton 1988) supports their common ancestry, but does not
greatly help with the definition of the group. A definition of Proetida is given below.

If the natant hypostomal condition is correctly regarded as a synapomorphy then the natural
group having it, or only secondarily losing it, needs a name. For this group I propose the name
Libristoma. The taxonomic status of the Libristoma is discussed in more detail below; it is a more
inclusive group even than the Order Ptychopariida.

HYPOSTOME ATTACHMENT AND HYPOSTOME SPECIALIZATION

One of the consequences of the natant hypostomal condition seems to have been that, upon the death
of the trilobite, or during moulting, the hypostome was easily separated from the rest of the
exoskeleton. I do not think it is a coincidence that most of the families listed above have very few
species for which the hypostome is definitely assigned. Of dozens of species of Ordovician trinucleids
only two to my knowledge have hypostomes associated, and the figure is similar for many Cambrian
families, and some (e.g. Dokimocephalidae) apparently have no hypostome yet associated.
Conversely, unassigned hypostomes, all very similar, are often figured at the end of monographs on
Cambrian trilobites (e.g. Palmer 1960, pi. 11). Cambrian families with conterminant hypostomal
condition, such as Redlichiidae, Paradoxididae or Corynexochidae, have numerous species with
definite assignment of hypostomes. The same applies to those families (Asaphidae, Nileidae) in
which secondary hypostomal attachment has occurred. Nonetheless in natant families when the
right kind of ventral-side-up preservation occurs the hypostomes are found in the appropriate
position, and I see no grounds to suppose that there were species which lacked hypostomes
altogether.

On gathering together information on the hypostomes which have been assigned to trilobites

FORTEY: TRILOBITE CLASSIFICATION

551

k

I

m

n

text-fig. 11. The conservative morphology of hypostomes from trilobites having natant hypostomal
condition, illustrated by hypostomes of species belonging to several families and ranging in age from Lower
Cambrian to Carboniferous (top left to bottom right), a, yimtid redlichiid Drepanopvge (after Zhang et al.
1980, pi. 51. fig. 9), showing how this primitive morphology is retained in natant trilobites, x 2. Natant
hypostomes: b , Crassifimbra (Ptychopariidae), Lower Cambrian (after Palmer 1958, pi. 25, fig. 13), x 10; c ,
Conocoryphe (Conocoryphidae), Middle Cambrian (after Snajdr 1958, pi. 34, fig. 4), x 4; d , Ptychoparia striata
(Ptychopariidae), Middle Cambrian, (after Snajdr 1958, pi. 40, figs. 3, 5), x2-5; e, Dunderbergia
(Alokistocaridae), Upper Cambrian (after Hu 1971, pi. 15, fig. 30), x 10; /, Aphelaspis (regarded as
Pterocephaliidae), Upper Cambrian (after Rasetti 1965, pi. 20, fig. 9), x 5; g, a hystricurid, early Ordovician
(after Ross 1951, pi. 19, fig. 3), x9; /;, Parabolinella (Olenidae), early Ordovician, x4; /, Bathyurellus
(Bathyuridae), early Ordovician (after Fortey 1979, pi. 31, fig. 9), x3; j, Paraproetus, (Proetidae), late
Ordovician (after Owens 1973, pi. 12, figs. 12a and b), x 7; k, Proetus (Proetidae), Silurian (after Owens 1973,
pi. 4, fig. 19), x 4; /, Koneprusites (Proetidae), Devonian (after Snajdr 1980, pi. 51, fig. 10), x 6; m, Voigtaspis
(Proetidae), Devonian (after Snajdr 1980, pi. 52, fig. II), x 3; «, Phillipsia (Phillipsiidae), Carboniferous (after

Woodward 1883, pi. 2, fig. 6), x3.

having natant hypostomal condition another fact becomes clear. The hypostomes of trilobites with
natant hypostomal condition tend to be as similar as they are unremarkable (Whittington 1988a,
p. 604; text-fig. I I herein). What they share is an elongate (sag.) oval middle body, carrying a pair
of usually rather short middle furrows often posterolaterally placed; rather narrow (tr) lateral
borders, and a rounded to transverse posterior border without a fork, or other marginal
modification; a pair of small posterolateral spines may or may not be present. Hypostomes of this
kind can be found from the Lower Cambrian to the Upper Palaeozoic. The hypostomes of some
redlichiid and dolerolenid trilobites are virtually identical (e.g. Dolerolenus , see Rasetti 1972, pi. 16)
and the obvious inference is that this ‘ptychoparioid’ hypostome is a retained primitive character.
As the redlichiid ancestor of the Libristoma acquired the natant hypostomal condition, the
hypostome became, as it were, morphologically frozen. Surface sculpture on many of these
hypostomes was lacking, or at most finely granulate, and only became at all complex on some
proetids. The cephalic doublures of most ‘ptychoparioids’ carry terrace ridges. We can contrast the
conservatism of the natant hypostome with the new features in morphology that arise in secondarily
conterminant and impendent hypostomes (text-fig. 12). Posterolateral hypostomal borders can
become very wide (Nileidae); posterior forks can develop (most Asaphidae, Hypodicrcmotus in the
Remopleurididae) ; the overall shape can become greatly elongate (Brachymetopidae) or transverse
(Cyclopygidae) ; maculae may become prominent (. Remopleurid.es ) or absent (the cyclopygid
Degamella). The terrace ridges hitherto largely confined to the cephalic doublure appear to ‘spread
onto’ the hypostome and may be as prominent, or even more prominent, as they are upon the

552

PALAEONTOLOGY, VOLUME 33

text-fig. 12. Hypostomes of trilobites in which the hypostome was attached to the doublure in either
impendent or conterminant mode, showing various kinds of specialization, in contrast to the natant
hypostomes shown in text-fig. 11. a-e shows primary conterminant examples, as interpreted herein;/-/ are
secondarily conterminant hypostomes belonging to Libristoma; j may be secondarily conterminant (see
discussion of Phacopida, p. 565). a , Fieldaspis (Corynexochida), Middle Cambrian (after Rasetti 1951, pi. 16,
fig. 17), x 1-5; 6, Palaeadotes (Damesallacea), Cambrian, Mindyallan (after Opik 1967, pi. 50, fig. 3), x4; c,
illaenid (lllaenacea). Middle Ordovician (after Whittington 1963, pi. 18, fig. 4), x3; d, Amphilichas (Lichida),
Middle Ordovician (after Tripp and Evitt 1981, pi. 2, fig. 28), x6; e, Primaspis (Odontopleurida), Silurian
(after Chatterton and Perry 1983, pi. 5, fig. 5), x 6;/, Lycophron (Asaphacea), early Ordovician (after Fortey
and Shergold 1984, pi. 43, fig. 5), x3; g , Remopleurides (Remopleuridacea), Middle Ordovician (after
Whittington 1959, pi. 17, fig. 18), x6; /;, Symphysurus (Cyclopygacea). Middle Ordovician (after Fortey 1986,
fig. 2), x 1; i, Degamella (Cyclopygacea), early Ordovician (after Fortey and Owens 1987, fig. 38 d), x2-5; j,
Salterocoryphe (Calymenacea), Middle Ordovician (after Hammann 1983, pi. 11, fig. 107), x6.

doublure. Such derived hypostomal characters at once become of taxonomic use, in defining
families, genera, or even species (asaphids are often best distinguished by hypostomal details).
Hence, natant hypostomes are of a generalized type and change little; attached hypostomes are
rapidly deployed into a variety of shapes and are as variable as any dorsal morphological features.
This has to be connected with the function of the hypostome in the life habits of the trilobites
included in the Libristoma.

If these facts are easily stated the possible explanations of the facts quickly become speculative.
What is certain is that ti , natant condition was both widespread and enduring. As a derived
character it was connected with life habits that the trilobites could ubiquitously and successfully
prosecute from the Cambrian to the Carboniferous, and it is unlikely, therefore, that it was related
to any peculiar specialization, such as parasitism. Because the hypostome remains remarkably
conservative it is possible to suggest that it was not actively involved in the mastication of food, and
this is supported by the lack of the kind of buttressed attachment found in the conterminant and
impendent hypostomal conditions. So it seems reasonable to argue further that it was the very
flexibility of the hypostomal attachment that was of adaptive importance. If Whittington’s (1988a)
view of the relationship between anterior hypostomal wings and anterior pits (or the homologous
points in the axial furrows) is correct then the hypostome was tethered rather than entirely free-
lying, but that the attachment was not rigid is shown by the ease with which it becomes displaced
on death to produce specimens like that shown in PI. 1, fig. 7.

If the mouth lay at the back end of the hypostome it is not unreasonable to assume that the ovate
middle body of the hypostome corresponded with some kind of foregut above it. In this case

FORTEY: TRILOBITE CLASSIFICATION

553

flexibility of the hypostome would mean that the foregut might be manipulated by the hypostome
itself. If we further assume that the hypostome was ‘tethered’ by the anterior wings this movement
would be confined to up-and-down movement of the adoral margin: more precisely the movement
would be upwards-and-backwards because of the tethering at the anterior end of the hypostome.
It would be a scooping motion. The musculature necessary to do this would be appropriately sited
if it attached to the middle furrows. It does not seem likely that the muscles were particularly
powerful because the middle furrows are usually rather weakly defined. Hence it is likely that if the
hypostome functioned in this scoop-like fashion the particles handled were small. This could either
mean that the natant trilobite lived on small organic particles extracted from the sediment by the
setate bases of the coxae (cf. Whittington and Almond 1987) and then passed forwards along the
ventral ‘food groove’, or that the trilobite directly ingested partially sorted sediment, the process
of nutrient extraction being carried out in the gut. In any case it seems unlikely that natant trilobites
handled large prey. The scooping motion of the hypostome/foregut could have functioned in
shovelling a considerable quantity of sediment and organic material for digestive processing. It may
be relevant that many ‘ ptychoparioid ’ (Ptychopariidae, Olenidae, aphelaspids, etc.) trilobites have
well-developed caecal networks, some of which have been interpreted as intestinal diverticulae
(other workers claim these as arterial, see Jell 1974). But the conservatism of the natant hypostome
attests to it not being directly involved with the manipulation and processing of food, however the
food was derived.

As soon as the hypostome became secondarily conterminant or impendent specializations
appeared. Whittington (1988a, b) has demonstrated that the hypostomes in all cases of conterminant
or impendent attachment he examined were rigidly attached at the sutural junction with the
doublure. This implies a functional difference compared with the natant condition. Such
hypostomes cannot be involved with the shovelling of food forwards, but it becomes possible for
them to become part of the mechanism of manipulating and processing the food gathered. The
powerful buttressing of the hypostomal wings and doublure that happens in, for example, asaphids
and nileids implies that mechanical strength was important in this function. Certain species with
attached hypostomes have conspicuous rasps on the adoral part of the hypostomal margin: this is
shown by the odontopleurids discussed by Chatterton and Perry (1985), and by the asaphid Isotelus
(text-fig. 3e, p. 534). Such rasps could have functioned in concert with the limbs in the shredding
or comminution of large food particles. It seems likely to me that the Libristoma having predatory
habits were also those having secondary conterminant or impendent hypostomes. The special-
izations of the hypostomes were the result of the various stratagems evolved to deal with bulkier
food items, including prey in at least some species. The function of rasps does not pose an
interpretative problem, but many of the other specializations do. Forks, or lateral spine-like
processes, are developed on many conterminant hypostomes, both primary and secondary
( Paradoxides , calymenids, asaphids, some remopleuridids). In asaphids the fork extended to either
side of the mouth and may carry strong terrace ridges on its outer surface (e.g. Fortey 1979, pi. 23,
fig. 4) could these have also helped in manipulating food? The maculae behind the middle furrows
on asaphids and nileids are modified, bevelled areas, and their function is not resolved, but if the
hypostomal retractor muscles remained they must have taken on a different function.
Remopleuridids have enormously extended maculae the function of which is unknown. The spread
of terrace ridges from the cephalic doublure onto the anterior part of the middle body of the
hypostome in the secondarily attached condition means that this part of the hypostome has become,
functionally speaking, part of the antero-venter. The function of terrace ridge systems in trilobites
is not without controversy, but anterior cephalic ones probably functioned as sediment grippers in
some instances (Schmalfuss 1981; Fortey 1986). In the secondarily conterminant families these
changes from natant hypostome morphology happened over an early Ordovician period of perhaps
ten million years. The functional interpretation of the specializations of attached hypostomes is a
fruitful field for further research. For the moment it is only possible to contrast these morphological
excursions with the conservatism of hypostomes separated from the doublure, for which one
presumes a single function sufficed for several hundred million years.

554

PALAEONTOLOGY, VOLUME 33

CLASSIFICATION OF AGNOSTIDA

The Agnostida are usually understood to comprise the agnostids themselves, Suborder Agnostina,
and the eodiscids and pagetiids (suborder Eodiscina sensu Moore (in Moore 1959); superfamily
Eodiscoidea sensu Jell (1975)). There is now much evidence that the hypostome of the Agnostida was
not attached ventrally (text-fig. 13). The apparently natant condition in Eodiscina has been
discussed by Jell (1975, pi. 1, fig. 1) and Whittington (1988(3), and is now known in several members
of the Agnostina (Robison 1972a; Jell 1975, pi. 1, fig. 3; Muller and Walossek 1987; Robison 1988),
and there is no reason to suppose it was other than general.

text-fig. 13. Natant hypostomal condition in Agnostida, showing no indication of rostrum. Agnostina: a,
Peronopsis ferox (Tullberg) (Robison 1972a, fig. Id) x21 ; b, Onymagnostus seminula (Whitehouse), (Robison
1982, pi. 4, fig. 5b), x 10; photographs kindly supplied by R. A. Robison. Eodiscina: c, Pagelia cephalic shield

from underside (after Jell 1980, pi. 1, fig. 1).

When I first recognized the natant condition as a synapomorphy 1 had considered that the
agnostids should belong in the same monophyletic group as libristomates because of their
unattached hypostome. This was wrong. The natant condition in Agnostida is not homologous with
that in Libristoma. Recognition of this also serves to place the Agnostida within the Trilobita (text-
fig. 14), which has been a long-standing problem. It is necessary to discuss this before considering
the ptychoparioids further.

Much more is known of agnostid anatomy following M tiller and Walossek's (1987) magnificent
work. Obvious agnostoid specializations have led them to be classified separately from the rest of
the Trilobita (Miomera of Jaekel 1909). However, the original emphasis placed on the paucity of
thoracic segments does not bear scrutiny because it is now known that several other groups of
trilobites may include species with only two or three thoracic segments, for example Corynexochida
(Robison and Campbell 1974) and Raphiophoridae (Zhang 1980). Small numbers of segments
result from a change in developmental programme (cf. McNamara 1986), and it is not a character
of great weight, taxonomically speaking, although it is stable in Agnostina. The fact remains that

FORTEY: TRILOBITE CLASSIFICATION

555

inset shows a second hypothesis which is not favoured by the author. The relationship of Libristoma to the
group termed ‘other polymerids’ is not resolved. ‘Other polymerids’ comprises primary conterminant non-
redlichiid trilobites: Corynexochida, Lichida, Odontopleurida, Scutelluina as discussed in text, and may not
be a natural group. Derived characters are: X, combination of seven, possibly nine, synapomorphies of
olenellids and other trilobites as discussed by Fortey and Whittington (1989) (presence of a pygidium; rostral
plate; calcite eye lens structure; calcite cuticle; circum-ocular sutures in adult; eye ridges; hypostome with
anterior wing directed dorsally to anterior pit; and possibly terrace ridges on doublure and spinose margin of
larval hypostome); 1, acquisition of dorsal ecdysial sutures; 2, reduction in thoracic segment number to three
or fewer; 3, acquisition of a calcified protaspis; 4, connective sutures present (section of doublure included on
free cheek); 5, natant hypostomal condition; 6, glabella sharply truncated at ledge-like border; 7, loss of
calcification of olenellid-like rostral plate; 8, occipital width greatly exceeds (tr.) width of preoccipital glabella
(expressed as triangular ‘basal lobes’ in Agnostina); 9, broad, ‘rolled’ cephalic border; 10, cephalic shield long
(sag.) (maximum width often in front of posterior margin); 11, genal spines reduced or absent; 12,
cephalothoracic aperture; 13, protaspis/early meraspis eye ridges close to anterior margin (‘ptychopariid
style’ protaspis); 14, fulcrum-and-facet thoracic articulation (flange-like articulation primitively present in
olenellids and at least some redlichiids, position of acquisition of advanced articulation may be oversimplified
here - probably accompanied by the capacity to enrol fully). Circles, some plesiomorphic characters; 15, dorsal
sutures lacking; 16, rostral plate extending around ventral cephalic margin; 17, olenelloid glabella form
(elongate, fusiform with round to acuminate frontal lobe). The illustrated genera are only token representatives
of their groups, and not necessarily those from which coding was made.

556

PALAEONTOLOGY, VOLUME 33

the Agnostina are the most specialized of trilobites, possibly being planktic (Robison 19726) and
spending much of their life in the enrolled condition, according to Muller and Walossek (1987). This
is not, however, relevant to their classification within the trilobites, which has to be based on shared
characters rather than distinctive autapomorphies. Their peculiarities have tended to dominate the
view about how they should be classified, for example: they ‘are best treated within the Trilobita
at present but... having a separate position within them’ (Muller and Walossek 1987, p. 54).

The relationships of the Agnostida suggested here are shown on the cladogram (text-fig. 14).
This places them firmly within the Trilobita, between the Olenellida and Redlichiida. If the reasons
advanced by Fortey and Whittington (1989) for the olenelloids as the most primitive trilobites are
accepted, this is the preferred position for the agnostoids based on the characters I consider relevant
to the problem.

The reasons for this phylogenetic position are threefold:

1. Students of Eodiscina agree that this group is related to Agnostina, of which they may be the
primitive sister group (Jell 1975; Zhang et al. 1 980). According to Zhang et at. ( 1 980, text-fig. 7) two
different groups of agnostids may have been closely related to different eodiscids - implying that the
latter is a paraphyletic group and the former polyphyletic. Howsoever these relationships are
worked out in detail, within the Eodiscina as understood at present there appear to be a series of
‘intermediates’ spanning some of the gaps between pagetiid organization and agnostid
organization: three and two segmented species within the same genus ( Pagetia as discussed by Jell
1975); blind eodiscids otherwise with pagetiid features (see Rasetti 1966); and some eodiscids with
pre-occipital median tubercles on the glabella ( Serrodiscus and Ladadiscus , see Rushton 1966, text-
figs. 3, 9). The agnostid hypostome is so far unique in construction (Robison 1972r/) with its
window-like openings, although according to Muller and Walossek (1987) this may only be a matter
of cuticular thinning (see also Robison 1988, fig. 9). The cephalothoracic aperture (a gap between
the thoracic axis and cephalon in the enrolled state produced by the absence of the articulating half-
ring on the first thoracic segment) is considered to be a good apomorphic character of Agnostina.
The hypostome of Pagetia is no different from the generalized ‘ ptychoparioid ' hypostomes shown
on text-fig. 1 1 (Jell 1975, pi. 28, figs. 1 and 2). Hence, for comparative purposes, the most primitive
pagetiids are probably the best candidates for comparison with other trilobites to determine the
relationships of Agnostida. They have dorsal facial sutures. The presence of such sutures was
considered by Fortey and Whittington (1989) to define a clade of ‘higher’ (i.e. non-olenelloid)
trilobites, and most authors agree (Bergstrom 1973; Lauterbach 1980) on the peculiarity of this
character in arthropod phylogeny.

2. Redlichiida, and all other trilobites, other than olenelloids and Agnostida, have a rostral plate
bounded laterally by connective sutures - or if the rostral plate has been lost (e.g. to produce fused
cheeks) it is known that their ancestors had such a rostral plate. In ventral view free cheeks of this
group show a truncated section of doublure. Olenelloids have a broad crescentic rostral plate,
without connective sutures. If the olenelloids, on other criteria, are the most primitive trilobites, this
form of rostral plate is also primitive. Ventral-side up preservation of numerous Agnostida,
including those with hypostomes in place (text-fig. 13), show no sign of a rostral plate. The primitive
agnostoids with dorsal sutures included in Pagetiidae do not show that part of the genal doublure
bounded by connective sutures. Hence, the rostral form must have been as in olenelloids, but
calcification has been lost (text-fig. 15). This is confirmed by an example where it appears to have
been silicified in one of the youngest agnostids (Middle Ordovician. Hunt 1966), but still retains
olenelloid form. The loss of calcification of the (olenelloid) rostral plate affords an autapomorphy
defining the Agnostida, one which is unique in the Trilobita.

3. It follows that the unattached condition of the hypostome in the Agnostida is the result of loss
of the rostral plate and not the same as in the Libristoma. Individual agnostoid ontogenies described
by Jell (1975) and Rushton (1966) show that the preglabellar field became progressively longer
during ontogeny. It is suggested that the most primitive member of the Eodiscina could be such a
species as Sinodiscus changyanensis S. G. Zhang in Zhang et a /., 1980, in which dorsal furrows and
eye ridges are well-defined (as in olenellids) and the preglabellar field is short.

FORTEY: TRILOBITE CLASSIFICATION

557

a bed

text-fig. 15. Series of cephalic shields viewed from the ventral side to illustrate why natant condition in
Agnostida is not homologous with that in Libnstoma, and, left to right, morphological steps in origin of
Agnostida. Only b is hypothetical, a, primitive condition illustrated by typical conterminant olenellid ( Holmia ,
see Whittington 1988a) showing long crescentic rostral form, confining other cephalic doublure to narrow
selvage; b, dorsal sutures have appeared, but calcified rostral plate lost; c , most primitive kind of pagetiid (well-
defined glabellar furrows, glabella extends comparatively far forwards) based on Sinodiscus (Zhang et al. 1980,
pi. 4, figs. 16-21); d, Pagetia (based on Whittington 1988a) illustrating advanced agnostoid morphology
triangular ‘basal lobes’ (possibly extraglabellar) and glabella further short of cranidial margin.

Another character may be cited in the question of agnostid relationships. No protaspis has been
described for an olenellid (Palmer 1957), nor yet for a member of the Agnostina. Since meraspides
are well known for both it seems likely that this is a consequence of calcification not having ‘spread
back' far enough in ontogeny in these groups, which again would be interpreted here as a primitive
feature. However, given the other specializations of agnostids, lack of protaspis could well be a
secondary loss. J. H. Shergold informs me that he has found an eodiscoid protaspis (Zhang 1989;
Hu 1971, pi. 7, fig. 23), which would confirm the latter view. Planktic ostracodes, resembling
agnostids in size and putative life habits, undergo accelerated development in the egg stage, and of
all the ostracodes only one genus, Manawa , has a free-living nauplius larva (Swanson 1989). An
alternative view might separate Eodiscina and Agnostina as distinct clades, influenced perhaps by
the more general ‘ptychoparioid ’ appearance of the former, is shown also on text-fig. 14. This
produces a less parsimonious arrangement of the characters I consider important, and is not
favoured.

If these characters are interpreted correctly the Agnostida appear above the olenelloids by virtue
of their (primitively present) dorsal sutures but below all other trilobites, including Libristoma,
because they had not acquired the advanced rostral form. This position does justify their high
taxonomic status, and is also consistent with their early stratigraphic appearance in the Lower
Cambrian of, for example, China. The loss of calcification of the crescentic rostral plate is one
character (and perhaps the most important) defining Agnostida. They are not part of Libristoma.

DIAGNOSIS AND STATUS OF LIBRISTOMA

The Libristoma may be diagnosed as follows : Trilobites having natant hypostomal condition , or with
secondarily conterminant or impendent hypostomal condition. Rostral plate bounded by connective
sutures primitively present, but may be lost by virtue of fusion of the free cheeks, or replaced by
median suture.

If Libristoma is a high level monophyletic taxon then it requires taxonomic recognition. A glance
at Harrington’s (in Moore 1959) review of high level trilobite taxonomy quickly shows that the fate
of most of the proposed high level taxa in the past has been oblivion. Lew authors use such orders
as Miomera, Epiparia, Polymera or Protoparia at the moment, and none of these is equivalent to
Libristoma. Miomera and Polymera are occasionally employed by Cambrian specialists (Robison
1988); Miomera is equivalent to Agnostida as used here, while Polymera is an unnatural taxon used
to encompass the rest of the non-agnostid trilobites. In dealing taxonomically with the new
phylogenetic concept implied by Libristoma there were two choices.

I . To redefine the Order Ptychopariida to equate with Libristoma. This would mean that major

558

PALAEONTOLOGY, VOLUME 33

monophyletic taxa within the redefined ptychopariids would become suborders. This would include
Proetida, which has been regarded as a separate order, although Liitke (1980) regarded Proetida as
a ptychopariid suborder. Also, for reasons explained below, the suborder Ptychopariina resulting
from this classification would be paraphyletic, and it is perhaps wiser to provide a clear
nomenclatorial distinction between an allegedly monophyletic group (libristomates as a whole) and
its least satisfactory subgroup.

2. To propose Libristoma as a Subclass. In this case the major groups within Libristoma can be
regarded as Orders, which does least violence to the existing classification. This runs the risk that
Libristoma, like the major subdivisions proposed in the past, will simply not be used. On the other
hand, if it is a good monophyletic taxon it should survive by virtue of the light it casts on major
phylogenetic events in trilobite history. The trouble with, say, Stormer’s (1942) Order Proparia
(including eodiscids, norwoodiids, burlingiids and Phacopida on the basis of proparian facial
sutures) was not so much the high level status of the taxon involved as the fact that it has
subsequently been proved to be a polyphyletic assemblage.

For the moment I prefer the second option. A stricter cladist would insist that a full phylogenetic
analysis of the whole group be carried out before the question of taxonomic status is decided. I
accept that this kind of study may alter the status of the subgroups within Libristoma. For the
moment a pragmatic option is to retain major monophyletic subgroups of Subclass Libristoma as
Orders or Suborders, which is what is done below. Suborders are retained within Ptychopariida for
no better reason than traditional usage, and this may well change in the future.

Orders and suborders within Libristoma and their diagnoses

The subgroups within Subclass Libristoma should be definable on autapomorphies, with the natant
hypostomal condition being the synapomorphy uniting the whole group. A list of primitive
characters within the accepted 'Ptychopariida' was given by Fortey and Chatterton (1988, table 2).
These include: rostal plate present (transverse width not exceeding that of glabella); tapering
glabella with three pairs of glabellar furrows which are shorter anteriorly; a relatively large number
of thoracic segments with unspecialized articulation; opisthoparian sutures; terrace ridges on
doublure; genal spines present; unspecialized hypostome of the kind discussed above. Most of these
characters are present also in redlichiids and are therefore simply characters of all trilobites of the
Redlichiid + Libristoma clade: for example, genal spines; opisthoparian sutures; large numbers of
thoracic segments; larval hypostomal margin with fringe of spines; glabellar form in many species
of redlichiid; terrace ridges on doublure (Kobayashi and Kato 1951, pi. 5, fig. 6; Hupe 1953, pi. 9,
fig. 9). With the exception of opisthoparian sutures and the glabellar form, most of these characters
are also present in olenellids, and if the olenellids are regarded as trilobites, such characters are
simply generalized characters of the trilobites as a whole (Fortey and Whittington 1989). So
diagnostic characters of subgroups of Libristoma must comprise characters outside this list.

Order proetida Fortey and Owens, 1975

Diagnosis. Libristoma having fusiform glabella in the protaspis stage, and natant hypostomal
condition attained early in ontogeny. Larval hypostome margin lacking spines. Most species natant
also in adult; some secondarily conterminant or impendent. Median occipital tubercle usually
present.

Fortey and Owens (1975) proposed Proetida for what they regarded as a major natural group of trilobites
which had been incorrectly classified in the Treatise (Moore 1959). Despite the criticisms of Bergstrom (1977)
Proetida has gained wide acceptance since, and some additional taxa have been added to the group (e.g.
Telephinidae by Chatterton 1980). Liitke (1980) recognized the group as a suborder of Ptychopariida, and
therefore part of the 'ptychopariid problem’. Fortey and Owens’ original list of attributes of Proetida was
based on what they regarded as overall similarity of the taxa included, without critical assessment of what
characters were plesiomorphic and what derived. Part of their argument for monophyly stemmed from a
stratigraphical/morphological hypothesis that all Proetida could be derived from the early Ordovician

FORTEY: TRILOBITE CLASSIFICATION

559

Hystricurinae. Now that it has been recognized that the majority of Proetida (and all primitive ones) had natant
hypostomal condition it is possible to apply a more critical judgement to the characters on which the group
is based. If it is a natural group there should be a synaponrorphy defining it, whether or not the hypothesis of
hystricurine derivation is correct. Many of the proetide characters Fortey and Owens listed (1975, pp. 235-6)
are generalized ‘ptychopariid ’ (cf. Fortey and Chatterton 1988, tabie 2) or even belong to the whole
Libristoma: they will not suffice for the definition of the group. Bergstrom’s (1977) criticisms were justified on
this point. For example, the hypostomal form described by Fortey and Owens is simply that of the typical
natant trilobite (text-fig. 11), while the presence of terrace ridges on the doublure is an even more
symplesiomorphic character, being present in Redlichiida and even Olenellida also.

Fortunately for the concept of Proetida there is a derived character uniting the group, but this does not
become apparent until the hypostomal condition is brought into the argument. Fortey and Owens (1975) did
point out that early growth stages of Proetida resembled one another, and illustrated early meraspide cranidia
of various families showing tapering, rather elongate glabellas with a deeply parabolic outline. A preglabellar
field was present even at this early growth stage. Chatterton (1980) and Fortey and Chatterton (1988) noted the
similarity of proetide protaspides. These differ from the typical 'ptychopariid’ (text-figs. 9 and 10) in having
a fusiform, anteriorly rounded or pointed glabella. That this is general for the Proetida is demonstrated (text-
fig. 16) by the occurrence of proetide type protaspides in Proetidae ( Proetus , see Chatterton 1971, pi. 16, fig.
1), Bathyuridae ( Licnocephala , see Ross 1953; Bathyurus , see Chatterton 1980), Aulacopleuridae ( Harpidella ,
see Chatterton 1971, pi. 18, fig. 20; Scharyia , see Snajdr 1981), Dimeropygidae ( Dimeropyge , see Tripp and
Evitt 1983), and Telephinidae (Phorocephala ( Carrickia ), see Chatterton 1980; also Carolinites, see Fortey
1975, pi. 36, figs. 12-15 meraspides). It is reasonable to assume that the proetide protaspis type was present
throughout the group.

text-fig. 16. Late protaspides of various families within Proetida to show synapomorphy of tapering glabella
(which falls short of border in these late protaspides); compare text-fig. 10 /?— m. a , Scharyia (Scharyiinae ;
Aulacopleuridae) (after Snajdr 1980, text-fig. It/), x 20; b, Phorocephala (Telephinidae) (after Chatterton 1980,
fig. 5), x40; c, Dimeropyge (Dimeropygidae) (after Tripp and Evitt 1983, pi. 31, figs. 16-21), x40; d,
Bathyurus (Bathyuridae) (after Chatterton 1980, fig. 5, and pi. 6, fig. 4), x25; e, Proetus (Proetidae) (after
Chatterton 1971, fig. 15c and d), x20 ;/, Aulacopleura ? (Aulacopleurinae ; Aulacopleuridae) (after Hu 1971,
pi. 23, fig. 8; note: Hu assigned the species from which this protaspis is taken to the Proetidae and the genus
Phaseolops \ it seems to the writer far more likely to be an aulacopleurid ; also Hu’s drawing of the protaspides
(his text-fig. 53 a-d) misinterprets the glabellar shape, and the illustration here is based on the photograph
cited), g, heterochronic hypothesis for the origin of the proetide type protaspis from the conventional

libristomate development (see text-fig. 9).

560

PALAEONTOLOGY, VOLUME 33

Because the protoglabella in this kind of protaspis falls somewhat short of the border the proto-hypostome
corresponding with the forward part of the glabella must lie a little to the posterior when compared with the
ptychoparioid protaspis. At the smallest protaspis stages (at which the rostral plate has not yet developed)
Tripp and Evitt (1983) show the proto-hypostome attached to the doublure in Dimeropyge , and Chatterton
(1971) has indicated that the same applied to the earliest proetid growth stages. Snajdr's (1981) full ontogeny
of Scharyia shows that the preglabellar field lengthens during protaspis ontogeny. By the earliest meraspis the
cranidial border is clearly developed and Snajdr's plate 2, fig. 7 also shows that the genal doublure is both
coincident with the border, and much narrower (sag.) than the preglabellar field separating the tapering
glabella from the border. The hypostomal condition must have been natant. The same argument can be
repeated for the meraspis cranidia of the Proetida shown in Fortey and Owens (1975, fig. 3), spanning most
of the proetide families. A full ontogeny of hystricurids is not described, but a very small cranidium of
Parahystricurus oculirotundus Ross (1951, pi. 12, fig. 33) shows identical features, and, because the cephalic
doublure is known to be narrow in that species, it is likely that the early onset of the natant hypostomal
condition applied also to the hystricurids. It is possible that the natant condition extended even into the
protaspis in Scharyia , which has a wide area in front of the proto-glabella from the earliest protaspis stage. The
spinose margin of the larval hypostome is a character shared by almost all trilobites (Fortey and Whittington
1989), and could be argued to be a synapomorphy of the whole group. But in Proetida for which larval
hypostomes have been described (e.g. Proetus, see Fortey and Chatterton 1988, fig. 13; Dimeropyge , see Tripp
and Evitt 1983) the spinose margin is apparently lacking. In this respect these larval hypostomes are more like
that of the adult, and differ from immature hypostomes of all other trilobites.

The Proetida can thus be defined within the Libristoma by their protaspis type, and by the ‘displacement
back’ of the natant hypostomal condition, and probably the non-spinose hypostomal margin, into early
ontogeny. Because some stratigraphically late Proetida (above) become secondarily conterminant and
impendent it is likely that there were reversals (i.e. natant to conterminant) in the ontogeny of such forms, but
because they are undisputed proetids this does not pose a problem for the definition of the group as a whole.
What is less certain is whether the neat ‘cut-off’ at the Cambrian-Ordovician boundary suggested by Fortey
and Owens (1975) applies with the definition of the Order given here. Fortey and Owens based their
stratigraphic definition of the group on the hypothesis that Proetida were all derived from hystricurines. While
this is reasonable for dimeropygids and bathyurids (and probably telephinids) for which there are a series of
intergrading species between hystricurids and the derived families, this is much less clear for proetids and
aulacopleurids. It now seems perfectly possible that one or more Upper Cambrian families will prove to have
a proetide protaspis - and if this is the case they, too, should be assigned to Proetida. Solenopleuridae is an
obvious candidate; some Cambrian solenopleurids seem to me indistinguishable from Ordovician Hystricurus.
Ontogenetic studies on trilobites in the later Cambrian, or even the mid Cambrian, might well reveal an
extension of Proetida from its present post-Cambrian assemblage of families. Dimeropygidae are better placed
in the superfanuly Bathyuracea, rather than in Proetacea as in the 1959 Treatise.

Order asaphida Salter, 1864, emend. Fortey and Chatterton, 1988

Diagnosis. Libristoma having ventral median suture. Only primitive forms retain natant hypostomal
condition, the majority of the group being conterminant or impendent. Higher Asaphida also have
a distinctive inflated protaspis - the asaphoid protaspis.

This major group was defined at length by Fortey and Chatterton (1988) to which paper the reader is referred
for details. It was treated as a suborder there, pending further discussion of the ptychopariid problem. If the
concept of the Libristoma is accepted its rank can be elevated to Order as discussed previously. Shergold and
Sdzuy (1984) have already employed Asaphida as an order, although with a more restricted view of what
trilobite groups should be placed within it than that of Fortey and Chatterton (1988). The superfamilies
Anomocaracea, Asaphacea (incorporating Ceratopygacea), Remopleuridacea, Dikelocephalacea, and Tri-
nucleacea were included in Asaphida in that work, together with a number of additional families, such as
Pterocephahidae and Dikelokephalimdae. To this list can be added the family Parabolinoididae. Westrop
(1986) has recently described a ventral median suture for this group, and because this is the main
synapomorphy of Asaphida this is where the family belongs unless it can be shown that the median suture was
independently derived therein. Parabolinoididae are interesting in that the bulk of their morphology is almost
an inventory of primitive and generalized ptychoparioid characters (Fortey and Chatterton 1988, table 2). As
Westrop has pointed out they resemble olenids in many features - but we know that olenids were derived from

FORTEY: TRILOBITE CLASSIFICATION

561

ptychoparioids with a rostral plate, not a median suture, and on our analysis of the characters, therefore, such
resemblances are because of symplesiomorphies, and misleading.

The placing of Asaphida within the larger group Libristoma is proved by the apparently universal occurrence
of the natant hypostomal condition in the primitive members of the group, which are also the stratigraphically
earliest. These include the families Anomocaridae, Auritamiidae, Pterocephaliidae, Parabolinoididae, listed
above, and probably Liostracimdae (Fortey and Chatterton 1988, p. 879). For the moment these have been
retained in a paraphyletic group Anomocaracea - based on the retention of the natant condition in
combination with the median suture - but further analysis should ally these families with one or another of the
derived monophyletic groups.

Order ptychopariida Swinnerton, 1915
Suborder olenina Burmeister, 1843

Diagnosis. Ptychopariida with thin dorsal exoskeleton; free cheeks yoked together as single unit
connected by narrow (sag.) doublure (rostral plate present only in the most primitive members of
Olenus).

Olenids comprise an accepted monophyletic group (Henningsmoen 1957). Most occurred in dysaerobic
palaeoenvironments. Probably more than in any other trilobite group many relationships within the olenids
have been determined stratigraphically, especially because of the work of Westergard, Flenningsmoen and
Kaufmann in Scandinavia. This is possible because their record in the Upper Cambrian ‘olenidskiffer’ is
exceptionally continuous. I see no reason to doubt the phylogenies resulting from this work, even though it has
to be said that subfamily relationships are not clear. It is a paradox that the very completeness of the olenid
phylogenetic history has probably hindered phylogenetic study of other Cambrian groups. The olenid model
dictates that if enough collections are made and enough stratigraphy is done the phylogeny will simply ‘fall
out’ from sequence, allowing a direct reconstruction of the historical tree of descent. However the bulk of the
olenids occupied a distinct, stable and peculiar environment, certainly dysaerobic and possibly beneath the
thermocline (Fortey 1974; Cook and Taylor 1975), which favours the preservation of a continuous phyletic
succession of species. This does not apply to the greater part of the shelf record, nor even to deeper
environments in which oxygenation was normal.

As Henningsmoen (1957) noted, the morphological excursions taken by olenids in their long Cambro-
Ordovician history makes the framing of a diagnosis very difficult. I introduced the fused free cheeks into the
diagnosis (Fortey 1974) and this has been subsequently adopted also by Nikolaisen and Henningsmoen (1985).
This fusion is true of all olenids for which the material is adequate to see it, other than the earliest species of
Olenus , which still have the rostral plate. These species of Olenus form the sister group of the rest of the
Olenidae. Because they include the type species O. gibbosus (Wahlenberg) it is obviously necessary to ‘draw
the line’ for the family below these species. An important point is that the fused cheeks of olenids are not
homologous with those of Asaphida which develop them (Kainellidae, Nileidae and Cyclopygidae) because
they result from the loss of a rostral plate, not a median suture. All olenids except possibly the stratigraphically
youngest Triarthrus (Whittington and Almond 1987) apparently retain the natant hypostomal condition, but
it would not be surprising to find other examples of secondary conterminant hypostomal condition among
pelturines with reduced preglabellar fields. In any case (cf. Nikolaisen and Henningsmoen 1985) one cannot
incorporate the natant condition into the diagnosis as it is the general character of Libristoma.

There is now the greater problem of whether families other than Olenidae can be assigned to Olenina.
Various attempts have been made in this direction. One problem is that unrelated trilobites may have
‘olenimorph’ features (such as numerous narrow thoracic segments with wide pleurae, long genal spines,
caecate dorsal surface) because they, too, became adapted to the dysaerobic environment. Good examples
would be Hedincispis among the Asaphacea, Aulacopleura among Proetida, Seleneceme among Trinucleacea.
In the Treatise (Poulsen in Moore 1959) the family Papyriaspididae was included ; as we have noted above there
is no evidence to support this, apart from generalized ptychopariid similarity. Palmer (1965) has stated that
Aphelaspis, which is not a pterocephaliid, might be related to Olenidae. Shergold (1980) included Elviniidae in
Olenacea. Westrop (1986) included not only this family, but Idahoiidae, Parabolinoididae, and Ptero-
cephalhdae in the same superfamily. In the analysis of Fortey and Chatterton (1988) these last three would be
part of Asaphida, having a median suture. The problem is that I can find no synapomorphic characters to link
any of these families to Olenidae in particular. The characters they share are all those of primitive
ptychopariids. Westrop (1986) states that he followed Ludvigsen and Westrop (1983) in the reasons for
associating his selection of families with Olenidae, but in the latter paper there are no characters listed that

562

PALAEONTOLOGY, VOLUME 33

might unite such a group. One has to conclude that the association is made on ‘general' resemblance and on
the stratigraphic presumption that they are all similar enough in age to be likely to constitute a natural group.
Obviously Olenidae had to have a sister group, and I am certain that a major libristomate group including the
olenids will be recognizable. At the moment I can see no characters to define such a group. For this reason
Olenina are retained as a suborder of Ptychopariida faux de mieux. I think it wiser to restrict the scope of
Olenina to a genuinely monophyletic group, rather than cram in other families on spurious grounds with the
risk of creating a polyphyletic taxon.

Suborder harpina Whittington, 1959

Diagnosis. Libristoma having small eyes placed anteriorly relative to glabella; cephalon broadly
convex (sag. tr.) with long preglabellar area, and often a brim, with stout genal spines or genal
prolongations of the brim; free cheeks narrow (tr.) or lost altogether, fused together; thoracic
segments numerous (12 or more) and characteristically narrow (sag.) with long pleurae in contact
along their length; pygidium relatively small and often transverse.

The natant condition of the hypostome of Harpides is clearly illustrated on the holotype of the type species
illustrated by Barrande (1872), and was suggested for Harpetidae by Whittington (19886, p. 328); I know of
no harpinid in which the hypostome has become secondarily conterminant. Even the earliest harpidinids
(which are found at an horizon close to the Cambrian-Ordovician boundary) are specialized in most respects.
There is no better hypothesis available for their Cambrian ancestry than that they are related to the Upper
Cambrian to early Ordovician genera Loganopeltis and Loganopcdtoides. Loganopeltoides still has narrow free
cheeks, defined by sutures which are subparallel and close (Rasetti 1945). Rasetti (1948) demonstrated how a
harpid-like cephalon could result from ‘closure’ of these sutures leaving the eyes isolated within the genal
fields, a situation seen in the Harpides- like genus Loganopeltis.

All Harpina have the doublure fused as a single piece, as do loganopeltids and entomaspids (Rasetti 19526;
Ludvigsen 1982) which were placed in Harpina in Moore (1959). However, if the Harpina belong within the
Libristoma the ancestor from which the group derived presumably carried a rostral plate, which was lost in
the same fashion as in Olenina. At some stage also there must have been a change from a normal opisthoparian
facial sutures. Ludvigsen (1982) assigned the genera Heterocaryon and Bowmania to Entomaspididae, both of
which have normal opisthoparian sutures. He also placed the Entomaspididae in the superfamily
Solenopleuracea, and hence presumably regarded them as unrelated to Harpina. If this were the case then
presumably the harpinid features of entomaspids would have to be evolved in parallel, which would be hard
to prove because Solenopleuracea has no autapomorphies. Ludvigsen also drew attention to the resemblance
between entomaspids and the older genus Doremataspis Opik, 1967 -which Opik placed in another
superfamily, Liostracinacea. This is some measure of the confusion pertaining in Cambrian trilobite
classification. It is worth mentioning that there are some similarities also between Harpina and the mostly
Upper Cambrian Superfamily Norwoodiacea : anterior position of small eyes, stout genal prolongations, wide
and flat thoracic pleurae, and a tendency for the reduction of free cheeks (Norwoodia has strongly proparian
sutures which have become almost parallel in the fashion of Loganopeltoides). Again, homoplasy will no doubt
be invoked to explain these similarities. The best that can be said is that the Harpina have no recognized sister
group, and to follow existing usage in retaining it as a suborder of Ptychopariida.

Suborder ptychopariina Swinnerton, 1915

Diagnosis. A paraphyletic libristomate group retaining primitive characters; a combination of:
relatively narrow rostral plate present, glabella barrel-shaped to tapering; natant hypostomal
condition (possibly some secondarily conterminant); hypostome unspecialized; protaspis with
forward-expanding axis reaching far forwards; with few exceptions micro- to heteropygous.

It is now possible to return to the ‘ptychopariid problem’ in the context of the recognition of the monophyletic
groups listed previously. Having ‘hived off’ these natural groups there remain a collection of families which
do not have the derived characters typifying these groups (or in which such characters are not known because
of the poor state of knowledge of the ventral sutures and/or ontogeny of the group). What characters there
are are often those of Ptychoparia itself, and primitive for Libristoma as a whole. Because we cannot yet assign
such trilobites as more closely related to one or another of the derived groups there is little choice but to retain

FORTEY: TRILOBITE CLASSIFICATION

563

a Suborder Ptychopariina for their reception. Because they include the primitive libristomates they are
considered here to have descended from a common ancester; because they also will include the sister taxa of
Proetida, Olenina, Asaphida, and presumably Harpina, the Ptychopariina constitutes a paraphyletic
assemblage. Ptychopariina can only be defined as being libristomate, but not having one or another of the
synapomorphies that define the monophyletic Orders or Suborders. This is scarcely satisfactory, and to a strict
cladist not acceptable at all. I would regard it only as a convention to accommodate primitive libristomates
until they can be related to one or another of the derived groups. Before this can be done, I believe more
characters will have to be brought into play than I have considered in this paper.

Such a concept of the Suborder Ptychopariina would result in the inclusion of the following superfamilies
(uncritically culled from Moore 1959): at least some Ellipsocephalacea (transferred from Redlichiida, see
above); Ptychopariacea ; Conocoryphacea; Crepicephalacea; Illaenuracea ; Solenopleuracea ; Asaphiscacea;
Komaspidacea (excluding Telephinidae which are Proetida); Raymondinacea ; Norwoodiacea (unless they
prove to belong in the Harpina); and Marjumiacea. Some groups included in Ptychopariida in the Treatise
(Moore 1959) would be excluded because they were probably not natant, especially Leiostegiacea and
Damesellacea. I return to these below.

One can be critical of almost all the superfamilies of Ptychopariina as natural groups, and this has been
reflected in a reluctance to employ these higher categories by more cautious authors such as Robison (1971)
and Zhang and Jell (1987). The whole group needs proper phylogenetic analysis. It can be noted that:

1. Conocoryphacea is almost certainly polyphyletic (contrast Eldredge 1977). The uniting character is loss
of eyes. This character is known to be derived many times in the history of the trilobites, for example, in
Proetida, Phacopida, Trinucleacea and Agnostina. A comparison of the contrasting glabellar structures of such
conocoryphacean genera as Bailiella , Conocoryphe, Hospes , Hartshillia and Shumardia, shows that they are
more likely to be more closely related to disparate (but eye-bearing) ptychopariids than to one another. For
example, Bailiaspis might be related to Solenopleura, Conocoryphe to Ptychoparia , on the basis of their cephalic
morphology. The conocoryphaceans do not even ‘lose’ the eyes in the same way: Bailiella has short,
‘marooned’ eye ridges and comparatively wide fixed cheeks, while Conocoryphe and Meneviellci have long
genal ridges and marginal sutures. The Conocoryphacea has simply been used as a taxon to collect together
blind ptychopariids which are merely the result of (especially Middle Cambrian) adaptations to off- or outer-
shelf atheloptic situations.

2. It is scarcely possible to produce a differential diagnosis of Ptychopariacea, Solenopleuracea and
Marjumiacea which presumably should be possible if any one of them is a natural group. As it is, they
encompass only primitive ptychoparioid morphologies.

3. Illaenuracea and Raymondinacea are strikingly heterogeneous groups, which are not definable; some of
the families within these groups have already been reassigned (e.g. by Westrop 1986), and the coherence of
those remaining is in doubt.

4. Komaspidacea and Norwoodiacea are probably natural groups, but as with the Olenina their sister
group relationships are uncertain.

These problems are enough to show that we have no idea at present on how to dismantle the Ptychopariina
as a paraphyletic group, even with the restriction on the definition of the group I have introduced in this paper.
Further stratigraphical studies may allow for the recognition of monophyletic subgroups, but if the lack of
success in the past is any guide, it seems most unlikely that stratigraphic series will do more than resolve a few
generic relationships, leaving the broader issues untouched. The future seems to me to he in the recognition
of monophyletic groups, and the recognition of their sister groups within the Ptychopariina, which will allow
those groups to be ‘hived off’ into the derived group (similar to the way in which Fortey and Chatterton (1988)
placed the ‘ptychoparioid’ anomocaraceans into the Asaphida). Even so there will probably remain a core of
ptychopariids sharing only the primitive grade of libristomate organisation, and differing only in characters of
the most trivial kind (see, for example, the ptychopariid subfamily Antagminae, Rasetti in Moore 1959, p. 236).
These will likely remain classified together, as perhaps they should.

CLASSIFICATION OF TRILOBITE GROUPS FORMERLY REGARDED AS

PTYCHOPARIIDS

If it is correct to regard the Libristoma as a monophyletic group, trilobites which do not have natant
hypostomal condition (or did not have a natant ancestor) are classified elsewhere. They are not part
of the ‘ptychopariid problem’ except insofar as they have been classified with Ptychopariida in the
past. This applies to the following superfamilies in the 1959 Treatise : Damesellacea, Leiostegiacea

564

PALAEONTOLOGY, VOLUME 33

(including Tsinaniidae), Illaenacea (which excludes proetids and relatives) and Calymenacea by
virtue of their supposed ptychopariid origins. The important question about non-natant groups
where their ancestry is uncertain is whether they were primitively conterminant or whether they
were secondarily so, having been derived from a natant ancestor in the manner of most Asaphina.
Unless there is evidence to the contrary it is reasonable to assume that they were primarily
conterminant, this being, of course, the primitive condition. This assumption may not prove to be
justified, especially with regard to the Phacopida.

The superfamilies removed from Ptychopariida ( sensu 1959 Treatise ) can be accommodated
within other non-libristomate groups, as follows. Detailed evidence will be reviewed elsewhere.

1. Superfamily Illaenacea. The families Illaenidae, Styginidae ( = Scutelluidae) and Phillipsinellidae
belong in a monophyletic group termed the suborder Scutelluina (emend, from Superfamily
Scutelloidea Hupe, 1953) by Lane and Thomas (1983), for which I retain the older name Illaenacea.
This may be part of a larger group to be included within the Order Corynexochida, as noted by
Pribyl and Vanek (1971). Corynexochida is usually considered to be Cambrian, and inclusion of
Illaenacea would extend its stratigraphical scope to the Devonian.

2. Superfamily Leiostegiacea. With the inclusion of Annamitella in the Leiostegiacea (Fortey and
Shergold 1984) the range of this superfamily extends from the Cambrian until well into the Middle
Ordovician. Few articulated leiostegiaceans are known and our knowledge of the ventral surface is
accordingly limited; articulated material of Leiostegium was described by Jell (1985), who
concluded that a rostral plate was present. The likely conterminant hypostomal condition of
leiostegiaceans is shown by the sharp termination of the glabella at the border, at which there are
prominent pits to articulate with the anterior wings of the hypostome (text-fig. 17), and by the fact
that where the anterior cephalic border becomes wider, as it does in Palocorona Shergold, 1975,
backward-curving paradoublural lines are developed comparable with those on Proceratopyge (see
Jago 1987) which are associated with increase in width of doublure medially in conterminant
condition. Hence there is no evidence that Leiostegiacea belong to Ptychopariida (or Libristoma).

a be

text-fig. 17. Leiostegiacea and possible relatives, a, cranidium of Leiostegium , early Ordovician, showing
prominent anterior pits at border suggesting conterminant hypostomal attachment, x3; i, cranidium of late
Cambrian Peichiashania (after Shergold 1980, pi. 27, fig. 1), x2-5; c, early Ordovician Theamataspis
(Scutelluidae) (after Fortey 1980), x 5. If preoccipital muscle impressions in b and c are homologous, this may
provide evidence of leiostegiacean ancestry of Illaenacea via tsinaniids (see text).

It seems possible that the Leiostegiacea also belongs within the Corynexochid/Illaenacean group.
Many leiostegiaceans show a corynexochoid glabellar form: Leiostegium (Jell 1985, pi. 22, fig. 11;
Shergold 1975, pi. 45, fig. 7), Pagodia ( Oreadella ) (Shergold 1975, pi. 36, fig. 1) Annamitella (Fortey
and Shergold 1984, pi. 34, figs. 1-5, 7) and Parakoldinioidia (see Taylor and Halley 1974) from
various families. The last-named is a member of the family Missisquoiidae, which was assigned to
the Leiostegiacea by Shergold (1980), but which has been claimed (Ludvigsen 1982) as lying at the
‘root’ of the post-Cambrian radiation of the styginids (a contrary view was expressed by Lane and
Thomas 1983). Shergold (1975) assigned the effaced family Tsinaniidae to the Leiostegiacea; in the
Treatise (Lochman-Balk in Moore 1959) this family was considered to belong to Asaphacea (and

FORTEY: TRILOBITE CLASSIFICATION

565

hence to Libristoma). I think Shergold is correct in removing tsinaniids from Asaphacea. For all
their general dorsal effacement, tsinaniids show prominent muscle impressions lying outside the
axial furrow opposite the posterior part of the palpebral lobes. These are like the ‘lunettes’ of
illaenids, and the homologous impression on styginids, and may be a synapomorphy. A similar
impression exists on some more conventional leiostegiaceans, such as Peichiashania (Shergold 1980,
pi. 27, fig. 1 ; see text-fig. 17 herein). These structures are distinctive enough to suggest they may be
homologues. All the evidence combines to indicate that leiostegiaceans may be part of an enlarged
illaenid/styginid/corynexochoid group.

3. Superfamily Damesellacea. The damesellaceans are regarded as the primitive sister group of the
Odontopleurida, and should be placed in that Order. As in that group the hypostomal condition
(see, for example, Chatterton and Perry 1983) was conterminant, and there is no evidence to suggest
natant ancestry. Characters shared between odontopleurids and damesellaceans include: 1, narrow,
(sag.) ledge-like anterior cranidial border, which the glabella abuts sharply; 2, deeply scribed and
inflated eye ridges; 3, transverse hypostome with relatively wide posterior border; 4, spinosity,
especially pygidium and anterolateral margins of free cheeks (not the odontopleurid Selenopeltis);
5, occipital tubercle without thoracic homologues (where known, apparently of Fortey and
Clarkson’s 1976 type D); 6, third glabellar lobe reduced or absent. The autapomorphy of
Odontopleuridae is the unique bi-spinose thoracic pleural tip, not present on Damesellacea {fide Lu
et al. 1965, pi. 72, fig. 6). The apparent stratigraphic gap between Damesellacea (mid to late
Cambrian) and Odontopleuridae (early Ordovician to Devonian) has now been bridged with the
recognition by Bruton (1983) and Thomas and Holloway (1988) of Acidaspides praecurrens
Lermontova, 1951, from the Upper Cambrian of Kazakhstan, as a true odontopleurid; the
pygidium of this species (Thomas and Holloway 1988, pi. 16, fig. 352) shows more axial segments
than Ordovician odontopleurids, but is identical to certain damesellaceans in this regard. Thomas
and Holloway (1988) joined many previous authors since Swinnerton (1915) in regarding
Odontopleurida and Lichida as together constituting a monophyletic group (as repeated on text-fig.
19). This is not the only possibility, because as many characters support the Styginidae as the sister
group of Lichida. Further investigation of damesellids should cast light on this matter.

THE PROBLEM OF THE ORDER PHACOPIDA

The Order Phacopida Salter, 1864 is the largest group outside Libristoma, and may eventually
prove to belong within it. In current usage it comprises the Suborders Cheirurina, Calymenina and
Phacopina. The Cheirurina and Phacopina are accepted as monophyletic groups, and the
Calymenina usually so, but there is no consensus about their mutual relationships. Both
Henningsmoen (in Moore 1959) and Bergstrom (1973) indicated ‘ptychopariid’ origins for
Phacopida. The diagnosis of Phacopida by Henningsmoen (in Moore 1959) includes no characters
peculiar to the group. Eldredge (1977, fig. 8) claimed that Calymenina were unrelated to Phacopina
and Cheirurina, and placed them instead as the sister group of ‘some Ptychopariina ’. This certainly
brings the Calymenina within the remit of the ‘ptychopariid problem’. The alliance of Calymenina
with Phacopida and Cheirurina was based on the close resemblance between their protaspides
(Whittington and Evitt in Whittington 1954; Whittington in Moore, 1959; Chatterton 1971),
including the presence of three pairs of prominent marginal spines, of which only two are cephalic
(the ‘phacopoid protaspis’). Bergstrom (1973) placed all three suborders (demoted to superfamilies)
in different orders: Odontopleurida (Cheiruracea), Phacopida (Phacopacea) and Ptychopariida
(Calymenacea). Again, this shows how different are opinions about high level trilobite classification.
But whatever classification is closest to the truth, Phacopida have to be discussed in relation to
Libristoma.

A list of possible synapomorphies can be cited as supporting the concept of Phacopida used in
the Treatise (Henningsmoen in Moore 1959), shared between Calymenina, Cheirurina and
Phacopina. These are:

566

PALAEONTOLOGY, VOLUME 33

1. The phacopoid protaspis type.

2. Loss of concentric terrace ridges on the cephalic doublure. As discussed above, cephalic
doublural terrace ridges are primitive for the Trilobita and retained in Libristoma; their loss is
significant.

3. Presence of hypostomal wing process. Although not widely figured this has been shown on
representatives of all three groups by Whittington (19886), and on an early cybeline by Fortey (1980).

4. Granulose surface sculpture. A characteristic densely granulose surface sculpture appears to
be ubiquitous on early representatives of all three major groups (later Phacopina have much
modified sculpture, but all stratigraphic evidence shows this to be secondary).

5. Absence of pygidial border in more primitive members of the group. Pleural and interpleural
furrows usually extend very close to the margin and do not stop at a marginal upturned pygidial
rim or flattened border so common in Libristoma. A ‘pseudoborder’ may appear on certain
homalonotids such as Plaesiacomia (Henry 1980, pi. 23, fig. 4), but this is connected with co-
aptation, and is not homologous with borders of other trilobites.

6. Five or more pygidial segments. This number is known to be secondarily reduced in later
Cheiruridae.

This list is sufficient to suggest that the Treatise (Moore 1959) Phacopida does constitute a natural
group. I cannot find any of the alternative suggestions which have been made regarding ordinal
relationships of Calymenina and Cheirurina supported by more characters which are likely to be
synapomorphies. For example, of Bergstrom’s (1973, p. 41) list of odontopleurid characters there
are only two (nature of thoracic articulation and loss of cephalic doublure terrace ridges) which
would support the inclusion of Cheiruracea in Odontopleurida.

If Phacopida includes Calymenina, Cheirurina, and Phacopina it is likely that the suborder
Calymenina is the primitive sister group of Cheirurina + Phacopina, as shown by the distribution of
characters summarized in text-fig. 18. Detailed discussion of these characters is beyond the scope
of this paper, except to note that calymenaceans retain several primitive characters, such as the
circum-ocular suture (uniquely among post-Cambrian trilobites) and tapering glabella. The
question is whether any calymenines show evidence of having an ancestor with natant hypostomal
condition, which would imply that they, like the Asaphida, should be included within Libristoma.
All Phacopina and Cheirurina have attached, conterminant hypostomes, and we would expect natant
ancestry to be revealed in the most primitive calymenaceans. Neseuretus is close to the common
ancestor of Calymenidae and Homalonotidae, and, as noted above (text-fig. 3 a), was conterminant.
Probably still more primitive calymenaceans have clear preglabellar fields, and include Prionocheilus
Rouault, 1847 (= Pharostoma of authors) and Bavarilla Barrande, 1868. Whittard (1960, pi. 18, fig.
4) illustrated a posterior extension of the doublure/rostral plate in Prionocheilus (after the manner
of text-fig. 6), and in this form the hypostome was conterminant. So all Phacopida from the early
Arenig onwards apparently have attached hypostomes, and the likelihood is that the relevant
trilobites will be found in rocks of Tremadoc age or older. Bavarilla is Tremadocian, and was
discussed in detail by Sdzuy, 1955. Most of its characters are primitive ones, even for libristomates :
its hypostome is like that of generalized libristomates (Sdzuy 1955, pi. 6, fig. 50); it retains a pygidial
border; it has a pygidium with only two (three?) segments; it does not show basal (‘alar’) muscle
impressions in the cephalic axial furrows. In fact, its only likely phacopide character is the epifacetal
type of pleural furrow (Sdzuy, pi. 6, fig. 55). Given that it also had narrow genal doublure (Sdzuy,
pi. 6, fig. 52) and a preglabellar field it is not unreasonable to infer that the hypostomal condition
was natant, unless the rostral plate were extended in the manner of Prionocheilus. So the position
of Bavarilla as sister taxon of the Phacopida relies on a single character - and stratigraphialc
position, which may mislead. Since its hypostome is not known in situ its natant character is
inferred, and on this alone would reside the placing of the Phacopida within the Subclass
Libristoma.

I would be more enthusiastic about including Phacopida in Libristoma if some trilobites existed
which had more of the phacopide synapomorphies whilst retaining the natant hypostomal
condition. There are some late Cambrian trilobites which show other similarities to Calymenidae.

FORTEY: TRILOBITE CLASSIFICATION 567

emphasizing synapomorphies regarded as important, with Bavarilla as the sister group of higher Phacopida ;
retained plesiomorphic characters are shown as circles. Whether or not this group is a member of Libristoma
depends on whether the sister taxon was natant (see text, p. 566). Primitive members of each group are
illustrated : Calymenina, Neseuretus; Cheirurina, Pliomerops\ Phacopina, Ormathops. Characters: 1, phacopoid
protaspis type (protaspis of Bavarilla is unknown and may prove of this type); 2, loss of concentric terrace
ridges on cephalic doublure; 3, presence of hypostomal wing process; 4, densely granulose surface sculpture;
5, loss of pygidial border; 6, five or more pygidial segments (this character undergoes reversal in later
Cheiruracea) ; 7, proparian suture; 8, forward-expanding glabella (there are a very few exceptions to this in
Cheirurina + Phacopina); 9, loss of circum-ocular facial suture; 10, spongy-reticulate genal prosopon; 11,
gonatoparian style sutures; 12, muscle pad in ‘alar’ position; 13, muscle pad at inner end of SI (which may
otherwise bifurcate); 14, loss of falcate pleural tips; 15, epifacetal pleural furrow; 16, schizochroal eye; 17,
characteristic sigmoidal form of 3S; 18, loss of rostral plate by fusion of free cheeks (except the primitive
phacopide Gyrometopus Jaanusson, 1975, which retains it); 19, spinose pygidial margin; 20, cheiruroid type
of thoracic articulation (see Bergstrom 1973); 21, only two deep pygidial pleural, no interpleural furrows.
Retained primitive characters: 22, circumocular sutures; 23, tapering glabella; 24, opisthoparian sutures; 25,
generalized ptychoparioid glabellar furrows; 26, strong pygidial border.

568

PALAEONTOLOGY, VOLUME 33

Calymenidius Rasetti, 1944 has granulose surface sculpture, and similar glabellar structure to
calymenids. There are no ‘alar’ muscle impressions on Cambrian Calymenidius spp., although they
were present on an early middle Ordovician species described as aff. Calymenidius sp. ind. by
Whittington (1965, pi. 59, figs. 10, 12-15), a species which might now be referred to Protocalymene
Ross, a calymenid. Cliffia (see also Kirengina Ogienko, 1974) is worth mentioning because its
pygidium (e.g. Westrop 1986, pi. 27, figs. 4, 6) is at least superficially like those of early calymenids
and Phacopina, and the cranidium can be compared with that of Neseuretus in the anterior position
of the eye, glabellar shape, and the like. It does not show any distinctive cephalic autapomorphy
of Calymenina, and nothing is known of the hypostomal attachment. Search for Cambrian sister
taxa of Phacopida may not be fruitless. In the meantime the evidence of Bavarilla and
lonchocephalids such as Calymenidius is hardly enough to place Phacopida within Libristoma, to
which it might well belong.

HISTORY OF HYPOSTOME ATTACHMENT

The history of hypostome attachment among the non-olenellid trilobites can now be summarized
(text-fig. 19). It seems likely to be of considerable importance in understanding aspects of feeding
and functional morphology bearing on the adaptive history of the group as a whole. Early
Cambrian redlichiids had conterminant hypostomes. On other early Cambrian trilobites, like
Dolerolenus , the preglabellar field became extended, the cranidial border was extended backwards
as a plectrum, but the hypostome may have become detached from the doublure at later stages in
the ontogeny. Redlichiid conterminant condition was retained and modified in paradoxidids, in
which both the frontal lobe of the glabella, and the middle body of the hypostome became inflated.
Spinose projections on the posterior margin of the hypostome, such as those on paradoxidids, are
a common feature of trilobites with attached hypostomes throughout the Palaeozoic. If it is correct
to identify the ancestry of corynexochid-styginid-illaenid and the odontopleurid-lichid groups
within what would be currently classified as Redlichiida these groups originated in the Lower
Cambrian and retained the primitive conterminant mode of hypostomal attachment until the very
end of their history in the Devonian. By the mid Cambrian corynexochids included superficially
asaphid-like trilobites ( Glossopleura ), and others that resembled Ordovician or later styginids, but
in all of these rostral plate and hypostome were fused together.

The first Libristoma also appeared in the Lower Cambrian among the protolenids, trilobites
which otherwise resemble contemporary redlichiids closely. The major ‘radiation’ of libristomates
was in the middle part of the Cambrian, when highly effaced forms, or genera with eyes reduced or
lost (polyphyletically), are found in different palaeoenvironments, sometimes to the exclusion of
other trilobites. These libristomates may have been the first trilobites to exploit particle-feeding in
deep sea environments. Libristomate hypostomes remained unspecialized for much of the
Palaeozoic. One kind of conservative libristomate morphology persisted with little change from
Elrathia- like forms in the mid Cambrian to Au/acop/eura- like forms in the Devonian. Before the end
of the Middle Cambrian the ventral median suture had developed in ‘anomocaracean ’ asaphide
libristomates, and the earliest secondarily conterminant hypostomal attachment appears in the
asaphide group ( Proceratopyge ) shortly afterwards. In the Upper Cambrian these secondarily
conterminant forms became important components of shelf faunas (Dikelokephalacea) but many
different kinds of libristomates with unattached hypostomes persisted, notably Olenidae, which
successfully occupied dysaerobic environments. Secondarily conterminant Asaphida started to
develop modifications of the hypostome, such as wide borders, and posterior forks. It is interesting
that the diversification of secondarily conterminant forms followed the eclipse of the primarily
conterminant Corynexochida, some of whose morphologies they broadly reproduced. The Upper
Cambrian record is overwhelmingly dominated by Libristoma, but if the relationships suggested in
this paper are correct, primarily conterminant forms related to the lichid-odontopleurid and
stygi nid-illaenid groups persisted through this interval. The Leiostegiacea and Damesellacea are
regarded as primarily conterminant groups that bridge this ‘gap’.

FORTEY: TRILOBITE CLASSIFICATION

569

text-fig. 19. Simplified summary of history of hypostome attachment and the phylogeny of Trilobita, drawn
as a tree. Secondary conterminant and impendent forms are derived from proetide sources after the
disappearance of other conterminant clades. This may have happened first in the Silurian, but is not shown here

for reasons of clarity.

570

PALAEONTOLOGY, VOLUME 33

By the early Ordovician natant hypostomes were confined to Proetida, and to a number of
specialized groups persisting from the Cambrian (Trinucleacea, Harpina, Agnostina and Olenidae).
Secondarily conterminant asaphids and remopleurids were important and characteristic com-
ponents of Ordovician faunas, but did not survive the Ordovician-Silurian boundary, at which level
most of those natant families persisting from the Cambrian also became extinct (trinucleids,
agnostids, olenids). At the same time the secondarily conterminant Asaphida also died out.
Phacopida were important from the Ordovician onwards; all those we know had attached
hypostomes, but the question of whether they had a natant, or primary conterminant ancestor is
not satisfactorily resolved (see p. 565). The Odontopleuridae, Lichidae, Styginidae and lllaenidae
with primary conterminant hypostomes diversified in the Ordovician and passed through to the
Silurian and Devonian. Impendent hypostomal condition was common in the Ordovician
(Cylopygacea, Illaenina, later remopleuridids) for the first time. Those groups with attached
hypostomes developed hypostomal specializations in the Ordovician: forks, wide borders, enlarged
maculae, varied surface sculpture. Natant hypostomes remained similar to those of the Cambrian.

From the Silurian onwards the natant condition persisted in Proetida alone, but within this
group certain forms ( Warburgella , Cordania) became secondarily conterminant. Again, one might
speculate that these forms were filling niches unoccupied since the extinction of conterminant
asaphides. In the Silurian impendent phacopids became widespread, and these continued into the
Devonian.

After the end-Frasnian extinction of major trilobite groups, Proetida alone survived. Some of
these were natant. But secondarily conterminant genera, such as Paladin , are also common trilobites
in the Carboniferous - these developed hypostomal specializations like those of other conterminant
trilobites. Certain phillipsiids ( Paraphillipsia , Cyphinioides) became impendent, the first proetides to
be so. I suspect that only secondarily conterminant and impendent hypostomal attachment
persisted into the Permian, and the end of trilobite history (Owens 1983 has illustrated the relevant
genera, none of which has a preglabellar field). So by the Carboniferous one group of Proetida were
exhibiting the whole range of hypostomal attachment modes, although this group as a whole was
primitively, and persistently, natant. As has been recognized for some time, some of these
Carboniferous trilobites show parallelisms to earlier and unrelated groups - such as Phacopida. It
does not seem likely to be coincidence that the proetids show such morphological novelty only after
the demise of the conterminant groups (or impendent groups derived therefrom) which may have
persisted all the way from the Lower Cambrian.

REMAINING PROBLEMS

Further problems in trilobite classification are legion. A number arise out of my analysis.

1. The relationships between the major subgroups of Libristoma, here treated as Orders and
Suborders, are not resolved.

2. The relationships between the other high level taxa having primary conterminant hypostomal
attachment are not resolved, neither between themselves, nor with the Libristoma. These appear
merely as ‘other trilobites’ on text-fig. 14. These comprise, at the least: the Corynexochid/
Illaenid/Leiostegiacean group; the Lichid/Odontopleurid group (if it is one); the Phacopida. Such
evidence as there is, and it is feeble, places the last-named within Libristoma.

3. Redlichiida is only defined by not having one of the advanced characters of the other major
trilobite groups with a rostral plate bounded by connective sutures. It is certainly a primitive group,
and probably a paraphyletic group, which requires further analysis. Stability on the taxonomic
status of higher level taxa cannot be achieved without resolution of some of these problems.

4. All these problems would benefit from phylogenetic analysis using one of the parsimony-based
computer programmes for the production of dichotomous trees. The change from natant to
secondarily conterminant is a major character reversal according to my treatment, and it would be
interesting to see whether this was supported by other methods of analysis.

FORTEY: TRILOBITE CLASSIFICATION

571

Acknowledgements . I cannot adequately express my indebtedness to the various colleagues and friends with
whom I have discussed the ideas in this paper. Early drafts of all or part of the text were read by A. W. A.
Rushton and R. M. Owens. Photographs were supplied by R. A. Robison, W. T. Chang and R. M. Owens;
unpublished information was made available to me by H. B. Whittington, A. T. Thomas and M. D. Brasier.
They do not share the responsibility for the ideas expressed herein, with which they may disagree, although all,
I am sure, would sympathize with the problems that are raised. Claire Mellish drew text-fig. 19.

REFERENCES

ax, f. 1987. The phylogenetic system. J. Wiley and Son, Chichester, 340 pp.
barrande, J. 1868. Reapparition du genre Arethusina Barr. Prague, 29 pp. 1 pi.

- 1872. Systeme silurien du centre de la Boheme. lere partie. Recherches paleontologiques , supplement au
vol. 1. Prague, xxx + 647 pp, 35 pis.

bassett, m. G., Owens, R. m. and rushton, a. w. a. 1976. Lower Cambrian fossils from the Hell’s Mouth Grits,
St Tudwal’s Peninsula, North Wales. Journal of the Geological Society of London, 132, 623-644.
bergstrom, j. 1973. Organisation, life and systematics of trilobites. Fossils and Strata, 2, 1-69.

1977. Proetida - a disorderly Order of trilobites. Lethaia, 10, 95-105.

bornemann, j. G. 1883. Palaontologisches aus dem cambrischen Gebiete von Canalgrande in Sardinien.

Zeitschrift der Deutschen Geologischen Gesellschaft , 35, 270-274.
boxshall, G. a. 1982. On the anatomy of the misophrioid copepods, with special reference to Benthomisophria
palliata Sars. Philosophical Transactions of the Royal Society of London. Series B, 297, 125-181.
brasier, m. d. 1989. Towards a biostratigraphy of the earliest skeletal biotas. 117-165. In cowie, j. w. and
brasier, m. d. (eds.). The Precambrian-Cambrian Boundary. Oxford Monographs on Geology and
Geophysics, 12, 213 pp.

briggs, d. e. G. and fortey, R. a. 1989. The early radiation and relationships of the major arthropod groups.
Science, New York, 246, 241 243.

— and Clarkson, e. n. k. 1988. Extinction and the fossil record of the arthropods. 171-209. In
larwood, G. p. (ed. ). Extinction and survival in the fossil record. Systematics Association Special Volume
No. 34, 365 pp.

bruton, D. L. 1983. Cambrian origins of the odontopleurid trilobites. Palaeontology , 26, 875-885.
burmeister, h. 1843. Die organisation der Trilobiten Berlin, 147 pp., 6 pis.

Campbell, k. s. w. 1977. Trilobites of the Haragan, Bois d’Arc and Frisco Formations (Early Devonian),
Arbuckle Mountains region, Oklahoma. Bulletin of the Oklahoma Geological Survey, 123, 1-139, 40 pis.
chatterton, b. d. e. 1971. Taxonomy and ontogeny of Siluro-Devonian trilobites from near Yass, New South
Wales. Palaeontographica Abteilung A, 137, 1-108, pis. 1-24.

— 1980. Ontogenetic studies of Middle Ordovician trilobites from the Esbataottine Formation, Mackenzie
Mountains Canada. Palaeontographica Abteilung A, 171, 1-74, pis. 1-19.

— and ludvigsen, r. 1976. Silicified Middle Ordovician trilobites from the South Nahanni River area.
District of Mackenzie, Canada. Palaeontographica Abteilung A, 154, 1-106, 22 pis.

— and perry, d. G. 1983. Silicified Silurian odontopleurid trilobites from the Mackenzie Mountains.
Palaeontographica Canadiana, 1, 1-55, 36 pis.

cisne, j. l. 1974. Trilobites and the origin of arthropods. Science, New York, 186, 13-18.
cook, h. e. and taylor, m. e. 1975. Early Paleozoic continental margin sedimentation, trilobite biofacies and
the thermochne, western United States. Geology, 3, 559-562.
dean, w. t. 1972. The trilobite genus Holasaphus Matthew 1895 in the Middle Cambrian rocks of Nova Scotia
and eastern Turkey. Canadian Journal of Earth Sciences, 9, 266-279.

EGOROVA, L. L, SHABANOV, YU YA., PEGEL, T. V., SAVITSKY, V. E., SUCHOV, S. S. and TSCHERNYSHEVA, N. E. 1982.
Mayan Stage of the type locality (Middle Cambrian of the Siberian Platform). Trudy Mezhvedomstvennyi
Stratigraficheskii Komitet CCCP, 8, 1-146, pis. 1-64. [In Russian.]
eldredge, N. 1977. Trilobites and evolutionary patterns. 305-332. In hallam, a. (ed.) Patterns of evolution.
Elsevier, Amsterdam, 591 pp.

— and cracraft, i. 1982. Phylogenetic patterns and the evolutionary process. Columbia University Press,
Columbia, 349 pp.

fortey, R. a. 1974. The Ordovician trilobites of Spitsbergen. I. Olenidae. Skrifter norsk Polarinstitutt , 160,
1 81, 24 pis.

572

PALAEONTOLOGY, VOLUME 33

— 1975. The Ordovician trilobites of Spitsbergen. II. Asaphidae, Nileidae, Raphiophoridae and
Telephinidae of the Valhallfonna Formation. Skrifter norsk Polarinstitutt, 162, 1-125, pis. 1—41.

— 1979. Lower Ordovician trilobites from the Catoche Formation (St George Group), western
Newfoundland. Bulletin of the Geological Survey of Canada , 321, 61-114, pis. 23-37.

— 1980. The Ordovician trilobites of Spitsbergen. III. Remaining trilobites of the Valhallfonna Formation.
Skrifter norsk Polarinstitutt , 171, 1-113, 25 pis.

— 1986. The type species of the Ordovician trilobite Symphysurus : systematics, functional morphology and
terrace ridges. Palaeontologische Zeitschrift , 60, 255-275.

— and chatterton, b. d. e. 1988. Classification of the trilobite suborder Asaphina. Palaeontology , 31,
165-222, pis. 17-19.

— and clarkson, E. n. k. 1976. The function of the glabellar ‘tubercle’ in Nileus and other trilobites.
Lethaia , 9, 101-106.

— and owens, R. m. 1975. Proetida - a new order of trilobites. Fossils and Strata, 4, 237-239.

1979. Enrollment in the classification of trilobites. Lethaia, 12, 219-226.

— 1978. Early Ordovician (Arenig) stratigraphy and faunas of the Carmarthen district, south-west
Wales. Bulletin of the British Museum ( Natural History ) (Geology), 30, 225-294.

— and shergold, j. h. 1984. Early Ordovician trilobites from the Nora Formation, central Australia.
Palaeontology, 27, 315-366, pis. 38—46.

— and Whittington, h. b. 1989. The Trilobita as a natural group. Historical Biology, 2, 125-138.
hammann, w. 1983. Calymenacea (Trilobita) aus dem Ordovizium von Spanien; lhre Biostratigraphie,

Okologie und Systematik. Abhandlungen von der Senckenbergischen Naturforschenden Gesellschaft, 542,
1-177.

hawle, i. and CORDA, a. j. c. 1847. Prodrom einer monographic der bohmischen trilobiten. Prague, 167 pp, 7 pis.
henningsmoen, g. 1951. Remarks on the classification of Trilobites. Norsk Geologisk Tiddskrift, 29,
174-217.

— 1957. The trilobite family Olenidae. Skrifter utgitt av det Norske Videnskaps-Akadami i Oslo. 1.
Mathematics and Natural Sciences, 1, 1-303, pis. 1-31.

henry, j.-l. 1980. Trilobites ordoviciens du Massif Armoricain. Memoires de la Societe Geologique et
Mineralogique de Bretagne, 22, 1-250, pis. 1-48.

hessler, r. r. and newman, w. a. 1975. A trilobitomorph origin for the Crustacea. Fossils and Strata, 4,
437-460.

hu chung-hung 1971. Ontogeny and sexual dimorphism of Lower Palaeozoic Trilobita, Palaeontographica
Americana, 44, 29-115, pis. 7-26.

— 1984. Ontogenesis of Ehmaniella burgessensis Rasetti (Trilobita) from Burgess Shale, Middle Cambrian,
Yoho Park, British Columbia. Transactions and Proceedings of the Palaeontological Society of Japan, 135,
395—400, pi. 76.

— 1986. Ontogenetic development of Cambrian trilobites from British Columbia and Alberta, Canada (Part
II). Journal of the Taiwan Museum, 39, 1-44.

hughes, c. p., ingham, j. k. and addison, r. 1975. The morphology, classification and evolution of the
Trinucleidae (Trilobita). Philosophical Transactions of the Royal Society of London, Series B, 272, 537-607.
hunt, a. s. 1966. Submarginal suture and ventral plate of an agnostid trilobite. Journal of Paleontology, 40,
1238-1240.

hupe, p. 1953. Contribution a l’etude du Cambrien inferieur et du Precambrien III de l’Anti-Atlas marocain.

Notes et Memoires, Service dies Mines et de la Carte Geologique du Maroc, 103, 1-402, pis. 1-24, 99 figs.
jaanusson, v. 1975. Evolutionary processes leading to the trilobite suborder Phacopina. Fossils and Strata, 4,
209-218.

jaekel, o. 1909. Uber die Agnostiden. Zeitschrift der Deutschen Geologischen Gesellschaft, 61, 380-401.
jago, j. b. 1987. Idamean (Late Cambrian) trilobites from the Denison Range, south-west Tasmania.
Palaeontology . 30, 207-232.

Jefferies, R. p. s. 1986. The ancestry of the vertebrates. British Museum (Natural Flistory), London, 474 pp.
jell, p. a. 1975. Australian Middle Cambrian eodiscoids with a review of the superfamily. Palaeontographica
Abteilung A, 150, 1-97, 29 pis.

— 1977. Penarosa netenta, a new Middle Cambrian trilobite from northwestern Queensland. Memoirs of the
Queensland Museum, 18, 199-123, pi. 21.

— 1978u. Asthenopsis Whitehouse, 1939 (Trilobita, Middle Cambrian) in northern Australia. Memoirs of the
Queensland Museum, 18, 219-231, pis. 31-38.

— 19786. Trilobite respiration and genal caeca. Alcheringa, 2, 251-260.

FORTEY: TRILOBITE CLASSIFICATION

573

1985. Tremadoc trilobites of the Digger Island Formation, Waratah Bay, Victoria. Memoirs of the

Museum of Victoria, 46, 53-88, pis. 19-33.

kobayashi, t. 1935. The Cambro-Ordovician faunas of South Chosen. Palaeontology, Part III. Cambrian
faunas of South Chosen with special study on the Cambrian trilobite genera and families. Journal of the
Faculty of Science , Imperial University of Tokyo , (2), 4, 49-344.

- and k.ato, f. 1951. On the ontogeny and the ventral morphology of Redlichia chinensis with description
of Alutella nakamurai new gen. and sp. Journal of the Faculty of Science , Imperial University of Tokyo, (2),
8, 100-143, 5 pis.

lane, p. d. 1971. British Cheiruridae (Trilobita). Palaeontographical Society Monograph, 95pp, 16 pis.

— and thomas, a. t. 1983. A review of the trilobite suborder Scutelluina. Special Papers in Palaeontology,
30, 141-160.

lauterbach, k.-e. 1980. Key events in the evolution of the groundplan of the Arachnata (arthropods)
Abhandlungen des Naturwissenschaftlichen Vereins in Hamburg , n.f., 23, 163-327.

Lermontova, e. v. 1951. Upper Cambrian trilobites and brachiopods from Boshche-Kulya (north-east
Kazakhstan). Trudy Vsesoyuznogo Nauchno-Issledovatel'skogo Geologicheskogo Instituta. (VSEGEI), 175
pp., 21 pis.

lu yan-hao et al. 1965. Trilobites of China. 2 vols. Beijing, 766 pp, 135 pis.

ludvigsen, r. 1979. A trilobite zonation of Middle Ordovician rocks, southwestern District of Mackenzie.
Bulletin of the Geological Survey of Canada, 312, 1-100.

— 1982. Upper Cambrian and Lower Ordovician trilobite biostratigraphy of the Rabbitkettle Formation,
western District of Mackenzie. Contributions. Life Sciences Division, Royal Ontario Museum, 134,
1-188.

— and westrop, s. R. 1983. Franconian trilobites of New York State. Memoirs of the New York State
Museum, 23, 1^15, pis. 1-19.

lutke, f. 1980. Zur evolution der altpalaozoischen Proetina (Trilobita). Senckenbergiana lethaea, 61, 73-144.
mcnamara, k. j. 1986. The role of heterochrony in the evolution of Cambrian trilobites. Biological Reviews, 61,
121-156.

moore, r. c. (ed.) 1959. Treatise on invertebrate paleontology. Part O. Arthropoda 1 Geological Society ol
America and University of Kansas Press, Boulder and Lawrence, xix + 560 pp.

Muller, k. j. and walossek, d. 1987. Morphology, ontogeny and life habit of Agnostus pisiformis from the
Upper Cambrian of Sweden. Fossils and Strata , 19, 1-123.

Nicosia, m. l. and rasetti, f. 1971. Revisione dei trilobiti del Cambriano delFIglisiente (Sardegna) descritti da
Meneghini. Memorie. Atti della Accademia Nazionale dei Lincei, (8), 10, 1-18, 3 pis.
nikolaisen, F. and henningsmoen, g. 1985. Upper Cambrian and lower Tremadoc olenid trilobites from the
Digermul peninsula, Finnmark, northern Norway. Bulletin Norges Geologiske Undersokelse, 300, 1-49.
ogienko, l. v. 1974. Early Ordovician trilobites. 109-130. In ogienko, l. v., byalyi, v. l and kolosnitsyna,
G. R. Biostratigraphy of the Cambrian and Ordovician deposits of the southern Siberian platform.
VOSTSIBNIIGGIMS, Moscow, 207 pp. [In Russian.]

opik, a. a. 1958. The Cambrian trilobite Redlichia : organization and generic concept. Bulletin of the Bureau of
Mineral Resources , Geology and Geophysics , Australia, 42, 1-38, 6pls.

— 1961. The geology and palaeontology of the headwaters of the Burke River, Queensland. Bulletin of the
Bureau of Mineral Resources, Geology and Geophysics, Australia, 53, 1-249, 24 pis.

— 1963. Early Upper Cambrian fossils from Queensland. Bulletin of the Bureau of Mineral Resources,
Geology and Geophysics, Australia, 64, 1-107.

— 1967. The Mindyallan fauna of northwestern Queensland. Bulletin of the Bureau of Mineral Resources ,
Geology, and Geophysics, Australia, 74, 1-404, 51 pis. 2 vols.

Owens, r. m. 1973. British Ordovician and Silurian Proetidae (Trilobita). Paleontographical Society Monograph,
98 pp.

— 1983. A review of Permian trilobite genera. Special Papers in Palaeontology, 30, 15-42.

— and thomas, a. t. 1975. Radnoria, a new Silurian proetacean trilobite, and the origins of the
Brachymetopidae. Palaeontology, 18, 809-822, pis. 95-96.

palmer, A. R. 1957. Ontogenetic development of two olenellid trilobites. Journal of Paleontology, 31, 105-128,
pi. 19.

— 1958. Morphology and ontogeny of a Lower Cambrian ptychoparioid trilobite from Nevada. Journal of
Paleontology, 32, 154-170.

— 1960. Trilobites of the Upper Cambrian Dunderberg Shale, Eureka District. Nevada. Professional Papers
of the U.S. Geological Survey , 334-C, 53-105, 1 1 pis.

574

PALAEONTOLOGY, VOLUME 33

- 1962. Comparative ontogeny of some opisthoparian, gonatoparian and proparian Upper Cambrian
trilobites. Professional Papers of the U.S. Geological Survey, 36, 87-96, pi. 19.

- 1965. Trilobites of the late Cambrian pterocephaliid biomere in the Great Basin, western United States.
Professional Papers of the U.S. Geological Survey, 493, 1-105, 23 pis.

1968. Cambrian trilobites of east central Alaska. Professional Papers of the U.S. Geological Survey, 559-
B, 1-115, 15 pis.

and halley, r. b. 1979. Physical stratigraphy and trilobite biostratigraphy of the Carrara Formation
(Lower and Middle Cambrian) in the southern Great Basin. Professional Papers of the U.S. Geological
Survey, 1047, 1-131, pis. 1-16.

pocock, K. J. 1970. The Emuellidae, a new family of trilobites from the Lower Cambrian of South Australia.
Palaeontology, 13, 522-562, pis. 106-110.

pribyl, a. and vanek, j. 1971. Studie uber die familie Scutelluidae Richter et Richter (Trilobita) und ihre
phylogenetische entwicklung. Acta Universitatis Carolinae Geologica, 4, 361-394.

RASETTi, F. 1944. Upper Cambrian trilobites from the Levis conglomerate. Journal of Paleontology, 18,
229-258.

- 1945. Evolution of the facial sutures in the trilobites Loganopeltoides and Loganopeltis. American Journal
of Science, 243, 44—50, pi. I

- 1948. Cephalic sutures in Loganopeltoides and the origin of ‘hypoparian’ trilobites. Journal of
Paleontology, 22, 25-29, pi. 7.

-1951. Middle Cambrian stratigraphy and faunas of the Canadian Rocky Mountains. Smithsonian
Miscellaneous Contributions, 116, 1-277, 34 pis.

- 1952a. Ventral cephalic sutures in Cambrian trilobites. American Journal of Science, 250, 885—
898.

19526. Cephalic sutures in the Upper Cambrian genus Entomaspis. Journal of Paleontology, 26, 797-802,
pi. 117.

1954. Phylogeny of the Cambrian trilobite family Catillicephalidae and the ontogeny of Welleraspis.
Journal of Paleontology, 28, 599-612.

1965. Upper Cambrian trilobite faunas of northeastern Tennessee. Smithsonian Miscellaneous Collections,
148, 127 pp„ 21 pis.

- 1966. New Lower Cambrian trilobite faunule from the Taconic sequence of New York. Smithsonian
Miscellaneous Collections, 148, 52 pp., 12 pis.

- 1972. Cambrian trilobite faunas of Sardinia. Memorie. Atti Accademia Nazionale dei Lincei, (8), 11,
1-100, 19 pis.

robison, r. a. 1964. Late Middle Cambrian faunas from western Utah. Journal of Paleontology , 38, 510-566,
pis. 79-92.

- 1967. Ontogeny of Bathyuriscus fimbriatus and its bearing on the affinities of corynexochid trilobites.
Journal of Paleontology, 41, 213-221, pi. 24.

- 1971. Additional Middle Cambrian trilobites from the Wheeler Shale of Utah. Journal of Paleontology,
45, 796-804, pis. 89-91.

1972a. Hypostoma of agnostid trilobites. Lethaia, 5, 239-248.

- — 19726. Mode of Life of agnostid trilobites. 24th International Geological Congress, Montreal, section 7,
33-40.

1982. Some Middle Cambrian agnostoid trilobites from western North America. Journal of Paleontology,
56, 132-160.

- 1988. Trilobites of the Horn Dal Formation (late Middle Cambrian), central North Greenland.
Meddelelser om Gronland, Geoscience, 20, 23-103.

- and Campbell, d. p. 1974. A Cambrian corynexochoid trilobite with only two thoracic segments. Lethaia,
7, 273-282.

romanenko, E. v. 1977. Cambrian trilobites from sections on the River Bolshaya Ishe (north-east Altai). Trudy
Instituta Geologii i Geofiziki. Sibirskoe Otdelenie, 313, 161-183. [In Russian ]
ross, R. J. 1951. Stratigraphy of the Garden City Formation in northeastern Utah, and its trilobite faunas.
Bulletin of the Peabody Museum of Natural History, 6, 161 pp., 36 pis.

- 1953. Additional Garden City (Early Ordovician) trilobites. Journal of Paleontology, 27, 633-646.
rouault, m. 1847. Extrait du memoire sur les trilobites du departement d’Hle-et-Vilaine. Bulletin de la Societe
Geologique de France, 4, 309-328.

rushton, a. w. a. 1966. The Cambrian trilobites from the Purley Shales of Warwickshire. Palaeontographical
Society Monographs, 120, 1-55, 6 pis.

FORTEY: TRILOBITE CLASSIFICATION

575

— 1983. Trilobites from the Upper Cambrian Olenus Zone in central England. Special Papers in
Palaeontology , 30, pis. 14-19.

Salter, j. w. 1864. A monograph of British trilobites. Part 1. Palaeontographical Society Monograph , 1-80,
pis. 1-6.

sanders, h. l. 1963. The Cephalocarida : functional morphology, larval development, comparative external
anatomy. Memoirs of the Connecticut Academy of Arts and Sciences , 15, 1-80.
schmalfuss, h. 1981. Structure, patterns and function of cuticular terraces in trilobites. Lethaia, 14, 331 -341
SCHOCH, R. M. 1986. Phytogeny reconstruction in paleontology. Van Nostrand Reinhold, New York,
339 pp.

schram, F. R. 1986. Crustacea. Oxford University Press, Oxford, xiv + 616 pp.

sdzuy, K. 1955. Die fauna der Leimitz-Schiefer (Tremadoc). Abhandlungen von der Senckenbergischen
Naturforschenden Gesellschaft , 492, 1-74.

shergold, j. h. 1975. Late Cambrian and early Ordovician trilobites from the Burke River Structural Belt,
western Queensland, Australia. Bulletin of the Bureau of Mineral Resources, Geology and Geophysics,
Australia, 153, 1-251, pis. 1-58 (2 vols).

— 1980. Late Cambrian trilobites from the Chatsworth Limestone, western Queensland. Bulletin of the
Bureau of Mineral Resources, Geology and Geophysics, Australia, 186, 1-111. pis. 1-35.

— and sdzuy, k. 1984. Cambrian and early Tremadocian trilobites from Sultan Dag, central Turkey.
Senckenbergiana lethaea, 65, 53-135.

snajdr, M. 1958. Trilobiti ceskeho stredniho kambria. Rozpravy Ustredniho Ustavu Geologickeho , 24, 1 -279,
pis. 1 -46.

— 1980. Bohemian Silurian and Devonian Proetidae (Trilobita). Rozpravy Ustredniho Ustavu Geologickeho,
45, 1-323, 64 pis.

1981. Ontogeny of some representatives of the tnlobite genus Scharyia. Sbornik Geologickyck Ved,
Praha, 24, 1-35, pis. 1-11.

stdrmer, l. 1942. Studies in trilobite morphology. II. The larval development, the segmentation and the
sutures, and their bearing on trilobite classification. Norsk Geologisk Tiddskrift , 21, 49-164.

Stubblefield, c. j. 1936. Cephalic sutures and their bearing on current classification of trilobites. Biological
Reviews, 11,407-440.

swanson, k. m. 1989. Manawa staceyi n.sp. (Puncidae, Ostracoda): soft anatomy and ontogeny. Courier
Forschunginstitut Senckenberg, 113, 235.

swinnerton, h. h. 1915. Suggestions for a revised classification of trilobites. Geological Magazine , 52,
407-496, 538-545.

taylor, m. e. and halley, r. b. 1974. Systematics, environment, and biogeography of some late Cambrian and
early Ordovician trilobites from eastern New York State. Professional Papers of the U.S. Geological Survey,
834,’ 1-38, pis. 1-4.

thomas, a. T. and holloway, D. L. 1988. Review of the trilobite suborder Lichida. Philosophical Transactions
of the Royal Society of London, Series B, 321, 179-262.
tripp, r. p. and evitt, w. R. 1983. Silicified trilobites of the genus Dimeropyge from the Middle Ordovician of
Virginia. Special Papers in Palaeontology, 30, 229-240.

— 1986. Silicified trilobites of the family Asaphidae from the Middle Ordovician of Virginia.
Palaeontology, 29, 705-724.

westrop, s. r. 1986. Trilobites of the Upper Cambrian Sunwaptan Stage, southern Canadian Rocky
Mountains, Alberta. Palaeontographicci Canadiana , 3, 1-97, pis. 1-41.
whittard, w. f. 1960. The Ordovician trilobites of the Shelve Inlier, West Shropshire, part IV.

Palaeontographical Society Monograph, 117-162, pis. 1 6—2 1
Whittington, H. b. 1954. Status of invertebrate paleontology, 1953. VI. Arthropoda: Trilobita. Bulletin of the
Museum of Comparative Zoology, Harvard , 112, 193-200.

1959. Silicified Middle Ordovician trilobites: Remopleurididae, Trinucleidae, Raphiophoridae, Endy-
mioniidae. Bulletin of the Museum of Comparative Zoology, Harvard, 121, 369-496, pis. 1-36.

— 1963. Middle Ordovician trilobites from Lower Head, western Newfoundland. Bulletin of the Museum of
Comparative Zoology, Harvard, 129, 1-118, pis. 1-36.

— 1965. Trilobites of the Ordovician Table Head Formation, western Newfoundland. Bulletin of the
Museum of Comparative Zoology, Harvard , 132, 275—442, pis. 1-68.

1988a. Hypostomes and ventral cephalic sutures in Cambrian trilobites. Palaeontology, 31, 577-610.

— 19886. Hypostomes of post-Cambrian trilobites. Memoirs of the New Mexico Bureau of Mines and
Mineral Resources, 44, 321-339.

576

PALAEONTOLOGY, VOLUME 33

— and almond, j. e. 1987. Appendages and habits of the Upper Ordovician trilobite Triarthrus eatoni.
Philosophical Transactions of the Royal Society, Series B, 317, 1-46.

— and Campbell, K. s. w. 1967. Silicified Silurian trilobites from Maine. Bulletin of the Museum of
Comparative Zoology , Harvard , 135, 447-483, pis. 1-19.

whitworth, p. h. 1970. Ontogeny of the Upper Cambrian trilobite Leptoplastus crassicornis (Westergaard)
from Sweden. Palaeontology, 13, 100-111, pis. 22-24.

WILEY, E. o. 1981. Phylogenetics, the theory and practice of phylogenetic systematics. Wiley Interscience, New
York, 419 pp.

woodward, h. 1883. A monograph of the British Carboniferous trilobites. Part 1. Palaeontographical Society
Monograph , 1-38, pis. 1-6.

zhang wentang 1980. On the Miomera and Polymera (Trilobita). Scientia Sinica, 23, 223-234.

— ZHANG SENGUI, LIN HUANLING, ZHU ZHAOLING, QIAN YIYUAN, ZHOU ZHIYI, YUAN JINLIANG and LU YANHAO

1980. Cambrian trilobite faunas of southwestern China. Palaeontologica Sinica (n.s. B), 16, 1-497, pis.
1-134.

— and jell, p. A. 1987. Cambrian trilobites of North China - Chinese Cambrian trilobites housed in the
Smithsonian Institution. Science Press, Beijing, 332 pp, 125 pis.

zhang xiguang 1989. Ontogeny of an early Cambrian eodiscoid trilobite from Henan, China. Lethaia, 22,
13-29.

R. A. FORTEY

Department of Palaeontology
Natural History Museum

Typescript received 6 April 1989 Cromwell Road

Revised typescript received 21 August 1989 London SW7 5BD

BEDDING PLANE ASSEMBLAGES OF PROMISSUM
PULCHRUM , A NEW GIANT ASHGILL CONODONT
FROM THE TABLE MOUNTAIN GROUP,
SOUTH AFRICA

by J. N. THERON, R. B. RICKARDS and R. J. ALDRIDGE

Abstract. Promissum pulchrum Kovacs-Endrody, 1986, originally described as the earliest vascular land plant
or an ancestor of the same, is shown to be a conodont. The species is represented by several spectacular bedding
plane assemblages which occur in the Soom Shale Member of the Cedarberg Formation, Table Mountain
Group, near Clanwilliam, South Africa. Invertebrate fossils from the Soom Shale indicate a late Ashgill (late
Rawtheyan to Hirnantian) age. The conodont assemblages are the largest known, reaching 17 mm in length,
with individual ramiform elements as long as 14 mm. The apparatuses are octomembrate with 2Pa, 2Pb, 2Pc,
2M, ISa, 4Sb, 2Sc and 2Sd elements. The possession of three pairs of P elements and a fully developed Sa-Sd
symmetry transition series distinguishes Promissum from any other known conodont genus. The arrangement
of the elements is unlike that shown by Carboniferous polygnathacean assemblages and implies a significantly
different apparatus architecture. Promissum probably represents a cold-water Gondwanan lineage that
survived the late Ordovician extinction event, and may well have given rise to the radiation of the
Pterospathodontidae in the warmer Llandovery seas.

The name Promissum pulchrum is unlike those commonly used for conodonts, and reflects the fact
that Kovacs-Endrody (in Theron and Kovacs-Endrody 1986) originally described these fossils as
the earliest known vascular land plants or their ancestors. The binomen means 'beautiful promise',
an allusion to the perceived potential of the species in interpreting early land plant evolution. The
interpretation of Promissum as a plant was not readily accepted by Rayner (1986), a referee of the
paper, whose objections were published immediately following the Theron and Kovacs-Endrody
(1986) preliminary report. These largely revolved around the lack of information normally required
to substantiate a claim for vascular land plant status, principally the presence of tracheids, a cuticle
with stomata, and sporangia containing trilete spores with sporopollenin-impregnated walls.
Kovacs-Endrody (1986, p. 103) pointed out that her conclusions were based on comparative
morphology and that no attempt had been made to obtain tracheids because the specimens were so
few in number (twelve, including parts, counterparts and fragments). Scanning electron microscopy
had revealed no preserved cell structure, but she took as important criteria the facts that the fossils
were distinct, clearly outlined and typical of plants with hard cortical tissues, and that they had well
developed spines.

It should be said that identification of these fossils is far from easy, and the attribution of them
to the plant kingdom on grounds of general comparative morphology is not so surprising. Only
after painstaking study did the present authors manage to establish them unequivocally as conodont
apparatuses. The possibility was raised at one stage that they might be graptolites, again not an
unreasonable suggestion as some do superficially resemble monograptids (see, for example, Theron
and Kovacs-Endroy 1986, fig. 5). Rayner (1986) in his review of Kovacs-Endrody's conclusion
referred to one of us (R.B.R.) as having decided that the fossils were not graptolites. The same
comment is made by Kovacs-Endrody (1987) in her fuller description of Promissum pulchrum. These
statements are not absolutely correct. R.B.R. examined only one specimen, figured by Kovacs-
Endrody (1987) as pi. 3.5 (PI. 3, fig. 2 herein): his service identification file records it as ‘not a

Palaeontology, Vol. 33, Part 3, 1990, pp. 577-594, 4 pis.)

© The Palaeontological Association

578

PALAEONTOLOGY, VOLUME 33

graptolite ... looks like a primitive jaw apparatus’. On publication of the Theron and Kovacs-
Endrody (1986) paper, R.B.R. borrowed, through the kindness of the Geological Survey of South
Africa and J.N.T., the remainder of the collection, quickly confirmed that none were graptolites,
and rather more slowly concluded that they were conodonts. At this stage the help of R.J.A. was
enlisted.

STRATIGRAPHY AND BIOSTRATIGRAPHY

In the rugged mountain ranges of the Southwestern Cape Province of South Africa a laterally persistent
argillaceous unit, the Cedarberg Formation, forms a conspicuous and critical marker horizon (c. 1 20 m) among
predominantly arenaceous rocks (3000 m) of the Table Mountain Group (text-fig. 1). The Table Mountain
Group is the basal unit of the Cape Supergroup and rests unconformably on Precambrian metasediments and
the Cape Granite Suite, which has been dated radiometrically at 500 to about 610 Ma (Burger and Coertze
1973; Schoch. Leygoine and Burger 1975). It is overlain conformably by the fossilit'erous Emsian Bokkeveld
Group.

The Cedarberg Formation thins northwards with concurrent thickening of the underlying glacial Pakhuis
Formation, and is broadly divisible into two members: a lower Soom Shale Member and an upper Disa
Siltstone Member (text-fig. 2). The dark grey, predominantly thinly laminated, micaceous Soom Shale Member
seldom exceeds 15 m in thickness. It grades into or is intercalated with the underlying tillite of the Pakhuis
Formation. Towards the north these basal units of the Cedarberg Formation in places display
varvites/rhythnutes and dropstones.

The Disa Siltstone Member is a thin-bedded argillaceous, fine-grained, sandy siltstone. Rare thin sandstone
lenses become more prominent upwards, resulting in a gradational contact with the overlying Goudini
Formation of the Nardouw Subgroup. Approximately in the middle of the Disa Siltstone Member in the
Porterville- Wellington area occurs the ‘brachiopod zone’ in which Rust discovered a shelly fauna (Cocks et
al. 1970; Cocks and Fortey 1986).

Body fossils from the Cedarberg Formation are as yet restricted to the western outcrops, and those on the
farm Keurbos, about 115 km south-east of Clanwilliam, represent the northernmost fossiliferous occurrence
known (text-fig. 3). The specimens of Promissum described in this paper all come from the base of the Soom
Shale Member on Keurbos and were collected by J. N. Theron, J. H. Bredell and D. J. Barnardo of the
Geological Survey in 1981. The associated fossils include rare scolecodonts, unidentified trace fossils, and a
dark-coloured tube-like fossil which Kovacs-Endrody (1987) referred to the plant Eohostimella Schopf.

Quarrying for road metal has revealed that the Soom Shale locally attains a thickness of 6-5 m and contains
a variety of invertebrate fossils. A 50 cm thick layer of partly bedded reworked glacial sediment interposes here
between the underlying Pakhuis tillite and the shale unit. Systematic investigation of the quarry face has
revealed two distinct levels at which Promissum and ‘ Eohostimella ' preferentially occur, T2 and 2-5 m above
the Pakhuis/Soom contact. Promissum fragments are also present in the uppermost 5 cm of the reworked
glacial bed, only 45 cm above the Pakhuis tillite. The Soom Shale Member at Keurbos is often conspicuously
finely laminated with the individual laminae extensively characterized by linear and/or sinuous surface trails.
Promissum and ‘ Eohostimella ' are infrequently present on the same laminar surface as brachiopods.

At least two major advances of the Gondwana ice sheet are registered in the glaciogenic deposits of the
Pakhuis Formation (Rust 1967, 1981). The concurrent glacioeustatic fall in sea level resulted in partial
reworking of glaciogenic deposits by coastal processes in a shallow marine embayment. The Cedarberg
Formation was deposited during the succeeding glacioeustatic rise in sea level and represents the outwash silt
and mud from the retreating ice sheets. Downwasting of the ice would have uncovered local embayments (ice-
dammed?) along the northern shoreline which were probably flooded with fresh glacial melt-waters. The fine
mud was clearly deposited towards the north in glaciolacustrine to shallow marine environments as shown by
the presence of chitinozoans, marine phytoplankton, brachiopods, algae and spore tetrads in the same samples
or in close proximity (Cramer et al. 1974; Gray et al. 1986). Varvites do not form in seawater (Flint 1971;
Wright and Moseley 1975), but are commonly deposited seasonally in glacial lakes as graded layers. The finely
laminated sediments at Keurbos with their biota of Promissum , 'Eohostimella' and brachiopods manifestly
accumulated in a protected still-water marine environment rather than in a proglacial lacustrine setting as
previously envisaged (Theron and Kovacs-Endrody 1986).

Theron and Kovacs-Endrody (1986) regarded the basal Soom Shale Member at Keurbos as probably early
Silurian (Llandovery) on the basis of spore tetrads, phytoplankton and chitinozoans (Gray et al. 1986).
However, Cocks and Fortey (1986) demonstrated that the Soom Shale Member at Buffels Dome in the Hex

THE RON ET A L. : SOUTH AFRICAN ASHGILL CONODONT

579

text-fig. 1 . Locality map showing the distribution of the Table Mountain Group, the microfossil site reported
by Gray et al. (1986) and the sample site for the specimens of Promissum pulchrum.

River Mountains is late Ashgill (late Rawtheyan to Hirnantian) in age. The crucial element in the fauna is the
trilobite Mucronaspis olini , which on all known records is late Ashgill; none of the other elements of the fauna
contradict this. Buffels Dome is more than 100 km from Keurbos and diachronism cannot yet be ruled out,
but the same authors reviewed the evidence for the age of the overlying Disa Siltstone Member and concluded
from the presence of brachiopods, including Trematis taljaardi Rowell, Marklandella africana Cocks and

580

PALAEONTOLOGY, VOLUME 33

metres

Grey to light grey fine-grained quartzitic sandstone
with thin siltstone lenses.

GOUDINI

FORMATION

Grey fine-grained laminated micaceous sandstone and siltstone
characterised by upward coarsening and increase in bedding thickness.
Intermittently characterised by ripple mark cross-bedding
and load casting.

Bioturbated fine-grained siltstone and mudstone, massively bedded I Brachiopod
with large brachiopod fauna. J Zone

Disa

> Siltstone
Member

Grey to black, thin-bedded, fine-grained siltstone
with few thin shale lenses.

> CEDARBERG
FORMATION

Black fine-grained thinly laminated micaceous shale with Invertebrate fossils,
microfossils, spore tetrads, Promlssum and Eohostimella .

Soom
■ Shale
Member

Dlamictites, varvites and poorly stratified pelitic units;
polished, striated and faceted clasts present.

Fold zone and dlamictite.

PAKHUIS

FORMATION

Light grey, medium to coarse-grained massively bedded quartzitic sandstone
with thin lenses of quartz pebbles; trough cross-bedding well developed.

PENINSULA

FORMATION

text fig. 2. Schematic stratigraphic profile.

Brunton, Eostropheodonta discumbata Cocks and Brunton and Plectothyrella haughtoni Cocks and Brunton,
that it is Hirnantian. Thus we regard Promissum pulchrum as more likely to be oflate Rawtheyan to Hirnantian
age, rather than Silurian.

PALAEONTOLOGY

Introduction

Our examination of the specimens of Promissum pulchrum has shown that they are bedding-plane
assemblages and isolated elements of conodonts. They are remarkable in several ways, not least in
their spectacular size. The ramiform elements in the largest assemblage are more than 14 mm long,
with the whole assemblage having a length of 17 mm. The elements are robust with peg-like
denticles, between some of which finer denticles are apparent. Preservation of the original apatite
is poor, with several of the elements represented only by internal and external moulds. In others the
phosphate has been leached or replaced by clay minerals to give a pale greenish appearance. On
analysis, this material gave peaks in aluminium, silicon and potassium. Peaks in calcium and
phosphorus were obtained from some relatively unaltered amber-coloured areas of one of the
ramiform elements (text-fig. 4).

The anterior and posterior ends of the assemblages can be determined from the anterior cusps and
posterior processes of the ramiform elements. Beyond this, orientation is problematical, as almost
all the elements have their ‘upper’ surfaces directed inwards and it is impossible to tell whether the

THE RON ET AL. : SOUTH AFRICAN ASHGILL CONODONT

581

text-fig. 3. Geological map of the Clanwilliam area, showing the sample site.

582

PALAEONTOLOGY, VOLUME 33

assemblages are being viewed from above or below. The only guide is provided by the Pa and Pb
elements in the central area of the holotype. On the counterpart (C.2b-I, text-fig. 5), these are
preserved as deep external moulds, showing that their oral (conventionally ‘upper’) surfaces were
directed into the bedding plane; they are thus being viewed from below, which indicates that they
are situated on the sinistral side of the assemblage. This designation has been adopted in the
description and discussion of specimens in this paper, but it should be emphasized that this is based

dextral

text-fig. 5. Camera lucida drawing of the counterpart of the holotype of Promissum pulchrum , specimen C.2b-
I, for comparison with PI. 1, fig. 2. The anterior of the apparatus is to the right.

THERON ET AL.\ SOUTH AFRICAN ASHGILL CONODONT

583

on conventional orientation of the Pa and Pb elements alone and does not necessarily relate to
actual left and right sides of the Promissum animal.

Systematic palaeontology

Genus promissum Kovacs-Endrody, 1986

Type species. Promissum pulchrum Kovacs-Endrody, 1986, from the basal Soom Shale Member, Cedarberg
Formation, of Keurbos, near Clanwilliam, South Africa; by monotypy.

Diagnosis (emend.). Apparatus octomembrate, with 2Pa, 2Pb, 2Pc, 2M, ISa, 4Sb, 2Sc and 2Sd
elements. Pa and Pb elements pastinate with deep basal cavities, M element arched, S elements with
long processes.

Remarks. The apparatus is unlike that of any other known genus in possessing three pairs of P
elements in conjunction with a fully developed Sa-Sd symmetry transition series.

Promissum pulchrum Kovacs-Endrody, 1986
Plates 1-3; text-figs. 5-8

1986 Promissum pulchrum Kovacs-Endrody in Theron and Kovacs-Endrody, p. 102, figs. 4 and 5.

1987 Promissum pulchrum Kovacs-Endrody; Kovacs-Endrody, p. 99, pis. 3.1-3.22.

Holotype. Specimen C.2a, Geological Survey of South Africa, Pretoria.

Material. We have examined most of the specimens figured by Kovacs-Endrody (1987): the holotype C.2a (PI.
2, text-fig. 6), its counterpart, C.2b-I (PI. 1, fig. 2, text-figs. 5, 7, 8), assemblages C.l (PI. 3, figs. 1 and 4) and

sinistral

text-fig. 6. Camera lucida drawing of the part of the holotype of Promissum pulchrum , specimen C.2a. The

anterior is to the left.

584

PALAEONTOLOGY, VOLUME 33

text-fig. 7. Scanning electron micrograph of latex cast of portion of counterpart of holotype, specimen C.2b-
I, showing Pa element (above, centre), Pb element (below, centre) and portions of the processes of ramiform

elements; posterior towards top, x35.

EXPLANATION OF PLATE 1

Figs. 1 and 2. Promissum pulchrum Kovacs-Endrody. 1, latex cast of specimen C.2b-II. Anterior to the right,
x 13.5. 2, specimen C.2b-I, counterpart of holotype. Anterior to the right, x 7.5. For annotation of the
elements see text-fig. 5.

PLATE 1

THERON, RICKARDS and ALDRIDGE, Promissum

586

PALAEONTOLOGY, VOLUME 33

C.2b-II (PI. 1, fig. 1), and isolated element C.6 (PI. 3, fig. 2). The poorly preserved specimen numbered C.81
(Kovacs-Endrody 1987, pi. 3.6) and the unfigured specimen C.3 have not been included in our study.

Diagnosis. As for genus.

Description. Pa element: pastinate, high, with deep basal cavity; probably thin-walled. Cusp apical, stout;
anterior, posterior and primary lateral processes with discrete nodose median denticles. Primary lateral process
long, secondary lateral process short and apparently adenticulate.

Pb element: pastinate, high, with deep basal cavity; probably thin- walled. Anterior, posterior and primary
lateral processes with discrete nodose median denticles. Primary lateral process long; distribution of secondary
lateral processes unclear, but process disposition appears to differ from that shown by Pa element.

Pc element: indistinct, but with at least two denticulate processes diverging widely from a prominent
triangular cusp.

M element: symmetrically arched, with a prominent cusp and discrete, erect, robust denticles on each
process. Basal cavity at apex of arch, with lip on outer side.

Sa element : alate, probably with long posterior process. Lateral processes diverge at 40-50° to form an
anterior arch ; very long, but broken about I mm from apex on all specimens, distal portions extend posteriorly
for several millimetres. Proximal denticles stout, widely-spaced with u-shaped separations and no evidence of
mtercalatory smaller denticles; denticulation beyond 1 mm indistinct. Basal cavity apparently conical with
small lips.

Sb element: tertiopedate, with very long posterior process bearing regular, stout, erect, large denticles
between which smaller, slender denticles are irregularly recognizable. Antero-lateral process also long, curved
proximally to become sub-parallel to posterior process, with widely-spaced, stout, conical denticles of subequal
size. Outer lateral process not clearly preserved on any specimens, may be short and adenticulate or perhaps
broken. Basal cavity small and conical beneath small erect cusp.

Sc element : bipennate, with very long posterior process displaying similar denticulation to that of the Sb
element (text-fig. 8). Antero-lateral process strongly curved proximally to become nearly parallel to posterior
process; very long and bearing widely-spaced conical denticles that are directed inwards and somewhat
downwards. Basal cavity conical beneath a relatively small, erect cusp.

Sd element: quadriramate, with all four processes probably very long. Posterior process with similar
denticulation to that of the Sb and Sc elements; other processes may have widely-spaced conical denticles, but
they are difficult to distinguish. Cusp small and erect.

text-fig. 8. Camera lucida drawing of dextral Sc element of specimen C.2b-I, showing details of denticulation;

scale bar 1 mm.

Remarks. Since Kovacs-Endrody (1987) considered Promissum to be a plant, she provided
interpretations of various features of the specimens on that basis. These features are highlighted on
the twenty-two illustrations of Promissum in her paper. Our recognition of the specimens as
conodonts necessitates a re-interpretation, and we would caption her figures as follows:

EXPLANATION OF PLATE 2

Promissum pulchrum Kovacs-Endrody, specimen C.2a, holotype, photographed under alcohol. Anterior at
the top, x 9.

PLATE 2

THERON, RICKARDS and ALDRIDGE, Promissum

588

PALAEONTOLOGY, VOLUME 33

pi.

3.1

pi.

3.2

pi.

3.3

pi.

3.4

pi.

3.5

pi.

3.6

pi.

3.7

pi.

3.8

pi.

3.9

pi.

3.10

pi-

3.11

pi-

3.12

n'.

3.13

Pi

3.14

pl-

3.15

P'-

3.16

pi.

3.17

pi.

3.18

pi.

3.19

pi-

3.20

pi-

3.21

n'-

3.22

holotype of P. pulchrum, C.2a; complete octomembrate bedding-plane assemblage (see PI. 2 herein),
dextral portion of counterpart of holotype, C.2b-I; elements mostly preserved as external moulds. To
the left are the P elements, the arrow points to the M element and the long processes of the S elements
are evident in the centre and right of the picture. The displaced dextral Sc element is at top left,
faintly preserved complete assemblage, C2b-II; (a) cusps of ramiform S elements, (b) P element, (c)
distal portion of S elements (see PI. 1, fig. 1 herein),
complete assemblage, C.l (see PI. 3, fig. 1 herein).

isolated quadriramate Sd element, only proximal portion preserved (see PI. 3, fig. 2 herein),
poorly preserved processes of an S element.

anterior portion of assemblage C.l, arched Sa element at the top of the picture (see PI. 4 herein),
cusps and proximal parts of two dextral S elements, C.2a.

holotype, C.2a; (1) sinistral Pb element, (2) sinistral Pa element, (3) dextral Pa element, (4) dextral
Pb element, (a) dextral Sb element, (b) axial portion of ramiform element group, (c) displaced dextral
Sc element.

dextral Pa element, C.2a.
dextral Pb element, C.2a.

portion of specimen C.2a, to show process disposition on Pa elements,
sinistral Pa element, C.2a.

Pa element, C.l.

Pc element, C.l.

central portion of complete assemblage, C.l.

proximal portion of dextral Sb element, C.2a; the process on the right side of the photograph is part
of the lateral process of the displaced dextral Sc element and shows robust denticles in lateral view,
axial portion of the ramiform element group, showing opposition of denticulate surfaces, C.2a.
proximal portion of dextral Sc element and dextral Pc element, C.2a.

part of a process on the quadriramate Sd element C.6, showing robust denticles in apical view.

enlargement of processes on S elements of C.l, showing robust denticles in apical view.

cusp of dextral Sb element and portion of lateral process of displaced dextral Sc element, C.2b-I.

Arrangement of the assemblages and the architecture of the Promissum apparatus

The arrangement of the elements is most clearly displayed by the counterpart and part of the

holotype (text-figs. 5 and 6), in which two integrated sets are evident:

(1) The elements of the first symmetry transition series ( sensu Barnes et al. 1979). The Sb, Sc and
Sd elements are orientated with their posterior processes parallel; other processes of these elements
diverge at the cusp, but become parallel to the posterior processes distally. On the sinistral side of
the assemblage these elements are piled on each other, with the cusps and denticulate surfaces of
the posterior processes directed inwards. On the dextral side the elements are more widely separated.
One of the dextral Sb elements and the probable Sd element are orientated in opposition to the
sinistral elements, whereas the second Sb element has become overturned to face outwards. The
dextral Sc element is also overturned and the proximal portion is displaced so that the long
processes are directed inwards and somewhat anteriorly. The Sa element is positioned to the
anterior of the other elements, displaced dextrally from the axis with the cusp pointing dextrally;
the lateral processes are broken, with the distal portions parallel or subparallel to the processes of
the other S elements.

(ii) the elements of the second transition series ( sensu Barnes et ah 1979). The P elements are
arranged in two parallel rows, aligned along an axis at 30° to that of the S elements. Those of the
sinistral side are superposed on the processes of the S elements, but those on the dextral side are

EXPLANATION OF PLATE 3

Figs. 1 and 2. Promissum pulchrum Kovacs-Endrody. 1, specimen Cl. Anterior at the top of the picture, x9.
Specimens of ‘ Eohostimella ’ are apparent around and across the apparatus. 2, specimen C.6. Single,
isolated quadriramate Sd element. Anterior at the top, x22.

PLATE 3

THERON, RICKARDS and ALDRIDGE, Promissum

590

PALAEONTOLOGY, VOLUME 33

positioned largely outside the members of the first transition series. On the dextral side the Pa and
Pb elements are in lateral aspect, whereas on the sinistral side of the part they are in upper view.
The M element on each side is immediately to the left of its companion Pb element, with the cusp
directed dextrally. As with the M elements, the Pc elements are in similar positions but apparently
in opposing orientation on the two sides of the assemblage. The dextral Pc element on the
counterpart has its denticulate posterior process directed away from the bedding surface while the
equivalent process on the sinistral member is preserved as an external mould. The revealed face of
the sinistral M element has a clear lip to the basal cavity that is not apparent on that of the dextral
element.

The bilateral symmetry retained by the two sets of elements shows that the assemblage has been
compacted in close to a dorso-ventral configuration. The displacement of the outer Sb and Sc
elements on the dextral side, together with the apparent dextral shift of the axial Sa element, indicate
that the apparatus was tilted a little to that side before it collapsed onto the bedding plane. This
tilting has also displaced the P and M elements towards the right and may account for the different
orientations of the Pa and Pb elements on the two sides.

The other two assemblages we have examined add little information about the relative
dispositions of the elements in the apparatus. The second, less well preserved, assemblage (C.2b-II,
PI. 1, fig. 1) on the same slab as the counterpart of the holotype displays a similar arrangement of
the two sets of elements, except that the upper surfaces of the Pa and Pb elements face inwards. The
elements of the third assemblage (C.l, PI. 3, fig. 1) are superposed on each other, giving a much
more crowded arrangement in which the P and M elements cannot all be clearly distinguished; the
anterior portion of the Sa element is separated from the others at the front of the apparatus in
exactly the same manner as in the holotype (PI. 4). The fact that all three specimens are preserved
in close to dorso-ventral compaction may suggest that the Promissum animal was dorso-ventrally
flattened in life, in contrast to the lateral flattening deduced for Carboniferous polygnathaceans
(Aldridge et al. 1987).

In the absence of an assemblage compacted in lateral aspect, it is impossible to produce a
reconstruction of the three-dimensional architecture of the Promissum apparatus. However, the
evidence from these three specimens does place constraints on that architecture and reveals several
important differences from the structure of polygnathacean apparatuses, as deduced from the study
of several Carboniferous bedding-plane assemblages preserved in different orientations by Aldridge
et al. (1987). In the polygnathaceans, the elongate, ramiform S elements were aligned at a steep
angle to the long axis of the conodont animal, with their cusps to the anterior and an M element
flanking them on each side; to the posterior of this group were one pair each of Pb and Pa elements,
lying nearly vertically and nearly normal to the trunk. In Promissum , the M elements are clearly
associated geometrically with the P elements, rather than with the S elements. In addition, unless
the head of the Promissum animal has collapsed anteriorly on to the bedding-plane in all three
known specimens, it appears that the P and M elements may not have been positioned posterior to
the S elements in this apparatus. The superposition of the P and S elements suggests that the former
were situated above or below the denticulate processes of the latter, although with the currently
limited material it cannot be discounted that the two sets of elements may have been brought into
juxtaposition by muscular contraction. Whether the cusps of the P elements were directed inwards
or towards the S elements cannot be determined, but their position at least suggests that an analogy
with the palatal tooth of the myxinoids is worthy of investigation.

EXPLANATION OF PLATE 4

Promissum pulchrum Kovacs-Endrddy, specimen Cl. Anterior area of apparatus, showing proximal portions
of ramiform S elements. Anterior at the top, x 30.

PLATE 4

THERON, RICKARDS and ALDRIDGE, Promissum

592

PALAEONTOLOGY, VOLUME 33

RELATIONSHIPS OF PROMISSUM AND EVOLUTIONARY IMPLICATIONS

One of the striking features of Promissum is its extremely large size. Gigantism is not uncommon
in cold-water taxa, although the controls in this case are uncertain. In other features, Promissum
differs less spectacularly from other coeval conodont taxa. The apparatus structure broadly
conforms to a prioniodontacean plan, although differing somewhat from all other genera whose
apparatuses are known. The Sa-Sd elements compare in general pattern with those of the
Ordovician genera Prioniodus Pander and Amorphognathus Branson and Mehl (Bergstrom 1971,
1983; see text-fig. 9), but the very long processes are distinctive. The M element is totally unlike that
of Prioniodus , which is falodiform, or that of Amorphognathus , which is holodontiform. The Pa and
Pb elements do resemble specimens that have been assigned to Prioniodus , but are most similar to
the Pa element of Sagittodontina Kniipfer. The apparatus of Sagittodontina has been reconstructed
by Bergstrom (1983), who included ramiform elements with very short processes, dissimilar from
those displayed by the South African specimens. The material illustrated by Kniipfer (1967) from
the same collections as the original specimens of Sagittodontina does include some broken specimens
that may be comparable with some of the ramiform elements in Promissum , but nothing resembling
the arched M element is reported. There have been no suggestions that Sagittodontina , Prioniodus
or any other Ordovician genus possessed three pairs of P elements.

Some Silurian genera may well have had Pa, Pb and Pc elements homologous with those of
Promissum. A quinquemembrate apparatus with three pairs of P elements has been reconstructed
for Pterospathodus Walliser by Mannik and Aldridge (1989), who also recognized Pa, Pb and Pc
elements in their new genus, Pranognathus. Three P elements, termed Par Pa., and Pb were described
in Apsidognathus Walliser by Uyeno and Barnes (1983), and comparable elements occur in

text-fig. 9. Element types in the apparatus of Amorphognathus tvaerensis Bergstrom for comparison with
Promissum pulchrum; redrawn after Bergstrom (1983). a. Pa element, upper view; b, Pb element, lateral view;
c, M element, lateral view; d. Sc element, lateral view; e, Sb element, lateral view; f, Sa element, lateral view;

G, Sd element, lateral view.

THERON ET AL.: SOUTH AFRICAN ASHG1LL CONODONT

593

Astropentagnathus Mostler. None of these genera have first transition series comparable with that
of Promisswn and quadriramate Sd elements, in particular, are unknown in Silurian taxa. Only in
Astropentagnathus does the M element have a symmetrically arched morphology.

Pterospathodus . Astropentagnathus and Apsidognathus were all included by Klapper (1981 ) in the
family Pterospathodontidae, which appears cryptically in the mid- Llandovery (Mannik and
Aldridge 1989). The earliest representatives are currently assigned to Pranognathus , which possesses
thin-walled Pa, Pb and Pc elements, a dolabrate M element and a first transition series of Sa, Sb,
Sc and Sc., elements; related later Llandovery genera have more robust P elements, and in
Pterospathodus the first transition series comprises only an Sa/Sb element. It is possible that
Promissum represents a stock from which this group of genera descended by reduction of the first
transition series. The radiation of the Pterospathodontidae may thereby reflect a spreading into
warm Llandovery seas of a cold-water Gondwanan lineage that survived the late Ordovician
extinction event.

Acknowledgements. This reappraisal of very exciting fossils would not have been possible without the very open
and friendly cooperation of the original authors and the Geological Survey of South Africa. We especially wish
to thank Dr C. MacRae for all his help. Dr A. Buckley carried out the probe analysis at Cambridge and Dr
S. Conway Morris gave helpful advice on the fossils in the Soom Shale Member. Part of this work was funded
by NERC Research Grant GR3/5105 to RJA.

REFERENCES

aldridge, r. j., smith, m. p., norby, r. d. and briggs, d. e. g. 1987. The architecture and function of
Carboniferous polyngnathacean conodont apparatuses. 63-75. In aldridge, r. j. (ed. ). Palaeobiology of
conodonts. Ellis Horwood Limited, Chichester, 180 pp.
barnes, c. R., Kennedy, d. j., McCracken, a. d., nowlan, G. s. and tarrant, G. a. 1979. The structure and
evolution of Ordovician conodont apparatuses. Lethaia , 12, 125-151.
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.). Conodont biostratigraphy. Geological Society
of America Memoir 127, 83-161.

- 1983. Biogeography, evolutionary relationships, and biostratigraphical significance of Ordovician
platform conodonts. Fossils and Strata , 15, 35-58.

burger, a. j. and coertze, F. j. (compilers). 1973. Radiometric age measurements on rocks from southern
Africa to the end of 1971. South African Geological Survey . Bulletin 58, 46 pp.
cocks, l. r. m. and fortey, r. a. 1986. New evidence on the South African Lower Palaeozoic: age and fossils
reviewed. Geological Magazine , 123, 437-444.

— , brunton, c. H. c., rowell, a. j. and rust, i. c. 1970. The first Lower Palaeozoic fauna proved from
South Africa. Quarterly Journal of the Geological Society of London , 125, 583-603.
cramer, f. h., rust, i. c. and diez de cramer, m. d. c. r. 1974. Upper Ordovician chitinozoans from the
Cedarberg Formation of South Africa. Preliminary note. Geologische Rundschau , 63, 34r 15.

flint, R. F. 1971. Glacial and Quaternary geology. John Wiley and Sons, New York, 892 p,
gray, j., theron, J. n. and boucot, a. j. 1986. Age of the Cedarberg Formation, South Africa and early land
plant evolution. Geological Magazine , 123, 445-454.

klapper, g. 1981. Family Pterospathodontidae Cooper, 1977. 135-136. In robison, r. a. (ed.). Treatise on
invertebrate paleontology. Part W. Supplement 2 Conodonta. Geological Society of America, Inc. and the
University of Kansas.

knupfer, j. 1967. Zur Fauna und Biostratigraphie des Ordoviziums (Grafenthaler Schichten) in Thiiringen.

Freiberger Forschungshefte, C220 Palaontologie, 1-119.
kovacs-endrody, e. 1987. The earliest known vascular plant, or a possible ancestor of vascular plants in the
flora of the Lower Silurian Cedarberg Formation, Table Mountain Group, South Africa. Annals of the
Geological Survey of South Africa , 20, 93-118.

mannik, p. and aldridge, r. j. 1989. Evolution, taxonomy and relationships of the Silurian conodont genus
Pterospathodus. Palaeontology. 32, 893-906.

rayner, R. j. 1986. Promissum pulchrum : the unfulfilled promise? South African Journal of Science. 82,
106-107.

594

PALAEONTOLOGY, VOLUME 33

rust, i. c. 1967. On the sedimentation of the Table Mountain Group in the western Cape Province.
Unpublished D.Sc. thesis. University of Stellenbosch.

- 1981. Lower Palaeozoic rocks of southern Africa. 165-187. In Holland, c. h. (ed.). Lower Palaeozoic
rocks of the world , volume 3. Wiley Interscience, New York, 331 pp.
schoch, a. e., laygonie, F. E. and burger, a. j. 1975. U-Pb ages for Cape granites from the Saldanha
batholith: a preliminary report. Transactions of the Geological Society of South Africa, 78, 97-100.
theron, j. N. and kovacs-endrody, E. 1986. Preliminary note and description of the earliest known vascular
plant, or an ancestor of vascular plants, in the flora of the Lower Silurian Cedarberg Formation, Table
Mountain Group, South Africa. South African Journal of Science , 82, 102-105.
uyeno, t. t. and barnes, c. r. 1983. Conodonts of the Jupiter and Chicotte Formations (Lower Silurian),
Anticosti Island, Quebec. Geological Survey of Canada , Bulletin 355, viii + 49 pp.
wright, a. E. and moseley, f. (eds). 1975. Ice ages : ancient and modern. Geological Journal Special Issue 6. Seel
House Press, Liverpool, 320 pp.

j. N. theron

Geological Survey
P.O. Box 572
Bellville 7535
South Africa

R. B. RICKARDS

Sedgwick Museum
Department of Earth Sciences
University of Cambridge
Cambridge CB2 3EQ

Typescript received 5 July 1989

Revised typescript received 7 December 1989

R. }. ALDRIDGE
Department of Geology
University of Leicester
Leicester LEI 7RH

DRILLING AND PEELING OF TURRITELLINE
GASTROPODS SINCE THE LATE CRETACEOUS

by WARREN D. ALLMON, JAMES C. NIEH and RICHARD D. NORRIS

Abstract. Frequencies of predation on turritelline gastropods by drilling and peeling predators have not
changed significantly during the course of the Cenozoic. Rates of drilling in the Cretaceous are lower than
Cenozoic rates, but not significantly so. Conversely, rates of peeling and repair in the Late Cretaceous reach
or exceed Cenozoic values.

Turritelline shell form is not correlated with predation intensity. Highly sculptured species are not more
immune to drilling and peeling predation than are less sculptured taxa. Shell geometry in these gastropods does
not show progressive trends during the Cenozoic. Sculpture strength and most aspects of shell form and
sculpture strength are evidently not adaptations to resisting peeling and drilling predation in turritellines.
Turritellines have not evolved during the Cenozoic in an arms race to build more predation-resistant shells,
although behavioural or other non-shell characteristics may have changed over time. Thus, in this group, any
‘marine revolution’ and adaptive response of prey to the evolution of durophagous predators must have
occurred prior to the Late Cretaceous.

the notion that predation influences the evolution of prey has been widely discussed (e.g. Vermeij
1977, 1978, 1982 b, 1983, 1987; Hughes 1980; Bayne 1981 ; Kitchell et al. 1981; Bakker 1983; Taylor
1984) but has seldom been tested within a single prey clade. Turritelline gastropods (i.e. members
of the family Turritellidae, subfamilies Turritellinae and Protominae, sensu Marwick [1957]) are
common to abundant fossils in many Cretaceous and Cenozoic horizons, and many bear the marks
of attack by drilling and shell-peeling (i.e. aperture-breaking) predators. Previous studies (e.g.
Dudley and Vermeij 1978; Vermeij and Dudley 1982) have discussed trends in predation in this
group, at least as represented by these traces. In this paper we expand this earlier work by ( 1 )
considering a much larger data set, (2) making use of more adequate views of both turritelline
ecology (Allmon 1988c/) and of the stratigraphic and systematic relationships of species from the
southeastern United States (Toulmin 1977; Allmon 1988//), which have figured prominently in
previous work, and (3) exploring possible evolutionary consequences of predation on turritelline
shell form.

Vermeij (1987, and references therein) has long argued that durophagous predation has been at
least partly responsible for some trends in gastropod shell morphology. Whether this has been the
case for turritellines is of particular interest since almost nothing is known about the functional
significance of most features of their shells. Correlation or lack of correlation between predation
intensity and shell form over the history of this group may indicate evolutionary origins of
particular features. Although their origin is obscure, turritellines appear to have arisen in the Late
Jurassic or Early Cretaceous (Merriam 1941), a time of rapid morphological and taxonomic
diversification among gastropods (Vermeij 1977; Taylor et al. 1980, 1983). An understanding of the
factors controlling the occurrence and significance of predation in turritellines may thus also
contribute to a better understanding of the evolutionary dynamics of this interval.

MATERIALS AND METHODS

New data presented here are derived from examination of specimens in the collection of the Department of
Invertebrate Paleontology of the Museum of Comparative Zoology (MCZ), Harvard University. All are from

IPalaeontology, Vol. 33, Part 3, 1990, pp. 595-611. |

© The Palaeontological Association

596

PALAEONTOLOGY, VOLUME 33

text-fig. 1. Traces of drilling and peeling predation in fossil turritellines. All specimens from the Pliocene
Pinecrest Beds of southern Florida, a. Turritella cf. T. apicalis Heilprin, showing naticid drill hole in the middle
of the whorl, b. Turritella cf. T. apicalis Heilprin, showing the less common positioning of a drill hole astride
a suture between whorls. Peeling/repair scar indicated by arrow, c. Turritella pontoni Mansfield, showing
peehng/repair scar, indicated by arrow. Scale bar = 1 cm.

text-fig. 2. Longitudinal distribution of drill holes on
intact turritelline shells of all species in our data set
(Appendix 1). Figure in upper right shows the base of
a turritelline shell looking toward the apex parallel to
the axis of coiling, indicating division into quadrants
for recording the longitudinal position of drill holes.
Intact bases and apertures are distinguished from
broken and incomplete specimens by the presence of
parietal callus on outside of last whorl.

Cenozoic deposits of the New World and most were personally collected by one of us (WDA). A total of 1097
specimens representing 27 species were examined. For each specimen the following observations or
measurements were made: length (or estimated length if broken), maximum whorl diameter, presence/absence
of drill holes (text-fig. 1a, b), hole diameter, position of the hole vertically on the whorl and relative to the
aperture if this could be determined (text-fig. 2), diameter or drilled whorl, presence/absence and number of
repaired shell breaks (text-fig. lc), and diameter of broken-and-repaired whorl. All measurements were made
with digital calipers to the nearest 0-1 mm. Frequency of drilling in each species is the number of drilled shells
divided by the total number of shells of that species examined, expressed as a percentage. Frequency of
peeling/repair is the total number of scars divided by the number of shells examined, expressed as a decimal

ALLMON ET A L.: DRILLING OF GASTROPODS

597

value. These data (Appendix 1), together with additional observations from the literature (Appendix 2)
represent a total of 10,387 specimens of 68 species, ranging in age from Early Cretaceous to Recent.

As noted by Dudley and Vermeij (1978), the generic and subgeneric taxonomy of turritellines is unresolved,
particularly for fossil species (Marwick 1957; Allmon 1988b). Consequently we refer all species to Turritella
sensu lato. Despite this remaining uncertainty, however, it is reasonable to make use of recent and relatively
uncontroversial opinions on the position of some taxa, and we have therefore excluded from consideration
three Recent species included in the genus by Dudley and Vermeij (1978). 1 T. erosa Couthouy’ and
'T. reticulata Mighels’ belong to the genus Tachyrhynchus Morch (e.g. Abbott 1974), which Marwick (1957)
places in the subfamily Pareorinae. "T. duplicata Linnaeus’ is the type species for the genus Zaria Gray
(Marwick 1957). We have also excluded fossil and living species assigned to the genus Mesalia Gray, including
the Eocene species M. regularis Deshayes and M. amekiensis Eames, both of which were considered by
Dudley and Vermeij. Mesalia and Zaria are placed in Pareorinae by Marwick. Whatever the generic place-
ment of species within Turritellinae, these taxa are almost certainly only distantly related to the other species
considered here.

In their study of turritelline specimens in the collection of the U.S. National Museum, Dudley and Vermeij
(1978) assigned several strictly Palaeocene species (e.g. T. mortoni , T. praecincta) to the Eocene (see Toulmin
[1977] and Allmon [1988b] for further discussion of stratigraphic relations of early Tertiary species).
Reassignment of these species and examination of others provide the first estimates of drilling frequencies in
turritellines from the Palaeocene of the southeastern U.S.

Drilled and undrilled shells may exhibit differential preservation potential (Dudley and Vermeij 1978 ; Taylor
et al. 1983), but biases may operate in both directions. Drilled shells may be more easily fragmented or
transported or, because many naticid predators pull their prey into the sediment to feed, they may increase the
probability of preservation of drilled shells (Edwards 1974). Here we assume that all shells have equal
preservation potential.

Most of the holes observed in these turritelline shells resemble the ‘truncated spherical paraboloid’ typically
produced by naticacean gastropods (text-figs. 1a, b), rather than the usually more straight-sided holes
characteristic of muricaceans (e.g. Sohl 1969; Carriker 1981). The great majority of holes examined here fall
into types C, E and F of Arua and Hoque (1989), which these authors suggest as belonging to naticids.
Specification of an exact percentage is difficult since many holes are eroded and cannot be assigned to one type
over the other. Probably no more than 5-10% of the holes we examined are attributable to muricids, and we
assume that the majority of drilling predators were naticids. We have also assumed that most of the observed
breakage/repair scars were produced by shell-peelmg predators, such as calappid crabs (cf. Vermeij 1982«,
1983), rather than accidentical breakage not associated with predation (cf. Vermeij 1987, p. 227).

RESULTS

Temporal patterns

Early and Late Cretaceous drilling frequencies are not significantly different (/-test, P = 0-32),
although both are well below almost all Cenozoic values (text-fig. 3a). Drilling frequency for all
Cretaceous species taken together is significantly lower than that for Palaeocene, Miocene or Plio-
Pleistocene species (Mann-Whitney C-tests, 0-025 > P < 0 05), but not significantly different from

Recent
Pliocene
Miocene
Oligocene
Eocene
Palaeocene
Late Cretaceous
Early Cretaceous

B

23.5 20.7 (17)

27.6 19.3 (9)
24.1 17.6 (11)

1.5 2.1 (2)

22.8 18.5 (12)
20.3 18.2 (8)

4.5 7.0 (5)

1.9 3.7 (2)

Pliocene

Miocene

Oligocene

Eocene

Palaeocene

Late Cret.

0.39 0.25 (6)
0.1 1 0.14 (5)
0.05 0.03 (2)
0.28 0.34 (5)
0.19 0.22 (9)
0.52 0.15 (3)

text-fig. 3. Distribution of rates of a, drilling, and b, peeling predation in turritelline gastropods in the
Cretaceous and Cenozoic. First of three numbers on right is mean frequency (in % for drilling) of all samples
for each time interval; second number is one standard deviation; third number is number of species sampled.
Figures derived from all data presented in Appendices 1 and 2.

598 PALAEONTOLOGY, VOLUME 33

that for Eocene species (P = 0-25). Drilling frequency for the Oligocene is lower than that for either
the Eocene or the Miocene (probably a result of small sample size), but the differences are not
significant. In fact, no epoch of the Cenozoic shows a drilling frequency significantly different from
any other (Mann-Whitney (7-tests, 010 < P > 0-05).

Late Cretaceous values for peeling/repair frequency, are higher, but not significantly so, than
those of any epoch of the Cenozoic (text-fig. 3b). As is the case for drilling, peeling/repair values
for each Cenozoic epoch are not significantly different from any other (Mann-Whitney (7-tests,
0-10 > P > 0-05).

Geographic patterns

Dudley and Vermeij (1978) stated that their data showed ‘a distinct latitudinal trend in drilling
predation for Recent species of Turritella with tropical and subtropical shells showing drilling
frequencies roughly three times those of temperature shells. Exclusion of the temperate
Tachvrhvnchus species from the group in our data set does not greatly alter this pattern. Of 212
shells collected above 30° latitude, 33 (15-6%) were drilled; of 766 shells collected below 30°, 226
(29-5%) were drilled.

Among fossil species only limited latitudinal comparisons are possible because of a paucity of

text-fig. 4. Whorl profiles of the 27 species examined in the present study (Appendix 1). a. T. abrupta,
b. 7. alabamiensis , C. T. aldrichi , d. T. altilira , e. T. alveata , F. T. apicalis , G. T. carinata , h. T. cumberlandia ,
i. T. rina , J. T. rubicollis , K. T. eurynome , l. T. femina , M. T. gilberti , N. T. gladeensis, o. T. humerosa ,
p. T. indent a, q. T. larensis , R. T. postmortoni, s. T. praecincta, t. T. mississippiensis , u. T. mortoni,
v. T. multilira , w. T. perattenuata , x T. perdita, Y. T. plebeia, z. T. pontoni, aa. T. wagneriana.

ALLMON ET AL .: DRILLING OF GASTROPODS

599

approximately isochronous and geographically widespread samples. Turritella mortoni,
T. praecincta and T. humerosa from the Palaeocene Aquia Formation of Maryland and Virginia lived
at approximately the same time as T. praecincta, T. postmortoni, T. multilira and T. eurynome
from the Palaeocene Tuscahoma and Nanafalia Formations of Alabama (e.g. Hazel et al. 1984;
Ward 1985). All of these species are relatively large forms with moderate to well-developed spiral
sculpture (text-fig. 4). Of 137 shells of the four lower latitude Alabama species, 11 (8-0%) were
drilled and the peeling/repair frequency was 0-299. Of 179 shells of the three higher latitude Aquia
species, 33 (18-4%) were drilled and the peeling/repair frequency was 0 078. The single species in
common to the two areas, T. praecincta, showed a lower frequency of drilling (17% vs. 38%) and
higher frequency of peeling/repair (0-36 vs. 0-035) in Alabama than in Virginia.

Predation and shell form

When ranked on a subjective scale of strength of sculpture, highly and moderately sculptured
species show relatively low frequencies of drilling (text-fig. 5a). Sculpture and drilling frequency are
significantly negatively correlated (Me st, 0 025 > P < 0 05). Despite this relationship, the nine most
highly sculptured species as a group are not significantly less frequently drilled than all other species

Most sculptured Least sculptured Most sculptured Least sculptured

60

# 60

>,

o

c

e

or 40
®

o»

c

®

® 20
Q.

0

0 20 40 0Oq 10 20 30

n*52 y=9.02*0.382X FT0.34

# 60

DO ■ »

n- 24 y.18,76*0,64x R.0.18 0

B

a

l H Q Hi

IV

a

® o

a

O Q

0

0 0

0 °

0 0 i o

j OO 0

Species

Species

text-fig. 5. a. Drilling frequency vs. development of external spiral sculpture in 52 of the 67 turriteliine species
listed in Appendices 1 and 2. Species are arranged subjectively from least to most sculpture development along
the horizontal axis as follows: 1. T. praecincta , 2. T. postmortoni , 3. T. mauryana , 4. T. mortoni, 5. T. larensis,
6. T. abrupta, 7. T. rina, 8. T. berjadinensis, 9. ‘ T. carinifera Lamarck’ (Recent, South Africa), 10. T. gatunensis,
11. T. bicarinata , 12. T. altilira, 13. T. exoleta, 14. T. pagoda, 15. T. apicalis, 16. T. mariana, 17. T. eurynome ,
18. T. multilira, 19. T. leucostoma, 20. T. variegata, 21. T. carinata, 22. T. trilira, 23. T. symmetrica,
24. T. femina, 25. T. vertebroides, 26. T. badensis, 27. T. cumberlandia, 28. T. nodulosa, 29. T. perattenuata,
30. 7. pontoni, 31. T. gladeensis, 32. T. variabilis , 33. T. mississippiensis , 34. T. humerosa, 35. T. rubricollis,
36. T. aldrichi, 37. T. subangulata , 38. T. imbricataria , 39. T. bcinksi, 40. T. alabamiensis, 41. T. gilberti,
42. T. wagneriana , 43. T. howelli, 44. T. alveata, 45. T. bieniaszi , 46. T. indenta, 47. T. gonostoma, 48.
T. perdita, 49. T. acropora, 50. T. annulata, 51. T. communis, 52. T. plebeia. Open squares represent our
observations (Appendix 1); solid squares represent previously published data (Appendix 2).

b. Peeling frequency vs. development of external spiral sculpture in the 27 turriteliine species we have examined.
Species are arranged subjectively along the horizontal axis, and numbered as follows: 1. T. praecincta, 2.
T. postmortoni, 3. T. mortoni, 4. T. abrupta, 5. T. larensis , 6. T. rina, 7. T. altilira, 8. T. apicalis, 9. T. eury-
nome, 10. T. multilira, 11.7’. carinata, 12. T. femina, 13. T. cumberlandia, 14. T. perattenuata, 15. T. pontoni,
16, T. gladeensis, 17. 7. mississippiensis, 18. T. humerosa , 19 .7. rubricollis, 20. T. aldrichi, 21. 7. alabamiensis,
22. 7. gilberti, 23. 7. wagneriana, 24. 7. alveata, 25. 7. indenta, 26. 7. perdita , 27. 7. plebeia. In an attempt to
increase interpretability, groups of species indicated by Roman numerals were designated a priori on the basis
of general similarity of degree of sculpture development; I = highly sculptured, II = moderately sculptured,
III = weakly sculptured, IV = unsculptured.

600

PALAEONTOLOGY, VOLUME 33

together (Mann-Whitney U- test, P = 048). Similarly, the five most highly sculptured species as a
group show lower peeling/repair frequency, although not significantly so (Mann-Whitney (7-test,
/> = 045; text-fig. 5b). Peeling/repair frequency declines with increased sculpture, but not
significantly so (t- test, P = 0-135). Frequencies of peeling/repair and drilling among less sculptured
species range from very high to very low. (WDA and RDN independently arranged the 27 species
in text-figure 5b according to their own judgments of ‘strength of sculpture'; results for
peeling/repair and drilling frequencies were essentially identical for both arrangements, suggesting
confidence in the patterns despite subjectivity of the method.)

To assess the influence of spire height on peeling/repair frequency we computed length: width
ratios for each shell and compared the distribution of peeled shells with the total sample. The two
distributions are not significantly different (G-test, P = 0-15). Drilling frequency and peeling/repair
frequency decline with increasing size of the largest whorl (text-fig. 6).

drilling frequency (%) peeling/repair frequency

text-fig. 6. Mean size of largest whorl measured for each species plotted against frequency of a, drilling and
b, peeling/repair. Size is negatively correlated with predation frequency in both cases, but not significantly so
with peeling/repair. The five most strongly sculptured species are represented by solid squares. Elimination of
these species does not change the correlation between size and drilling (P = 0-027), but does further reduce the
significance between size and peeling/repair {P = 0-48).

A a o pt £1# 4 tip 4 6 a a& H &P&

RECENT-
PLEISTOCENE

PLIOCENE
MIOCENE
OLIGOCENE
EOCENE

PALAEOCENE
CRETACEOUS

text-fig. 7. a. Classification of 197 Cenozoic and 43 Cretaceous turritelline species from North, South and
Central America and the Caribbean into nine whorl profile types proposed by Ida (1952) and used by Marwick
(1971). This figure includes approximately 75% of the described species of Turritella from the New World, b.
Predation frequencies (drilling in dark shading, peeling/repair in lighter shading) for turritelline species
according to whorl classification used in text-fig. 4. Text-figure based on only those species in common between
the compilation used in text-fig. 4 and Appendices 1 and 2.

ALLMON ET A L.. DRILLING OF GASTROPODS

601

In turritellines, whorl profile can be examined separately from the external, mostly spiral
sculpture (Marwick 1971 ; Allmon 1988/?). In a sample of New World species, both the distribution
of whorl profile types (text-fig. 7a) and the distribution of predation frequencies (text-fig. 7b), show
little or no temporal pattern in drilling and peeling/repair data.

Size

Size distributions of drilled and undrilled shells in our data set are basically similar (text-fig. 8). For
the 27 species we examined (although not for turritellines as a whole) size and sculpture
development are significantly positively correlated (Mest, P < 0-001) (text-fig. 9).

text-fig. 8. Comparison of size distribution of drilled
and undrilled shells in 1097 shells examined in the
present study.

text-fig. 9. Plot of size (represented by maximum
whorl diameter on Y axis) against the subjective
ranking of sculpture development used in text-fig. 5b.

Incomplete drill holes

Only nine incomplete drill holes were observed in our sample: three each on specimens of T. ahrupta
and T. eurynome , and one each on specimens of T. praecincta , T. larensis and T. humerosa.

Predator and prey behaviour

Data on position of the drill hole relative to intact apertures indicate a high degree of selectivity on
the part of the predators (text-fig. 2). Most holes are located on the half of the shell just behind the
aperture (quadrants numbered II and III). This pattern is significantly different from the null
hypothesis of a uniform distribution of holes around the circumference of the shell (G-test,
TcO-Ol). Drill holes vary in their vertical position on the whorl, but most (54%, N = 170) are
located near the centre, rather than closer to either upper or lower sutures. Occasional holes are
observed straddling a suture (text-fig. 1 b). Diameter of the drill hole (a measure of predator size [e.g.
Wiltse 1980]) plotted against maximum whorl diameter (a measure of prey size) shows that size of
predator and prey are highly correlated (text-fig. 10).

602

PALAEONTOLOGY, VOLUME 33

text-hg. 10. Plot of drill hole diameter against
diameter of drilled whorl.

DISCUSSION

Problems of pattern recognition

Studies of predation in the fossil record are potentially biased by a large number of factors. In any
ecosystem, predation frequency on a given species will vary with abundance of predators, number
of alternative prey, and the ranked preference relative to other prey. Only some of these factors can
be assessed in the fossil record (e.g. Stanton and Nelson 1980; Stanton et al. 1981); for those that
can be studied (e.g. abundance of naticids vs. potential prey species), few data are so far available
from the coastal plain.

Preservable traces of predation may represent only a small fraction of actual predation on
preservable prey species (Signor 1985; Vermeij 1987). Ansell and Morton (1987) have shown, for
example, that some bivalves may be suffocated by naticid predators, leaving an incomplete borehole
or no trace at all. Naticids are also known to attack gastropods through the aperture (Edwards
1969; Hughes 1985). Bottom-feeding fishes and asteroid echinoderms may be locally important
predators of Recent turritellines (Allmon 1988a, and references therein), and will leave little or no
record of such behaviour in the fossil record.

Interpreting possible evolutionary consequences of predation can be complicated in turritellines
by the potential importance of non-shell characters in resisting predation. These include: (1) deep
withdrawal into the high-spired shell (Vermeij et al. 1980; Vermeij 19826, p. 708; 1987, pp. 1950'.;
Allmon 1988a), (2) active escape by burrowing or crawling (Allmon 1988a), (3) seeking shelter
among rocks (e.g. T. banksi - Dudley and Vermeij 1978) or in sponges (e.g. ‘ T. carinifera
Lamarck’ - Kilburn and Rippey 1982), and (4) predator avoidance by small, patchily distributed
populations or predator saturation by very large populations (Allmon 1988a).

Sample size must also be considered. Although our total data set is large, the sample of
Oligocene and Cretaceous species is small. Both of these time periods are characterized by the lowest
drilling frequencies we observe. Clearly one must wonder if these periods of reduced drilling
intensity are real or are sampling artefacts.

Some or all of these factors may be responsible for the high variability in most of our results.
Variability occurs within single species from the same area (e.g. T. plebeia from the Miocene of
Maryland and Virginia, which shows drilling frequencies of 0-30%), within single species from
different areas (e.g. T. badensis is 17% drilled in a Polish sample and 40-4% in a Bulgarian sample),
among species within single time periods (e.g. high standard deviations in text-fig. 3), and among
species within the group as a whole (text-fig. 5). Substantial variability also exists between species
occurring in the same formation, and even the same outcrop.

Traces of drilling and peeling can represent two different phenomena (Vermeij 19826, 1987).
Frequency of incomplete drill holes (on shells without complete holes; see Kitchell et al. 1986;
Vermeij et al. 1989) and frequency of peeling/repair represent frequency of unsuccessful predation.

ALLMON ET AL.. DRILLING OF GASTROPODS

603

On the other hand, frequency of complete drill holes represents the frequency of successful
predation. High frequency of unsuccessful predation indicates that predators are exerting strong
selection pressure on prey, and that prey are successfully resisting most attacks (Vermeij 1 982c/, b).
Low frequency of unsuccessful predation indicates either that few prey are being attacked (either
because there are few predators or because by hiding or escape the prey avoid even the initiation
of attack), or that most attacks are lethal. Low frequency of successful predation could indicate
either that predators are rare or that prey avoid detection and capture. It is important to add that
high frequency of successful predation may be evidence of intense demographic pressure (high
predation-induced mortality) on a prey population, but not necessarily intense selective pressure for
antipredatory adaptations (Vermeij 1 982/?).

Predator and prey behaviour

Although much is known about how naticids attack other molluscs (e.g. Ziegelmeier 1954; Sorensen
et al. 1955; Fretter and Graham 1962; Gonor 1965; Edwards 1969; Taylor 1970; Berry 1982;
Hughes 1985) naticid predation behaviour on turritellines is unknown. Observations of living
turritellines (Allmon 1 988c/) show that the half of the shell comprising the two most-drilled
quadrants in this study (II and III, text-fig. 2) is normally uppermost when the animal is actively
crawling (aperture parallel to the substrate); quadrants I and II are normally uppermost when the
animal is in a sedentary feeding position (aperture perpendicular to the substrate). Studies with
living naticids and turritellines are required to determine whether naticids prefer the dorsal side of
the turritelline shell, and actively manipulate them to this end, or whether the dorsal side is drilled
because turritellines are frequently active crawlers.

The vertical distribution of holes over the whorls suggests that predators actively select the
thinnest part of the shell for drilling. The correlation between size of drilled whorl and maximum
whorl size of the drilled shell (text-fig. 1 1 ) suggests that for prey of any size, drilling predators tended
to choose a whorl that was in the same relative position, usually two to three whorls behind the
aperture (text-fig. 1a).

text-fig. 1 1 . Plot of diameter of drilled whorl against
maximum shell diameter of the drilled shell.

Maximum shell diameter (mm)

Geographic patterns

Geographic patterns in predation are more difficult to study in the fossil record than in Recent
assemblages (e.g. Dudley and Vermeij 1978; Vermeij et al. 1989). In the case of the Palaeocene
species from Virginia and Alabama discussed above, the situation is exactly the reverse from that
seen among living species, with lower latitude species showing lower rates of drilling. Other factors
may be involved in this case, including a relatively small latitudinal difference between the two areas
(< 10°), a weaker latitudinal temperature (and predation?) gradient during the early Tertiary
compared to today (cf. Vermeij et ai 1980), and the small number of species considered.

604

PALAEONTOLOGY, VOLUME 33

Evolutionary significance

The idea that predation intensity in some way affects gastropod shell morphology has been
considered by many authors (e.g. Vermeij 1978, 1982a, c, 1987; Raffaelli 1978; Hughes and Elner
1979; Palmer 1979; Hughes 1980; Bertness and Cunningham 1981 ; Johannesson 1986; Thomas and
Himmelman 1988), but we know little about the actual involvement of predation in the
morphological evolution of a single prey clade. In the case of turritellines, Dudley and Vermeij
(1978) have suggested that extreme development of carinae in T. postmortoni and T. praecincta from
the Palaeocene of Alabama conferred protection from predation, since these species showed low
drilling frequencies in their data. Our results show that well-developed sculpture confers only slight,
if any, protection from either drilling or peeling attacks. While a significant negative relationship
exists between sculpture and drilling frequency overall (text-fig. 5a), the most highly sculptured
species are not significantly less drilled as a group than are less sculptured species as a group.
Furthermore, the weak relationship that is observed between sculpture and drilling intensity can be
explained as an effect of prey size.

The most sculptured species are larger than less sculptured species (text-fig. 9). The larger size of
strongly sculptured taxa contributes to their lower frequencies of drilling and peeling (text-fig. 6).
Since large size is correlated with reduced predation intensity, any correlations between sculpture
development and predation are confounded by size effects (A. R. Palmer, pers. comm.). The size
effect, however, should strengthen the correlation between highly sculptured, large species and low
levels of predation. In fact, this correlation is quite poor, and shows that despite the help size effects
proved to boost the fit, sculpture development is very poorly related to predation frequency.

Anecdotal indications of a relationship between sculpture and predation come from examination
of co-occurring species. T. mortoni and T. humerosa , for example, occur in the Upper Palaeocene
Aquia Formation of Maryland and Virginia. T. mortoni has more pronounced sculpture, consisting
of several carinae, the largest of which is very strong and occurs near the base of the whorl.
T. humerosa has fainter sculpture, consisting of fine spiral lines over the entire whorl and a
pronounced but rounded subsutural collar. T. humerosa shows a drilling frequency of 25% and a
peeling/repair frequency of 0 03, while T. mortoni shows 5-5% drilling and a peeling/repair
frequency of 0 115. In the Matthews Fanding Member of the Upper Palaeocene Porters Creek
Formation in Alabama, T. alabamiensis shows a drilling frequency of 10% and a peeling/repair
frequency of 0-03, while T. aldrichi is 59 % drilled and has a peeling/repair frequency of 0-78. These
two species are of similar size, but late whorls of T. alabamiensis are basally convex while those of
T. aldrichi are more straight-sided and usually bear a weak adapical carina. Neither species shows
pronounced spiral sculpture.

These individual examples (all of which need more detailed study) notwithstanding, there is little
convincing evidence that sculpture yields any consistent advantages in resisting predation on
turritellines.

Shell geometry is also uncorrelated with peeling and repair frequencies in turritellines; slender
and robust species suffer statistically equivalent frequencies. This finding is contrary to Signor’s
(1985) results for Recent terebrid species in Guam, where robust forms sustained not only much
higher rates of both successful and unsuccessful predation than did slender forms, but also higher
rates of repair. Slender species are somehow better able to avoid detection or capture by peeling
predators. Signor tentatively concludes that smaller aperture size in slender forms is responsible, by
preventing access to a crab’s cheliped.

It is possible that the turritellid shell shape itself is an adaptation against peeling predation (since
it allows deep withdrawal of the body mass); this might help explain its recurrence in a variety of
gastropod groups since the Devonian (e.g. Signor 1984).

Two temporal patterns stand out in our results. The first is the Cenozoic and apparent Fate
Cretaceous stability in predation frequencies. The second is the timing of establishment of this
stability. While frequencies of drilling and peeling/repair do fluctuate from the Late Cretaceous to
the Recent, none of these changes is significant at the five percent level of confidence. Late

ALLMON ET AL. : DRILLING OF GASTROPODS

605

Palaeocene drilling frequencies are approximately as high as those at any other time in the Cenozoic,
and modern peeling frequencies are no different from those in the Late Cretaceous. If predation on
turritellines did substantially increase, as suggested by previous workers, this increase must have
happened prior to the Late Cretaceous.

The low incidence of incomplete drill holes in our total sample (9 of 1097 specimens) is consistent
with previous findings (e.g. Vermeij and Dudley 1982), and suggests that, as in most other
gastropods, turritellines are not very successful at resisting drilling once subjugated (cf. Vermeij
19826, 1987, p. 210). Late Cretaceous peeling/repair values (text-fig. 3b) are higher than mean
values for any epoch in the Cenozoic. This agrees with earlier findings for turritellines and other
Cretaceous gastropod groups (Vermeij and Dudley 1982; Vermeij 1987, p. 229).

The overall temporal patterns we see in turritellines agree with patterns observed in some
gastropod groups but not in others. As in turritellines, Terebridae show little or no temporal trend
in predation frequency during the Cenozoic (Vermeij et al. 1980). On the other hand, Conidae
display increased incidence of peeling and repair from the Eocene to the Miocene (Vermeij 1987,
p. 231). Neither Conidae nor Terebridae have a Cretaceous record (Taylor et al. 1980), so if
predation significantly affected the evolution of these groups, it did so at varying rates and times.

Vermeij (1977) proposed the ‘Mesozoic marine revolution’ as an arms race between newly
evolved durophagous predators and their prey, and an explanation for the largely Mesozoic
appearance of anti-predatory shell structures in many groups of marine invertebrates. Our results
for turritellines (and Vermeij’s own findings for terebrids) suggest that after the appearance and
initial diversification of a prey group, there may be few long-term trends in shell form or structure
aimed at resisting drilling or peeling predators. The evolution of anti-predatory shell structures may
have occurred relatively rapidly in some prey groups and then advanced no further. In the case of
turritellines, any gradual evolution of anti-predatory shell structures, or increase in predation
intensity, must have been a completely Mesozoic phenomenon because long-term trends in shell
form, sculpture, and predation intensity are absent in the Cenozoic. Any ‘arms race’ appears to be
at a standstill in so far as turritelline shell design is concerned. These conclusions are consistent with
previous results (Vermeij et al. 1981; Vermeij 1987, pp. 227ff.) that suggested that predation
intensities attained essentially modern levels near the end of the Mesozoic, and remained essentially
unchanged thereafter.

None of these results contradicts the basic notion of the marine revolution. The history of drilling
and peeling/repair in turritellines may suggest that they switched from shell-based defence to
behavioural or soft-anatomy defences against predators. Alternatively, both predators and prey
may have reached an impasse early on, with neither group able to achieve a morphological
breakthrough that might permit a renewal of the race. Although the ‘marine revolution' has
resulted in many highly armoured prey species, not all of these species have gradually evolved
greater predation resistance after their initial appearance.

CONCLUSIONS

We find remarkable stasis in the overall spectrum of shell design in turritellines, a situation mirrored
in the absence of temporal trends in the frequency of drilling and peeling attacks on these
gastropods. Frequencies of drilling and peeling/repair have not changed significantly since the Late
Cretaceous. Unfortunately, low sample sizes in the Cretaceous prevent us from determining exactly
when these frequencies reached Cenozoic levels. The strength of shell sculpture is uncorrelated with
drilling and peeling/repair frequencies, suggesting that sculpture itself is currently of little value as
a predator defence. The lack of correlations between shell form and predation, and the absence of
temporal trends in predation intensity, suggest that turritelline shell structure has not evolved
continuously in response to predation, at least not during the Cenozoic. Any gradual evolutionary
trends in turritelline shell evolution that may have occurred could only have taken place in the
Mesozoic.

606

PALAEONTOLOGY, VOLUME 33

Acknowledgements. We are grateful to R. L. A. Allmon, M. J. Benton, A. R. Palmer, P. W. Signor, J. D.
Taylor, R. D. Turner, G. J. Vermeij and an anonymous reviewer for comments on earlier drafts of the
manuscript. This research was supported by grants from the Geological Society of America, Sigma Xi, and the
Department of Earth and Planetary Sciences of Harvard University, funds from the Department of
Invertebrate Paleontology of the Museum of Comparative Zoology, and a National Science Foundation
Graduate Fellowship to WDA. Order of authorship is alphabetical.

APPENDIX I

Frequencies of drilling and peeling in 27 fossil species of Cenozoic turritellines. All specimens are in the
collection of the Department of Invertebrate Paleontology of the Museum of Comparative Zoology.

Frequency

Species

Drilling

Peeling/

(Age)

Locality

Formation

N

(%)

repair

T. abrupta Spieker
Miocene

T. alabamiensis Whitfield

Venezuela

?

23

4

000

Palaeocene
T. aldrichi Bowles

Alabama

Porters Creek

30

10

0-03

Palaeocene

Alabama

Porters Creek

32

59

0-31

T. altilira Conrad
Pliocene

T. a Iveat a Conrad

Panama

Gatun

20

25

0-30

Eocene

Mississippi

Moodys Branch

15

13

0-93

Eocene

Louisiana

Moodys Branch

30

60

0-37

T. apicalis Heilprin
Pliocene

T. carinata I. Fea

Florida

Pinecrest

58

17

0-21

Eocene

T. cumberlandia Conrad

Alabama

Gosport

30

13

0-23

Miocene

Maryland

Calvert

15

13

0-33

T. eurynome Whitfield
Palaeocene
T. femina Stenzel

Alabama

Nanafalia

30

10

0-60

Eocene

T. gilberti Bowles

Texas

Weches

30

3

000

Palaeocene

T. gladeensis Mansfield

Alabama

Bashi

45

29

002

Pliocene

Florida

Pinecrest

22

27

0-00

22

23

0-41

T. humerosa Conrad
Palaeocene
T. indenta Conrad

Virginia

Aquia

32

25

003

Miocene

T. larensis Hodson

Maryland

Calvert

18

67

0T7

Miocene

Venezuela

?

50

6

006

T. mississippiensis Conrad
Oligocene
T. mortoni Conrad

Mississippi

Byram

30

3

0-03

Palaeocene

Virginia

Aquia

80

4

0-03

15

17

0-20

ALLMON ET A L.: DRILLING OF GASTROPODS

607

APPENDIX I (com.)

Frequency

Species Drilling Peeling/

(Age)

Locality

Formation

N

(%)

repair

T. multilira Whitfield

Palaeocene

Alabama

T uscahoma

30

7

0-30

Palaeocene

Alabama

Nanafalia

16

0

0-25

18

6

0-22

T. perattenuata Heilprin

Pliocene

Florida

Pinecrest

48

2

0-51

T. perdita Conrad

Eocene

Mississippi

Moodys Branch

30

7

013

T. plebeia Conrad

Miocene

Maryland

Calvert

30

0

000

T. pontoni Mansfield

Pliocene

Florida

Pinecrest

49

31

0-25

27

37

0-70

T. postmortoni Harris

Palaeocene

Alabama

Nanafalia

15

0

0-20

T. praecincta Conrad

Palaeocene

Virginia

Aquia

30

40

007

22

36

000

Palaeocene

Alabama

Tuscahoma

7

0

0-29

4

50

0-75

19

0

005

T. rina Palmer

Eocene

Alabama

Lisbon

38

13

0-05

T. rubricollis MacNeil

Oligocene

Mississippi

Mint Spring

15

0

0-07

T. wagneriana Olsson & Harbison

Pliocene

Florida

Pinecrest

72

19

0-74

APPENDIX 2

Data from previously published sources

on drilling and peeling frequencies

in fossil and

living turritellines.

Species

Age

Locality

N

Drilling

(%)

T. acropora Dali

Recent

Bermuda1

41

12

T. annulata Kiener

Recent

Ghana2

436

62-6

T. badensis Sacco

Miocene

Poland3

1229

17-0

Miocene

Bulgaria1

1921

40-4

T. banksi Reeve

Recent

Panama1

12

16-7

Recent

Ecuador1

9

0

T. berjadinensis Hodson

Miocene

Venezuela1

35

90

T. bicar inata (Eichwald)

Miocene

Poland3

112

28-0

T. bieniaszi Friedberger

Miocene

Bulgaria4

532

36-6

‘ T. carinifera Lamarck ’

Eocene

France5

51

141

608

PALAEONTOLOGY, VOLUME 33

APPENDIX 2 (cont.)

Species

Age

Locality

N

Drilling

(%)

' T. carinifera Lamarck ’

Recent

South Africa1

6

0

T. communis Risso

Recent

Shetlands1

107

110

T. erronea Cossmann

Miocene

Poland3

120

25-0

T. exoleta (Linnaeus)

Recent

Tobago1

48

4-3

T. funiculosa Deshayes

Eocene

France5

77

58-4

T. gatunensis Conrad

Pliocene

Panama1

70

640

T. gonostoma Valenciennes

Recent

Mexico1

70

40-0

T. gramdata Sowerby

Cretaceous

(Albian)

England6

704

3-7

T. howelli Harbison

Cretaceous

(Campanian)

Mississippi1

83

10

(0)*

Cretaceous

(Campanian)

Mississippi7

16

6-3

(0-47)*

T. imbricataria Lamarck

Eocene

France5

45

17-8

T. leucostoma Valenciennes

Recent

Mexico1

35

260

T. mariana Dali

Recent

Mexico1

36

360

T. mauryana Newton

Eocene

Nigeria8

6

16-7

T. mortoni Conrad

Palaeocene

Virginia1

14

7-0

T. nodulosa King & Broderip

Recent

Panama1

151

260

T. pagoda Reeve

Recent

New Zealand1

50

2-0

T. perdita Conrad

Eocene

Mississippi1

70

210

T. plebeia Conrad

Miocene

Maryland1

101

27-7

Miocene

Maryland9

416

210

T. postmortem i Harris

Palaeocene

Alabama1

12

0

T. praecincta Conrad

Palaeocene

Alabama1

27

0

T. subangulata d’Orbigny

Miocene

Bulgaria4

189

21-7

T. symmetrica Hutton

Recent

New Zealand1

57

19 3

T. tricarinata

Pliocene

Italy10

19

52-6

T. trilira Conrad

Cretaceous

(Campanian)

Alabama1

13

0

T. triplicata (Brocchi)

Recent

Libya1

34

47 1

T. unisulcata Lamarck

Eocene

France5

14

500

T. variabilis Conrad

Pliocene

Florida1

40

200

T. variegata (Linnaeus)

Recent

Gulf of Mexico1

53

120

T. vertebroides Morton

Cretaceous

(Campanian)

Mississippi1

1 1

90

T. sp.

Recent

Mozambique1

47

19-0

T. sp.

Recent

Philippines1

138

68-0

T. sp.

Recent

India1

61

10-0

T. sp.

Cretaceous

(Campanian)

Mississippi7

17

120

(0-41)*

T. sp.

Cretaceous

(Campanian)

Mississippi7

13

7-8

(0 69)*

T. sp.

Cretaceous

England6

2

0

Key. ‘Dudley and Vermeij (1978); 2Buchanan (1958); 3Hoffman et al. (1974); ‘Kojumdjieva (1974); 3Fischer
(1966) 'Taylor et al. (1983); 7 Vermeij and Dudley (1982); 8Arua (1982); “Kelley (1982); 10Robba and Ostinelli
(1975);

* Percent peeling.

ALLMON ET AL.: DRILLING OF GASTROPODS

609

REFERENCES

abbott, r. r. 1974. American seashells. Van Nostrand Reinhold, New York, 663 pp.

a decoke, o. s. and tevesz, M. J. s. 1974. Gastropod predation patterns in the Eocene of Nigeria. Lethaia , 7,

1 7-24.

allmon, w. d. 1987. Multiple modes of homeomorphy in Cenozoic turritelline gastropods and their
evolutionary implications. Geological Society of America, Abstracts with Programs , 19, 570.

— 1988a. Ecology of Recent turritelline gastropods (Prosobranchia, Turritellidae) : current knowledge and
paleontological implications. Palaios , 3, 259-284.

— 19886. Evolution and environment in turritelline gastropods. (Mesogastropoda, Turritellidae), Lower
Tertiary of the U.S. Gulf and Atlantic coastal plains. Unpublished PhD thesis, Harvard University, 818 pp.

ansell, a. d. and morton, b. 1987. Alternative predation tactics of a tropical naticid gastropod. Journal of
Experimental Marine Biology and Ecology, 111, 109-119.
arua, i. 1982. Borings and shell damage in Eocene gastropods: southeastern Nigeria. Palaeogeography ,
Palaeoclimatology, Palaeoecology , 38, 269-282.

— and hoque, m. 1989. Predatory gastropod boreholes in an Eocene molluscan assemblage from Nigeria.
Lethaia, 22, 49-59.

barker, r. t. 1983. The deer flees, the wolf pursues: incongruencies in predator-prey coevolution. 350-382. In
futuyma, d. j. and slatkin, m. (eds. ). Coevolution. Sinauer Associates, Sunderland, Massachusetts.
bayne, b. L. 1981. Theory and observation: benthic predator-prey relationship. 127-158. In longhurst, a. r.

(ed.). Analysis of marine ecosystems. Academic Press, London.
berggren, w. a., Kent, d. v., flynn, j. J. and van couvering, ). a. 1985. Cenozoic geochronology. Geological
Society of America Bulletin, 96, 1407-1418.

berry, a. j. 1982. Predation by Natica maculosa Lamarck (Naticidae: Gastropoda) upon the trochacean
gastropod Umbonium vestiarum (L.) on a Malaysian shore. Journal of Experimental Marine Biology and
Ecology, 64, 71 -89.

bertness, m. d. and Cunningham, c. 1981. Crab shell-crushing predation and gastropod architectural defense.

Journal of Experimental Marine Biology and Ecology, 50, 213-230.
buchanan, j. b. 1958. The bottom fauna communities across the continental shelf of Accra, Ghana (Gold
Coast). Proceedings of the zoological Society of London, 130, 1-56.
carriker, m. r. 1981. Shell penetration and feeding by naticacean and muricacean predatory gastropods: a
synthesis. Malacologia, 20, 403-422.

Dudley, e. c. and vermeij, g. j. 1978. Predation in time and space: drilling in the gastropod Turritella.
Paleobiology , 4, 436-441.

edwards, d. c. 1969. Predators on Olivella biplicata, including a species-specific predator avoidance response.
Veliger, 11, 326-333.

— 1974. Preferred prey of Polinices duplicatus in Cape Cod inlets. Bulletin of the American Malacological
Union , 40, 1 7-20.

fretter, v. and graham, a. 1962. British prosobranch molluscs. Ray Society, London, 755 pp.
gonor, J. J. 1965. Predator-prey reactions between two marine prosobranch gastropods. Veliger, 7,
228-232.

hazel, j. e., edwards, j. e. and bybell, l. m. 1984. Significant unconformities and the hiatuses represented by
them in the Paleogene of the Atlantic and Gulf coastal plain province. 59-66. In schlee, j. s. (ed.).
Interregional unconformities and hydrocarbon accumulation. American Association of Petroleum Geologists
Memoir, 34.

hoffman, a., pisera, a. and ryszkiewicz, m. 1974. Predation by muricid and naticid gastropods on the Lower
Tortonian mollusks from the Korytnica Clays. Acta Geologica Polonica, 24, 249-260.
hughes, r. n. 1980. Predation and community structure. 699-728. In price, j. h., irvine, d. e. g. and farnham,
w. f. (eds.). The shore environment ecosystems. Academic Press, London.

— 1985. Predatory behavior of Natica unifasciata feeding intertidally on gastropods. Journal of Molluscan
Studies, 51, 331-335.

— and elner, r. w. 1979. Tactics of a predator, Carcinus mcienas, and morphological responses of the prey,
Nucella lapillus. Journal of Animal Ecology, 48, 65-78.

ida, k. 1952. A study of fossil Turritella in Japan. Geological Survey of Japan Report, 150, 1-64.
johannesson, b. 1986. Shell morphology of Littorina saxati/is Olivi : the relative importance of physical factors
and predation. Journal of Experimental Marine Biology and Ecology, 102, 181-195.
keen, a. M. 1971 . Sea shells of tropical west America. Stanford University Press, Stanford, California, 1064 pp.

610

PALAEONTOLOGY, VOLUME 33

kelley, r>. 1982. The effect of predation on Miocene mollusc populations of the Chesapeake Group. 35^18. In
SCOTT, T. M. and upchurch, s. B. (eds.). Miocene of the Southeastern United States. Florida Department of
Natural Resources , Bureau of Geology , Special Publication 25.

KILBURN, R. and rippey, E. 1982. Sea shells of southern Africa. Macmillan South Africa (Publishers) (Pty.) Ltd.,
Johannesburg, 249 pp.

kitchell, j. a., boggs, c. h„ kitchell, j. F. and rice, j. a. 1981. Prey selection by naticid gastropods:
experimental tests and application to the fossil record. Paleobiology, 7, 533-552.

KITCHELL, J. A., BOGGS, C. H., RICE, J. A., KITCHELL, J. F., HOFFMAN, A. and MARTINELL, J. 1986. Anomalies in
naticid predatory behavior: a critique and experimental observations. Malacologia , 27, 291-298.
kojumdjieva, E. 1974. Les gasteropodes perceurs et leurs victimes du Miocene de Bulgarie de nord-ouest.

Bulgarian Academy of Sciences, Bulletin of the Geological Institute (Series Pate ontology), 23, 5-24.
marwick, J. 1957. Generic revision of the Turritellidae. Proceedings of the Malacological Society of London, 32,
144-166.

— 1971. New Zealand Turritellidae related to Zeacolpus Finlay. New Zealand Geological Survey
Palaeontological Bulletin, 44, 1-87.

merriam, c. H. 1941. Fossil turritellas from the Pacific coast region of North America. University of California
Publications in Geological Sciences, 26, 1-214.

palmer, a. r. 1979. Fish predation and the evolution of gastropod shell sculpture: experimental and
geographic evidence. Evolution , 33, 697—713.

raffaelli, d. G. 1978. The relationship between shell injuries, shell thickness and habitat characteristics of the
intertidal snail Littorina rudis Maton. Journal of Molluscan Studies, 44, 166-170.
robba. e. and ostinelli, f. 1975. Studi paleoecologici sul Pliocene ligure. I. Testimonianze de predazione sui
molluschi Pliocenici di Albenga. Rivista Italiana di Paleontologia e Stratigrafia, 81, 309-372.
signor, p. w., m. 1984. Abundance of turritelliform genera through the Phanerozoic: another predation-
related pattern? Annual Report of the Western Society of Malacologists (for 1984), 17, 19-20.
signor, p. w., in. 1985. The role of shell geometry as a deterrent to predation in terebrid gastropods. Veliger,
28, 179-185.

sohl, N. f. 1969. The fossil record of shell boring by snails. American Zoologist, 9, 725-734.

Sorenson, A., strohbeen, j. p. and whelchel, h. p. 1955. Molluscan predators and their prey. Annual Report
of the American Malacological Union (for 1955), 34.

stanton, R. J., jr. and nelson, p. c. 1980. Reconstruction of the trophic web in paleontology: community
structure in the Stone City Formation (Middle Eocene, Texas). Journal of Paleontology, 54, 118-135.
stanton, r. j., powell, e. N. and nelson, p. c. 1981 . The role of carnivorous gastropods in the trophic analysis
of a fossil community. Malacologia, 20, 451-469.

taylor, j. d. 1970. Feeding habits of predatory gastropods in a Tertiary (Eocene) molluscan assemblage from
the Paris Basin. Palaeontology, 13, 255-260.

— , cleeveley. r. j. and morris, n. j. 1983. Predatory gastropods and their activities in the Blackdown
Greensand (Albian) of England. Palaeontology, 26, 521-553.

— , morris, n. j. and taylor, c. n. 1980. Food specialization and the evolution of predatory prosobranch
gastropods. Palaeontology, 23, 375-409.
taylor, r. j. 1984. Predation. Chapman and Hall, New York. 166 pp.

thomas, m. L. H. and himmelman, j. H. 1988. Influence of predation on shell morphology of Buccinum unda-
tum L. on Atlantic coast of Canada. Journal of Experimental Marine Biology and Ecology, 115, 221-236.
toulmin, l. d. 1977. Stratigraphic distribution of Paleocene and Eocene fossils in the Eastern Gulf Coast
region. Alabama Geological Survey Monograph, 13, 1-602.
vermeij, g. j. 1 977 . The Mesozoic marine revolution : evidence from snails, predators and grazers. Paleobiology,
3, 245-258.

— 1978. Biogeography and adaptation. Patterns of marine life. Harvard University Press, Cambridge,
Massachusetts. 332 pp.

1982m Gastropod shell form, repair, and breakage in relation to breakage by the crab Calappa.
Malacologia, 23, 1-12.

1982 b. Unsuccessful predation and evolution. American Naturalist, 120, 701-720.

1982c. Phenotypic evolution in a poorly dispersing snail after arrival of a predator. Nature, 299, 349-350.
1983. Shell-breaking predation through time. 649-669. In tevesz, m. j. and mccall, f. (eds.). Biotic
interactions in Recent and fossil benthic communities. Plenum Press, New York.

1987. Evolution and escalation. An ecological history of life. Princeton University Press, Princeton, New
Jersey. 527 pp.

ALLMON ET A L.: DRILLING OF GASTROPODS

611

— and Dudley, E. c. 1982. Shell repair and drilling in some gastropods from the Ripley Formation (Upper
Cretaceous) of the South-eastern U.S.A. Cretaceous Research , 3, 397-403.

— , and zipser, e. 1989. Successful and unsuccessful drilling predation in Recent pelecypods. Veliger ,

32, 266-273.

— , schindel, d. e. and zipser, e. 1981. Predation through geological time: evidence from gastropod shell
repair. Science , 214, 1024-1026.

— , zipser, e. and Dudley, e. c. 1980. Predation in time and space: peeling and drilling in terebrid
gastropods. Paleobiology , 6, 352-364.

ward, l. w. 1985. Stratigraphy and characteristic mollusks of the Pamunkey Group (Lower Tertiary) and the
Old Church Formation of the Chesapeake Group - Virginia Coastal Plain. U.S. Geological Survey
Professional Paper , 1346, 1-78.

wiltse, w. i. 1980. Predation by juvenile Polinices duplicatus (Say) on Gemma (Totten). Journal of
Experimental Marine Biology and Ecology , 42, 187-199.
ziegelmeier, E. 1954. Beobachtungen liber den Nahrungsserwerb bei der Naticide Lunatia nitida Donovan
(Gastropoda Prosobranchia). Helgolander Wissenschaftliche Meeresuntersuchungen , 5, 1-33.

WARREN D. ALLMON

Department of Geology
University of South Florida
Tampa, FL 33620, U.S.A.

JAMES C. NIEH

Department of Organismic and
Evolutionary Biology
Flarvard University
Cambridge, MA 02138, U.S.A.

RICHARD D. NORRIS

Museum of Comparative Zoology
Flarvard University
Cambridge, MA 02138, U.S.A.

Typescript received 10 June 1989
Revised typescript received 29 August 1989

A DEDUCTIVE ENQUIRY SYSTEM FOR A
PALAEONTOLOGICAL DATABASE OF MUSEUM

MATERIAL

by M. J. ROGERS, D. T. DONOVAN and M. H. ROGERS

Abstract. The use of database management systems for cataloguing museum material is becoming
widespread. Recent advances in database management systems indicate that in future such systems will have
a range of increased facilities, including deductive capabilities and natural language interfaces for casual
enquiries. In order to explore some of the consequences of these developments for museum catalogues, a
prototype system has been written and examples of its use demonstrated. These show that such systems will
have significant research potential in a number of areas in geology in addition to taxonomy.

Our museums contain a wealth of information which can be used in taxonomic, stratigraphic,
palaeoecological and other geological studies. If these valuable resources are to be fully exploited,
a sophisticated catalogue system is necessary.

The first catalogues of fossils, or registers of specimens, were usually impressive books containing
a list of specimens numbered according to some local system, together with information concerning
geographic, stratigraphic and collection details. The entries were made in beautiful copperplate
writing but there was little attempt at ordering the entries. To obtain information from such
registers is almost impossible unless the registration number of the required material is known. The
register might be supplemented by card-index systems, one based on genus, one on stratigraphic
horizon, one of type and figured specimens and so on. Such indexes increase the usefulness of
museum material, but they are onerous to produce and bulky to store.

During the past twenty years progress has been made towards storing museum catalogue
information on computer databases. Price (1984) described how a hierarchical database manage-
ment system was designed by Cutbill et al. (1971) for material in the Sedgwick Museum, Cambridge.
Several museums now use improved versions of this early system, such as GOS (Porter 1982, 1983),
which was released by the Museums Documentation Association, or a more sophisticated version
MUSCAT. Other museums, including Bristol University Museum, use commercially available
database management systems, for example the relational databases dBASE III or Oracle.
International progress in this field was reviewed by Roberts and Light (1980), and Brunton et al.
(1985) discuss computing procedures in their Guidelines for the curation of geological material.

These systems have enormously increased the availability of information stored in our museums
because explicit queries can be made on any attribute recorded in the database, and lists are easily
prepared in which specimens are ordered alphamerically on a particular attribute such as name,
collector, geographic location or stratigraphic horizon.

Research into database management systems has continued at a high level, and during the past
few years advances have been made in several areas including database query languages, user
interfaces, deductive databases and recursive query processing. At the same time there has been an
increase in the use of remote databases and distributed databases, and many of these developments
will be incorporated into the next generation of commercial systems. Much of this work has been
done using logic programming languages, in particular Prolog, which have greatly facilitated the
construction of prototype systems incorporating some or all of these developments. Thus it is
possible to explore the consequences of these advances before commercial systems become widely

(Palaeontology, Vol. 33, Pari 3, 1990, pp. 613-622.|

© The Palaeontological Association

614

PALAEONTOLOGY, VOLUME 33

available. The relevance of some of these developments has already been recognised in a wide range
of disciplines; Rawlings (1988) gives a discussion of applications to molecular biology. Here we
investigate the consequences of applying these concepts to the cataloguing of museum material. We
describe our experience in designing a prototype system for the large Jurassic ammonites in the
University Museum, Bristol, and give some illustrations of the power of such a system in other fields
of geology, not just as a museum tool.

PROLOG AND DATABASES

In this section we give a brief explanation of why Prolog is particularly suitable as a research tool
in this area; for a full description of the language the reader is referred to one of the many excellent
books on the subject, for example Sterling and Shapiro (1986). The recent book edited by Gray and
Lucas (1988) gives a good account of recent developments in the use of Prolog in conjunction with
databases.

Prolog (programming in logic) is a programming language based on logic and was developed by
Colmerauer and his colleagues at Marseille in 1970 (Colmerauer et al. 1973). The use of the
language has grown rapidly, especially when it became clear in 1981 that logic programming would
play a fundamental role in the Japanese Fifth Generation Project. Prolog is now used in a wide
variety of applications including expert systems, artificial intelligence, natural language processing,
deductive databases and general problem solvers. An example in palaeontology is described by
Brough and Alexander (1986) who use Prolog to construct an expert system which provides
computer assisted identification of fossils. It has become clear that there is a close relationship
between Prolog and relational databases (Gallaire et al. 1984), and more generally between Prolog
and database query languages. As an example the Prolog predicate

f(X,Z): - g(X,Y),h(Y,Z)

has the declarative meaning that f(X, Z) is true for particular values of X and Z if g is true for X
and some value Y, and h is true for Y and Z. But in a relational database language this clause has
the interpretation that the relation f is the join of relations g and h. Nevertheless, there are major
differences between these two approaches. For example, Prolog possesses great flexibility in the data
structures it can handle, and allows recursion in both rules and facts. Relational database
management systems, on the other hand, have advantages in such areas as updating, efficient
handling of transactions, and concurrency.

There is currently a considerable amount of interest in combining these facilities in order to
produce a new generation of database query systems. One approach is to use Prolog to construct
a ‘front end’ which transforms a query into a standard form for execution on an existing database
management system (DBMS), and an example of this is given in Ghosh et al. (1988). Clearly,
developments along these lines which allow queries to be formulated in restricted natural language
will be of great benefit for casual enquiries to existing databases. Another approach has been to
establish deductive database systems containing predicates (or rules) which can be used to produce
facts (or tuples) which are not stored explicitly in the database.

In the present work we are examining a prototype system written in Prolog, where the data have
come partly from an existing relational database system containing the University of Bristol
collection of large ammonites, while the remaining data, the knowledge base, consisting of
appropriate chronostratigraphic and ammonite classification schemes and other relevant data, have
been entered ab initio. The aim is to illustrate, by means of a few examples, the type of query which
can be handled and to indicate some of its potential uses. No attempt has been made in this work
to produce a natural language front end; as explained above there is intense commercial activity in
this area and such products will undoubtedly become widely available in the future.

ROGERS ET AL : DEDUCTIVE ENQUIRY SYSTEM

615

THE KNOWLEDGE BASE: PROLOG DATA STRUCTURES FOR GEOLOGICAL

DATA

General

The particular aspects of geology which are discussed here are chronostratigraphy, palaeontological
classification, lithological horizon and locality. These aspects are of two main types. Chrono-
stratigraphic and palaeontologic classifications are both attempts to impose order on imperfectly
understood phenomena and are of a hierarchical form. They can be represented by a tree structure
in Prolog, but the quadtree, a particular form of tree used in geographic information systems
(Williams 1988), is too inflexible for the present purpose. The data structures defined in this paper
are more general as is necessary to accommodate the complexities of a chronostratigraphic and
palaeontologic information system. Both classifications have been modified over the years - and
will no doubt continue to be modified in the light of future research. In consequence, both types of
classification contain many obsolete terms. In contrast, lithological horizon and locality are
observed facts. Again, the same stratum or the same outcrop may be known by more than one
name, but in each case a name is associated with an observation, not an abstraction.

In the case of the classification schemes, not only is it necessary to define the structures, but also
the data must be internally consistent. As noted above, these schemes are continuously evolving.
However, in order to maintain internal consistency, it is necessary to choose a scheme and adhere
to it. The choice of a classification for a particular taxon or stratigraphic system is a task for a
specialist who should ensure that it is modern, widely accepted and readily available in the
literature. The Special Reports published by the Geological Society could be an appropriate source
for British chronostratigraphy and the Jurassic volumes (Cope, Getty et al. 1980; Cope, Duff el al.
1980) are selected here for the Jurassic system. The recent revision of the Treatise on invertebrate
paleontology is used for the ammonite classification (Donovan et al. 1981).

Because the lithological horizon and location are factual records, there is not the same problem
of internal consistency. In this prototype only one set of lithostratigraphic terms is used at
each locality, and no attempt has been made to incorporate a hierarchical classification for
lithostratigraphy.

Data structures for chronostratigraphic classification

The chronostratigraphic classification is a tree structure with two additional properties:

1 . The sub-trees (which can contain an arbitrary number of elements) are ordered in a time sense,
where the latest appears first in the list and the earliest (oldest) last.

2. Each level of the tree is referred to by a descriptor, which is used by geologists as a
chronostratigraphic division.

Thus, for example, a small subset of the scheme can be represented as:

Division Descriptor

cainozoic erathem

quaternary tertiary system

(LATE< EARLY) series

At the system level. Quaternary is later than Tertiary.

The actual representation used in this report has the following chronostratigraphic divisions (the
descriptors) :

erathem - system - series - stage - substage - zone - subzone

Each division is not always used. For example, the Aalenian Stage is divided into zones, whereas
the Bajocian Stage has two substages, the Upper Bajocian and Lower Bajocian, both of which are
divided into zones. The Prolog representation has been chosen so that it can accommodate such

616

PALAEONTOLOGY, VOLUME 33

variations of the basic data structure, and accordingly contains three arguments. The first is the
name of the node and will be a particular system or series etc. The second is the name (descriptor)
of its subdivisions, and the third is a list of those subdivisions with the convention that the elements
in the list are ordered, with the latest first and the earliest last. Thus, the general definition is of the
form

tree(Node, Subdivision, List)

and the particular example illustrated above is written as

tree(cainozoic,system,[quaternary,tertiary]).

A chronostratigraphic scheme based on this representation has been written down to the level of
subzone divisions for the Jurassic, using data given in Cope, Getty et al. (1980) and Cope, Duff
et al. (1980).

A further complication in chronostratigraphy is that obsolete terms are met when studying
specimens from old collections, or in older literature. The following simple structure is used to
incorporate such terms.

obsolete(Name, Division, Youngest_di vision, 01dest_division)

where Name is the obsolete name. Division is the chronostratigraphic level in the structure ‘tree’
of the terms YoungesUdivision and 01dest_division which are the upper and lower boundaries of
Name according to Cope, Getty et al. (1980) and Cope, Duff et al. (1980). For example, the Prolog
fact defining the Charmouthian is written as

obsolete(charmouthian,zone,davoei,jamesoni)

Using this representation we can now formulate in Prolog several rules of varying complexity
which can be used in a query system. Some examples are given below.

1. stratum(Y,X).

This states that Y is a division of X so that for example stratum(quaternary, cainozoic) means that
Quaternary is a division of the Cainozoic. The Prolog rule for this is:

stratum(Y,X) : - tree(X,_L),member(Y,L).

The query ?-stratum(Y, cainozoic). will, if asked repeatedly, generate all the divisions of the
Cainozoic, while the query ?-stratum(tertiary,X). will produce the single answer

X = cainozoic

2. substratum(Y,X).

This is similar to stratum(Y,X) defined above except that it spans an arbitrary number of
‘generations’. The Prolog definition for this involves two rules:

substratum(Y.X) : - stratum(Y,X).

substratum(Y,X) : - stratum(Y,Z), substratum(Z,X).

The query ?-substratum(Y, cainozoic). will, if asked repeatedly, generate all the divisions down to,
and including, the level of subzones which are in the Cainozoic Erathem. These rules are useful
building blocks to which other rules can be added to formulate powerful queries. For example

?-list xDo_y(X,Y,Lr).

lists all the elements, including X and Y, between X and Y, irrespective of the division to which they
belong, or their order. Thus

?-list_x„to_y(obtusum, bucklandi,L),list(L).

ROGERS ET AL.\ DEDUCTIVE ENQUIRY SYSTEM

617

will yield the following result

obtusum

turned

semicostatum

bucklandi

even though the obtusum Zone is in the Upper Sinemurian and the remaining zones are in the Lower
Sinemurian.

Data structures and Prolog representation for palaeontological classification

The palaeontological classification is a tree structure which is similar to that adopted for the
chronostratigraphic scheme. The main difference is that the subtrees at each stage are unordered,
whereas in the chronostratigraphic case there is a temporal ordering, with the latest first and the
earliest last in each subtree. Each level of the tree is referred to by a descriptor, for example order,
family, genus etc. A small example of the scheme is given by:

Division Descriptor

arietitidae family

arietitinae agassiceratinae sub-family

coroniceras arnioceras genus

The hierarchical ordering of the levels or taxa is:

suborder - superfamily - family - sub-family - genus - subgenus

Some of these levels are absent in particular cases, so that, for example, a family could be composed
of a set of genera. The general Prolog representation is of the form

ammclass(Node, Subdivision, List)

where Node is the name of a particular palaeontological category and Subdivision is the
classificatory level of the items in the list. The small example above is accordingly represented as

ammclass(arietitidae,subfamily, [arietitinae, agassiceratinae ...]).

Using this representation, a query pedigree(X) has been written which will list all the ancestors
of the argument X. The Prolog definition of pedigree is

pedigree(ammonitina) : - !

pedigree(X): - division(X,D),write(X),tab(3),write(D),nl,stratum(X,Y),pedigree(Y).

and the query

?-pedigree(arietites).

yields the result

arietites subgenus

coroniceras genus

arietitinae subfamily

psilocerataceae superfamily

An additional complication in the palaeontological classification scheme arises through the

618 PALAEONTOLOGY, VOLUME 33

widespread occurrence of synonyms. To deal with this, a further Prolog representation has been
introduced of the form

synonym(Name,Level,List_of_synonyms)

so that for example

synonym(schlotheimia,genus,[anguliferites,scamnoceras]).

indicates that Anguliferites and Scamnoceras are all synonyms of the genus Schlotheimia. This
structure is used when dealing with specimens which were identified according to earlier
classifications.

Link between the chrono stratigraphic and palaeontologic classification

The two schemes described are independent, and indeed each is independent of any classification
scheme. However, the link between the two is of paramount importance and accordingly a further
Prolog structure has been designed to represent this link. This is of the form

pal_range(Name,' Taxon, Division, Upper_occurrence,Lower_occurrence)

As an example of this

pal_range(caloceras,genus,zone,liasicus,planorbis).

represents the fact that the genus Caloceras occurs in the zone range liasicus to planorbis. It should
be noted here that it is only necessary for entries in this category to be made at the generic or
subgeneric level; the range for higher categories can easily be derived using simple Prolog
constructs.

A query gen_range has been written which, by searching both the chronostratigraphic and
classificatory trees, lists all genera found in the chronostratigraphic range R1 to R2. The program
is as follows

gen_range( R 1 , R2 ) : - list_x_to_y( R 1 ,R2,L), !,pal_range(N,genus,_,X, Y),
member(X,L),member(Y,L),
nl,tab( 1 5),write(N),fail.

so that for example the query

?-gen^range(obtusum,bucklandi).

gives the result

sulciferites

coroniceras

arnioceras

metarnioceras

tmaegoceras

agassiceras

euagassiceras

asteroceras

aegasteroceras

caenisites

epophioceras

xiphcroceras

microderoceras

promicroceras

ROGERS ET AL DEDUCTIVE ENQUIRY SYSTEM

619

Prolog representation for observed data

The observed data of lithological horizon and location are intimately related, so the two types of
information are combined in one data structure.

The general form is

place(Grid_easting,Grid_northing,Locality_name,List)

List elements are of the form

correlation(Lithologic„horizon,Thickness,Chronostratigraphic_division,Upper_occurrence,
Lower occurrence)

The fact

place(368300, 165000,corston_field_quarry,[correlation(blueJias,2-4,zone,angulata,angulata)]).

records the fact that the outcrop at Corston Field Quarry (368300,165000), consists of 2-4 metres
of Blue Lias, all in the angulata Zone.

In this Prolog structure, the observed data of the lithological horizons recorded at a section is
linked to the chronostratigraphy according to the current interpretation. A revision of this
correlation in the light of new evidence can easily be incorporated by a simple modification of the
data relating to that section.

Many queries can be asked using the data structures so far described which are of use to
geologists in many fields and not only the taxonomist or museum curator.

One such is a query which lists the locations and ranges of a lithostratigraphic unit and its
thickness in metres in the overall chronostratigraphic range R1 to R2.

?-thick(obtusum,bucklandi,blue Jias).

yields

saltford_midland_railway_cutting 3 semicostatum bucklandi
kelston_park_quarry 3-93 semicostatum bucklandi

This information is essential for the construction of palaeogeographic maps, geologic cross-sections
and other visual representations of stratigraphy, and indeed the answers could be transferred
directly as data to an appropriate graphics program.

THE DATABASE: DATA STRUCTURE FOR MUSEUM SPECIMENS

The data recorded about each specimen in a catalogue should be a list of facts including horizon,
locality and collector and also the subsequent history of the material. A specimen may be moved
from one collector or collection to another and it may be identified and cited or figured by several
different workers at different times. Each of these events should be noted because they are often
important in identifying type material.

In practice it is rare for all these facts to be known about a specimen, particularly if it was
collected many years ago and has been moved from collection to collection, but it is often possible
to make inferences about its origin and history. For example, if the handwriting on a specimen is
recognised as that of a well known collector, it may be assumed that it was collected by that person.
These inferences should be clearly noted in the catalogue.

In the prototype described here we make no attempt to create structures to accommodate all the
complexities which should be recorded in a museum catalogue. Instead we have selected a few facts,
which we assume to be correct, about each specimen so that we can indicate how the museum data
can be used in conjunction with the knowledge base. Initially, these facts formed the columns in a

620

PALAEONTOLOGY, VOLUME 33

table of a relational database. This information has been transferred into a file and edited into a
series of Prolog facts of the form

ammonite(V 1 ,V2,V3,V4,V 5,V6,V7,V8,V9,V 1 0,V 1 1 ,V 1 2,V 13)

where VI and V2 are identifying numbers, V3,V4 are the generic and specific name, V5 = number
of specimens, V6 = matrix, V7 = biological class or order, V8 = size in cm, V9 = stratigraphic
system, V10 = horizon, VI I = locality, VI 2 = geographic region, and VI 3 = quality of specimen.

The relational database was designed and the ammonite data recorded before the current research
was started. No attempt has been made so far to test its Prolog representation for internal
consistency, or to ensure that all the terms in the ammonite file are present in the knowledge base.
However, with very little editing, it has been possible to search for all specimens belonging to a
particular taxon. In this query, the knowledge base is searched and all the members of the same
taxon and their synonyms are listed, followed by all such specimens recorded in the database.

In the example, the genus Schlotheimia belongs to the family Schlotheimiidae and all the genera
in this family and their synonyms are listed. Thus

?-taxon(schlotheimia).

yields

schlotheimia
synonyms are:
anguliferites
scamnoceras

angulaticeras
synonyms are:
argoceras
boucaulticeras
etc.

{the museum specimens belonging to the family Schlotheimiidae are:}

28 5001 schlotheimia
9 5001 schlotheimia
etc.

The answer to such a query is invaluable to the taxonomist. Further queries should be constructed
which use the horizon and location data from the ammonite structure in conjunction with the
knowledge base. The palaeobiologist could easily use these results to test hypotheses which would
be very onerous to test by conventional means.

It is pertinent to ask whether the information contained in a computerized museum catalogue is
of a quality suitable for research purposes. When using any secondhand material, whether from
a museum collection or from published information, the researcher always has to balance the
accuracy of the information about a specimen against the availability of alternative specimens and
the time and expense of collecting and investigating fresh material.

In a museum catalogue the identification may be wrong, the identifier unknown and the locality
and the horizon may have been inferred because of the lithology or some other characteristic feature
of the specimen. The computer catalogue should contain all available evidence to enable a
researcher to assess the quality of the information about a specimen. Also, if necessary, the specimen
itself can be examined.

In the case of published information, it is not always possible to find the source material,
particularly referenced material, and in such circumstances future researchers depend entirely upon
the professional skill and integrity of the author(s).

ROGERS ET AL.\ DEDUCTIVE ENQUIRY SYSTEM

621

SUGGESTIONS FOR FURTHER DEVELOPMENTS

The deductive database described in this paper could form the basis of a powerful information
retrieval/on-line query system, but it will need to be extended and developed in several main areas.
The knowledge base should be extended to include, for example, all stratigraphic systems and
biological phyla. Facilities should be provided so that stratigraphic classifications appropriate to
other parts of the world can be incorporated. The structures described in this paper are sufficiently
general to achieve this, but the linkage between data from different provinces must be designed with
care so that new evidence on correlation can easily be included.

The source of lithostratigraphic information should be added to the description of sections and
research, and collection details should be included in the database of museum specimens.

A data entry facility should be provided which checks all new data for consistency and accuracy,
as far as this is possible. A ‘user-friendly’ interface is essential for the casual user and this can be
achieved either by a menu-driven query system or a limited natural language processor. The full
power of the system described here can only be realized if it is integrated with geologically orientated
program packages.

SUMMARY AND CONCLUSIONS

The primary aim of this paper is to show that the valuable information contained in geological
museum catalogues can be made readily available for a wide range of investigations. This can be
achieved by constructing a knowledge base of geological information and using it in conjunction
with a conventional database of museum specimens. The knowledge base itself is an important
reference source for many geological investigations. The information in the knowledge base is held
in a series of simple Prolog structures which are linked at only a few essential points. Consequently,
any modifications which may arise through the availability of new evidence or new interpretations
can be made easily, and additional facts can be added without corrupting the existing information.

The geological information included in the conventional museum database can be restricted to
observed facts, while the presence of checking routines, although not included in this prototype,
would ensure that the internal consistency and accuracy of the database is maintained. This is
particularly important when data are entered by relatively unskilled personnel and allows the
specialist to concentrate on the complex problems.

The deductive capabilities which are available in the knowledge base allow a more general class
of queries to be made than is possible with a conventional database management system.

The results presented in this paper are based on the use of local databases, but it is already clear
from the experience currently being gained with remote access to library catalogues that the
techniques described here can be extended to include remote databases, both at national and
international sites.

REFERENCES

brough, D. r. and Alexander, I. F. 1986. The fossil expert system. Expert Systems , 3, 76-83.

brunton, c. h. c., besterman, t. p. and cooper, j. a. 1985. Guidelines for the curation of geological materials.
Miscellaneous Papers of the Geological Society , 17, 1-174.

colmerauer, a., kanoui, h., pasero, r. and roussel, p. 1973. Un systeme de communication homme-machine
en frangais. Rapport de recherche sur le contract CRI , 72-81. Groupe d’lntelligence Artificielle, Universite
d’Aix-Marseille II, France.

cope, j. c. w., getty, t. a., HOWARTH, m. k., morton, n. and torrens, h. s. 1980. A correlation of Jurassic
rocks in the British Isles. Part 1. Introduction and Lower Jurassic. Geological Society of London. Special
Report , 14, 1-73.

cope, j. c. w., duff, k. l., parsons, c. f., torrens, h. s., wimbledon, w. a. and WRIGHT, J. R. 1980. A
correlation of Jurassic rocks in the British Isles. Part 2. Middle and Upper Jurassic. Geological Society of
London. Special Report , 15, 1 109.

cutbill, j. l.. hallen, a. j. and lewis, G. d. 1971. A format for the machine exchange of museum data.
31 1-320. In cutbill, j. l. (ed.). Data processing in biology and geology , Academic Press, London, 346 pp.

622

PALAEONTOLOGY, VOLUME 33

donovan, d. t., callomon, j. h. and howarth, m. k. 1981. Classification of the Jurassic Ammonitina.

101- 155. In house, m. r. and senior, j. r. (eds. ). The Ammonoidea , Academic Press, London, 593 pp.
gallaire, h., minker, j. and nicolas, j. m. 1984. Logic and databases: A deductive approach. Computing
Surveys, 16, 153-185.

ghosh, s., lin, c. c. and sellis, t. 1988. Implementation of a Prolog-INGRES interface. Sigmod Record , 17,

gray, p. m. d. and lucas, R. j. 1988. Prolog and databases. Ellis Horwood, Chichester, 358pp.
porter, M. F. 1982. GOS: A package for making catalogues. Information Technology: Research and
Development, 1, 113-129.

1983. Information Retrieval at the Sedgwick Museum. Information Technology : Research and
Development , 2, 169-186.

price, d. 1984. Computer-based storage and retrieval of palaeontological data at the Sedgwick Museum,
Cambridge, England. Palaeontology , 27, 393-405.
rawlings, c. J. 1988. Designing databases for molecular biology. Nature, 334, 477.

Roberts, d. a. and light, R. B. 1980. Progress in documentation. Museum documentation. Journal of
Documentation, 36, 42-84.

sterling, l. and Shapiro, e. 1986. The art of Prolog: Advanced programming techniques. MIT Press,
Cambridge, Massachusetts, 427 pp.

williams, r. 1988. The Goblin quadtree. Computer Journal, 31, 358-363.

77-88.

M. JENNIFER ROGERS
DESMOND T. DONOVAN

Department of Geology
University of Bristol
Bristol BS8 1RJ

Typescript received 23 November 1988
Revised typescript received 7 November 1989

MICHAEL H. ROGERS

Department of Computer Science
University of Bristol

Bristol BS8 1TR

THE AFFINITIES OF EARFY ONCOCERID
NAUTIFOIDS FROM THE LOWER ORDOVICIAN OF
SPITSBERGEN AND SWEDEN

bv DAVID H. EVANS Cltld ANDREW H. KING

Abstract. Phthanoncoceras oelandense gen. et sp. nov. from the Kundan (Llanvirn) of Sweden and
Valhalloceras floweri gen. et sp. nov. from the Cassinian (Arenig) of Ny Friesland, Spitsbergen are described
and a new family, the Phthanoncoceratidae (Oncocerida) is proposed. The relationships of the Phthanonco-
ceratidae to the Oncocerida and Ellesmerocerida are discussed.

Collections of Ordovician cephalopods from the lower Kundan of Oland, Sweden (held in the
Swedish Museum of Natural History, Stockholm) and from the Cassinian of Ny Friesland,
Spitsbergen (made available by Dr R. A. Fortey) contain two genera showing features intermediate
between the Ellesmerocerida and the Oncocerida. These genera, Phthanoncoceras gen. nov. and
Valhalloceras gen. nov., possess thick, layered connecting rings and primary siphonal diaphragms,
features considered to be characteristic of the predominantly endogastric Ellesmerocerida. However,
both these new genera are exogastric, have relatively narrow, marginal to submarginal siphuncles
and conch cross-sections which strongly support an affinity with the Oncocerida. Since it is clear
that these genera exhibit mixed characters, the presence of an exogastric curvature is here deemed
sufficient to assign Phthanoncoceras and Valhalloceras to the Oncocerida.

The Graciloceratidae were previously thought to represent the most primitive oncocerids. They
have thin connecting rings and lack diaphragms. Phthanoncoceras and Valhalloceras clearly are not
graciloceratids, and the family Phthanoncoceratidae fam. nov. is proposed for these two genera.
This new family includes the earliest known oncocerids, and exhibits the retention of primitive
features more typical of the Ellesmerocerida.

SYSTEMATIC PALAEONTOLOGY

Order oncocerida Flower in Flower and Kummel, 1950
Family phthanoncoceratidae nov.

Diagnosis. Exogastric cyrtocones with depressed to compressed section and a narrow marginal to
sub-marginal tubular siphuncle with thickened and layered connecting rings; septal necks
orthochoanitic. Siphonal diaphragms present.

Type genus. Phthanoncoceras gen. nov.

Remarks. The Phthanoncoceratidae possess primary siphonal diaphragms and thickened,
differentiated connecting rings. They therefore should be assigned to the Ellesmerocerida,
themselves diagnosed partly on the possession of such structures (Furnish and Glenister 1964, p.
K140). However, in their exogastric curvature, combined with the narrow marginal to sub-marginal
siphuncle, the two new genera Phthanoncoceras and Valhalloceras differ from all other
ellesmerocerids. Because of these differences, these two genera are placed in the Oncocerida rather

[Palaeontology, Vol. 33, Part 3, 1990, pp. 623-630, 1 pi. |

© The Palaeontological Association

624 PALAEONTOLOGY, VOLUME 33

than the Ellesmerocerida. The origin of the Oncocerida is unclear and two possibilities require
consideration.

Thickened and differentiated connecting rings are a persistent feature throughout most of the
Ellesmerocerida. They are also present in the Proterocameroceratidae of the Endocerida and many
tarphycerids, including the Bassleroceratidae. Primary siphonal diaphragms are almost entirely
associated with late Cambrian cephalopods as well as the Ellesmeroceratidae and Cyclostomi-
ceratidae of the Ellesmerocerida. The possession of thickened connecting rings combined with the
presence of primary siphonal diaphragms may indicate that the Phthanoncoceratidae are related to
the Ellesmeroceratidae or Cyclostomiceratidae.

Flower ( 1964, 1968, 1976) and Sweet et al. (1964) considered that the Oncocerida originated from
the Bassleroceratidae, but the systematic position of the Bassleroceratidae has been disputed.
Flower (in Flower and Teichert 1957) and Flower (1964) placed the Bassleroceratidae in the
Tarphycerida, but Furnish and Glenister (1964) and Balashov (1962) assigned the family to the
Ellesmerocerida. Later Flower (1968) argued that placing the Bassleroceratidae in the Ellesmero-
cerida broadened that order beyond definition. The present authors agree with this latter view, as
the Ellesmerocerida already form a large and diverse order.

The Phthanoncoceratidae are similar to the Bassleroceratidae in possessing thick differentiated
connecting rings and an exogastrically curved conch. However, to date, it is unknown whether or
not the Bassleroceratidae possess primary siphonal diaphragms. The Phthanoncoceratidae differ
from the Bassleroceratidae in their narrower siphuncles and more breviconic, rapidly expanding
conchs. In the latter feature the Phthanoncoceratidae are thought to have more in common with the
Ellesmeroceratidae and Cyclostomiceratidae.

The Phthanoncoceratidae extend the stratigraphical range of the Oncocerida and also modify the
way in which the early evolution of the Oncocerida is viewed (see text-fig. 1). Since there is
stratigraphical overlap between the Phthanoncoceratidae, Graciloceratidae and Oncoceratidae, it is
not clear when the last originated and this family may have a longer history than currently
recognized. The existence of thick-ringed oncocerids in strata as young as the early Kundan may
also bring into question the relationship of the Graciloceratidae to later oncocerid families. While
it is simpler to derive the Oncoceratidae from the Graciloceratidae (e.g. Flower 1976), the same may
not be true for the Valcouroceratidae which possess thick, actinosiphonate connecting rings. If the
Valcouroceratidae were derived from the Graciloceratidae the connecting rings would have been
secondarily thickened, but, if descended from the Phthanoncoceratidae, a thickened connecting ring
was already present. Thus it is possible that the Phthanoncoceratidae may, at different times have
given rise to more than one oncocerid lineage.

In conclusion, the initial development of the Oncocerida may be the reverse of that which is
currently accepted (i.e. derivation from the Bassleroceratidae, e.g. Flower 1976, text-fig. 9) and
instead they may have originated from the Ellesmerocerida (quite possibly the Ellesmeroceratidae)
through the appearance of exogastric curvature. The Bassleroceratidae and the Tarphycerida
formed a lineage derived early in the history of the Phthanoncoceratidae. The Phthanoncoceratidae
may themselves have given rise to many of the older oncocerid families through several independent
branches (text-fig. 1 ).

Genus phthanoncoceras gen. nov.

Type species. Phthanoncoceras oelandense sp. nov.

Diagnosis. Moderately expanding exogastric cyrtocones; body-chamber slightly inflated adapically
with shallow constriction near aperture. Conch section laterally compressed with broadly rounded
venter and obtusely rounded dorsum. Siphuncle narrow, tubular and sub-ventral. Septal necks
orthochoanitic, connecting rings thick and layered; siphonal diaphragms present apically. External
shell smooth, growth lines forming a weak ventral sinus.

Derivation of name. From Greek, phthanein = to come before.

EVANS AND KING: ORDOVICIAN NAUTILOIDS 625

text-fig. 1. Stratigraphical and phylogenetic relationships of the Phthanoncoceratidae. Chronostratigraphical
data based upon Fortey and Bruton (1973), and Jaanusson (1982). Stratigraphical distributions: a, possible
range of Phthanoncoceras oelandense; b, range of the Graciloceratidae in Oland; c, occurrence of
Graciloceratidae (Leonardoceras and Ikesoceras ) in Nevada (Flower 1968); d, occurrence of Leonardoceras in
Spitsbergen (pers. comm. Dr R. H. Flower to Dr R. A. Fortey 1975); e, occurrence of Valhalloceras floweri in
Spitsbergen. Synapomorphies : 1, loss of siphonal diaphragms leading to the Bassleroceratidae; 2, loss of
siphonal diaphragms and the appearance of thin connecting rings leading to the Graciloceratidae; 3, loss of
siphonal diaphragms; appearance of thin connecting rings and development of suborthochoanitic and
cyrtochoanitic septal necks (leading to the Oncoceratidae?); 4, appearance of suborthochoanitic and
cyrtochoanitic septal necks (leading to the Oncoceratidae?); 5, development of cyrtochoanitic septal necks
combined with an actinosiphonate connecting ring leading to the Valcouroceratidae.

Remarks. Phthanoncoceras is a monotypic genus, presently recorded only from Sweden. Oncoceras ,
Beloitoceras , Miamoceras and Neumatoceras (all Oncoceratidae) resemble Phthanoncoceras in
exhibiting similar conch forms and possessing a constriction on the body-chamber. They differ in
having suborthochoanitic to cyrtochoanitic septal necks combined with thin connecting rings.
Valhalloceras differs from Phthanoncoceras only with respect to conch section, which in the former
is depressed and subtriangular.

Occurrence. Ordovician, Hunderumian (uppermost Arenig to lowermost Llanvirn) of Oland, Sweden.

Phthanoncoceras oelandense sp. nov.
Plate 1, figs. \-^. ; text-fig. 2

626

PALAEONTOLOGY, VOLUME 33

text-fig. 2. Phthanoncoceras oelandense gen. et sp. nov. Camera lucida drawing of
adapical portion of preserved phragmocone of Mol 58453 showing details of
siphuncle and development of diaphragms, a, septal neck; b, spicular layer
of connecting ring; c, nacreous layer of connecting ring; d, diaphragm; e,
phragmocone wall, x 25.

Diagnosis. As for genus.

Holotype. Swedish Museum of Natural History, Stockholm, Mol58453, collected by G. Holm in 1894 from the
'glaukomt. gra vaginatumkalk ’ (lower Kundan) at Halludden. The species is only known from the holotype.

Derivation of name. From Latin, oelandense = from Oland.

Description. The type consists of an exogastric, weakly cyrtocomc conch with a moderate expansion rate of 20°.
The total length preserved is 53 mm, representing the body-chamber and a portion of phragmocone. Adorally
the conch section is compressed with a broadly rounded venter and narrowly rounded dorsum; apically the
section is less compressed. External shell ornament consists only of very fine growth lines which form a faint
ventral sinus.

The body-chamber has a length of 26 mm with adoral and dorsoventral diameters of 16-9 mm and 14-6 mm
respectively. A shallow but broad constriction is present on the internal mould 8 mm from the aperture. The
body-chamber is slightly inflated adapically.

EXPLANATION OF PLATE 1

Figs. 1^1, Phthanoncoceras oelandense gen. et sp. nov. Hunderumian Substage, lower Kundan at Halludden,
northern Oland, Sweden. 1-3, dorsal, left lateral, and ventral aspects of Mol58453, whitened, x 15. 4,
sagittal section of Mol58453 showing ventral siphuncle, x3-75.

Figs. 5 11, Valhalloceras floweri gen. et sp. nov. Olenidsletta Member of the Valhallfonna Formation. Adjacent
to Hinlopenstretcd, North Ny Friesland, Spitsbergen. 5-7, dorsal, right-lateral and ventral aspects of PMO
NF2276 and 2285, whitened, x 1 4. 8, 1 thick ’ section of NF2276, right-hand side viewed under reflected light
showing submarginal ventral siphuncle, x 3-5. 9, apical surface of NF2285 showing depressed subtriangular
section and submarginal septal foramen, venter up, x4-4. 10, details of siphuncle of NF2276 showing short
septal necks and thickened connecting rings represented by dark bands, x 1 1-7. I I . left-hand side of NF2276,
apical camerae with siphonal diaphragms crowded into the siphuncle, x 11-7.

PLATE 1

EVANS and KING, Phthanoncoceras, Valhalloceras

628

PALAEONTOLOGY, VOLUME 33

The phragmocone consists of fifteen camerae and has an apicad dorsoventral diameter of 6-9 mm. Sutures
exhibit weak dorsal and ventral lobes with corresponding lateral saddles. Cameral depth is 20% of the
dorsoventral diameter adapically and 12% adorally where septa are approximated. The siphuncle is narrow
(about 1 5 % of the phragmocone diameter) and submarginal (0-5 mm to 1 0 mm from the venter). Septal necks
are orthochoanitic and 0-3-04 mm long. Connecting rings are tubular and layered. Nine closely spaced
diaphragms are present in the siphuncle adjacent to the most adapical camerae (see text-fig. 2).

Occurrence. Hunderumian Substage, lower Kundan (uppermost Arenig to lowermost Llanvirn) at Halludden,
northern Oland, Sweden.

Genus valhalloceras gen. nov.

Type species. Valhalloceras floweri sp. nov.

Diagnosis. Small exogastric conch with suhtriangular section, venter obtusely rounded, lateral sides
more acutely curved and dorsum broadly rounded. Siphuncle narrow, subventral, orthochoanitic;
connecting rings thickened and differentiated ; siphonal diaphragms present. Sutures consist of weak
dorsal and ventral lobes with lateral saddles. Shell smooth, weak sinus over venter.

Derivation of name. After the Valhallfonna Formation, from which the type species was collected.

Remarks and occurrence. Valhalloceras is at present known only from Spitsbergen in facies
belonging to the Nileid Association of Fortey (1975). Valhalloceras shows many similarities to
Ikesoceras Flower in terms of gross conch morphology; however, it is unknown whether the latter
possesses siphonal diaphragms. The holotype of Ikesoceras (Flower 1968, p. 25, pi. 25, figs. 1-5)
consists of a body-chamber and the four most adoral camerae, too far adoral to exhibit siphonal
diaphragms. The paratype (Flower 1968, pi. 26, figs. 6 and 7) is partly crushed and silicified and
represents an adapical portion of phragmocone which, on the basis of Valhalloceras floweri, might
contain siphonal diaphragms. None have so far been demonstrated. Valhalloceras came from a
horizon near the top of the Olenidsletta Member of the Valhallfonna Formation, dated as Cassinian
J Zone (Cooper and Fortey 1982, fig. 2) and is older than Ikesoceras (early Whiterockian zone L
sensu Cooper and Fortey 1982; Fortey and Owens 1987). Ikesoceras appears to occur in a similar
facies to Valhalloceras, considered to represent the outer lip of the shelf of Ross (1975) and
equivalent to the upper slope of Shaw and Fortey (1977), or the Nileid Association of Fortey (1975).
Until Ikesoceras is better known its systematic position will remain uncertain. Flowever, should
Ikesoceras prove to possess siphonal diaphragms good reason will exist for regarding it as a senior
synonym of Valhalloceras.

Valhalloceras floweri sp. nov.

Plate 1, figs. 5-1 1 ; text-fig. 3.

Diagnosis. As for genus.

Holotype. Palaeontological Museum. Oslo NF2276, 2285. These represent complementary fragments of the
same specimen. NF2276 was prepared as a thin section, but has been left thick as the siphuncle was displaced
in the most apical portion. It was this portion which contained siphonal diaphragms, further preparation of
the section would have destroyed these structures. The species is only known from the holotype.

Derivation of name. In honour of the late Dr R. H. Flower.

Description. PMO NF2276 and 2285 represent a portion of phragmocone 44 mm long, slightly crushed
adorally. Adapically the cross-section of the conch is almost elliptical with lateral and dorsoventral diameters
of 6-4 mm and 5-0 mm respectively. At a position 17-5 mm adorally the cross-section is subtriangular, with the
maximum lateral diameter (lit) mm) 40% of the distance from venter to dorsum (dorso-ventral diameter

EVANS AND KING: ORDOVICIAN NAUTILOIDS

629

text-fig. 3. Valhalloceras floweri gen. et sp. nov. Camera lucida drawing of adapical
portion of phragmocone of NF2276 showing details of siphuncle and development of
diaphragms, a, septal neck; b, connecting ring; c, diaphragm. x9.

91 mm). A further 15 mm adorally the lateral diameter is estimated to be 16 0 mm. Cameral depth is 17% of
the dorsoventral diameter adapically and 13% adorally. The suture comprises weak dorsal and ventral lobes
and lateral saddles. Where the conch wall is preserved, it is smooth, consisting of very fine growth lines which
trace out a shallow ventral sinus.

The siphuncle is submarginal, its centre being positioned 1 5 % of the distance from the venter to the dorsum ;
it is 9% of the dorso-ventral diameter. Septal necks are orthochoanitic and about 0-3 mm long. Connecting
rings are tubular and consist of two layers; a very thin light coloured dense layer lining the siphonal cavity and
a thicker, darker, more diffuse layer forming the cameral side of the connecting ring. In the four most adapical
camerae eleven closely spaced diaphragms are present in the siphuncle.

Occurrence. From the outcrop of the Olenidsletta Member between Lundehuken and Papegoyneset, adjacent
to Hinlopenstreted, North Ny Friesland, Spitsbergen. The stratigraphical horizon is c. 100 m above the base
of the Olenidsletta Member of the Vallhalfonna Formation at the top of V2 in the terminology of Fortey (1980)
and Arenig Series Ca 1 in terms of graptolites ( = North American D. bifidus Zone, see Cooper and Fortey
1982, fig. 2).

Acknowledgements. The authors thank Professor C. H. Holland and Dr J. C. W. Cope for reading early
versions of the manuscript. Thanks go to Dr R. A. Fortey for making the Spitsbergen material available and
to Professor V. Jaanusson and Dr H. Mutvei for providing the material from Oland. This work was carried
out partly under NERC grants GT4/83/GS/135 and GT4/86/GS/ 1 37 which are gratefully acknowledged.

REFERENCES

balashov, z. G. 1962. Podtip Ellesmeroceratida. 73-77. In orlov, yu. a. (ed.). Osnovy paleontologii :
Nautiloidei. Endoseratoidei. Aktinoseratoidei, Baktritoidei, Ammonoidei. Nedra Press, Moscow, 438 pp. [In
Russian]

cooper, r. a. and fortey, r. a. 1982. The Ordovician graptolites of Spitsbergen. Bulletin of the British Museum
( Natural History) (Geology), 36, 157-302, pis. 1-6.

flower, R. H. 1964. The Nautiloid Order Ellesmerocerida (Cephalopoda). New Mexico Bureau of Mines and
Mineral Resources , Memoir, 12, 243 pp.

— 1968. Some additional Whiterock cephalopods. New Mexico Bureau of Mines and Mineral Resources ,
Memoir, 19, 17-55.

— 1976. Ordovician cephalopod faunas and their role in correlation. 523-552. In bassett, m. g. (ed.). The
Ordovician System. University of Wales Press and National Museum of Wales, Cardiff, 696 pp.

— and kummel, b. 1950. A classification of the Nautiloidea, Journal of Paleontology , 24, 604-616.

— and teichert, c. 1957. The cephalopod order Discosorida. University of Kansas, Paleontological
Contributions. Mollusca, 6, 1-114.

630

PALAEONTOLOGY, VOLUME 33

furnish, w. m. and glenister, b. f. 1964. Nautiloidea-Ellesmerocerida. K129-K159. In moore, r. c. (ed.).
Treatise on invertebrate paleontology. Part K. Mollusca 3. Geological Society of America and University of
Kansas Press, Boulder, Colorado and Lawrence, Kansas, 519 pp.
fortey, r. a. 1975. Early Ordovician trilobite communities. Fossils and Strata , 4, 339-360.

1980. Trilobites of Spitsbergen III. Remaining trilobites of the Vallhalfonna Formation. Skrifter om norsk
Polarinstitutt , 171, 113 pp., 25 pis.

— and bruton, d. l. 1973. Cambrian-Ordovician rocks adjacent to Hinlopenstretet, north Ny Friesland,
Spitsbergen. Bulletin of the Geological Society of America , 84, 2227-2242.
jaanusson, v. 1982. Introduction to the Ordovician of Sweden. Palaeontological Contributions of the University
of Oslo. 279, 1-9.

ross, r. j. 1975. Early Palaeozoic trilobites, sedimentary facies, lithospheric plates, and ocean currents. Fossils
and Strata , 4, 307-329.

shaw, f. c. and fortey, r. a. 1977. Middle Ordovician facies and trilobite faunas in North America. Geological
Magazine , 114, 409 496.

sweet, w. c., teichert, c. and kummel, b. 1964. Phylogeny and evolution. K106-K1 14. In moore, r. c. (ed.).
Treatise on invertebrate paleontology. Part K. Mollusca 3. Geological Society of America and University of
Kansas Press, Boulder, Colorado and Lawrence, Kansas, 519 pp.

DAVID H. EVANS

Department of Earth Sciences
University College
Swansea SA2 8PP
Present address:

7 Gentian Court
Braiswick

Colchester C04 5UF

ANDREW H. KING

Department of Earth Sciences
University College
Swansea SA2 8PP
Present address:

Typescript received 2 May 1989
Revised typescript received 14 July 1989

Cambridge Arctic Shelf Program
West Building
Huntingdon Road
Cambridge CB3 0DJ

THE SOLUTE D EN D ROC YSTO 1 D ES SCOTICUS
FROM THE UPPER ORDOVICIAN OF SCOTLAND
AND THE ANCESTRY OF CHORDATES AND

ECHINODERMS

by R. P. S. JEFFERIES

Abstract. A study of the solute Dendrocystoides scoticus (Bather), from the Upper Ordovician Ashgill Series,
near Girvan, Scotland produces much new anatomical information. The positions of head, tail, the tube feet
in the feeding arm, hydropore, gonopore, gonad, pharynx, stomach, anus, brain, left trigeminal ganglion, gill
slit (at posterior left in the head), notochord, and tail muscles are deduced. In their basic structure, solutes
resemble cornutes, especially the primitive cornute Ceratocystis , but differ in retaining a water vascular system,
and in other ways. They can be compared in detail with the inferred latest common ancestor of chordates and
echinoderms, visualized as resembling the hemichordate Cephalodiscus lying on its right side. The tail of solutes
is almost certainly homologous with that of cornutes, and therefore with the tail of chordates in general. It is
also probably homologous with the locomotory stalk of Cephalodiscus. If so, the tail ( = stalk) was lost early
in the evolution of the Echinodermata, while the water vascular system was lost in the early evolution of the
Chordata. The Soluta, if characterized as retaining both a tail and a water vascular system, would therefore
have included the latest common ancestor of chordates and echinoderms.

Dendrocystoides scoticus probably belonged to the stem group of the Chordata, on the basis of several
advanced features shared with Cornuta including, especially, the posterior left position of the gill slit. The
‘carpoids’ placed in the group Cincta, on the other hand, probably belonged to the echinoderm stem group
but were primitive for echinoderms in possessing a gill slit, located in an anterior, presumably primitive,
position like the left gill slit of Cephalodiscus. If the tail of solutes is homologous with the tail of cornutes, then
the latter organ, contrary to the ‘aulacophore’ interpretation of Ubaghs, cannot be a feeding arm.

The aim of this paper is to determine the phylogenetic, and therefore systematic, position of the
Soluta. These are a group of strange fossils which are characterized by a feeding arm at one end of
the animal and a different appendage (in my view a locomotory tail) at the other end. They have
a calcite skeleton of echinoderm type but lack any trace of radial symmetry and for this reason are
traditionally placed in the ‘carpoid echinoderms’. They are found in marine rocks of Lower
Cambrian to Devonian age and comprise 15 described species placed in 13 genera.

I approach the problem of the systematic position of the solutes by a detailed anatomical study
of one species, the late Ordovician Dendrocystoides scoticus (Bather), chosen because it is abundant
in accessible collections. It is the only species in its genus but is very similar to Dendrocystites
Barrande 1887 from the Ordovician of Czechoslovakia, differing mainly in having a better
developed anti-brachial process. My studies convince me that the solutes are built on the same basic
plan as the stem-group chordates known as Cornuta. I also believe, however, that the latest common
ancestor of the echinoderms and chordates, if it were ever found, would be regarded as a solute. This
means that the Soluta, characterized by the possession of a locomotory tail and a feeding arm,
straddle the phylogenetic separation between chordates and echinoderms, which gives them the
status of an invalid stem group of the Dexiothetica (Tmechte Stammgruppe’ to use the expression
of Hennig 1969). Further work will be needed to decide which of them belong to the true stem group
of the Dexiothetica, which of them can be assigned to the stem group of the chordates or of the
echinoderms, and which of them, being close in systematic position to the latest common ancestor
of chordates and echinoderms, should remain in the nodal group of the Dexiothetica.

IPalaeontology, Vol. 33, Part 3, 1990, pp. 631-679, 7 pls.|

© The Palaeontological Association

632

PALAEONTOLOGY, VOLUME 33

The conclusion, presented in this paper, that the solute tail is homologous with the cornute tail,
is important to the general morphological interpretation of cornutes, since the solutes possess,
in addition to this tail, an appendage clearly homologous with an echinoderm feeding arm. In such
case, as Paul (1977, p. 126) pointed out, the cornute tail, contrary to the view of Ubaghs (1968)
and Parsley (1988), cannot itself be a feeding arm. In a different direction, I shall argue that the tail
of solutes is homologous with the locomotory and prehensile stalk of hemichordates such as
Cephalodiscus. If so, and if the hemichordates include, or else constitute, the extant sister group of
the Dexiothetica, then the latest common ancestor of echinoderms and chordates would have had
such a tail. And the echinoderms, as an autapomorphy of the group Echinodermata, would have lost
this organ (contrary to the view expressed in Jefferies 1988, p. 11). It follows that the mere
possession of a stalk or tail does not characterize an animal as a chordate.

Which way up the solutes lay in life is a matter of dispute. Bather (1913, p. 505) and Caster (1968)
oriented the animal with the arm on the left, if the tail is towards the observer, and accordingly
spoke of an obverse face (presumably upper) and a reverse face (presumably lower). Kolata,
Strimple and Levorson (1977), however, considered it more likely, on functional grounds, that the
solutes lay the other way up and used ‘aboral’ for the presumed upper and ‘oral’ for the presumed
lower surface. I agree with Kolata's orientation of the solutes, partly on functional grounds and
partly because it agrees with the likely orientation, accepted by all workers, of the cornutes. I use
‘dorsal' for the presumed upper and ‘ventral’ for the presumed lower surface. These terms, for me,
are intended to imply homology with ‘dorsal' and ‘ventral’ in chordates, but not with ‘dorsal’ and
‘ventral’ in echinoderms where the terms mean the exact opposite to what they do in chordates
(Jefferies, Lewis and Donovan 1 987, p. 474). For me, therefore, ‘ left ’ and ‘ right ' in solutes mean the
opposite of what they did for Caster (1968).

I studied Dendrocystoides scoticus by reconstructing it on a drawing board, as with previous
studies done in the British Museum (Natural History) (Jefferies 1968, 1969, 1973; Jefferies and
Lewis 1978; Jefferies and Prokop 1972; Jefferies et al. 1987; Craske and Jefferies 1989; Cripps
1988, 1989 a, b, in press). Six different projections (dorsal, ventral, right, left, posterior, anterior)
were drawn simultaneously, firstly of the outside and then of the superficial internal anatomy of the
animal. The reconstruction was difficult since the fossils, although numerous, are all incomplete.
Moreover D. scoticus changed shape as it grew and the plating was largely irregular.

The reconstruction of the external shape was largely based on latex casts, but silicone rubber casts
were sometimes taken by pressing part and counterpart together against silicone rubber (Wacker
RTV-M531 with catalyst T461). The superficial internal anatomy was reconstructed on the basis of
internal moulds which, being negatives of the skeleton, can be regarded as positives of the soft parts.
The in vivo shape of certain parts of the skeleton was reconstructed using large cardboard replicas
of the plates and fixing them together with masking tape.

SYSTEMATIC PALAEONTOLOGY

Superphylum deuterostomia Grobben, 1908
Subsuperphylum dexiothetica Jefferies, 1979
?Phylum chordata Bateson, 1886

(If a chordate, then stem group of the Chordata)

soluta Jaekel, 1900 (invalid stem group of the Dexiothetica)

Genus Dendrocystoides Jaekel 1918
Species Dendrocystoides scoticus (Bather 1913)

Remarks. Dendrocystoides scoticus is probably a stem-group chordate, as argued below, though at
present this conclusion is tentative. It is certain, on the other hand, that D. scoticus is a solute, if
the Soluta are characterized as a group retaining both a feeding arm and a tail. On the other hand,

JEFFERIES: DENDROCYSTOIDES AND CHORDATE ANCESTRY

633

the position of the Soluta, straddled across the phylogenetic separation of the Echinodermata and
Chordata, makes them an invalid stem group of the Dexiothetica which further research will
probably break down. These uncertainties explain the disorderliness of the above systematic
‘address’. Ax (1984, 1987) has argued that Linnaean categorial ranks above the species level are
arbitrary and ought to be discontinued, except for the genus which is necessary for nomenclatorial
reasons. I agree with Ax on this (Craske and Jefferies 1989) and the ranks assigned above are purely
traditional (superphylum, phylum) or got by interpolation (subsuperphylum). The group Soluta, set
up by Jaekel (1900) as a suborder, is exactly coextensive with the Homoistelea, set up by Gill and
Caster (1960) as a subclass and raised by Caster (1968) to the rank of class. Since Linnaean
categorial ranks are arbitrary, not real, the difference between Homoiostelea and Soluta is purely
conventional and the term ‘Homoiostelea’, being younger than ‘Soluta’, ought to be abandoned.

Dendrocystoides scoticus (Bather 1913)

Plates 1-7; text-figures 2-11, 21

1913 Dendrocystis scotica Bather; Bather, p. 391.

1913 Dendrocystis scotica ; Woodward, p. 421, 423
1918 Dendrocystoides scoticus Bather sp.; Jaekel, p. 123.

1920 Dendrocystis scotica \ Hawkins, p. 134.

1925 Dendrocystis scotica ; Bather, p. 5.

1928 Dendrocystis scotica ; Bather, p. 5.

1929 Dendrocystis scotica ; Bather, p. 34.

1932 Dendrocystis scotica ; Hennig, p. 170.

1934 Dendrocystoides scotica (Bather); Dehm, p. 39.

1937 Dendrocystis scotica; Woods, p. 175.

1938 Dendrocystis scotica ; Bassler, p. 85, as type species of Dendrocystoides Jaekel 1918.

1941 Dendrocystis scotica Bather; Chauvel, p. 241.

1943 Dendrocystoides scoticus (Bather); Bassler and Moody, p. 150.

1947 Dendrocystis scotica Bather; Begg, p. 30.

1948 Dendrocystis scotica., Cuenot, p. 14.

1960 Dendrocystis scotica ; Spinar, p. 668.

1960 Dendrocystis scotica ; Termier and Termier, p. 87.

1965 Dendrocystoides scotica ; Parsley and Caster, p. 162.

1968 Dendrocystoides scoticus (Bather); Caster, p. S587.

1982 Dendrocystoides scoticus (Bather); Melendez, p. 688.

Comments on synonymy. Dendrocystoides scoticus is the only known species of its genus and is also the type
species by original designation (Jaekel 1918, p. 123). Dendrocystis Bather 1889 is an objective junior synonym
of Dendrocystites Barrande 1887, of which the type species is Cystidea sedgwicki Barrande 1867.

D. scoticus is a famous animal and Bather’s reconstruction of it (1913, text-fig. 8, p. 374) has been quoted
in many text books and general works without mentioning the trivial name. Such citations are not included
in the above list.

History of study. The anatomy of D. scoticus has previously been dealt with in two important works: Bather
(1913) and Caster (1968). Bather’s most notable contribution was to discover the anus for the first time in
solutes. Caster was responsible for a number of striking advances: he was the first to discover the pore nodes
(hydropore node and gonopore node) in this species and he noted prominences which he called ‘horns’, though
he gave no clear account of their position and relations. He also greatly clarified the systematic situation by
distinguishing between D. scoticus Bather and the very different solute Girvanicystis batheri Caster which
occurs in the same bed.

Material. The material examined is as follows: British Museum (Natural History): Gray Collection,
El 66220-El 6223, E23195-E23200, E23202-E2321 5 (including holotype E23207), E23217-E23231, E23233-
E23244, E23246-E23247, E23249-E23255, E23257-E23259, E23289, E23339, E23346-E23348, E23351,
E23459, E23461, E23466, E23472, E23700, E23714, E23718, E23759, E23762, E23764, E23765, E23768,

634

PALAEONTOLOGY, VOLUME 33

E23769, E28410-E28416, E28418, E28422, E28423, E28426, E28428, E28432-E28435, E28438-E28450,
E28452-E28456, E28459, E28460, E28463-E28478, E28481, E28483-E28507, E2851 1, E28513-E28515,
E28517, E28519, E28521-E28529, E28531-E28538, E28540-E28550, E28556, E28557, E28576-E28583,
E28585-E28589, E28592-E28596, E28598-E28602, E28760-E28821, E28823, E28825-E28830, E28832-
E28836, E28838-E28851, E28853, E28854, E28856-E28860, E28863-E28872, E28874-E28880, E29602-
E29609, E52763, E52765, E52870, E62619, E62620, E62622-E62626, E62754, E62755; L. R. M. Cocks
Collection, E29365; A. B. Smith Collection, E62521-E62533, E62537; R. P. Tripp Collection, E62778-
E62779. Hunterian Museum, Glasgow: E5718, E5719, E5722-E5725, E5729-E5732, E5734-E5736, E5738,
E5777, E5778, E5782, E5789, E5792, E5794, E5795, E5797, E5798, E5801-E5803, E5820-E5823, E5826,
E5827, E5842, E5856, E5890, E5892, E5893, E5914, E5915, E5947. The total number of specimens examined
was 399 and the number of individuals was rather greater than this.

Horizon and locality. All of the specimens examined were from the famous Starfish Bed of the Upper
Ordovician Ashgill Series, near Girvan, Scotland. The locality is S6 of Harper (1982, p. 265) which is at
National Grid reference NS 2503 0379 on the south bank of the stream called Ladyburn. The valley in which
the Ladyburn flows is sometimes called Threave (or Thraive) Glen. For details see Harper (1981, p. 28) and
text-fig. 1 herein. This locality is about 8-5 km NE of the church at Girvan (formerly in Ayrshire, now in
Stratchclyde Region, Scotland).

24 25 26

text-fig. 1. General position and detailed map of the Starfish Bed locality (S6 of Harper 1982). Locality map
traced from Ordnance Survey 1 : 25 000, sheet NS20, Dailly. The dotted lines are field boundaries (stone walls).

According to Harper (1981), the Ladyburn Starfish Bed at this locality is 48 cm thick and divided into three
beds (a, b, and c). The lowest of these (Bed a) is 22 cm thick, the famous Starfish Bed of the Gray family and
it probably yielded most of the known fossils. Bed b is about 10 cm thick and relatively unfossiliferous, while
Bed c is 15 cm thick and contains abundant fossils, though it is not so productive as Bed a. Stratigraphically,
the Starfish Bed is near the top of the Farden Member of the South Threave Formation of the Drummuck
Group and belongs to the topmost part of the Rawtheyan Stage of the Ashgill Series (Harper 1984, p. 3).

The matrix of the fossils is a greenish silty sandstone. A considerable clay-grade admixture accounts for the
good histological detail of the preserved fossils. The calcite of the skeleton of D. scoticus has nearly always been
dissolved so that the plates are now represented by rust-lined cavities in the rock.

As to conditions of deposition, the beautifully articulated condition of many of the fossils of the Starfish Bed

JEFFERIES: D E N D RO C YSTO I D ES AND CHORDATE ANCESTRY

635

suggests that they were killed by burial (obrution in Seilacher’s terminology; see Goldring and Stephenson
1972, p. 612) or perhaps were buried immediately after death. Goldring and Stephenson considered that the
Starfish Bed was laid down in shallow water, but Harper (1981) believed that the fossils did not normally live
where they are now found, and that the rocks of the Ashgill Series near Girvan were deposited as a turbidite
fan at the foot of the continental slope of a continent which lay NW of the lapetus Ocean. This implies that
the fossils of the Starfish Bed normally lived on the continental shelf north-west of where they are now found
and died when they were carried south-eastwards, downslope and out to sea by turbidity currents.

It seems quite possible that the animals of the Ladyburn Starfish Bed were transported by turbidity currents,
as Harper believes, shortly before they died. It seems unlikely that they were usually dead before burial,
however. A particular reason for believing that they were killed by burial, as Goldring and Stephenson
proposed, is seen in the cornute Cothurnocystis elizae (Jefferies 1968, p. 255), where the dorsal and ventral
integuments are often separated from each other by a thin layer of rock behind the mouth and right of the strut
in the head, but are in contact with each other farther away from the mouth, left of the strut. This suggests
that the animal ingested a last mouthful of mud, before being killed. If after death it had then been tumbled
about by turbidity currents on the sea floor, it seems almost certain that the mud inside the animal would have
become evenly distributed, instead of being preferentially right of the strut as observed. I therefore believe that
the articulated fossils of the Starfish Bed, such as C. elizae and D. scoticus , lived, at least for a few seconds,
where they are now found, even if they had been transported downslope by a turbidity current just before
death.

ANATOMICAL DESCRIPTION

General morphology

Dendrocystoid.es scoticus consists of a head, an arm and a tail. In the following description, I shall
deal with the external anatomy first and the internal soft-part anatomy later. The main features of
D. scoticus , in my view, can readily be compared with those of cornutes, as will gradually become
clear from what follows. So far as possible, I therefore apply the cornute terminology which I have
used in my previous work (see e.g. Jefferies 1986, Chap. 7). In this sense, for example, I use ‘right’,
‘left’, ‘dorsal’, ‘ventral’, ‘anterior’, ‘posterior’ and ‘fore tail’, ‘mid tail’ and ‘hind tail’. When
naming particular plates I use an objective notation, similar to that formerly applied to cornutes and
mitrates (e.g. Jefferies 1968). In this notation: CP = cover plate of arm; D = dorsal plate of arm;
M = marginal plate (of tail insertion); P = proximal plate of arm base; suffixes i 2 3 etc. refer to
position in a proximo-distal sequence; suffix L = left; suffix R = right; suffix M = median; suffix A
= anterior; suffix P = posterior; suffix D = dorsal; suffix v = ventral.

The head is trapezium-shaped in dorsal aspect (text-fig. 2). The trapezium is widest near the
posterior end of the head and becomes narrower anteriorly. The tail is inserted somewhat left of the
middle of the posterior edge.

The size of D. scoticus can never be precisely measured in all aspects, partly because the specimens
are incomplete and partly because the head was not rigid. One of the best-preserved individuals is
E28768 in the BM(NH) (pi. 6, fig. 2). The total length was about 128 mm as preserved (the tail is
incomplete), the head width was about 32 mm (the left side is missing), the arm was about 28 mm
long, the head length was about 36 mm, and the tail length, so far as preserved, was 76 mm.

External features of the head. The head is approximately, but not accurately, bilaterally symmetrical.
The most obvious departures from symmetry are that the arm is at anterior right, an antibrachial
process is at anterior left, the right posterior lobe of the head (bearing the anus) is larger and more
rounded than the left posterior lobe (the latter bearing, as I believe, a gill slit), and the left side of
the head in dorsal aspect shows an obtuse lateral prominence, while no such prominence exists on
the right.

The dorsal surface of the head is less flat than the ventral surface, though, in this respect, the
difference between the two surfaces is less obvious in very large specimens (on which the
reconstruction in text-fig. 2 is based; cf. pi. 1, figs. 1-3, text-fig. 4) than in smaller ones (pi. 1, figs.
4-8, text-fig. 4). On the dorsal surface there are an anterior and a posterior dorsal prominence. The
anterior dorsal prominence is situated in the mid line between the base of the arm (on the right) and

636

PALAEONTOLOGY, VOLUME 33

text-fig. 2. Dendrocystoides scoticus Bather. Reconstruction of external features: (a) right lateral aspect; (b)
dorsal aspect and transverse section of hind tail; ( c ) ventral; (d) posterior (omitting tail); (e) anterior (omitting

arm); (/) left lateral.

the antibrachial process (on the left). In middle-sized individuals it is thimble-shaped and slopes
equally in all directions (pi. 1, fig. 5, text-fig. 4). In large individuals, however, it projects strongly
forwards, so that its anterior face is actually an overhang and visible in ventral aspect (text-fig. 2;
pi. 1, fig. 3, text-fig. 4; pi. 6, figs. 2, 6, text-fig. 9), (text-fig. 2; pi. 1, fig. 3, text-fig. 4; pi. 6, figs 2,
6, text-fig. 9). The posterior dorsal prominence is situated somewhat right of the mid line. It is
capped by a hemispherical plate (pi. 3, fig. 9; text-fig. 6) and usually slopes away from this plate in
all directions at about 40°. Like the anterior prominence, the posterior prominence is most obvious
in large individuals.

On the ventral surface, there are likewise anterior and posterior prominences. Especially in small
and medium-sized individuals, however, these protrude much less than the dorsal prominences (pi.
1, fig. 5, text-fig. 4; pi. 6, fig. 4, text-fig. 9). Their summits approximately underlie the positions of
the corresponding dorsal prominences, but are somewhat more posterior. Caster (1968, p. S591)

JEFFERIES: D E N D ROC YSTO I D ES AND CHORDATE ANCESTRY

637

referred to the anterior and posterior ventral prominences respectively as the distal and proximal
obverse tumescences.

In transverse section through the head, dorsal and ventral surfaces meet at a sharp angle on the
left, posterior to the obtuse lateral prominence already mentioned, and the plates of the respective
surfaces meet at this edge along a rather straight line (text-fig. 2; pi. 7, figs. 1 5, text-fig. 10). This
line probably functioned as a hinge, allowing the roof of the head to be raised and lowered slightly
in this region. On the right side of the head, dorsal and ventral surfaces meet at a rounded contact
in transverse section, with no sign of such a hinge.

The plates of the head skeleton are in general unspecialized and irregular. There are, however,
some specialized and nameable plates, particularly at the arm base, at the tail insertion, in the anal
region right of the tail and in the region of the gill slit left of the tail. These specialized plates will
be described below when the respective parts are discussed.

The arm. The arm projects forwards and somewhat downwards from the anterior right extremity
of the head (text-fig. 2; pi. 2, fig. 1, text-fig. 5; pi. 6, fig. 2, text-fig. 9). Its skeleton consists of four
series of plates - two series of cover plates ventrally and two series of larger ‘flooring’ plates
dorsally which, to avoid ambiguity, I shall refer to as dorsal plates. The exact length of the arm
cannot be determined because it is never complete in the available specimens. The cover plates are
shorter and more numerous than the dorsal plates, except in the proximal part of the arm, and there
is no numerical correspondence between dorsal-plate and cover-plate series. The right and left cover
plates exactly alternate with each other, whereas the right and left dorsal plates do not. The right
and left cover-plate series meet at a smooth curve externally, without interdigitating, though the two
series do interdigitate internally where the plates are complexly sculptured. Analogy with
echinoderms suggest that the cover plates could open in life so that tube feet could emerge. The right
and left dorsal series of plates, on the other hand, would have been flexibly but permanently held
together at the dorsal edge of the arm. In transverse section (text-fig. lie) the arm was not
symmetrical for its dorsal edge was consistently leftward of the line of contact of right and left cover
plates. The cover plates abut against each other, whereas the dorsal plates imbricate, each one being
overlapped posteriorly by its more proximal neighbour. This imbrication would allow the arm to
bend in life, especially upwards, rightwards and leftwards.

The plates of the arm-base region, as already mentioned, are specialized (text-fig. 3; pi. 2, fig. 2;
pi. 5, figs. 3, 5, 6, text-fig. 8; pi. 6, figs. 5-7, text-fig. 9). The most proximal pair of cover plates, given
the notation CPlL and CP1R, are much larger than more distal cover plates, some of which they
strongly overlap. Right and left proximal cover plates meet each other at a flat suture and it is fairly
certain that they could not open in life. This is shown by a pathological specimen (pi. 6, fig. 7, text-
fig. 9) in which an extra plate is inserted distally between the two first cover plates and has formed
with them a typical triple junction where the sutures meet at 120°. Such a junction shows that all
three plates had always grown in mutual contact, without relative movement. The most proximal
pair of dorsal plates (Dn , D1R) make broad contact with the most proximal cover plates, and differ
from more distal dorsal plates in not meeting at the dorsal edge of the arm. Indeed, they stop
dorsally far short of this edge, since the right one abuts dorsally against the right dorsal proximal
plate of the arm base (PRD) and the left one against the left dorsal proximal plate (PLD). A third
proximal plate, ventral and somewhat right of median in position with respect to the arm, is named
the median ventral proximal plate (PVM). It makes contact anteriorly with the two most proximal
cover plates (CP1L, CP1R) and D1R, while dorsally on the right it abuts against Pl{n. On the left of
the arm base, the gap between PVM and PLD is filled, behind D1L, with two or three irregular plates.
The right dorsal proximal plate ( P RD) carries a boss penetrated by small holes. The histological
nature of this boss suggests that it is a node for hydropores (Caster 1968, fig. 372, p. S584) since
its porosity resembles the madreporic plate of an echinoid. This suggestion is confirmed by the
position of the boss near the mouth, since the latter presumably lay at the base of the arm, and the
hydropore is the opening nearest the mouth in cystoids such as Glyptosphaerites (Jefferies 1986, fig.
7.18).

638

PALAEONTOLOGY, VOLUME 33

gonopore node

2 L

L D

1 Omm

text-fig. 3. Dendrocystoides scoticus. External anatomy of arm-base region: ( a ) anterior aspect (arm omitted
distal to D2R); (b) dorsal; (c) ventral; (d) right lateral.

JEFFERIES: D EN D RO C YSTO 1 D ES AND CHORDATE ANCESTRY

639

Anterior part of the head. Another set of openings is situated slightly above and left of the hydropore
node, at the right base of the anterior dorsal prominence. These openings are variable. Each one is
located at the summit of a small paraboloidal calcitic process. The number of these processes ranges
from one to about six or seven and they are joined at their bases. The processes, with their pores,
are sometimes grouped around the suture between two plates, but sometimes are situated in the
middle of a plate. When best developed, the openings are obvious, circular and about 150 //m in
diameter (pi. 5, fig. 3, text-fig. 8). Sometimes, however, a circular slight concavity of the same
diameter is present at the distal end of a process but is not penetrated by a hole (pi. 5, figs 5, 6, text-
fig. 8; pi. 6, fig. 5, text-fig. 9). Such blind processes give the impression that an opening which
previously existed had been occluded with skeletal calcite during life. If the mouth was situated at
the base of the arm, as seems likely on echinoderm analogies, then the pores now being discussed
would be farther from the mouth than the hydropore is, since they do not penetrate one of the
standardized plates of the arm base but are situated on, or between, the irregular plates just behind
and above these.

The openings at the ends of the paraboloidal processes are probably gonopores. This is suggested
by their position, since in cystoids such as Glyptosphaerites (Jefferies 1986, fig. 7-18, p. 217) the
gonopore is consistently farther away from the mouth than is the hydropore. Also, if the openings
were gonopores, their frequent occlusion could represent closure when not reproducing. If these
openings were gonopores, then the anterior dorsal prominence of the head, at the base of which
their canals emerge, probably represents the position of a gonad. As discussed later, the limits of
the gonad seem to show on the surface of the natural internal mould. Despite special search, no
feature was observed in these presumed gonopores which might have suggested separate sexes. I
therefore cannot say whether this solute was dioecious or hermaphrodite.

The antibrachial process, situated at the left anterior extremity of the head, is, like the dorsal
prominences, variable in shape and relative size. It is best developed in the largest specimens as
shown in the reconstructions (text-figs. 2, 1 1) and in pi. 1, fig. 1 (text-fig. 4), and in these specimens
it projects downwards, leftwards and forwards from the head. In some smaller specimens it is
relatively small and projects leftwards (pi. 1, fig. 7, text-fig. 4). It is more prominent in D. scoticus
than in any other known species of solute, although Bather (1889) records its presence, more weakly
developed, in Dendrocystites sedgwicki (Barrande 1867) from the Upper Llandeilo Letna Formation
of Czechoslovakia. By projecting downwards and forwards in Dendrocystoides scoticus, its shape
and orientation are: (1) grossly the same as the arm, but on the opposite side of the head; and (2)
reminiscent of the anterior appendages of cornutes, which probably functioned in locomotion to
hinder forward movement and facilitate rearward movement (Jefferies 1986, p. 193). During
locomotion, the arm and the antibrachial process of D. scoticus may well have functioned in a
similar way to these cornute appendages, preventing the head from moving forwards during
locomotion and helping it to slide rearwards. The position of the antibrachial process on the left side
of the head suggests that it counterbalanced the arm in locomotion, resisting forces acting leftwards
and forwards, whereas the arm would resist those acting rightwards and forwards.

The teal. The tail consists of three parts, the fore, the mid and the hind tail, and I believe that these
regions are probably homologous with the like-named parts of the cornute tail.

The skeleton of the hind tail (text-fig. 2; pi. 6, figs. 2, 6, text-fig. 9; pi. 7, figs. 1, 2, text-fig. 10)
consists of a dorsal and a ventral series of plates which between them enclose a distally tapering
canal, circular in transverse section (text-fig. 2b). Except anteriorly towards the mid tail, the plates
of the ventral series are thicker walled than those of the dorsal series and occupy much more of the
external surface. Because of this, the canal is not central in a transverse section of the hind tail but
located entirely dorsal to the centre. Dorsal and ventral plates tend to alternate in position and
therefore would have interlocked. Consequently it seems almost certain that, in life, the hind tail was
rigid in all directions.

At the posterior end of the hind tail there seems to have been a ventrally-directed rigid hook,
ending abruptly, and the axis of the hook is at about 70° to the axis of the rest of the hind tail (text-

640

PALAEONTOLOGY, VOLUME 33

fig. 2a, f; pi. 1, fig. 4). In the hook, the dorsal and ventral series of plates are more equally massive
than anteriorly and the canal more nearly central in transverse section. A reason for doubt is that
the hook as reconstructed is based on a single specimen only and this, as preserved, is not now
connected to the rest of the tail. All the other specimens of the hind tail either end abruptly (pi. 6,
fig. 6) or else, much more commonly, the tail runs over the edge of the block and is not seen. I
believe, however, that the hook probably belongs to D. scoticus since it clearly is part of a solute
and since Caster (1968, p. S709, fig. 382/4) has published a photograph of a similar terminal hook
in Dendrocystites barrandei (Bather 1913), a species which in most respects is very similar to
Dendrocystoides scoticus. Also Ubaghs (in Ubaghs and Robison 1985, figs. 1, 2, 8, 9; text-fig. 16a
herein) has recorded a specialized terminal hook, in my view directed ventrally, in the tail of the
American Middle Cambrian solute Castericystis vali Ubaghs and Robison 1985. Anteriorly in the
hind tail of D. scoticus the ossicles shorten, both in the dorsal and the ventral series.

In the fore tail (text-fig. 2) the skeleton consisted of major and intercalary plates surrounding a
large lumen. The major plates are in four series - right, dorsal and ventral, and left, dorsal and
ventral. Each dorsal major plate is united at a suture (pi. 3, fig. 9, text-fig. 6; pi. 7, fig. 7, text-fig.
10) with a ventral major plate on the same side to form a C-shaped element built of two plates.
Successive C-shaped elements have curved edges where they meet the dorsal and ventral mid-line
of the tail, and these edges overlap, or are overlapped by, those of the C-shaped elements of the
opposite side (pi. 3, fig. 9, text-fig. 6; pi. 6, figs. 2, 4, text-fig. 9). The number of C-shaped elements
is about nine on both sides. The intercalary plates of the fore tail (pi. 1, fig. 9; pi. 6, fig. 2, text-fig.
9) are small and irregular and were evidently carried in a pliable, small-plated integument which
connected the major plates together. They are most obvious near the anterior end of the fore tail,
especially at the dorsal and ventral mid-lines. The exact shape of the major plates suggests that the
foretail was able to flex, particularly rightwards and leftwards, but also up and down. The lumen
of the fore tail was presumably largely filled with muscle and some anticompresional structure,
probably a notochord, would be needed to allow flexion without telescoping. The quadriserial
skeleton of the fore tail is comparable with that in most cornutes, particularly with the two least
crownward of known cornutes Ceratocystis perneri and Protocystites menevensis , in which right and
left ventral plates strongly overlap each other in the ventral mid line (text-fig. 14b; Jefferies 1986,
p. 214; Jefferies et al. 1987, fig. 10b, p. 442).

The mid tail of D. scoticus has a complicated and rather irregular skeleton, transitional between
the fore tail and hind tail (text-fig. 2). The dorsal hind-tail series of ossicles can be followed
continuously forward into the right dorsal series of major plates in the fore tail (pi. 6, fig. 2, text-
fig. 9). And the ventral hind-tail series of ossicles can likewise be followed forward into the right
ventral series of major fore-tail plates (pi. 6, fig. 3, text-fig. 9; pi. 7, fig. 1, text-fig. 10). The left side
of the mid tail is not comparable for it begins posteriorly with a single half-cone-shaped plate (the
posterior left mid-tail plate; pi. 6, fig. 3, text-fig. 9; pi. 7, fig. 1, text-fig. 10) which is neither dorsal
nor ventral in position, but left lateral. A dorsal and a ventral plate abut against this anteriorly, and
are the posterior members of series leading respectively into the dorsal left and ventral left major
plates of the fore tail.

This description of the transition between hind tail and fore tail differs from that of Caster (1968,
pp. S583ff.) who considered that, in all solutes, the left ventral hind-tail series passed forward into
either the left or right ventral fore-tail series, while the dorsal hind-tail series would pass forward
into the diagonally opposite dorsal fore-tail series. In other words, if the ventral hind-tail series
passed into the left ventral fore-tail series, then the dorsal hind-tail series would pass forward into
the right dorsal fore-tail series and vice versa. Perhaps Caster's description is true for some solutes.
It does not hold for D. scoticus.

Ubaghs (1981) has argued that the tails of solutes are not homologous with those of cornutes and
mitrates or, as he would say, the solute stele is not homologous with the cornute and mitrate
aulacophore. This assertion reflects his view that the ‘aulacophores’ of cornutes and mitrates were
feeding arms. In accordance with this view, he believes that what he calls the cover plates of
cornutes could open to reveal a water-vascular system inside the appendage. (The Ubaghs school

JEFFERIES: D END RO C Y STO I D ES AND CHORDATE ANCESTRY

641

formerly asserted the same of the mitrate ‘cover plates’ but have since decided that such opening
in the mid-line could not happen [Parsley 1982, 1988].) In support of his opinion that the tails of
solutes are not homologous with those of cornutes, Ubaghs (1981, p.3flf.) has cited some genuine
differences between them: (1) the hind-tail skeleton of solutes consists of an upper and lower series
of ossicles, rather than a lower series of ossicles with paired ‘cover plates’; (2) the mid-tail skeleton
of solutes is made of many somewhat irregular plates whereas the mid tail of cornutes is dominated
by a single element, the stylocone, with ‘cover plates’ above it; and, (3) the canal of the hind tail
of solutes expands gradually into the funnel-shaped lumen of the mid tail which passes gradually,
in turn, into the broad lumen of the fore tail, whereas in cornutes the lumen of the stylocone is
sharply distinct from the canal of the more distal parts of the tail (although continuous with it). In
my view these real differences do not show that the tails of solutes, cornutes and mitrates are not
basically homologous. They merely show, as might be expected, that these organs are not in all ways
the same.

The plates of the tail insertion (text-fig. 2; pi. 7, figs. 6, 7, text-fig. 10) do not seem to be completely
standardised and the region was difficult to reconstruct since very few specimens show it in both
ventral and dorsal aspect - most of the known specimens of D. scoticus are not preserved as part
and counterpart. Normally there seem to have been seven plates: two large dorsal plates, which I
classify as left and right first dorsal marginals with the notation M1LD and M1RD, one large posterior
median ventral plate (Mpv), and sometimes two smaller marginal plates on the left and the right,
dorsal and ventral to each other - M2RD and MRV on the right and M2LD and MLV on the left. I am
doubtful whether these smaller marginal plates always number two on each side, or whether there
may not sometimes be one less or one more. The internal structure of the plates of the tail insertion
will be discussed below. Caster (1968, p. S592) referred to the plates of the tail insertion as the
adsteleal girdle.

Posterior part of the head. As to the head openings, the gonopores and hydropores have been
described already and the position of the mouth, at the base of the arm, has been implied. Two
openings remain to be discussed, the anus on the posterior face of the posterior right lobe of the
head, and a gill slit, as I believe, at the end of the posterior left lobe of the head.

I describe the anus first. As already mentioned, it was discovered by Bather (1913, p. 393) in D.
scoticus for the first time in solutes. It is surrounded by an anal pyramid of long spike-shaped plates
and this pyramid is set in a horizontally elongate periproct which was evidently flexible in life, and
is armoured with numerous small plates (pi. 5, figs. 2, 4, text-fig. 8). The periproctal plates left of
the anus are markedly elongate in a horizontal direction which suggests that the periproct here
would sometimes bow outwards as part of a horizontal cylinder, which probably implies that the
roof of the head could be pulled here towards the floor. Right of the anal pyramid, the periproctal
plates are also horizontally elongate, but not to the same extent, which suggests that the roof here
could be depressed, but not so much as left of the pyramid. The frame of the periproct is not well
defined, since the integument plates tend to decrease in size towards the periproct. These framing
plates tend to be imbricate, whereby the large ones outside the periproct overlap the peripheral plates
of the periproct itself. Just to the right of the anus, both dorsally and ventrally, an individualised
framing plate larger than its neighbours is present, which Bather called the sugar-loaf plate because
of its outline. Hence there are a dorsal and a ventral sugar-loaf plate (pi. 5, figs. 2, 4, text-fig. 8; pi.
6, figs. 2, 3, 6, text-fig. 9; pi. 7, figs. 6, 7, text-fig. 10). At the sugar-loaf plates the frame of the
periproct projects posteriorly. Perhaps these two plates helped the anus to protrude posteriorly
during defaecation.

If a gill slit existed in D. scoticus we would expect, on cornute analogies, that it would be located
near the left posterior angle of the head. As it happens, this region is difficult to reconstruct because
few specimens show both part and counterpart and the region has usually been crushed on burial.
Some specimens, however, show that there was an area of small plates in this region surrounded by
a frame of larger plates (pi. 3, figs. 1, 2, text-fig. 6; pi. 7, fig. 5, text-fig. 10). The frame was made
up ventrally of two or three rather thick, ventrally convex plates (subbranchial plates, pi. 3, fig. 2,

642

PALAEONTOLOGY, VOLUME 33

text-fig. 4. Dendrocystoides scoticus. Half-scale outline drawings of figs. 1-3 and 5-8 of pi. 1. Each drawing
has the same number as the corresponding photograph on the plate; a.-br. pr. = antibrachial process; a. d. pr.
= anterior dorsal prominence; a. lim. per. m. = anterior limit of peripheral muscle; a. v. pr. = anterior ventral
prominence; lim. g. = limit of gonad; lim. ph. = limit of pharynx; p. d. pr. = posterior dorsal process; p. lim.
ph. = posterior limit of pharynx; p. v. pr. = posterior ventral process; r. lim. per. m. = right limit of peripheral

muscle; st. a.-br. = stump of antibrachial process.

EXPLANATION OF PLATE I

Figs. 1 13. Dendrocystoides scoticus. 1-3, Hunterian Museum, E5890; x L2. Internal mould of head and
external mould of part of fore tail; 1, ventral aspect; 2, dorsal aspect; 3, left lateral aspect; see also text-
figs. 4/1, 2, 3.4, BM(NH), E29609; x 1-2. Latex of terminal hook of hind tail in left lateral aspect
(determination as D. scoticus not certain); th = terminal hook; cr = crinoid stem on same block. 5-8,
BM(NH) E23339, x 2. Natural internal mould of head (figs. 5-7) and latex of anterior left portion of head
(fig. 8). See text-fig. 4 (5-8). The presence of an antibrachial process and the anterior dorsal prominence on
one and the same specimen shows that these structures are different. 9, BM(NH) E28529, x 5. Natural
internal and external mould of posterior part of head and adjacent fore-tail region in ventral aspect to show,
especially, the left pyriform body in ventral aspect (lpb). 10, BM(NH) E28470, x 5. Natural internal mould
of posterior part of head and adjacent fore-tail region in ventral aspect to show, especially the brain (br) in
undissected condition. 1 1, BM(NH) E28494, x 5. Natural internal mould of posterior part of head (plate
Mpv) to show cast of brain in contact with Mpv, so far as it can be revealed by dissection. 12,13, BM(NH)
E284 16; 12 x 2 ; 13, detail, x 5. Natural mould, mainly external, of posterior part of head, fore tail and mid
tail in dorsal aspect, to show especially the left pyriform body (lpb) as an internal mould.

PLATE I

JEFFERIES, Dendrocystoides scoticus

644

PALAEONTOLOGY, VOLUME 33

text-fig. 5. Two-thirds-scale outline drawings of figs. I and 3-6 of pi. 2. Each drawing has the same number
as the corresponding photograph on the plate; br. = brain; d. pi. = dorsal plate of arm; d. r. = dorsal ridge
in infilling of arm; f. t. = fore tail; gonop. d. = gonopore duct; gr. = groove of unknown significance; hyd. d.
= hydropore duct; hyd. n. = hydropore node; inf. DL = infilling of left dorsal arm plates; lim. gon. = limit
of gonad; lim. stom. = limit of stomach; Ig. = long tube foot of a triad; med. = medium-length tube foot of
a triad; or. part. = oral partition; p. d. pr. = posterior dorsal prominence; sh. = short tube foot of a triad; t.
w. v. = terminal water vessel; CPR, PID, D, etc. = plates of arm and arm-base region; cf. text-fig. 3.

EXPLANATION OF PLATE 2

Figs. 1-6. Dendrocystoides scoticus. All specimens in BM(NH). Compare text-fig. 5. 1,2, E28803. 1, Dorsal
aspect of natural internal mould of head, arm and fore tail, x 2. 2, arm-base region x 5. 3, E28507, x 5.
Natural internal mould of arm-base region in antero-ventral aspect x 5. 4, 5, E28787. Latexes of arm-base
region, x 5; 4, ventral aspect. 5, dorsal aspect. 6, 7, E28790. Natural internal mould of arm in ventral aspect
to show especially the terminal water vessel and the triads of lube feet. 6, x 10; 7, x 40, to show detail.

PLATE 2

JEFFERIES, Dendrocystoides scoticus

646

PALAEONTOLOGY, VOLUME 33

text-fig. 6. Reduced-scale outline drawings ( x 0.6) of pi. 3, figs. 1-14. Each drawing has the same number as
the corresponding photograph on the plate; br. = brain; br. op. = branchial opening; br. pi. = branchial
plate; curv. med. e. = curved median edge of dorsal plate of fore tail; interc. = intercalary plate of fore tail;
1. perip. = left region of periproct; p. lim. ph. = posterior limit of pharynx; r. p. lim. st. = right posterior limit
of stomach; subbr. pi. = subbranchial plate; sut. d./v. = suture between dorsal and ventral plate of fore tail;

Mild’ M1rd, M2rd = plates of tail insertion; cf. text-fig. 2.

EXPLANATION OF PLATE 3

Figs. 1-14. Dendrocystoides scoticus . to show the gill slit (branchial opening) and associated structures.
Compare text-fig. 6. Specimens in BM(NH). 1-7, E28454. 1 , 2 latexes of dorsal and ventral aspects; x2;3-7,
cardboard model based on the latexes in figs. 1,2; x 036 of model, c. x4-9 of specimen. 3, Posterior aspect.
4, Anterior aspect. 5, Dorsal aspect. 6, Ventral aspect. 7, Left lateral aspect. 8-14, E23239. 8, General view
of natural mould in dorsal aspect ( x 2) and the brachiopod shell which the Dendrocystoides scoticus was
squashed upon at death - this brachiopod probably explains the inflation with mud of the posterior parts
of the head and therefore the good preservation of the branchial region. 9, General view of latex of dorsal
surface of head and fore tail ( x 2). Figs. 10-14, details of branchial region ( x 5). 10, Posterior aspect. 11,
Lateral aspect. 12, Dorsal aspect. 13, Ventral aspect. 14, Left antero-lateral aspect.

PLATE 3

JEFFERIES, Dendrocystoides scot ic us

648

PALAEONTOLOGY, VOLUME 33

text-fig. 7. Half-scale outline drawings of pi. 4, figs. 1-6. Each drawing has the same number as the
corresponding photograph on the plate; a. lim. ph. = anterior limit of pharynx; ant. d. pr. = anterior dorsal
prominence; ap. a. d. pr. = apex of anterior dorsal prominence; ap. p. d. pr. = apex of posterior dorsal
prominence; br. = brain; f. t. = fore tail; gon. d. = gonadial duct; gon. str. = gonadial striae on surface of
gonad; hyd. d. = hydropore duct; lim. gon. = limit of gonad; 1. fac. f. t. = lateral facet of dorsal plate of fore
tail (making suture with corresponding ventral plate); or. part. = oral partition; p. lim. ph. = posterior limit
of pharynx; p. lim. st. = posterior limit of stomach; r. p. lim. st. = right posterior limit of stomach; st. =

stomach.

EXPLANATION OF PLATE 4

Figs. 1-6. Dendrocystoides scoticus. Natural internal moulds in dorsal aspect. Compare text-fig. 7. Specimens
in BM(NH). 1, E28434, x 5. 2, E28769, Head and foretail x 2. 3, E28532, Left anterior part of head to show
gonadial striae x 5. 4, E28601, Right anterior part of head x2. 5, E23199, Posterior left part of head x 2.
6, E28439, Posterior part of head and fore tail x 3.

PLATE 4

JEFFERIES, Dendrocystoides scoticus

650

PALAEONTOLOGY, VOLUME 33

text-fig. 8. Reduced-scale outline drawings ( x0.6) of pi. 5, figs. 1-6. Each drawing has the same number as
the corresponding photographs on the plate; a. d. pr. = anterior dorsal prominence; anus = anus; d. sug.
If. = dorsal sugar-loaf plate; f. t. = fore tail; hyd. d. = hydropore duct; hyd. n. = hydropore node; lat.
pr. = lateral prominence; lim. gon. = limit of gonad; 1. perip. = left region of periproct; occ. gon. = occluded
gonopore; op. gon. = open gonopore; op. part. = oral partition; r. perip. = right region of periproct; v. sug.
If. = ventral sugar loaf; CP1L, M1RD etc. = plates of arm, arm base and tail base as shown in text-figs. 2, 3.

EXPLANATION OF PLATE 5

Figs. 1-6. Dendrocystoides scoticus. Compare text-fig. 8. Specimens in BM(NH). 1, E28439, x 3. Natural
internal mould of head and fore tail in dorsal aspect. 2, E23229, x 2. Latex of dorsal surface, to show
especially the anus, periproct and dorsal sugar-loaf plate. 3, E23219, x 5. Latex of arm-base region with
gonopore node and hydropore mode; one of the gonopores is occluded. 4, E23237, x 5. Latex of anal region
in postero-ventral aspect. 5, E28592, x 5. Latex of arm-base region with gonopore node, hydropore node
and the first cover plates in dorsal aspect. The gonopores are all occluded. 6, E23718, x 5. Latex of arm-base
region with gonopore node and hydropore node. Some of the gonopores are occluded.

PLATE 5

JEFFERIES, Dendrocystoides scoticus

652

PALAEONTOLOGY, VOLUME 33

text-fig. 9. Half-scale outline drawings of pi. 6, figs. 1-7. Each drawing has the same number as the
corresponding photograph on the plate; a-br. pr. = antibrachial process; a. d. pr. = anterior dorsal
prominence; adven. = adventitious plate between CP1L and CP1R; as. tr. = asymmetrical transition between
mid tail and fore tail; a. v. pr. = anterior ventral prominence; curv. med. e. = curved median edge of dorsal
plates of fore tail; d. sug. If. = dorsal sugar-loaf plate; d. pi. = dorsal plate of hind tail; gon. n. = gonopore
node; hyd. n. = hydropore node; int. CPL, int. CPR = interior of left and right cover plates; interc. intercalary
plate of fore tail; lat. pr. = lateral prominence; 1. d. pi. = left dorsal plate of fore tail; I. v. pi. = left ventral
plate of fore tail; 1. perip. = left region of periproct; occ. gon. = occluded gonopore; op. gon. = open
gonopore; p. 1. mid. t. = posterior left plate of mid tail; p. d. pr. = posterior dorsal prominence; r. perip. =
right region of periproct; subbr. pi. = subbranchial plate; tr. end. = truncated end of tail; v. pi. = ventral plate
of hind tail; v. sug. If. = ventral sugar-loaf plate; CPU, D1R, PLD etc. = plates of arm and arm base (see text-
fig. 3); M1LD, Mpv etc. = plates of tail insertion (see text-fig. 2).

EXPLANATION OF PLATE 6

Figs. 1-7. Dendrocystoides scoticus. Latexes of BM(NH) specimens. Compare text-fig. 9. 1,2, 3, E28768. 1,
Arm-base region, x 5. in dorsal aspect, with the hydropore node and gonopore node; one of the gonopores
is occluded. 2, General view in dorsal aspect, x 1-2. 3, ventral aspect, with the asymmetries of the mid tail.
4, E28776, x 2, ventral aspect. 5, 6, E23700, dorsal aspect. 5, Detail of arm-base region, x 5. 6, General view
note truncated tail and overhanging dorsal prominence. 7, E28794, x 10. Arm-base region in ventral aspect
to show pathological adventitious plate between CP1R and CP1L.

PLATE 6

JEFFERIES, Dendrocystoides scoticus

654

PALAEONTOLOGY, VOLUME 33

text-fig. 10. Reduced-scale outline drawings ( x 0.6) of pi. 7, figs. 1-7. Each drawing has the same number as
the corresponding photograph on the plate: a-br. pr. = antibrachial process; as. tr. = asymmetrical transition
between mid tail and fore tail; br. op. = branchial opening (gill slit); hr. pi. = branchial plate; curv. med. e.
= curved median edge of ventral plate of fore tail; d. pi. = dorsal plate of hind tail; d. sug. If. = dosal sugar-
loaf plate; interc. = intercalary plate of fore tail; lat. pr. = lateral prominence; I. h. 1. = left hinge line; 1. perip.
= left region of periproct; p. 1. mid. t. = posterior left plate of mid tail; r. perip. = right region of periproct;
subbr. pi. = subbranclnal plate; v. pi. = ventral plate of hind tail; v. sug. If. = ventral sugar-loaf plate; CPR,
PRD etc. = plates of arm and arm insertion (see text-fig. 3); M1LD, Mpv etc. = plates of tail insertion (see text-

fig. 2).

EXPLANATION OF PLATE 7

Figs. I 7. Dendrocystoides scoticus. BM(NH) specimens. Compare text-fig. 10. 1, 2, 3. E23198. To show
structure of presumed left hinge line. I, Latex of ventral aspect, x 2. 2, Latex of dorsal aspect, x2. 3,
Cardboard model (x 0-5 = c. x 8 for specimen) of the plates stippled in text-fig. 21/1, 2 to show how the
dorsal and ventral plates met at a straight edge - the left hinge line. 4, 5, E23207, holotype, x 1-2. 4, Ventral
aspect. 5, Dorsal aspect 42a, b. 6, 7, E28781. Latex ( x 2) of posterior part of head and fore tail, to show
especially the plates of the tail insertion. 6, Dorsal aspect. 7, Ventral aspect.

PLATE 7

JEFFERIES, Dendrocystoides scoticus

656

PALAEONTOLOGY, VOLUME 33

text-fig. 6; pi. 6, fig. 4, text-fig. 9; pi. 7, fig. 4, text-fig. 10). These would probably have raised the
posterior left angle above the sea floor. Dorsally, the frame consisted of two or three large, rather
irregular plates. In one specimen (pi. 3, figs. 8-12, text-fig. 6) the small plates of the posterior left
angle of the head seem to be preserved in their proper articulation and form a cone pointing leftwards
and rearwards. (Under the anterior part of the head of this specimen there is a large brachiopod
perhaps the mud inside the head was squashed rearwards when the D. scoticus fell on top of this
shell so that the posterior part of the head was preserved in an unusually inflated condition.) There
is no single plate capping the cone. Indeed, the arrangement of the plates seems to be such that a
small hole in the skeleton could open between the plates at the tip of the cone (pi. 3, fig. 11, text-
fig. 6). I suggest that this small hole represents an outlet opening and, on cornute analogies, that
it was a gill slit.

Internal anatomy

A number of grooves and other details can be seen on the internal moulds (text-fig. 11). Such
grooves would correspond to ridges on the internal surface of the skeleton but it is more
illuminating to consider them as indicating grooves in the surface of the soft parts and thus to use
them directly in reconstructing the soft anatomy. Once again, I am influenced, at the outset, by the
thought that, in basic plan, D. scoticus seems to be similar to the cornute Ceratocystis. None of the
available specimens showed a complete internal mould. The diagrams of such a mould shown in
text-figure I I are highly composite.

One of the most obvious features in the internal mould is in the arm-base region. Dorsally this
feature, called the oral partition (pi. 2, figs. 2, 3, text-fig. 5; pi. 4, fig. 1, text-fig. 7; pi. 5, fig. 1, text-
fig. 8), is expressed as a deep groove in the natural internal mould, obliquely transverse to the arm
and situated on the inner face of the dorsal arm-base plates P1D and PRD. Ventrally it passes into
a shallow groove or break in slope, where it crosses the inner face of plate PVM. In the skeleton, it
would correspond to a thin, sharp-topped wall on PRD and PLD which would partly separate the
cavity of the arm form the rest of the head. On echinoderm analogies, the arm would contain a
radial water vessel and tube feet and was probably used to pick food particles from the mud.
Probably the arm would graze food from the mud by sweeping left and right through or over the
superficial layer of the sea bottom. If so, the food particles would be passed back towards the base
of the arm and there enter the head, through the mouth. The oral partition would delimit the
brachial cavity posteriorly and would arch over the mouth.

A probable indication of the tube feet is preserved on the ventral surface of the internal mould
of the arm (text-fig. Ilf; pi. 2, figs. 6, 7, text-fig. 8): such moulds reveal, namely, a series of
approximately sausage-shaped structures equal in number to the cover plates. In the terminology
which I shall adopt, the proximal end of the sausage is situated where the latter makes contact with
the dorsal plate, at right or left of the arm, while the distal end is near the median line of the arm.
Each sausage is obliquely disposed so that its distal end is anterior to its proximal end. The long
axis of the sausage runs diagonally across the relevant cover plate. The sausage itself overlaps
distally on to the next cover plate anterior, near to the median line of the arm, while proximally it
overlaps on to the next cover plate posterior, away from the median line of the arm. The distal part
of each sausage is divided into three lobes by two grooves which run transverse to the length of the
arm. Because of these grooves, each sausage is clearly divided into three lobes towards its distal end.

A comparison of these structures with the tube feet of some crinoids is suggestive. Thus in many
crinoids, such as Antedon (Nichols 1960), the tube feet are arranged in groups of three (text-fig. 12),
each group arising from a terminal water vessel which itself branches from the radial water vessel.
The three tube feet of a group differ in length, there being a long one distally, a short one proximally,
and a medium-length one between the two others (‘distally’ and ‘proximally’ here refer to the
position in the pinnule of the crinoid). The short and medium tube feet are in contact externally with
a fold of soft tissue which is called a lappet and contains an ambulacral plate (a cover plate), whereas
the long tube feet emerge between successive lappets. The lappets open when the tube feet in contact
with them expand and they close when these tube feet contract. Because of their contained

JEFFERIES: D EN D ROC Y STO I D ES AND CHORDATE ANCESTRY

657

text-fig. 1 1. Dendrocystoides scoticus. Reconstructed internal mould representing the superficial anatomy of
the soft parts: (a) left lateral aspect of head; ( b ) dorsal; ( c ) ventral; (d) right lateral; (e) transverse section, seen
from in front, of an arm near, but not at. its base; (/) soft parts of arm in ventral aspect.

text-fig. 12. The tube feet of the crinoid Antedon bifida in oral aspect
(after Nichols 1960, text fig. 1, modified), corresponding to the ventral
aspect of Dendrocystoides. Note the arrangement of the tube feet (black)
in triads, each with a long, a medium and a short tube foot, and the
relation of the triads to the cover plates (dotted) and to the radial and
terminal water vessels.

1 00 micrometres

658

PALAEONTOLOGY, VOLUME 33

musculature, the long tube feet can contract suddenly towards the ambulacral groove, the medium
tube feet can contract suddenly towards or away from the groove, and the short tube feet can move
m any direction. Paul and Smith (1984) have suggested that primitive echinoderms such as
edrioasteroids could open the ambulacral plates or plating by pushing with their tube feet against
the adradial faces of these plates, very much as crinoids do. The surfaces of the tube feet of crinoids
are provided with sensitive outgrowths called papillae which, when stimulated, squirt out a stream
of mucus to entangle microscopic prey. Such prey particles will normally bump first into the papillae
of the long tube feet and, when once entangled, will then be passed to the medium tube feet, and
from there to the short tube feet which will hand them, still wrapped in mucus, onto the ciliated tract
of the ambulacral groove. By the beat of cilia in this groove, they will then be carried to the mouth.

The sausage-shaped structures observed in D. scoticus may each correspond to the terminal water
tube of a crinoid, together with the three attached tube feet. The three terminal lobes of the sausage
would, in such case, correspond to the three tube feet of a group - short, medium and long. If this
suggestion is correct, then the long tube foot, as in crinoids, will be the most distal member of its
group (distal with respect to the arm), the medium tube foot will be in the middle of the group, and
the short tube foot will be the most proximal. Moreover, as in crinoids, the longest tube foot will
coincide in position with the contact between successive cover plates (lappets) whereas the short and
medium tube foot will be in contact with the inner surface of the main part of a cover plate.
Presumably the cover plate would open when the contiguous terminal water tube distended. These
various correspondences seem too complex to be accidental and suggest that we are dealing here
with genuine homologues. If so, then the grouping of tube feet into triads, and their observed
relationship to cover plates, would probably be primitive for echinoderms, and this is true whether
D. scoticus is a stem-group chordate (as is likely) or a stem-group dexiothete or even a stem-group
echinoderm. This supports Nichols’ (1972, p. 536) assumption that the tube feet of eleutherozoans,
which are always single rather than arranged in triads, were derived from triadic tube feet of crinoid
type. There was no previous clear evidence for this assumption, however, since in hemichordates,
which are the closest living relatives of the dexiothetes, the ciliated outgrowths on each arm of the
mesocoel (which are perhaps homologous to tube feet and function in producing a feeding current
[Lester 1985]) are never arranged in triads.

The position of the radial water vessel in D. scoticus is uncertain. It presumably ran the length
of the arm, giving rise to branches connected with the terminal water vessels. It may be represented
by a dorsal ridge on the internal mould which follows the dorsal edge of the arm (text-fig. lie; pi.
2, fig. 1, text-fig. 5), or it may have been more ventral than this, somewhere inside the arm. The
obvious flexibility of the arm suggests that there would have been large longitudinal muscles in it
in life, probably filling much of its volume.

Beneath the hydropore, immediately behind the oral partition, a small upwardly pointed
projection is present on the internal mould (hydropore duct; pi. 2, figs. 2, 3, text-fig. 5; pi. 4, fig.
1, text-fig. 7; pi. 5, fig. 1, text-fig. 8). This presumably represents the dorsal end of the axial sinus
which opens into the hydropore in echinoderms (Fedotov 1924). By analogy with extant
echinoderms, the water vascular system would have joined the axial sinus inside the animal and
would thus have been connected, indirectly, with the outside world. However, any such internal
connection in D. scoticus does not show in the internal mould. The heart of D. scoticus, or in
echinoderm terms the dorsal sac ( = madreporic vesicle), would also probably have been situated
near the hydropore, but its position cannot be identified in the fossils.

The position of the gonad, inside the anterior dorsal prominence of the head, has been suggested
already. Its likely boundary on the internal mould is indicated by a slight groove which almost
encircles the dorsal anterior prominence and passes at the left from the dorsal to the ventral surface
(text-fig. 1 1 ; pi. 1, fig. 2, text-fig. 4; pi. 2, fig. 1, text-fig. 5; pi. 4, figs. 1, 3, 5, text-fig. 7). Corrugations
on the surface of the internal mould (pi. 4, figs. 1, 3) run forwards from this groove, in some
specimens, and suggest that the surface of the gonad was plicated in life. How the gonad terminated
in a ventro-posterior direction is unclear: it may, in the largest individuals, have projected rearwards
and downwards into the pharynx behind it (whose disposition will be explained below) so as to fill

JEFFERIES: D EN D ROC Y STO 1 D ES AND CHORDATE ANCESTRY

659

a large part of the ventral anterior prominence. Such a morphology would explain why this
prominence is relatively large in the biggest of all observed individuals (pi. 1, fig. 1, text-fig. 4). On
the other hand, it may be that the ventral anterior prominence corresponded to some other organ
which expanded disproportionately in old age, perhaps part of the pharnyx. It is clear, from the
inferred position of the gonad in the internal mould, that the most anterior part of it, at least, was
dorsal to the pharynx.

The pharynx would be expected to run, as in cornutes, from the gill slit at posterior left, to the
velar mouth, and the buccal cavity would run from the velar mouth to the true mouth. The likely
position of the pharynx in D. scoticus , or of the pharynx posteriorly and the buccal cavity anteriorly,
in indicated by a sinuous tract with this oblique disposition, limited at anterior left and posterior
right by grooves on the dorsal and ventral surfaces of the internal mould. There is no way of saying
whether D. scoticus had a buccal cavity or, if it did have one, where the boundary between this cavity
and the pharynx may have lain.

On the dorsal surface of the internal mould, the anterior limiting groove of the presumed pharynx
plus buccal cavity coincided in its most anterior and rightmost part with the posterior limit of the
probable gonad, as proposed above (pi. 1, fig. 2, text-fig. 4; pi. 4, fig. 5, text-fig. 7). More leftward,
the anterior dorsal pharyngeal groove departs from the groove bounding the gonad, and bends to
pass leftward and posteriorly, parallel to the posterior left margin of the head (pi. 4, figs. 2, 5, text-
fig. 7). Perhaps the groove in this region separated the pharynx from muscles which pulled the roof
downwards about the left peripheral hinge line (already mentioned) and which thus squirted water
through the gill slit. The groove bounding the pharynx plus buccal cavity to the right and behind
on the dorsal surface (pi. I, fig. 2, text-fig. 4; pi. 2, fig. 1, text-fig. 5; pi. 3, fig. 8, text-fig. 6; pi. 4,
figs. 2, 5, text-fig. 7) has a sinuous course parallel to the left anterior pharyngeal groove. Its exact
course gives the strong impression of being deflected by an inflated, approximately spherical organ
occupying the posterior prominence, possibly the stomach. This interpretation is suggested by
comparing D. scoticus in dorsal aspect with the recent hemichordate Cephalodiscus in left aspect
(text-fig. 13), on the assumption, already several times presented (e.g. Jefferies 1986, pp. 50fT, 22 Iff.,
and herein below) that the Dexiothetica are descended from a Cephalodiscus- like ancestor which lay
down on its right side. For the stomach is the largest organ in the viscera in the trunk region of
Cephalodiscus and, like the organ filling the posterior prominence of D. scoticus , is situated behind
the pharynx and obliquely right of it in chordate terms (behind and dorsal to it in hemichordate
terms). The posterior boundary of the stomach on the dorsal surface of the internal mould may be
indicated by two grooves, approximately concentric to the highest point of the posterior dorsal
prominence (pi. 2, fig. 1, text-fig. 5; pi. 3, fig. 8, text-fig. 6; pi. 4, figs. 2, 4, 6, text-fig. 7; pi. 5, fig.
1, text-fig. 8). One of these grooves runs leftward and forward from the left anterior margin of the
likely position of the brain, while the other one runs rightwards and forwards from the right anterior
part of the brain. The rest of the non-pharyngeal gut, probably an intestine ending in a rectum,
presumably ran from the stomach to the anus but there is no indication of it in the internal moulds.

On the ventral surface of the internal mould (pi. 1, fig. 1, text-fig. 4), the likely position of the
pharynx plus buccal cavity is, once again, indicated by two weak sinuous grooves. The more
posterior of these, limiting the presumed pharynx posteriorly, runs around the front of the suggested
position of the stomach and then passes leftward and rearward to the region of the gill slit (‘left’
and ‘right’ here refer always to the animal, not to the reader). The groove bounding the pharynx
left anteriorly is very weak in its anterior part, but becomes stronger as it approaches the left border
of the head, and then turns to run leftward and posteriorly parallel to that border. Perhaps this
groove represents the right edge of muscles associated with the left hinge line. A somewhat similar
weak groove runs parallel to the right border of the head and possibly delimits muscle which could
lower the roof of the head on the right - as already explained, the elongate shape of the periproctal
plates, especially left of the anus, suggests, although there was no right hinge line, that the right part
of the head roof could be depressed, presumably by the action of muscles beneath it.

The distinction between buccal cavity and pharynx, as already implied, is uncertain in D. scoticus.
In embryological terms, this distinction ought to coincide with that between endoderm and

660

PALAEONTOLOGY, VOLUME 33

tentacle

arm

terminal knob

protocoel pore
heart

protocoel
buccal diverticulum
right mesocoel

right metacoel

attachment
sucker

bud

stalk nerve
stalk coelom

1 mm

mesocoel pore
right gonad

branchial slit
rectum
stomach

protocoel
pore

mouth

food groove

text-fig. 13. Cephalodiscus after SchepotiefT (1907) and Andersson (1907); (a) sagittal section with the organs

of the right side; (b) external features.

ectoderm. It is fairly certain that the epithelium lining the food groove of D. scoticus would be
ectodermal, on echinoderm analogies, and also almost certain that the gill slit would emerge from
endoderm, and therefore from pharynx. Moreover, in cornutes the distinction between buccal cavity
and pharynx can usually be plausibly recognised (e.g. Jefferies 1986, in Cothurnocystis p. 198;
Scotiaecystis p. 211 ; Ceratocystis p. 217; Nevadaecystis p. 227; Reticulocarpos p. 225; Galliaecystis
p. 235) and the same is sometimes true in mitrates (Jetfferies 1986, Plcicocystites p. 269; Lagynocystis
p. 295). In D. scoticus , however, there is no such visible distinction in the internal moulds. At one
extreme, it seems possible that the mouth identified above was, in fact, the velar mouth, all gut inside
it being endodermal. At the other extreme, it may be that the velar mouth was situated some
distance internal to the identified mouth, but has left no trace on the internal moulds. The difficulty

JEFFERIES: D E N D ROC Y STO 1 D ES AND CHORDATE ANCESTRY

661

is compounded by ambiguities in the hemichordate situation, for Lester ( 1988, p. 116) has recorded
that the region usually called pharynx in Rhabdopleura is ectodermal in origin. It would be
interesting to know whether the same is true in Cephalodiscus where, unlike Rhabdopleura , the walls
of the ‘pharynx' are penetrated by gill slits.

As for the brain of D. scoticus , on cornute and mitrate analogies this would be expected to lie at
the anterior end of the tail. (In mitrates the identification of the brain as being located here is
confirmed by detailed anatomical resemblances with the brain of fishes.) The internal moulds of
plates M1LD and M1RU dorsally and of M,,v ventrally in D. scoticus carry strong grooves which seem
to delimit the anterior border of the brain (pi. 1, figs. 10, 11; pi. 2, fig. 1, text-fig. 5; pi. 4, figs. 2,
4, text-fig. 7 ; pi. 5, fig. I , text-fig. 8). These moulds suggest that the brain would be an approximately
hemispherical structure which made contact with the skeleton on the inside of these plates. It would
be directly comparable, in these respects, with the brain of the most anticrownward cornute
Ceratocystis perneri.

The actual anterior border of the brain on the internal moulds, if it has not been exposed by
dissection, appears asymmetrical, on the right projecting farther forward than on the left. Dissection
of the internal moulds reveals, however, that this asymmetry is largely illusory, at least ventrally (pi.
1, figs. 10, 11) and is due to the way in which an approximately symmetrical brain intersects the
asymmetrical surface of the posterior part of the internal mould of the head. The presumed brain
does not occupy the whole extent of the tail insertion, being confined to the central part of the latter.
Thus the natural mould of the tail insertion is strongly bipartite in appearance. In this way it
resembles the tail insertion of mitrates which, in natural moulds, shows a positive replica of the
bipartite brain, the two portions there corresponding to prosencephalon and deuterencephalon
(Jefferies 1986, Chap. 8). It is unlikely, however, that the bipartition of the tail insertion has this
meaning in D. scoticus for three reasons: (1) in mitrates and cornutes the stereom in contact with
the nervous tissue is characteristically smooth whereas, in D. scoticus , only the central anterior
portion of the tail insertion has a similar smoothness; (2) the left pyriform body of D. scoticus (no
right one, if it existed, has been observed) seems to emerge medially just anterior to the presumed
brain (text-fig. 11 ; pi. 1, figs. 9, 12, 13) and the pyriform body, being the trigeminal ganglion, would
be related to nerves coming from the deuterencephalon, not from the prosencephalon; and (3) as
shown by Cripps (1989b) the highly visible bipartition of the mitrate brain seems to have been
evolved in the crownward parts of the chordate stem lineage since it is incipient in crownward
cornutes and is not developed in anticrownward cornutes such as Ceratocystis.

The features of the internal mould of the tail have been discussed already. They are simple,
however, and can conveniently be repeated here. In the hind tail of the fossils, there is a strand of
rock, circular in cross section, corresponding to a canal in the skeleton, and situated dorsal to the
centre of the hind tail in transverse section. This strand expands forwards conically in the mid-tail
region, to join the approximately cylindrical infilling of the fore tail. As previously suggested, the
infilling of the fore tail probably represents the position of muscles, presumably segmentally divided
like the skeleton, and of a notochord. The muscles of the fore tail would end posteriorly by insertion
into the conical skeleton of the mid tail. The notochord would continue rearwards into the hind tail
and would correspond in position and shape to the central canal of the hind tail i.e. to the
longitudinal strand of rock preserved in this region in the fossils between upper and lower hind-tail
plates. I have argued elsewhere (Jefferies 1986, p. 202) that the notochord of cornutes and mitrates
probably had a blood vessel along its central axis. The same conclusion applies to the presumed
notochord in the hind tail of D. scoticus , since some blood supply to the posterior end of the hind
tail would be needed and it could only have been delivered along the central canal between dorsal
and ventral series of plates.

The pyriform body of D. scoticus has already been mentioned. It has been observed only on the
left side of the brain and only in two specimens. One of these (pi. 1, figs. 12, 13) reveals it plainly
in dorsal aspect inside plate M2LD and the other, less clearly, in ventral aspect (pi. 1, fig. 9) inside
plate Mlv. I have argued elsewhere that the left and right pyriform bodies of cornutes and mitrates
are homologous respectively with the left and right trigeminal ganglia of vertebrates (Jefferies 1986,

662

PALAEONTOLOGY, VOLUME 33

p. 252, etc.) and the same would be true of this left pyriform body in a solute. It could easily be that
a pyriform body existed also on the right in D. scoticus but has not been observed because it was
never coated with calcite in life, i.e. it was never in contact with the nearest plates - in most
specimens of D. scoticus the left pyriform body is as invisible as the possible right one.

ANATOMICAL COMPARISONS
Comparison of Dendrocystoides scoticus with cornutes

Ceratocystis perneri (text-fig. 14) is the only species of cornute known to have a hydropore (Jefferies
1969; 1986, p. 21 3ff ). For this reason it is assigned to the least crownward plesion of the cornutes
and is therefore the most suitable species for comparison with D. scoticus. (In the species
Ceratocystis vizainoi , described by Ubaghs [1987], a hydropore may perhaps exist but has not been
observed.)

ventral prominence
keel

^7 incipient hinge line

I J I U V, UIIC

ventral ossicle

text-fig. 14. Ceratocystis perneri , external anatomy, to show especially the tail, the head openings and the
hinge lines at right and left of the head; (u) dorsal aspect; ( b ) ventral; (c) posterior; (ci) right lateral.

JEFFERIES: D EN D ROCYSTO 1 D ES AND CHORDATE ANCESTRY

663

D. scoticus resembles Ceratocystis perneri in the following respects: (1) it is divided into a head
and a tail; (2) the ventral surface of the head is flatter than the dorsal surface; (3) the mouth is at
anterior right of the head; (4) the anus is at right of the tail on the posterior surface of the head;
(5) the gonopore and hydropore are right of the mid line and the gonopore is farther from the mouth
than is the hydropore (against this, the hydropore and gonopore are anterior in position in D.
scoticus , while the hydropore, if considered with respect to the centre of the head, is clockwise of
the gonopore in D. scoticus , but anticlockwise of it in C. perneri ); (6) there was an opening of the
pharynx at posterior left of the head (in D. scoticus this seems to have been a single branchial
opening or gill slit, whereas in C. perneri there are regularly seven gill slits); (7) the pharynx runs
from anterior right towards the posterior left angle of the head - the extent of the buccal cavity, if
there was one, is unknown in D. scoticus ; (8) the non-pharyngeal gut would have lain posterior to,
and right of, the pharynx, as indicated by the position of the anus in D. scoticus; (9) there was a
brain at the tail insertion with a pyriform body (so far as observed on the left only in D. scoticus ,
but left and right of the brain in C. perneri) ; (10) there are indications that the roof of the head could
be lowered in life - in D. scoticus these indications are the seeming hinge line on the left of the head
and the horizontal elongation of the periproctal plates, while in C. perneri (Jefferies et al. 1987, p.
443) they are a probable hinge line on the right side of the head and accessory gaps, probably filled
with muscle, on the left side of the head (text-fig. 14); ( 1 1) the tail is divided into fore, mid and hind
regions; (12) the more massive skeletal elements of the hind tail are ventral, while the less massive
are dorsal; (13) the skeleton of the fore tail is quadriserial and surrounds a large lumen, and the fore
tail seems to be adapted to wave the mid and hind tail mainly from side to side. These resemblances
are enough to show that Dendrocystoides scoticus can be regarded as built on the same basic plan
as Ceratocystis perneri and other cornutes in general.

Moreover, the distal part of the hind tail was probably permanently and rigidly curved
ventralwards in D. scoticus. Similarly, the distal part of the hind tail was able to curve ventrally in
the cornute Cothurnocystis elizae (Jefferies 1986, p. 202). The same was probably true of the cornute
Protocystites (Jefferies et al. 1987, p. 470). (Concerning C. perneri there is no evidence.) Both in
primitive cornutes and in solutes, the downturned end of the hind tail probably served to grip the
sea bottom during locomotion. It was probably used much like a punt pole in pulling the head
rearwards.

However, there are obvious differences between Dendrocystoides scoticus and cornutes, especially
Ceratocystis perneri. These have partly been mentioned already, but for clarity I here list them all,
even when this involves repetition: (1) D. scoticus, like all other solutes, had a feeding arm which
seems to have been constructed on a standard echinoderm pattern with cover plates and triadic
groups of tube feet - there is no such arm in the cornutes; (2) D. scoticus had an antibrachial process
which is not represented in any cornute (though it is also absent in nearly all other solutes); (3) there
was an anterior prominence, probably containing a gonad, near the anterior end of the head in D.
scoticus - there was no such prominence in any cornute and the gonad in all cornutes was probably
situated near the posterior right corner of the head; (4) the hydropores, on the hydropore node,
were anterior in position and just right of and behind the mouth in D. scoticus where they penetrated
the skeleton to form a madreporic plate - in cornutes the hydropore is only known to exist in C.
perneri where it is slit-shaped and situated near the posterior right of the head; (5) the gonopores
in D. scoticus are anterior, just left of the mouth, multiple and sometimes occluded in C. perneri,
which probably represents the primitive condition for cornutes, it is a single circular hole across a
suture, is never occluded and is situated posteriorly in the head, just right of the anus; (6) D. scoticus
had presumably only one gill slit (unless the gill slits were internal and the external branchial
opening was an atrial opening) whereas C. perneri had seven such slits and other cornutes, except
the most crownward ones where the number is irregular, had more than seven; (7) the head skeleton
of D. scoticus consisted of a large number of plates, mostly irregular, whereas that of C. perneri was
constructed of a smaller number of regular plates (none of the plates can reliably be homologized
between the two species); (8) the brain of D. scoticus did not occupy the whole of the tail insertion,
had no auditory groove or median eye, and was linked, so far as known, to a single pyriform body

664

PALAEONTOLOGY, VOLUME 33

(trigeminal ganglion) on the left, whereas the brain of C. perneri occupied the whole tail insertion,
was associated with a pair of pyriform bodies and with an auditory groove and a median eye; (9)
as to head chambers, there is no sign of a posterior coelom (left epicardium) in D. scoticus , whereas
such seems to have existed in C. perneri and probably in all other cornutes; also it is uncertain
whether or not D. scoticus had a buccal cavity; (10) the alternation of fore-tail plates on right and
left is more irregular in D. scoticus than in C. perneri and the ventral overlap of right and left ventral
plates is less extensive; (11) the mid-tail region of D. scoticus is altogether more irregular than that
of C. perneri or any other cornute - in particular in D. scoticus there is no massive ossicle
corresponding to the stylocone, while in cornutes nothing corresponds to the serial continuity of
hind-tail plates with the right dorsal and right ventral plates of the fore tail in D. scoticus ; (12) the
hind tail of D. scoticus was rigid and its dorsal skeleton was constructed of a single unpaired series
of rather long approximately hemicylindrical plates, not the right and left series of short imbricating
plates seen in C. perneri and other cornutes.

The phylogenetic meaning of these differences from cornutes will be discussed below. Here I shall
only repeat that, despite some differences, the fundamental identity of plan between solutes and
cornutes seems obvious.

Comparison of Dendrocystoides scoticus with Cephalodiscus

The latest common ancestor of echinoderms and chordates, in my opinion, resembled the recent
hemichordate Cephalodiscus but had become dexiothetic, i.e. had been modified to lie on its right
side (Jefferies 1969, 1979, 1986, Chaps. 2 and 9; text-fig. 15 herein). In this respect, a comparison

Cephalodiscus-like ancestor

text-fig. 15. The dexiothetic rotation. In the origin of the Dexiothetica, it is likely that a Cephalodiscus-Y\ke
ancestor fell right-side downwards onto the mud of the sea floor and lost the openings and arms of the primitive
right side in consequence. The stalk was retained and became the chordate tail; an = anus; ax = axocoel
(protocoel); bs = branchial slit; g = gonopore; h = hydropore (left and right mesocoel pores); 1. hyd, r. hyd
= left and right hydrocoels (mesocoels); 1. som, r. som = left and right somatocoels (metacoels).

JEFFERIES: D E N D ROC YSTO I D ES AND CHORDATE ANCESTRY

665

between Dendrocystoides scoticus and Cephalodiscus is illuminating. I discuss the tail first, and then
the head, beginning at the front and working rearwards.

An evolutionary change of orientation, such as dexiothetism, brings great danger of muddle
in the description. I find it necessary to use two equivalent sets of terms. Thus, with reference to
Cephalodiscus , I write of hemichordate-ventral, hemichordate-dorsal, etc., and with reference to D.
scoticus I write of chordate-ventral, etc. The reader will understand this most easily by remembering
that: hemichordate-left = chordate-dorsal; hemichordate-right = chordate-ventral; hemichordate-
dorsal = chordate-right; hemichordate-ventral = chordate-left. Posterior and anterior mean
approximately the same in hemichordates and chordates. In using these terms, I do not imply that
D. scoticus was a chordate, though this was probably true as discussed later. I merely intend to assert
that, for example, the dorsal surface of D. scoticus was homologous with the dorsal surface of Man.
The hemichordate terminology for orientation is the same as that traditionally used in describing
echinoderm larvae.

The tail is the first noteworthy point of comparison. D. scoticus , like all other solutes and also the
cornutes and mitrates, was divided into a tail and a head. In like manner, Cephalodiscus is divided
into a stalk and the rest of the body. This stalk can be markedly lengthened or shortened and is used,
in prehensile fashion, in collaboration with the head shield, for climbing up and down a horny
skeletal process which projects upwards from the coenoecial cup. (In the tubeless hemichordate
Atuharia the stalk is similarly used for clambering over hydroids [Sato 1936].) It is likely that the
tail of D. scoticus was likewise locomotory, serving to pull the head rearwards over the sea floor like
the tail of cornutes and mitrates, probably by side-to-side motion as in primitive cornutes. The stalk
of Cephalodiscus would differ from the tail of D. scoticus in its mode of action, however, since the
latter could bend proximally but was probably in all parts constant in length - this constancy of
length certainly held for the rigid mid- and hind-tail regions but, in addition, the plates of the fore
tail, because of their dorsal and ventral overlaps and sutures at right and left, suggest that the fore
tail could bend right and left but not shorten or lengthen. This in turn suggests that the fore tail
contained an incompressible but flexible notochord, as was probably true in cornutes and mitrates
also. The notochord probably continued rearwards from the fore tail of D. scoticus to fill the canal
between the dorsal and ventral plates of the hind tail, just as the notochord of the cornutes probably
continued posteriorly into the median groove of the ventral hind-tail ossicles. There is no notochord
in the stalk of Cephalodiscus , and its very absence allows the stalk to stretch and shorten.

An attachment sucker exists at the distal end of the stalk of Cephalodiscus (text-fig. 13). By means
of this sucker, the animal fixes itself, at will, to the surface of the horny coenoecium - usually to the
inner face of the coenoecial cup. There is no sign of such a sucker on the end of the tail of D. scoticus ,
though this may be due to ignorance, since the end of the tail has seldom been seen. It is fascinating
to note, however, that Ubaghs (in Ubaghs and Robison 1985) observed in the Middle Cambrian
solute Castericystis vali Ubaghs and Robison 1985 (text-fig. 16) that the adults often carry on their
surface small, and therefore presumably young, individuals of the same species. These small animals
are attached to the adult by distal ends of their tails. This strongly suggests that, at least in juveniles,
the tail of G. vali had an attachment sucker at its distal end, perhaps followed in time by temporary
cementation. (An alternative hypothesis, that the juveniles arose by budding from the associated
adult, is unlikely because the juveniles are disposed seemingly at random on the adult body, whether
on the head or the tail, with no orderly relationship between the size of the juvenile and its position
on the adult.) If the tail of C. vali had a sucker on its end in the juvenile, a comparison with the stalk
of Cephalodiscus is reinforced. I therefore propose that the stalk of hemichordates is homologous
with the tail of solutes, and also with the tail of chordates in general. I first implied that the stalk
of Cephalodiscus was homologous with the tail of cornutes (then called by me the stem) in Jefferies
(1969). Gradually I abandoned this view for lack of evidence and in my book of 1986 I did not even
mention it as a possibility. Ubaghs’ interesting observations now make it likely again. I therefore
now agree with the percipient remark of Eaton (1970, p. 977) when he said that: '...it is highly
probable that the tail [of cornutes and mitrates] is a specialized version of the pterobranch
peduncle.’ Henceforth I shall stress this presumed homology by referring to the stalk of

666

PALAEONTOLOGY, VOLUME 33

arm

text-fig. 16. Castericystis vali Ubaghs and Rob-
ison 1985 -a solute from the Middle Cambrian
Marjum Formation of Utah (USA), (a) Recon-
struction redrawn after fig. 1 of Ubaghs and
Robison (1985) -the reconstruction is in an ana-
tomically possible position for a corpse, but the
hind tail is in right aspect, the head in ventral
aspect and the arm in left aspect; ( b ) drawing
traced from a photograph (Ubaghs and Robison
1985, fig. 11/5) to show the presence of young
individuals of the species fixed to a presumed adult
by the ends of their tails ( x 3 5).

hemichordates as the tail. In this 1 follow the usage of Burdon-Jones (1952) when he described the
homologous organ of post-larval enteropneusts.

The planes of bilateral symmetry in the tails of D. scoticus and Cephalodiscus do not correspond
to each other, if D. scoticus is supposed to have undergone a dexiothetic rotation : in Cephalodiscus
the plane of symmetry in the tail is sagittal, which ought to correspond to chordate-horizontal, but
the plane of symmetry in the tail of D. scoticus, admittedly imperfect in the mid and fore tail, is
vertical. The position of the plane of symmetry has therefore rotated through 90°, either clockwise
or anticlockwise when seen from behind, from hemichordate-sagittal to chordate-sagittal, thus
counteracting the dexiothetic rotation.

The feeding arm of D. scoticus , with its ventrally-opening cover plates and its presumed
ambulacral groove, water vascular canal and tube feet in probable triads, projects from the head
at the opposite pole to the tail. The obviously comparable structure in Cephalodiscus would be part
of the arm apparatus. More exactly, the best comparison would be with a single arm of the left side
of Cephalodiscus. For the feeding arm of D. scoticus is presumably homologous with part of the
water vascular system of echinoderms, and this system, in turn, is homologous with the left arm
apparatus (left mesocoel = left hydrocoel) of Cephalodiscus (Jefferies 1986, p. 318).

There are difficulties in this comparison, however. For in a Cephalodiscus seen from the left side,
the left arms present their food grooves toward the observer, whereas the food groove of D. scoticus
in dorsal aspect faced away from the observer. If Cephalodiscus represents the primitive condition,
therefore, the feeding arm of D. scoticus has rotated through 180° so that the food groove opens
not towards chordate-dorsal but towards chordate-ventral. The right and left mesocoels of

JEFFERIES: D EN D ROC Y STO I D ES AND CHORDATE ANCESTRY

667

Cephalodiscus open to the outside each by a mesocoel pore. In common with echinoderms, D.
scoticus had no trace of any such pore so perhaps, as in echinoderms, the left mesocoel
communicated with the axocoel ( = protocoel) by a stone canal while the right mesocoel, together
with its pore, had disappeared. Rehkamper and Welsch (1988) have stressed the fine-structural
similarities between the stone canal of echinoderms and the mesocoel pores of Cephalodiscus. These
resemblances suggest to me that perhaps the stone canal is homologous with the left mesocoel pore
of hemichordates but has come to open into the protocoel ( = axocoel) instead of direct to the
outside.

The mouth of Cephalodiscus opens, in a hemichordate-ventral position, behind the head shield
and faces posteriorly, toward the tail. The mouth of D. scoticus , on the other hand, is anterior and
at chordate-right. In passing from the Cephalodiscus condition to the D. scoticus condition,
therefore, the mouth would have rotated through some 160°, clockwise in chordate-dorsal aspect,
about a vertical axis. Such a rotation would shift the left protocoel pore (equivalent to the
hydropore of D. scoticus and the echinoderms) from a position anterior to the mouth to one where
it lay, as actually observed in D. scoticus , chordate-right of the mouth. This suggests that, in
addition to the dexiothetic rotation, D. scoticus differs from Cephalodiscus in having undergone a
clockwise rotation of the anterior part of the head about a chordate-vertical axis. I shall refer to this
additional rotation as anteriorisation of the mouth. The same rotation would perhaps explain
another difference between Cephalodiscus and D. scoticus , namely the absence from the latter of any
externally distinguishable part of the body corresponding to the head shield (protosome).
Presumably, D. scoticus had an equivalent of the protosome, in that it would have possessed a
protocoel (axocoel) buried inside the head and opening upwards at the hydropore. The burial of the
protosome in the head might easily result from anteriorization of the mouth since the mouth, in
moving chordate-rightwards and anteriorly, would press on the protosome. Such burial of the
protosome is presumably a derived feature which D. scoticus shared with echinoderms and also with
definite chordates, since the cornute Ceratocystis had a hydropore but no externally distinct
protosomal body region. The heart and pericardium of D. scoticus would probably be near the
hydropore, like the equivalent head process of the axial organ and dorsal sac of echinoderms. These
structures are equivalent to the heart and pericardium in the protocoel of hemichordates, which lie
anterior to the mouth in Cephalodiscus. The rotation involved in anteriorizing the mouth of C.
scoticus would have shifted the heart and pericardium to their presumed new positions also.

As for the gonad, I have argued above that the anterior prominence of D. scoticus contained a
gonad which issued by a group of gonopores at the chordate-anterior-right of the prominence. The
gonad, being chordate-dorsal in position, presumably corresponds to the left gonad of Cephalodiscus
and the gonopores to the left gonopore. In general position at the anterior end of the body farthest
from the tail, the gonad of D. scoticus is comparable with the left gonad of Cephalodiscus and
contrasts in both these animals with the cornute situation where the gonad was at chordate-right
and posterior in the head. However, there are also differences between Cephalodiscus and D. scoticus
since in the latter the gonad and gonopore lay chordate-left of the arm whereas in Cephalodiscus
they were at chordate-right of the arms (dorsal to them in hemichordate terms). In passing from
Cephalodiscus to D. scoticus , therefore, the gonopore and gonad would move chordate-leftwards.
I cannot decide whether this was part of the clockwise rotation of the anterior part of the head
implied by the anteriorization of the mouth, or whether it was an independent process.

The embryological origin of the gonad of echinoderms is interesting here: the primordial germ
cells are said to arise from the wall of the left somatocoel in crinoids (MacBride 1914, p. 552), in
asteroids (Gemmill 1914) in crinoids (Hyman 1955, p. 502) and in ophiuroids (Hyman 1955, p. 636).
(No comparable statement can be made for the holothuroids, perhaps because the distinction
between left and right somatocoels breaks down at an early stage.) It seems likely, therefore, that
the gonad or gonads of echinoderms, like that of D. scoticus , correspond to the left gonad of
Cephalodiscus.

The pharynx of D. scoticus seems to have been elongate, stretching from the mouth or velar
mouth at anterior chordate-right to the gill slit at posterior chordate-left, with presumably an

668

PALAEONTOLOGY, VOLUME 33

oesophageal opening posterior and right of it somewhere along its course. (A dubious point, as
already mentioned, is whether D. scoticus had a buccal cavity.) The position of the pharynx in D.
scoticus is broadly the same as in Cephalodiscus seen from the left, but the two pharynges differ in
shape. The differences correspond to extension of the pharynx chordate-leftwards and rearwards
and chordate-rightwards and forwards in D. scoticus, with the loss of a gill slit (the hemichordate
right gill slit) and the migration of the remaining gill slit to a chordate-left-posterior position. In
these respects, the pharynx of D. scoticus was more cornute-like than that of Cephalodiscus.

Concerning the non-pharyngeal gut, the stomach of D. scoticus probably coincided in position
with the posterior dorsal and ventral prominences. This position, near the centre of the head, is
comparable with the stomach of Cephalodiscus as seen from the left. The position of the anus in
Cephalodiscus and D. scoticus is comparable in being, in both animals, chordate-right of the tail, but
it differs in being more posterior in D. scoticus, nearer to the proximal end of the tail. This difference
in the position of the anus implies that the intestine and rectum must also have been differently
disposed, but nothing precise can be said about this since there is no evidence in D. scoticus of the
course of the gut between stomach and anus.

Not much can be said about the positions of the coeloms in D. scoticus, compared with
Cephalodiscus. If D. scoticus has undergone a dexiothetic rotation with respect to Cephalodiscus,
then the right somatocoel ( = right metacoel) would be downward and the left somatocoel upward,
as is primitively the case in echinoderms (Jefferies 1986, p. 51) and as seems also to have been true
in cornutes and mitrates (Jefferies 1986, p. 198ff. in discussing left and right anterior coeloms = left
and right somatocoels). The mesentery between the two somatocoels would thus be horizontal. The
left hydrocoel (= left mesocoel) would be located in the feeding arm while the right hydrocoel
(= right mesocoel) would have disappeared. And the axocoel (= protocoel), buried inside the
head but single and unpaired as in hemichordates, would open upwards at the hydropore, just right
of the mouth and right of the arm base, the hemichordate right protocoel pore having
disappeared.

Finally as for size, D. scoticus, like all other solutes, was about 30 mm in head length and
therefore much bigger than Cephalodiscus, whose greatest head length is about 5 mm.

To summarize, if D. scoticus is compared with a Cephalodiscus lying on its right side, and if the
latter is assumed, in all respects, to represent the primitive condition, then the following changes
would have occurred in passing from Cephalodiscus to D. scoticus.

(1) The evolution of a calcite skeleton and the formation of specialized plates in some regions,
especially in the tail, the tail insertion, the arm and the arm insertion.

(2) The evolution of a vertical plane of symmetry in the tail and probably the origin of the
notochord so that locomotion took place by lateral bending of the tail, rather than by shortening
and lengthening of it; loss of the power of budding, which had been located in Cephalodiscus at the
distal end of the tail.

(3) Loss of all the hemichordate-right arm apparatus and all but one arm of the hemichordate-
left arm apparatus; rotation of the remaining left arm through 180° so that the food groove opened
chordate-downwards; development of a triadic arrangement of tube feet regularly related to the
cover plates; loss of the hemichordate-left and hemichordate-right mesocoel pores. (Perhaps the
hemichordate left mesocoel pore became the stone canal of dexiothetes, which in extant forms is
retained in echinoderms but absent in chordates.)

(4) Anteriorization of the mouth, involving some degree of rotation of the anterior part of the
body clockwise about a chordate-vertical axis, so that the hydropore came to lie chordate-right of
the mouth rather than anterior to it; loss of the head shield (protosome) as an externally distinct
body region, by burial of the protocoel (axocoel) in the body chordate-right of the mouth (this
burial may have resulted from the anteriorization of the mouth, that is from the clockwise rotation
of the anterior part of the head which such movement of the mouth implies); loss of the
hemichordate-right, chordate- ventral protocoel pore; perhaps the evolution of a stone canal linking
the hemichordate-left hydrocoel (= left mesocoel) with the axocoel (= protocoel).

(5) Loss of the hemichordate-right gonad and right gonopore; migration of the hemichordate-left

JEFFERIES: D EN D ROC Y STO I D ES AND CHORDATE ANCESTRY

669

gonad and left gonopore chordate-leftwards, so as to lie left of the arm when D. scoticus is viewed
in dorsal aspect.

(6) Movement of the hemichordate-left gill slit from a chordate-anterior-central to a chordate-
left-posterior position, with corresponding elongation of the pharynx.

(7) Migration of the anus from chordate-anterior-right to chordate-posterior-right in the head.

(8) Evolution of a brain and a left trigeminal ganglion at the anterior end of the tail.

(9) Loss of the hemichordate-right hydrocoel (= right mesocoel); repositioning of the
hemichordate-right and -left metacoels, as a result of dexiothetic rotation of the animal, to become
respectively chordate-ventral and chordate-dorsal.

(10) Increase in size.

Not all these changes would actually have happened since Cephalodiscus is probably not in all
ways more primitive than D. scoticus.

Comparison of Cephalodiscus with Kinzercystis

The ‘eocrinoid’ Kinzercystis (text-fig. 17), from the Lower Cambrian Kinzers Formation of
Pennsylvania in the U.S.A., has been well described and reconstructed by Sprinkle (1973) and Paul
and Smith (1984). The latter authors regard it as belonging to the crown group of the echinoderms,
seeing it as a primitive member of the Pelmatozoa. I choose it for comparison here for three reasons :
because it is well studied; because, unlike most primitive echinoderms, the hydropore is separate
from the gonopore; and because I wish to simplify the phylogenetic argument by concentrating on
one quinqueradiate echinoderm only.

Kinzercystis was a small animal about 30 mm in height and 25 mm in greatest width. It was an
inverted truncated cone in shape and plated all over with calcite plates. The wide end of the cone,
which would have faced upward in life, was the oral surface while the truncation of the cone,
downward in life, was developed as a circular attachment area by which the animal could fix itself
to objects on the sea floor.

The oral surface was approximately circular with an elongate mouth near the centre, roofed over
by cover plates. Three ambulacral grooves radiate from the mouth, and two of these split into two
not far from the mouth. This produces five food grooves in all, with a 2 + 1+2 arrangement
suggesting a vestigial triradiality. The undivided food groove is conventionally labelled A and the

text-fig. 17. The eocrinoid Kinzercystis durhami
Sprinkle 1973 from the Lower Cambrian Kinzers
Formation of Pennsylvania (USA). Copied from

Paul and Smith (1984, fig. 8) and relabelled.

670

PALAEONTOLOGY, VOLUME 33

others, in clockwise direction seen from above, are numbered B, C, D and E. Conventionally the
A food groove is taken as anterior and the C-D interradius as posterior. In the CD interradius,
just behind and to the right of the mouth when seen from above, there are two openings very close
to each other. The opening nearer to the mouth is spout-shaped and is identified by Paul and Smith
as the hydropore while the other opening, guarded by a pyramid of small spike-shaped plates and
therefore intermittently closable in life, is identified by them as the gonopore. (Sprinkle [1973, p. 74]
labelled hydropore and gonopore conversely, but Paul and Smith’s view is preferable because it
agrees with the situation in cystoids and because it is reasonable to suppose that a gonopore could
be closed intermittently.) Also in the C-D interradius, near the edge of the oral surface and left of
the mouth as seen from above, was the anus guarded by an anal pyramid.

Each ambulacral groove rested on two series of flooring plates and was covered on each side by
a plated lip the plates of these lips are called cover plates. On either side of each ambulacral
groove, alternating in position, was a series of flexible plated appendages called brachioles which
Paul and Smith interpret, for good reasons, as plated tube feet. The interambulacra were covered
with large plates which, at their sutures, carried openings called epispires. These may have had a
respiratory function, each one marking the position of a thin-called external gill or papula.

The aboral suface of Kinzercystis was in two parts - the distal attachment area and the rest. The
attachment area was approximately circular and plated with polygonal plates, whereas the
remainder of the aboral surface, vaguely divided into an upper wider theca and a lower almost
cylindrical holdfast, is covered with imbricating plates (Sprinkle, 1973, p.70, fig. 22). The
imbrication is such that, seen from outside, lower plates overlap the lower parts of higher ones. This
imbrication suggests that the aboral surface was able to elongate and shorten.

Kinzercystis can be interpreted, in its main anatomical features, by comparison with crinoids (for
references see Jefferies 1986, pp. 47ff.). It is highly probable that, at least in the juvenile stages, a
right and left somatocoel would have been present, with the right somatocoel located toward the
attachment and the left somatocoel lying orally. Much of the gut would have been situated between
the two somatocoels. The ambulacral grooves would have been occupied by radial water vessels
with brachioles representing the tube feet, and this water vascular system would correspond to the
left hydrocoel of the larva. The hydropore would be the opening of the axocoel and would
communicate with the water vascular system by a stone canal. The attachment area would be
homologous with the axocoel attachment area of the larva and thus with the ventral suface of the
head shield of Cephalodiscus. The gonad, opening at the gonopore, would probably be homologous
with the left gonad of Cephalodiscus as suggested above.

In comparing Kinzercystis with Cephalodiscus I shall assume, as when discussing Dendrocystoides
scoticus , that Cephalodiscus in most ways represents the primitive condition. Some of the differences
between Cephalodiscus and Kinzercystis are shared by the latter with D. scoticus and can be seen as
synapomorphies of D. scoticus and Kinzercystis in a three-taxon comparison. Such include:

(1) The dexiothetic rotation, with the oral surface of Kinzercystis corresponding to the dorsal
surface of D. scoticus and to the left side of Cephalodiscus - other results of dexiothetism include the
absence of the hemichordate-right arm system and right mesocoel pore and of the hemichordate-
right gill slit, gonad and gonopore.

(2) The calcitic skeleton.

(3) The disappearance of the protosome as an externally distinct body region, presumably by
becoming buried in the body though still opening outwards by the hydropore.

(4) The disappearance of the right mesocoel pore, presumably because the water vascular system
communicated with the protosome by a stone canal inside the body.

All four of these advanced resemblances between Kinzercystis and D. scoticus can be seen as
autapomorphies of the Dexiothetica.

Some features which differentiate Kinzercystis from Cephalodiscus are not shared with D. scoticus.
These include:

(1) The absence of a tail.

(2) Attachment by the middle of the aboral (chordate-ventral) surface to some external object.

JEFFERIES: D E N D RO C YSTO / D ES AND CHORDATE ANCESTRY

671

(3) The position of the mouth at the centre of the oral surface.

(4) The recumbent position of the water vascular system on the surface of the body and its
division into five branches with a distinct 2+1+2 arrangment.

(5) The absence of any gill slit.

All five of these distinguishing features of Kinzercystis in the three-taxon comparison can be seen
as autapomorphies of the crown echinoderms, except for attachment by the aboral surface to an
outside object which is perhaps an autapomorphy of the pelmatozoans.

THE SYSTEMATIC POSITION OF D E N D RO C Y STO I D ES

The analysis so far shows that D. scoticus was a member of the Dexiothetica, that it did not belong
to the echinoderm crown group, nor to the chordate crown group, nor to the crownward members
of the chordate stem group known as Cornuta. The methodology for placing it more exactly within
the Dexiothetica is clear: if it lacked some autapomorphy of the Dexiothetica, it belongs in the stem
group of that group; if it shared a synapomorphy with the chordates, particularly with the stem-
group chordates of the Cornuta, it belongs in the chordate stem group; if it shared a synapomorphy
with the echinoderms, it would belong in the stem group of the echinoderms; and if none of these
conditions applied, it would have to remain in the nodal group of the Dexiothetica along with the
latest common ancestor of chordates and echinoderms.

There are, in fact, many advanced features connecting D. scoticus with cornutes, which are not
present in such a primitive echinoderm as Kinzercystis for example. In the tail these comprise: (1)
the division into fore, mid and hind tail; (2) the quadriserial plating of the fore tail; and (3) the large
lumen of the fore tail and its adaptation for bending from side to side, suggesting the presence of
a notochord. Just anterior to the tail, D. scoticus shows advanced resemblances to cornutes, and to
chordates in general, in having a brain and at least one pyriform body. Elsewhere in the head,
advanced resemblances to cornutes include the position of the gill slit at posterior left and the
corresponding leftward and posterior elongation of the pharynx.

Unfortunately, all these features relate either to the tail or to the gill slit, and both these organs
are absent in echinoderms, probably having been lost. Loss of complex structures is intrinsically more
likely than their origin. All these advanced resemblances are therefore dubious as synapomorphies
between cornutes and D. scoticus , because of the possibility that the latest common ancestor of
chordates and echinoderms may have had them but that they were then lost in the echinoderm stem
lineage when the tail and the gill slit were lost.

To eliminate this possibility, or perhaps to confirm it, will require study of the echinoderm stem
group, to which, up till now, only one group of fossils has been assigned. These are the
Helicoplacoidea (text-fig. 18) which Paul and Smith (1984) place as stem-group echinoderms on
account of their triradial ambulacral system incorporated in the wall of the body, the mouth being,
as Derstler (1981) suggested, at the junction of three ambulacral rays. The helicoplacoids, so far as

text-fig. 18. The helicoplacoid Helicoplacus
curtisi Durham and Caster 1963. Recon-
struction copied from Durham and Caster
(1966, fig. 109a). The presence of spikes over
the whole surface of the body suggests that
the animal was totally buried in normal life,
presumably with the long axis horizontal.

About x 2.

672

PALAEONTOLOGY, VOLUME 33

known, probably constitute a monophyletic group because of their strongly spiral skeleton. They
have no gill slit or tail and these absences confirm their position as echinoderms. Unlike Paul and
Smith, I believe that helicoplacoids probably lived entirely buried in the sea bottom, burrowing
earthworm-fashion with the long axis horizontal, by dilating or extending different parts of the
obviously flexible body. Their spiral structure suggests that they rotated about their long axis as
they moved. My reason for suggesting that they lived entirely buried is the presence of low spikes
on all parts of the external surface in Helicoplacus curtisi (text-fig. 18), H. gilberti (Durham and
Caster 1966, fig. 107), Waucobella nelsoni (Durham 1967, pi. 14, fig. 5), and Polyplacus kilmeri
(Durham 1967, pi. 14, fig. 1). These spikes look like sediment-gripping structures which could only
have functioned when the whole surface of the animal made contact with mud. The helicoplacoids
suggest that a triradiate ambulacral system recumbent on the body surface, rather than borne on
arms, was evolved in the echinoderm stem lineage and that later this triradiate system gave rise to
the quinqueradiate system with a 2+1+2 structure. An important conclusion is that the
recumbency of the ambulacral system represents an autapomorphy of echinoderms, not shared by
Cephalodiscus with its many arms nor by solutes with their single arm.

The Cincta may also belong to the echinoderm stem group. These fossils need restudy since
Ubaghs’ (1968) comprehensive treatment of them is now rather old. They can be exemplified by
Trochocystites (text-fig. 19) which is shaped much like a tennis racquet with a theca and an
appendage (stele). The animal lay with one face of the theca on the sea bottom and the other
upwards. The theca is surrounded by a frame of marginal plates. The appendage or stele, at least
proximally, would be stiff in a right-left direction, but perhaps able to bend up and down. In its
plate arrangement it resembles an extension of the frame - the major series of plates, at right and

operculum

text-fig. 19. Thecinctan Trochocystites bohemicus Barrande 1887. Reconstruction redrawn and modified from
Ubaghs (1968, fig. 363): (a) dorsal aspect, to show anus at anterior left; ( b ) ventral aspect. x2-5.

JEFFERIES: D EN D ROC Y STO 1 D ES AND CHORDATE ANCESTRY

673

left, resemble and pass into frame plates, whereas intercalated dorsal and ventral plates in the stele
resemble dorsal and ventral integument plates in the theca, and probably represent serial
homologues of these integument plates. Opposite the stele, two openings have so far been recorded
in the literature. The opening more to the right, when the stele is towards the observer, had a pair
of ambulacral grooves converging upon it, these grooves coming to it along the marginal frame
from right and left. This opening was therefore the mouth which was thus at the centre of a biradial
recumbent water vascular system. The two rays presumably correspond to the rays B/C and D/E
of Kinzercystis and other quinqueradiate echinoderms. The other opening recorded in the literature
was covered by a plate, the operculum, hinged at the top. This opening is conventionally
interpreted as the anus. However, there is a third opening, not yet recorded in the literature, just
to the left of the operculum (Prof. K. Sdzuy, personal communication). This third opening, which
is visible in Ubaghs (1968, fig. 367/1), is guarded by a pyramid of plates and is likely to be the true
anus. And, in that case, the operculate opening, which in its structure suggests an outlet valve, is
probably a gill slit. The recumbency of the water vascular system of cinctans, if interpreted as a
synapomorphy with the recumbent water vascular system of helicoplacoids, suggests that Cincta
are stem-group echinoderms. This is confirmed by the biradiality of the water vascular system in
Cincta, since the 2+1+2 pattern of water vascular system, can be construed as fundamentally
biradial.

However, the Cincta are probably less crownward than helicoplacoids in the echinoderm stem
group. This is for several reasons: (1) they retain a gill slit; (2) they have a biradial, rather than a
triradial water vascular system; and (3) the mouth is in the hemichordate-sagittal plane, as in
Cephalodiscus - noj in the centre of the chordate-dorsal surface as in crown-group echinoderms
and, probably, in helicoplacoids. The position of the gill slit, anterior in the body and near to the
mouth, resembles that of the left gill slit in Cephalodiscus and is therefore probably primitive for
echinoderms. This means that the gill slit of the latest common ancestor of chordates and
echinoderms would also probably have been anterior in position. And in that case, the left posterior
position of the gill slit of D. scoticus is probably a synapomorphy with cornutes. If so, then D.
scoticus is a stem-group chordate, less crownward than any cornute.

The stele of Cincta is difficult to interpret and there are two main possibilities: (1) its position in
the animal, as far as possible away from the mouth and therefore posterior, and the fact that it is
biserial in its major plating, might suggest that it is homologous with the tail of hemichordates,
solutes and chordates; but, (2) the fact that the two series of plates are at right and left, not as in
solutes dorsal and ventral, and the fact that in some cinctans the two series are continuous with the
marginal plates of the frame (e.g. Trochocystites , Gyrocystis), might suggest that the stele arose
within the Cincta, after the evolution of a marginal frame, as an extension of that frame, and that
therefore it is not homologous with a tail. Further work is needed to establish which of these two
alternatives is nearer the truth.

If the appendage of Cincta is genuinely a tail, then the chordate nature of D. scoticus is almost
certain since the species, would then share a large number of advanced tail features with the Cornuta
which are absent in the Cincta. Such include: the brain; the left pyriform body; the fore, mid and
hind regions of the tail; the large lumen of the fore tail, adapted for right-left bending and probably
containing muscles and a notochord; and the ventral series of plates in the hind tail, corresponding
to the ventral hind-tail ossicles of cornutes. At the moment, however, I am inclined to think that
the stele of cincta is an autapomorphy of that group. If so, then the chordate nature of D. scoticus
remains doubtful, based, as it is, on nothing but the posterior left position of the gill slit.

Whether or not D. scoticus is a chordate, it is almost certainly more closely related to the chordate
crown group than is Castericystis. The conclusion is based on resemblances to Ceratocystis (a
definite stem-group chordate) which Castericystis does not share: i.e., the distinctly tripartite tail of
D. scoticus , with its stylocone-like mid tail and quadriserial fore tail. Moreover, some solutes
resemble Ceratocystis still more closely. Thus Iowacystis (Caster 1968, p. S620) has the fore tail
sharply distinct from the rest of the tail and the dorsal surface of the head is formed of large plates,
in both respects like Ceratocystis. And Belemnocystites is still more like Ceratocystis in having a big-

674

PALAEONTOLOGY, VOLUME 33

plated ventral (as well as dorsal) surface to the head and a very short arm (Caster 1968, p. S623)
whereas cornutes have no arm at all. All this suggests that these four solutes are successively closer
related to the chordate crown group in the sequence: Castericystis , Dendrocystoides, Iowacystis,
Belemnocystites. I summarize the results of these arguments in a cladogram (text-fig. 20).

text-fig. 20. Cladogram of the deuterostomes, to show the phylogenetic position of the solutes within the
Dexiothetica : Be = Belemnocystites \ Ca = Castericystis; Cc = Ceratocystis\ Dc = Dendrocystoides ; He =
Helicoplacus ; Io = Iowacystis ; Ki = Kinzercystis ; Tr = Trochocystites.

As to the timing of the phylogenetic split between echinoderms and chordates, I formerly believed
that this happened in the earliest Cambrian (Jefferies 1979, p. 475). This view was based on the
assumption that the calcitic skeleton is homologous between echinoderms and chordates and on the
further assumption that skeletons in general arose at the base of the Cambrian. The latter belief now
seems to be mistaken, however, since Gehling (1987) has described an edrioasteroid-like fossil
(Arkarua adami ) from the late Precambrian Pound Quartzite of Australia. This form has a 2+ 1 +2
symmetry and probably possessed a lightly built skeleton, since it seems to show a marginal ring of
plates. If Arkarua is correctly regarded as a quinqueradiate echinoderm, as seems likely, then the
split between echinoderms and chordates must be older than it. Perhaps the separation occurred in
the late Precambrian. This raises the interesting possibility that solutes will be found in the late
Precambrian.

CONCLUSIONS

The main conclusions of this paper are as follows.

(1) Dexiothetism. Text-figure 21 summarizes the evidence for dexiothetism on the basis of the five
animals particularly discussed above. In the diagrams all five are shown in chordate-dorsal aspect,
meaning that Cephalodiscus is seen from the left while the other animals are seen from above. The
view from above is called the dorsal aspect in Dendrocystoides and Ceratocystis , the upper surface
in Trochocystites and the oral, or sometimes ‘ventral’ surface in quinqueradiate echinoderms such
as Kinzercystis.

(duu)

JEFFERIES: DENDROCYSTOIDES AND CHORDATE ANCESTRY

675

text-fig. 21. The topological evidence for dexiothetism as shown by five deuterostomes in chordate-dorsal aspect, i.e. Cephalodiscus seen from
the left and four dexiothetes seen from above. Tabulated below are the sequences of body openings and appendage passed in moving clockwise
around each sketch, beginning at the mouth, m = mouth; h = hydropore (left protocoel pore of Cephalodiscus ; mp = mesocoel pore (of
Cephalodiscus only); g = gonopore; an = anus; t = tail (stalk of Cephalodiscus) ; bs = branchial slit; st = cinctan stele.

676

PALAEONTOLOGY, VOLUME 33

A clockwise journey around each of the five sketches, beginning at the mouth, produces the
tabular statements below the sketches, in which the body openings and the tail are listed in sequence.
Cephalodiseus and Ceratocystis have identical sequences, except for the absence of the mesocoel
pore in Ceratocystis. Dendrocystoides is the same as Ceratocystis except that the gonopore is out of
sequence. Trochocystites agrees in sequence with Ceratocystis , although with no hydropore,
gonopore or tail the sequence has only three members and the random probability of agreement is,
in any case, an unimpressive 05. With Kinzercystis , without tail or gill slit, the random probability
for four items is 1 / !3 = 0-17. This in itself is no more than suggestive of dexiothetism, but for
quinqueradiate echinoderms in general the evidence for dexiothetism does not depend on the
topology of the adults, being firmly based on embryology. The major conclusion from text-figure
21 and from embryology is that all the five animals except Cephalodiseus have undergone the
dexiothetic rotation in their ancestry.

(2) Failure of the aulacophore theory. The homology of the tail of Dendrocystoides with that of
cornutes is confirmed by evidence of the brain and left trigeminal ganglion (see especially pi. 1, figs.
12, 13) at the proximal end of the tail in Dendrocystoides , as well as by the division of its tail into
fore, mid and hind tail and the quadriserial arrangement of the fore-tail plates. If the tail of solutes
is homologous with that of cornutes, then Ubaghs’ interpretation of the tail as an aulacophore or
feeding-arm cannot be correct. This conclusion is confirmed by difficulties in interpreting the mitrate
tail as a feeding arm since one of the leading advocates of the aulacophore theory now accepts that
the hind-tail plates of mitrates could not open at the mid line in life (Parsley 1988).

(3) Homology of the hemichordate stalk and the chordate tail. Evidence that the early solute
Castericystis could attach when young to hard surfaces by the distal end of its tail strongly suggests
that this tail was homologous with the stalk of Cephalodiseus (which has a sucker at the distal end).
If the pterobranch stalk is homologous with the chordate tail, and if echinoderms are the extant
sister group of chordates, then the echinoderms must have lost the tail in their early phylogeny.

(4) Similarity of body plan in cornutes and solutes. If the solutes are orientated as here suggested
with the arm at anterior right, then their basic body plan is seen to resemble that of cornutes with
the mouth at anterior right, the anus posterior in the head and right of the tail, the gill slit at
posterior left in the head, and the flatter surface of the head ventral.

(5) The phvlogenetic position of the solutes. Solutes had a branchial slit (at least in
Dendrocystoides), a water vascular system and a tail and in all these respects resembled
hemichordates (in which the water vascular system is referred to as the left mesocoel). They were,
however, dexiothetes, since they show signs of dexiothetism and had a calcite skeleton. They differed
from the stem-group chordates called cornutes by retaining the water vascular system. And they
differed from recognized stem-group echinoderms (except possibly the cinctans) in retaining the tail.
The latest common ancestor of echinoderms and chordates would have had a tail, a branchial slit
and an arm and therefore, if it were ever found, would be placed in the Soluta. The solutes,
therefore, straddled the phylogenetic separation between echinoderms and chordates and so, to use
Hennig’s term, were an invalid stem group of the Dexiothetica. Further work will be needed to put
the known solutes in their correct stem groups.

Acknowledgements. I am grateful to my colleague Andrew Smith for telling me, forcibly, that Dendrocystoides
scoticus was the next animal to study. I am also grateful to a small group of devoted Scottish collectors - Mrs
Eliza Gray and her daughters Alice, Agnes and Edith Gray, and Dr Archie Lamont and Mr James Begg - who,
between 1880 and about 1940, accumulated the material on which this study is based (Cleevely, Tripp and
Howells, 1989). Dr Keith Ingham kindly gave me access to the collections of the Hunterian Museum. Glasgow,
and David Lewis was a constant help at the British Museum (Natural History) and made up the plates. I thank
my students Tony Cripps, Paul Daley, Fritz Friedrich, and Ian Woods for their moral support and their patient
attempts to educate me.

JEFFERIES: D EN D ROC Y STO I D ES AND CHORDATE ANCESTRY

677

REFERENCES

andersson, k. a. 1907. Die Pterobranchier der schwedischen Siidpolarexpedition 1901-1913, nebst
Bemerkungen liber Rhabdopleura normani Allman. Wissenschaftliche Ergebnisse der schwedischen Siidpolar-
expedition , 5 (Zoologie), 1-122.

ax, p. 1984. Das phylogenetische System. Gustav Fischer, Stuttgart.

1987. The phylogenetic system. Wiley, Chichester and New York.

barrande, j. 1867. Systeme silurien du centre de la Boheme , Vol. 3. Barrande, Prague.

- 1887. Systeme silurien du centre de la Boheme. Vol. 7. Classe des echinodermes, ordre des cystidees. Rivnac,
Prague and Gerhard, Leipzig.

bassler, r. s. 1938. Pelmatozoa Palaeozoica (Generum et Genotyporum Index et Bibliographia). Fossilium
Catalogus I Animalia , 83, 1-194.

and moody, m. w. 1943. Bibliographic and faunal index of Paleozoic pelmatozoan echinoderms. Special

Papers , Geological Society of America , 45, 1-733.

bateson, w. 1886. The ancestry of the Chordata. Quarterly Journal of Microscopical Sciences, 26, 535-571.
bather, f. a. 1889. The Cystoidea of Bohemia. Nature 40, 267-270.

— 1913. Caradocian Cystidea from Girvan. Transactions of the Royal Society of Edinburgh, 49, 359-529.

1925. Cothurnocystis : a study in adaptation. Palaontologische Zeitschrift, 7, 1-15.

1928. Dendrocystis in North America. Bulletin of the Geological Survey of Canada, 49, 5-8.

1929. Imagination and fossils. Proceedings of the South-Western Naturalists' Union, 2, 27-41.

begg, j. l. 1947. Some new fossils from the Girvan District. Transactions of the Geological Society of Glasgow,
21, 29-47.

burdon jones, c. 1952. Development and biology of the larva of Saccoglossus horsti (Enteropneusta).

Philosophical Transactions of the Royal Society of London, Series B, 236, 553-590.
caster, K. E. 1968. Homoiostelea. S581-S623. In moore, r. c. (ed.). Treatise on invertebrate paleontology. Part
S. Echinodermata 1, (2). Geological Society of America and University of Kansas, Lawrence, Kansas.
chauvel, j. 1941. Recherches sur les cystoi'des et les carpoi'des armoricains. Memoires de !a Societe geologique
et miner alogique de Bretagne, 5, 1-286.

cleevely, r. j., tripp, r. and howells, y. 1989. Mrs Robert Gray (1831-1924): a passion for fossils. Bulletin
of the British Museum ( Natural History), (Historical Series), 17, 167-258.
craske, a. j. and Jefferies, r. p. s. 1989. A new mitrate from the Upper Ordovician of Norway, and a new
approach to subdividing a plesion. Palaeontology, 32, 69-99.
cripps, a. p. 1988. A new species of stem-group chordate from the Upper Ordovician of Northern Ireland.
Palaeontology , 31, 1053-1077.

1989n. A new stem-group chordate (Cornuta) from the Llandeilo of Czechoslovakia and the

cornute-mitrate transition. Zoological Journal of the Linnean Society of London, 96, 49-85.

19896. A new genus of stem chordate (Cornuta) from the Lower and Middle Ordovician of

Czechoslovakia and the origin of bilateral symmetry in the chordates. Geobios, 22, 215-245.

— (in press). A new stem-craniate from Morocco and the search for the sister group of the Craniata.
Zoological Journal of the Linnean Society of London.

cuenot, L. 1948. Anatomie, ethologie et systematique des echinodermes. 1-363. In grasse, p.-p. (ed.). Traite
de zoologie, Vol. 11. Masson, Paris.

dehm, t. 1934. Untersuchungen iiber Cystoideen des rheinischen Unterdevons. Sitzungsberichte der bayerischen
Akademie der Wissenschaften, 1934, 19-43.

derstler, K. 1981. Morphological diversity of early Cambrian echinoderms. 71-75. In taylor, m. e. (ed.).
Short papers for the 2nd International Symposium on the Cambrian System. United States Geological
Survey, Open File Report, 81-743.

Durham, J. w. 1967. Notes on the Helicoplacoidea and early echinoderms. Journal of Paleontology, 41, 97-102.

and caster, k. e. 1963. Helicoplacoidea: a new class of echinoderms. Science, 140, 820-822.

1966. Helicoplacoids. U1 31—1 36. In moore, r. c. (ed.). Treatise on invertebrate paleontology. Part U.

Echinodermata 3, (I). Geological Society of America and University of Kansas, Lawrence, Kansas.

Eaton, T. h. 1970. The stem-tail problem and the ancestry of chordates. Journal of Paleontology, 44,
969-979.

fedotov, D. m. 1924. Zur Morphologic des axialen Organkomplexes der Echinodermen. Zeitschrift der
wissenschaftlichen Zoologie, 123, 209-304.

gehling, J. G. 1987. Earliest known echinoderm - a new Ediacaran fossil from the Pound Subgroup of South
Australia. Alcheringa, 1 1, 337-345.

678

PALAEONTOLOGY, VOLUME 33

gemmill, J. F. 1914. The development and certain points in the adult structure of the starfish Asterias rubens
L. Philosophical Transactions of the Royal Society of London, Series B , 205, 213-294.
gill, E. d. and caster, k. e. I960. Carpoid echinoderms from the Silurian and Devonian of Australia. Bulletin
of American Paleontology, 41 (185), 1—71.

goldring, R. and stephenson, d. G. 1972. The depositional environment of three starfish beds. Neues Jahrbuch
fur Geologie und Paldontologie , Monatshefte, (1972), 611-624.
grobben, k.. 1908. Die systematische Einteilung des Tierreiches. Verhandlungen der zoologisch-botanischen
Gesellschaft in Wien, 58, 491-511.

harper, d. a. t. 1981. The stratigraphy and faunas of the Upper Ordovician High Mains Formation of the
Girvan district. Scottish Journal of Geology, 17, 247-255.

1982. The stratigraphy of the Drummock Group (Ashgill), Girvan. Geological Journal, 17, 256-277 .

- 1984. Brachiopods from the Upper Ardmillan succession (Ordovician) of the Girvan district, Scotland.
Monographs of the Palaeontographical Society , 565, 1-78.

hawkins, H. L. 1920. Invertebrate palaeontology. Methuen, London.

hennig, e. 1932. Wesen und Wege der Paldontologie. Borntraeger, Berlin.

hennig, w. 1969. Die Stammesgeschichte der Insekten, Kramer, Frankfurt-am-Main.

hyman, L. h. 1955. The invertebrates : Echinodermata, the coelomate Bilateria. vol. 4. McGraw-Hill, New York,
763 pp.

jaekel, o. 1900. Uber Carpoideen, eine neue Klasse von Pelmatozoen. Zeitschrift der deutschen geologischen
Gesellschaft, 52, 661-677.

1918. Phylogenie und System der Pelmatozoen. Palaontologische Zeitschrift, 3, 1-128.
jefferies, r. p. s. 1968. The subphylum Calcichordata (Jefferies 1967) - primitive fossil chordates with
echinoderm affinities. Bulletin of the British Museum (Natural History), (Geology), 16, 243-339.

- 1969. Ceratocvstis perneri - a Middle Cambrian chordate with echinoderm affinities. Palaeontology, 12,
494-535.

1973. The Ordovician fossil Lagynocystis pyramidalis (Barrande) and the ancestry of amphioxus.
Philosophical Transactions of the Royal Society of London, Series B, 265, 409—469.

1979. The origin of chordates - a methodological essay. 443-477. In house, m. r. (ed.). The origin of major
invertebrate groups. Systematics Association Special Volume, 12, 1-515.

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

1988. How to characterize the Echinodermata - some implications of the sister-group relationship
between echinoderms and chordates. 3-12. In PAUL, c. r. C. and smith, a. b. (eds.). Echinoderm phytogeny
and evolutionary biology. Liverpool Geological Society and Clarendon Press, Oxford.

— and lewis, d. n. 1978. The English Silurian fossil Placocystites forbesianus and the ancestry of the
vertebrates. Philosophical Transactions of the Royal Society of London, Series B, 282, 205-323.

— , lewis, m. and donovan, s. k. 1987. Placocystites menevensis - a stem-group chordate (Cornuta) from the
Middle Cambrian of South Wales. Palaeontology, 30, 420-484.

— and prokop, r. j. 1972. A new calcichordate from the Ordovician of Bohemia and its anatomy,
adaptations and relationships. Biological Journal of the Linnean Society of London, 4, 69-115.

kolata, d. r., strimple, H. l. and levorson, c. o. 1977. Revision of the carpoid family Iowacystidae.
Palaeontology , 20, 529-557.

lester, s. m. 1985. Ceplialodiscus sp. (Hemichordata : Pterobranchia) around Bermuda. Marine Biology, 85,
263-268.

1988. Settlement and metamorphosis of Rhabdopleura normani (Hemichordata: Pterobranchia). Acta
Zoologica (Stockholm), 69, 111-120.

m acbride, E. w. 1914. Textbook of embryology, Vol. I. Invertebrata. Macmillan, London.
melendez, B. 1982. Paleontologia. Paraninfo. Madrid, xvi + 722 pp.

nichols, D. I960. The histology and activities of the tube feet of Antedon bifida. Quarterly Journal of
Microscopical Sciences, 101, 105-117.

1972. The water-vascular system in living and fossil echinoderms. Palaeontology, 15, 519-538.
parsley, r. l. 1982. Functional morphology of mitrate homalozoans (Echinodermata). Abstracts and
Programs of the Geological Society of America, 14 (7), 583.

1988. Feeding and respiratory strategies in Stylophora. 347-361. In paul, c. r. c. and smith, a. b. (eds.).
Echinoderm phytogeny and evolutionary biology. Liverpool Geological Society and Clarendon Press,
Oxford.

and caster, k. e. 1965. North American Soluta (Carpoidea, Echinodermata). Bulletin of American
Paleontology, 49, 109-174.

JEFFERIES: D EN D ROC Y STO I D ES AND CHORDATE ANCESTRY

679

paul, c. R. C. 1977. Evolution of primitive echinoderms. 123-158. In hallam, a. (ed.). Patterns of evolution as
illustrated by the fossil record. Elsevier, Amsterdam.

-and smith, a. b. 1984. The early radiation and phylogeny of echinoderms. Biological Reviews , 59,
443-481.

rehkamper, g. and welsch, u. 1988. Functional morphology of the stone canal in the sea urchin Eucidaris
(Echinodermata : Echinoidea). Zoological Journal of the Linnean Society , 94, 259-269.
sato, t. 1936. Vorlaufige Mitteilung iiber Atubaria heterolopha gen. et sp. nov., einen in freiem Zustand
aufgefundenen Pterobranchier aus dem Stillen Ozean. Zoologischer Anzeiger, 1154, 97-106.
schepotieff, a. 1907. Die Pterobranchier. Anatomische und histologische Untersuchungen iiber Rhabdopleura
normani Allman und Cephalodiscus dodecalophus MTnt. 2. Teil. Cephalodiscus dodecalophus M’lnt. 1.
Abschnitt. Die Anatomie von Cephalodiscus. Zoologische Jahrbiicher (Abteilung Anatomic , etc.), 24,
553-608.

spinar, z. 1960. Zaklady Paleontologie Bezobratlych. Ceskoslovenska Akademia Ved, Praha.
sprinkle, J. 1973. Morphology and evolution of blast ozo an echinoderms. Harvard, Cambridge, Mass.
termier, h. and termier, g. 1960. Paleontologie stratigraphique. Masson, Paris.

ubaghs, G. 1968. Homostelea. S565-S581. In moore, r. c. (ed.). Treatise on invertebrate paleontology. Part S.
Echinodermata 1. Geological Society of America and University of Kansas, Lawrence, Kansas.

— 1981. Reflexions sur la nature et la fonction de l’appendice articule des carpoides Stylophora
(Echinodermata). Annales de Paleontologie (Invertebres), 67, 33-48.

— 1987. Echinodermes nouveaux du Cambrien moyen de la Montagne Noire (France). Annales de
Paleontologie, 73, 1-27.

— and robison, r. a. 1985. A new homoiostelean and a new eocrinoid from the Middle Cambrian of Utah.
Paleontological Contributions, University of Kansas, 115, 1-24.

woods, h. 1937. Palaeontology-invertebrate. Cambridge University Press. London.

woodward, h. [as H W. ’]. 1913. Review of ‘ Caradocian Cystidea from Girvan’. By F. A. Bather. Geological
Magazine, (5) 10, 4 1 8 — 423.

R. P. S. JEFFERIES
Department of Palaeontology

Typescript received 10 April 1989 British Museum (Natural History)

Revised typescript received 20 August 1989 Cromwell Road, London, SW7 5BD

STROM ATOPOROID PA LA EO B I O LOG Y AND
TAPHONOMY IN A SILURIAN BIOSTROM E ON
GOTLAND, SWEDEN

by STEPHEN KERSHAW

Abstract. A well-exposed stromatoporoid biostrome in the Hemse Group (middle Ludlow) of Gotland,
Sweden, displays a wide range of stromatoporoid morphologies distributed amongst 16 species. The most
abundant species, Clathrodictyon mohicanum Nestor has a laminar to low domical form, often of large size,
distributed throughout the biostrome. Fast lateral growth is suggested to account for its abundance and
commonly large size, its profile being suited to the normally low to moderate energy environment envisaged
for the biostrome. Several other species adopted a similar growth style but were less successful; yet others, such
as Plectostroma intermedium Yavorskyi, show a range of growth form from low to high domical shape,
suggesting a phenotypic plasticity of growth form of each of these species. However, the range simply reflects
the fact that individuals grew taller with age. Most laminar to low domical stromatoporoids at Kuppen are
intact, but taller forms were frequently damaged and overturned by periodic storm action, indicating that a
lower profile was clearly an advantage. Some species show variations in distribution of undamaged forms
horizontally between, and vertically within, localities; there is evidence that some responded to environmental
gradients, notably turbulence and sedimentation. Most grew on a stable substrate provided by dead skeletons
of other stromatoporoids. Under cathodoluminescence stromatoporoid skeletons show speckled dull and
bright luminescence, an identical signal to the pelmatozoan debris found in the biostrome, circumstantial
evidence for an original mineralogy of high Mg calcite in both.

Recent advances in studies of stromatoporoid palaeoecology have focused attention on the
importance of recognizing the degree to which genetics controlled growth form (Cornet 1975;
Kershaw 1981 ; Stearn 1984) and this study examines these features in an excellent example of a
stromatoporoid biostrome at Kuppen on southeast Gotland, Sweden (text-fig. I). The biostrome
was the subject of an earlier paper (Kershaw 1981) which demonstrated that low profile and high
profile forms were created by different stromatoporoid species. However, that study was biased in
its collecting, by selecting particular morphologies and at only one site on a biostrome which is
exposed continuously for nearly one km laterally. Thus it was not possible to determine whether
species were truly limited to particular growth forms or instead were capable of producing a range
of morphologies.

In this study, a total sample of approximately 400 stromatoporoids collected from several points
along the length of the biostrome provides comprehensive information on the distribution of
stromatoporoids and the relationship between species and their morphotypes. Morphotypic
variation is shown to be greater than that discovered in the earlier study, but is related to growth
histories of individual stromatoporoids, and the interpretation of a genetic control on growth form
is confirmed. Also, as the biostrome accumulated, storms selectively damaged some growth forms
and certain species, and allow useful observations on the palaeoenvironment and palaeobiology of
the stromatoporoids to be made. Lithification history of the biostrome and cementation of
stromatoporoid skeletons are briefly examined to enhance the taphonomic interpretations. The
information obtained from the study provides an insight into growth styles of different
stromatoporoid species, controls on their forms, and indicates the use of stromatoporoids in
determining features of the palaeoenvironment, particularly with regard to sedimentation and
turbulence.

jPalaeontology, Vol. 33, Part 3, 1990, pp. 681-705. |

© The Palaeontological Association

682

PALAEONTOLOGY, VOLUME 33

text-fig. 1 . Location of the Kuppen biostrome within the Hemse Beds (stippled) on Gotland. Sampling
localities are shown, details of which are filed at the Allekvia Field Station on Gotland and based on the locality
system devised by Laufeld ( 1 974or). Material used for this study is deposited with the Swedish Geological

Survey, Uppsala.

GEOLOGICAL SETTING

The Silurian sediments of Gotland have been viewed as deposited in a carbonate platform setting
(Kaljo 1970, Laufeld and Bassett 1981), but with evidence of a carbonate ramp (Frykman 1989) in
the Baltic area, and record several transgressive and regressive sequences in which argillaceous
limestone formations alternate with biohermal and biostromal limestones (Riding 1981).

The Hemse Beds (middle Ludlow) of Hede (1960) comprise calcareous mudstones towards the
west, the Hemse Marl, and stromatoporoid-dominated biostromal and biohermal limestones in the
east. The precise stratigraphic relationship between the muds and limestones is unclear. A generally
southwest dipping palaeoslope has been envisaged by Martinsson (1967, p. 383) for much of the
Gotland sequence, and the Hemse Marl facies are estimated to have been deposited in deeper water
than the limestones (Laufeld 19746). However, work with orthoconic nautiloids has shown
conflicting probable shoreline trends in the Hemse Beds. Laufeld (19746) interpreted current flow
in their southwestern outcrops to be perpendicular to the Ludlovian shoreline which he interpreted
to have an ENE-WSW trend, with open sea generally to the south. However, Sundquist (1982u)
presented rose diagrams of bipolar nautiloid orientations for the Grogarnshuvud localities, 3 km
NW of Kuppen, which show orientations at approximately 90° to Laufeld’s. This suggests an
approximate NW-SE shoreline in this area. These results are not necessarily incompatible because
there may have been local variations in shoreline trends or current flow directions.

Shallow water facies are common in the eastern outcrops of the Hemse Beds. Ripple marks at

KERSHAW: SILURIAN STROM ATOPOROID PALAEOBIOLOGY

683

Gannes, 5 km west of Kuppen (Sundquist 19826) are interpreted to have formed in water of about
10m depth; Manten (1971, pp. 353, 379) reported erosion surfaces at Gannes. Cherns (1982)
identified palaeokarsts, erosion surfaces and stromatolites in the overlying Eke Formation, and
Cherns (1983) indicated shallow water facies at the Hemse-Eke boundary.

Biostromes 0-5 - c. 10 m thick are widespread in the southeast Hemse Beds, interbedded with
shales, skeletal grainstones and rudstones. Such thinly bedded deposits, with shallow water
associations, exhibit common vertical facies changes illustrative of a sea level fluctuating relative to
sea bed. Lateral facies changes occur also, clearly demonstrable at Kuppen; there a facies mosaic
is developed, termed here the Kuppen facies complex. A brief description of the complex follows,
providing a background to interpretation of the stromatoporoid distributions and growth forms.

THE KUPPEN FACIES COMPLEX

The complex consists of several biostromes interbedded with argillaceous limestones, stromato-
poroid conglomerates and crinoidal grainstones (text-fig. 2); the lowest biostrome is studied here.
Stromatoporoid biostromes crop out in neighbouring areas, but only those at Kuppen were
examined. Riding (1981, p. 65) recognized the biostromes as distinct from other organic build-ups
on Gotland and applied the term ‘Kuppen reef type’. He provided a log of the vertical facies
changes at locality Kuppen 2 (see text-fig. 2), and noted lateral variations in the complex. Text-fig.
3 illustrates important selected features of the facies and displays the general character of the
stromatoporoid morphologies.

As shown in text-fig. 2, the sequence begins with a poorly exposed argillaceous limestone
containing a sparse stromatoporoid fauna. The overlying biostrome (from which the sample was
collected) has, in most places, a crinoidal limestone basal bed, 30-50 cm thick, with stromatoporoid
debris and some large stromatoporoids in situ. Pressure solution has extensively affected the
stromatoporoids in places, such that in many the original morphology is unrecognizable. Biostrome
sediment is most micritic fill between stromatoporoids, with lesser amounts of bioclastic material.
The biostrome varies in thickness from 2-5 4 m and is everywhere terminated by an erosion

surface, much of which is prominent and planar (text-fig. 3a and c; Munthe 1910, fig. 28);
stromatoporoids are truncated, indicating that they were cemented into the matrix prior to erosion.
An indeterminate amount of the upper section of the biostrome has been removed, and the erosion
surface becomes more irregular towards the ESE, (text-fig. 3a). Minor sea level fluctuations could
account for the vertical facies changes. The sequence is similar to the model provided by James
(1984, p. 220) of shallowing-upwards carbonate units and to the example studied by Wilson (1967)
in Devonian limestones of the Williston Basin of N. Dakota. Gypsum crystals in the upper part of
the biostrome suggest the possibility of hypersaline conditions during a stage in the development of
the biostrome (Kershaw 1987c/), but no corroborative sedimentological evidence exists due to
erosion of the top of the biostrome.

Dipping beds at Kuppen 1 are attributed to deposition on a slope associated with the palaeocliff
(text-fig. 2), while undulation of bedding in the remainder of the complex is probably due to settling
of the sediments on the shale at the base of the complex. There is only a slight tectonic dip across
Gotland and the complex evidently formed on an approximately flat-lying sea bed. Lack of
siliciclastics suggests a low hinterland, and consequently little terrigenous sediment transport into
the area during the development of the complex. The SE area contains coarse skeletal limestones,
whereas the NW areas with thin, interbedded, finer-grained limestone facies were quieter.

The current energy affecting the complex clearly varied. Most tall domical stromatoporoids are
now on their sides, indicating that energy was relatively low to moderate while they were alive, but
periodically sufficient to topple them. Substantial amounts of stromatoporoid debris also indicate
higher energy events. The biostrome has the appearance of a mixed fauna (text-fig. 3), with many
large stromatoporoids not disturbed from growth position, while debris was piled around them.
Further thin biostromes interleaved with bedded limestones overlie the lower biostrome (text-fig. 2),
but were not studied in detail. However, there are some differences in fauna; for example, the

684

PALAEONTOLOGY, VOLUME 33

KERSHAW: SILURIAN STROM ATOPO ROI D PALAEOBIOLOGY 685

biostrome overlying that studied here has a greater abundance of tabulate corals, particularly
favositids.

Diagenesis

A full description of diagenesis of the biostrome is inappropriate here, but some aspects pertain to
a taphonomic assessment of the stromatoporoids, so an outline of the important events is given.
Three cement stages are recognized in both bioclastic fill and stromatoporoid galleries (text-fig. 4).

First stage cement forms non-luminescent (under cathodoluminescence - CL), non-ferroan
syntaxial calcite on pelmatozoan debris. It is also present sporadically in stromatoporoid skeletal
galleries (text-fig. 4b, d) and is most likely to be meteoric cement formed during uplift. Syntaxial
cement on pelmatozoans is commonly regarded as meteoric because high Mg levels in sea water
prevent its growth there; meteoric water is low in Mg. In the present sample, such cement fills most
of the pore space of pelmatozoan grainstones, and leaves little doubt that it is of meteoric origin.
There was apparently little marine cement, and this is corroborated by the common occurrence of
toppled tall stromatoporoids. As noted by Kershaw (1981) some were toppled during life and show
reorientated growth directions.

Second stage cement comprises very thin alternating bright and non-luminescent bands
suggestive of subsequent burial near the redox boundary, with cyclic recharge of oxygenated water,
a common situation in limestones (Scoffin 1987, p. 129). Staining of the same thin sections that were
used for CL examination shows that these areas of the sections are also non-ferroan, which is
consistent with the interpretation of such banded cement (under CL) being generated by the
presence (bright) and absence (dark) of manganese, but with no iron present. The absence of iron
is usually regarded as being due to its combination with sulphides liberated by action of sulphate-
reducing processes on porewater sulphates in shallow burial, so that it is unavailable for inclusion
into calcite lattices (Scoffin 1987, p. 129). This cement is again best developed in the porous
pelmatozoan grainstone deposits and to a lesser extent in stromatoporoid skeletal galleries.

The final cement is dull luminescent and ferroan, where iron, now incorporated into the calcite
cement following depletion of sulphides in pore waters, partially counteracts the bright CL effect
of manganese (referred to as quenching). This cement, which probably indicates deeper burial,
occupies the small remaining pore space in grainstones. However, most of the pore space of
stromatoporoid galleries is usually filled with dull luminescent, ferroan cement, suggesting that it
grew slowly at depth, after the biostrome had been buried. Since they represent largely closed
porosity, stromatoporoid galleries are likely to have been less accessible to earlier pore waters.
Comparable luminescent zones were observed by Dorobek (1987) in late Silurian - early Devonian
carbonates in the Appalachians.

Stromatoporoid skeletons show a similar CL response to the pelmatozoan debris in the biostrome
(text-fig. 4b, d). Also, all stromatoporoids in the assemblage are partially recrystallized. General
observations by myself and other authors (Stearn, pers. comm. 1988; Watts 1981; Wood 1987)
indicate that many stromatoporoids did not recrystallize as completely as aragonitic shells of
molluscs, yet are consistently more recrystallized than the low Mg calcite of brachiopods. The
inference that these stromatoporoids were therefore high Mg calcite originally is supported by a) the
similarity of response in CL to pelmatozoans (high Mg calcite) and b) the widespread occurrence
of microdolomite rhombs in stromatoporoid skeletons at Kuppen. Lohmann and Meyers (1977)
discussed a burial diagenetic process whereby Mg expelled from high Mg calcite skeletons form
localized dolomite. CL response and dolomite occurrence may of course be due to other diagenetic
processes, but the partial recrystallization widespread amongst Silurian and Devonian stromato-
poroids argues against low Mg calcite or aragonite original compositions. There is a wider issue of
variation of stromatoporoid skeletal mineralogy through geological time and in relation to
taxonomy, which requires investigation.

686

PALAEONTOLOGY, VOLUME 33

text-fig. 3. For legend see opposite

KERSHAW: SILURIAN STROM ATOPOROI D PALAEOBIOLOGY

687

text-fig. 4. Thin section views of typical material from the lower biostrome. a. Plane polarized light view of
crinoidal grainstone from Kuppen 3A. b. Cathodoluminescence response of the same area, showing three zones
of cement growth: 1. non-luminescent syntaxial cement on crinoids; 2. alternating bright and non-luminescent
cements; 3. dull cement. Note the bright, dull and dark-speckled appearance of the crinoid fragments, c. Plane
polarized light view of recrystallized stromatoporoid skeletal material of Stromatopora venukovi with crystal
boundaries of neomorphic cement cutting across the stromatoporoid structure. The preserved laminae and
pillar structure are difficult to recognize at this magnification, but galleries are discernible as sparry patches.
d. Cathodoluminescence response of the same area shows the three cement zones in the galleries, plus a
speckled luminescent character of bright, dull, and dark areas in the stromatoporoid skeleton. See text for

discussion. All photos x 35.

FOSSIL ASSEMBLAGE OF THE LOWER BIOSTROME

Three accessible localities selected for intensive sampling are Kuppen 3, 4 and 5 ( text-figs. 1 and 2),
where the biostrome is 3-4, 3 and 2-5 m thick, respectively.

At each locality a grid was drawn on the vertical cliff face, 10 m long and of variable height
depending on the thickness of the biostrome at that place. Each was divided into three parts parallel

text-fig. 3. Photographs of selected parts of the Kuppen biostrome. a. Cliff face at Kuppen 5 showing basal
bioclastic bed and densely packed stromatoporoids in the biostrome; length of view 8 m. b. Interbedded shales
and limestones overlain by thin biostromes and limestone bands at the northwestern end of Kuppen; length
of view 3-5 m. c. Biostrome at Kuppen 1, undercut by the sea where shale underlies the biostrome, and sharp
erosion surface overlain by crinoidal gravel. Note the large, low domical stromatoporoid centre left in
biostrome; length of view 5 m. d. Eroded stack (top left in hollow) at the biostrome top at Snabben 1,
surrounded by stromatoporoid conglomerate and crinoidal gravel ; length of view 3 m. e. Vertical section of
central part of biostrome at Kuppen 4 (see text-fig. 5), showing laminar, low to high domical, broken and

stylolitized stromatoporoids; length of view 0-95 m.

688

PALAEONTOLOGY, VOLUME 33

A

KUPPEN 4 sample grid

Kuppen 4 reference point

T

^erosion surface

.CM

PS

.CM

.CM

’CM .CM
>T .

CM

CM

PI

*LL

CM.,

SB

CC*

•CM

CM.

CM*

PT

SC
CM l

• CM
• CC

PT SB
CC

.CM

.CM

Level B

PT

SBo*

,PT

CM

SC.

PI

#.CM

P|r..CM

CM

CC

SBo.. p|
•SB

.CC

CM

PT.

PT

•LS SC
CC

CM. •

LS c^.SBo

CM. PI

CM.
CM *
CM*

CC

•CM

CC-

PI.

CC
Level C

CM

.CM

CM-

CM. .SC
*SV

.SC

biostrome base

B

KUPPEN 4 STROM AT0P0R0IDS

text-fig. 5. Schematic vertical section of the lower biostrome at Kuppen 4 to demonstrate sampling method.
a. Sample locations of identified species, determined by vertical and lateral coordinates derived from random
number tables, b. Morphology of the stromatoporoids sampled, illustrated in their preserved attitudes and
positions. Three categories of preservation of growth form are recognized: c = complete specimens; b =
broken skeletons, including small fragments; 5 = specimens so badly affected by pressure solution (stylolites)
to make impossible any distinction between originally complete or broken. Specimens not identified are
omitted from the diagram. Text-fig. 3e (inset) shows the general character of the outcrop. Stromatoporoid
species are abbreviated in this and other diagrams, as follows. LL = Labechia lepida. CC = Clathrodictyon
convictum, CM = C. mohicanum , EM = Ecclimadictyon macro tuberculatum, PK = Plexodictyon katriense , DY
= Diplostroma yavorskyi, PS = Plectostroma scaniense , PI = P. intermedium , SB = Stromatopora bekkeri , SC
= 5. carteri , SL = S. lamellosa, SV = 5. venukovi , SBo = Syringostromella borealis , PT = Parallelo stroma

typicum , LS = Lophiostroma schmidti.

to bedding, here termed levels, of approximately equal thickness and labelled a, b, c, respectively
from top to bottom (text-fig. 5). Bedding planes do not exist within the biostrome and so the level
boundaries are artificial. Therefore levels are not necessarily time equivalent between localities.

Within each level, stromatoporoids were collected using random number tables to determine the
precise collecting point and to obtain a representative sample of the assemblage (text-fig. 5). The

KERSHAW: SILURIAN STROM ATOPOROI D PALAEOBIOLOGY

689

intention to collect 50 samples from each level at every locality was not possible and the final sample
consisted of 404 specimens. Kuppen 5A was the worst affected with only 18 pieces retrieved. Of the
total, 60 were unidentifiable, but these were evenly spread across sample sites (average of five in
each) except in the case of Kuppen 3B, where 13 of 48 collected could not be classified. Thus the
overall distribution of identified stromatoporoids is not unduly affected by the presence of
unidentified samples. Almost all samples collected are stromatoporoids, with few corals. Heliolitids,
however, are common in places, whereas favositids are rare. Kershaw (19876) showed that 10% of
stromatoporoids in this biostrome contain symbiotic corals; further samples revealed that in fact
20% of the stromatoporoid assemblage contains a variety of rugose corals and syringoporids,
although this is unevenly distributed within the biostrome. Various other taxa, particularly
echinoderins, are present as debris.

Since collecting was random, the data set provides a representative sample of the assemblage, and
allows comparison within and between localities to furnish information on vertical and horizontal
variation.

Stromatoporoid assemblage

Sixteen species distributed amongst several orders of stromatoporoids were found. Taxonomic
revision by Stearn (1980) produces differences from the scheme applied by Mori (1969, 1970) to
Gotland stromatoporoids which was used also by Kershaw (1981). Steam’s division into several
orders is adopted here, giving the following list for the biostrome:

Order Lophiostromatida ; Lophiostroma schmidti (Nicholson)

Order Labechiida; Labechia lepida Mori

Order Actinostromatida : Plectostroma scaniense Mori, P. intermedium Nestor, Pseudolabechia

granulata Yabe and Sugiyama

Order Clathrodictyida : Clathrodictyon mohicanum Nestor, C. convictum Yavorsky, Ecclimadictyon

macro tuberculatum (Riabinin), Plexodictyon katriense Nestor, Diplostroma yavorsky i Nestor
Order Stromatoporida : Stromatopora bekkeri Nestor, S. carteri Nicholson, S. lamellosa Yavorsky,

S. venukovi Yavorsky, Syringostromella borealis (Nicholson), Parallelostroma typicum (Rosen).

These species are all well-known Palaeozoic forms and since they are fully described by Nestor
( 1966), Mori (1969, 1970), Stock (1979) and Stock and Holmes (1986), there is no need to illustrate
them here. However, worth noting is that stromatoporoid classification is based on microscopic
characters of laminae and pillar structure and arrangement. Two thin sections (vertical and
horizontal) are required per specimen. The shape of the entire skeleton is not a diagnostic character,
although some species are limited to certain shapes. Most species in the present sample are
distinctive, but some raise taxonomic problems that are briefly discussed below.

Most specimens.of Stromatopora bekkeri and Parallelostroma typicum are distinct, but some show
such variability of structure of skeletal elements that occasional difficulty was faced in deciding into
which species they should be grouped. Fagerstrom and Saxena (1973) recognized that variations of
stromatoporoid structure within single thin sections can be high, but all within a single distinct
species. Stock (1979), Stock and Holmes (1986) distinguished several species of Parallelostroma in
the Upper Silurian of eastern America, with minor, but consistent, differences in skeletal structure
defining the species. This is not possible in the present specimens, where variation, often in different
parts of the same thin section, produces laminae and pillar arrangements resembling both species.
The taxonomic uncertainty generated by these observations means that the distinction between
these two species is unreliable. The possibility that they represent phenotypic variants of a single
species cannot be discounted, although they could represent two species with overlapping skeletal
morphology, clearly causing havoc with the morphospecies concept as applied to these species. The
validity of other species at Kuppen is not questioned since they show consistent differences of
skeletal structure. For the present purposes, P. typicum and S. bekkeri have been distinguished as
separate species in text-figs. 5, 6, 8, 9 and 10, and the uncertain examples omitted, pending a separate
analysis of this problem.

table 1. Distribution of numbers of specimens of all identified stromatoporoids from Kuppen, arranged according to locality and level, with the

690

PALAEONTOLOGY, VOLUME 33

"O

G

"O

X

T3

G

a

o

a

6

o

o

bO

a

D-

O

H

U

00

c

aj

a

a

g

U

CG

u

CQ

a

a

— (N <N <N

— X

I <N I

tJ- (N (N <N

<U>

a

C/3

Tf I ^ ro

M D —

I rn

rn ON OO (N |

Tf <N

— I X

Tf oo n

I I ^ I

— I m <N

ro <n
' m

5 § s ^ S ' -

Qs ^ ^

co «c> -C» be •

^ ~C)

2 J(' ? J
^ 12 ^ s
~ a 'r h

^Q^cbcocbcdcb^^ c cd q od kj '"dd^

. lepida

KERSHAW: SILURIAN STROM ATOPOROI D PALAEOBIOLOGY

691

Plectostroma scaniense is distinguished from P. intermedium only by possessing astrorhizae
(Nestor 1966, pi. XI; Mori 1970, pi. X). Many authors have discussed the taxonomic significance
of these structures, and there is disagreement about their importance (see Stearn 1980, Wood 1987
for reviews). Both species exhibit the same growth form characteristics. 1 would accept that they
could be the same species, but have left them separate in case future work should show them to be
distinct.

Stromatoporoid growth form

Table 1 shows the total identified sample, in each locality. Text-fig. 6 and Table 2 summarize the
data into localities and levels for visual comparison. Division into three categories of preservation
of the growth form (c, b, s; see Table 1) is important for morphological analysis. The most useful
for morphological study are complete specimens (category c). Fragmented specimens (category b)
are useful for studying biostratinomy, and specimens affected so badly by pressure solution that
their shape cannot be determined (category s), were included to provide full information on specific
distribution. Pressure solution effects are more pronounced at Kuppen 5 than at the more southerly
localities. Dimensions of broken and stylolitized specimens were not recorded in the field; both are
generally small, of the order of 10 cm in size, although larger specimens occur.

Growth forms fall into a relatively simple range from laminar to extended domical and bulbous
of Kershaw and Riding (1978, 1980), illustrated in text-fig. 7. Most stromatoporoids have smooth
outlines and rarely show ragged edges (where sediment periodically covered the flanks of skeletons).
Within smooth outlines, laminae show different growth patterns. Some are completely enveloping ,
in which successive laminae cover all earlier laminae; others are non-enveloping , where growth
produced additions at the apex and extended only a short distance down the flanks. Others have a
combination of the two, and in these, lower stages of growth were enveloping while later growth was

table 2. Distribution of numbers and percentages of complete specimens at Kuppen; note the large decrease
of Clathrodictyon mohicanum, and increase of Plectostroma intermedium from Kuppen 4 to 5.

Numbers of

com

plete specimens

Kuppen 3

Kuppen 4

Kuppen 5

Total

Species

No.

%

No.

%

No. %

No.

%

L. schmidti

4

51

5

6-8

6 10

15

7-1

P. typicum

1

1-3

6

8-2

4 6-7

1 1

5-2

S. borealis

8

10-3

3

41

1 1

5-2

S. venukovi

3

3-8

2

2-7

1 1-7

6

2-8

S. lamellosa

1

1-3

—

—

1

0-5

S. carteri

2

2-6

2

2-7

—

4

1-9

S. bekkeri

13

16-7

2

2-7

7 1 16

22

10 4

P. granulata

—

—

—

—

P. inter-

3

3-8

2

2-7

15 25

20

9-5

medium
P. scaniense

1

1-3

i

1-4

2 3-3

4

1-9

D. yavorskyi

2

2-6

—

—

2

0-9

P. katriense

1

1-3

—

—

1

0-5

E. macrotuber-

—

i

1-4

—

1

0-5

culatum
C. mohicanum

33

42-2

41

56 3

15 25

89

42-3

C. convict um

—

8

1 1

10 16-7

18

8-5

L. lepida

6

7-7

—

—

6

2-8

78

73

60

211

692

PALAEONTOLOGY, VOLUME 33

KUPPEi

STR0SV1AT0P0R01D DISTRIBUTION

A. levels

B. localities

6
3

LEVEL A
100

S BOREALIS

S VENUKOVI
S LAMEILOSA
S.CARTERI

S BEKKERI
PGRANUIATA

P INTERMEDIUM

PSCANIENSE

D YAVORSKYI

P KATRIENSE

E. MACROTUBER-
CULATUM

14

10

1

9

38

1

69

10

2

2

1

KUPPEN 4
125

C.MOHICANUM 107
C CON VICTUM 31

1

KUPPEN 5
101

L.LEPIDA

9

344

text-fig. 6. For legend see opposite.

KERSHAW: SILURIAN STROM ATOPOROI D PALAEOBIOLOGY

693

Kuppen stromatoporoid morphotypes

EXTENDED DOMICAL

BULBOUS

morphotype classes:

LAMINAR- V/B UP TO 0-1

LOW DOMICAL- V/B 01-05

HIGH DOMICAL- V/B 0-5-1

EXTENDED DOMICAL -V/B>1 L

NON- ENVELOPING

SMOOTH

ENVELOPING

NON -ENVELOPING

text-fig. 7. Stromatoporoid morphotypes in the lower biostrome, showing a range from laminar to extended
domical and bulbous, together with morphotype classification and terminology applied to the forms. These
shapes are typical, simple, Silurian forms, shown in vertical section except where indicated. All growth forms
can possess either enveloping or non-enveloping laminae. Very few are ragged and some show well developed

mamelons.

non-enveloping. All three variations of smooth outlines are found in each of the more abundant
species.

Attempts to quantify these features proved unrewarding; to determine the extent of enveloping
and non-enveloping laminae in individual stromatoporoids requires medial vertical sections
through skeletons. Only some specimens are suitably broken open in outcrop, and many individuals
are displayed entire, obscuring the internal patterns. Therefore it has not been possible to determine
the degree to which each style is developed at Kuppen, and graphical representations of morphology
in the present study necessarily ignore these patterns in smooth stromatoporoids.

Because of the range of simple morphologies at Kuppen, the use of the parameterization scheme
devised by Kershaw and Riding for graphically displaying stromatoporoid morphotypes is

text-fig. 6. Graphical representation of the distribution of specimens of identified stromatoporoid species in
the lower biostrome. The same data are arranged in two ways to illustrate vertical and horizontal variations.
a. All specimens in each level, b. All specimens in each locality. Specimens are grouped into categories c, b and
s according to the preservation of their growth form; see caption of text-fig. 5. Note abundance of
Clathrodictyon mohicanum , and lesser amounts of Plectostroma intermedium , Stromatopora bekkeri ,
Parallelostroma typicum , Clathrodictyon conviction , Syringostromella borealis and Lophiostroma schmidti.

694

PALAEONTOLOGY, VOLUME 33

text-fig. 8. Triangular displays of morphology of all complete stromatoporoids in the lower biostrome,
plotted using the Kershaw and Riding (1978) scheme. Morphology varies little between localities and levels in
the biostrome, so species are illustrated grouped together. Inset shows measurements of basal diameter (B),
vertical height (V) and diagonal length (D), and the morphological fields generated by the display. Partial rose

diagrams of attitudes of skeletons are provided.

KERSHAW: SILURIAN STROM ATOPOROID PALAEOBIOLOGY

695

appropriate. Text-fig. 8 shows the plotted forms of all complete specimens; fragmented and
stylolitized examples cannot be represented.

Text-fig. 8 shows that while many species, such as the ubiquitous C. mohicanum , are restricted to
the lower profile fields, other species show a range of form from low to extended domical and
bulbous. This suggests a phenotypic plasticity of the growth form of these species. S. bekkeri also
seems not to be restricted to tall growth form as suggested by Kershaw (1981).

However, these displays take no account of the overall dimensions of skeletons. Since most
growth forms plot on a line emanating from the basal corner, the most important components of
the stromatoporoid morphology at Kuppen are B and V, with D having less effect. Therefore the
V/B ratio in these specimens provides a crude measure of shape, and when plotted against B (text-
figs. 9, 10), two clear trends are revealed in the more abundant species. For low profile forms, low
B values are usually accompanied by low V/B ratios; but at higher B values, V/B stays low. For
those species which can have taller forms, the graphs show that low B values are also accompanied
by low V/B ratios, but at higher V/B ratios, B is still low (text-fig. 9). Furthermore, examination
of early growth stages in the stromatoporoids shows those of tall forms to be low profile. Thus
Plectostroma intermedium , Clathrodictyon convictum Parallelostroma typicum and Stromatopora
bekkeri have a range of growth form from low to extended domical (text-fig. 9). Such ranges are
present in most places sampled, and there is no evidence that they are restricted to low profile forms
in particular sites on the biostrome. They are interpreted here as merely younger individuals.
Therefore within these specimens, the early stages of growth are low domical, and the V/B ratio
increased as the specimen grew, so that its morphology changed from lower to higher domical. Since
ragged forms are rare at Kuppen, the taller morphologies are not an artifact generated by
sedimentation, as can be the case in stromatoporoids in other environments. Graphs in text-fig. 9
consequently indicate that species adopted either lateral or vertical growth styles, giving low or high
profile forms, respectively. These features are summarized in text-fig. 1 1 .

Although the two-fold growth response pattern is clear, there is not as sharp a dichotomy of
growth style as this discussion may imply. In text-figs. 9 and 10 a few specimens of Clathrodictyon
convictum , Parallelostroma typicum , Stromatopora bekkeri and Plectostroma intermedium are large
low profile stromatoporoids, which therefore do not fit the pattern. In the field, these
stromatoporoids were shown to have more than one growth centre, so that these forms can be
accounted for by coalescence of several specimens of one species which happened to grow near each
other, and merged as they grew. Coalescence is common amongst conspecifics in stromatoporoids,
so presumably there was no immune system response of individuals to members of the same species.
Not all coalesced examples became large, low domical forms; cases of extended domical coalesced
specimens were found at Kuppen, and the final growth form appears to have been controlled by the
number and spatial distribution of conspecifics involved in its generation (text-fig. 1 1). The same
process probably accounts for the larger specimens of C. mohicanum , but cannot be recognized
satisfactorily in the field because laminae are usually undulose and identification of separate early
growth stages is difficult.

Little information is available on morphologies of these stromatoporoids in other localities.
Combining information from reports on Silurian stromatoporoids from elsewhere on Gotland,
Estonia (Mon 1970; Nestor 1966), Canada (Savelle 1979) and New York (Stock 1979), it is clear
that morphologies of Labechia lepida, Clathrodictyon convictum , C. mohicanum , Stromatopora
carteri , S. bekkeri , Lophiostroma schmidti , and Syringostromella borealis are broadly similar to these
species at Kuppen, but several species have also been found with other growth forms. Ecclimadictyon
macrotuber culatum can be high domical and bulbous in addition to the low domical specimen at
Kuppen; Plexodictyon katriense is laminar at Kuppen, but domical elsewhere; Diplostroma
yavorskyi is extended domical as well as the low domical form found at Kuppen; Plectostroma
scaniense and Parallelostroma typicum can be laminar as well as the domical forms at Kuppen.
Stromatopora venukovi has been reported as domical elsewhere, but at Kuppen there are also
laminar forms. These descriptions are not based on comprehensive data, but imply differences in
response of these species in other environments.

LEVEL C LEVEL B LEVEL

696

PALAEONTOLOGY. VOLUME 33

KUPPEN

STR0MAT0P0R0IDS

morphology & size

O S.BEKKERI 22

a P. INTERMEDIUM 20

. C. MOHICA NUM 89
A C. CON VIC TUM 18

i

2q

.1-

V/B

B

T— | ! 1 1 1 1 1 ! 1 T“

50 100 cm

-i — i — i — i — | — i — i — i — i — j — i — i — i — r

am

A

1

-i — i — I — i — | — i — i — I — i — | — i — i — i — r-

4a o

. t

KUPPEi 3

KUPPEN 4

1 — i — r- 1 — l — l — i — rnr

KUPPEN 5

text-fig. 9. Graphical representation of stromatoporoid morphology, using V/B ratio in relation to basal
diameter for the four most common species at Kuppen, arranged according to levels and localities to illustrate
abundance and variation in size and morphology. Inset shows scales and morphologies represented by the
graphs. Two distinct trends are present, vertical and lateral. See text for discussion.

LEVEL C LEVEL B LEVEL

KERSHAW: SILURIAN STROM ATOPOROID PALAEOBIOLOGY

697

KUPRIN

STR0MAT0P0R0IDS

morphology

• L. LEPIDA 6

© E MACROTUBER- H
CULATUM »

ePKA TRIENSE 1

® D.YAVORSKYI 2

□ P SCANIENSE 4

0 S.CARTERI 4

m size

a S.LAMELLOSA 1
* S. BOREALIS 11
^ S.VENUKOVI 0
o P TYPICUM 1 1
▼ L. SCHMIDT! 15

B

1 — j — J — b — i — ! — i — i — » — i — r

50 100 cm

KUPPEN 3 KUPPEi 4 KUPPEN 5

text-fig. 10. Graphical representation of stromatoporoid morphology, using V/B ratio in relation to basal
diameter for the less common species at Kuppen, to illustrate abundance and variation in size and morphology.
Inset shows scales and morphologies represented by the graphs.

698

PALAEONTOLOGY, VOLUME 33

Kuppen stromatoporoid growth styles

C COALESCENCE

text-fig. 1 1 . Stromatoporoid growth styles in the lower biostrome, interpreted from text-fig. 9. a. lateral style,
developed in low profile forms, such as Clathrodictyon mohicanum. b. Vertical style, formed by high profile
forms, such as Plectostroma intermedium. Many develop enveloping laminae, but others are non-
enveloping, while retaining the smooth outline. See text for discussion, c. Forms produced by coalescence of
several specimens of the same species which grew near each other, (i) some large low domical forms may be
due to this; (ii) high and extended domical forms may contain a small number of early growth stages; (iii) many
specimens coalescing may mask the vertical growth style and produce large, low to high domical forms of
species which would otherwise be extended domical. Some examples of Stromatopora bekkeri and Plectostroma

intermedium in text-fig. 9 are due to this feature.

Morphology of Plectostroma scaniense and Parallelo stroma typicum was examined in more detail
by Mori ( 1970, figs, 23 and 28). His graphs of height plotted against diameter for specimens of these
species, in several stratigraphic units on Gotland, reveal considerable variety of morphology and
size. Using Mori's data, V/B was calculated, and plotted against B, as in text-figs. 9 and 10. The
results are not presented here, but show that, in the Hemse Beds, both species have a range of form
from laminar to high domical. Some low profile skeletons are large and may be due to coalescence,
although this is not determinable from Mori's data. In the Sundre Beds (upper Ludlow), however,
Mori’s data show that almost all specimens of both species fall into the low domical field (V/B =
01 0 5), and some large forms are laminar, including a specimen of Plectostroma scaniense c.

70 cm in diameter ( = B). Although Mori’s graphs combine data from reef and non-reef localities,
Mori (1970, p. 45) noted that stromatoporoid reefs in the Hamra and Sundre Beds are dominated

KERSHAW: SILURIAN STROM ATOPO ROI D PALAEOBIOLOGY

699

by low profile forms, and my own field examination revealed a remarkable lack of tall
stromatoporoids. Therefore coalescence cannot be responsible for the low profiles, and there is
clearly a lateral growth style in these specimens from the Sundre Beds.

STROMATOPOROID DISTRIBUTION AND BIOSTR ATINOM Y

Stromatoporoid morphology is not uniformly distributed in the biostrome. Visual examination
along the length of the outcrop shows that large, low profile forms are most abundant between
Snabben 1 and Kuppen 4, but from Kuppen 4 towards Herrvik Harbour such forms decrease in
abundance and taller stromatoporoids are more common. Also, the number of stromatoporoids
with symbiotic (in the broad sense of the term) syringoporids and rugosans dramatically increases
just north of Kuppen 4, and in places is much greater than the 20% average for the biostrome
quoted earlier. Sampling the entire length of the biostrome was impractical because of access to the
cliff, but the sample sites chosen are clearly representative of the variations present. The only
obvious vertical change in the biostrome is that the upper 0 5 m between Kuppen 2 and Kuppen 4
contains stromatoporoids commonly of smaller size, with a more rubbly general appearance than
the lower portions.

Text-fig. 6 shows heterogeneity in the vertical and lateral abundance of several species in the
biostrome, and analysis of this variation is important to determine palaeoecological characters of
the stromatoporoids. The more abundant species show variations as follows. Lophiostroma schmidti
and Stromatopora venukovi are present only in the basal part of the biostrome, but in all localities;
Parallelostroma typicum and Stromatopora bekkeri occur throughout (including Kuppen 2;
Kershaw 1981) but rarely in the upper portion of the unit; Syringostromella borealis occurs only at
Kuppen 3 and 4 with more specimens in the upper part; Stromatopora carteri also occurs only at
Kuppen 3 and 4, but rarely in the upper part; Plectostroma intermedium is widespread throughout
the biostrome with a lot of fragments at Kuppen 3, and was also found at Kuppen 2 (Kershaw
1981); Clathrodictyon mohicanum is abundant everywhere (including Kuppen 2: Kershaw 1981),
but diminishes slightly at Kuppen 5; C. conviction is more abundant in the lower part, but is missing
in Kuppen 3 (one specimen at Kuppen 2: Kershaw 1981); and lastly Labechia lepida is uncommon,
but is missing from the lower level, and rare at Kuppen 5.

Most laminar to low domical stromatoporoids (e.g. C. mohicanum , Lophiostroma schmidti and
Syringostromella borealis ) are complete, upright and probably did not suffer much damage (effects
of abrasion have not been recognized because of pressure solution on the stromatoporoid margins).
Laminar to low domical forms of all species are usually upright, irrespective of basal diameter. All
the species with these morphologies comprise a well developed calcite skeletal structure, with robust
laminae and pillars, expected to provide considerable strength to their skeletons. Most of the
higher profile species, including Parallelostroma typicum , Stromatopora bekkeri and Clathrodictyon
conviction , are also complete, but are commonly overturned (text-fig. 8), and in some cases more
damaged than stromatoporoids of lower domical shapes (text-fig. 6). Plectostroma intermedium
suffered considerable damage in places.

The lack of open reef framework and early cementation permitted ready movement of debris
across the biostrome, and this has clearly happened to a substantial number of stromatoporoid
fragments. Comparisons of stromatoporoid abundance and growth form between localities can be
usefully studied only using complete specimens. Distribution of fragments, however, provides
information on severity of damage to stromatoporoids and is a necessary part of ecological study.
Since fragments are generally small, around 10 cm across, a large number can be derived from a
single stromatoporoid. However, assessment of the number of missing whole stromatoporoids
represented by the fragmented material is unrealistic because of the variation in size of complete
specimens and of the unknown degree of transport into, and out of, the Kuppen area. Nevertheless,
in many species the degree of damage is small, especially in Clathrodictyon mohicanum , giving the
impression that only a small number of these were broken up during high energy. In the case of
Plectostroma intermedium, although no broken specimens were recorded in the sampling grid at

700

PALAEONTOLOGY, VOLUME 33

Kuppen 5, half the sample there is affected by pressure solution. Its effects distort the relative
numbers of complete and broken samples, because most specimens in category s were of small
dimensions, and are likely to be fragments; consequently damage to P. intermedium is most
probably even greater than text-fig. 6 implies. The difference in damage between C. mohicanum and
P. intermedium can be partly attributed to the more easily damaged taller form of the latter, but P.
intermedium is the most damaged of all the stromatoporoids in the suite and especially so in the
upper part of Kuppen 3. The skeleton of this species consists of a network of thin pillars and
processes instead of the strong skeleton characteristic of other species with tall forms at Kuppen.
Its skeleton appears delicate in thin section and, as has been shown earlier, the majority of cement
filling stromatoporoid galleries is dull luminescent, suggesting late (burial) cement. Thus the
observed degree of damage is consistent with interpretation of an easily damaged, brittle structure
whilst on the sea bed. Consequently the growth forms and skeletal structures of the stromatoporoids
possess different preservation potentials. Although most showed non-significant results when using
chi-squared tests, comparisons between numbers of specimens of P. intermedium and C. mohicanum
show clearly than the former is significantly more damaged at Kuppen 3 at the 5% level. Chi-
squared tests between P. intermedium and combined numbers of Parallelostroma typicum and
Stromatopora hekkeri showed the same result. This confirms the impression of a delicate structure
to Plectostroma intermedium compared to the more robust skeletons of Parallelostroma typicum-
Stromatopora hekkeri. All these species have similar morphologies (text-fig. 8).

Statistical comparisons are also instructive to assess the general character of the distribution of
complete stromatoporoids. Statistical tests were used to compare stromatoporoids between
localities, and between levels within localities, but not between levels combined from different
localities since level boundaries are arbitrary and no reliable correlation can be made between them.
Tests on the distribution of specimens, and of growth forms, show a variety of results. Chi-squared
tests were used to compare numbers of complete specimens of pairs of species between localities to
determine distributional relationships, and therefore lateral variation. Importantly, Plectostroma
intermedium increased significantly from Kuppen 4 to Kuppen 5, and over the larger distance from
Kuppen 3 to Kuppen 5 when compared with C. mohicanum (1 % level). C. conviction increases
significantly from Kuppen 3 to Kuppen 4 in contrast to C. mohicanum , and also from Kuppen 3 to
Kuppen 5 when tested with C. mohicanum and Syringostromella borealis at the 5% level. Tests
involving Parallelostroma typicum and Stromatopora hekkeri were considered unreliable in view of
taxonomic problems and were omitted. Tests using other species were not particularly informative
in view of the small numbers, although much variation in distribution is evident from Table 1 and
text-fig. 6.

The above tests illustrate large scale lateral changes, but variations can be recognized between
levels within localities. Thus each level is considered to be time-averaged and compared as a unit
with other levels to illustrate temporal changes in the development of the stromatoporoids at any
site. With use of the Fisher exact probability test, the visual decrease in basal diameter (a measure
of size) of C. mohicanum from Kuppen 3B to 3A (text-fig. 9) is significant only at the 13% level,
and the increase from 4B to 4A at the 9% level. With levels combined to compare localities, the
visual reduction in size from Kuppen 4 to 5 reveals significance at the I 1 % level. These results are
therefore not regarded as significantly different. The Fisher test also reveals a non-significant result
when the V/B ratio (a measure of shape) in Plectostroma intermedium is compared between Kuppen
4 and 5 (P = 0-48) and when Kuppen 5 is compared to both Kuppen 3 and 4 combined (P = 0-24).
Thus specimens of this species are not significantly taller at Kuppen 5 than at Kuppen 4 or Kuppen
3, as might be expected in the quieter environment of Kuppen 5.

To summarize important features of the tests: C. mohicanum (a low profile form) is significantly
less abundant at Kuppen 5 compared with Plectostroma intermedium (a tall profile form), which
increases sharply at Kuppen 5; C. mohicanum varies in size between levels and localities, but not
significantly, and some less abundant species show fluctuations in abundance laterally in the
biostrome. It can be argued that the striking dilference in distribution of numbers of whole samples
of C. mohicanum and P. intermedium is an artifact caused by extensive damage to P. intermedium

KERSHAW: SILURIAN ST RO M ATOPO RO I D PALAEOBIOLOGY

701

at Kuppen 3A. However, it should be noted that the number of whole specimens of P. intermedium
which would need to have been destroyed to generate the fragments is relatively small, because
debris is much smaller than whole specimens, even if the stylolitized samples (most of which were
probably originally fragments) are included. Also, no fragments of C. mohicanum were collected
from Kuppen 3A, and the complete specimens were of smaller size than in most other sites (text-
fig. 9). Other species with a variety of growth profiles are represented largely by complete specimens
in Kuppen 3A (Table I), and the overall impression is of localized movement of debris. Variations
in distribution of stromatoporoid species vertically and laterally throughout the biostrome support
this (text-fig. 6 and Tabic 1) and reflect the life distribution of the stromatoporoids, thus broadly
relating to their palaeobiology.

Lastly, a curious feature of the Kuppen stromatoporoids is a notable lack of bioerosion.
Stromatoporoids from the Upper Visby Beds (lower Wenlock, Gotland), for example, are heavily
bored by Trypanites , but not at Kuppen. Thus the damage to stromatoporoids at Kuppen can be
safely attributed to physical processes.

CONTROLS ON STROMATOPOROID GROWTH

The very high abundance of stromatoporoids in the biostrome was an important aid to the
development of its fauna, because of the provision by taphonomic processes of a substrate of dead
skeletons of earlier individuals. Particularly significant is the dominance of low profile
stromatoporoids. C. mohicanum must have acted as a key substrate stabilizer, and Lophiostroma
schmidti , with its characteristic encrusting habit, would have assisted in immobilizing areas of
substrate. Some specimens of this latter species are large (text-fig. 10), and occur in the lower portion
of the biostrome as well as in the marl beneath. In the absence of a cemented frame, the stability
of laminar and low domical stromatoporoids in a regime of low sedimentation, allowed repeated
colonization of the surface. Encrusting crinoids, bryozoans, and rugose and tabulate corals have
also made use of this facility. L. schmidti was also commonly found encrusting, and was encrusted
by, other stromatoporoid species, a feature noted by Mori (1970).

Distributions of species in the lower biostrome are clearly heterogeneous. The south end of the
Kuppen complex shows coarser grained sediments overlying the biostrome, and deep palaeo-
erosion into it. The northwestern end (at Kuppen 5) shows a thinning of the unit and frequent
interfingering of sediment in the area between Kuppen 5 and Herrvik. Lower energy water
permitting a higher sedimentation rate in the Kuppen 5 area would have selected taller forms, and
prejudiced survival of low profile stromatoporoids because of greater chance of them being buried.
This could account for the distributional differences between P. intermedium and C. mohicanum, but
paucity of ragged forms implies an overall low depositional rate. Most specimens of Clathrodictyon
conviction show a smooth non-enveloping form, suited to a lower energy situation with an increased
sedimentation rate at Kuppen 5, so that growth was concentrated at the apex of individual
stromatoporoids. The small numbers and size of C. mohicanum at Kuppen 5 could be taken to
indicate inhibition of growth by slow continuous sedimentation, rather than periodic deposition
which would have led to development of ragged forms.

Species-morphology relationships in the Kuppen suite show a pronounced taxonomic control on
growth form (text-figs. 8-10), and illustrate ontogenetic patterns in growth form development (text-
fig. 11). The reasons why different species developed different growth forms, however, remain
undiscovered. While sedimentation rate is invoked here to play a part in controlling distribution,
the abundance of low profile forms and rarity of ragged shapes indicates an overall low rate of
deposition for the biostrome, and therefore sedimentation cannot be inherently responsible for the
differences in form. In conditions of low energy, some species may have grown upwards to create
local turbulence as suggested by Kershaw ( 1981 ). This speculative point could be invoked to explain
why tall forms grew in the NW area, but since high and low profile forms occur upright and in close
proximity in the southern area at Kuppen 2 (Kershaw 1981) and can be satisfactorily regarded as
being in situ , such an interpretation cannot be generally applied. Also, apart from those at the base

702

PALAEONTOLOGY, VOLUME 33

of the biostrome, there was no need for low profile forms to develop as a response to avoid sinking
in soft sediment, because their substrate was mostly dead skeletons of stromatoporoids, suitable
also for taller forms because of the stability of a hard surface.

Work on Devonian stromatoporoids illustrates the importance of lateral growth as a competitive
feature (Meyer 1981), which would account for the abundance of C. mohicanum. However,
competition for space between stromatoporoids at Kuppen, whereby fast growing, low profile
forms would have commanded large areas of substrate, forcing others to grow upwards, cannot
explain the dichotomy of form because of the different distributions of Plectostroma intermedium
and Clathrodicyon mohicanum (text-fig. 6). Abundance of low profile shapes could be related to a
competitive feature, but also to adaptation to an occasionally turbulent environment with storm
action selectivity removing taller forms periodically. Since energy was lower most of the time, a low
profile shape would only be necessary for that purpose in times of high energy.

Recent work on other stromatoporoid assemblages also indicates the selective advantage of
lateral growth. Harrington (1987) describes the usefulness of a lateral growth habit for colonizing
unstable substrates in Devonian stromatoporoids. Bjerstedt and Feldmann (1985) show how
stromatoporoid morphology varies from stabilizing, low profile forms in the lower parts of a Middle
Devonian reef to more erect shapes once stabilization has been achieved. Therefore the development
of a low profile form can be explained in a number of ways, with difficulty in distinguishing between
them. Thus at Kuppen the presence of higher profile forms could be attributed to fluctuations of
environmental energy, allowing periodic colonization by stromatoporoids with high profile shapes
which were selectively damaged during storms.

Stromatoporoids which normally formed tall shapes are occasionally large, low domical
morphologies (text-fig. 9). These are demonstrably due to coalescence of several individuals, and the
simultaneous growth of a number of stromatoporoids near each other is circumstantial evidence for
growth on dead areas of substrate since closely packed fossils were not necessarily alive
simultaneously. In the case of Stromatopora hekkeri (text-fig. 9), the location of these particular
specimens at the base of the biostrome at Kuppen 3 is also notable, their substrate being the
underlying bedded marls (text-fig. 2) poorly populated by stromatoporoids. Development of a low
profile form in this situation would also help to spread weight on a soft substrate. Kershaw (1984)
suggested substrate control for low profile forms in the Upper Visby Beds (Wenlock) on Gotland.

Data presented by Mori (1970), noted earlier, show that stromatoporoid reefs in the Sundre Beds
are dominated by low profile stromatoporoids. This reinforces the importance of a low profile in
reef-builders. The principal species in these Sundre reefs, Plectostroma scaniense and Parallelostroma
typicum , are tall forms at Kuppen, while the dominant stromatoporoid at Kuppen. Clathrodictyon
mohicanum , is missing from the Sundre Beds (upper Ludlow, younger than the Hemse Beds). There
is clearly a need to investigate the reasons for these differences, which may involve environmental
changes, selecting for some species and excluding others. However, the possibility of adaptive
evolution towards a more efficient growth form in some species also needs to be considered.

Stromatoporoid growth irregularities

Some stromatoporoids at Kuppen exhibit small protrusions, called mamelons, 0-5 - 2 cm in height,
and not associated with astrorhizae as mamelons sometimes are. They occur rarely in Stromatopora
hekkeri , Parallelostroma typicum , Plectostroma scaniense, and P. intermedium , but they covered the
upper surface of the skeleton in 7 of the 107 specimens of C. mohicanum. They were vertically
orientated in specimens which lay with their basal dimension horizontally on the substrate, but one
specimen lying at an angle of c. 30° from horizontal in the upper part of Kuppen 3 still had vertically
orientated protrusions (text-fig. 7). This latter feature was also found in several unidentified
stromatoporoids from other parts of the biostrome and a recrystallized stromatoporoid from a
biostrome in the Klinteberg Beds at locality Vivungs 1 of Laufeld (1974u), and lead to speculation
as to their cause.

Boyajian and La Barbera (1987) investigated mamelons experimentally in conjunction with
astrorhizae in model stromatoporoids and Ceratoporella nicholsoni and showed their importance in

KERSHAW: SILURIAN STROM ATOPOROI D PALAEOBIOLOGY

703

directing water currents through the structure to maximize feeding. This work is based on the well
established Bernoulli’s principle of water flow from high pressure, slow moving water near the
stromatoporoid skeleton to low pressure faster moving water around the tips of protrusions (Vogel
1978).

Mamelons in the present specimens could have had a similar function, or as Kershaw (1981)
suggested, caused turbulence close to the skeleton surface to assist feeding. The absence of
astrorhizae on surfaces of many stromatoporoids does not necessarily negate a relationship with
water flow, because in many modern calcified sponges the astrorhizal system exists entirely in the
soft tissue coat and makes no impression on the skeleton. Nevertheless, in many stromatoporoids
at Kuppen, and elsewhere in Silurian and Devonian rocks, there is a calcified astrorhizal system that
is not associated with mamelons, and so the importance of mamelons in controlling water currents
is questionable. Also, the presence of vertically orientated mamelons on sloping stromatoporoids
suggest that other factors are responsible for mamelon development.

Suggestions have been made that stromatoporoids were photoresponsive. Kazmierczak (1969)
proposed symbiotic ?algae, and also that stromatoporoids are themselves cyanobacteria (e.g.
Kazmierczak and Krumbein 1983). Since only a few specimens of C. mohicanum have such
protrusions, they are unlikely to represent a photoresponse in this case. Surface morphology of the
biostrome was suggested by Kershaw ( 1 98 1 ) to cause mamelon growth in hollows. Although surface
depressions on the biostrome are not identifiable due to biostratinomic effects, their inferred
presence could explain the specimens with mamelons. They are present in several localities and
levels and consequently not limited to particular parts of the biostrome. They could therefore be a
response to low energy causing an increase in stromatoporoid surface area and enhanced eddy
formation for feeding in quiet spots. However, they cannot reflect increased sedimentation where
they occur since the 'valleys’ between mamelons only rarely contain sediment, and laminae of the
stromatoporoid are continuous over the tops of mamelons and the valleys between them.

CONCLUSIONS

This study reveals two major growth styles in stromatoporoids in a low diversity assemblage,
favoured by one species, in a low to moderately turbulent environment with a low sedimentation
rate. Environmental gradients involving water turbulence and sedimentation appear to have played
an important role in controlling distributions of species and growth forms. The abundance of low

text-fig. 12. Reconstruction of a small part of the
lower biostrome, showing many stromatoporoids in
life attitude, and taller forms overturned, plus
fragments. Large, low profile shapes are based on
Clathrodictyon mohicanum , and taller forms on Plecto-
stroma intermedium. One specimen contains rugose
corals, while an extended domical form on its side has
continued to grow following overturning. A large
domical form has developed by coalescence. Matrix is
micrite, with skeletal debris.

704

PALAEONTOLOGY, VOLUME 33

profile stromatoporoids, in particular Clathrodictyon mohicanum, can be attributed to a number of
factors including substrate colonization, stability of the substrate during times of higher energy,
competitive ability, and coalescence, and is favoured by a low sedimentation rate. Most
stromatoporoids in the suite grew on dead skeletal material, providing an important feedback to the
living assemblage. Biostratinomic movement of skeletons, selective removal of taller growth forms
and the pervasive effects of pressure solution, have distorted the life assemblage. A composite
reconstruction of a small area of typical biostrome surface is shown in text-fig. 12, to illustrate the
biostrome surface at an unspecified point in its development.

Acknowledgements . I gratefully acknowledge funding from the West London Institute, with additional support
from the Swedish Geological Survey, and access to the Allekvia Field Station on Gotland. I thank Robert
Riding (Cardiff, U.K.). Michael Keeling, Julian Harrigan and Nicholas Palaus (West London Institute) for
discussions on the Kuppen Complex, and Peter Sutcliffe (Kingston Polytechnic, U.K.) for assistance with
cathodoluminescence. I am grateful to Colin Stearn (McGill, Canada), Michael Bassett (National Museum of
Wales, U.K.) and reviewers for many helpful comments in the manuscript. This work is dedicated to Ruth
and Reginald Kershaw.

REFERENCES

bjerstedt, t. w. and feldmann, r. m. 1985. Stromatoporoid paleosynecology in the Lucas Dolostone (Middle
Devonian) on Kelleys Island, Ohio. Journal of Paleontology , 59, 1033-1061.
boyajian, G. e. and labarbera, m. 1987. Biomechanical analysis of passive flow of stromato-
poroids morphologic, paleoecologic and systematic implications. Lethaia , 20, 223-229.
cherns, l. 1982. Paleokarst, tidal erosion surfaces and stromatolites in the Silurian Eke Formation of Gotland,
Sweden. Sedimentology, 29, 819-833.

1983. The Hemse-Eke boundary; facies relationships in the Ludlow Series of Gotland, Sweden. Sverige
Geologiska Undersokning , Series C, 800, 45 pp.

cornet, p. 1975. Morphogenese , caracteres ecologiques , et distribution des stromatoporoides Devoniens au bord
du sud du Bassin de Dinant (Belgique). These de doctorat (inedite), Universite Catholique de Louvain. 195

pp.

dorobek, s. l. 1987. Petrography, geochemistry and origin of burial diagenetic facies, Siluro-Devonian
Helderberg Group (carbonate rocks). Central Appalachians. American Association of Petroleum Geologists
Bulletin , 71, 492-514.

fagerstrom, J. a. and saxena, k. m. l. 1973. Intracoenosteal variation in a Devonian stromatoporoid. Lethaia ,
6, 155-162.

frykman, p. 1989. Carbonate ramp facies of the Klinteberg Formation Wenlock-Ludlow transition on
Gotland, Sweden. Sverige Geologiska Undersokning Series C. 820, 79 pp.

Harrington, r. J. 1987. Lithofacies and biofacies of the Middle and Upper Devonian Sultan Formation at
Mountain Springs, Clark County, Nevada: implications for stromatoporoid paleoecology. Journal of
Paleontology , 61, 649-662.

hede, j. e. 1960. The Silurian of Gotland, pp. 44-87. In regnell, g. and hede, j. j. (eds.). The Lower Palaeozoic
of Scania; The Silurian of Gotland; Guide to Excursions nos. A22 and Cl 7. International Geological
Congress 21, Norden. Stockholm, 44—89.

james, n. p. 1984. Shallowing-upwards sequences in carbonates, pp. 213-218. In r. g. walker (ed). Facies
Models. Geoscience Canada Reprint Series 1, 317 pp.

kaljo, D. L. (ed.) 1970. Silur Estonii. Eesti NSV Teaduste Akadeemia Geoloogia Instituut. Tallinn, 343 pp.
kazmierczak, J. 1969. A new interpretation of astrorhizae in the stromatoporoidea. Acta Palaeontologica
Polonica , 14, 499-525.

— and krumbein, w. e. 1983. Identification of coccoid cyanobacteria forming stromatoporoid stromatolites.
Lethaia , 16, 207-213.

kershaw, s. 1981. Stromatoporoid growth form and taxonomy in a Silurian biostrome, Gotland. Journal of
Paleontology , 55, 1284-1295.

1984. Patterns of stromatoporoid growth in level-bottom environments. Palaeontology, 27, 113-130.

- 1987m Gypsum evaporites in the Ludlovian of Gotland. Geologiska Foreningens i Stockholm
Forhandlingar , 109, 331-333.

KERSHAW: SILURIAN ST ROM ATOPO RO I D PALAEOBIOLOGY

705

1987b. Stromatoporoid-coral intergrowths in a Silurian Biostrome. Letluiia , 20. 371-380.

— and riding, R. 1978. Parameterization of stromatoporoid shape. Lethaia , 11, 233-242.

1980. Stromatoporoid morphotypes of the Middle Devonian Torbay Reef Complex at Long Quarry
Point, Devon. Proceedings of the Ussher Society , 5, 13-23.
laufeld, s. 1974a. Reference localities for palaeontology and geology in the Silurian of Gotland. Sverige
Geologiska Undersokning Series C, 705, 172 pp.

— 1974b. Preferred orientation of orthoconic nautiloids in the Ludlovian Hemse Beds of Gotland.
Geologiska Foreningens i Stockholm Forhandlingar , 96, 1 57 1 62.

— and bassett, M. G. 1981 . Gotland : the anatomy of a Silurian carbonate platform. Episodes , ( 1981 ) 23-27.
lohmann, k. c. and meyers, w. j. 1977. Microdolomite inclusions in cloudy prismatic calcites: a proposed

criterion for former High-magnesium calcites. Journal of Sedimentary Petrology , 47, 1078-1088.
manten, a. a. 1971 . Silurian reefs of Gotland. Developments in Sedimentology 13, 539 pp. Elsevier, Amsterdam.
martinsson, a. 1967. The succession and correlation of ostracode faunas in the Silurian of Gotland. Geologiska
Foreningens i Stockholm Forhandlingar , 89, 350-386.
meyer, f. 1981. Stromatoporoid growth rhythms and rates. Science , 213, 894-895.

mori, K. 1969. Stromatoporoids from the Silurian of Gotland, I. Stockholm Contributions in Geology , 19, 100

pp.

1970. Stromatoporoids from the Silurian of Gotland, II. Ibid. 22, 152 pp.
munthe, h. 1910. On the sequence of strata within southern Gotland. Geologiska Foreningens i Stockholm
Forhandlingar , 32, 1397-1453.

nestor, h. 1966. Stromatoporidei wenlocka i ludlowa Estonii. Eesti NSV Teaduste Akadeemia Geoloogia
Instituut , 1 -87.

riding, R. E 1981. Composition, structure and environmental setting of Silurian bioherms and biostromes in
Northern Europe, pp. 41-83. In d. f. toomey (ed.) : European Fossil Reef Models, SEPM. Special
Publication 30, 546 pp.

savelle, j. M. 1979. Upper Silurian stromatoporoids from Somerset Island, Arctic Canada. Canadian Journal
of Earth Science , 16, 364-372.

SCOFFIN, T. p. 1987. An introduction to carbonate sediments and rocks. Blackie, Glasgow and London, 274 pp.
stearn, c. w. 1980. Classification of the Paleozoic stromatoporoids. Journal of Paleontology, 54, 881-902.

— 1984. Growth form and macrostructural elements of the coralline sponges. Palaeontographica Americana,
54, 315-325.

stock, c. w. 1979. Upper Silurian (Pridoli) Stromatoporoidea of New York. Bulletins of American
Paleontology, 76, 291-381.

— and holmes, a. e. 1986. Upper Silurian/Lower Devonian Stromatoporoidea from the Keyser Formation
at Mustoe, Highland County, West-Central Virginia. Journal of Paleontology, 60, 555-580.

sundquist, b. 1982a. Wackestone petrography and bipolar orientation of cephalopods as indicators of littoral
sedimentation in the Ludlovian of Gotland. Geologiska Foreningens i Stockholm Forhandlingar, 104, 81-90.

— 1982b. Palaeobathymetric interpretation of wave ripple-marks in a Ludlovian grainstone of Gotland.
Geologiska Foreningens i Stockholm Forhandlingar, 104, 157-166.

VOGEL, s. 1978. Organisms that capture currents. Scientific American, 239, 108-115.

watts, N. R. 1981. Sedimentology and diagenesis of the Hdgklint reefs and their associated sediments. Lower
Silurian, Gotland , Sweden. Unpublished Ph.D. thesis. University of Wales, U.K. 407 pp.
wilson, j. l. 1967. Carbonate-evaporite cycles in Lower Duperow Formation of Williston Basin. Bulletin of
Canadian Petroleum Geology, 15, 230-312.

wood, r. a. 1987. Biology and revised systematics of some late Mesozoic stromatoporoids. Special Papers in
Palaeontology, 37, 89 pp.

STEPHEN KERSHAW

Palaeobiology Research Unit
West London Institute
Borough Road
Isleworth

Middx. TW7 5DU
U.K.

Typescript received 7 March 1989

Revised typescript received 20 December 1989

REVIEW OF THE CAENOZOIC HETERODONT
BIVALVE SUPERFAMILY DREISSENACEA

by C. P. NUTTALL

Abstract. The Dreissenacea consists of one family, the Dreissenidae, with subfamilies Dreisseninae for
Dreissenci (with three subgenera), Congeria and Mytilopsis ; and Dreissenomyinae for Dreissenomya (two
subgenera). Poorly preserved Prodreissensia is provisionally placed in the Dreisseninae, but may be wrongly
assigned to the superfamily. Dreissenids live in fresh and brackish water. Most are epifaunal and byssally
attached but Dreissenomya was infaunal and unattached. Mytilopsis first appeared in the European Eocene and
invaded the Western Hemisphere in the late Oligocene. Dreissena , Congeria and Dreissenomya arose in the late
Miocene of the Paratethys. Two genera survive: Mytilopsis occurs naturally only in the Western Hemisphere
and Dreissena in Europe and Western Asia. European Mio-Pliocene boundary extinctions coincide with the
break up of Paratethys and the Messinian salinity crisis. Often anomalous increases in the distribution of both
Recent genera are attributed to human activity, but some records from eastern Asia are due to confusion with
Sinomytilus (Mytilacea).

The largely freshwater, eulamellibranchiate, anisomyarian superfamily Dreissenacea first appears
in the Eocene and is still living. It was placed in the Veneroida (Heterodonta) in the classification
of Newell (1965) which was adopted by McCormick and Moore in Moore (1969). Confirmation of
this position was provided by Taylor, Kennedy and Hall ( 1973) who stressed the strong resemblance
between the shell structures of Dreissena itself and the Corbiculacea, another largely freshwater
superfamily which dates back into the Mesozoic.

Examination of the fossil record and the Recent distribution of the Dreissenacea shows that the
generic ranges, geographical distribution and arrangement of the genera given by Keen in Moore
(1969) (see Table 1) in the Treatise and by Eames in Morley Davies (1971) are in need of
modification. The Treatise classification (Moore 1969) had some influence on the views of both
Morton (1970) and Marelli and Gray (1985) on evolution the superfamily: in consequence, several
of their conclusions also need re-examination.

It is understandable how much of the confusion arose. Firstly for generic determination it is
necessary to examine internal features of the shell: one such feature is the presence or absence (pi.
1, figs. 6, 14, 16) of the apophysis which may be hidden under the septum and can easily be
overlooked except in good specimens and good illustrations. Secondly, many species were originally
described either as Mytilus or Dreissena. In addition, this situation is further complicated by the
almost universal practice of European workers of placing all species possessing an apophysis in
Congeria , regardless of the fact that most resemble the type species of Mytilopsis (pi. 1, figs. 11 14)
rather than that of Congeria (pi. 3). Thirdly, it is becoming clear that many Recent occurrences of
both Dreissena and Mytilopsis are the result of human introduction.

Oppenheim (1891, pi. 51) gave excellent illustrations of the internal details of several European
Eocene species, showing that all whose internal features were known possessed an apophysis (pi. 2,
figs. 12-18). These species had been originally described as Dreissena and have subsequently been
assigned to various subgenera of Congeria , but are here regarded as belonging to Mytilopsis. Most
subsequent European studies, however, have concentrated on the varied dreissenid faunas of the
Neogene of the Paratethys, stretching from Austria eastwards through the southern USSR as far
as the Aral Sea. The basis of the more recent Russian classifications (Starabogatov 1970; Babak
1983) is the work of Andrussov (1897-8; 1900) whose very comprehensive monograph of fossil and

| Palaeontology, Vol. 33, Part 3, 1990, pp. 707-737, 6 pls.|

© The Palaeontological Association

708

PALAEONTOLOGY, VOLUME 33

table 1 . Summary of modern classifications of the Dreissenacea

Keen in Moore (1969)
Fam. DREISSENIDAE

Babak (1983)

Fam. DREISSENIDAE
Subfam. dreisseninae

Herein

Fam. dreissenidae
Subfam. dreisseninae

Dreissena

Dreissena

Dreissena

( Dreissena )

( Dreissena )

( Dreissena )

—

( Pontodreissena)

( Pontodreissena )

—

( M odiolodreissena)

( M odiolodreissena)

( Prodreissensia )

—

Prodreissensia

Dreissenomya)

—

( Sinucongeria )
Congeria

Congeria

Congeria

( Congeria )

(= Congeria)

( Rhombocongeria )

( = Congeria)

(Eocongeria)

( = Mytilopsis)

( Trigonipraxis )

( = Mytilopsis)

(Andrusoviconcha)

( = Mytilopsis)

Mytilopsis

( Mytilopsis )

Mytilopsis

Subfam. dreissenomyinae

Subfam. dreissenomyinae

Dreissenomya

Dreissenomya

( Dreissenomya )

( Dreissenomya )

( Sinucongeria )

(Sinucongeria)

table 2. Relationship between informal groups of the genus Congeria proposed by Andrussov (1897-8) and
the subgeneric arrangements adopted by Starabogatov (1970) and Marinescu (1973)

Andrussov, 1897-8*

Starabogatov, 1970

Marinescu, 1973

eocenae

Eocongeria nov. subgen.

—

eocena (sic) = eocenica
Oppenheim

mytiliform.es

Mytilopsis Conrad

Mytilia nov. subgen.

basteroti Deshayes
modioliformes

Andrusoviconcha nov. subgen.

Modiola nov. subgen.

amygdaloides Dunker
subglobosae

Congeria Partsch

Congeria Partsch

no type species
rhomboidea

Rhombocongeria nov. subgen.

Rhomboidea nov. subgen.

no type species
triangulares

Trigonipraxis nov. subgen.

Triangularia nov. subgen.

triangularis Partsch

* Andrussov’s informal groups and their ‘type species’ are shown in this column.

living Dreissenidae of Eurasia also dealt with species from the rest of the world. Andrussov (1898)
divided both Congeria and Dreissena into a number of non-binomial ‘groups’, and gave ‘type
species ’ for some but not all of them (see Table 2). Later authors have subsequently formalized these
groups as nomenclatorially valid subgenera, sometimes with type species different to those
originally suggested by Andrussov. In the present work all such subgenera of Congeria are

NUTTALL REVIEW OF DREISSENACEA

709

synonymized either with Congeria or Mytilopsis (Table 1). The aim of the present study is to revise
the stratigraphic and geographic distributions of the various genera and to suggest improvements
to the classification adopted by earlier authors. Text-fig. I shows the correlation adopted herein for
the European post-Lower Miocene.

m.y

0

1.8

6.3

12.0

16.5

W. EUROPE

PARATETHYS
Central Eastern

RECENT

PLEISTOCENE

PLIOCENE

yy

x

UJ

o

o

Piacenzian
( Astian)

Zanclian

Messinian

I ortonian

Serravillian

Langhiart

Rumanian

Akchagylian

Dacian

Kimmerian

Pont

i a m

Pannonian

Meotian

(Malvensian)

Sarmatian

Sarmatian

(sensu Barbot)

(sensu Suess )

Konkian

Badeniam

Karaganian

Tchokrakian

text-fig. 1. European post-lower Miocene correlation and generic ranges. Mytilopsis first occurs in the
Eocene. Key, asterisk, Mytilopsis reintroduced after becoming extinct in eastern Hemisphere. Based on

Steininger et a! ., 1985.

The generic synonymies given herein concentrate on names proposed in modern taxonomic
works. They omit the numerous variations in the spelling of Dreissena. Also omitted are the many
nineteenth-century names which have long since fallen into disuse. This information is readily
available elsewhere (Keen in Moore, 1969; Vokes 1980).

The scope of this study is somewhat limited by the comparatively small collections in the
Palaeontology and Zoology departments, British Museum (Natural History) (BMPD; BMZD) and
is therefore largely based on reinterpretation of the literature. Probably the most useful source of
reliable specific identifications and natural groupings is Andrussov (1897-8). Other important
works consulted include the following concerned with classification: Orlov (1960); Keen in Moore
(1969); Starabogatov (1970); Babak 1983.

Works summarizing regional distribution and containing illustrations and references sufficiently
adequate for reinterpretation include: for Europe:- Archambault-Guezou (1976u, b): Iljina et al.
(1976); Locard (1893); Marinescu (1973, 1975); Oppenheim (1891, 1892); Papp (1953, 1954);
Rozanov (1986); Zelinskaya et al. (1968): for the Western Hemisphere:- Keen (1971); Marelli and
Gray (1983); Olsson (1961); Nuttall (1990, pp. 275-287): for West Africa:- Pilsbry and Becquaert
(1927). Anatomy and other aspects of the biology of living Dreissena are dealt with by Morton
(1970) and by Yonge and Campbell (1968). Escarbassiere and Almeida (1976) and Morton (1981)

710

PALAEONTOLOGY, VOLUME 33

dealt similarly with living Mvtilopsis. Several, mainly early, papers have been largely ignored
because it is not possible to interpret the sparse stratigraphical information they contain. For
example, among those that have proved too difficult to use is the important study of Yugoslav
Miocene dreissenids by Kochansky-Devide and Sliskovic (1978). They figured many apparently
rather crushed species and introduced (pp. 35, 60) a new informal group, the Tuciniformes’. Their
range chart (p. 76) differs from current views in correlating the Tortonian with the Badenian,
regarding both as pre-Sarmatian. They placed both the Otnangian and Karpathian within the
Helvetian, which they regarded as pre-Badenian. Nowadays the general concensus is to consider the
Helvetian as equivalent to the Tortonian.

MORPHOLOGY AND MODE OF LIFE

Inequivalvity

Slight inequivalvity in the Dreissenacea arises from two factors. In all genera except Dreissenomya ,
some articulation, or rather means of relative location between the two valves, is achieved by
differences in the ventral shell margin or commissure immediately posterior to the umbones (pi. I,
figs. 2, 3). In the left valve a sinus followed by a broader upswelling is developed just posterior to
the umbo. In Dreissena ( Pontodreissena ) rostriformis (Deshayes 1838) upswelling is very
pronounced, forming a strong beak (pi. 1, figs. 8 10). In the right valve, an upward angular
projection is present immediately posterior to the umbo. The sinus developed behind this projection
is rather weaker than the upswelling in the left view. These structures, which are purely
modifications of the commissure, are not true hinge teeth, even though they are concerned with
articulation. The other cause of inequivalvity is that the byssal notch is not always developed evenly
in both valves. Yonge and Campbell (1968, p. 5) stated that in Dreissena polymorpha (Pallas) the
umbo of one valve overlaps that of the other ventrally, and this one alone bears the byssal notch.
They did not state which valve is which and their observation on the byssal notch cannot be
confirmed. This species lives in clusters with the result that both shell shape and also the orientation
of the byssal anchorage could well be influenced by crowding against adjacent shells. This habit of
living in tight clumps adopted by most dreissenids (text-figs 2a, e) can cause considerable
morphological variation due to pressure from adjacent individuals. This has led to the description
of numerous unnecessary nominal species, as witnessed by the lengthy synonymies in Marelli and
Gray (1983) and in Nuttall (1990, pp. 280, 285) of, respectively. Recent and Neogene Mvtilopsis.
Similar reservations must be expressed about the large number of other fossil species described as
occurring either together or at particular horizons. Internal inequivalvity may occur in the septal
region. The septa of the two valves are often not mirror-images of each other. Variations occur in
the size and shape of the muscle attachments on the septa and also the way in which the septa may,
or may not be, supported or buttressed from below.

Anterior muscle attachments

The form of the septum and its muscle attachments have provided the basis not only of generic
separation but also much of the evidence of the evolution of the family. The anterior adductor
muscle is always attached to the septum. In Dreissena (pi. I , fig. 6) and Dreissenomya (pis. 5, 6; text-
fig. 3) the pedal/byssal retractor scar occurs on the dorsal part of the septum. In Mvtilopsis (pi. 1,
figs. 14, 16) the pedal/byssal retractor scar is situated on a myophore placed sometimes slightly
underneath, sometimes posterior to, the dorsal part of the septum. In Congeria (s.s.) this myophore
tends to be fused to the dorsal margin of the septum (pi. 3). This fusion may be interpreted as
resulting from massive thickening of the septum rather than a fundamental difference between
Congeria and Mvtilopsis. The taxonomic usefulness of the septum and apophysis has been
questioned (Archambault-Guezou 1982, with further references). Previously (1976a, pi. 9, figs. 16,
17) she had illustrated juvenile valves of the southern French ‘infra-Pliocene’ (= Messinian)

‘ Congeria ' rhodanica (Fontannes 1882) (here transferred to Mvtilopsis). In the right valve the

NUTTALL: REVIEW OF DREISSEN ACEA

711

apophysis is fused to the septum and the pedal/byssal scar is deep and oval. In the left valve the
apophysis is mounted behind and below the septum.

Archambault-Guezou (1976/?) was clearly of the opinion that ‘C’. rhodanica was extremely
closely related to ‘C’. subcarinata (Deshayes 1838) originally described from the Pliocene of the
Crimea. Morton (1970) figured a left valve (BMPD LL18540) of ‘C’. subcarinata botenica
Andrussov, 1897 from the Middle Pontian of Rumania in which the apophysis is fused to the
septum. Examination of a further sample (BMPD LL9022-4) from another Rumanian locality of
Upper Pontian age (pi. 6, figs. 8-10) confirms his observation on the left valve, but also shows that
the apophysis in two accompanying right valves is in the normal position for Mytilopsis. This is a
mirror image of the situation in M. rhodanica. Unfortunately, lack of material makes it impossible
to discover how constant the difference between valves is in these two species or subspecies. The
possibility that the difference is random, caused by the shells living at different attitudes in clumps,
cannot be dismissed.

Examples of reduction of the apophysis in apparently different lineages during the late Miocene
are illustrated by Papp (1950) and Pana (1962). This suggests that Dreissena may be polyphyletic
and is, perhaps, some justification for the recognition of its three described subgenera.

All these aberrations in the arrangement of the septum and apophysis are perhaps to be expected
at a period during which the Dreissenacea were evolving rapidly. If Mytilopsis gave rise to Congeria
and Dreissena , it seems highly likely that the quite good fossil record should reveal fossils, such as
those discussed above, which may be interpreted as intermediate forms. I will accept, therefore, the
presence of the apophysis, or lack of it, as being a useful taxonomic distinction in most instances.
Morton (1981, p. 39), in discussing the anterior pedal retractor in Dreissena , Congeria and
Mytilopsis , inferred that the introduction of the apophysis was a progressive improvement, and his
fig. 1 1 certainly supports his view that there is an increase in the mechanical efficiency of the
pedal/byssal retractor muscle. It is, however, a reverse of the evolutionary sequence. Mytilopsis ,
with an apophysis, appeared some 40 million years before Dreissena dispensed with this structure,
and is still successful.

Articulation

The family appears to be edentulous. Morton (1970) described and figured the interiors of left valves
of Dreissenomya ( Sinucongeria ) aperta (Deshayes 1838) (BMPD LL18541) and Congeria (C.)
subglobosa Partsch (BMPD LL22134), stating that the former has two rudimentary cardinal teeth
and the latter a poorly developed cardinal covered in life by the anterior adductor. In fact, both
valves of LL18541 possess a single irregular rugose ridge crossing the hinge plate (pi. 6, figs. 1, 2,
5, 6). These ridges, however, are not only far too low to articulate as teeth, but are also in the wrong
position to do so. They are situated more or less opposite each other rather than being offset like
true teeth which fit into corresponding sockets in the other valve. Another specimen (BMPD
LI 9502; pi. 6, figs. 3, 4, 7) and numerous illustrations (Archambault-Guezou 1976u, pis. 4, 5;
Marinescu 1975, pis. 6-8) show wide variation in both size and position of these ridges on the
septum of D. ( S .) aperta. Similar ridges occur in other species of Sinucongeria. Examination of
LL22134 and other specimens of Congeria subglobosa (PI. 3) shows that no cardinal tooth, however
poorly developed, exists. The ridge is not always present, and, as in Sinucongeria , cannot have acted
as a tooth. These ridges are here interpreted, therefore, as some modification to the adductor muscle
attachment area rather than as rudimentary dentition. It is possible that they mark the boundary
between the 'catch' and 'quick' components of the anterior adductor muscle. Yonge and Campbell
(1968), however, in rather briefly discussing the musculature of Dreissena polymorpha (Pallas), do
not mention finding these two separate components of the adductor muscle. An alternative
explanation is that ridges or corrugations might help to make a more secure anchorage between the
muscle and shell.

Marinescu (1975, p. 78, text-figs. 4, 5, 6, 8) stated that there was a rudimentary cardinal tooth in
the right valve of Dreissenomya (s.s.), in contrast to D. ( Sinucongeria ), which he described as
edentulous. His text-fig. 5 illustrates a series of juvenile shells (0-6-1 -3 mm long). The structure.

712

PALAEONTOLOGY, VOLUME 33

situated on the commissure just below the umbo, is in the correct position for a tooth, but (his text-
fig. 5e) also occurs in the left valve. In adult shells (BMPD L19501, L71745) (pi. 5, figs. 1-5) it
appears to match a slight indentation of the commissure below the umbo, and is unlikely to have
been an effective form of articulation. This structure could be equivalent to the flexures of the
commissure in the umbonal region of more typical members of the family, rather than a cardinal
tooth.

The umbones are terminal in the majority of Dreissenacea, including the earliest known. It is only
in the late Tertiary Dreissenomya (s.s.) that they are situated some distance behind the anterior of
the shell, but still well forward. The change in umbonal position can be regarded as a secondary
modification connected with adaptation to an epifaunal mode of life (text-fig. 2b, c).

text-fig. 2. Mode of life of Dreissenacea. Arrows indicate inhalent and exhalent siphons, a. Recent Dreissena
and Mytilopsis , epifaunal, byssally attached to hard object such as pebble or piece of wood. Dead shells at bases
of cluster shown without siphons, b, extinct Dreissenomya (Sinucongeria) and c, D. ( Dreissenomya ), both
infaunal, powerful burrowers, with foot extended. D, extinct Congeria subglobosa Partsch, epifaunal, attached
to hard object by massive byssus. E, extinct Mytilopsis spathulata (Partsch) epifaunal, bysally attached cluster
on dead valve of Congeria subglobosa , dead shells shown without siphons, or with valves not closed. Based on
LL 28243. Upper Miocene, Pannonian, Hennersdorf Brick Pit, 10 km S. of Vienna. All extinct examples from
late Miocene and Pliocene of Paratethys. b and c modified from Marinescu, 1975. All figures approximately

x 0-5.

Mode of life

All known Recent species of both Dreissena and Mytilopsis are epifaunal, living in clusters attached
to each other by the byssus (text-fig. 2a). The cluster or colony (text-fig. 2e, BMPD LL28243) of
the Austrian Pannonian Mytilopsis spathulata (Partsch) on a single valve of Congeria subglobosa
Partsch is evidence that some fossil dreissenids were adapted to a similar mode of life.

NUTTALL REVIEW OF DREISSEN ACEA

713

The habit of forming clusters is probably a reflection of the fact that byssally attached epifaunal
bivalves, such as many Dreissenacea and Mytilacea, often need to exploit every suitable piece of
hard substrate. The pallial line in Dreissena (s.l.), Mytilopsis and Congeria is normally described as
being entire, and the siphons of D. polymorpha (Pallas) were described as short (Yonge and
Campbell 1968; Morton 1969). Such features are indicative of an epifaunal existence. Morton
(1970, p. 567, text-fig. 7) described the pallial line of Mytilopsis sallei (Recluz 1849) as not being
indented by a sinus. Later, on studying living material of the same species, he described it ( 1981, p.
30) as having a slight pallial sinus, an unusual feature in an epifaunal species (p. 37), and (p. 39)
mentioned that its extensible siphons helped it cope with a high sediment load. However, Morton
did not state how far the siphons could extend. These interesting observations led him (1981, p. 30)
to reconsider his views on Congeria zsigmondyi Halavats, from the Pontian of Hungary (1970, p.
565, text-fig. 4), which he had previously described as lacking a pallial sinus and which he had
interpreted as being a byssally attached, infaunal species like the quadrate and globose Congeria
subglobosa Partsch from the Pannonian of the Vienna Basin (1970, p. 565, text-fig. 3). Morton had
argued (1970, p. 569) that, in moving water, the size of the latter species would put a great strain
on the byssal apparatus. It seems unlikely that Congeria was completely infaunal. It has no pallial
sinus, so it may be inferred that the siphons were short as in Dreissena. It would almost certainly
prefer not to be buried sufficiently for the siphons to be permanently covered by sediment, even
though Yonge and Campbell (1968) had described the normally epifaunal D. polymorpha as being
singularly oblivious to the presence of silt because of its ability to expel massive collections of
pseudofaeces. The large size of the byssal notch in C. subglobosa (pi. 3, fig. 4) and other large species
suggests that the byssus may have been extremely well developed and able to cope with bottom
currents, which in the muddy clay facies of the Pannonian were probably not strong. It is difficult
to imagine that these large, heavy Congeria lived in clumps: they were almost certainly solitary and
byssally attached to hard objects such as dead shells (text-fig. 2f).

The pallial sinus which Morton (1981) described in M. sallei (pi. 1, fig. 19, herein) and in C.
zsigmondyi is barely noticeable, consisting of a slight truncation of the posterior portion of the
pallial line. These pallial lines are not indented, and it is difficult to justify Morton’s comparison
(1981, pp. 30, 37) between them and the fully sinupalliate Dreissenomya ( Sinucongeria ) aperta
(Deshayes) (pi. 6, figs. 2, 4). By no means all species assigned to Mytilopsis possess this weak pallial
sinus. Species in which it may be detected include M. leucophaetus (Conrad 1831) (pi. 1, fig. 1 1), the
living type species, and M. basteroti (Deshayes in Lamarck, 1836) from the Lower Miocene of
Aquitaine (pi. 5, fig. 9). It is almost certain that all species of Mytilopsis , whether or not this weak
pallial sinus is detectable, were basically epifaunal and byssally attached. Dreissenomya (s.l ), in
which a definite pallial sinus is present, is thought to have been infaunal. Morton (1970, pp. 564,
568, text-fig. 2) suggested that D. (S.) aperta was infaunal and byssally attached as it possessed a
byssal notch. However, the commissure in this species is so irregular and variable that it is
impossible positively to identify a byssal notch as indicated in Morton’s figure. Marinescu (1975,
p. 90) stated that the gapes in D. (Sinucongeria) were either less important than in Dreissenomya
(s.s.) or absent altogether. He made no mention of the presence of a byssal notch in either subgenus.
Dreissenomya (D.) schroeckingeri (Fuchs 1870c/), on the other hand, possesses pronounced anterior
(pedal) and posterior (siphonal) gapes (Marinescu 1975, p. 78, text-fig. 3). Marinescu also (pp.
105-107) discussed the mode of life of these subgenera. Single valves were less common than
specimens with both valves together, found nearly perpendicular to the bedding plane, presumably
in life position. The indications are that the Pontian D. (D.) schroeckingeri lived at depths of up to
30cms below the sediment surface and that D. ( S .) aperta at about 5-7 cm (text-fig. 1b, c). He
considered that the earlier, Meotian, species of both subgenera were barely covered, as they often
occur as isolated valves, and that juveniles, lacking the posterior gape were also surface dwellers.
According to Prof. W. Russell-Hunter ( pers . comm, in Yonge and Campbell, 1968), juveniles of
living Dreissena use the foot to move about from time to time and produce a weak byssus with only
a few threads for temporary attachment. With growth, the foot becomes less important as an organ
of locomotion and a much stronger, more permanent byssus is formed. It seems likely that juvenile

714

PALAEONTOLOGY, VOLUME 33

fossil dreissenids had similar habits. It is reasonable to suppose that the earlier species of
Dreissenomya (s.l.) followed this pattern, but the later, more deeply buried species differed in not
being byssally attached as adults. They were possibly strong and rapid burrowers, in a manner
comparable to members of the Tellinacea.

Other characters

The opisthodetic ligament and soft parts of D. polymorpha (Pallas) are described in detail (Yonge
and Campbell 1968; Morton 1969). Mytilopsis sallei (Recluz 1849), a member of the only other
living genus, is similarly dealt with (Escarbassiere and Almeida 1976; Morton 1981). The shell
structure of D. polymorpha was studied by Taylor, Kennedy and Hall, (1973) who noted its strong
resemblance to that found in the Corbiculacea, a predominantly freshwater superfamily, whilst
Archambault-Guezou (1982) noted cylindrical perforations which she considered to be tubules in
two other living species, which she placed in Congeria , but are here assigned to Mytilopsis as M.
leucophaetus (Conrad) and M. coch/eatus (Kickx in Nyst, 1835).

Planktotrophic larvae are reported in both living genera (Morton 1969, with further references)
for D. polymorpha (Pallas), and by Escarbassiere and Almeida (1976, p. 167, fig. 14) for Mytilopsis
sallei (Recluz). As Morton pointed out, the retention of a free-swimming larva is unique among
freshwater bivalves. Morton (1969) showed that D. polymorpha possessed a very efficient filtering
system that enabled it to live in both fast-running and in relatively static freshwaters. M. sallei is
similar in this respect and, additionally, is extremely tolerant to salinity changes between 0-50 %G
(Escarbassiere and Almeida 1976; Morton 1981). This must be an important factor in its ability to
be transported across oceans and also in its survival in tropical estuaries and harbours where salinity
can vary enormously between wet and dry seasons.

SYSTEMATIC PALAEONTOLOGY

Repositories. With the exception of figures reproduced from other works, all figures are of specimens in the
British Museum (Natural History). Fossils are in the Palaeontology Department (BMPD) and their
registration numbers are prefixed by either L or LL. Recent specimens are in the Zoology Department
(BMZD), whose registration numbers frequently refer to a whole sample rather than a single specimen.

Subclass HETERODONTA
Order veneroida

Superfamily dreissenacea Gray in Turton, 1840

Diagnosis. Normally mytiliform to modioliform; umbones terminal or nearly so; umbonal septum
or shelf present, but sometimes vestigial ; edentulous; heteromyarian with anterior adductor muscle
attached to septum; gills eulamellibranchiate ; shell aragonitic, outer layer of simple and crossed
lamellae; inner layer of complex crossed lamellae.

Distribution. Lower Eocene to Recent, Europe, Asia Minor and as far east as Kazakhstan (Aral Sea region).
Late Oligocene, Central and tropical South America, spreading northwards into southern and Atlantic
seaboard areas of United States during the Neogene. Introduced to India, Hong Kong and Japan. Probably
introduced to Fiji and West Africa.

Remarks. The characters mentioned in this diagnosis are discussed in the section on morphology
and mode of life, whilst the distribution of the family is dealt with under each genus in turn and in
the discussion.

Family dreissenidae Gray in Turton, 1840
Diagnosis. As for superfamily.

NUTTALL: REVIEW OF DREISSENACEA

715

Subfamily dreisseninae. Gray in Turton, 1840
Diagnosis. Umbones terminal; pallia I sinus absent or weak; byssal gape usually marked.

Remarks. Living Dreisseninae are usually mytiliform, sometimes triangular or quadrate, and have
short, separate, siphons. They are epifaunal, byssally attached in clusters. Distinctions between
them and the Dreissenomyinae are discussed under the latter.

Distribution. As for superfamily.

Genus dreissena van Beneden, 1835u
(text-fig. 2a).

Type species. Mytulus (sic) polymorphus Pallas, 1771, by monotypy. Recent, Europe, (pi. 1, figs. 1-6).

Diagnosis. Mytiliform, smooth except for growth lines, with simple septum on which anterior
adductor and pedal-byssal retractor muscles are mounted; apophysis lacking; pallial line entire.

Distribution. Late Miocene (Meotian)-Recent, Europe, northern Asia Minor and as far east as the Aral Sea,
and the River Euphrates.

Remarks. Numerous fossil species have been recognized particularly in the Paratethys. Many have
very restricted distributions both geographically and stratigraphically and the suspicion remains
that some are of little more than local significance. Carinate mytiliform Dreissena (s.s.) may be easily
distinguished from the non-carinate, more modioliform D. ( Pontodreissena ) and D. ( Modio -
lodreissena ), but the differences between these do not seem great.

Subgenus dreissena van Beneden, 1835a
Plate I, figs. 1-6.

Diagnosis. Dreissena with straight dorsal margin and markedly angular diagonal carina.
Distribution. As for genus (s.l.)

Remarks. Dreissena (s.s.) corresponds to the carinatae- group of Andrussov (1897-8). The living type
species, D. polymorpha is long ranging, first occurring in the Meotian of the eastern Paratethys
(Iljina et al. 1976). Archambault-Guezou (19766) and Locard (1893) gave useful summaries of its
spread from its apparent area of origin. Much of its dispersal through the rivers and canals of
western Europe has occurred very recently, being recorded from the Danube in 1769, England in
1824 and the south of France by 1865-6. This rapid spread probably coincided with increasing trade
and the interconnections of inland waterways by canals and locks. Locard (1883) described two
species from the upper reaches of the Euphrates river system in northern Syria, and later (1893)
recorded the genus from above Baghdad. The Elolotype (BMZD 1984243) of Mytilus tenebrosus
Reeve, 1858, described from the Mississippi, is a damaged D. polymorpha. It is the only record of
the genus from the Western Hemisphere and may be discounted, having either been mislocalized or
transported as ballast.

Subgenus pontodreissena Logvinenko and Starabogatov, 1966
Plate 1, figs. 7-10. plate 2, figs. 6-8.

Type species. Mytilus rostriformis Deshayes, 1838, by original designation. Pliocene, Crimea, (pi. 1, figs. 7-10,
pi. 2, figs. 6-8).

716

PALAEONTOLOGY, VOLUME 33

Diagnosis. Dorsal margin convex; ventral margin slightly concave behind umbone; not tumid,
diagonal ridge much weaker than in D. ( Dreissena ) and mainly confined to umbonal region; septum
broad.

Distribution. Late Miocene (Meotian) to Recent. Paratethys and bordering Mediterranean (see below).

Remarks. The work in which this subgenus was erected has not been seen. Pontodreissena is used
by Starabogatov (1970), Babak (1983) and Rozanov (1986). Pontodreissena corresponds to part of
the modioliformes- group of Andrussov (1897-8): other members of this group were assigned to D.
(Modiolodreissena) by Babak (1983). D. (P.) rostriformis is first recorded from the Meotian of the
eastern Paratethys (Rozanov 1986) though other authors do not recognize it before the Pontian.
Archambault-Guezou (19766) summarizes its distribution. In her opinion, it originated in the
Dacian Basin during the Lower Pontian and persisted there through the Dacian, invading the
eastern Paratethys during the Middle Pontian, continuing there throughout the Pliocene and into
the Quaternary, and living now in the Caspian. It invaded the Mediterranean, being found in strata
of probable Messinian age from Greece, Italy, Sicily, southern France and southern Spain. Other
species are shorter ranging, mainly confined to the eastern Paratethys, and all extinct (Babak 1983).

Subgenus modiolodreissena Babak, 1983
Plate 2, figs. 3, 4.

Type species. Dreissensia theodori Andrussov, 1893, by original designation. Pliocene, Crimea, (pi. 2, figs. 3,
4).

Diagnosis. Like D. (Pontodreissena) but amygdaloidal with both dorsal and ventral margins
convexly curved ; diagonal ridge lacking.

Distribution. Late Miocene (Meotian) - Late Pliocene (Akchagylian), Paratethys only.

EXPLANATION OF PLATE 1

Figs. 1-6. Dreissena (D.) polymorpha (Pallas), sample BMZD 1984242, Recent, Commercial Docks, London,
G.B. Sowerby Colin. 1 , right valve. 2, 3, ventral views (left valve on right of figures) to show different profiles
of commissures in beak region (compare with fig. 8, below), and also byssal notch in fig. 3. 4, left valve. 5,
6, right valve, internal view showing septum, x 2, except 2, 6, both x 5.

Figs. 7-10. Dreissena ( Pontodreissena ) rostriformis (Deshayes 1838), LL 28238/1 left valve. Pliocene, exact
horizon unknown, Kamych-Burun. Kerch Peninsula, Crimea, U.S.S.R. 7, internal, x 1-5, 8, 10, details of
septum and rostrate process from different angles, both x 4. 9, ventral view, showing rostrate process
(compare with fig. 2 above), x L5.

Figs. II 14. Mytilopsis leucophaetus (Conrad 1831) sample BMZD 1984239. Recent, Green Cove Springs,
Black Creek (tributary of St John’s River), Florida, attached to submerged wood, collected live by Messrs
D. C. Marelli and M. J. Greenberg. 11. 12, left valve internal, showing slight pallial sinus, x 5, external view
x 3. 13, right valve internal, x 3. 14, umbonal area of another right valve, tilted to show apophysis, x 10.
(N.B. The umbones of this species invariably seem to be eroded).

Figs. 15, 16, 18, 19. Mytilopsis sal/ei (Recluz 1849). 15, sample BMZD 1984230, left valve of shell figured as
Mytilus sallei by Reeve, 1858, pi. 10, fig. 44, Central America, Cuming Colin, x2-5. 16, left valve of shell
from same sample, detail of septum and apophysis, x 10. 18, sample 1984236, left valve of shell figured as
Mytilus domingensis Recluz (1852) by Reeve, 1858, pi. 100, fig. 44, Dominican Republic, Cuming Colin, x 2.
19, right valve of shell from same sample showing extremely weak pallial sinus, x2-5.

Fig. 17. Mytilopsis africanus (van Beneden 18356), sample BMZD 1984238. left valve of shell figured as
Mytilus africanus by Reeve, 1858, pi. 10, fig. 47, Senegal, Cuming Colin, x F5, cf. figs. 18 and 19.

PLATE I

NUTT ALL, Dreissena, D. (. Pontodreissena ) and Mytilopsis

718

PALAEONTOLOGY, VOLUME 33

Remarks. The broad septum often extends some distance along the dorsal margin. The
modioliformes- group of Andrussov ( 1897-8) is split between this subgenus and D. ( Pontodreissena ):
the distinctions between them do not seem great, but may be useful in the context of Paratethys
geology. Rozanov (1986) records D. rimestiensis (Fontannes 1887) from the Meotian, the species
survives to the top of the Pontian in Rumania. The type species first occurs in the late Pontian
(Kamyschburun, Kertsch Peninsula) and persists into the Akchagylian.

Genus congeria Partsch, 1835
Plate 2, figs 1-3, Plate 3, figs. 1-6, text-fig. 2d.

Type species. Congeria subglobosa Partsch, 1835, subsequent designation by Pilsbry, 1911, Upper Miocene,
Pannonian, Vienna Basin, (pi. 3, figs. 1-6). = Globocongeria Taktakishvili, 1973, p. 14, (objective synonym),
type species Congeria subglobosa Partsch, 1835, by original designation, — Rhombocongeria Starabogatov,
1970, type species Congeria rhomboidea M. Hornes, 1870, by original designation, Upper Miocene, Pontian,
Hungary, (pi. 2, figs. 1-3) = Rhomboidea Marinescu, 1973, (objective synonym of Rhombocongeria ), type
species Congeria rhomboidea Hornes, 1870, by original designation.

Diagnosis. Quadrate to rhomboidal; smooth except for growth lines; apophysis fused to septum;
pallial line entire.

Distribution. Late Miocene (Pannonian to Lower and Middle Pontian), western Paratethys only.

Remarks. There is more than one approach to the classification of dreissenids possessing an
apophysis as well as a septum. The first, with variations, adopted by most Eastern Hemisphere
authors (Andrussov 1897-8; Orlov 1960; Starabogatov 1970; Taktakishvili 1973 ; Marinescu 1973;

EXPLANATION OF PLATE 2

Figs. 1, 2, 5. Congeria rhomboidea M. Hornes, presumed Upper Miocene, Pontian, 1, left valve internal,
Arpad, S.E. of Fiinfkirchen, southern Hungary, from M. Hornes, 1867, pi. 48, fig. 4a. 2, 5, right valve and
anterior view, from Andrussov, 1897, pi. 10, figs. 8, 10. 2, Ogrugliak, Croatia, Yugoslavia, 5, Arpad,
Hungary, all xl.

Figs. 3, 4. Dreissena ( Modiolodreissena ) theodori Andrussov (1893), Pliocene, late Kimmerian or early
Akchagylian, Eisenerzschichton, Kamych-Burun, Kerch Peninsula, Crimea, U.S.S.R., from Andrussov,
1897, pi. 13, figs. 16, 15. 3, right valve internal. 4, left valve, both x 2.

Figs. 6-8. Dreissena ( Pontodreissena ) rostriformis (Deshayes 1838), presumed Pliocene, exact horizon and
locality unknown, Crimea, U.S.S.R. from Deshayes, 1838, pi. 4, figs. 14—16. 6, left valve. 7, ventral view, with
left valve on right of drawing. 8, right valve internal, all x 1 .

Fig. 9. Prodreissensia perrandoi Rovereto, Lower Oligocene, Tongrian, Mioglia, Liguria, Italy, left valve of the
figured syntype, from Rovereto, 1990, pi. 4, fig. 14, x 1-5.

Figs. 10, 11. ‘I Prodreissensia euchroma (Oppenheim 1891), Lower Eocene, Mt. Pulli, near Valdagno, Vicentin,
Italy, figured syntypes, from Oppenheim, 1891, pi. 51, figs. 5a, 6a. 10, x4; 11, x 2.

Figs. 12, 13. Mytilopsis chonioides (Cossmann), Upper Eocene, Priabonian, Le Ruel, Seine-et-Oise, France,
right valve of Cossmann’s (1887, pi. 6, figs. 32, 33) figured syntype, as refigured by Oppenheim, 1891, pi. 51,
figs. 1, la, approx, x 6, xl.

Figs. 14, 15. Mytilopsis curvirostris (Cossmann), Eocene, Priabonian, Marines, Seine-et-Oise, France, right
valve of Cossmann’s (1887, pi. 6, figs. 29, 30) figured syntype, as refigured by Oppenheim, 1891, pi. 51, figs.
2, 2a, approx, x 5, x L67.

Figs. 16, 17, 18. Mytilopsis eocenica (Oppenheim, 1891), Lower Eocene, Dorogh, near Gran, northern
Hungary, from Oppenheim, 1891, pi. 51, figs. 8b, d, f, 16 internal view of right valve umbonal region; 17, left
valve; 18, right valve, magnifications unknown.

Figs. 19, 20. Mytilopsis modiolopsis (Andrussov, 1897-8), Upper Miocene, Meotian, Akmanajsky, Crimea,
U.S.S.R., left valve, internal, external, from Andrussov; 1897, pi. 9, figs. 10, 9, x 3.

PLATE 2

NUTT ALL, Dreissenidae

720

PALAEONTOLOGY, VOLUME 33

Archambault-Guezou 1976a, 6, 1982; Babak 1983) is to place them all in a varying number of
subgenera or ‘groups’ of Congeria Partsch, the first valid name available. The second view is that
adopted by Keen (in Moore 1969) who recognized Mytilopsis as a separate genus with a basically
Western Hemisphere distribution. A modification of these two views, along with, in some cases
radical, revision of the stratigraphic and geographic ranges of the subgenera concerned is adopted
herein. The basic argument for this decision relies on the fact that mytiliform species resembling
living Mytilopsis are the first dreissenids to appear in the Eocene, being recorded from all provinces
in which dreissenids are found at that time. In contrast, the large, heavy and often quadrate rather
than mytiliform species, typified by C. subglobosa , occur only at about the Miocene-Pliocene
boundary in the western Paratethys. Congeria is here regarded as a short-lived offshoot of
Mytilopsis.

Andrussov’s (1897-8) divisions of Congeria into various non-binomial ‘groups’, for some of
which he designated ‘type-species’ have been validated subsequently, and some have junior
synonyms (see Table 2).

Comparatively few species remain in Congeria. Most are illustrated and listed as members of the
subglobosae and rhomboidea groups by Andrussov (1897-8). They are recorded from the Vienna
Basin, Yugoslavia, Hungary and Rumania. C. subglobosa Partsch was first described from the
Plattensee (Lake Balaton) district of Hungary and is common in the Pannonian of the Vienna Basin.
C. rhomboidea M. Hornes, was first described from Hungary and, in common with other species
assigned to C. ( Rhomboidea ) by Marinescu (1973), is confined to the Lower and Middle Pontian
(Odessian and Portaferrian).

Some confusion may exist about the status of Globocongeria. Taktakishvili (1973, p. 14)
designated C. subglobosa Partsch as type species, thus making it an objective synonym of Congeria
(s.s.) itself. Taktakishvili’s interpretation of Congeria (s.s.) was that it corresponded to the
mytiliformes and eocenae groups of Andrussov, both of which are here placed in Mytilopsis.
Unfortunately the Zoological Record ( Mollusca ) for 1978, p. 425 quoted the type species of
Globocongeria as being Congeria digitifera Andrussov, 1897-8. This, in fact, was merely the first
species dealt with by Taktakishvili (1973, p. 36) in his formal descriptions of the Pliocene dreissenids
of western Georgia. All the species assigned by Taktakishvili to Globocongeria belong to
Andrussov’s subglobosae group, and thus to Congeria (s.s.).

Genus mytilopsis Conrad, 1858

Plate 1, figs. 11-19; Plate 2, figs. 12-20; Plate 4, figs. 1-16; Plate 5, figs. 6-13; Plate 6, figs. 8-10; text-figs.

2a, e.

Type species. Mytilus leucophaetus Conrad, 1831, subsequent designation by Dali, 1898, Recent, eastern
U.S.A., (pi. 1, figs. 11-14). = Andrusoviconcha Starabogatov, 1970, type species Congeria modiolopsis
Andrussov, 1897-8, by original designation. Upper Miocene, Meotian, Crimea, (pi. 2, figs. 19, 20). =
Eocongeria Starabogatov, 1970, type species Tichogonia ( Congeria ) eocenica Oppenheim, 1891 (see remarks),
by original designation. Lower Eocene, Hungary, (pi. 2, figs. 16-18). = Mytilia Marinescu, 1973, type species
Congeria neumayri Andrussov, 1897-8, by original designation (preoccupied by Mytilia Gray, 1858, Reptilia
and by Mytilia Gosse in Hudson and Gosse, 1886, Rotifera), = Modiola Marinescu, 1973, type species
Congeria czizeki Hornes, 1867, by original designation (preoccupied by Modiola Lamarck, 1801, Mollusca), =

EXPLANATION OF PLATE 3

Figs. 1-6. Congeria subglobosa Partsch, Upper Miocene, Pannonian, Austria. 1, 2, 3, 5, 6. LL 28239., Brunn,
near Vienna, (ex K.K. Mineralien-Kabinet), 1, left valve, x 1. 2, right valve, internal, x 1. 3, left valve
internal, x 1. 5, left valve, tilted, to show apophysis, x L5. 6, right valve, tilted to show apophysis, x L5.
4, LL 28240/1. Hennersdorf Brick Pit, 10 km S. of Vienna, F. Rogl, B. R. Rosen and J. G. Darrell Colin,
front view showing byssal notch, x 1.

PLATE 3

NUTT ALL, Congo ia subglobosa

722

PALAEONTOLOGY, VOLUME 33

Trigonipraxis Starabogatov, 1970, type species Congeria triangularis Partsch, 1835, by original designation.
Upper Miocene, Pannonian, Hungary, (pi. 4, figs. I -5). = Triangularia Marinescu, 1973, type species Congeria
ornithopsis Brusina, 1892, by original designation, (preoccupied by Triangularia Freeh, 1894, Mollusca).

Diagnosis. Like Dreissena , but possessing an apophysis, usually behind, but sometimes partly
underneath and partially fused to the septum; pallial sinus, if present, very weak.

Distribution. Lower Eocene - Pliocene, Europe; late Oligocene - Recent, central and northern South America
and Caribbean, spreading to North America bordering Gulf of Mexico and Atlantic. Introduced (Recent),
southern India, Japan and Hong Kong, probably introduced, Fiji, West Africa and Rhine - Scheldt Delta.

Remarks. The generic synonymy and its relationship to Andrussov’s (1897-8) ‘groups’ is best
explained by reference to Table 2, whilst the main reasons for placing these taxa in Mytilopsis rather
than Congeria are given in the remarks on the latter genus. Andrussov (1897-8) placed all the living
Western Hemisphere nominal species of Mytilopsis , including the type species M. leucophaetus and
M. sallei (Recluz 1849) (pi. 1, figs. 15, 16, 18, 19), into his ' mytiliformes ’-group, whilst he placed the
living European M. cochleata (Kickx in Nyst, 1835) and the West African M. africanus (van
Beneden 18356) (pi. 1, fig. 17) into his ‘ modioliforntes ’.

It would appear, however, (Nuttall 1990, pp. 282, 284, fig. 335) that M. africanus is synonymous
with the form of M. sallei originally described as M. domingensis (Recluz 1852) (pi. 1, fig. 19) from
the Dominican Republic. It is suggested (S. Morris, BMZD, pers. comm.) that this African
occurrence may well be the result of introduction by man, perhaps during the period of the slave
trade between West Africa and the West Indies. M. cochleata is also likely to have been introduced
to the Rhine-Scheldt Delta from the Western Hemisphere. Variation within species is such that there
is no apparent reason for the separation of Recent ‘ mytiliformes ’ from ‘ modioliformes'.

Starabogatov (1970, p. 83) designated, as type species of Eocongeria , a new subgenus of Congeria ,
Drevssensia (Congeria) eocenica Munier-Chalmas in Hebert and Munier-Chalmas (1877, p. 126),
even though the name appears only in a faunal list and is clearly a nomen nudum. Oppenheim (1891,
p. 953) was the first to describe the species and is here credited with its authorship. Members of the
' eocenae ’-group of Andrussov, upon which Eocongeria is based, are all of Eocene age and exhibit
a wide variation in shell shape, but most are typical of Mytilopsis. D. eocenica from the Lower
Eocene (Eames in Morley Davies, 1975, p. 146) of Dorogh, Hungary, possesses a low angular fold
not normally seen in Mytilopsis , but the form of its septum and apophysis is indistinguishable from
that of modern Mytilopsis. Oppenheim (1891, pi. 51) gave excellent illustrations, including the
internal characters, of this and three other Eocene species, C. chonioides (pi. 2, figs. 12, 13) and C.
curvirostris (pi. 2, figs. 14, 15) both described by Cossmann (1887) from the French upper Eocene,

EXPLANATION OF PLATE 4

Figs. 1-5. Mytilopsis triangularis (Partsch), Upper Miocene, Pontian, Radmanest, Banate, Hungary. 1, 4, 5,
LI 9499/1, left valve, 1, interior, tilted, showing typical Mytilopsis apophysis, x4. A, 4 interior, x 1-5. 5,
exterior, x 1-5. 2, 3, L19499/2, right valve. 2, interior, x 1-5. 3, interior, tilted, x4.

Figs. 6-1 1. All from Miocene, Pebasian, Peru.

Fig. 6. Mytilopsis sallei Recluz (1849), LL 27195, left valve, Canama, x3.

Fig. 7. Mytilopsis scripta (Conrad 1874), LL 27956, left valve, Pichana, x 2-5.

Figs. 8-1 I. Mytilopsis cf. scripta (Conrad 1874), LL 27914, right valve, Canama. 8, external; 9, anterior; 10,
internal, all x 5. 1 1, internal, showing apophysis, x20.

Figs. 12-16. Mytilopsis sowerbyi d’Orbigny (1850), Upper Eocene, Priabonian, Lower Headon Beds, all
Hordwell (except 15, Mead End) Hampshire. 12, LL 28131, Edwards Colin, right valve, x4. 13, LL 28130.
left valve internal x 4. 14, 16, LL 28242/1, 2, Hastings Colin, internal views of umbonal region, tilted to
show apophysis lying under septum. 14, left valve; 16, right valve, both x 10. 15, LL 28244, Edwards Colin,
left valve, x 2.

PLATE 4

NUTT ALL, Mytilopsis

724

PALAEONTOLOGY, VOLUME 33

and of C. euchroma Oppenheim, 1891, (pi. 2, figs. 10, 11) from the northern Italian Eocene.
Andrussov (1897) recorded C. curvirostris as the basal member of his 1 mytiliformes'- group
( = Mytilia ) and C. euchroma along with the British Priabonian D. sowerbyi (d'Orbigny 1850) (pi. 4,
figs. 12-16) as one of the earliest 1 modioliformes' . In the present work C. euchroma is provisionally
placed in Prodreissensia because it possesses radiating ribbing whilst the other species mentioned
above are considered to be typical of Mytilopsis. Thus, there seems as little reason for either generic
or subgeneric separation of these early members of the family, as there is for separating the living
species discussed in the preceeding paragraph.

During the Eocene Mytilopsis is represented by several species (Oppenheim 1891; Andrussov
1 897-8) including M. eocenica (Oppenheim 1891), Lower Eocene, Dorogh, Hungary ; M. curvirostris
and M. chonioides (both Cossmann 1887) rare in the Sables Moyen (Bartonian = Priabonian) of the
Paris Basin (Cossmann and Pissarro 1904-6) and M. sowerbyi (d’Orbigny 1850) from the Lower
Headon Beds (Priabonian), of Hampshire and the Isle of Wight, England (BMPD).

Oligocene species include M. nystiana (d’Orbigny 1852), Tongrian (= Lattorfian), Belgium
(Glibert and de Heinzelin 1954 as Congeria nysti ) and M. aralensis (Merklin 1974), Middle
Oligocene, Aral Sea Region, USSR. M. brardi (Brongniart 1823) was recorded (Wenz 1921) from
the Cerithien-schichten of the Mainz Basin, now dated as Chattian (Martini in Steininger et at. 1985,
p. 334). This species persists into the Miocene, occurring in the Aquitanian Corbicula- Schichten and
overlaying Hydrobien-Schichten (Wenz 1921). Widespread Miocene occurrences include M.
amygdaloides (Dunker 1848), (pi. 5, figs. II 13) probably late Lower Miocene, Wurttemburg,
Germany (BMPD); M. alta (Sandberger 1874), Middle Miocene (Pontilevian), Faluns of Touraine,
France (Glibert and van de Poel 1967); several species from the Aquitanian to Helvetian of the
Aquitaine Basin, France, including (pi. 5, figs. 8-10) M. basteroti (Deshayes in Lamarck, 1836) (see
Cossmann and Peyrot 1914; BMPD); M. sandbergeri (Andrussov 1897-8), Miocene, Poland
(Friedberg 1936); M. sandbergeri , Miocene (Konkian and Tortonian), and several species up to the
Meotian, of the Ukraine (Zelinskaya et al. 1968); M. soceni and M. moesia (both Jekelius, 1944)
Sarmatian, Rumania and also Vienna Basin (Papp 1954; Sieber 1955); M. spathulata (Partsch 1835)
(pi. 5, figs. 6, 7, text-fig. 2e) and several other species, Pannonian, Vienna Basin (Papp 1951, 1953;
Sieber 1955); several species, Meotian, Paratethys (Rozanov 1986); M. subcarinatus (Deshayes
1838), Pontian and Lower Kimmerian, Paratethys, and other species of which the last surviving
were M. mirabilis (Seninsky 1905) and M. caucasica (Seninsky 1905) (see Babak 1983, p. 87), both
of which died out in the late Kimmerian.

In Western Europe, the last surviving, naturally occurring Mytilopsis appears to be M. rhodanica
(Fontannes 1882) from the ‘Infra-Pliocene’ (Messinian) of the Rhone Valley. According to
Archambault-Guezou (1976/5) this species was derived from M. subcarinatus (Deshayes 1838) of the
Paratethys, having migrated into the Mediterranean in the same way as Dreissena (Pontodreissena)
rostriformis (Deshayes 1838) (see above).

EXPLANATION OF PLATE 5

Figs. 1-5. Dreissenomya (D.) schroeckingeri , (Fuchs, I870«), Upper Miocene, Pontian, Radmanest, Banate,
Hungary. 1, 2, L71745, left valve interior, 1, x4; 2, x 15. 3-5, LI 9501. right valve, 3, 4, x 15; 5, x 4.

Figs. 6, 7. Mytilopsis spathulata (Partsch), L23908/1, left valve. Upper Miocene, Pannonian, Brunn, near
Vienna, Austria. 6, detail of internal umbonal region, tilted, apophysis attached to posterior of septum (right
valve has similar arrangement), x 10. 7, x 2.

Figs. 8-10. Mytilopsis basteroti (Deshayes 1836), Lower Miocene, Aquitanian or Burdigalian, Dax, Landes,
France, Deshayes Colin. 8, 9, LL 28240/1, left valve, 8, internal view of worn umbonal region, x 10. 9, to
show weak pallial sinus, x2-5. 10, LL 28240/2, left valve, x 2.

Figs. 1 1-13. Mytilopsis amygdaloides (Dunker, 1848), probably late Lower Miocene, Unterkirchberg, Bavaria,
L1266/1, left valve, II, 12, x2-5. 13, tilted to show apophysis, x 10.

PLATE 5

NUTT ALL, Dreissenomya and Mytilopsis

726

PALAEONTOLOGY, VOLUME 33

The Western Hemisphere distribution of Mytilopsis , the only dreissenid to reach the New World,
is discussed in detail by Nuttall (1990, pp. 278-280, figs. 319-325). The first occurrences, some thirty
million years later than the appearance of the genus in Europe, are of M. trigalensis (Olsson 1931)
from the late Oligocene of the Pacific coastal region of Peru and of M. dalli (Clerc in Joukowsky,
1906) from rocks of similar age in Western Panama. M. trigalensis may well be conspecific with the
later (Miocene, but now extinct) M. scripta (Conrad 1874) (pi. 4, fig. 7) from the Pebas Beds of the
Upper Amazon Valley. Until the appearance of the Panama landbridge in the Pliocene, South
America was an island with the result that the Caribbean and tropical eastern Pacific areas formed
one faunal province. The presence of the living Caribbean M. sallei (Recluz 1849) (pi. 4, fig. 6) in
the Pebas Beds and other Miocene occurrences of M. scripta in the Magdalena Valley and at La
Tagua in the Oriente of Colombia suggest that Mytilopsis invaded the Upper Amazon Valley from
the Caribbean. M. dalli is very similar to M. sallei and may be synonymous. M. sallei also occurs
in the Miocene of the Dominican Republic. The living type species of Mytilopsis , M. leucophaetus
(Conrad) is not known before the Pliocene. It is recorded as M . jamaicensis (Woodring 1925) from
the Bowden Beds (now known to be Pliocene) of Jamaica. A series of shells (LL28 109-29) from the
Plio-Pleistocene Caloosahatchee Formation of Florida originally identified as Congeria lamellata
Dali (1898) appears to be a mixture of M. sallei and M. leucophaetus. Detailed distribution maps
of the two species at the present day are given by Marelli and Gray (1983). They show that they
overlap in this region nowadays and that M. leucophaetus now occurs as far north as the Hudson
River. The apparent lack of fossil records from the eastern seaboard of the United States points to
this northward extension of its distribution being the result of introduction.

A distribution map given by Starabogatov (1970, fig. 12) erroneously shows Mytilopsis as
occurring over much of north-eastern South America including the Amazon Basin and the northern
Brazilian coastal strip as far south-east as Recife (Pernambuco). Dunker (1853), who described
Tichogonia rossmaessleri (a junior synonym of M sallei ) from Pernambuco, was himself very
dubious about the locality information, and this record can safely be discounted. Although M. sallei
is widespread on Caribbean islands (Marelli and Gray 1983), it is not encountered further eastward
and its presence is not mentioned in the very thorough review of the Surinam fauna (Altena 1971).

On the Pacific coast, Mytilopsis is known only from northern South America (Olsson 1961 ; Keen
1971) at the present day. It is probable that Fijian occurrences (Dali 1898; Hertlein and Hanna
1949) are the result of introduction, though clearly not primarily as a result of the opening of the
Panama canal (in 1914) as suggested by Morton (1981). Morton’s suggestion that M. sallei has been
more recently introduced to southern India is supported. The species has subsequently been
reported in Victoria Harbour, Hong Kong (Huang and Morton 1983) and in Shimizu Harbour,
Honshu, Japan (T. Habe, pers. comm, in Huang and Morton), and in Tokyo Bay (Furuse and
Hasegawa 1984). There is the possibility, however, that more than one species of Mytilopsis has
been introduced westwards across the Pacific from the Americas. Marelli and Gray (1985) placed
the Fijian M. alleyniana Hertlein and Hanna ( 1949) in the synonymy of M. adamsi Morrison ( 1946),
originally described from the Pearl Islands, Panama.

The occurrences of M. cochleatus (Kickx in Nyst, 1835) in the Rhine-Scheldt Delta (Adam 1960;
Wolff 1969) and of several nominal West African species (Pilsbry and Bequaert 1927, Congo;
Binder 1958, 1968, Ivory Coast; Reeve 1857-8, Senegal) are also thought to be the result of
introduction.

Records of Mytilopsis from the East Indies (Keen in Moore, 1969) and Indonesia (Eames in
Morley Davies, 1971) are probably based on the septate Sinomytilus Thiele ( 1 934, p. 80 1 ), described
as a section of Mytilus whose type species is Dreissenci crosseana L. Morlet (1884). Specimens
(BMZD 1906.9.19.76-77) of this species and of D. massiei L. Morlet (1892) (BMZD 1902.3.22.55),
both from Cambodia, confirm Thiele’s assignment to the Mytilidae. Their inner shell layer appears
to be nacreous.

Trigonipraxis , (pi. 4, figs. 1 -5) (the ’ triangulares ’ group of Andrussov 1897-8) is well covered in
other literature (Brusina 1892; Jekelius 1944; Papp 1953; Sieber 1955; Marinescu 1973). It is
confined to the Pannonian and Pontian of the Pannonian and Dacic Basins. Differences between

NUTTALL: REVIEW OF DREISSENACEA

727

nominal species assigned by these authors to this group are mostly slight and there seems to be no
compelling reason for subgeneric separation from Mytilopsis. This view is supported by the
occurrence of rather similar trigonal-shaped Mytilopsis (pi. 4, figs. 8- 1 1 ) in Miocene brackish-water
deposits in South America (Nuttall 1990, pp. 285-6, figs. 347, 348).

Genus prodreissensia Rovereto, 1898
Plate 2, fig. 9.

Types species. Dreissensia ( Prodreissensia ) perrandoi Rovereto, 1898, by original designation. Lower Oligocene,
Tongrian, northern Italy, (pi. 2, fig. 9).

Diagnosis. Resembling Dreissena but with sculpture of radiating costules; septum present, shell
interior otherwise unknown.

Distribution. ?Lower Eocene and Lower Oligocene (Tongrian), northern Italy only.

Remarks. Rovereto (1898, pp. 159, 174) described his single, poorly preserved species from the
Tongrian of Mioglia (Liguria) as the type of a new ‘section' of Dreissena characterized by
longitudinal ribbing. The only known illustration is Rovereto’s later (1900, pi. 4, fig. 14) external
view of his type specimen. Both this and the original description were copied by Sacco (1904, p. 152,
pi. 29, fig. 12) who was unable to add any new information about this taxon. The illustration clearly
shows both the faint traces of radiating ribs and a strong carinate diagonal ridge. The original
specific description mentioned the presence of an area (septum), but gave no other details of the
interior. Rovereto (1898, p. 159), presumably because of its strong carination, had compared his
species with the much younger Dreissena gnezdai Brusina, 1884, which was described from near
Zagreb and placed by Andrussov (1897-8) in his Congeria ‘ triangulares ’ group. All other members
of Rovereto’s rich bivalve fauna are undoubtedly marine, raising the suspicion that this species
might be a septate member of the Mytilacea.

Tiehogonia (Congeria) euchroma Oppenheim (1891) (pi. 2, figs. 10, 11) from the Lower Eocene,
Lignite Marls of Mount Pulli (Vicentin, North Italy) seems to be the only definite species of the
Dreissenidae with similar radiating ribbing. Oppenheim's description mentioned the presence of a
septum but stated that the Schloss (hinge) and apophysis were as yet unknown. Andrussov (1900,
pi. 3, figs. 23, 24) provided photographic illustrations of the exterior of two of Oppenheim’s original
specimens. The ribs are faint and the diagonal ridge is low and definitely not carinate. The outline
is mytiliform as in typical Mytilopsis and Dreissena. It is possible that T. ( C .) euchroma and P.
perrandoi may be related in spite of the differences in shape.

The systematic position of Prodreissensia perrandoi cannot be properly assessed because of lack
of knowledge of internal details of its shell. Its Palaeogene age, however, would suggest that it is
more likely to be an offshoot of Mytilopsis rather than Dreissena , if it is, in fact, a member of the
family. T. (C.) euchroma , on the other hand, is definitely a dreissenid.

Subfamily dreissenomyinae Babak, 1983

Diagnosis. Shell modioliform, compressed; umbones not terminal; apophysis lacking; sinupalliate;
posterior siphonal gape present.

Distribution. Late Miocene-early Pliocene (Meotian-late Pannonian) to top Pliocene (Rumanian), Paratethys
only.

Remarks. In comparison with nearly all Dreisseninae the shell is modioliform rather than
mytiliform and is laterally compressed. Exceptions include species such as Mytilopsis amygda/oides
(Dunker) (pi. 5, figs. II 13), which is slightly modioliform but not compressed, whilst Dreissena

728

PALAEONTOLOGY, VOLUME 33

(Pontodreissena) (pi. 1, figs. 7-10) and D. ( Modiolodreissena ) (pi. 2, figs. 3, 4), though fairly
compressed, are basically mytiliform with terminal umbones. In Dreissenomyinae a byssal notch is
lacking and an antero-ventral pedal gape may sometimes be observed. The commissure, however,
is so irregular that these features, interpreted as adaptations to an infaunal mode of life, are not
always readily apparent in most specimens, particularly separated valves. They are therefore
omitted from the subfamilial diagnosis.

Babak (1983, p. 60) referred to Dreissenomyinae Taktakishvili, 1973. I have been unable to find
any reference to this new subfamilial name either in Taktakishvili's work, or in the 1978 Zoological
Record which covers his paper. Authorship is therefore tentatively ascribed to Babak (1983).

Genus dreissenomya Fuchs, 18706

Types species. Congeria schroeckingeri Fuchs, 1870a, by original designation. Upper Miocene, Pontian,
Hungary, (pi. 5, figs. 1-5, text-figs. 2b, c; 3a, b, d, e).

Diagnosis. As for subfamily.

Distribution. As for subfamily.

Subgenus dreissenomya Fuchs, 18706
Plate 5, figs. 1-5, text-figs. 2c, 3a, b, d, e.

Diagnosis. Umbones well back from anterior of shell; thin-shelled; both antero-ventral and
posterior gapes marked; septum reduced to strong ridge, with anterior adductor and pedal/byssal
scars lying partly on and partly anterior to ridge; pallial sinus deep.

Distribution. Late Miocene - early Pliocene (Meotian - late Pannonian) to late Pontian, Paratethys.
Remarks. See under D. ( Sinucongeria ).

Subgenus sinucongeria Lorenthey, 1894
Plate 6, figs. 1-7, text-fig. 2b. 3c, f.

Type species. Congeria arcuata Fuchs, 1870a, by monotypy. Upper Miocene, Pontian, Hungary, (text-fig 3c,
f).

Diagnosis. Similar to Dreissenomya s.s. but thicker shelled; umbones almost terminal; valves more
tumid; true septum present, usually bearing irregular ridges; pallial sinus not very deep.

EXPLANATION OF PLATE 6

Figs. 1-7. Dreissensomya (Sinucongeria) aperta (Deshayes 1838). 1, 2, 5, 6, LL 18451. Upper Miocene, Upper
Pontian ( Bosphorian), Stoichitza Valley, Cocorova-Mehedintzi, Rumania, F. Marinescu Colin, (left valve
figured Morton 1970. text-fig. 2). 1, 2, left valve interior, x4, x F5. 5, 6, right valve interior, x 4, x L5. 3,
4, 7, L19502, probably Pliocene (Kimmerian or Akchagylian), Crimea, U.S.S.R., right valve. 3, 4, interior
x 4, x F5. 7, exterior, x F5.

Figs. 8-10. Mytilopsis subcarinata botenica Andrussov (1897-8), ?Top Miocene-basal Pliocene, Upper
Pontian, Basinul Govorei, Rumania, ex Muzuel Istoria Naturala “Gr. Antipa”, Bucharest, 8, 9, LL 9023,
8, left valve internal, tilted, with anterior pedal/byssal attachment fused to posterior wall of septum as in
Congeria , x 3. 9, same external, x I -5. 10, LL 9022, right valve internal, tilted, with 'normal' Mytilopsis
apophysis, x 3.

PLATE 6

NUTTALL, Dreissenomya (Sinucongeria) and Mytilopsis

text-fig. 3. All from Upper Miocene, Pontian, Radmanest, Banate, Hungary, a, b, d, e. Dreissenomya (D.)
schroeckingeri (Fuchs), from Fuchs, 1870/5, pi. 16, figs. 5-8, x 1. c, f. Dreissenomya ( Simtcongeria ) arcuata

(Fuchs), from Fuchs. 1870c/, pi. 16, figs. 12, 13, x2.

Distribution. Late Miocene (Meotian - late Pannonian) to top Pliocene (Rumanian), Paratethys.

Remarks. Dreissenomya (s.s.) barely penetrates into the Euxinic Basin, the eastern limit of its
distribution being the western Black Sea Coast. Simtcongeria , however, is represented there by the
long-lived (Pontian-Rumanian), and comparatively widely distributed, D. (S'.) aperta (Deshayes
1838) (pi. 6, figs. 1-7) which occurs in the Crimea, Kerch Peninsula and as far east as western
Georgia, as well as in the Dacian and Pannonian Basins. Marinescu (1975) assigned eight species
to Dreissenomya (s.s.) and seven to Simtcongeria. Although Dreissenomya (s.s.) with its markedly
non-terminal umbo, greatly reduced septum and deeper pallial sinus, would appear to be further
removed from the standard dreissenid pattern than Simtcongeria , both subgenera first appear in the
Meotian. Marinescu postulated descent from Congeria (Andrnssoviconcha) novorossica Sinsow
(1877). Archambault-Guezou (1976/7, pis. 4, 5) provided numerous illustrations of D. (S.) aperta
(Deshayes) including syntypes, and of D. (D.) shroeckingeri.

DISCUSSION

The classifications set out by Keen in Moore (1969), by Babak (1983, p. 60) and that proposed
herein may be compared in Table 1. So little is known about Prodreissensia, based on poorly
preserved material, that it is idle to speculate on its relationship with the other genera. Its
Palaeogene age, however, does suggest that it is either synonymous with, or an early offshoot from,
Mytilopsis, if indeed it is a member of the Dreissenidae. Keen’s placement of Dreissenomya and
Simtcongeria as subgenera of Dreissena , on the grounds that they lack the apophysis, is not
accepted. It is, however, worth remembering that all three were probably derived from species of
Mvtilopsis living at more or less the same time and place: the late Miocene ot the Paratethys.
Babak’s classification is an emended version of that proposed by Starabogatov (1970).

The stratigraphic ranges of genera are summarized in Text-fig. 1. Only the two surviving genera,
Mvtilopsis and Dreissena are comparatively long-ranging. The remainder are confined to the
Paratethys during the Neogene and have short ranges: it is estimated that none survived beyond 5
million years, and that Dreissenomya (s.s.) may have existed for as little as 2 million years.

NUTTALL REVIEW OF DREISSENACEA

731

Similarities in shell structures (Taylor, Kennedy and Hall 1973) provide strong evidence that the
Corbiculacea and Dreissenacea shared common ancestry. No early Tertiary fossils showing
intermediate characters have been found.

There are two possible and inter-related explanations. First, the changes from a shallow infaunal
to byssally attached epifaunal mode of life could have occurred quite rapidly. The later sudden
radiation shown by the superfamily in the late Miocene of the Paratethys, including re-adaptation
to an infaunal mode of life by Dreissenomya , suggests that such an explanation is plausible.
Although Mytilus has small, crenulate, dysodont teeth (Nuttall, fig. 46, 5 in Moore, 1969) situated
at the anterior extremity of the shell, loss of dentition in dreissenids may well be connected to the
adoption of a heteromyarian, mytiliform shell suitable for epifaunal life. The narrow, pointed,
umbonal region allows little space for hinge teeth, whose efficiency would be likely to be very limited
in such an anterior position. In addition, the possession of a ligament analogous to that of Mytilus ,
which Trueman (p. N64 in Moore, 1969) described as being among the most powerful known, may
also reduce the need for dentition.

The second possible explanation is that, as a general rule, the fossil record of freshwater and
marginal marine faunas is far less continuous than those of shelf seas. In the late Cretaceous and
Palaeogene of Europe, there were no extensive freshwater basins of comparable extent to the
Neogene Paratethys.

text-fig. 4. a. Palaeogene distribution of Mytiiopsis. Key: circles. Eocene, squares, Oligocene. b. Lower and
Middle Miocene distribution of Mytiiopsis in Europe and western Asia. Key: circles, western Europe; stipple,
throughout Paratethys. c. Late Miocene (Messinian) distribution of Dreissenacea in Europe and western Asia.
Key: circles, Dreissena ; triangles, Mytiiopsis ; stipple, Paratethys with Mytiiopsis and Dreissena throughout;
Dreissenomya and Dreissenomya ( Sinucongeria ) mainly in western and central areas. D. Recent world

distribution of Dreissena , shown stippled.

732

PALAEONTOLOGY, VOLUME 33

Mytilopsis sowerbyi (d'Orbigny) the only known English Caenozoic species of Dreissenidae
occurs quite commonly in freshwater horizons of the English Priabonian Headon Beds, but these
beds are confined to comparatively small areas in Hampshire and the Isle of Wight. This sporadic
distribution pattern is repeated elsewhere throughout most of the Tertiary (text-fig. 4). Mytilopsis
brardi (Brongniart), accompanied by small hydrobiid snails, crowd some bedding planes in
freshwater Oligo-Miocene deposits of the Mainz Basin of Germany. Glibert and de Heinzelin ( 1954,
pp. 388-9) found only 54 individuals of M. nystiana (d’Orbigny) among 90,000 specimens from two
marine horizons in the Belgian Lower Oligocene. Most of the other rarities were also non-marine
species. Similarly, M. basteroti (Deshayes) is a comparatively rare constituent of the rich marine
inshore faunas of the Lower Miocene in the Aquitaine Basin, occurring at localities thought to have
been situated only a few kilometres from the then shoreline. All the specimens of this species that
I have examined are more worn than those of marine species in these faunas, suggesting transport
after death (pi. 5, figs. 8-10). The genus occurs in much the same way in the Middle Miocene Faluns
of Touraine.

Recognition of chance invasions and local extinctions is of importance in interpreting the history
of the superfamily. Such events include:

1. The invasion by Mytilopsis of tropical America by the late Oligocene (text-fig. 5), some thirty
million years after its first appearance in Europe, probably as adult shells attached to logs, rather
then as veliger larvae.

text-fig. 5. Fossil occurrences of
Mytilopsis in western Hemisphere.

Key: stars, late Oligocene; circles,

Miocene; squares. Pliocene.

2. The local extinction of Mytilopsis in western Europe during the Middle Miocene, preceding
its eventual disappearance from the Paratethys.

3. The short-lived invasion, during the Messinian, of the Mediterranean region from the Black
Sea, by Paratethyan species of both Mytilopsis and Dreissena ( Pontodreissena ). Although Mytilopsis
survived well into the Kimmerian (Lower Pliocene) in the eastern Paratethys and D. (P.) rostriformis
is still living there, neither is represented by descendants in the western European Pliocene and
Pleistocene. As neither genus survived along the borders of the Mediterranean, their disappearance
is unlikely to have been due to competition between them. It seems far more likely that their local
extinction was due to the removal of suitable fresh and brackish-water habitats.

4. The Recent considerable extension of the geographical distribution of dreissenids as a result
of human activity. The industrial revolution, linking of waterways and increase in water-born trade
appears to be responsible for the invasion of much of Europe, via the Danube, since the mid-
eighteenth century. It is suggested that Mytilopsis probably increased its range from Florida
northwards along the United States eastern seaboard in similar fashion, whilst shipping has led to
its appearances in Fiji, India, Hong-Kong, Japan, West Africa and the Rhine-Scheldt delta (text-
fig. 6).

Offshore marine molluscs occupy an extremely stable environment in comparison to the habitats
of freshwater molluscs. The history of the Dreissenidae illustrates the success of the highly

NUTTALL: REVIEW OF DREISSEN ACEA

733

text-fig. 6. Recent distribution of Mytilopsis. Key: stipple, naturally occurring; black circles, presumed

introduced.

conservative but, most importantly, salinity-tolerant Mytilopsis which has survived since the early
Tertiary. Nevertheless, its invasion of the Western Hemisphere during the Oligocene appears to
have been chance, whilst its disappearance from Europe (except by re-introduction) shows its
vulnerability to changed conditions. The rapid diversification of the family in the Paratethys during
the late Miocene when Dreissena , Congeria and infaunal Dreissenomya all appeared, is a startling
example of fast evolution. The equally rapid extinction of all but Dreissena as conditions changed
with the break-up of Paratethys shows how vulnerable freshwater molluscs can be when their
habitats are altered. Both Mytilopsis and Dreissena appeared to colonize new regions very slowly
unless influenced by man. This inability to spread quickly in natural freshwater conditions is
another source of long-term vulnerability. The veliger larva, which allows rapid dispersal in the sea,
and even in estuarine conditions successfully exploited by Mytilopsis , is still of some advantage in
lakes. However, in stretches of river unaffected by tides, veligers are unlikely to be nearly as
successful in migrating upstream as the unionacean glochidian larvae attached to the gills of fish.

Acknowledgements . I am indebted to Dr D. Kadolsky (Texaco, London) for first drawing my attention to the
work of Andrussov (1897-8). I would like to thank my colleagues at BM(NH), in particular Dr N. J. Morris
(Palaeontology) and Dr J. D. Taylor (Zoology) for helpful discussions, Mr R. Croucher, Palaeontology
Laboratory for his dissections of very fragile Dreissenomya and Sinucongeria , and members of the
photographic studio who were responsible for all the illustrations other than S.E.M. pictures, taken by myself.
Finally, I would like to acknowledge my appreciation of the late Dr Myra Keen, one of the group of major
contributors to the bivalve volume of the Treatise on Invertebrate Paleontology , without whose efforts that
work would never have been completed.

REFERENCES

adam, w. 1960. Faune de Belgique , Mollusques , /, Mollusques terrestre et dulcioles. 402 pp. Institut Royal des
Sciences Naturelles de Belgique; Brussels.

altena, c. o. van regteren 1971. The marine Mollusca of Suriname (Dutch Guiana), Holocene and Recent.

Part II. Bivalvia and Scaphopoda. Zoologische Verhandelingen , Leiden , 119, 99+1 pp.
andrussov, n. 1893. Bemerkungen fiber die Familie der Dreissensiden. Zapiski Novorossiiskago Obshchestva
Estestvoispytatelei, 18, 69-94. [In Russian],

734

PALAEONTOLOGY, VOLUME 33

1897-8. Fossile und lebende Dreissensiidae eurasiens. (25 plates, 1897; text. 1898) Trudy St.
Petersbwgskago Obshchestva Estestvoispytatelei , (Otdyelenie Geologii i Mineralogii) 25, 683+115 pp. [In
Russian with German summary].

1900. Fossile und lebende Dreissensiidae Eurasiens. Erstes Supplement. Trudy St. Petersburgskago
Obshchestva Estestvo isp it at el e i , (Otdyelenie Geologii i Mineralogii) 29 (5), 60-132. [In Russian with German
summary.]

arch ambault-guezou, J. 1976m Etude de Dreissenidae du Neogene europeen et revue stratigraphique des
niveaux correspondants de la Paratethys. Travaux du Laboratoire de Paleontologies Faculte des Sciences ,
Universite de Paris , Orsay , 361 pp.

1976/5. Presence de Dreissenidae euxiniques dans les depots a Congeries de la vallee du Rhone et sur
pourtour du bassin mediterraneen. Implications biogeographiques. Bulletin de la Societe Geologique de
Frances (7) 18, 1267-1276.

— 1982. Comparaison microstructurale des tests de diverses especes actuelles des genres Dreissena et
Congeria (Dreisseniidae, Mollusca Bivalvia). Malacologia , 22, 325-332.

babak, e. v. 1983. Pliotseno i Chetverti'chnye Dreissenid (Dreissenidae, Bivalvia) Evksinskogo Bassema
(Pliocene and Quarternary Dreissenidae of the Evksin Basin) Trudy Paleontologicheskogo Instituta,
Akademiya Nauka SSSRs 204, 104 pp. [In Russian],

beneden, p. j. van. 1835m Histoire naturelle et anatomique du Dreissena polymorpha genre nouveau dans

la famille de mytilacees. Bulletin de I' Academic Royale des Sciences, des Lettres et des Beaux Arts de Belgique ,
2, 25-26, 44-47.

1853/). Sur une nouvelle espece du genre Dreissena. Bulletin de I’Academie Royale des Sciences, des Lettres
et des Beaux Arts de Belgique, 2, 166-9.

binder, e. 1958. Mollusques aquatiques de Cote d’Ivoire. Bulletin de Tlnstitut Frangais d’Afrique Noire, ser. A.
20, 82-89.

1968. Repartition des mollusques dans la lagune Ebrie (Cote d’Ivoire). Cahiers de I'Office de la Recherche
Scientifique et Technique Outre-Mer, series Hydrobiologie , 2 (3^1), 2-34.
brongniart, a. 1823. Memoire sur les terrains du sediment superieur calcareo-trappeens de Vicentin , Paris,
iv + 86 pp.

brusina, s. 1884. Die Fauna der Congerienschichten von Agram in Kroatien. Beitrage :ur Paldontologie
Osterreich-Ungarns und des Orients, 3, 125-187.

1892. Ueber die Gruppe der Congeria triangularis. Zeitschrift der Deutschen Geologischen Gesellschaft, 44,
488-497.

conrad, t. a. 1831. Description of fifteen new species of Recent and three of fossil shells, chiefly
from the coast of the United States. Journal of the Academy of Natural Sciences of Philadelphia , 6,
256-268.

1858. Description of a new genus of the family Dreisseniidae. Proceedings of the Academy of Natural
Sciences of Philadelphia, (1857), 167.

— 1874. Remarks on the Tertiary clay of the upper Amazon. Proceedings of the Academy of Natural Sciences
of Philadelphia, (1874), 25-32.

cossmann, m. 1887. Catalogue illustre des coquilles fossiles de l’Eocene des environs de Paris. Annales de la
Societe Royale Malacologique de Belgique, 22, 5-2 1 1 .

— and peyrot, a. 1914. Conchologie neogenique de l’Aquitaine, Anisomyaria. Actes de la Societe Linneenne
de Bordeaux, 68, 5-210.

and pissarro, G. 1904-6. Iconographie complete des coquilles fossiles de /' Eocene des environs de Paris. I .
Pelecypodes. Pissarro, Paris, 45 pis.

dall, w. H. 1898. Tertiary fauna of Florida. Transactions of the Wagner Free Institute of Philadelphia, 3, vii,
viii, 591-947.

deshayes, g. p. 1836. in lamarck, j. b. p. a. de. Histoire naturelle des animaux sans vertebres (Ed. 2 by
Deshayes, G. P. and Milne- Edwards , E.), 7, Histoire des mollusques. Bailliere, Paris, vi + 735 pp.

1838. in verneuil, e. de and deshayes, g. p. Memoire geologique sur la Crimee: observations sur les
fossiles de cette peninsule. Memoires de la Societe Geologique de France, 3 (1), 69 pp.
dunker, w. 1848. Ueber die in der Mollasse bei Giinzburg vorkommenden Conchylien und Pflanzenreste.
Palaeontograpliica, 1, 155-168.

1853. Neue Mytilaceen. Zeitschrift fur Malakozoologie , 10, 82-92.
escarbassiere, r. m. and almeida, p. 1976. Aspectos biologicos y ecologicas de Mytilopsis sallei Recluz
(Bivalvia-Eulamellibranchia) en areas adyacentes a la Laguna de Unare. (Estadi Anzoategui, Venezuela).
Acta Biologica Venezuelica 9, 165-193.

NUTTALL: REVIEW OF DREISSENACEA

735

fontannes, F. 1879-1882. Les invertebres du Bassin Tertiaire du Sud-Est de la France. Les mollusques Pliocenes
de la vallee du Rhone et du Roussillon. 2. Acephales. 320 pp, Georg, Lyon and Savy, Paris.

— 1887. Contribution a la faune malacologiques des terrains Neogenes de la Roumanie. Archives du
Museum d’Histoire Naturelle de Lyon , 4, 321-365.

frech, f. 1894. Ueber das Devon der Ostalpen. Ill (Die Fauna des unterdevonischen Rilfkalkes. 1). Zeitschrift
der Deutschen Geologischen Gesellschaft , 46, 446-479.

friedberg, w. 1936. Mieczaki Miocenskie ziem Polskich (Mollusca miocaenica Poloniae). //, Lamellibranchiata.

Polskie Towarzystwo Geologiczne, Krakow, (2), 159-283.
fuchs, T. 1870«. Beitrage zur Kenntniss fossiler Binnenfaunen. Ill, Die Fauna der Congerienschichten von
Radmanest im Banate. Jahrbuch der Kaiserlich- Koniglichen Geologischen Reichsanstalt , 20, 343-364.

18706. Ueber Dreissenomya ein neues Bivalvgenus aus der Mytilaceen. Verhandlungen der
Kaiserlich - Koniglich Zoologischen-Botanischen Gesellschaft in Wien , 20, 997-998.
furuse, k. and hasegawa, k. 1984. Mytilopsis sallei found in Tokyo Bay. The Chirobotan , 15 (1), 18.
glibert, m. and de heinzelin de braucourt, j. 1954. L’O/igocene inferieur Beige, in Volume Jubilaire Victor
van Straelen, 1, Institute Royal des Sciences Naturelles Belgique, Brussels, 281-438.

-and van de poel, l. 1967. Les Bivalvia fossiles du Cenozoique etranger. V. Oligodontina. Memoires de
I'Institut Royal des Sciences Naturelles de Belgique , 83, 152 pp.
gray, J. e. 1858. On a new genus and several new species of Uropeltidae, in the collection of the British
Museum. Proceedings of the Zoological Society of London , 26, 260-265.
halavats, J. 1883. Palaeontologische Daten zur Kenntniss der Fauna der sudungarischen Neogen -
Ablagerungen, I, Die pontische Fauna von Langenfeld. Mitteilungen aus dem Jahrbuch der Koniglich
Ungarisclien Geologischen Anstalt , 6, 162-173.

hebert, e. and munier-ch almas, e. c. p. a. 1877. Recherches sur les terrains tertiares de l’Europe meridionale.
Premiere partie. Terrains tertiare de la Hongrie (Bakony, Gran, Buda-Pesth). Comptes rendus hebdomadaires
des Seances de I'Academie des Sciences , 25 (Juilet-Decembre 1877), 125-128.
hertlein, L. G. and hanna, G. D. 1949. Two new species of Mytilopsis from Panama and Fiji. Bulletin of the
Southern California Academy of Sciences , 48, 13-18.

hornes, m. 1859-1870. Die fossilen Mollusken des Tertiaer-Beckens von Wien, II : Bivalven. Abhandlungen der
Kaiserlich- Koniglich Geologischen Reichsanstalt , 4, 478 pp and Atlas, (pp. 343-430, pis. 45-67, 1867).
huang, z. G. and morton, b. 1983. Mytilopsis sallei (Bivalvia : Dreissenoidea) established in Victoria Harbour,
Hong Kong. Malacological Review , 16, 99-100.

Hudson, c. t. and gosse, p. h. 1886. The Rotifera; or wheel animacules, 2, Longmans, Green, London, 144 pp.
iljina, L. b., nevesskaja, l. a. and paramanova, n. p. 1976. Regularities of mollusc development in the
Neogene semimarine and brackish water basins of Eurasia (late Miocene - early Pliocene). Trudy
Paleontologicheskogo Instituta , Akademiya Nauk SSSR , 155, 288 pp. [In Russian],
jekelius, E. 1944. Sarmat und Pont von Soceni (Banat). Memoriile Institutuli Geologic al Romaniei , 5, 167 pp.
joukowsky, e. 1906. Sur quelques affleurements nouveaux, de roches tertiares dans lTsthme de Panama.

Memoires de la Societe de Physique et d'Histoire Naturelle de Geneve , 35, 155-178.
keen, m. 1971. Sea shells of tropical West America , Ed. 2. Stanford University Press, California, x + 1064 pp.
kickx, p. in nyst, p.-h. 1835. Notice sur une nouvelle espece du moule trouvee a Anvers. Bulletin de I'Academie
Royale des Sciences , des Lettres et des Beaux- Arts de Belgique , 2, 235-6.
kochansky-devide, w. and sliskovic, t. 1978. Miocenske kongerije Hrvatske, Bosne i Hercegovine (Miozane
kongerien in Kroatien, Bosnien und Herzegowina.). Palaeontologia Jugoslavia , 19, 1-98. [In Croatian with
German summary, pp. 85-98].

lamarck, j. p. b. A. de M. 1801. Systeme des animaux sans Vertebres. The author, and Deterville, Paris,
viii + 432 pp.

locard, a. 1883. Malacologie des Lacs de Tiberiade, d'Antioche et d'Homs. Archives du Museum d'Histoire
Naturelle de Lyon , 3, 195-203.

1893. Les Dreissensia du systeme europeen. Revue suisse de zoologie, 1, 113-185.

Logvinenko, b. m. and starabogatov, ya. i. 1966. Novye dvust-orchatye mollyuski iz Kaspiiskoi profindali.

Nauchnye Doklady Vyssnei Shkoly Biologicheski Nauki , 1966 (2), 13-16. Not seen. [In Russian],
lorenthey, e. 1894. Die pontische Fauna von Kurd im Comitate Tolna. Fbldtani Kozlony. A Magyar Foldtani
Tarsulat Folyoirata , 24, 73-88, 148-161.

marelli, d. c. and gray, s. 1983. Conchological redescriptions of Mytilopsis sallei and Mytilopsis leucophaeta
(Bivalvia: Dreissenidae) of the brackish Western Atlantic. Veliger , 25, 185-193.

— 1985. Comments on the status of Recent members of the genus Mytilopsis (Bivalvia: Dreissenidae).
Malacological Review , 18, 117-122.

736 PALAEONTOLOGY, VOLUME 33

marinescu, f. 1973. Les mollusques pontiens de Tirol (Banat Roumain). Memoriile Institutuli Geologic al
Romaniei, 18, 7-56.

— 1975. Genre Dreissenomya Fuchs (Bivalvia, Heterodonta). Memoires, Institut de Geologie et de
Geophysique , Bucarest , 6, 75-118.

merklin, r. l. 1974. Handbook of Oligocene Bivalvia of the south of the USSR. Trudy Paleontologicheskogo
Instituta , Akademiya Nauka SSSR. 145, 189 pp. [In Russian],
moore, R. c. (ed.). 1969. Treatise on invertebrate paleontology. Part N, Mollusca, 6, Geological Society of
America and University of Kansas Press, Boulder, Colorado and Lawrence, Kansas (1), i-xxxviii, N1-N489;
(2), i-ii, N491-N951 . ’

morlet, l. 1884. Description d’especes nouvelles de coquilles, recuellies par M. Pavie, au Cambodge. Journal
de Conchyliologie, 32, 386-403.

— 1892. Diagnoses molluscorum novarum in Indo-China collectorum. Journal de Conchyliologie, 40, 82-86,
315-329.

morley-davies, a. 1971-1975. Tertiary faunas. (Ed. 2, revised by Eames, F. E.), George Allen and Unwin,
London, 1 (1971), 571 pp.; 2 (1975). 447 pp.

morrison, j. p. e. 1946. The nonmarine mollusks of San Jose Island, with notes on those of Pedro Gonzalez
Island, Pearl Islands, Panama. Smithsonian Miscellaneous Collections, 106 (6), 49 pp.
morton, b. 1969. Studies on the biology of Dreissena polymorpha Pall. I. General anatomy and morphology.
Proceedings of the Malacological Society of London, 38, 301-321.

— 1970. The evolution of the heteromyarian condition in the Dreissenacea (Bivalvia). Palaeontology, 13,
563-572.

— 1981. The biology and functional morphology of Mytilopsis sallei (Recluz) (Bivalvia: Dreissenacea)
fouling Visakhapatnam Harbour, Andrhra Pradesh, India. Journal of Molluscan Studies, 47, 25-42.

Newell, n. d. 1965. Classification of the Bivalvia. American Museum Novitates, 2206, 1-25.
nuttall, c. p. 1990. A review of the Tertiary non-marine molluscan faunas of the Pebasian and other
inland basins of north western South America. Bulletin of the British Museum ( Natural History ) (Geology),
45, 165-371.

olsson, A. A. 1931. Contributions to the Tertiary paleontology of northern Peru: Part 4, the Peruvian
Oligocene. Bulletin of American Paleontology, 17, 99-260.

1961. Panamic- Pacific Pelecypoda, Paleontological Research Institute, Ithaca, N.Y., 574 pp.
orbigny, A. d\ 1850-52. Prodrome de paleontologie stratigraphique universelle des animaux mollusques et
rayonnes. Masson, Paris, 2, 427 pp. (1850); 3, 190 pp. (1852).
oppenheim, p. 1891. Die Gattungen Dreyssensia van Beneden und Congeria Partsch, ihre gegenseitigen
Beziehungen und ihre Vertheilung in Zeit und Raum. Zeitschrift der Deutschen Geologischen Gesellschaft, 43,
923-966.

1892. Ueber emige Brackwasser - und Binnen - mollusken aus der Kreide und dem Eocan Ungars.
Zeitschrift der Deutschen Geologischen Gesellschaft , 44, 696-818.
orlov, Y. a. (ed.) 1960. Osnovy paleontologii : Mollusca, Bivalvia, Akademii Nauka SSSR, Moscow, 300 pp. [In
Russian],

pallas, L. p. 1771. Reise durch verschiedene Provinzen des Russischen Reichs, 1, Kayserlichen Academie der
Wissenschaften, St Petersburg, vi + 504 pp.

pana, i. 1962. Contributii la cunoastere a molustelor de talie mica din depositele Pliocene. Analele Universitatii
Bucuresti , 31, 105-119.

papp, a. 1950. Ubergangsformen von Congeria zu Dreissena aus dem Pannon des Wiener Beckens. Annalen des
Naturhistorischen Museums, Wien, 57, 148-156.

1951. Uber das Vorkommen von Dreissenomya im Pannon des Wiener Beckens. Verhandlungen der
Geologischen Bundesanstalt , Wien, (1949), 128-133.

— 1953. Die Molluskenfauna des Pannon im Wiener Becken. Mitteilungen der Geologischen Gesellschaft in
Wien, 44, (1951), 85-222.

— 1954. Die Molluskenfauna im Sarmat des Wiener Beckens. Mitteilungen der Geologischen Gesellschaft in
Wien, 45, (1952), 2-1 12.

partsch, p. 1835. Uber die Sogenannten Versteinerten Ziegenklauen ausdem Platensee in Ungarn, und ein
neues, urweltliches geschlecht zweischaliger Conchylien. Annalen des Wiener Museums der Naturgeschichte,
1, 95-101.

pilsbry, H. A. 1911. On the type of Congeria. Nautilus, Philadelphia, 25, 95-96.

— and becquaert, j. 1927. The aquatic mollusks of the Belgian Congo, with a geographical and ecological
account of Congo malacology. Bulletin of the American Museum of Natural History, 53, 69-602.

NUTTALL: REVIEW OF DREISSENACEA 737

recluz, c. A. 1849. Description de quelques nouvelles especes de coquilles. Revue et Magazin de Zoologie Pure
et Applique , (2) 1, 64 71.

— 1852. Description de coquilles nouvelles. Journal de Conchyliologie, 3, 249-256.

reeve, l. a. 1857-1858. Monograph of the genus Mytilus. Conchologica Iconica: or, illustrations of the shells
of molluscous animals, 10, 11 pis., 61 species.

rovereto, G. 1898. Note preventive sui pelecipodi del tongriano ligure. II. Atti della Societd Ligustica di Science
Naturali e Geografiche , 9, 153-187.

- 1900. Illustrazione dei molluschi fossili tongriani. Atti della (Rea/e) Universitd di Genova, 15, 31-210.
rozanov, a. yu. (ed. ) 1986. Istoriya neogenvykh mollyuscov Paratetisa. (History of the Neogene molluscs of

the Paratethys.) Trudy Paleontologiclieskogo Instituta, Akademiya Nauka SSSR, 220, 208 pp. [In Russian].
sacco, F. 1898-1904. / molluschi dei terreni terziarii del Piemonte e delta Liguria. Carlo Clausen, Turin, 25
(1898), 77 pp.; 30 (1904), 203 + xxxvi pp.

sandberger, c. l. f. 1870-75. Land- und Siisswasser-Conchylien der Vorwelt. C. W. Kreidel, Weisbaden,
viii + 1000 pp.

seninski, k. 1905. Neogen ablagerungen im District Suchum des Siid-Westlichen Kaukasus. Schriften der
Naturforscher-Gesellschaft bei der Universitat Jurjeff ( Dorpat ), 16, 80 pp. [In Russian with German
summary].

sieber, r. 1955. Systematische Ubersicht der jungtertiaren Bivalven des Wiener Beckens. Annalen der
Naturhistorischen Museums Wien. 60, 169-201.

sinsov, i. 1877. Opisanie o novyi i malo-izlyedovannykhi formakhi rakovini izi tretichnykhi obrazovanii
Novorossil. Zapiski Novorossiiskago Obshchestva Estestvoispytatelei, 5, 1-23. [In Russian],
starabogatov, ya. I. 1970. Fauna mollyuskov i zoogeographicheskoe rayonirovanie kontinentalvnvikh vodoemov.
(Mollusc fauna and zoogeographical partitioning of continental water reservoirs of the world.) Izdatelstvo
Nauka Leningradskoe Otdelenie, Leningrad, 372 pp. [In Russian].

STEININGER, F. F., senes, J., kleemann, K. and rogl, F. (eds. ). 1985. Neogene of the Mediterranean Tethys and
Paratethys. Stratigraphic correlation tables and sediment distribution maps. University of Vienna, 1, xiv+ 189
pp., 10 folding maps, 2, xxvi + 536 pp.

taktakishvili, i. G. 1973. Pliotsenovye Dreissenidy zapadnoi Gruzii (Pliocene dreissenids of western Georgia ).

Metsniereba Tbilisi (Tiflis), 149 pp. [In Russian with French summary, p. 136],
taylor, j. d., Kennedy, w. j. and hall, a. 1973. The shell structure and mineralogy of the Bivalvia. II
Lucinacea - Clavagellacea conclusions. Bulletin of the British Museum (Natural History) (Zoology), 22,
253-294.

thiele, j. 1934-5. Handbuch der systematischen Weichtierkunde. 2, Gustav Fischer, Jena, 779-1022. (1934);
1023-1153 (1935).

turton, w. 1840. A manual of the land and fresh-water shells of the British Islands with figures of each of the
kinds (Ed. 2, by Gray, J. E.), Longman, Orme, Brown, Green and Longmans, London, ix + 324 pp.
vokes, H. E. 1980. Genera of the Bivalvia '. A systematic and bibliographic catalogue (Revised and updated).

Paleontological Research Institute, Ithaca, N.Y., xxvii + 307 pp.
wenz, w. 1921. Das Mainzer Becken und seine Randgebeite. W. Ehrig, Heidelberg, 351 pp.
wolff, w. J. 1969. The Mollusca of the estuarine region of the rivers Rhine, Meuse and Scheldt in relation to
the hydrography of the area. II. The Dreissenidae. Basteria, 33, 93-103.
woodring, w. p. 1925. Miocene mollusks from Bowden, Jamaica. Pelecypods and scaphopods. Publications,
Carnegie Institution of Washington, 361, 221 pp.

yonge, c. m. and Campbell, J. I. 1968. On the heteromyarian condition in the Bivalvia with special reference
to Dreissena polymorpha and certain Mytilacea. Transactions of the Royal Society of Edinburgh, 68, 21^13.
ZELINSKAYA, V. A., KULICHENKO, V. G., MAKARENKO, D. E., and SOROCHAN, E. A. 1968. PaleontologicheskH
spravochnik. I. Bivalvia. Mollusca, Palaeogene and Miocene, Ukraine. Naukova Dumka, Kiev, 297 pp. [In
Russian].

C. P. NUTTALL

Department of Palaeontology
The Natural History Museum
Cromwell Road
London SW7 5BD

Typescript received 16 January 1989
Revised typescript received 24 November 1989

PREDATION ON KOSMOCERAS BY SEMIONOTID
FISH IN THE MIDDLE JURASSIC LOWER OXFORD

CLAY OF ENGLAND

by DAVID M. MARTILL

Abstract. Broken ammonite debris is common in the Lower Oxford Clay and is thought to be attributable
in part to the activities of predators. A new specimen of Kosmoceras (Zugokosmoceras) cf. obductum
(Buckman) displays a series of bite marks in the region of the peristome suggestive of the dentition of the fishes
Lepidotes, Heterostrophus (Semionotidae) and Mesturus (Pycnodontidae).

The ammonite fauna of the Lower Oxford Clay (Middle Callovian, Middle Jurassic) of
Peterborough, Cambridgeshire, UK, is remarkable for its exceptional preservation. Although
specimens are usually crushed flat, most are preserved in original iridescent white aragonite
(Callomon 1968). Species of Kosmoceras are most abundant, and include both macroconch and
microconch morphs. Many are intact to the peristome, and specimens with extended and complete
lappets are numerous. Besides numerous species of Kosmoceras , there are also Erymnoceras spp.,
Binatasphinctes spp., Choffatia sp., Reineckia sp., Hecticoceras spp. Sigaloceras sp. and
Pseudocadoceras sp. Many of the ammonites are complete, but broken ammonite shell material also
occurs in quantity. Shells of the various species of Kosmoceras are thin and delicate, and could easily
be broken by continued current or storm activity. However, ammonite tests found in storm
generated shell beds are often intact and have suffered little from abrasion, suggesting that they are
more robust than would at first appear. Much of the broken ammonite debris found must therefore
be attributable to the activities of predators rather than to non-biological processes.

DESCRIPTION

The new specimen is a small, laterally compressed phragmacone of Kosmoceras cf. obductum
(Buckman) (text-fig. 1), with the major portion of the body chamber, and a scallop-shaped margin
on a slightly damaged peristome. The shell is well preserved, consisting of iridescent white aragonite
with a purple sheen. The maximum diameter at the aperture is 50 mm. The specimen was discovered
in olive brown bituminous shales at the base of the K. obductum Subzone of the E. coronation Zone,
at the Dogsthorpe Brick Pit, near Peterborough (Grid reference TF 219019). The specimen has now
been protected by a coating of polyvinyl butyral, and is deposited in the collection of the
Department of Geology, University of Leicester, accession number LEIUG 99163.

The damaged peristome displays five arcuate incisions, arranged in a semicircle. Each incision is
separated by an arc of 3-5 mm from its neighbours. The margin of one incision flaked away after
the specimen was collected, reducing the visual impact of the specimen. The incisions have an
estimated average diameter of approximately 3 mm, but the shape if complete would have been
oval, rather than circular. The arrangement of these incisions is similar to the distribution of teeth
(text-fig. \e) on the dentary of the ganoid fishes Lepidotes and Heterostrophus (Semionotidae) and
possibly the maxilla of Mesturus (Pycnodontidae).

|Palaeontology, Vol. 33, Part 3, 1990, pp. 739-742. |

© The Palaeontological Association

740

PALAEONTOLOGY, VOLUME 33

text-fig. 1. a-d, Kosmoceras cf. obductum from the Lower Oxford Clay, Peterborough, with bitten peristome.
LEIUG 99163. a , phragmacone and body chamber with predator damaged peristome; arrow indicates end of
phragmacone, x L5. b , detail of peristome, approximately x 3-5. c, outline of bitten peristome with position of
teeth incisions indicated, x 3-5. d , position of five teeth which made contact with the ammonite test, x 3-5. e,
occlusal surface of right lower jaw (here reversed to simulate left jaw) of Lepidotes macrocheirus , BM(NH)

P6839, Lower Oxford Clay, Peterborough, x 3.

MARTILL: FISH PREDATION ON AMMONITES

741

AMMONITE PREDATION

Predation on ammonites has been described by a number of authors (see Lehmann 1976 for a
review). Ample evidence that ammonites were victims of predators exists in the form of broken,
bitten and damaged tests, but it is often difficult to identify the predators with certainty. Arthropods
have been considered to be predators on some Lower Jurassic ammonites, while Kauffman and
Kesling (1960) have demonstrated predation by the marine reptile Mosasaurus. Small ammonites
have been recorded from coprolitic material attributed to plesiosaurs (Wetzel 1960).

Potential predators of ammonites in the Lower Oxford Clay can be found amongst both
vertebrates and invertebrates. The large number of belemnites of considerable size (guards up to
30 cm) might be candidates for ammonite predators, as may some of the arthropods. The vertebrate
fauna of the Lower Oxford Clay includes a rich diversity of marine reptiles and fishes, but it is only
the fishes which have dentitions modified to any degree for coping with hard shell material.

The fish fauna of the Lower Oxford Clay formed the subject of numerous papers by Woodward
(1888, 1892 a, 6, 1893, 1896, 1928, 1929), but has received little attention since. Fortunately the
dentitions of many of the numerous taxa are available for analysis, most housed in the British
Museum (Natural History).

Among the fish with durophagous dentitions are the ganoid fishes Lepidotes and Heterostrophus
[ = Dapedium ?] (Semionotidae), Mesturus leedsi Woodward (Pycnodontidae), the hybodont shark
Asteracanthus ornatissimus Agassiz (Hybodontidae), and the chimaeroids Brachymylus , Pachymylus
and Ischyodus. The dentition of the chimaeroids is robust, and incorporates large tritoral areas
designed to crush shell material. Their sharp mandibular dental plates may have been used to dig
for shelled prey. These teeth are unlikely to leave scallop-shaped incisions on any invertebrate test.
The dentition of Asteracanthus is complex, and comprises a highly heterodont assemblage, although
most teeth comprise a broad tritoral area with a well developed central ridge. It is unlikely that
Asteracanthus would have been capable of producing a crescentic series of incisions. The dentition
of the pycnodont Mesturus consists of robust, rounded teeth, often with tritoral cusps. Teeth on the
splenial and vomer of Mesturus are arranged in straight rows (Woodward 1896), and would not
produce a crescent-shaped bite. There are four teeth on each premaxilla and dentary of Mesturus.
They are directed forward and have expanded occlusal surfaces ideal for nibbling small shelly
organisms. This dental pattern could produce the pattern of bites seen in the ammonite if all
premaxillary teeth from one side, plus one of the premaxillary teeth from the other side of the jaw
had made contact with the ammonite test. The dentitions of Heterostrophus phillipsi Woodward.
1928, and the three species of Lower Oxford Clay Lepidotes ( L . leedsi Woodward, 1895, L.
macrocheirus Egerton, 1845 and L. latifrons Woodward, 1893) are robust, sometimes peg-like with
slightly expanded tips, and well-spaced along the jaw. The dentary is crescentic in occlusal view. Any
of these four species could have produced the crescentic bite seen in the ammonite (text-fig. \e).

Alternatives to fish predators are difficult to find preserved as macrofossils in the Lower Oxford
Clay. Two arthropods, Mecochirus and Goniochirus , are common in the same bed from which the
ammonite was collected, but both have rather small claws, and those of Mecochirus are extremely
elongate and delicate, and appear somewhat specialized. Other cephalopods, especially the various
naked forms, were possible predators on ammonites, but details of the jaw apparatus in these are
poorly known.

DISCUSSION

Lepidotes spp. and Heterostrophus phillipsi are the most likely candidates for producing the
damaged peristome of the ammonite, but Mesturus leedsi cannot be ruled out. Lepidotes and
Heterostrophus are common in the Lower Oxford Clay, but Mesturus appears to be rare. Complete
specimens are not common, but isolated scales are very abundant. Lepidotes is possibly the most
abundant large fish in the Lower Oxford Clay, but this is difficult to assess as the robust enamel
scales are easy to detect, and have a relatively high preservation potential. No information is
available on the level at which Lepidotes spp. or Heterostrophus lived in the water column, but their

742

PALAEONTOLOGY, VOLUME 33

general robust fusiform shape, and the massive form of the skull might suggest that they were part
of the near benthos biota rather than a part of the surface living community. Perhaps ammonites
were also part of the benthic community.

The trophic status of a number of vertebrates from the Lower Oxford Clay has been discussed
by Martill (1985, 1987, 1988), but as yet no single account of the interrelationships of the biota is
available. This stems mainly from a lack of reliable data. The top of the food chain was dominated
by the giant carnivorous reptiles Liopleurodon and Pliosaurus. There were a number of opportunistic
feeders, including the marine crocodilian Metriorhynchus which fed on giant fish and hookleted
cephalopods (Martill 1987). Many of the vertebrates were highly specialized feeders; the ichthyosaur
Ophthalmosaurus is thought to have fed exclusively on naked cephalopods, while the marine
crocodile Steneosaurus and the long necked plesiosaurs were probably mainly fish and naked
cephalopod feeders. The fish fauna displays similar trophic niche partitioning. The giant
pachycormid Leedsichthys was a plankton feeder (Martill 1988), while the other pachycormids
Hypsocormus and Asthenocormus were active predators on smaller fish. Clearly the Lower Oxford
Clay food web was highly complex.

Acknowledgments. Thanks go to the London Brick Company for permission to visit the Dogsthorpe Brick Pit,
to Alan Dawn for help in excavating the site, and to Colin Patterson (BMNH) for permission to examine
material in his care.

REFERENCES

callomon, j. h. 1968. The Kellaways Beds and the Oxford Clay. 264—290. In Sylvester Bradley, p. c. and
ford, t. d. (eds. ). The Geology of the East Midlands. University of Leicester Press, Leicester, 400 pp.
egerton, p. G. 1845. On some new species offish from the Oxford Clay at Christian Malford. Proceedings of
the Geological Society of London , 4, 446-449.

rauffman, e. G. and kesling, R. 1960. An Upper Cretaceous ammonite bitten by a mosasaur. Contributions
from the Museum of Paleontology , the University of Michigan , 15, 193-248.
lehmann, u. 1976. The ammonites: their life and their world. Cambridge University Press, Cambridge, 245 pp.
martill, d. m. 1985. The preservation of marine vertebrates in the Lower Oxford Clay (Jurassic) of central
England. Philosophical Transactions of the Royal Society of London, Series B, 311, 155-165.

— 1987. The diet of Metriorhynchus , a Mesozoic marine crocodile. Neues Jahrbuch fur Geologie und
Palaontologie , Monatshefte, 1987. 621-625.

1988. Leedsichthys problematicus , a giant filter-feeding teleost from the Jurassic of England and France,
Neues Jahrbuch fur Geologie und Palaontologie, Abhandlungen, 171, 670-680.
wetzel, w. 1960. Nachtrag zum Fossilarchiv der Fuiriquina-Schichten. Neues Jahrbuch fur Geologie und
Palaontologie, Monatshefte, 1960, 46-53.

woodward, a. s. 1888. On some remains of the extinct selachian Aster acanthus from the Oxford Clay of
Peterborough, preserved in the collection of Alfred Leeds esq., of Eyebury. Annals and Magazine of Natural
History, 2, 336-342.

1892a. On the skeleton of a chimaeroid fish (Ischyodus) from the Oxford Clay of Christian Malford,
Wiltshire. Annals and Magazine of Natural History, 9, 94-96.

1 8926. On some teeth of new chimaeroid fishes of the Oxford and Kimmeridge Clays of England. Annals
and Magazine of Natural History, 10, 13-16.

1893. On the cranial osteology of the Mesozoic ganoid fishes Lepidotes and Dapedius. Proceedings of the
Zoological Society of London, 38, 559-565.

1895. Catalogue of fossil fishes in the British Museum (Natural History ), pt. Ill, Actinopterygia, Teleostomi.
British Museum (Natural Elistory), London, xxxix + 544 pp., 18 pis.

1896. On the remains of the pycnodont fish Mesturus, discovered by Alfred N. Leeds, Esq., in the Oxford
Clay of Peterborough. Annals and Magazine of Natural History , 17, 1-15.

— 1928. A new ganoid fish. 56th Annual Report of the Peterborough Natural History Society, 59-60, pis. 1-5.
1929. The Upper Jurassic ganoid fish Heterostrophus. Proceedings of the Zoological Society of London,

99, 561-566.

DAVID M. MARTILL

Department of Earth Sciences
The Open University
Walton Hall

Milton Keynes, MK7 6AA, UK

Typescript received 12 June 1989
Revised typescript received 31 July 1989

NON-PREDATORY DRILLING OL M I SS I SS I PP I AN
CRINOIDS BY PLATYCERATID GASTROPODS

by TOMASZ K. BAUMILLER

Abstract. The conical hole in the tegmen of a Mississippian crinoid (Macrocrinus mundulus ) directly
underneath the shell of a platyceratid gastropod, Platyceras ( Ortlionychia ) sp., and similar holes in other
Mississippian crinoids ( Batocrinus irregularis and B. icosidactylus) demonstrate the drilling abilities of
platyceratids. This is the first case of drilling by an archaeogastropod. Drilling on the crinoid host by the
gastropod was non-predatory ; the relationship was probably parasitic. This study, the first to identify
unequivocally a Palaeozoic borer, supports the notion that gastropod drilling has evolved several times.

Evidence implying predation on Palaeozoic invertebrates is well documented (Fenton and
Fenton 1931; Brunton 1966; Rohr 1976; Ausich and Gurrola 1979). Examples include conical
holes, mostly in brachiopods, which have been attributed to Palaeozoic gastropods because of their
similarity to drillholes produced by Recent predatory gastropods (Cameron 1967; Sheehan and
Lesperance 1978; Smith el a/. 1985). This interpretation is not universally accepted, since the
Palaeozoic taxon responsible for them has not been conclusively identified (Carriker and Yochelson
1968; Sohl 1969). The discovery of a borehole in a Mississippian crinoid, directly beneath an
attached platyceratid gastropod, is the first definitive demonstration of drilling activity by a
Palaeozoic gastropod.

PLATYCERATI D-CRINOID INTERACTION

The archaeogastropod family Platyceratidae (Knight et al. 1960) is well known for its association
with crinoids (Meyer and Ausich 1983). Generally, the gastropod is firmly attached to the crinoid
calyx usually above the base of the arms. Its position, commonly covering the crinoid anus, and the
close conformity of its margin to the morphology of the calyx, has suggested commensalism and
that platyceratids were coprophages on crinoid wastes (Keyes 1890; Bowsher 1955). Recently it has
been argued that platyceratids were parasites feeding on crinoid gametes (Lane 1984). The parasitic
interpretation is supported by the smaller size of gastropod-infested crinoids as compared to
uninfested ones (Rollins and Brezinski 1988).

Commonly, the gastropod sits directly over the crinoid anus, which is flush with the tegmen, or
upper portion of the calyx. However, several platyceratids have been found situated on the crinoid
tegmen but well away from the anus (Lane 1978). In such instances the infested crinoids have a long
stout anal tube, with the anus at its apex, while the platyceratid rests at the base of the tube. In this
position the gastropod could not extract anything from the crinoid without drilling into the host.
Therefore, to test the hypothesis that platyceratids were capable of boring their tube-bearing hosts,
I serially ground a crinoid with Platyceras ( Ortlionychia ) sp. at the base of its long tube.

PLATYCERATID DRILLHOLES

In the collections of the Field Museum of Natural History (FMNH) I found a well-preserved
specimen of Macrocrinus mundulus (Batocrinidae) with a species of Platyceras ( Ortlionychia ) at the
base of the anal tube. Specimen P. 19426, which was collected at Canton, Indiana, USA, from rocks
of Tournaisian age, was ground perpendicular to the oral-aboral axis (long axis of the anal tube)

I Palaeontology, Vol. 33, Part 3, 1990, pp. 743-748.|

© The Palaeontological Association

PALAEONTOLOGY, VOLUME 33

text-fig. 1. Specimen of Macrocrinus mundulus (FMNH P.19426) with a long anal tube and Platyceras on its
tegmen. (a) Lateral view of specimen with lines B, C, D, and E indicating positions at which photographs
B, C, D, E of ground sections were taken. Ground surfaces are perpendicular to the plane of the page and the
long axis of the specimen, x 1-4. (b) Section through the arms, anal tube and the anterior of the Platyceras.
Note that the anal tube plate directly beneath the platyceratid appears intact, x 3-0. (c) Section 1-0 mm below
b. Arrow points to a small concavity in the anal tube plate beneath the platyceratid, x 3 0. (d) Section 10 mm
below C. Arrow points to drill hole. Note contrast between hole filled with dark matrix and light colour of anal
tube plates, x 3-0. (e) Section through centre of drill hole. Note that hole penetrates to centre of anal tube and
the tight fit between the test of gastropod and the anal tube, x 2-9.

BA UMILLER: GASTROPOD DRILLING OF CRINOIDS

745

at 0-5 to 10 mm intervals, starting at the apex of the tube (text-fig. 1). The serial sections revealed
a circular, conical hole (3 mm outer diameter) beneath the central, anterior portion of the gastropod
shell. The hole is easily recognized from the colour contrast between the white plates of the anal tube
and the red matrix filling the body cavity of the gastropod and the hole underneath it.

In addition, 27 specimens of Visean Batocrinus icosidactylus and B. irregularis from the
collections of the Field Museum and Indiana University, Bloomington, USA, also possessing a long
anal tube, were found with previously undescribed holes in their tegmina. The location of these
holes is highly stereotyped: they are either just below or at the base of the anal tube (text-figs. 2 and
3). The holes also exhibit little variation in morphology; they are round in plan view, cylindrical to
conical in cross-section, with a mean outer diameter of 2-2 mm (range from T4 to 3-2 mm) (text-fig.
2h). The morphology and position of these holes match those of the hole in Macrocrinus beneath
Platyceras (text-figs. 1e, 2h).

Platyceratids occur with the aforementioned batocrinids in localities of the Salem Limestone
(Cummings et al. 1905), a fact consistent with the interpretation that these gastropods were
responsible for the drillholes. The lack of an association between Platyceras and Bcitocrimts in these
localities may be a taphonomic artifact. When platyceratids are associated with crinoid tegmina,
articulated crinoid arms are also preserved. This implies either that the arms prevented the post-
mortem separation of the two organisms or that the process leading to the disarticulation of the
arms from the crown also led to the detachment of the gastropod from the crinoid. Batocrinus with
preserved arms have not been described. The absence of platyceratids on Batocrinus is thus not
inconsistent with this mode of preservation.

Thesq results establish the boring abilities of the Mississippian Platyceras ( Orthonychia ), the first
unequivocal example of a drilling Palaeozoic gastropod.

DISCUSSION

N on-predatory nature of drillholes

The boring by Platyceras does not necessarily imply predation on the crinoid. Several lines of
evidence suggest that the relationship was non-predatory. The tight fit between the margin of
Platyceras and the tegmen of Macrocrinus (text-figs. I d, 1 e), U-shaped attachment scars
surrounding the boreholes (text-fig. 2g), instances of multiple boreholes, and healed (or incomplete)
boreholes on Batocrinus all indicate that the association was long term and that drilling was not
fatal to the crinoid. Furthermore, Recent parasitic capulid gastropods, which have a shell
morphology very like the platyceratids, also drill their host (Orr 1962; Kosuge and Hayashi 1967).

Whether this relationship was beneficial, neutral, or detrimental to the crinoid cannot be
determined conclusively. Although drilling itself was probably detrimental, the feeding strategy of
platyceratids is most relevant in distinguishing between the three alternatives. If the gastropods fed
exclusively on crinoid wastes, then their position at the apex of the anal tube, directly over the
crinoid anus, would have sufficed. It is possible, however, that these relatively large gastropods
would have been unstable in this position due to the small size of the distal end of the tube. Drilling
at the base of the tube would have permitted feeding in a more stable orientation. More probably
the gastropod extracted from the crinoid more than just waste matter. Camerate crinoid food
grooves converge subtegminally and carry within them food particles collected by the dense
filtration fans. By inserting its snout into the tegmen the gastropod could have extracted food from
these grooves. This would have involved penetrating the crinoid gut wall, since in camerates the
subtegminal food grooves form closed tubular conduits leading directly to the fore-gut (Haugh
1975); a similar situation has been described for the parasitic prosobranch, Echineulima , whose
proboscis was found to pierce the intestinal wall of its host echinoid (Liitzen and Nielsen 1975). In
crinoids without an anal tube this could have been accomplished without drilling, simply by using
the existing anal opening through which to insert the snout beneath the tegmen. Such a strategy
would not have worked with crinoids possessing a long tube, due to the greater distance of the anus
from the food grooves lying deep below the tegmen. Drilling would be essential. This may be an

746

PALAEONTOLOGY, VOLUME 33

text-fig. 2. Drillholes in bactocrinids from the Salem Limestone, (a) Batocrinus icosidactylus (FMNH
P.19393), x 1-9. (B) Anal tube of B. icosidactylus (FMNH P.19392), x 2-5. (c) B. icosidactylus (FMNH
P.19394), x 2 2. (D) Anal tube of B. icosidactylus (FMNH 19394), x 1-4. (e) B. icosidactylus (FMNH P.21506),
x 2 0. (f) B. irregularis (FMNH P.19402), x2-l. (G) SEM of B. icosidactylus (FMNH P.19402). Arrows
point to U-shaped scar around hole, x 8 8. (h) A latex cast of drillhole in B. icosidactylus (FMNH PE. 1957).
Orientation of cast is with the top corresponding to the outer surface of the anal tube and the bottom to the

inner surface, x 2-3.

BAUMILLER: GASTROPOD DRILLING OF CRINOIDS

747

text-fig. 3. A reconstruction of the calyx of Batocrinus with part of the anal tube.
Actual drillholes projected onto reconstruction. Note stereotypic position of holes,

‘arms-race’ (Vermeij 1987) in which crinoids evolved an anal tube in response to gastropod
parasitism, which in turn led to the evolution of platyceratid drilling. Though highly conjectural,
this hypothesis may be tested by determining whether anal tubes evolved preferentially in
gastropod-infested lineages, and whether platyceratid drillholes are found in crinoids without anal
tubes.

The drilling habit and ‘ evolutionary radiations ’

Other evolutionary implications of the drilling habit of Platyceras hinge upon our knowledge of the
systematics of this group. Although archaeogastropods are probably not a true clade (Hickman
1988), if the taxortomic scheme proposed by Knight et al. (1960) is correct with regard to
platyceratids, Platyceras represents an example of independent evolution of drilling in gastropods
that did not lead to a major diversification (Smith et al. 1985; Fiirsich and Jablonski 1984).
Platyceratid generic diversity did not undergo major changes during or after the Mississippian
(Bowsher 1955), and their extinction at the end of the Permian (Bowsher 1955, but see Bandel 1988)
coincides with the extinction of other invertebrates, including most of their crinoid hosts.

A pre-Mississippian origin of the platyceratid drilling habit is possible, for platyceratids as well
as bored echinoderms are found in rocks of Ordovician to Permian age (Bowsher 1955; Brett 1978;
Paul 1971). In most instances, however, these borings are in the form of pits which do not penetrate
the skeletal plates of the host echinoderm, and their origin has been attributed to various epizoans.
Whether platyceratids were responsible for any of these pits remain to be examined.

It has been argued that evolutionary ‘failures’ of taxa with drilling abilities resulted either from
deterministic causes related to the adaptive value of this innovation (e.g. inability to drill rapidly)
or from factors unrelated to the innovation itself but affecting rates of speciation and extinction
(Smith et al. 1985; Fiirsich and Jablonski 1984). The identification of the innovation-bearing taxon
provides an excellent opportunity for testing these ideas. Thus the suggestion that the lack of a hard
skeleton (Smith et al. 1985) contributed to the evolutionary failure of Palaeozoic drillers must be
rejected for the shelled Platyceras. Furthermore, the non-predatory nature of these drillholes forces
us to make a distinction between predatory and parasitic drilling, as well as between their
evolutionary implications.

Acknowledgements . I thank M. Foote, D. Jablonski, M. LaBarbera, M. Listokin, D. Miller, M. Nitecki, and
an anonymous reviewer for comments on the manuscript, and D. Miller for technical assistance. Partial
funding for this project was provided by the Gurley Fund of the University of Chicago.

REFERENCES

ausich, w. i. and gurolla, r. a. 1979. Two boring organisms in a Lower Mississippian community of southern
Indiana. Journal of Paleontology, 53, 335-344.

bandel, k. 1988. Early ontogenetic shell and shell structure as aids to unravel gastropod phylogeny and
evolution. Malacological Review Supplement , 4, 267-272.
bowsher, a. l. 1955. Origin and adaptation of platyceratid gastropods. University of Kansas Paleontological
Contributions , Mollusca Article 5, 1-11.

748

PALAEONTOLOGY, VOLUME 33

brett, c. E. 1978. Host-specific pit-forming epizoans on Silurian crinoids. Lethaia , 11, 217-232.
brunton, H. 1966. Predation and shell damage in a Visean brachiopod fauna. Palaeontology 9, 355-359.
cameron, b. 1967. Oldest carnivorous gastropod borings found in Trentonian (Middle Ordovician)
brachiopods. Journal of Paleontology , 41, 147-150.

carriker, m. r. and yochelson, e. t. 1968. Recent gastropod boreholes and Ordovician cylindrical borings.

U.S. Geological Survey Professional Paper , 593- B, 1-26.

Cummings, E. r., beede, J. w., branson, e. b. and smith, e. a. 1905. The fauna of the Salem Limestone of
Indiana. 30th Annual Report of the Department of Geology and Natural Resources of Indiana, 1187-1486.
fenton, c. L. and fenton, M. A. 1931. Some small borings of Paleozoic age. American Midland Naturalist, 12,
522-528.

fursich, f. t. and jablonski, d. 1984. Late Triassic naticid drillholes: carnivorous gastropods gain a major
adaptation but fail to radiate. Science, 224, 78-80.

haugh, b. n. 1975. Digestive and coelomic systems of Mississippian camerate crinoids. Journal of Paleontology,
49, 472-493.

hickman, c. s. 1988. Archaeogastropod evolution, phylogeny and systematics: a re-evaluation. Malacological
Review Supplement, 4, 17-34.

keyes, c. r. 1890. Synopsis of American Carbonic Calyptraeidae. Proceedings of the Academy of Natural
Sciences of Philadelphia, (1890) 150-181.

KNIGHT, J. B., COX, L. R., KEEN, M. A., BATTEN, R. L., YOCHELSON, E. L. and ROBERTSON, R. 1960. Systematic
descriptions. In MOORE, R. c. (ed.). Treatise on invertebrate paleontology. Part I. Mollusca, 1169-1310.
Geological Society of America and University of Kansas Press, Lawrence, Kansas. 351 pp.
kosuge, s. and hayashi, s. 1967. Notes on the feeding habits of Capulus dilatatus A. Adams, 1860
(Gastropoda). Science Report of the Yokosuka City Museum, 13, 45-54.
lane, n. g. 1978. Mutualistic relations of fossil crinoids. In moore, r. c. and teichert, c. (eds.). Treatise on
invertebrate paleontology. Part T. Ecliinodermata 2(1), T345-T347. Geological Society of America and
University of Kansas Press, Lawrence, Kansas, 1027 pp.

1984. Predation and survival among inadunate crinoids. Paleobiology, 10, 453-458.
lutzen, j. and nielsen, k. 1975. Contributions to the anatomy and biology of Echineulima n.g. (Prosobranchia :
Eulimidae), parasitic on sea urchins. Videnskabelige Meddele/ser fra Dansk Naturhistorisk Forening , 138,
171-199.

meyer, d. l. and ausich, w. i. 1983. Biotic interactions among Recent and fossil crinoids. In tevesz, m. j. s. and
McCall, p. L. (eds.). Biotic interactions in Recent and fossil benthic communities, 378-427. Plenum, New York.
837 pp.

orr, v. 1962. The drilling habit of Capulus danieli (Crosse) (Mollusca: Gastropoda). Veliger , 5, 63-67.
paul, c. R. c. 1971. Revision of the Ho/ocystites fauna (Diploporita) of North America. Fieldiana: Geology,
24, 166 pp.

rohr, p. m. 1976. Silurian predator borings in the brachiopod Dicoelosia from the Canadian Arctic. Journal
of Paleontology 50, 1 175-1 179.

rollins. h. b. and brezinski, d. k. 1988. Reinterpretation of crinoid-platyceratid interaction. Lethaia, 21,
207-217.

sheehan, p. m. and lesperance, p. j. 1978. Effect of predation on the population dynamics of a Devonian
brachiopod. Journal of Paleontology, 52, 812-817.

smith, s. a., thayer, c. w. and brett, c. e. 1985. Predation in the Palaeozoic: gastropod-like drillholes in
Devonian brachiopods. Science, 230, 1033-1037.
sohl, n. f. 1969. The fossil record of shell borings by snails. American Zoologist, 9, 725-734.
vermeij, g. j. 1987. Evolution and escalation, xv + 527 pp. Princeton University Press, Princeton, New Jersey.

tomasz k. baumiller
Department of Geophysical Sciences
University of Chicago
5734 S. Ellis Ave.

Chicago, IL 60637
USA

Typescript received 19 June 1989
Revised typescript received 18 August 1989

NOTES FOR AUTHORS

The journal Palaeontology is devoted to the publication of papers on all aspects of palaeontology. Review articles are
particularly welcome, and short papers can often be published rapidly. A high standard of illustration is a feature of the
journal. Four parts are published each year and are sent free to all members of the Association. Typescripts should conform
in style to those already published in this journal, and should be sent to Dr Dianne Edwards, Department of Geology,
University of Wales College of Cardiff CFI 3YE, who will supply detailed instructions for authors on request (these are
published in Palaeontology 1985, 28, pp. 793-800).

Special Papers in Palaeontology is a series of substantial separate works conforming to the style of Palaeontology.

SPECIAL PAPERS IN PALAEONTOLOGY

In addition to publishing Palaeontology the Association also publishes Special Papers in Palaeontology. Members may
subscribe to this by writing to the Membership Treasurer: the subscription rate for 1990 is £45 00 (U.S. $66) for Institutional
Members, and £20 00 (U.S. $32) for Ordinary and Student Members. A single copy of each Special Paper is available on a
non-subscription basis to Ordinary and Student Members only, for their personal use, at a discount of 25 % below the listed
prices: contact the Marketing Manager. Non-members may obtain Nos 35-40 (at cover price) from Basil Blackwell Ltd.,
Journal Subscription Department, Marston Book Services, P.O. Box 87, Oxford OX2 0DT, England, and older issues from
the Marketing Manager. For all orders of Special Papers through the Marketing Manager, please add £1-50 (U.S. $3) per
item for postage and packing.

PALAEONTOLOGICAL ASSOCIATION PUBLICATIONS

Special Papers in Palaeontology

For full catalogue and price list, send a self-addressed, stamped A4 envelope to the Marketing Manager. Numbers 2-1 1 and
13-30 are still in print and are available together with those listed below:

31. (for 1984): Systematic palaeontology and stratigraphic distribution of ammonite faunas of the French Coniacian, by
w. j. Kennedy. 160 pp., 42 text-figs., 33 plates. Price £25 (U.S. $38).

32. (for 1984): Autecology of Silurian organisms. Edited by M. o. bassett and j. d. lawson. 295 pp., 75 text-figs., 13 plates.
Price £40 (U.S. $60).

33. (for 1985): Evolutionary Case Histories from the Fossil Record. Edited by j. c. w. cope and p. w. skelton. 202 pp., 80
text-figs., 4 plates. Price £30 (U.S. $45).

34. (for 1985): Review of the upper Silurian and lower Devonian articulate brachiopods of Podolia, by o. l. Nikiforova,
t. l. modzalevskaya and m. g. bassett. 66 pp., 6 text-figs., 16 plates. Price £10 (U.S. $15).

35. (for 1986): Studies in palaeobotany and palynology in honour of N. F. Hughes. Edited by d. j. batten and d. e. g.
briggs. 178 pp., 29 plates. Price £30 (U.S. $50).

36. (for 1986): Campanian and Maastrichtian ammonites from northern Aquitaine, France, by w. j. Kennedy. 145 pp., 43
text-figs ., 23 plates. Price £20 (U.S. $35).

37. (for 1987): Biology and revised systematics of some late Mesozoic stromatoporoids, by rachel wood. 89 pp., 31 text-
figs., 1 plates. Price £20 (U.S. $35).

38. (for 1987): Taxonomy, evolution, and biostratigraphy of late Triassic-early Jurassic calcareous nannofossils, by p. r.
bown. 118 pp., 19 text-figs., 15 plates. Price £30 (U.S. $50).

39. (for 1988): Late Cenomanian and Turonian ammonite faunas from north-east and central Texas, by w. j. Kennedy,
131 pp., 39 text-figs., 24 plates. Price £30 (U.S. $50).

40. (for 1988): The use and conservation of palaeontological sites. Edited by p. r. crowther and w. a. Wimbledon. 200 pp.,
31 text-figs. Price £30 (U.S. $50).

41. (for 1989): Late Jurassic-early Cretaceous cephalopods of eastern Alexander Island, Antarctica, by p. j. howlett.
72 pp., 9 text-figs., 10 plates. Price £20 (U.S. $35).

42. (for 1989): The Palaeocene flora of the Isle of Mull, by m. c. boulter and z. kvacek. 149 pp., 23 text-figs., 23 plates.
Price £40 (U.S. $70).

Field Guides to Fossils and Other Publications

These are available only from the Marketing Manager. Please add £100 (U.S. $2) per book for postage and packing plus
£1-50 (U.S. $3) for airmail. Payments should be in Sterling or in U.S. dollars, with all exchange charges prepaid. Cheques
should be made payable to the Palaeontological Association.

1. (1983): Fossil Plants of the London Clay, by m. e. collinson. 121 pp., 242 text-figs. Price £7-95 (U.S. $12) (Members £6

or U.S. $9).

2. (1987): Fossils of the Chalk, compiled by E. OWEN; edited by a. b. smith. 306 pp., 59 plates. Price £11-50 (U.S. $18)

(Members £9-90 or U.S. $15).

3. (1988): Zech stein Reef fossils and their palaeoecology, by n. hollingworth and t. pettigrew. iv-t-75 pp. Price £4-95

(U.S. $8) (Members £3.75 or U.S. $6).

1982. Atlas of the Burgess Shale. Edited by s. conway morris. 31 pp., 24 plates. Price £20 (U.S. $30).

1985. Atlas of Invertebrate Macrofossils. Edited by j. w. Murray. Published by Longman in collaboration with the
Palaeontological Association, xiii + 241 pp. Price £13-95. Available in the USA from Halsted Press at U.S. $24-95.

© The Palaeontological Association, 1990

Palaeontology

VOLUME 33 • PART 3

CONTENTS

The classification of the Foraminifera - a review of historical and
philosophical perspectives

J. R. HAYNES 503

Ontogeny, hypostome attachment and trilobite classification

R. A. FORTEY 529

Bedding plane assemblages of Promissum pulchrum, a new giant
Ashgill conodont from the Table Mountain Group, South Africa
J. N. THERON, R. B. RICKARDS and R. J. ALDRIDGE 577

Drilling and peeling of turritelline gastropods since the Late
Cretaceous

W. D. ALLMON, J. C. NIEH and R. D. NORRIS 595

A deductive enquiry system for a palaeontological database of
museum material

M. J. ROGERS, D. T. DONOVAN and M. H. ROGERS 613

The affinities of early oncocerid nautiloids from the Lower
Ordovician of Spitsbergen and Sweden

D. H. EVANS and A. H. KING 623

The solute Dendrocystoides scoticus from the Upper Ordovician of
Scotland and the ancestry of chordates and echinoderms

R. P. S. JEFFERIES 63]

Stromatoporoid palaeobiology and taphonomy in a Silurian
biostrome on Gotland, Sweden

S. KERSHAW 681

Review of the Caenozoic heterodont bivalve superfamily
Dreissenacea

C. P.NUTTALL 7Q7

Predation on Kosmoceras by semionotid fish in the Middle
Jurassic Lower Oxford Clay of England

D. M. MARTILL 739

Non-predatory drilling of Mississippian crinoids by platyceratid
gastropods

T. K. BAUMILLER 743

Printed in Great Britain at the University Press , Cambridge

ISSN 0031-0239

Palaeontology

VOLUME 33 • PART 4 NOVEMBER 1990

Published by

The Palaeontological Association • London
Price £30

THE PALAEONTOLOGICAL ASSOCIATION

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

COUNCIL 1990-1991

President : Professor J. W. Murray, Department of Geology, The University, Southampton S09 5NH
Vice-Presidents : Dr M. Romano, Department of Geology, University of Sheffield, Sheffield S3 7HF
Dr P. R. Crowther, City of Bristol Museum and Art Gallery, Queen’s Road, Bristol BS8 1RL
Treasurer: Dr M. E. Collinson, Department of Biology, King’s College, London W8 7AH
Membership Treasurer : Dr H. A. Armstrong, Department of Geology, University of Newcastle,

Newcastle upon Tyne NE1 7RU

Institutional Membership Treasurer: Dr A. W. Owen, Department of Geology and Applied Geology,

University of Glasgow, Glasgow G12 8QQ

Acting Secretary : Dr J. A. Crame, British Antarctic Survey, High Cross, Madingley Road, Cambridge CB3 OET
Circular Reporter : Dr D. Palmer, Department of Geology, Trinity College, Dublin 2
Marketing Manager: Dr C R. Hill, Department of Palaeontology, British Museum (Natural History), London SW7 5BD
Public Relations Officer : Dr M. J. Benton, Department of Geology, University of Bristol, Bristol BS8 1RJ

Editors

Dr M. J. Benton, Department of Geology, University of Bristol, Bristol BS8 1RJ
Dr J. E. Dalingwater, Department of Environmental Biology, University of Manchester, Manchester M13 9PL
Dr D. Edwards, Department of Geology, University of Wales College of Cardiff, Cardiff CF1 3YE
Dr P. D. Lane, Department of Geology, University of Keele, Keele, Staffordshire ST5 5BG (co-opted)

Dr P. A. Selden, Department of Extra-Mural Studies, University of Manchester, Manchester M13 9PL
Dr P. D Taylor, Department of Palaeontology, British Museum (Natural History), London SW7 5BD

Other Members

Dr E. A. Jarzembowski, Brighton Dr W. J. Kennedy, Oxford Dr D. M. Martill, Milton Keynes
Dr A. R. Milner, London Dr R. A. Spicer, Oxford

Overseas Representatives

Argentina: Dr M. O. Mancenido, Division Paleozoologia invertebrados, Facultad de Ciencias Naturales y Museo, Paseo del
Bosque, 1900 La Plata. Australia: Dr K. J. McNamara, Western Australian Museum, Francis Street, Perth, Western
Australia 6000. Canada: Professor S. H. Williams, Department of Earth Sciences, Memorial University, St John’s,
Newfoundland A1B 3X5. China: Dr Chang Mee-mann, Institute of Vertebrate Palaeontology and Paleoanthropology,
Academia Sinica, P.O. Box 643, Beijing. Dr Rong Jia-yu, Nanjing Institute of Geology and Palaeontology, Chi-Ming-Ssu,
Nanjing. France: Dr J.-L. Henry, Institut de Geologie, Universite de Rennes, Campus de Beaulieu, Avenue du General
Leclerc, 35042 Rennes Cedex. Iberia : Prof. F. Alvarez, Departamento de Geologia, Universidad de Oviedo, C/. Jesus Arias
de Velasco, s/n. 33005 Oviedo, Spain. Japan: Dr I. Hayami, University Museum, University of Tokyo, Hongo 7-3-1, Tokyo.
New Zealand: Dr R. A. Cooper, New Zealand Geological Survey, P.O. Box 30368, Lower Hutt. Scandinavia: Dr R.
Bromley, Fredskovvej 4, 2840 Holte, Denmark. U.S.A. : Prof. A. J. Rowell, Department of Geology, University of Kansas,
Lawrence, Kansas 66044. Prof. N. M. Savage, Department of Geology, University of Oregon, Eugene, Oregon 97403. Prof.
M. A. Wilson, Department of Geology, College of Wooster, Wooster, Ohio 44961. Germany : Prof. F. T. Fursich,
Institut fur Palaontologie, Universitat, D8700 Wurzburg, Pliecherwall 1

MEMBERSHIP

Membership is open to individuals and institutions on payment of the appropriate annual subscription. Rates for 1990 are:
Institutional membership . . £60 00 (U.S. $108) Student membership . . £1 1-50 (U.S. $20)

Ordinary membership . £28-00 (U.S. $50) Retired membership .... £14 00 (U.S. $25)

There is no admission fee. Correspondence concerned with Institutional Membership should be addressed to Dr A. W. Owen,
Department of Geology and Applied Geology, The University, Glasgow G12 8QQ. Student members are persons receiving full-
time instruction at educational institutions recognized by the Council. On first applying for membership, an application form
should be obtained from the Membership Treasurer: Dr H. A. Armstrong, Department of Geology, University of Newcastle,
Newcastle upon Tyne NE1 7RU. Subscriptions cover one calendar year and are due each January; they should be sent to the
Membership Treasurer. All members who join for 1990 will receive Palaeontology, Volume 33, Parts 1-4. Enquiries
concerning back numbers should be directed to the Marketing Manager.

Non-members may subscribe, and also obtain back issues up to 3 years old, at cover price through Basil Blackwell Ltd,
Journal Subscription Department, Marston Book Services, P.O. Box 87, Oxford OX2 0DT, England. For older issues
contact the Marketing Manager.

Cover: Framboidal pyrite within lumen of a tracheid of the early land plant, Gosslingia breconensis, from the Lower Old Red
Sandstone of South Wales, x 700. See P. Kenrick and D. Edwards. Botanical Journal of the Linnean Society , 97, 95-123.

BIOMECHANICS OF TRILOBITE EXQ SKELETONS

ns

by NADINE V. WILMOT

0

Abstract. Most skeletal materials, such as bone and insect cuticle, are viscoelastic;. but heavily mine/alized
structures such as mollusc shell, are linearly elastic. The type of microstructure used in making a,sk€leton is
related to required mechanical strength and to the metabolic cost involved in construction. The effects of
composition, microstructure, and architecture on mechanical properties are discussed, and then related to
trilobite exoskeletons. Due to their composition and internal organization, trilobite cuticles can be regarded
as ceramics that behaved in a linearly elastic manner. The small size of the calcite crystals and the presence of
an organic framework reduced the risk of crack formation and slowed the progress of fractures. As a result
of its crystal arrangement, the thin outer prismatic layer would have had good compressive strength, but only
poor crack-stopping abilities, whereas the underlying principal layer added bulk to the cuticle and deflected
cracks. Structurally, trilobite exoskeletons are analogous to monocoque shells, that is, they are strong ‘thin
shells’ with the same composition throughout and behave as a ‘stressed skin’. The overall architecture of the
cephalon and pygidium is of a series of modified domes, strengthened by the presence of the doublure, whereas
thoracic segments are compromise structures which allow articulation as well as conferring mechanical
strength.

The ways in which materials respond to forces acting upon them are determined by their
mechanical properties. Accommodation of the resulting stresses is a function of the composition,
microstructural organization, and overall form (or architecture) of the specimen. The study of the
mechanical characteristics of biological materials is part of the discipline known as biomechanics,
a relatively new field with most research occurring from the 1970s onwards. Such research has
concentrated on investigating the biomechanics of Recent skeletal materials such as bone (Currey
1969, 1975, 1979), insect cuticle (Hepburn et al. 1975; Vincent 1980), and mollusc shell (Taylor and
Layman 1972; Currey and Taylor 1974; Currey 1976, 1980; Jackson et al. 1988) in order to compare
their competencies. Various mechanical tests have been developed to measure the tensile,
compressive, bending, and hardness characteristics of these materials by modification of standard
engineering procedures. However, although the tests were basically the same, authors each followed
their own individual techniques. Currey (1980) argued that sample preparation, size, and shape,
would all have a bearing on the results obtained, and so demonstrated the desirability of a standard
methodology.

These studies have demonstrated that not only do materials with the same composition have
different mechanical properties, but also that biomechanics may not be the most important factor
in their design ; the metabolic cost involved in constructing a skeleton is also very important (Currey
1980). Some of the common terms used in biomechanical studies are explained at the end of this
section.

The mechanical properties of biological materials have generally been related to composition and
microstructure. The significance of the form of the organism has only been studied in depth for the
Ostracoda (Benson 1974, 1975, 1981, 1982). Most testing of arthropod exoskeletons has been on
insect cuticle, mainly because they are the most common arthropods today (see Vincent 1980 for
review). However, such cuticles are uncalcified and behave as viscoelastic materials, in contrast to
heavily mineralized exoskeletons which behave as ceramics and are linearly elastic. Consequently
the biomechanics of trilobite exoskeletons, which were strongly calcified, are better compared with
other mineralized skeletons such as mollusc shells, rather than insect cuticle. In the following
sections the effects of composition, ultrastructure, and architecture on mechanical properties are
discussed in turn, and the principles involved then related to trilobite cuticles.

© The Palaeontological Association

IPalaeontology, Vol. 33, Part 4, 1990, pp. 749-768.|

750

PALAEONTOLOGY, VOLUME 33

Terminology

Stress (a): force per unit area.

Strain (e): a ratio to express the deformation created within a specimen when subjected to stress.
It is the ratio of the change in size to the basic size.

Stiffness is a measure of the resistance of a material to deformation. It is measured by performing
tensile tests (Currey and Taylor 1974; Joff'e et al. 1975; Ker 1980), or compression tests (Taylor and
Layman 1972; Currey 1976); both types of test have to be repeatable, that is, not incurring
permanent damage to the specimen such as fracture.

Hardness measures the ease with which a material flows under a stress, and is related to the stiffness
of the material and its plasticity. For details of tests see Taylor and Layman (1972) and Hillerton
et al. (1982). It is useful for comparing viscoelastic materials, or ceramics with grain sizes of about
100 /mi (Craig and Vaughan 1981).

Linearly elastic materials, such as ceramics, react immediately to the application of a stress, for as
long as it is present, and on its removal immediately revert back to their pre-stressed state (Text-
fig. 1a (i)). In a linearly elastic material, strain is directly proportional to stress and the stress/strain
plot is linear:

£ = o/E,

where E is a constant (Young’s modulus) and is a measure of the stiffness of the material.
Viscoelastic materials, such as insect cuticle, increasingly deform the longer a force is applied. Once
the stress is removed, recovery is also gradual, so that any measurement of strain is time-dependent
(Text-fig. 1a (ii)).

See Wainwright et al. (1976), Dorrington (1980), Gordon (1980), and Vincent (1980, 1982) for
more details about linear elastic and viscoelastic materials.

COMPOSITION

Every material has a unique response to both tension and compression, which is largely due to its
composition and the strength of the bonds that maintain its atomic structure. For example,
ceramics, characterized by ionic or covalent bonding, have great compressive strength.

Organisms are subject to both tensile and compressive forces, and their skeletons usually
comprise a mixture of components, that is, they are composites. Tensile forces usually occur within
the structure of the skeleton, or are created by support of the viscera, whereas compressive forces
can be induced by walking, the surrounding water pressure, or predation.

Insect cuticle is one of the most efficiently constructed naturally occurring composites, being
composed of chitin fibrils weakly bonded to a protein matrix. Chitin is very strong when subject to
tensile forces, whereas the type of protein matrix determines stiffness (Hillerton 1984). The
arrangement of these two components is also important (see below). Stiff and pliant insect cuticles
owe their differences to their protein matrix compositions; the properties of the chitin fibrils are
always the same (Hillerton 1984). Additional hardening can be generated by sclerotization, for
example in locust incisors (Hillerton 1980). This involves the incorporation of phenols which
become tightly bound within the cuticle, coupled with some loss of water (Vincent and Hillerton
1979).

Some marine arthropods harden their exoskeletons by reducing the amount of protein present
(compared with terrestrial forms) and mineralizing their exoskeletons with calcium salts. It has been
demonstrated for crustaceans that the greater the proportion of calcium salts, the harder the cuticle
becomes (Welinder 1974; Abby-Kalio 1982). The type of mineralization is also important, for
example the cuticle of the mantid shrimp Gonodactylus is mainly composed of calcium carbonate,
but the harder outer surface of the smashing limb is predominantly calcium phosphate (Currey et
al. 1982).

Sometimes organic matter is incorporated within the crystal lattice and alters the way in which

WILMOT: BIOMECHANICS OF TRILOBITE EXOSKELETONS

751

A (i)

YAVAYAYA

I

(iii)

*

(ii)

text-fig. I . Responses of different materials to stress, based on Wainwright et al. ( 1976, fig. 2 . 12), and Gordon
(1980, figs 6 and 7). a (i), linearly elastic materials ; a (ii), viscoelastic materials. B (i), continuous material is weak
in tension, as cracks can advance unimpeded; b (ii), isolated elements are stronger in tension as cracks are
unable to spread once the first element has broken, c, weak interfaces prevent the spread of cracks
(Cook-Gordon mechanism): (i), crack formation begins; (ii), weak interface opens out in advance of the crack;
(iii), progression of the crack is stopped, d, lateral bonds increase compressive strength: (i), isolated elements
bend under compression; (ii), weak lateral bonds restrict movement.

752

PALAEONTOLOGY, VOLUME 33

the mineral fractures. This is exemplified by proteins within the calcite crystals of echinoderm plates
(Berman et al. 1988). Although the concentration of protein is very small (only approximately ten
molecules per 1 x 1 0‘5 unit cells) these crystals are less brittle than inorganic calcite.

MICROSTRUCTURE

The size and arrangement of skeletal materials are of fundamental importance to the mechanical
behaviour of the structure as a whole. In general, composites are stronger than pure materials.
Cracks, generated in tensile conditions, are unable to propagate between isolated elements (Text-
fig. 1b), hence chitin fibrils are extremely strong as they are composed of individual long chains of
chitin molecules. If one chain is broken, the fracture does not spread to the others.

Weak interfaces within a material can also give strength due to the Cook-Gordon mechanism
(Cook and Gordon 1964). As a crack propagates through a material, the weak interface opens up
in advance. When the crack reaches the hole, its energy is dissipated and so is unable to continue
(Text-fig. lc). Wood behaves in this manner. Bonding between components can also increase
compressive strength by creating greater resistance to bending (Text-fig. Id).

Arthropod cuticle is a very efficient composite as it contains tension- and compression-resistant
members. The chitin/protein fibrils are arranged in sheets parallel to the cuticle surface. Each sheet
contains fibrils orientated in the same direction, but successive layers are rotated by a few degrees
to produce a helicoidal structure. The typical laminated appearance of arthropod cuticle is due to
this internal arrangement; each unit of 180° rotation of the helicoid corresponding to a single
lamina (Bouligand 1965; Neville and Luke 1969; Neville 1970; Livolant et al. 1978). Consequently
the cuticle is strong in tension along any plane parallel to the surface. Sometimes the sheets show
preferred orientation along the direction of greatest stress (Wainwright et al. 1976), for example in
the walking legs of spiders (Barth 1973), the hind limbs of coleopterans (Dennell 1978), the
pedipalps of scorpions (Mutvei 1977; Dalingwater 1980), in Limulus cuticle (Mutvei 1977;
Dalingwater 1980), and the legs of the Carboniferous eurypterid Mvcterops (Dalingwater 1985). In
such cases the cuticle is very strong along the plane of preferred fibre orientation but much weaker
in other directions (Harris 1980).

The organic matrix of heavily mineralized skeletons is one of the most important factors in
determining its mechanical properties. Most biomineralization is an ‘organic matrix-mediated’
process (Lowenstam 1981) where the organic matter forms a framework not only to control the
nucleation, size, and orientation of the inorganic crystals, but also the physical behaviour of the
shell (Krampitz et al. 1983). The organic matrix of crustaceans is a chitin-protein complex, and the
proteins can be subdivided into water insoluble, and soluble fractions (Richards 1951) which are
species-specific (Hackman 1974). This is also true in other marine invertebrates such as molluscs.
Weiner et al. (1983) demonstrated that molluscan insoluble matrix (including chitin fibrils) forms
a framework to limit the size and orientation of the crystals, which is very important mechanically,
and also supports the soluble matrix that controls nucleation sites (see also the review by Mann 1988
concerning organic matrices).

The quantity and mechanical characteristics of the organic framework compared with the
crystalline component can greatly influence the strength of a ‘stony’ (Wainwright et al. 1976)
skeleton. As the matrix is much less brittle than the crystals it encloses, when sufficiently thick it can
absorb some of the energy involved in crack propagation by plastic flow. If the organic matrix is
thin, fractures are prevented from developing mainly by deflecting cracks as they reach crystal
boundaries.

The size of the crystalline component of a stony skeleton (as controlled by the organic
framework) is also very important. The larger the component, the more likely it is to contain flaws
such as internal deformation or cracks. For linearly elastic solids at a given stress level, there is a
‘critical crack length’ or defect size at which fracture occurs. Hence it is advantageous for a ceramic
to be constructed from small components to reduce the chances of containing defects. For calcium
carbonate minerals in tension at 50 MN m“2 and 100 MN itT2, the critical crack length is 2-8 pm

WILMOT: BIOMECHANICS OF TRILOBITE EXOSKELETONS

753

(Wainwright et aL 1976). In general, heavily mineralized skeletal materials have grain sizes with one
dimension smaller than 3 //m which minimizes fracture occurrence. Uniformity in size is also
advantageous, to prevent weak links generated by larger components. The less organic matrix
present, the more important it is for the crystal sizes to be small, so that if cracks do form, they
frequently have to change direction at crystal boundaries: a process that requires additional energy.
Hence fine-grained ceramic skeletons that contain only a small amount of organic matrix, are much
stronger than a large inorganic crystal. This has been demonstrated for mollusc shells, which
contain only 01-5% matrix (Currey 1980) and yet are stronger than inorganically precipitated
calcite (Taylor and Layman 1972). Currey (1980, fig. 5) has illustrated the extremely good crack-
stopping capabilities of molluscan nacre, which is composed of many thin sheets of aragonite
separated by thin organic layers. A crack has to proceed along a very tortuous path in order to cross
this type of microstructure. Therefore it is not surprising that most ceramic skeletons are fine
grained and contain very few voids. When voids do occur, such as ducts, their function outweighs
the structural weaknesses they induce (Wainwright et al. 1976).

It can therefore be seen that the type of microstructure used to construct a skeleton will influence
its mechanical characteristics. Mollusc microstructural types (described by Watabe 1984) have been
found to exhibit different mechanical properties despite all being constructed from calcium
carbonate. For example, nacre is the strongest, and crossed lamellar structure the hardest. However,
mechanical strength may not be the most important factor involved in the design of a skeleton.
Although nacre is the strongest of all types of molluscan microstructure and occurred first in the
fossil record, weaker structural units are more commonly used today (Currey and Taylor 1974).
There is a general correspondence between molluscan microstructure and mode of life (Taylor and
Layman 1972; Currey and Taylor 1974; Currey 1976; Gabriel 1981). The development of other
shell types may have been due to a necessity to protect against abrasion or chemical attack prevalent
in some habitats, at which nacre is relatively poor (Gabriel 1981). The metabolic cost involved in
secreting the material may also be more important than its overall strength (Currey 1980). It has
been demonstrated that microstructures with a relatively high organic content are more ‘expensive’
to generate (Palmer 1983). A metabolically cheaper type of microstructure to construct, although
of inferior quality to nacre, may be adequate for a particular lifestyle. Oyster shells, for example,
are constructed from relatively weak microstructures (foliated structure and chalky deposits) and
are prone to attacks from boring organisms; however, they can grow very quickly, which suits these
animals’ particular mode of life (Currey 1980).

ARCHITECTURE

The general shape of a structure will also influence the way in which it responds to stresses.
Structural mechanics have only been applied in any depth to the design of ostracode carapaces,
where exoskeletal features have been discussed in relation to common engineering or architectural
structures (Benson 1974, 1975, 1981, 1982). Impact and compression testing of ostracode carapaces
(Whatley et al. 1982) did not reveal any one factor that had overriding importance in conferring
strength, though it has been demonstrated that the architecture of cirripede exoskeletons is more
important than microstructure (Murdock and Currey 1978).

The importance of shape to the overall mechanical strength of a structure is best explained with
reference to common architectural structures. The concept of ‘strength through form’ (Nervi 1951)
has been successfully exploited in architecture from the 1950s. The various man-made constructional
forms which have been developed follow well-understood engineering principles, and have many
analogies with biological designs. (For an introduction to structural mechanics see Buckle (1977)
and Cowan ( 1 980). ) Although buildings are many times larger in scale than biological constructions,
the forces acting on skeletons generate stresses acting in similar directions.

The type of material to be used usually determines the type of structure that is produced. Due
to their different mechanical properties, materials lend themselves to be used in certain ways: for
example, stone is relatively weak in tension but possesses great compressive strength.

754

PALAEONTOLOGY, VOLUME 33

compression boom

tension boom

text-fig. 2. Space-spanning structures, a, hemisphere, b, dome, c, stone vault; thick walls counteract outward
thrusts from the arch and ensure that the forces are deflected to the ground within the structural form (or else
the building would collapse), d, portal frames (pf) and purlins (p); portal frames always have a shallow pitch,
and the knee is strengthened by increased thickness or reinforcement; the weight and outward thrusts are
absorbed by the foundations. E, truss: the girders of the truss act either in tension (T) or compression (C) to
take up the load, and have flexible joints (based on Buckle 1977, fig. 8.1). f, space frames: (i), geodesic dome,
based on triangular or diamond-shaped grids with flexible joints (from Buckle 1977, fig. 15.3); (ii), lamellar
roofs of various shapes are based on horizontal or diagonal grids with rigid joints.

Some of the more common three-dimensional architectural structures are illustrated in Text-
figure 2. A catenary is the shape a cable takes when suspended equally at both ends, and can be
expressed mathematically as:

y = a/2(ex + e x)

WILMOT: BIOMECHANICS OF TRILOBITE EXOSKELETONS

755

where x and y are coordinates, e = exponential constant, a = variable. When inverted, this form is
a catenary arch, on which vertical loads are directed evenly over the whole length. Its three-
dimensional extension is the catenary dome (Text-fig. 2b). Although a hemisphere (Text-fig. 2a)
encloses a given volume with less surface area than a catenary-shaped dome, the dome is structurally
superior at resisting forces directed horizontally, and normally, to the crown. In both structures, the
weight of the material is directed downwards, generating horizontal thrusts at the margins. These
have to be counteracted either by ring beams, or some other sort of reinforcement such as increased
thickness, which confines outward movement of the base.

A vault is another three-dimensional version of an arch, and also has outward thrusts at its base.
For stone vaults (Text-fig. 2c), thick adjoining walls are necessary to counteract these horizontal
moments and ensure that the thrusts are directed to the foundations within the building structure.
This buttressing is essential to prevent collapse. Portal frames are space-spanning structures
composed of straight members (Text-fig. 2d). The greatest shear and bending moments occur at the
joint between the beam and the column, so the ‘knee’ is usually strengthened, either by thickening
or adding reinforcing material. As with arches, shallower frames are subject to greater horizontal
thrusts, and ties can be introduced between the supports to attain equilibrium. Purlins are beams
that span between portal frames, and on which additional material can be placed such as roofing
tiles. The load acts on the joints between the purlins and the portal frames.

Folded plates and corrugated sheets (Text-fig. 3) are more economical in material than flat plates
spanning the same area, as their shape imparts strength and so they can be thinner. A vertical load
acting on corrugations is divided into two components: a force R acting at right-angles to the slab
surface, and a force P acting parallel to it. P forces are resisted by ‘skin stresses’ within the slab,
so only force R will cause bending. Consequently, much greater forces can be withstood than by a
horizontal slab of the same thickness, on which all the force is directed perpendicularly to the
surface. A corrugated sheet can withstand up to a hundred times its own weight in load (Cowan
1980): its strength is a function of its sinusoidal cross-section and its overall depth.

As a method of spanning space, a truss (Text-fig. 2e) is more economical in material than a solid
beam, and there are many different types. Simply, the stresses from a vertically acting load are taken
up by a series of members which act either in tension or compression. Space frames (Text-fig. 2f)
can be likened to three-dimensional trusses. Variously shaped, strong but lightweight lattices
provide the necessary support for thin coverings. As in trusses, the members of the grid act either
in tension or compression to transmit the stresses acting on the structure. The covering ‘membrane’
or ‘skin’ lies passively between the struts. An alternative type of space frame is the monocoque shell
(Benson 1974, 1975), which does not have an internal grid. It is composed of the same material
throughout to produce a strong thin shell, and all the load is transmitted through this thin ‘stressed
skin '.

INFLUENCES OF TRILOBITE CUTICLE COMPOSITION AND MICROSTRUCTURE

ON MECHANICAL BEHAVIOUR

Since the organic matrix has so much influence on the mechanical behaviour of heavily mineralized
exoskeletons, no mechanical tests were performed in this study on fossil material, as any remaining
organic matter will have been degraded. This has been proved for ostracodes by compression and
impact tests on fossil and Recent specimens of the same species (Whatley et al. 1982), which were
found to have different strengths. However, knowledge of trilobite microstructure (Wilmot 1988)
can still provide much information on the likely mechanical properties of the exoskeleton.

Trilobite exoskeletons were heavily calcified, composed predominantly of low-magnesian calcite
(Wilmot and Fallick 1989) with only a small proportion of organic matter. In this respect they can
be categorized with other stony skeletons as behaving as ceramics. Remnants of the organic matrix
have been obtained by decalcifying the cuticles in EDTA (Dalingwater 1969, 1973; Teigler and
Towe 1975; Miller 1976; Dalingwater and Miller 1977). Although the composition remains
uncertain, it may well have consisted of proteins and chitin fibrils as in all other arthropods.

756

PALAEONTOLOGY, VOLUME 33

A

text-fig. 3. Corrugations, a, folded sheet: there is only a small local span (Is) so only thin slabs are necessary
to resist the R and P forces; thin, end diaphragms (d) resist the thrusts (T) from the whole roof (based on
Buckle 1977, fig. 14.7). b, the more folds in the folded sheet, the less local bending, enabling thinner slabs to
be used, therefore sinusoidal corrugations are the strongest, c, examples of various profiles of corrugations
used in buildings: (i), sinusoidal; (ii), trapezoidal (symmetrical and asymmetrical); (iii), stiffened trapezoidal.

Like all exoskeletons, trilobite cuticles were used both for protecting and supporting the viscera,
and would have been subject to both tensile and compressive forces. Tensile forces parallel to the
surface would have been present within the domed structure of the skeleton, and to a lesser extent
generated by the suspension of the viscera. Compressive stresses would have been imposed by water
currents and predators. To resist these, the exoskeleton contained tension elements, probably long
chitin fibrils orientated parallel to the cuticle surface, and short calcite crystals strong under
compression.

Unmineralized insect cuticle is stronger and stiffer than calcified crustacean cuticle; crustacean
exoskeletons have to be thicker to compensate (Wainwright et al. 1976). However, calcification is
very common in the marine environment as calcium ions are readily available. Hence it is much
more economical for a marine invertebrate to construct an exoskeleton consisting of less protein
and more calcium carbonate in comparison with its terrestrial counterpart (Wainwright et al. 1976).
This probably explains why trilobite exoskeletons were heavily calcified.

Apart from agnostid exoskeletons which may have been constructed of only a thin, prismatic
layer (Wilmot 1990o), most calcified trilobite cuticles comprise an outer prismatic layer with a
thicker principal layer below. Sometimes the thin outer layer is finely laminated (Mutvei 1981;
J. E. Dalingwater pers. comm.). Prismatic layer is a relatively thin layer of calcite crystals,
approximately I /mi in diameter, orientated with their longer c-axes perpendicular to the outer

WILMOT: BIOMECHANICS OF TRILOBITE EXOSKELETONS 757

surface (Text-fig. 4a). It is not analogous to the molluscan prismatic layer which can have crystals
up to several millimetres long, with each surrounded by a thick (5 /mi) layer of organic matrix
(Currey 1980). Trilobite principal layer forms 85-95% of the total cuticle thickness. It is finer
grained, and has a much less regular crystal arrangement than the prismatic layer (Text-fig. 4b).
Parallel laminations are sometimes preserved within this layer which may mark the former positions
of organic material. The amount of organic matrix was low, with much less than 1 pm thickness
between crystals. As in molluscan nacre (Currey 1980), fracture of trilobite exoskeletons mainly
occurred through the thin layers of organic matter rather than through individual calcite crystals
(unpublished observation, see Text-fig. 4). Except for eye lenses which were necessarily always of
extremely high quality calcite, none of the cuticular crystals have two axes more than 3 /mi long,
so minimizing the risk of dangerously large internal defects.

text-fig. 4. Scanning electron micrographs of trilobite cuticular structure, exposed on a vertical fracture
surface. Phacops rana crassituberculata Stumm, NMW.88.22G.41, Silica Shale, Silica, Ohio (Middle
Devonian), x 1 500. a, prismatic layer with calcite crystals (1x10 //m) orientated with their c-axes perpendicular
to the surface; note that fracture has occurred between crystals, not through them, b, principal layer, with

calcite crystals aligned parallel to the surface.

It has been argued above that the type of microstructure used is a compromise between the
mechanical properties necessary for a particular mode of life, and the metabolic price that has to
be paid in order to construct it. Trilobites, like other arthropods, regularly had to shed their
exoskeletons in order to grow. Immediately after moulting, the organism would have been unable
to move or feed properly, as well as being very vulnerable to predation, and so it would have been
advantageous to mineralize the cuticle as soon as possible. It is therefore likely that rapidity of
calcification was one of the main priorities involved in the selection of microstructure. Modern
mollusc shells are stronger than crustacean cuticles (Wainwright et al. 1976), but such shells are
gradually secreted throughout life. When trilobite cuticle was secreted, the prismatic layer formed
first, with only a very thin principal layer present; as secretion continued, cuticle thickness increased
by addition to the principal layer (Miller and Clarkson 1980). This suggests that the prismatic layer
was relatively easy to generate, being calcified rapidly after ecdysis to give some degree of hardening
to the cuticle. Bulk was added later to the exoskeleton to give greater strength. Prismatic layer may
also have been effective against shell-boring organisms, as with molluscs (Gabriel 1981). Trilobite
prismatic layer with its regular crystal arrangement was probably quite strong under compressive
forces acting normal to the cuticle surface, but would have had poor crack-stopping capabilities. A

758

PALAEONTOLOGY, VOLUME 33

crack would have been able to travel unimpeded between crystals, and would only have been
deflected on reaching the principal layer: hence the need for a thin principal layer even in newly
secreted cuticle. The crack-stopping abilities of the principal layer can still be effective after millions
of years: prismatic layer is often lost from trilobite cuticles when they are extracted from the rock
matrix, whereas the principal layer remains intact (Miller 1976).

Unlike molluscs, trilobites only developed two kinds of exoskeletal microstructure. However, the
overall thickness of the cuticles and the relative proportions of the different layers do vary
considerably, which would have been of mechanical significance. Different types of cuticular
ultrastructure may have been related to mode of life, or may just have been a phylogenetic trait, e.g.
proetides have relatively thin prismatic layers.

EFFECTS OF TRILOBITE ARCHITECTURAL DESIGN ON MECHANICAL

STRENGTH

When assessing the design of an exoskeleton, it is important to consider the main function it
provides for the animal. In trilobites, it seems likely that the exoskeleton developed primarily as a
means of protection against predators, rather than just for muscle support. The high degree of
calcification of the cuticle compared with other arthropods suggests an evolutionary bias for thicker
exoskeletons and hence increased protection. Indeed, in low-oxygenated environments such as those
existing in the ‘Olenid Sea’, predation was rare and the predominant trilobites only had very thin
cuticles (Fortey 1985).

Other evidence suggesting a defensive function for the cuticle is the development of sophisticated
enrolment mechanisms. This was a great advantage to trilobites, as they were able to curl up into
‘balls’ that enclosed vulnerable soft parts within the calcified exoskeleton. The cuticle then
effectively formed a hard, protective outer casing with a much greater diameter than before,
and the animal was therefore more difficult to attack. As cuticle thickness was fairly uniform
throughout (increasing at muscle insertion areas), there were no major weak points in the ‘ball’ that
could be exploited by predators. The margins of the exoskeleton, such as the borders of the cephalon
and pygidium were often locked tightly together and, due to the doublure, were also thicker and
stronger regions. Pleurae overlapping articulating facets greatly increased ‘ball’ thickness laterally.
Detailed methods of trilobite articulation and enrolment are well known and will not be discussed
further here: see Harrington (1959, pp. 070-073, figs 49-51), Bergstrom (1973), and Fortey and
Owens (1979) for further information.

Although it can be seen that the trilobite exoskeleton was well developed for enrolment, the
general shape was also mechanically strong for ‘normal’ life situations, and it is this aspect which
will be discussed further below. For example, although the doublure may sometimes be involved in
enrolment, incorporating coaptative devices (Clarkson and Henry 1973), this alone cannot explain
its function for all trilobites. Schmalfuss (1981) suggested that doublures with terrace ridges
supported a respiratory and feeding chamber below the animal, but this is unlikely to be true of
pelagic species, or groups with long pygidial doublures (Fortey 1985). It is proposed here that the
doublure acted primarily as a structural strengthening device, necessary for a dome-shaped
exoskeleton.

Like most ostracode carapaces, the trilobite exoskeleton structurally most resembles a series of
modified domes based on catenary arches (Text-fig. 5). The cephalon and pygidium can be regarded
simply as half-domes. The horizontal thrusts at the margins of the domes are resisted by the
doublure, which confers strength by increased thickness and a change of direction. The inner edge
of the doublure marks the point of attachment of the ventral membrane, but although the latter was
probably tough, its flexibility ( Muller and Walossek 1987) would have given it negligible structural
strengthening properties. Its main function would have been to constrain the positions of the body
organs. As with the knee of a portal frame constructed of the same material throughout, the
exoskeleton also had to be thicker at the doublure to resist shear forces acting at the change ot
direction. On average, cuticle thickness at least doubles at the doublure (Text-fig. 6). As slightly

WILMOT: BIOMECHANICS OF TRILOBITE EXOSKELETONS 759

text-fig. 5. General form of cephala and pygidia. a (i), trilobite cephala and pygidia can be regarded simply
as half-domes; a (ii), generalized transverse section through a cranidium; the basic shape of a half-dome has
been modified by addition of extra domes, b (i), the doublure gives strength to the domed structure by
increasing thickness at the base, and a change of direction; b (ii), the shape of the doublure is also important;
cylindrical doublures such as those of certain proetids are very strong, c, flatter domes exert greater thrusts at
the margins and therefore need longer doublures; (i), (ii), diagrammatic representation of the relationship
between convexity and doublure length; (iii), transverse section through the pygidium of Proetus ( Proetus )
concinnus (Dalman) (based on specimen NMW . 88 . 22G .42); (iv), transverse section through the pygidium of
Warburgella ( Warburgella ) stokesii (Murchison) (based on specimen NM W . 88 . 22G .43); both these trilobites
belong to the Proetacea yet have slightly different convexities, and hence dissimilar doublure lengths.

oblique sections through the doublure would give the impression of increased cuticle thickness, the
minimum values recorded are the most significant. The width of the doublure is proportional to its
strength (longer doublures affording more resistance to thrusts), therefore flatter domes which exert
greater horizontal thrusts have to have longer doublures (Text-fig. 5c). However this general
relationship is not necessarily valid for all trilobites, as factors such as doublure shape,
ultrastructure, and overall cuticle thickness also exert an effect. Different types of trilobite doublure
are illustrated in Text-figure 7. The addition of extra domes to parts of the exoskeleton such as the
glabella (Text-fig. 5a (ii)) is also a recognized practice in architecture.

760

PALAEONTOLOGY, VOLUME 33

400 i

w

w

a)

c

o

300

200 -

100 -

X

X

X X V X

X x^ *

X

X

0 100 200 300 400 500

Cuticle thickness ( pm)

text-fig. 6. Increase in cuticle thickness across the doublure.

Although, in sagittal section, a trilobite is much more elongated and therefore resembles a
shallower dome than in transverse section, the thorax is well jointed, and dorso-ventral forces could
have been taken up by flexure (Text-fig. 7a (iv)). However, the main purpose of the jointed thorax
was for locomotion and enrolment.

Trilobite thoracic segments are a compromise in structure between the necessary mechanical
strength to protect and support the underlying organs, and permitting articulation. Some trilobites,
such as certain illaenimorphs and homalonotids are characterized by their high convexity and
effacement. Their thoracic segments resemble single arches (Text-fig. 8b) and are therefore
structurally strong, and articulate only at the two fulcra (Lane and Thomas in Thomas 1978;
Thomas and Lane 1984). The segments are arranged in an imbricate manner, and articulation is
achieved by them sliding underneath one another (Thomas and Lane 1984, Text-fig. 2d). However,
this is not common for trilobites as a whole.

Most trilobites have various structures on the anterior and posterior margins of the thoracic
segments which enable them to articulate. These may include an articulating half-ring anterior to
the axial ring, flanges from the proximal parts of the pleurae to the fulcra, fulcral processes with
corresponding sockets, and articulating facets on the distal parts of the pleurae (Harrington 1959).
Additionally, coaptative structures may exist on the cephalon and pygidium (Clarkson and Henry
1973; Henry and Clarkson 1975).

Many thoracic segments have the proximal parts of the pleurae extending horizontally from the
axial furrows (Text-fig. 9). Although this is not very strong structurally, it is of great importance in
articulation as the pleurae form a hinge plane. Strengthening of these thoracic segments is generated
by the axial and pleural furrows (Bergstrom 1973; Miiller and Walossek 1987) which are expressed
ventrally as ridges (Text-fig. 9g), thus turning the cuticle into a folded sheet. Additional thickening
of the exoskeleton in these areas also increases the strength of the structure, especially in areas of
muscle attachment such as the axial furrows.

Other thoracic segments are intermediate in form between these two end members and resemble
arches on arches, but with several articulating processes (Text-fig. 10). Therefore, although the
cephalon and pygidium of a trilobite are well designed for mechanical strength, the shape of
thoracic segments are also a function of their role in articulation.

Occasionally a network of polygonal structures can be seen on the surface of trilobite cuticles, for

WILMOT: BIOMECHANICS OF TRILOBITE EXOSKELETONS

761

text-fig. 7. Different types of doublure, a, Illaenus sp. (JED) 2 Feb. 1966, pygidial doublure, Boda Limestone,
Dalarna, Oland (Ordovician), xl5. b, Cybantyx anaglyptos Lane and Thomas, NMW.88.22G.21,
longitudinal section through pygidium. Much Wenlock Limestone Formation, Wren’s Nest, Dudley
(Wenlock), x 7. c, Dalmanites myops (Konig), NMW . 88 . 22G . 20, longitudinal section through pygidium. Hill
End Farm Borehole, Walsall (Wenlock), x 10. d, Warburgella ( Warburgella ) stokesii (Murchison),
NMW . 88 . 22G . 5, transverse section through the lateral border of a free cheek (note the canals opening
out at the crests of the terrace ridges), Coalbrookdale formation. Daw End railway cutting, Walsall

(Wenlock), x 70.

example Homagnostus obesus (Wilmot 1990r/) which gives the appearance of space frames, such as
geodesic domes. However, as trilobite exoskeletons were constructed predominantly from low-
magnesian calcite, their space-enclosing structure is more analogous to the monocoque shell, which
resists forces throughout its ‘skin’, not just in members of a grid system.

The polygonal structures may simply be the external expression of epidermal cells which
generated the cuticle (Wilmot 1990h), but the network of ridges they create can give the ‘thin shell’
some additional strength. When examined in detail, the cell polygons are essentially curved plates
with reinforced ridges on their edges (Text-fig. I I a, b). In section, the ridges are acting as T-beams

762

PALAEONTOLOGY, VOLUME 33

text-fig. 8. Trilobite thoracic segments, a, generalized forms of trilobite thoracic segments : (i), single arch ; (ii),
arch on arches; (iii), horizontal pleurae; (iv), simplified sagittal section through a trilobite, emphasizing the
relatively low convexity compared with transverse sections, and the jointed nature of the thorax, b, single arch
thoracic segments: (i), (ii), Bumastus ( Bumastoides ) lenzi Chatterton and Ludvigsen (Illaenidae), anterior and
posterior views of thoracic segments, x 3-8 (based on Chatterton and Ludvigsen 1976, pi. 5, figs 27 and 31).
(iii)-(vii), Failleana calva Chatterton and Ludvigsen (Styginidae), thoracic segments (based on Chatterton and
Ludvigsen 1976, pi. 6): (iii), anterior view, x 3 (fig. 18); (iv), posterior view, x 4-3 (fig. 21 ); (v), posterior view,
x 2-9 (fig. 24); (vi), posterior view, x4-3 (fig. 22); (vii), anterior view, x 3-5 (fig. 27).

(Text-fig. 1 lc), which are much stronger than ordinary beams with columns since there are no
joints. Although in architecture these are normally inverted, the principles involved remain valid for
trilobites. Altogether, a strengthening meshwork has been produced over the external surface, the
structural significance of which is most marked in thin cuticles such as those of agnostid trilobites,
where the ridges can form as much as 15% of the total cuticle thickness. Such grids are rarely used
in man-made structures as external meshes are subject to corrosion; as a compromise, they have to
be placed within the structure. In this respect therefore, trilobite exoskeletons are better designed
than most buildings, although an external mesh was applied to the Philips Pavilion, designed by Le

WILMOT: BIOMECHANICS OF TRILOBITE EXOSKELETONS

763

A

text-fig. 9. Thoracic segments with horizontal pleurae, a, Ceratocephala lacinata Whittington and Evitt
(Odontopleuridae), posterior view of thoracic segment, x 10 (based on Whittington and Evitt 1954, pi. 8, fig.
8). B, c, Acidaspis ( Acidaspis ) lesperancei Chatterton and Perry (Odontopleuridae), posterior and anterior views
of thoracic segment, x 10 (based on Chatterton and Perry 1983, pi. 21, figs 15 and 16). d, Nanillaenus
mackenziensis Chatterton and Ludvigsen (Illaenidae), anterior view of thoracic segment, x 2-9 (based on
Chatterton and Ludvigsen 1976, pi. 4, fig. 15). E, Dolichoharpes aff. D. reticulata Whittington (Harpedidae),
posterior view of thoracic segment. x7 (based on Chatterton and Ludvigsen 1976, pi. 7, fig. 24). f-h,
Ceraurinella typa Cooper (Cheiruridae), anterior, ventral and posterior views of thoracic segment, x 2 (based

on Whittington and Evitt 1954, pi. 11, figs 1, 4, 5).

Corbusier at the Brussels Exhibition in 1958. As the building was a temporary exhibit, corrosional
effects were unimportant.

Additional structures on trilobite exoskeletons acting as T-beams occur on the ventral surfaces
of some effaced cuticles (Text-fig. 1 1e). For example, the pygidium of Leiolichas is strengthened by
a series of radiating ribs (Thomas and Holloway 1988, pi. 9, fig. 208). As mentioned previously,
ridges on the ventral surface are generally more prominent than the corresponding dorsal furrows,
although this is probably primarily for muscle attachment rather than for strengthening purposes.
Other larger-scale features such as terrace ridges will have strengthened the exosksleton by
increasing the thickness of the cuticle. As most terrace ridges only have relief on the external surface
of the exoskeleton, they acted mechanically as reinforcing ridges to the ‘sheet’, and often occur on
the margins or doublure where strains were greatest. Those that have relief on both dorsal and
ventral surfaces of the cuticle (Wilmot 1988) will have acted as folded plates.

764

PALAEONTOLOGY, VOLUME 33

text-fig. 10. Arch-on-arch thoracic segments, a-d, Dimeropyge virginensis Whittington and Evitt
(Dimeropygidae). thoracic segments, x 15 (based on Whittington and Evitt 1954, pi. 3, figs 8, 11, 1, 5). e,
Dimeropyge clintonensis Shaw (Dimeropygidae), posterior view of thoracic segment, x 15 (based on Chatterton
and Ludvigsen 1976, pi. 18, fig. II). f, g, Isotelus parvirugosus Chatterton and Ludvigsen (Asaphidae),
posterior and dorsal views of thoracic segment, x 2-9 (based on Chatterton and Ludvigsen 1976, pi. 2, figs 19
and 20). H, Acanthoparypha chiropyga Whittington and Evitt (Cheiruridae), posterior view of thoracic segment,
x 3-2 (based on Whittington and Evitt 1954, pi. 29, fig. 25). I, J, Encrinuroides rams (Walcott) (Encrinuridae),
posterior and dorsal views of thoracic segments, x 5-5 (based on Chatterton and Ludvigsen 1976, pi. 15, figs

40 and 41).

CONCLUSIONS

1 . For the first time, the mechanical characteristics of a trilobite exoskeleton have been related
to its composition, microstructure, and architecture.

2. Trilobite cuticles are best regarded as ceramics which are linearly elastic, as they were
predominantly composed of low-magnesian calcite with only a small proportion of organic matter.

3. Calcification was probably the most economical method of strengthening the cuticle in the
marine environment, due to the abundance of the relevant ions.

4. Trilobite exoskeletons were composites that could resist both tensile and compressive forces.

5. The small size of the calcite crystals reduced the risk of crack formation, and the progression
of fractures was slowed by the changes of direction at each crystal boundary.

6. Prismatic layer was probably the easiest microstructure to generate, being rapidly calcified
after ecdysis, and would have been strong under compressive forces acting normal to the cuticle
surface. The underlying principal layer functioned as a crack-stopper and added bulk to the
exoskeleton.

7. The relative proportions of prismatic layer to principal layer would have been of mechanical
significance.

8. Structurally, the trilobite exoskeleton is a monocoque shell in the form of a series of modified

WILMOT: BIOMECHANICS OF TRILOBITE EXOSKELETONS

765

B

plate

principal

layer

TTTTTtTrrTTTT 1 1 1 1 1 II llTTTTi i iTurunTTm

<t

C

D

E

text-fig. 1 1. T-beams. a, plan view of cell polygons on trilobite cuticle, b, schematic transverse section through
trilobite cuticle, not to scale, c, T-beam, as used in architecture, d, transverse section through an agnostid
cuticle, e, schematic transverse section through an effaced pygidium showing ribs on ventral surface.

domes. The doublure, cell polygons, and terrace ridges all strengthened the cuticle. Strongly convex
trilobites had shorter doublures than shallower forms.

9. The shape of thoracic segments is a compromise between mechanical strength and their
function as articulating structures.

Acknowledgements. This work was carried out as part of a Ph.D. thesis funded by an Aston University
Research Studentship Award. I thank Dr A. T. Thomas for his encouragement and supervision throughout the
project, and my father for valuable discussions on structural mechanics and architectural design. The
manuscript has also benefited from the comments of Professor J. D. Currey. All material is housed at the
National Museum of Wales, Cardiff, accession number NMW 88.22G, or as part of the collections of Dr
J. E. Dalingwater (JED), University of Manchester, also housed in Cardiff.

766

PALAEONTOLOGY, VOLUME 33

REFERENCES

abby-kalio, n. j. 1982. Aspects of feeding in Carcinus maenus (Linnaeus): functional anatomy and food-
harvesting strategies. Unpublished Ph.D. thesis. University of Reading.
barth, f. G. 1973. Microfibcr reinforcement of an arthropod cuticle. Zeitschrift fur Zellforschung und
mikroskopische Anatomie , 144, 409-433.

BENSON, R. H. 1974. The role of ornamentation in the design and function of the ostracode carapace. Geoscience
and Man , 6, 47-57.

- 1975. Morphologic stability in Ostracoda. Bulletin of American Paleontology , 65 (282), 13^46.

1981 . Form, function and architecture of ostracode shells. Annual Review of Earth and Planetary Sciences ,
9. 59-80.

1982. Comparative transformation of shape in a rapidly evolving series of structural morphotypes of the
ostracod Bradleya. 147-164. In bate, r. h., robinson, e. and Sheppard, l. m. (eds). Fossil and Recent
ostracods. Ellis Horwood, Chichester, 493 pp.

bergstrom, j. 1973. Organisation, life and systematics of trilobites. Fossils and Strata , 2, 1-69, pis 1-5.
berman, a., addadi, l. and weiner, s. 1988. Interactions of sea-urchin skeleton macromolecules with growing
calcite crystals - a study of intracrystalline proteins. Nature , 331, 546-548.
bouligand, y. 1965. Sur une architecture torsadee repandue dans de nombreuses cuticules d’arthropodes.

Compte Rendu hebdomadaire des Seances de I'Academie des Sciences , Paris , 261, 3665-3668.
buckle, i. G. (ed.). 1977. W. Morgan's ‘ The elements of structure ' (Second edition). Pitman, London, xi + 252

pp.

chatterton, b. d. e. and ludvigsen, r. 1976. Silicified Middle Ordovician trilobites from the South Nahanni
River area. District of Mackenzie, Canada. Palaeontographica , Abteilung A, 154, Lfg. 1-3, 1-109, pis 1-22.

- and perry, d. G. 1983. Silicified Silurian odontopleurid trilobites from the Mackenzie Mountains.
Palaeontographica Canadiana, 1. 1-127, pis 1 36.

clarkson, e. n. k. and henry, j.-l. 1973. Structures coaptatives et enroulement chez quelques Trilobites
ordoviciens et siluriens. Lethaia , 6, 105-132.

cook, t. and Gordon, j. e. 1964. A mechanism for the control of crack propagation in all-brittle systems.

Proceedings of the Royal Society , Series A , 282, 508-520.
cowan, h. j. 1980. Architectural structures: an introduction to structural mechanics (Metric edition). Pitman,
London, xvi + 320 pp.

craig, J. R. and vaughan, D. J. 1981. Ore microscopy and ore petrography. John Wiley and Sons, New York,
xii + 406 pp.

currey, j. d. 1969. The mechanical consequences of variation in the mineral content of bone. Journal of
Biomechanics , 2, 1 11.

- 1975. The effects of strain rate, reconstruction and mineral content on some mechanical properties of
bovine bone. Journal of Biomechanics , 8. 81-86.

— 1976. Further studies on the mechanical properties of mollusc shell material. Journal of Zoology, London ,
180, 445-453.

1979. Mechanical properties of bone tissues with greatly differing functions. Journal of Biomechanics, 12,
313-319.

— 1980. Mechanical properties of mollusc shell. Symposia of the Society for Experimental Biology, 34, 75-97 .

— nash, a. and bonfield, w. 1982. Calcified cuticle in the siomatopod smashing limb. Journal of Materials
Science, 17, 1939-1944.

— and taylor, J. d. 1974. The mechanical behaviour of some mollusc hard tissues. Journal of Zoology,
London, 173, 395-406.

dalingwater, j. E. 1969. Some aspects of the chemistry and fine structure of the trilobite cuticle. Unpublished
Ph D. thesis. University of Manchester.

1973. Trilobite cuticle microstructure and composition. Palaeontology, 16, 827-839.

- 1980. SEM observations on the cuticles of some chclicerates. Proceedings 8tli International Congress of
Araclmology, Vienna, 285-289.

1985. Biomechanical approaches to eurypterid cuticles and chelicerate exoskeletons. Transactions of the
Royal Society of Edinburgh (Earth Sciences), 76, 359-364.

and miller, j. 1977. The laminae and cuticular organisation of the trilobite Asaphus raniceps.
Palaeontology, 20, 21-32.

dennell, R. 1978. Balken orientation in three coleopteran cuticles. Zoological Journal of the Linnean Society,
64, 15-26.

WILMOT: BIOMECHANICS OF TRILOBITE EXOSKELETONS

767

dorrington, K. l. 1980. The theory of viscoelasticity in biomaterials. Symposia of the Society for Experimental
Biology , 34. 289-314.

fortey, r. a. 1985. Pelagic trilobites as an example of deducing the life habits of extinct arthropods.
Transactions of the Royal Society of Edinburgh ( Earth Sciences ), 76, 219-230.

and Owens, R. m. 1979. Enrolment in the classification of trilobites. Lethaia. 12, 219-226.

Gabriel, j. m. 1981 Differing resistance of various mollusc shell materials to simulated whelk attack. Journal
of Zoology, London, 194, 363-369.

Gordon, J. E. 1980. Biomechanics: the last stronghold of vitalism. Symposia of the Society for Experimental
Biology, 34, 1-11.

hackman, r. h. 1974. The soluble cuticular proteins from three arthropod species: Scylla serrata (Decapoda:
Portunidae), Boophilus microplus (Acarina: Ixodidae), and Agrianome spinicollis (Coleoptera:

Cerambycidae). Comparative Biochemistry and Physiology, 49 B, 457^464.

Harrington, h. J. 1959. General description of Trilobita. 038-0117. In moore, r. c. (ed. ). Treatise on
invertebrate paleontology. Part O. Arthropoda 1. Geological Society of America and University of Kansas
Press, New York and Lawrence, Kansas, 560 pp.

Harris, B. 1980. The mechanical behaviour of composite materials. Symposia of the Society for Experimental
Biology , 34, 41-74.

henry, j.-l. and Clarkson, E. n. K. 1975. Enrolment and coaptations i n some species of the Ordovician
trilobite genus Placoparia. Fossils and Strata, 4, 87-95, pis 1-3.
hepburn, h. R., joffe, L, green, n. and nelson, k. j. 1975. Mechanical properties of a crab shell. Comparative
Biochemistry and Physiology, 50 A, 551-554.

hillerton, j. e. 1980. The hardness of locust incisors. Symposia of the Society for Experimental Biology, 34,
483-484.

1984. Cuticle: mechanical properties. 627-637. In bereiter — hahn, j., matoltsy, a. g. and Richards, s.
k. (eds). Biology of the integument I . Invertebrates. Springer-Verlag, Berlin, Heidelberg, New York, 841 pp.

Reynolds, s. E. and Vincent, j. F. v. 1982. On the indentation hardness of insect cuticle. Journal of
Experimented Biology, 96. 45-52.

jackson, a. p., Vincent, j. F. v. and turner, r. m. 1988. The mechanical design of nacre. Proceedings of the
Royal Society, Series B, 234, 415 440.

joffe, l, hepburn, h. r. and anderson, s. o. 1975. On the mechanical properties of Limulus solid cuticle.
Journal of Comparative Physiology. 101, 147-160.

ker, r. f. 1980. Small-scale tensile tests. Symposia of the Society for Experimental Biology, 34, 487-489.
krampitz, G., drolshagen, h„ hausle, j. and hof-irmscher, k. 1983. Organic matrices of mollusc shells.
231-247. In westbroek, p. and de jong, e. w. (eds). Biomineralization and biological metal accumulation.
Reidel, Dordrecht, Holland, 533 pp.

lane, p. d. and thomas, a. t. 1978. Family Scutelluidae. 8-25. In thomas, a. t. British Wenlock Trilobites
(Part 1). Monograph of the Palaeontographical Society, 132 (552), 1 56, pis 1-14.
livolant, f., giraud, m-m. and bouligand, y. 1978. A goniometric effect observed in sections of twisted
fibrous materials. Biologie Cellulaire, 31, 159-168.
lowenstam, h. a. 1981. Minerals formed by organisms. Science, 211, 1126-1131.
mann, s. 1988. Molecular recognition in biomineralisation. Nature, 332, 119-124.

miller, j. 1976. The sensory fields and mode of life of Phacops rana (Green 1832) (Trilobita). Transactions of
the Royal Society of Edinburgh ( Earth Sciences), 69, 337-367, pis 1-4.

— and clarkson, e. n. k. 1980. The post-eedysial development of the cuticle and the eye of the Devonian
trilobite Phacops rana milleri Stewart 1927. Philosophical Transactions of the Royal Society of London, Series
B, 288, 461-480, pis 1-7.

muller, k. j. and walossek, d. 1987. Morphology, ontogeny, and life habit of Agnostus pisiformis from the
Upper Cambrian of Sweden. Fossils and Strata, 19, 1 124, pis 1-33.
murdock, g. r. and currey, j. d. 1978. Strength and design of shells of the two ecologically distinct barnacles
Balanus balanus, Semibalanus ( Balanus ) balanoides (Cirripedia). Biological Bulletin. Marine Biological
Laboratory , Woods Hole, Mass 155, 169-192.

mutvei, h. 1977. SEM studies on arthropod exoskeletons. 2. Horseshoe crab Limulus polyphemus (L.) in
comparison with extinct eurypterids and recent scorpions. Zoologica Scripta, 6, 203-213.

- 1981. Exoskeletal structure in the Ordovician trilobite Flexicalymene. Lethaia, 14, 225-234.
nervi, l. 1951. II ferro-cemento : sue carattieriche c possibilita. L'Ingegnere, 1.

Neville, a. c. 1970. Cuticle ultrastructure in relation to the whole insect. Symposia of the Royal Entomological
Society of London, 5, 17-39.

768

PALAEONTOLOGY, VOLUME 33

— and luke, b. m. 1969. A two-system model for chitin-protein complexes in insect cuticles. Tissue and Cell ,
1. 689-707.

palmer, a. r. 1983. Relative cost of producing skeletal organic matrix versus calcification: evidence from
marine gastropods. Marine Biology , 75, 287-292.

richards, a. G. 1951. The integument of arthropods. University of Minnesota Press, Minneapolis, xvi + 41 1 pp.
schmalfuss, H. 1981. Structure, patterns and function of cuticular terraces in trilobites. Lethaia , 14, 331-341.
taylor, J. d. and layman, m. 1972. The mechanical properties of bivalve (Mollusca) shell structures.
Palaeontology , 15, 73-87.

teigler, d. j. and towe, k. m. 1975. Microstructure and composition of the trilobite exoskeleton. Fossils and
Strata , 4, 137-149, pis 1-9.

thomas, a. T. 1978. British Wenlock Trilobites (Part 1). Monograph of the Palaeontographical Society, 1, 1-56,
pis 1-14.

— and holloway, d. j. 1988. Classification and phylogeny of the trilobite Order Lichida. Philosophical
Transactions of the Royal Society of London, Series B , 321, 179-262, 16 pis.

— and lane, p. d. 1984. Autecology of Silurian trilobites. Special Papers in Palaeontology , 32, 55-69.
Vincent, j. F. v. 1980. Insect cuticle: a paradigm for natural composites. Symposia of the Society for

Experimental Biology , 34, 183-210.

1982. Structural biomaterials. The Macmillan Press Ltd, London, xi + 206 pp.

— and hillerton, J. E. 1979. The tanning of insect cuticle - a critical review and revised mechanism. Journal
of Insect Physiology, 25. 653-658.

wainwright, s. A., biggs, w. d., currey, j. D. and gosline, J. M. 1976. Mechanical design in organisms. Edward
Arnold, London, xii + 423 pp.

watabe, n. 1984. Shell (Mollusca). 448-485. In bereiter-hahn, j., matoltsy, k. and richards, s. k. (eds).

Biology of the integument I. Invertebrates. Springer-Verlag, Berlin, Heidelberg, New York, 841 pp.
weiner, s., traub, w. and lowenstam, h. a. 1983. Organic matrix in calcified exoskeletons. 205-224. In
westbroek, p. and de jong, E. w. (eds). Biomineralization and biological metal accumulation. Reidel,
Dordrecht, Holland, 533 pp.

welinder, b. s. 1974. The crustacean cuticle. 1. Studies on the composition of the cuticle. Comparative
Biochemistry and Physiology, 47 A, 779-787.

whatley, r. c., trier, k. and dingwall, p. m. 1982. Some preliminary observations on certain mechanical and
biophysical properties of the ostracod carapace. 76-104. In bate, r. fl, robinson, e. and sheppard, l. m.
(eds). Fossil and Recent ostracods. Ellis Horwood, Chichester, 493 pp.

Whittington, H. B. and evitt, w. R. 1954. Silicified Middle Ordovician trilobites. Memoirs. Geological Society
of America, 59. 1 135, pis 1-33.

wilmot, n. v. 1988. Design and function of trilobite exoskeletons. Unpublished Ph.D. thesis, University of
Aston in Birmingham.

— 1990m Cuticular structure of the agnostine trilobite Homagnostus obesus. Lethaia , 23, 87-92.

1990 b. Primary and diagenetic microstructures in trilobite exoskeletons. Historical Biology 4, 51-65.

— and fallick, a. e. 1989. Original mineralogy of trilobite exoskeletons. Palaeontology, 32, 297-304.

N. v. WILMOT
Department of Zoology

Typescript received 15 May 1989 The Natural History Museum

Revised typescript received 15 February 1990 Cromwell Road, London SW7 5BD, UK

THE SILURIAN RUGOSE CORAL GENUS
ENTELOPHYLLUM AND RELATED GENERA IN
NORTHERN EUROPE

by john s. jell and Patrick k. Sutherland

Abstract. Restudy of Entelophyllum from Gotland (including the type species) and Great Britain indicates
restriction of Entelophyllum to phaceloid forms with peripheral, parricidal increase. Typical forms also have
smooth or carinate septa, well-developed biserial tabularia, and dissepimentaria composed of globose
interseptal dissepiments. On Gotland the genus ranges from the late Telychian to the Ludfordian.
Stereoxylodes Wang and Carinophyllum Strelnikov are considered junior synonyms of Entelophyllum. Species
from Gotland with nonparricidal budding are referred to Donacophyllum Dybowski. Cerioid forms internally
similar to Entelophyllum are referred to Prohexagonaria Merriam. Petrozium Smith is retained for some Early
Silurian forms. Newly described taxa are: Entelophyllum articulation anglicum subsp. nov., E. dendroides sp.
nov., E. lauense sp. nov. E. hamraense sp. nov., E. sp. A, Prohexagonaria favia sp. nov., P. gotlandica sp. nov.,
Donacophyllum neumani sp. nov., and D. wallstenense sp. nov.

Entelophyllum was proposed by Wedekind in September 1927, and Lang and Smith described
Xylodes in October of the same year, both for Silurian rugose corals similar to Cyathophyllum
articulation Wahlenberg, 1821. Since 1927, many forms worldwide have been referred to
Entelophyllum and other genera have been based on species referred to it at one time or another.
Several of the early species were based on unsectioned specimens, some of which have since been
lost or destroyed. Variation and stratigraphic distribution of most species were unknown.
Consequently, many of the species have been misinterpreted and some identifications away from the
type areas are clearly inaccurate.

The Silurian sequence on Gotland, Sweden (Text-fig. 1a) is the type area for several species and
provides the best succession of entelophylloid corals worldwide. This restudy is based on collections
in the Riksmuseet, Stockholm, Dr B. Neuman’s collection in Bergen, Professor D. Hill’s collection
in the University of Queensland, and on large collections made by the authors in 1972. We have
studied the available English and Scottish material in the collections of various British museums and
have collected from many of the British sections. Our descriptions and interpretations of other
northwestern European occurrences of entelophylloid corals are based on the published data and
photographs of type meterial sent to us by various colleagues. It is hoped that this study will provide
a significant basis for the re-evaluation of such corals from other regions in the world.

STRATIGRAPHY AND DISTRIBUTION

Great Britain

The earliest known occurrence of entelophylloid corals in Britain is Petrozium dewari Smith in the
Llandovery (possibly late Rhuddanian: based on correlations by Cocks and Toghill 1973), of
southwestern Scotland. In Shropshire, it occurs in the Pentamerus Beds, the graptolites of which are
attributable to the early Telychian turriculatus Biozone, and in the overlying Hughley Shale (Ziegler
et al. 1968) (Text-figs 2 and 3).

Entelophyllum has not been recorded from the Sheinwoodian Stage in Britain as this interval is
generally noncalcareous but as the sequence becomes more calcareous, E. articulation anglicum

IPalaeontology, Vol. 33, Part 4, 1990, pp. 769-821, 12 pls.|

© The Palaeontological Association

770

PALAEONTOLOGY, VOLUME 33

text-fig. 1 . a. Distribution of entelophylloid corals in northern Europe, b, Palaeogeography for mid-Silurian
times based mainly on Cocks and Fortey (1982) and McKerrow (1988). Localities: 1, Girvan; 2, Wenlock
Edge; 3, Dudley; 4, Malverns; 5, May Hill; 6, Dalecarlia; 7, Gotland; 8, Estonia; 9, Holy Cross Mountains;

10, Podolia; 11, Bohemia.

JELL AND SUTHERLAND: SILURIAN RUGOSE CORALS

771

subsp. nov. and E. pseudodicmthus (Weissermel) appear together in the Homerian Much Wenlock
Limestone at Dudley, Worcestershire. In Shropshire, specimens from the middle Homerian Farley
Member of the Coalbrookdale Formation have been referred to both of the above species. The
overlying Much Wenlock Limestone, which along Wenlock Edge is late Homerian (Bassett et al.
1975), yields both species. Both E. articulation anglicum and E. pseudodiant hus are known from the
lower Elton Beds of early Gorstian age (Holland et al. 1980).

text-fig. 2. Correlation of standard Silurian chronostratigraphic units and the principal entelophylloid
bearing sequences of northern Europe; correlations based on Holland and Bassett (1989).

Gotland

Where possible, Gotland collections are related to localities of Laufeld (1974) and to geological
mapping of Hede (1960). The earliest occurrence is Entelophyllum dendroides sp. nov. (Text-fig. 3)
from the Telychian Roda Lagret, which crops out just below sea level north of Visby.

Prohexagonaria favia sp. nov. occurs in the Sheinwoodian Upper Visby Beds. Wedekind (1927)
described E. visbyense and E. anschutzi from early Sheinwoodian Hogklint Beds at Visby. Neuman
and Hanken (1979) listed E. visbyense in the top of the Upper Visby Beds southwest of Visby. We
consider these two species synonymous and probably not referable to Entelophyllum s.s.

The neotype of E. articulation is from an unknown locality and horizon on Gotland but it is
closely similar to specimens occurring in the early Homerian Slite Beds at Bogeklint 1 . It also occurs
in the Homerian Mulde and Klinteberg Beds and its highest occurrence is in the Gorstian lower part
of the Hemse Beds at Snoder 1.

Entelophyllum proliferum (Dybowski) was described from the island of Stora Karlso in the upper
part of the Slite Group (late Sheinwoodian to early Homerian age). From the northwestern coast
of Stora Karlso we describe Donacophyllum neumani sp. nov. and D. wallstenense sp. nov., which
are similar to E. articulation except in mode of increase and in development of lonsdaleoid
dissepiments.

772

PALAEONTOLOGY, VOLUME 33

E. fasciculatum Wedekind, which descended from E. articulation , is common in the upper Hemse
and Eke Beds, the shelly faunas of which suggest an early Ludfordian age (Bassett and Cocks 1974).
Of heavily carinate forms, E. lauense sp. nov. appears in the Eke Beds and E. hamraense sp. nov.
occurs in the late Ludfordian Hamra and Sundre Beds.

Central Sweden

The only suggestion of Entelophyllum from mainland Sweden is one specimen described by
Lindstrom (1880) as Cyathophyllum dalecarlicum. No further material has been found and its
assignment is questionable. It is from the Telychian (based on graptolites listed by Hede 1958) in
central Sweden.

Estonia

Petrozium losseni (Dybowski) from the late Rhuddanian, upper Juuru Horizon is the oldest known
entelophylloid from Estonia (Text-fig. 3). Kaljo (1970) listed E. articulation in the Telychian
Adavere Horizon, E. articulation and E. pseudodianthus in the late Ludfordian Kuressaare Horizon
and in the Pridoli Ohesaare Horizon (Text-fig. 2). These forms are almost certainly not conspecific
with English and Gotland species. Fedorowski and Gorianov (1973) referred one specimen from
Gorstian-Ludfordian Paadla Horizon and two specimens from the late Ludfordian Kuressaare
Horizon to E. confusion (Pocta).

Podolia

Bulvanker (1952) described Xy lodes nikiforovae sp. nov. from the Malinovsty and Skala Horizons
but this was later referred to Spongophylloides by Sytova (1968). We have compared specimens from
the Gorstian lower part of the Malinovtsy Horizon to E. fasciculatum Wedekind.

Poland

Rozkowska (1962) described E. pseudodianthus (Weissermel) from the upper part of the greywacke
sequence of the lower Rzepin Beds in the Lezyce-Belczoross section in the Holy Cross Mountains.
This unit equates with the lower Pridoli (Bassett et al. 1982), lowermost part of the Podlasie Horizon
(Tomczyk 1970). The form may well belong to Entelophyllum but positive identification from
available line drawings is not possible.

Czechoslovakia

Entelophylloid corals from Bohemia, first described by Pocta (1902) and revised by Prantl (1940),
are similar to the English E. articulation anglicum - E. pseudodianthus forms and became even more
carinate. Prantl (1940) recorded the less carinate E. articulation and E. pseudodianthus transiens
from the upper Motol Formation (Homerian) extending into the lower Kopanina Formation
(Gorstian); specimens from the ‘Amerika’ quarries are now thought to be from the Kopanina
Formation (Galle pers. comm.). E. prosperum and E. prosperum crassum are common in the
Kopanina Formation. Prantl’s material was from Tachlovice but today the precise locality is not
known and it is thought to have been in a mine (Galle pers. comm.). E. confusum is known from
the upper Motol and Kopanina Formations.

PHYLOGENY

In northern Europe, the earliest members of this group are referred to Petrozium (Text-fig. 3) and
are characterized by a less well developed biserial tabularium than in later forms, and by stout
straight long septa, and a relatively narrow dissepiinentarium. The first species with a well-
developed biserial tabularium occurs in the Telychian of Gotland but it still shows simple septa and
a narrow dissepimentarium. Typical Entelophyllum is not known from the early Sheinwoodian of
northern Europe.

E. articulation with well developed long septa, biserial tabularium, and wide uniform

JELL AND SUTHERLAND: SILURIAN RUGOSE CORALS

773

RHUDDANIAN

AERONIAN

TELYCHIAN

SHEINWOODIAN

HOMERIAN

GORSTIAN

LUDFORDIAN

PRIDOLI

SERIES

STAGES

‘Ft

’. dewan •’

‘E. articulatum anfilicum
— E. pseudodianthus

BRITAIN

'£. dt

ndroide

—

E. pro

iferum — ■ — ■

E?'visbyense

? datecarticum
— - F.favia

——— D. new
T>. wait

I 1

g 1

rticu(atum

E. fasciculatum

E. [quense

— E. hamraense
— E. sp. !A
— F. £ot(andica

to

£

m

a

m

z

/ m

h

i

fz

f O

r_

>

Te. Cos

seni

• • •

2

. articufatum

' E . "pseudodia

thus •••••

\E. pseudc

dianthus !

E. cf. fast

icuCatum

|

• • «

• ••■£. pseudodianthus transiens

; •••• E. prosperum

• ••••■£. prosperum crassum

BOHEMIA

text-fig. 3. Stratigraphic distribution of species of Entelophyllum and related genera in northern Europe. Bold
solid lines indicate confirmed identifications and stratigraphic ranges; bold dashed lines indicate confirmed
identifications but stratigraphic ranges doubtful; dotted lines indicate identifications and stratigraphic ranges

not confirmed by the authors.

dissepimentarium appears in the early Homerian in Gotland. It continues through the Gorstian and
is thought to give rise to E. fasciculatum in the early Ludfordian, with the withdrawal of the septa
from the axis, a reduction in the width of the dissepimentarium, an increase in the size of
dissepiments, and a less well defined separation of axial and periaxial tabellae. The middle to late
Ludfordian E. lauense and E. hamraense , with heavily carinate septa, possibly evolved from this
stock. In England, E. articulation anglicum , with light carination of some septa, occurs later in the
Homerian than E. articulation articulation in Gotland. The English subspecies also dies out earlier

774

PALAEONTOLOGY, VOLUME 33

and again this may be related to changes in facies. In England, forms with different colony shapes
and with strongly carinate septa, referred to E. pseudodianthus occur throughout the range of E.
articulation anglicum. Similar variation is found in the Bohemian forms of the E. articulation - E.
pseudodianthus transiens - E. prosperum - E. prosperum crassum group. Here they have a range
slightly longer than the English forms but within the range of E. articulation in Gotland. In contrast,
the latter species is recorded in Estonia in Telychian as well as late Ludfordian and Pridoli intervals
(Text-fig. 3). These presumed occurrences and others from elsewhere in the world may have had the
name E. articulation loosely applied.

Cerioid species with typical entelophylloid internal features ( Prohexagonaria favia and P.
gotlandica ) occur at widely separated horizons on Gotland (Text-fig. 3). Whether these represent
iterative evolution from the Baltic entelophylloid stock or successive migrations from an evolving
Prohexagonaria lineage elsewhere, must await detailed study of Prohexagonaria.

Two new species of Donacophyllum in Gotland (Text-fig. 3) have similar internal features to E.
articulation but show non-parricidal rather than parricidal increase and commonly have well
developed lonsdaleoid dissepiments. This generic similarity suggests affinities between Entelo-
phyllidae and Kyphophyllidae.

PALAEOBIOGEOGRAPHY

The first occurrences of entelophylloid forms are from opposite sides of the Iapetus Ocean. In
Llandovery times Petrozium dewari , from Girvan, Scotland, would have been on the southeastern
seaboard of the North American Continent, and P. losseni , from Estonia, on the Anglo-Baltic Plate
(Text-fig. 1b; Cocks and Fortey 1982). In later Llandovery times as the Iapetus Ocean closed P.
dewari is known from England and the first species of Entelophyllum occurs in Gotland. Although
on the same plate, England and Gotland were separated by a broad sea on the site of the earlier deep
Tornquist’s Sea. This probably led to differentiation of subspecies in E. articulation , earlier
development of carinate forms in Britain, and evolution of E. fasciculatum in Gotland. Bohemian
and English forms show a similar development, reflecting the closer proximity of England to the
northern Gondwanan Continent in the Llandovery as the Rheic Ocean opened.

SYSTEMATIC PALAEONTOLOGY

Terminology and suprageneric classification follow Hill (1981 ). Measurements are standard, with the diameters
of corallites being the mean of the diameters of the circumscribed and inscribed circles in transverse section.

Septal microstructure is similar in all species described. Septa are monacanthate with the monacanths
arranged in broad half fans over the dissepimentarial floors and inclined upwards and inwards towards the axis
in the tabularium. Monacanths vary from 0-01 to 0 02 mm in diameter and the fibres appear to diverge only
slightly from the axis of the trabeculae. Thus axes of trabeculae and junctions between trabeculae are not
obvious so that in median longitudinal section septa have a fibrous appearance (PI. 1, fig. 2; PI. 5, fig. 20).

Variation and taxci discrimination

Large samples of E. articulation anglicum subsp. nov. and of E. yassensis (Etheridge) from Australia, to be
described in a future paper, indicate considerable intra-colony and intra-specific variation. Previously,
characters showing intra-colony variation have been used to differentiate species and even genera by some
authors.

In E. articulatum anglicum , variation in colony form appears to be dependent on environmental conditions.
Flattened colonies are most common in bedded limestones and marls, whereas tall globular heads are found
in more massive biohermal limestones. Ranges of variation in corallite diameter and septal number are almost
as great for the holotype (10-19 mm, 21-32 major septa) as for the populations from Dudley (twenty-four
specimens: 8-19 mm, 21 32 major septa) and Wenlock Edge (forty specimens: 9-20 mm, 20-32 major septa).
Mean colony corallite diameters are likewise variable but mean colony tabularium diameters remain fairly
constant. There seems to be no relationship between corallite diameter or septal number and corallum form.
Major septa are generally long, straight, smooth and extend to the axis, but within one corallum they may tend

JELL AND SUTHERLAND: SILURIAN RUGOSE CORALS

775

to withdraw from the axis, become discontinuous at the periphery, and carry zigzag carinae. Discontinuity in
the dissepimentarium appears to be connected with widening of the dissepimentarium associated with
connecting processes. In several sections of one corallum, septal carination is confined to the oldest part of
the corallite just before budding. However, in other specimens development of carinae is spasmodic throughout
the corallite. In contrast, septal microstructure, tabularial characters, and dissepimentarium apart from its
width, all show little variation.

E. pseudo dianthus (Weissermel), which occurs with and resembles E. articulation anglicum typically but not
invariably has heavily carinate septa. Similarly, on Gotland, most specimens of E. fasciculatum are readily
distinguished from E. articulation but in some colonies a few corallites show characters typical of the other
species. The ranges of variation of these stratigraphically separated forms also show considerable overlap.

It is tempting, after admitting such considerable variation, to lump together all the above mentioned forms
into one species. This would obscure the biostratigraphic and palaeogeographic usefulness of these groups of
corals in northern Europe. Assignments are subjective and may well need future modification but they will
provide a basis for the careful description of forms from elsewhere. Diagnoses formulated are for typical
coralla and corallites within coralla, and the exceptions are noted but not included.

At generic level, the fasciculate and cerioid forms are differentiated, as are those with parricidal and
nonparricidal colony increase. But even at this level intracolony variation has produced exceptions with some
cerioid forms becoming locally fasciculate and some colonies with typical parricidal increase having one or two
corallites that produced lateral nonparricidal offsets. Petrozium is retained for the simpler early forms without
well-developed biserial tabularia and dissepimentaria, since it appears to have stratigraphic significance.

Synonymy

We have used many of the symbols proposed by Matthews (1973) to indicate degree of confidence in synonymy
listings. These lists include only the most important previous citations, all of which we have examined and
evaluated.

Collections

The repositories in which the material described herein is housed are indicated by the following
prefixes: BMNH, British Museum (Natural History), London; BU, University of Bristol, Bristol; EGM, Eesti
NSV Teaduste Akadeenna, Geoloogia Instituudi, Tallinn; GSM, Institute of Geological Sciences, Keyworth;
NM, Narodni Museum, Prague; OU, University of Oklahoma, Norman; RM, Paleozoologiska sectionen,
Naturhistoriska Riksmuseet, Stockholm; SM, Sedgwick Museum, Cambridge; SMF, Natur-Museum
Senckenberg, Frankfurt-am-Main; UB, University of Birmingham, Birmingham; UO, University of Oxford,
Oxford; UQ, University of Queensland, Brisbane.

Subclass rugosa Milne-Edwards and Haime, 1850
Order stauriida Verrill, 1865
Suborder arachnophyllina Zhavoronkova, 1972
Family entelophyllidae Hill, 1940
Genus entelophyllum Wedekind, 1927

1927 Entelophyllum Wedekind, p. 11.

1927 Xylodes Lang and Smith, p. 461 (pre-occupied by Xylodes Waterhouse, 1876, a Recent
coleopteran).

1944 Stereoxylodes Wang, p. 25.

1964 Carinophyllum Strelnikov, p. 59.

Type species. Chosen by subsequent designation of Lang et al. 1940, p. 57, Entelophyllum articulation
(Wahlenberg, 1821) from the Slite, Mulde, Klinteberg, and lower Hemse Beds of Gotland; Homerian to early
Ludfordian Stages, late Wenlock to early Ludlow.

Diagnosis. Phaceloid or dendroid rugosans with peripheral parricidal increase; septa long, generally
radially arranged, counter-cardinal septa rarely distinguishable, smooth or asymmetrically carinate;
major septa slightly withdrawn from axis; minor septa contraclined or contratigent in some;
tabularium wide, broadly domed commonly with depressed axial area and marginal periaxial

776

PALAEONTOLOGY, VOLUME 33

trough formed by small subhorizontal or concave tabellae; dissepiments numerous, small, globose
with lonsdaleoid dissepiments in some.

Remarks. Wedekind placed eight species in Entelophyllum , but named no type species. Among these
was E. articulation , cited without any author’s name or statement that it was new. Lang et a/. (1940)
considered that Madreporites articulatus Wahlenberg was implied, and with this we would agree. By
selecting E. articulation (Wahlenberg) as type, the invalid homonym Xylodes Lang and Smith
became an objective synonym of Entelophyllum.

Soshkina and Dobrolyubova (1962, p. 333) listed Evenkiella Soshkina, 1955, as having
Madreporites articulatus Wahlenberg as the type species, and gave Xylodes as a synonym. Soshkina
(1955) had, however, named as type species of Evenkiella nom. nov., Evenkiella helenae Soshkina,
1955 (p. 126, pi. 13, fig. 1) from the Wenlock of the Stony Tunguska River, Siberia. In a footnote
she explained ‘name Evenkiella for the genus given instead of Xylodes Smith and Tremberth,
previously used for an insect'. In our opinion the designation of E. helenae as type species is of
stronger force than the citation ‘nom. nov. ’ instead of ‘gen. nov. ’ and the explanatory footnote, and
we uphold E. helenae as type in spite of Soshkina and Dobrolyubova’s later reference to M.
articulatus as the type. We consider the cerioid Evenkiella distinct from Entelophyllum in that it has
lonsdaleoid dissepiments, flat complete tabulae, septa composed of fine trabeculae, and we agree
with its placement in the Kyphophyllidae by McLean and Pedder (1984, p. 18).

Stereoxylodes was proposed by Wang (1944) as a subgenus of Entelophyllum , with type
Cyathophyllum pseudodianthus Weissermel, 1894, for entelophyllids with dilated and carinate septa.
The type specimen of C. pseudodianthus chosen by Lang and Smith (1927, p. 473) as the original
of Weissermel (1894, pi. 47, fig. 3) from the glacial drift of Lauth, Germany, is lost. Ivanovskii
(1976, p. 165) invalidly proposed as a neotype of E. pseudodianthus , the holotype of Stereoxylodes
pseudodianthus var. sinensis Wang, 1944 from the ‘middle’ Silurian of Malung, eastern Yunnan,
China. To stabilize this species, we have selected a lectotype from Weissermel’s syntypes, see
discussion of E. pseudodianthus. In England, E. pseudodianthus commonly occurs as small loose
clumps of ceratoid corallites formed by peripheral parricidal budding and with strongly carinate
and commonly thickened septa, quite distinct from the typical phaceloid coralla of E. articulation
anglicum with cylindrical corallites and typically little or only light carination of the septa. As
described in detail later, some specimens of both species show variation that liken individual
corallites or whole coralla to those of the other. Prantl (1940) also recognized a similar tendency and
proposed E. pseudodianthus transiens for forms showing smooth, slightly carinate, dilated and
asymmetrical carinate septa combining characters of E. articulation and E. pseudodianthus.
Similarly, Strelnikov (1964, p. 59) erected Carinophyllum for species that have septa thickened in the
inner dissepimentarium and are heavily carinate, and designated E. confusion (Pocta, 1902) as type
species. Like Schouppe (1951), we believe that there is a continuum of variation in the thickening
and carination of the septa in the dissepimentarium of such entelophylloid corals and that they all
belong to the one group. They may be arbitrarily divided into subgenera but there does not appear
to be any biogeographic or stratigraphic significance to such divisions. If such a subdivision was
practical, then Nanshanophyllum Yu (1956; type N. typicum Yu, Middle Silurian, China) might be

EXPLANATION OF PLATE 1

Figs 1-10. Entelophyllum articulation (Wahlenberg). 1-6, RM Cn54823 (neotype), unknown horizon and
locality on Gotland ; I , lateral view showing tall corallites, x 1 ; 2 and 3, longitudinal sections showing typical
biserial tabularium and small globose dissepiments, x4, x 2, respectively; 4-6, transverse sections showing
long smooth septa, x 2, x 2, x4 respectively. 7-10, Slite Beds, Bogeklint 1, 2 km SE of Boge church,
Gotland; 7 and 8, UQ F34300; 7, longitudinal section, x 2; 8, transverse section, x 2; 9 and 10, UQ F34298;
9, longitudinal section, x2; 10, transverse section, x2.

PLATE I

JELL and SUTEIERLAND, Entelophyllum

778

PALAEONTOLOGY, VOLUME 33

regarded as another subgenus of the group for solitary forms with thickened and carinate septa, in
the same way as Pedder (1976) used it as a subgenus of Stereoxylodes.

Although some species contain both cerioid and fasciculate forms (such as E. polymorphum
Shurygina, 1977, pi. 6, figs 1-3 [= E. articulation of Sytova 1952, p. 140, pi. 4, figs 1-5] from the
‘upper’ Silurian of the Urals), we refer typical cerioid species to Prohexagonaria Merriam, 1973 (see
discussion of that genus).

We include in Entelophyllum only species that show parricidal budding. One apparent exception
is E. articulation anglicum. It almost invariably has obvious parricidal budding but we have
observed one corallum, which, in addition to common parricidal budding, includes a single example
of lateral budding. Similarly, lonsdaleoid dissepiments are uncommon in most species of
Entelophyllum but do occur in some corallites in coralla of various species especially in areas of
lateral outgrowths and in the late growth stages. Forms showing lateral budding and well-developed
lonsdaleoid dissepiments are included in Donacophyllion Dybowski, 1874. The two species here
included in it are similar to Entelophyllum in their septa and tabularia.

Entelophyllum articulation (Wahlenberg, 1821)

Plate 1, figs I 10; Plate 2, figs 1-11

(1821)

v*1837

1874

1927

vl929

vl933
1940
( ?) 1 940

Madreporites articulatus Wahlenberg, p. 97.

Cyathophyllum articulation (Wahlenberg); Hisinger, p. 102, pi. 29, fig. 4.

Cyathophyllum articulation (Wahlenberg); Dybowski, p. 435, pi. 3, fig. 1, 1 a, 1 b.
Entelophyllum articulation (Wahlenberg); Wedekind, p. 22.

Xylodes articulatus (Wahlenberg); Smith and Tremberth, p. 363, text-figs 1 and 2; pi. 7, fig. 5
[non figs 1-4, 6].

Xylodes articulatus (Wahlenberg); Smith, p. 513, pi. I, figs 4 and 5 [ non figs 1-3],
Entelophyllum articulation (Wahlenberg); Lang et al., p. 57.

Xylodes articulatus (Wahlenberg); Prantl, p. 6, pi. 1, figs 1-3; pi. 2, figs 1, 3, 4.

Neotype. Chosen by Smith and Tremberth (1929, p. 363), one of the original specimens of Hisinger labelled
'Cyathophyllum articulation e Gottlandia’; RM Cn54823, from an unknown locality and horizon on Gotland.

The exterior side view of the neotype was figured at half size by Smith and Tremberth (1929, p. 365, text-
fig. 1 ). Three thin sections cut from the neotype with the numbers R49348a, 6, and c are in the British Museum
(Natural History). The two transverse sections, apparently cut from the base of the specimen, include two
corallites each and there is a longitudinal cut from a single corallite, apparently from high on the side of the
specimen. Smith and Tremberth (1929, p. 365, text-fig. 1 ) figured one of the four transverse corallites and Smith
(1933, pi. 1, figs 4 and 5) refigured that picture and also figured the longitudinal section. At a later time two
additional transverse sections were taken from the base of the neotype, which include two corallites each, and
a longitudinal section was taken from one of these corallites. These sections carry the same number as the
neotype, Cn54823, in the Riksmuseet, Stockholm (PI. 1. figs 3-6).

Material studied. The neotype; similar material occurring in the Slite Beds at Bogeklint 1, northeast coast of
Gotland (6 specimens; figured: UQ F34298, F34300). Other Gotland occurrences; upper Mulde Beds, from
Blahall 1, on west coast of Gotland (one specimen; figured: RM Cn66044); Klinteberg Beds, from

EXPLANATION OF PLATE 2

Figs 1-11. Entelophyllum articulation (Wahlenberg), showing variation in Gotland material. 1 and 2, RM
Cn66040, Klinteberg Beds, Klinleberget 1, cliff exposure at Klinte, west coast of Gotland; 1, transverse
section, x2; 2, longitudinal section, x 2. 3-9, lower Hemse Beds, Snoder 1, 2 5 km NW of Slite Church,
southwest Gotland; 3-5, RM Cn66041 ; 3, longitudinal section, x2; 4, transverse section showing lateral
expansions of corallites, x2; 5. lateral view showing parricidal increase, x 1; 6 and 7, RM Cn66042; 6,
transverse section, x2; 7, longitudinal section, x2; 8 and 9, RM Cn66043; 8, transverse section, x2; 9,
longitudinal section, x 2. 10 and II, RM Cn66044, upper Mulde Beds, Blahall 1, 0 5 km NE of Djupviks
fislage, west coast of Gotland; 10, transverse section, x2; 11. longitudinal section, x 2.

PLATE 2

JELL and SUTHERLAND, Entelophyllum

780

PALAEONTOLOGY, VOLUME 33

Klinteberget 1, central western Gotland (seven specimens; figured: RM Cn66040); Lower Hemse Beds,
from Snoder 1, southwest Gotland (four specimens: figured RM Cn66041, Cn66042, Cn66043).

Distribution. Gotland: Homerian and Gorstian; Estonia: ?Telychian to Pridoli; Bohemia: ?late Homerian to
early Gorstian.

Diagnosis. Phaceloid, increase peripheral, parricidal; corallite diameter 10-15 mm. Major septa
typically 23-28, long, thin, extending to or almost to axis, rarely irregularly zigzag to flexuose in
outer dissepimentarium, rarely carinate; minor septa two-thirds the length of major, may be
irregularly contratigent, Tabularium one-third to one-half diameter of corallite, distinctly biserial
with well developed periaxial trough. Dissepiments small, globose, to medium and irregular in size.

Description. Coralla are typically large, tall and phaceloid. Corallites are subcylindrical, closely packed, sub-
parallel and up to 16 mm in diameter with irregular growth contractions and expansions (PI. 1, fig. 1). The
neotype is a fragment about 130 mm in length and 50 mm in diameter (PI. 1, fig. 1). It includes parts of four
corallites that represent one generation, approximately 100 mm in length. At the top of the fragment ten
corallites are to be seen. Increase is peripheral and parricidal with buds extending almost straight up from the
dissepimental zone of the parent corallite. Commonly, budding is more frequent than in the neotype and the
corallites are less than 100 mm in length.

In transverse section major septa are typically long and thin but are highly variable even in the neotype.
Major septa number 23-32 at diameters of 12-16 mm. Septa typically extend to the axial region but do not
form an axial structure. A small open axial area about 1 mm in diameter occurs in some corallites. Most septa
are thickened in the outer dissepimentarium and taper evenly towards the axis but in some corallites they thin
in the tabularium to about half their thickness in the dissepimentarium. Some are irregularly zigzag to sinuous
in the dissepimental area and are locally carinate. Some corallites show no carinae at all. Minor septa are about
two-thirds the length of the major septa and commonly extend just past the inner margin of the dissepimental
zone. However, in two specimens they are four-filths the length of the major septa and extend into or across
that part of the tabularium that forms the outer trough. Some minor septa are irregularly contratingent but
they form no pattern. In some corallites, minor septa are discontinuous. In such cases, herringbone
dissepiments are generally present. Locally, a few major and minor septa become discontinuous near the
periphery and lonsdaleoid dissepiments are developed (PI. 1, fig. 10). These discontinuous septa coincide with
a lateral extension of the margin of the corallite that may represent a lateral outgrowth of attachment and the
two developments are presumably related.

The tabularium is typically 4-5 mm in diameter. Longitudinal sections show two series tabellae (PI. 1, figs
2 and 3; PI. 2, fig. 1 1). Those in the axial region are generally flat but sag axially in those corallites with the
septa slightly withdrawn from the axis, and have sharply downturned marginal edges. Between this zone and
the dissepiments lies a narrow zone occupied by flat, sagging or inclined tabellae that number 5-8 in 5 mm
vertical spacing, and form a periaxial trough. Small arched tabellae may be situated at the margins of the flat
central region, forming levy-like margins to the central platform (PI. 1, fig. 9). Dissepiments are small, globose,
and form a wide peripheral zone that varies considerably in relative width as a result of rejuvenescence, from
about one half to two-thirds the corallite radius. The number of rows of dissepiments varies from 3 to 12 and
dissepiments are uniform in size and distribution and number 13-14 in 5 mm vertically.

Remarks. In Gotland, E. articulatum occurs in the Slite, Mulde, Klinteberg and Lower Hemse Beds.
Even in the Slite, in comparison with the neotype, some specimens have a comparatively narrower

EXPLANATION OF PLATE 3

Figs 1-10. Entelophyllum articulatum anglicum subsp. nov., showing variation in holotype, SM A5143, Much
Wenlock (Dudley) Limestone, Dudley, Worcestershire. 1, transverse section of typical corallites, x 2. 2 and
3, longitudinal sections, x 4, x 2 respectively. 4, transverse section showing corallite expansion, x 4. 5,
transverse section showing some septa slightly thickened and zigzag, x 4. 6, transverse section of an early
growth stage, x 4. 7, lateral view of corallum, x 1 8, transverse section, showing lightly carinate septa, x 4.
9, lateral view showing parricidal increase, x 2. 10, transverse section in the distal part of corallum showing
lonsdaleoid dissepiments, x 4.

ma w

PLATE 3

JELL and SUTHERLAND, Entelophyllum

782

PALAEONTOLOGY. VOLUME 33

dissepimental zone that includes some larger dissepiments (PI. 1, fig. 9). Stratigraphically higher,
budding patterns are more irregular, offsets are shorter, dissepiments are comparatively larger and
more irregular in size, and septa tend to withdraw slightly from the axial region leaving a small open
axial area. Beginning in the lower Hemse (PI. 2, figs 2-4) there is an apparent shift in growth form
in that there is a more frequent development of new parricidal offsets but adult corallites remain
typical.

Carinae are lacking in most specimens from Gotland. Where present they occur in the
dissepimental zones of only a few septa. None of the specimens from Gotland develops complex
vacuolar septa as occur in some corallites of the English subspecies E. articulation anglicum.

In Estonia, Kaljo (1970) listed but did not describe or illustrate E. articulation in coral faunas
from the Adavere (Telychian), Kuressaare (late Ludfordian), Kaugatuma (early Pridoli) and
Ohesaare (late Pridoli) Horizons (Text-fig. 2).

Smith and Tremberth (1929, p. 366) cited E. articulation in the Upper Silurian of the Island of
Bjerkoy, Christiania Fjord, Norway, but B. Neuman (pers. comm.) has been unable to find
additional specimens in that area.

Smith and Tremberth (1929, p. 363) suggested that many Bohemian specimens (Pocta 1902)
figured as Cyathophyllum prosperum might be synonymous with E. articulation. Prantl (1940) did
not include any of Pocta’s figured material in E. articulation but described E. articulation from the
'Amerika’ quarries, near Morina (Budnany). Dr A. Galle (pers. comm., 1989) considers these to be
from the Kopanina Formation (Text-fig. 2). Occurrences in the Motol Formation (Prantl 1940)
need to be confirmed. The septa are smooth, straight, extending almost to the axis and the
tabularium and dissepimentarium are very similar to the Gotland material.

Material from the USSR, China, and Middle East, in Flandovery to Early Devonian strata,
previously referred to E. articulation , will now have to be reinterpreted.

Entelophyllum articulation anglicum subsp. nov.

Plate 3, tigs 1-10; Plate 4, figs 1-1 1

v*1855 Cyathophyllum articulation (Wahlenberg); Milne-Edwards and Haime, p. 282, pi. 67, fig. 1, la.

vl929 Xylodes articulatus (Wahlenberg); Smith and Tremberth, 1929, p. 363, pi. 7, figs 1M, 6; pi. 8,

fig. 2 ( non pi. 7, fig. 5).

Holotype. Milne-Edwards and Haime, 1855, pi. 62, fig. 1, la; SM A5134a and thin sections SM A51346-y from
the Fletcher Collection, Much Wenlock (Dudley) Limestone, Dudley, Worcestershire; Homerian Stage.

Material studied. We examined sixty-four sectioned specimens from the collections of the BMNH, SM, UO,
UB, and UQ from the Much Wenlock Limestone at Dudley, Worcestershire, and along Wenlock Edge,
Shropshire; figured; SM A5134 (holotype), UQ F3531 1, F35315, F35319, F36966. Other English occurrences:
Farley Member, Coalbrookdale Formation, Wenlock Edge (figured: UQ F41347); lower Elton Beds,
Leintwardine, Shropshire; several specimens labelled lower Elton Beds, Wenlock Edge; several specimens
labelled lower Ludlow (presumably Elton Beds), Ledbury Quarry, Malverns, Herefordshire.

EXPLANATION OF PLATE 4

Figs 1-11. Entelophyllum articulation anglicum subsp. nov. showing variation in English material. 1-4, UQ
F55311, Much Wenlock (Dudley) Limestone, Dudley, Worcestershire; 1-3, transverse sections, x2, x4,
x4, respectively; 4, longitudinal section, x 2. 5 and 6, UQ F41347, Farley Member, Coalbrookdale
Formation, Presthope tunnel cutting, Wenlock Edge, Shropshire; 5, transverse section showing lateral
dissepiments, x2; 6, longitudinal section, x2. 7 and 8, UQ F35315, Much Wenlock Limestone, Wenlock
Edge, Shropshire; 7, transverse section, x4; 8, longitudinal section, x4. 9-11, Much Wenlock Limestone,
Lilleshall Quarry, Wenlock Edge, Shropshire; 9 and 10, UQ F36966; 9, transverse section, x2; 10,
longitudinal section, x2; 11, UQ F35319, transverse section showing thickened and carinate septa, x 4.

PLATE 4

JELL and SUTHERLAND, Entelophyllum

784

PALAEONTOLOGY, VOLUME 33

Distribution. England: middle Homerian to early Gorstian.

Diagnosis. Coralla tall phaceloid to squat fasciculate; increase typically peripheral parricidal;
corallite diameter typically 10-12 mm, maximum 20 mm; major septa 21-24, maximum 30, straight
to flexuose, smooth in tabularium, commonly reaching axis but not confluent; septa vary from
straight, smooth or slightly thickened to zigzag and distinctly carinate, becoming very ragged and
interrupted by lonsdaleoid dissepiments in the lateral expansions; minor septa commonly
contraclined to contratigent around an obscure cardinal-counter plane. Tabularium biserial with
well developed periaxial trough; dissepiments small, globose, regularly arranged in 5-8 vertical
rows.

Description. Colony form varies from small clump-shaped coralla, 60 mm high and 40^45 mm in diameter
showing two generations of budding, to squat or flat cake-shaped coralla several times greater in diameter than
in height, up to 300 mm diameter, to tall large globular heads up to 1 m diameter. The holotype is a fragment
of a tall, phaceloid colony 160 x 80 mm in cross section and 130 mm in height (PI. 3, fig. 7).

Protocorallites, where preserved, vary from low petalate, up to 35 mm in diameter, to tall ceratoid corallites,
15 mm in diameter and 80 mm tall. Six to ten buds are generally produced at the first generation. In most
coralla, budding occurs at the same level throughout the colony with the corallites in each generation being
of the one length. In flattened colonies divergence of daughter corallites from the axis of the parent is rapid.
In several flattened colonies growth increases on one side of the corallum. Budding is parricidal but in two
cases, UB 42 of Holcroft Collection and UQ F36996, it is nonparricidal. Corallum increase appears to be
peripheral, with one to six daughter corallites originating in the peripheral parts of the parent corallite generally
at an expansion. The buds abruptly expand so as to occupy most of the calyx of the parent. Thin sections show
buds originating at or just axial to the dissepimentarium/tabularium boundary.

Corallites are long, slender, cylindrical, and closely spaced, up to 10 mm. Corallites show regular
contractions and expansions (PI. 3, fig. 11). At intervals, a corallite may expand laterally to abut onto an
adjacent corallite forming a connecting process 5-10 mm high. This pattern of outgrowths is not regularly
spaced or level specific within a colony. Thus in transverse section the corallites show considerable variation
in diameter (8-20 mm) but are commonly 10-13 mm. In the holotype first generation corallites are 20-30 mm
long, those of the second 50-60 mm, and of the third 50-80 mm.

Septa are typically long and radially arranged in two orders (PI. 3, fig. 1); one order commonly has 21-24
septa in adult stages at diameters 10-13 mm. At diameters 1 7-20 mm, major septa number 28-32. Major septa
typically almost reach the axis; in some corallites an open axial area may be 1-2 mm in diameter (PI. 3, figs
4-6). In a few corallites axial ends of the major septa deflect towards the cardinal-counter plane and the
counter-cardinal septa may bisect the axial space. In others axial ends may be irregularly twisted. Minor septa
are 0-6-0-8 of major septa and extend a short distance into the tabularium. Some minor septa may be
contraclined and even contratigent. Septa are smooth and straight, curved or flexuose in the tabularium but
are rather variable in the dissepimentarium even within one corallite. They may be straight or flexuose with
smooth, rough or ragged sides (PI. 3, fig. 8); slightly thickened at the junction of the septa with the curved
dissepimental plates and carrying irregular zigzag carinae (PI. 3, fig. 5); thickened in the inner dissepimentarium
and outer tabularium so that they are fusiform (PI. 4, fig. 11); or uncommonly discontinuous with irregular
segments being interrupted by lonsdaleoid dissepiments (PI. 3, fig. 10). Discontinuity of septa in the peripheral
parts occurs in areas of broad expansion of the dissepimentarium (i.e. in a connecting process or just before
budding). Septa may become more ragged or carinate in the expansions (PI. 4, fig. 2).

The tabularium is relatively constant in diameter in adult stages, being 0-6 the corallite diameter. It shows
an irregular separation into an axial series and a periaxial series of tabellae (PI. 3, fig. 2). The axial series forms
tabularial floors that are flat or slightly sagged axially and with steep downturned marginal slopes. Between
this zone and the dissepimentarium is a periaxial series of flat, slightly sagging, or inclined tabellae (up to 30
per 10 mm vertically) forming a periaxial trough of the tabularial floors. The dissepimentarium varies in width
as a result of the contractions and expansions of the corallite. Dissepiments are small, globose, and uniform
in size and distribution (20-28 in 10 mm vertically and 6-8 total rows horizontally).

New offsets arise (BMNH R267) from the outer half of the periaxial series of tabellae and the axial edge of
the dissepimentarium. The buds arise within the mterseptal loculii between major and minor septa and thus
budding is marginal and not truly tabularial. Having developed in the inner part of the marginarium they
diverge towards the periphery and increase rapidly occupying the total dissepimentarium.

The twenty-five peels from UQ F36996 show close relationship between carinae and stages of parricidal

JELL AND SUTHERLAND: SILURIAN RUGOSE CORALS

785

increase. In this corallum, growth of a coralhte stops when offsets develop. The septa are simple during growth
but, near the base of the calice, just before budding occurs, septa become thickened, develop rough margins,
become zigzag, and develop rudimentary, short carinae. New offsets almost invariably have simple thin septa
in their early stages of growth. Earlier development of carinae is seen in some corallites in other coralla where
they are not related to the development of offsets. In UQ F8388B, an individual coralhte 50 mm in length has
simple septa in the earliest stage, then carinate septa in a section 10 mm higher, then simple septa still higher,
followed in the highest growth stages by the development of carinate and complex septa with vacuoles. Most
of the other corallites in this same corallum have simple septa throughout their growth.

Remarks. Entelophyllum articulatum anglicum subsp. nov. is close to E. articulation (Wahlenberg,
1821) from Gotland and many corallites cannot be differentiated. With sufficient material anglicum
is seen to have more commonly carinate septa. Both subspecies show parricidal budding but in
anglicum tabularia of the offsets may originate from the peripheral trough of the parent as
illustrated by Smith and Tremberth (1929, text-fig. 2a), whereas in articulatum the indications are
that offsets arise from wholly within the dissepimentarium. Minor differences are the more frequent
occurrence among the English specimens of contratigent minor septa, development of an obscure
cardinal-counter plane and less frequent development of discontinuous minor septa and lonsdaleoid
dissepiments. Sections of E. articulatum anglicum showing more carinate septa resemble E.
pseudodianthus (Weissermel, 1894).

Entelophyllum fasciculatum Wedekind, 1927

Plate 5, figs 1-21

vl927

vl927

vl927

vl927

non 1927
vpl927

71927

71927

71973

Entelophyllum fasciculatum Wedekind, p. 24, pi. 2, figs 1 1 and 12; pi. 29, figs 30, 31, 50; ?pl. 29,
figs 34-49, 51 [not sectioned]; ?pl. 30, figs 1-8 [not sectioned],

Entelophyllum proliferum typus Wedekind, p. 23, pi. 29, figs 21 and 22 [ non Cyathophyllum
proliferum Dybowski, 1874, p. 445, pi. 3, fig 2a, b],

Entelophyllum proliferum var. elongata Wedekind, p. 23, pi. 29, fig. 23; ?pl. 29, figs 24, 25, 33 [not
sectioned].

Entelophyllum proliferum var. brevis Wedekind, p. 23, pi. 29, figs 27 and 728; ?pl. 29, figs 26 and
29 [not sectioned],

Entelophyllum culmiforme Wedekind, p. 23, pi. 29, fig. 32 [= tryplasmid coral],

Entelophyllum roemeri Wedekind, p. 23, pi. 30, fig. 15; ?pl. 30, figs 9-12, 16 [not sectioned] [non
pi. 30, figs 13 (= holotype of E. roemeri) or 14 (both specimens = tryplasmid corals)].
Entelophyllum rhizophorum Wedekind, p. 23, pi. 30, fig. 17.

Entelophyllum confer rhizophorum Wedekind, p. 23, pi. 29, figs 18-20 [not sectioned],
Entelophyllum articulatum (Wahlenberg); Fedorowski and Gorianov, p. 19, pi. 4, figs 1-4.

Lectotype. Here chosen, original of Wedekind, 1927, pi. 29, fig. 30; RM Cn54855, Rhizophyllum limestone,
lower Eke Beds, Lau Backar 1, southeast Gotland; middle Ludfordian Stage.

Material studied. All of Wedekind’s (1927) original specimens of E. fasciculatum were from Eau Backar 1,
southeast Gotland. Additional material examined by us from this locality are in the RM, UQ, OU, and Dr B.
Neuman’s collections. Sectioned specimens examined from Lau Backar number about thirty-five (figured
specimens; RM Cn54855 (lectotype); RM Cn54847, Cn54849, Cn54850, Cn54852, Cn54856, Cn54880,
Cn54881). The species also occurs at the same approximate horizon at Hallsarve 1, 0 4 km east of Lau Backar
(figured RM Cn66045); Kauparvegard I, inland cliff north of Kauparve farmhouse, southern Gotland; and
possibly from an unknown locality, Ostergarn parish, central east coast of Gotland.

Distribution. Gotland: middle Ludfordian.

Diagnosis. Phaceloid, increase peripheral, parricidal; coralhte diameters 10-12 mm. Major septa
typically 20-22, extending into tabularium but not reaching axis, leaving distinct open area in axial
region. Septa typically straight throughout, not zigzag or carinate. Tabularium about one-half
diameter, flat axially, sharply downturned to margin of dissepimentarium producing irregular
periaxial trough. Dissepiments variable in size, width of zone medium.

786

PALAEONTOLOGY, VOLUME 33

Description. Specimens available include fragments of loosely dendroid to phaceloid coralla and numerous
loose broken offsets. The latter are 30-70 mm long and ceratoid to cylindrical. Narrow distal ends of some
loose offsets are turned abruptly laterally representing areas of attachment to parent. Increase is peripheral and
parricidal with 4-6 offsets in each generation but 7 have been observed. Calices are shallow consisting primarily
of a flat-bottomed axial pit coinciding with the tabularium and a marginal platform.

Septa typically number 20-22 of each order at diameters of 1 0— 11-5 mm. Septa are relatively thin being
thicker in the dissepimentanum compared to the tabularium. Septa are variable in the dissepimentarium. Most
are straight and smooth, others are zigzag to sinuose. Isolated carinae are rarely developed in the zigzag septa
of some specimens. Septa become discontinuous in the dissepimental zone only near a lateral extension (PI. 5,
fig. 5). Major septa are irregular in length and are 0- 7-0-8 corallite radius in length leaving an open axial space
up to 2 mm in diameter. Protosepta are not distinguishable and fossulae are not present. Minor septa are
irregular in length and extend past the inner margin of the dissepimental zone.

The tabularium occupies half the corallite diameter. It consists of a broad axial region in which the tabulae
are flat or with a wide, shallow median depression and a narrow periaxial region where edges of the tabulae
are irregularly turned sharply down and in some cases back up again. Thus, an irregular periaxial trough may
be developed that may contain a few horizontal tabellae (PI. 5, fig. 20). Axial tabulae are irregularly spaced
(12-16 vertically in 5 mm). The dissepimentarium is irregular due to variation in size and convexity of medium
to large globose dissepiments that develop between the septa. There are 4-6 rows of dissepiments in a zone
typically 3-4 mm in diameter and there are 8-10 in 5 mm vertically.

Remarks. E. fasciculatum apparently evolved in the Gotland area from E. articulation. The former,
in its common occurrences in the Upper Hemse/Lower Eke differs from forms similar to the
neotype of E. articulation in the Slite Beds in having a distinctly narrower dissepimental zone of
relatively larger and more irregular globose dissepiments, in having septa withdrawn from the axial
region, and in having a poorly defined separation of the tabularium into two zones with a less
distinctly developed periaxial trough. However, E. articulation in the Klinteberg and lower Hemse
is similar to E. fasciculatum with which it could be considered to be gradational, but it has distinctly
more open axial area and less well defined periaxial tabular trough.

Wedekind ( 1927) confused his definition of E. fasciculatum by basing the species primarily on the
growth form of unsectioned broken offsets. He illustrated only two transverse sections. Also, he
named as new (from the same locality, Lau Backar 1) an additional five species and varieties, all
based on unsectioned broken offsets. The holotype or at least one syntype of each of these has been
subsequently sectioned. Each of the following specimens is here named lectotype of the listed
subspecies or variety of Wedekind (1927) and placed in the synonomy of E. fasciculatum :

EXPLANATION OF PLATE 5

Figs 1-21. Entelophyllum fasciculatum Wedekind. 1 19, Rhizophyllum limestone. Eke Beds, Lau Backar 1,
10 km NE of Lau, southeast Gotland; 1-3, RM Cn54855 (lectotype), original of Wedekind (1927, pi. 29,
fig. 30); 1, lateral view; 2, transverse section; 3, longitudinal section; 4 and 5, RM Cn54852, original of
Wedekind (1927, pi. 29, fig. 50); 4, longitudinal section; 5, transverse section; 6 and 7, RM Cn54857, original
of Wedekind (1927, pi. 30, fig. 17), holotype of E. rkizophorum Wedekind; 6, transverse section; 7,
longitudinal section; 8 and 9, RM Cn54853, original of E. roemeri Wedekind (1927, pi. 30, fig. 15); 8,
transverse section; 9, longitudinal section; 10 and 11, RM Cn54856, original of Wedekind (1927, pi. 29,
fig. 3); 10, transverse section; 1 1, longitudinal section; 12 and 13, RM Cn54849, original of Wedekind (1927,
pi. 29, fig. 23), lectotype of E. proliferum elongata Wedekind; 12, longitudinal section; 13, transverse section;
14, RM Cn54881, original of Wedekind (1927, pi. 2, fig. 12), transverse section; 15, RM Cn54880, original
of Wedekind (1927, pi. 2, fig. 1 1), transverse section; 16 and 17, RM Cn54850, original of Wedekind (1927,
pi. 29, fig. 27), lectotype of E. proliferum brevis Wedekind; 16, longitudinal section; 17, transverse section;
18 and 19, RM Cn54847, original of Wedekind (1927, pi. 29, fig. 21), lectotype of E. proliferum typus
Wedekind; 18. longitudinal section; 19, transverse section. 20 and 21, RM Cn66045, Eke Beds, Hallsarve
I, 1 -35 km NE of Lau, Gotland; 20, longitudinal section; 21, transverse section.

All x 4, except Fig. 1 which is x 1 .

PLATE 5

JELL and SUTHERLAND, Entelophyllum

788 PALAEONTOLOGY, VOLUME 33

1. Entelophyllum proliferum typus Wedekind, 1927, pi. 29, fig. 21 (RM Cn54847) (= PI. 5, figs 18
and 17).

2. E. proliferum var. elongata Wedekind, 1927, pi. 29, fig. 23 (RM Cn54849) (= PI. 5, figs 12 and
13).

3. E. proliferum var. brevis Wedekind, 1927, pi. 29, fig. 27 (RM Cn54850, = PI. 5, figs 16 and 1 7).

The following specimens are tryplasmid corals and they must be transferred from Entelophyllum :

1. Entelophyllum culmiforme Wedekind, 1927, pi. 29, fig. 32 (RM Cn54854, holotype).

2. E. roemeri Wedekind, 1927, pi. 30, fig. 13 (RM Cn54858, holotype), fig. 14 (RM Cn54859).
Specimen RM Cn54853, the original of E. roemeri Wedekind, 1927, pi. 30, fig. 15, is referable to E.
fasciculatum (= PI. 5, figs 8 and 9).

The holotype of E. rhizophorum Wedekind (1927, pi. 30, fig. 17; RM Cn54857, = PI. 5, figs 6 and
7) is questionably placed in the synonomy of E. fasciculatum as it is slightly larger and has some
carinae. It is from Ostergarn but the exact stratigraphic horizon is not known in the Hemse Beds
and could be equivalent to or slightly older than Lau Backar 1. Wedekind (1927, pi. 29, figs 8-20)
compared three specimens from Lau Backar to this species but the internal structures of these are
still unknown.

Eichwald (1861) described but did not illustrate specimens from Estonia that he referred to E.
articulation. In their revision of Eichwald’s collection, Fedorowski and Gorianov (1973) listed
twenty-two specimens as E. articulation (Wahlenberg) and figured two from the Paadla Horizon
(Gorstian to early Ludfordian), one from the Kuressare Horizon (late Ludfordian) and one from
the Kaugatuma Horizon (PridoK), all from the Island of Saaremaa. From their description and
illustrations (Fedorowski and Gorianov 1973, pi. 4, figs 1^4) these Eichwald specimens appear to
compare more closely with E. fasciculatum in the smaller corallite size, smaller average number of
septa, the development of a small open axial area, and the more common development of
lonsdaleoid dissepiments that are larger and more irregular in size. The two figured longitudinal
sections (Federowski and Gorianov, pi. 4, figs 3 h and 4b) are not median and it is difficult to
compare their tabularia with those of the Gotland material.

Entelophyllum cf. fasciculatum Wedekind, 1927
Text-fig. 4a-f

Material studied. Material given to us at the Geological Institute in Kiev collected from the lower part of the
Malinovtsy Horizon, Grate Sloboda, Dnister River, Podolia (four specimens; figured: OU 10668-10671).

Distribution. Podolia: Gorstian Stage.

Description. Cylindrical corallites 6-8 mm in diameter and up to 25 mm in length. Two show peripheral
parricidal increase with four and five offsets arising from the calical platform of the parent (Text-fig. 4e).

Septa are radially arranged with 22-24 in each order. Major septa are irregularly withdrawn from the axis
leaving a distinct axial space. In the tabularium septa are thin, smooth and typically straight. They thicken in
the dissepimentarium where they may be zigzag and carry rare carinae. Minor septa are typically O6-0-8 as
long as the major but may be less than 0-3 or may be discontinuous being interrupted by larger interseptal
dissepiments.

The tabularium is half corallite diameter. It consists of either complete tabulae that are strongly arched with
flattened axial region and upturned peripheral edges, or more commonly closely spaced flat axial tabellae
downturned sharply at their margins to rest on the ones below or supplemented by globose tabellae, with the
space between the downturned edges and the dissepimentarium spanned by flat or concave tabellae.

Dissepiments are globose, vary from small to large, and are not arranged in vertical rows (Text-fig. 4a). In
longitudinal section, offsets appear to be initiated in the periaxial trough of the tabularium and then expand
with the dissepimentarium of the parent being reduced and the tabularium of the offset expanding outwards
(Text-fig. 4a).

Remarks. Material from Podolia resembles E. fasciculatum from Gotland in having relatively simple

JELL AND SUTHERLAND: SILURIAN RUGOSE CORALS

789

text-fig. 4. a-f, Entelophyllum ci.fasciculatum Wedekind, lower part of Malinovtsy Horizon, Grate Sloboda,
Dnister River, Podolia; a and b, OU 10668; a, longitudinal section showing offset in top left corner, x4; b,
transverse section, x4; c, OU 10669, transverse section, x4; d, OU 10670, transverse section x4; E and F,
OU 10671; E, lateral view showing peripheral parricidal increase, x2; f, transverse section, x 4. G and h,
Entelophyllum sp. A., Hamra Beds, Kattelviken 3, coast below the road at Kattelviken, 3-5 km N of Hoburgen,
southern tip of Gotland; G, transverse section abutting a tryplasmid corallite, x 4; h, longitudinal section, x 4.
i-l, Entelophyllum proliferum (Dybowski), EGM Col 352 (lectotype), Slite Beds, Stora Karlso, Gotland; i,
longitudinal section showing offset, x4; j, reproduction of Dybowski (1874, pi. 3, fig. 2a), line drawing of
longitudinal section, x2; k, transverse section, x4;l, reproduction of Dybowski (1874, pi. 3, fig. 2), showing

corallum, x 2.

septa withdrawn from the axis, a wide tabularium with a moderately developed periaxial trough,
and irregular sized dissepiments. It differs in that the corallites are slightly narrower, more
cylindrical, and the offsets do not diverge outwards from the lip of the parent calice as abruptly as
in E. fasciculatum. Our material closely resembles material from Podolia in Dr V. A. Sytova’s
collection, which we examined in Leningrad, which she had tentatively referred to E. proliferum
(Dybowski). Her material showed slightly more variation with one corallite having 30 major septa,
and the septa slightly more thickened. The material from Podolia is certainly like the holotype of

790

PALAEONTOLOGY, VOLUME 33

E. proliferum (Text-figs 4i-l) but the latter species is known only from the one specimen from the
Slite Beds of Stora Karlso, Gotland and its range of variation is unknown. E. proliferum may be
a senior synonym of E. fasciculatum.

Material from Podolia is similar to Petrozium losseniformis Zheltonogova (1965, p. 41, pi. 8, fig.
2) from the Gorstian Chagyr Formation in northwest Altaya, central Asia. We examined six
topotypes lent to Dr V. A. Sytova by Dr V. A. Zheltonogova and these are referable to
Entelophvllum and may well be conspecific with the European E. fasciculatum.

Entelophyllum sp. A

Text-fig. 4g-h

Material studied. RM Cn66046 from the Hamra Beds at Kattelviken 3, southern tip of Gotland.
Distribution. Gotland: late Ludfordian Stage.

Description. Small ceratoid coralhte 20 mm tall and 1 1 mm in diameter immediately below the calice. With the
proximal tip broken off, it does not indicate whether it is colonial or not.

Septa are radially arranged in two orders with 22 in each. Major septa are slightly withdrawn leaving an axial
space 2 mm by 1 mm. Minor septa are one-half to three-fifths the length of the major. The septa of both orders
are variable in the dissepimentarium, from slightly flexuose to straight. They are thickened and moderately
cardinate in the cardinal quadrants only.

The longitudinal section is not median but indicates that the tabularium is a third to a half the diameter of
the corallite wide and has a centrally raised area and probably a periaxial trough. The tabellae are globose and
moderately spaced. The dissepiments are globose, up to 1 mm across, and not arranged in regular series.

Remarks. In growth form, size, septal number, open axial space, and longitudinal section, it resembles E.
fasciculatum Wedekind, 1927, from the Eke Beds. It differs only in that the septa in the cardinal quadrants are
thicker and more carinate than in E. fasciculatum.

Entelophyllum proliferum (Dybowski, 1874)

Text-fig. 4i-l

1874 Cyathophyllum proliferum Dybowski, p. 445, pi. 3, fig. 2, 2a , 2b.
non 1927 Entelophyllum proliferum (Dybowski); Wedekind, p. 23, pi. 29, figs 21-29, 33.

Lectotype. Chosen Wedekind (1927, p. 23), original of Dybowski, 1874, pi. 3, fig. 2 and 2 a [non 2b]\ EGM
Col352, Slite Beds, Stora Karlso, Gotland; late Sheinwoodian or early Homerian.

Material. Two specimens (Dybowski 1874, pi. 3, fig. 2, 2a, 2b) from Stora Karlso and that of fig. 2b are only
questionably assigned to this species as thin sections are not available.

Distribution. Gotland : late Sheinwoodian or early Elomerian.

Diagnosis. Small bushy fasciculate corallum, increase peripheral parricidal; corallites up to 10 mm
in diameter, major septa smooth, straight, number 20-22, leave an open axial space; minor septa
weak; tabularium not distinctly biserial and periaxial trough not well developed; dissepiments
variable in size, not regularly arranged.

Description. The lectotype is a small bushy corallum with a lower ceratoid to subcylindrical corallite producing
five offsets one of which has a small bud in its calice. Budding in this mode is peripheral parricidal. The
longitudinal thin section (Text-fig. 4i) suggests that the tabularium of the offset commenced as an extension of
the periaxial trough of the parent and the neo-wall or part of it was inserted within the tabularium of the
parent. The corallites vary from 8—9-5 mm in diameter before budding and the early corallite is 1-7 cm high and
those of the second generation 1 -2—1 -3 cm. The epitheca is thin and shows very fine growth lines on slightly
expanded growth bands. The second specimen figured by Dybowski is a tall tapering corallite 33 mm tall and

JELL AND SUTHERLAND: SILURIAN RUGOSE CORALS

791

8 mm in diameter at the calice with four offsets developed in the calice. Dybowski’s drawing shows concentric
growth constrictions and low longitudinal ridges and furrows.

The septa are radially arranged, straight, thin and smooth. There are 20-22 major septa that extend three-
quarters of the way to the axis leaving an open axial space almost 2 mm wide. Minor septa are weakly
developed or discontinuous within the dissepimentarium.

The longitudinal section shows a well developed tabularium 0-6 the corallite diameter. In places it is not well
delineated from the dissepimentarium. It consists of broad tabulae and supporting tabellae spanning the
tabularium forming a raised flat axial area strongly downturned 1 mm from the tabularium boundary and then
outturned to meet the dissepimentarium almost horizontally. There are twelve tabulae in 5 mm vertically.
Dissepiments are highly variable in size and are irregularly arranged. They are globose with some three times
the size of others. Some are horizontally based while others slope downward and inward.

Remarks. Dybowski (1874) gave the locality of his two specimens as Stora Karlso on which is
exposed the Slite Beds. In searching previous collections and in collecting from Stora Karlso and
other outcrops of the Slite Beds on Gotland, we have been unable to obtain further material. The
possibility must be considered that Dybowski’s specimens may not have come from Stora Karlso.
Of the Gotland species Dybowski's specimens of E. proliferum most closely resemble E. fasciculatum
from the much higher upper Hemse and lower Eke Beds. Both forms have a wide axial area depleted
of septa, a generally similar tabularium with a poorly defined periaxial trough, and a peripheral
zone of irregularly distributed medium and large dissepiments. In the longitudinal section of E.
proliferum the tabularium of the otfset appears to originate as an extension of the periaxial trough
of the parent (Text-fig. 4i), while in E. fasciculatum offsets originate from the dissepimental zone
only. E. proliferum differs from E. articulation , which also occurs in the Slite Beds, in having larger
more irregular dissepiments and a tabularium that is much less well defined into two series.

Description and interpretation of this species are based on Dybowski’s description and drawings,
and photographs of the longitudinal section figured by Dybowski (1874, pi. 3, fig. 2a) and a
transverse section on the same slide, kindly sent to us by Dr D. Kaljo. Wedekind (1927, p. 23)
designated Dybowski’s figures 2 and 2 a as the lectotype, regarding the thin section 2a as being from
part of 2. The thin longitudinal section is on the same slide as the transverse section but it is not
known which specimen the transverse is from (D. Kaljo pers. comm. 1972). If the thin sections are
from different specimens and are different from that figured as figure 2 by Dybowski, we select the
longitudinal section as the primary type.

Entelophyllum dendroides sp. nov.

Text-fig. 5a-f

Holotype. RM Cn2560. Roda Lagret, Visby coast, Gotland; Telychian.

Material studied. Holotype (RM Cn2560, including two thin sections), and a paratype of four thin sections
(RM Cn55554-Cn55557); both are given as from the Roda Lagret (probably from pebbles on the beach),
Visby, west coast of Gotland (both figured).

Distribution. Gotland: Telychian.

Diagnosis. Dendroid colonies of closely-spaced small corallites 4-7 mm in diameter; septa typically
straight, noncarinate; major septa number 20-23; 1 mm axial space into which some septa may
extend; tabularium biserial with prominent periaxial trough; dissepiments small, globose; increase
peripheral parricidal.

Description. The holotype is part of a dendroid colony and measures 30 x 40 mm in cross-section at its widest
part and is 75 mm long. It is composed of small subcylindrical corallites 4-7 mm in diameter, typically
separated by less than 5 mm. The corallites arise mainly in groups of four from the dissepimentarium of the
parent by parricidal increase. They diverge rapidly outward soon after separation from each other and expand
in diameter. They then tend to turn vertically and produce lateral extensions that abut adjacent corallites.

792

PALAEONTOLOGY, VOLUME 33

text-fig. 5. Ente/opliyllum dcndroides sp. nov.. Roda Lagrct. from pebbles on the beach at Visby, Gotland.
a-d, RM Cn2560 (holotype); a-c, transverse sections, x 2, x4, x 4 respectively ; d, longitudinal section, x4.
e, RM Cn55555, longitudinal section, x 4. f, RM C55556, transverse section, x4.

Budding across the colony appears irregular; one corallite is 30 mm tall before budding, another is in excess
of 60 mm. Lateral expansions are not produced at the same level across the corallum.

The figured transverse section (Text-fig. 5f) shows the initial stages of increase where the neo-walls form as
four scallops at the inner dissepimentarium, slightly expanding into the tabularium. A dissepiment is based on
the neo-wall, and neo-septa project inwards from the dissepiment towards the atavo-septa. In the longitudinal
section (text-fig. 5e) there is a suggestion of one offset being produced by nonparricidal increase but this cannot
be definitely established.

Septa are typically long, radially arranged in two orders. The major septa number 20-23 in mature corallites
and leave a narrow axial space (Text-fig. 5b and c). The septal ends are typically irregularly bunched near the
axis in groups of two to three but two or more septa may extend irregularly to the axis. In immature corallites
many of the major septa are confluent at the axis. Minor septa are half major septa in length and variously
developed. They may be straight, slightly thinner than the major septa, projecting radially a short distance into
the tabularium, and discontinuous.

Longitudinal sections are typically entelophylloid with the tabularium and dissepimentarium distinctly
delineated and the tabularium separated into an axial and a periaxial series of tabellae (Text-fig. 5d). The
tabularium is typically 0-6 the diameter of the corallite, and the axial series of tabellae 0-75 to 0 8 the diameter
of the tabularium. The axial series consists of flat-topped tabellae that have steeply downturned edges and

JELL AND SUTHERLAND: SILURIAN RUGOSE CORALS

793

text-fig. 6. Entelophyllum pseudodianthus (Weissermel), Much Wenlock Limestone, Wenlock Edge,
Shropshire; GSM 6575 (neolectotype). A, longitudinal section, x 4. b, transverse section, x 4.

number 10-15 in 5 mm vertically. The periaxial series are flat or slightly concave upwards producing a
pronounced periaxial trough with 20-22 tabellae in 5 mm vertically. The dissepiments are small, globose or
slightly elongated and sloping downward and inward.

Remarks. This species is distinguished by its relatively small sized corallites and dendroid habit. It
differs from the Llandovery Petrozium dewari Smith, 1930 and P. losseni (Dybowski 1874) by the
more marked separation of the tabularium into axial and periaxial series and wider dissepi-
mentarium.

Entelophyllum pseudodianthus (Weissermel, 1894)

Text-figs 6a, b and 7a-e

vl839 Cyathophyllum dianthus Goldfuss; Lonsdale, p. 690, pi. 16, figs 12 b, 12 c, 12 d[non Cyathophyllum
dianthus Goldfuss, 1826].

1894 Cyathophyllum ( Heliophyllum ) pseudodianthus Weissermel, p. 591, pi. 47, figs 2 and 3.
vl927 Xylodes pseudodianthus (Weissermel); Lang and Smith, p. 475, pi. 35, fig. 9.
v!929 Xylodes pseudodianthus (Weissermel); Smith and Tremberth, p. 366, pi. 8, figs 3 and 4.

794

PALAEONTOLOGY, VOLUME 33

Neolectotype. Here chosen, original of Lonsdale, 1839, pi. 26, fig. 126, c, d\ GSM 6575 and thin section PF 4617
from the Geological Society Collection, Much Wenlock Limestone, Wenlock Edge, Shropshire, England;
Homerian.

Type material. Weissermel (1894, p. 591) described Cyathophyllum ( Heliophyllum ) pseudodianthus as a new
name for a species represented by the four specimens which had been figured as Cyathophyllum dianthus by
Lonsdale (1839, pi. 16, fig. 12, 12 a-e). Lang and Smith (1927, p. 473) stated that one of Lonsdale’s four
specimens (fig. I2e) is lost and that the other three represent three distinct species, each belonging to a different
genus. They state that only Lonsdale’s figures 126-r/(all from the one specimen: GSM 6575) represents the
species described by Weissermel as C. ( H .) pseudodianthus and selected as lectotype the specimen figured by
Weissermel (1894, pi. 47, fig. 3) from a boulder in the glacial drift of Lauth, Germany. That specimen, in the
East Prussian Provincial Museum at Konigsbcrg (now Kaliningrad), was never sectioned and was destroyed
during the Second World War (Dr V. A. Sytova pers. comm.).

We have thus selected one of Weissermel’s original syntypes as the neolectotype, the specimen figured by
Lonsdale (1839, p. 16, fig. 12 b-d)\ GSM 6575, and herein refigured (Text-figs 6a, b and 7e).

Material studied. Neolectotype (GSM 6575) and twenty-four specimens from the collections of the GSM,
BMNH, UO, SM, UB, and UQ from the Much Wenlock Limestone, Wenlock Edge, Shropshire; figured:
GSM 6575 (neolectotype), BMNH R2022. Farley Member, Coalbrookdale Formation; Much Wenlock
Limestone of Dudley, Worcestershire, including Wren’s Nest (numerous specimens from same collections as
the Wenlock Edge material above; figured: UB 124); Much Wenlock Limestone of May Hill, Gloucestershire
(figured BU 8715); and lower Elton Beds, from Ledbury Quarry, Malverns, Herefordshire.

Distribution. England: Homerian to early Gorstian; Estonia: Ludlow to ?PridoK; Poland: ?early Pridoli.

Diagnosis. Typically small bushy coralla of turbinate to rapidly expanding subcylindrical corallites
up to 25 mm in diameter; increase peripheral parricidal with up to 10 offsets. Septa 20-33 in each
order, long, thickened in dissepimentarium and heavily carinate to aerolate; major septa slightly
withdrawn from axis; minor septa rarely contratigent. Tabularium biserial with periaxial trough;
dissepiments globose, horizontally based at periphery and steeply inclined axially.

Description. The coralla are small and bushy, up to 200 mm in diameter and 150-200 mm in height, consisting
of two to five generations of corallites (Text-fig. 7d). The coralla are fasciculate, generally formed of rapidly
expanding corallites 60 mm in height and 15-20 mm in diameter but in some coralla they may be cylindrical
and up to 100 mm in height. The neolectotype is a small bushy clump 20 mm in diameter and 40 mm high,
consisting of 8-10 corallites (Text-fig. 7e). Budding is peripheral, parricidal with up to ten corallites originating
at the extreme margin of the calyx. Corallites rapidly expand and quickly diverge giving the corallum a
radiating appearance (Text-fig. 7d); they have strong contractions and expansions but few connecting
processes. Corallite walls are thin and epithecate.

Major septa number 21-33 (mean 27 at diameter 12 mm) in adult stages (Text-fig. 6b). Some major septa
extend to the axis while others are withdrawn leaving an axial space up to 3 mm in diameter. Axial ends are
commonly turned aside or twisted. Minor septa typically do not project into the tabularium, are commonly
0-6-0-7 the length of the major, and are only rarely contratigent. In the dissepimentarium both orders of septa
are thickened, heavily carinate and commonly aerolate, but in parts of some corallites they are relatively
smooth. Carinae are zigzag and parallel the trabeculae. In the outer dissepimentarium they diverge at a small
angle to the vertical and are fanned over the inner dissepimentarium so that at the dissepimentarium/
tabularium boundary they are 30° to the horizontal.

In longitudinal section the dissepimentarium and tabularium are distinctly delineated. The tabularium is
6-7 mm wide and differentiated into two series of tabellae. In forms with septa extending to the axis the axial
series are domed and relatively high (Text-fig. 6a), while in those with the septa withdrawn from the axis they
are flat or sagging and of low profile (Text-fig. 7b). A narrow series of small, flat or slightly saucered tabellae
span the area between the axial series and the dissepimentarium forming a periaxial trough. Dissepiments are
small, globose, occurring in almost horizontal rows at the periphery and steeply inclined inwards and
downwards at the tabularium boundary. In some specimens broad zones of more globose dissepiments
alternate vertically with thinner zones of smaller and slightly thicker dissepiments (Text-fig. 7b). The former
correspond to expansions of the corallite diameter and the latter to the contractions.

JELL AND SUTHERLAND: SILURIAN RUGOSE CORALS

795

text-fig. 7. Entelophyllum pseudo dicmthus (Weissermel). A and B, BU 8715, Much Wenlock Limestone, May
Hill, Gloucestershire; a, transverse section showing typical carinate and thickened septa, x4. c and d. Much
Wenlock (Dudley) Limestone, Dudley, Worcestershire; c, UB 124, calical view showing peripheral parricidal
increase, x I; d, UO C 1 75 1 6, oblique view showing typical loose fasciculate corallum, x 1. e, GSM 6575
(neolectotype). Much Wenlock Limestone, Wenlock Edge, Shropshire, polished surface, x 2.

Remarks. In England, Entelophyllum articulatum anglicum and E. pseudo diant hus have the same
distribution and commonly occur together. We have included those forms with their septa typically
thickened, highly carinate and in some cases aerolate in E. pseudodianthus, as did Smith and
Tremberth (1929). These specimens generally have squatter coralla, more turbinate corallites,
shorter minor septa that are rarely contratigent, and less regular biserial tabularia. However, some
have corallites that are slender and cylindrical, minor septa that project into the tabularium and are
contratigent, and have regular biserial tabularia. On the other hand corallites in some colonies

796

PALAEONTOLOGY, VOLUME 33

referred to E. articulation anglicum are turbinate or have some thickened and carinate septa. Thus
some specimens seem to be gradational between the two forms and their specific designation is
arbitrary.

Entelophyllum pseudodiant hus transiens (Prantl, 1940)

Plate 6, figs 5-7

1940 Xy lodes pseudo dianthus transiens Prantl, p. 13, pi. 1, fig. 4; pi. 3, figs 1, 2, 4.

Holotype. Prantl, 1940, pi. 1, fig. 4, from the Kopanina Formation, ‘ Amerika' quarries near Morina (Budnany
of Prantl 1940), Czechoslovakia; Gorstian.

Material. Interpretation based on photographs of thin sections figured by Prantl (1940, pi. 3, figs 1, 2, 4).
Distribution. Bohemia: ?late Homerian to ?Gorstian.

Diagnosis. Small bushy coralla with ceratoid to subcylindrical corallites 11-17 mm in diameter;
increase peripheral, parricidal. Major septa 31-34, withdrawn up to 1 mm from axis, cardinal
septum projecting into axial space; septa thin in tabularium; continuous, some thickened and
carinate in dissepimentarium. Tabularium distinctly biserial with axial tabellae flat topped;
dissepiments small, globose, inclined axially.

Remarks. Prantl (1940) proposed this species for forms that show both long thin smooth septa
typical of E. articulation and thickened carinate septa typical of E. pseudodianthus . Considering the
variation in E. articulation anglicum and E. pseudodianthus described previously, the Bohemian form
could well belong to this series. The lack of heavily carinate and cavernous septa suggest similarities
to E. articulation anglicum rather than E. pseudodianthus. In size and septal number, it is more like
specimens of the subspecies from the Elton Beds of similar age.

Entelophyllum prosperum (Pocta, 1902)

Plate 6, figs 1-4

1902 Cyathophyllum prosperum Pocta, p. 105, pi. 43, figs 1, 2, 1019, 23-29, 36-41 ; pi. 44, figs 1-33;

pi. 45, figs 1-5, 18-39; pi. 46, figs 8-24; pi. 103, figs 6-8; pi. 109, fig. 8.

1902 Cyathophyllum minusculum Pocta, p. 104, pi. 42, figs 1-8.

1940 Xylodes prosperus prosperus (Pocta); Prantl, p. 8, pi. 1, figs 5 and 6; pi. 2, figs 2, 5, 7.

Lectotype. Chosen Prantl (1940, p. 8), original of Pocta, 1902, pi. 44, figs 1-4 from Budnaner Kalksteine,
Kopanina Formation at Tachlovice, Czechoslovakia; Gorstian.

EXPLANATION OF PLATE 6

Figs 1-4. Entelophyllum prosperum (Pocta), Kopanina Formation, Tachlovice, Czechoslovakia. 1, NM No. 768,
original of Pocta (1902, pi. 103, fig. 7), transverse section, x 2. 2, NM unnumbered, original of Prantl (1940,
pi. 2, fig. 2), transverse section x 2. 3, NM No. 855, original of Pocta (1902, pi. 103, fig. 8), longitudinal
section, x 2. 4, NM No. 473, original of Pocta (1903, pi. 103, fig. 6), longitudinal section, x 2.

Figs 5-7. Entelophyllum pseudodianthus transiens (Prantl), Kopanina Formation, ‘Amerika’ quarries, near
Morina, Czechoslovakia. 5, NM 26329, original of Prantl (1940, pi. 3, fig. 2), transverse section, x 2. 6, NM
26329, original of Prantl (1940, pi. 3, fig. 4), transverse section, x 2. 7, NM 26329, original of Prantl (1940,
pi. 3, fig. 1), longitudinal section, x 2.

Figs 8 10. Entelophyllum confusion (Pocta), Upper Liten Group [Motol Formation], ‘V Kozle' between
Beroun and Srbsko, Czechoslovakia. 8 and 9, NM unnumbered, transverse sections, x 7-5, x 4, respectively.
10, NM unnumbered, longitudinal section, x4.

PLATE 6

JELL and SUTHERLAND, Entelophyllum

798

PALAEONTOLOGY, VOLUME 33

Material. Interpretation based on photographs of thin sections figured by Pocta (1902, pi. 103, figs 6—8) and
Prantl (1940, pi. 2, fig. 2); all from Kopanina Formation at Tachlovice.

Distribution. Bohemia: Gorstian.

Diagnosis. Small bushy coralla with trochoid to subcylindrical coralhtes 18-19 mm, maximum
35 mm, in diameter; increase peripheral parricidal. Major septa 28-40, withdrawn from axis in the
cardinal quadrants leaving axial space elongate in counter-cardinal plane; minor septa 06 the
length of major; septa considerably thickened, heavily carinate to aerolate, continuous in
dissepimentarium. Tabularium biserial; axial series varying from small tabellae forming high domes
to large complete tabellae slightly sagged axially; dissepiments small, globose, horizontally based at
periphery and steeply inclined axially.

Remarks. Prantl (1940) in redescribing the entelophylloid corals in the collection of Pocta (1902),
recognized two subspecies for the material described as Cyathophyllum prosperum and the
synonymy given above is that of the nominate subspecies. E. prosperum is characterized by
thickened and heavily carinate septa with the outer dissepimentarium appearing reticulate in
transverse section and resembling E. pseudo dianthus from Britain. These two species are very
similar, the Bohemian species only differing in its slightly greater size and more numerous septa, and
the more noticeable fossula. Offsets arising right at the margin of the parent in the specimen figured
by Prantl ( 1940, pi. 1, fig. 6) are identical to that figured herein for E. pseudodianthus (Text-fig. 7c).
The Bohemian form might best be considered a subspecies of E. pseudodianthus. However, not
enough material is available to us and the type locality ‘Tachlovice' is today unknown but thought
to have been in a mine (A. Galle, pers. comm.).

Entelophyllum prosperum crassum (Prantl, 1940)

1902 Cyathophyllum prosperum Pocta, p. 105, pi. 43, figs 3-9, 20-22, 30-35; pi. 44, figs 34-40; pi. 45,
figs 40-42; pi. 46, figs 1-7.

1940 Xylodes prosperus crassus Prantl, p. 11, pi. 1, figs 7-9; pi. 2, fig. 6; pi. 3, fig. 3.

Holotype. Pocta, 1902, pi. 44, figs 33-34, from Budnaner Kalksteine, Kopanina Formation at Tachlovice,
Czechoslovakia; Gorstian.

Material. No material was available for this study.

Distribution. Bohemia: Gorstian.

Remarks. This subspecies was described by Prantl (1940) for the larger specimens of E. prosperum
from the same strata. The originals of Prantl ( 1940, pi. 2, fig. 6; pi. 3, fig. 3) have not been located.

Entelophyllum confusum (Pocta, 1902)

Plate 6, figs 8 10

1902 Cyathophyllum confusion Pocta, p. 103, pi. 99, figs 3-11.

1940 Xylodes confusus (Pocta); Prantl, p. 16, pi. 3, figs 5 and 6.

1981 Carinophyllum confusion (Pocta); Hill, fig. 127, 2 d, e.

Lectotype. Chosen Prantl (1940, p. 16), original of Pocta, 1902, pi. 99, fig. 3, from the Budnaner Kalksteine,
Kopanina Formation of Tachlovice, Czechoslovakia; Gorstian.

Material. Interpretation based on photographs of the thin sections figured by Prantl ( 1940, pi. 3, figs 5 and 6)
from the upper Liten Group (Motol Formation), ‘V Kozle’ between Beroun and Srbsko (Homerian).
Fedorowski and Gorianov (1973) referred three specimens of the Eichwald collection from Estonia to this

JELL AND SUTHERLAND: SILURIAN RUGOSE CORALS

799

species, one from the Gorstian-Ludfordian Paadla Horizon of Pilguse and two from the late Ludfordian
Kurcssaare Horizon of Lode.

Distribution. Bohemia: Homerian and Gorstian; Estonia: Ludlow.

Diagnosis. Small dendroid to phaceloid coralla; corallites cylindrical, 8-14 mm in diameter;
increase peripheral parricidal. Septa fusiform, carinate in dissepimentarium ; 22-24 major septa
extend almost to axis leaving small axial space; cardinal septum projects slightly into axial space.
Tabularium biserial; axial series high, flat topped; dissepimental zone broadly arched; dissepiments
small, globose.

Remarks. This species is distinguished by the fusiform and carinate nature of septa, characters
Strelnikov (1964, p. 59) used to typify his new genus Carinophyllum with E. confusion as type species.
These features and the corallite size seem to distinguish the Bohemian material as a distinct species
but we do not consider the carination and thickening to be of generic significance.

Entelophyllum lauense sp. nov.

Plate 7, figs 1-12

Holotype. RM Cn66047, Rhizophyllum limestone, lower Eke Beds, Lau Backar I, southeast Gotland; middle
Ludfordian.

Material studied. E. lauense is known only from the type locality (ten specimens; figured: RM Cn66047
(holotype); UQ F35165, 35169, 35170, 35281 (paratypes)).

Distribution. Gotland: middle Ludfordian.

Diagnosis. Presumably fasciculate, corallites flaring after early trochoid stage, up to 25 mm in
diameter and 20 mm high; major septa 30-32, cardinal fossula confluent with axial space; minor
septa 0-6-0-7 length of major; septa in dissepimentarium thickened axially and heavily carinate to
retiform peripherally. Tabularium biserial with axial tabellae domed typically with axial sag;
dissepiments small globose; broad dissepimentarial zone peripherally domed.

Description. Broken isolated corallites are trochoid initially but flare rapidly producing wide, everted calices
14-25 mm in diameter, while the corallites are less than 20 mm high. The dissepimental zone forms a broad
calical platform that may show several rejuvenescenes. They surround moderately deep tabularial pits that
have a shallow trough about the outside and a central depression which may be connected to the trough by
a depressed cardinal fossula (PI. 7, fig. 3). Early stages of several specimens show either broken early tips or
crescent-shaped broken walls that are interpreted as fragments from a parent corallite. Some corallites show
a supporting buttress (PI. 7, fig. 4). One specimen, UQ F35165, shows two generations of budding with one
offset arising at a rejuvenescene by peripheral non-parricidal budding and the parent has two small offsets
developing in the dissepimentarium 8 mm higher. The holotype has at least three offsets beginning by the
development of scalloped neo-walls near the inner margin of the dissepimentarium (PI. 7, fig. 1).

There are 30-32 major septa arranged pinnately about a long cardinal septum which bisects a distinct fossula
(PI. 7, fig. 11). They are slightly withdrawn abaxially leaving a narrow axial space 1-2 mm in diameter, into
which the cardinal fossula opens. The septa in the tabularium are thin, straight or slightly flexuose at their axial
extremities. The minor septa are well developed only slightly thinner than the major and 0-6-0-7 their length,
projecting a short distance into the tabularium. In the dissepimentarium both orders are variably thickened,
zigzag, irregularly carinate to retiform. The thickening is more pronounced in the inner dissepimentarium and
the carination increases toward the periphery. Carinae typically alternate, arising from the angles of the zigzag
septa but in places they are so numerous and the segments of the septa so short and turned away from the plane
of the septum that the carinae appear opposite each other.

The tabularium expands rapidly with growth and is 0-6-07 the diameter of the corallite immediately below
the calice. The tabularial floors are domed centrally with a slight axial depression, and downturned in the outer
tabularium producing a periaxial trough (PI. 7, fig. 2). The axial tabellae are relatively large and globose, and
their downturned edges rest on tabellae below or outturn to the dissepimentarium. Periaxial tabellae are flat
or concave upwards. Dissepiments are small, globose, peripherally domed, and not arranged in vertical series.

800

PALAEONTOLOGY, VOLUME 33

Septa appear to be composed of a uniserial row of monacanthine trabeculae 0 01-0-02 mm in diameter.
Extension of the fibres of the trabeculae form the carinae. Both the trabeculae and carinae show a fanning over
the dissepimentarial floor, becoming flatter near the dissepimentarium/tabularium boundary and then
becoming steep again axially.

Remarks. All specimens are isolated but they are interpreted as being broken offsets from loosely
fasciculate colonies. The colonial nature is also suggested by UQ F35165 which shows two
generations of budding.

E. lauense differs from other described Gotland species of Entelophyllum by its corail um form, flat
flaring nature of the corallites, and heavy carination and thickening of the septa. In the latter regard,
it resembles E. pseudo dianthus but differs from it in that the corallites are much flatter and have well
developed fossulae.

There is some reservation in referring E. lauense to Entelophyllum because it is not clearly as
fasciculate, the cardinal septum and fossulae are more pronounced, and the septa more heavily
carinate than in the type species. However, as discussed above, it appears to be compound with at
least two generations of buds. The development of the cardinal septum is variable in other species
of Entelophyllum and this is not regarded as a generic character. The carination and retiform nature
of the septa are comparable with that of E. pseudo dianthus and E. prosperum which we included in
Entelophyllum rather than separating out the carinate forms as Stereoxylodes Wang, 1944. There are
similarities to Nanshanophyllum typicum Yu, 1956, a solitary form, heavily carinate in the
dissepimentarium, showing a cardinal fossula opening into an axial space. Pedder (1976) considered
Nanshanophyllum Yu to be a subgenus of Stereoxylodes.

Entelophyllum hamraense sp. nov.

Plate 8, figs 1-9

Holotype. RM Cn66048; Hamra Beds, Nars fyr 1, southeast Gotland; late Ludfordian.

Material studied. Hamra Beds of Gotland: the type locality (one specimen; figured: RM Cn66048 - holotype);
Kiittelviken 3, southern tip of Gotland (one specimen; figured: UO 10672); Hoburgen 2, southern tip of
Gotland (six specimens; figured: UQ F34001); ?Hanira or Lower Sundre Beds at Juves 3, southern tip of
Gotland (one specimen; figured: RM Cn66049).

Distribution. Gotland: late Ludfordian.

Diagnosis. Fasciculate corallites trochoid to ceratoid, maximum diameter 25 mm, height 40 mm,
peripheral parricidal increase. Major septa number 27-31, withdrawn from axis, axial ends slightly
thickened, septal segments and long cardinal septum common in axial space; minor septa
commonly contratigent ; septa heavily carinate and thickened in inner dissepimentarium ; tabularium
regularly biserial; dissepimentarium broad, dissepiments small, globose.

EXPLANATION OF PLATE 7

Figs 1-12. Entelophyllum lauense sp. nov., Rhizophyllum limestone. Eke Beds, Lau Backar 1,1-0 km NE of Lau,
southeast Gotland. 1 and 2, RM Cn66047 (holotype); I, calical view showing incipient buds, x2;
2, longitudinal section, x 4. 3 and 4, UQ F35 1 70 ; 3, calical view, x 2 ; 4, lateral view, x 2. 5-8, UQ F3528 1 ;
5, lateral view, x2; 6, transverse section from base, x4; 7, transverse section from central part, x4;
8, transverse section of calice, x4. 9 and 10, UQ F35165; 9, transverse section, x4; 10, longitudinal section
showing edge of second corallite, x 4. 1 1 and 12, UQ F35169; 11, transverse section x4; 12, longitudinal
section, x 4.

PLATE 7

JELL and SUTHERLAND, Entelophyllum

802

PALAEONTOLOGY, VOLUME 33

Description. Corallites are trochoid to ceratoid commonly occurring as isolated broken offsets up to 25 mm in
diameter and 40 mm in height. The holotype consists of three offsets, 18, 18 and 20 mm in diameter and 35 mm
in height, arising from the weathered edge of a calice. One of the offsets has at least three offsets in its calice.
Another specimen, UO 10672, shows three small trochoid corallites arising from the periphery of a corallite
1 7 mm in diameter. Increase is peripheral parricidal.

Major septa number 27-30 and are smooth and straight in the tabularium, commonly withdrawn from the
axis leaving an axial space 2-4 mm in diameter (PL 8, fig. 6). The cardinal septum and rarely the counter are
longer than the others and project well into the axial space. Axial ends of septa are commonly thickened and
discrete septal segments are common in the axial space based on the axial tabellae. Minor septa are 0-5-0-75
the length of the major and are commonly contratingent. In the dissepimentarium they are commonly as thick
as the major and both are dilated in the inner parts and strongly carinate in the outer parts (PI. 8, fig. 4).
Carinae on the thickened inner parts are short and stout, and in places appear to have short prickles coming
off them as though they were extensions of secondary trabeculae. In the outer parts the carinae are longer and
more numerous, coming off the angles of the zigzags of the septa. The long carinae give a rather ragged
appearance to, the septa and only rarely are they retiform or aerolate.

The tabularium is 0-4-0-5 the diameter of the corallite, domed centrally with the tabulae sloping steeply
downward peripherally and then turning outwards giving a shallow periaxial trough (PI. 8, figs 5 and 7). The
central area is composed of numerous globose tabellae. In places where the outer globose tabellae do not reach
the dissepimentarium, flat periaxial tabellae span the space. Dissepiments are small, globose, not arranged in
vertical series, and commonly flat peripherally but inclined axially downward near the tabularium.

Remarks. Specimens from Hoburgen are mainly isolated corallites but have similar internal
structures to the holotype. Some may be initial protocorallites while others may be offsets broken
from loosely fasciculate colonies.

E. hamraense differs from E. lauense from the Eke Beds, which also is heavily carinate, in that the
corallites are taller, no fossula is developed, the septa are thicker in the inner dissepimentarium, and
the septal ends commonly extend into the axial area. The occurrence of short prickles on the carinae
suggests that the trabeculae may bear secondary trabeculae, which is not an entelophylloid
character.

This species differs from E. pseudodianthus in that its axial space is not as open, the axial ends of
the septa are not grouped as in the latter species and there is not as regular a biserial tabularium.

Entelophyllum1 * * * 5 * * 8. visbyense Wedekind, 1927
Plate 9, figs 1-9

v*1927 Entelophyllum visbyense Wedekind, 1927, p. 24, pi. 7, figs 9 and 10.

v*1927 Entelophyllum anschutzi Wedekind, 1927, p. 24, pi. 7, figs 7 and 8.

Holotype. Wedekind, 1927, pi. 7, figs 9 and 10; RM Cn54873, Hogklint Beds, Gutevagen 3 [= Cement factory],
Visby, Gotland; Sheinwoodian.

Material studied. Four specimens from the type locality; figured: RM Cn54873 (holotype), SMF Wdkd.
10308-10309 (holotype of E. anschutzi ), RM Cn66196, Cn66197. Neuman and Hanken (1979) recorded this

EXPLANATION OF PLATE 8

Figs I -9. Entelophyllum hamraense sp. nov. 1 and 2, UQ F34001, upper Hamra Beds, Hoburgen 2, southern

tip of Gotland; 1, transverse section, x2; 2, longitudinal section, x2. 3-5, RM Cn66048 (holotype), lower

Hamra Beds, Nars fyr 1, coast at Nar, southeast Gotland ; 3 and 4, transverse sections, x 2, x 4, respectively;

5, longitudinal section, x 4. 6 and 7, OU 10672, Hamra Beds, Kattelviken 3, coast below road, 3-5 km N

of Hoburgen, southern tip of Gotland; 6, transverse section, x 4; 7, longitudinal section, x 4. 8 and 9, RM

Cn66049, Sundre Beds, Juves 3, western side of road at Juves, 2 km E of Hoburgen, southern tip of Gotland;

8, transverse section, x4; 9, longitudinal section, x 4.

PLATE 8

JELL and SUTHERLAND, Entelophyllum

804 PALAEONTOLOGY, VOLUME 33

species from the Upper Visby and lowermost Hogklint Beds in the Vattenfallsprofilen 1, near the ferry harbour
of Visby.

Distribution. Gotland: Sheinwoodian.

Diagnosis. Phaceloid, corallites 10-12 mm in diameter, closely spaced, increase parricidal; major
septa number 20-23, straight, thickened in inner tabularium, withdrawn from the axis leaving a
broad axial space; minor septa very short; tabularium broad, axial tabulae flat-topped or sagging,
weakly developed axial trough; dissepimentarium irregularly narrow, in one to four vertical series.

Description. The hololype is of a subcylindrical corallite 1 1 mm in diameter. Other specimens are fragments of
large phaceloid coralla consisting of closely packed cylindrical corallites 10-12 mm in diameter, some are
crushed due to later compaction. Calices are moderately deep with inverted conical sides and a wide flat axial
floor. Increase is parricidal (PI. 9, figs 8 and 9) with up to 12 offsets originating in the one calice. The neo-wall
is initiated in the tabularium and it appears as though part of the tabularium of the parent is continuous with
that of the offset.

In transverse section major septa are prominent, thickened in the outer tabularium and taper axially and
some peripherally (PI. 9, fig. 4). They are 0-6-0-8 the radius of the corallite in length and number 20-23 in
mature corallites. In some the cardinal septum is shortened and may lie in a shallow fossula. The counter and
counterlateral septa may be slightly extended. In the dissepimentarium major septa become irregular in places
with the stereome that thickened them in the tabularium continuous onto the dissepiments. They may be
thickened into triangular bases at the wall. Minor septa are short, 0-2-03 the length of the major, much thinner
than the major, and in places interrupted by interseptal dissepiments.

The tabularium is 0-65-0-75 the diameter of the corallite in width, composed of tabulae that are flat or axially
sagged with the edges downturned initially, then turned outward to meet the dissepimentarium nearly
horizontally (PI. 9, fig. 7). There are 8- 1 5 tabulae in 5 mm vertically. Dissepiments are irregular in size and vary
from small globose forms to more elongate ones, arranged in one to four rows (PI. 9, fig. 8), three and rarely
four in the holotype (PI. 9, fig. 2).

Septa are recrystallized in the holotype but appear to be monacanthate, with the monacanths sloping in and
upward toward the axis at about 45° peripherally, and steepening axially to 55°-60°. There are approximately
3 trabeculae per 0-4 mm in vertical section.

Remarks. We consider E.l visbyense to belong to some genus other than Entelophyllum because of
the striking differences in character in both transverse and longitudinal section when compared to
E. articulation. The tabularium is not distinctly divided into two series with a well developed
periaxial trough, and the septa are much more thickened and more typically withdrawn from the
axis. The material at hand, however, is not considered adequate as a basis for defining a new genus.

The holotypes of E.l visbyense and E.l anschutzi, both from the same locality and described by
Wedekind ( 1 927), show only minor differences. In transverse section E. 1 anschutzi has thinner major
septa that are more withdrawn from the axis, leaving a broader axial space, and in longitudinal
section the tabularium sags compared with more flat tabulae in E.l visbyense. Sections through
many corallites of the one corallum show similar variations (PI. 9, figs 8 and 9). For these reasons,
we are treating them as conspecific as did Neuman and Hanken (1979, p. 89).

EXPLANATION OF PLATE 9

Figs 1-9. Entelophylluml visbyense Wedekind, Hogklint Beds, Gutevagen 3, T54 km SW of Visby Cathedral,
Gotland. 1 and 2, RM Cn54873 (holotype), original of Wedekind (1927, pi. 7, fig. 10); 1, transverse section,
x4; 2, longitudinal section, x 4. 3 and 4, from holotype of E. anschutzi Wedekind, (1927, pi. 7, figs 7 and
8); 3, SMF Wdkd. 10308, longitudinal section, x4; 4, SM Wdkd. 10309, transverse section, x4. 5 and 6,
RM Cn66196; 5, longitudinal section, x2; 6, transverse section, x 2. 7-9, RM Cn66197; 7 and 8,
longitudinal sections, x 4, x 2, respectively ; 9, transverse section showing parricidal increase, x 2.

PLATE 9

JELL and SUTEIERLAND, Ente/ophylluml

806

PALAEONTOLOGY, VOLUME 33

text-fig. 8. Entelophyllum ? dalecarlicum (Lindstrom), RM Cn57149 (holotype), Styggforsen Limestone,
Styggforsen, Siljan district, Dalecarlia, central Sweden, a, reproduction of Lindstrom (1880, pi. 2, tig. 8), x 3.
B, thin section from which Lindstrom’s figure was drawn, showing accuracy of the drawing, x 2. c,
reproduction of Lindstrom (1880, pi. 1, fig- 22), x 3.

Entelophyllum ? dalecarlicum (Lindstrom, 1880)

Text-fig. 8a-c

1880 Cyathophyllum dalecarlicum Lindstrom, p. 34, pi. 1, fig. 22; pi. 2, fig. 8.

Holotype. Lindstrom, 1880, pi. 2, fig. 8; RM Cn 57149, Styggforsen Limestone, Styggforsen, Siljan district,
Dalecarlia, central Sweden; Telychian.

Material studied. The description is based on Lindstrom’s (1880, pi. 1, fig. 22; pi. 2, fig. 8) illustrations of the
type specimen and on the longitudinal thin section of the holotype (RM Cn57149); his transverse section is
presumed lost. Dr B. Neuman (pers. comm.) has searched for more material in the type area without success.

Distribution. Central Sweden; Telychian.

Diagnosis. Fasiculate, possibly peripheral nonparricidal increase; corallites slender, less than 10 mm
in diameter containing 22-23 major septa; prominent axial space; minor septa weakly developed;
tabularium regularly arranged broad, flat-topped axial series and narrow, closely spaced, sagging
periaxial series of tabellae; dissepiments elongate in one or two rows.

Description. The specimen from which the slide was cut is not available. The longitudinal section has the central
corallite abutted on one side by the dissepimentarium of another and on the other side by the apparent margin
of an offset (text-fig. 8b). This presumed offset shows the early stage of what is possibly a narrow row of flat
tabellae beside the neo-wall, which is inserted on top of a dissepiment and which is not continuous with that
of the parent. If these relations are correctly interpreted then the budding is peripheral nonparricidal.
Immediately before the budding the corallite is 10 mm wide and at the top of the section it is 9-5 mm.

The transverse section is not available. Lindstrom (1880, pi. 1, fig. 22) shows 22 or 23 major septa that are
smooth, straight and extend 0-8 the corallite radius to the axis where their axial tips may be turned aside,
leaving an axial space. Minor septa are poorly developed as segments on the interseptal dissepiments between
some major septa.

In longitudinal section the tabularium is very regularly developed and is well differentiated from the
dissepimentarium. It shows axial series of broad flat tabellae with strongly downturned peripheral parts that
are based on the tabellae below. There are 14-18 in 5 mm vertically. Between this zone and the
dissepimentarium is a narrow periaxial series of slightly concave upward tabellae (17-19 per 5 mm) forming
an exceptionally regular periaxial trough (Text-fig. 8b). Dissepiments are very irregular in size and shape, some
twice as broad as others and varying from globose to elongate. The tabularia are 0-6-0-7 the corallite diameters
and the axial series is 0-8 the diameter of the tabularium.

JELL AND SUTHERLAND: SILURIAN RUGOSE CORALS

807

Remarks. E.l dalecarlicum cannot be adequately evaluated on the basis of the single longitudinal
section of the holotype available for study. The possible occurrence of nonparricidal budding
indicates a generic assignment other than Entelophyllum.

Genus petrozium Smith, 1930
1930 Petrozium Smith, p. 307.

Type species. By original designation, P. dewari Smith, 1930, from the Pentamerus Beds of Shropshire,
England; early Telychian.

Diagnosis. Phaceloid, corallites slender, cylindrical; increase presumably peripheral, nonparricidal. Septa long,
grouped at axis or slightly withdrawn, thin or slightly thickened, commonly zigzag but rarely carinate in
dissepimentarium. Tabularium strongly domed, tabulae complete with central area flattened or sagged and
with outturned edges, or commonly incomplete with outermost tabellae forming a distinct series producing a
periaxial trough. Dissepimentarium relatively narrow, dissepiments small, interseptal.

Remarks. Smith (1930) introduced Petrozium for one species described from a limited number of
specimens and no further material has become available (see discussion of P. dewari). The genus has
been variously interpreted by subsequent revisors. In his description of the type species. Smith
(1930, p. 308) recognized the similarity, in longitudinal section, of Petrozium to Entelophyllum but
did not elaborate on how they were to be distinguished and some authors have considered them
synonymous.

Smith (1930) described the increase as marginal nonparricidal but we have not been able to
confirm this in the material still available. If this is correct, then it contrasts with the typical
parricidal mode of increase in Entelophyllum and resembles Donacophyllum Dybowski, 1876, but
lacks the lonsdaleoid dissepiments characteristic of the latter genus.

Kaljo (1958) and Fedorowski and Gorianov (1973) both recognized the similarity between P.
dewari and Donacophyllum losseni Dybowski, 1874, from the Llandovery of Estonia. Kaljo referred
the Estonian form to the genus Petrozium , and Fedorowski and Gorianov questionably included P.
dewari in the synonymy of the Estonian species which they referred to Entelophyllum. The latter
authors described six specimens of P. losseni from Eichwald’s collection and noted that increase was
peripheral nonparricidal. Apart from this, P. losseni resembles other species of Entelophyllum in
which the major septa are thin, straight, and somewhat withdrawn from the axis leaving a
conspicuous axial space such as that found in E. fasciculatum. In its early growth stages the septa
may join at the axis as shown by Fedorowski and Gorianov (1973, fig. 6). The relatively narrow
dissepimentarium and the incomplete development of the periaxial series of tabellae separate both
P. dewari and P. losseni from E. articulation but those features are found in varying degrees in other
species of Entelophyllum.

Smith (1930, p. 307) described the development of stereome between the axial ends of the longer
major septa as forming with them an axial structure. Merriam (1972, 1973) considered the
development of an incipient axial structure as a significant generic character of Petrozium. As seen
in the Gotland species of Entelophyllum the withdrawal of the axial ends of the major septa from
the axis is very variable within species as well as between species. Examination of the type material
of P. dewari shows that the grouping of septa near the axis is common but the development of an
incipient axial structure is rare.

Until the type species is better known Petrozium is maintained for those species similar to
Entelophyllum but with nonparricidal increase, relatively narrow dissepimentarium, incomplete
development of the periaxial series of tabellae, and lacking lonsdaleoid dissepiments.

Petrozium dewari Smith, 1930
Text-fig. 9a-m

v*1930 Petrozium dewari Smith, p. 307, pi. 26, figs 20-28.

808

PALAEONTOLOGY. VOLUME 33

text-fig. 9. Petrozium dewari Smith, a-f, Pentamerus Beds, right bank Morrells Wood Brook, 275 m NNE
of Morrells Wood Farm, 1-6 km NNW of Buildwas, Shropshire; a and b, GSM PF4618 (holotype); transverse
sections; x 2, x 4, respectively; c, GSM PF4622u, transverse section in upper part of corallite, x4; d, GSM
PF46226, transverse section, x4; e, GSM PF4622c/, transverse section, x4; f, GSM PF4622e, transverse
section showing septa slightly thickened in early growth stage, x4. g, GSM PF4619. longitudinal section, x4.
h, GSM PF4623, longitudinal section, x 4. i-m, Saugh Hill Group, Woodland Point, Girvan, Ayrshire; i-k,
SM A15217; i, longitudinal section, x4; j and k, transverse sections, x4, x2 respectively; l and m, SM
A 1 52 1 5 ; L, transverse section, x4; m, longitudinal section, x4.

Holotype. Smith, 1930, pi. 26, fig. 22; GSM 48674 (thin section PF4618), Pentamerus Beds, Morrells Wood
Brook, F6 km NNW of Buildwas, Shropshire, England; early Telychian.

Material studied. The holotype and most of Smith’s (1930) material is from the Calostylis limestone, a thin
10-15 cm thick bed of limestone in the Pentamerus Beds, at the type locality (seven specimens; figured: GSM

JELL AND SUTHERLAND: SILURIAN RUGOSE CORALS

809

PF4618 (holotype); GSM PF4619, PF4622a-a', PF4623 (paratypes)); Smith’s collection from the upper part of
the Hughley Shale in the north bank of the Severn River, 366 m southeast of the church, Buildwas, Shropshire
(one specimen); Saugh Hill Group, Woodland Point, Girvan, Ayrshire, Scotland (two specimens; figured: SM
A 152 15, 15217).

Distribution. England: Telychian; Scotland: Llandovery ( ?Rhuddanian).

Diagnosis. Dendroid, almost placeloid; corallite diameter typically 8-10 mm; most of the 28-30
major septa meet at the axis or join in groups just short of the axis; minor septa 0-4—0-6 the length
of major; septa slightly thickened or lightly carinate in outer dissepimentarium ; tabularium biserial,
axial tabellae arched, periaxial series incompletely developed; dissepiments globose, irregular in
size, in two or three vertical rows.

Description. The holotype is a thin upper part of a dendroid to phaceloid corallum showing eight corallites
imbedded in a coarse bioclastic limestone so that no calices are exposed. The corallites are 8-10 mm in diameter
and appear to be cylindrical. Adjacent corallites are separated by as much as 5 mm but in section they are
touching in places suggesting they have lateral expansions that spread out to support the corallite against
adjacent ones. The mode of increase is not evident.

The other specimens are all fragments of isolated cylindrical corallites. Some show slight lateral expansions
of the corallite as though they were part of a phaceloid coralla.

Septa are commonly thin, straight, radially arranged in two orders with 28-30 in each. Major septa extend
well into the tabularium and may meet at the axis or join in groups of three or four just short of the axis (Text-
fig. 9c, d). In the holotype, an axial plane is evident across which opposing septa join so that the corallite has
a bilateral symmetry. Only in the largest corallite of the holotype, the axial ends of the septa are slightly
thickened and in places confluent, forming a small open axial structure (Text-fig. 9b). Minor septa are 04— 0 6
the length of the major with a few contraclined. In the dissepimentarium, both orders of septa are: commonly
straight and thin; rarely straight, slightly thickened and noncarinate; or rarely thin, slightly zigzag with short
carinae.

The tabularium and dissepimentarium are clearly differentiated with the former consisting of an axial and
periaxial series of tabellae. Axial tabellae are domed with the central part flattened or depressed in some and
their peripheral edges sharply downturned. A series of horizontal or slightly sagging tabellae, closely spaced
with 20-26 per 5 mm vertically, span the space between the downturned peripheral edges of the axial tabellae
and the dissepimentarium, forming a periaxial trough (Text-fig. 9h). The dissepimentarium is composed of
irregular sized, globose dissepiments, horizontally based at the periphery and steeply declined axially, and
arranged in two or three irregular rows with 8-10 in each row per 5 mm vertically.

Septa are composed of fine monacanths arranged in one series. Trabeculae are directed upwards at a small
angle to the vertical at the periphery and flatten axially so that at the tabularium/dissepimentarium boundary
they are within 20° of the horizontal and then towards the axis they steepen considerably. Thickenings of the
septa and the short carinae are only the lengthening of fibres of the trabeculae.

Remarks. Description is based on the same material as described by Smith (1930), excluding
specimen GSM 48721. We have twice visited the type locality without finding any additional
material referable to this species. Also, we have not recognized this species in our collections from
other upper Llandovery localities of Shropshire.

Smith (1930) included in this species specimen GSM 48721 (PF4529-30) from the type locality.
It is a fragment of a much larger corallite (20 mm in diameter) and whether or not it is a corallite
from a colony or a worn solitary form is not apparent. It seems unlikely that it is conspecific with
the smaller cylindrical corallites. A fragment of a phaceloid corallum (Text-fig. 9i- k) and an isolated
cylindrical corallite (Text-fig. 9l and m) from the Saugh Hill Group, Woodland Point, Girvan,
Ayrshire (SM A5215, A5217) are very similar to the Shropshire material. The Scottish material may
be Rhuddanian in age and thus somewhat older than the type material. We have searched the
Woodland Point sequence for additional material but without success. Several larger specimens are
known from the same collection but whether solitary or from fasciculate coralla is not known and
they are not included in the species at this time.

810

PALAEONTOLOGY, VOLUME 33

text-fig. 10. Petrozium losseni (Dybowski), EGM Col282, Tamsalu Formation, Juuru Horizon, Suurcmyiza
(near Grossenkhof), Isle of Hiiumaa, Estonia, a and b, longitudinal sections, x4. c, transverse section, x4.

Petrozium losseni (Dybowski, 1874)

Text-fig. 10a-c

1874 Donacophyllum lossenii Dybowski, p. 464, pi. 4, fig. 6, 6a, b.

1958 Petrozium losseni (Dybowski); Kaljo, p. 114, pi. 4, figs 11-17.

1973 Entelophyllum losseni (Dybowski); Fedorowski and Gorianov, 1973, p. 20, pi. 4, fig. 5 a-c, text-
figs 6 and 7.

Holotvpe. Dybowski, 1874, pi. 4, fig. 6, 6a, b\ EGM Co 1282; Tamsalu Formation, Juuru Horizon, at
Suuremyiza (near Grossenkhof)- Isle of Hiiumaa, Estonia; Rhuddanian.

Material studied. This species is interpreted from published data and from photographs of the holotype kindly
supplied by Dr D Kaljo (figured: EGM Co 1281, holotype). It is known from several localities on Hiiumaa
Island. Kaljo (1970) listed it as only occurring in the Juuru Horizon (Rhuddanian) in Estonia. Two specimens
in the Eichwald Collection are labelled as from Lode on Saaremaa Island and thus from the Kuressaare Horizon
(Ludfordian) but Fedorowski and Gorianov (1973, p. 23) suggest this is an error. The same authors list another
of Eichwald's specimens from Viljandi in southern Estonia and suggest it is probably from an erratic boulder.

Distribution. Estonia: Rhuddanian.

Diagnosis. Phaceloid, peripheral nonparricidal increase; corallite diameters 6-7 mm, maximum
9-5 mm; major septa number 23-26, straight, smooth, withdrawn from the axis leaving axial space
or extending to or almost to axis, several adjacent septa may be grouped just short of axis; minor
septa half length of major, not contratigent. Tabularium broad, biserial with narrow periaxial
trough; dissepiments small, globose in one to three rows.

Remarks. P. losseni is very similar to P. dewari , and Fedorowski and Gorianov (1973) suggested that
the two might be synonymous. P. losseni is here distinguished from P. dewari by the thinner septa,
especially in the tabularium, absence of any septal carination, thicker corallite walls, and typically
a narrower dissepimentarium. Tabularia of both species are very similar but there appears to be
more complete tabellae and less complete development of the periaxial series of tabellae in P.
losseni. The Estonian P. losseni is from slightly older strata (Rhuddanian) than the type material of

JELL AND SUTHERLAND: SILURIAN RUGOSE CORALS

P. den ari of England (early Telychian). P. losseni may be of similar age to the Scottish form referred
to P. dewari and resembles it in that the major septa may be withdrawn from the axis and are
similarly grouped. However, the septa of the Scottish material are thicker and more like the type
material of P. dewari.

P. losseni differs from the Llandovery E. dendroides from Gotland by its phaceloid habit,
withdrawal of septa from the axis, more complete and less densely spaced tabulae, and narrower
dissepimentarium.

Genus prohexagonaria Merriam, 1973

1973 Entelophylloides (Prohexagonaria) Merriam, p. 50.

Type species. By original designation, Entelophylloides ( Prohexagonaria ) occidentalis Merriam, 1973, from
bottom of unit 3, Roberts Mountains Formation, northwest side Roberts Creek Mountain, Nevada, U.S.A.;
?Pridoli Series.

Diagnosis. Cerioid or subcerioid ; increase peripheral nonparricidal ; septa thin with inconspicuous
asymmetrical carinae, long, some reaching the axis; tabulae closely spaced, complete or typically
incomplete, axially arched and outturned at their periphery; dissepiments globose, small, increasing
in axial inclination away from the periphery, typically interseptal, rarely lonsdaleoid.

Remarks. Merriam (1973) originally defined Prohexagonaria as a subgenus of Entelophylloides
Rukhin, 1938 (type species Columnaria inequalis Hall, 1852), but that genus is a ptenophyllid coral
and possibly synonymous with Xystriphyllum Hill, 1939. Prohexagonaria is generally similar to
Entelophyllum in internal character but is cerioid and has nonparricidal increase. In the latter
feature it shows resemblance to the phaceloid Donacophyllum but in Prohexagonaria lonsdaleoid
dissepiments are rare except in the extended corners of some corallites. Until the modes of increase
in all of the cerioid forms with typical entelophylloid internal structure are better known,
Prohexagonaria is included in the Entelophyllidae.

Prohexagonaria may be congeneric with Tenuiphyllum Soshkina, 1937 (type species T. ornatum
Soshkina, 1937, from the Elkinskii Horizon, right bank of River Vya, near Elkino, eastern slopes
of the Urals, USSR; see Ivanovskii and Shurygina 1975, p. 23). Pedder (1976, p. 290) indicates that
the two genera may not be synonymous as T. ornatum shows significantly expanded septal bases.

Prohexagonaria favia sp. nov.

Plate 10, figs 1-4

vl929 IXylodes sp. Smith and Tremberth. p. 366, pi. 8, fig. I
vl973 Entelophylloides (Prohexagonaria) sp., Merriam, pi. 9, fig. 5.

Holotype. Original of Smith and Tremberth, 1929, pi. 8, fig. 1 ; RM Cn57005, two parts of specimen plus three
thin sections; two parts of the holotype plus two thin sections are in the BMNH collection ( R26 1 68—261 71);
from upper Visby Beds, Norderstrand, Visby, Gotland; early Sheinwoodian.

Material. The holotype is the only known specimen.

Distribution. Gotland: early Sheinwoodian.

Diagnosis. Cerioid, increase peripheral nonparricidal; corallites 10-15 mm in diameter; major septa
number 21-25, long, extending to or almost to axis, only rarely confluent, commonly interrupted
at periphery; minor septa weakly developed, discontinuous in places; tabularium weakly biserial
with incomplete periaxial trough; dissepiments irregular in shape and size.

Description. The holotype is a fragment of a cerioid corallum and measures 15 x 10 cm in cross section and
5 cm high. Corallites are five- to six-sided and measure 10-15 mm in diameter. Offsets common along

812

PALAEONTOLOGY, VOLUME 33

boundaries between corallites especially on corners between three adjacent corallites (PI. 10, fig. 2) so that it
is difficult to determine from which corallite they have arisen. In longitudinal view they appear to arise from
the extreme outer edge of the dissepimentarium. Increase is nonparricidal.

Septa are long, thin, radially arranged in two orders. Major septa number 22-25 and extend right to or
almost to the axis (PI. 10, fig. 2). In a few cases one and even two major septa project into axial area further
than the others. Major septa thin slightly towards periphery and in extended corners before budding they may
be incomplete resting on larger dissepiments. Minor septa may be thin and complete, 0-3-04 the length of the
major septa, but more commonly they are incomplete and represented as short segments based on herringbone
dissepiments and more strongly developed in inner dissepimentarium than in outer parts. The septa are
noncarinate. In small offsets some of the septa are discontinuous in the dissepimentarium.

Tabularium is 0-6-0-7 diameter of corallite in width and is well marked from the dissepimentarium. It
consists of small globose tabellae arranged in a domed or flattened domed axial series 0-7 diameter of
tabularium, and flatter and smaller tabellae forming a periaxial series abutting dissepimentarium (PI. 10, fig.
4). Periaxial series does not form a continuous periaxial trough. There are up to 12 tabellae in both series in
5 mm vertically. Dissepimentarium is composed of irregular sized dissepiments varying from small globose, less
than 0-5 mm across, to elongated moderately inclined up to 3 mm in length spanning the complete
dissepimentarium.

Septa are composed of one series of fine trabeculae (5-6 in 1 mm in median longitudinal section) that are
inclined upwards and inwards in the dissepimentarium and steepening to near vertical in the tabularium.

Remarks. This species is not similar to any known European entelophylloid ceroid coral. It is similar
to the type species of the genus, which is from considerably younger strata in western North
America, differing only in that P. favia has noncarinate septa and its tabularium is considerably
wider.

Prohexagonaria gotlandica sp. nov.

Plate 10, figs 5-8

Holotype. RM Cn66198, Hamra Beds, Storburg 1, Hoburgen, southern tip of Gotland; late Ludfordian.
Material. Only the holotype is known.

Distribution. Gotland: late Ludfordian.

Diagnosis. Cerioid, increase peripheral nonparricidal; corallites 8-12 mm diameter; major septa
number 20-25, extending almost to axis, thickened in the inner dissepimentarium, zigzag and
carinate in outer dissepimentarium; minor septa well developed, carinate; tabularium biserial with
distinct periaxial trough; dissepiments globose, irregularly arranged with outer series horizontally
based and inner ones steeply inclined axially.

Description. Holotype is a fragment (8- 10 cm in section and up to 5 cm in height) of a larger corallum.
Corallites are five-to six-sided and measure 8-12 mm in diameter. Offsets arise in the expanded corners of the
corallites (PI. 10, fig. 6) and quickly expand so that the corallum resumes a regular cerioid form. Twenty calices
are preserved and from these and the sections, all increase appears to be peripheral nonparricidal.

EXPLANATION OF PLATE 10

Figs 1-4. Prohexagonaria favia sp. nov. RM Cn57005 (holotype), upper Visby Beds, Nordenstrand, Visby,
Gotland. 1 and 2, transverse sections, x 2, x4, respectively. 3 and 4, longitudinal sections, x 2, x4,
respectively.

Figs 5-8. Prohexagonaria gotlandica sp. nov. RM Cn66198 (holotype), Hamra Beds, Storburg 1, Hoburgen,
southern tip of Gotland. 5 and 6, transverse sections, x 2, x 4, respectively. 7 and 8, longitudinal sections,
x 2, x 4, respectively.

PLATE 10

JELL and SUTHERLAND, Prolw xagonaria

814

PALAEONTOLOGY, VOLUME 33

Major septa number 20-25 with one corallite having 30. They are slightly thickened in the inner
dissepimentarium and thin in the tabularium where they are straight to flexuose and extend almost to the axis,
leaving a narrow axial space (PI. 10, fig. 6). Minor septa are slightly thinner than the major, 0-7-0-8 the length
of major, and project into tabularium a short distance. All septa are zigzag in outer dissepimentarium and may
carry short carinae. Carination varies from absent to light in outer dissepimentarium, light to heavy in middle
dissepimentarium, and light on thickened parts in the inner dissepimentarium. Only in extended corners where
offsets are forming do septa become discontinuous.

The longitudinal section is oblique but locally shows a typical entelophylloid tabularium half the radius of
the corallite (PI. 10, fig. 8). Axial series consists of flattened tabellae strongly downturned at their margins; the
downturned edges may flatten out to meet inner dissepimentarium horizontally or they are surrounded by a
series of flat to sagging tabellae. A periaxial trough is well-developed. Dissepiments are globose, varying from
0-5-1 -0 mm across, not arranged in regular vertical rows.

Remarks. This species differs from P. favia in having distinctly smaller corallites, better developed
septa that are commonly carinate, and a narrower tabularium with a well developed periaxial
trough.

Family kyphophyllidae Wedekind, 1927
Genus donacophyllum Dybowski, 1874

1874 Donacophyllum Dybowski, p. 460.

1927 Kyphophyllum Wedekind, p. 19.

Type species. Chosen by subsequent designation of Wedekind, 1927, p. 34, Donacophyllum schrenkii Dybowski,
1874, from Raikkiila Horizon of Estonia; Aeronian, middle Llandovery.

Diagnosis. Phaceloid, typically with nonparricidal peripheral increase; periodic rejuvenescence of
corallites, reflecting expansions and subsequent contractions of dissepimentarium; major septa long
and thin, a little withdrawn from axis, one may be longer; tabularium wide, domed, commonly with
depressed axial area and marginal periaxial trough; dissepiments commonly lonsdaleoid.

Remarks. We follow Hill (1981) in placing Kyphophyllum in the synonomy of Donacophyllum. We
have not made a comprehensive study of these genera. Gotland species here included in
Donacophyllum are close to Entelophyllum in septal and tabularial characters but increase is lateral
nonparricidal and lonsdaleoid dissepiments are much better developed.

Donacophyllum neumani sp. nov.

Plate 1 1, figs 1-8

Holotype. RM Cn66199, from a limestone of the Slite Beds, towards top of cliff section of Lerberget 1, Stora
Karlso, Gotland; late Sheinwoodian or early Homerian.

Material studied. This species is known only from the type locality (five specimens; figured: RM
Cn66199 (holotype), Cn66200, Cn66201).

Distribution. Gotland: late Sheinwoodian or early Homerian.

EXPLANATION OF PLATE 1 I

Figs 1-8. Donacophyllum neumani sp. nov., Slite Beds, Lerberget 1, south ot lighthouse, western shore of Stora
Karlso, Gotland. 1^4, RM Cn66199 (holotype); 1 and 2, transverse sections, x4, x2 respectively; 3,
longitudinal section, x 2; 4, lateral view of corallum, x I. 5 and 6, RM Cn66200; 5, transverse section, x 2;
6, longitudinal section x 2. 7 and 8, RM Cn66201; 7, transverse section, x2; 8, longitudinal section, x 2.

PLATE 1 1

JELL and SUTHERLAND, Donacophyllum

816

PALAEONTOLOGY, VOLUME 33

Diagnosis. Fasciculate, flat, laterally flaring coralla ; increase lateral nonparricidal ; corallites up to
20 mm in diameter. Major septa 25-31, long, generally straight, smooth in tabularium, extending to
within 1 mm of axis; slightly thickened in outer tabularium and inner dissepimentarium ; may be
carinate in dissepimentarium, interrupted by lonsdaleoid dissepiments in outer dissepimentarium;
minor septa absent to weakly developed. Tabularium half diameter of corallite, biserial with well
developed periaxial trough, axial part of domed tabularial floor sagged. Dissepiments medium to
large sized, highly irregular, not in regular vertical rows with larger lonsdaleoid dissepiments at
periphery.

Description. Holotype is a flat laterally flaring corallum 100 mm in diameter and 70 mm high. Corallites are
ceratoid up to 20 mm in diameter and 20-50 mm in height. Increase is lateral, nonparricidal with offsets arising
from extreme peripheral rim of the calice (PI. 1 1, fig. 3). Offsets turn outwards soon after initiation and diverge
up to 30° from parent (PI. 11, fig. 4). No lateral expansions between corallites have been observed. Other
specimens from type locality show same corallum form and corallite shape and size.

Major septa number 25-31, are moderately straight and smooth in tabularium, and extend to within 1 mm
of the axis. In a few corallites, a single septum extends into axial space and in some, two or three adjacent septa
are confluent giving small groupings of septal ends about the axial space. Some show a slight thickening in
outer tabularium and/or inner dissepimentarium. Septa are commonly discontinuous in outer dis-
sepimentarium where they are interrupted by irregular lonsdaleoid dissepiments. They may become zigzag and
carry short carinae, especially along their peripheral parts. Minor septa mostly poorly developed, much thinner
and more irregular than adjacent major septa and commonly lacking from parts of some corallites. They
project only slightly into tabularium. In young, small corallites septa of both orders are smooth, straight and
interrupted peripherally by large dissepiments.

Wide tabularium, 0-4— 0-5 corallite diameter in width, is distinctly separate from dissepimentarium. It is
biserial with axial tabulae centrally domed but with an axial depression, downturned steeply at margin of this
central area and then flattening out or even concave peripherally to give a prominent periaxial trough. Central
tabellae are flat or slightly concave upwards and number 8-10 in 5 mm vertically. They are surrounded by more
globose tabellae that have steeply downturned outer edges that rest on the ones immediately below. Periaxial
tabellae may span interval between domal part of tabularium and dissepimentarium. These are flat or concave
upwards and number 10-12 for 5 mm vertically. Dissepiments are mostly elongated, adaxially inclined,
variable in size, and not arranged in regular vertical series. Lonsdaleoid dissepiments common in peripheral
parts.

Septa consist of one series of monacanthine trabeculae that are directed upwards and inwards across the
dissepimentarium at approximately 45°. They tend to flatten to nearly horizontal in periaxial region of
tabularium but steepen to within 30° of vertical in axial region. Fibers within trabeculae diverge at only a slight
angle from axis so that fibers appear to parallel trabeculae, and boundaries and axes of individual trabeculae
cannot be distinguished. Carinae are lateral extensions of the fibers of the trabeculae and, where present, are
0-2-0-5 mm apart, centre to centre.

Remarks. D. neumani is distinguished from E. articulatum, also from the Slite Beds, primarily by its
lateral nonparricidal increase and flaring growth form of the corallum, and major development
of lonsdaleoid dissepiments. It differs from D. wallstenense in the flaring growth form of the
corallum and much larger size of the corallites.

EXPLANATION OF PLATE 12

Figs 1—11. Donacophyllum wallstenense sp. nov. 1—4, RM Cnl0637, Slite Beds, Stora Karlso, Gotland; 1,
transverse section, x4; 2, longitudinal section, x4; 3 and 4, lateral view of corallite; 3, showing offset
arising from expansion of corallite, x 4; 4, showing series of expansions, x 1. 5-8, RM Cnl 1963 (holotype),
Slite Beds, Stumra stenbrott, Wallstena Parish, Gotland; 5 and 6, lateral view of corallum showing lateral
nonparricidal increase, x 1, x 2, respectively ; 7, longitudinal section, x 2; 8, transverse section, x 2. 9-11,
Slite Beds, Stora Karlso, Gotland; 9, RM Cnl0643, showing lateral expansion, x4; 10 and 11, RM
Cnl 0635; 10, transverse section, x4; 11, longitudinal section, x4.

PLATE 12

JELL and SUTHERLAND, Donacophyllum

818

PALAEONTOLOGY, VOLUME 33

Donacophyllum wallstenense sp. nov.

Plate 12, figs 1-1 1

Holotype. RM Cnl 1963, Slite Beds, Stumra stenbrott (stone quarry), Wallstena Parish, Gotland; Slite Group;
late Sheinwoodian to early Homerian.

Material studied. Holotype (RM Cn 1 1963) is the only specimen from Wallstena. Additional specimens referred
to this species are also apparently from the Slite Group, from an unknown locality on the island of Stora
Karlso, Gotland (14 Specimens, RM Cnl0635- 10647 ; figured : RM Cnl0635, Cnl0637, Cnl0643). Some of the
specimens from Stora Karlso show small remnants of gray marl and thus they may have come from the
lithology of that type observed in the lower part of the Slite Group below the more greenish limestone from
which were collected the type specimens of D. neumani sp. nov. An additional occurrence is questionable: 2 km
southeast of ferry at Broa, Isle of Faro, Gotland.

Distribution. Gotland: late Sheinwoodian to early Homerian.

Diagnosis. Phaceloid, tall coralla; increase lateral nonparricidal; corallites up to 15 mm in diameter.
Major septa number 23-27, typically smooth leaving a small open axial area into which a single
septum may extend; may be distinctly thickened in outer tabularium and/or in dissepimentarium ;
may be interrupted by lonsdaleoid dissepiments in outer dissepimentarium. Tabularium half
diameter of corallite, biserial with a well developed periaxial trough. Dissepiments variable in size
and distribution.

Description of holotype. Holotype is a fragment of a tall, phaceloid corallunr up to 90 mm in height and 75 mm
across, consisting of close to 20 corallites. Corallites are long, subcylindrical and closely packed. They show
semi-regular expansions and contractions at a spacing of 10-15 mm apart (PI. 12, fig. 5). Method of increase
is lateral and nonparricidal, with the offset diverging at outer lip of expansion but becoming immediately
parallel to parent (PI. 12, fig. 6).

Major septa number 23-27 and they extend to within I mm of axis. Commonly one long septum bisects the
narrow axial space (PI. 12, fig. 8). Septa mostly smooth and straight throughout their length but a few are
zigzag in dissepimentarium and a few are carinate. In some corallites they are rather markedly thickened in the
outer tabularium and less commonly in dissepimentarium. Septa thickening is lacking in some corallites.
Peripherally septa may be discontinuous or interrupted locally by lonsdaleoid dissepiments, especially at lateral
expansions. Minor septa extend slightly into tabularium and tend to have a similar pattern in dissepimentarium
to major septa or they may be more discontinuous.

Tabularium is half diameter of corallite and is distinctly delineated from dissepimentarium. Central highly
domed area may or may not have a central depression. Tabulae steeply downturned periaxially to form a
distinct periaxial trough (PI. 12, fig. 7). Axial tabellae number about 8 in 5 mm vertically. Axial part of domal
area consists of flat or slightly depressed tabellae bordered by more globose plates whose peripheral edges are
steeply downturned. Flat to deeply sagging tabellae span space between central domal area and
dissepimentarium. Dissepiments are variable in size and shape and range from globose to more elongate,
peripherally lonsdaleoid dissepiments are common. Width of dissepimentarium varies markedly because of
expansions and contractions of corallite.

Microstructure is similar to that described in D. neumani.

Description of other material. Specimens from the island of Stora Karlso that are referred to this species consist
of 14 broken individual corallites (PI. 12, figs 1-4, 9-1 1 ). They are similar to the smaller corallites of holotype
in number of septa, nature of biserial tabularium and variable dissepimentarium. They differ from holotype
primarily in lack of pronounced thickening in outer tabularium in some corallites, in greater tendency of septa
to extend to axis, and in more common occurrence in dissepimentarium of zigzag septa and carinae.

Remarks. D. wallstenense differs from D. neumani in the upright growth form of the corallum, the
distinctly smaller size of the corallites, and the greater development of thickening at the outer
tabularium and dissepimentarium.

JELL AND SUTHERLAND: SILURIAN RUGOSE CORALS

819

Acknowledgements . We wish to thank the many persons who have aided us in obtaining a better understanding
of the entelophylloid corals. In particular, we thank Bjorn Neuman (University of Bergen) and the late Anders
Martinsson (University of Uppsala) for having shown us the Silurian stratigraphy of Gotland and for having
taken us to many collecting localities; Bjorn Neuman for the loan of some of his coral collections from
Gotland; Dr V. A. Sytova for having shown us in Leningrad her extensive coral collections and for extensive
information on the distribution of entelophylloid corals in the Soviet Union; Dr Dorothy Hill (University of
Queensland) for her knowledgeable discussions on the relationships of Silurian coral genera; Dr D. L. Kaljo
(Institute of Geology, Estonia) for information on Estonian coral distributions; Dr A. Galle (Ustfedni ustav
geologicky) for photographs and information on the Czechoslovakian forms; and Dr Michael Bassett
(National Museum of Wales) for information on the international correlations of Silurian stages. For the loan
of specimens for study we are indebted to the curators of the museums listed at the beginning of the section
on Systematic Palaeontology. We owe a particular debt of thanks to William W. Clopine (University of
Oklahoma) who assisted us during the final preparation of the manuscript by both the preparation of critical
thin sections and by photographing many of the specimens.

REFERENCES

bassett, m. G. 1974. Review of the stratigraphy of the Wenlock Series in the Welsh Borderland and south
Wales. Palaeontology , 17, 745-777.

— and COCKS, L. r. m. 1974. A review of Silurian brachiopods from Gotland. Fossils and Strata , 3, 1-56.

— Holland, c. h., rickards, r. b. and warren, p. t. 1975. The type Wenlock Series. Report of the
Institute of Geological Sciences , 75/13. 1 19.

— lawson, j. d. and white, d. e. 1982. The Downton Series as the fourth series of the Silurian System.
Lethaia , 15, 1-24.

bulvanker, E. z. 1952. Korally rugoza silura Podolii. Trudy Vsesoyuznogo Nauchoissledovatelskogo
Geologicheskogo Instituta (VSEGEI) Ministerstra Geologii. Gosudarstrennoe Izdatelstvo Geologicheskoy
Literatury, Moscow, 33 pp. [In Russian],

cocks, l. r. m. and fortey, r. a. 1982. Faunal evidence for oceanic separations in the Palaeozoic of Britain.
Journal of the Geological Society of London , 139, 465-478.

— and toghill, p. 1973. The biostratigraphy of the Silurian rocks of the Girvan district, Scotland. Journal
of the Geological Society of London, 129, 209-243.

dybowski, w. n. 1874. Monographic der Zoantharia sclerodermata rugosa aus der Silurformation Estlands,
Nord-Livlands und der Insel Gotland. Archiv fur die Naturkunde Liv -, Ehst- u. Kurlands , Series 7, 5, 415-532.
eichwald, c. e. d’ 1861. Paleontologiya Rossii: Drevniy period. St Petersburg, 520 pp. [Not seen by the
authors],

fedorowski, j. and gorianov, v. b. 1973. Redescription of tetracorals described by E. Eichwald in
‘Palaeontology of Russia'. Acta Palaeontologica Polonica, 18, 3-70.
hall, j. 1852. Palaeontology of New York. Volume 2. In Natural history of New York, Part VI. Carrol and
Cook, Albany, vii + 362 pp.

hede, i. E. 1958. Styggfordsen Limestone. Lexique Stratigraphique International, 1, (2c), 369.

— 1960. The Silurian of Gotland. 44-89. In regnell, g. and hede, j. e. (eds). The Lower Palaeozoic of
Scania. The Silurian of Gotland. International Geological Congress, XXI Session, Norden. Guidebook
Sweden d, Stockholm, 89 pp.

hill, d. 1939. The Middle Devonian rugose corals of Queensland, 1: Douglas Creek, Clermont district.
Proceedings of the Royal Society of Queensland , 50, 55-65.

— 1940. The Silurian Rugosa of the Yass-Browning district, N.S.W. Proceedings of the Linnean Society of
New South Wales, 65, 388-420.

- 1981. Rugosa and Tabulata. In teichert, c. (ed.). Treatise on invertebrate paleontology. Part F.
Coelenterata, Supplement I. Geological Society of America and University of Kansas Press, Boulder,
Colorado and Lawrence, Kansas, 762 pp.

hisinger, w. 1837. Lethaea Svecica seu Petriftcata Sveciae, iconibus et characteribus illustrata. P. A. Norstedt,
Holmiae, iv+112 pp., supplement pp. 113-124.

Holland, c. H. and bassett, M. G. (eds). 1989. A global standard for the Silurian System. National Museum of
Wales, Geological Series No. 10, Cardiff, 325 pp.

— lawson, j. d., walmsley, v. g. and white, d. e. 1980. Ludlow Stages. Lethaia, 13, 268.

820

PALAEONTOLOGY, VOLUME 33

ivanovskii, a. b. 1976. Ukazatel rodov rugoz. Trudy Institute! Geologii i Geofiziki. Sibirskoe Otdelenie, 217,
1-255. [In Russian].

— and shurygina, M. V. 1975. Reviziya rugoz Urala. Trudy Institute! Geologii i Geofiziki. Sibirskoe Otdelenie ,
218. 1-66. [In Russian],

kaljo, d. L. 1958. Nekotorye novye i maloizvestnye rugozy Pribaltiki. Eesti NSV Teaduste Akadeemia,
Geoloogiei Instituudi , Uurimused 3, 101-123. [In Russian],

1970. Silur Estonii. Valgus, Tallin, 343 pp. [In Russian],
lang, w. d. and smith, s. 1927. A critical revision of the rugose corals described by W. Lonsdale in
Murchison’s ‘Silurian System’. Quarterly Journal of the Geological Society of London, 83, 448—491.

— and thomas, h. d. 1940. Index of Palaeozoic coral genera. British Museum (Natural History),
London, 231 pp.

laufeld, s. 1974. Reference localities for palaeontology and geology in the Silurian of Gotland. Sveriges
geologiska Undersoknig Afhandlingar och Uppsaster, Stockholm , Series C, 705, 1-172.
lindstrom, G. 1880. iv + 60 pp. In angelin, N. p. Fragmentd Silurica. Samson and Wallin, Holmiae, 60 pp.
lonsdale, w. 1839. Corals. 675-694. In Murchison, r. i. The Silurian System. John Murray, London,
i-xxxii + 768 pp.

mckerrow, w. s. 1988. The development of the Iapetus Ocean from the Arenig to the Wenlock. 405-412. In
Harris, A. L. and fettes, d. j. (eds). The Caledonian - Appalachian Orogen. Special Publication of the
Geological Society of London, Blackwell Scientific Publications, Oxford, 643 pp.
mclean, r. a. and pedder, A. e. h. 1984. Frasnian rugose corals of Western Canada. Part 1 : Chonophyllidae
and Kyphophyllidae. Palaeontographica , Abteilung A, 185, 1-38.

Matthews, s. c. 1973. Notes on open nomenclature and on synonymy lists. Palaeontology , 16, 713-719.
merriam, c. w. 1972. Silurian rugose corals of the Klamath Mountains region, California. Professional Paper.
United States Geological Survey , 738, 1-50.

— 1973. Silurian rugose corals of the central and southwest Great Basin. Professional Paper. United States
Geological Survey , 777, 1-66.

milne-edwards, h. and haime, j. 1850-1855. A monograph of the British fossil corals. Monograph of the
Palaeontographical Society , pp. i -lxxxv + 1-71 , pis 1-11 ( 1850); pp. 147-210, pis 31-46 ( 1852); pp. 21 1-244,
pis 47-56 (1853); pp. 245-299, pis 57-72 (1855).

neuman, b. and hanken, n-m. 1979. Rugose corals. 86-91. In jaanusson, v., laufeld, s. and skoglund, r.
(eds). Lower Wenlock faunal and floral dynamics - Vattenfallet Section, Gotland. Sveriges Geologiska
Under sokning Afhandlingar och Uppsaster , Stockholm , Series C, 762, 1-294.
pedder, a. E. h 1976. First records of five rugose coral genera from Upper Silurian rocks of the Canadian
Arctic Islands. Paper. Geological Survey of Canada , 76-1B, 287-293.
pocta, p. 1902. Anthozoaires et Alcyonaires. In barrande, j. Systeme Silurien du centre de la Boheme. Part
1 , v. 8 , t. 2. The Author, Paris and Prague, viii + 347 pp.
prantl, F. 1940. Korallengattung Xylodes Lang and Smith (Rugosa) in bohmischen Silur. Mitteilungen der
Tschechischen Akademie der Wissenschaften, 1940, 1-21.
rukhin, l. b. 1938. Nizhnepaleozoiskie korally i stromatoporoidei verkhney chasti basseyna R. Kolymy.
Materialy po izucheniyu Kolymsko - Indigirskogo kraya, series 2, geologiya i geomorfologiya, Vyrusk 10,
State Trust Dalstroy, Leningrad, Moscow, 119 pp. [In Russian],
rozkowska, m. 1962. Gornosylurskie tetracoralla z warstw rzepinskich w profilu Lezyce-Belcz (Gory
Swietokrzyskie). Biuletyn Instytutu geologicznego, 174, 115-150.
schouppe, a. von 1951. Kritische Betrachtungen und Revision des Genus Entelophyllum Wdk. nebst einigen
Bemerkungen zu Wedekins “Kyphophyllidae” und “Kodonophyllidae”. Sitzungsberichte der
Osterreichischen Akademie der Wissenschaften. Mathematisch-naturwissenschaftliche klassee, (1), 160,
243-256.

shurygina, m. v. 1977. Rugozy verkhnego silura Ufimskogo Amfiteatra. 49-60. In Novye materialy po
paleontologii Urala. Akademiya Nauk SSSR, Uralskii Nauchnyi Tsentr, Instituta Geologii Geochemica,
Sverdlovsk. [In Russian],

smith, s. 1930. Some Valentian corals from Shropshire and Montgomeryshire, with a note on a new
stromatoporoid. Quarterly Journal of the Geological Society of London , 86, 291-330.

— 1933. On Xylodes rugosus sp. nov., a Niagaran coral. American Journal of Science , (5), 26, 512-522.

— and tremberth, r. 1929. On the Silurian corals Madreporites articulatus Wahlenberg and Madrepora
truncata Linnaeus. Annals and Magazine of Natural History , (10), 3, 361-376.

soshkina, e. d. 1937. Korally verkhnego silura i nizhnego devona vostochnogo i zapadnogo sklonov Urala.
Trudy Paleontologicheskogo Instituta , 6, 1-155. [In Russian].

JELL AND SUTHERLAND: SILURIAN RUGOSE CORALS

821

1955. Korally. 118-128. In ivanova, e. a., soshkina, e. d., astrova, g. g. and ivanova, v. a. Fauna
ordovika i gotlandiya nizhncgo techniya R. Podkamcnnoy Tungusku, ee ekologiya i stratigrapicheskoe
znachenie. Trudy Paleontologicheskogo Instituta 56, 118-128. [In Russian],

— and Dobrolyubova, T. a. 1962. Olryad Evenkiellida. 33-339. In orlov, yu. a. (ed.). Osnovy Paleontologii.
sokolov, b. s. (ed.). V. 2 , Gubki , Arkheotsiaty , Kishechnopolostnye, chervl. Akademiya Nauk, SSSR,
Moscow, 486 pp. [In Russian],

strelnikov, s. i. 1964. Ob obeme semeystva Kodonophyllidae (Rugosa). Paleontologicheskiy Zhurnal, 4,
49-60. [In Russian],

sytova, v. a. 1952. Korally semeystva Kyphophyllidae iz verkhnego silura Urala. Trudy Paleontologicheskogo
Instituta , 40. 127-158. [In Russian].

Siluriisko-Devonskaya fauna Podolii. Nauchno Issledobatelskii Institut Zemnoi Kory. Paleontologicheskaya
Laboratoriya, Izdatelstvo Leningradskogo Universiteta, Leningrad, 124 pp. [In Russian],
tomczyk, H. 1970. The Silurian. 237-319. In sokolowski, s. (ed.). Geology of Poland. 1. Stratigraphy , Pt 1 ,
Pre-Cambrian and Palaeozoic. Wydawnictwa Geologiczne, Warsaw.
wahlenberg, g. 1821. Petrificata telluris Svecanae. Nova Acta Regiae Societatis Scientiarum Upsaliensis , 8,

wang, H. c. 1944. The Silurian rugose corals of northern and eastern Yunnan. Bulletin of the Geological Society
of China , 24. 21-32.

Waterhouse, c. o. 1876. New species of Coleoptera from the Island of Rodriguez, collected by the Naturalists
accompanying the Transit-of-Venus Expedition. Annals and Magazine of Natural History , (4), 18, 105-121.
wedekind, r. 1927. Die Zoantharia Rugosa von Gotland (bes. Nordgotland). Sveriges Geologiska
Under sokning Afhandlingar och Uppsaster, Stockholm , Series C, 19, 1-94.
weissermel, w. 1894. Doe Korallen der Silurgeschiebe Ostpreussens und des ostlichen Westpreussens.

Zeitschrift der Deutschen geologischen Gesellschaft , 46, 580-674.
yu chang-ming 1956. Some Silurian corals from the Chiuchuan Basin, western Kansu. Acta Palaeontologica
Sinica , 4, 599-620.

zheltonogova, v. a. 1965. Znachenie rugoz dlya stratigrafii Silura gornogo Altaya i Salaira. 33 — 44. In
sokolov, B. s. and ivanovskii, a. b. (eds). Rugozy Paleozoya SSSR. Tr. I Vsesoyuznogo simpozioma po
izuchenie iskopaemykh korallov SSSR , pt. 3. Akademiya Nauk SSSR, Institut Geologii Geofiziki Sibirskoe
Otdelenie, Moscow, 90 pp. [In Russian],

ziegler, a. m., cocks, l. r. m. and mckerrow, w. s. 1968. The Llandovery transgression of the Welsh
Borderland. Palaeontology , 1 1. 736-782.

1-116.

JOHN S. JELL

Department of Geology and Mineralogy
University of Queensland
St Lucia, QLD 4067, Australia

Typescript received 9 May 1989

Revised typescript received 1 February 1990

PATRICK K. SUTHERLAND

School of Geology and Geophysics
University of Oklahoma
Norman, OK 73019, USA

%

BRACHIOPODS AND THE STRATIGRAPHY OF THE
UPPER CAMPANIAN AND LOWER
MAASTRICHTIAN CHALK OF NORFOLK,

ENGLAND

by m. b. johansen and f. surlyk

Abstract. Thirty-four species of brachiopods are described from the Upper Campanian and Lower
Maastrichtian Chalk of Norfolk. The Campanian-Maastrichtian boundary sequence in Norfolk is very poorly
exposed and is condensed. However, the existence of a highly diverse micromorphic brachiopod fauna allows
correlation with more complete and well-exposed sections elsewhere in northwest Europe, especially northwest
Germany and Denmark. Most of the brachiopod zones established in continental northwest Europe can be
recognized in Norfolk. These are from below: the tenuicostata-longicollis Zone (corresponding to the bulk of
the Upper Campanian of Norfolk), the longicollis-jasmundi Zone (uppermost Campanian), the acutirostris-
spinosa Zone (basal Maastrichtian), the spinosa-pulchellus Zone (middle part of the Lower Maastrichtian) and
the pulchellus-pulchellus Zone (upper Lower Maastrichtian). The stratigraphical scheme of previous workers
was essentially based on faunal and subordinate lithological criteria. This is revised within a strictly
lithostratigraphical framework to avoid ambiguities. Eight members are recognized comprising, from below:
Eaton Chalk, Weybourne Chalk (including Catton Sponge Bed at its top), Beeston Chalk, Paramoudra Chalk,
Sidestrand Chalk, Trimingham Sponge Beds, Little Marl Point Chalk, and Beacon Hill Grey Chalk Members.

High diversity faunas of micromorphic brachiopods are well known from the Upper Cretaceous
chalks of Denmark, Sweden, Germany and Poland (Steinich 1965, 1967, 1968c/, b ; Surlyk 1970a, b ,
1972, 1973, 1974, 1975, 1982, 1984; Surlyk and Birkelund 1977; Surlyk and Johansen 1984; Bitner
and Pisera 1979; Ernst 1984; Johansen 1986, 1987a, b , 1988), and a detailed brachiopod biozonation
has been established for the Maastrichtian Stage. Washing of a large number of bulk samples
collected in the Upper Campanian - Maastrichtian Chalk of Norfolk has yielded about 2200
specimens of brachiopods representing approximately thirty-five species. The occurrence of this
fauna in the English Chalk makes it possible to correlate parts of the often poorly exposed
Maastrichtian outcrops in Norfolk with the more completely exposed sections of continental
northwest Europe, and to support previous correlations based on other faunal groups.
Palaeoecological conclusions obtained by the study of the micromorphic fauna (Surlyk 1972) can
also be extended to these parts of the English Chalk.

METHODS

Thirty-seven bulk samples were washed after boiling in a supersaturated Glauber-Salt solution
following the method described by Surlyk ( 1 972). Sample dry weight varied between 1 and 5 kg. The
washed residues were sieved into 0-25-0-5 mm, 0-5—10 mm, and > TO mm fractions, and the
brachiopods were picked from the two latter fractions under a binocular microscope. The
0-25-0-5 mm fraction was in some instances checked for the earliest juvenile stages.

All specimens including the smallest juveniles were determined to species, and the minimum
number of individuals was computed for each sample by adding to the number of complete shells
the highest number of either dorsal or ventral valves. Ultrasonic cleaning was necessary when the
details of shell ornamentation were hidden by adherent chalk. This was particularly the case with

| Palaeontology, Vol. 33, Part 4, 1990, pp. 823-872, 11 pls.|

© The Palaeontological Association

824

PALAEONTOLOGY, VOLUME 33

material from the more indurated lithologies. SEM photographs were used for illustrative purposes,
as normal photography is inadequate for illustrating finer details of rib sculpture and the pedicle
foramen of the very small species. Preparation was done under a binocular microscope with a fine
needle and brush, after attaching the specimen to a glass plate with ‘Lakeside’ cement.

The catalogued material is housed at the Geological Museum, University of Copenhagen.
Material not catalogued belongs to a private collection of the second author of this paper (F.S.).

STRATIGRAPHY

The Campanian-Maastrichtian part of the Chalk in Norfolk is not very well known because of its
incomplete exposure (Text-fig. 1 ). For much of the succession the sections are limited to widely
scattered, often abandoned, chalk pits, and larger, glacially transported masses in coastal cliff's. This
chalk was subdivided into a number of units by Brydone (1906, 1908, 1909, 1930, 1938). These units
form the basis for the detailed geological map of Peake and Hancock (1970, pi. 2).

The nature of Brydone’s stratigraphic units is somewhat unclear. Most of the subdivisions are
named after a locality, although the fossil content is normally the main defining element. According
to Wood (1988) the scheme is essentially palaeontological and the units of Brydone, as revised and
updated by Peake and Hancock (1961, 1970), can be mapped out as faunal belts. The units are,
however, in most cases characterized by a distinct lithology, or topped by a thin horizon of different
lithology such as a hardground. They are redefined as lithostratigraphic units in the present paper,
but it should be noted that the units recognized here are essentially equivalent to those of earlier
authors even if they are conceptually different. Our aim in this revision is thus to introduce some
consistency to the stratigraphical terminology.

The rank of the units is difficult to establish since the Upper Cretaceous Chalk of the North Sea
and surrounding land areas has received little study from a lithostratigraphic point of view. Several
formations have been recognized for the subsurface chalks of the North Sea (e.g. Deegan and Scull
1977; Svendsen 1979). These units are, however, of a very broad nature, and they are defined on
the basis of petrophysical logs and relatively sparse core material from boreholes. They are difficult
to correlate with exposed onshore chalk sequences.

The lithological differences between the individual units of the Chalk of Norfolk are relatively
small and the thickness of the individual units usually amounts to a few metres only. The units can
thus conveniently be treated as having member rank. The Chalk of Norfolk is generally of a more
shallow-water nature than the North Sea Chalk, as witnessed by the common occurrence of single
or compound hardgrounds and richer benthic faunas.

Lit ho stratigraphy

All localities described in the present paper expose Chalk of Late Campanian and Maastrichtian age; older
sediments, including the ‘Basal mucronata chalk’ of Norfolk (Peake and Hancock 1970), are thus not dealt
with in the following account. Since the lithostratigraphic units are given member rank (Text-figs 1 and 3) they
should logically be grouped in one or more formations. This we leave, however, to students of the regional
geology of eastern England. The description of the individual members is mainly based on Peake and Hancock
(1970), Wood (1967, 1988), and our own observations.

Eaton Chalk Member

Name. After underground workings at the village of Eaton (National Grid Reference TG 208063).

Type section. The great part of the member is unexposed, so a satisfactory type section cannot be designated.

Reference sections. Small exposures may be found in the vicinity of Cley (TG 054440) and Drayton
(TG 175132). Coastal sections at Weybourne Hope show the upper boundary, which is taken at Flint Z of
Peake and Hancock (1970, fig. 5, p. 339D-E; Text-fig. 1 herein).

Thickness. Approximately 15 m (Peake and Hancock 1970, fig. 3).

Lithology. Soft chalk with irregularly scattered flints, sometimes forming ‘open’ bands.

JOHANSEN AND SURLYK: CRETACEOUS BRACHIOPODS

825

0- J=

- o

U x CD O
J)

lu Q 5 cr
t- z < <

Sssy

5UJ 2 h

o9it
owh j

— ST To- 1 I -*1 |

v,

wm

iViAViyi1

it. i 1 1 1.«

TU)!1!'

IT ITMPjM

j .* T i_ ^ ij 4' i ii i t 1 \i\' iHii y 8 1 >

t 1

text-fig. 1. Geological sketch showing the Upper Campanian-Lower Maastrichtian localities and sections of Norfolk described in this paper. The map

and the letter key to individual beds are modified from Peake and Hancock (1970).

EE 3WmOA ADO'lOlNOHV'lVd

938

826

PALAEONTOLOGY, VOLUME 33

sijb/oiubjo jje 9
eie/ndai y
HiajaiiidM 7
ds pipiapoam
eioejiaj Q
Bsouids y
opunjiq \/

ds snosjUBJQ
Bioumuoo JJB

S Biouniuoo v

ipunuusBl 9

spisoj unos y

xapuJBO o

shiodiGuoi i

ds sninujnos

ds Binujay

iisi6 9

ds piinjejqajaj

uuuojq v

USBlnBI ±

smqou o
c BaoBiajo 7
ds o $ BisqLuii o
si/BuipjBoqns Q
BiBpsnui \/
simqns i
BiB)sooiuids y
BiBinbuBiuad >/
Bieisoomuai y
snoiuo/od 7
siiujopuoiiqo iv
siissAjqo ±

SfllOBJd ±

Bptdu \\b ±
ds pinauoqouAqj
BIBISOO 7

CO CO C\J *— T- X

0)2

c

E £

I s

m O

-C-CO O (0 [f) O Tt Cj) o '

= 2 I CD § i o § I I

H\U\M\ \

* >.
$ 2
* o

S3N0Z QOdOlHOVda

snnaqoind-\

snnaqoind\

sniiaqo/nd - ssouids

BSOUfdS-

suisojpnoB

Sfnooiduoi - BiBisoomuai

NVIlHOIdiSVVW HSMCH

NVINVdWVO d3ddn

. Range chart of brachiopod species in the Upper Campanian-Lower Maastrichtian Chalk sequence of Norfolk.

JOHANSEN AND SURLYK: CRETACEOUS BRACHIOPODS

827

Boundaries. The member overlies the 'Basal mucronata chalk’, but the boundary to this unit is ill-defined and
possibly mainly based on palaeontological criteria. The latter unit may thus in the future be included in the
Eaton Chalk Member. The member is overlain by the Weybourne Chalk Member, the base of which is taken
at the top of the sponge-bed hardground above Flint Z of Peake and Hancock (1970, fig. 5).

Distribution. The member is presently only recognized in Norfolk (Peake and Hancock 1970, pi. 2), although
probably present at Studland and the Isle of Wight (J. M. Hancock, written communication, June 1986).
Geological age. The member is probably contemporaneous with the lower part of the Ballintoy Chalk Member
of Northern Ireland which corresponds to the lower Upper Campanian basiplana - spiniger Zone of northwest
Germany (Fletcher and Wood 1978; Schulz 1985). This level was chosen as stratotype for the neotype of
Belemnitella mucronata mucronata , and falls within the Hoplitoplacenticeras ‘ vari ’ Zone of the international
ammonite zonal scheme (Christensen et al. 1975).

Samples. One sample was collected from an excavation for a house foundation in Cley.

Weybourne Chalk Member

Name. After the village of Weybourne (TG 110430), situated 700 m south of the coastal outcrops of the
member.

Type section. The member is exposed in an interrupted line of low chalk bluffs on the north coast of Norfolk
at Weybourne Hope. A composite section is shown in Peake and Hancock (1970, fig. 5; Text-fig. 1 herein).

Reference sections. The lower part of the section at Catton Grove (TG 229109) (Peake and Hancock 1970, fig.
6), Keswick (TG 212048), and a former section at Stoke Holy Cross (TG 235016; Text-fig. 1).

Thickness. Approximately 25 m (Peake and Hancock 1970, fig. 3).

Lithology. Chalk with bands of nodular flint, strong and more or less continuous in the lower part. Fossils have
a pinkish tinge in the middle 15 m of this unit.

Boundaries. This member overlies Eaton Chalk, the base taken at the top of the sponge-bed hardground above
Flint Z of Peake and Hancock (1970, fig. 5) in the section at Weybourne Hope. The upper boundary is drawn
at the top of hardground forming the top of the Catton Sponge Bed.

Distribution. The member has only been recognized in Norfolk (Peake and Hancock 1970, pi. 2).

Geological age. Highest part of the lower Upper Campanian, approximately corresponding to the German
roemeri Zone (Schulz 1985).

Samples. Four samples from Catton Grove, one from Keswick, and one from Stoke Holy Cross (Text-figs 1
and 3).

Catton Sponge Bed

Name. After Catton Grove (TG 229109) in the northern outskirts of Norwich.

Type section. Catton Grove (Peake and Hancock 1970, fig. 6; Text-fig. 1 herein).

Reference sections. Stoke Holy Cross. Foreshore at Sheringham, north coast of Norfolk (see Peake and
Hancock 1970, p. 317).

Thickness. From about 0-3 m to a few metres.

Lithology. The member comprises one to three yellow-stained hardgrounds and the intervening soft chalk with
bands of often huge flint nodules. The member contains a characteristic 'hardground preservation fauna’
comprising casts of originally aragonite shelled bivalves, gastropods and ammonites.

Boundaries. The Catton Sponge Bed forms the top unit of the Weybourne Chalk Member. The lower boundary
is placed at the base of the sponge-bed (B in Peake and Hancock 1970, fig. 6) overlying the unlithified
Weybourne Chalk Member, while the upper boundary is placed at the top of the hardground which is overlain
by the unlithified Beeston Chalk Member. This definition conforms with that of Wood (1988).

Distribution. The member is only known from Norfolk.

Geological age. Boundary between lower and upper Upper Campanian, corresponding to the boundary
between the German roemeri and polyplocum Zones (Schulz 1985). A contemporaneous hardground sequence
has been described from the Glenarm Chalk Member of Northern Ireland as the North Antrim Hardground

828

PALAEONTOLOGY, VOLUME 33

(Fletcher and Wood 1978). Many of the Norfolk specimens of B. polyplocum probably come from this member
(J. M. Hancock, written communication, June 1986).

Samples. Two samples from Catton Grove (Text-hgs I and 3).

Beeston Chalk Member

Name. After Beeston Hill, east of Sheringham on the north coast of Norfolk.

Type section. Caistor St Edmunds (TG 238046) (Peake and Hancock 1970, fig. 6; Text-fig. I herein).
Reference section. Stoke Holy Cross (TG 235016).

Thickness. Approximately 25 m (Peake and Hancock 1970, fig. 3).

Lithology. White chalk characterized by irregular bands of large flints many of which appear as flint circles
0-5—3 m or more in diameter when seen on bedding planes. Sometimes two or more circles are concentric, and
there may be a hollow core of the trace fossil Bathichnus paramoudrae at the centre (Peake and Hancock 1970,
pp. 318, 339f; Bromley et al. 1975).

Boundaries. The member overlies the highest hardground of the Catton Sponge Bed (Weybourne Chalk
Member) and is overlain by the poorly exposed Paramoudra Chalk Member. The upper boundary of the
member is defined at the top of a hardground which appears on the shore about 200 metres east of West
Runton (Peake and Hancock 1970, p. 339f).

Distribution. The Beeston Chalk Member has only been recorded from Norfolk (Peake and Hancock 1970, pi.
2).

Geological age. Lower part of the upper Upper Campanian roughly equivalent to the German polyplocum and
langei Zones (Schulz 1985). It correlates well with the Portrush Chalk Member of Northern Ireland (Fletcher
and Wood 1978).

Samples. Two samples from Caistor St Edmunds and one from Stoke Holy Cross (Text-figs 1 and 3).

Paramoudra Chalk Member

Name. After the name applied to the barrel-shaped flints which are characteristic of the unit (Peake and
Hancock 1970, p. 318; Bromley et al. 1975). The name thus does not meet the strict demands of a
lithostratigraphic term, but since it refers to the lithology and as it has no biostratigraphic connotations there
is no strong reason to abandon it.

Reference sections. The member is very poorly exposed and a typfe section cannot be proposed. Whitlingham
(TG 267078), Frettenam (TG 246173) and Postwick (TG 270080) may serve as reference sections. The
reference section at Bramerton is no longer exposed (J. M. Hancock, written communication, June 1986).

Thickness. Approximately 25 m (Peake and Hancock 1970, fig. 3).

Lithology. White chalk with repeated hardgrounds and characterized by the occurrence of ‘paramoudras’. A
‘paramoudra’ is a vertically orientated barrel-shaped or cylindrical flint with a semi-lithified core of chalk
through which passes a vertical dark tube like burrow which may extend upwards and downwards for many
metres. The burrow and the associated ‘paramoudra’ were described in detail by Bromley et al. (1975) who
named the trace fossil Bathichnus paramoudrae .

Boundaries. The base of the member is defined at the top of a hardground, which crops out on the shore about
200 m east of West Runton (Peake and Hancock 1970, p. 339f).

The upper boundary cannot be precisely defined, as younger beds are only exposed in glacially disturbed
masses. The overlying ‘pr e-Porosphaera Beds’ of Wood (1967) comprise relatively hard, sometimes yellowish
chalk largely without ‘paramoudras’. According to J. M. Hancock (pers. comm. 1988) there is a marker
bed - the Overstrand Pyramidata hardground - below the Overstrand Hotel Lower Mass and above
undoubted chalk with paramoudras. The top of this hardground may serve as the upper boundary of the
member.

Distribution. The member is not known outside Norfolk (Peake and Hancock 1970, pi. 2).

Geological age. The member represents the highest Campanian of Norfolk. It is equivalent to the Ballymagarry
Chalk Member of Northern Ireland (Fletcher and Wood 1978), and corresponds to the German grimmensis-
granulosus Zone and perhaps the basal lanceolata Zone of Schulz (1978, 1979, 1985).

Samples. One from Whitlingham at the level of the top flint, probably of uppermost Campanian age.

JOHANSEN AND SURLYK: CRETACEOUS BRACHIOPODS

829

Sidestrand Chalk Member

History. The member encompasses the ‘pr e-Porosphaera Beds’ of Wood (1967) and the ' Porosphaera Beds' of
Brydone (1906, 1908, 1938) (see also Peake and Hancock 1970).

Name. After the village of Sidestrand on the northeast coast of Norfolk.

Type section. The sequence exposed in the glacially induced anticlines at Sidestrand (TG 255404) (Text-fig. 1 ;
see also Peake and Hancock 1970, figs 7 and 8). The eastern (left) anticline especially shows a good section
through the main part of the member, which comprises the chalk containing flints X to P in the composite
section of Peake and Hancock (1970, fig. 7).

Reference sections. Overstrand Hotel Upper Mass (TG 255406).

Thickness. Probably about 10-12 m.

Lithology. White, in places nodular chalk with grey lenticular streaks representing burrows, and about 15
bands of nodular flint. Flint band S is particularly characteristic as it comprises huge cylindrical, black nodules
(up to 30 cm thick and I nr in diameter) with chalk filled holes.

Boundaries. The lower boundary with the Paramoudra Chalk Member is conventionally placed at the top of
the Overstrand Pyramidata hardground. The upper boundary is defined at the top of the hardground labelled
O in Peake and Hancock (1970, fig. 7).

Distribution. The member has only been recognized along the northeast coast of Norfolk (Peake and Hancock
1970).

Geological age. Lowermost Maastrichtian (Wood 1967). Schulz (1978) referred the Overstrand Hotel Upper
Mass to the lanceolata Zone, and the part of the member exposed at Sidestrand to the pseudobtusa and obtusa
Zones.

Samples. One sample from Overstrand Hotel Lower Mass, one sample from Bramerton, six samples from
Overstrand Hotel Upper Mass and six samples from Sidestrand (Text-figs 1 and 3).

Trimingham Sponge Beds Member

History. The member corresponds to the ‘Sponge Beds’ of Brydone (1908).

Name. After the village of Trimingham on the northeast coast of Norfolk.

Type section. Trimingham, where the top is occasionally exposed (Peake and Hancock 1970, fig. 7, beds D to
G). The lower beds can also be seen at Sidestrand.

Thickness. 2-9 m.

Lithology. Lithified chalk with several erosion surfaces with green-coated pebbles. The chalk is yellow in
outcrop, but grey on fresh surfaces. Impressions of large masses of hthistid and hexactinellid sponges are
characteristic. Masses of pyrite are not infrequent (Peake and Hancock 1970). Four bands of nodular flint with
thick white cortices (K, J, I, H) occur in the upper part of the member.

Boundaries. The lower boundary of the Trimingham Sponge Beds Member is placed at the top of the sponge-
bed labelled O in Peake and Hancock (1970, fig. 7). The upper boundary is defined by the top of the thick
‘greasy’ marl band labelled G in Peake and Hancock (1970, fig. 7).

Distribution. Only recorded in the coastal sections of the northeast Norfolk (Peake and Hancock 1970).

Geological age. Top of the lower Lower Maastrichtian. The member probably mainly belongs to the obtusa
Zone of Schulz (1978, 1985).

Samples. Three samples from the top of the member at Little Marl Point.

Little Marl Point Chalk Member

History. The member comprises the so-called ‘White Chalk without Ostrea lunata' and the overlying ‘White
Chalk with Ostrea lunata' (Brydone 1906; Peake and Hancock 1970). Ostrea lunata was originally proposed
as zonal index fossil for the whole of the Maastrichtian exposed at Trimingham. Brydone (1906) demonstrated
that O. lunata did not occur throughout the sequence and suggested that Terebratulina gracilis should take its
place as index fossil. Brydone (1906, 1938) suspected that several horizons with O. lunata were present, but
Peake and Hancock (1970) demonstrated that there is only one horizon with O. lunata.

830

PALAEONTOLOGY, VOLUME 33

O

N

Q

O

CL

O

X

o

<

cr

CD

3 Q.
a i

CL

CL

=>

I

o

ac

i-

ff)

<

<

Q)

■C

o

3

a

/. -/.

<

0

1
CL
<
DC

a

<

cr

c—

c P
O

X

Beacon Hill

Grey Chalk

c. 5 m

Little Marl Point
Chalk

5 m

Trimingham Sp.
Beds 2.9 m

Sidestrand Chalk
12 m

Param_Ch_22_ m

Beeston Chalk
22 m

Catton Sp B
22 m

Weybourne Chalk
22 m

Eaton Chalk
15 m

basal mucronata
beds "

Trimingham 3
2

Little Marl Point

Trimingham 1
Little Marl Point

7

6

5

4

3

2 A
2

1, 1 A

1 B

Sidestrand 4
3 A
3
2

1 A
1

Overstrand Hotel Upper Mass 4
3

Bramerton

Overstrand Hotel Lower Mass

Whitlingham

Caistor 1

— ,i — 3

Stoke Holy Cross 1
Catton 6
— — 5

Stoke Holy Cross 2
Catton 4

— „ — 3A
— — 3

— ii—l

Keswick

Cley

: longicoliis-jasmundi Zone

text-fig. 3. Lithostratigraphy, biostratigraphy, and brachiopod diversity and density of the Norfolk sections.

JOHANSEN AND SURLYK: CRETACEOUS BRACHIOPODS

831

Name. After the bluff Little Marl Point (TG 298380) on the coast east of Trimingham (Peake and Hancock
1970, fig. 9) where the member is well exposed.

Type section. Little Marl Point (Text-fig. 1 ; Peake and Hancock 1970, fig. 7, sequence with beds F to B).
Thickness. About 5-5-6 m (Peake and Hancock 1970).

Lithology. Soft white chalk with four horizons (F, D, C, B) of nodular flint and two thin marl bands (below
F, E). The top contains a horizon densely packed with the dark blue grey shells of the small oyster Ostrea
lunata.

Boundaries. The base of the member is placed at the top of the ‘greasy' marl band labelled G in Peake and
Hancock (1970, fig. 7). The upper boundary is placed at the base of the first soft grey chalk beds of the Beacon
Hill Grey Chalk Member labelled A by Peake and Hancock (1970, fig. 7).

Distribution. The member is only known from northeast Norfolk.

Geological age. Lower part of the upper Lower Maastrichtian. Schulz (1979, 1985) correlated it with the
sumensis Zone of Germany.

Samples. Five samples from Little Marl Point and one from Trimingham mass C of Brydone (1908) (Text-figs
1 and 3).

Beacon Hill Grey Chalk Member

History. The member corresponds to the ‘Grey Beds’ of Brydone (1906, 1908) and Peake and Hancock (1970).
Name. After Beacon Hill (TG 289383) on top of the cliff overlooking the type locality.

Type section. Trimingham Mass C (Peake and Hancock 1970, figs 7 and 9). The sequence is only occasionally
exposed and the fragmentary section figured by Peake and Hancock (1961, fig. 7) was based on a fortuitous
exposure opposite Beacon Hill (Peake and Hancock 1970, p. 339h) (approximately TG 292386). Their data
have been integrated in the synthetic section shown in Text-figures 1 and 3.

Thickness. About 5 m, but the upper limit is not known (Peake and Hancock 1970, p. 339h).

Lithology. Grey chalk with five flint bands. A fawn-coloured calcarenite 01 m thick occurs below the second
flint band from the top (Peake and Hancock 1970, p. 339h).

Boundaries. The lower boundary of the member is defined by the base of the first soft grey chalk bed labelled
A by Peake and Hancock (1970, fig. 7). The upper boundary is not known since the member comprises the
youngest Cretaceous rocks exposed in England.

Distribution. Only known from northeast Norfolk.

Geological age. Lower part of the upper Lower Maastrichtian. The unit belongs to the upper part of the
German sumensis Zone (Schulz 1978, 1985). A highly characteristic form of Echinocorys occurs above the
calcarenite band and can be matched exactly in a section near ‘die Zeven Wegen' in southern Limburg
(Netherlands), where it accompanies B. sumensis (N. B. Peake, written communication, June 1986).

Samples. Two samples from Trimingham Mass C, and one sample from Little Marl Point (Text-figs 1 and 3).

BR ACHIOPOD STRATIGRAPHY

The Upper Campanian and Maastrichtian Chalk of Norfolk is exposed in small quarries, road cuts,
and strongly disturbed coastal cliff sections. The majority of the sections are only a few metres thick.
The succession is furthermore characterized by hardgrounds reflecting periods of non-deposition
and perhaps erosion. It is thus clear that a chart showing the full range of brachiopod species
through the time interval represented by the investigated sequence cannot be constructed. Exposure
of the Campanian-Maastrichtian boundary strata are particularly fragmentary.

The Chalk of Norfolk is thus here dated within the framework of the brachiopod zonation
worked out for the thick continuous Campanian-Maastrichtian boundary section at Kronsmoor,
northwest Germany (Surlyk 1982, 1984). Schulz (1978, 1979, 1985) presented a detailed belemnite
zonation for the Kronsmoor sequence and this zonation is well integrated with the brachiopod
zonation. It is thus possible to use the combined evidence from brachiopods and belemnites in the
correlation of the Norfolk sequence.

832

PALAEONTOLOGY, VOLUME 33

The localities of Text-figure 1 were first arranged in ascending stratigraphic order on the basis of
evidence from macrofossils, notably belemnites and to some extent echinoids (Schulz 1978, 1979,
1985; Peake and Hancock 1970; Wood 1967) and field mapping (Peake and Hancock 1970)
supplemented by scattered information in the literature on the Cretaceous of the British Isles
(Fletcher and Wood 1978).

The brachiopod species were then plotted on this scheme and their vertical distributions
compared with ranges in the Kronsmoor section (Surlyk 1982, 1984) and to more limited
successions in Denmark (Surlyk 1984) and Poland (Bitner and Pisera 1979).

Seven brachiopod zones were recognized in the Kronsmoor section within the interval
represented by the langei through sumensis belemnite zones (Text-fig. 3). The Maastrichtian portion
of the Norfolk and Kronsmoor sequences can be roughly correlated, while the Norfolk Campanian
investigated here reaches lower than the part exposed in Kronsmoor.

It is important to stress that both the brachiopod and belemnite zones are defined by first or last
occurrences of species with Kronsmoor as the key locality. This means, for example, that a slightly
later appearance of a brachiopod index species in Norfolk relative to a belemnite index species is
reflected by an upwards shift of a brachiopod zone boundary in respect to the belemnite zone
boundary. Schemes showing the correlation of the same brachiopod and belemnite zones in
northwest Germany and Norfolk respectively are thus somewhat different. This is shown in Text-
figure 4.

The brachiopod zones are defined and named following the system proposed by Murphy (1977).
The lower boundary of each zone is defined in a stratotype by a biostratigraphic event such as the
first appearance or last occurrence of a species. The upper boundary is defined by the base of the
next, higher zone to avoid gaps or overlap of zones. A binomial nomenclature is used to name the
zone. The name of the species which defines the lower boundary is followed by the name of the
species which defines the lower boundary of the following zone. Thus the longicollis-jasmundi Zone
is from the first occurrence of Terebratulina longicollis to immediately below the first occurrence of
Gisilina jasmundi. The following jasmundi-acutirostris Zone is defined by the first occurrences of
Gisilina jasmundi and Rugia acutirostris respectively.

Most of the brachiopod zones recognized in Kronsmoor are also represented in Norfolk. The
tenuicostata-longicollis Zone includes the localities Cley, Stoke Holy Cross, Catton and Caistor
(Text-fig. 2). This zone can undoubtedly be subdivided, when the brachiopod faunas from
continuous sections through the lower Upper Campanian are investigated.

The longicollis-jasmundi Zone corresponds to the highest Campanian in Kronsmoor and seems
to be represented by the Whitlingham locality.

The basal Maastrichtian jasmundi-acutirostris Zone in Kronsmoor is defined by the first
occurrence of Gisilina jasmundi and Rugia acutirostris. G. jasmundi has its first occurrence in
Kronsmoor sample F10 which was sampled 20 cm below flint band F600. The lowest Belemnella
lanceolata , which defines the base of the Maastrichtian (e.g. Schulz 1978, 1979), was found at the
level of F600. This means that the bases of the lanceolata Zone and the jasmundi-acutirostris Zone
essentially coincide. For convenience, Schulz (1978), however, placed the Campanian-Maastrichtian
boundary precisely at the midpoint of F600. The jasmundi-acutirostris Zone thus seems to have its
base immediately below the Campanian-Maastrichtian boundary (Surlyk 1982, fig. 1), but this only
reflects the greater precision of sample localization in respect to the lithological marker beds. G.
jasmundi and R. acutirostris occur for the first time in the Norfolk material in Overstrand Hotel
Lower Mass which is the next locality stratigraphically above Whitlingham (Text-fig. 2). It is thus
highly probable that the jasmundi-acutirostris Zone is represented by unexposed strata
stratigraphically between Whitlingham and Overstrand Hotel Lower Mass.

The acutirostris-spinosa Zone is defined by the first occurrence of Rugia acutirostris and the last
occurrence of Rugia spinosa. In Norfolk it is represented by the localities Overstrand Hotel Lower
Mass, Bramerton, and Overstrand Hotel Upper Mass samples 1-2A (Text-fig. 2).

The spinosa-subtilis Zone is defined by the last occurrence of Rugia spinosa and the first
occurrence of Terebratulina subtilis. The latter species has a much longer vertical range in Norfolk

JOHANSEN AND SURLYK: CRETACEOUS BRACHIOPODS

833

c

(/) o

CO O

® E

STAGES

5 o

t o CO
Z “ UJ

BRACHIOPOD ZONES

.y o

2 - z

LU X O

o N
UJ UJ
QQ

Kronsmoor

Norfolk

50 -
40 -
30 -

z

<

i—

i

o

CC

UJ

>

upper

B sumensis

pulchetlus - pulchellus
subtil is -pulchellus

pulchellus - pulchellus
spmosa -pulchellus

20-

OC
1 —
CO
<
<

>

o

—I

B obtusa

spmosa- subtil is

10-

<D

$

O

B pseud-
obtusa

acutirostris - spmosa

6

o <13
m <D

acutirostris-spmosa

0 =

QQ

lasmundi- acutirostris

?

co

3

CO

O

2 a

longicollis - / asmundi

longicollis - jasmundi

10 -

co c

C ctj
CD

r~ O)
1°

20 -
30 -

z

CD

5

z

£

<

o

UPPER

upper

5

a

c

-2

tenuicosta ta - longicollis

tenuicostata- longicollis

40 -

s

m

-fig. 4. Scheme showing the correlation of the belemnite and brachiopod zones of Kronsmoor (northwest
Germany) and Norfolk based on Schulz (1978, 1979, 1985), Surlyk (1982, 1984) and this paper.

834

PALAEONTOLOGY, VOLUME 33

than in northwest Europe generally (Text-fig. 2) and its first appearance is thus of no value in
regional correlation. The next higher stratigraphic event used in the northwest European
brachiopod zonation is the incoming of Trigonosemus pulchellus. This species has not been found
in our material but is reported from the Grey Beds (the Beacon Hill Grey Chalk Member of this
paper) by Wood (1967). This is confirmed by the study of the museum collections of the British
Geological Survey by one of us (F.S.). The Beacon Hill Grey Chalk Member includes the samples
Little Marl Point 6 and Trimingham 2 and 3.

The lowest occurrence of T. pulchellus is not known, but is provisionally taken to correspond to
the base of the Beacon Hill Grey Chalk Member. A spinosa-pulchellus Zone can thus be recognized
as including the localities Overstrand Hotel Upper Mass 3 and 4, Sidestrand, Little Marl Point 1-5,
and Trimingham 1 .

It is followed by the pulchellus-pulchellus Zone defined by the first and last appearance of
Trigonosemus pulchellus. The top of the zone may not be exposed in Norfolk.

SYSTEMATIC PALAEONTOLOGY

Repositories. Specimens from Norfolk have MGUH as prefix to their registration numbers. MGUH refers to
the formal numbering system of Geological Museum, the University of Copenhagen, Denmark. Specimens
from Lagerdorf and Kronsmoor are prefixed SPGIH, referring to the formal numbering system of the Institute
of Geology and Palaeontology, the University and Museum of Hamburg, Germany. The abbreviation in the
plate captions WKC refers to specimens collected by Walter Kegel Christensen, Geological Museum, the
University of Copenhagen.

Family lingulidae Menke, 1828
Genus lingula Bruguiere, 1797

Type species. Lingula anatina Lamarck, 1801, p. 141.

Lingula cretaceal Lundgren, 1885

Plate 1, figs 1 and 2

1885 Lingula cretacea Lundgren, p. 20, pi. 1, fig. 1.

1894 Lingula cretacea Lundgren; Posselt, p. 14.

1909 Lingula cretacea Lundgren; Nielsen, p. 17.

1972 Lingula cretacea Lundgren; Surlyk, p. 27, figs 5 and 12.

1982 Lingula cretacea Lundgren; Surlyk, fig. 1.

Material. The shell of this species is very fragile, and when found in the washed residues is highly fragmented.
The Norfolk material consists of one dorsal and one ventral valve, and twenty-five fragments. Significant
measurements have not been obtained. Catalogue numbers MGUH 16841, 16842.

EXPLANATION OF PLATE 1

Figs 1 and 2. Lingula cretacea ? Lundgren, 1885. 1, MGUH 16841, lower Lower Maastrichtian, Overstrand
Hotel Upper Mass 2A, Norfolk; interior-posterior view of ventral valve, x 35. 2, MGUH 16842, lower
Lower Maastrichtian, Little Marl Point IB, Norfolk; exterior of broken valve, x 30.

Figs 3 and 4. Isocrania costata (J. de C. Sowerby, 1823). 3, MGUH 16843, upper Lower Maastrichtian,
Trimingham (coll. WKC, 1971), Norfolk; interior of adult ventral valve, x 12. 4, MGUH 16844, lower
Lower Maastrichtian, Sidestrand 1, Norfolk; exterior of juvenile ventral valve showing stem-shaped
attachment surface, x 27.

Figs 5 and 6. Craniscus sp. MGUH 16845, lower Lower Maastrichtian, Little Marl Point 1A, Norfolk. 5,
interior of fragmented dorsal valve, x 50. 6, oblique lateral view showing median septum, x 50.

Fig. 7. Lacazella ( Bifolium ) wetherelli Morris, 1851, MGUH 16846, lower Lower Maastrichtian, Little Marl
Point IB, Norfolk; interior of adult ventral valve, x 10.

PLATE 1

JOHANSEN and SURLYK, Lingula , Isocrania , Craniscus , Lacazella

836

PALAEONTOLOGY, VOLUME 33

Description. Shell small and thin. Outline irregular elongate, oval, almost parallel-sided, gently biconvex. Beak
of ventral valve contains broad triangular groove for passage of pedicle. Posterior margins of shell are broadly
pointed, anterior margins broadly rounded. Shell surface smooth apart from numerous growth lines. Shell
composed of a brown to whitish material. Interior details not seen.

Remarks. The material is too sparse and fragmented to give a more precise statement of the species
than Lingula cretaceal

Occurrence. Catton 1 and 3A, Bramerton, Overstrand Hotel Upper Mass 2A, Sidestrand 3A, Little Marl Point
IB and 2A and Trimingham 2?

Family craniidae Menke, 1828
Genus crania Retzius, 1781

Type species. Anomia craniolaris Linnaeus, 1758, p. 700.

Crania afif. craniolaris (Linnaeus, 1758)

1972 Crania aff. craniolaris (Linnaeus); Surlyk, p. 27, fig. 5.

Material. Only one fragmented ventral valve, estimated length 4 4 mm.

Description. Outline rounded; valve attached to substrate by entire outer surface. Margin thickened and
pustulose forming rather wide limbus. Anterior and posterior adductor scars close together in posterior part
of valve. Surface of attachment area shows prominent growth lines and very fine radial striation, which may
reflect nature of substrate.

Remarks. The specimen is too fragmentary for a reliable species determination, but it belongs to a
fairly common group of rather small craniaceans, which occur in the chalk of northwest Europe and
which resemble Crania craniolaris (Linnaeus 1758, p. 700) in virtually all features except for their
much smaller size.

Occurrences. Trimingham 1 and elsewhere in the Upper Campanian and Maastrichtian of northwest Europe.

Genus isocrania Jaekel, 1902
Type species. Crania egnabergensis Retzius, 1781, p. 52.

Isocrania costata (J. de C. Sowerby, 1823)

Plate I , figs 3 and 4

1823 Crania costata J. de C. Sowerby, pi. 35, fig. 6.

1973 Isocrania costata (J. de C. Sowerby); Surlyk, figs 1, 2, 5-10, 12; pi. 1, figs 1-1 1, 16-19; pis 3, 4,
6; fig. 12.

1975 Isocrania costata (J. de C. Sowerby); Nestler, p. 42, text-fig. 52.

1979 Isocrania costata (J. de C. Sowerby); Bitner and Pisera, p. 70, fig. 2.

1982 Isocrania costata (J. de C. Sowerby); Surlyk, fig. 1.

1984 Isocrania costata (J. de C. Sowerby); Ernst, p. 65-67, pi. 6, figs 3 and 4; pi. 7, figs 1-6; pi. 8, figs
1, 2, 9-11.

1984 Isocrania costata (J. de C. Sowerby); Surlyk and Johansen, fig. I.

1987a Isocrania costata (J. de C. Sowerby); Johansen, fig. 7.

1988 Isocrania costata (J. de C. Sowerby); Johansen, fig. 2, pi. 4, fig. 4.

Material. Three dorsal, thirteen ventral valves and seven fragments; largest specimens all somewhat
fragmentary. Dorsal valve from Sidestrand 1 is 5-8 mm wide. Catalogue numbers MGUH 16843, 16844.

JOHANSEN AND SURLYK: CRETACEOUS BRACHIOPODS

837

Description. Outline rectangular to circular. Dorsal valve low conical with sculpture of 10-30 primary ribs. A
few additional ribs form by intercalation increasing to a total of about 35. Ribs are strong, prominent,
protruding somewhat beyond the valve margin. Anterior and posterior ribs straight; lateral ribs curve slightly
in posterior direction. Surface of valve between ribs covered with coarse, radially directed spines. Interior
surface has two closely situated anterior adductor scars. Posterior adductors separated by septum-like
extension of limbus and also limited anterio-laterally by such extensions. Limbus wide and finely pustulose.
Ventral valve similar to dorsal valve in outline and sculpture except for flat protegulal node in some cases
showing imprints of tiny attachment substrates. Anterior adductor scars drop-shaped, converging anteriorly
towards midline of valve. Posterior adductor scars oval, of same size and in same position as in dorsal valve.
A pointed septum is situated between anterior adductors.

Remarks. I. costata can be distinguished from earlier forms known from the English Chalk by the
low number of ribs which only protrude slightly beyond the valve margin, and by the subcentral
position of the umbo.

Occurrence. Cley, Catton 5, Overstrand Hotel Lower Mass, Sidestrand 1, 1A and 3 A, Little Marl Point IB,
1, 1A, 2 and 4, and Trimingham 2. In addition the species is recorded from the Upper Campanian and
Maastrichtian of northern Germany and eastern Poland, and from the Maastrichtian and Lower Danian of
Denmark.

Genus craniscus Dali, 1871

Type species. Crania tripartita Munster, 1837, p. 297 .

Craniscus sp.

Plate 1, figs 5 and 6

Material. Three fragmentary dorsal valves. Largest valve 3 6 mm in diameter, and from Bramerton sample.
Catalogue number MGUH 16845.

Description. Outline irregular subrectangular to subrounded. Dorsal valve conical, surface smooth. Anterior
adductor scars in dorsal valve situated on two strong ridges extending posterio-laterally from apex, united with
the median septum or ridge. The three ridges divide the valve into three chambers. Margins of valve not
thickened. Inner surface with densely spaced minute cavities.

Remarks. The material is too poorly preserved to warrant any specific determination; generic
assignment is based on the three characteristic ridges dividing the dorsal valve into three
compartments.

Occurrence. Bramerton and Little Marl Point 1A.

Order rhynchonellida Kuhn, 1949
Family rhynchonellidae Gray, 1848
Unidentified rhynchonellids

Material. Twenty-two dorsal, forty-nine ventral valves and more than two hundred and fifty fragments. The
material is too fragmentary for any further description but belongs to the family Rhynchonellidae, showing the
characteristic glossy, fibrous structure of the thick and impunctate shell. Some fragments probably belong to
Cretirhynchia sp. described below.

Occurrence. Cley, Keswick, Catton 3 A and 4, Stoke Holy Cross 2, Catton 5 and 6, Caistor 1, Whitlingham,
Bramerton, Overstrand Hotel Upper Mass 1A and 2, Sidestrand I A and 2, Little Marl Point IB, 1, 1A and
2, Trimingham I, Little Marl Point 6 and 7, and Trimingham 2 and 3.

838

PALAEONTOLOGY, VOLUME 33

Subfamily cyclothyridinae Makridin, 1955
Genus cretirhynchia Pettitt, 1950

Type species. Terebratula plicatilis J. Sowerby, 1816, p. 37.

Cretirhynchia sp.

Plate 2, figs I and 2

Material. Eight complete shells, nine dorsal and twelve ventral valves and a large number of fragments. All
specimens represent early juvenile stages and no measurements have been obtained. Catalogue numbers
MGUH 16847, 16848.

Description. Shell smooth, pointed oval in juvenile stages. Beak suberect to erect; deltidial plates have a
characteristic growth pattern which results in two laterally directed wing-like doublings of the plates. This
pattern is a result of changes in growth direction of deltidial plates during ontogeny as described by Steinich
(1965). Material does not allow adequate description of brachidium or hinge, although inner socket ridges are
strong, low, short and converge strongly anteriorly. Posteriorly they almost meet. Crura strongly built, long,
converging ventrally.

Remarks. A large number of species of Cretirhynchia are recorded from the Upper Campanian and
Lower Maastrichtian of Norfolk (e.g. Peake and Hancock 1961, 1970; Pettitt 1950, 1953). From the
Upper Campanian C. lentiformis (Woodward) of the ‘ limbata series’ is present below the
Weybourne Chalk Member, and C. arcuata Pettitt of the ‘ plicatilis series’ above this member. C.
norvicensis Pettitt and C. woodwardi (Davidson) appear at the base of the Weybourne Chalk
Member. In the Lower Maastrichtian C. magna Pettitt occurs in the Sidestrand Chalk Member
along with C. aff. arcuata Pettitt and C. retracta (Roemer). C. limbata s.s. (Schlotheim) appears in
the upper Lower Maastrichtian Beacon Hill Grey Chalk Member. Note that Pettitt’s (1950, 1953)
species are in need of revision.

It is not at present possible to recognize juveniles of these species and thus not possible to assign
the present material to any known species.

Occurrence. Catton 1 and 3, Caistor 1 and 3, Overstrand Hotel Lower Mass, Bramerton, Overstrand Hotel
Upper Mass 2A, Sidestrand 1 and 3A, Little Marl Point IB, 2A, 5 and 7, and Trimingham 2 and 3.

Cretirhynchia limbata (Schlotheim, 1813)
Plate 2, figs 3-5

(1798) Terebratulus fossiles Faujas, pi. 26, fig. 4.

1813 Terebratulites limb at us Schlotheim, p. 113.

EXPLANATION OF PLATE 2

Figs 1 and 2. Cretirhynchia sp., MGUH 16847, lower Lower Maastrichtian, Little Marl Point IB, Norfolk. 1,
juvenile specimen in dorsal view, x 20. 2, MGUH 16848, lower Lower Maastrichtian, Little Marl Point 1 A,
Norfolk; small juvenile specimen in dorsal view, x 35.

Figs 3-5. Cretirhynchia limbata (Schlotheim, 1813), MGUH 16849, upper Upper Campanian, Stoke Holy
Cross 1, Norfolk. 3, interior of medium-sized ventral valve, x 10. 4, frontal view showing uniplicate sinus
with low rounded costae at anterior shell margin, x 10. 5, lateral view showing convexity of ventral valve
and curvature of sinus and rostrum, x 8.

Figs 6 and 7. Carneithyris subcardinalis (Sahni, 1925). 6, MGUH 16850, lower Lower Maastrichtian,
Overstrand Hotel Upper Mass 2A, Norfolk ; juvenile specimen in dorsal view, x 10. 7, MGUH 16851, upper
Lower Maastrichtian, Little Marl Point 6, Norfolk; posterior part of adult dorsal valve showing fused
cardinalia, x 10.

PLATE

JOHANSEN and SURLYK, Cretirhynchia , Carneithyris

840

PALAEONTOLOGY, VOLUME 33

1841 Terebratula subplicata Mantell; Roemer, p. 38, fig. 10.

1856 Rhynchonella limbata (Schlotheim) ; Boll, p. 47.

1894 Rhynchonella limbata (Schlotheim); Posselt, p. 27, fig. 1; pi. 2, fig. 16.

1953 Cretirhynchia limbata (Schlotheim); Pettitt, p. 27, fig. 16; pi. 1, fig. 1 a-c; pi. 2, fig. \2a-c; text-
figs 7a-c, 8, 9.

1965 Cretirhynchia limbata (Schlotheim); Steinich, p. 24, fig. 13; pi. 2, fig. 4 a-d.

1972 Cretirhynchia limbata (Schlotheim); Surlyk, p. 24, fig. 17.

Material and occurrence. One bivalved, fragmentary adult specimen found in the Upper Campanian sample
Stoke Holy Cross 1. Dimensions; length, 1 1-4 mm; width, 109 mm; length of dorsal valve, 100 mm; width of
hinge line, 2-5 mm; width of foramen, 0-5 mm. Fragments of this species are probably also present among the
fragments of the Cretirhynchia sp. described elsewhere in this paper. Catalogue number MGUH 16849.

Description. Shell relatively pointed subtriangular in outline. Shell surface smooth except for weak rib pattern
developed along anterior shell margin. Uniplicate sinus with low, rounded costae developed in anterior
commissure. Shell almost plano-convex with dorsal valve strongly convex. Ventral valve fragmentary but
probably only slightly convex. Beak low, pointed incurved; foramen is small, oval to subcircular and limited
by collar-like deltidium. Hinge strong, consists of strongly built, blunt, triangular hinge teeth, inner sockets
which continue directly into the crura. Cardinal process present but material does not allow study of
brachidium.

Remarks. Both external features and nature of the hinge are consistent with descriptions of
Cretirhynchia limbata (Schlotheim) by Steinich (1965) and others. This is the first record of
C. limbata in the Upper Campanian of Norfolk. It is widely distributed in the Maastrichtian of
England, the Netherlands, northern Germany and Denmark.

Family terebratulidae Gray, 1840
Genus carneithyris Sahni, 1925

Type species. Carneithyris subpentagonalis Sahni, 1925, p. 364.

Carneithyris subcarciinalis (Sahni, 1925)

Plate 2, figs 6 and 7

1842 Terebratula carnea J. Sowerby; Hagenow, p. 539, fig. 13.

1894 Terebratula carnea J. Sowerby; Posselt, p. 38, fig. 29.

1909 Terebratula carnea J. Sowerby; Nielsen, p. 163, pi. 2, figs 68-77.

1925 Chatwinothyris subcardinalis Sahni, p. 369, pi. 23, fig. 9; pi. 24, fig. 4, 4 a; pi. 26, fig. 4, 4 a.

1929 Chatwinothyris subcardinalis Sahni; Sahni, p. 40, pi. 5, figs 20-22; pi. 6, figs 10-12; pi. 10, figs

1-4.

1965 Chatwinothyris subcardinalis Sahni; Steinich, p. 37, figs 24-34; pi. 5, figs la— d, 2 a-d, 3, 4; pi. 7,
figs la, b, 2; pi. II, figs 1, 2 a-b, 3.

1972 Carneithyris subcardinalis (Sahni); Surlyk, p. 24, figs 5, 11, 16-18; pi. 5, fig. c.

1975 Carneithyris subcardinalis (Sahni); Asgaard, p. 320, pi. 8, figs 1-4.

1982 Carneithyris subcardinalis (Sahni); Surlyk, fig. 1, pi. 1, fig. a.

1984 Carneithyris subcardinalis (Sahni); Surlyk and Johansen, fig. I.

1987a Carneithyris subcardinalis (Sahni); Johansen, p. 13, pi. 2, fig. 2.

1988 Carneithyris subcardinalis (Sahni); Johansen, fig. 2.

Material. Five complete shells, ten dorsal, twenty-two ventral valves and about ninety fragments. No usable
measurements have been obtained. Catalogue number MGUH 16851.

Description. Adult shell large, relatively thin, with large punctae. The outline changes during ontogeny from
elongated pointed oval to oval - subpentangonal. Biconvex except for earliest juvenile stages which are almost
plano-convex. Shell surface smooth, anterior commissure straight. Hinge line short; beak changes from acute
and suberect to incurved in adult stages. Foramen very small through ontogeny, due to presence of tw’o

JOHANSEN AND SURLYK: CRETACEOUS BRACHIOPODS

841

characteristically large triangular deltidial plates. In large forms foramen is circular and of submesothyridid
to mesothyridid pin-hole type. Hinge strong in adults with dorsal cardinalia so that socket ridges, cardinal
process and crural bases are fused. Cardinal process typically lacking. Brachidium simple, consists, in juveniles,
of short, rather widely separated crura. During growth descending branches develop from dorsal part of crura,
ultimately to fuse and thereby form a transverse bridge.

Remarks. The material is consistent with descriptions of Carneithyris subcardinalis (Sahni) by Sahni
(1925, 1929), Steinich (1965), Asgaard (1975) and Johansen (1987a). A more detailed description of
the ontogenetic stages is given by these authors. The juveniles of C. subcardinalis are distinguished
from other small smooth-shelled related species by the very small foramen limited by two triangular,
deltidial plates.

Occurrence. The species was first described by Sahni (1925) from the Lower Maastrichtian Ostrea lunata Zone
of Trimingham, Norfolk. In the present paper the species is recorded from the following samples: Cation 1,
3, 5? and 6, Stoke Holy Cross 1, Caistor I?, Overstrand Hotel Lower Mass, Bramerton, Overstrand Hotel
Upper Mass 1A, Sidestrand 1 and 3A, Little Marl Point 1 A, 4, 5 and 6, and Trimingham 2 and 3. Carneithyris
subcardinalis is in addition recorded from the Maastrichtian and Upper Campanian of Denmark, northern
Germany and Poland.

Order terebratulida Waagen, 1883
Family terebratulidae Gray, 1840

Undetermined terebratulid

Remarks. Fragments of a punctate, smooth and thin-shelled form occur in the samples Overstrand
Hotel Upper Mass 2, Caistor 3 and 1, all of Early Maastrichtian age. The state of the material does
not allow any generic assignment of the fragments but it is clearly a juvenile terebratulid.

Family cancellothyrididae Thomson, 1926
Genus terebratulina d'Orbigny, 1847

Type species. Anomia craniolaris Linnaeus, 1758, p. 700.

Terebratulina chrysalis (Schlotheim, 1813)

Plate 3, figs 1-5

1813 Terebratulites chrysalis Schlotheim, p. 113.

1894 Terebratulina striata Wahlenberg; Posselt, p. 32, fig. 19; p. 5, footnote 3; p. 10.

1909 Terebratulina striata Wahlenberg; Nielsen, p. 159, fig. 20; pi. 1, figs 28-32; p. 134, fig. 20; p. 137,

fig. 20; pp. 138-139, fig. 20; p. 141, fig. 20; p. 144.

1965 Terebratulina chrysalis (Schlotheim); Steinich, p. 53, figs 44-61 ; pi. 8, fig. 1 a-d\ pi. 9, figs 1-5,
9a, 6, 1 0a, b.

1972 Terebratulina chrysalis (Schlotheim); Surlyk, p. 21, figs 12-18; pi. 4, figs 5-8.

1979 Terebratulina chrysalis (Schlotheim); Bitner and Pisera, p. 73, pi. 3, figs 12-15.

1982 Terebratulina chrysalis (Schlotheim); Surlyk, fig. 1.

1984 Terebratulina chrysalis (Schlotheim); Surlyk and Johansen, fig. 1.

1987a Terebratulina chrysalis (Schlotheim); Johansen, p. 14, fig. 14a-d; pi. 4, figs 1-5.

1988 Terebratulina chrysalis (Schlotheim); Johansen, fig. 2; pi. 2, figs 7-10.

Material. At least forty-five complete shells, one hundred and thirty-five dorsal and seventy ventral valves and
a large number of fragments. The size of fully grown specimens has not been measured or estimated. Catalogue
numbers MGUH 16852-16855.

Description. Shell relatively large. Outline changes during ontogeny from elongated subtriangular to elongated
subpentagonal. The juveniles possess large, well-defined auricles which in adults become small and oblique.
Shell surface with 8 to more than 60 straight ribs, increasing by intercalation. Ribs strong and coarsely

842

PALAEONTOLOGY, VOLUME 33

tuberculate in the juveniles, sculpture characteristically consists of nodes placed on top of ribs, never forming
transverse half rings as is seen in Terebratulina faujasii and T. longicollis. Rib width constant through
ontogeny, interspaces very wide in large specimens. Beak short, suberect; foramen changes from subtriangular
hypothyridid to oval submesothyridid with growth. Hinge relatively weak. Brachidium development as in other
terebratulinids (e.g. Steinich 1965) forming strong crura; short, robust descending branches terminate in closed
brachial ring, rather wide in T. chrysalis. The spicular skeleton well-mineralized; plectolophe.

Remarks. The material is consistent with descriptions of Terebratulina chrysalis by a large number
of authors, e.g. Steinich (1965) and Johansen (1987a). T. longicollis Steinich differs from T. chrysalis
in the lack of intercalated ribs or branching of ribs, the coarse rib sculpture and very elongated
outline. T. faujasii (Roemer) differs from T. chrysalis in possessing single undivided ribs ornamented
by coarse transverse half rings.

Occurrence. Cley, Catton 3 and 4, Stoke Holy Cross 2. Caistor 1 and 3, Whitlingham, Overstrand Hotel Lower
Mass, Bramerton, Overstrand Hotel Upper Mass 1A, 1, 2, 2A, 3 and 4, Sidestrand 1, 1A, 2, 3, 3 A, 4 and 5,
Little Marl Point IB, 1, 1 A, 2A, 3, 4 and 7, and Trimingham 2 and 3. Terebratulina chrysalis is widespread and
very common throughout the Upper Cretaceous and Danian chalks of northwest Europe.

Terebratulina faujasii (Roemer, 1841)

Plate 4, figs 8 and 9

1841 Terebratula faujasii Roemer, p. 40, fig. 24; pi. 7, fig. 8.

1965 Terebratulina faujasii (Roemer); Steinich, p. 72, figs 76-94; pi. 9, figs 6-8; pi. 10, figs 1 a~d, 2-5.

1972 Terebratulina faujasii (Roemer); Surlyk, p. 18, figs 5, 12, 14-16, 18; pi. 2, figs c-g.

1982 Terebratulina faujasii (Roemer); Surlyk, fig. 1, pi. 1, figs c and d.

1984 Terebratulina faujasii (Roemer); Surlyk and Johansen, fig. 1.

1987a Terebratulina faujasii (Roemer); Johansen, p. 15, pi. 7, fig. 3.

1988 Terebratulina faujasii (Roemer); Johansen, fig. 2.

Material. Seventy-five complete shells, 296 dorsal, 268 ventral valves, and approximately 300 fragments.
Largest specimen is a fragmented dorsal valve from Sidestrand 3A: length of dorsal valve, 4-5 mm; width,
50 mm; 12 ribs. Catalogue numbers MGUH 16861, 16860.

Description. Outline changes during ontogeny from broad subtriangular to elongated subtriangular with
maximum width at anterior shell margin. Auricles large. Shell surface with 8-12 straight, coarse, broad single
ribs, with sculpture of very coarse prominent transverse half rings on rib crests. Beak low, suberect to erect;
foramen large, triangular, submeso- to mesothyridid in the adults. Brachidium consists of strong crura, distinct
descending arms and closed brachial ring. Spicular skeleton well-mineralized; plectolophe.

Remarks. Terebratulina faujasii differs from other Terebratulina species in possessing single ribs with
a sculpture of coarse transverse half rings. Juveniles of Gisilina jasmundi Steinich differ from T.

EXPLANATION OF PLATE 3

Figs 1-5. Terebratulina chrysalis (Schlotheim, 1813). I, MGUH 16852, lower Lower Maastrichtian, Sidestrand
1A, Norfolk; exterior of juvenile dorsal valve, x 15. 2 and 3, MGUH 16853, Lower Maastrichtian, Rordal,
Denmark (coll. MBJ, 1981); 2, adult specimen in dorsal view, x 10; 3, lateral view. 4, MGUH 16854,
lower Lower Maastrichtian, Overstrand Hotel Upper Mass 2A, Norfolk; small juvenile specimen in dorsal
view, x35. 5, MGUH 16855, upper Upper Campanian, Caistor I, Norfolk; fragment of large specimen
showing characteristic rib pattern, x 10.

Figs 6-8. Argyrotheca hirundo (Hagcnow, 1842). 6, MGUH 16856, Campanian-Maastrichtian boundary
strata?, Bramerton, Norfolk; exterior of adult ventral valve, x22. 7, MGUH 16857, lowermost
Maastrichtian, Bramerton, Norfolk; interior of adult ventral valve, x 22. 8, SPGIH 3534, upper Lower
Campanian, G 79, Lagerdorf; adult specimen in dorsal view, x 20.

PLATE 3

JOHANSEN and SURLYK, Terebratulina , Argyrotheca

844

PALAEONTOLOGY, VOLUME 33

faujasii in possessing a larger number of weaker sculpted ribs. Juveniles of T. chrysalis (Schlotheim)
differ from T. faujasii in having intercalated ribs at a very early growth stage.

Occurrence. Catton 3A, 4 and 5, Caistor 1, Whitlingham, Bramerton, Overstrand Hotel Lower
Mass, Overstrand Hotel Upper Mass 1, 1A, 2, 2A, 3 and 4, Sidestrand 1, 1A, 2, 3, 3A, 4 and 5, Little
Marl Point 1, 1A, IB, 2, 2A and 3, Trimingham 1, Little Marl Point 6 and 7, and Trimingham 2
and 3. In addition, the species is recorded from the Upper Campanian - Maastrichtian of Denmark,
northern Germany and Poland.

Terebratulina gracilis (Schlotheim, 1813)

Plate 4, figs 1-7

1813 Terebratulites gracilis Schlotheim, p. 113, pi. 3, fig. 3a, b.

1860 Terebratulina gracilis (Schlotheim); Bosquet, p. 390.

1866 Terebratulina gracilis (Schlotheim); Schloenbach, p. 287, pi. 38, figs 18a, b , 19a, b, 20 a, b.

1894 Terebratulina gracilis (Schlotheim); Posselt, p. 33, fig. 20; pi. 3, figs 5-7.

1908 Terebratulina gracilis (Schlotheim); Brydone, p. 403.

1909 Terebratulina gracilis (Schlotheim); Nielsen, p. 161, fig. 22; p. 134, fig. 22; p. 137, fig. 22.

1965 Terebratulina gracilis (Schlotheim); Steinich, p. 81, figs 95-114; pi. 11, fig. la-afi pi. 12, figs 1,

2a, b; pi. 13, figs 1-3.

1972 Terebratulina gracilis (Schlotheim); Surlyk, p. 24, figs 5, 9, 11, 13, 156, 18-22.

1982 Terebratulina gracilis (Schlotheim); Surlyk, fig. 1.

1984 Terebratulina gracilis (Schlotheim); Surlyk and Johansen, fig. 1.

1987a Terebratulina gracilis (Schlotheim); Johansen, p. 16, pi. 7, figs 7-9.

1988 Terebratulina gracilis (Schlotheim); Johansen, fig. 2; pi. 1, fig. 5a, b.

Material. Fifty complete shells, 320 dorsal and 170 ventral valves, and more than 1200 fragments. Largest
specimen from Sidestrand 2 (lower Lower Maastrichtian) representing ventral valve medium sized adult:
length, 3-4 mm; width, 3-2 mm; dorsal valve length, 3-4 mm; width of foramen, 0-3 mm; 30 ribs. Material also
contains fragments from larger specimens, but no usable measurements have been obtained from these.
Catalogue numbers MGUH 16858, 16859; SPGIH 3535-3538.

Description. Shells large, thick. Outline changes from elongated oval to subcircular during growth. Juvenile
shells are biconvex, ventral valve of highest convexity; adults plano-convex. Very early growth stages possess
8-10 straight finely knobbed strong radial ribs forming characteristic fanlike pattern. At shell length of
approximately 1-5 mm lateral deflection of ribs begins; at this stage new ribs form by intercalation. Large
specimens may possess 50-60 ribs. Hinge line very short, oblique; auricles small. Beak short, erect to somewhat
incurved; foramen oval, submesothyridid to permesothyridid and small throughout ontogeny. Hinge strong;
brachidium consists of thin crura and closed brachial ring. Probably plectolophe.

EXPLANATION OF PLATE 4

Figs 1-7. Terebratulina gracilis (Schlotheim, 1813). 1, MGUH 16858, upper Lower Maastrichtian, Little Marl
Point 7, Norfolk; broken adult dorsal valve in exterior view showing lateral bend of ribs, x 15. 2, MGUH
16859, upper Lower Maastrichtian, Little Marl Point 3, Norfolk; large juvenile in dorsal view, x 30. 3 and
4, SPGIH 3535, lower Lower Maastrichtian, G 148-150, Kronsmoor; 3, adult specimen in dorsal view, x 10;
4, lateral view showing reclining profile, x 10. 5, SPGIH 3536, lower Lower Maastrichtian, G 148-150,
Kronsmoor; small juvenile specimen in dorsal view, x 5. 6, SPGIH 3537, lower Lower Maastrichtian,
G 148-150, Kronsmoor; interior of adult dorsal valve showing ring-formed brachidium, x 8. 7, SPGIH 3538,
lower Lower Maastrichtian, G 148-150, Kronsmoor; large juvenile in dorsal view, x 8.

Figs 8 and 9. Terebratulina faujasii (Roemer, 1841). 8, MGLIH 16860, upper Lower Maastrichtian,
Trimingham 1, Norfolk ; adult specimen in dorsal view, x 22. 9, MGUH 16861, lower Lower Maastrichtian,
Overstrand Hotel Upper Mass 2A, Norfolk; adult specimen in dorsal view, x 30.

PLATE 4

JOHANSEN and SURLYK, Terebratulina

846

PALAEONTOLOGY, VOLUME 33

Remarks. The species is consistent with descriptions of Terebratulina gracilis (Schlotheim) by, for
example, Steinich (1965) and Johansen (1987a). Terebratulina gracilis differs from other species of
Terebratulina in the plano-convex shell, short incurved beak, very small foramen and the high
number of deflected ribs. Later forms of T. gracilis (e.g. from the Upper Maastrichtian of Denmark)
possess almost concavo-convex shells and highly incurved beaks; these two features represent
modifications for a free living mode of life on soft chalk substrates and presumably had not yet
reached optimum development in the Upper Campanian-Lower Maastrichtian.

Gisi/ina gisii (Roemer) differs from juvenile T. gracilis in having straight, faintly nodate or almost
smooth ribs.

Throughout the literature T. gracilis has been confused with Terebratulina rigida (J. Sowerby).
Schloenbach (1866) described and compared T. gracilis and T. rigida and from both his descriptions
and illustrations it is clear that T. rigida is smaller and possesses a less incurved beak and much less
reflexed ribs than T. gracilis. T. rigida or a close relative is present in the Norfolk material. It is
described below as T. cf. rigida and is slightly biconvex, possesses finely knobbed ribs which are only
very slightly reflexed in large forms and possesses a distinct hinge line. It has, however, not in all
cases been possible to separate the two species, but the material is dominated by T. gracilis.

Occurrence. Cley?, Keswick, Catton 3A and 4, Stoke Holy Cross 1, Catton 5 and 6, Caistor 3 and 1?,
Whitlingham, Overstrand Hotel Lower Mass, Bramerton?, Overstrand Hotel Upper Mass 1A, 1, 2?, 2A, 3?
and 4, Sidestrand 1 A?, 2, 37, 3A and 47, Little Marl Point IB, 1, 1A, 2, 2A? and 3, Trimingham 1, Little Marl
Point 4, 5, 6 and 7, and Trimingham 2 and 3. Terebratulina gracilis is in addition known from the Upper
Campanian and Maastrichtian of northern Germany, and the Maastrichtian of the Netherlands, Denmark,
southern Sweden and eastern Poland.

Terebratulina longicollis Steinich, 1965
Plate 5, figs 4 and 5

1965 Terebratulina longicollis Steinich, p. 66, figs 62-75, pi. 8, fig. 2 a-d.

1972 Terebratulina longicollis Steinich; Surlyk, p. 18, figs 5, 12-15, 17, 18.

1979 Terebratulina longicollis Steinich; Bitner and Pisera, p. 75, pi. 3, figs 9-1 1.

1982 Terebratulina longicollis Steinich; Surlyk, fig. I, pi. 1, fig. e.

1984 Terebratulina longicollis Steinich; Surlyk and Johansen, tig. I.

1987a Terebratulina longicollis Steinich; Johansen, p. 16, tigs 15a-f; pi. 7, figs 4-6.

1988 Terebratulina longicollis Steinich; Johansen, fig. 2, pi. 2, fig. II.

Material. Seven complete specimens, four dorsal and six ventral valves. Largest specimen is from Little Marl
Point 3: length, 2-8 mm; width, 1-7 mm; length of dorsal valve, 2-2 mm; thickness, 10 mm; 9 ribs. Catalogue
numbers MGUH 16865; SPGIH 3539.

Description. Outline elongate subpentagonal. Auricles small, but clearly defined. Biconvex, ventral valve the
more convex. Commissure is straight. At 0-5 mm width there are 6-8 radial ribs; larger specimens normally
have 8 10 ribs. Width and height of ribs increases during growth; new ribs occasionally form in the auricular

EXPLANATION OF PLATE 5

Figs 1-3. Terebratulina cf. rigida (J. de C. Sowerby, 1821), lower Upper Campanian, Catton 1, Norfolk.
I, MGUH 16862, exterior of medium-sized dorsal valve, x 20. 2, MGUH 16863, medium-sized specimen in
dorsal view, x 30. 3, MGUH 16864, interior of adult ventral valve showing recrystallized spicular skeleton
of a plectolophe, brachidium itself is broken off, x 25.

Figs 4 and 5. Terebratulina longicollis Steinich, 1965. 4, SPGIH 3539, upper Upper Campanian, G 142,
Kronsmoor; adult specimen in dorsal view, x 25. 5, MGUH 16865, upper Lower Maastrichtian, Little Marl
Point 3, Norfolk; interior of adult ventral valve showing large foramen, x 30.

PLATE 5

JOHANSEN and SURLYK, Terebratulina

848

PALAEONTOLOGY, VOLUME 33

regions external to first formed ribs; no intercalated ribs. Ribs straight; sculpture comprises small crescentic
nodes. Rib sculpture strongest on the ventral valve. Rib width is L5 times the width of interspaces; 1-2 growth
lines. Beak suberect, short, the cardinal area very narrow, sharply delimited. Foramen initially hypothyridid,
later mesothyridid. Deltidial plates converge anteriorly during growth resulting in subcircular foramen. Pedicle
collar well developed. Inner socket ridges high, thin, long; outer socket ridges low. Hinge teeth relatively
strong, pointed. Brachidium begins to form at 0 6 mm width. Fully developed terebratulinid brachidium
attained at width about 1-5 mm (see Steinich 1965, tig. 71).

T. longicollis possesses a well-developed spicular skeleton in mantle, anterior body wall, and lophophore
arms. Plectolophe at maturity.

Inner shell surface has smooth topography matching that of exterior. Shell relatively thick.

Remarks. T. longicollis can be distinguished from T. chrysalis (Schlotheim) by its undivided ribs.
Juvenile T. chrysalis , in which intercalation of new ribs has not appeared, are always relatively
wider, and the ribs lack the crescentic sculpture of T. longicollis. The ribs of T. longicollis are
furthermore much more closely-spaced. T. longicollis resembles T. faujasii (Roemer) in that both
have undivided ribs and similar rib sculpture, but they can always be distinguished by the very
elongate outline of T. longicollis.

Occurrence. Whitlingham, Bramerton, Overstrand Hotel Upper Mass I A and 2A, Little Marl Point 2 and 3,
and Trimingham 3. The species is in addition recorded from the Upper Campanian and Maastrichtian of
northern Germany and eastern Poland, and from the Maastrichtian and possibly also Lower Danian of
Denmark.

Terebratulina cf. rigida (J. Sowerby, 1829)

Plate 5, figs 1-3

cf. 1829 Terebratula rigida J. Sowerby, p. 69, pi. 536, fig. 2.

cf. 1866 Terebratulina rigida (J. Sowerby); Schloenbach, p. 287, pi. 38, figs 10-17.

Material. Eleven complete shells, twelve dorsal, thirty-two ventral valves, and a large number of fragments.
Largest complete specimen is juvenile from Catton 1 : length, 3-8 mm; width, 31 mm; length of dorsal valve,
2-8 mm; thickness, 1-3 mm; width at hinge line, 10 mm; width of foramen, 0-5 mm; 21 ribs. Catalogue
numbers MGUH 16862-16864.

Description. Shell relatively large, biconvex, outline subtriangular, rather thick and punctuate. Auricles small
but distinct; hinge line straight. At shell width 20 mm, shell bears 10 ribs, at 2-5 mm, 18-20 ribs and at 3 0 mm
23-25 ribs. New ribs arise by intercalation; ribs straight to slightly reflexed in larger specimens. Rib sculpture
consists of closely-spaced transverse half rings, most prominent on ventral valve. Beak erect, blunt to slightly
incurved; foramen rather large, broad, triangular to elongate oval. Deltidial plates high, triangular,
converging; internal morphology otherwise poorly known. Plate 5, fig. 3 shows a recrystallized spicular
skeleton from a plectolophous lophophore.

Remarks. T. gracilis and T. rigida have been confused since their descriptions in 1813 and 1821
respectively (Schloenbach 1866; J. Sowerby 1829). Steinich (1965) and Johansen (1987a) both
mention this problem. T. rigida is in general smaller, possesses a larger foramen, a less incurved beak
and ribs much less reflexed than in T. gracilis. The Norfolk specimens correspond to T. rigida , but
as the material does not permit detailed observations of internal morphology, we refer to the
material as T. cf. rigida.

T. chrysalis differs from T. cf. rigida in possessing a more elongated outline, and more ribs, which
retain their width during growth.

Occurrence. Catton 1, 3, 4 and 6, Sidestrand 1A? and 2?, Little Marl Point IB?, and Trimingham 2.

Terebratulina subtilis Steinich, 1965

Plate 6, figs I and 2

JOHANSEN AND SURLYK: CRETACEOUS BRACHIOPODS

849

1965 Terebratulina subtilis Steinich, p. 93, figs 1 15-129; pi. 9, fig. 2 a-d.

19706 Terebratulina subtilis Steinich; Surlyk, p. 13, fig. 3.

1972 Terebratulina subtilis Steinich; Surlyk, p. 18, fig. 2.

1982 Terebratulina subtilis Steinich; Surlyk, fig. 1.

1988 Terebratulina subtilis Steinich; Johansen, fig. 2.

Material. Eleven complete shells, twenty-six dorsal, twenty-five ventral valves, and thirty-five fragments. The
largest specimen is from Little Marl Point 6: length, 2-5 mm; width, 20 mm; thickness, 0-8 mm; length of
dorsal valve, 2 1 mm; 48 ribs. Catalogue numbers MGUH 16866, 16867.

Description. Shell thin with smooth interior. Outline very elongate subpentagonal to subtriangular in juvenile
stages, changing into more pointed oval with growth. Biconvex to almost plano-convex; dorsal valve normally
convex umbonally, otherwise flat. Auricles poorly-defined; commissure straight. Ornament consists of
numerous delicate radial ribs. At a width of c. 05 mm, 3-10 relatively wide ribs appear. New ribs arise by
intercalation and are always narrower than earlier formed ribs. Specimens 2 mm or more wide have 40-50 ribs
of very low amplitude bearing a characteristic sculpture of small granules often somewhat coarser on ventral
than on dorsal valve, where they may be very difficult to detect. Growth lines indistinct or absent. Beak short,
suberect, cardinal area narrow, but well-defined. Foramen limited by two narrow triangular deltidial plates;
develops from hypothyridid to submesothyridid during growth. Pedicle collar wide, well-developed. Hinge
rather weakly — developed ; inner socket ridges thin. Brachidium develops at a width 04 mm; normal, rather
delicate terebratulinid brachidium present at width c. 13 mm.

Spicular skeleton well-developed, normally strongly recrystallized. Plectolophe. Mantle spicules not found.

Remarks. Terebratulina subtilis shows great resemblance to Rugia tenuicostata Steinich, and the
juveniles of the two species are particularly difficult to separate. T. subtilis, however, always has a
larger number of ribs, and a less pointed beak. The most important difference is in the rib
morphology where R. tenuicostata has rather high, relatively widely-spaced serrate ribs, while T.
subtilis has flat densely-spaced ribs with only small rather rounded tubercles.

Occurrence. Keswick?, Catton 1, 3 and 5, Overstrand Hotel Lower Mass, Overstrand Hotel Upper Mass 1A,
I and 2, Sidestrand 2 and 3A, Little Marl Point 2A, Trimingham 1, Little Marl Point 6 and 7, and Trimingham
3. In addition, the species is known from the Lower Maastrichtian of northwest Germany, East Germany and
Denmark.

Genus gisilina Steinich, 1963
Type species. Terebratula gisii Roemer, 1841, p. 40, fig. 23.

Gisilina gisii (Roemer, 1841)

Plate 6, figs 3-5

1841 Terebratula gisii Roemer, p. 40, fig. 23.

19636 Gisilina gisii (Roemer); Steinich, p. 732, figs 1-8.

1965 Gisilina gisii (Roemer); Steinich, p. 100, figs 130-148; pi. 14, figs 2 and 3 a~d\ pi. 15, figs 1-7.

1972 Gisilina gisii (Roemer); Surlyk, p. 18, figs 2, 12, 20-22.

1982 Gisilina gisii (Roemer); Surlyk, fig. 1, pi. 2, fig. d.

1988 Gisilina gisii (Roemer); Johansen, fig. 2.

Material. Three complete shells, eleven dorsal, six ventral valves, and one fragment. Largest specimen is a
dorsal valve from Caistor 1: length of dorsal valve, 31 mm; width, 3-2 mm; 12 ribs. Catalogue numbers
MGUH 16868; SPGIH 3540.

Description. Outline rounded subtriangular; auricles small, indistinct. Shell rather thick, strongly biconvex,
ventral valve having highest convexity. Shell surface with 7-12 straight, coarse, broad, smooth or faintly
nodate ribs. New ribs arise by intercalation. Beak changes from acute and suberect to slightly incurved or erect
during ontogeny. Foramen large, initially hypothyridid, thereafter submesothyridid and in some cases

850

PALAEONTOLOGY, VOLUME 33

mesothyridid in large forms, and thus partly absorbed in ventral valve. Inner socket ridges short and teeth
strong. Little can be said of the brachidium in present material : crura very short. Probably a plectolophe.

Remarks. Large specimens of G. gisii are easily distinguished from G.jasmundi Steinich because they
are more strongly biconvex, have smooth to faintly nodate ribs, and a mesothyridid foramen. Small
specimens are very difficult to distinguish from juvenile Terebratulina chrysalis (Schlotheim) and G.
jasmundi. G. gisii has fewer ribs than T. chrysalis and is more biconvex. G. jasmundi is less biconvex
than G. gisii and has distinctly coarser sculpted ribs. Ribs are single in G. jasmundi.

Occurrence. Caistor 1, Overstrand Hotel Lower Mass, Overstrand Hotel Upper Mass 1A, 2 and 3, Little Marl
Point 7, and Trimingham 2. The species is also recorded from the Upper Campanian and Maastrichtian of
northwest Germany, and the Maastrichtian of East Germany, southern Sweden and Denmark.

Gisilina jasmundi Steinich, 1965
Plate 6, fig. 6

1965 Gisilina jasmundi Steinich, p. 110, figs 149-161; pi. 16, figs 1 a-d and 2.

1972 Gisilina jasmundi Steinich; Surlyk, p. 18, figs 2, 5, 13, 15a, 16, 17.

1982 Gisilina jasmundi Steinich; Surlyk, fig. 1, pi. 2, figs a-c.

1984 Gisilina jasmundi Steinich; Surlyk and Johansen, fig. 1.

1987a Gisilina jasmundi Steinich; Johansen, p. 21, pi. 8, fig. 3.

1988 Gisilina jasmundi Steinich; Johansen, fig. 2.

Material. At least one complete shell, twelve dorsal, two ventral valves, and four fragments. Largest specimen
represented by ventral valve from Sidestrand 1 : length, 3-8 mm; width, 30 mm; length of dorsal valve, 30 mm;
width of hinge line, 24 mm; 13 ribs. Catalogue number MGUH 16869.

Description. Norfolk specimens have subtriangular to subpentangonal outline and small but distinct auricles.
Beak suberect, foramen large, submesothyridid. Hinge line short, straight. Characteristic for this species are
12-18 straight, radial, and coarsely nodate single ribs, and a relatively flat shell. Rib sculpture varies between
coarse knobs on top of ribs and prominent transverse half rings arranged across the ribs. Inner socket ridges
high, teeth short. Crura short, other details of brachidium not seen. Probably plectolophe.

Remarks. G. jasmundi is thoroughly described by Steinich (1965). Large specimens of G. jasmundi
recall Terebratulina faujasii (Roerner), but differ in the larger number of ribs, flatter shell, and finer

EXPLANATION OF PLATE 6

Figs 1 and 2. Terebratulina subtilis Steinich, 1965. 1, MGUH 16866, lower Upper Campanian, Keswick,
Norfolk; dorsal view of adult specimen, x 30. 2, MGUH 16867, lowermost Maastrichtian, Overstrand Hotel
Upper Mass 1A, Norfolk; dorsal view of juvenile specimen, x 30.

Figs 3-5. Gisilina gisii (Roerner, 1841). 3 and 4, SPGIH 3540, lower Upper Campanian, G 140, Lagerdorf; 3,
adult specimen in dorsal view, x 18. 4, lateral view showing strongly biconvex shell, x 18. 5, MGUH 16868,
upper Lower Maastrichtian, Trimingham 2, Norfolk; medium-sized adult ventral valve in exterior view,
x 25.

Fig. 6. Gisilina jasmundi Steinich, 1965. MGUH 16869, uppermost Campanian, Overstrand Hotel Lower Mass,
Norfolk; medium-sized adult ventral valve in exterior view, x25.

Fig. 7. Rugia tenuicostata Steinich. 1963. MGUH 16870, lower Upper Campanian, Keswick, Norfolk; adult
specimen in dorsal view, x 25.

Figs 8-10. Rugia acutirostris Steinich, 1965, lower Lower Maastrichtian, Overstrand Hotel Upper Mass 2A,
Norfolk. 8, MGUH 16871, juvenile in dorsal view, x 27. 9, MGUH 18226, Norfolk; small juvenile in dorsal
view, x 20. 10, MGUH 18227, adult specimen in dorsal view, x25.

PLATE 6

mm

JOHANSEN and SURLYK, Terebratulina , Gisilina , /?wg/a

852

PALAEONTOLOGY, VOLUME 33

sculptured rib surface. Small specimens of G. jasmundi with more faintly sculptured ribs are
extremely difficult to distinguish from juveniles of G. gisii (Roemer) and T. chrysalis (Schlotheim).
Juvenile G. jasmundi possess coarse sculptured, single ribs; T. chrysalis is elongate and has
intercalated ribs.

Occurrence. Overstrand Hotel Lower Mass, Overstrand Hotel Upper Mass 1A and 3?, and Sidestrand 1. In
addition, the species is recorded from the Maastrichtian of northwest Germany, East Germany and Denmark.

Genus rugia Steinich, 1963

Type species. Rugia tenuicostata Steinich, 19636, p. 735, fig. 3.

Rugia tenuicostata Steinich, 1963
Plate 6, fig. 7

19636 Rugia tenuicostata Steinich, p. 738, figs 6, 7, 8 a-d.

1965 Rugia tenuicostata Steinich; Steinich, p. 116, figs 162-174; pi. 11, 3 a-d and 4.

1970 Rugia tenuicostata Steinich; Surlyk, figs 2 and 3.

1972 Rugia tenuicostata Steinich; Surlyk, p. 8, figs 2, 5, 12, 16.

1979 Rugia tenuicostata Steinich; Bitner and Pisera, p. 77, pi. 4, figs 3-6; pi. 5, fig. 3.

1982 Rugia tenuicostata Steinich; Surlyk, fig. 1, pi. 1, figs f-h.

1984 Rugia tenuicostata Steinich; Surlyk and Johansen, fig. 1.

1987a Rugia tenuicostata Steinich; Johansen, p. 22, figs 18f, g, pi. 7, fig. 2.

19876 Rugia tenuicostata Steinich; Johansen, p. 159, figs 3, 14a-f; pi. 6, figs 1-5.

1988 Rugia tenuicostata Steinich; Johansen, fig. 2, pi. 2, fig. 1.

Material. Seventeen complete shells, eighty-one dorsal, sixty-nine ventral valves, and one hundred fragments.
Largest specimen from Keswick: length, 2-48 mm; width, 204 mm; length of dorsal valve, 212 mm; thickness,
0-96 mm; 41 ribs. Catalogue number MGUH 16870.

Description. Outline very elongate subtriangular in early growth stages, changing through ontogeny to more
pointed oval. Shell is biconvex; dorsal valve always less convex than ventral valve. Inner shell smooth. Some
juvenile specimens may have completely flat dorsal valve. Auricles indistinct. Commissure straight. Shell
ornamentation of straight radial ribs with serrate sculpture. Initial ribs are formed at width 0-4 mm, number
increases by intercalation from about 6-7 to 28-30. In rare cases large forms have 40 ribs, but last 10 ribs are
very narrow low intercalatories close to shell margin. Ribs are of uniform height and width through ontogeny.
Rib density varies considerably, decreasing markedly during ontogeny. Rib sculpture highly characteristic,
comprising small triangular tubercles aligned on top of ribs, giving saw-tooth appearance. New intercalated
ribs often form an acute angle to adjacent older ribs but change to more parallel alignment during growth.
Growth lines uncommon, beak short, suberect, pointed in juvenile stages; more rounded in adult forms.
Cardinal area narrow, well-defined. Foramen delineated by two narrow deltidial plates, which grow in ventro-
median direction in the earliest stages; at shell width of 0-6 mm they grow outwards in ventro-lateral direction;
at width of 0-7 mm growth direction changes to median. Pedicle collar increases relative length in growth to half
length of beak. Foramen triangular, hypothyridid. Hinge close to that of Terebratulina. Brachidium consists
of long, thin crura, converging strongly anteriorly. Crural processes short, spoon-shaped, not fused during
growth. Descending branches thin, fusing under acute angle. Mantle and arms heavily spiculate; lophophore
plectolophe.

Remarks. R. tenuicostata can be distinguished from R. acutirostris Steinich by the less acute beak,
lower rib density, and characteristic rib sculpture. Juveniles are difficult to separate from
Terebratulina subti/is Steinich, other than by the rib characters; T. subtilis has numerous weakly
sculptured very low ribs.

Occurrence. Cley, Keswick, Catton 1?, 3, 3 A and 4, Stoke Holy Cross 1, Catton 5 and 6, Caistor 3 and 1,
Whitlingham, Overstrand Hotel Upper Mass 1 A, Little Marl Point 1 ?, 1 A and 5?. Also known from the Upper
Campanian to Maastrichtian of northwest and East Germany, Denmark and Poland.

JOHANSEN AND SU R LY K : CRETACEOUS BRACHIOPODS

853

Rugia acutirostris Steinich, 1965

Plate 6, figs 8-10

1965 Rugia acutirostris Steinich, p. 122, figs 175-178; pi. 14, I a~d.

1970 Rugia acutirostris Steinich; Surlyk, figs 2 and 3.

1972 Rugia acutirostris Steinich; Surlyk, p. 18, figs 2, 5, 14, 18.

1979 Rugia acutirostris Steinich; Bitner and Pisera, p. 77, pi. 3, figs 1-3.

1982 Rugia acutirostris Steinich; Surlyk, fig. 1, pi. I, figs i and j.

1984 Rugia acutirostris Steinich; Surlyk and Johansen, fig. I.

1987a Rugia acutirostris Steinich; Johansen, p. 22, pi. 7, fig. 1a, b.

19876 Rugia acutirostris Steinich; Johansen, p. 163, figs 3, 15a-f; pi. 5, fig. 4.

1988 Rugia acutirostris Steinich; Johansen, fig. 3.

Material. Thirty-eight complete shells, sixty dorsal, thirty-seven ventral valves, and sixty fragments. Largest
specimen is from Overstrand Hotel Upper Mass 1A: length, 2 0 mm; length of dorsal valve, 1-7 mm; width,
1-5 mm; thickness, 0-8 mm; 26 ribs. Catalogue numbers MGUH 16871, 18226, 18227.

Description. Outline changes from elongate subpentagonal to pointed oval during ontogeny. Auricles small,
blunt. Shell biconvex, the valves of equal convexity. 6-8 radial ribs are present at shell width of 0-2-0- 3 mm.
Ribs increase in number by intercalation to about 40. Rib width constant during growth. Rib sculpture consists
of small, scale-like tubercles; in some specimens ribs are smooth. Lateral ribs curve slightly outwards. Beak
very acute, suberect. Cardinal area wide, clearly delineated, triangular, slightly concave. Foramen very small,
elongate triangular, hypothyridid, limited by two narrow slightly elevated deltidial plates, often granular.
Pedicle collar very long. Hinge and brachidium development as in R. tenuicostata. Inner surface of shell is
smooth. The mantle and arms heavily spiculate. Plectolophe.

Remarks. R. acutirostris is characterized by its pointed oval outline, its very acute beak, and small
foramen.

Occurrence. Overstrand Hotel Lower Mass, Bramerton, Overstrand Hotel Upper Mass 1A, 1, 2A, 3 and 4,
Sidestrand 1, 1A, 3A and 4, and Little Marl point 5 and 7. Also known from both the Lower and Upper
Maastrichtian of northwest and East Germany, Denmark and Poland.

Rugia tegulata Surlyk, 1970

Plate 7, figs 1 and 2

1970a Rugia tegulata Surlyk, p. 152, pi. 1, figs a-e , fig. 4a-d.

1972 Rugia tegulata Surlyk; Surlyk, p. 18, fig. 5.

1984 Rugia tegulata Surlyk; Surlyk, fig. 2.

1988 Rugia tegulata Surlyk; Johansen, fig. 2.

Material. Five dorsal, three ventral valves, and four fragments. Largest specimen a dorsal valve: length,
1-6 mm; width, T9 mm; 21 ribs. Catalogue numbers MGUH 18228-9.

Description. Outline subtriangular to pointed oval; length to width ratio decreases rapidly during growth.
Auricles blunt; shell slightly biconvex, dorsal valve being a little less convex than ventral valve. Ornament very
characteristic: radial ribs with imbricate, plate-like scales. Rib number increases rapidly by intercalation from
about 5 at width 0-5 mm to about 20 in largest specimens. Beak suberect; cardinal area small. Foramen is
triangular, hypothyridid, limited by two deltidial plates which often take the form of irregular ridges with
constriction at middle of each plate. Inner socket ridges short, strong. Crurae long, thin; full brachidium is not
preserved. Shell is rather thick; gentle undulations on inner surface correspond to the ribs on outer.

Remarks. Although limited the material falls well within the range of the type material from the
lower Upper Maastrichtian of Denmark. The characteristic rib sculpture distinguishes it from
other Chalk brachiopods.

854

PALAEONTOLOGY, VOLUME 33

Occurrence. All specimens are from Little Marl Point 1A. The species has previously been described only from
the lower Upper Maastrichtian of Denmark.

Rugia spinosa Surlyk, 1970

1970a Rugia spinosa Surlyk, p. 157, fig. 3 d-f; pi. 2. figs a-d\ table 1.

1972 Rugia spinosa Surlyk; Surlyk, p. 18, figs 2, 5.

1982 Rugia spinosa Surlyk; Surlyk, p. 262, fig. 1.

1984 Rugia spinosa Surlyk; Surlyk, p. 219, figs 2 and 3.

1 987/? Rugia spinosa Surlyk; Johansen, p. 157, figs 3, 11a-f, 12a, b, 13.

1988 Rugia spinosa Surlyk; Johansen, fig. 2.

Material. One fragmentary valve and two fragments.

Description. Shell small and thin, outline oval to pointed oval. Auricles are blunt. Shell strongly biconvex,
anterior commissure straight. Shell sculpture comprises long, pointed spines of variable form and size in many
cases extending from irregular, poorly-developed radial ribs. They tend to be at right angles to the shell surface
on the posterior part of the shell, while larger and directed forwards in the anterior part of shell. Spines may
be arranged in rows, on a more-or-less distinctly developed rib. Ribs few in number, irregularly distributed,
often best seen on inside of shell as strong, well-defined smooth ribs. Beak very short, pointed and suberect.
Cardinal area and deltidial plates narrow. Foramen very small, submesothyridid. Inner socket ridges short with
anterior lateral bend, close together. Teeth short, pointed. Crura thin, cylindrical, originating from position
anterior to inner socket ridges. Inner surface of the shell smooth apart from a few irregularly distributed ribs.

Remarks. R. spinosa is characterized by its rib sculpture; it differs from R. spinicostata Johansen and
Surlyk in having only a few irregularly distributed ribs. Bitner and Pisera (1979) emended the
original diagnosis of R. spinosa to include forms with regularly distributed ribs, but study of
abundant stratigraphically well-dated material of this form from eastern England, northwest
Germany and Denmark shows that lower Lower Maastrichtian forms are R. spinosa , while Upper
Campanian forms belong to R. spinicostata , described below (Surlyk 1982; Johansen 1 987/?). Bitner
and Pisera (1979) assigned the Mielnik chalk section, eastern Poland to the Lower Maastrichtian
because of the presence of R. spinosa in the lower part of the section, but the Polish material clearly
belongs to R. spinicostata. The lower part of the Mielnik section is thus most probably of Late
Campanian age.

Occurrence. Overstrand Hotel Upper Mass 1A and 2A. Also known from the lower Lower Maastrichtian of
Denmark and northwest and East Germany.

EXPLANATION OF PLATE 7

Figs 1 and 2. Rugia tegulata Surlyk, 1970. 1, MGUH 18228, lower Lower Maastrichtian, Little Marl Point 1 A,
Norfolk; exterior of broken ventral valve, x 30. 2, MGUH 18229, lower Lower Maastrichtian, Little Marl
Point 1A, Norfolk; exterior of dorsal valve, x 20.

Figs 3-10. Rugia spinicostata Johansen and Surlyk in Johansen, 19876. 3, MGUH 18230, upper Upper
Campanian, Catton 3A, Norfolk; interior of broken adult dorsal valve, x 25. 4, MGUH 18231, upper Upper
Campanian, Cley, Norfolk; ventral valve in exterior view, x 25. 5, SPGIH 3541, upper Lower Campanian,
G 81, Lagerdorf; adult specimen in dorsal view, x 30. 6 and 7, SPGIH 3542A, lower Upper Campanian, G
103, Lagerdorf; 6, interior of adult dorsal valve showing almost complete spoon-shaped brachidiunr and
recrystallized plectolophous lophophore, x 30; 7, oblique lateral view showing long recrystallized spicular
skeleton of lophophore filaments, x 30. 8, SPGIH 3542B, interior of ventral valve corresponding to dorsal
valve shown on figs 6 and 7, x 30. 9, SPGIH 3543, upper Lower Campanian, G 78, Lagerdorf ; juvenile in
dorsal view, x 20. 10, SPGIH 3544, lower Upper Campanian, G 103, Lagerdorf; dorsal view, x 30.
dorsal view, x 30.

PLATE 7

JOHANSEN and SURLYK, Rugia

856

PALAEONTOLOGY, VOLUME 33

Rugia spinicostata Johansen and Surlyk, 1987 in Johansen, 1987/?

Plate 7, figs 3-10

1979 Rugia spinosa Surlyk; Bitner and Pisera, p. 78, pi. 3, figs 4-8.

1982 Rugia spinosa s.l. Surlyk; Surlyk, p. 262, fig. 1.

1984 Rugia spinosa s.l. Surlyk; Surlyk, p. 219, fig. 2.

19876 Rugia spinicostata Johansen and Surlyk; Johansen, p 152, figs 3 and 1 1 a— F ; pi. 5, figs 1-3.

1988 Rugia spinicostata Johansen and Surlyk; Johansen, fig. 2, pi. 2, fig. 2.

Material. Five dorsal, six ventral valves, and three fragments. Largest Norfolk specimen a ventral valve from
the upper Upper Campanian of Cley: length, 1-6 mm; width, 1-4 mm; length of dorsal valve, L3 mm; width
of hinge line, 0-8 mm; foramen diameter, 02 mm; 22 ribs. Catalogue numbers MGUH 18230, 18231 ; SPGIH
3541, 3542A and B, 3543, 3544.

Description. Shell small, pointed oval to pointed subtriangular in outline. Strongly biconvex, anterior
commissure straight. Auricles are blunt; stumpy. Surface sculpture of long, pointed spines extending from
regularly distributed radial ribs, slightly reflexed in larger forms. Spines are cylindrical, pointed solid, variable
in form and size. Spines arranged in rows on distinct ribs; in most cases spines are directed anteriorly. About
16 ribs present at shell width 10 mm; about 26 at 2 0 mm. Inside of the shell typically with very prominent,
strong, smooth straight median rib and several straight and smooth ribs lateral to this. Beak short, suberect;
cardinal area narrow. Deltidial plates narrow, triangular, foramen small, subtriangular, submesothyridid.

Structure of hinge and brachidium similar to R. acutirostris and R. tenuicostata. Inner socket ridges strong,
short and close together, forming an anterio-lateral bend. Cardinal process small, bulbous, protrudes well
beyond posterior shell margin together with inner socket ridges. Teeth short, pointed, triangular. Crura long,
thin, developing from anterior edge of inner socket ridges. Descending branches converge rapidly and fuse in
strong bridge. Ascending branches short. Shell relatively thin. Plectolophe when adult.

Remarks. R. spinicostata differ from other Rugia species in possessing a surface sculpture of long,
pointed spines extending from regularly distributed and distinct radial ribs; it was first recognized
and described as a distinct form Rugia spinosa s.l. by Surlyk (1982).

The surface sculpture of Rugia spinosa Surlyk also consists of long, pointed spines, but in this
species spines often extend from poorly-developed, irregular, radial ribs, which are fewer in number
than in R. spinicostata. Ribs of R. spinosa often are recognized only from the inside of the shell as
strong, smooth ridges (Surlyk 1970n).

R. spinosa is generally smaller than R. spinicostata and possesses a shorter and more pointed beak,
and very close-lying, anterio-laterally bent inner socket ridges (Johansen 19876).

Occurrence. Cley, Keswick, and Catton 1, 3, 3A and 4. Also known from the Lower to the upper Upper
Campanian of northwest Germany, and from possible Upper Campanian strata of eastern Poland (see
discussion under Rugia spinosa).

Genus dracius Steinich, 1967

Type species. Dracius carnifex Steinich, 1967, p. 1146, pi. 1, fig. I.

Dracius carnifex Steinich, 1967

Plate 9, figs 5-8

1967 Dracius carnifex Steinich, p. 1145, figs 1-7; pi. 1, figs I^J.

1972 Dracius carnifex Steinich; Surlyk, figs 5, 16, 18.

1982 Dracius carnifex Steinich; Surlyk, fig. 1, pi. 2, figs e and /.

1988 Dracius carnifex Steinich; Johansen, fig. 2.

Material. One ventral valve and two fragments. Catalogue numbers MGUH 18241, 18242; SPGIH 3548, 3849.

JOHANSEN AND SU R L Y K : C RET ACEOUS BRACHIOPODS

857

Remarks. Limited material prevents detailed description, but characteristic ribbing of the ventral
valve points to D. carnifex. Ribs at a shell length c. 05 mm are seen as 7-8 highly irregular rows
of small spines. Ribs become stronger during growth and separation increases markedly. Ribs
ornamented by small often spiny tubercles. Interspaces bear a dense cover of identical tubercles.

Occurrence. Whitlingham (?) and Overstrand Hotel Lower Mass. Also known from the Maastrictian of East
and northwest Germany and Denmark.

Family megathyrididae Dali, 1870
Genus argyrotheca Dali, 1900

Type species. Terebratula cuneata Risso, 1826, p. 388.

Argyrotheca bronnii (Roemer, 1841)

Plate 8, figs 1-4, 7

1841 Terebratula bronnii v. Hagenow; Roemer, p. 41, fig. 31.

1965 Argyrotheca bronnii (Roemer); Steinich, p. 124, figs 179-196; pi. 17, figs 1 a-d and 2 a-d.

1972 Argyrotheca bronnii (Roemer); Surlyk, p. 20, figs 5, 14—16; pi. 4 g.

1979 Argyrotheca bronnii (Roemer); Bitner and Pisera, p. 78, pi. 4, figs 1 and 2.

1982 Argyrotheca bronnii (Roemer); Surlyk, fig. 1, pi. 3, figs / and/

1984 Argyrotheca bronnii (Roemer); Surlyk and Johansen, fig. 1.

1987a Argyrotheca bronnii (Roemer); Johansen, p. 27.

1988 Argyrotheca bronnii (Roemer); Johansen, fig. 2.

Material. Three complete shells, seventeen dorsal and twenty ventral valves, and twenty-seven fragments.
Largest specimen is from Sidestrand 1 : length, 2-52 mm; width, 3 32 mm; length of dorsal valve, 208 mm;
thickness, 1-56 mm; eight ribs. Catalogue numbers MGUH 18232-3; SPGIH 3543, 3546.

Description. A very variable species. Outline semicircular to subpentagonal. Hinge margin often extended into
small wings. Juveniles clearly biconvex, but plano-convex shape is rapidly attained. Rib number increases
during growth from four to about eight. Ribs divided into two wide bundles, separated by shallow median
sinus. Single ribs can develop in sinus at all stages of growth. Ribs rather flat, extending only slightly beyond
shell margin; there are up to eight strong growth lines. Beak pointed; nearly straight to suberect. Cardinal area
plane to slightly concave. Deltidial plates are narrow, foramen large, triangular. Pedicle collar well-developed
with concave anterior margin. Inner socket ridges diverge strongly anteriorly. Cardinal process developed as
low knob between posterior edges of inner socket ridges. Two concave circular separated hinge plates are
situated anterior to the cardinal process. Brachidium is not preserved in present material. Dorsal median
septum high, extended in ventral direction by attached distal ends of descending branches giving septum a
characteristic split appearance. Ventral septum very thin, short. Inner surface of the shell is ornamented by a
weak negative of external sculpture. Margin slightly elevated, finely tuberculated.

Remarks. The great variability of A. bronnii results in morphological overlap with several other
species. It can normally be separated from these on the basis of its rib number, split median septum,
disjunct hinge plates and low median sinus.

Occurrence. Catton 5, Caistor 3 and 1, Overstrand Hotel Lower Mass, Overstrand Hotel Upper Mass 1A, 2A
and 3, Sidestrand 1, 1 A, 4 and 5, Little Marl Point 1, 5 and 6?, and Trimingham 2. The species is also recorded
from the Upper Campanian and Maastrichtian of England, the Netherlands, northwest and East Germany,
eastern Poland and Denmark. In Denmark it also occurs in the Lower Danian.

Argyrotheca coniuncta Steinich, 1965

Plate 8, figs 5, 6, 8 and 9

858

PALAEONTOLOGY, VOLUME 33

1965 Argyrotheca coniuncta Steinich, p. 138, figs 200-206; pi. 18, fig. 2 a-d.

1972 Argyrotheca coniuncta Steinich; Surlyk, p. 20, figs 5-12.

1979 Argyrotheca coniuncta Steinich; Bitner and Pisera, p. 79, pi. 5, figs 5-7.

1982 Argyrotheca coniuncta Steinich; Surlyk, fig. 1, pi. 4, figs a-d.

1984 Argyrotheca coniuncta Steinich; Surlyk and Johansen, fig. 1.

1987a Argyrotheca coniuncta Steinich; Johansen, p. 28, pi. 12, fig. 5a, b.

1988 Argyrotheca coniuncta Steinich; Johansen, fig. 2.

Material. One complete shell, nine dorsal, eight ventral valves, and six fragments. Largest specimen a dorsal
valve from Overstrand Hotel Upper Mass 2: length of dorsal valve, 3-4 mm; width (which corresponds to the
width of the hinge line), 7 0 mm; 12 ribs. Catalogue numbers MGUH 18234-18236; SPGIH 3547.

Description. Outline characteristically semicircular very often with wing-like extensions of the hinge line.
Anterior commissure straight, shell flat biconvex, shell surface bears 6-12 low, straight ribs. Beak erect and
characteristically strongly withdrawn. Foramen very large, hypothyridid, triangular; hinge strong; cardinal
process small, anterior margins of inner socket ridges much longer than posterior. Hinge plates fused mid-
dorsally into closed platform, sometimes seen as individual plates separated by median furrow. Dorsal median
septum high, strong, not split. Shell surface with large punctae; rather thick.

Remarks. The specimens investigated correspond in many details to the originals of A. coniuncta
described by Steinich (1965). Minor differences in the development of the hinge plates are, however,
observed as the hinge plates in specimens from Riigen always fuse to form a closed platform
(Steinich 1965), while the hinge plates in the specimens from Norfolk form both closed platforms
and separate plate-like hinge plates.

Large forms of A. bronnii (Roemer) differ from A. coniuncta in always possessing two distinctly
separate concave disc-like hinge plates and a split dorsal median septum.

Occurrence. Overstrand Hotel Lower Mass, Bramerton, Overstrand Hotel Upper Mass 1A, 1, 2 and 2A,
Sidestrand 1 and 3A, and Little Marl Point 2. Also known from the Upper Campanian and Maastrichtian of
northwest Germany and from the Maastrichtian of East Germany, Denmark and eastern Poland.

Argyrotheca hirundo (Hagenow, 1842)

Plate 3, figs 6-8

1842 Orthis hirundo Hagenow, p. 545, pi. 9, fig. 9 a-d.

1965 Argyrotheca hirundo (Hagenow); Steinich, p. 152, figs 232-249; pi. 16, fig. 3; pi. 17, fig. 3 a-d.

1972 Argyrotheca hirundo (Hagenow); Surlyk, p. 20, figs 5, 17, 18.

1979 Argyrotheca hirundo (Hagenow); Bitner and Pisera, p. 79, pi. 4, fig. 7.

1982 Argyrotheca hirundo (Hagenow); Surlyk, fig. 1, pi. 3, figs f-h.

1984 Argyrotheca hirundo (Hagenow); Surlyk and Johansen, fig. 1.

1987a Argyrotheca hirundo (Hagenow); Johansen, p. 34, fig. 23a-h; pi. 14, figs 1-8.

1988 Argyrotheca hirundo (Hagenow); Johansen, fig. 2, pi. 3, fig. 2.

EXPLANATION OF PLATE 8

Figs 1-4, 7. Argyrotheca bronnii (Roemer, 1841). 1, MGUH 18232, upper Lower Maastrichtian, Trimingham
(coll. WKC, 1971 ), Norfolk; large juvenile in dorsal view, x 30. 2, SPGIH 3543, lower Lower Maastrichtian,
G 147, Kronsmoor; interior of large adult dorsal valve, x 15. 3 and 4, MGUH 18233, upper Lower
Maastrichtian, Trimingham (coll. WKC, 1971), Norfolk; 3, interior of broken adult dorsal valve, x 15. 4,
oblique lateral view showing profile of median septum, x 50. 7, SPGIH 3546, lower Lower Maastrichtian,
G 147, Kronsmoor; large adult specimen in dorsal view, x 10.

Figs 5, 6, 8, 9. Argyrotheca coniuncta Steinich, 1965. 5, SPGIH 3547, upper Lower Campanian, G 79,
Lagerdorf; large adult specimen in dorsal view, x 30. 6, MGUH 18234, upper Lower Maastrichtian,
Trimingham (coll. WKC, 1971), Norfolk; juvenile specimen in dorsal view, x25. 8, MGUH 18235,
lowermost Maastrichtian?, Overstrand Hotel Lower Mass, Norfolk; broken adult dorsal valve showing
separated hinge plates and high median septum, x 25. 9, MGUH 18236, lower Lower Maastrichtian,
Overstrand Hotel Upper Mass 2A, Norfolk; small juvenile in dorsal view, x20.

PLATE 8

JOHANSEN and SURLYK, Argyrotheca

860

PALAEONTOLOGY, VOLUME 33

Material. Two dorsal valves, one ventral valve, and one fragment. Largest specimen a ventral valve from
Bramerton: length, l-92mm; width, I 84 mm; length of dorsal valve (measured from hinge line to anterior
margin), 1-28 mm. Catalogue numbers MGUH 16856. 16857; SPGIH 3534.

Description. Outline subtriangular. Shell biconvex with rather flat dorsal valve. Hinge margin enlarged in some
specimens as lateral pointed extensions. Surface has very strong growth lines and 4-6 radial ribs in two anterio-
laterally directed bundles divided by wide median furrow. Beak pointed, suberect, cardinal area plane to slightly
concave. Foramen large, triangular, hypothyridid, limited by two narrow deltidial plates. Pedicle collar
strongly developed with concave anterior margin. Inner socket ridges strong, anteriorly divergent. Cardinal
process knob-shaped. Hinge plates small, converging anteriorly, delimiting a triangular chamber together with
socket ridges and cardinal process. Dorsal median septum is high. Its anterior margin forms right angle with
inner surface of the shell, posterior margin forms gentle, concave-upward slope. Ventral septum thin, low.
Relatively thin shelled.

Remarks. A. hirundo is an extremely variable species, but can normally be distinguished by its
triangular outline, and the bundled arrangement of the ribs.

Occurrence. Bramerton and Sidestrand 4. Also known from the Upper Campanian and Maastrichtian of
northwest Germany and from the Maastrichtian of East Germany, Poland, the Netherlands and Denmark. In
Denmark it also occurs in the Lower Danian.

Family terebratellidae King, 1850
Subfamily magasinae Dali, 1870
Genus magas J. Sowerby, 1816

Type species. Magas pumilus J. Sowerby, 1816, p. 30.

Magas chitoniformis (Schlotheim, 1813)

Plate 10, figs 1-6

(1798) Terebratulus fossiles Faujas, pi. 26, fig. 6.

1813 Terebratula chitoniformis Schlotheim, p. 133.

1818 Magas pumilus J. Sowerby, p. 40, pi. 1 19, figs 3-5.

1965 Magas chitoniformis (Schlotheim); Steinich, p. 183, figs 280-294; pi. 19, fig. 2; pi. 20, fig. 1 a-d.
1972 Magas chitoniformis (Schlotheim); Surlyk, p. 26, figs 5, 11, 12; pi. 5, fig. C.

1982 Magas chitoniformis (Schlotheim); Surlyk, fig. 1, pi. 4, figs e-i.

1984 Magas chitoniformis (Schlotheim); Surlyk and Johansen, fig. 1.

1987a Magas chitoniformis (Schlotheim); Johansen, p. 43.

1988 Magas chitoniformis (Schlotheim); Johansen, fig. 2.

EXPLANATION OF PLATE 9

Figs 1^1. Aemula inusitata Steinich, 1968. 1, MGUH 18237, upper Lower Maastrichtian, Trimingham (coll.
WKC, 1971), Norfolk; exterior view of ventral valve, x25. 2, MGUH 18238, lower Lower Maastrichtian,
Overstrand Hotel Upper Mass 3, Norfolk; interior of slightly broken dorsal valve, x20. 3, MGUH 18239,
upper Lower Maastrichtian, Trimingham (coll. WKC, 1971), Norfolk; interior of ventral valve showing
recrystallized schizolophe, x 30. 4, MGUH 18240, lower Lower Maastrichtian, Overstrand Hotel Upper
Mass 3, Norfolk; juvenile dorsal valve in exterior, x 30.

Figs 5-8. Dracius carnifex Steinich, 1967. 5, MGUH 18241, lowermost Maastrichtian, Overstrand Hotel Lower
Mass, Norfolk; exterior of juvenile ventral valve, x 30. 6, MGUH 18242, lowermost Maastrichtian,
Overstrand Hotel Lower Mass, Norfolk; fragment of large ventral valve showing characteristic surface
sculpture, x25. 7, SPGIH 3548, lower Lower Maastrichtian, G 148-150, Kronsmoor; adult specimen in
dorsal view, x25. 8, SPGIH 3549, lower Lower Maastrichtian, G 148-150, Kronsmoor; exterior of large
ventral valve, x 25.

PLATE 9

JOHANSEN and SURLYK, Aemula , Dracius

862

PALAEONTOLOGY, VOLUME 33

Material. Twenty-one complete shells, the majority juvenile specimens, ninety-six dorsal, twenty-seven ventral
valves, and one hundred and sixty fragments. Largest specimen a slightly damaged complete adult from Little
Marl Point 7: length, 9-0 mm; length of dorsal valve, 7-0 mm; width, 9-0 mm; width of hinge line, 6-4 mm;
width of foramen, 10 mm and thickness, 4 0 mm. Catalogue numbers MGUH 18243-18246.

Description. Shell relatively large, smooth, with large punctae. Outline changes during growth from elongated
oval to subcircular. Shell flattened biconvex in juveniles, becoming plano-convex with very high convex ventral
valve as size increases. Hinge line straight, relatively long, giving posterior margin of dorsal valves of juveniles
a pair of characteristic shoulders. Beak low, blunt, erect; changes to distinctly incurved during growth.
Foramen subtriangular, large, hypothyridid in juveniles; later triangular, submesothyridid. Hinge strongly
built; brachidium characteristically complex as shown in the illustrations of Steinich (1965, figs 287-292).
Dorsal median septum high, strong, ventral valve contains low median ridge in anterior part of valve floor.

Remarks. Juvenile M. chitoniformis characteristically possess shoulders on the posterior margin of
the dorsal valve but otherwise resemble Dalligas nobilis Steinich and Scumulus inopinatus Steinich.
D. nobilis is distinugished from M. chitoniformis in possessing a much simpler brachidium. Larger
forms of D. nobilis bear weak radial ribs; their shells are flat biconvex. S. inopinatus differs from
juvenile M. chitoniformis in often possessing an amphithyridid foramen and an irregular outline.
Leptothyrellopsis polonicus Bitner and Pisera is distinguished from juvenile M. chitoniformis by its
high hypothyridid foramen and long, high, plate-like median septum.

Occurrence. Keswick, Catton 1, 3 and 3A, Stoke Holy Cross 27, Catton 47, 5? and 67, Caistor 1, Bramerton,
Overstrand Hotel Upper Mass 1A, 1, 2, 3 and 47, Sidestrand 1, 1A7, 3A and 5, Little Marl Point 1A, IB, 17,
2, 2A, 37, 4, 5, 6 and 7, and Trimingham 1, 2 and 3. Also widely distributed in the Upper Campanian and
Maastrichlian of northwest Germany and the Maastrichtian of East Germany, Poland, France, the
Netherlands and Denmark.

Subfamily kingeninae Elliott, 1948
Genus kingena Davidson, 1852

Type species. Terebratula lima DeFrance, 1828, p. 156.

Kingena pentangulata (Woodward, 1833)

Plate 1 1, figs 5-8

1833 Terebratula pentangulata Woodward, p. 49, pi. 6, fig. 10.

1970 Kingena pentangulata (Woodward); Owen, p. 64, pi. 7, figs 6 and 7; pi. 8, figs 2-6; text-fig. 13.

1970 Kingena blackmorei Owen, p. 63, pi. 7, figs 1-3; pi. 8, fig. 1.

19686 Kingena sp. ; Steinich, p. 342, figs 4 and 5.

1972 Kingena pentangulata (Woodward); Surlyk, p. 17.

EXPLANATION OF PLATE 10

Figs 1-6. Magas chitoniformis (Schlotheim, 1813). I and 2, MGUH 18243, lower Upper Campanian, Catton
1, Norfolk; 1, interior of large slightly broken dorsal valve, x 7. 2, oblique lateral view, x 7. 3 and 4, MGUH
18244, upper Lower Maastrichtian, Little Marl Point 7, Norfolk; 3, slightly broken large adult specimen in
dorsal view, x 10; 4, oblique lateral view showing reclining profile, x 13. 5, MGUH 18245, lower Upper
Campanian, Catton 3 A, Norfolk; small juvenile specimen in dorsal view, x40. 6, MGUH 18246, lower
Upper Campanian, Catton 1, Norfolk; juvenile specimen in dorsal view, x25.

Fig. 7. Dalligas nobilis Steinich, 19686, MGUH 18247, lower Upper Campanian, Catton 1, Norfolk; adult
specimen in dorsal view, x 25.

Fig. 8. Scumulus inopinatus Steinich, 19686, SPGIH 3550, lower Upper Campanian, G 109, Lagerdorf; adult
specimen in dorsal view, x 30.

PLATE 10

JOHANSEN and SURLYK, Magas , Dalligas , Scumulus

864

PALAEONTOLOGY, VOLUME 33

1979 Kingena sp.; Bitner and Pisera, p. 81, pi. 5, figs 1 and 2.

1984 Kingena pentangulata (Woodward); Surlyk and Johansen, fig. 1.

1987a Kingena pentangulata (Woodward); Johansen, p. 42, pi. 20, figs 5 and 6.

1988 Kingena pentangulata (Woodward); Johansen, fig. 2.

Material. Seven shells, thirty-seven dorsal, thirty-four ventral valves, and two hundred and fifty-six fragments.
Largest valves have following dimensions: fragment of ventral valve from Overstrand Hotel Upper Mass 2A:
length at least 4-4 mm; dorsal valve from Cley: length, 2-9 mm; width, 3-4 mm; shell from Overstrand Hotel
Upper Mass 3: length, 2-6 mm; width, 3 0 mm; length of dorsal valve, 2-2mm; thickness at least 0-6 mm.
Catalogue numbers MGUH 18252-18255.

Description. Adults not seen; medium-sized specimens are highly fragmented. Juvenile stages less than 5 mm in
length. The outline changes during growth from subtriangular to circular to subpentagonal. Small forms plano-
convex to weakly biconvex; larger forms more uniformly biconvex. Commissure straight in early stages. Shell
surface characteristically bears densely-spaced small regular tubercles. Beak almost straight to suberect in
earliest stages, changing to suberect and finally erect. Cardinal area wide, sharply delimited. Foramen large,
trapezoidal; submesothyridid in forms less than 5 mm long. It becomes permesothyridid as beak becomes more
incurved in larger forms. Foramen limited laterally by rather large triangular deltidial plates. Teeth strong,
pointed, supported by well-developed dental plates which converge posteriorly. Larger specimens have well-
defined hinge plate. A low median septum is formed at c. 0-8 mm length. At c. L4 mm length two pointed
anterio-laterally directed ascending branches form at top of anterior margin of the septum. At the same lime
two short crurae are formed as extensions of inner socket ridges. Strongly divergent descending branches
extend from end of crurae. Latest stages of brachidium development not seen in present material. Shell is
thin and almost always rather deformed. Inner surface is smooth.

Remarks. The present material is referred to K. pentangulata with slight reservation. It corresponds,
however, in all details to abundant material of this species from the Danish Maastrichtian.
Furthermore the neotype of K. pentangulata (Woodward) selected by Owen (1970) comes from the
transition between the Eaton and Weybourne Chalk Members in Norfolk, while its closest relative
K. blackmorei Owen is said to characterize the somewhat older chalk of the Goniotheuthis quadrata
Zone in South Wiltshire and Hampshire (Owen 1970). The differences between the two species are
extremely slight, and mainly reflect one main feature, the less incurved umbo of K. blackmorei. We
prefer to include K. blackmorei in K. pentangulata.

The surface sculpture, hinge plate and brachidium distinguish K. pentagulata from all other
known Upper Campanian-Maastrichtian species.

Occurrence. Cley, Keswick, Catton 1, 3, 3 A, 4, 5 and 6, Caistor 3 and 1, Whitlingham, Overstrand Hotel Lower
Mass, Bramerton, Overstrand Hotel Upper Mass 1A, 1, 2, 2A, 3 and 4, Sidestrand 1, 1A, 2, 3, 3 A and 4, and
Little Marl Point 1A. The species is known from the Upper Campanian-Maastrichtian and possibly older
Chalk strata of northwest Europe.

Family platidiinae Thomson, 1927
Genus aemula Steinich, 1968

Type species. Aemula inusitata Steinich, 1968a, p. 193.

Aemula inusitata Steinich, 1968

Plate 9, figs 1^4

1968a Aemula inusitata Steinich, p. 192, figs 1-5; pi. 1, fig. 1.

1972 Aemula inusitata Steinich; Surlyk, fig. 5, pi. 3, figs e, f h.

1974 Aemula inusitata Steinich; Surlyk, figs 1 and 3, pi. 2.

1979 Aemula inusitata Steinich; Bitner and Pisera, p. 79, pi. 5, fig. 4.
1982 Aemula inusitata Steinich; Surlyk, fig. 1, pi. 2, figs g and h.

1984 Aemula inusitata Steinich; Surlyk and Johansen, fig. 1.

JOHANSEN AND SURLYK: CRETACEOUS BRACHIOPODS

865

1987a Aemula inusitata Steinich; Johansen, p. 39, pi. 18, figs 1-5.

1988 Aemula inusitata Steinich; Johansen, fig. 2, pi. 3, figs 4a, b. 5, 6.

Material. Five complete shells, ten dorsal, sixty-three ventral valves and ten fragments. Largest specimen is a
dorsal valve from Overstrand Hotel Upper Mass 3: length, 2-32 mm; width, 2-64 mm. Catalogue numbers
MGUH 18237-18240.

Description. Outline oval to circular, commonly very irregular. Shell plano-convex to slightly concavo-convex.
Commissure straight in regular specimens. Surface of dorsal valve smooth, surface of the ventral valve bears
short pointed tubercles. Shells with width of less than 10 mm rarely have tubercles, but larger specimens show
no correlation between shell size and size and density of tubercles. There are several growth lines, clearly
developed on the dorsal valve. Beak very short, suberect. Cardinal area small, clearly delineated. Hinge line
almost straight. Foramen relatively large, amphithyridid ; laterally delineated by two narrow deltidial plates.
Short, slightly elevated pedicle collar is delineated from deltidial plates by furrow. Teeth are very short, obtuse.
Inner socket ridges well-developed, extending somewhat outside posterior shell margin. Dorsal valve has low
triangular median septum, with two ventrally directed short triangular ascending branches. A. inusitata has a
strongly developed spicular skeleton; schizolophe. Inner side of the rather thin shell is smooth.

Remarks. A. inusitata is easily distinguished from all other Chalk brachiopods. Very regular and
smooth forms may, however, show considerable overlap with a new species of Aemula illustrated
from the Maastrichtian chalk of Hemmoor, northwest Germany by Surlyk (1974, pi. 2b). The
present material does not help in the separation of the two forms.

Occurrence. Keswick, Caistor 1, Overstrand Hotel Lower Mass, Bramerton, Overstrand Hotel Upper Mass
1A, 2, 2A, 3 and 4, Sidestrand 3, 3A, 4 and 5, and Little Marl Point IB, 6 and 7. Also known from the
Santonian to the Maastrichtian of northwest Germany and from the Maastrichtian of East Germany,
Denmark and eastern Poland.

Aemula sp.

Material. One complete shell and one ventral valve. Both specimens are juveniles with shell widths around
F5 mm.

Description. Shell smooth, thin, plano-convex pointed, oval, slightly irregular outline. Beak very short, acute,
suberect; foramen large, amphithyridid. Teeth small, blunt, inner socket ridges protrude beyond posterior shell
margin. Internal morphology not known.

Remarks. A. inusitata Steinich differs from this species in possessing a shell sculpture of low, densely-
spaced tubercles concentrated on the ventral valve. This smooth Aemula sp. may be conspecific with
the Aemula n. sp. illustrated from the Maastrichtian chalk of Hemmoor, northwest Germany
(Surlyk, 1974, pi. 2b). The material from Norfolk is, however, too small to allow this problem to
be resolved.

Occurrence. Caistor 1.

Genus scumulus Steinich, 1968
Type species. Scumulus inopinatus Steinich, 1968a, p. 199, pi. 1, fig. 2.

Scumulus sp.

Material. One fragmented ventral valve is present, but this is not illustrated here due to its poor state of
preservation. For comparison a specimen of S. inopinatus from the lower Upper Campanian of Lagerdorf is
shown (PI. 10, fig. 8). Catalogue number SPGIH 3550.

Description. Valve small, smooth, elongated oval in outline. Beak short, erect, blunt; area characteristically
broad and triangular with lateral edges formed by shallow ridges. Deltidial plates very narrow, hinge teeth are
strong, acute.

866 PALAEONTOLOGY, VOLUME 33

Remarks. The species can be assigned to the genus Scumulus Steinich, but the material does not
allow specific determination.

Juveniles of Magas chitoniformis (Schlotheim) show great resemblance to this species, but M.
chitoniformis differs in possessing a small, distinctly limited cardinal area and large punctae.
Leptothyrellopsis polonicus Bitner and Pisera and Dalligas nobilis Steinich both differ from the
Norfolk species in their very narrow cardinal areas.

Occurrence. Caistor 1.

Family dallinidae Beecher, 1893
Genus dalligas Steinich, 1968

Type species. Dalligas nobilis Steinich, 1968 b, p. 336, pi. 1, fig. 1.

Dalligas nobilis Steinich, 1968

Plate 10, fig. 7

19686 Dalligas nobilis Steinich, p. 336, figs 1-3; pi. 1, figs 1 and 2.

1972 Dalligas nobilis Steinich; Surlyk, p. 20, figs 5, 13-15, 17, 18.

1987a Dalligas nobilis Steinich; Johansen, p. 43, pi. 18, figs 6 and 7.

Material. Two bivalved specimens, three ventral and one dorsal valve. All specimens fragmentary, largest
specimen, from Overstrand Hotel Upper Mass 4, is 2-8 mm long. Catalogue number MGUH 18247.

Description. Outline circular, shell flatly biconvex. Auricles not delimited. Commissure straight. Numerous
very fine radial ribs appear at 0-5-10 mm width. Rib number about 30. Cardinal area is narrow, triangular.
Deltidial plates narrow, well-separated posteriorly. Foramen triangular to trapezoidal. Pedicle collar wide,
well-defined. Inner socket ridges subparallel, forming characteristic triangular area together with well-defined
hinge plates, and posterior margin of the median septum. Septum high, bearing two short ascending branches.
Remaining elements of the brachidium not developed. Shell relatively thin.

Remarks. The specimen closely resembles D. nobilis as described by Steinich (19686) from Riigen,
although the ribbing seems to appear somewhat earlier in ontogeny. Juvenile specimens, where
ribbing has not yet developed, cannot be distinguished form the smooth D. mielnicensis Bitner and
Pisera. Young specimens differ from Leptothyrellopsis polonicus Bitner and Pisera in their more
circular outline, larger foramen, triangular area and wider, less pointed beak.

Occurrence. Keswick, Catton 1, Overstrand Hotel Lower Mass, and Overstrand Hotel Upper Mass 2A and 4.
The stratigraphic and geographic distributions of this almost featureless species are not yet well known.

EXPLANATION OF PLATE 1 1

Figs 1-5. Leptothyrellopsis polonicus Bitner and Pisera, 1979. 1, MGUH 18248, lower Lower Maastrichtian,
Sidestrand 3 A, Norfolk; broken adult specimen in dorsal view, x 30. 2, MGUH 18249, uppermost
Campanian?, Whitlingham, Norfolk; juvenile in dorsal view, x25. 3, MGUH 18250, uppermost
Campanian?, Whitlingham, Norfolk; large juvenile specimen in dorsal view, x 30. 4 and 5, MGUH 18251,
lower Lower Maastrichtian, Sidestrand 2, Norfolk; 4, interior of juvenile dorsal valve, x 25; 5, 4 in lateral
view showing profile of median septum, x 50.

Figs 6-9. Kingena pentangulata (Woodward, 1833). 6, MGUH 18252, lower Lower Maastrichtian, Overstrand
Hotel Upper Mass 2A, Norfolk; cardinal area of medium-sized dorsal valve, x 25. 7, MGUH 18253, lower
Lower Maastrichtian, Overstrand Hotel Upper Mass 2A, Norfolk; small juvenile specimen in dorsal view,
x25. 8, MGUH 18254, upper Upper Campanian, Caistor 3, Norfolk; interior of broken medium-sized
dorsal valve, x 25. 9, MGUH 18255, upper Upper Campanian, Catton 3A, Norfolk; juvenile specimen in
dorsal view, x 30.

PLATE 11

JOHANSEN and SURLYK, Leptothyrellopsis , Kingena

868

PALAEONTOLOGY, VOLUME 33

Family thecideidae Gray, 1840
Subfamily lacazellinae Backhaus, 1959
Genus lacazella Munier-Chalmas, 1880

Type species. Thecidea mediterranea Risso, 1826, p. 393.

Subgenus bifolium Elliott, 1948
Lacazella (Bifolium) wetherelli (Morris, 1851)

Plate 1, fig. 7

1851 Thecidea wetherelli Morris, p. 86, pi. 4, figs 1-3.

1852 Thecidea wetherelli Morris; Davidson, p. 14, pi. 1, figs 15-26.

1874 Thecidium wetherelli Morris; Davidson, p. 22.

1908 Thecidium wetherelli Morris; Rowe, p. 320.

1959 Lacazella ( Bifolium ) wetherelli (Morris); Backhaus, p. 35, pi. 2, figs 1 3.

1965 Lacazella (Bifolium) wetherelli (Morris); Steinich, p. 15, pi. 4, figs 2 and 3 a, b.

1972 Bifolium wetherelli (Morris); Surlyk, p. 27, fig. 5.

1974 Praelacazella wetherelli (Morris); Pajaud, p. 333.

1988 Lacazella (Bifolium) wetherelli (Morris); Johansen, fig. 2.

Material. Four dorsal valves and three fragments, all of which probably belong to ventral valves. Largest
specimen a dorsal valve from Lower Maastrichtian sample Little Marl Point IB. Length of dorsal valve,
5-5 mm; width, 6-5 mm. Catalogue number MGUH 16846.

Description. Outline of the dorsal valve subtriangular to trapezoidal. Shell margins broad, steep, levelling out
towards anterior shell margin. Valves thick, stout. Shell surface smooth with several growth rings. Dorsal valve
planar. Shell margin sculptured with numerous fine, minute pustules and radiating, low ridges. Posterior shell
margin short, straight. Cardinal process protrudes well beyond posterior shell margin and is broad and short.
Ascending part of brachial apparatus consists of two separate laminae which rise anteriorly in dorsal valve.
Laminae follow depression in floor of valve and run obliquely posteriorly towards jugum, where laminae
abruptly change direction of growth deflecting towards aperture. Folded laminae unite mid dorsally above
jugum. A second folding of laminae may occur but characteristically only one or two folds are present.
Ptycholophe.

Remarks. Backhaus (1959) mentions that the dorsal valve of this species is typically very thin, but
this is not the case in the specimens from Norfolk. When present, the ventral valve of this species
is characterized by being attached (cemented) to the substrate with the entire outer surface, and by
possessing a distinct radial rib pattern in the inner surface of the ventral valve. Lacazella (Bifolium)
wetherelli is distinguished from other closely-related species of Lacazella in the form of the brachial
apparatus. In L. ( B .) wetherelli the laminae are folded once or in some cases twice.

Occurrence. Little Marl Point IB, 1,4 and 7?. Also known from Upper Campanian-Maastrichtian Chalk
strata of northwest Germany and from the Maastrichtian of East Germany and Denmark.

Family uncertain

Genus leptothyrellopsis Bitner and Pisera, 1979

Type species. Leptothyrellopsis polonicus Bitner and Pisera, 1979, p. 82, pi. 7, fig. 2.

Leptothyrellopsis polonicus Bitner and Pisera, 1979
Plate 1 1, figs 1-4

1979 Leptothyrellopsis polonicus Bitner and Pisera, p. 82, pi. 7, figs 1-4.

1982 Leptothyrellopsis polonicus Bitner and Pisera; Surlyk, fig. 1, pi. 3, figs c-e.

1988 Leptothyrellopsis polonicus Bitner and Pisera; Johansen, fig. 2.

JOHANSEN AND SURLYK: CRETACEOUS BRACHIOPODS

869

Material. Seventy complete shells, one hundred and ninety dorsal, one hundred and fifty-three ventral valves,
and fifty-eight fragments. Largest specimen from Cley: length, 3-4 mm; width, 2-8 mm; length of brachial
valve, 2-9 mm; thickness, 06 mm. Catalogue numbers MGUH 18248-18251.

Description. Outline oval to elongate oval. No clearly delimited auricles. Shell flatly biconvex; commissure
straight. Surface smooth apart from a few growth lines ; beak erect ; shell is punctate ; area small ; deltidial plates
narrow with subparallel sides; elevated above shell surface as low ridges. Foramen large, subtriangular. There
is no pedicle collar but high inner socket ridges. Dorsal valve has prominent plate-like median septum.

Remarks. The almost total absence of diagnostic features makes it difficult to distinguish L.
polonicus from other small smooth brachiopods. It differs from juvenile Dalligas Steinich and
Magas chitoniformis (Schlotheim) in the absence of a pedicle collar and of a prominent hinge plate.
Scumulus inopinatus Steinich can be distinguished from L. polonicus by its well-developed
brachidium, its amphithyridid foramen and irregular outline.

Occurrence. Cley, Keswick, Catton 3A, 5 and 6, Caistor 1, Whitlingham, Overstrand Hotel Lower Mass,
Overstrand Hotel Upper Mass 1A, 1,2, 2A, 3 and 4, Sidestrand 1, 1A, 2, 3 A and 4, Little Marl Point IB, 2
and 3, Trimingham 1, and Little Marl Point 4 and 6. This species is known from Upper Campanian-
Maastrichtian Chalk strata of northwest Germany, and from the Maastrichtian of eastern Poland and
probably also Denmark.

SUMMARY

The Upper Cretaceous Chalk of Norfolk is not very well known due to its poor exposure in
scattered, commonly abandoned chalk pits, and in glacially transported masses exposed in coastal
cliffs. The sediments consist of white to yellowish chalk with scattered bands of nodular or barrel-
shaped flints (‘paramoudras’). Calcarenites, nodular cemented chalks with sponge impressions, and
hardgrounds are also common.

The Chalk has been subdivided into a number of units, but the stratigraphic nature of these is
rather vague, although they seem to be mainly palaeontologically defined (see also Peake and
Hancock 1961, 1970; Wood 1988). In the present paper we have attempted to present a coherent
lithostratigraphic scheme based on the older units, and eight lithostratigraphical units of member
rank are proposed.

A detailed biozonation based on the micromorphic brachiopod fauna has been established for the
Upper Campanian and Maastrichtian of northwest Europe (Surlyk 1984). A large number of bulk
samples collected in the Upper Campanian-Lower Maastrichtian of Norfolk, England has, after
Glauber-Salt treatment and freeze-thawing, yielded a brachiopod fauna of about 2200 specimens
representing at least thirty-five species. All of the species present are known also from the Upper
Cretaceous Chalk of Denmark, Sweden, Germany and Poland.

The existence of this fauna in the English Chalk makes it possible to correlate parts of the
Campanian and Maastrichtian outcrops in Norfolk with more complete and better exposed sections
of northwest continental Europe, especially the Campanian-Maastrichtian boundary section at
Kronsmoor, West Germany.

In spite of the condensed and poorly exposed nature of the Norfolk section, most of the
brachiopod zones established in northwest Europe can be recognized. These zones are from below;

1. The tenuicostata-longicollis Zone corresponding to the bulk of the Upper Campanian of
Norfolk ;

2. the longicollis-jasmundi Zone corresponding to the uppermost Upper Campanian;

3. the acutirostris-spinosa Zone corresponding to the basal Maastrichtian;

4. the spinosa-pulchellus Zone corresponding to the middle part of the Lower Maastrichtian;

5. the pulchellus-pulchellus Zone of the upper Lower Maastrichtian.

The jasmundi-acutirostris Zone which includes the lowermost Maastrichtian in Kronsmoor is
undoubtedly represented in unexposed strata between the highest Campanian locality, Whitlingham
and the lowest Maastrichtian exposed at Overstrand Hotel Lower Mass.

870

PALAEONTOLOGY, VOLUME 33

In spite of its highly fragmentary exposure, the Upper Campanian-Maastrichtian of Norfolk is
now correlated in detail on the basis of the brachiopods with the standard boundary sequences at
Kronsmoor and Lagerdorf in northwest Germany, and with the detailed macrofossil zonation
established for these localities by German workers.

Acknowledgements . The samples which form the basis for the present study were collected in 1971 by W. Kegel
Christensen and F. Surlyk under the excellent and highly inspiring guidance of J. M. Hancock and N. B.
Peake. Some additional samples were collected by R. G. Bromley in 1971. This paper was read critically by J.
M. Hancock, N. B. Peake, W. J. Kennedy and one anonymous referee. The typing was done by B. Larsen and
A. Brantsen and drafting by B. Sikker Hansen. Stereoscan work was undertaken in the Scanning Electron
Microscope Laboratory of the Geological Central Institute, University of Copenhagen, and darkroom work
was done by J. Aagaard and O. B. Berthelsen. The paper was written while Surlyk was a recipient of a research
professorship awarded by the Danish Natural Science Research Council.

REFERENCES

asgaard, u. 1975. A revision of Sahni’s types of the brachiopod subfamily Carneithyridinae. Bulletin of the
British Museum ( Natural History), (Geology), 25, 320-362.
backhaus, e. 1959. Monographic der cretacischen Thecideidae (Brach.). Mitteilungen aus dem geologischen
Staatsinstitut Hamburg , 28, 5-90.

bitner, m. a. and pisera, a. 1979. Brachiopods from the Upper Cretaceous chalk of Mielnik (Eastern Poland).
Acta Geologica Polonica , 29, 67-88.

boll, E. 1856. Die Brachiopoden der Kreideformation in Mecklenburg. Archiv des Vereins der Freunde der
Naturgeschichte Mecklenburg , 10, 29 48.

bosquet, j. 1860. Monographic des Brachiopodes fossiles du terrain Cretace superieur des Duche de
Limbourg. Part 1. Memoirs geologiques de la Neerlande, 3, 1-50.
bromley, r. G., schulz, M. G. and peake, N. b. 1975. Paramoudrae, giant flints, long burrows and the early
diagenesis of chalks. Det Kongelige danske Videnskabernes Selskab , 20 (10), 1-31.
brydone, r. M. 1906. Further notes on the stratigraphy and fauna of the Trimingham Chalk. Geological
Magazine , 43, 13-22, 72-78, 124-31, 289-300.

1908. On the Subdivisions of the Chalk of Trimingham (Norfolk). Quarterly Journal of the Geological
Society of London , 64, 401-12.

- 1909. The Trimingham Chalk - South Bluff. Geological Magazine , 46, 189-90.

— 1930. The 'Norwich Chalk’. Transactions of the Norfolk and Norwich Natural Society , 13, 47-49.

— 1938. On correlation of some of the Norfolk exposures of Chalk with Belemnitella mucronata. London,
15 pp.

Christensen, w. k., ernst, G., schmid, f., schulz, m.-g. and wood, c. J. 1975. Belemnitella mucronata
mucronata (schlotheim 1813) from the Upper Campanian: Neotype, biometry, comparisons and
biostratigraphy. Geologisches Jahrbuch , A28, 27-57 .

davidson, T. 1852. A monograph of British Cretaceous Brachiopoda. Monograph of the Palaeontographical
Society , 20, 5-54.

— 1874. A monograph of the British fossil Brachiopoda, supplement to the British Cretaceous Brachiopoda.
Monograph of the Palaeontographical Society, 27, 17-72.

deegan, c. e. and scull, b. J. 1977 (Compilers). A standard lithostratigraphic nomenclature for the Central
and Northern North Sea. Report of the Institute for Geological Sciences, 77/25, 1-36.
ernst, h. 1984. Ontogenie, Phylogenie und Autokologie des inartikulaten Brachiopoden Isocrania in der
Schreibkreidefazies NW-Deutschlands (Coniac bis Maastricht). Geologisches Jahrbuch, All, 3-105.
faujas de saint-fond, b. 1798. Histoire naturelle de la Montagne de Saint-Pierre de Maastricht. H. J. Jansen,
Paris, 263 pp.

fletcher, t. p. and wood, c. J. 1978. Cretaceous rocks. In wilson, h. e. and manning, p. i. Geology of the
Causeway Coast - Memoir for one-inch geological sheet 7, 2, 84—1 15, Geological Survey of Northern Ireland,
Belfast.

hagenow, f. v. 1842. Monographie der Riigen’schen Kreide-Versteinerungen, III. Abt: Mollusken. Neues
Jahrbuch fur Mineralogie, 1842, 528-575.

johansen, m. b. 1986. Colonisation, mass extinction and adaptive radiation - an evolutionary and

JOHANSEN AND SURLYK: CRETACEOUS BRACHIOPODS

871

stratigraphical study of the Upper Cretaceous-Lower Tertiary brachiopod fauna of northwestern Europe.
Unpublished Ph.D. thesis. University of Copenhagen.

— 1987a. Brachiopods from the Maastrichtian-Daman boundary sequence at Nye Klov, Jylland, Denmark.
Fossils and Strata , 20, 1-99.

— 19876. The micromorphic brachiopod genus Rugia Steinich from the Middle Coniacian - Lower
Maastrichtian chalk of Lagerdorf and Kronsmoor, Northwest Germany. Mitteilungen aus dem geologischen
Staatsinstitut Hamburg , 163, 127-185.

1988. Brachiopod extinctions in the Upper Cretaceous to lowermost Tertiary Chalk of Northwest

Europe. Revista Espanola de Paleontologia , Numero Extraordinario, Palaeontology and Evolution:
Extinction Events, 41-56.

linnaeus, c. 1758. Systema naturae, l. Salrii, Stockholm, 823 pp.

lundgren, b. 1885. Undersokningar ofver Brachiopoderna i Sveriges Krit-system. Arsskrift Lunds Universitet ,
20, 1-72.

morris, J. 1851. Palaeontological Notes. Annals and Magazine of Natural History, (8), 2, 85-90.
munier-chalmas, c. 1880. Thecididae, d’Orbigny et Koninckidae Munier-Chalmas. Bulletin Societe Geologique
en France, 3, 8, 279.

murphy, M. A. 1977. On time-stratigraphic units. Journal of Paleontology, 51, 213-219.
nestler, h. 1975. Die Fossilien der Rtigener Schreibkreide. Die Neue Brehm-Biicherei, 486, 1-120.
nielsen, k. b. 1909. Brachiopoderne i Danmarks Kridtaflejringer. Det Kongelige danske Videnskabernes
Selskab, 7 (6), 129-178.

Owen, e. f. 1970. A revision of the brachiopod subfamily Kingeninae Elliott. Bulletin of the British Museum
( Natural History), (Geology), 19, 29-83.

pajaud, d. 1974. Biostratigraphie des Thecidees (Brachiopodes) donnees recentes ou inedites. Geobios, 7,
323-344.

peake, N. b. and Hancock, J. m. 1961. The Upper Cretaceous of Norfolk. 293-339. In larwood, g. p. and
funnell, B. M. (eds). The Geology of Norfolk. Transactions of the Norfolk and Norwich Natural History
Society , 19, 269-375.

- 1970. The Upper Cretaceous of Norfolk. 293-339. In larwood, g. p. and funnell, b. m. (eds). The
Geology of Norfolk. Transactions of the Norfolk and Norwich Natural History Society, 19, 269-375
(reprinted with supplement).

pettitt, n. e. 1950. A monograph on the Rhynchonellidae on the British Chalk. Part 1. Palaeontographical
Society Monograph, 1-26.

— 1953. A monograph on the Rhynchonellidae of the British Chalk. Part 2. Palaeontographical Society
Monograph , 27-52.

posselt, h. j. 1894. Brachiopoderne i den danske Kridtformation. Danmarks Geologiske Undersogelse, Second
Series , 4, 1-59.

roemer, F. a. 1841. Die Versteinerungen des norddeutschen Kreidegebirges. Elannover, 145 pp.
rowe, a. w. 1908. Zones of the White Chalk of the English Coast. V. The Isle of Wight. Proceedings of the
Geologists' Association , 20, 209-352.

sahni, m. r. 1925. Morphology and zonal distribution of some Chalk terebratulids. Annals and Magazine of
Natural History, (9), 15, 353-385.

1929. A monograph of the Terebratulidae of the British Chalk. Palaeontographical Society Monograph

(for 1927), 1-62.

schloenbach, u. 1866. Beitrage zue Palaontologie der Jura und Kreide Formation im NW-lichen Deutschland.

St. II Kritische Studien liber Kreide-Brachiopoden. Paldontographica, 12, 1-66.
schlotheim, e. f. v. 1813. Beitrage zur Naturgeschichte der Versteinerungen in geognostischer Hinsicht.
Taschenbuch Gesteinische Minerale , 1, 3-134.

Schulz, m.-g. 1978. Zur Litho- und Biostratigraphie des Obercampan-Untermaastricht von Lagerdorf und
Kronsmoor (SW-Holstein). Newsletters on Stratigraphy, 7, 73-89.

— 1979. Morphometrisch-variationsstatistische Untersuchungen zur Phylogenie der Belemniten-Gattung
Belemnella im Untermaastricht NW Europas. Geologisches Jahrbuch, A47, 3-157.

1985. Die Evolution der Echiniden-Gattung Galerites im Campan und Maastricht Norddeutschlands.

Geologisches Jahrbuch, A80, 3-93.

sowerby, j. 1818. The mineral conchology of Great Britain. 2. Sherwood and Co., London, 235 pp.

1829. The mineral conchology of Great Britain. 6. Longman and Co., London, 218 pp.

sowerby, j. de c. 1823. Genera of recent and fossil shells. 4. Sherwood and Co., London, 1 15-160.
steinich, g. 1963a. Fossile Spicula bei Brachiopoden der Rtigener Schreibkreide. Geologic, 12, 604—610.

872

PALAEONTOLOGY, VOLUME 33

- 19636. Drei neue Gattungen der Subfamilie Cancellothyrinae Thomson. Geologie , 12, 732-740.

1965. Die artikulate Brachiopoden der Riigener Schreibkreide (Unter-Maastricht). Paldontologische
Abhandlungen , A2 (1), 1-200.

1967. Neue Brachiopoden aus der Riigener. Schreibkreide (Unter-Maastricht). I. Draciinae - eine neue
Unterfamilie der Cancellothyrididae Thomson. Geologie , 16, 1145-1155.

1968a. Neue Brachiopoden aus der Riigener Schreibkreide (Unter-Maastricht). II. Die Platidiidae

Thomson. Geologie , 17, 192-209.

19686. Neue Brachiopoden aus der Riigener Schreibkreide (Unter-Maastricht). III. Dalligas nobilis gen.
et spec. nov. und Kmgena sp. Geologie, 17, 336-347.

surlyk, F. 1970a. Two new brachiopods from the Danish White Chalk (Maastrichtian). Bulletin of the
Geological Society of Denmark , 20, 152-161.

- 19706. Die Stratigraphie des Maastricht von Danemark und Norddeutschland auf grund von
Brachiopoden. Newsletters on Stratigraphy , 12, 7-16.

1972. Morphological adaptions and population structures of the Danish Chalk brachiopods
(Maastrichtian, Upper Cretaceous). Det Kongelige danske Videnskabernes Selskab , 19 (2), 1-57.

— 1973. Autecology and taxonomy of two Upper Cretaceous craniacean brachiopods. Bulletin of the
Geological Society of Denmark, 22, 219-243.

1974. Life habit, feeding mechanism and population structure of the Cretaceous brachiopod genus
Aemula. Palaeogeography, Palaeoclimatology and Palaeoecology, 15, 185-203.

1975. Die Brachiopoden der Hemmoorer Schreibkreide. Palaontologisches Gesellschaft, 45te
Jahresversammlung in Hannover, 1 p.

- 1982. Brachiopods from the Campanian-Maastrichtian boundary sequence, Kronsmoor (NW Germany).
Geologisches Jahrbuch , A61, 259-277.

- 1984. The Maastrichtian Stage in NW Europe and its brachiopod zonation. Bulletin of the Geological
Society of Denmark , 33, 217-223.

— and birkelund, t. 1977. An integrated stratigraphical study of fossil assemblages from the Maastrichtian
White Chalk of NW Europe. 259-281. In kauffman, e. g. and hazel, j. e. (eds). Concepts and methods of
biostratigraphy. Dowden, Hutchinson and Ross, Stroudsburg, Pennsylvania, 658 pp.

— and johansen, M. b. 1984. End-Cretaceous brachiopod extinctions in the chalk of Denmark. Science , 223,

svendsen, N. 1979. The Tertiary/Cretaceous chalk in the Dan Field of the Danish North Sea. 112-119. In
Christensen, w. K. and birkelund, T. (eds). Symposium on Cretaceous-Tertiary boundary events. University
ot Copenhagen, Copenhagen, 25U pp.

wood, c. J. 1967. Some new observations on the Maastrichtian Stage in the British Isles. Bulletin of the
Geological Survey of Great Britain, 27, 271-288.

- 1988. The stratigraphy of the chalk of Norfolk. Bulletin of the Geological Society of Norfolk, 38, 3-120.
woodward, s. p. 1833. Outline of the geology of Norfolk. Norwich, 54 pp.

1174-1177.

M. B. JOHANSEN

F. SURLYK

Geologisk Institut

Kobenhavns Universitet

Typescript received 3 July 1989

Revised typescript received 5 February 1990

0ster Voldgade 10

1350 Kobenhavn K, Denmark

VERTEBRATES FROM THE MIDDLE TRIASSIC
OTTER SANDSTONE FORMATION OF DEVON

by A. R. MILNER, B. G. GARDINER, N. C. FRASER and M. A. TAYLOR

Abstract. New vertebrate material from the Otter Sandstone Formation of Devon includes a cleithrolepidid
fish, temnospondyl amphibians, a procolophonid reptile, and a tooth which may be attributed to a
tanystropheid reptile. The Cleithrolepididae is represented by several specimens of Dipteronotus cyphus. The
temnospondyl amphibian material includes two skull roof fragments referable to Mastodonsaurus lavisi , a skull
and fragments assignable to the benthosuchid Eocyclotosaurus sp., and one mandibular fragment of an
indeterminate large capitosaurid. The procolophonid material comprises several tooth-bearing fragments and
a partial interclavicle, and appears to represent a relatively primitive form. A single tricuspid tooth most closely
resembles those of the reptile Tanystropheus. An enigmatic elongate bone may be the neural spine of a
Ctenosauriscus- like thecodontian or the rib of a dicynodont therapsid. No fish, capitosaurid, or procolophonid
material has previously been described from the Otter Sandstone, and the benthosuchid Eocyclotosaurus has
not previously been identified in the Triassic of the British Isles. Comparison of this vertebrate assemblage with
those of Western and Central Europe and North America supports an Anisian age for the Otter Sandstone.

The Otter Sandstone Formation is exposed along the south-east coast of Devon from Otterton
Point near Budleigh Salterton to Port Royal, east of Sidmouth. It is a 118 metre thick Triassic
sequence of largely fl uviatile sandstones with included breccias and thin mudstone layers. It overlies
the Budleigh Salterton Pebble Beds and is in turn overlain by ‘marls’ of the Mercia Mudstone
Group. Warrington et a/. (1980) and Laming (1982) reviewed the stratigraphy of the British Triassic
and discussed the problems of correlating the Otter Sandstone with other Triassic sequences. The
Otter Sandstone has proved difficult to date as it includes no volcanic rocks, and attempts to obtain
spores have not been successful (Warrington 1971, and pers. comm, to M.A.T. 1984). The
discontinuity at the base of the Budleigh Salterton Pebble Beds is usually taken as the base of the
local Triassic, while palynomorphs have provided evidence for a Carnian (late Triassic) age for beds
135 m above the base of the Mercia Mudstone Group (Warrington 1971). The only evidence for a
more precise dating of the Otter Sandstone Formation is the fossil vertebrates found within it.

In the nineteenth century, vertebrate material was collected from coastal exposures of the Otter
Sandstone in two localities at least. Whitaker (1869) recovered a rhynchosaur specimen (GSM
90494) from a ‘brecciated horizon’ in the lower part of the Otter Sandstone Formation exposed in
a low cliff on the west bank of the mouth of the River Otter. Later workers (Johnston- Lavis 1876;
Metcalfe 1884; Carter 1888) found material of amphibians, reptiles, and coprolites (all BMNH)
from fallen blocks ascribed to an ‘ossiferous zone... ten feet from the top of the Sandstone’ near
High Peak, west of Sidmouth. Recently a large collection of vertebrate material (all EXEMS) has
been made by Mr P. S. Spencer, principally from the north-eastern region of the exposure from
Smallstones Point towards Sidmouth (Spencer and Isaac 1983, fig. 1) but also from a scattering of
localities from Danger Point near the mouth of the Otter River eastwards to Port Royal at the
mouth of the River Sid. Spencer and Isaac confirmed the early impression that the upper part of
the Otter Sandstone is the most fossiliferous, specifying the top 20 m or so. A few specimens have
been found in situ , but most have been recovered from fallen blocks and cannot be readily attributed
to a specific horizon within the cliff section. Spencer and Isaac (1983) gave a preliminary account
of the material which has now been subject to more detailed study. The rhynchosaur material has
been described by Dr M. J. Benton (Bristol University) (Benton 1987, 1990), while much of the

(Palaeontology, Vol. 33, Part 4, 1990, pp. 873— 892.|

© The Palaeontological Association

874

PALAEONTOLOGY, VOLUME 33

remaining new fish, amphibian and reptile material is described in the following account. Following
the systematic description of this material, the age of the Otter Sandstone Formation is discussed.

Institutional abbreviations used in this work are: BMNH, British Museum (Natural History),
London; EXEMS, Royal Albert Memorial Museum, Exeter, Devon; GSM, British Geological
Survey, Keyworth, Nottinghamshire; NMI, National Museum of Ireland, Dublin; SMNS,
Staatliches Museum fur Naturkunde, Stuttgart; WARMS, Warwick County Museum, Warwick-
shire (WCM in earlier literature).

SYSTEMATIC PALAEONTOLOGY

Class OSTEICHTHYES
Subclass ACTINOPTERYGII
Infraclass actinopteri
Superdivision neopterygii
Plesion perleidiformes
Family cleithrolepididae Wade, 1935
Genus dipteronotus Egerton, 1854

Type species. Dipteronotus cyphus Egerton, 1854

Emended diagnosis. Deep-bodied fish in which the dorsal side is marked with a hump. Scaling
characteristic, with fewer scales below lateral line than above it, and with extremely enlarged and
pointed dorsal ridge scales between skull and dorsal fin. Head much deepened posteriorly, frontals
greatly enlarged, and suspensorium upright. Orbit large, and two supraorbitals. Preoperculum
vertical with a prominent pit-line ventrally and an infraorbital process which extends behind the
triangular maxillary. Body lobe of tail reduced, rays 12-16 with expanded bases.

Remarks. Prominent dorsal ridge scales midway between the skull and dorsal fin are confined to
Pseudobeaconia and Dipteronotus among cleithrolepidids. In D. cyphus they were originally
mistaken for the bases of fin-rays (Egerton 1854, p. 369). Two further species of Dipteronotus have
subsequently been described, namely : D. aculeatus from the Upper Buntsandstein (Scythian-
Anisian) of France and Germany (Jorg 1969; Gall et al. 1974) and D. gibbosus from the Upper
Triassic of Morocco (Martin 1980). These two species appear to be immediately related and
together they form the sister-taxon to D. cyphus (Gardiner 1988).

Dipteronotus cyphus Egerton, 1854
Text-fig. I

1854 Dipteronotus cyphus Egerton, p. 369, pi. 11, figs 1 and 2.

1910 Dipteronotus cyphus Egerton; Woodward, p. 322.

1973 Dipteronotus cyphus Egerton; Hutchinson, p. 321.

1974 Dipteronotus cyphus Egerton; Gall et al., pp. 7-8, pi. 5.

Holotype. GSM 18188 and 18189, counterparts of an almost complete specimen, from the Bromsgrove
Sandstone Formation, Middle Triassic of Bromsgrove, Hereford and Worcester (previously Worcestershire).

Referred material. NMI F.l 5677, a small fragment of body from the type locality. EXEMS 60/1985.293 and
294, two partial specimens in a fine mudstone from the Otter Sandstone Formation (Text-fig. 1). These
specimens are better preserved than the holotype and have permitted a more accurate count to be made of the
various fin-rays as well as a more detailed description of the scale ornamentation. Further specimens were
collected and are in the EXEMS collection.

Emended diagnosis (based on holotype and EXEMS 60/1985.293 and 294). A Dipteronotus with

MILNER ET AL.\ TRIASSIC VERTEBRATES FROM DEVON

875

text-fig. 1. Dipteronotus cyphus Egcrton. a, EXEMS 60/1985.293, entire specimen, x 1-5. b, EXEMS

60/1985.294, detail showing caudal fin, x3.

10-13 elongate, spinose, ridge scales in front of dorsal fin; ventral ridge scales similar, but smaller
than dorsals. Dorsal fin comprising 17 rays and 5 basal fulcra, and anal fin with about 14 rays and
2 fulcra. Tail with 27-30 rays of which the most dorsal 12 are epaxial and with three prominent basal
fulcra dorsally and ventrally; pelvic fin with 5-6 rays. Dermosphenotic tall and narrow, and 2
infraorbitals. A single suborbital probably present between top of preoperculum and dermo-
sphenotic. Scales with posterior crimping, those on dorsal hump having several rows of

876 PALAEONTOLOGY, VOLUME 33

backwardly-pointing tubercles. On rest of body, scales are pectinate posteriorly with some ten
projections.

Class AMPHIBIA
Order temnospondyli
Family mastodonsauridae Lydekker, 1885
Genus mastodonsaurus Jaeger, 1828

Type species. Salamandroides giganteus Jaeger, 1828

Mastodonsaurus lavisi (Seeley) Paton, 1974 nomen dubium

Text-fig. 2a-c

1876 Labyrinthodon lavisi Seeley, p. 278, pi. 19, figs 1-3.

1916 Labyrinthodon lavisi Seeley; Wills, p. 15.

1947 '’Labyrinthodon ' lavisi Seeley; Romer, p. 228.

1962 ‘ Labyrinthodon ’ lavisi Seeley; Watson, p. 252.

1965 Labyrinthodon lavisi Seeley; Welles and Cosgriff, p. 35 as a nomen vanum.
1974 Mastodonsaurus lavisi (Seeley); Paton, p. 282, figs 14b and 16b.

text-fig. 2. Mastodonsaurus lavisi (Seeley), a, EXEMS 60/1985.287, left postfrontal and parietal, b, c,
EXEMS 60/1985.309, left tabular, squamosal and part of supratemporal in (b) dorsal and (c) ventral view.
Abbreviations for Text-figs 2-4: FR, frontal; L, lacrimal; J, jugal; PAR, parietal; PO, postorbital; POF,
postfrontal; PP, postparietal ; PRF, prefrontal; QJ, quadratojugal ; SQ, squamosal; ST, supratemporal; TAB,

tabular. Scale bars, 10 mm.

Holotype. BMNE1 R.4215, posterior region of right mandible, collected by H. J. Johnston-Lavis from the Otter
Sandstone Formation below High Peak, west of Sidmouth, Devon (Johnston-Lavis, 1876). Most recently
figured by Paton (1974, figs 14b and 16b).

Remarks. Up to 1974, the specific name lavisi was restricted to a mandibular fragment from the Otter
Sandstone. Paton (1974) revised the British Triassic temnospondyl material and combined the Midlands and
Devon mastodonsaurid specimens as one species for which the senior specific name was lavisi borne by this
specimen. The following description is restricted to Otter Sandstone specimens and unity ol the Midlands
material with M. lavisi is not assumed, although some comparison is made with specimens from Warwickshire.

Diagnosis. No species-level diagnosis can be given for the Otter Sandstone Mastodonsaurus. Paton’s

MILNER ET AL.\ TRIASSIC VERTEBRATES FROM DEVON

877

(1974) diagnosis of M. lavisi is based on Mastodonsaurus generic characters which permit it to be
distinguished only from contemporaneous capitosaurid material. Pending a revision of the species-
level taxonomy of the German Mastodonsaurus material, there is no basis for diagnosing M. lavisi,
but neither can it be assumed to be indeterminable. It seems most practical to retain the name, albeit
as a nomen dubium, as a convenient label for the Otter Sandstone mastodonsaurid material.

Referred material from the Otter Sandstone. As Mastodonsaurus lavisi is a nomen dubium , the following
Mastodonsaurus material is referred to that species solely on the basis of locality and horizon: BMNH R.331,
posterior region of right mandible collected by H. J. Carter from the type locality (Carter 1888); four boxes
of fragments collected by Carter are registered as R.331, but the mastodonsaur mandible is the only significant
specimen; referred to M. lavisi by Paton (1974); EXEMS 60/1985.287, left postfrontal and parietal (Text-fig.
2a), collected by P. S. Spencer; EXEMS 60/1985.309, left tabular and parts of squamosal and supratemporal
(Text-fig. 2b, c), collected by P. S. Spencer, and referred to by Spencer and Isaac (1983) as a fragment of
Cyclotosaurus.

Description of new material. EXEMS 60/1985.287 (Text-fig. 2a) comprises the attached left postfrontal and left
parietal of a large temnospondyl. Scaling these bones against a skull of Mastodonsaurus cappelensis figured by
Wepfer (1923) suggests a midline skull length of 290 mm. The two bones bear coarse shallow sculpture with
narrow ridge-systems. Some of the confluent sculpture pits on the orbital margin of the postfrontal may
represent a lateral-line sulcus. The anterior end of the postfrontal is incomplete and it is impossible to ascertain
whether it contacted the prefrontal or not. The postfrontal is relatively elongate and is almost as large as the
parietal. The pineal foramen is situated about halfway along the parietals and hence well behind the posterior
edge of the orbits.

The dermal sculpture resembles that of mastodonsaurids rather than capitosaurids and the large elongate
postfrontal resembles that of the early mastodonsaurid Mastodonsaurus cappelensis from the Anisian of
Kappel near Villingen, Baden-Wiirttemberg, West Germany. Using WARMS Gz.20, a larger skull fragment
of the Warwickshire Mastodonsaurus material, Paton (1974, table 1 ) was able to show that it also corresponded
to M. cappelensis in orbit shape, size and position. In capitosaurids, the orbit length was 45-85% of the
interorbital width, in WARMS Gz.20 and M. cappelensis 120-125%, in the larger Ladinian Mastodonsaurus
species over 200%. Unfortunately, no comparative study has been undertaken of the various Mastodonsaurus
specimens from the Anisian, Ladinian and Carnian of Baden-Wiirttemberg and we do not know whether
proportional differences such as those described above are of systematic significance or simply related to
absolute size.

EXEMS 60/1985.309 (Text-fig. 2b, c) comprises the left tabular and parts of the squamosal and
supratemporal of a very large temnospondyl. Scaled against figures of Mastodonsaurus skulls, it appears to be
derived from a skull of about 600 mm midline length. The suture between the tabular and the other bones is
densely convoluted. The sculpture on the dorsal surface of the bones is coarse and shallow. The tabular is a
curved, slightly elongate bone thinning out at its distal extremity and clearly does not suture with any other
bone behind the otic notch. The curvature of the tabular is such that the distal end was laterally rather than
postero-laterally directed. The deep open otic notch, partly bordered by a long curved tabular bearing coarse
sculpture, is characteristic of large specimens of Mastodonsaurus, but not of any particular species. Some
specimens of M. giganteus from Gaildorf (e.g. SMNS 4707 and 54679) are very similar and the specimen is
closely comparable to WARMS Gz.9 from Warwickshire (Paton 1974, fig. 6) which Paton attributed to M.
lavisi. The new undescribed Mastodonsaurus material from the Lettenkeuper of Kupferzell (Wild 19806)
includes skulls over a wide size range which demonstrate that the tabular is small in skulls of 300 mm length
but elongates later in ontogeny (A.R.M. pers. obs.). A few capitosaurids have similar tabulars (e.g.
Parotosuchus nasutus Welles and Cosgriff 1965, fig. 19 and Parotosuchus megarhinus Chernin 1974, fig. 8) but
the dermal sculpture is finer, deeper and more reticulate.

Discussion. The Mastodonsauridae is at present a neglected family. The full range of German
Mastodonsaurus material has not been subjected to comprehensive comparative study this century
and meaningful comparisons are still difficult. Beyond the genus Mastodonsaurus, there is
uncertainty as to what constitutes a primitive mastodonsaurid. Some capitosauroids may in fact be
primitive mastodonsaurids with only one or two of the specializations of Mastodonsaurus.
Kestrosaurus dreyeri Haughton, 1925 has generally been considered to be an aberrant capitosaurid,
but Ochev (1966) placed it in the new family Bukobajidae which Shishkin (1980) later reduced in

878

PALAEONTOLOGY, VOLUME 33

synonymy with the Mastodonsauridae. ‘ Parotosaurus ' lapparenti Lehman, 1971 was also assigned
to the Mastodonsauridae by Shishkin (1980). Both of these forms have double anterior palatal
vacuities and laterally directed tabulars, and appear to be primitive mastodonsaurids with a
gradistic resemblance to Benthosuchus.

The following observations can be made, however. Firstly, no mastodonsaurid-producing locality
is known to have more than one species present. As the holotype of M. lavisi is a mandible from
the Otter Sandstone, other mastodonsaur fragments from this formation can be placed under the
name M. lavisi with confidence. Secondly, the proportions of the orbits and the circumorbital bones
of specimens EXEMS 60/1985.287 from Devon and WARMS Gz.20 from Warwickshire both show
the closest resemblance to the primitive Mastodonsaurus cappelensis Wepfer, 1923 from the Upper
Bunter of Kappel in Baden-Wiirttemberg, West Germany. This does not demonstrate the unity of
the British material from the two areas not does it demonstrate synonymy with the German species.
It suggests either that all three populations belong to a similar primitive evolutionary grade within
the genus Mastodonsaurus or alternatively that all three populations are simply of similar-size
specimens with similar orbital proportions. Because of the latter possibility, it is best to maintain
the Warwickshire material as Mastodonsaurus sp. indet. (as no species name uniquely applies to it),
the Devon material as M. lavisi, and the Kappel material as M. cappelensis , but the three
populations may represent from one to three species of mastodonsaurid. Consequently,
stratigraphical conclusions based on such material should be treated with great circumspection.

Family benthosuchidae Efremov, 1940
Genus eocyclotosaurus Ortlam, 1970

Type species. Eocyclotosaurus woschmidti Ortlam, 1970

Eocyclotosaurus sp.

Text-figs 3 and 4.

Material. EXEMS 60/1985.72, damaged posterior skull roof with internal cast of left cheek (Text-fig. 3), from
the Otter Sandstone below High Peak near Sidmouth, Devon and reported by Spencer and Isaac (1983) as a
skull of Cyclotosaurus ; EXEMS 60/1985.310, right tabular and parts of squamosal, supratemporal and
postorbital (Text-fig. 4a); EXEMS 60/1985.75, an incomplete left clavicle (Text-fig. 4b), referred here with
doubt; all collected by P. S. Spencer.

Description. EXEMS 60/1985.72 (Text-fig. 3), the most complete amphibian specimen yet collected from the
Otter Sandstone Formation, is a partial skull of a small capitosauroid temnospondyl. Regions represented by
bone include the skull table and parts of the interorbital area and cheeks; the rest of the cheeks and the right
posterior region of the snout are represented by a weathered endocast, the part exposed at the time of
collection. Most of the snout had already been weathered away prior to collection. The dermal bones of the
right cheek, skull roof and interorbital region were all present in situ but were unfortunately damaged during
excavation and have been reassembled. The palate is not exposed. The skull is estimated to have been 250 mm
long, 150 mm wide from quadrate to quadrate, and 30 mm deep at the back, hence very low in profile. In most
respects it is a typical capitosauroid skull and only key features need be discussed.

The dermal bones bear deep reticulate sculpture. There are no well-preserved lateral-line sulci although
elongate pits on the supratemporals may represent poorly expressed sulci. The otic notch is completely closed
by a stout extension of the tabular suturing with the squamosal. The length of this suture is about equal to the
diameter of the otic notch. The posterior edge of the skull table is concave but appears to consist of two straight
edges meeting medially at a distinct obtuse angle. The distal margin to the orbit is represented only by endocast
on both sides, but on the left side the suture-lines are clearly visible. They show that the jugal was excluded
from the orbit margin by an anterior extension of the postorbital extending forwards to the anterior margin
of the orbit. Although the prefrontal shape is not determinable, it is probable that the postorbital sutured with
the prefrontal ahead of the orbit; the postorbital has the same anterior extent as in German and French
Eocyclotosaurus specimens in which suturing with the prefrontal occurs. Part of the corresponding suture is

MILNER ET AL.: TRIASSIC VERTEBRATES FROM DEVON

879

text-fig. 3. Eocyclotosaurus sp. EXEMS 60/1985.72, incomplete skull roof in dorsal view. Abbreviations: see

Text-figure 2. Scale bar, 10 mm.

visible on the endocast on the right side as well (Text-fig. 3). The relationship between the prefrontal, the
postfrontal and the frontal is unfortunately difficult to observe because of damage to the bones. On the right
side, the pre- and postfrontals are represented only by damaged scraps of bone but the frontal appears to be
inset with an exposed sutural edge along its distal margin. This indicates that the pre- and postfrontals
probably met mesially to the orbit and excluded the frontal from the orbit margin. Although the snout is poorly
preserved, it can be seen that the sides converge anteriorly, which implies a narrow anterior snout, rather than
a broad parallel-sided snout.

Discussion. The specimen is clearly one of the capitosaurid or benthosuchid temnospondyls with a
completely enclosed otic notch. In the Triassic of western Europe, the genera Cyclotosaurus ,
Stenotosaurus, and Eocyclotosaurus all have such notches. Their characteristics and relationships
have recently been discussed by Kamphausen and Morales (1981), Kamphausen (1983) and
Morales and Kamphausen (1984). The apparently narrowing snout and the anteriorly extending
postorbital excluding the jugal from the orbit margin characterize the genera Eocyclotosaurus and
Stenotosaurus. Cyclotosaurus species have the jugal entering the orbit margin and have broad
snouts. The prefrontal-postfrontal contact identifies the specimen as Eocyclotosaurus rather than
Stenotosaurus in which the frontal enters the orbit margin. These character-states suggest that the
specimen is a skull of Eocyclotosaurus and all other visible features are consistent with this identity.

Kamphausen and Morales (1981) established the generic unity of the Eocyclotosaurus material
from the Upper Bunter of France and West Germany but did not attempt to provide separate
diagnoses for the two species E. lehmani (Heyler) and E. woschmidti Ortlam. There are no obvious
discriminating features between the species as described and the problem of whether this material

880

PALAEONTOLOGY, VOLUME 33

text-fig. 4. Eocyclotosaurus sp. a, EXEMS 60/1985.310, fragment of right skull table, b, EXEMS 60/1985.75,
incomplete left clavicle. Abbreviations: see Text-figure 2. Scale bar, 10 mm.

represents one or two species is still being studied by Kamphausen. For this reason, the Otter
Sandstone material is not assigned to a species but simply identified as Eocyclotosaurus sp.

Possible further referred material. The following two specimens are not in themselves diagnostic but are
consistent with attribution to Eocyclotosaurus sp. and can be referred here with various degrees of confidence.

EXEMS 60/1985.310 (Text-fig. 4a) is a fragment of dermal skull roof from the region of the right otic notch.
It is made up of an abraded right tabular and well-preserved parts of the right supratemporal, squamosal, and
postorbital. The sculpture is relatively deep and reticulate. The supratemporal has broken along the lateral-line
sulcus and only the distal part of the bone remains. The sulcus can be seen to be deep but its width cannot be
determined. The otic notch is closed, the tabular running behind it as a stout structure, broadening distally
towards the region where it would meet the squamosal. The specimen appears to be identical in all significant
respects with the corresponding region in EXEMS 60/1985.72 and may also be referred to Eocyclotosaurus sp.

EXEMS 60/1985.75 (Text-fig. 4b) is an incomplete left clavicle of a capitosauroid temnospondyl. In shape,
size, and texture of sculpture it is most consistent with attribution to the cranial material described above and
is cautiously referred here.

Family capitosauridae Watson, 1919
Capitosauridae incertae sedis

Text-fig. 5

Material. EXEMS 60/1985.78, posterior region of right mandible (Text-fig. 5a, b), collected by P. S. Spencer,
and referred to by Spencer and Isaac (1983) as a mandible of Cyclotosaurus.

Description. This specimen comprises the greater part of the post-dentary region of a massive right mandible.
The preserved elements are the surangular, prearticular, articular, and the posterior region of the angular. The
angular bears deep coarse sculpture which extends slightly on to the surangular. A broad, shallow lateral-line
sulcus, the mandibular sulcus, extends back from the ventrodistal side of the angular and curving dorsally
behind it up to the dorsal side of the retroarticular process (Text-fig. 5a). Two branches extend forward along
the external face of the surangular, the lower is the oral sulcus and the upper is the accessory sulcus
(terminology as in Jupp and Warren 1986). Behind the mandibular sulcus, the external face of the retroarticular
process has a few striations but does not have a sculptured surface. The ventral edge of the mandible is narrow

MILNER ET AL.: TRIASSIC VERTEBRATES FROM DEVON

881

text-fig. 5. Capitosauridae incertae sedis. EXEMS
60/1985.78, posterior region of right mandible, a,
buccal view, b, lingual view. Abbreviations: ANG,
angular; ART, articular; PRA, prearticular; SUR,
surangular. Scale bar, 10 mm.

and sharply concave. The retroarticular process is broadly rounded in lateral aspect, and is a slightly elongated
triangle in dorsal view. The articular surface with the quadrate is saddle shaped. The foramen for the chorda
tympani is present and set high on the side of the jaw.

Discussion. Comparison with the mandibular fragments described and tabulated by Paton (1974,
table 2) shows complete correspondence to the Warwickshire specimens which she assigned to
Cyclotosaurus pachygnathus. The nature of the sculpture and the lateral-line sulci, the high position
of the chorda tympani foramen, the absence of sculpture on the retroarticular process and the shape
of that process all serve to identify it as a capitosaurid or benthosuchid rather than a
mastodonsaurid. It is clearly not attributable to M. lavisi. The massive construction and the saddle-
shaped articular surface correspond to those of a large Cyclotosaurus or Parotosuchus rather than
a lightly built Stenotosaurus or a benthosuchid. It has all the characteristics of those Warwickshire
mandibles such as GSM 27964 that Paton assigned to Cyclotosaurus pachygnathus. However,
referring this specimen to C. pachygnathus is unsatisfactory because, by Paton’s (1974, p. 280) own
admission, C. pachygnathus is an undiagnosed aggregation of large cyclotosaur material from
Warwickshire placed under one name for systematic stability and general convenience. Some of it
certainly pertains to a Stenotosaurus or Cyclotosaurus of massive construction, but is of uncertain
unity, cannot be comparatively diagnosed with the German material, and is effectively
Capitosauridae incertae sedis , with the name C. pachygnathus a nomen dubium restricted to the
lectotype specimen designated by Paton. However, much of the Warwick material is not
demonstrably cyclotosaurian. In assuming the unity of this capitosaurid material, Paton did not
attempt to compare any of it with open-notched Parotosuchus and Eryosuchus , both of which are
present in Middle Triassic assemblages. A fragment of massive capitosaurid which does not include
the otic notch is as likely to be Parotosuchus or Eryosuchus as Cyclotosaurus. Morales (1987) has
noted that there is no certain Cyclotosaurus material from pre-Carnian beds and that ‘ Cyclotosaurus ’
randalli Welles, 1947 from the Middle Triassic of the Moenkopi Formation is indeterminate but
probably a benthosuchid. As the Otter Sandstone specimen is from a geographically distinct
locality, and as unwarranted stratigraphical assumptions might be drawn if it and the Warwickshire
material were assigned to the same genus or species, it is simply assigned to Capitosauridae incertae
sedis.

882

PALAEONTOLOGY, VOLUME 33

Indeterminate temnospondyl fragments

Several other specimens from the Otter Sandstone clearly derive from temnospondyl amphibians but are not
determinable. Six of these specimens which have numbers prefixed by EXEMS 60/1985, are as follows: .4, .79,
.183, labyrinthodont teeth or palatal tusks ; .96, .148, fragments of sculptured dermal bone, probably from skull
roof; .308, half of a small interclavicle. A seventh specimen is EXEMS 7/1986.2, the posterior half of the right
prefrontal of a benthosuchid or capitosaurid.

Class REPTILIA
Order procolophonia
Family procolophonidae Cope, 1889

Procolophonidae incertae sedis

Text-figs 6 and 7a, b

Material. EXEMS 60/1985.87, a right dentary fragment (Text-fig. 6a), referred to by Spencer and Isaac (1983)
as a specimen of the procolophonid Leptopleuron ; EXEMS 60/1985.154, a right dentary (Text-fig. 6b),
reported as a sphenodontid by Spencer and Isaac (1983); EXEMS 60/1985.311, an incomplete right dentary
(Text-fig. 6c), EXEMS 60/1985.3, a fragmentary left maxillary (Text-fig. 7a), reported by Spencer and Isaac
(1983) as a sphenodontid; EXEMS 60/1985.9, an incomplete interclavicle (Text-fig. 7b), correctly identified by
Spencer and Isaac (1983). All specimens were collected by P. S. Spencer.

text-fig. 6. Procolophonidae incertae sedis. a, EXEMS 60/1985.87, fragment ol right dentary. b, EXEMS
60/1985.154, juvenile right dentary. c, EXEMS 60/1985.311, juvenile right dentary. All in lateral view.
Abbreviations: CO. PR., coronoid process; D, dentary. Scale bar, 5 mm.

Description. EXEMS 60/1985.87 was the only jaw specimen correctly identified by Spencer and Isaac (1983)
as procolophonid, but whether it can be attributed to Leptopleuron , as they suggest, is questionable. The
dentary, which is predominantly exposed in lateral aspect (Text-fig. 6a), is 9 0 mm long and, in addition to the

MILNER ET AL.: TRIASSIC VERTEBRATES FROM DEVON

883

two posteriormost teeth, it retains the broken base of a coronoid process. Superficially at least, the broadly-
based obtusely conical teeth display an acrodont attachment. The posterior tooth is 2-7 mm high and that in
front of it is 2-4 mm high. Their bulbous appearance in lateral aspect is typically procolophonid. The teeth of
this specimen (the only one in which they are completely exposed from the matrix) are not transversely
broadened. However the maximum development of transverse broadening in procolophonids is seen most
frequently in the teeth immediately anterior to these positions.

Two further dentary fragments are more complete, but somewhat smaller than EXEMS 60/1985.87, and
they are regarded here as juveniles. EXEMS 60/1985.154 is 12 mm long and retains nine teeth and a
pronounced coronoid process (Text-fig. 6b). The element is complete anteriorly, with the mandibular
symphysis partly embedded in matrix. Spencer and Isaac (1983, p. 268) referred to this specimen as a
sphenodontid, but it lacks typical sphenodontid dental characters such as tooth flanges, distinct successional
hatchling and additional tooth series, and at least partial alternation in tooth size (Fraser 1986a; Whiteside
1986), nor is there a marked posteriorward increase in tooth height, nor any indication of extensive lateral wear
facets correlated with the shearing jaw action so typical of sphenodontids (Robinson 1973 ; Gorniak et al. 1982;
Fraser and Walkden 1983). Where exposed, wear facets appear to be confined to the apical surfaces of each
tooth. The prominent incisiform anterior teeth are inclined anteriorly and somewhat laterally, a characteristic
readily observed in Procolophon and Orenburgia.

EXEMS 60/1985.311 (Text-fig. 6c) is a third dentary fragment, also exposed from the right side. It is
incomplete posteriorly and lacks the coronoid process, but the six exposed teeth are rather better preserved
than in the previously described specimens. In all specimens, tooth attachment appears to be acrodont. The
first two teeth are equal in height to the more posterior members of the tooth row, yet they are markedly
narrower and thus incisiform in appearance. By contrast, the posterior three teeth are broader based, and in
lateral aspect they are widest at a point approximately mid-way up the crown. The resulting profile has been
described as barrel-shaped by Malan (1963) and characterizes most procolophonids (e.g. Procolophon ,
Thelegnathus , Tichvinskia).

text-fig. 7. a and b, Procolophonidae incertae sedis\ a, EXEMS 60/1985.3, left maxillary in lateral view; b,
EXEMS 60/1985.9, interclavicle in ventral view, c, ITanystropheus sp., EXEMS 60/1985.143, an isolated
tooth. Abbreviations: A.C., accessory cusps; MX, maxillary. Scale bar, 2 mm.

EXEMS 60/1985.3 is a maxillary fragment bearing four teeth (Text-fig. 7a) which Spencer and Isaac (1983)
referred to as a sphenodontid. They believed it to represent a right element in association with a partly
preserved jugal. Presumably they took a faint line running across the specimen as the suture between the two
bones. However the dental morphology is consistent with that of a procolophonid and, other than the acrodont
implantation, the dentition shows no sphenodontid characteristics. The enlarged orbit of procolophonids is
characterized by the great expansion of the posterodorsal corner so that the ventral margin, normally formed
by the jugal and (to a variable extent) the maxillary, gently slopes posterodorsally. The preserved orbital

884

PALAEONTOLOGY, VOLUME 33

margin of EXEMS 60/1985.3 is too abrupt for it to be anything but the anterior margin. Furthermore the slight
recurvature of the teeth supports the interpretation of this fragment as part of a left-sided element.
Consequently, if the line referred to above is a suture, then it is between the lacrimal and maxillary rather than
the jugal and maxillary. However the line is poorly defined and it seems more plausible to attribute it to post-
mortem damage.

EXEMS 60/1985.9 is an incomplete T-shaped interclavicle (Text-fig. 7b) which is characteristically
procolophonid. Although the interclavicles of sphenodontids are also T-shaped, they are usually much more
slender and they lack the broad median bar that is so prominent in EXEMS 60/1985.9.

Discussion. There is nothing to suggest that more than one procolophonid species is represented in
the Otter Sandstone, and the variation in size can be satisfactorily explained as ontogenetic. The
only characteristics which may be of systematic or stratigraphical value are those of the mandibular
dentition, which may be summarized as follows: the mandible possesses nine acrodont teeth (a
primitively high number within the Procolophonidae), the anterior incisiform teeth are inclined
anterolaterally, and the posterior molariform teeth are barrel-shaped (procolophonid characters)
but appear to lack the derived conditions of being cuspidate or transversely broadened. The Otter
Sandstone material thus appears to belong to a primitive procolophonid and its closest resemblance
is to the Middle Triassic (Anisian) Anisodontosaurus greeri from the Holbrook Member of the
Moenkopi Formation of Northern Arizona. Although Anisodontosaurus was recorded by Welles
(1947) as Reptilia incertae sedis, and by Murry (1987) as a primitive sphenodontid, it appears to
be a primitive procolophonid. The dentary bears twelve teeth differentiated into incisiforms and
broad-based molariforms. Welles (1947) described the two posteriormost teeth of Anisodontosaurus
as acrodont in implantation, but noted that the more anterior teeth project further down into the
dentary and approach a thecodont condition. However, it is sometimes impossible clearly to
distinguish between different forms of tooth implantation (Fraser and Shelton 1988). Although
procolophonids are normally considered to be acrodont, the teeth in some specimens appear to be
positioned in shallow sockets (Fraser 1986/?), not unlike the tooth implantation described in
Anisodontosaurus. Although the former possesses rather more teeth, Anisodontosaurus and the Otter
Sandstone procolophonid are primitively more similar to each other than to any other
procolophonid.

Typical early Triassic procolophonids such as Procolophon, Koiloskiosaurus, Sclerosaurus ,
Eumetabolodon , and Thelegnathus are more derived than the Otter Sandstone form in that they have
transversely broadened ‘molariform’ teeth, and a rudimentary bicuspidate condition is sometimes
present. The dentitions of Koiloskiosaurus and Sclerosaurus from the Upper Buntsandstein of
Germany are incomplete but, although they appear to possess between seven and nine teeth in each
jaw ramus, at least some of these are transversely broadened (Huene 191 1, 1943). In late Triassic
forms, the bulbous profile is accentuated and the molariform teeth are invariably transversely
broadened and either bicuspidate ( Leptopleuron ) or tricuspidate (forms in the English fissure
sediments (Fraser 1 986/?)).

Some early and most late Triassic procolophonids show a marked trend towards reduction in
overall tooth numbers. The third Buntsandstein genus, Anomoiodon (Huene 1939), is small with a
skull length just over 30 mm and hence close in size to the Otter Sandstone material. Anomoiodon
resembles the Otter Sandstone material in that all the marginal teeth are conical and no transverse
broadening is apparent. On the other hand, despite the incomplete preservation, there would appear
to have been no more than five teeth in the mandible in contrast to nine in EXEMS 60/1985.54.
Thus there is no clear basis for assigning the Devon specimens to any one of the contemporaneous
German genera. The Otter Sandstone procolophonid material is too fragmentary to be determinable
below family level, but it does appear to belong to a very primitive number of the family.

Division archosauromorpha
Order prolacertiformes
Family tanystropheidae Romer, 1945
Genus tanystropheus Meyer, 1852

885

MILNER ET A L. \ TRIASSIC VERTEBRATES FROM DEVON
cf. Tanystropheus sp.

Text-fig. 7c.

Material. EXEMS 60/1985.143, a tricuspid tooth (Text-fig. 7c), collected by P. S. Spencer, and reported as a
therapsid tooth by Spencer and Isaac (1983).

Description. A single tricuspid tooth, approximately TO mm high, bearing two small accessory cuspules on
either side of the main cusp. There is some resemblance to the large teeth of the Norian pterosaur
Eudimorphodon (Wild 1978), but the latter are considerably larger and more elongate. A better comparison
is with the teeth of the prolacertiform Tanystropheus (Wild 1973, 1980a). Wild (1980a, fig. 4) figures several
teeth from a Tanystropheus skull including the fourth dentary tooth (fig. 4c) which is almost identical to
60/1985.143 in overall shape and relative size of cusp and cuspules.

Class AMNIOTA incertae sedis

Material. EXEMS 60/1985.88, an elongate bone collected by P. S. Spencer. Referred to by Spencer and Isaac
(1983, p. 269) as a neural spine of ‘the pelycosaur Ctenosaurus

Description. This specimen is a long straight bone, 5 1 4-9 mm in length according to Spencer and Isaac, but
broken at both ends. It has a slightly flattened figure-of-eight cross-section and a surface pattern of fine ridges
and foramina. It seems to be part of the axial skeleton of a large reptile, but beyond this general identification,
it has proved difficult to place anatomically or systematically. As well as two of the authors (A.R.M. and
M.A.T.), several colleagues, including Drs M. J. Benton, A. R. I. Cruickshank, G. M. King, and D. B.
Norman have examined the specimen. There is no basis for attributing it to the pelycosaurs, and in fact the
Triassic ‘ Ctenosaurus ’ material, now renamed Ctenosauriscus , has been shown by Krebs (1969) to be one of
a series of long-spined Middle Triassic thecodontian-grade archosaurs. In the specimen from the late Middle
Buntsandstein figured by Krebs (1969, pi. 1), the tallest neural spine is a long straight flattened bone about
640 mm long. The Otter Sandstone specimen could be a long neural spine of such a ctenosauriscid or a rib of
a large dicynodont therapsid such as a kannemeyeriid. Its size is compatible with either identity but it appears
to be somewhat narrow in relation to its length for either interpretation. There is slight asymmetric curvature
at one end which seems to be too great for a ctenosauriscid spine and too little for a dicynodont rib. The
identity of this specimen is left open.

STRATIGRAPHY

Recent history

As noted in the introduction, the only available evidence for dating the Otter Sandstone is the
vertebrate fossil fauna itself. Previously, only the few fragments of rhynchosaur and temnospondyl
collected in the nineteenth century have been considered for the purposes of dating. Walker (1969)
noted the Sidmouth rhynchosaur as ‘a more primitive species’ of Rhynchosaurus than those known
from the Midlands and Shropshire. The Midlands Rhynchosaurus derives from the Bromsgrove
Sandstone Formation of the Sherwood Sandstone Group, while the Shropshire material derives
from the Helsby Sandstone Formation of the Sherwood Sandstone Group and the Tarporley
Siltstone Formation of the Mercia Mudstone Group (Warrington et al. 1980). However, no precise
independent dating of the Midlands or Shropshire reptiles had then been achieved and Walker
could only suggest an age within the range of late Anisian to earliest Carnian, probably early-mid
Ladinian. Unfortunately, Walker (1969) did not explicitly state whether this dating also applied to
the Otter Sandstone material. As a result. Laming (1982) cited Walker as ascribing an early-mid
Ladinian age to the Devon rhynchosaur, whereas Warrington et al. (1980) and Spencer and Isaac
(1983) took Walker’s statement that the Devon rhynchosaur was more primitive to imply an
Anisian date for the Otter Sandstone Formation. Walker later (1970) said of the Shropshire and
Midlands animals that ‘a late Anisian, or, preferably, a Ladinian age seems to be implied’. He also
repeated his opinion that the Devon material was more primitive and tentatively suggested that the

886

PALAEONTOLOGY, VOLUME 33

Otter Sandstone Formation was older than the Shropshire and Midlands horizons, and so
presumably an Anisian age was implied. Most subsequent assignments of an Anisian date to the
Otter Sandstone Formation can be traced back to Walker’s observation (see Benton 1990 for a
discussion). The English Rhynchosaurus material, including several new specimens from the Otter
Sandstone, has recently been restudied by Benton (1987, 1990). The Otter Sandstone Rhynchosaurus
represents a distinct species from the Midlands material but the characters of the various species do
not suggest any obvious stratigraphical sequence of the British Triassic sites (Benton 1990).

Paton (1974) reviewed the temnospondyl amphibians of the English Triassic and suggested that
the Otter Sandstone vertebrates were coeval with the Shropshire and Midlands animals. She
associated Devon and Midlands Mastodonsaurus material in the single species M. lavisi which
implied that the two horizons were contemporaneous. Paton compared the mastodonsaur and the
two cyclotosaur species from the English Triassic with German material and argued that M. lavisi
and Cyclotosaurus leptognathus indicated a Scythian/ Anisian age, while Cyclotosaurus pachygnathus
indicated a Carnian/Norian age. She appears to have taken an average of the two ages, ascribing
the Midlands and Devon amphibians to the early Ladinian, after taking Walker's (1969, 1970)
reptile evidence into account. Thus the previously published situation for the age of the Otter
Sandstone Formation vertebrates is that it is either Anisian (based on Walker’s suggestion that the
rhynchosaur is more primitive than the presumed Ladinian Midlands material) or Ladinian (based
on Paton’s conclusion that the material is contemporaneous and that two of the amphibians suggest
an Anisian age and one suggests a Carnian age).

New evidence

Attempts to date the British Triassic vertebrate-bearing assemblages by comparison with the French
and German sequences are handicapped by the presence of the Muschelkalk through much of the
Anisian and Ladinian in mainland Europe. Consequently, comparisons of continental vertebrates
must either be with the Upper Buntsandstein/Rot/ Voltzia Sandstone faunas of uppermost
Scythian/basal Anisian age, or with the Lettenkohle faunas of latest Ladinian age. Middle Anisian-
Middle Ladinian continental faunas are barely represented in the Muschelkalk, and there are no
sequential series of stratigraphically restricted taxa, and so only relatively crude comparisons and
correlations are possible. Of the new material reported in this work, the capitosaurid jaw is least
useful, merely supporting a Triassic age for the Otter Sandstone. All other forms can be used to
narrow the possibilities to give a more precise assessment.

The genus Dipteronotus is known from the Scythian of Europe to the Carnian/Norian of
Morocco but the Otter Sandstone species D. cyphus is otherwise known only from the fish
assemblage from the Building Stones, the upper member of the Bromsgrove Sandstone Formation
at Bromsgrove. Warrington (1967, 1970) dated the entire Bromsgrove Sandstone in a borehole at
Bromsgrove as Scythian to early Ladinian using acritarch and miospore evidence, and this would
imply an Anisian - early Ladinian age for D. cyphus. Comparisons of Bromsgrove vertebrates with
material from Germany suggest a date at the later end of this range, particularly the presence of the
shark Palaeobates keuperinus (Acrodus sp. of Wills 1907, 1910) and the palaeoniscoid Gyro/epis
albertii (see Appendix for further details). Palaeobates keuperinus (Murchison and Strickland)
occurs at several localities in the Bromsgrove Sandstone Formation and the Mercia Mudstone
Group (it may be noted that the material of Phoebodus brodei Woodward from Shrewley consists
of symphysial teeth of Palaeobates keuperinus). P. keuperinus resembles closely P. angustissimus
(Agassiz) from the Upper Muschelkalk and Lettenkohle of Germany, Poland, and France. G.
albertii is not only found in the Bromsgrove Sandstone Formation and Mercia Mudstone Group
of the Midlands, but also continues into the Rhaetic. This species also occurs in Baden-
Wiirttemberg, Bavaria, and Thuringia where it is almost entirely confined to the Upper
Muschelkalk. Thus Dipteronotus cyphus at Bromsgrove is in a horizon of Anisian to early Ladinian
age and is associated with some fish found in the Ladinian upwards elsewhere. A late Anisian - early
Ladinian age might be inferred for the Otter Sandstone on the basis of the presence of D. cyphus.

The genus Eocyclotosaurus has the most restricted stratigraphical range of any form described

MILNER ET AL.. TRIASSIC VERTEBRATES FROM DEVON

887

here. Eocyclotosaurus species have been described from two European formations, E. lehmani from
the Voltzia Sandstone of the Vosges in France and E. woschmidti from the Lower Rot of the
Schwarzwald in West Germany. Both formations are equated with the Upper Bunter and are either
late Spathian or early Anisian (uppermost Lower or lower Middle Triassic) in age (Morales 1987).
There is also an undescribed eocyclotosaur, which is being studied by S. P. Welles and M. Morales
(Morales 1987), from the Holbrook Member of the Moenkopi Formation of Arizona, which is
believed to be early Anisian in age (Welles and Estes 1969; Morales 1987). The presence of this
genus very positively suggests an Anisian age for the Otter Sandstone.

Mastodonsaunis species have been described from the Upper Bunter (Anisian) of Kappel in West
Germany (M. cappelensis) to the Schilfsandstein (Carnian) of Feuerbacher Heide, Stuttgart ( M .
keuperinus). As discussed above, the British material from both Devon and Warwickshire has some
resemblance in interorbital proportions to the Anisian M. cappelensis within the genus. However,
in the absence of detailed comparisons among the Anisian, Ladinian, and Carnian Mastodonsaunis
material from Germany, not much weight can be given to this observation and the material merely
indicates an Anisian-Carnian date.

The procolophonid, though not precisely determinable, is clearly primitive in the number and
narrowness of the mandibular teeth, and forms with such primitive features are found only in
Scythian and Anisian horizons. Other reptiles clearly indicate a Middle Triassic age. The
comparison of the isolated tricuspid tooth with the dentition of Tanystropheus supports a Middle
Triassic assignment, Tanystropheus being known from the Anisian and the Ladinian of central
Europe (Wild 1980a). If the elongate element proves to be a ctenosauriscid neural spine, this also
supports a Middle Triassic age. Known ctenosauriscids are Ctenosauriscus from the Anisian Upper
Buntsandstein of Germany (Krebs 1969), Hypselorhachis from the Ladinian Manda Formation of
Tanzania, and Lotosaurus from the Middle Triassic of China.

A consideration of the Otter Sandstone assemblage as an association supports an Anisian age.
The association of Dipteronotus cyphus (Anisian-basal Ladinian), Eocyclotosaurus (late Scythian-
Anisian), Mastodonsaunis (Anisian-Carnian), primitive procolophonid (Scythian-Anisian), ?cteno-
sauriscid (Anisian-Ladinian) and ?tanystropheid (Anisian-Ladinian) has the Anisian as the only
shared date. The fauna is in fact reminiscent of the Scythian/Anisian late Buntsandstein and Voltzia
Sandstone faunas of Germany and France (Jorg 1969; Krebs 1969; Ortlam 1970; Gall et at. 1974;
Kamphausen and Morales 1981), although Rhynchosaurus is conspicuously absent from the latter
assemblages. The resemblance suggests that the Otter Sandstone fauna is an extension of the
Anisian Upper Buntsandstein fauna.

THE OTTER SANDSTONE FAUNA

The Otter Sandstone appears to have been laid down on a flood-plain covered with small ephemeral
braided streams and lakes (Laming 1982). The climate was semi-arid with long dry periods
punctuated by seasonal rains and flash-floods. The vertebrate material is fragmentary, clearly
transported and current-sorted, which can be inferred from the composition of the assemblage;
teeth, tooth-bearing elements, and dense pieces of skull and pectoral girdle. The notable exception
is the Dipteronotus material which is found as intact specimens in a laminated mudstone. Because
of the derived nature of the assemblage, it bears an uncertain relationship to the original fauna (or
faunas) and only a few observations can be made.

There seems to be distinct aquatic and terrestrial components to the fauna. The three types of
temnospondyl amphibian were all superficially crocodile-like forms and all belong to families which
are believed to have been fundamentally aquatic. The presence of three types of aquatic carnivore,
some growing to at least 2 metres in length, suggests a rich fish fauna of which we currently have
only one genus. It also suggests that there must have been at least some permanent water bodies
among the aquatic environments on the flood-plain which gave rise to the Otter Sandstone.

The only frequent elements in the terrestrial fauna are the rhynchosaur Rhynchosaurus , a
medium-sized herbivore, and the procolophonid, a small herbivore. Both appear to have been

PALAEONTOLOGY, VOLUME 33

adapted for feeding on particularly durable vegetation. The long bone appears to have represented
a rarer medium-large herbivore, whether it belonged to a ctenosauriscid or a dicynodont.
Notwithstanding earlier reports, there are no sphenodontids or insectivorous/carnivorous
therapsids in the collected material. Spencer and Isaac (1983, p. 269) noted teeth and a skull
fragment of an Ornithosuchus- like medium-sized carnivore. At present, the terrestrial component of
the Otter Sandstone fauna does not appear to be diverse and perhaps consists of representatives of
a few abundant specialists living in a semi-arid flood-plain environment.

Acknowledgements . We particularly thank Mr Patrick S. Spencer for generously donating his collection to the
Royal Albert Memorial Museum. Exeter and thus making it available for this work. We are also grateful to
Mr Kelvin I. Boot (Royal Albert Memorial Museum), Dr Rupert Wild and Dr Ronald Boettcher (Staatliches
Museum fur Naturkunde in Stuttgart) and Dr Angela C. Milner (British Museum (Natural History)) for
permission to study material in their care. We further thank Dr M. J. Benton and Dr G. Warrington for
improving the manuscript and for providing information on Triassic stratigraphy , Dr Michael J. Benton, Dr
Arthur R. I. Cruickshank, Dr Gillian M. King and Dr David B. Norman for examining the long axial bone
and permitting use to cite their opinions on its identity and Mr David J. Ward and Mrs Sandra E. Sequeira
for photographic assistance.

REFERENCES

benton, m. j. 1987. The phylogeny of rhynchosaurs (Reptilia; Diapsida; Triassic), and two new species. 12-17.
In currie, P. J. and roster, e. h. (eds). 4th Symposium on Mesozoic Terrestrial Ecosystems, Short Papers.
Occasional Papers of the Tyrrell Museum of Paleontology, 3, 1-100.

— 1990. The species of Rhynchosaurus , a rhynchosaur (Reptilia, Diapsida) from the Middle Triassic of
England. Philosophical Transactions of the Royal Society of London , Series B, 320, 213-306.
browne, M. 1893. A contribution to the history of the geology of the Borough of Leicester. Transactions of the
Leicester Literary and Philosophical Society , 1893, 123-240.
carter, h. j. 1888. On some vertebrate remains in the Triassic strata of the south coast of Devonshire between
Budleigh Salterton and Sidmouth. Quarterly Journal of the Geological Society of London, 44, 318-319.
chernin, s. 1974. Capitosaurid amphibians from the Upper Luangwa Valley, Zambia. Palaeontologica
Africana , 17, 29-55.

egerton, p. m. g. 1854. On a fossil fish from the Upper Beds of the New Red Sandstone of Bromsgrove.

Quarterly Journal of the Geological Society of London, 10. 367-371.
fraser, n. c. 1986a. New Triassic sphenodontids from South-west England and a review of their classification.
Palaeontology , 29, 165-186.

19866. Terrestrial vertebrates at the Triassic-Jurassic boundary in South-west Britain. Modern Geology,
10, 147-157.

and shelton, c. G. 1988. Studies of tooth implantation in fossil tetrapods using high-resolution X-
radiography. Geological Magazine, 125, 117-122.

-and walkden, G. m. 1983. The ecology of a late Triassic reptile assemblage from Gloucestershire,
England. Palaeogeography, Palaeoclimatology, Palaeoecology, 42, 341-365.
gall, j. c., grauvogel, l. and Lehman, j. p. 1974. Faune du Buntsandstein. V- Les poissons fossiles de la
collection Grauvogel-Gall. Annates de Paleontologie ( Vertebres ), 60, 129-147.

Gardiner, b. 1988. A new Cleithrolepis from the Triassic of Central Cyrenaica, northeast Libya. 259-265. In
el-arnauti, a., Owens, b. and thusu, b. (eds). Subsurface palynostratigraphy of Northeast Libya. Garyounis
University Publications, Benghazi, xiv + 276 pp.

gorniak, G. C., rosenberg, H. i. and gans, C. 1982. Mastication in the Tuatara Sphenodon punctatus
(Reptilia: Rhynchocephalia). Structure and activity of the motor system. Journal of Morphology, 171,
321-353.

horwood, a. r. 1908. The flora and fauna of the Trias (Keuper only) in Leicestershire, with some notes on the
surrounding counties. Report of the British Association for the Advancement of Science, Leicester , 1907,
306-312.

huene, F. von. 1911. Uber die Procolophoniden, mit einer neuen Form aus dem Buntsandstein. Centralblatt
fur Mineralogie, Geologie und Paldontologie, 3, 78-83.

MILNER ET A L.: TRIASSIC VERTEBRATES FROM DEVON

889

1939. Ein neuer Procolophonide aus dem deutschen Buntsandstein. Neues Jahrbuch fur Mineralogie ,
Geologie unci Palaontologie , Beilband , Abteilung B , 81, 501-511.

1943. Zur Beurteilung der Procolophoniden. Neues Jahrbuch fur Mineralogie, Geologie und Palaontologie,

Monatshefte, Abteilung B, 7, 192-198.

Hutchinson, p. 1973. A revision of the redfieldiiform and perleidiform fishes from the Triassic of Bekker’s
Kraal (South Africa) and Brookvale (New South Wales). Bulletin of the British Museum ( Natural History )
(Geology), 22, 235-354.

johnston-lavis, h. j. 1876. On the Triassic strata exposed in the cliff sections near Sidmouth, and a note on
the occurrence of an ossiferous zone containing bones of a Labyrinthodon. Quarterly Journal of the
Geological Society of London, 32, 274—277.

jorg, E. 1969. Eine Fischfauna aus dem Oberen Buntsandstein (Unter-Trias) von Karlsruhe-Durlach
(Nordbaden). Beitrdge zur naturkundlichen Forschung in Siidwestdeutschland, 28, 87 102.
jupp, R. and warren, a. a. 1986. The mandibles of Triassic temnospondyl amphibians. Alcheringa, 10.
99-124.

kamphausen, d. 1983. Stenotosaurus gracilis, ein neuer Capitosauride (Stegocephalia) aus den Unteren
Rottonen Oberfrankens. Neues Jahrbuch fur Geologie und Palaontologie, Monatshefte, 1983, 119-128.

— and morales, M. 1981. Eocyclotosaurus lehmani , a new combination for Stenotosaurus lehmani Heyler
1969 (Amphibia). Neues Jahrbuch fur Geologie und Palaontologie, Monatshefte, 1981, 651-656.
krebs, b. 1969. Ctenosauriscus koeneni (v. Huene), die Pseudosuchia und die Buntsandstein-Reptilien. Eclogue
geologicae Helvetiae, 62, 697-714.

laming, d. j. c. 1982. The New Red Sandstone. 148-178 In durrance, e. m. and laming, d. j. c. (eds). The
Geology of Devon. University of Exeter, Exeter, 346 pp.
malan, m. e. 1963. The dentition of the South African Rhynchocephalia and their bearing on the origin of the
rhynchosaurs. South African Journal of Science, 59. 214—219.
martin, m. 1980. Dipteronotus gibbosus (Actinopterygii, Chondrostei), nouveau Colobodontide du Trias
superieur continental marocain. Geobios, 13, 445^449.

McCUNE, a. r. 1986. A revision of Semionotus (Pisces: Semionotidae) from the Triassic and Jurassic of Europe.
Palaeontology, 29, 213-233.

metcalfe, a. t. 1884. On further discoveries of vertebrate remains in the Triassic strata of the south coast of
Devonshire, between Budleigh Salterton and Sidmouth. Quarterly Journal of the Geological Society of
London , 40, 257-262.

morales, m. 1987. Terrestrial fauna and flora from the Triassic Moenkopi Formation of the southwestern
United States. Journal of the Arizona-Nevada Academy of Sciences, 22, 1-19.
morales, m. and kamphausen, d. 1984. Odenwaldia heidelbergensis, a new benthosuchid stegocephalian from
the Middle Buntsandstein of the Odenwald, Germany. Neues Jahrbuch fur Geologie und Palaontologie,
Monatshefte, 1984, 673-683.

murry, p. a. 1987. New reptiles from the Upper Triassic Chinle Formation of Arizona. Journal of
Paleontology, 61, 773-786.

OCHEV, v. G. 1966. ( Systematics and phytogeny of capitosauroid labyrinthodonts.) Saratov University Press, 184
pp. [In Russian],

ortlam, d. 1970. Eocyclotosaurus woschmidti n.g. n.sp. -ein neuer Capitosauride aus dem Oberen
Buntsandstein des nordlichen Schwarzwaldes. Neues Jahrbuch fur Geologie und Palaontologie, Monatshefte,
1970. 568-580.

owen, r. 1842. Description of parts of the skeleton and teeth of five species of the genus Labyrinthodon . . . with
remarks on the probable identity of Cheirotherium with this genus of extinct batrachians. Transactions of the
Geological Society of London, (2) 6, 515-543.

paton, R. L. 1974. Capitosauroid labyrinthodonts from the Trias of England. Palaeontology, 17, 253-289.
ROBINSON, P. L. 1973. A problematic reptile from the British Upper Trias. Journal of the Geological Society of
London , 129, 457-479.

romer, a. s. 1947. Review of the Labyrinthodontia. Bulletin of the Museum of Comparative Zoology, Harvard
College, 99, 1-368.

Seeley, h. g. 1876. On the posterior portion of a lower jaw of a Labyrinthodon (L. lavisi) from the Trias of
Sidmouth. Quarterly Journal of the Geological Society of London, 32, 278-284.
shishkin, m. a. 1980. The Luzocephalidae, a new Triassic labyrinthodont family. Paleontologicheskiy Zhurnal,
1980, 104—119 [In Russian]; Paleontological Journal, 23, 88-101 [In English],
spencer, p. s. and Isaac, k. p. 1983. Triassic vertebrates from the Otter Sandstone Formation of Devon,
England. Proceedings of the Geologists' Association, 94, 267-269.

890

PALAEONTOLOGY, VOLUME 33

swinnerton, h. H. 1925. A new catopterid fish from the Keuper of Nottingham. Quarterly Journal of the
Geological Society of London , 81, 87-89.

— 1928. On a new species of Semionotus from the Keuper of Nottingham. Geological Magazine , 65, 406-409.
wade, R. T. 1935. The Triassic fishes of Brookvale, New South Wales, xi v +110 pp. British Museum (Natural

History), London.

walker, a. d. 1969. The reptile fauna of the ‘Lower Keuper’ Sandstone. Geological Magazine , 106, 470-476.
1970. [Written contribution to discussion]. Quarterly Journal of the Geological Society of London, 126,
217-218.

Warrington, G. 1967. Correlation of the Keuper Series of the Triassic by Miospores. Nature , 214, 1323-1324.
1970. The stratigraphy and palaeontology of the ‘Keuper’ Series of the central Midlands of England.
Quarterly Journal of the Geological Society of London , 126, 183-223.

— 1971. Palynology of the New Red Sandstone sequence of the south Devon coast. Proceedings of the
Usslier Society, 2, 307-314.

— AUDLEY-CHARLES, M. G., ELLIOTT, R. E., EVANS, W. B., IVIMEY-COOK, H. C., KENT, P., ROBINSON, P. L.,
shotton, F. w. and taylor, F. M. 1980. A correlation of Triassic rocks in the British Isles. Special Report of
the Geological Society of London, 13, 1-78.

watson, D. M. s. 1962. The evolution of the labyrinthodonts. Philosophical Transactions of the Royal Society
of London, Series B, 245, 219-265.

welles, s. p. 1947. Vertebrates from the Upper Moenkopi Formation of Northern Arizona. University of
California Publications in Geological Sciences, 27, 241-294.

— and cosgriff, j. 1965. A revision of the labyrinthodont family Capitosauridae and a description of
Parotosaurus peabodyi, n. sp. from the Wupatki member of the Moenkopi Formation of Northern Arizona.
University of California Publications in Geological Sciences, 54, 1-148.

— and estes, r. 1969. Hadrokkosaurus bradyi from the Upper Moenkopi Formation of Arizona, with a
review of the brachyopid labyrinthodonts. University of California Publications in Geological Sciences, 84,
1-56.

wepfer, e. 1923. Der Buntsandstein des badischen Schwarzwalds und seine Labyrinthodonten. Monographien
zur Geologie und Palaeontologie, Berlin, (2) 1, 1-101.

whitaker, w. 1869. On the succession of beds in the ‘New Red’ on the South Coast of Devon, and on the
locality of a new specimen of Hyperodapedon. Quarterly Journal of the Geological Society of London, 25,
152-158.

white, E. i. 1950. A fish from the Bunter near Kidderminster. Transactions of the Worcestershire Naturalists'
Club, 10, 185-189.

whiteside, d. i. 1986. The head skeleton of the Rhaetian sphenodontid Diphyodontosaurus avonis gen. et. sp.
nov. and the modernizing of a living fossil. Philosophical Transactions of the Royal Society of London, Series
B, 312, 379-430.

WILD, r. 1973. Tanystropheus longobardicus (Bassani) (Neue Ergebmsse). In kuhn-schnyder, e. and peyer, b.
Die Triasfauna der Tessiner Kalkalpen XXIII. Schweizerische paldontologische Abhandlungen, 95, 1-162.

1978. Die Flugsauner (Reptilia, Pterosauria) aus der Oberen Trias von Cene bei Bergamo, Italien.

Bollettino della Societa Paleontologica Italiana, 17, 176-256.

1980m Tanystropheus (Reptilia: Squamata) and its importance for stratigraphy. Memoires de la Societe
Geologique de France, Nouvelle Serie, 109, 201-206.

19806. The fossil deposits of Kupferzell, southwest Germany. Mesozoic Vertebrate Life, 1, 15-18.

wills, L. J. 1907. On some fossiliferous Keuper rocks at Bromsgrove (Worcestershire). Geological Magazine,

(5) 4, 28-34. ...

1910. On the fossiliferous Lower Keuper rocks of Worcestershire, with descriptions of some ot the plants

and animals discovered therein. Proceedings of the Geologists' Association, 21, 249-322.

- 1916. The structure of the lower jaw of Triassic labyrinthodonts. Proceedings of the Birmingham Natural
History Society, 14, 1-16.

woodward, a. s. 1910. On Dipteronotus cyphus, Egerton. A ganoid fish from the Lower Keuper ot
Bromsgrove, Worcestershire. Proceedings of the Geologists' Association, 21, 322-323.

A. R. MILNER
Department of Biology
Birkbeck College
Malet Street
London WC1E 7HX

MILNER ET AL.: TRIASSIC VERTEBRATES FROM DEVON

891

B. G. GARDINER

Department of Biology
Kings College London
Campden Hill Road
London W8 7AH

N. C. FRASER

Virginia Museum of Natural History
1001 Douglas Avenue
Martinsville
Virginia 24112, USA.

Typescript received 5 July 1989

Revised typescript received 3 November 1989

M. A. TAYLOR

Earth Sciences Section
Leicestershire Museums, Arts
and Records Service
96 New Walk
Leicester LEI 6TD

APPENDIX

Check-list of the pre-Penarth Group Triassic fishes of the British Isles (all material examined by B.G.G).
MMG = Mercia Mudstone Group, SSG = Sherwood Sandstone Group.

Palaeobates keuperinus (Murchison and Strickland, 1840)

Bromsgrove Sandstone Formation (SSG) of Bromsgrove. Arden Sandstone Member (MMG) of Callow Hill;
Mousley End; Shrewley; Knowle; Rowington; Shelfield Square, WARWICKSHIRE; and Burghill; Ripple;
Pendock, HEREFORD and WORCESTER. Dane Hills Sandstone Member (MMG) of Shoulder of Mutton
Hill; Dane Hills; New Parks; Aylestone Road, LEICESTERSHIRE.

Gyrolepis albertii Agassiz, 1835

Bromsgrove Sandstone Formation (SSG) of Bromsgrove, HEREFORD and WORCESTER (WARMS
material). Colwick Formation (MMG) of Colwick Wood, NOTTINGHAMSHIRE. Arden Sandstone
Member (MMG) Leamington, WARWICKSHIRE (Owen 1842 pi. 46, fig. 15; see Walker 1969). Dane Hills
Sandstone Member (MMG) of Aylestone Road (G. quenstedti of Browne 1893) and Spinney Hills (BMNH
P.41626-9), LEICESTERSHIRE.

Dictyopyge catoptera (Agassiz, 1835)

Sherwood Sandstone Group of Dungannon, CO. TYRONE.

Dictyopyge superstes (Egerton, 1858)

Arden Sandstone Member (MMG) of Rowington, WARWICKSHIRE. This record is based on a partial body
and is probably part of a dictyopygid as Woodward (in Swinnerton 1925) surmized. It shows no resemblances
to Woodthorpea despite Swinnerton’s (1925, p. 97) comments, and closely resembles Dictyopyge socialis
(Berger) from the Upper Keuper of Colburg.

Dictyopyge sp. nov.

Colwick Formation (MMG) of Colwick Wood, NOTTINGHAMSHIRE. Known from a single specimen
(BMNH P.5288), this novel dictyopygid has scales with posterior margins bearing three prominent projections.

IPerleidus

Wildmoor Sandstone Formation (SSG), Aggborough, HEREFORD and WORCESTER. This record (White
1950) is based on a single specimen which lacks head and tail, but nevertheless closely resembles Perleidus.

892

PALAEONTOLOGY, VOLUME 33

Dipteronotus cyphus Egerton, 1854

Bromsgrove Sandstone Formation (SSG) of Bromsgrove, HEREFORD and WORCESTER and Otter
Sandstone of Sidmouth. DEVON.

4 Semionotus ' brodiei Newton, 1887

Arden Sandstone Member (MMG) of Shrewley, WARWICKSHIRE. This species is represented by five
syntypes (BMNH P.7615), which bear no resemblance to the genus Semionotus (McCune 1986, p. 229). It is
a small species with large dorsal ridge scales and deep flank scales. The tail is heterocercal, the suspensorium
more or less upright and supraorbitals seem to be present. It appears to be a catopterid or a redfieldiid.

Semionotus metcalfei Swinnerton, 1928

Colwick Formation (MMG) of Woodthorpe and Colwick Wood (BMNH P. 9818), NOTTINGHAMSHIRE.
Dane Hills Sandstone Member (MMG) of Spinney Hills, LEICESTERSHIRE ('semionotid fish’ of Horwood
1907). This species is probably synonymous with S. kapffi Fraas, 1861 (McCune 1986, p. 230). It has ridge
scales with spines and a single large suborbital.

Woodthorpea wilsoni Swinnerton, 1925

Colwick Formation (MMG) of Woodthorpe, NOTTINGHAMSHIRE. This is a small semionotid, apparently
with a single suborbital.

Ceratodus laevissimus Miall, 1878

Bromsgrove Sandstone Formation (SSG) of Coten End, WARWICKSHIRE ( Ceratodus cf. C. kurii of Wills
1910) and Bromsgrove, HEREFORD and WORCESTER. Mercia Mudstone Group of Ripple, HEREFORD
and WORCESTER.

A TINY MICROSAUR FROM THE LOWER PERMIAN
OF TEXAS: SIZE CONSTRAINTS IN PALAEOZOIC

TETRAPODS

by ROBERT L. CARROLL

Abstract. Quasicaecilia texana gen. et sp. nov., from the Lower Permian of Texas, is a very small microsaur
in which the nasal bones form the front margin of the skull and the cultriform process of the parasphcnoid is
not in contact with the sphenethmoid. The bones of the otic-occipital region are fused to one another, and the
otic capsule and stapes are very large relative to other cranial dimensions, giving a superficial resemblance to
modern caecilians. Many of the unusual cranial features of this species may be attributed to its small size. The
relatively large size of the sensory capsules in Palaeozoic amphibians with an adult skull length less than about
2 cm results in significant modification of the area of the otic capsule and jaw articulation.

Palaeozoic amphibians have customarily been divided into two large groups, the labyrinth-
odonts and the lepospondyls (Romer 1966; Carroll 1977, 1987). Labyrinthodonts are typified by a
suite of primitive characters inherited from their ancestors among the rhipidistian fish.

a. Squamosal embayment of cheek occupying position of rhipidistian operculum.

b. Marginal teeth exhibiting labyrinthine enfolding.

c. Large palatal fangs and associated replacement pits.

d. Occipital condyle (where present) narrow and circular.

e. Multipartite atlas-axis complex.

f. Vertebral centra primitively consisting of crescentic ventral intercentrum and paired
pleurocentra.

Lepospondyls are characterized by a suite of more derived characteristics.

a. Absence of squamosal embayment.

b. Lack of labyrinthine enfolding of marginal teeth.

c. Absence of large palatal fangs and adjacent replacement pits.

Most lepospondyls have a wide or double occipital condyle and a single cylindrical centrum per
segment.

Labyrinthodonts are certainly a paraphyletic group, including the ancestors of all more-derived
tetrapods. Several recent attempts have been made to establish the affinities of the many groups
termed lepospondyls, but as yet no consensus has been reached as to their specific relationships to
the labyrinthodont groups, or to one another (Thomson and Bossy 1970; Gaffney 1979; Gardiner
1983; A. R. Milner et al. 1986; Gauthier et al. 1988; Panchen and Smithson 1988). In the absence
of generally accepted relationships, it remains convenient to retain the descriptive terms
labyrinthodont and lepospondyl.

In addition to similar derived features by which all lepospondyls are distinguished from
labyrinthodonts, each individual group - aistopods, nectrideans, adelogyrinids, lysorophids and
microsaurs is distinguished from the others by numerous unique derived features. Lepospondyls
also show a much higher degree of variability within each order than that exhibited by the
labyrinthodonts in the pattern of the dermal bones of the skull, the configuration of the braincase,
the length of the vertebral column, and the proportions and structure of the limbs. Labyrinthodonts

© The Palaeontological Association

(Palaeontology, Vol. 33, Part 4, 1990, pp. 893-909. |

894

PALAEONTOLOGY, VOLUME 33

had a basically similar skeletal pattern from their first appearance in the fossil record in the Upper
Devonian to their last appearance in the Cretaceous.

Shared derived features of the skeleton demonstrate that all tetrapods share a common ancestry
above the level of the rhipidistian fish (Panchen and Smithson 1987). This implies that lepospondyls
must have evolved from tetrapods having an anatomical pattern similar to that of known
labyrinthodonts; however, no intermediates are recognized. This situation suggests that some
complex of evolutionary constraints acted to retain a stereotyped anatomical pattern among the
labyrinthodonts. The emergence of the various lepospondyl groups may be attributed to
circumvention of these constraints.

In addition to the anatomical differences already mentioned, all lepospondyls are further
distinguished from primitive labyrinthodonts by their smaller adult size. Some labyrinthodonts are
as small as large lepospondyls, indicating that small size is not a sufficient factor in itself to result
in the evolution of lepospondyl characters, but it seems a necessary factor, since none of the
lepospondyls is the size of large labyrinthodonts. The larger Permian lepospondyls are highly
derived members of this assemblage; all the early lepospondyls are much smaller.

Small size is an important attribute of the origin of many vertebrate groups: the modern
amphibian orders (Carroll and Holmes 1980; Milner 1988), amniotes (Carroll 1970), and some
lizard families (Rieppel 1988). Hanken (1983, 1989) has stressed its importance in the radical
reorganization of the limbs and skulls of tiny living salamanders. The peculiar anatomy of the
animal described in this paper provides a basis for considering the importance of skull size in
influencing the cranial anatomy of Palaeozoic amphibians.

MATERIALS AND METHODS

This paper deals primarily with a single specimen (Text-fig. 1) collected by C. H. Sternberg in 1917
from the classic Texas Permian redbeds. The specimen had long been identified as Cardiocephalus ,
a gymnarthid microsaur. The short antorbital region may have been attributed to breakage, and the
great width of the parietals may not have been recognized due to the presence of longitudinal cracks
that might have been identified as sutures separating relatively narrow parietals from more lateral
tabular bones.

As preserved, the skull is approximately 14-5 mm long, and equally as wide. The anterior
extremity is formed by the broad, short nasal bones. They show no area for attachment of the
premaxillae, and those bones are otherwise missing. The posterior surface of each occipital condyle
is broken off, precluding exact determination of the length of the skull. Most of the bones of the
palate are missing or displaced, exposing the braincase ventrally.

The skull roof and occiput were largely exposed as a result of natural weathering of the matrix.
The parasphenoid was also initially exposed. The remaining areas of the palate and braincase were
prepared with a fine needle. The matrix is a hard, fairly homogeneous limey clay. It varies in colour
from white to pink and in places was very difficult to differentiate from the bone.

SYSTEMATIC PALAEONTOLOGY

Class amphibia Linnaeus, 1758
Order microsauria Dawson, 1863
Suborder microbrachomorpha Carroll and Gaskill, 1978
?Family brachystelechidae Carroll and Gaskill, 1978
Genus quasicaecilia gen. nov.

Type species. Quasicaecilia texana sp. nov.

Derivation of generic name. The generic name refers to several similarities with modern caecilians that are
almost certainly not indicative of close relationship.

CARROLL LOWER PERMIAN MICROSAUR

895

text-fig. 1. Quasicaecilia texana , type, United States National Museum, 22079. a, dorsal view of skull, b,
obliquely anterior view of palate and braincase; postorbital bone that appears in drawings was removed prior
to photographing, c, left lateral view of skull, d, occipital view of skull. All x4.

Diagnosis. Microbrachomorph niicrosaur, possibly belonging to the family Brachystelechidae,
differing from Brachystelechus and Carrolla in having extremely short and broad nasals that show
no area of articulation for the premaxillae. Stapes apparently lacking stapedial foramen. Cultriform
process of parasphenoid separated from sphenethmoid by wide gap.

Quasicaecilia texana sp. nov.

Text-figs 1-5

Derivation of trivia! name. Refers to the state from which the type species was collected.

Holotype. United States National Museum (USNM) no. 22079.

896

PALAEONTOLOGY, VOLUME 33

Locality and horizon. ?Baylor County, Texas; Lower Permian ?Arroyo Formation.

Diagnosis. As for genus. This is the only recognized species.

Description. Both the orbital and pineal openings appear large relative to the length of the skull, as would be
expected in a small animal. The orbits are protected by large saucershaped palpebral bones. The surface of the
roofing bones is essentially smooth and shows no evidence of lateral line canals.

The pattern of the skull roof (Text-fig. 2) differs dramatically from that of all other Palaeozoic tetrapods.

text-fig. 2. Quasicae cilia texana , type. United States National Museum, 22079. a, dorsal view of skull, b,
reconstruction of skull in dorsal view, c, occipital view. All x4-5. Abbreviations: a, angular; ba, basicranial
articulation ; bo, basioccipital ; bs, basisphenoid ; c, coronoid ; eg, groove for carotid artery ; com, composite of
posterior jaw bones; d, dentary; dp, dorsal process of stapes; ect, ectopterygoid ; eo, exoccipital; f, frontal;
f-p, frontoparietal; j, jugal; 1, lacrimal; m, maxilla; n, nasal; ob, os basale; opis, ophisthotic; otic, position
of otic capsule; p, parietal; pal, palatine; pc, palpebral cup; pea, palatine canal; pf, postfrontal; pi,
pleurosphenoid; pm, premaxilla; po, postorbital; pp, postparietal ; prf, prefrontal; pro, prootic; ps,
parasphenoid ; pt, pterygoid; q, quadrate; qj, quadratojugal ; sa, surangular; sm, septomaxilla ; so,
supraoccipital ; sp, splenial; sph, sphenethmoid ; sq, squamosal; st, supratemporal ; sta, stapes; syt, synotic
tectum; t, tabular; t-sq, bone occupying position of tabular and squamosal; v, vomer; X, area of matrix left
for support of cultriform process of stapes; ?, bone of uncertain identity; I, II, V, foramina for cranial nerves.

CARROLL: LOWER PERMIAN MICROSAUR

897

It consists primarily of very wide parietal bones, rectangular frontals, and wide and extremely short nasals.
The nasals are rounded anteriorly to form the anterior margin of the skull. They extend ventrally so as to
surround the area that would have been occupied by the nasal capsules. No bone occupies the position expected
of the premaxillae at the anterior margin of the skull roof or palate. Since there is no area for their articulation
with the nasals, it is probable that they were not solidly attached to the skull. There is no trace of postparietal
bones. The parietals, as in modern caecilians, extend far posteriorly over the back of the braincase. Behind the
pineal opening they appear fused at the midline where the bone forms a ridge that extends nearly to the margin
of the foramen magnum between two areas that are recessed to receive the epaxial musculature. A transverse
ridge on the patietals clearly demarcates the anterior limit of this musculature. On each side, the ridge is bowed
posteriorly and slightly overlaps the occipital surface. There is no evidence of the large tabular bones that
characterize tuditanomorph microsaurs.

The frontal narrowly enters the orbital margin between the pre- and postfrontals. All three bones form a
wide medial margin to the orbit. The left postfrontal is missing, revealing a thick area of the parietal to which
it was attached. On the right side, a small triangular remnant of the postorbital is retained in place between
the postfrontal and the parietal. The remainder of the bone is broken off and appears in palatal view, with the
lateral surface exposed medially. The entire bone is triangular in outline. The left jugal appears in more or less
its natural position beneath the orbit. The remainder of the cheek region (Text-fig. 5) is badly crushed, making
determination of its original configuration extremely difficult. The great size of the otic capsule indicates that
the jaw articulation must have been far anterior in position. This is confirmed by the presence of the quadrate
just behind the orbital opening. Presumably the squamosal occupied a narrow vertical area behind the jugal
and wrapped around the quadrate posteriorly. The broken nature of the bone surface behind the jugal
precludes determination of the presence or absence of the quadratojugal. The quadrate is the shape of a
thickened vertical plate. The dorsal end is angled towards the posterior margin of the orbit. The lateral portion
of the lower end forms a hemispherical condyle. More medially, the distal surface is broadly embayed for a
distance that is as great as the width of the condyle. The medial edge is in contact with the quadrate ramus
of the pterygoid. The shortness of the cheek and the large size and orientation of the quadrate indicate that
the adductor chamber must have been very small, providing little space for jaw-closing musculature.

A thin crescent of bone above the distal end of the jugal may be the left palpebral cup. Beneath this is an
area of bone that may be the lacrimal, although no opening for a lacrimal duct is evident. No trace of the
maxillae can be identified.

A large triangular bone on the left side of the palatal area (Text-fig. 3) is identified as the pterygoid. Its

text-fig. 3. Quasicaecilia texana. a, palatal view of skull, b, reconstruction of palate; right side shows
braincase with elements of palate and stapes omitted. x4-5.

898

PALAEONTOLOGY, VOLUME 33

posterior end is in contact with the medial surface of the quadrate. Medially, it may have been in contact with
a small unfinished surface of the prootic. The posteromedial surface of the bone shows no evidence of the
basicranial articulation. Laterally, the bone angles sharply dorsally. A similar dorsal projection was noted in
Carrolla by Langston and Olson (1986). The anterior surface of the bone is broadly concave. This implies an
unusual outline for the interpterygoid vacuities, but this area might have accommodated retraction of the
relatively large eyes. Anterolaterally, the pterygoid ends in an enlarged surface that appears to terminate in
an oval area of articulation. This might have attached to the ectopterygoid or palatine.

Bone is visible on the right side between the orbit and the base of the braincase, but none of the surfaces
is sufficiently well preserved to permit identification. Anteriorly, on both sides, are smaller bones with two
rami meeting at nearly right angles. They are logically identified as the vomers. They lie adjacent to the
sphenethmoid; presumably they were displaced from a more ventral position.

Above the scattered and difficult to identify palatal elements is a heavily ossified braincase (Text-figs
4 and 5). Its degree of ossification is matched by that of aistopods, but its structure is unique among Palaeozoic
tetrapods. The braincase consists of two major elements, the anterior sphenethmoid, broadly comparable to
the sphenethmoid of Palaeozoic labyrinthodonts, and the posterior otic-occipital complex. The latter
incorporates all the bones recognized in other Palaeozoic tetrapods, fused into an essentially unipartite
structure. The basioccipital, exoccipitals, supraoccipital, basisphenoid, laterosphenoids (or pleurosphcnoids),
opisthotics, prootics and parasphenoid are all fused without trace of sutures, except possibly the dorsal contact
between the opisthotic and prootic. The resulting structure vaguely resembles the os basale of caecilians.

The articulating surface of the basioccipital and exoccipitals is lost, but would have been very wide as in
typical microsaurs. The ventrolateral surface of these bones is deeply grooved just above the area of the
parasphenoid. This groove is irregularly pitted, probably for insertion of hypaxial musculature to lower the
head. Nerves X-XII presumably exited in this area, but their foramina cannot be differentiated with certainty
from the other pits.

One of the most striking features of this tiny skull is the relatively enormous size of the otic capsules and
stapes. The otic capsules extend to the lateral extremities of the skull as preserved. The fenestrae ovales appear
huge in ventral view. They are almost 4 mm in dorsoventral extent. Their configuration differs somewhat on
the two sides. On the right side the opisthotic rim extends further laterally than the prodtic. On the left side

text-fig. 4. Quasicae cilia texana. a, oblique anteroventral view of skull, b, reconstruction of skull in
anteroventral view. Most palatal elements have been omitted to expose braincase. x4-5.

CARROLL: LOWER PERMIAN MICROSAUR

899

text-fig. 5. Quasicaecilia te.xana. a, right side of skull, b, restoration of the skull in right lateral view, with
cheek removed to show braincase. c, left side of skull, d, restoration of skull in left lateral view; configuration
of cheek is very speculative. Left and right sides have been independently restored. x4-5.

the condition is reversed. The left capsule is somewhat compressed from back to front. Posteriorly, the
opisthotics and exoccipitals are continuous with the supraoccipital, which extends forward a short distance
beneath the overlying parietals.

The stapes is present in association with the fenestra ovalis on both sides, but on neither side is it in natural
articulation. On the left side, the foot plate of the stapes is in continuity with the margin of the prootic in a
clearly unnatural position that suggests minor dissolution and redeposition of the calcium phosphate during
preservation. Otherwise the bone appears to retain its natural configuration. The foot plate extends for slightly
less than 3 mm in dorsoventral diameter. It is not possible to reconstruct its exact position in the fenestra, but
there may have been a medial unossified gap, as in several microsaurs (Carroll and Gaskill 1978). The stem of
the stapes is somewhat less than 2 mm in diameter and 4 mm in length including the base of the foot plate. The
stem appears to point anteroventrally, and ends in a shallow depression that may have been continued in
cartilage. No stapedial foramen is evident. From the stem extends a broad but thin dorsal process that is almost
parallel with the surface of the foot plate. What appears to be the right stapes is a simple columnar structure.
It is about the size of the stem of the left stapes. Damage prior to burial may have led to the loss of the dorsal
process and the margins of the foot plate.

Above the fenestra ovalis, the area of the inner ear continues to the skull roof. Medially, a further ossification
is encountered that is not solidly fused to other elements. It is tentatively identified as an ossification of the
synotic tectum. It bears grooves that may have accommodated the dorsal portion of the vertical semicircular
canals. In several microsaurs ( Pelodosotis , Pantylus, and Rhynchonkos [ Goniorhynchus ]), the supraoccipital
continues anteriorly to incorporate the area of the synotic tectum, but this area appears to be separate in
Quasicaecelia. In Microbrachis, in contrast, there is apparently no supraoccipital, but there is a more anterior
ossification that may represent the synotic tectum. DeBeer (1937) recognized the presence of both a posterior
tectum, which may ossify as the supraoccipital, and a more anterior synotic tectum in all major vertebrate

900 PALAEONTOLOGY, VOLUME 33

groups, but noted that their expression in individual species is extremely variable and appears to have little
taxonomic significance.

In anterior view, the prootic appears as a nearly vertical sheet of bone that forms the anterior wall of the
otic capsule. Ventrally, it is continuous with the parasphenoid-basisphenoid area. Just above the parasphenoid
plate, a short area of unfinished bone is evident on both sides. It may have been in contact with the quadrate
ramus of the pterygoid. Above this area the margin of the prootic is inflected posteriorly and somewhat
medially into the fenestra ovalis.

Posteriorly, the prootic terminates some distance below the skull roof, along a line that corresponds to the
dorsal margin of the fenestra ovalis. A very thin flange of bone, continuous with the surface of the right stapes,
may represent the anterior extension of the opisthotic towards the prootic in this area. Anteriorly, the prootic
extends further dorsally, narrowing to a point laterally. Anteromedially the prootic is notched for the prootic
foramen, an opening formed jointly with the pleurosphenoid. The opening, clearly visible with minimal
distortion on both sides, is approximately 0-9 mm in diameter, which is strikingly small relative to the other
cranial dimensions.

At the base of the braincase, at the anterior margin of the posterior plate of the parasphenoid, are the
basicranial processes which extend anteriorly as shelflike structures. No specific area of articulation is
distinguished by unfinished bone surface. The adjacent margin of the pterygoid does not show a specialized
area for the articulation between the braincase and the palate. Just anterior to the basicranial articulation are
foramina that are comparable in position to those in caecilians which serve for the palatine arteries.

Anterior to the prootic, the area of the pleurosphenoid, most completely visible on the right side, forms the
lateral wall of the braincase. It rises nearly vertically from the area of the basisphenoid-parasphenoid, and
extends as a narrow sheet of bone to within less than 2 mm of the underside of the skull roof. It is lower
posteriorly, where it is continuous with the prootic, and rises anteriorly. On the left side, it can be seen to form
the posterior margin of a large opening for the optic and possibly the oculomotor nerves. The opening is about
2 mm in diameter, substantially larger than the opening for the Vth nerve. The pleurosphenoid forms the
anterior margin of the oticoccipital portion of the braincase. Ventrally, there is a wide gap above the
parasphenoid. Dorsally, the pleurosphenoid lies lateral of the sphenethmoid, apparently without being in
sutural contract.

The anterior extent of the pleurosphenoid, beyond the level of the basicranial processes, would appear to
preclude the origin of the rectus eye muscles from a retractor pit in the position of that structure in amniotes
and anthracosaurian labyrinthodonts, as described by Clack and Holmes (1988). The high degree of
specialization of all aspects of the braincase makes specific comparison with any labyrinthodonts difficult,
however.

The posterior portion of the braincase is covered ventrally by the plate of the parasphenoid, mdistinguishably
fused to the overlying endoskeletal elements. The parasphenoid is marked on the left side by a strong ridge that
formed the medial border of a groove for the carotid artery. This ridge is missing on the right side. The
cultriform process of the parasphenoid extends forward like a dagger. It may have angled somewhat dorsally,
but could not have underlain the sphenethmoid directly.

The sphenethmoid is visible in ventral and lateral views as a tubular structure extending anteriorly from the
area of the pleurosphenoid to the front of the nasals. It appears to be closely appressed to the underside of the
skull roof, beneath the frontals. Posteriorly, it forms the upper border of the foramen for the optic nerve, where
it lies slightly medial to the pleurosphenoid. In contrast with caecilians, these bones do not abut against one
another. Rather, the sphenethmoid appears to fit within the anterior borders of the pleurosphenoids.

The sphenethmoid is pierced laterally by relatively large foramina for the olfactory nerve. The anterior end
of the bone is dorsoventrally compressed but widened laterally behind the front of the skull. This portion of
the bone presumably served primarily as a structural brace since the foramina for nerve I exit more posteriorly,
rather than at the end of the bone. The ventral border of the sphenethmoid is broadly rounded, but ends well
above the cultriform process of the parasphenoid. In this it differs from other microsaurs in which the
sphenethmoid is ossified, where it appears to be supported by the parasphenoid ventrally. The sphenethmoid
is very incompletely ossified in early amniotes, but in genera in which it is preserved, there is also a wide gap
between it and the cultriform process of the parasphenoid. The remainder of the braincase in this genus and
early amniotes is so different that this one similarity is not thought to be taxonomically significant.

On the left side, just anterior to the otic capsule, is the posterior end of the lower jaw. The articular bone
is in articulation with the condyle of the quadrate. Anteriorly it extends dorsally against the surface of the
quadrate. Medially, the articular extends well beyond the remainder of the jaw. There is no retroarticular
process. The anterior, tooth-bearing portion of the lower jaw is not preserved.

With no evidence of the postcranial skeleton, there is little basis for establishing the general adaptation of

CARROLL: LOWER PERMIAN MICROSAUR

901

text-fig. 6. a. b, skull of the burrowing snake Typhlops punctatus in lateral and ventral views, modified from
Parker (1977); small size of prootic foramen may be associated with reduction in size of jaws and dentition.

this species. A somewhat analogous modification of the cranial anatomy can be seen in burrowing typhlopid
snakes (Text-fig. 6) in which the marginal jaw elements have lost their tooth-bearing role, and are much
modified. As in caecilians, the fusion of the otic-occipital elements of the braincase might be attributed to a
burrowing habitus, but such fusion did not occur in primitive burrowing snakes.

DISCUSSION

Taxonomic position

Despite the loss of their articulating surfaces, the great width of the exoccipitals and basioccipital
in Quasicaecilia indicates the presence of wide condyles as in typical microsaurs. The pattern of the
bones of the skull table resembles that of members of the Microbrachomorpha in the great width
of the parietals and the progressively smaller size of the frontals and nasals. The greatest
resemblance is to the previously described members of the family Brachystelechidae (Carroll and
Gaskill 1987; Langston and Olson 1986) (Text-fig. 7). In both Brachystelechus and Carrolla , the

text-fig. 7. Skulls of the three genera assigned to the Brachystelechidae. a, Brachystelechus. b, Carrolla.

c, Quasicaecilia.

902

PALAEONTOLOGY, VOLUME 33

parietals overlap the occipital surface, and the tabulars are either greatly reduced or entirely
missing. All three genera have very wide skulls and possess palpebral cups.

Quasicaecilia resembles Carrolla in the high degree of ossification of the braincase; the cranial
endoskeleton is not known in Brachystelechus. Quasicaecilia definitely differs from Br achy stele chus
in the shortness of the nasals, and from Carrolla by the absence of sutural attachment between the
nasals and the premaxillae. Quasicaecilia also differs from Carrolla by the absence of a stapedial
foramen and the separation of the sphenethmoid and the cultriform process of the parasphenoid.

The degree of specialization of these three genera appears to correspond with their stratigraphic
succession. The relative age of specimens from the European Rotliegend succession is subject to
continuing dispute, but Boy (1987) suggested that the horizon from which Brachystelechus was
collected is roughly equivalent to the North American Admiral Formation. Carrolla comes from the
Belle Plains Formation. Quasicaecilia is said to come from the Arroyo Formation, but this is
indicated with a question mark on the label. No other described group of Palaeozoic tetrapods
shares more derived characters in common with Quasicaecilia than do the previously identified
members of the Brachystelechidae.

Affinities between goniorhynchid inicrosaurs and caecilians have been postulated on the basis of
derived features of the braincase, several of which are also exhibited by Quasicaecilia (Carroll and
Currie 1975; Walsh 1987; Jenkins and Walsh, in preparation). The fusion of the posterior portion
of the braincase into a unitary os basale, the large size of the otic capsule, the presence of a large
pleurosphenoid, the anterior position of the jaw articulation and the loss of dermal bones at the back
of the skull table are features expected in the ancestors of caecilians (Walsh 1987). On the other
hand, in all caecilians, premaxillae make up the anterior margin of the skull. Caecilians and their
putative ancestors, the microsaur family Goniorhynchidae, are characterized by the small size of the
orbits, in contrast with their large size in Quasicaecilia. Caecilians maintain the primitive pattern of
having the sphenethmoid supported by the parasphenoid, in contrast with the derived condition
seen in Quasicaecilia.

Many of the features in which Quasicaecilia resembles modern caecilians may be associated with
small size and so do not constitute a strong basis for assuming close relationship.

Anatomical significance

The most interesting features of this specimen are the anatomical peculiarities, especially of the
braincase. Among the most conspicuous are the great size of the otic capsule and stapes, but the
relative size of the orbits and the configuration of the nasal region are also unusual.

The relatively large size of the sensory structures can be attributed primarily to the absolutely
small size of the skull. The great importance of the sensory structures and the braincase in
constraining the proportions of other features of the skull in small tetrapods was discussed in a
compelling fashion by Hanken (1983, 1984) in his study of miniaturization in salamanders of the
genus Thorius.

Although some species of modern frogs, salamanders, and caecilians are even smaller than this
specimen, it is among the smallest Palaeozoic amphibians. Comparative dimensions of the braincase
and sense organs of Quasicaecilia and other small tetrapods are given in Table 1.

In Quasicaecilia and other Palaeozoic tetrapods, the dimensions of the bony otic capsule and
braincase can be measured directly. It is assumed that, as in other small tetrapods, the circumference
of the eyeball closely matches that of the bony orbital margin. This is confirmed in many small
Palaeozoic amphibians by the extent of the sclerotic ring which fills most of the orbital space. The
size of the nasal capsule is more difficult to estimate. It is never ossified in Palaeozoic tetrapods (in
contrast with their rhipidistian relatives). In some species, the extent of the capsule can be judged
from impressions in the undersurface of the overlying dermal bones, but this area is only rarely
exposed. In animals with small skulls, the probable size of the nasal capsule can be roughly estimated
by the space available between the front margin of the skull and the anterior limit of the orbit,
taking into consideration the position of the internal and external narial openings. In Quasicaecilia ,
the area of the nasal capsules appears to be outlined by the nasal bones that curve over the front

CARROLL: LOWER PERMIAN MICROSAUR

903

table 1 . Cranial dimensions of Palaeozoic and modern tetrapods. Area of sensory capsules and braincase are
expressed as a percentage of the total skull area following the model of Hanken (1983). Measurements of
Thorius and Pseudoeurycea from Hanken (1983). Measurements of Ambystoma from specimens in Redpath
Museum. Measurements of fossils from literature cited in text.

Species

Skull

length

(mm)

Olfactory

capsule

(%)

Optic

capsule

(%)

Otic

capsule

(%)

Braincase

(%)

Quasicaecilia lexana

15

1 1

13

20

18

Cardiocephalus stembergi

16

5

4

14

14

Brachydectes elongatus

18

7

9

19

23

Rhynchonkos stova/li

18

8

9

13

18

Doleserpeton annectens

1 1

8

25

9

18

Tersomius texensis

58

5

16

5

13

Phlegethontia longissima

23

6

15

12

20

Thorius pennatulus

3-3

14

27

17

37

Pseudoeurycea goebeli

10 6

17

18

11

23

A mby stoma jefferson ianum

15

9

1 1

1 1

22

Ambystoma tigrinum

19

14

12

1 1

15

margin of the skull, inclosing a circular space lateral to the opening for the olfactory tract. The
extent of the nasal capsule is the most difficult of the sensory organs to estimate, but judging from
modern salamanders, it is probably the least important in controlling the proportions of the skull.

Fortunately, the measures of the relative area of the entire skull and the sensory capsules used
by Hanken can be compared directly with small Palaeozoic amphibians. The adult salamander
specimens he measured range from 3-3 mm ( Thorius pennatulus) to 10-6 mm ( Pseudoeurycea goebeli)
in skull length - all smaller than Quasicaecilia. Study of living forms shows that there is slight
positive allometry shown by the nasal capsules, but strong negative allometry for the relative size
of the otic capsule, eye, and braincase. These allometric relationships apply to both ontogeny within
a particular species and to comparison between adults of different size within related groups.

The relative size of the nasal capsules, eyes, and braincase of Quasicaecilia are similar to those
of salamanders measuring 15 and 19 mm in skull length, but are much smaller than those of the
specimens examined by Hanken. In contrast, the otic capsules of Quasicaecilia are substantially
larger than those of even the smallest salamander measured by Hanken, and approximately twice
the size of those in salamanders in which the skull length is comparable to that of Quasicaecilia. In
fact, nearly all the Palaeozoic specimens that were examined have very large otic capsules. The
relatively large size of the semi-circular canals in Palaeozoic vertebrates has been discussed
previously (Bernacsek and Carroll 1981). Several hypotheses have been advanced, but the specific
physiological reason(s) for larger inner ear structures in early vertebrates has not yet been
established. Presumably, early Palaeozoic tetrapods inherited this condition from their immediate
ancestors among the rhipidistian fish. Although the data remain limited, it appears likely that the
size of the otic capsule is an even more important factor in controlling the configuration of the skull
in Palaeozoic tetrapods than it is in modern genera.

The otic capsules in Quasicaecilia dominate the back of the skull and extend to the margin of the
cheek. As a result of their large size, the jaw suspension occupies a more anterior position and the
adductor chamber is very small. This pattern is closely reminiscent of the early developmental stages
of modern amphibians (DeBeer 1937). Ossification presumably occurred in an individual retaining
essentially embryonic cranial proportions. The small size of the prootic foramen indicates a small
Vth nerve, and reduced importance of the adductor jaw musculature. This, however, cannot be
explained as a direct result of small skull size, since modern caecilians have skulls of even smaller
size, and yet retain a large foramen for the Vth nerve.

The relatively anterior position of the orbits may also be attributed to the large size of the otic

904

PALAEONTOLOGY, VOLUME 33

capsules. In contrast, the configuration of the nasal bones and the absence of articulating surfaces
for the premaxillae probably cannot be attributed simply to small size since the nasals are relatively
long in the related genus Brachystelechus , which has an even shorter skull (estimated length
84 mm). Carrolla , a member of the same family, is approximately 15 mm long and has short nasals,
but suturally attached premaxillary bones. The separation of the cultriform process from the
sphenethmoid cannot be directly attributed to small size either, since the bones are in their normal
position in Carrolla.

Examination of other genera (Text-figs 8 and 9) suggests that constraints related to the sense

A

1cm

qi

text-fig. 8. Skulls of amphibians less than 3 cm in length, all of which have large otic capsules and show
modification in the position of the jaw articulation, adductor chamber and/or bones of the skull table. All x 2.
a-f, microsaurs; a, b, Rhynchonkos [Goniorhynchus], in dorsal and palatal views; C, D, Cardiocephalus
sternbergi , lateral and palatal views; E, Hapsidopareion , in lateral view. F, Odonterpeton , in dorsal view. G, the
aistopod Phlegethontia. h, lateral view of the lysorophid Brachydectes elongatus. (a-f, from Carroll and
Gaskill 1978; G, from McGinnis 1967; h, from Wellstead 1985.)

1 cm

text-fig 9. Outline drawings of the skulls of small Palaeozoic tetrapods showing the areas of the sensory
capsules. Drawn to a common scale, a, Quasicaecilia. b, Cardiocephalus sternbergi. c, Rhynchonkos. D,
Brachydectes elongatus. E, Phlegethontia. F, Doleserpeton. G, Tersomius-, in this genus the extent of the nasal
and otic capsules have been established by serial sectioning.

CARROLL: LOWER PERMIAN MICROSAUR

905

organs may have a more general significance in determining the cranial anatomy of small animals
belonging to a variety of Palaeozoic groups. The size of the orbits is extremely variable in Palaeozoic
tetrapods, and the dimensions of the nasal capsules are difficult to estimate, but the otic capsule
appears to retain a nearly constant size in small adults (Text-fig. 8) - from 13% to 20% of skull
area in skulls less than 2 cm in length. Since the capsules extend nearly to the lateral margin of the
skull, the area of the jaw articulation and/or adductor chamber is typically much modified.

text-fig. 10. Skull roof of A, Adelogyrinus ; b, Asaphestera\ c, Microbrachis. Drawn to the same scale. The
skulls of these genera are all greater than 3 cm long, and none shows evidence of modification resulting from

a large otic capsule.

In the earliest and/or most primitive known members of the Adelogyrinidae and the two
Microsauria suborders, Tuditanomorpha and Microbrachomorpha, the skull is still three
centimeters or more in length (Text-fig. 10). The otic capsules are not known in these forms, but they
would not be expected to be of sufficiently large size to require accommodation in the area of the
jaw articulation or adductor chamber. In contrast, this has occurred separately within several
microsaur families, and in the Lysorophia, Nectridea, and Aistopoda.

Members of the tuditanomorph microsaur families Goniorhynchidae, Hapsidopareiontidae, and
Gymnarthridae, and the microbrachomorph Odonterpeton triangularis have skulls less than three
centimeters in length. In the Goniorhynchidae and small gymnarthrids, the jaw articulation is
anterior to the otic capsule. The jaw articulation remains in a posterior position in the
Hapsidopareiontidae, but in some members of this family the cheek region is emarginated, enabling
the jaw musculature to expand beyond the confines of the restricted adductor chamber. These three
families retain the primitive pattern of the bones of the skull table. In contrast, they are much
modified in Odonterpeton , in which the skull of the only known specimen is 6-7 mm in length. In this
genus, the parietals are greatly expanded and the posparietals are reduced to a tiny median element.

However, there are apparent exceptions to the association between small skull size and
modification of the otic region in two microsaur genera. Within the Hapsidopareiontidae,
Saxonerpeton is known from several skulls, none more than 1-5 cm in length, but the jaw suspension
and cheek region do not appear modified from the pattern of larger microsaurs. Although all the
specimens of Saxonerpeton are moderately well ossified, they may not be of adult size. Several
species of gymnarthrid microsaurs are known from the Lower Permian. They show a progressive
shift in the position of the jaw articulation with decreasing skull size. In Euryodus , with a skull
approximately 35 cm in length, the jaw articulation is posterior in position: essentially in the plane
of the occiput. In Cardiocephalus sternbergi , the skull is approximately T6 cm in length, and the jaw
articulation has shifted forward, 30% of the skull length (Carroll and Gaskill 1978, fig. 105). On

906

PALAEONTOLOGY, VOLUME 33

the other hand, a well-preserved skull of Sparodus , from the Upper Pennsylvanian, with a skull
3-3 cm long (Carroll 1988) also has the jaw articulation far forward. This may be attributed to a
change in jaw mechanics associated with the large size of the marginal and palatal dentition, rather
than to small size.

Among other lepospondyls, all members of the Lysorophia possess relatively small skulls,
dominated by sensory structures. The jaw articulation is anterior to the otic capsule and the
adductor chamber is open laterally (Wellstead 1985).

The configuration of the skull of aistopods (McGinnis 1967; Thomson and Bossy 1970;
Wellstead 1982) is highly modified. All have an open cheek, although the jaw articulation retains
a relatively posterior position in large individuals. In none of the described species is the adult skull
less than 2 cm long, but the bones of the braincase are fused to one another and the otic capsules
are large in Phlegethontia.

text-fig. 11. Nectridean skulls, a, Ptyonius marshii (from Bossy 1976). b, c, skull roof and palate of
Scincosaurus crassus (from Milner 1978). Drawn to the same scale.

Ptyonius marshii (Text-fig. 11) is an early nectridean that has a skull shape resembling that of
primitive labyrinthodonts except for the loss of the squamosal embayment, and retains all the
primitive roofing bones except the intertemporal. The largest known skull is only 19 mm long. In
contrast with some other nectrideans, the carpals and tarsals are not ossified, which suggests that
even the largest specimens may not have reached maturity. Scincosaurus crassus , from the Upper
Carboniferous of Czechoslovakia, has the smallest skull among possibly mature nectrideans -
approximately 7 mm in length. What is known of the otic capsule indicates very large relative size,
and the jaw articulation is displaced anteriorly. Postcranial remains are known of a related species
from North America, but they indicate an animal only twice the size of the European form (Milner
1980).

As in modern salamanders (Hanken 1984), the relatively small size of all animals termed
lepospondyls may contribute to the loss of dermal bones from the pattern seen in early
labyrinthodonts. The fact that different bones are absent in microsaurs, nectrideans, and
adelogyrinids suggests that they were lost independently in different lineages, each of which
separately reduced their body size. If all animals termed lepospondyls evolved from a single
common ancestor that had already reduced its skull size significantly, one would expect to observe
a common pattern of bone reduction.

A general problem in interpreting the significance of small size among Palaeozoic tetrapods is
the near impossibility of knowing whether the available specimens are of adult size. Frogs and

CARROLL: LOWER PERMIAN M1CROSAUR

907

salamanders, whose egg size is constrained to less than 10 mm in diameter by the absence of extra-
embryonic membranes (Carroll 1970), may increase their cranial dimensions many fold before
reaching adulthood. This was presumably the case for Palaeozoic amphibians as well (Boy 1987).
Since all amphibians are small at hatching, it is presumably the adult size that imposes the primary
constraint on the dimensions of the sensory capsules. Only measurements of adults are hence of
significance in evaluating the quantitative effects of small size. Unfortunately, there is no objective
way of establishing whether an individual specimen is an adult. Broadly speaking, large tetrapods
show progressive ossification as they approach adulthood, but species with very small adult size may
become heavily ossified at a much earlier growth stage (Hanken 1982). Hence, extensive ossification
of the ends of the limb bones and the braincase may be evidence that the specimen in hand
represents a species that reaches maturity at small size, but not necessarily that this particular
specimen is of adult dimensions.

None of the genera recognized as labyrinthodonts shows modification of the skull adjacent to
the otic region that can be attributed to small size. Many tiny individuals have been described (Boy
1987) but all appear to be larvae or later immature stages. Bolt (1969, 1977) proposed that the
dissorophoid labyrinthodont Doleserpeton might be close to the ancestry of the modern
'lissamphibian ' orders. Milner (1988) argued that small size was an important factor in the origin
of frogs, salamanders, and caecilians. The skull of Doleserpeton , although little more than 1 cm in
length, retains the proportions of larger dissorophoid amphibians, and the otic capsule occupies
no more than about 9% of its area, much less than that of Quasicaecilia and other small
Tepospondyls’.

Bolt (1979) has suggested that the material of Doleserpeton from Fort Sill might represent juvenile
individuals of a genus resembling Tersomius , which reached a skull length of approximately 6 cm.
The apparently small area of the otic capsule supports this interpretation. Even if Doleserpeton does
belong to an assemblage that is the sister group of frogs and possibly salamanders, most of the
osteological features that are characteristic of the modern orders must have evolved subsequently,
presumably in relationship to significant reduction in adult skull size.

Acknowledgements. I wish to thank the Division of Vertebrate Paleontology of the United States National
Museum for the long-term loan of the type of Quasicaecilia. The final impetus to complete this study resulted
directly from Robert Purdy’s request for the return of this specimen. Many drawings of this specimen were
carefully executed by Mrs Pamela Gaskill during the course of its preparation. Her artistry and patience are
much appreciated. Photographs of this very difficult material were prepared by Mr Murray Sweet. Funds for
this study were provided by the Natural Sciences and Engineering Research Council of Canada. Thanks are
also due to Drs Green, Holmes and Sues who read early drafts of the manuscript.

REFERENCES

bernacsek, G. m. and Carroll, r. L. 1981. Semicircular canal size in fossil fishes and amphibians. Canadian
Journal of Earth Sciences , 18, 150-156.

bolt, j. r. 1969. Lissamphibian origins: possible protolissamphibian from the Lower Permian of Oklahoma.
Science , 166, 888-891.

1977. Dissorophoid relationships and ontogeny, and the origin of the Lissamphibia. Journal of
Paleontology , 48, 434-458.

— 1979. Amphibamus grandiceps as a juvenile dissorophid: evidence and implications. 529-563. In nitecki,
m. H. (ed.). Mazon Creek fossils. Academic Press, New York, 581 pp.
bossy, k. v. h. 1976. Morphology, paleoecology and evolutionary relationships of the Pennsylvanian
urocordylid nectrideans (Subclass Lepospondyli, Class Amphibia). Unpublished Ph D. thesis, Yale
University.

boy, j. a. 1987. Studien fiber die Branchiosauridae (Amphibia: Temnospondyli ; Ober-Karbon-Unter-Perm).
2. Systematische Ubersicht. Neues Jahrbuch fur Geologie und Paldontologie , Abhandlungen , Stuttgart , 174,
75-104.

908

PALAEONTOLOGY, VOLUME 33

Carroll, R. L. 1970. Quantitative aspects of the amphibian-reptilian transition. Forma et Functio , 3, 165-178.

— 1977. Patterns of amphibian evolution: an extended example of the incompleteness of the fossil record.
405-437. In hallam, h. (ed.). Patterns of evolution. Elsevier, Amsterdam, 591 pp.

1987. Vertebrate paleontology and evolution. W. H. Freeman and Company, New York, xiv + 698 pp.

1988. An articulated gymnarthrid microsaur (Amphibia) from the Upper Carboniferous of
Czechoslovakia. Acta Zoologica Cracoviensia , 31, 441-450.

— and CURRIE, p. 1975. Microsaurs as possible apodan ancestors. Zoological Journal of the Linnean Society,
57, 229-247.

— and gaskill, p. 1978. The Order Microsauria. Memoirs of the American Philosophical Society, 126,
1-211.

— and holmes, r. 1980. The skull and jaw musculature as guides to the ancestry of salamanders. Zoological
Journal of the Linnean Society, 68, 1-40.

clack, j. a. and holmes, r. 1988. The braincase of the anthracosaur Archeria crassidisca with comments on
the interrelationships of primitive tetrapods. Palaeontology, 31, 85-107.
dawson, J. w. 1863. Air-breathers of the Coal Period. American Journal of Science, 36, 430-432.
debeer, g. r. 1937. The development of the vertebrate skull. Oxford University Press, Oxford, 552 pp.
Gaffney, E. s. 1979. Tetrapod monophyly : a phylogenetic analysis. Bulletin of the Carnegie Museum of Natural
History, 13, 92-105.

Gardiner, b. G. 1983. Gnathostome vertebrae and the classification of the Amphibia. Zoological Journal of the
Linnean Society, 79, 1-59.

Gauthier, j. a., kluge, a. G. and rowe, t. 1988. The early evolution of the Amniota. 103-155. In benton,
m. j. (ed.). The phylogeny and classification of the tetrapods. The Systematics Association Special Volume,
35 A. Clarendon Press, Oxford, x + 377 pp.

hanken, j. 1982. Appendicular skeletal morphology in minute salamanders, genus Thorius (Amphibia:
Plethodontidae) : Growth regulation, adult size determination, and natural variation. Journal of Morphology,
174, 57-77.

1983. Miniaturization and its effects on cranial morphology in plethodontid salamanders, genus Thorius
(Amphibia, Plethodontidae) : II. The fate of the brain and sense organs and their role in skull morphogenesis
and evolution. Journal of Morphology, 177, 255-268.

1984. Miniaturization and its effect on cranial morphology in plethodontid salamanders, genus Thorius
(Amphibia: Plethodontidae). I. Osteological variation. Biological Journal of the Linnean Society, 23,
55-75.

langston, w., jr. and olson, e. c. 1986. Carrolla craddocki, a new genus and species of microsaur from the
Lower Permian of Texas. Texas Memorial Museum Pearce-Sellards Series, 43, 1-20.
linnaeus, c. 1758. Systema naturae. (10th edn). L. Salvii, Stockholm, 823 pp.

mcginnis, h. j. 1967. The osteology of Phlegethontia, a Carboniferous and Permian aistopod amphibian.

University of California Publications in Geological Sciences, 71, 1 — 49 .
milner, a. c. 1978. Carboniferous Keraterpetontidae and Scincosauridae (Nectridea: Amphibia) - a review.
Unpublished Ph.D. thesis. University of Newcastle upon Tyne.

1980. A review of the Nectridea (Amphibia). 377^105. In panchen, a. l. (ed.). Environment and the origin
of land vertebrates. The Systematics Association Special Volume , 15. Clarendon Press, Oxford, xii + 633 pp.
milner, a. r. 1988. The relationships and origin of living amphibians. 59-102. In benton, m. j. (ed.). The
phylogeny and classification of the tetrapods. The Systematics Association Special Volume , 35 A. Clarendon
Press, Oxford, x + 377 pp.

— smithson, t. r., milner, a. c., coates, M. l. and rolfe, w. d. i. 1986. The search for early tetrapods.
Modern Geology, 10, 1-28.

panchen, a. l and smithson, T. R. 1987. Character diagnosis, fossils and the origin of tetrapods. Biological
Reviews, 62, 341-438.

- 1988. The relationships of the earliest tetrapods. 1-33. In benton, m. j. (ed.). The phylogeny and
classification of the tetrapods. The Systematics Association Special Volume, 35 A. Clarendon Press, Oxford,
x + 377 pp.

Parker, H. w. 1977. Snakes: a natural history. 2nd edition, revised and enlarged by A. G. C. Grandison. British
Museum (Natural History) and Cornell University Press, Ithaca and London, 108 pp. + 16 plates.
rieppel, o. 1988. A review of the origin of snakes. Evolutionary Biology, 22, 37-130.
romer, a. s. 1966. Vertebrate Paleontology (3rd edn). Chicago University Press, Chicago, vii + 468 pp.
THOMSON, K. T. and bossy, K. H. 1970. Adaptive trends and relationships in early Amphibia. Forma et Functio,
3,7-31.

CARROLL: LOWER PERMIAN MICROSAUR 909

walsh, n. m. 1987. Systematics of the caecilians (Amphibia: Gymnophiona). Unpublished Ph.D. thesis,
McGill University.

wellstead, c. f. 1982. A Lower Carboniferous aistopod amphibian from Scotland. Palaeontology , 25,
193-208.

- 1985. Taxonomic revision of the Permo-Carboniferous lepospondyl amphibian families Lysorophidae
and Molgophidae. Unpublished Ph.D. thesis, McGill University.

ROBERT L. CARROLL

Redpath Museum
McGill University

Typescript received 24 July 1989 859 Sherbrooke Street

Revised typescript received 12 January 1990 Montreal, Canada, H3A 2K6

rT

THE EYES OF LOWER CAMBRIAN EODISCID

TRILOBITES

by ZHANG XI-GUANG AND EUAN N. K. CLARKSON

Abstract. The oldest known well-preserved trilobite eyes are described in the Lower Cambrian eodiscids
Neocobboldia chinlinica Lee and Shizhudiscus longquanensis S. G. Zhang and Zhu, the latter being slightly older.
The material, from south central China, is preserved as moulds or by partial replacement in phosphate, and
gives fine details of lens structure. Shizhudiscus eyes have biconvex lenses, polygonal and closely packed as in
normal holochroal eyes, whereas those of Neocobboldia have round, separated lenses, and compare with
Middle Cambrian eodiscid eyes previously described as ‘abathochroal’. The validity of ‘abathochroal eyes’ as
a separate eye-type is discussed. The lenses of Neocobboldia , like those of some phacopids, conform to the ideal
model of an aplanatic lens originally described by Des Cartes. Ontogenetic changes in the eyes are described
for both species. These show that the programmes for growth of the visual surface and lens emplacement are
separate and that the two sometimes got out of phase. The eyes of eodiscids are similar enough to those of other
trilobites to suggest that this group is truly trilobitan. They also indicate how advanced the visual organs were
at this early stage in trilobite history.

Most trilobites bear a pair of laterally placed compound eyes, which are of particular interest
because the trilobite eye is the most ancient visual system so far known. Although well-preserved
trilobite eyes showing details of structure are comparatively rare in the fossil record, it has proved
possible to build up a fair understanding of their structure and function. Following the initial
studies of Clarke (1889) and Lindstrom (1901), two types of eye, holochroal and schizochroal have
been recognized; a third type, abathochroal, was proposed by Jell (1975) for Cambrian eodiscids.
Recent work is summarized by Clarkson (1975, 1979) and Miller and Clarkson (1980).

Schizochroal eyes are believed to have been derived from a holochroal-eyed ancestral type and
holochroal eyes are the earliest to appear in the fossil record. But although the record of different
kinds of trilobite eyes in the Ordovician and later is good, our knowledge of Cambrian eyes is
sparse. In olenelloids, which are considered to be the earliest trilobites found in continuous rock
sequences spanning Late Precambrian to early Cambrian times (e.g. in Morocco, North America
and Siberia: Fortey and Whittington 1989), the eye may occasionally be preserved. For example,
Palmer and Halley (1979, pi. 1, fig. 12) figured the visual surface in Bristolia and lenses have been
reported in a few meraspides of Olenellus gilberti Meek from Alberta (Clarkson 1973, text-fig. 1).
Eyes are also present in some Australian Middle Cambrian eodiscids described by Jell (1975), and
in Upper Cambrian leptoplastine olenids described by Clarkson (1973). Occasional librigenae with
eyes attached belonging to other Upper Cambrian taxa have been figured in various papers, though
they have not been studied in detail. The principal reason for the paucity of trilobite eyes in the
Cambrian was originally suggested by Opik (1967): it is that in Cambrian trilobites the visual
surface has been bounded by a circumocular suture, comprising the palpebral suture above the eye
and the ocular suture below. This circumocular suture, according to Fortey and Whittington (1989),
is a primitive character for all trilobites, including agnostoids and olenelloids, and loss of this suture
by retention of the eye on the cheek is a derived character. When a trilobite whose eye was bordered
by such a suture moulted or died, the visual surface would dehisce entirely and thus would not be
preserved. This accords with the observations of Clarkson (1973) who showed that in the olenids
the ocular suture became functional only in the adult, and that its retention in many later olenids,
and in the vast majority of post-Cambrian trilobites, was due to paedomorphosis. Only because of
this fortunate circumstance does the eye in more advanced trilobites remain attached to the
librigena, so that it has a good chance of being preserved.

| Palaeontology, Vol. 33, Part 4, 1990, pp. 911-932, 4 pis.)

© The Palaeontological Association

912

PALAEONTOLOGY, VOLUME 33

The Cambro-Ordovician Agnostida are generally considered as blind and lacking facial sutures,
and ocular structures are known only in Oculagnostus frici , as reported by Ahlberg (1988).
However, in the Cambrian eodiscid Family Pagetiidae Kobayashi, eyes and proparian sutures have
been known for some time, though knowledge of these was limited until Jell’s (1975) description of
the eyes of Pagetia which he proposed as a new type of eye in trilobites - the abathochroal eye.
Abathochroal eyes superficially resemble small schizochroal eyes, having a relatively small number
of lenses (usually less than 100); these are biconvex and separated from one another. There is,
however, no indication of deep intralensar sclera between the lenses, and in this respect they differ
from schizochroal eyes proper. Jell (1975) suggests that each lens had its own separate corneal
membrane, fixed to the intralensar area round the lens margin. If these differences can be confirmed,
then the abathochroal eye can truly be regarded as a distinct type, but our material does not permit
a final resolution of this issue.

The two eodiscid genera discussed in this paper both come from the Lower Cambrian of south
central China but from different horizons. For both Neocobboldia and Shizhudiscus large numbers
of phosphatized specimens with fine details of structure, and representing several growth stages,
have been freed from the enclosing limestone by etching with acid. Neocobboldia is restricted to the
Canglangpu Stage. Its eyes share a number of distinguishing features with the ‘abathochroal’ eyes
of Pagetia described by Jell (1975). Shizhudiscus is found at a lower horizon than Neocobboldia , in
the Qiongzhuzi Stage. The eyes of Shizhudiscus are the holochroal. They are the oldest known from
China, they show clearly many primary characters, and indeed are amongst the most ancient
trilobite eyes of all.

LOCALITIES AND STRATIGRAPHY

All specimens described in this paper came from two localities: Xichuan, Henan, and Pengshui,
Sichuan (Text-fig. 1a). Since the former locality has been already described and the stratigraphy
tabulated in detail (Zhang 1987), only a stratigraphical section (Text-fig. 2) for the latter is given
here - the Longquanxi section, which is about 80 km north of Pengshui, in south-east Sichuan
(Text-fig. 1b).

Shizhudiscus is widely distributed in China, and has been recorded from Shaanxi, Guizhou, and
Xinjang as well as from Pengshui, but is always associated with Eoredlichia , Wutingaspis , and
Zhenbaspis. All of these have been regarded as of Early Cambrian (Qiongzhusi Stage) age (Lu et
al. 1982), and this is clearly earlier than Neocobboldia. These genera appeared almost at the

where all specimens of Shizhudiscus longquanensis S. G. Zhang and Zhu investigated in this study were

collected.

913

ZHANG AND CLARKSON: EODISCID TRILOBITE EYES

beginning of the trilobite record in China, directly following a horizon of small shelly fossils.
Although Brasier (1989) suggests that the base of the trilobite-bearing Cambrian in China is
younger than in Morocco and North America, where olenelloids occur, the phosphatized eyes of
Shizhudiscus that have been obtained are the earliest really well-preserved trilobite eyes so far
known from anywhere in the world, and clearly represent a most important stage in their evolution.

In the Longquanxi section there are about ten thin beds or flat lenses of limestone within black
shale. Most of these limestones yield Shizhudiscus , but they obviously cover a certain time span. To
avoid the possible influence of variation that might arise from a mixed population of individuals
that lived at different times, only specimens that came from a single horizon were used in this study.

text-fig. 2. Columnar section through the Lower Cambrian at Longquanxi, Pengshui, Sichuan, China.

914

PALAEONTOLOGY, VOLUME 33

PRESERVATION

The Lower Cambrian trilobites discussed here, Neocobboldia and Shizhudiscus , have all been
preserved by phosphatization and most of them have been extracted from the rock matrix by
solution with 3-5% acetic acid. It is very probable that the cuticle was originally composed mainly
of calcium carbonate as is the case with other trilobites, but there may have been an additional
primary phosphate component. The exuvia are preserved either by encrustation with secondary
phosphate, or by replacement of the cuticle itself by phosphate; sometimes there may be both
replacement and encrustation. These modes of preservation are evident in specimens dissolved out
of the rock, but the extent of phosphatization in our material, even within the same sample is very
variable. When a block whose surfaces are replete with trilobite exuvia (Text-fig. 3a) is dissolved in
acid, the number of specimens actually recovered as three-dimensional objects may be many fewer
than expected. This is because many specimens are hardly phosphatized at all, the others in varying
degrees. Thin sections confirm that only some of the specimens have phosphate coats or patches of
secondary phosphate within the cuticle. Whether, and to what extent an individual exuvium was
preserved by encrustation depended upon a delicate balance of surface chemical conditions.

text-fig. 3. a, mode of occurrence of Shizhudiscus on bedding planes within concretions, Longquanxi section,
Gr. I. 69317, x9. b, thin-section showing both phosphatized black rims on the trilobite cuticle, and invasion
of cuticle by phosphate, Gr. I. 69318, x 85. c, thin section showing phosphatized black rims on the cuticle and
phosphatic overgrowth on lower surface, Gr. I. 69319, x 85.

Preservation of Shizhudiscus. Where preservation by encrustation has taken place, the secondary
phosphate always coats both the internal and external surfaces. What is visible from the outside
therefore depends largely upon the relative thickness of the secondary phosphate cover. Where the
phosphate is very thin and forms only a light coating on the outer surface, the surface of the eye
with its convex lenses is more or less perfectly replicated by the thin phosphate film. Both the
internal and the external surface may be preserved in this way. If the secondary phosphate coating
is thicker, however, the lenses may be partially or totally obscured.

ZHANG AND CLARKSON: EODISCID TRILOBITE EYES

915

For most specimens of Shizhudiscus replacement is common, and specimens usually have both the
outer and inner surfaces of the cuticle phosphatized, so that fine details of ornamentation on the
outer surface of the carapace is exceptionally well preserved (Text-fig. 4a-e; PI. 1, figs 8 and 9). This
is quite different from that on the surface of the internal layer. In thin section the phosphate
replacing both the outer and inner layers shows as a thin opaque film on both the outer and inner
surfaces (Text-fig. 2b, c). It seems that the outermost and innermost cuticular layers are easily
replaced by phosphate, since it has been found that in many specimens only these layers are altered,
leaving the inner parts unchanged (PI. 1, figs 2, 5, 6, 7). Although most cuticles seen in thin section
normally show only black phosphatic rims, there are some instances in which the phosphate shows
as a black amorphous mass invading the inner part of the cuticle itself (Text-fig. 3b). In such cases
the main internal part of the cuticle could be preserved by replacement. While such phosphatization
of the internal cuticle is less common, confirmatory evidence that it does happen comes from acid-
etching. The application of a 10% aqueous solution of ethylenediaminetetracetic acid (disodium
salt) to a polished surface of the rock lightly dissolves the carbonate matrix so that the phosphatized
outer and inner layers stand up from the surface. While these phosphatic shells replacing the outer
and inner surfaces of an individual specimen appear quite distinct, the occasional partial filling or
presence of irregular phosphate bars connecting the two layers indicate that the cuticle itself has
been phosphatized.

Some specimens, however, have a phosphate coating on their outer surface (PI. 2, fig. 2), and it
is not clear whether the cuticle below it is replaced or not. From a specimen (PI. 2, fig. 1) directly
examined in its rock matrix it is apparent that the outer layer of the eye, which has been partly flaked
off, seems too thick to be considered as a single corneal surface. Its thickness may suggest the
presence of a phosphate coating. It is of interest to note that this unusual preservation of the eye
of Shizhudiscus is quite different from that of other known modes of preservation in trilobite eyes,
for the inner surface of the lenses has been preserved as a continuous thin layer, running smoothly
into the inner surface of the cuticle.

Because the outer layer of the lens is flatter than the inner layer, and the outer surface of the lens
may be partially or totally obscured by a phosphate coating (PI. 2, fig. 2), or to some extent irregular
(PI. 1, fig. la), most observations on lens number, morphology, arrangement, and ontogeny have
been made with reference to internal moulds of the lentiferous surface.

Preservation of Neocobboldia. Only a few specimens of Neocohboldia have been preserved by
phosphatic replacement of the cuticle, but these exhibit many details of fine structure on the outer
surface (PI. 3, fig. 1). Most specimens, however, were preserved by encrustation, and the secondary
phosphate always coats both the internal and external surfaces. Unfortunately, we have obtained
only a very few really small librigenae, and these are almost covered by a thick coating of phosphate.
The smaller the specimens, the more thickly they tend to be covered, so that the early stages in
ontogeny of the eye are less clear than would be desirable. If the external shell of secondary
phosphate is removed, the inner surface of the lens array can be clearly seen as an internal mould
(PI. 3, fig. 8), and likewise the external mould of the surface is visible from the inside. On the whole,
the replication of the visual surface as moulds is excellent, and can give near-perfect details of the
lens surface, but this depends upon the relative completeness of phosphatization. This is apparent
in the series of specimens of Neocobboldia illustrated on Plate 3, figures 3-5, which show different
degrees of secondary phosphatization. In figure 3 only the proximal parts of the lens surfaces are
well preserved, elsewhere the phosphate forms only a dense meshwork; in figure 4 it is more
complete, though replication of the whole lens surface is still not quite perfect, whereas in figure 5,
the surface of the lenses is an exact mould of the original. In such instances the true shape of the
lenses can be appreciated ; in the case of Neocobboldia they conform to the ideal model of a thick
but aplanatic lens, as designed by Des Cartes in 1637 (Clarkson and Levi-Setti 1975).

The presence of so much diagenetic phosphate in these Lower Cambrian rocks is associated with
a major global phosphogenic event around the Precambrian/Cambrian boundary (Cook and
Shergold 1984). The highest concentrations of phosphate in rocks of this age are actually found in

916

PALAEONTOLOGY. VOLUME 33

text-fig. 4. Shizhudiscus longquanensis. S. G. Zhang and Zhu, Lower Cambrian, Longquanxi, Pengshui. a, Gr.
I. 69251, oblique lateral view of a protaspid cranidium, x 128. b, Gr. I. 69252, oblique lateral view of a holaspid
cranidium, showing outline of facial suture, x 70 5. c, Gr. I. 69253, oblique lateral view of a ineraspid
pygidium, x 128. d, e, Gr. I. 69254, protaspid pygidium; d, dorsal view; E, oblique lateral view, x 128.

southeast Asia, whence our material comes, and Cook and Shergold refer specifically to the Lower
Cambrian of southwestern China as one of the major phosphorite areas in the world. The presence
of a rich and specialized Early Cambrian eodiscid fauna in southwestern China (Zhang et al. 1980)
may be associated with very high productivity at this time. The abundance of phosphate was related
by these authors to a period of enhanced oceanic overturn, following a long time period when such
overturn was slight. This brought nutrients up into the photic zone, and whereas much of the

EXPLANATION OF PLATE 1

Figs 1-9. Shizhudiscus longquanensis S. G. Zhang and Zhu, Lower Cambrian, Longquanxi, Sichuan, la, b , Gr.
I. 69255; a, lateral view of smallest right librigena, x 123; b, internal view of same specimen, showing
nineteen lenses (including half-lenses) and their pattern of arrangement, partially eroded but still
recognizable, x410. 2, Gr. I. 69256, internal view of a right librigena with twenty-two lenses, x 123. 3, Gr.
I. 69257, internal view of a right librigena with thirty-four lenses, x 123. 4, Gr. I. 69258, internal view of a
right librigena with forty-two lenses, x 123. 5, Gr. I. 69259, internal view of a right librigena with fifty-one
lenses, x 123. 6, Gr. I. 69260, internal view of a right librigena with sixty-nine lenses, x 123. 7, Gr. I. 69261,
lateral view of the largest left librigena with ninety-six lenses, x 123. 8, Gr. I. 69262, lateral view of a
damaged left librigena showing a half-lens (arrowed) at the posterior margin of the eye. x 123. 9, Gr. I.
69263, lateral view of a left librigena showing the regular arrangement of small lenses, x 123.

PLATE 1

ZHANG and CLARKSON, Shizhudiscus longquanensis

918

PALAEONTOLOGY, VOLUME 33

phosphate so released was used up in animal and algal metabolism, and for building skeletons, some
was made available for diagenesis. The type of preservation we see in the Chinese material is very
similar to that noted by Jell (1975a, b ) in small trilobites preserved during the somewhat later
(Middle Cambrian) phosphate ‘boom’ in Australia.

In general terms, the high-concentration phosphorite episode, which was probably associated
with major evolutionary changes at this time, and has allowed such superb preservation, was over
long before the Upper Cambrian.

GEOMETRY OF THE VISUAL SURFACE IN TRILOBITES

Our material includes a gradational size series of eyes showing various stages in development for
both Shizhudiscus and Neocobboldia. This has enabled us to work out how the visual surface grows
and how the lenses were emplaced. While this is generally similar to that of other trilobite eyes there
are some differences in detail, so a brief review of trilobite eye ontogeny and visual surface geometry
is given here so that our new information on development of eodiscid eyes can be interpreted in
terms of what is already known.

All trilobite eyes known at present, whether holochroal or schizochroal, grow in the same general
way (Beckmann 1951 ; Clarkson 1971, 1975, 1979). They begin their growth as a thin strip of lens-
generating cuticle, the generative zone, just below the palpebral lobe. The first horizontal row of
lenses is produced from the generative zone which moves downward as the visual surface enlarges,
so that new lenses are added at the base of the eye as the generative zone advances, and once
produced are permanently ‘frozen’ into place.

The genetic programme governing the expansion of the visual surface (dominantly downward but
also anteriorly and posteriorly) is independent of that governing the emplacement of the lenses,
though the two are obviously related. When new space is made at the base of the eye by the
generative zone new lenses are formed upon it, but their emplacement follows a defined programme
so that they are normally fixed to the bottom of pre-existing files and they will grow until the
proximity of neighbouring lenses prevents their further expansion.

In nearly all cases the resultant pattern is a hexagonal close packing system but there are many
different versions of this, some involving a change of lens size over the visual surface. Such patterns,
many of which have been analyzed in detail (Clarkson 1975) are different answers to the problem
of packing lenses in a regular pattern from a single marginal generative zone onto a curving visual
surface (Clarkson 1979, p. 6). This is a fundamental problem for the growth of all trilobite eyes
including those of the eodiscids, which, while conforming generally to the normal growth patterns,
exhibit some striking peculiarities all of their own.

EXPLANATION OF PLATE 2

Figs 1-8. Shizhudiscus longquanensis S. G. Zhang and Zhu. Lower Cambrian, Longquanxi, Pengshui. 1, Gr.
I. 69264, a librigena within the rock matrix, showing internal moulds of small lenses and partial preservation
of their outer surfaces, probably the result of encrustation by secondary phosphate, x 128. 2, Gr. I. 69265,
an isolated librigena, showing internal mould of small lenses and secondary phosphate on the outer surface,
x 176. 3, Gr. I. 69266, internal view of a right librigena, showing large lenses, x 324. 4, Gr. I. 69267, oblique
internal view of a broken eye, showing the inner (black arrowed), and outer (white arrowed) layers of small
lenses, x 400. 5a, b , Gr. I. 69268; a, internal view of a left librigena showing the irregular size and
arrangement of the small lenses and the sensory pits below the eye, x 272; b, detail of the same eye, showing
the two layers of lenses, the separate rounded lenses of the upper row, the polygonal lenses in the centre of
the eye, and half-lenses along the lower edge, x 680. 6, Gr. I. 69269, internal view of a left librigena, showing
half-lenses along the lower margin of the eye, unequally developed, x 400. 7, Gr. I. 69270, internal view of
a right librigena, showing half lenses towards the anterior of the eye, x 400.

PLATE

ZHANG and CLARKSON, Shizhudiscus longquanensis

920

PALAEONTOLOGY, VOLUME 33

MORPHOLOGY AND GROWTH OF THE EYE IN SHIZHUDISCUS

Many librigenae of Shizhudiscus were available for this study. Unfortunately none of these are
preserved in their original position, attached to the cranidium. The cranidium, however, clearly
shows the outline of the facial suture, which slopes anteriorly at about 45°, is nearly flat below the
palpebral lobe, and descends steeply posteriorly. This constant shape enables isolated librigenae
that are bounded by the facial suture to be identified as belonging to the left- or right-hand side of
the trilobite (Text-fig. 5). In addition, variation in size of these librigenae indicates that they
represent different growth stages. The smallest we have, according to its size and outline, may have
separated from a cranidium of late meraspid or early holaspid stage, while some of the larger ones
can undoubtedly be referred to the adult stage.

text-fig. 5. Reconstruction of an early holaspid
cephalon of Shizhudiscus longquanensis S. G. Zhang
and Zhu in oblique-lateral view, showing the left
librigena in its original position. Scale bar = 0 2 mm.

The eye of Shizhudiscus forms a relatively narrow, elongated band with subparallel upper and
lower surfaces, somewhat higher at the rear. It is set at the summit of the librigena, which is about
twice the height of the eye itself. The lenses are contiguous and are usually of polygonal form. It
has been observed that as the eye grows, it stays much the same shape, always retaining an elongated
kidney-shaped outline. Some general principles of growth are evident. Firstly, it is readily seen that
as the librigenae increase progressively in size, so the lenses gradually become larger. For some small
librigenae, the average lens diameter ranges from 19-25 //m and for some of the largest ones the
range is from 30—45 //m. Secondly, the smaller the specimen, the fewer the lenses it bears. A
comparison of the different growth stages represented in the series shows how the lenses are
emplaced (Text-fig. 6).

EXPLANATION OF PLATE 3

Figs 1-8. Neocobboldia chinlinica Lee. Lower Cambrian, Xichuan, Henan. I, Gr. I. 69300, oblique lateral view
of a meraspid cranidium with minutely ornamented outer layer, showing the arched palpebral suture, x 104.

2, Gr. I. 69301, oblique dorsal view of a holaspid cranidium, showing the arched palpebral suture, x 105.

3, Gr. I. 69302, poor phosphatization of eye, x 264. 4, Gr. 1. 69303, incomplete phosphatization of eye,
x27L 5, Gr. I. 69304, perfect phosphatization of eye, x268. 6, Gr. I. 69305, outer surface of somewhat
damaged eye, coated with very thin phosphate and showing separated rounded lenses, also well-preserved
internal mould below the flaked-off outer surface, x 268. 7, Gr. I. 69306, external mould of a right librigena,
showing the smooth outer surface of the rounded lenses, and minutely ornamented librigena, x 1 12. 8, Gr.
I. 69307, a left librigena with the outer layer and phosphatic coating flaked off; the irregular polygonal
packing of the lenses on the internal mould is due to erosion of the originally separated lenses, x 104.

PLATE 3

ZEIANG and CLARKSON, Neocobboldia chinlinica

922

PALAEONTOLOGY, VOLUME 33

text-fig. 6. Schematic diagram of early growth stages
of the eye of Shizhudiscus longquanensis S. G. Zhang
and Zhu. a, the smallest librigena with nineteen lenses
(including about six half-lenses), b, a somewhat larger
librigena with twenty-two lenses: the filled circles or
part circles correspond to the lenses or half-lenses in
librigena a; the blank parts represent the increase of
lenses with growth, c, a still larger librigena with
thirty-four lenses, showing lenses or half-lenses
corresponding to b (filled circles or part-circles), and
the increase in number of lenses with growth; the
arrow indicates the anterior. Scale bar = 0-2 mm. See
also Text-fig. 7.

The earliest stage in the development of the visual surface must have been the emplacement of
a single horizontal row of five or six lenses directly below the palpebral suture. In our smallest
specimen, though the phosphate coating is lumpy and irregular so obscuring the details, nineteen
lenses (including half-lenses) can be recognized, of which the middle ones are separated laterally
from each other by small gaps. These lenses and gaps remain a constant feature throughout the
subsequent development of the eye, and can be used as a reference point for tracing the fate of
individual lenses. Following the emplacement of the upper horizontal row a second horizontal row
is then laid down below it, each of the lenses set between and below the lenses of the upper
horizontal row, with the vertical and lateral spacing of the lens centres being about the same. At
this stage the upper margin of the eye, along the palpebral suture is virtually flat. As the eye grows,
further rows of lenses are added below in regular sequence, but at the same time the eye expands
laterally both at the front and the rear. The gradational size series of specimens allows the mode of
addition of extra lenses to be established. For while the visual surface is expanding downwards, the
upper horizontal row receives new lenses both anteriorly and posteriorly, thus extending itself in
both directions, though curving downwards to follow the contour of the facial suture. The upper
horizontal row thus retains its apparent identity, though it has grown by accretion in both directions
away from the centre.

The lenses that lie along the lower margin of the eye are truncated by the margin of the visual
surface so that they appear as semicircular half-lenses. Such half-lenses can sometimes be found also
at the posterior (PI. 1, fig. 8) or anterior (PI. 2, fig. 8) edges. It has been observed that such half lenses
are present in varying degree on every librigena (PI. 2, figs 3, 5, 7). We consider these to be immature,
which implies that, as in other trilobites, the generative zone lies along the anterior, lower, and
posterior edges of the eye, and that lenses were emplaced along this generative zone as the eye grew.
Two alternative models for growth may be considered. The first is that a single lens could only be
fully formed after two moult stages. If this were so, then it may be presumed that the genetic
programme responsible for producing new lenses along this margin went into action before each
moult. While new, soft cuticle was being formed prior to ecdysis, a number of new lenses and
associated visual units below were developing along the lower margin of the eye. The timing of
ecdysis was controlled by an over-riding and different programme, which operated when these
marginal lenses were imperfectly formed, as represented by the ‘half-lens’ stage. At the first moult,
therefore, they were not fully functional; they only became so after the next moult, by which time
they were round and probably were fully functional. The second possibility is that the genetic
programme for lens-development was independent of ecdysis so that new lenses continually
appeared along the lower margin of the visual surface at any time during growth. This is very

*••• a •••

a ' •

ZHANG AND CLARKSON: EODISCID TRILOBITE EYES

923

unlikely, however, since the visual surface is shed as a whole unit, and the hardness of the cuticle
would inhibit the eruption of new lenses. A zone of half-lenses on the lower margin of the eye is
characteristic not only of Shizhudiscus but also of Neocobboldia and the two specimens of Pagetia
figured by Jell (1975, fig. 36d). It is not known from other trilobite eyes and seems to be a feature
unique to the eodiscids.

text-fig. 7. Schematic diagram of growth stages of
the eyes of Shizhudiscus longquanensis S. G. Zhang
and Zhu, showing increase in size and number of
lenses and expansion of the visual surface with
growth. The regular arrangement of the lenses in
curving files or rows can be recognized, but irregu-
larities become more pronounced with growth. The
arrow indicates the anterior. Scale bar = 0-2 mm.

•AV.y.v

A * • '

''.VtVtVtV

'VvWV*

••• • ••

Over most of the eye the system of lens packing is more-or-less regular, and approximates a
pattern of cubic close packing (Text-fig. 7). In such a packing system the lens-centres are equidistant
from one another, so that the immediately evident appearance is of two intersecting diagonal rows
inclined at 45° to the horizontal but at 90° to each other. The vertical files which form the third
intersecting set consist of more widely spaced lenses and are harder to recognize. Cubic close
packing is not common in trilobites. Hexagonal close packing on the other hand is normal: it is a
more economical system which allows better use of the available space, and the accommodation of
relatively more lenses per unit area. The only other recorded example of cubic packing is in the
schizochroal-eyed Phacops turco aff. praecedens Haas, from the Eifelian of Morocco, discussed by
Fortey and Morris (1977), which represents a departure from the normal phacopid packing. It is
interesting, however, that cubic close packing is found in the eodiscids, which bear the earliest
known well-preserved eyes. A cubic system, in which all lens-centres are equidistant is easy to
generate. It can be done by a somewhat simpler genetic developmental programme than a hexagonal
close packing system. Might this be the reason for cubic packing in the Lower Cambrian? Although
the cubic packing system is recognizable, it is never absolutely regular; the vertical files are often
sinuous and in places extra lenses seem to be intercalated. Moreover the lenses are larger and less
regularly spaced posteriorly, which we interpret as a consequence of developmental constraints
imposed by the shape of the visual surface. This must have been controlled by a separate genetic
programme to that governing lens-emplacement, and the irregularities seen have developed because
the separate genetic ‘blueprints’ had slightly conflicting instructions.

924

PALAEONTOLOGY, VOLUME 33

Anteriorly the facial suture curves downwards fairly gently, whereas at the rear it swings much
more sharply downwards in a semicircular curve. It seems that the shape of the facial suture at this
point is the primary determinant of the geometry of lens packing towards the rear of the eye, for
it imposes a constraint on the number and arrangement of the lenses that can be packed in below
it. There is only a limited amount of space in the expanding area enclosed beneath the downwardly
curving upper horizontal row, though the curvature of the visual surface in this region, as seen in
plan, is higher here than in any other part of the eye, and this allows more room for the lenses that
are present to grow. While the regularity of spacing of lens centres breaks down to some extent,
these same centres are more widely spaced than in other parts of the eye, and the lenses make use
of the available space to grow larger (PI. 2, figs 3 and 5). The rather unusual lens arrangement is
based first on the cubic close packing system, with its 45° inclination of the diagonal rows, rather
than the more usual 30° inclination, and secondly on irregularities in the posterior part, mainly
determined by geometrical constraints consequent on the form of the facial suture. The later stages
in the growth of the eye are illustrated in Text-figure 7. When the maximum depth of the visual
surface has been achieved, usually after three or four horizontal rows have been emplaced, the eye
still expands laterally, and the final lenses are added at the lower anterior and lower posterior
margins. The development of the eye, as in other trilobites, must have operated upon a genetically
controlled rolling programme basis. The largest eyes of all still have half-lenses along the base of
the eye; presumably these were ‘frozen’ before their full development had been completed. This
happened at each moult stage, and may also have been the case when the trilobite was fully grown,
and when the genetic programme which finally ended the development of the eye came into
operation.

Although it has not proved possible to measure the angular range of vision in so small an eye as
that of Shizhudiscus nor to make an accurate stereogram of its visual field (cf. Clarkson 1966a, b ,
1975), an eye such as this has a horizontal range of about 70°, the fields of the two eyes not quite
meeting at front and rear, and a latitudinal range from just below the equator to about 30° above
the sea floor.

MORPHOLOGY AND GROWTH OF THE EYE IN NEOCOBBOLDIA

The abundant phosphatized material of Neocobboldia chinlinica Lee, from Xichuan, Henan, has
enabled all stages in the ontogeny to be worked out in detail and statistically processed (Zhang
1989). The smallest librigena we have (Text-fig. 8a), according to its size and outline, corresponds
to a cranidium that has been assigned to an early holaspid stage. The specimens are preserved
mainly by phosphatic encrustation and details of structure are normally best established by studying
internal and external moulds. In some instances, where the phosphate forms a very thin film or in
rare instances of replacement, the external morphology of the visual surface is easily appreciated.

In specimens of average size the eye is an oval body, the dorsal edge of which, bounded by the
palpebral suture, is the more strongly curved. Such eyes have about fifty lenses, though some very
large specimens, in which the palpebral suture is flatter, may have up to eighty (Text-fig. 8e). The
lenses are arranged in a packing system, analyzed below, which is semi-regular. There are three
respects in which the eye of Neocobboldia differs from that of Shizhudiscus ;

1. The palpebral suture forms an arch rather than a horizontal line (PI. 3, figs 1 and 2); a factor
which imposes certain constraints upon the geometry of the visual surface.

2. The lenses are separated one from another, as is clear both from the outer surface and the
internal mould. The eye is thus like that described by Jell (1975) in the eodiscid Pagetia oculata as
‘abathochroal’.

3. The proximal surface of each lens, in well preserved specimens, has a centrally placed nipple,
marked on the internal mould by a corresponding dimple (PI. 4, figs 3 and 6a, b). Such a lens is
similar in form to the ideal thick, but aplanatic lenses designed by Des Cartes in 1637, and detected
in the Ordovician phacopid Dalmanitina (Clarkson and Levi-Setti, 1975). The latter genus, however,
has an additional structure, the intralensar bowl, which in Neocobboldia seems to be lacking.

ZHANG AND CLARKSON: EODISCID TRILOBITE EYES

925

text-fig. 8. a-e, a series of librigenae of Neocobboldia chinlinica Lee, probably representing different holaspid
growth stages. Lower Cambrian, Xichuan, Henan; a, Gr. I. 69308, a left librigena with nineteen lenses
(including half-lenses); B, Gr. I. 60309, a left librigena with thirty-six lenses, outer surface partially preserved;
c, Gr. I. 69310, an external surface of a left librigena with fifty-five lenses; d, Gr. I. 6931 1, internal mould of
a right librigena with sixty-two lenses; E, Gr. I. 69312, an internal mould of a right librigena with eighty-four
lenses; the outline of the eye has changed to elongate ovoid, f, g, two librigenae of an unknown genus; f, Gr.
I. 69400, a librigena with an elongate eye, composed of eighty-three small lenses in close contact; G, Gr. I.

69401, a larger eye with many small lenses. All x 112.

926

PALAEONTOLOGY, VOLUME 33

The visual surface seems to be separable into three zones: an upper row, a central semi-regular
zone, and a basal zone of half-lenses. It is possible, by studying individual growth stages, to work
out how the lenses were emplaced as the eye grew, but because lens-packing is quite variable from
one eye to another (Text-fig. 9), this imposes some limits on interpretation.

B

text-fig. 9. Schematic diagram showing growth stages of the holaspid eye of Neocobboldia chinlinica Lee,
exhibiting the irregular arrangement of small lenses and change in the outline of the eyes, a, b, e, left eyes, c,

D, right eyes. Scale bar = 02 mm.

The upper row, lying just below the palpebral suture and equivalent to the upper horizontal row
of other trilobite eyes, consists of an arched string of lenses. The central three or four lenses of this
chain are usually much smaller than the others of the same row (e.g. PI. 3, figs 5 and 7; PI. 4, figs
3 and 6b)-, they are here termed the ‘initial set’. Sometimes the upper margin of these lenses is
truncated so that they form half-lenses. In many cases this initial set of lenses is separated by a sharp
discontinuity in size or level from the neighbouring lenses, anteriorly, posteriorly, or indeed in both
directions. In rare instances, the lenses of this initial set are fused together (PI. 4, fig. la) as if they
had not separated properly during development. The initial set must represent the lenses which
formed first, at a larval stage earlier than any represented in our material (our smallest specimens
have about twenty lenses). It is clear, both from our gradational size series, and from the
morphology of the adult eye, that new lenses were added both to the front and rear of the initial
set, so that as the visual surface grew downwards, it also expanded anteriorly and posteriorly, more
or less symmetrically, by addition of new lenses below the curving palpebral suture. The upper row
in the adult eye has lenses increasing in size in both directions from the centre, and especially where
the initial set is well marked, is clearly defined and thus seems rather distinct from the less regular
lens-rows below.

EXPLANATION OF PLATE 4

Figs 1-6. Neocobboldia chinlinica Lee. Lower Cambrian, Xichuan, Henan, la, b , Gr. I. 69313, internal mould
of a left librigena; a, showing the rounded lenses, and the ‘fusing’ of some of the lenses (including the ‘initial
set’; arrowed), of the upper horizontal row, x254; b, anterior view, showing the lenses arranged upon a
moderately curved spherical surface, x 170. 2, Gr. I. 69306, detail of an external surface, showing rounded,
separate lenses, which seem to be covered by a common cornea, x 560. 3, Gr. I. 69314, several small lenses
of the ‘initial set’ (including arrowed lens and adjacent lenses to the right), located in the middle of the upper
horizontal row, x 880. 4, Gr. I. 69315, fine annular groove surrounding each lens, on phosphatized internal
mould, x 560. 5, Gr. I. 69316, imperfectly developed half-lenses (arrowed) truncated by later half-lenses of
the basal zone, x 625. 6a, b , Gr. I. 69312, well-preserved internal mould; a, showing centrally placed dimple
for each lens, x 750; b, showing small lenses of the ‘initial set' in the middle of the upper horizontal row,
x 750.

PLATE 4

ZHANG and CLARKSON, Neocobboldia chinlinica

928

PALAEONTOLOGY, VOLUME 33

Under the upper row is a broad, convex zone in which the lenses are arranged in more complex
patterns, which vary quite markedly between one eye and another. Perhaps the most immediately
evident features are second, third, and other rows, roughly concentric with the upper row, but
intercalated lenses usually break down the regularity of arrangement in the lower parts of the eye.
The lenses of the second row are usually of more uniform size than those of the upper row,
emphasising the apparent difference between the two. Such concentric rows, however, form part of
a hexagonal close packing system, which though not fully regular is still fairly well defined, and
which seems to have begun to operate automatically, after the formation of the initial set. It is
important to note that the visual surface retains its ovoid shape throughout development (at least
until the latest stages when it has about eighty lenses and acquires a more elongate form). Thus
concentric ‘fronts’ or zones of growth spread away from the initial set and adjacent lenses, though
the variability of the eye of this species makes it difficult to establish the fate of individual lenses by
direct comparison.

All the specimens, representing the various growth stages, have a zone of half-lenses along the
downwardly bowed lower margin of the visual surface (PI. 3, figs 3-8; PI. 4, figs 1 a and 5),
equivalent to those of Shizhudiscus , and formed in the same way. There are some examples where
lenses towards the base of the eye have themselves been truncated by later lenses. Plate 4, figure 5
clearly shows otherwise rounded lenses with their lower edges overprinted by the last, semicircular,
lenses to be formed. Presumably the developmental programme which terminated the growth of the
visual surface was not fully synchronized with the programme for lens emplacement; the latter
continued for a time after the growth of the visual surface had stopped.

The irregularities which are so evident in all examples studied are of two kinds: those which arise
as a normal consequence of packing lenses of fairly uniform size on a visual surface of this shape
and those which seem to be abnormal. ‘Normal’ growth irregularities are widespread in the eyes
of trilobites, and arise because of conflicting requirements of different programmes which govern
growth (Clarkson 1975, 1979). A more evident malformation, however, is seen in Plate 4, figure la.
This is an example of an eye in which the lenses of the initial set are not properly separated, and
this imperfect separation extends some way anteriorly along the upper row. But it is posteriorly from
the initial set that the lenses become more haphazardly packed, the irregularity affecting the
posterior part of the upper row and also the lenses below it. It is quite probable that this abnormal
arrangement may relate to the initial malformation.

In a very few particularly well-preserved specimens a fine annular groove is visible on the
phosphatized internal mould of the visual surface surrounding each of the lenses (PI. 4, fig. 4). These
thin rings are preserved on the inner surface of the eye only, and probably represent the site of
attachment of sublensar units, presumably of cylindrical form. If this is so, the interior of the eye
may well have possessed a series of short, cylindrical capsules, one attached below each lens, as has
been envisaged for the schizochroal eyes of Phacopina. Perhaps, then the eye is less ‘abathochroal '
(as envisaged by Jell (1975), whose model suggests that the cornea was anchored at the margin of
each lens, on the upper surface) than ‘schizochroal’ and more similar to the eyes of Phacopina.

As with Shizhudiscus , the small size of the eyes of Neocobboidia precludes accurate measurement
of the angular range of vision. Yet, in profile, the eye is much more highly curved (PI. 4, fig. 1/)),
and subtended a latitudinal range of over 90° from the animal's equator to the pole. The
longitudinal range as in Shizhudiscus was above 70° for each eye.

THE ROLE OF PAEDOMORPHOSIS AS AN EVOLUTIONARY CONTROL

Heterochrony, that is ‘ the change through time in the appearance, cessation, or rate of development
of ancestral characters’ (McNamara 1988), is increasingly recognized as a frequent and important
control of trilobite evolution, from Lower Cambrian times onwards. Both paedomorphosis, and
more rarely, peramorphosis have been demonstrated. Paedomorphosis (the retention of the juvenile
characters of the ancestor as adult characters of the descendant) has recently been reviewed by

ZHANG AND CLARKSON: EODISCID TRILOBITE EYES

929

McNamara (1986, 1988), who noted three different categories of paedomorphosis and showed how
their separate effects enable them to be distinguished.

The eye complex in trilobites was subject to paedomorphic changes and in particular to alteration
by neoteny, a reduced rate of development of particular characters relative to that of the rest of the
trilobite. Two such examples of neoteny have so far been demonstrated affecting the eye. The first
is the retention of the visual surface by the librigena in some leptoplastine olenids as a derived
condition (this is a juvenile feature of Olemis and allied genera which becomes fairly common in the
later leptoplastines : see Clarkson 1973). Secondly, there is the origin of the schizochroal eye from
a holochroal ancestral type. It is known that the meraspid eyes of Paladin have comparatively few
round lenses, large with respect to the visual surface and separated from each other by intralensar
sclera. Although this meraspid eye is like a miniature schizochroal eye, this aspect of its morphology
is lost in the adult, where the lenses become closely appressed, and the eye becomes holochroal. It
is from the juvenile condition that the schizochroal eye is considered to have been derived (Clarkson
1975, 1979).

The process of neoteny, so widespread in the history of the trilobites, may be invoked to account
for the differences between the eye of Shizhudiscus and the closely related Neocobboldia. The eye of
Shizhudiscus seems to be holochroal and has fairly thin biconvex lenses, all in contact, except for
the central ones of the upper horizontal row, and of polygonal form. In contrast, the lenses of
Neocobboldia are generally rounded and separate. The Shizhudiscus eye is here regarded as the
‘ancestral’ adult holochroal type. That of Neocobboldia , however, is interpreted as derived from the
juvenile eye of a precursor, whose adult eye was of normal holochroal type, more like that of
Shizhudiscus. In other words, the process is directly analogous to the envisaged derivation of the
schizochroal eyes of phacopid trilobites, though with a lesser degree of specialization. It would be
desirable to have confirmation of this from the juvenile eye of Shizhudiscus , but the early stages are
normally too thickly encrusted by phosphate to allow the relative contiguity of the lenses to be seen.
It is clear, however, that the first-formed lenses of the eye of Shizhudiscus in the centre of the upper
row and the second row are separated one from another; which suggests that in the earliest stages
of development the lenses were not in contact.

If it is correct to assume a neotenic origin for the eyes of Neocobboldia (and implicitly for the
similar ‘abathochroal’ eyes of Pagetia and Opsidiscus described by Jell (1975)), it is of particular
interest that the lenses conform to an idealized aplanatic Des Cartes model, in which the lower
surface of each lens has a central nipple, as do the upper units of the lenses of the early schizochroal
eyes of the Ordovician Dalmanitina , likewise believed to have risen by neoteny (Clarkson and Levi-
Setti 1975). Possibly a similar larval lens structure may have become neotenously modified in the
same way; this is a rather critical area where more evidence may be sought in the future.

COMPARISONS WITH MODERN COMPOUND EYES

Compound eyes in present-day insects and crustaceans consist of radial bundles of cylindrical visual
units, the ommatidia. Each of these is capped by a lens, beneath which is a crystalline cone, the main
dioptric element. Under the crystalline cone, though not always in contact with it, is the
photoreceptive element, or rhabdom. Nerves from the individual rhabdoms link with a large optical
ganglion, lying below the ommatidia, in which the overall mosaic image is processed and enhanced.

Exner (1891) proposed that there were two distinct kinds of compound eye, known as apposition
and superposition eyes. In apposition eyes the rhabdom is directly connected with the base of the
crystalline cone. Each ommatidium is optically isolated from its neighbours and receives light
through its individual lens only. These eyes, which are normal in diurnal organisms (and amongst
marine crustaceans typical of brachyuran crabs, decapod larvae, and some amphipods and
stomatopods) have good resolution, though photon capture is low and the image they form is not
very bright. The lenses of apposition eyes are round or polygonal.

Superposition eyes have short spindle-shaped rhabdoms, each set towards the base of its
ommatidium, and separated from the lower end of the crystalline cone by a clear zone. In these, light

930

PALAEONTOLOGY, VOLUME 33

coming through several adjacent lenses can be focussed on a single rhabdom, and a combined single
image is formed at the level of the photoreceptors. Though this image is bright, it is poorly resolved
by comparison with that of apposition eyes. Such eyes are characteristic of dark-adapted animals,
and visual brightness is achieved at the expense of acuity. In bright light the individual rhabdoms
can be isolated from each other by pigment which migrates along the length of the ommatidium to
form a sleeve, so that the visual system can also function as an apposition eye.

Superposition eyes could not function unless the dioptric apparatus were able to bend the light
rays more strongly than is normal between two media of different refractive index. Since this cannot
be achieved by normal refraction, some kind of supplementary dioptric apparatus is needed. It is
here that there may be some prospect of useful comparison with trilobite eyes, for the operation of
these dioptric elements in crustaceans is directly related to the shapes of the lenses. Three kinds of
superposition eyes have now been recognized in crustaceans:

1. Superposition refracting eyes. In these eyes, which are typical of mysids and euphausids, and
also of insects, the crystalline cone is a ‘lens cylinder’ (Exner 1891), in which the refractive index
is graded, highest along the axis and lowest along the margins. This system permits light to be bent
to a much higher degree than if the crystalline cone had been homogeneous. The cones are circular
in cross-section, and the lenses round or hexagonal.

2. Superposition reflecting eyes. These eyes are found only in macruran decapods (lobsters,
shrimps, etc.). The crystalline cones are ‘square-sided plugs of jelly ’ (Vogt 1975; Land 1976). These
act as orthogonal mirrors, and light reflects from their surfaces at twice its incident angle to produce
an upright image at the level of the photoreceptor. Land has described them as acting optically as
a ‘silvered four-sided prism’. These dioptric elements work entirely by mirror-box reflection at their
surfaces. They have to be square in cross-section to operate, and eyes of this kind are invariably
distinguished by having square lenses.

3. Superposition parabolic eyes. Eyes of this kind have recently been described from portunid
crabs (Nilsson 1988). They have inwardly-curved parabolic crystalline cones, round or square in
cross section. They could easily have originated from apposition eyes.

Recognizing that there are several kinds of modern compound eyes, which can in some respects
be distinguished by the shapes of the lenses as seen externally, is there any basis for comparison with
eodiscids, or indeed with other holochroal-eyed trilobites? One’s first impression might be that if the
holochroal eyes of trilobites can be considered as at all similar to those of recent crustaceans, they
are likely to have been of apposition type; this is because apposition eyes are the basic type from
which various pathways led to superposition eyes of various kinds. The real problem in interpreting
them further is that we do not know if the eyes of any trilobites had crystalline cones below the
lenses. In crustaceans the cones form the main part of the dioptric system; in trilobites the whole
optical function may well have been performed by the lens array alone. If there were cones our
interpretation of how the eyes of trilobites worked is likely to be significantly different than if they
were not.

Some holochroal eyes had square lenses in the lower part of the eye: Paralejurus (Clarkson 1975)
and, as shown here, Shizhudiscus. Such square lenses originated as a consequence of lens-packing
geometry alone. Whereas their shape might suggest that they were pre-adapted for mirror-box
optics, the eye needs to be uniformly curved in order for this to work, and this is not the case in
trilobites. There is an interesting, though remote parallel here between the development of the larval
eye of the decapod Palaemon serratus and the eye of Shizhudiscus. Land (1981) has shown that in
the larval form the eye has hexagonal facets, but as the eye develops these gradually square off. This
happens late in development when the animal is two-thirds its adult size, probably relating to a
change of habitat and an adult preference for dimmer deeper water. The possibility that the eyes of
Shizhudiscus and similar forms were other than of apposition type is suggestive, though not as yet
compelling. If internal structures of trilobite eyes are one day to be found the whole question can
then be re-assessed.

ZHANG AND CLARKSON: EODISCID TRILOBITE EYES

931

Acknowledgements. The eight month visit of Zhang Xi-guang to Edinburgh was financed by the Sino-British
Fellowship Trust and the Royal Society of London, for which we are most grateful. The SEM photographs
were taken in the Faculty of Science Electron Microscope Unit, Edinburgh University, and we thank Colin
MacFarlane and John Findlay for assistance. Cecilia Taylor read the manuscript and suggested a number of
improvements and an anonymous referee gave helpful criticism of an earlier draft of the manuscript.

REFERENCES

ahlberg, p. 1988. Ocular structures in an Ordovician agnostid trilobite. Lethaia , 21, 115-120.
beckmann, h. 1951. Zur Ontogenie der Sehflache grossaugiger Phacopiden. Pa/aontologisches Zeitschrift , 24,
126-41.

brasier, m. d. 1989. China and the Palaeotethyan Belt (India, Pakistan, Iran, Kazakhstan and Mongolia).
42-77 (Chapter 3). In cowie, j. w. and brasier, m. d. (eds). The Precambrian-Cambrian boundary. Clarendon
Press, Oxford, 213 pp.

Clarke, J. 1889. The structure and development of the visual area in the trilobite Phacops rana Green. Journal
of Morphology, 2, 253-270.

Clarkson, e. n. k. 1966a. Schizochroal eyes and vision of some Silurian acastid trilobites. Palaeontology, 9,
1-29.

— 1966 h. Schizochroal eyes and vision in some phacopid trilobites. Palaeontology , 9, 464-487.

1971. On the early schizochroal eye of Ormathops (Trilobita, Zeliszkellinae). Memoires du Bureau de
Reclierches Geologiques et Minieres , 73, 51-63.

1973. Morphology and evolution of the eye in Upper Cambrian Olenidae. (Trilobita). Palaeontology, 16,
735-763.

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

1979. The visual system of trilobites. Palaeontology, 22, I -22.

— and levi-setti, r. 1975. Trilobite eyes and the optics of Des Cartes and Huygens. Nature, 254, 663-667.
cook, p. j. and shergold, j. h. 1984. Phosphorus, phosphorites, and skeletal evolution at the Precambrian-

Cambrian boundary. Nature, 308, 231-236.

des cartes, r. 1637. Oeuvres de Des Cartes. La Geometrie, Livre 2. J. Maire, Leyden, 134 pp.
exner, s. 1891. Die Physiologie der facettirten Augen von Krebsen und Insekten. Deuticke, Leipzig und Wien,
98 pp.

fortey, r. a. and morris, s. f. 1977. Variation in lens packing of Phacops (Trilobita). Geological Magazine ,
114, 25-32.

— and Whittington, h. b. 1989. The Trilobita as a natural group. Historical Biology, 2, 125-138.

jell, p. a. 1975a. Australian Middle Cambrian eodiscoids, with a review of the superfamily. Palaeontographica,
150, 1-97.

— 19756. The abathochroal eye of Pagetia, a new type of trilobite eye. Fossils and Strata, 4, 33-43.
land, M. f. 1976. Superposition images are formed by reflection in the eyes of some oceanic decapods. Nature,

263, 764-765.

— 1981. Optical mechanisms in the higher Crustacea with a comment on their evolutionary origins. 31^18.
In laverack, m. s. and cosens, d. j. (eds). Sense organs. Blackie, Glasgow, 394 pp.

lindstrom, g. 1901. Researches on the visual organs of the trilobites. Kungliga Svenks Vetenskaps Akademiens
Handlingar, 34, 1-85.

lu yan-hao, zhu zhao-ling, qian yi-yuan, lin huan-ling and yuan jin-liang. 1982. Correlation chart of
the Cambrian in China with explanatory text. 28-54. In su zong-wei (ed.). Stratigraphic correlation chart
of China, with explanatory text. Science Press, Beijing, 154 pp.
mcnamara, k. j. 1986. A guide to the nomenclature of heterochrony. Journal of Paleontology, 60, 4-13.

1988. The abundance of heterochrony in the fossil record. 287-325. In mckinney, m. l. (ed.).
Heterochrony in evolution. Plenum, London and New York, 427 pp.
miller, j. and clarkson, e. n. k. 1980. The post-ecdysial development of the cuticle and the eye of the
Devonian trilobite Phacops rana milleri Stewart, 1927. Philosophical Transactions of the Royal Society of
London, Series B, 288. 161-172.

muller, k. J. and walossek, D. 1987. Morphology, ontogeny, and life habit of Agnostus pisiformis from the
Upper Cambrian of Sweden. Fossils and Strata, 19, 1-124.
nilsson, d-e. 1988. A new type of imaging optics in compound eyes. Nature, 332, 76-78.

932

PALAEONTOLOGY, VOLUME 33

opik, A. A. 1967. The Mindyallan fauna of north-western Queensland. Bureau of Mineral Resources of Australia
Bulletin , 74 (2 vols) 1 : 1-404, 2: 1-167.

palmer, a. r. and halley, r. b. 1979. Physical stratigraphy and trilobite biostratigraphy of the Carrara
Formation (Lower and Middle Cambrian) in the Southern Great Basin. United States Geological Survey
Professional Paper , 1047, 1—131.

vogt, K 1975. Optik des Flusskrebsauges. Zeitschrift fur Naturforschung, 30C, 691.

ZHANG, WEN-TANG, LU YAN-HAO, ZHU ZHAO-LING, QIAN YI-YUAN, LIN HUAN-LING, ZHOU ZHI-YI, ZHANG SEN-GUI
and yuan jin-ling. 1980. Cambrian trilobite faunas of southwestern China. Palaeontologia Sinica , New
Series C, 16, 1-497.

zhang xi-guang. 1987. Moult stages and dimorphism of early Cambrian bradoriids from Xichuan, Henan,
China. Alcheringa , 11, 1-19.

- 1989. Ontogeny of an early Cambrian eodiscoid trilobite from Henan, China. Lethaia , 22, 13-29.

ZHANG XI-GUANG

Chengdu Institute of Geology and

Mineral Resources

North Renmin Road, Chengdu, Sichuan 610082
People’s Republic of China

E. N. K. CLARKSON

Department of Geology and Geophysics
University of Edinburgh

Typescript received 4 August 1989 King's Buildings, West Mains Road

Revised typescript received 10 December 1989 Edinburgh EH9 3JW, Scotland

A NEW GENUS OF ABROGRAPTID GR APTOLITE
FROM THE ORDOVICIAN OF SOUTHERN

SCOTLAND

by ISLES STRACHAN

Abstract. The new genus and species, Metabrograptus scoticus , of the family Abrograptidae is described from
the Upper Ordovician (Glenkiln Shales) of southern Scotland. The new genus bears the same relationship to
Abrograptus as Dicranograptus bears to Dicellograptus. The recognition of this further type in the family
perhaps adds some weight to Finney’s (1980) reference of Reteograptus to the Abrograptidae. Although there
are problems of homologizing stipe structure, it is clear that the biserial retiolitoid condition in Ordovician
forms can be derived in more than one way.

The Family Abrograptidae was erected in 1958 by Mu to accommodate his new genus Abrograptus
and Dinemagr aptus Kozlowski, 1951. The family was characterized by the extreme reduction of
periderm in forms with a modified dichograptid type of development and, in conformity with his
views on trends in the evolution of graptolites, the family was seen as a different line from the
Retiolitidae, although similarities with Reteograptus were noted. In particular. Mu (1958, p. 265)
mentioned that the ‘split specimen referred to Retiograptus geinitzianus Hall by Elies and Wood
(1908, p. 316, fig. 209c) resembles closely Abrograptus in general aspect’.

This specimen (BUI 3366) is on a card from the Lapworth Collection with two other specimens.
One of these (BUI 336a) was figured by Elies and Wood (1908, pi. 34, fig. Id) as ‘part of ventral
lattice’ of R. geinitzianus, and the third specimen (BUI 336c) I presumed was the counterpart of the
first. When I had them photographed recently, however, it was clear that this was a second specimen
with the stipes united proximally. The card is labelled in Lapworth’s writing as Clathrogr aptus
cuneif ormis .

In his detailed description and revision of Hall’s Reteograptus geinitzianus, Finney (1980) pointed
out the differences between Reteograptus and the other members of the subfamily Archiretiolitinae,
in particular the simple clathrial framework and sicular details in Reteograptus which are very
similar to those of Abrograptus. He therefore transferred the genus to the Abrograptidae which he
emended to include biserial forms.

The genera Parabrograptus Mu and Qiao, 1962 (with a pair of threads from the sicular apex
forming a third ‘branch’) and Protabrograptus Ni, 1981 (with the sicula lying along one stipe) have
been added to the family, the former genus being also recorded from Canada (Jackson 1966). Mu
and Qiao (1962) also described another new genus, Jianshanites, which has single filaments like
Dinamagraptus but which shows branching. Its sicula is much stouter than in other genera of
Abrograptidae.

In the absence of fusellar details of the proximal end, it is difficult to decide on the developmental
type present. Kozlowski (1951) regarded Dinemagraptus as having dichograptid development in
keeping with its simple structure and Arenig age. Mu and Qiao (1962) have added a second species,
D. sinicus , which is of Caradocian age ( N . gracilis zone), associated with Abrograptus. In their
description of Parabrograptus, from the same horizon, they mention the occurrence of ‘an
additional oblique filament in the crossing canal region’ which possibly indicates the presence of a
second crossing canal as in isograptid or leptograptid development. The third ‘branch’ could then
be interpreted as similar to that seen in a number of specimens of Leptograptus (centribrachiate

(Palaeontology, Vol. 33, Part 4, 1990, pp. 933— 936.|

© The Palaeontological Association

934

PALAEONTOLOGY, VOLUME 33

forms of Elies and Wood 1903, pis 14 and 15). Protabrograptus could similarly be interpreted as
showing the position of the sicula seen in Dicellograptus exilis Elies and Wood (Strachan 1986).
Finney regards Reteograptus as having isograptid development since he recognizes two crossing
canals but notes that ‘it is peculiar because all proximal thecae grow upward’ (Finney 1980). In the
absence of fusellar structure, it would, I think, be difficult to prove that th22 was derived from th 1 2
rather than th21? i.e. that the development was isograptid rather than diplograptid or dicellograptid.
Unfortunately, not enough is known about the development in various Dicellograptus and
Dicranograptus to know how they relate to the diplograptid story.

There is a further problem with Finney’s (1980) placing of Reteograptus in the Abrograptidae.
This relates to the nature of the lists of the stipes. Finney calls them ventral lists in Abrograptus but
Bulman (1970) regarded them as dorsal, although lateral might be the best term since they are paired
on each theca. At any rate, they are simple straight or slightly curved structures with the apertural
loops separating off each section of the stipe. In R. geinitzianus , the ventral lists of Finney are curved
or angular structures which are clearly not strictly ventral since the apertural loops occur in the
middle of each curved list. The upper part of the curve, above the aperture, is a cross structure
between the aperture and the succeeding transverse list, while the lower part below the aperture
represents the lateral boundary between the thecae. This is clearly seen in his figure of the lectotype
(Finney 1980, fig. 12c) where the thecal outline is preserved. It is therefore difficult to homologize
the list structure in Abrograptus and Reteograptus.

The specimen figured by Elies and Wood (1908, p. 316, fig. 209c) shows the stipe characters of
Abrograptus and cannot easily be regarded as a broken specimen of Reteograptus since there is no
trace of the transverse or septal lists. Their illustration of this specimen unfortunately does not show
the position of the sicula. In Abrograptus the sicula projects above the line of the stipes as in
Dicellograptus but another specimen in the Lapworth Collection shows a proximal end with the
sicula in the lower half exactly as in Dicranograptus irregularis Hadding where the biserial portion
consists of only two or three pairs of thecae (Strachan 1986).

While checking through the collections of the British Geological Survey in Edinburgh, I was
shown a partially exposed fragment identified as R. geinitzianus which matched the one on the
Monograph plate. At my suggestion some matrix was removed by Peter Brand and a second stipe
was revealed with the proximal end like a dicranograptid. We thus have four specimens showing this
character and it seems reasonable to consider them as a new genus of Abrograptidae paralleling
Dicranograptus in the Dicranograptidae.

SYSTEMATIC PALAEONTOLOGY

Specimens numbered BU are housed in the School of Earth Sciences, University of Birmingham,
the GSE specimen is in the British Geological Survey, Murchison House, Edinburgh.

Genus metabrograptus gen. nov.

Diagnosis. Rhabdosome small, represented only by clathrial framework of a biserial proximal part
(of ?four thecae in the type species) and two uniserial stipes diverging at 80 to 110°; sicula
sclerotized normally.

Type species. Metabrograptus scoticus gen. et sp. nov.

Derivation of name. Meta (Greek - with) and Abrograptus as other new genera in the family bear Greek
prefixes. The specifit epithet indicates the country of origin.

Stratigraphical age and distribution. Upper Ordovician, Glenkiln Shales, probably Nemagraptus gracilis Zone,
latest Llandeilian - earliest Caradocian, from the south of Scotland.

935

STRACHAN: ABROGRAPTID GRAPTOLITE FROM SCOTLAND

Remarks. This genus bears the same relation to Abrograptus as Dicranograptus does to
Dicellograptus, that is, Metabrograptus has a short biserial portion whereas Abrograptus does not.

Metabrograptus scoticus gen. et sp. nov.

Text-fig. 1a-d, f

v 1908 Retiograptus Geinitzianus Hall; Elies and Wood (pars), p. 316, pi. 34, fig. Id. fig. 209c.

text-fig. 1. a-d, f, Metabrograptus scoticus gen. et sp. nov. ; a, holotype, BUI 3366; b, BUI 336c; c, GSE 14649;
d, BU2150a; F, reconstruction of proximal end. e. Reteograptus geinitzianus Hall. Young growth stage,
redrawn from Finney 1980, fig. 16j. All figures approximately x 5, except f which is x8.

Types. Holotype, BU13366, figured Elies and Wood, 1908, p. 316, fig. 209c. Glenkiln Shales, Birnock.
Paratypes, BUI 336c, same locality and horizon as holotype; BU2150u, b, locality unknown, Glenkiln Shales;
GSE14649, Glenkiln Shales, Portslogan.

Description. The rhabdosome is quite small, the length of the biserial portion being 10 to I 5 mm, and the stipes
up to 8 mm long. The sicula appears to be normally sclerotized but details of the proximal end are poor in all
specimens although they can be matched to some extent with Finney's (1980) figures of young stages of
Reteograptus which supports its interpretation as biserial. The sicula is about 0-4 mm long. There is no distinct
trace of a separate nema. The stout thread from the apex of the sicula forms part of the clathrial framework
and ends at the level of the axil. A single strand from about the aperture of the sicula apparently represents
the ventral line of the first theca and passes into an apertural loop. A second shorter thread connects the loop
to the apex of the sicula. Above the loop, there are a pair of threads marking the lateral walls of th2'. On the
opposite side of the sicula, a thread arising a short distance above the sicular aperture represents the ventral
wall of the second theca and its apertural loop is connected on the obverse side by a short thread to the ‘nema’
and on the reverse side by a longer oblique thread to the centre of the axil. Above the loop there are again a
pair of threads marking the walls of th22.

The uniserial stipes consist of a pair of longitudinal threads divided into curved segments with apertural
loops (appearing mainly as cross-bars) between the segments. The largest specimen has nine thecae on one of
the stipes. The thecal spacing is 9-10 per cm. The stipes diverge at 80-1 10° from a broad axil which is clearly
defined by paired threads in all specimens, thus differing from Abrograptus and Par abrograptus where there is
only a single thread. The breadth of the axil varies from 1 4—1 6 mm, the narrower axil of BUI 336c (IT mm)
being due to oblique compression.

Remarks. Although the biserial part is badly preserved in the specimens available, comparison with
Finney’s (1980) growth stages of Reteograptus shows some strong similarities. A basic pattern of

936

PALAEONTOLOGY, VOLUME 33

four unequal meshes seems likely, ending in a pair of threads forming the axil which in one specimen
at least can be divided into three subsections. The position of the thecal apertures is obscure but it
is likely that the second pair of apertures is at the base of the uniserial stipes. An attempted
reconstruction is given but much of it is speculative.

The curvature of the thecal lists shown in BU2150 is only hinted at in the Chinese figures of
abrograptids which mainly show what are scalariform views of the stipes with the apertural lists
appearing as cross bars. Similar views are seen in the Birnock specimens. The Portslogan example
appears to be very heavily sclerotized although fragmentary but shows the same sort of contrast
with the other specimens as is shown by Finney’s old and young growth stages of Reteograptus
(Finney 1980, cf. figs 13f and 16j).

The four specimens appear to be conspecific although they are from different localities. They are
all clearly abrograptid in their uniserial stipe characters and distinct from previously described
forms in the stout proximal end. Thus a new generic name is justified since the difference is
paralleled in the Dicranograptidae. Whether the Abrograptidae should include fully biserial forms
as Finney has advocated remains I think open to question. While the proximal clathrial structure
which he has demonstrated in Reteograptus agrees well with that in Abrograptus , there is the
problem that functionally when the periderm is reduced to a simple series of lists there is little scope
for variation if the lists have a supporting function.

Associated fauna. The Birnock specimens are associated with a young specimen of Cryptograptus tricornis
(Carruthers) and a slender stipe probably of Nemagraptus. The other Lapworth specimen, which is labelled
Coenograptus ( Nema .) pertenuis , has slender stipes presumably of this form as well as Orthograptus cf.
apiculatus Elies and Wood, Dicellograptus exilis Elies and Wood, Hallograptus cf. mucronatus (Hall) and
Corynoides calicularis Nicholson, indicating a horizon in the Glenkiln Shales.

Acknowledgements. I thank Jim Allan (St Andrews University) for photographing the specimens to provide a
basis for the drawings and Peter Brand for looking out possible material from the Survey collections in
Edinburgh.

REFERENCES

bulman, o. m. b. 1970. Graptolithina. In teichert, c. (ed. ). Treatise on invertebrate paleontology. Part V
(revised). Geological Society of America and University of Kansas Press, Boulder, Colorado and Lawrence,
Kansas, xxxii+ 163 pp.

elles, G. l. and wood, E. m. r. 1901-18. A monograph of British graptolites. Monograph of the
Palaeontographical Society, 1-539, 52 pis.

finney, s. c. 1980. Thamnograptid, dichograptid and abrograptid graptolites from the Middle Ordovician
Athens Shale of Alabama. Journal of Paleontology, 54, 1184-1208.

JACKSON, D. E. 1966. On the occurrence of Par abrograptus and Sinograptus from the Middle Ordovician of
Western Canada. Geological Magazine, 103, 263-268.

kozlowski, R. 1951 . Sur un remarquable graptolithe ordovicien. Acta Geologica Polonica, 2, 86-93 [in French],
291-299. [In Polish],

mu, a. t. (Mu Enzhi) 1958. Abrograptus, a new graptolite genus from the Hulo Shale (Middle Ordovician) of
Kiangshan, western Chekiang. Acta Palaeontologica Sinica, 6, 259-267, I pi. [In Chinese and English],

mu, a. T. 1974. Evolution, classification and distribution of Graptoloidea and graptodendroids. Scientia Sinica,
17, 227-238.

mu, a. t. and qiao, x. d. 1962. New material of Abrograptidae. Acta Palaeontologica Sinica , 10, 1-11, 2 pis.
[In Chinese and English].

ni yu-nan 1981 . Two new genera from the Ningkuo Formation (Lower Ordovician) of Wuning, North Jiangxi.
Special Paper of the Geological Society of America, 187, 203-206, 1 pi.

strachan, I. 1986. The Ordovician graptolites of the Shelve District, Shropshire. Bulletin of the British Museum
( Natural History ), (Geology), 40, 1-58.

ISLES STRACHAN

12 St Nicholas Steading

Typescript received I August 1989 St Andrews, Fife, KYI 6 8LD

Revised typescript received 23 October 1989 Scotland

ON THE GRAPTOLITES DESCRIBED BY
BA I L Y (1871) FROM THE SILURIAN OF
NORTHERN IRELAND AND THE GENUS
STREPTOGRAPTUS YIN

by DAVID K. LOYDELL

Abstract. The graptolite faunas described by Baily in 1871 from County Down, N. Ireland are re-examined
and assigned to the middle part of the Monograptus turriculatus Biozone of the Telychian (Upper Llandovery).
One of Baily’s new species, Graptolithus plumosus , is redescribed and a neotype selected. It is suggested that
G. plumosus is the valid type species of the genus Streptograptus Yin, 1937, but was misidentified by Yin as M.
nodifer Tornquist, 1881. An emended generic diagnosis for Streptograptus is given.

In 1871 William Hellier Baily described the fossils collected during the mapping by the Geological
Survey of Ireland of Sheets 49, 50 and part of 61 (including the country around Downpatrick, and
the shores of Dundrum Bay and Strangford Lough, County of Down). These were exclusively
graptolites and were collected from three localities only (Baily 1871): (a) Loc. 1. Railway cutting,
0-75 mile SE of Annacloy Bridge, 3 miles NW of Downpatrick; (b) Loc 2. Ballytrustan, 4 miles E
of Downpatrick ; (c) Loc. 3. Tieveshilly, a little N of Carrstown Burn, 2 miles SE of Portaferry.

Baily assigned the graptolites to three species, Graptolithus priodon, G. plumosus and G. gradatus ,
the latter two of which he considered to be new species. Baily’s work subsequently received scant
attention in the literature, especially after Lapworth (1876) placed G. plumosus into synonymy with
Monograptus exiguus Nicholson, and did not include it or G. gradatus in his ‘ Graptolites of County
Down , (Lapworth 1877).

The author has recently re-examined some ot the specimens described by Baily after their location
in the Ulster Museum (abbreviated BELUM) by Mr John Wilson. Apart from one slab, from Loc.
1, all the material comes from the Tieveshilly locality (Loc. 3). This re-examination has shown that
G. plumosus is not a synonym of M. exiguus , and the resurrection of G. plumosus will solve a number
of complex taxonomic problems which have been hampering the progress of Upper Llandovery
graptolite research for some time. In addition, a fairly precise age for the graptolite assemblage
(which includes more species than Baily originally identified) can be determined.

DESCRIPTION OF THE GRAPTOLITE FAUNA

All the graptolites are preserved in very low relief in pyrite in dark grey mudstone. There is no
evidence of any tectonic distortion.

The single slab from Locality 1 bears three specimens of Stimulograptus halli (Barrande, 1850) (PI.
1, fig. 7) (misidentified by Baily as G. priodon ), two specimens of Streptograptus plumosus (see
Systematic Palaeontology below), a proximal end of a robust pristiograptid (probably Pristiograptus
bjerringus (Bjerreskov, 1975)), and one specimen of Monograptus cf. barrandei sensu Bjerreskov,
1975, (PI. 1, fig. 2).

The Tieveshilly fauna comprises six species: M. turriculatus (Barrande, 1850) sensu lato; M.
planus (Barrande, 1850) (PI. 1, fig. 1) ( = Baily’s G. gradatus ); an indeterminate small diplograptid ;

(Palaeontology, Vol. 33, Part 4, 1990, pp. 937-943, 1 pi. |

© The Palaeontological Association

938

PALAEONTOLOGY, VOLUME 33

S’, plumosus ; and Stimulograptus halli and M. tuvaensis Obut in Kul’kov and Obut, 1973 (PI. 1, figs
5 and 6) (both of which had been identified by Baily as G. priodon).

SYSTEMATIC PALAEONTOLOGY
Streptograptus plumosus (Baily, 1871)

Plate 1, figs 3 and 4; Text-fig. 1
v* 1871 Graptolithus plumosus ; Baily, pp. 22-23, fig. 1 a-c.

vp 1913 Monograptus nodifer Tornquist; Elies and Wood, pp. 454-6, pi. 46, fig. 2 a-d\ text-fig. 3136
(non u, c, d).

1937 Streptograptus nodifer (Tornquist); Yin, p. 297.

1943 Monograptus ( Streptograptus ) exiguus primulas Boucek and Pribyl, p. 7, pi. 1, fig. 4;
text-fig. 3 e-f.

1975 Monograptus exiguus primulus Boucek and Pribyl; Bjerreskov, p. 62, pi. 9, fig. d; text-fig. 18h.
1986 Streptograptus nodifer (Tornquist, 1881); Chen, pp. 134-6, pi. 1, figs 1-12; pi. 2, figs 1-12;
pi. 3, figs 1-12.

Type specimen. The specimen illustrated by Baily (1871, fig. 1 a-c) is no longer present within his collection. A
neotype has therefore been selected : BELUM : K12275a, a proximal end with thecae up to th7 (figured herein
as PI. 1, fig. 4 and Text-fig. 1). This is from the Monograptus turriculatus Biozone of Tieveshilly, County Down,
Northern Ireland.

Material. Approximately 100 specimens on slabs BELUM: K12274, 12275, 12277 and 12280. Many are
fragmentary, but a few are more complete and include proximal ends.

Diagnosis. Rhabdosome hook-shaped ; straight or very gently dorsally curved proximally, strongly
ventrally curved mesially, becoming more gently ventrally curved or even straight distally.
Prothecae with folds at their bases, otherwise parallel-sided. Metathecae retroverted, terminating in
an upturned lip causing the central part of the thecal aperture to face proximally. Laterally, the
aperture is slit-like. Dorso-ventral width increases from c. 04 mm at th 1 to c. 0-7 mm at th30.

Description. The rhabdosome is hook-shaped, although the degree of curvature is somewhat variable. From
th 1 to 2 or 3 it is straight or very gently dorsally curved : to th6— 9 it is strongly ventrally curved : thereafter
ventral curvature is more gentle and distally the rhabdosome may appear straight. The sicula has a length of
0-95-E2 mm and its apex reaches to from just above the top of thl to half way up th2. Its apertural width is
0-21-0-22 mm. The thecae are uniform throughout the rhabdosome and take up approximately half of its
dorso-ventral width. The bases of the thecae are expanded laterally into prothecal folds (poorly seen on this
low relief material), after which the prothecae are parallel-sided. The metathecae are retroverted, but terminate
in an upturned lip which causes the central part of the thecal aperture to face proximally. Laterally the aperture
is very narrow and slit-like. Thecal overlap is negligible. Details of dorso-ventral and thecal spacing are given
in Table 1 .

EXPLANATION OF PLATE 1

Fig. 1. Monograptus planus (Barrande, 1850), BELUM: K12274.

Fig. 2. Monograptus cf. barrandei sensu Bjerreskov, 1975, BELUM: K12277.

Figs 3 and 4. Streptograptus plumosus (Baily, 1871). 3, BELUM: K122746. 4, BELUM: K12274r/, neotype.

Figs 5 and 6. Monograptus tuvaensis Obut in Kul'kov and Obut, 1973, BELUM: K12280. 5, proximal end. 6,
distal end of same specimen.

Fig. 7. Stimulograptus halli (Barrande, 1850), BELUM: K12276.

All figures x 5. All specimens, except Stimulograptus halli (Fig. 7), are from the middle part of the Monograptus
turriculatus Biozone, Tieveshilly, County Down, Northern Ireland (Baily’s 1871 Loc. 3); the specimen of S.
halli is from the same horizon from the railway cutting, SE of Annacloy Bridge, County Down, Northern
Ireland (Baily’s 1871 Loc. 1).

PLATE 1

LOYDELL, Monograptus, Streptograptus , Stimulograptus

940

PALAEONTOLOGY, VOLUME 33

text-fig. I . Neotype of Streptograptus plumosus
(Baily, 1871), BELUM: K 1 221 Ad, middle Mono-
graptus turriculatus Biozone, Tieveshilly, County
Down, Northern Ireland, x 20.

table 1. Measurements of dorso-ventral width and thecal spacing (2TRD, Howe 1983) for Streptograptus
plumosus (Baily, 1871), based on the ten specimens with proximal ends preserved in Baily’s collection.

Theca

Width (mm)

2TRD (mm)

1

2

0-39-0-52

0-36-0-44

1-22-1-74

3

0-38-0-48

1-30-1-52

5

0-44-0-58

1-34—1-46

10

0 51-0-56

1-30

20

0-63-0-66

1-36—1-59

30

0-70

1-60

Horizon. Bjerreskov (1975) records an almost identical fauna from the middle part of the Monograptus
turriculatus Biozone (Lower Telychian, Llandovery) on Bornholm. The author has also found the same
assemblage of species (except for M. tuvaensis) at this level in central Wales. It would seem very likely,
therefore, that Baily’s material is of middle turriculatus Biozone age.

THE GENUS STREPTOGRAPTUS YIN

The history of this genus is highly complex, involving, in particular, the misidentification of the type
species by the author of the genus and the misinterpretation of thecal morphology in imperfectly
preserved material.

Lapworth’s (1876) incorrect synonymy of Streptograptus plumosus with his Monograptus exiguus
is mentioned above. The species differ in terms of thecal morphology, rate of increase of
rhabdosome width and stratigraphical horizon ( S . plumosus does not range into the uppermost
turriculatus Biozone and M. crispus Biozone, whereas in all correct references to M. exiguus the
associated fauna is indicative of these horizons).

Far more serious, however, particularly when viewed in the light of later events, was the
misidentification of Monograptus nodifer Tornquist, 1881 by Elies and Wood (1901-18) in their
enormously influential A Monograph of British Graptolites. This work became the standard guide
for graptolite identification over much of the world for the next fifty years and is still widely used.
In their text-figures Elies and Wood illustrated more than one species as M. nodifer , but all 4 figures
on their plate 46 are undoubtedly S. plumosus from the turriculatus Biozone of Aberystwyth, Wales.
M. nodifer s.s. does not occur in this zone, but is characteristic in particular of the uppermost
Telychian Monoclimacis crenulata Zone. Although rhabdosome form is occasionally similar to that
seen in S. plumosus , thecal morphology in M. nodifer s.s. has been shown by Rickards et al. (1977,
text-fig. 32) to be quite different.

LOYDELL: IRISH SILURIAN GR APTOL1TES

941

Yin (1937, p. 297) erected the genus Streptograptus, for which he gave the following diagnosis:
‘Polypary with dorsal, ventral or more frequently dorso- ventral curvature, thecae essentially
uniform, being tubes with the whole apertural region coiled into a definite lobe and more or less
twisted; overlap usually insignificant: thecal aperture visible from the obverse view.’ He designated
Monograptus nodifer Tornquist the type species of the genus. However, his specimens from Shihtien
had been collected from a horizon also yielding Monograptus turriculatus and were thus clearly not
M. nodifer s.s. Chen (pers. comm.) has stated that in China all references to M. nodifer are sensu
Elies and Wood and not sensu Tornquist, and indeed this is borne out by examination of the
Chinese graptolite literature. Mu et al. (1962), N.I.G.P. (1974), Wang et al. (1977), Wang (1978),
Chen (1984) and Chen (1986) have all based their identifications upon Elies and Wood’s (1901-18)
description.

The situation was further complicated by Boucek and Pribyl (1943). They considered
Streptograptus to be a subgenus of Monograptus and emended Yin’s diagnosis (Boucek and Pribyl
1943, p. 3): ‘Polyparium schwacher, nach innen gebogen oder s-artig gekriimmt, seltener gerade.
Zellen wenigstens teilweise auf der ventralen Seite der Stocke. Die Zellenspitzen nach innen
eingerollt.’ Unfortunately, the imperfectly preserved material on which they based their diagnosis
resulted in a misinterpretation of the thecal apertural structure. In addition, they created a
subspecies of Monograptus exiguus which they named Streptograptus exiguus primulus. This is a
junior synonym of M. plumosus and is thus, therefore, identical to Elies and Wood’s M. nodifer. It
is accordingly surprising, as Strachan (1971, p. 54) points out, that Boucek and Pribyl placed Elies
and Wood’s M. nodifer into synonymy with M. nodifer sensu Tornquist (i.e. sensu stricto).

Streptograptus has subsequently been widely used outside Britain and North America, either as
a genus or subgenus, for a wide variety of species having or appearing to have lobed, coiled or
retroverted metathecae. These undoubtedly belong to a number of different phyletic groupings.

In Britain a very conservative attitude has quite rightly been followed with regard to the erection
and use of monograptid genera and subgenera. Bulman and Rickards (1970, p. VI 50) state: ‘The
main objection to most such genera is that their erection was not accompanied by any addition to
our imperfect knowledge of their morphology and phytogeny; their context is ill-defined and their
application correspondingly uncertain.’ This has been very much the case with Streptograptus.
However, the time would now seem to be ripe for a re-evaluation of the situation, particularly in
the light of the redescription of Baily’s collection above and the description by Chen (1986) of
chemically isolated specimens of S. plumosus (refigured herein as Text-fig. 2) originally identified by
him as Streptograptus nodifer but now (Chen pers. comm.) recognized as S. plumosus.

As Yin clearly based his generic diagnosis for Streptograptus on Monograptus plumosus and not
on Monograptus nodifer s.s., G. plumosus should be made the type species of the genus
Streptograptus Yin, 1937, and the generic diagnosis should, on the basis of both Baily’s (1871) and
Chen’s (1986) material, be emended. I propose to submit an application to the International
Commission on Zoological Nomenclature to designate the nominal type species Graptolithus
plumosus Baily, 1871 as the type species of Streptograptus Yin, 1937 since the original type was
based on misidentified material.

text-fig. 2. Metathecal morphology of Strepto-
graptus plumosus (Baily, 1871) (modified from
Chen 1986, text-fig. 4) in a chemically isolated
specimen, x75.

942

PALAEONTOLOGY, VOLUME 33

Emended generic diagnosis. Rhabdosome ventrally, dorsally or dorsoventrally curved, rarely
straight. Metathecae retroverted, terminating in an upturned lip causing the central part of the
thecal aperture to face proximally. Laterally the aperture is slit-like. The bases of the prothecae are
usually expanded into prothecal folds. Thecal overlap insignificant. Sicula small.

This emended diagnosis would result in the genus including the following species: Graptolithus
plumosus Baily, 1871; Monograptus pseudoruncinatus Bjerreskov, 1975; Streptograptus filiformis
Chen, 1984; M. petilus Hutt, 1975; M. ansulosus Tornquist, 1 892 ; M. runcinatus sensu Perner, 1897;
Monoclimacis labialis Chen, 1984 ( = Monograptus exiguus A of Bjerreskov, 1975); S. linearis Chen,
1984; Pernerograptus sidiachenkoi Obut and Sobolevskaya, 1965; and a number of as yet unnamed
species which have been compared to Monograptus barrandei sensu Elies and Wood (e.g. Hutt et
al. 1970; Bjerreskov 1975).

M. ( S .) pseudobecki Boucek and Pribyl, 1943 and M. exiguus Lapworth, 1876 (see Loydell 1989)
probably also belong in this genus, but examination of well-preserved material will be necessary
before this can be confirmed. It is worth noting here that the specimens described by Hutt et al.
(1970) and by Bulman and Rickards (1970) as M. exiguus are not this species but are an as yet
unnamed species with apertures laterally expanded and different to those of Streptograptus as
defined herein. The objections of Bulman and Rickards to the genus, based on this material, are thus
unfounded.

Acknowledgements. The author is very grateful to Mr John Wilson of the Ulster Museum, Belfast for arranging
the loan of the specimens and for much useful information. Dr A. G. Sleeman of the Geological Survey of
Ireland initially suggested to me that Baily’s graptolites might be housed there. Useful discussions were had
with Prof. Chen Xu, whilst I was in Nanjing in November, 1988 regarding the Streptograptus problem. He also
made his collections available for study. Petr Storch kindly sent me material of M. (S.) exiguus primulus. Dr
P. D. Taylor and an anonymous referee made useful criticisms of the original manuscript. Dr J. D. D. Smith
gave advice on ICZN matters. Research for this paper was carried out while in the tenure of a University of
Wales Postgraduate Studentship.

REFERENCES

baily, w. h. 1871. Palaeontological remarks. 22-23. In traill, w. a. and egan, f. w. Explanatory memoir to
accompany Sheets 49. 50 and part of 61 of the maps of the Geological Survey of Ireland including the country
around Downpatrick , and the shores of Dundrum Bay and Strangford Lough. County of Down. Alexander
Thom, Dublin and London, 71 pp.

barrande, J. 1850. Graptolites de Boheme. Prague, vi + 74 pp., pis 1-4.

bjerreskov, m. 1975. Llandoverian and Wenlockian graptolites from Bornholm. Fossils and Strata , 8, 1-94,
pis 1-13.

boucek, B. and pribyl, a. 1943. Uber bohmische Monograpten aus der Untergattung Streptograptus Yin.

Mitteilungen der Tschechischen Akademie der Wissenschaften , 52, 1-23, pis 1-3.
bulman, o. M. b. and rickards, R. b. 1970. Classification of the graptolite family Monograptidae Lapworth,
1873. V 1 49—1 57. In teichert, c. (ed.). Treatise on invertebrate paleontology. Part V. Graptolithina , with
sections on Enteropneusta and Pterobranchia. Geological Society of America and University of Kansas Press,
Boulder, Colorado and Lawrence, Kansas, xxxii+163 pp.
chen xu 1984. Silurian graptolites from Southern Shaanxi and Northern Sichuan with special reference to the
classification of Monograptidae. Palaeontologica Sinica, New Series B. 166, No. 20, 1-102, pis 1-19.

- 1986. On Streptograptus and its paleoautecology. Selected papers from the 1 3th and 14th Annual
Conventions of the Palaeontological Society of China , 115-137, pis 1-3.

elles, G. L. and wood, e. m. r. 1901-18. A monograph of British graptolites. Edited by C. Lapworth.

Monograph of the Palaeontographical Society, i-clxxi + a-m + 539 pp., 52 pis.
howe, m. p. a. 1983. Measurement of thecal spacing in graptolites. Geological Magazine, 120, 635-638.
hutt, j. e. 1975. The Llandovery graptolites of the English Lake District. Part 2. Monograph of the
Palaeontographical Society, 57-137, pis 11-26.

— rickards, r. b. and skevington, d. 1970. Isolated Silurian graptolites from the Bollerup and Klubbudden
stages of Dalarna, Sweden. Geologica et Palaeontologica, 4, 1-23.

LOYDELL: IRISH SILURIAN GR APTOLITES

943

kul’kov, n. p. and obut, a. m. 1973. New finds of Lower Silurian graptolites and Chitinozoa in Tuva. Doklady
of the Academy of Sciences U S S R. Earth Sciences Sections , 209, 226-30.
lapworth, c. 1876. On Scottish Monograptidae. Geological Magazine , (2), 3, 308-321, 350-360, 499-507,
544-552, pis 10-13, 20.

- 1877. On the graptolites of County Down. Proceedings of the Belfast Naturalists' Field Club , Appendix,
1876-7, 125-147, pis 5-7.

loydell, d. k. 1989. Monograptus exiguus (Graptolithina) : proposed conservation of accepted usage by the
citation of Lapworth (1876) as author. Bulletin of Zoological Nomenclature , 46, 33-34.

MU, a. T., li ji-jin, GE mei-yu and yin ji-xiang 1962. Graptolite fauna from the Chilien Mountain. Geology of
Cliilienslian , 4, W— 1 68.

NANJING INSTITUTE OF GEOLOGY AND PALAEONTOLOGY, ACADEMIA SINICA (NIGP) 1974. Handbook of Stratigraphy
and palaeontology of Southwest China. Science Press, Beijing, 454 pp., 202 pis.
nicholson, h. a. 1868. On the graptolites of the Coniston Flags; with notes on British species of the genus
Graptolites. Quarterly Journal of the Geological Society of London, 24, 521-545, pis 19-20.
obut, a. m., sobolevskaya, r. f. and bondarev, r. f. 1965. Graptolitit silura Taymira. Akademii Nauk SSSR ,
Sibirskoe Ot dele me, Institut Geologii Geofiziki. Ministerstvo Geologii SSSR, Nauchno-issledovatelskie Institut
Geologii Arktiki, 1-120, pis 1-19. [In Russian].

perner, j. 1897. Etudes sur les Graptolites de Boheme. Illieme Partie. Monographic des Graptolites de VEtage
E. Section a, 1-25, pis 9-13, Prague.

rickards, R. b., hutt, j. e. and berry, w. b. n. 1977. Evolution of the Silurian and Devonian Graptoloids.

Bulletin of the British Museum (Natural History), (Geology), 28, 1-120, pis 1-6.
strachan, i. 1971 . A synoptic supplement to ‘A monograph of British graptolites by Miss G. L. Elies and Miss
E. M. R. Wood’. Monograph of the Palaeontographical Society, 130 pp.
tornquist, s. L. 1881. Om nagra graptolitarter fran Dalarne. Geologiska Foreningens i Stockholm
Forhandlingar, 5, 434-445, pi. 17.

- 1892. Undersokningar ofver Siljansomradets graptoliter. II. Lunds Universitets Arsskrifter , 28, 1-47, pis
1-3.

wang xiao-feng 1978. A restudy of the graptolites and the age of the Wentoushan Formation from Liantan,
Kwangtung. Acta Geologica Sinica, 4, 303-17, pis 1-4.

jin yu-qin, wu zhao-tong, fu han-yin, li zho-chong and ma guo-gan 1977. A palaeontological atlas

of central-south China. Geological Publishing House, Beijing, 470 pp., 116 pis.
yin, t. h. 1937. Brief description of the Ordovician and Silurian fossils from Shihtien. Bulletin of the Geological
Society of China, 16, 281-302, pis 1-2.

DAVID K. LOYDELL
Institute of Earth Studies

Typescript received 15 October 1989 University College of Wales, Aberystwyth

Revised typescript received 19 February 1990 Dyfed, SY23 3DB, UK

-vT

f

EDIACARAN FOSSILS FROM THE SEKWI BROOK
AREA, MACKENZIE MOUNTAINS, NORTHWESTERN

CANADA

bv GUY M. NARBONNE and JAMES D. AITKEN

Abstract. Ediacaran body fossils and trace fossils occur sporadically throughout more than one kilometre of
strata in the upper part of the Windermere Supergroup in the western Northwest Territories of Canada. The
oldest fossiliferous unit, the Sheepbed Formation, contains body fossils (Beltanella gilesi, Charniodiscus ? sp..
Cyclomedusa plana , Cyclomedusa sp., Eoporpita sp., Kullingia sp. and Medusinites asteroides ) and very rare
trace fossils (Planolites montanus). The Blueflower Formation contains rare body fossils ( Charniodiscusl sp.,
Ediacaria sp., Inkrylovia sp., Pteridinium sp. and Sekwia excentrica ) and abundant trace fossils (Aulichnites
ichnosp., Helminthoida ichnosp., Helminthoidichnites tenuis , Helminthopsis aheli. Helminthopsis irregularis ,
Helminthopsis ? ichnosp., Lockeia ichnosp., Neonereites ichnosp., Palaeophycus tubularis , Planolites montanus ,
Torrowangea rosei and a knotted burrow). Overlying dolostones and thick-bedded sandstones of the Risky
Formation contain only simple trace fossils (Palaeophycus tubularis , Planolites montanus). The Sekwi Brook
biota lived in a deep-water, basin slope setting below storm wave-base. Most of the body fossils probably
represent benthic polypoid and frond-like organisms. The body fossil assemblage is broadly similar to that
described from correlative shallow shelf deposits in the Wernecke Mountains, Flinders Ranges and Russian
Platform. The trace fossil assemblage is dominated by simple and irregularly meandering burrows, but contains
some patterned meanders typical of the Nereites ichnofacies. The occurrence of this relatively diverse, mainly
deep-water assemblage of benthic body fossils and infaunal burrows indicates that the initial radiation of
metazoans extended to the deep sea, and that some aspects of Phanerozoic-style marine ecosystems were
initiated during the earliest stages of metazoan evolution.

The Ediacaran fauna represents the oldest diverse assemblage of megascopic animals and trace
fossils in the world, and consequently the rootstock from which all subsequent metazoans evolved.
Ediacaran animal remains are predominantly discoid and pennate fossils (commonly interpreted as
primitive cnidarians) with fewer arthropods, echinoderms and problematic taxa. Trace fossils are
chiefly simple subhorizontal burrows with rarer meandering forms. Recent summaries of this fauna
can be found in Glaessner (1984) and Sokolov and Ivanovskiy (1985).

The Ediacaran assemblage is of latest Precambrian age, and consistently occurs between the
highest Proterozoic tillites and the lowest Cambrian-type shelly fossils and trace fossils (Jenkins
1981 ; Cloud and Glaessner 1982; Sokolov and Fedonkin 1984). It was first reported from Namibia
and South Australia, but subsequently has been documented from more than twenty-five localities
around the world (Glaessner 1984, fig. 1/8; Hofmann 1987, fig. 13). The first description of
Ediacaran fossils from Laurentia (= ancestral North America) was by Hofmann (1981), who
collected body fossils and trace fossils from the Sekwi Brook area of the Mackenzie Mountains
(Text-fig. 1) during a brief visit of the I.U.G.S. Precambrian-Cambrian Boundary Working Group
(Fritz 1980). Hofmann reported the body fossils Inkrylovia sp. and Sekwia excentrica , the trace
fossils Torrowangea sp. and Gordia sp. and several problematica and dubiofossils. Aitken (1989u)
subsequently figured a specimen of the megafossil Pteridinium sp. from the same locality, Baudet
et al. (1989) studied the acritarchs from this section, and Aitken and Narbonne (1989) described
cryptalgal structures from a nearby section. Ediacaran body fossils and ichnofossils have also been
described from age-equivalent shallow shelf deposits in the Wernecke Mountains (Text-fig. 1 ),
250 km west north-west of Sekwi Brook (Hofmann et al. 1983; Narbonne and Hofmann 1987).

IPalaeontology, Vol. 33, Part 4, 1990, pp. 945-980, 4 pls.|

© The Palaeontological Association

946

PALAEONTOLOGY, VOLUME 33

During June 1988, the present authors collected Ediacaran fossils from Hofmann’s ( 1981 ) locality
and two others in the Sekwi Brook area of the Mackenzie Mountains (Text-fig. 1) as part of a
regional study of the upper Precambrian strata. These collections considerably extend the known
diversity and stratigraphic range of Ediacaran fossils in the Mackenzie Mountains. Previous reports
of Ediacaran fossils have been mainly from shallow shelf settings (e.g. Flinders Ranges, Russian
Platform, Wernecke Mountains) with fewer reports from deep slope/fan environments (e.g.
Newfoundland, Charnwood Forest). The occurrence of a relatively diverse assemblage of body

fossiliferous Ediacaran sections described previously are in the June Lake Anticline (J) and in the Wernecke

Mountains (W).

NARBONNE AND AITKEN: CANADIAN EDIACARAN FAUNA

947

text-fig. 2. Regional stratigraphic setting of the Ediacaran biota from Sekwi Brook. E = Ediacaran fossils;
T = Chuaria-Tawuia assemblage; MMSG = Mackenzie Mountains Supergroup.

fossils and trace fossils in basin slope deposits at Sekwi Brook permits more detailed evaluations of
the palaeoenvironmental affinities of Ediacaran organisms. In this paper, description of the fossils
and discussion of their significance are by Narbonne; information on the stratigraphy and
depositional environments of the strata is by Aitken.

STRATIGRAPHIC AND PALAEOENVIRONMENTAL SETTING

The terminal Proterozoic formations of the Mackenzie and Wernecke Mountains (Text-figs 1 and 2) are
assigned to the upper part of the Windermere Supergroup. The base of the Windermere is marked by the onset
of rifting at 770-780 Ma (Armstrong et al. 1982; Eisbacher 1981, 1985; Jefferson and Ruelle 1986). The top
is the base of Cambrian strata.

The terminal Windermere Formations make up four, kilometre-scale, shallowing-upward, carbonate-
capped Grand Cycles (sensu Aitken 1966, 1978), as follows (Aitken 1989a):

4. Ingta Formation (230 m)

3. Blueflower (550 m) and Risky (150 m) Formations
2. Sheepbed (900 m) and Gametrail (300 m) Formations
1. Twitya (900 m) and Keele (500 m) Formations

These cyclical deposits post-date, and pass without noticeable effect, across the sites of the rift-depressions in
which the older Windermere deposits, the Coates Lake and Rapitan groups, were deposited (Text-fig. 2).
Evidence for a Rapitan glaciation is well known (Young 1976; Yeo 1981 ; Eisbacher 1981, 1985). The deposits
of a second Late Proterozoic glaciation, which directly underlie the Sheepbed Formation, have thus far been
described only in oral presentations with published abstracts (e.g. Aitken 1987, 19896).

The middle two of the upper Windermere Grand Cycles (Sheepbed-Gametrail, Blueflower-Risky) contain
abundant Ediacaran megafossils (this study). The uppermost Windermere Grand Cycle, the Ingta Formation,
yields abundant, well-preserved trace fossils, but lacks both diagnostically Cambrian traces and body
impressions of Ediacaran type. We interpret it as latest Precambrian. The Ingta, and in other places all of the
other terminal Proterozoic formations, are unconformably overlain by correlatives of the Backbone Ranges
and Vampire Formations bearing Early Cambrian trace fossils and small shelly fossils (Fritz et al. 1983;
Narbonne et al. 1985; Nowlan et al. 1985). The position, in basinward areas such as Sekwi Brook, of the major
unconformity beneath the typical Backbone Ranges Formation is in dispute (reviewed by Aitken 1989a).

Sheepbed Formation

The Sheepbed Formation of the Sekwi Brook area is much coarser, on average, than in the ‘typical’ belt of
exposures above the Plateau Thrust. The lower part, below the lowest fossils found, contains rare packets of
thick and very thick beds of quartzose sandstone that are normally graded and in part coarse-grained (Bouma

948

PALAEONTOLOGY, VOLUME 33

Ta). Rare packets of ribbon-bedded lime mudstone are also present. Because of their association with and
involvement in mass-transport deposits, and normal grading seen in minor beds with a content of coarser
allochems, they are interpreted as turbidites. The upper (fossiliferous) part is dominated by laminated to finely
laminated, dolomitic siltstone and very fine-grained sandstone, with partings and packets of dark grey to brown,
silty shale and mudstone. Sole marks are present, but rare. The thin sandstones are Bouma Tb, Tbc, and Tc
units. The content of dolormtized and undolomitized, ribbon-bedded lime mudstone (turbidites), monomict,
matrix-supported carbonate breccias (debris flows), and penecontemporaneous folds and slide masses is highly
variable along strike. The thickness in the Sekwi Brook area is not determinable, but is almost certainly greater
than the maximum of 800 m in the ‘typical’ outcrop belt.

Interpretation of the depositional environment begins with the observation that the ‘typical’ Sheepbed to
the north-east contains very little sand-grade or coarser siliciclastic material. The Gametrail Formation
(carbonates, there partly in a shallow-water facies) that intertongues with and overlies the upper Sheepbed, also
lacks siliciclastic detritus. The ‘typical’ Sheepbed overlies the deeply drowned Keele platform, whereas Sekwi
Brook lies outboard of the Keele platform margin, above a previously-existing bathymetric basin or trough.
It is precisely at Sekwi Brook that coarse-grained, thick-bedded, turbiditic sandstones occur in the lower
Sheepbed. These sandstones appear to be deep-sea fan-channel deposits (Mutti and Normark 1987;
Shanmugam and Moiola 1988), transported axially along the inherited off-platform trough. The upper part of
the formation is readily interpreted as a complex of slope deposits that accumulated below storm wave-base.
This is consistent with the character of the overlying Gametrail carbonates at Sekwi Brook, which are of
ribbon-bedded facies, interrupted only by monomict, matrix-supported, debris-flow breccias and slide masses.
Thus, in the Sekwi Brook area, southwestward-prograding slope deposits (upper Sheepbed, Gametrail) appear
to have buried axially prograding deep-sea fan deposits (lower Sheepbed).

Blueflower Formation

The Blueflower Formation is not preserved above the Sheepbed and Gametrail formations of the ‘typical’
outcrop belt (Text-fig. 2). At Sekwi Brook it is 550 m thick, and consists of dark-coloured, deep-water shale
and mudstone (59%), turbiditic sandstones (15%), and limestones of turbiditic and grain-flow origin
(26%). It is the lower part of a shoaling-upward cycle that terminates with the peritidal dolomites of the
overlying Risky Formation. Other upward trends within the Blueflower are: decrease and virtual disappearance
of carbonate beds; increase in the proportion of sandstone; and the appearance, near the top, of non-turbiditic
sandstones with low-angle cross-stratification.

Sandstones in the lower three-quarters of the formation are predominantly thin-bedded, fine-grained
turbidites (Bouma Tb, Tc, and Tbc), with a few thicker and generally coarser beds that are Bouma Ta and Tab
units, and structureless beds interpreted as liquefied-flow deposits (Lowe 1982). Flame structures are common,
flute and groove casts less so. Thin sandstone beds are persistent across the typical widths of the gully outcrops
(50 m), but bundles of such beds are markedly lenticular over distances of 200-400 m.

Limestones occur in packets of two types: ribbon-bedded lime mudstone, in part displaying flame structures
and normal supply-grading of a sparse content of quartz sand grains; and subordinate, markedly lenticular
packets of medium and thick beds of intraclast grainstone and wackestone with a variable content of quartz
sand grains. Most of the latter display normal grading, and are Bouma Tab and Tb units. The packets of
ribbon-bedded limestone are distinctly lenticular over distances of less than one kilometre.

Evidence of deposition on a slope is common. It includes penecontemporaneous folds, detached fold-noses
in dismembered sandstone beds, scoop-shaped slide surfaces, contorted and mixed mudstone units with
deformed pods of sandstone, matrix-supported debris-flow breccias, at least two slide masses or debris-flow
paraconglomerates with boulders of shallow-water carbonate rocks, rare exotic blocks and boulders, and, high
in the type section, an exotic raft of stromatolitic dolomite.

The overall aspect of the lower three-quarters of the Blueflower at Sekwi Brook is that of a slightly
channelled slope below storm wave-base. The content of thin-bedded turbidites, with thickening-upward and
less common thinning-upward sequences, might suggest the middle to lower parts of a deep-sea fan. Indeed,
a lens of thick-bedded, massive sandstone cropping out less than one kilometre north of the type section
resembles a fan-channel deposit. The abundant evidence of slope deposition is, however, inconsistent with a
fan interpretation overall. Because it characterizes the middle and upper parts of the formation, it probably
records a depositional slope prograding above a fan-complex.

Towards the top of the Blueflower Formation, carbonate rocks reappear as blocks, boulders and rafts of
limestone and dolomite of shallow-water origin. These document the progradational approach of a re-
established carbonate platform, the Risky Formation, which appears at a gradational contact.

NARBONNE AND AITKEN: CAN ADI AN EDIACARAN FAUNA

949

Summary

Thus it is apparent that most of the Ediacaran fossils at Sekwi Brook occur in strata deposited in a deep-water,
basin slope setting considerably below storm wave-base. This is in marked contrast with the shallow-shelf
assemblages in the nearby Wernecke Mountains (Narbonne and Hofmann 1987). The occurrence of
assemblages of shallow- and deep-water Ediacaran fossils in localities only 250 km apart, within a single
depositional and structural province, permits analysis of their bathymetric affinities, without the
palaeogeographic uncertainties inherent in comparisons between distant areas.

FOSSIL OCCURRENCE AND PRESERVATION

Fossils were collected from three sections (Text-fig. 1) in the Sekwi Brook structural panel (sensu Aitken
1989a). The stratigraphy and distribution of fossils in the two main sections, Sekwi Brook North and Sekwi
Brook South, is shown in Text-figures 3 and 4. Sekwi Brook North contains the type sections of the Gametrail,
Blueflower and Risky Formations (Aitken 1989a); Sekwi Brook South was one of the sections investigated
during the 1979 visit of the PreCambrian-Cambrian Boundary Working Group (Fritz 1980), and is the section
where Hofmann (1981) described the first Ediacaran fossils discovered in western North America. In addition
to these two main sections, trace fossils were collected from the lower half of the Blueflower Formation (locality

29) , and trace fossils and body fossils were collected from the upper half of the Blueflower Formation (locality

30) during reconnaissance investigation of the Majesty Pb-Zn-Ag Property (Text-figs 1 and 4).

Some of the collections were made from outcrop, but most were from very local float (frost-heaved but
effectively untransported slabs). The only exception is the figured specimen of Ediacaria sp., which was in float
that had slid a considerable distance downhill, but which was tied into the section on the basis of the unique
lithology and the presence of a fragmentary specimen of the same taxon in the same lithology at locality 27.
Float that could not be tied down definitely stratigraphically was not included in the study material. The type
of collection at each locality is shown in Text-figure 3.

Ediacaran body fossils occur predominantly as raised features on the bottoms of beds (positive hyporelief)
and very rarely as depressed features on soles (negative hyporelief). Trace fossils also exhibit the same modes
of preservation, or occur as raised or depressed features on the tops of beds (positive/negative epirelief).

DISCUSSION

Body fossils

Previous reports of Ediacaran body fossils from the Sekwi Brook area by Hofmann (1981) and
Aitken (1989a) were from the lower and middle parts of what is presently regarded as the Blueflower
Formation. Our discoveries of specimens of Charniodiscus and Ediacaria near the top of this
formation and of a relatively diverse assemblage of Ediacaran body fossils approximately 225
metres below the top of the Sheepbed Formation considerably extends the known diversity and
stratigraphic range of these fossils in the Mackenzie Mountains. Ediacaran body fossils have been
reported by Narbonne and Hofmann (1987) from the ‘Goz Siltstone' in the Wernecke Mountains,
a unit tentatively correlated with the Sheepbed Formation (Narbonne et al. 1985), but the present
paper represents the first confirmed report of an Ediacaran biota from the Sheepbed Formation.
This is significant, as this unit is geographically widespread (Eisbacher 1981) and immediately
overlies the highest Proterozoic glacial deposits in northwestern Canada (Aitken 1987, 1989/)).

Ediacaran body fossils occur sporadically throughout approximately one kilometre of strata
(upper Sheepbed Formation to the top of the Blueflower Formation) in the Sekwi Brook area of
the Mackenzie Mountains. A total of ten genera comprising eleven species are now known from
these strata (Text-fig. 4). The assemblage is broadly similar to that reported by Narbonne and
Hofmann (1987) from equivalent strata in the Wernecke Mountains 250 kilometres to the west
(Text-figs 5 and 6). Six genera ( Beltanella , Charniodiscus f Cyclomedusa , Ediacaria , Kullingial, and
Medusinites) are common to both areas. Six genera (Beltanelliformis, Nadalina, Rugoconites‘1,
Spriggia , Tirasiana , and Vendotaenial ) are reported only from the Wernecke Mountains, and four
genera ( Eoporpita , Inkrylovia , Pteridinium , and Sekwia ) are thus far known only from Sekwi Brook.
Of the geographically restricted taxa, only Beltanelliformis is common, and the absence of the other

□lira

950

PALAEONTOLOGY, VOLUME 33

LEGEND

ISJ

<74. m)

13

THICK-BEDOED CARBONATE

RIBBON-BEDDED CARBONATE

SHALE, MUDSTONE

SANDSTONE. THIN-BEDDED
TURBIDITES

SANDSTONE. MEDIUM- AND
THICK-BEDDED

BOULDER PARACONGLOMERATE

DEBRIS FLOW. BRECCIA

COVERED INTERVAL

EXOTIC BLOCK, RAFT

SLUMP FOLD

STROMATOLITE

NORMAL FAULT

COLLECTION FROM OUTCROP

COLLECTION FROM LOCAL FLOAT

SEKWI SEKWI

BROOK BROOK

NORTH SOUTH

1250

1000

500

250

text-fig. 3. Stratigraphic sections and sample localities. B.R. FM. = Backbone Ranges Formation.

NARBONNE AND AITKEN: CANADIAN EDIACARAN FAUNA

951

TAXA

STRATIGRAPHY

Sheep-

bed

Blue-

llower

Risky

LOCALITIES

Sekwi Brook North
1 2 34 5 6 789 10 11 1213 14 15 16

Sekwi Brook South [Majesty
17 18 19 20 21 22 23 24 25 26 27 28 | 29 30

BODY FOSSILS

Beltanella gilesi
Charniodiscus? sp.
Cyclomedusa plana
Cyclomedusa sp.
Ediacaria sp
Eoporpita sp
Inkrylovia sp
Kullingia 7 sp.
Medusinites asteroides
Pteridinium sp
Sekwia excentrica

Plumose problematicum

ICHNOFOSSIIS

Aulichnites ichnosp
Helminthoida ichnosp.
Helmintholdichmtes tenuis
Helminthopsis abeli
Helminthopsis irregularis
Helminthopsis ? ichnosp
Lockeia ichnosp
Neonereites 7 ichnosp
Palaeophycus tubularis
Planolites montanus
Torrowangea rosei
Knotted burrow

text-fig. 4. Synopsis of Ediacaran macrobiota from the Sekwi Brook region.

taxa may simply reflect the relatively short periods of time devoted to collecting in these remote
regions. Thus, it would appear that the faunas from the Wernecke and Mackenzie Mountains can
be regarded as part of a single ‘Windermere fauna’ that characterizes the Ediacaran of
northwestern Canada, and may also include the low diversity fauna described by Hofmann et al.
(1985) from the southern Canadian Rocky Mountains.

A tentative correlation of the formations and body fossils of the Wernecke and Mackenzie
Mountains is shown in Text-figure 5. With the exception of the middle siliciclastic unit of the Risky
Formation, which at Sekwi Brook consists entirely of thick-bedded dolomitic sandstone and thus
represents an unfavourable facies for preservation, all units that contain Ediacaran fossils in the
Wernecke Mountains have fossiliferous correlatives in the Mackenzie Mountains. However, fossil
assemblages in correlative units are not particularly similar, and the greatest faunal similarity is
between the youngest fossiliferous unit in the Wernecke Mountains and the oldest fossiliferous unit
in the Mackenzie Mountains! This suggests that most of the apparent geographical and

952

PALAEONTOLOGY, VOLUME 33

CORN CREEK AREA,
WERNECKE MINS

SEKWI BROOK AREA,
MACKENZIE MINS

I

Lithology

ribbon-bedded carbonate

thick-bedded carbonate

shale/mudstone

shale, with minor turbidite/storm sandstone beds

medium-bedded sandstone

conglomeratic channel

text-fig. 5. Correlation of Ediacaran strata and megafossils in the eastern Wernecke Mountains and in the
Sekwi Brook area of the Mackenzie Mountains. Lithological correlations are from Aitken (1989u, fig. 7);
faunal lists are from Narbonne and Hofmann (1987) and the present study.

NARBONNE AND AITKEN: CANADIAN EDIACARAN FAUNA

953

stratigraphic variation in the Windermere fauna reflects palaeoenvironmental and preservational
factors, and an original patchy distribution, rather than palaeogeography and evolution. This view
is supported by the great variation in the faunal content of equivalent strata in adjacent sections (cf.
Narbonne and Hofmann 1987, text-figs 2 and 3; this study, Text-figs 3 and 4), even where the
sections are within sight of each other. For this reason, we believe that any attempt to subdivide the
Windermere fauna into biostratigraphic zones would be premature at the present time.

Global correlation of the Windermere body fossils is summarized in Text-figure 6. Close
similarities are evident between the faunas of the Wernecke and Mackenzie Mountains, the
Ediacara Member of the Rawnsley Quartzite in South Australia, and the Redkino ‘Series’ of the
Vendian in northern Europe and Podolia. The Windermere biota is considerably less diverse than
those reported from the other two regions, and apparently lacks segmented forms such as Spriggina
and Dickinsonia , which are present in South Australia and on the Russian Platform. This may relate
to the relatively early stages of investigation in northwestern Canada, a palaeogeographic
separation between Laurentia and the latter-mentioned two areas (see McMenamin and
McMenamin 1990), or some other palaeoecological or stratigraphic factor. Nevertheless, the high
proportion of Windermere taxa that are also known from the type Ediacaran and the type Vendian
(Text-fig. 6), emphasizes the considerable correlation potential of the Ediacaran fauna.

Baudet et al. (1989) concluded that the acritarchs of the Sheepbed and Blueflower Formations
best correlate with the Kotlin ‘Series’ of the Vendian on the Russian Platform. This contradicts the
above correlation, as the macrofossils of these formations are much more similar to the fossil
assemblage of the Redkino ‘Series’ (see above) than they are to those of the overlying Kotlin
‘Series’, which lacks identifiable megafossils (Sokolov and Fedonkin 1984). The stratigraphic
setting of these units, immediately overlying a diamictite unit and two or more Grand Cycles
(400-1200 m) below the lowest Cambrian-type fossils (Phycodes pedum Zone), is also quite similar
to that of the Redkino ‘Series’ and supports the correlation based on macrofossils. On the basis of
published descriptions of the body fossils, ichnofossils, small shelly fossils, and the overall
stratigraphic sequence, Narbonne et al. (1987, fig. 8) concluded that the probable equivalents of the
Kotlin ‘Series’ in the Mackenzie Mountains are the upper part of the Risky Formation and the
overlying Ingta Formation. Baudet et al. (1989) were unable to date these latter units, as they did
not yield diagnostic acritarchs in their study. Thus, the majority of the evidence suggests that the
Sheepbed, Gametrail, Blueflower and lower Risky Formations are equivalent to the Redkino
‘Series’ of the Vendian, and that the upper Risky and Ingta formations are equivalent to the Kotlin
‘Series’.

Until recently, it was generally believed that the Ediacaran fauna was dominated by discoidal
pelagic medusoids (e.g. Sprigg 1947, 1949; Glaessner and Wade 1966). Fedonkin (1985«) and
Jenkins (1988) have reinterpreted many of the forms as benthic ‘polyps’, and this interpretation also
seems to be appropriate to the Windermere biota. With the possible exception of Kullingial , the
fauna is entirely dominated by benthic organisms, both in terms of numbers and diversity (see
Systematic Palaeontology below). The shapes of many Ediacaran organisms and their common
occurrence in shallow-water facies has led some authors to suggest that they may have been partly
or completely photoautotrophic (e.g. McMenamin 1986), but this is not consistent with their
occurrence in muddy slope deposits considerably below storm wave-base in the Sekwi Brook area.

The Sekwi Brook fauna represents the first report of an Ediacaran assemblage in a basin slope
setting. The Sekwi Brook fauna is markedly different from those reported from the deep slope/fan
‘Avalon assemblage’ of Newfoundland (e.g. Anderson and Conway Morris 1982), Charnwood
Forest (e.g. Ford 1980), and the Carolina Slate Belt (Gibson et al. 1984), with only two or three of
the reputedly more than twenty forms from these settings also present at Sekwi Brook. It is presently
uncertain how much of this difference is palaeoenvironmental, and how much reflects
palaeogeography. A closer similarity exists between the basin slope faunas from Sekwi Brook and
those reported from shallow shelf deposits in the Wernecke Mountains, Flinders Ranges, Podolia,
northern Europe and elsewhere (Text-fig. 6). This implies that some elements of the Ediacaran
fauna extended over a considerable depth range.

954

PALAEONTOLOGY. VOLUME 33

TAXA

NORTHWESTERN

CANADA

SHALLOW

SHELF

Wernecke

Mtns.

BASIN

SLOPE

Sekwi

Brook

GLOBAL COMPARISONS

SHALLOW

SHELF

0) ®

£ &

EDIACARAN

I!

_ 03
03 —
Zz 03
C Zz
Q) (/)

° Z

DEEP

SLOPE/FAN

OTHERS

2 "2
c -

5 g

O LU

to

o

< n
d

V)

c n
O

LL

>

Q

O

m

Beltanella

Beltanelliformis

Charniodiscus?

Cyclomedusa

Edlacaria

Eoporpita

Inkrylovia

Kullingia ?

Medusinites

Nadalina

Pteridinium

Rugoconites?

Sekwi a

Spriggla

Tirasiana

Vendotaenia?

□

O

□

□

□

o

□

□

CO

d

00

00

o

I

o

Aulichnites

Helminthoida

Helminthoidichnites

Helminthopsis

Lockeia

Neonereites ?

Palaeophycus

Planolites

Torrowangea

□

□

□

O

text-fig. 6. Regional and global correlation of the Windermere biota. Localities with less than two
Windermere-type taxa are omitted. Abundance symbols for northwestern Canada: narrow bar = rare (present
in < 10% of fossiliferous units); intermediate bar = common (present in 10-50% of fossiliferous units); broad
bar = abundant (present in > 50% of fossiliferous units). For global comparisons, same (square) or similar

(circle) forms indicated.

Ichnofossils

A relatively diverse assemblage of trace fossils is present in the Ediacaran strata of the Sekwi Brook
area (Text-figs 4 and 6). Previously, the oldest known trace fossils in northwestern Canada were

NARBONNE AND AITKEN: CANADIAN EDIACARAN FAUNA

955

from the Blueflower Formation (Hofmann 1981; Aitken 1989a) and its stratigraphic equivalent in
the Wernecke Mountains (Narbonne and Hofmann 1987). The occurrence of specimens of
Planolites montanus in the upper part of the Sheepbed Formation extends the stratigraphic range
of trace fossils downwards by approximately 500 metres. The Blueflower Formation contains a
more diverse assemblage of trace fossils, comprising at least twelve ichnospecies referable to nine
ichnogenera. All taxa known from the Ediacaran of Sekwi Brook also range the Phanerozoic.
Forms restricted to the Ediacaran (e.g. Harlaniella, Palaeopascichnus , Nenoxites , etc.) have not yet
been documented from Sekwi Brook.

Seilacher (1964, 1967) first suggested that trace fossils show a consistent ecological zonation
throughout the Phanerozoic which he subdivided into several 'universal ichnofacies’. Although
numerous studies have supported and refined Seilacher’s model (see Ekdale et al. 1984 and
references therein), it remains uncertain when and how ichnofacies first developed. Most of the
ichnotaxa at Sekwi Brook also occur in shallower water deposits in the Ediacaran of the Wernecke
Mountains, eastern and northern Europe, and Australia, and in deep slope/fan deposits in the
Carolina Slate Belt (Text-fig. 6). This supports the view of Crimes and Anderson (1985) that upper
Precambrian-Lower Cambrian trace fossil assemblages show a lower degree of palaeo-
environmental differentiation than do those in post-Cambrian strata.

Basin slope deposits at Sekwi Brook also contain abundant simple burrows (e.g. Planolites and
Palaeophycus ), and non-directed meanders (especially Torrowangea and Helminthoidichnites)\
patterned meanders occur rarely (Helminthoida) to commonly ( Helminthopsis ). The presence of an
assemblage dominated by meandering forms and apparently lacking vertical burrows and
arthropod traces is reminiscent of the Nereites ichnofacies of Seilacher (1964, 1967). The Nereites
ichnofacies typifies turbiditic slope deposits of Phanerozoic age but had not previously been
reported from the Precambrian. The comparatively low diversity, the apparent absence of
graphoglyptids, and the dominance of unpatterned and simple patterned forms among the
meandering burrows of the Blueflower (basin slope) assemblage implies that this Precambrian
assemblage represents a relatively simple version of the ichnofacies, and is consistent with
Seilacher’s (1974, 1977) observation that the Nereites ichnofacies exhibits a step-wise increase in all
three of these features throughout the Phanerozoic. Previous evolutionary models for trace fossils
(e.g. Crimes 1974; Seilacher 1974, 1977; Frey and Seilacher 1980) have generally concluded that
deep-water slopes of Ediacaran/Vendian age were all but completely devoid of infaunal animal life,
but the new evidence from the Carolina Slate Belt (Gibson 1989), the Avalon Zone of
Newfoundland (pers. comm. M. M. Anderson 1989) and the Sekwi Brook area (this study) indicates
that the initial radiation of metazoans also extended to the sediments of the deep-sea.

Previous workers (e.g. Crimes and Anderson 1985) suggested that complexly meandering
burrowers first evolved in shallow-water settings in the early Cambrian, and later migrated into
deeper-water settings. It now appears more likely that meandering burrowers originated in the
Precambrian, where they occurred in both shallow marine (Glaessner 1969; Fedonkin 1985c) and
deep marine (this study) settings. The explosive diversification of shallow-water burrowers in the
early Cambrian, including the first appearance of abundant complexly branching, spreiten-bearing,
and deep vertical burrows (cf. Alpert 1977; Crimes 1987; Narbonne et al. 1987), gradually
eliminated meandering burrowers, especially forms producing complex meanders, from shallow-
water settings. However, a low diversity assemblage of meandering burrows persisted in deeper-
water slope settings, and gradually increased in diversity and burrow complexity throughout the
remainder of the Phanerozoic.

SYSTEMATIC PALAEONTOLOGY

Although genus- and species-level identification of Ediacaran organisms are relatively stable, no
classification of family and higher taxonomic levels has yet been accepted. Conway Morris (1985)
has argued that attempts to relate Ediacaran organisms to extant phyla and classes may obscure the
relationships among these early metazoans, and even the assignment of some Ediacaran fossils to

956

PALAEONTOLOGY, VOLUME 33

the Metazoa has been questioned by Seilacher (1984, 1989). Rival classification schemes are
presented in Pflug (1972), Glaessner (1979), Seilacher (1989), Fedonkin (19856), Gureev (1987) and
other papers. A full evaluation of these proposals is beyond the scope of this study, and the Sekwi
Brook fossils are simply classified as 'body fossils’ or ‘ichnofossils’ (= trace fossils). Dubiofossils
have been described by Hofmann (1981), and are not included in the present study. Acritarchs from
these strata have also been discussed previously (Baudet et al. 1989).

In the following section, synonymies include only Precambrian forms; the occurrence (if any) of
Cambrian or younger forms is discussed under 'Remarks’. All type and figured specimens have
been deposited in the National Type Collection of Invertebrates and Plants, Geological Survey of
Canada (GSC) in Ottawa, Canada.

BODY FOSSILS
Genus beltanella Sprigg, 1947

Type species. Beltanella gilesi Sprigg, 1947.

Beltanella gilesi
Plate 1, fig. 1

For synonymy up to 1987, see Narbonne and Hofmann (1987).

1987 Beltanella gilesi Narbonne and Hofmann, pp. 653-654, pi. 73, fig. 6.

1987 Planomedusites grandis Gureev, pp. 40-42, fig. 16.

Description. Single incomplete disc preserved in convex hyporelief. Smooth disc 40 mm in diameter with a
prominent central tubercle 4-8 mm in diameter; disc surrounded by a flange 5-11 mm wide with irregularly
spaced marginal indentations. Poorly preserved, tapering stalk 20 mm wide at its base extending 40 mm
beyond the outer edge of the flange.

Remarks. The specimen is strikingly similar to one illustrated by Narbonne and Hofmann (1987, pi.
73, fig. 6) from the Wernecke Mountains of northwestern Canada. Beltanella has traditionally been
interpreted as a pelagic medusoid (e.g. Sprigg 1947; Harrington and Moore 1956), but Jenkins
(1988) recently reinterpreted it as a benthic polyp attached to the sea-bottom by a short stalk. The
specimen of Beltanella from Sekwi Brook exhibits a poorly preserved stalk which extends outward
from the outer margin of the disc, thereby implying that it was above rather than below the disc.
As this structure is known from a single specimen, it is uncertain whether or not it is accidental.

Genus charniodiscus Ford, 1958
Type species. Charniodiscus concentricus Ford, 1958

EXPLANATION OF PLATE 1

Fig. 1. Beltanella gilesi Sprigg, hyporelief. Locality 17, Sekwi Brook South. GSC 95895, x I.

Figs 2 and 6. Cyclomedusa sp., hyporelief. 2, locality 3, Sekwi Brook North, GSC 95896, x 1. 6, three
specimens that interfered with each other during growth, locality 17, Sekwi Brook South, GSC 95898-95900
(clockwise from top right), x 1.

Fig. 3. Plumose problematicum, epirelief. Locality 3, Sekwi Brook North, GSC 95897, x 1-5.

Fig. 4. Cyclomedusa plana Glaessner and Wade, hyporelief. Locality 3, Sekwi Brook North, GSC 95902, x F5.
Fig. 5. Ediacaria sp., hyporelief. Float, probably from locality 27. Sekwi Brook South, GSC 95903, x0-5.
Fig. 7 Charniodiscus ? sp., hyporelief. Locality 30, Majesty Section, GSC 95904, x 1.

PLATE 1

NARBONNE and AITKEN, Ediacaran fauna

958

PALAEONTOLOGY, VOLUME 33

Chamiodiscus ? sp.

Plate 1, fig. 7

Description. Five circular discs preserved in positive hyporelief (0-5— 2-5 mm relief). Central tubercle 3-8 mm in
diameter surrounded by 3-5 concentric rugae. Disc diameter 26-46 mm (mean = 37-2 mm). Gently tapering
stalk, 12-21 mm wide at its base, extending up to 85 mm from the outer margin of the disc.

Remarks. Ford (1958) originally described Chamiodiscus on the basis of a single concentric disc with
an attached stalk, and later (1963) figured the complete specimen with a frond attached to the stalk.
Jenkins and Gehling (1978) and Glaessner (1979) subsequently regarded the frond as the most
diagnostic part of the fossil. The absence of an attached frond in any of the Sekwi Brook specimens
thus far collected precludes more specific identification. Two specimens of the frond of Chamiodiscus
cf. arboreus and numerous basal discs of Chamiodiscus ? were described from the Wernecke
Mountains by Narbonne and Hofmann (1987, fig. 5c).

Genus cyclomedusa Sprigg, 1947
Type species. Cyclomedusa davidi Sprigg, 1947.

Cyclomedusa plana Glaessner and Wade, 1966
Plate 1, fig. 4

For synonymy to 1987, see Narbonne and Hofmann (1987).

1987 Cyclomedusa plana Narbonne and Hofmann, pp. 656-658, pi. 73, fig. 3.

1987 Glaessneria plana Gureev, fig. 14.

1987 Glaessneria imperfecta Gureev, p. 40, fig. 15.

Description. Two bipartite discs preserved in convex hyporelief. Central cone (5 2 mm relief) or concentrically-
ringed disc (< 0-5 mm relief) 51-71 mm in diameter superimposed on a flat disc approximately 34 mm in
diameter. Outer edge of the fossil exhibiting a sharp annulus 0-8-F2 mm wide and 0-5 mm high. Surface of the
outer disc smooth, or marked by very fine radial grooves.

Remarks. The status of C. plana is controversial. Until recently, most authors have followed
Glaessner and Wade (1966), and regarded it as a species of Cyclomedusa. Sun (1986) questioned
whether C. plana should be included in Cyclomedusa , as it is poorly understood and exhibits some
significant differences with the type species. Gureev (1987) specifically removed C. plana from
Cyclomedusa , and regarded it as the type species of his new genus Glaessneria. Alternatively, Jenkins
(1988, 1989) has pointed out that Cyclomedusa and other discoid Ediacaran genera have been
oversplit, and that there is evidence of intergradation among some of the species of the Cyclomedusa
plexus and perhaps even among some related genera. Further systematic revision of Cyclomedusa
is obviously needed, but in the interim we follow Glaessner and Wade (1966) in regarding C. plana
as a valid species of Cyclomedusa.

Most figured specimens of C. plana , including material from the Wernecke Mountains (Narbonne
and Hofmann 1987) and from Sekwi Brook (e.g. GSC 95901), exhibit a series of concentric wrinkles
near the centre of the disc; this presumably represents a conical apex that was folded flat during
compaction. One specimen from Sekwi Brook (PI. 1, fig. 4) resembles two unfigured specimens from
South Australia (collected in 1988 and now in the collections of R. J. F. Jenkins) in exhibiting an
uncompressed cone at its apex. This implies that the cone was buried in relatively cohesive mud
while the flat outer ring rested on the sea floor.

Cyclomedusa plana is a cosmopolitan taxon (Wade 1972), not known from Australia, eastern
Europe and western North America (see synonymy in this paper and in Narbonne and Hofmann
1987).

NARBONNE AND AITKEN: CANADIAN EDIACARAN FAUNA

959

Cyclomedusa sp.

Plate 1, figs 2 and 6; Plate 2, figs 3 and 5

Description. Twenty-one measured specimens and numerous field identifications, all preserved in positive
hyporelief. Circular discs 13-4-58-5 mm (mean = 26-2 mm) in diameter with central tubercle, 11 7-7 mm;
tubercle surrounded by irregularly-spaced concentric markings. Very fine radial ridges locally present (e.g. PI.
1, fig. 2; PI. 2, fig. 5). Relief 0-3— 3-5 mm, with maximum relief in centre. Specimens deformed laterally where
in contact (PI. 1 . fig. 6).

Remarks. Specimens are mostly at the small end of the size range for Cyclomedusa , which has a
maximum diameter of more than 100 mm (Sun 1986). They closely resemble the specimens
illustrated by Wade (1972, pi. 41, fig. 1) as C. davidi Sprigg, and like them show evidence that they
were originally highly conical. Sun (1986) questioned whether these specimens should be included
in C. davidi as the radial striations typical of this species are not well developed in Wade’s specimens,
an objection that applies equally well to the specimens from Sekwi Brook and elsewhere in
northwestern Canada (Narbonne and Hofmann 1987). However, the presence of fine radial striae
on the bases of Cyclomedusa specimens has been attributed by Sun (1986) to composite molding of
features on the oral and aboral surfaces. Composite molding is prevalent in the Flinders Ranges but
is relatively rare in most other Ediacaran deposits, and it would seem unwise to base a generic
difference on a preservational feature. Further systematic revision of Cyclomedusa is obviously
needed, but in the interim we refer these specimens to Cyclomedusa sp.

The interpretation of discoid Ediacaran fossils such as Cyclomedusa is uncertain, with various
authors regarding them as impressions of pelagic medusae (e.g. Wade 1968; Sun 1986), benthic
polyps (e.g. Fedonkin 1985«; Jenkins 1988), sea-pen attachment structures (Jenkins 1989), or
burrows similar to those produced by modern actinians (Seilacher 1984). The absence of specimens
preserving the oral surface among any of the specimens from northwestern Canada is not consistent
with the orientations of medusae stranded at the strand line or in deeper water (cf. Wade 1968). This
implies that the aboral surface permanently rested on or in the substrate, and thus that the
Cyclomedusa organism was benthic. Closely-crowded specimens on the same bedding plane tend to
be relatively similar in size (e.g. Hofmann el at. 1983, fig. 2a; Narbonne and Hofmann 1987, text-
fig. 5a; this study, PI. 1, fig. 6), and may represent products of a single spatfall. The fact that closely-
crowded specimens deform but do not cross-cut each other, and evidence of increasing mutual
deformation during growth (PI. 1, fig. 6), support the interpretation of Cyclomedusa sp. as the
impression of a sessile benthic organism. Most likely, the prominent central tubercle represents the
point where the pelagic larva settled onto the bottom and dug a small pit for attachment to the sea
floor. The organism was firm-bodied, and apparently grew out concentrically in the shape of a cone.
With the possible exception of the central tubercle there is no evidence of active burrowing; however
passive sedimentation of mud around the bases of some of the cones may have resulted in the
relatively high relief of some specimens (e.g. PI. 1, figs 2 and 6). The organism was later buried by
a sandy turbidite or storm bed, and was cast by the subsequent collapse of the sand into the space
formerly occupied by the organism.

This model implies that Cyclomedusa sp. and perhaps other Ediacaran discs with a prominent
central tubercle, represent impressions of firm-bodied organisms that were sessile except during
their larval stages. The nature of the oral surface remains uncertain, with various authors showing
evidence for tentacles (Wade 1972, pi. 41, fig. 2) or stalks (Jenkins 1989, fig. 2e) in some specimens.
Elucidation of this is important, as Cyclomedusa is one of the most widely recognized Ediacaran
fossils.

Genus ediacaria Sprigg, 1947

Type species. Ediacaria flindersi Sprigg, 1947.

960

PALAEONTOLOGY, VOLUME 33

Ediacaria sp.

Plate 1, fig. 5

Description. One nearly complete specimen and one fragment, both preserved in positive hyporelief. Complete
specimen a slightly elliptical bipartite disc, 160-170 mm in diameter. Central disc 110-120 mm in diameter,
with concentric grooves and discontinuous radial ridges particularly well-developed towards its periphery.
Poorly preserved outer flange 25 mm wide, marked by irregular texture and discontinuous radial ridges.
Maximum relief of T5 mm in the central disc. Possible central circular depression 25 mm in diameter largely
covered by sediment.

Remarks. The large size and bipartite to tripartite organization are typical of Ediacaria , but the
relatively poor quality of preservation does not permit definite comparison with E.flindersi, the only
named species of this genus. Ediacaria , the largest discoid fossil of the Ediacara assemblage, was
previously interpreted as a 'medusoid' (Sprigg 1947; Glaessner and Wade 1966) but Fedonkin
(1985a) and Jenkins (1988, 1989) have reinterpreted it as an attached polyp.

Genus eoporpita Wade, 1972
Type species. Eoporpita medusa Wade, 1972

Eoporpita sp.

Plate 2, fig. 1 ; Plate 3, fig. 4

Description. Three discoid specimens preserved in positive hyporelief. Bipartite central cup consisting of a
conical knob 4-9-71 mm in diameter and 2-14-5 mm high surrounded by an upwards-flaring flange
9-2-13 0 mm in diameter and 1-3 — 1 -8 mm high; surface of the flange covered with approximately 12-24
hemispherical to radially elongate pustules 1 -5-2-5 mm wide. Flange surrounded by a radiating pattern of
approximately twenty-five club-shaped tubes 2- 1^1-2 mm wide passing off the pustules and extending outwards
to the margin of the fossil. Tubes predominantly straight to slightly sinuous, but locally strongly curved. Total
fossil diameter 30-66 mm.

Remarks. The two specimens illustrated in Plate 3, fig. 4 are approximately 25 % smaller than the
smallest specimens reported by Wade (1972), but the largest specimen is well within the described
range of variation of Eoporpita. The central cone, radially elongate pustules, and radiating clavate
tubes are all diagnostic of Eoporpita. However, in contrast with E. medusa Wade, which exhibits
overlapping whorls of tubes, the Sekwi Brook specimens exhibit only an incipient overlapping near
the central zone. For this reason, we refer our material to Eoporpita sp.

Wade (1972) originally interpreted Eoporpita as a primitive chondrophorian with a chambered
float (pneumatophore) and whorls of radiating tentacles, a view supported by most subsequent
workers (e.g. Glaessner 1979; Stanley 1986). According to this interpretation, the Sekwi Brook
specimens would reflect preservation of the oral surface; the central cone, elongate pustules, and

EXPLANATION OF PLATE 2

Fig. 1. Eoporpita sp., hyporelief. One kilometre north of locality 3, Sekwi Brook North, GSC 95905, x 1.
Fig. 2. Pteridinum sp., hyporelief. Locality 18, Sekwi Brook South, GSC 68463, x 1.

Figs 3 and 5. Cyclomedusa sp., hyporelief. Locality 3, Sekwi Brook North. 3, GSC 95907, x 1. 5, GSC 95908,
x I .

Fig. 4. Kullingia? sp., hyporelief. Locality 3, Sekwi Brook North, GSC 95909, x 1.

Figs 6 and 7. Medusinites asteroides (Sprigg), hyporelief. 6, Locality 2, Sekwi Brook North, GSC 95910, x 2.

7, Locality 3, Sekwi Brook North, GSC 959 i 1, x 2.

Fig. 8. Sekwi excentrica Hofmann, hyporelief. Locality 23, Sekwi Brook South, GSC 95912, x 2.

PLATE 2

NARBONNE and AITKEN, Ediacar an fauna

962

PALAEONTOLOGY, VOLUME 33

clavate tubes would respectively represent the gastrozoid, incipient (gonozoid) tentacles and outer
(dactylozoid) tentacles. However, experimental taphonomic studies led Norris (1989) to conclude
that fossil chondrophorians are very unlikely to preserve a pattern of tentacles similar to that shown
by Eoporpita. Evidence that the tubes were originally pliable can be seen in GSC 95916, where three
of the tubes are deflected against the margin of a large, unidentified discoid fossil, but their original
function remains uncertain.

Eoporpita is also known from the Flinders Ranges in Australia and from the White Sea region
of the Russian Platform, but had not previously been reported from North America.

Genus kullingia Glaessner in Foyn and Glaessner, 1979
Type species. Kullingia concentrica Glaessner in Foyn and Glaessner, 1979.

Kullingia? sp.

Plate 2, fig. 4

Description. One incomplete disc, partially overlying an indeterminate (?)stalked organism, preserved in convex
hyporelief. Disc 160 mm in diameter and less than I mm in relief, marked by concentric ridges uniformly
spaced 4 mm apart.

Remarks. The specimen is similar in size and morphology to the type species, K. concentrica ,
differing mainly in the wider spacing between concentric ridges. Foyn and Glaessner (1979) pointed
out the similarity between the chambered structure of Kullingia and the concentrically-chambered
float (pneumatophore) of a chondrophorian cnidarian. Alternatively, Seilacher (1984, 1989)
suggested that some chambered Ediacaran discs may represent sedentary, benthic ‘quasi-
autotrophs’ that were segmented to facilitate metabolic processes. Close similarity in morphology
between Ediacaran, Palaeozoic, and modern chondrophorians supports the view that they are
related (Stanley 1986; Narbonne et al. 1990).

Genus medusinites Glaessner and Wade, 1966
Type species. Medusina aster oides Sprigg, 1949.

Medusinites aster oides (Sprigg), 1949
Plate 2, figs 6 and 7

For synonymy up to 1987, see Narbonne and Hofmann (1987).

1987 Medusinites asteroides Narbonne and Hofmann, p. 660, pi. 73, figs 7-9
1987 Medusinites asteroides Gureev, pp. 30-31, fig. 10.

Description. Nine bipartite discs preserved in convex hyporelief. Smooth outer ring 1 3-4-24-8 mm (mean
1 8-7 mm) in diameter and less than 1-5 mm high with an incomplete outer flange 0-2-1 -2 mm wide preserved
on some specimens. Smooth inner disc 4-2-12-4 mm (mean 7 0 mm) in diameter and 0-5—2- 1 mm high separated
from outer ring by a sharp annulus. Diameter of the inner disc one quarter of one half of total diameter of the
fossil. Fossil surface generally smooth, locally with very faint radial ridges and/or grooves on the outer ring.

Remarks. Specimens closely resemble those described by Glaessner and Wade (1966, pi. 97, figs 1
and 2) from Ediacara and by Narbonne and Hofmann (1987, pi. 73, figs 7-9) from the Wernecke
Mountains. Sprigg (1949) and Glaessner and Wade (1966) regarded Medusinites as the impression
of a medusoid. Alternatively, Fedonkin (1985a) interpreted Paliella , a very similar form possibly
synonymous with Medusinites (Narbonne and Hofmann 1987), as a benthic ‘polyp’. The absence

NARBONNE AND AITKEN: CANADIAN EDIACARAN FAUNA

963

of specimens that preserve the oral surface suggests that Medusinites represents the base of a
sedentary organism which lived with its mouth facing upwards. This is supported by the relatively
high relief of some specimens, which implies that their bases were already buried in the mud at the
time of sand deposition.

Genus pteridinium Giirich, 1933
Type species. Pteridinium simplex Giirich, 1933.

Pteridinium sp.

Plate 2, fig. 2

Description. One specimen figured by Aitken (1989a, fig. 6a) but not previously described; the scale bar on
Aitken’s figure is incorrect, and should read 3 75 cm. Specimen at least 92 mm long (incomplete at both ends),
60-62 mm wide, and with 4-5 mm of relief. Straight, smooth ribs 3-5-4 mm wide passing laterally from median
furrow (V-angle = 100-120°) to the margin of the fossil. No evidence of a third vane.

Remarks. Pflug (1970) interpreted Pteridinium as a three- vaned, leaf-like structure. Distortion of
these flexible organisms led to the wide range in morphology evident in Pteridinium (cf. Richter
1955; Glaessner and Wade 1966; Pflug 1970; Fedonkin 1981), a problem which hinders the
recognition of species and comparisons with related forms such as Inkrylovia. Pteridinium was a
cosmopolitan taxon that occurred over a broad bathymetric range (Gibson et al. 1984). The
specimen from Sekwi Brook South in the first Ediacaran body fossil reported from deep-water
carbonates.

Genus sekwia Hofmann, 1981
Type species. Sekwia excentrica Hofmann, 1981.

Sekwia excentrica Hofmann, 1981
Plate 2, fig. 8

1981 Sekwia excentrica Hofmann, pp. 305-307, fig. 4a-g.

Remarks. Sekwia was originally defined by Hofmann (1981) on the basis of twenty-four specimens collected
from strata now assigned to the middle part of the Blueflower Formation at Sekwi Brook South. Our fieldwork
in 1988 resulted in the discovery of two additional specimens from the same bed. Both specimens are well
within the range of variation described by Hofmann (1981).

Sekwia excentrica is similar in size and eccentricity to the laterally deformed specimens of Cyclomedusa sp. (PI.
1, fig. 7; Wade 1972, pi. 41, fig. 1) and Sekwia kaptarenkoe Gureev, 1987, but exhibits much lower relief and
apparently lacks a central tubercle. One high-relief specimen questionably identified as Sekwia by Hofmann
(1981, fig. 4h) was later referred to Beltanelliformis by Narbonne and Hofmann (1987). Re-examination of this
specimen by the present authors suggests that it represents an indeterminate fossil that has been badly
deformed by loading, and cannot definitely be ascribed to either of these genera.

Sekwia probably represents the impression of a highly conical cnidarian (Hofmann 1981), most likely a
‘polyp’ similar in some respects to Cyclomedusa sp., that was flattened and deformed laterally during burial.

Plumose Problematicum
Plate 1, fig. 3

Description. One incomplete specimen preserved in negative hyporelief. Preserved length 51 mm, with a

964

PALAEONTOLOGY, VOLUME 33

maximum relief of 2-3 mm. Specimen petaloid, with a sharp outer margin. Plumose arrangement of sharp
ridges 1—2-5 mm wide and up to 1-4 mm high extending the length of the specimen.

Remarks. The specimen superficially resembles tectonic features such as plumose fractures, but this
is ruled out by the presence of a thin carbonaceous lamella that evenly covers the entire bedding
surface, including the sides and tops of the fine ridges. This indicates that they represent very sharp
folds of the bedding surface (prelithification) rather than tectonic fractures of the rock surface
(postlithification). The presence of a sharp outer margin is also atypical of plumose fractures.
Preservation of the structure in negative hyporelief, and the occurrence of undeformed specimens
of indeterminate discoid megafossils preserved in positive hyporelief elsewhere on the same slab
(GSC. 95906), both imply that the structure does not represent a load mark, a pattern of rill marks,
or some other inorganic sedimentary structure.

The specimen resembles the oral surface of a single lobe of Inaria karli Gehling, 1988 in size,
preservation in negative hyporelief, and the plumose arrangement of longitudinal ridges, but differs
in exhibiting a single rather than double border and in several other respects. Broad comparisons
can also be made with Lomosovis Fedonkin, which, however, is parallel-sided rather than petaloid
and also differs in the pattern of ridge development. The range of morphology of the Plumose
Problematicum is unknown, as it is presently known only from a single, fragmentary specimen.

The Plumose Problematicum probably represents a bag-shaped organism that was folded
longitudinally during burial.

ICHNOFOSSILS

Ichnogenus aulichnites Fenton and Fenton, 1937
Type ichnospecies. Aulichnites parkerensis Fenton and Fenton, 1937.

Aulichnites ichnosp.

Plate 3, figs 1 and 3

Description. Ten specimens preserved in positive relief. Unbranched, horizontal, sinuous to irregularly
meandering, 10-1 1 mm wide bilobate trails; one specimen passing laterally into a trilobate trail. Smooth lobes
separated by a median furrow 2-3 mm wide. Cross-sectional shape indeterminate, despite sectioning.

Remarks. Specimens are approximately twice as large as those figured by Fedonkin (1980, pi. 1, figs
3-5; 1985c, pi. 26, figs 1 and 8) from the Vendian of the White Sea area, but are otherwise similar.
They are similar to the type species, A. parkerensis , in size and overall morphology, but differ in the
absence of curved striae on the surface of the lobes. Some of the type specimens of A. parkerensis

EXPLANATION OF PLATE 3

Figs 1 and 3. Aulichnites sp., epirelief. Locality 13, Sekwi Brook North. 1, GSC 95913, x 1. 3, GSC 95914, x 1

Fig. 2. Helminthopsis abeli Ksiazkiewicz, 1977, hyporelief. Locality 26, Sekwi Brook South, GSC 95915,
x 0-75.

Fig. 4. Eoporpita sp., hyporelief. Locality 3, Sekwi Brook North, GSC 95916 (lower left), GSC 95917 (upper
right), x I .

Fig. 5. Planolites montanus Richter, hyporelief. Locality 30, Majesty Section, GSC 95918, x I.

Figs 6-8. Palaeophycus tubularis Hall. 6, hyporelief, locality 13, Sekwi Brook North, GSC 95919, x 1. 7,
hyporelief, locality 30, Majesty Section, GSC 95920, x 1. 8, cross-section of 7 showing thin clay lining and
partial collapse.

Fig. 9. Neonereitesl sp. and Planolites montanus. hyporelief. Locality 1 1, Sekwi Brook North, GSC 95921 and
95922, x F5.

PLATE 3

NARBONNE and AITKEN, Ediacaran fauna

966

PALAEONTOLOGY, VOLUME 33

exhibit a unilobed lower surface (Hakes 1977), a feature not determinable on the present material.
These preservational features preclude identification of the Sekwi Brook specimens to the
ichnospecies level.

Aulichnites is generally interpreted as the grazing trail of a gastropod (Fenton and Fenton 1937;
Hantzschel 1975; Howard and Frey 1984) or similar organism. Yochelson and Schindel (1978)
doubted this interpretation, but did not offer specific reasons for their view. Aulichnites also ranges
throughout the Phanerozoic (Hantzschel 1975). It occurs most commonly in shallow marine
deposits (Hakes 1977), but is also found in deep-water turbidites (Hill 1981).

Ichnogenus helminthoida Schafhautl, 1851

Type ichnospecies. Helminthoida labyrinthica Heer, 1865.

Helminthoida ichnosp.

Plate 4, fig. 1

Description. One specimen preserved in part and counterpart, and a second uncollectable specimen studied in
the field. Unbranched, unlined, horizontal meandering burrows 2- 1-2-9 mm wide, with smooth surface and fill
similar to host lithology. Meanders increasing in amplitude distally from 100-150 mm; wavelength 40-60 mm.
Adjacent burrow segments parallel to sub-parallel.

Remarks. Helminthoida is a regularly meandering, hemicylindrical trace that shows a strong
tendency towards development of parallel meanders (Hantzschel 1975; Ksiazkiewicz 1977). As
such, it can readily be distinguished from Taphr helminthoida Ksiazkiewicz (which is bilobate) and
Helminthopsis Heer (which is less regularly meandering). The specimens of Helminthoida from
Sekwi Brook display the meandering habit diagnostic of the ichnogenus, but are considerably larger
and exhibit more open meanders than most Phanerozoic specimens (but see the Early Cambrian
specimens of H. miocenica in Crimes and Anderson 1985, figs 7. 3-7. 5). Seilacher (1974, 1977) has
documented similar trends in the evolution of other meandering trace fossils such as Nereites, and
has related it to a gradual decline in body size and increase in behavioural complexity in deep-sea
environments throughout the Phanerozoic.

Helminthoida is a characteristic element of Cretaceous-Tertiary flysch deposits (e.g. Seilacher,
1964; Ksiazkiewicz 1970, 1977; Crimes 1977) and also occurs in Recent deep-sea sediments
(Chamberlain 1975). It has been reported from Ordovician and Silurian flysch (Pickerill 1980, 1981 )
and from Early Cambrian storm deposits of eastern Newfoundland (Crimes and Anderson 1985).
Glaessner (1969) figured complexly meandering burrows from Ediacara under the name ‘Form C’;
these have been compared with Helminthoida by Crimes (1987) and are also quite similar to
Yelovichnus gracilis Fedonkin, 1985c from the Vendian of the White Sea area. Fedonkin (1985c) has
also figured specimens of Helminthoida ichnosp. from the same strata.

EXPLANATION OF PLATE 4

Fig. I. Helminthoida sp., hyporelief. Locality 10, Sekwi Brook North, GSC 95923, xO-5.

Figs 2 and 3. Helminthoidichnites tenuis Fitch, epirelief. 2, locality 6, Sekwi Brook North, GSC 95924, x 0 5.

3, locality 19, Sekwi Brook South, GSC 95925, x0-5.

Fig. 4. Helminthopsis ? sp., epirelief. Locality 11, Sekwi Brook North, GSC 95926, xO-5.

Figs 5 and 6. Helminthopsis irregularis (Schafhautl), hyporelief. Locality 26, Sekwi Brook South. 5, GSC
95927, 1. 6, GSC 95928, x 0-75.

PLATE 4

NARBONNE and AITKEN, Ediacar an fauna

968

PALAEONTOLOGY, VOLUME 33

Ichnogenus helminthoidichnites Fitch, 1850

Type ichnospecies. Helminthoidichnites tenuis Fitch, 1850

Helminthoidichnites tenuis

Plate 4, figs 2 and 3

1969 ‘Form B' Glaessner, p. 381, fig. 5b.

71969 ‘Form E’ Glaessner, p. 382, fig. 5f.

1976 ‘crawling trails’ Palij, pi. 26, figs 1 and 2.

1979 ‘crawling trails, first variety’ Palij et al. , p. 77, pi. 53, figs 2 and 4.

1981 Gordia sp. Hofmann, p. 307, fig. 5b.

1981 Gordial sp. Hofmann, p. 307, fig. 5d.

1983 Gordia sp. Fritz et al., pi. 44.1, fig. 3.

1985 Gordia sp. Crimes and Anderson, p. 321, fig. 5.8.

1985c Gordia sp. Fedonkin, pi. 23, fig. 1.

1987 Gordia marina Narbonne and Hofmann, pp. 668-670, text-fig. 10a
1989a Planolites sp. Aitken, fig. 6c.

1989 Planolites beverleyensis Gibson, p. 5, figs 3.8 and 4.1.

Description. Numerous specimens (47 measured) preserved in both positive and negative relief on the tops and
soles of beds. Irregularly sinuous to meandering, smooth, unlined, cylindrical burrows 1 - 1—3-3 mm (mean =
21 mm) in diameter; diameter constant within a single specimen. Burrows commonly crossing adjacent
specimens, and rarely (three examples) crossing previously constructed portions of the same specimen.
Specimens preserved in negative epirelief with narrow levees. Specimens preserved in positive relief with fill
similar to host lithology, and flanked by narrow grooves. True branching absent, but offset of crossing
specimens commonly producing pseudobranches.

Remarks. Helminthoidichnites is herein used in the sense of Hofmann and Patel (1989) for irregularly
sinuous to meandering burrows with random crossings of the same and adjacent burrows. As such,
it differs from Gordia Emmons, 1844 which exhibits long parallel limbs with numerous over-
crossings, and from Helminthopsis Heer, 1877 which is irregularly meandering but avoids level
crossings of the same or adjacent burrows. The Sekwi Brook specimens are most similar to material
from Ediacara described by Glaessner (1969) as ‘Form B’, and to Gordia sp. from the late
Precambrian of the Burin Peninsula, Newfoundland (Crimes and Anderson 1985). Specific points
of similarity include the size, presence of narrow levees, development of pseudobranches, and the
tendency of burrows commonly to cross adjacent burrows but only rarely to cross themselves.

Helminthoidichnites is common in the Ediacaran, where it previously has been referred to Gordia
or Planolites (see synonymy), and also ranges throughout the Phanerozoic (Hofmann and Patel
1989). It is eurybathic, and probably represents the crawling or feeding trail of a vermiform
organism.

Ichnogenus helminthopsis Heer, 1877

Type ichnospecies. Helminthopsis magna Heer, 1877 (but see Ksiazkicwicz 1977, p. 116).

Helminthopsis aheli Ksi§zkiewicz, 1977
Plate 3, fig. 2

Description. Three specimens preserved in positive hyporelief. Smooth, unbranched, unlined, hemicylindrical
burrows 6 0-6-3 mm in diameter, with burrow-fill similar to host lithology. Burrows loosely winding, with a
tendency to meandering.

NARBONNE AND AITKEN: CANADIAN EDIACARAN FAUNA

969

Remarks. The specimens closely conform in size and morphology to those illustrated by
Ksiazkiewicz (1977, pi. 12, fig. 5; text-fig, 21a-/?). H. abeli ranges throughout the Phanerozoic, in
rocks of both shallow-marine (e.g. Crimes and Anderson 1985) and especially deep-sea (e.g.
Ksiazkiewicz 1977) aspect. It had not previously been reported from the Precambrian.

Helminthopsis is an irregularly meandering burrow that avoids level-crossings (Ksiazkiewicz
1977). The ichnogenus has been reported from the Precambrian of the Cassiar Mountains,
northwestern Canada (Fritz and Crimes 1985) and the Carolina Slate Belt of the southeastern USA
(Gibson 1989), and also ranges throughout the Phanerozoic (Hantzschel 1975). Helminthopsis is
eurybathic, but is most commonly reported from deep-water flysch deposits (Pickerill 1981). It
represents the feeding or grazing burrow of a vermiform organism, most likely an annelid
(Ksiazkiewicz 1977).

Helminthopsis irregularis (Schafhautl), 1851
Plate 4, figs 5 and 6

Description -. Numerous specimens preserved in positive hyporelief on eleven slabs. Burrows ranging from
irregularly winding to tightly meandering with parallel arms and second-order windings, even within the same
specimen. Burrows unbranched and unlined, with fill similar to host lithology, predominantly smooth but with
local fiansverse markings. Burrow diameter 1-4-31 mm, with diameter constant within any single specimen.
Cress-overs of the same or adjacent burrows absent.

Remarks. This ichnospecies was originally assigned to Helminthoida Schafhautl. Ksiazkiewicz
(1977) regarded it as a very regular version of Helminthopsis rather than an irregular version of
Helminthoida, a view with which we concur. Helminthopsis irregularis exhibits two behavioural
approaches to avoid level-crossings (PI. 4, figs 5 and 6). The most common approach is to cross at
a stratigraphically higher or lower level than the previous burrow ; this avoids crossings, but requires
that the feeding organism leave the organic-rich bedding plane briefly. A less common, but more
efficient method is the local development of subparallel meanders to avoid crossing over its own or
an adjacent burrow. This represents a primitive version of phobo taxis (sensu Richter 1928), which
characterizes meandering burrows in Phanerozoic deep-sea deposits (Seilacher 1967, 1974; Raup
and Seilacher 1968).

H. irregularis has previously been described only from Mesozoic and Tertiary, deep-water flysch
deposits (Ksiqzkiewicz 1977). Paczesna (1988) listed, but has not yet figured or described, a
possible specimen from the upper Proterozoic Lublin Formation of Poland.

Helminthopsis ? ichnosp.

Plate 4, fig. 4

Description. One specimen preserved in negative epirelief. Horizontal burrow 2 mm in diameter consisting of
a series of connected arcuate loops. Burrow flanked by narrow levees.

Remarks. The single specimen is most similar to one illustrated by Gibson (1989) from the deep-
water Carolina Slate Belt of the southeastern U.S.A. Broad comparisons can also be made with
Gordia arcuata Ksiqzkiewicz, 1977, but Helminthopsis sp. is considerably less arcuate with turnings
confined to the horizontal plane.

Ichnogenus lockeia James, 1879

Type ichnospecies. Lockeia siliquaria James, 1879.

970

PALAEONTOLOGY, VOLUME 33

text-fig. 7. a— c, Torrowangea rosei Webby, hyporelief; a, locality 29, Majesty Section, GSC 95929, x 1 ; b,
locality 1 5, Sekwi Brook North, GSC 95930, x 1 ; c, locality 29, Majesty Section, GSC 9593 1 , xl.D, Knotted
circular burrow, hyporelief. Locality 1 1, Sekwi Brook North, GSC 95932, x 0 5. e-h, Lockeia ichnosp., locality
13, Sekwi Brook North; e-g, hyporelief, GSC 95933-95937, x2; h, photomicrograph cross-section, GSC

95938, x 5,

NARBONNE AND AITKEN: CANADIAN EDIACARAN FAUNA

971

Lockeia ichnosp.

Text-fig. 7e-h

Description. Eight specimens preserved in positive hyporelief. Oblong bodies 4-3— 5 6 mm (mean = 5-2 mm)
long and 2'4-2-9 mm (mean = 2-7 mm) wide, with a constant length/width ratio of approximately 2:1.
Specimens smooth and almond-shaped, with one rounded and one pointed end. Maximum relief 2-2 mm;
burrow-fill structureless, and similar in composition to the overlying quartzarenite bed. Sandstone sole sharply
upturned in the immediate vicinity of the burrow.

Remarks. This ichnogenus is referred to under the name Pelecypodichnus Seilacher by many
workers, but we agree with Osgood (1970), Hantzschel (1975) and Maples and West (1989) that
Lockeia is a valid name and the subjective senior synonym of Pelecypodichnus. The specimens from
Sekwi Brook are similar to L. amygdaloides (Seilacher) in exhibiting an almond-shaped cross-section
and a consistent 2:1 length/width ratio, but lack the keel present on some specimens of L.
amygdaloides.

Based on its symmetry and comparisons with the burrows of modern bivalves, Seilacher (1953)
concluded that Lockeia (‘ Pelecypodichnus ’) represents the resting burrow of a bivalve. Seilacher’s
view has been supported by virtually all subsequent workers (e.g. Osgood 1970; Eager 1974; Hakes
1976, 1977; Bromley and Asgaard 1979; Thoms and Berg 1985; Wright and Benton 1987), many
of whom suggested specific bivalve genera as the originators of the Lockeia in their sections. The
Sekwi Brook specimens of Lockeia exhibit typical bivalve symmetry, with a single plane of mirror
symmetry that passes longitudinally through the structure. Another feature consistent with a
bivalve origin is the upturning of the sole of the bed in the immediate vicinity of the burrow, and
the associated upward pinch of clay drapes around the burrow (Text-fig. 7h). This feature has been
described from Phanerozoic specimens of Lockeia amygdaloides (e.g. Bromley and Asgaard 1979,
fig. 5b), and implies that the mode of burrowing was bivalve-like, and consisted of probing and
penetration rather than excavation (Pojeta 1987).

Bivalves are the only extant organisms that exhibit both a cross-sectional shape and a mode of
burrowing consistent with the Sekwi Brook specimens of Lockeia. However, the oldest confirmed
bivalve body fossils are Early Cambrian (Pojeta et al. 1973; Pojeta 1985; Runnegar 1985). The
presence of Lockeia in the Ediacaran implies either that very thin-shelled/poorly calcified bivalves
evolved in the Late Precambrian, or that Lockeia could also be constructed by a now extinct group
of organisms similar to bivalves in some morphological and behavioural aspects.

Lockeia is a facies-crossing ichnogenus that is known from a wide range of non-marine (e.g.
Bromley and Asgaard 1979) and shallow to deep marine (e.g. Crimes 1977, table 4) settings.

Ichnogenus neonereites Seilacher, 1960
Type ichnospecies. Neonereites biserialis Seilacher, 1960.

Neonereites! ichnosp.

Plate 3, fig. 9

Description. Two specimens preserved in positive hyporelief. Each specimen comprising a gently curved,
uniserial string of hemispheres, 0-4-0-7 mm in relief and composed of fine sandstone similar to host lithology;
adjacent hemispheres in contact. Hemispheres progressively increasing in size distally. Larger specimen
16-0 mm long, consisting of 7 hemispheres 1 -2—3-5 mm in diameter; smaller specimen 8T mm long, consisting
of 4 hemispheres 1 -7-2-7 mm in diameter.

Remarks. The specimens are most similar to Neonereites uniserialis Seilacher, but differ in that
hemispheres progressively increase in size along the chain. No similar forms of Neonereites have
previously been described.

972

PALAEONTOLOGY, VOLUME 33

Neonereites is commonly reported from Ediacaran strata (Text-fig. 6), and also ranges
throughout the Phanerozoic (Hantzschel 1975). It is a facies-crossing form (Seilacher 1964; Crimes
1977) which represents the feeding burrow of a vagile vermiform organism, probably an annelid
(Hakes 1976).

Ichnogenus palaeophycus Hall, 1847
Type ichnospecies. Palaeophycus tubularis Hall, 1847.

Palaeophycus tubularis Hall, 1847
Plate 3, figs 6-8

Description. Thirty-four specimens preserved in convex hyporelief and full relief. Straight, curved, to gently
sinuous, horizontal, predominantly unbranched, hemicylindrical to cylindrical burrows, 3-7-7 T mm (mean =
5 8 mm) in diameter. Preserved length up to 140 mm. Burrows smooth-walled and very thinly lined with clay,
locally partially collapsed, and with fill similar to host lithology.

Remarks. The taxonomy of Palaeophycus has been reviewed by Pemberton and Frey (1982), who
distinguished it from the similar ichnogenus Planolites Nicholson by the presence of a burrow-lining
fill similar to the host lithology, and local evidence of partial burrow collapse. These features reflect
passive filling of originally open burrow systems constructed by suspension-feeding or predaceous
organisms, predominantly annelids (Pemberton and Frey 1982). Palaeophycus is common in
Phanerozoic strata, where it occurs in virtually all sedimentary facies (Pemberton and Frey 1982).
It is commonly reported in compilations as ranging in age from ‘Precambrian to Recent’ (e.g.
Hantzschel 1975), but the only primary report of Precambrian Palaeophycus known to us is from
the Risky Formation of the Wernecke Mountains (Nowlan et a/., 1985). The second variety of
‘crawling traces’ described by Palij et al. (1979, pp. 77-78, pi. 53, fig. 5; pi. 54, fig. 1) from the
Vendian of Podolia is broadly similar, but further details of the nature of the burrow fill and the
presence or absence of a burrow lining are necessary to determine whether it represents
Palaeophycus or Planolites.

Ichnogenus planolites Nicholson, 1873
Type ichnospecies. Planolites vulgaris Nicholson and Hinde, 1875.

Planolites montanus Richter, 1937
Plate 3, figs 5 and 9

1970 ‘hypichnial and exichnial casts’ Banks, p. 26, pi. \b,d.

1970 ‘trails’ Webby, pp. 87-88, fig. 3b-d, ?fig. 4 a-b.

1972 a ‘threadlike trials' Germs, p. 208, pi. 26, figs 5 and 7, pi. 27, fig. 1.

19726 ‘threadlike trails’ Germs, p. 866, pi. 1, figs 5 and 7, pi. 2, fig. 1.

1973 ‘hypichnial and endichnial burrows’ Banks, p. 4, fig. 4 a.

1977 Planolites sp. Fedonkin, p. 184, pi. 2 cl.

1979 ‘crawling traces, third variety’ Palij et al ., pp. 77-78, pi. 54, fig. 2.

1979 Planolites cf. serpens Palij et al.. p. 73, pi. 42, fig. 6.

1984 Planolites sp. Glaessner, p. 70, fig. 2/7.

1985c Planolites cf. serpens Fedonkin, pi. 28, figs 3 and 6.

1987 Planolites montanus Narbonne and Hofmann, pp. 670-671, text-fig. 10 b.d.
1989 Planolites montanus Gibson, p. 5, fig. 4.1.

NARBONNE AND AITKEN: CANADIAN EDIACARAN FAUNA

973

Description. Numerous specimens preserved in convex hyporelief and full relief. Burrows cyclindrical,
unbranched and unlined, and with structureless fill dissimilar to host lithology. Burrows highly sinuous and
undulatory, occurring on bedding surfaces as small knobs and as discontinuous burrow segments typically less
than 5 mm long (but ranging up to 55 mm long). Burrow diameter 03-1-7 mm (mean = 0-8 mm; n = 120).

Remarks. The taxonomy of Planolites was reviewed by Pemberton and Frey (1982), who recognized
only three valid ichnospecies from among the thirty-three formally defined forms; several new
ichnospecies have been named subsequently. The Mackenzie Mountain specimens fall within the
diagnosis of P. montanus as ‘relatively small, curved to contorted burrows’ (Pemberton and Frey
1982, p. 870). Planolites is a eurybathic ichnogenus that could be produced by a wide variety of
vermiform organisms (Alpert 1975; Pemberton and Frey 1982).

Planolites ranges in age from Ediacaran to Recent (Hantzschel 1975). P. montanus occurs
throughout the Phanerozoic (Pemberton and Frey 1982, pp. 869-870) and is perhaps the most
widely distributed Ediacaran ichnospecies (see synonymy).

Ichnogenus torrowangea Webby, 1970
Type ichnospecies. Torrowangea rosei Webby, 1970.

Torrowangea rosei Webby, 1970
Text-fig. 7a-c

1981 Torrowangea sp. Hofmann, p. 309, fig. 5c.

1985 Torrowangea rosei Paczesna, pi. 1, figs 1, 2, 5.

1986 Torrowangea rosei Paczesna, p. 35, pi. I, figs 1 and 4

1987 Torrowangea sp. Narbonne et al ., fig. 6c.

1 989a Torrowangea sp. Aitken, fig. 6b.

Description. Numerous specimens preserved in positive hyporelief and rarely in negative epirelief. Sinuous to
irregularly meandering, horizontal burrows. Burrows unbranched and unlined; cross-overs of the same or
adjacent individuals occurring commonly, in some instances forming figure-8 pattern. Burrows 0-8-2-7 mm
(mean = 18 mm; n = 160), with irregularly-spaced transverse constrictions resulting in a marked ‘pinch-and-
swell’ appearance. Burrow-fill slightly coarser and better sorted than host lithology.

Remarks. Torrowangea sp. was described from the Sekwi Brook South section by Hofmann (1981)
on the basis of relatively few specimens. Our analysis of several hundred specimens confirms
Hofmann’s generic identification, and further suggests that the material can be referred to T. rosei.
Irregularly meandering and crossing specimens similar to Webby’s (1970, fig. 18c) holotype occur
rarely (Text-fig. 7a, c), but most specimens are sinuous (Text-fig. 7b) and are best compared with
two of the paratypes figured by Webby (1970, fig. 18a, b) and with the specimens illustrated by
Hofmann (1981) and Paczesna (1985, 1986).

The present specimens slightly extend the range of burrow diameters for Torrowangea in both
directions, but the mean is well within the range of the type material. The marked ‘pinch-and-swell’
appearance of Torrowangea was interpreted by Webby ( 1970) as a backfill structure, but this is not
consistent with the apparent absence of internal structure in longitudinal thin sections of the
Mackenzie Mountain specimens. Most likely, it reflects peristalsis.

Torrowangea was originally described from the Lintiss Vale Formation of New South Wales,
Australia (Webby 1970). The age of this unit is controversial, with some authors favouring a latest
Precambrian (probably Kotlin-equivalent) age and others favouring an earliest Cambrian (probably
Rovno-equivalent) age (see review in Webby 1984). Narbonne and Myrow (1988) and Walter et al.
(in press) believed that the presence of Phycodesl and other complex burrows favoured the latter

974

PALAEONTOLOGY, VOLUME 33

view. Torrowangea has been described from undoubted Precambrian strata (e.g. Paczesna 1985,
1986), and from subtrilobite Lower Cambrian deposits (e.g. Acenolaza and Durand 1973).
Pemberton and Frey (1982) tentatively referred IPlanolites octichnus Chamberlain, 1971 from the
Carboniferous of Oklahoma to Torrowangea ; our examination of Chamberlain’s specimens
indicated that the pattern of meandering is similar but that the irregular transverse constrictions
diagnostic of Torrowangea are not present.

Torrowangea represents the feeding burrow of a worm-like organism, probably an annelid. It
occurs in both shallow-water and deep-water deposits, and is most common in fine-grained, muddy
sandstones. In the Sekwi Brook area, specimens also occur in ribbon-bedded lime mudstone.

Knotted Burrow
Text-fig. 7d

Description. Single specimen preserved in positive hyporelief. Unlined(?) hemicylindrical burrow 3-1 mm in
diameter forming an irregular circle 59-79 mm in diameter, and forming a crude ‘knot’ at the point of
overcrossing. Three arcuate burrows radiating outwards more than 100 mm. Burrow fill similar to host
lithology.

Remarks. The overall structure is unlike any known to us. The central circle is somewhat larger and
less regular than Circulichnis montanus Vyalov, 1971, but is otherwise similar. Circulichnis ranges
throughout the Phanerozoic (Fillion and Pickerill 1984), and is also figured under the name Gordia
arcuatal from the late Precambrian of the Carolina Slate Belt (Gibson 1989, fig. 3.1). It is
eurybathic, but is most commonly reported from deep-water settings (Fillion and Pickerill 1984).

Similarity between the diameters of the circle and the branches, termination of all the branches
at the circle, and the extreme scarcity of arcuate burrows of similar size elsewhere in the Blueflower
Formation, implies that central circular burrow and the radiating arcuate burrows are related. The
arcuate burrows may represent different activity by the same species or even organism, or (less
likely) may represent branches that pass off the circular knotted burrow at a higher stratigraphic
level and descend rapidly to the level of the bedding surface.

CONCLUSIONS

1. Ediacaran fossils occur sporadically throughout approximately one kilometre of deep-water
strata, encompassing the upper Sheepbed Formation to the top of the Blueflower Formation, in the
Sekwi Brook area of northwestern Canada. At least eleven body fossil species and twelve
ichnospecies are present in this interval.

2. Ediacaran megafossils at Sekwi Brook comprise mainly sessile, benthic polypoid and frond-like
organisms along with a single, presumably pelagic, chondrophorian. The occurrence of a
predominantly in situ benthic fauna in muddy slope deposits below storm wave-base is not
consistent with the hypothesis that these taxa functioned exclusively as photoautotrophs.

3. The Sekwi Brook megafossil assemblage is broadly similar to that reported from coeval shallow
shelf deposits in the Wernecke Mountains, Flinders Ranges and Russian Platform, but apparently
lacks the (?)actinian impression Beltanelliformis which is abundant in Ediacaran shallow shelf
deposits. Similarity of Ediacaran assemblages in shelf and slope settings implies that the organisms
had broad palaeoenvironmental ranges, and enhances their correlation potential.

4. The occurrence of a relatively diverse trace fossil assemblage at Sekwi Brook implies that the
initial radiation of infaunal organisms extended into slope settings. The ichnocoenoses includes
irregularly and regularly meandering burrows, and may represent a primitive version of the Nereites
ichnofacies, which characterizes deep-water Phanerozoic deposits.

NARBONNE AND AITKEN: CANADIAN EDIACARAN FAUNA

975

Acknowledgements. We gratefully acknowledge excellent field assistance by Paul Fejer and Lascelles Gayle, and
the skill of our Okanagan helicopter pilots Lois Hill, Tony Duckworth and Ray Portlock. Critical comments
by M. M. Anderson, M. A. Fedonkin, H. J. Hofmann, R. J. F. Jenkins and R. K. Pickerill greatly improved
the manuscript. Olga Iljewiwn translated critical articles, and Clinton Cowan, Steve Forrester, John Milne, and
Ela Mazur assisted with drafting and photography. T. E. Bolton (Geological Survey of Canada), R. J. F.
Jenkins (University of Adelaide), N. Pledge (South Australia Museum), and K. Westphal (University of
Winsconsin, Madison) provided access to relevant type specimens. Narbonne also acknowledges financial
assistance from the Natural Sciences and Engineering Research Council of Canada (NSERC Grant A2648)
and the Queen’s University Advisory Research Committee. This is Geological Survey of Canada Publication
No. 28089.

REFERENCES

acenolaza, f. G. and durand, f. 1973. Trazas fosiles del basanrento cristalino del noroeste Argentino. Boletin
Asociacion Geologica de Cordoba , 2, 45-55.

aitken, j. D. 1966. Middle Cambrian to Middle Ordovician cyclic sedimentation, southern Rocky Mountains
of Alberta. Bulletin of Canadian Petroleum Geology , 14, 405 44 1

1978. Revised models for depositional Grand Cycles, Cambrian of the southern Rocky Mountains,

Canada. Bulletin of Canadian Petroleum Geology , 26, 515-542.

1987. First appearance of trace fossils and small shelly fossils in Mackenzie Mountains, northwest

Canada, in relation to the highest glacial deposits (abs.). New York State Museum , Bulletin, 463, 8.

1989a. Uppermost Proterozoic formations in the central Mackenzie Mountains, Northwest Territories.

Canada. Geological Survey of Canada Bulletin, 368, 1-26.

19896. A “sea level” curve for the terminal Proterozoic, Mackenzie Mountains, N.W.T. - progress

report. Program with Abstracts, Geological Association of Canada, 14, A99.

— and narbonne, G. m. 1989. Two occurrences of Precambrian thrombolites from the Mackenzie
Mountains, northwestern Canada. Palaios, 4, 384-388.
alpert, s. p. 1975. Planolites and Skolithos from the upper Precambrian-Lower Cambrian, White-Inyo
Mountains, California. Journal of Paleontology, 49, 509-521.

1977. Trace fossils and the basal Cambrian boundary. Geological Journal Special Issue, 9, 1-8.
anderson, M. m. and conway morris, s. 1982. A review, with descriptions of four unusual forms, of the soft-
bodied fauna of the Conception and St. John’s Groups (Late Precambrian), Avalon Peninsula,
Newfoundland. Proceedings of the Third North American Paleontogical Convention, I, 1-8.

Armstrong, R. L., eisbacher, G. h. and evans, p. d. 1982. Ages and stratigraphic-tectonic significance of
Proterozoic diabase sheets, Mackenzie Mountains, northwestern Canada. Canadian Journal of Earth
Sciences, 19, 316-323.

banks, n. l. 1970. Trace fossils from the late Precambrian and Lower Cambrian of Finnmark, Norway.
Geological Journal Special Issue, 3, 19-34.

1973. Trace fossils in the Halkhavarre section of the Dividal Group (?Late Precambrian-Lower
Cambrian), Finnmark. Norsk Geologisk Undersokelse , 288, 1-6.
baudet, d., aitken, j. d. and vanguestaine, m. 1989. Palynology of uppermost Proterozoic and lowermost
Cambrian formations, central Mackenzie Mountains, northwestern Canada. Canadian Journal of Earth
Sciences, 26, 129-148.

bromley, R. G. and asgaard, u. 1979. Triassic freshwater ichnocoenoses from Carlsberg Fjord, East
Greenland. Palaeogeography, Palaeoclimatology, Palaeocology , 28, 39-80.
chamberlain, c. k. 1971. Morphology and ethology of trace fossils from the Ouachita Mountains, southeast
Oklahoma. Journal of Paleontology, 45, 212-246.

1975. Trace fossils in DSDP cores of the Pacific. Journal of Paleontology, 49, 1074-1096.
cloud, p. e. and glaessner, m. f. 1982. The Ediacarian Period and System: Metazoa inherit the earth. Science,
217, 783-792.

conway morris, s. 1985. The Ediacaran biota and early metazoan evolution. Geological Magazine, 122, 77-81.
crimes, t. p. 1974. Colonization of the early ocean floor. Nature, 248, 328-330.

1977. Trace fossils of an Eocene deep-sea sand fan, northern Spain. Geological Journal Special Issue, 9,
71-90.

976

PALAEONTOLOGY. VOLUME 33

1987. Trace fossils and correlation of late Precambrian and early Cambrian strata. Geological Magazine,
124, 97-119.

— and anderson, m. m. 1985. Trace fossils from Late Precambrian-Early Cambrian strata of southeastern
Newfoundland (Canada): temporal and environmental implications. Journal of Paleontology, 59, 310-343.
eager, r. m. c. 1974. The shape of shell of Carbonicola in relation to burrowing. Lethaia, 7, 219-238.
eisbacher, g. h. 1981. Sedimentary tectonics and glacial record in the Windermere Supergroup, Mackenzie
Mountains, northwestern Canada. Geological Survey of Canada, Paper, 80-27, 1-40.

1985. Late Proterozoic rifting, glacial sedimentation, and sedimentary cycles in the light of Windermere

deposition, western Canada. Palaeogeography, Palaeoclimatology, Palaeoecology, 51, 231-254.
ekdale, a. a., bromley, R. G. and pemberton, s. G. 1984. Ichnology. The use of trace fossils in sedimentology
and stratigraphy. Society of Economic Paleontologists and Mineralogists, Short Course, 15, 1-317.
emmons, m. a. 1844. The Taconic System. Caroll and Cook, Albany, New York, 68 pp.
fedonkin, m. a. 1977. Precambrian-Cambrian ichnocoenoses of the East European Platform. Geological
Journal Special Issue, 9, 183-194.

1980. Fossil traces of Precambrian Metazoa. Izvestia Akademiya nauk SSSR, Geology Series, 1, 39 — 46 .
[In Russian].

1981. White Sea biota of Vendian. Trudy Geologiduskiy Institut. Akademiya nauk SSSR , 342, 1-100.

[In Russian],

— 1985a. Soft-bodied Vendian fauna : promorphological analysis. 112-116. In sokolov, b. s. and Ivanovskiy,
m. a. (eds). The Vendian System: historic-geological and paleontological basis, 1, Nauka, Moscow. [In
Russian],

— 19857>. Systematic descriptions of Vendian metazoa. 70-107. In sokolov, b. s. and Ivanovskiy, m. a. (eds).
The Vendian System: historic-geological and paleontological basis, 1. Nauka, Moscow. [In Russian].

1985c. Paleoichnology of Vendian metazoa. 112-116. In sokolov, b. s. and Ivanovskiy, m. a. (eds). The
Vendian System: historic-geological and palaeontological basis, 1. Nauka, Moscow. [In Russian].
fenton, c. l. and fenton, m. a. 1937. Burrows and trails from Pennsylvanian rocks of Texas. American
Midland Naturalist, 18, 1079-1084.

fillion, d. and PiCKERii L, r. k. 1984. Systematic ichnology of the Middle Ordovician Trenton Group, St.

Lawrence Lowland, eastern Canada. Maritime Sediments and Atlantic Geology, 20, 1-41.
fitch, a. 1850. A historical, topographical and agricultural survey of the County of Washington. New York
Agricultural Society Transactions, 9, 753-944.

ford, T. D. 1958. Precambrian fossils from Charnwood Forest. Proceedings of the Yorkshire Geological Society,
31, 211-217.

1963. The Pre-Cambrian fossils of Charnwood Forest, Leicestershire. Transactions of the Leicestershire
Literary and Philosophical Society, 57, 57-62.

1980. The Ediacarian fossils of Charnwood Forest, Leicestershire. Proceedings of the Geologists'
Association , 91, 81-83.

FOYN, s. and glaessner, m. e. 1979. Platysolenites, other animal fossils and the Precambrian-Cambrian
transition in Norway. Norsk Geologisk Tidsskrift, 59, 25-46.
frey, r. w. and seilacher, a. 1980. Uniformity in marine invertebrate ichnology. Lethaia , 13, 183-207.
fritz, w. h. 1980. International Precambrian-Cambrian Boundary Working Group’s 1979 field study to
Mackenzie Mountains, Northwest Territories, Canada. Geological Survey of Canada, Paper, 80-1 A, 41-45.

-and crimes, t. p. 1985. Lithology, trace fossils, and correlation of Precambrian-Cambrian boundary
beds, Cassiar Mountains, north-central British Columbia. Geological Survey of Canada, Paper, 83-14, 1-24.

— narbonne, g. m. and gordey, s. p. 1983. Strata and trace fossils near the Precambrian-Cambrian
boundary, Mackenzie, Selwyn and Wernecke Mountains, Yukon and Northwest Territories. Geological
Survey of Canada, Paper, 83-1 B, 365-375.

gehling, j. G. 1988. A cnidarian of actinian-grade from the Ediacaran Pound Subgroup, South Australia.
Alcheringa. 12, 299-314.

germs, g. j. b. 1972a. The stratigraphy and paleontology of the Lower Nama Group, South West Africa.
Chamber of Mines Precambrian Research Unit, University of Cape Town, Department of Geology Bulletin, 12,
1-250.

— \912b. Trace fossils from the Nama Group, south-west Africa. Journal of Paleontology, 46, 864-870.
gibson, G. G. 1989. Trace fossils from Late Precambrian Carolina slate belt, south-central North Carolina.

Journal of Paleontology, 63, 1-10.

tetter, s. a. and fedonkin, m. a. 1984. Ediacarian fossils from the Carolina Slate Belt, Stanly County,

North Carolina. Geology , 12, 387-390.

NARBONNE AND AITKEN: CANADIAN EDIACARAN FAUNA

977

glaessner, m. f. 1969. Trace fossils from the Precambrian and basal Cambrian. Lethaia , 2, 369-393.

1979. Precambrian. A79-A1I8. In robison, r. a. and teichert, c. (eds). Treatise on invertebrate
paleontology , Part A. Geological Society of America and University of Kansas Press, Boulder, Colorado and
Lawrence, Kansas, 569 pp.

1984. The dawn of animal life. A biohistorical study. Cambridge University Press, Cambridge, 244 pp.

— and wade, m. 1966. The Late Precambrian fossils from Ediacara, South Australia. Palaeontology , 9,
599-628.

gureev, Y. a. 1987. Morphological analysis and systematics of Vendiata. Akademiya Nauk Ukrainian SSR
Institute of Geological Science , Preprint 87-15, 1-54. [In Russian],
gurich, G. 1933. Die Kuibis-Fossilien der Nama Formation von Sudwestafrika. Paldontogische Zeitschrift , 15,
137-154.

hakes, w. G. 1976. Trace fossils and depositional environment of four clastic units, Upper Pennsylvanian
megacyclothems, northeast Kansas. University of Kansas Pale ontological Contributions , 63, 1 —46.

1977. Trace fossils in Late Pennsylvanian cyclothems, Kansas. Geological Journal Special Issue , 9,
209-226.

HALL, j. 1847. Palaeontology of New York , vol. I. C. Van Benthuysen, Albany, 338 pp.
hantzschel, w. 1975. Trace fossils and problematica. W1-W269. In teichert, c. (ed. ). Treatise on
invertebrate paleontology. Part IV, Supplement 1. Geological Society of America and University of Kansas
Press, Boulder, Colorado and Lawrence, Kansas, 269 pp.

Harrington, h. j. and moore, r. c. 1956. Scyphomedusae ; Trachylmidae; Medusae incertae sedis and
unrecognizable forms. F38-F53, F68-F76, F153-F161. In MOORE, r. c. (ed.). Treatise on invertebrate
paleontology. Part F. Geological Society of America and University of Kansas Press, Boulder, Colorado
and Lawrence, Kansas, 498 pp.

heer, o. 1865. Die Urwelt der Schweiz. F. Schulthess, Zurich, 622 pp.

1877. Flora Fossilis Helveticae. Die vorweltliche Flora der Schweiz. J. Wiirster and Co., Zurich, 182 pp.
hill, G. w. 1981. Ichnocoenoses of a Paleocene submarine-canyon floor. In frizzell, v. (ed.). Upper Cretaceous
and Paleocene turbidites, central California coast. Society of Economic Paleontologists and Mineralogists ,
Field Guide Book , 6. 93-104.

hofmann, H. j. 1981. First record of a Late Proterozoic faunal assemblage in the North American Cordillera.
Lethaia , 14, 303-310.

1987. Precambrian biostratigraphy. Geoscience Canada , 14. 135-154.

fritz, w. h. and narbonne, g. m. 1983. Ediacaran (Precambrian) fossils from the Wernecke Mountains,

northwestern Canada. Science , 221, 455-457.

mountjoy, e. w. and teitz, m. w. 1985. Ediacaran fossils trom the Miette Group, Rocky Mountains,

British Columbia, Canada. Geology , 13, 819-821.

— and patel, i. m. 1989. Trace fossils from the type ‘Etcheminian Series’ (Lower Cambrian Ratcliffe Brook
Formation), Saint John area. New Brunswick, Canada. Geological Magazine , 126, 139-157.

Howard, J. d. and frey, r. w. 1984. Characteristic trace fossils in nearshore to offshore sequences. Upper
Cretaceous of east-central Utah. Canadian Journal of Earth Sciences , 21, 200-219.
james, u. p. 1879. Description of new species of fossils and remarks on some others from the Lower and LJpper
Silurian rocks of Ohio. The Paleontologist , 3, 17-24.

jefferson, G. w. and ruelle, j. c. l. 1986. The Late Proterozoic Redstone Copper Belt, Mackenzie Mountains,
Northwest Territories. 154-168. In morin, j. s. (ed). Mineral deposits of the Northern Cordillera. Special
Volume of the Canadian Institute of Mining and Metallurgy No. 37, 377 pp.
jenkins, r. J. f. 1981. The concept of an Ediacaran Period and its stratigraphic significance in Australia.
Transactions of the Royal Society of South Australia, 105, 179-194.

1988. Ediacaran succession and fossil occurrences. 1 17. In jenkins, r. j. f. and gravestock, d. i.
(compilers). Proterozoic Ediacara fauna and Cambrian Archaeocyatha of the Flinders Ranges, South
Australia. Fifth International Symposium on Fossil Cnidaria , Handbook for Excursion A2, 49 pp.

1989. The ‘supposed terminal Precambrian extinction event' in relation to the Cnidaria. Memoirs of the
Association of Australasian Palaeontologists, 8, in press.

and gehling, j. g. 1978. A review of the frond-like fossils of the Ediacara assemblage. Records of the

South Australia Museum, 17, 347-359.

ksiazkiewicz, m. 1970. Observations on the ichnofauna ot the Polish Carpathians. Geological Journal Special
Issue, 3, 283-322.

1977. Trace fossils in the flysch of the Polish Carpathians. Palaeontologica Polonica, 36, 1-208.
lowe, d. r. 1982. Sediment gravity flows. II. Depositional models with special reference to the deposits of high

978

PALAEONTOLOGY, VOLUME 33

density turbidity currents. Journal of Sedimentary Petrology , 52, 279-298.
maples, C. G. and west, r. r. 1989. Lockeia , not Pelecypodichnus. Journal of Paleontology , 63, 694-696.
McMENAMiN, m. a. s. 1986. The garden of Ediacara. Palaios , 1, 178-183.

— and McMENAMiN, D. L. s. 1990. The emergence of animals. The Cambrian breakthrough. Columbia
University Press, New York, 217 pp.

mutti, e. and normark, w. r. 1987. Comparing examples of modern and ancient turbidite systems. Problems
and concepts. 1-38. In leggett, j. k. and zuffa, g. g. (eds). Marine clastic sedimentology , Concepts and case
studies. Graham and Trotman, London, 221 pp.

narbonne, g. m. and hofmann, h. j. 1987. Ediacaran biota of the Wernecke Mountains, Yukon, Canada.
Palaeontology , 30, 647-676.

— and aitken, j. d. 1985. Precambrian-Cambrian boundary sequence, Wernecke Mountains, Yukon
Territory. Geological Survey of Canada , Paper , 85-1 A, 603-608.

— and myrow, p. 1988. Trace fossil biostratigraphy in the Precambrian-Cambrian boundary interval. New
York State Museum Bulletin, 463, 12-16.

landing, e. and anderson, m. m. 1987. A candidate stratotype for the Precambrian-Cambrian
boundary. Fortune Head, Burin Peninsula, southeastern Newfoundland. Canadian Journal of Earth
Sciences, 24, 1277-1293.

— 1990. Kullingia delicata (Fedonkin): A chondrophorian that crosses the Precambrian-

Cambrian boundary. Journal of Paleontology, 64, in press.
nicholson, h. a. 1873. Contributions to the study of the errant annclides of the older Palaeozoic rocks.
Proceedings of the Royal Society, 21, 288-290.

— and hinde, G. J. 1875. Notes on the fossils of the Clinton, Niagara and Guelph Formations of Ontario,
with descriptions of new species. Canadian Journal of Science, Literature and History, New Series 2, 14,
137-160.

norris, R. d. 1989. Cnidarian taphonomy and the affinities of the Ediacara biota. Lethaia, 22, 381-393.
nowlan, g. s., narbonne, g. m. and fritz, w. h. 1985. Small shelly fossils and trace fossils near the
Precambrian-Cambrian boundary in the Yukon Territory, Canada. Lethaia, 18, 233-256.

OSGOOD, R. G. 1970. Trace fossils of the Cincinnati area. Palaeontographica Americana, 6, 279-444.
paczesna, j. 1985. Upper Vendian and Lower Cambrian trace fossils of southern Lublin district. Kwartalnik
Geologiczny , 29, 255-270.

1986. Upper Vendian and Lower Cambrian ichnocoenoses of Lublin region. Biuletyn Instytutu
Geologicznego, 355, 31^17.

— 1988. Upper Vendian and Lower Cambrian trace fossils from southeastern Poland. New York State
Museum Bulletin, 463, 16.

palij, v. m. 1976. Remains of soft-bodied animals and trace fossils from the Upper Precambrian and Lower
Cambrian of Podolia. 63-76. In Paleontologiya i stratigrafiya verkhnego dokembriya i nizhnego kembriya
yugo-zapada Vostochno-Europeiskoi plalformy. Naukova dumka, Kiev. [In Ukrainian],

— posti, e. and fedonkin, m. a. 1979. Soft-bodied metazoa and animal trace fossils in the Vendian and Early
Cambrian. 56-93. In Upper Precambrian and Cambrian paleontology of the East European Platform, 49-82.
‘Nauka’, Moscow. [In Russian; English translation published in 1983 under joint sponsorship of the
Institute of Geological Sciences of the Polish Academy of Sciences and the Geological Institute of the
Academy of Sciences of the U.S.S.R. by Publishing House Wydawnictwa Geologiczne, Warsaw].

pemberton, s. G. and frey, r. w. 1982. Trace fossil nomenclature and the Planolites-Palaeophycus dilemma.
Journal of Paleontology, 56, 843-881.

PFLUG, h. d. 1970. Zur Fauna der Nama-Schichten in Sudwest-Afrika. I. Pteridinia, Bau und systematische
Zugehorigkeit. Palaeontographica, Abteilung A, 134, 153-262.

1972. Systematik der jung-prakambrischen Petalonamae. Paldontologische Zeitschrift, 46, 56-67.
pickerill, r. k. 1980. Phanerozoic trace fossil diversity - observations based on an Ordovician flysch
ichnofauna from the Aroostook-Matapedia Carbonate Belt of northern New Brunswick. Canadian Journal
of Earth Sciences, 17, 1259-1270.

— 1981. Trace fossils in a Lower Paleozoic submarine canyon sequence - the Siegas Formation of
northwestern New Brunswick, Canada. Maritime Sediments and Atlantic Geology, 17, 36-58.

pojeta, j. 1985. Early evolutionary history of diasome molluscs. 102-121. In broadhead, t. w. (ed.). Mollusks.
Notes for a short course. University of Tennessee, Department of Geological Sciences, Studies in Geology,
13, 305 pp.

1987. Class Pelecypoda. 386-435. In boardman, r. s., cheetham, a. h. and rowell, a. j. (eds). Fossil
invertebrates. Blackwell Scientific Publications, Oxford, 713 pp.

NARBONNE AND AITKEN: CANADIAN EDIACARAN FAUNA

979

runnegar, b. and kriz, j. 1973. Fordilla troyensis Barrande: the oldest known pelecypod. Science , 180,

886-888.

raup, D. m. and seilacher, a. 1968. Fossil foraging behavior: computer simulation. Science , 166, 994-995.
994—995.

richter, r. 1928. Psychische Reaktionen fossiler Tiere. Palaeobiologica , 1, 225-244.

1937. Marken und Spuren aus alien Zeiten. I- 1 1 Senckenbergiana , 19, 150-169.

1955. Die altesten Fossilien Siid-Afrikas. Senckenbergiana Letliaea , 36, 243-289.

runnegar, b. 1985. Origin and early history of mollusks. 17-32. In broadhead, t. w. (ed.). Mollusks. Notes
for a short course. University of Tennessee, Department of Geological Sciences, Studies in Geology, 13, 305

pp.

schafhautl, K. E. 1851. Geognostische Untersuchungen des sidbayrischen Alpengebirges. Literarisch-artistische
Anstalt, Mtichen, 208 pp.

seilacher, a. 1953. Studien zur Palichnologie. IF Die fossilen Ruhespuren (Cubichnia). Neues Jahrbuch fur
Geologie und Paldontologie , Abhandlungen, 98, 87-124.

1960. Lebensspuren als Leitfossilien. Geologische Rundschau , 49, 41-50.

— 1964. Biogenic sedimentary structures. 296-316. In imbrie, j. and Newell, n. d. (eds). Approaches to
paleoecology. Wiley, New York, 432 pp.

1967. Bathymetry of trace fossils. Marine Geology , 5, 413^128.

— 1974. Flysch trace fossils: evolution of behavioural diversity in the deep sea. Neues Jahrbuch fur Geologie
und Paldontologie , Monatshefte , 1974, 233-245.

1977. Evolution of trace fossil communities. 359-376. In hallam, a. (ed.). Patterns of evolution, as
illustrated by the fossil record. Developments in Palaeontology and Stratigraphy, 5, 1-591.

1984. Late Precambrian and early Cambrian metazoa: preservational or real extinctions. 159-168. In
Holland, H. D. and trendall, a. F. (eds). Patterns of change in earth evolution , Dahlem Konferenzen,
Springer- Verlag, Berlin, 431 pp.

1989. Vendozoa: organismic construction in the Proterozoic biosphere. Lethaia , 22, 229-239.
shanmugam, G. and moiola, r. j. 1988. Submarine fans. Characteristics, models, classification and reservoir
potential. Earth Science Reviews , 24, 383-428.

sokolov, b. s. and fedonkin, m. a. 1984. The Vendian as the terminal system of the Precambrian. Episodes, 7,
12-19.

— and Ivanovskiy, m. a. (eds). 1985. The Vendian System : historic-geological and palaeontological basis.
Nauka, Moscow 1, 221 pp.; 2, 237 pp.

sprigg, r. c. 1947. Early Cambrian (?) jellyfishes from the Flinders Ranges, South Australia. Transactions of
the Royal Society of South Australia , 71, 212-224.

1949. Early Cambrian jellyfishes of Ediacara, South Australia and Mount John, Kimberley District,

Western Australia. Transactions of the Royal Society of South Australia , 73, 72-99.

Stanley, G. d. 1986. Chondrophorine hydrozoans as problematic fossils. 68-86. In hoffman, a. and nitecki,
m. h. (eds). Problematic fossil ta.xa. Oxford University Press, Oxford, 267 pp.
sun, w. G. 1986. Precambrian medusoids: The Cyclomedusa plexus and Cyclomedusa- like pseudofossils.
Precambrian Research, 31, 325-360.

thomas, R. e. and berg, t. M. 1985. Interpretation of bivalve trace fossils in fluvial beds of the basal Catskill
Formation (Late Devonian), eastern U.S.A. 13-20. In curran, h. a. (ed ). Biogenic structures: their use in
interpreting depositional environments. Society of Economic Paleontologists and Mineralogists. Special
Publication, 35, 347 pp.

vyalov, o. s. 1971. Rare Mesozoic problematica from Pamir and the Caucasus. Paleontologicheskiy Sbornik,
7, 85-93. [In Russian],

wade, m. 1972. Hydrozoa and Scyphozoa and other medusoids from the Precambrian Ediacara fauna, South
Australia. Palaeontology, 15, 197-225.

water, m., elphinstone, r. and heys, G. R. 1990. Proterozoic and Early Cambrian trace fossils from the
Amadeus and Georgina Basins, Central Australia. Alcheringa, in press.
webby, b. d. 1970. Late Precambrian trace fossils from New South Wales. Lethaia, 3, 79-109.

1984. Precambrian-Cambrian trace fossils from western New South Wales. Australian Journal of Earth
Sciences , 31, 427 — 437.

wright, a. d. and benton, m. j. 1987. Trace fossils from Rhaetic shore-face deposits of Staffordshire.
Palaeontology, 30, 407 — 428.

yeo, g. m. 1981. The Late Proterozoic Rapitan glaciation in the northern Cordillera. 25-46. In Campbell,
f. h. a. (ed.). Proterozoic Basins of Canada. Geological Survey of Canada, Paper, 81-10, 1^443.

980

PALAEONTOLOGY, VOLUME 33

yochelson, e. L. and schindel, d. e. 1978. A re-examination of the Pennsylvanian trace fossil Olivellites.

Journal of Research , United States Geological Survey, 6, 789-796.
young, G. m. 1976. Iron-formation and glaciogenic rocks of the Rapitan Group, Northwest Territories,
Canada. Precambrian Research , 3, 137-158.

GUY M. NARBONNE

Department of Geological Sciences
Queen’s University
Kingston, Ontario
Canada K7L 3N6

JAMES D. AITKEN

Typescript received 15 August 1989
Revised typescript received 19 January 1990

Geological Survey of Canada
3303 33rd St N.W.
Calgary, Alberta
Canada T2L 2A7

A NEW SPECIES OF LATE CRETACEOUS WOOD-
BORING BIVALVE FROM NEW ZEALAND

by J. S. CRAMPTON

Abstract. A new species of late Campanian or Maastrichtian (Late Cretaceous) pholadid bivalve, Pholadidea
( Hatasia ) wiffenae , is described from shallow marine transgressive Maungataniwha Sandstone, northwestern
Hawkes Bay, New Zealand. It represents the earliest confirmed record for both genus and subgenus. The new
species is known from exceptionally well-preserved material, which permits detailed knowledge of external
morphology, including all accessory plates, and to a lesser degree, internal morphology. It is assigned to
Pholadidea ( Hatasia ) based largely on the nature of accessory plates. In life-habit and some details of
morphology, however. P. (H.) wiffenae resembles Opertochasma and Martesia , and may be an evolutionary
intermediate between Cretaceous Opertochasma and early Tertiary Pholadidea and Martesia. The present
record highlights the need for a comprehensive systematic review of Mesozoic pholadids.

The late Cretaceous Maungataniwha Sandstone (Moore 1986), outcropping west and southwest of
Lake Waikaremoana, northwestern Hawkes Bay, New Zealand (Text-fig. 1), contains the most
diverse faunas of this age known from New Zealand. Included are terrestrial and freshwater
invertebrates (Craw and Watt 1987; Moore et al. 1988), marine invertebrates (Glaessner 1980;
Wiffen 1980; Crampton 1988; Moore et al. 1988; Crampton and Moore 1990), and terrestrial
and marine vertebrates (Keyes 1977; Wiflen 1980, 1981, 1983, 1986; Molnar 1981; Scarlett and
Molnar 1984; Wiffen and Moisley 1986; Wiffen and Molnar 1988).

Of special note are some exceptionally well-preserved wood-boring pholadid bivalves. A recent
discussion of Antarctic Cretaceous pholadids (Kelly 1988) noted the poorly documented record of
these bivalves in the Southern Hemisphere, and the present paper describes the first Cretaceous
species known from New Zealand.

The boring bivalves were collected from a single float concretion in Mangahouanga Stream, a
tributary of Te Hoe River. Fossiliferous calcareous concretions are concentrated in Mangahouanga
Stream, and are derived from fine sandstone of the late Campanian-Maastrichtian Maungataniwha
Sandstone which crops out in both banks of the stream. Bored fossil wood is abundant in the
concretions, but to date only the single concretion has yielded shells of the bivalves responsible for
the borings. The borings themselves have been identified as ichnospecies Teredolites clavatus
Leymerie (PI. 3, figs 6-7; see also Kelly and Bromley 1984).

Study of the fauna and sediments in the region of Te Hoe River indicate that the Maungataniwha
Sandstone was deposited during a regional marine transgression in a shallow marine nearshore
environment on an embayed coastline and in close proximity to a river mouth (Crampton and
Moore 1990). Lower Maungataniwha Sandstone includes minor lagoonal sediments, but was
deposited mainly on the shoreface or foreshore above fair-weather wave-base. The bulk of this
formation was deposited on the lower shoreface to offshore-transition, between fair-weather and
storm wave-bases (Crampton and Moore 1990). The stratigraphical position of the fossils
described herein is unknown. Fossilized wood occurs as discrete logs and branches which may have
borings that enter from all directions, indicating that the wood is probably the same age as the
sediment, and was not derived from older, already lithified, lignite beds (see later discussion). The
bivalves were collected from araucariacean wood (J. I. Raine, pers. comm. 1989).

Morphological terminology used in this description follows that of Turner (1969) and Kennedy
(1974), as modified by Kelly (1988). Major features of the shell exterior of Pholadidea (Hatasia)

|Palaeontology, Vol. 33, Part 4, 1990, pp. 981-992, 3 pis. |

© The Palaeontological Association

982

PALAEONTOLOGY, VOLUME 33

text-fig 1. Distribution of Maungataniwha Sandstone (late Campanian-Maastrichtian : shaded) in
northwestern Hawkes Bay, New Zealand, showing locality of the fossils described herein (geology after

Grindley 1960; Moore 1986; Moore et al. 1988).

EXPLANATION OF PLATE 1

Figs 1-6. Pholadidea ( Hatasia ) wiffenae sp. nov., V19/f6909u, Mangahouanga Stream, northwestern Hawkes
Bay, New Zealand (late Campanian-Maastrichtian). All SEM photographs. Note that apparent variations
in dimensions are due to orientation of specimen and parallax distortion of the SEM. 1^, TM 6942
(holotype), adult; 1, lateral aspect, left valve, x 51 ; 2, anterior aspect, valves in occlusion, x4-6; 3, ventral
aspect, valves in occlusion, x 51 ; 4, dorsal aspect, valves in occlusion, x 51. 5, TM 6940, mesoplax showing
periostracal folds, anterior up, x 8-7. 6, TM 6951, mesoplax, anterior up, x 8-2.

PLATE 1

CRAMPTON, Pholadideci ( Hcitasia )

984

PALAEONTOLOGY, VOLUME 33

wiffenae sp. nov. are illustrated in Text-figure 2. In the following discussion, the bivalves are
described as juvenile if the anterior gape is entirely open, immature if the gape is partially closed
by callum, and adult if the gape is entirely closed by callum.

Two specimens of the bivalve were transversely serial sectioned (following the recommendation
of Kelly 1988) at 500 pm intervals using a Leitz annular saw. These sections were ground to a
thickness of 50 pm for examination. In the text individual thin sections are referred to by distance
(in millimetres) from the anterior margin of the shell.

Dorsal

Anterior

Mesoplax

sulcus

Posterior

Ventral

text-fig. 2. External morphology of Martesiinae, after Turner (1969), Kennedy (1974), and Kelly (1988).

SYSTEMATIC PALAEONTOLOGY

Family pholadidae Lamarck, 1809
Subfamily martesiinae Grant and Gale, 1931
Genus pholadidea Turton, 1819

Type species. Pholadidea loscombiana Turton, 1819 (original designation).

Subgenus hatasia Gray, 1851

Type species. Pholas melanura Sowerby, 1834 (subsequent designation, Stoliczka 1870).

Discussion. The genus Pholadidea is distinguished from other genera in Martesiinae by its single
umbonal-ventral sulcus, complete closure of the anterior gape in the adult by a callum which is
extended dorsally to cover the beaks, the presence of a mesoplax which is longitudinally divided at
some growth stage, and the presence or absence of a metaplax and a hypoplax which, it present, are
not separate plates, but result from the deposition of calcite in the periostracum uniting the valves
posterior to the umbos (see Turner 1969, p. 716). Two subgenera have been described; P. ( Hatasia )

CRAMPTON: CRETACEOUS WOOD-BORING BIVALVE

985

is distinguished by a relatively closely appressed umbonal reflection, a comparatively large flat
mesoplax in the juvenile overlain dorsally in the adult by a longitudinally divided plate, and a
variable siphonoplax composed largely of periostracum (Turner 1969, p. 716). Most Pholadidea are
shale, soft rock, and coral borers (Turner 1969, see discussion below).

Turner (1969) stated that Pholadidea is known from the Eocene-Recent, and the subgenus P.

( Hatasia ) from the Recent only. In his review of Mesozoic pholadids, Kelly (1988) did not list any
Cretaceous or older records of Pholadidea. Stephenson (1923, 1941) referred a number of Late
Cretaceous species from North America to the genus, but his species are known from incomplete
specimens or require reinterpretation in the light of the family taxonomy described by Turner
(1969). Campbell et al. (in press) list Pholadidea n. sp., based on well-preserved material, from
Palaeocene strata on the Chatham Islands, New Zealand. Re-examination of their material,
however, suggests that this species may belong in Jouannetia ( Pholadopsis ) Conrad. P. (//.) wiffenae
sp. nov., therefore, apparently represents the first fossil record for the subgenus, and probably one
of the oldest records for the genus.

The genus Pholadidea closely resembles both Martesia Sowerby and Opertochasma Stephenson.
Martesia are wood-boring pholadids distinguished from Pholadidea by an undivided mesoplax, the
absence of a dorsal extension of the callum, and by a metaplax and a hypoplax which are separate
calcified plates (Turner 1969; Kennedy 1974). Many Cretaceous pholadids formerly assigned to
Martesia have subsequently been referred to Opertochasma , and the earliest confirmed record of
Martesia is Palaeocene (Speden 1970; Kennedy 1974; Kelly 1988). Opertochasma are also wood-
borers distinguished by the presence of two umbonal-ventral sulci, periostracal flaps on the
posterior slope, and incomplete closure of the anterior gape in the adult (Turner 1969; Speden 1970;
Kennedy 1974; Kelly 1988). As noted by Kelly (1988), the second umbonal-ventral sulcus shows
considerable intra-population variation in strength, and some specimens may be difficult to
distinguish from both Martesia and Pholadidea (for example, see Speden 1970, pi. 38, figs 2, 4). The
first record of Opertochasma is in the late Jurassic (Kelly 1988).

Pholadidea ( Hatasia ) wiffenae sp. nov. is referred to the Martesiinae on the basis of its gross
morphology, in particular the complete closure of the anterior gape by a calcareous callum in the
adult, and by the absence of protoplax. It is referred to Pholadidea because of its dorsally-extended
callum, longitudinally divided adult mesoplax, and weakly calcified metaplex and hypoplax within
the periostracum uniting valves. The closely appressed umbonal reflection and nature of the
mesoplax are consistent with placement in P. [Hatasia).

One significant difference between P. (H.) wiffenae and typical Pholadidea is its wood-boring life-
habit. This character, and the well-defined prora (see Turner 1969) suggest affinities with Martesia ,
whereas the life-habit and comparatively narrow commarginal ridges suggest affinities with
Opertochasma. P. (H.) wiffenae , therefore may represent an intermediate between Cretaceous
Opertochasma , and early Tertiary Pholadidea and Martesia. It should be stressed, however, that
morphological similarities will be the result, in part at least, simply of convergent life habits, which
may or may not indicate evolutionary relationships. The influence of substrate type, for example,
on shell shape and ornament is outlined in Kennedy (1974). Conclusions about evolutionary
relationships, therefore, must await a comprehensive revision of all Cretaceous pholadids, a task
hampered by the poor fossil record of this group and the typically incomplete nature of the fossils
and early descriptions.

Pholadidea ( Hatasia ) wiffenae sp. nov.

Plates 1-3; Text-figs 3 and 4

Name. Named after Mrs Joan Wiffen who discovered these bivalves, in recognition of her major contribution
to our knowledge of New Zealand late Cretaceous palaeontology.

Type specimens. Holotype: TM 6942, V19/f6909n, GS 14241 ; entire articulated phosphatized specimen with
all accessory plates preserved. Fifteen paratypes: TM 6938-6941, TM 6943-6953, all from V19/f6909u, GS

986

PALAEONTOLOGY, VOLUME 33

m.

text-fig. 3. Interpretive drawings of selected vertical transverse thin sections of Pholadidea ( Hatasia )
wiffenae sp. nov., specimen TM 6953. Sections drawn from photographs, all enlarged x 3-4. Individual
drawings oriented dorsal up and arranged with respect to distance from the anterior margin, given below in
millimetres. Note that interpretation of sections is hampered by the replacement of originally calcitic shell
material by collophane, and the presence of interstitial and cavity-filling collophane. a, I -5 mm. b, 2-0 mm. c,
2-5 mm. d, 3 0 mm. E, 3-5 mm. f, 4-0 mm. g, 8 0 mm. Legend: a = apophysis; as = anterior slope; c = callum;
ch = chondrophore; d = disk; h = hypoplax; m = mesoplax; mt = metaplax; p = prora; ur = unrbonal

reflection ; uvr = umbonal-ventral ridge.

14241 . All specimens phosphatized. TM 6952 and TM 6953 serial sections at 0 5 mm intervals through juvenile
and adult respectively.

Type locality. Mangahouanga Stream, float concretion 470 m east southeast of Te Hoe forestry road bridge
over stream and approximately 200 m downstream from first major true left tributary above bridge, north
western Hawkes Bay, New Zealand (see Text-fig 1). New Zealand national Fossil Record locality V19/f6909a,
grid reference NZMS 260 sheet V19 42134700. New Zealand Geological Survey collection GS 14241 and
Wiffen collection. Collected by J. and M. Wiffen, February 1987.

Description. Small, less than 20 mm long, equivalve, inequilateral. Subconical in lateral and vertical profiles,
tapering evenly to posterior; subcircular in transverse profile (PI. 1, figs 1-4; PI. 2, figs 1-4). Prora well-defined,
antero- ventral margin with sharp re-entrant (with sides at approximately 100°) resulting in Teredo-Wke anterior
(PI. 1, fig. 1 ; PI. 2, fig. I ; PI. 3, fig. 1 ). Posterior-dorsal margin gently convex, posterior margin bluntly rounded,
posterior-ventral margin straight to weakly concave. Umbo low, strongly incurved, prosogyrous, at
approximately anterior fifth of shell (PI. 2, fig. 2). Unrbonal reflections lower than umbones. Prominent
anterior conical cavity beneath unrbonal reflection (PI. 3, figs 1 and 3). Anterior pedal gape confined to ventral

EXPLANATION OF PLATE 2

Figs 1—4. Pholadidea ( Hatasia ) wiffenae sp. nov., V19/f6909n, Mangahouanga Stream, northwestern Hawkes
Bay, New Zealand (Campanian-Maastrichtian). All SEM photographs. Note that apparent variations in
dimensions are due to orientation of specimens and parallax distortion of the SEM. 1 and 3, TM 6941, adult;

1, lateral aspect, left valve, x4-9; 3, dorsal aspect, valves in occlusion, x4-9. 2 and 4, TM 6938, immature;

2, dorsal aspect, mesoplax removed, valves in occlusion, x 51 ; 4, ventral aspect, x 5T.

PLATE 2

CRAMPTON, Pholadidea ( Hatasia )

988

PALAEONTOLOGY, VOLUME 33

text-fig. 4. Reconstruction of Pholadidea (Hatasia) wiffenae sp. nov. Note that juvenile mesoplax (omitted
here) may underlie longitudinally divided adult mesoplax shown, and may extend posteriorly well beyond adult
mesoplax (see discussion in text). (Drawn by R. C. Brazier.)

half of height, subtriangular, dorsal margin of gape concave, almost perpendicular to commissure. Anterior
gape entirely closed in adults by a callum, the two halves of which are inferred to have been connected by
periostracum medially, are slightly raised above level of anterior slope, and extend dorsally over the umbonal
reflection (PI. 1, figs 2 and 3). A few prominent commarginal growth lines present on callum of all specimens
examined. Protoplax lacking. Mesoplax of two pear-shaped plates which cover umbones and become fused

EXPLANATION OF PLATE 3

Figs 1-7. Pholadidea (Hatasia) wiffenae sp. nov., V10/f6909a, Mangahouanga Stream, northwestern Hawkes
Bay, New Zealand (Campanian-Maastrichtian). All SEM images except Fig. 6. 1 and 3, TM 6950, juvenile,
anterior aspect of prora and umbonal reflection, dorsal up; 1, x 10-5; 3, x34-7. 2, TM 6940, detail of
callum at commissure, showing texture resulting from incorporation of clastic grains into periostracum,
x 22-4. 4, TM 6941, denticulate ornament on anterior slope immediately adjacent to umbonal-ventral sulcus,
dorsal up, anterior to left, x4F6. 5, TM 6949, interior of juvenile, showing apophysis and umbonal-ventral
ridge, anterior to right, x 191. 6, TM 6942 (holotype) in life position in araucariacean wood, x 1-7. 7, TM
6954, iclmospecies Teredolites clavatus Leymerie, resulting from boring by Pholadidea (Hatasia) wiffenae sp.
nov., aperture to right, x 3-6.

PLATE 3

CRAMPTON, Pholadidea ( Hatasia )

990

PALAEONTOLOGY, VOLUME 33

posteriorly (PI. 1, figs 2, 4-6; PL 2, fig. 3). Mesoplax closely associated with, and possibly partially fused to,
dorsal projections of the callum. The juvenile portion of the mesoplax characteristic of subgenus Hatasia (see
above), is not clearly visible on any specimens, but may be represented by the (low flange surrounding and
possibly underlying the longitudinally divided portion of the mesoplax (PI. 1, figs 4-6). In thin section,
specimen TM 6953, interpreted as an adult on the basis of its complete callum, displays a single large flat plate
which extends 3-O^TO mm behind the umbo, and which is interpreted as the juvenile mesoplax (Text-fig. 3).
This specimen apparently lacks the divided mesoplax observed on other individuals. Metaplax and hypoplax
comprising anteriorly tapering, partially calcified periostracum, uniting valves across dorsal and ventral gapes
respectively (PI. 1, figs 3 and 4; PI. 2, figs 2^4). Umbonal-ventral sulcus well-defined, narrow. Matched on shell
interior by narrow upstanding umbonal-ventral ridge (PI. 3, fig. 5; Text-fig. 3). Sculpture on prora and anterior
slope of between 50 and 70 sharply-defined, narrow, rounded commarginal ridges and broader, flat-floored
troughs. Ridges finely denticulate on anterior slope (PI. 3, fig. 4). Ridges broader and weaker on disc, fading
out over posterior slope of many specimens. Periostracum present over entire shell, thicker on disc and
posterior slope than elsewhere. In some specimens clastic grains incorporated into periostracum over callum
(PI. 3, fig. 2) and, more commonly, over posterior slope. Siphonoplax short, composed of periostracum or
partly calcified periostracum (PI. 2, fig. 1 ; PI. 3, fig. 6). Apophysis relatively short, rod-like, not markedly
spatulate at distal end (PI. 3, fig. 5; Text-fig. 3).

Dimensions (undistorted specimens, mm).

Specimen

TM 6938 (immature)

TM 6941 (adult)

TM 6942 (adult, holotype)
TM 6943 (adult)

TM 6944 (adult)

TM 6945 (?)

TM 6946 (adult)

TM 6947 (juvenile)

TM 6948 (adult)

TM 6953 (adult)

Length

(excluding

siphonoplax)

Height

Width

(both valves)

17-6

10-8

110

17-1

9-7

10-4

17 6

1 1-3

1 1-3

17-2 +

10-7

10 6

?

8-9

8-8

?

c. 110

111

15-5

90

8-9

14-7 +

9-8

9-7

18-9

101

110

c. 12-6

8-1

c. 8-0

Discussion. A number of late Cretaceous pholadids have been described from the Southern
Hemisphere, but incomplete knowledge of these, in most cases, prevents detailed comparison with
the present species. Martesia leali Stinnesbeck, 1986 (pp. 185-186, pi. 5, fig. 9) from the
Maastrichtian of Chile, resembles P. (H.) wiffenae in gross form and life habit, but possesses a
coarser ornament (although, as mentioned above, the nature of the substrate can have a marked
effect on ornament, see Kennedy 1974). The nature of the accessory plates in M. leali is unknown.
Martesia cazadoriana Wilckens, 1907 (pp. 10-1 1, pi. 8, fig. 1 1), from the Campanian-Maastrichtian
(Riccardi 1988) of Patagonia, is known only from a single internal mould; and Martesial sp. of
Rennie (1930, pp. 205-206, pi. 23, figs 13 and 14) from the late Cretaceous of South Africa, is known
from a single incomplete and distorted specimen, and probably does not belong in Martesiinae
(Kelly 1988). Many late Cretaceous pholadids have been described from the Northern Hemisphere,
and although superficially some of these resemble P. (H.) wiffenae , detailed comparisons should be
made within the framework of a comprehensive systematic review of the family.

Acknowledgements. I would like to thank Joan and Pont Wiffen who found and extracted these bivalves, and
donated them to the NZ Geological Survey. Skilled technical assistance was given by J. Simes (SEM
photography), W. St George (light photography), N. Orr (thin sectioning), R. C. Brazier (drawings), and S.
Nepe (typing). Earlier drafts of this paper were reviewed and greatly improved by Drs J. G. Begg (New Zealand
Geological Survey, Lower Hutt), S. R. A. Kelly (British Antarctic Survey, Cambridge), P. A. Maxwell (New
Zealand Geological Survey, Lower Hutt), N. J. Morris (British Museum of Natural History), and R. D.
Turner (Museum of Comparative Zoology, Harvard University).

CRAMPTON: CRETACEOUS WOOD-BORING BIVALVE

991

REFERENCES

CAMPBELL, H. J., ANDREWS, P. B., BEU, A. G., EDWARDS, A. R., HORNIBROOK, N. DE B., LAIRD, M. G., MAXWELL, P.

a., mildenhall, d. c. and watters, w. a. in press. Cretaceous-Cenozoic geology of the Chatham Islands,
New Zealand. New Zealand Geological Survey Bulletin.
crampton, j. s. 1988. A Late Cretaceous near-shore rocky substrate macrofauna from northern Hawkes Bay,
New Zealand. New Zealand Geological Survey Record , 35, 21-24.

- and moore, p. R. 1990. Environment of deposition of the Maungataniwha Sandstone (Late
Cretaceous), Te Hoe River area, western Hawkes Bay, New Zealand. New Zealand Journal of Geology and
Geophysics , 33, 333-348.

craw, r. c. and watt, j. c. 1987. An Upper Cretaceous beetle (Coleoptera) from Hawkes Bay, New Zealand.

Journal of the Royal Society of New Zealand , 17, 395-398.
glaessner, m. f. 1980. New Cretaceous and Tertiary crabs (Crustacea: Brachyura) from Australia and New
Zealand. Transactions of the Royal Society of South Australia , 104, 171-192.
grant, v. s. and gale, h. r. 1931. Catalogue of the marine Pliocene and Pleistocene Mollusca of California.

San Diego Society of Natural History Memoir , 1, 1-1036.
gray, J. E. 1851. An attempt to arrange the species of the Pholadidae into natural groups. Annals and Magazine
of Natural History , (2), 8, 380-386.

grindley, G. w. 1960. Sheet 8 Taupo. Geological map of New Zealand 1 : 250000. DSIR. Wellington.
kelly, s. r. a. 1988. Cretaceous wood-boring bivalves from western Antarctica with a review of the Mesozoic
Pholadidae. Palaeontology , 31, 341-372.

and bromley, r. G. 1984. Ichnological nomenclature of clavate borings. Palaeontology , 27, 793-807.

Kennedy, g. l. 1974. West American Cenozoic Pholadidae (Mollusca: Bivalvia). San Diego Society of Natural
History Memoir, 8, 1-127.

keyes, i. w. 1977. Records of the northern hemisphere Cretaceous sawfish genus Onchopristis (order Batoidea)
from New Zealand. New Zealand Journal of Geology and Geophysics , 20, 263-272.
lamarck, J. b. p. a. DE M. de 1809. Philosophic zoologique , ou exposition des considerations relatives a Vhistoire
naturelle des animaux , la diversite de leur organisation et des facultes qu'ils en obtiennent , aux causes physiques
qui maintiennent en eux la vie , et donnent lieu aux mouvements qu’ils executent ; enfin a cedes qui produisent
les unes les sentiments , et les autres V intelligence de ceux qui en sont doues. Paris, 2 volumes, 422 pp., 463 pp.
molnar, r. e. 1981. A dinosaur from New Zealand. 91-96. In cresswell, m. m. and vella, p. (eds).
Gondwana Five. Fifth International Gondwana Symposium , Wellington , New Zealand , February 1980. A. A.
Balkema, Rotterdam, x + 339 pp.

moore, p. r. 1986. Stratigraphy and structure of the Te Hoe-Waiau River area, western Hawkes Bay. New
Zealand Geological Survey Record , 18, 4-12.

- Isaac, m. j., crampton, j. s. and mazengarb, c. 1988. Late Cretaceous sediments west of Lake
Waikaremoana, Urewera National Park. New Zealand Geological Survey Record, 35, 55-60.

rennie, m. a. 1930. New Lamellibranchia and Gastropoda from the Upper Cretaceous of Pondoland (with an
appendix on some species from the Cretaceous of Zululand). Annals of the South African Museum, 28,
159-260.

riccardi, a. c. 1988. The Cretaceous system of southern South America. Geological Society of America
Memoir, 168. 1-161.

scarlett, r. j. and molnar, r. e. 1984. Terrestrial bird or dinosaur phalanx from the New Zealand Cretaceous.
New Zealand Journal of Zoology, 11, 271-275.

sowfrby, G. b. 1834. Characters of new genera and species of Mollusca and conchifera, collected by Mr
Cuming. Proceedings of the Zoological Society of London, 2 (19), 68-72.
speden, i. G. 1970. The type Fox Hills Formation, Cretaceous (Maestrichtian), South Dakota. Part 2.

Systematics of the Bivalvia. Peabody Museum of Natural History , Yale University, Bulletin, 33, 1-222.
Stephenson, l. w. 1923. The Cretaceous formations of North Carolina. Part 1. Invertebrate fossils of the upper
Cretaceous formations. North Carolina Geological and Economic Survey, 5, xi + 408 pp.

- 1941. The larger invertebrate fossils of the Navarro Group of Texas. University of Texas Publication,
4101. 1-641.

stinnesbeck, w. 1986. Zu den faunistischen und palokologischen verhaltnissen in der Quiriguina Formation
and (Maastrichtium) zentral-Chiles. Palaeontographica, Abteilung A, 194, 99-237.
stoliczka, f. 1870. Cretaceous fauna of southern India. Volume III. The Pelecypoda, with a review of all
known genera of this class, fossil and Recent. Memoir of the Geological Survey of India, Palaeontologia Indica,
Series 6, 3 [for 1870-1], 1-222.

992

PALAEONTOLOGY, VOLUME 33

turner, R. D. 1969. Superfamily Pholadacea. N702-N741. In moore, r. c. (ed.). Treatise on invertebrate
paleontology. Part N. Mollusca 6. Bivalvia. Geological Society of America and University of Kansas Press,
Boulder, Colorado and Lawrence, Kansas, 3 volumes, 1224 pp.
turton, w. 1819. A conchological dictionary of the British Islands. John Booth, London, iii-xxvi + 272 pp.
wiffen, j. 1980. Moanasaurus , a new genus of marine reptile (Family Mosasauridae) from the Upper
Cretaceous of North Island, New Zealand. New Zealand Journal of Geology and Geophysics, 23, 507-528.

1981. The first Late Cretaceous turtles from New Zealand. New Zealand Journal of Geology and
Geophysics , 24, 293-299.

1983. The first record of Pachyrhizodus caninus Cope (Order Clupeiformes) from the Late Cretaceous
of New Zealand. New Zealand Journal of Geology and Geophysics, 26, 109-119.

- 1986. Flying reptiles from the Late Cretaceous. [Abstract], In Geological Society of New Zealand, 16th
Annual Conference, 1-5 December 1986. Programme and abstracts. Geological Society of New Zealand
Miscellaneous Publication , 35A, 113.

— and moisley, w. 1986. Late Cretaceous reptiles (Families Elasmosauridae and Pliosauridae) from the
Mangahouanga Stream, North Island, New Zealand. New Zealand Journal of Geology and Geophysics, 29,
205-252.

and molnar, r. e. 1988. First pterosaur from New Zealand. Alcheringa, 12, 53-59.
wilckens, o. 1907. Die Lamellibranchiaten, Gastropoden etc. der oberen Kreide Siidpatagoniens. Bericht der
Naturforschenden Gesellschaft zu Freiburg i. Breisgau , 15, 1-70.

J. S. CRAMPTON

New Zealand Geological Survey

Typescript received 18 September 1989 P.O. Box 30368

Revised typescript received 4 January 1990 Lower Hutt, New Zealand

THE PALAEONTOLOGICAL ASSOCIATION

Notes for authors submitting papers for publication in
PALAEONTOLOGY

and

SPECIAL PAPERS IN PALAEONTOLOGY

PUBLICATION POLICY AND PRACTICE

Scope of publications. Typescripts on any aspect of palaeontology will be considered for publication.
Papers on Recent material may be acceptable if their palaeontological relevance is clear. Preference
is given to typescripts with more than local significance; those which describe only one or two new
species of common genera are not usually published. Short papers and reviews are particularly
welcome; publication of lengthy papers (more than 25 printed pages) may take longer. The series
Special Papers in Palaeontology is for papers longer than those normally accepted for Palaeontology ,
or for collections of shorter papers with a theme.

Palaeontology and Special Papers in Palaeontology provide opportunities for illustration using
plates and text-figures. A high standard of illustration is a feature of both journals. Texts should
be brief and illustrations economical. The style should be simple. Care should be taken to avoid long
and complicated sentences and unusual words or phrases, so that papers may be read more easily
by those whose first language is not English.

Submission. Typescripts should be in English, with British rather than American spelling, and
should be in final form, i.e. complete and in the style of the journal. Three complete copies of the
typescript (including explanatory text for illustrations and tables) should be submitted. Originals of
all illustrations must accompany the typescript. Two additional high-quality copies of all
illustrations at suggested publication size are also required. Submissions should be made to the
Secretary of the Publications Committee at the address given on the inside cover of the current issue
of Palaeontology. The Secretary will assign each paper to an editor who will handle all subsequent
communication with the author. Papers will be sent to at least two referees. The editor will inform
authors of referees’ suggestions to be implemented. Editorial changes will be requested at this stage,
but these can be kept to a minimum if authors follow the instructions given below. Adherence to
these rules will shorten publication time.

Copyright. Authors of papers that are accepted for publication will be asked to sign over the
copyright to the Association.

THE FORMAT OF THE TYPESCRIPT

Typing. Typescripts should be submitted on good-quality paper, preferably of International A4 size
(297 x 210 mm). All pages of scripts, including references and explanations of illustrations, should
be numbered consecutively. Plate and text-figure explanations should be on separate sheets
following the references. Double spacing is required throughout and there should be a margin of 3
cm on both sides.

IPalaeontology, Vol. 33, Part 4, 1990, pp. 993-1000, I pl.|

© The Palaeontological Association

994

NOTES FOR AUTHORS

Style and arrangement . Authors should consult recently published issues of Palaeontology and
construct their papers in accordance with the format used.

The Title should be short but informative and should normally include fossil group, age, and
general location. It should not include the names of new taxa or contain parentheses.

A concise Abstract of not more than 200 words is required at the beginning of all papers. An
abstract reaches a far wider audience than the journal and therefore it should summarize results
(rather than contents) of the paper and must mention all new systematic names.

The main text contains three orders of heading. In typescript these should be represented as
follows: lst-order headings, upper case letters centred; 2nd-order headings, lower case underlined,
flush to the left-hand margin, on a line of their own; 3rd-order headings, as 2nd-order, but followed
by a full stop and with the text continuing on the same line. Nothing else should normally be
underlined in the text except generic and specific names, and foreign words and phrases that are not
anglicized.

In the script, references should be cited by the author’s name and the date of the publication,
without a comma in between. Parentheses are used as appropriate, e.g. ‘...as discussed by Dickens
(1963)’; ‘as discussed previously (Dickens 1963)’. The page reference to any quotation must be
given, e.g. Dickens (1963, p. 20). A work with three or more authors (Carton, Pickwick and Squeers
1964), should be shortened (Carton et al. 1964) unless ambiguity results. Consecutive references
within the same parentheses should be arranged chronologically and separated by a semicolon; for
those by the same author, dates should be separated by commas, e.g. (Dickens 1963, 1965; Carton
et al. 1964). Wherever authority and date are cited, including those of higher-level taxa in the
systematic section, a full reference must be given. Note that the ampersand (&) is never used in the
text.

Reference in the text to the author’s own plates should be cited as (PI. 1, fig. 3) and to those in
other publications as (Bloggs 1967, pi. 2, fig. 4). References to author’s own text-figures should be
cited as (Text-fig. 3a) or ‘in Text-figure 3a’. Again, the lower case should be used when text-figures
from other papers are mentioned. Note that figure and plate numbers, etc. that were originally given
in Roman numerals should be transliterated into arabic figures in the main text and the references.

Explanations of plates , text-figures and tables should be typed (double-spaced and in journal style)
and placed at the end of the typescript. Appropriate positions desired for insertion of plates, text-
figures and tables should be indicated on the typescript. Explanations should be brief but adequate.
Magnifications of each figure or text-figure should be stated. The museum number of the specimen
should be given. For example:

Fig. 1. Ammonia beccarii (Linne). Repository and catalogue number, description, locality, horizon,
magnification.

Figs 2-6. Ephidium crispum (Linne). 2, repository and catalogue number, description, locality, horizon,
magnification. 3, repository, etc. 4, etc. 5, etc. 6, etc.

Systematic work is always introduced by the lst-order heading: ‘SYSTEMATIC PALAE-
ONTOLOGY’. Second-order headings are often not appropriate in this section. The conventions
of the journal regarding a marginal or a central position on the page, the order of the different
sections, the format for synonymies, references to illustrations, etc. can be found by reference to
recent issues of Palaeontology and to the example given below.

Note that accessory information such as etymology, type data, stratigraphical position (which
should be detailed), curatorial information, description, etc., is set in small type. Care should be
taken to ensure that diagnoses, descriptions, and discussion are kept distinct.

The level of the highest taxon used is at the discretion of the author, but must always be
accompanied by authority and date. All references to authorities cited in such headings should be
included in the reference list. Only the most important synonymies should be given. Note that figure
and plate numbers, etc. that were given in Roman numerals are always transliterated into arabic
figures.

NOTES FOR AUTHORS

995

Authors are encouraged to use symbols to indicate the degree of confidence with which synonymy
entries are referred to the species under discussion. Such symbols are given in Matthews, s.c. 1973.
Notes on open nomenclature and on synonymy lists. Palaeontology , 16, 713-719. It may also be
useful to consult bengston, p. 1988. Open nomenclature. Palaeontology , 31, 223-227. Additional
information at the end of synonymy entries is enclosed in square brackets.

The recommended style is:

Order phacopida Salter, 1864
Suborder cheirurina Harrington and Leanza, 1957
Family encrinuridae Angelin, 1854
Subfamily encrinurinae Angelin, 1854
Genus encrinurus Emmrich, 1844

Type species. Entomostracites punctatus Wahlenberg, 1818 from the Wenlock of Gotland, Sweden.

Subgenus encrinurus (encrinurus) Emmrich, 1844
Type species. As for genus.

Encrinurus ( Encrinurus ) macrourus Schmidt, 1859

Plate 41, figs 1-10; Plate 42, figs 1-11; Text-fig. 5
*1859 Encrinurus punctatus var. macrourus Schmidt, p. 438.

.1941 Encrinurus punctatus (Wahlenberg) 1821 [a/'c ] ; Rosenstein, text-fig. 4a, pi. 2, fig. 4-4 b.

v.1962 Encrinurus macrourus Schmidt; Tripp [partim], p. 469, pi. 65, figs 1, 3, 4; pi. 66, fig. 1 a-c\ pi. 67,

figs 2—4: pi. 68, figs 1, 3, ?9; non pi. 65, fig. 2 [ = E. (E.) punctatus ]; pi. 67, fig. 1 ; pi. 68, fig. 2 [ = E.

(E.) cf. intersitus sp. nov.].

.1972 Encrinurus (E.) cf. punctatus 2; Schrank [partim], p. 38, pi. 11, fig. 4 [figs 1, 2, 7 = paratypes of
E. (E.) ruhneunsis Mannil, 1978; figs 3, 5, 6 indeterminable].
non 1972 Encrinurus (E.) cf. punctatus macrourus Schmidt, 1859; Schrank, p. 42, pi. 12, fig. 6 [? = £.(£.)
ruhneunsis Mannil, 1978], fig. 7 [? = E. (E.) punctatus}.

References should be listed at the end of the text in alphabetical order of authors’ names. Each name
should be typed in capitals, with the initials after the surname. The year of publication should follow
and the title of the paper should be given in full. In all titles capital letters should be used only for
proper nouns and for all nouns in German. For papers the name of the journal should be cited in
full and underlined; for books the title should be underlined. Volume number (part or fascicule
number, in parentheses, only if really necessary) and pagination should be given in arabic figures
with the items separated by commas only. Volume numbers only should be given a wavy underline
to indicate bold type. Illustration numbers should be given only where these fall outside the cited
pagination. For books (including volumes containing collections of papers), publisher, place of
publication and number of pages (e.g. 560 pp.) should be given in that order after the title. When
a title has been translated or transliterated, the original language should be stated in square brackets
at the end.

Examples :

boucot, a. j. 1981. Principles of benthic marine paleoecology. Academic Press, New York and London,
xv + 463 pp.

- and Johnson, i. c. 1979. Pentamerinae (Silurian brachiopods). Palaeontographica , Abteilung A, 163,
87-129.

cocks, l. r. m. and mckerrow, w. s. 1978. Silurian. 93-124. In mckerrow, w. s. (ed. ). The ecology of fossils.
Duckworth, London, 383 pp.

996

NOTES FOR AUTHORS

— 1984. Review of the distribution of the commoner animals in Lower Silurian marine benthic
communities. Palaeontology , 27, 663-669.

Holland, c. H. 1971. Some conspicuous participants in Palaeozoic symbiosis. Scientific Proceedings of the
Royal Dublin Society , Series A , 4 (2), 15-26.

jaanusson, v. 1979. Ecology and faunal dynamics. 253-294. In jaanusson, v., laufeld, s. and skoglund, r.
(eds). Lower Wenlock faunal and floral dynamics - Vattenfallet section, Gotland. Sveriges Geologiska
Undersokning Afhandlingar och Uppsatser , Stockholm , Series C , 762, 1-294.

Northrop, s. a. 1939. Paleontology and stratigraphy of the Silurian rocks of the Port Daniel-Black Cape
Region, Gaspe. Special Papers of the Geological Society of America , 21, i-ix, 1-302.
obut, a. m. 1964. Podtip Stomochordata. 279-337. Stomokhordovye. In orlov, y. a. (ed. ). Osnovv
paleontologii : Echinodermata, Hemichordata, Pogonophora, Chaetognatha. Nedra Press, Moscow, 383 pp. [In
Russian],

orbigny, a. c. v. d. d\ 1853. Note sur le nouveau genre Hvpotrema. Journal de Conchyliologie, 4, 432-438, pi.
10.

rasmussen, H. w. 1978. Articulata. T813-T928. In moore, r. c. and teichert. c. (eds). Treatise on invertebrate
paleontology. Part T. Echinodermata 2(3). Geological Society of America and University of Kansas Press,
Boulder, Colorado and Lawrence, Kansas, 215 pp.

selden, p. a. 1979. Functional morphology of Baltoeurypterus. Unpublished Ph.D. thesis. University of
Cambridge.

smith, a. b. and wright, c. w. 1988. British Cretaceous echinoids. Part 1, general introduction and Cidaroidea.

Monograph of the Palaeontographical Society , 141 (578), 1—101, i-vi, 32 pis.
stormer, L. 1934«. Merostomata from the Downtonian sandstone of Ringerike, Norway. Skrifter utgitt av det
Norske Videnskaps-Akademi i Oslo. 1. Matematisk-Naturvidenskapelig Klasse. 1933 (10), 1-125.

— 1934b. Downtonian Merostomata from Spitzbergen, with remarks on the suborder Synziphosura.
Skrifter utgitt av det Norske Videnskaps-Akademi i Oslo. I. Matematisk-Naturvidenskapelig Klasse, 1934 (3),
1-26.

whittard, w. f. 1934. A revision of the trilobite genera Deiphon and Onycopyge. Annals and Magazine of
Natural History (10), 14, 505-533.

Whittington, h. b. 1977. The Middle Cambrian trilobite Naraoia , Burgess Shale, British Columbia.
Philosophical Transactions of the Royal Society of London, Series B, 280, 409-443.

Authors’ names and addresses should be typed on the right-hand side of the page after the reference
list.

Footnotes are not allowed. Permission to publish, for example, should be included with the other
acknowledgements.

LINE ILLUSTRATIONS AND TABLES

Text-figures. Original drawings, in black on good-quality white card or on good-quality drawing
film, should be submitted with the typescript. They should be no larger than double the final size.
When reduced they must not exceed the type area of a page, 200 x 147 mm. If the caption of a full-
page figure is long, allowance for its inclusion on the same page should be made by reducing the
height of the figure. In composing smaller text-figures, space on the page is best used if the figure
is wide rather than long, and the full width of the page should be used where possible. Individual
line drawings within one text-figure should be identified by capitals (a, b, c, etc.) added by the
author. All descriptive lettering should be inserted by the author and should be readable when
reduced to publication size. Typewriting is not acceptable on text-figures. A linear scale may be
included.

Tables. Tables that cannot be typeset should be carefully drafted to the same standards as text-
figures. Fold-out tables are not acceptable, because of prohibitive cost. If a table of larger than page
size is essential, it should be arranged for two facing pages.

NOTES FOR AUTHORS

997

PLATES AND TEXT-PHOTOGRAPHS

Maintaining a high standard of plate quality is an editorial priority. Authors are reminded that the
published plate cannot be of better quality than the original. Plate X shows some of the common
defects noted in plates submitted for publication. If your plates show any of these defects, they
should be corrected before submitting them to the journal, or they will be returned.

Size. The size of plates is 200 x 147 mm. Where authors have photographs which fill less than a full
page, they will be referred to as text-figures. These will be reproduced by the same process as plates.
Every effort should be made to ensure that no page space is wasted, particularly in the width of text-
figures composed of photographs. However, wherever possible plates rather than photographic
text-figures should be submitted.

Plates and photographic text-figures should be submitted at publication size. Magnifications of
individual figures should be stated in the caption, and not indicated by scale-bars on the
photographs.

Lighting. The convention of lighting fossils from the top left should be followed.

Preparation of photographic prints. Photographs should be sharply in focus and printed on glossy
paper. They should be of medium contrast, using the range of shades of grey but avoiding extremes
of black and white. Artefacts such as dirt or scratches should be absent from the prints. All prints
on a plate should be of even tone and contrast. For this reason, it is often better to avoid mixing
conventional photographs with photomicrographs. Where possible, remove labels from macro-
fossils before photography. Avoid unsightly backgrounds.

Mounting. Mount the prints on clean, white board of A4 size, using dry mounting tissue or a non-
aqueous glue. Maximum thickness of board should be ^ inch or 1-5 mm. Use the same method for
all prints. Avoid glue on the background or on the print surface. Place prints as close together as
is reasonable, making full use of plate space. Rectangular prints should not only be cut rectangular
but also mounted parallel to each other and to the sides of the plates. Spacing should be as uniform
as possible: Plates 24-30 in Volume 32, Part 1 are good examples to follow. Leave space for figure
numbers either directly below each print or by cutting off a small corner at 45° to the print edge.

When it is desirable to remove the background from around a fossil, the print should be carefully
trimmed to the edge of the fossil and mounted on clean white board. Black backgrounds are
permitted, but the author is responsible for accurately delimiting the edge of the fossil; roughly
trimmed prints can be precisely outlined using black drafting ink.

Each plate should be protected by a tracing paper overlay attached along the upper edge. The
author’s name, title of the paper and the number, should be written on the back of the original of
each text-figure and plate.

Numbering. For plates, figure numbers should normally run consecutively from left to right and
from top to bottom of each plate. Numbers should be written in appropriate positions on the
overlays but not on the plates. All numbers are added in required positions by the printers, although
space for such numbers should be allowed when the plate is made up by the author. Lettering and
arrows should be avoided, but where absolutely necessary should be added to the overlay. For
photographic text-figures, lettering will also be added by the Press and should be indicated on the
overlay. It should be kept to a minimum.

ADMINISTRATION

Preservation of types and other specimens. In accordance with recommendations of the International
Codes of Botanical and Zoological Nomenclature, all illustrated and described fossils must be

998

NOTES FOR AUTHORS

registered and deposited in an appropriate permanent institution, with staff and facilities capable
of ensuring their conservation and availability for future reference in perpetuity. The registered
numbers should be quoted.

Proofs. Authors will normally receive one proof ; this proof is for the purpose of correcting printer’s
errors and not for altering the wording or substance of the paper. Authors will be charged for
excessive alterations. The editors will be responsible only for authors’ corrections notified by return
of post. Whenever possible, photographic proofs of publication quality will also be sent to the
authors.

Offprints. Fifty offprints of each paper will be sent free of charge; further copies may be purchased
at prices shown on the order form which will be sent with the proofs to the author (or corresponding
author in the case of multi-author papers).

Deposition of data. The Association makes use of the scheme run by the British Library, Lending
Division, whereby tables of data and other reference material can be deposited with the British
Library rather than printed. The deposited material is stored on microfiche, and either microfiche
or full-size copies may be obtained from the British Library by applicants (preferably using British
Library prepaid coupons) on a standard scale of charges which allows for postage, etc.

The British Library will accept deposited material only through the Publications Committee of
the Association; all such material will be referred to as part of a published paper. The published
paper will bear a reference to deposited copy with full details of its pagination and means of
acquisition, e.g. ‘...have been deposited with the British Library, Boston Spa, Yorkshire, UK, as
Supplementary Publication No. SUP 14003 (26 pages)’. Prepaid coupons for such purposes are held
by many technical and university libraries throughout the world. They may be purchased from the
British Library, Lending Division, Boston Spa, Wetherby, Yorkshire, LS23 7BQ, UK. Association
policy is that neither plates nor formal taxonomic data will be considered for deposition. Authors
should indicate and separate those parts of their papers that they propose for deposition; the
Publications Committee may also recommend that part of a paper should be deposited rather than
be printed.

Preparation of copy for deposition. Copy must be prepared by the author according to the following
specifications. Editors will not undertake the preparation of copy.

i. Copy must be camera-ready.

ii. Maximum page size for text or tables in typescript or computer printout is 330 mm
high x 240 mm wide, including margins. Preferred page size is A4.

iii. Tabular matter should be headed descriptively on the first page, with column heading
recurring on each page.

EXPLANATION OF PLATE X

Figs I -9. How not to make a plate. Overall, this plate is wasteful of space, individual figures are inaccurately
trimmed and spaced, and are parallel neither to each other nor to the edges of the plate; cut-offs to
accommodate numbers are excessive and variable, the sequence of numbers is inconsistent, and the numbers
themselves are inconsistently and, in some cases, ambiguously placed. The plate includes illustrations both
of hand specimens and electron micrographs: these rarely blend harmoniously. Faults on individual figures
are as follows. 1, scratches on print (possibly from negative). 2, lacks contrast, bottom left corner missing.
3, out of focus, bottom of specimen too near edge of print. 4, lacks depth of field, so that lowest portions
of specimen are out of focus. 5, data zone left on, contrast too high so that some areas are completely white
and show no detail, central canal clumsily retouched, right-hand canal chopped in half by insensitive
cropping of print. 6, butt-edged (badly) to 5, resolution poor leading to fuzziness, scan-lines produced by
poor earthing of specimen. 7, specimen edge clumsily blacked-out. 8, specimen edge clumsily trimmed. 9,
poor lighting has resulted in bottom edge of specimen merging with background.

PLATE X

l

PALAEONTOLOGY, How not to make a plate

1000

NOTES FOR AUTHORS

iv, Prefatory text, which should contain the abstract from the parent paper, should be included.

v, All pages must be consecutively numbered.

Authors with large sections for deposition are advised to consult the Secretary of the Publications
Committee for further information.

SPECIAL PAPERS IN PALAEONTOLOGY
Preparation of papers for this series should be in the same style as for Palaeontology.

Submission. Prospective authors should consult the Publications Secretary well in advance of
submission, supplying as much information as possible.

Cost of publication. The Association’s funds for this series are limited. Authors are asked to obtain
grants wherever possible. Scripts should not normally exceed the equivalent of 1 10 published pages.

Offprints. A small number of free offprints will be supplied, and further copies may be obtained at
a special reduced charge. For multiple-author volumes such as conference proceedings, authors will
receive one free copy of the volume, but no free offprints. Offprints of individual papers may be
ordered at the standard charge. Details will be supplied at the time by the editor concerned.

GRANTS IN AID OF PUBLICATION

Palaeontology has no compulsory page or plate charges. However, authors are requested to seek
grants in aid of publication from their institutions or from research funds, or to apply for
publication costs in research grants. Such financial support is particularly welcome, and may be
essential, for long papers.

Although acceptance of a paper for publication will not depend on the receipt of such grants,
authors will appreciate that the funds available to the Association for publication are limited. Every
grant or donation will therefore directly help the Association's publication programme.

1990

The Editors

THE PALAEONTOLOGICAL ASSOCIATION

ANNUAL REPORT OF COUNCIL FOR 1989

Membership and Subscriptions. Membership totalled 1,285 on 31 December 1989, an increase of 10 over the
previous year. There were 808 Ordinary Members, an increase of 4; 83 Retired Members, an increase of 10;
1 13 Student Members, an increase of 4, and 281 Institutional Members, a decrease of 8. Total individual and
institutional subscriptions to Palaeontology through Basil Blackwell’s agency numbered 438, an increase of 2.

Subscriptions to Special Papers in Palaeontology numbered 100 individuals, a decrease of 9, and 114
Institutions, a decrease of 3. Orders through Basil Blackwell’s agency for Special Papers totalled 108 volumes.

Sales of backparts of Palaeontology via the Membership Treasurer realized £370 05. Sales of back numbers
of Special Papers in Palaeontology to individuals yielded £771 • 50, and to Institutions £219 00.

Almost all sales of the Field Guides to Fossils series resulted from the activities of the Marketing Manager,
Fossils of the Chalk yielding £2,381, Fossil Plants of the London Clay , £413 and Zechstein Reef Fossils and their
Palaeoecology, £587. The Atlas of Invertebrate Macrofossils yielded £959 (£870 of which was royalties from
sales), the Burgess Shale Portfolio yielded £88 and sales of the Malta Field Guide , £2-50.

Finance. During 1989 the Association published Volume 32 of Palaeontology at an estimated cost of £72,208
(including postage and distribution). Special Papers 41 and 42 were published at an estimated cost of £4,600
and £9,300 respectively (including postage and distribution). The Association is grateful to all those who made
donations to offset the cost of publishing Palaeontology and Special Papers.

Publications. Volume 32 of Palaeontology , published in 4 parts during 1989, contained 917 pages and 107
plates. Special Papers 41 and 42 were published. Due to the sudden closure of the Oxford University Press
Printing House, the Association had to appoint new printers; Cambridge University Press were chosen and
Parts 2-4 of Palaeontology were printed there.

Meetings. Nine meetings were held in 1989. The Association is indebted to the organizers, hosts and field
leaders of these.

a. Review Seminar on ‘The Cretaceous/Tertiary Boundary’ held on 15 February at the University Museum,
Oxford and convened by Dr R. A. Spicer.

b. Lyell Meeting held jointly with the Geological Society, on ‘Palaeoclimates’ on 22 February at Burlington
House and convened by Dr C. P. Summerhayes.

c. Thirty-second Annual General Meeting held in the Department of Geology, Imperial College on 1 5 March.
Dr Dianne Edwards delivered the Annual Address on ‘Pioneering plants’. The Sylvester-Bradley Award
was made to Patrick Wyse Jackson.

d. Field Meeting to Yorkshire Jurassic plant localities, led by Dr C. R. Hill on 7-9 April. 21 people
participated.

e. Progressive Palaeontology meeting, an open meeting for presentations by research students was held at
University of Wales College of Cardiff on 17 May. It was convened by Andrew King.

/. Seminar and Workshop on ‘Physical Modelling in Palaeontology’ was held at the Department of Earth
Sciences, Cambridge on 9 October. It was convened by Sue Rigby and Clare Milsom.

g. Joint One Day Meeting with the Malacological Society on ‘The functional morphology of the mollusc
shell’, held on 25 October in the British Museum (Natural History). The meeting was convened by Dr
J. D. Taylor and Dr J. A. Crame. Approximately 75 people attended.

h. Review Seminar on ‘ Macroevolution ’ held at the Open University on 15 November and convened by Dr
P. W. Skelton. Over 180 people attended.

i. Review Seminar held jointly with the Geological Society, on ‘Carbonate buildups’ on 30 November at
Derbyshire College of Higher Education. The convenor was Dr P. Bridges and there were 80 participants.

j. The Annual Conference , held at the University of Liverpool on 18-21 December was an open meeting,
with a thematic session on ‘Functional Morphology’. The Local Secretary was Dr C. R. C. Paul and the
meeting was attended by about 1 50 people.

1002

THE PALAEONTOLOGICAL ASSOCIATION

Council. The following members served on Council following the Annual General Meeting on 15 March 1989.
President , Dr J. D. Hudson; Vice-Presidents , Dr P. W. Skelton, Dr M. Romano; Treasurer , Dr M. E.
Collinson; Secretary , Dr P. Wallace; Membership Secretary , Dr H. A. Armstrong; Institutional Membership
Treasurer , Dr A. W. Owen; Editors , Dr M. J. Benton (also Public Relations Officer), Dr J. E. Dalingwater, Dr
D. Edwards, Dr C. R. C. Paul, Dr P. A. Selden, Dr P. D. Taylor; Marketing Manager , Dr C. R. Hill; Circular
Reporter , Dr D. Palmer; Other Members , Dr J. A. Crame, Dr G. B. Curry, Dr E. A. Jarzembowski, Dr R. A.
Spicer. Dr D. M. Martill was co-opted in 1989.

Council Activities. In conjunction with other Societies, Council lobbied the Post Office to issue a set of
commemorative stamps in 1991 to mark the 150th anniversary of the first use of the term ‘dinosaur’ by Sir
Richard Owen. The suggestion was accepted and four postage stamps bearing pictures of four British dinosaurs
representing four major types of dinosaur will be issued in 1991.

Many amateur palaeontologists make highly significant and often entirely altruistic contributions to
palaeontology, ranging from the making of collections to their study, care and conservation. Council has
decided to recognize this contribution by an Award for Amateur Palaeontologists: the first award will be
made to Mr Alan Dawn at the AGM in March 1990.

The new A5 format Palaeontology Newsletter continues to be published, although problems with printers
made it necessary to combine issues 3 and 4 in one volume. It is attracting favourable comment and varied
contributions are beginning to be submitted.

BALANCE SHEET AND ACCOUNTS FOR THE
YEAR ENDING 31 DECEMBER 1989

Balance Sheet as at 31 December 1989

1988

£ £
52,359

6,596

28,966

2,476

450

6,811

Investments at Cost (see schedule) .
Current Assets
Sundry Debtors .

Cash at bank . . . . .

Sylvester- Bradley Fund .

Loans ......

Stocks at valuation

45,299

Current Liabilities

3,060 Subscriptions received in advance .
Provision for cost of :

Palaeontology . . . .

Special Papers . . . .

7,880 Sundry creditors . . . .

10,940

34,359

£ £

55,691

. 3,213

. 56,706

. 2,587

450
4,525

67,481

3,752

16,000

13,900

1,305

34,957

32,524

86,718

88,215

Represented by:

Publications Reserve Account

99,726

Balance brought forward ........

Excess of income over expenditure for the year transferred from

83,025

83,025

(16,701)

from Income and Expenditure Account (1988) deficit .
Sylvester- Bradley Fund

1,386

1,885

Balance brought forward ........

2,476

86

Interest ............

169

(200)

Grant awarded ..........

(200)

705

Donations ...........

142

2,476

1,217

—

Meeting Reserve

86,718

88,215

1003

Income and Expenditure Account for the Year Ended 31 December 1989

INCOME

1988

£

Subscriptions

1989

1988

39,833

Palaeontology

Sales .......

Donations ......

34,226

Special Papers

Sales .......

Donations ......

11,292

Burgess Shale Portfolio

53 Sales .......

Fossil Plants of the London Cla y

951 Sales .......

Fossils of the Chalk

6,340 Sales .......

Zechstein Reef Fossils

82 Sales

Atlas of Invertebrate Macrofossils

811 Sales

Palstat

Income .......

Expenses ......

82

(201) Offprints

Profit on Sales of Investments
10,153 Investment Income (see schedule)

54 Sundry Income

£

38,663

717

34,024

518

5,558

2,000

£

39,380

34,542

7,558

413

2,427

587

959

1,178

3,454

11,297

3

£103,676

£101,886

1004

1988

£

82,420

18,038

5,029

1,366

2,499

250

3,867

6,908

£120,377

(16,701)

EXPENDITURE

Cost of Publication of Palaeontology
Volume 32 - Part I
Part 2
Part 3

Part 4 (prov.)

Paper purchased from OUP

Cost of Publication of Special Papers
No. 41 (prov.) ....

No. 42 (prov.) ....

Fossils of the Chalk
Cost of books sold

Z ec h stein Reef Fossils
Cost of books sold

Cost of Scientific Meetings .
Advertising Field Guides Series
Warehousing of Publications

Grants

Cost of Newsletters

Preparation .....

Postage

Credit ......

Administrative Costs

Postage, stationery, telephone
Editorial expenses ....
Meeting expenses ....
Audit fee ....
Stockbroker fee
Membership of Societies

£ £

18,272

18,336

14,316

16,000

5,284

72,208

4,600

9,300

13,900

1,889

1,889

397

397

579

449

(3,193)

50

4,860

3,259

(1,125)

6,994

3,150

965

2,155

275

564

118

7,227

£100,500

Excess of Income over Expenditure for the Year Transferred to
Publications Reserve Account (1988 deficit) . . . .

£1,386

1005

Schedule of Investments and Investment Income as at 31 December 1989

Gross Income

Cost

for Year

£

£

£2,000

11% Exchequer Stock 1991

1,991

220

£1,000

9% Treasury Stock 1992/96

992

90

£10,000

9% Treasury Stock 1994 ........

9,943

900

£4,000

8% Treasury Stock 2002/6

2,192

320

£5,357

13|% Treasury Stock 1997

5,000

710

£3,280

13|% Exchequer Stock 1996 .......

3,000

434

5,270

M & G Charifund Units

4,073

1,945

1,872

M.E.P.C. p.l.c. 25p shares .

4,939

177

2,280

National Westminster Bank p.l.c. £1 Ordinary Shares

3,929

451

10,150

Agricultural Mortgage Corporation Ltd. 7|% Debenture Stock

1991/93

8,251

786

10,000

New Throgmorton Trust (1983) p.l.c. 25p Income Shares .

1,706

739

1,460

Saatchi & Saatchi 6-3 % Convertible Cumulative Redeemable

Preference £1 Shares ........

1,994

123

£2,100

Hanson Trust p.l.c. 10% Convertible Unsecured Loan Stock

2007/12

2,974

210

£4000

Cable & Wireless 7% convertible unsecured loan stock 2008

4707

140

7,369

Bank Interest ..........

3,924

55,691

11,297

Market Value at 31 December 1989 ( 1 988— £9 1 ,93 1 )

£103,579

Report of the Auditor to the Members of
The Palaeontological Association

In my opinion, the Accounts as set out above give a true and fair view of the state of the affairs of the
Association at 31 December 1989 and of its income and expenditure for the year ended on that date.

March 1990 G. R. Powell

Market Harborough, Leicestershire Chartered Accountant

1006

INDEX

Pages 1-256 are contained in Part 1 ; pages 257-502 in Part 2; pages 503-748 in Part 3; pages 749-992 in Part 4.

A

Acanthoceras amphibolum, 98; bellense , 92
Acritarchs: Proterozoic, Norway, 287
Aemula inusitata , 864; sp., 865
Aitken, J. D. See Narbonne, G. M. and Aitken, J. D.
Aldridge, R. J. SeeTheron, J. N., Rickards, R. B. and
Aldridge, R. J.

Allmon, W. D., Nieh, J. C. and Norris, R. D. Drilling
and peeling of turritelline gastropods since the late
Cretaceous, 595

Alzadites alzadensis sp. nov., 398; incomptus sp. nov.,
399; westonensis sp. nov., 398; sp. A, 400
Alzadites! sp., 400

Ammonites: Cretaceous, USA, 75, 379; Kosmoceras
preyed upon by fish, Jurassic, England, 739
Amphibians: frog, Jurassic, England, 299; microsaur,
Permian, USA, 893; Triassic, England, 873
Anagaudrvceras involulum, 84
Anechocrinus nalbiaensis sp. nov., 67
Anisoceras cf. plicatile , 142
Arachnidium smithii, 20
Arachnoidella abusensis sp. nov., 26
Argyrotheca bronnii, 857 ; coniuncta , 857 ; hirundo , 858
Aulichnites ichnosp., 964

Australia: Permian crinoids, 49; Quaternary birds,
447

B

Baker, P. G. The classification, origin and phylogeny
of thecideidine brachiopods, 175
Baird, R. F. and Rowley, M. J. Preservation of avian
collagen in Australian Quaternary cave deposits
Baumiller, T. K. Non-predatory drilling of Missis-
sippian crinoids by platycerid gastropods, 743
Beltane l la gilesi, 956
Bioimmuration, 1, 19
Biomechanics: trilobite exoskeletons, 749
Birds: Quaternary, Australia, 447
Bivalves: Caenozoic heterodont superfamily Dreis-
senacea, 707 ; gryphaeate oysters, Jurassic, western
Europe. 453; wood-boring, Cretaceous, New Zea-
land, 981

Borissiakoceras orbiculatum , 85, 385
Brachiopods: thecideidine classification, origin and
phylogeny, 175; late Cainozoic, South Africa, 313;
Cretaceous, England, 823

Brunton, C. H. C. and Hiller, N. Late Cainozoic

brachiopods from the coast of Namaqualand,
South Africa

Bryozoa: Jurassic, England and France, 19
Buccinammonites minimus sp. nov., 414

C

Cainozoic: brachiopods. South Africa, 313; drilling
and peeling predation of gastropods, 595; hetero-
dont bivalves, 707
Calceolispongia abundans , 68
Calycoceras (Newboldiceras) sp., 106
Cambrian: eocrinoid, Spain, 249; trilobite, Wales,
429; trilobite eyes, China, 911
Canada: Cretaceous dinoflagellate, 35; Precambrian
ediacaran fossils, 945
Cancellothyris platys sp. nov., 335
Capitosauridae incertae sedis , 880
Carboniferous: gastropod drilling of crinoids, USA,
743; hyoliths, USA, 343

Cardoarachnidium bantai sp. nov., 28; Cardoarach-
nidium bantai sp. nov., 28; voigti sp. nov., 30
Carneithyris subcardinalis, 840

Carroll, R. L. A tiny microsaur from the Lower
Permian of Texas: size constraints in Palaeozoic
tetrapods, 893

Carthaginites aquilonius , 417
Cenozoic. See Cainozoic

Cephalopods: orientation of shells in illustrations,
243

Cermatops discoidalis, 431
Char nio discus! sp., 958
China: Cambrian trilobite eyes, 91 1
Clarkson, E. N. K. See Zhang Xi-Guang and Clark-
son, E. N. K.

Classification: Foraminifera, 503; thecideidine brach-
iopods, 175; trilobites, 529
Cobban, W. A. See Kennedy, W. J. and Cobban,
W. A.

Collagen: preservation in Quaternary birds, 447
Communities: Silurian, Welsh Borderland, 209
Computer-aided restoration, 429
Colinoceras tarrantense, 107; sp., Ill
Conodont : Ordovician, South Africa, 577
Corals: Silurian, northern Europe, 769
Crampton, J. S. A new species of late Cretaceous
wood-boring bivalve from New Zealand, 981
Crania afif. craniolaris , 836
Craniscus sp., 837

1008

INDEX

Cretaceous: ammonites, USA, 75, 379; bivalve. New
Zealand, 981; brachiopods, England, 823; dino-
flagellate, Canada, 35; drilling and peeling pre-
dation of gastropods, 595; spiders, Spain, 257;
woods, USA, 225
Cretaraneus vilaltae sp. nov., 270
Cretirhynchia limbata , 838; sp., 838
Crinoids : drilled by gastropods. Carboniferous, USA,
743, Permian, Australia, 49
Critical point drying, 423
Cryptometoicoceras mite sp. nov., 412
Cunningtoniceras inerme, 128; johnsonanum , 124;

lonsdalei , 121 ; sp. juv., 391
Cyclomedusa plana , 958, sp., 959

D

Dalligas nobilis, 866

Database: deductive enquiry system for museum
material, 613

Dendrocystoides scoticus , 633

Dichocrinusl gerringgongensis sp. nov., 53

Dinoflagellate: Cretaceous, Canada, 35

Dipteronotus cyphus, 874

Donacophyllum neumani sp. nov., 814; wallstenense
sp. nov., 818

Donovan, D. T. See Rogers, M. J., Donovan, D. T.
and Rogers, M. H.

Doyle, P. Teuthid cephalopods from the Lower
Jurassic of Yorkshire, 193

Dracius carnifex, 856

Dunveganoceras pondi , 403

E

Ediacaran fossils: Canada, 945
Ediacaria sp., 960

England: Cretaceous brachipods, 823; Jurassic Bryo-
zoa, 19; Jurassic fish preying on ammonites, 739;
Jurassic frog, 299; Jurassic teuthids, 193; Silurian
marine communities, 209; Triassic fish. 873
Entelophyllum articulation , 778 ; articulatum anglicum
subsp. nov., 782; confusum , 798; dendroides sp.
nov., 791 ; fasciculatum, 785; cf . fasciculatum, 788;
hamraense sp. nov., 800; lauense sp. nov., 799;
proliferum , 790; prosperum, 796; prosperum cras-
sum , 798; pseudodianthus , 793; pseudodianthus
transiens , 796; sp. A, 790
Entelophyllum ? dalecarlicum , 806; visbyense , 802
Eocrinoid: Cambrian, Spam, 249
Eocyclotosaurus sp., 878
Eodiscoglossus oxoniensis sp. nov., 301
Eoindocrinus praecontignatus, 64
Eoporpita sp., 960

Ericiasphaera spjeldnaesii sp. nov., 291

Escuillie, F. See Hugueney, M., Tachet, H. and
Escuillie, F.

Evans, D. H. and King, A. H. The affinities of early
oncocerid nautiloids from the lower Ordovician of
Spitsbergen and Sweden, 623
Evans, S. E., Milner, A. R. and Mussett, F. A new
discoglossid frog from the Middle Jurassic of
England, 299

Evolution : gryphaeate oysters, Jurassic, western
Europe, 453

Eyes: Cambrian eodiscid trilobites, 911

F

Fish: Triassic, Italy, 155; application of critical point
drying, 423; semionotids preying on ammonites,
Jurassic, England, 739; Triassic, England, 873
Foraminifera : classification, 528
Forbesiceras brundrettei, 91; conlini , 92; cf. chevillei,
91

Fortey, R. A. Ontogeny, hypostome attachment and
trilobite classification, 529

France: Jurassic Bryozoa, 19; Miocene caddisfly
pupae, 495

Fraser, N. C. See Milner, A. R., Gardiner, B. G.,
Fraser, N. C. and Taylor, M. A.

Frog: Jurassic, England, 299

G

Gardiner, B. G. See Milner, A. R., Gardiner, B. G.,
Fraser, N. C. and Taylor, M. A.

Gastropods: drilling and peeling predation of turri-
tellines, Cretaceous and Cainozoic, 595; drilling
into crinoids, Carboniferous, USA, 743
Gerkella humboldti sp. nov., 351
Gisilina gisii , 849; jasmundi , 850
Gogia ( Alanisicvstis ) andalusiae subgen. et sp. nov.,
250

Graptolites: Ordovician, Scotland, 933; Silurian,
Northern Ireland, 937

H

Hamites cimarronensis, 140, 414; salebrosus , 415
Harding, I. C. Palaeoperidinium cretaceum : a brack-
ish-water peridiniinean dinoflagellate from the
early Cretaceous, 35

Harper, L. See Martill, D. M. and Harper, L.
Haynes, J. R. The classification of the Foraminifera
- a review of historical and philosophical perspec-
tives, 503

Helminthoida ichnosp., 966
Helminthoidichnites tenuis , 958
Helminthopsis abeli, 968 ; irregularis , 969
Helminthopsisl ichnosp., 969
Hiller, N. See Brunton, C. H. C. and Hiller, N.

INDEX

1009

Hughes, N. C. and Rushton, A. W. A. Computer-
aided restoration of a late Cambrian ceratopygid
trilobite from Wales, and its phylogenetic impli-
cations, 429

Hugueney, M., Tachet, H. and Escuillie, F. Caddisfly
pupae from the Miocene Indusial Limestone of
Saint-Gerard-le-Puy, France, 495
Hyolithesl aculeatus, 351; carbonaria , 352; milleri ,
353; parvulus , 353; waverliensis, 354
Hyoliths: Carboniferous, USA, 343

I

Idiohamites bispinosus sp. nov., 416; pulchellus sp.
nov., 416

Illustration: orientation of cephalopod shells, 243
Indusia tubulosa , 496

Insects: caddisfly pupae, Miocene, France, 495
Isocrania costata , 836
Italy: Triassic fish, 155

J

Jefferies, R. P. S. The solute Dendrocystoides scoticus
from the Upper Ordovician of Scotland and the
ancestry of chordates and echinoderms, 631
Jeletzkyteuthis agassizi, 198

Jell, J. S. and Sutherland, P. K, The Silurian rugose
coral genus Entelophyllum and related genera in
northern Europe, 769
Jimbacrinus minilyaensis sp. nov., 70
Johansen, M. B. and Surlyk, F. Brachiopods and the
stratigraphy of the Upper Campanian and Lower
Maastrichtian Chalk of Norfolk, England, 823
Johnson, A. L. A. and Lennon, C. D. Evolution of
gryphaeate oysters in the mid- Jurassic of western
Europe, 453
Johnsonites sp., 85

Jurassic: Bryozoa, England and France, 19; fish
predation on ammonite Kosmoceras , England, 739;
frog, England, 299; gryphaeate oysters, western
Europe, 453; teuthid cephalopods, Yorkshire, 193

K

Kastanoceras spiniger sp. nov., 394
Kennedy, W. J. and Cobban, W. A. Cenomanian
ammonite faunas from the Woodbine Formation
and lower part of the Eagle Ford Group, Texas, 75
Kennedy, W. J. and Cobban, W. A. Cenomanian
micromorphic ammonites from the Western In-
terior of the USA, 379

Kershaw, S. Stromatoporoid palaeobiology and taph-
onomy in a Silurian biostrome on Gotland,
Sweden, 681

King, A. H. See Evans, D. H. and King, A. H.
Kingena pentangulata , 862
Knotted Burrow, 974

Kraussina cnneata sp. nov., 335; laevicostata , 333;

lata , 325; rotunda sp. nov., 331; rubra , 324
Kullingial sp., 952

L

Lacazella ( Bifolium ) we therein, 868
Lennon, C. D. See Johnson, A. L. A. and Lennon,
C. D.

Leptothyrellopsis polonicus , 868
Lesperance, P. J. Cluster analysis of previously des-
cribed communities from the Ludlow of the Welsh
Borderland, 209
Lingula cretaceal , 834
Lockeia ichnosp., 971
Loligosepia aalensis , 196

Loydell, D. K. On the graptolites described by Baily
(1871) from the Silurian of Northern Ireland and
the genus Streptograptus Yin, 937

M

Macryphantes cowdeni sp. nov., 275
Magas chitoniformis , 860

Malinky, J. M. and Sixt, S. Early Mississippian
Hyolitha from northern Iowa, 343
Martill, D. M. Predation on Kosmoceras by semi-
onotid fish in the Middle Jurassic Lower Oxford
Clay of England, 739

Martill, D. M. and Harper, L. An application of
critical point drying to the comparison of modern
and fossilized soft tissues of fishes, 423
Mastodonsaurus lavisi , 876
Medusinites asteroides , 962
Metabrograptus scoticus gen. et sp. nov., 935
Met apty choc eras spp., 415
Metengonoceras dumbli , 90

Metoicoceras latoventor, 139; mosbyense , 408; swal-
lovi, 138; aff. praecox , 408; sp. A, 404
Microsulcatoceras crassum sp. nov., 401 ; puzosiiforme
sp. nov., 401 ; texanum sp. nov., 402
Microsulcatoceras sp.?, 402

Milner, A. R., Gardiner, B. G., Fraser, N. C. and
Taylor, M. A. Vertebrates from the Middle Triassic
Otter Sandstone Formation of Devon, 873
Milner, A. R. See Evans, S. E., Milner, A. R. and
Mussett, F.

Miocene: caddisfly pupae, France, 495
Moremanoceras costatum, 388; montanaense sp. nov.,
390; straini , 86, 386

Museum material : deductive enquiry system for
databases, 613

Mussett, F. See Evans, S. E., Milner, A. R. and
Mussett, F.

1010

INDEX

N

Nannometoicoceras nemos , 412
Nannometoicoceras ? glabrum sp. nov., 413
Narbonne, G. Y. and Aitken, J. D. Ediacaran fossils
from the Sekwi Brook area, Mackenzie Mountains,
northwestern Canada, 945

Nautiloids: Ordovician, Spitsbergen and Sweden, 623
Neocamptocrinus occidentalis sp. nov., 54; sp. nov., 57
Neonereites ? ichnosp., 971

New Zealand: Cretaceous wood-boring bivalve, 981
Nieh, J. C. See Allmon, W. D., Nieh, J. C. and
Norris, R. D.

Norway: Proterozoic acritarchs, 287
Norris, R. D. See Allmon, W. D., Nieh, J. C. and
Norris, R. D.

Northern Ireland: Silurian graptolites, 937
Nuttall, C. P. Review of the Caenozoic heterodont
bivalve superfamily Dreissenacea, 707

O

Occiducrinus australis sp. nov., 60
Ordovician : trilobite, Portugal, 487 ; conodont. South
Africa, 577 ; nautiloids, Spitsbergen and Sweden,
623; solute, Scotland, 631; graptolite, Scotland,
933

Ostlingoceras ( Ostlingoceras ) brandi , 144; davisense ,
145

P

Palaeoperidinium cretaceum , 44
Palaeophycus tubular is, 972
Palaeouloborus lacasae sp. nov., 263
Papillomembrana compta, 294
Paraconlinoceras barcusi , 114; leonense, 118
Parrish, J. T. See Spicer, R. A. and Parrish, J. T.
Pelagodiscus (?) sp., 322

Permian: crinoids, Australia, 49; microsaur amphib-
ian, USA, 893

Petrozium dewari , 807; losseni , 810
Pholadidea (Hatasia) wiffenae sp. nov., 985
Phylogeny: gymnolaemate b'-yozoans, 19; thecidei-
dine brachiopods, 175; chordate and echinoderm
ancestry, 631

Phthanoncoceras oelandense sp. nov., 625
Planolites montanus , 972
Plants: Alaskan Cretaceous woods, 225
Plesiacanthoceras bellsanum , 135; cl. bellsanum , 403
Plesiacanthoceratoides vetula, 137
Plumose Problematicum, 963
Portugal: Ordovician trilobite, 487
Precambrian : acritarchs, Norway, 287 ; ediacaran
fossils, Canada, 945
Procolophonidae incertae sedis, 882
Prohalecites porroi, 156

Prohexagonaria favia sp. nov., 811; gotlandica sp.
nov., 812

Promissum pulchrum , 583
Protolloydolithus sp. nov., 490
Pseudoplankton, 359
Pteridinium sp., 963
Puzosia ( Pusozia ) sp., 90

Q

Quasicaecilia texana gen. et sp. nov., 895
Quaternary: birds, Australia, 447

R

Reptiles: Triassic, England, 873
Rhenocrinidae gen. et sp. nov., 63
Rickards, R. B. See Theron, J. N., Rickards, R. B.
and Aldridge, R. J.

Rogers, M. H. See Rogers, M. J., Donovan, D. T.
and Rogers, M. H.

Rogers, M. J., Donovan, D. T. and Rogers, M. H. A
deductive enquiry system for a palaeontological
database of museum material, 613
Romano, M. The trilobite Protolloydolithus from the
Middle Ordovician of north Portugal, 487
Rowley, M. J. See Baird, R. F. and Rowley, M. J.
Rugia acutirostris, 853; spinicostata , 856; spinosa,
854; tegulata, 853; tenuicostata , 852
Rushton, A. W. A. See Elughes, N. C. and Rushton,
A. W. A.

S

Scaphites (Scaphites) sp., 418

Sciponocerasl sp., 144

Scotland: Ordovician solute, 631; Ordovician grap-
tolite, 933

Scumulus sp., 865

Sekwia excentrica, 963

Selden, P. A. Lower Cretaceous spiders from the
Sierra de Montsech, north-east Spain, 257

Silurian: communities, Welsh Borderland, 209; grap-
tolites, Northern Ireland, 937; rugose corals,
northern Europe, 769; stromatoporoids, Gotland,
Sweden, 681

Simms, M. J. See Wignall, P. B. and Simms, M. J.

Simms, M. J. See Wignall, P. B. and Simms, M. J.

Sixt, S. See Malinky, J. M. and Sixt, S.

Skaiocrinus granulosus gen. et sp. nov., 66

Solute: Ordovician, Scotland, 631

South Africa: Cainozoic brachiopods, 313; Ordo-
vician conodont, 577

Spain: Cambrian eocrinoid, 249; Cretaceous spiders,
257

Spicer, R. A. and Parrish, J. T. Latest Cretaceous
woods of the central North Slope, Alaska, 225

INDEX

1011

Spiders: Cretaceous, Spain, 257
Spitsbergen: Ordovician nautiloid, 623
Stomiocrinus ferruginus sp. nov., 58
Strachan, I. A new genus of abrograptid graptolite
from the Ordovician of southern Scotland, 933
Streptograptus plumosus , 938

Stridsberg, S. Orientation of cephalopod shells in
illustrations, 243

Stromatoporoids : Silurian, Gotland, Sweden, 681
Sumitomoceras spp. juv., 395
Surlyk, F, See Johansen, M. B. and Surlyk, F.
Sutherland, P. K. See Jell, J. S. and Sutherland, P. K.
Sweden : Ordovician nautiloid, 623 ; Silurian stromato-
poroids, 681

T

Tachet, FI. See Hugheney, M., Tachet, H. and
Escuillie, F.

cf. Tanystropheus sp., 885

Taphonomy: Silurian stromatoporoids, 681

Tapinocrinus spinosus , 65

Tarrantoceras cuspidium , 134, 392; exile, 394; multi-
costatum , 134; sellar dsi, 130
Taylor, M. A. See Milner, A. R., Gardiner, B. G.,
Fraser, N. C. and Taylor, M. A.

Taylor, P. D. Preservation of soft-bodied and other
organisms by bioimmuration - a review, 1
Taylor, P. D. Bioimmured ctenostomes from the
Jurassic and the origin of the cheilostome Bryozoa,
19

Terebratulina chrysalis, 841 ; faujasii, 842; gracilis,
844; longicollis, 846; subtilis, 848; cf. rigida, 848
Teudopsis schuebleri, 201 ; subcostata, 202
Teuthids: Jurassic, Yorkshire, 193
Theron, J. N„ Rickards, R. B. and Aldridge, R. J.
Bedding plane assemblages of Promissum pulchrum ,
a new giant Ashgill conodont from the Table
Mountain Group, South Africa, 577

Tintori, A. The actinopterygian fish Prohalecites from
the Triassic of northern Italy, 155
Torrowangea rosei, 973

Triassic: fish, Italy, 155; vertebrates, England, 873
Trilobites: Cambrian, Wales, 429; classification, 529;
exoskeleton biomechanics, 749; eyes, Cambrian,
China, 91 1 ; Ordovician, Portugual, 487
Turrilites ( Turrilites ) acutus, 145; dearingi , 146

U

USA: Carboniferous crinoids drilled by gastropods,
743; Carboniferous hyoliths, 343; Cretaceous am-
monites, 75, 379; Cretaceous woods, 225; Permian
microsaur amphibian, 893

Ubaghs, G. and Vizcaino, D. A new eocrinoid from
the Lower Cambrian of Spain, 249

V

Valhalloceras flower i sp. nv., 628

Vidal, G. Giant acanthomorph acritarchs from the
Upper Proterozoic in southern Norway, 287

Vizcaino, D. See Ubaghs, G. and Vizcaino, D.

W

Wales: Cambrian trilobite, 429; Silurian marine
communities, 209

Webster, G. D. New Permian crinoids from Australia,
49

Wignall, P. B. and Simms, M. J. Pseudoplankton, 359

Wilmot, N. V. Biomechanics of trilobite exoskeletons,
749

Z

Zhang Xi-Guang and Clarkson, E. N. K. The eyes of
Lower Cambrian eodiscid trilobites, 91 1

VOLUME 33

Palaeontology

1990

PUBLISHED BY THE
PALAEONTOLOGICAL ASSOCIATION
LONDON

Dates of Publication of Parts of Volume 33

Part 1, pp. 1-256
Part 2, pp. 257-502
Part 3, pp. 503-748
Part 4, pp. 749-1000

21 March 1990
31 May 1990
7 August 1990
16 November 1990

THIS VOLUME EDITED BY M. J. BENTON, J. E. DALINGWATER, D. EDWARDS,
P. D. LANE, C. R. C. PAUL, P. A. SELDEN AND P. D. TAYLOR

Date of Publication of Special Paper in Palaeontology
Special Paper No. 43. 5 April 1990

© The Palaeontological Association , 1990

Printed in Great Britain
by Cambridge University Press

CONTENTS

Part

Aitken, J. D. See Narbonne, G. Y. and Aitken, J. D.

Aldridge, R, J. See Theron, J. N., Rickards, R. B. and Aldridge, R. J.

Allmon, W. D., Nieh, J. C. and Norris, R. D. Drilling and peeling of turritelline gastropods
since the late Cretaceous 3

Baird, R. F. and Rowley, M. J. Preservation of avian collagen in Australian Quaternary cave
deposits 2

Baker, P. G. The classification, origin and phylogeny of thecideidine brachiopods 1

Baumiller, T. K. Non-predatory drilling of Mississippian crinoids by platyceratid gastropods 3
Brunton, C. H. C. and FIiller, N. Late Cainozoic brachiopods from the coast of Namaqualand,
South Africa 2

Carroll, R. L. A tiny microsaur from the Lower Permian of Texas: size constraints in
Palaeozoic tetrapods 4

Clarkson, E. N. K, See Zhang Xi-Guang and Clarkson, E. N. K,

Cobban, W. A. See Kennedy, W. J. and Cobban, W. A.

Crampton, J. S. A new species of late Cretaceous wood-boring bivalve from New Zealand 4
Donovan, D. T. See Rogers, M. J., Donovan, D. T. and Rogers, M. H.

Doyle, P. Teuthid cephalopods from the Lower Jurassic of Yorkshire I

Escuillie, F. See Hugueney, M., Tachet, H. and Escuillie, F.

Evans, D H. and King, A. H. The affinities of early oncocerid nautiloids from the lower
Ordovician of Spitsbergen and Sweden 3

Evans, S. E., Milner, A. R. and Mussett, F. A discoglossid frog from the Middle Jurassic of
England 2

Fortey, R. A. Ontogeny, hypostome attachment and trilobite classification 3

Fraser, N C. See Milner, A. R., Gardiner, B. G., Fraser, N. C. and Taylor, M. E.

Gardiner, B. G. See Milner, A. R., Gardiner, B. G., Fraser, N. C. and Taylor, M. E.
FIarding, I. C. Palaeoperidinium cretaceum : a brackish-water peridiniinean dinoflagellate from
the early Cretaceous 1

Harper, L. See Martill, D. M. and Harper, L.

Haynes, J. R. The classification of the Foraminifera - a review of historical and philosophical
perspectives 3

Hiller, N. See Brunton, C. H. C. and Hiller, N.

Hughes, N. C. and Rushton, A. W. A. Computer-aided restoration of a late Cambrian
ceratopygid trilobite from Wales, and its phylogenetic implications
Hugueney, M., Tachet, H. and Escuillie, F. Caddisfly pupae from the Miocene Indusial
Limestone of Saint-Gerand-le-Puy, France

Jefferies, R. P. S. The solute Dendrocystoides scoticus from the upper Ordovician of Scotland

and the ancestry of chordates and echinoderms 3

Jell, J. S. and Sutherland, P. K. The Silurian rugose coral genus Entelophyllum and related
genera in northern Europe 4

Johansen, M. B. and Surlyk, F. Brachiopods and the stratigraphy of the Upper Campanian
and Lower Maastrichtian of Norfolk, England 4

Johnson, A. L. A. and Lennon, C. D. Evolution of gryphaeate oysters in the mid-Jurassic of
western Europe 2

Kennedy, W. J. and Cobban, W. A. Cenomanian ammonite faunas from the Woodbine
Formation and lower part of the Eagle Ford Group, Texas I

Kennedy, W, J. and Cobban, W. A. Cenomanian micromorphic ammonites from the Western
Interior of the USA 2

Kershaw, S. Stromatoporoid palaeobiology and taphonomy in a Silurian biostrome on
Gotland, Sweden 3

King, A. H. See Evans, D. H. and King, A. H.

Page

595

447

175

743

313

893

981

193

623

299

529

35

503

429

495

631

769

823

453

75

379

681

Lennon, C. D. See Johnson, A. L. A. and Lennon, C. D.

Lesperance, P. J. Cluster analysis of previously described communities from the Ludlow of the
Welsh Borderland 1 209

Loydell, D. L. On the graptolites described by Baily (1871) from the Silurian of Northern

Ireland and the genus Streptograptus Yin 4 937

Malinky, J. M. and Sixt, S. Early Mississipian Hyolitha from northern Iowa 2 343

Martill, D. M. Predation on Kosmoceras by semionotid fish in the Middle Jurassic Lower

Oxford Clay of England 3 739

Martill, D. M. and Harper, L. An application of critical point drying to the comparison of

modern and fossilized soft tissues of fishes 2 423

Milner, A. R., Gardiner, B. G., Fraser, N. C. and Taylor, M. E. Vertebrates from the
Middle Triassic Otter Sandstone Formation of Devon 4 873

Milner, A. R. See Evans, S. E., Milner, A. R. and Mussett, F.

Mussett, F. See Evans, S. E., Milner, A. R. and Mussett, F.

Nieh, J. C. See Allmon, W. D., Nieh, J. C. and Norris, R. D.

Narbonne, G. Y. and Aitken, J. D. Ediacaran fossils from the Sewki Brook area, Mackenzie

Mountains, northwestern Canada 4 945

Norris, R. D. See Allmon, W. D., Nieh, J. C. and Norris, R. D.

Nuttall, C. P. Review of the Caenozoic heterodont bivalve superfamily Dreissenacea 3 707

Parrish, J. T. See Spicer, R. A. and Parrish, J. T.

Rickards, R. B. See Theron, J. N., Rickards, R. B. and Aldridge, R. J.

Rogers, M. H. See Rogers, M. J., Donovan, D. T. and Rogers, M. H.

Rogers, M. J., Donovan, D. T. and Rogers, M. H. A deductive enquiry system for a

palaeontological database of museum material 3 613

Romano, M. The trilobite Protolloydolithus from the middle Ordovician of north Portugal 2 487

Rowley, M. J. See Baird, R. F. and Rowley, M. J.

Rushton, A. W. A. See Hughes, N. C. and Rushton, A. W. A.

Selden, P. A. Lower Cretaceous spiders from the Sierra de Montsech, north-east Spain 2 257

Simms, M. J. See Wignall, P. B. and Simms, M. J.

Sixt, S, See Malinky, J. M. and Sixt, S.

Spicer, R. A. and Parrish, J. T. Latest Cretaceous woods of the central North Slope, Alaska 1 225

Strachan, I. A new genus of abrograptid graptolite from the Ordovician of southern Scotland 4 933

Stridsberg, S. Orientation of cephalopod shells in illustrations 1 243

Surlyk, F. See Johansen, M. B. and Surlyk, F.

Sutherland, P. K. See Jell, J. S. and Sutherland, P. K.

Tachet, H. See Hugueney, M., Tachet, H. and Escuillie, F.

Taylor, M. E. See Milner, A. R., Gardiner, B. G., Fraser, N. C. and Taylor, M. E.

Taylor, P. D. Preservation of soft-bodied and other organisms by bioimmuration - a review 1 1

Taylor, P. D. Bioimmured ctenostomes from the Jurassic and the origin of the cheilostome

Bryozoa 1 19

Theron, J. N., Rickards, R. B. and Aldridge, R. J. Bedding plane assemblages of Promissum
pulchrum , a new giant Ashgill conodont from the Table Mountain Group, South Africa 3 577

Tintori, A. The actinopterygian fish Prohalecites from the Triassic of northern Italy I 155

Ubaghs, G. and Vizcaino, D. A new eocrinoid from the Lower Cambrian of Spain I 249

Vidal, G. Giant acanthomorph acritarchs from the Upper Proterozoic in southern Norway 2 287

Vizcaino, D. See Ubaghs, G. and Vizcaino, D.

Webster, G. G. New Permian crinoids from Australia 1 49

Wignall, P. B. and Simms, M. J. Pseudoplankton 2 359

Wilmot, N. V. Biomechanics of trilobite exoskeletons 4 749

Zhang Xi-Guang and Clarkson, E. N. K. The eyes of Lower Cambrian eodiscid trilobites 4 911

NOTES FOR AUTHORS

The journal Palaeontology is devoted to the publication of papers on all aspects of palaeontology. Review articles are
particularly welcome, and short papers can often be published rapidly. A high standard of illustration is a feature of the
journal. Four parts are published each year and are sent free to all members of the Association. Typescripts should conform
in style to those already published in this journal, and should be sent to Dr Dianne Edwards, Department of Geology,
University of Wales College of Cardiff CF! 3YE, who will supply detailed instructions for authors on request (these are
published in Palaeontology 1990, 33, pp. 993-1000).

Special Papers in Palaeontology is a series of substantial separate works conforming to the style of Palaeontology.

SPECIAL PAPERS IN PALAEONTOLOGY

In addition to publishing Palaeontology the Association also publishes Special Papers in Palaeontology. Members may
subscribe to this by writing to the Membership Treasurer: the subscription rate for 1990 is £45 00 (U.S. $66) for Institutional
Members, and £20-00 (U.S. $32) for Ordinary and Student Members. A single copy of each Special Paper is available on a
non-subscription basis to Ordinary and Student Members only , for their personal use, at a discount of 25 % below the listed
prices: contact the Marketing Manager. Non-members may obtain Nos 35—40 (at cover price) from Basil Blackwell Ltd.,
Journal Subscription Department, Marston Book Services, P.O. Box 87, Oxford OX2 0DT, England, and older issues from
the Marketing Manager. For all orders of Special Papers through the Marketing Manager, please add £1-50 (U.S. $3) per
item for postage and packing.

PALAEONTOLOGICAL ASSOCIATION PUBLICATIONS

Special Papers in Palaeontology

For full catalogue and price list, send a self-addressed, stamped A4 envelope to the Marketing Manager. Numbers 2-1 1 and
13-30 are still in print and are available together with those listed below:

31. (for 1984): Systematic palaeontology and stratigraphic distribution of ammonite faunas of the French Coniacian, by
w. j. Kennedy. 160 pp ., 42 text-figs., 33 plates. Price £25 (U.S. $38).

32. (for 1984): Autecology of Silurian organisms. Edited by m. g. bassett and s. d. lawson. 295 pp ., 75 text-figs ., 13 plates.
Price £40 (U.S. $60f

33. (for 1985): Evolutionary Case Histories from the Fossil Record. Edited by j. c. w. cope and p. w. skelton. 202 pp ., 80
text-figs ., 4 plates. Price £30 (U.S. $45).

34. (for 1985): Review of the upper Silurian and lower Devonian articulate brachiopods of Podolia, by o. i. Nikiforova,
t. l. modzalevskaya and m. g. bassett. 66 pp ., 6 text-figs ., 16 plates. Price £10 (U.S. $15).

35. (for 1986): Studies in palaeobotany and palynology in honour of N. F. Hughes. Edited by d. j. batten and d. e. g.
briggs. 178 pp., 29 plates. Price £30 (U.S. $50).

36. (for 1986): Campanian and Maastrichtian ammonites from northern Aquitaine, France, by w. j. Kennedy. 145 pp., 43
text-figs., 23 plates. Price £20 (U.S. $35).

37. (for 1987): Biology and revised systematics of some late Mesozoic stromatoporoids, by rachel wood. 89 pp., 31 text-
figs., 7 plates. Price £20 (U.S. $35).

38. (for 1987): Taxonomy, evolution, and biostratigraphy of late Triassic-early Jurassic calcareous nannofossils, by p. r.
bown. 1 18 pp., 19 text-figs., 15 plates. Price £30 (U.S. $50).

39. (for 1988): Late Cenomanian and Turonian ammonite faunas from north-east and central Texas, by w. j. Kennedy,
131 pp., 39 text-figs., 24 plates. Price £30 (U.S. $50).

40. (for 1988): The use and conservation of palaeontological sites. Edited by p. r. crowther and w. a. Wimbledon. 200 pp.,
31 text-figs. Price £30 (U.S. $50).

41. (for 1989): Late Jurassic-early Cretaceous cephalopods of eastern Alexander Island, Antarctica, by p. j. howlett.
72 pp., 9 text-figs., 10 plates. Price £20 (U.S. $35).

42. (for 1989): The Palaeocene flora of the Isle of Mull, by m. c. boulter and z. kvacek. 149 pp., 23 text-figs., 23 plates.
Price £40 (U.S. $70).

Field Guides to Fossils and Other Publications

These are available only from the Marketing Manager. Please add £100 (U.S. $2) per book for postage and packing plus
£1-50 (U.S. $3) for airmail. Payments should be in Sterling or in U.S. dollars, with all exchange charges prepaid. Cheques
should be made payable to the Palaeontological Association.

1. (1983): Fossil Plants of the London Clay, by m. e. collinson. 121 pp., 242 text-figs. Price £7-95 (U.S. $12) (Members £6

or U.S. $9).

2. (1987): Fossils of the Chalk, compiled by e. owen; edited by a. b. smith. 306 pp., 59 plates. Price £11-50 (U.S. $18)

(Members £9-90 or U.S. $15).

3. (1988): Zechstein Reef fossils and their palaeoecology, by N. hollingworth and t. pettigrew. iv + 75 pp. Price £4-95

(U.S. $8) (Members £3.75 or U.S. $6).

1982. Atlas of the Burgess Shale. Edited by s. conway morris. 31 pp., 24 plates. Price £20 (U.S. $30).

1985. Atlas of Invertebrate Macrofossils. Edited by j. w. Murray. Published by Longman in collaboration with the
Palaeontological Association, xiii + 241 pp. Price £13-95. Available in the USA from Halsted Press at U.S. $24-95.

© The Palaeontological Association, 1990

Palaeontology

VOLUME 33 • PART 4

CONTENTS

Biomechanics of trilobite exoskeletons

N. v. wilmot 749

The Silurian rugose coral genus Entelophyllum and related genera
in northern Europe

J. S. JELL and P. K. SUTHERLAND 769

Brachiopods and the stratigraphy of the upper Campanian and
Lower Maastrichtian Chalk of Norfolk, England

M. B. JOHANSEN and F. SURLYK 823

Vertebrates from the Middle Triassic Otter Sandstone Formation
of Devon

A. R. MILNER, B. G. GARDINER, N. C. FRASER and M. A. TAYLOR 873

A tiny microsaur from the Lower Permian of Texas: size
constraints in Palaeozoic tetrapods

R. L. CARROLL 893

The eyes of Lower Cambrian eodiscid trilobites

ZHANG XI-GUANG and E. N. K. CLARKSON 911

A new genus of abrograptid graptolite from the Ordovician of
southern Scotland

I. STRACHAN 933

On the graptolites described by Baily (1871) from the Silurian of

Northern Ireland and the genus Streptograptus Yin

D. K. LOYDELL 937

Ediacaran fossils from the Sekwi Brook area, Mackenzie
Mountains, northwestern Canada

G. M. NARBONNE and J. D. AITKEN 945

A new species of late Cretaceous wood-boring bivalve from
New Zealand

J. S. CRAMPTON 981

Instructions to authors

THE EDITORS 993

Printed in Great Britain at the University Press , Cambridge

ISSN 0031-0239

INSTITUTION

S3 I H VH 8 IT

INSTITUTION

z

o

S3 1 H VH s n

INSTITUTION

siiavnan

INSTITUTION

o

'institution2

z
o

S3 1 u va an

r

m ^ xouusgx rn

NOlinillSNI NVIN0SH1IINS S3 I U VH 8 n”LI B R AR I ES^SMITHSONIAN^INSTITUTION

2 CO Z . CO ^

&

LI B RAR I ES“SMITHS0NIAN^ INSTITUTI0NW/N0lini!ISNrNVIN0SHiIWS</>S3 I 8 V3 8 11

MT> — CO — CO

_ J °

NOlinillSNI JNVIN0SH1IW$2S3 I UVH a nJLI BRAR I ES^SMITHSONIAN^INSTSTUTION^N
r“ v. z r* z r" z

o

LI BRAR I ES^SMITHSONIAN INSTITUTION^NOliniLLSNI NVINOSHlIlNS^Si I HVH an L

.• fn "3? Itfi 21

<

g 2

NOlinillSNI ^NVINOSHIIWS^SH I MVM SIl^LIB RAR S ES^SMITHSONIAN INSTITUTION N

CO “ M ™ CO

LIBRARIES SMITHSONIAN INSTITUTION NOlinillSNI NVINQSHilWS SlIUVaaiT L
^ 8=3 z ^ r- . z Z

m v ^ ^

N0linillSNrNVIN0SHiIlNSU/S3 I U Vil 3 11~LI BRAR! ES^SMITHSONIAN^NSTITUTION N

05 Z to Z ,v. to

< - s

5 % >

LI B RAR I ES^SMITHSONIAN^ INSTITUTION^NOIinillSNI^NVINOSHlIINS^Sa I U V8 8 II L

<o

NOlinillSNI NVINOSHilWS SBIBVaail LIBRARIES SMITHSONIAN INSTITUTION

O

z

IVASV''

-

'Vis''.

LIBRARIES SMITHSONIAN INSTITUTION NOlinillSNI NVIN0SH1IINS S3iava8ll L

2

0-J ✓ x . x — ^ „-v »* .XXX\ \ .'.' ’ — yX!' ' ■ .X m ^

70
>

SI

m xii^Dcx v> rn ^ ^5 N^>jiB!iX m

__ to _ w \ ~ to

NS NVIN0SH1IINS $3!9Vaan LIBRARIES SMITHSONIAN INSTITUTION NOlinillSNI NVINOSHlllAIS

S CO 2 . to 2 >v. OT 2

/•I ,.//.*'! I /^N 1 - -s

§ «| § %% i '?f' |

_ > 2 2 > 'JWg/ 2

LS^SMITHSONIAN2 INSTITUTION NOlinillSNI NVINOSHIIIAJS^SS I HViJ a II^LI B RAR I ES SMITHSONIAN
co ^ 5 ., . to „ ^ ~ to

= 3

2 ^fF' 5 5 V £ o

Nl“WlN0SHlt!ASS^S3 I HVH 9 n”*LI B RAR I ES^ SMITHSONIAN J INSTITUTION ^NOlIfUllSN! NVINOSHlllAIS
r* N, 2 r* 2 2 f~.

O .✓fSTiroX “ /^ateX O jwZfo,. m O /

:» 03

> M

^0'-™

m <■ 2 sL*va^x m x^osvi^x ^ m 1" m

:SWSMITHSONIAN”lNSTSTUTIONWNOIinillSNl"NVINOSHII!AiS S3 I MVH a II ~L! B R AR I ES SMITHSONIAN

to Z \ CO 2 to 2 > w ^

3 g >%%, a tpfe § /#% a wm § ■%% a

jr g W %/ffZ i m4l If^' 1 W i TOsSr

> X3^X s 'VVV >’ 2 Xfoost^ >■ 2 • .^

Nl NVINOSHHIAIS^SH I B VH 0 IT^LI B RAR I ES^SMITHSONIAN INSTITUTION NOIinillSNI_NVINOSHlllAIJ

5 \ 4/5 .. — . 5 to = to — ^ ^

m yfnaT^s. W m /T^X to to

8 O

o

2

c ^51 r: &s/’ if"' '■ *“ -'Win ~ \x&. c

:sw/ =!

•s SMITHSONIAN INSTITUTION NOlinillSNI NVINOSHlllAIS SlIHVHail LIBRARIES SMITHSONIAN
2 /ye •■*•“ 1 P /

s '•* S' K -r« /_ iS?t > f Fja.'iX f™1 nVj?M f m a X I— j v -n / A7) F ^ -"A e- ./ r..f

03
SI
>

70

m ^ ns^uu5^ m W Nv/uTIs'jX m

INI NVINOSHIIINS S3 I a VH a ll“=UBRARfE$t/,SMITHSONIAN“lNSTITUTION NOlinillSNI NVINOSHIIIAI!
CO Z .,. to ... CO 2

< /

o

CO

2 _. A5C44T// t " ^ 5 i

ES^SMITHSONIAN^ INSTITUTION^NOliniliSNI^NVINOSHIUAIS^Sa I aVH 9 ll^LI B R AR I ES^SMITHSONIAr
to ^ to ™ ._ . CO ^ to

UJ

^ 1 ilf' < (C'il i &f"fa s

^ £U \T««5y ^ XV^ £ 5 X^0y £ ^

INI^NVINOSHIIIAIS^SS IHVaail^UBRARI ES^ SMITHSONIAN^’lNSTITUTION Z NOlinillSNI^NVINOSHlIlAI!
r* > 2 r- 2 r- z >

1 1 ^

m

to — CO

CC CMITUCAMU M IMtTITI ITIAM llnl 1 fill IPkll