/ »v

saiavaan libraries Smithsonian institution Nosiruiisw nvinoshiiins saiavaan u

CO 4 00 2 >^®32r35v m O ^

2

I? /$ew&\ S / a -4 w ASitffwJ'X ~ /wasjs\ co x

33

>

50

m xgy.p.cx ^ m m

in — to i to \ <£- {/)

NSTITUTION^ NOlinillSNI NVIN0SH1IINS S3 I HVH 0 11 LIBRARIES SMITHS0NIAN“iNSTITUT!0N ' NC

z 15 z - j*’ . w z ... co

X

to ii

o £

C/> 2 CO

| /sSSSk §

X VTfer^T^I/ P

^ i > W" 2 ’T > N^iiis^x 2 >

sa i avaa n librari es^smithsonian institution NouniiiSNi nvinoshuiais^sb i avaa ii^li

“ CO ~ 0 0 — f/i —

„ d Vsfc ^fj 5 teF Jo] 5 wm. ' <

a hi \rt%fALlir§) cc

O vw„c.y m H CO \M

NSTITUTION^NOlinillSNI^NVINOSHimS^Sa lavaan^LIBRAR! E$Z SMITHSONIAN^ INST I TUT ION 2 NC

sGSa§?\ 2 /^g&v 5^2 2 5 °

<o

< >

. _ , *4*^-2

m s N^vasv^ m N^osvajJX 2 m

~ c/> — CO 2 CO _

saiavaan libraries Smithsonian institution NouniiiSNi nvinoshiiws saiavaan u

2 <5 Z 00 2 £0 2

NSTITUTION NOIinilJLSNI__ NVINOSHIISAIS^SB lavas ll^L! B RAR I ES^SMITHSONIAN^INSTITUTION ^NC

CO z \ to „ = co — CO

CcS s. «k. UJ ZilSTIT(y>^ . - Ll! ^ 11 H i

c .. „

O ' Z -^VASV^Z q ^ • O “

-J 2 ~J Z _j 2 J

saiavaan libraries Smithsonian institution NouniiiSNi nvinoshiiws S3iavaan n

m ■vO^.dcJV

CO — co

INSTITUTION NOlinillSNI NVIN0SH1IINS S3IHVH8

co

TO
>

TO

>^C4Pxy - _

ioiiisjjx m 5£J X^iiisjjX m

c/> \ 5 </>

1 LIBRARIES SMITHSONIAN INSTITUTION N(

2 .-,. tO

— ^\<c §• <

z -X

o ^ 5 ; > X; 2

S3 1 avaa ii libraries Smithsonian institution NouniiiSNi nvinoshiiws S3iavaa

if) 7Z if) — if)

pz*sm va P - Li - /fT^r ma\ ri vX\K - /S’

-v . NJv> q; /£>>

<
cc

■vy m 5 v/^rvy m \p^ma^t/ - W m

INSTITUTION"' NOlinillSNI ^NVINOSHIIWS 2 S 3 I H V a a !1 J LI B R A R I E SMITHSONIAN^ INSTITUTION 2 NC

Z r- > Z r- Z r~ Z

CQ

O X^osv^y _ ' O

SMITHSONIAN ^ INSTITUTION NOlifUllSNI^NVINOSHlIINS S3 I H VH S 11^ LIBRARIES^ SMITHSONIAN J

§ ✓ „ >" /4^\ 1 /^Sx 5 Jfes I 2 ° .../

• CD
7Q
>

<%$*% 50

m ^ m rn Xo^dc^ ^ ™

"■ - A. CO =—

nvinoshiiins saiavasn libraries Smithsonian institution, Noiiniiism nvinoshiiws j

Z W Z ^ |2 1^1 § xSHSvTX s

«c£ S / . ^ aa\ .-• ^ ^ v.- 5,6

**** / tT* e“!

- :/i

co ir#

_ ^ |%r g

^SMITHSONIAN^INSTITUTlON^NOlinillSNI NVIN0SH1IWSWS3 I H VM 8 I1_ II B R AR I ES SMITHSONIAN _

1 — CO “ C/> _____ 2 Z

H /«X » .f/J /,

A ^ /-»,r*A\ or fc&*r&A d K [§£? &$ \ 3 /yp “ ■.yj's:

q X^H pc, ^ _ N^ojixsgx O " 2 X^XIlS^X^ Q Vok.qi^ _ '^5?^ O

JNVIN0SHimS^S3 I ava a n JU B RAR I ES^SMITHSONIAN^INSTITUTION^NOlinillSNI^NVINOSHimS^

£ 5

£ \3» £

- W XQ^/ - " • rn | X^VA^ rn

;WSMITHSONIAN“lNSTITUTIONi0NOIinilISNrNVINOSHimS S3 I HVB 8 11 U B RAR I ES SMITHSONIAN
CO Z \ CO 2 £ 5 \ z ^

i jfikirii jsm mi

ȴ7'w: vJ Pi

JfiW’ffi:. to (pj

| ,#/^ | "

> | £ S5 z a

NViNOSHiiws^sa i ava a n libraries smithsonian institution NoiiniiiSNi_NviNOSHims

■ ' w 5 ___ .2 ~ 5 ^ 2 \ U

SMITHSONIAN INSTITUTION NOIlfUllSNI NVIN0SH1IINS S3IBVH8n LIBRARIES^ SMITHSONIAN^
z *” ^ z r* z __ 2 o

m Nt^-pc^ rn ^ m 2 r

b \ «“ (/) •*“ ^

ii_nvinoshiiws saiavaan "libraries smithsonian“institution NoiiniusNi nvinoshiiws

2 ~ 2 .< 1 /^rx 1 4jg-s 1 ..< '

./ A/jr * z — < /Mt'A -t- -i ..y ^

2 -.«X. =g

S^SMITHSONIAN^INSTITUTION^NOIinillSNI NVINOSHilSNS^SB I HVH 0 11 LI B RAR I ES SMITHSONIAN
— co — co ^ :

q pcy — N^UllS^x q v _ x-Miiiv^ O

JI_,NVIN0SH1IINSZS3 I ava 8 ll^LI B RAR I ESZSMITHSONIAN“J|NSTITUTION“NOIinilISNrNVINOSHlllNS

r- > z r- z n _____ z v

Xx O xScccDX “ O *A — /®N O

CD

CO

%

Palaeontology

VOLUME 36 • PART 3 • SEPTEMBER 1993

Published by

The Palaeontological Association • London
Price £40-50

THE PALAEONTOLOGICAL ASSOCIATION

(Registered Charity)

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

COUNCIL 1993-1994

President : Dr W. D. I. Rolfe, National Museums of Scotland, Chambers Street, Edinburgh EH1 1JF
Vice-Presidents'. Dr A. W. Owen, Department of Geology and Applied Geology, The University, Glasgow G12 8QQ
Dr M. E. Collin son, Department of Geology, Royal Holloway and Bedford New College,

University of London, Egham, Surrey, TW20 OEX
Treasurer-. Mr P. S. Clasby, 12 Haglane Copse, Lymington, Hants S041 8DT
Membership Treasurer : Dr M. J. Barker, Department of Geology, University of Portsmouth, Barnaby Road,

Portsmouth POl 3QL

Institutional Membership Treasurer : Dr J. E. Francis, Department of Earth Sciences, The University, Leeds LS2 9JJ
Secretary. Dr J. A. Crame, British Antarctic Survey, High Cross, Madingley Road, Cambridge CB3 OET
Newsletter Reporter'. Dr D. Palmer, c/o Department of Earth Sciences, Downing Street, Cambridge CB2 3EQ
Marketing Manager (Sales): Dr L. Cherns, Department of Geology, University of Wales College of Cardiff CF1 3YE
Marketing Manager (Publicity): Dr P. B. Wignall, Department of Earth Sciences, The University, Leeds LS2 9JJ
Public Relations Officer: Dr P. A. Allison, P.R.I.S., The University, P.O. Box 227, Whiteknights, Reading RG6 2AB
Newsletter Editor: Dr R. B. Rickards, Department of Earth Sciences, University of Cambridge, Cambridge CB2 3EQ
Editor of Field Guides to Fossils : Dr E. A. Jarzembowski, Booth Museum of Natural History, 194 Dyke Road,

Brighton BN1 5AA

Editors

Dr J. E. Dalingwater, School of Biological Sciences, Williamson Building, The University, Manchester M13 9PL (co-opted)
Dr P. D. Lane, Department of Geology, University of Keele, Keele, Staffordshire ST5 5BG
Dr P. Doyle, School of Earth Sciences, University of Greenwich, Brewhouse Lane, London El 9NU
Dr A. R. Milner, Department of Biology, Birkbeck College, Malet Street, London WC1E 7HX
Dr C. J. Cleal, 37 Lipscombe Close, Newbury, Berkshire, RG14 5JW
Dr R. M. Owens, Department of Geology, National Museum of Wales, Cardiff CF1 3NP

Other Members: Dr B. M. Cox, British Geological Survey, Keyworth, Nottingham NG12 5GG
Dr P. Smith, School of Earth Sciences, University of Birmingham, Birmingham B15 2TT

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 Palaeoanthropology,
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: Professor F. Alvarez, Departamento de Geologia, Universidad de Oviedo, Cl. Jesus
Arias de Velasco, s/n. 33005 Oviedo, Spain. Japan: Dr I. Hay ami, 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, Fredskowej 4, 2840 Holte, Denmark. USA : Professor A. J. Rowell, Department of Geology, University of
Kansas, Lawrence, Kansas 66044. Professor N. M. Savage, Department of Geology, University of Oregon, Eugene, Oregon
97403. Professor M. A. Wilson, Department of Geology, College of Wooster, Wooster, Ohio 44961. Germany: Professor
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 1993 are:
Institutional membership . . . £80 00 (U.S. $160) Student membership . . . . £11 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 J. E.
Francis, Department of Earth Sciences, The Lfniversity, Leeds LS2 9JJ. 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 M. J. Barker, Department of Geology, University of Portsmouth,
Barnaby Road, Portsmouth POl 3QL. Subscriptions cover one calendar year and are due each January; they should be
sent to the Membership Treasurer. All members who join for 1993 will receive Palaeontology, Volume 36, 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, UK. For older issues contact
the Marketing Manager.

US Mailing: Second class postage paid at Rahway, New Jersey. Postmaster: send address corrections to Palaeontology,
c/o Mercury Airfreight International Ltd, 2323 EF Randolph Avenue, Avenel, NJ 01001, USA (US mailing agent).

Cover: Reconstruction of Paraostenia voultensis Secretan from the Middle Jurassic of the Ardeche, France, preying upon
a coleoid. This is a typical thylacocephalan, a recently recognized arthropod group of uncertain, but probably crustacean,
affinity, ranging from Silurian to Cretaceous, x 1.25. Reproduced by permission of the Royal Society of Edinburgh and
Dr W. D. I. Rolfe from Transactions of the Royal Society of Edinburgh, 76, 398, fig. 4.

^momAhr

OCT 2& **
_ librabSS^

NEOPROTEROZOIC (VENDIAN) PHYTOPLANKTON
FROM THE SIBERIAN PLATFORM, YAKUTIA

£y MALGORZATA MOCZYDLOWSKA, GONZALO VIDAL
and VALERIA A. RUDAVSKAYA

Abstract. Seven new species of comparatively large Neoproterozoic organic-walled acritarchs ( Appendi -
sphaera grandis, A. fragilis, A. tenuis , A.? tabifica, Cavaspina basiconica, Tanarium irregulare, T. tuberosum )
are reported from two drilling sites in Yakutia, eastern Siberia. Two previously known form-species (Cavaspina
acuminata comb. nov. and Tanarium conoideum) are also emended. The acritarchs derive from siliciclastic rocks
of the Khamaka Formation, the lowermost part of the Vendian to Cambrian sedimentary succession in the
central part of the Siberian Platform. By comparison with assemblages from the Ediacaran Pertatataka
Formation, the acritarchs in the Khamaka Formation are considered to indicate an Ediacaran age. The study
confirms the broad environmental and geographical distribution of Neoproterozoic (late Vendian) plankton,
and that they are diverse in rocks reflecting a range of depositional settings.

The biotic changes near the Proterozoic- Phanerozoic boundary are currently the subject of intense
debate (Cowie and Brasier 1989; Brasier 1990, 1992). While the causes and effects of these changes
remain problematic (Brasier 1992; Kaufman et al. 1992), their magnitude is being revealed with
increasing clarity by a steadily growing record of well-preserved, diagnostic acritarchs from strata
straddling the Neoproterozoic/Lower Cambrian boundary (Awramik et al. 1985; Yin 1985a,
19856, 1987; Zang 1988, 1992; Zang and Walter 1989, 1992; Moczydlowska 1991; Vidal and
Moczydlowska 1992). Here we report on Neoproterozoic organic-walled phytoplankton (acritarchs)
from two drilling sites in Yakutia, eastern Siberia, including seven new species and two previously
described form-species that are emended.

GEOLOGICAL FRAMEWORK

Eastern Siberia offers exceptionally well-developed Neoproterozoic to Lower Cambrian successions
that contribute substantially to the understanding of Neoproterozoic and early Cambrian biotic
change (Sokolov and Ivanovskij 1985). Proterozoic sedimentary rocks overlie the crystalline
basement with major disconformity. The lower part of the Mesoproterozoic to Cambrian
sedimentary succession formed on a carbonate platform that developed above siliciclastic deposits.
It is well exposed along the rivers Lena, Aldan, Maya, Olenek, Kotuj, Kotujkan, Nemakit-Daldyn,
Khorbusuonka and Miroedikha (Rozanov et al. 1969; Rozanov and Sokolov 1984; Khomentovski
1985) and is also widely known from subsurface borehole sections on the Siberian Platform. During
Riphean time, detrital and carbonate deposition was largely in troughs and pericratonic basins. The
Neoproterozoic (Yudomian or Vendian) succession developed on a stable carbonate platform and
lies unconformably or disconformably on Riphean strata and/or basement rocks. From more or
less reliable biostratigraphical evidence, Yudomian strata are generally interpreted as time-
equivalent to the Vendian of the East European Platform (Khomentovski 1985). By comparison
with the East European successions, recent chemostratigraphical data indicate that lower Yudomian
rocks are of latest Neoproterozoic age (Knoll, pers. comm. 1992).

Irrespective of their depositional settings, a general feature of these deposits is the almost total
lack of metamorphism. This has resulted in the generally excellent preservation of organic-walled
fossils.

IPalaeontology, Vol. 36, Part 3, 1993, pp. 495-521, 1 pl.|

© The Palaeontological Association

496

PALAEONTOLOGY, VOLUME 36

Shields

Fold belts

m

Riphean-

Palaeozoic

Mesozoic-
C e n o z o ic

Platforms

□

Precambrian

Palaeozoic

Main structural
divisions

Regional tectonic
boundaries

Investigated
dr illcores

text-fig. 1 . Simplified geological sketch-map of Siberia showing main structural units and location of
investigated drillholes within regional tectonic units. H, Helsinki; T, Tallinn; SP, St Petersburg;
N, Novosibirsk; Y, Yakutsk; M, Magadan; UB, Ulan Bator. Figures indicate: 1, Baltic Shield; 2, Anabar

MOCZYDLOWSKA ET AL.: VENDIAN PHYTOPLANKTON

497

This paper is concerned with assemblages from the lowermost part of the sedimentary succession
in the Nepa-Botuoba region (Text-fig. 1) in the central part of the Siberian Platform. Rocks in this
region have been dated as Vendian and Early Cambrian. This subsurface sequence was penetrated
by numerous hydrocarbon exploration boreholes, resulting in the geology of the area being
relatively well documented, largely in internal company reports.

The lower part of the succession in the study area (Text-fig. 2) consists of siliciclastic rocks of
variable thickness succeeded by thick carbonates, which are, in turn, overlain by evaporites. The
siliciclastic deposits consist of sandstones, mudstones and shales ranging in thickness from
approximately 30 to 300 m (Rudavskaya and Vasileva 1989). Occasionally, conglomerates occur at
the base of the sequence (Borehole V-Ch 96; Rudavskaya and Vasileva 1989). Sandstones are
predominant, forming thick units with numerous discontinuity surfaces and unconformities.
Mudstones occur as thin intercalations in the sandstones or as discrete beds reaching 20 to 40 m in
thickness (for example, boreholes 845 and 611; Rudavskaya and Vasileva 1989). Carbonates
overlying the siliciclastic portion of the sequence range in thickness from 170 to 400 in. They consist
largely of dolostones or more rarely limestones, and clayey dolostones (Grausman and Zhernovskij
1989; Rudavskaya and Vasileva 1989). The carbonate succession is overlain by thick (250 in)
evaporites, comprising mainly of halite, with lesser amounts of anhydrite.

Although variably complete in different parts of the basin, the succession constitutes a single
transgressive-regressive depositional cycle. At its greatest development, the succession is more than
700 m thick (Borehole 606; Rudavskaya and Vasileva 1989). Detailed facies reconstruction and
basin analysis is not yet possible, since drillcores are available for only a limited portion of the
sequence, and some of the data are not accessible. As a consequence, the present level of knowledge
is not enough to allow comparison with other depositional grand cycles (Aitken 1978). However,
from the available evidence it is clear that the sequence was deposited under shallow marine
conditions on a stable platform, on which almost exclusively fine-grained detrital deposits and
carbonates accumulated. The depositional cycle began with a transgression over the extensively
peneplained crystalline basement, and extended through late Vendian-early Cambrian times.

The overlying carbonates represent a thick sequence that formed part of an extensive carbonate
piatform occupying a vast area of present-day Siberia. Shallowing-up interrupted carbonate
sedimentation on the platform, which was followed by accumulation of evaporites (Text-fig. 2).

FOSSIL RECORD

The fossil record in the studied sequence comprises abundant acritarchs, abandoned cyanobacterial
sheaths and small shelly fossils. The lowermost recorded occurrence of small shelly fossils, including
hyolithids ( Conotheca mammilata Missarzhevskij, Turcutheca sp.), problematic shells resembling
obolellid brachiopods and possible archaeocyathans, is in the upper part of the Yuryakh Formation
(Text-fig. 2; Grausman and Zhernovskij 1989; Rudavskaya and Vasileva 1989); these fossils were
considered by the latter authors to indicate an early Cambrian age. However, their taxonomic
assignment and age are uncertain; thus, the chronostratigraphical position of the Yuryakh
Formation remains questionable (Grausman and Zhernovskij 1989; Text-fig. 2).

In the Bilir Formation, which overlies the Yuryakh Formation, there is a rich association of
various shelly fossils, including hyolithids, hyolithelmintids, brachiopods, gastropods, chancellorids,
tommotids and archaeocyathans. These fossils are more convincingly early Cambrian (Tommotian
to early Atdabanian; Grausman and Zhernovskij 1989).

Acritarchs and cyanobacterial microfossils occur at specific levels in the siliciclastic and carbonate
succession and numerous taxonomically and stratigraphically undiagnostic cyanobacterial sheaths

Shield; 3, Aldan Shield; 4, Siberian Platform; 5, West Siberian Platform; 6, East European Platform;
7, Verkhoyansk-Chukotka-Kamchatka area; 8, Primorye area; 9, Baikal fold zone; 10, Altaj-Sayan fold zone;
11, Urals; 12, Scandinavian Caledonides. Compiled after Rundquist (1984) and Geological-Prospecting Oil
and Gas Review Map of Yakutian ASSR 1 : I 000000, Ministry of Geology of the USSR, 1990.

498

PALAEONTOLOGY, VOLUME 36

OZERO 761

ZAPAD 742

♦ A

* A

□ Sandstone
[ | Granitoid

Shale and mudstone

A Acritarch occurrence

Proterozoic

&

Archean

text-fig. 2. Generalized stratigraphical sections of the investigated sequences penetrated at the Zapad 742 and
Ozero 761 drilling sites in the Siberian Platform, Yakutia.

and spheromorphic microfossils occur throughout (Rudavskaya and Vasileva 1989; Grausman and
Zhernovskij 1989; Text-fig. 2).

MATERIAL AND TAPHONOMY

Microfossils were studied in permanent strew slides prepared by conventional palynological
maceration techniques. The samples investigated are from dark-grey, thin-bedded, kerogen-rich
mudstones and shales of the Khamaka Formation at depths betweeri 1887 and 1894 m in Borehole
Zapad 742, and between 1876 and 1884 and 1770 and 1790 m in Borehole Ozero 761 (Text-fig. 2).
Additional specimens of acritarchs from the earlier collections of one of us (V.A.R.) marked with
the acronym VNIGRI have been re-examined and taxonomically re-assigned. These additional
specimens originated from boreholes Talakan 806, Talakan 823 and Zapad 844 in the principal
study region, from Dyudan 291-0 and Nakyn 295-0 boreholes situated in the Syugdzher Saddle to
the northeast of the Nepa-Botuoba region, and from the Borehole Charchyk 1 from the Lena-
Anabar Depression (near the mouth of the Olenek River; Text-fig. 1 ). Microfossils from the Zapad
742 and Ozero 761 successions are exceptionally well-preserved, showing particularly well
morphological elements such as processes. Nevertheless, rare instances of corrosion (Text-figs 6a-b,
16) occur, and vesicle collapse following the loss of cell turgescence and ensuing sediment
compaction has resulted in the formation of folds and wrinkles.

The colour of organic vesicles ranges from pale yellow to very light brown, a feature that indicates
low-grade thermal alteration. The palaeotemperature to which the host rocks were heated is inferred

MOCZYDLOWSKA ET AL.: VENDIAN PHYTOPLANKTON

499

to have been approximately 50-70 °C, corresponding to the diagenesis and protokatagenesis
stages of lithogenesis (Rovnina 1981; Hayes et al. 1983). However, no induced fluorescence is
observed in the organic residues, which could indicate thermal alteration beyond the oil generation
window ( c . 90 °C). While the large acanthomorphic acritarchs are light yellow, spheromorphs are
darker (brown).

PALAEOBIOLOGY

Siliclastic and carbonate rocks were initially thought to represent different environmental or
taphonomic settings yielding different microfossils. However, the occurrence of encysted or motile
life stages of planktonic protists in both silicified carbonate and detrital shelf deposits is now amply
documented (Knoll 1984, 1992; Zhang 1984; Awrainik et al. 1985; Yin 1985a, 1985 6; Knoll and
Ohta 1988). Despite the extensive literature dealing with environmental and climatological factors
affecting the lateral and vertical distribution of acritarchs (see Moczydlowska and Vidal 1992,
pp. 30-36 for a comprehensive review), few conclusions have been generated. The distribution of
extant marine planktic protists responds to the complex interaction of water masses of different
temperatures, salinity and nutrient-availability. In contrast, populations of early Palaeozoic cyst-
forming protists appear to have displayed remarkable taxonomic homogeneity during relatively
short intervals of time, a feature that has made them so useful in biostratigraphy (Moczydlowska
and Vidal 1992). It could be argued that the unevenness observed in the populations of modern
marine environments does not apply to the populations in the fossil record, due to the compression
inherent in geological data making them appear substantially more stable and constant. On the
basis of previous studies (e.g. Knoll 1992) and the present material, the wide environmental (Vidal
and Nystuen 1990a) and geographical distribution of late Vendian (Ediacaran) plankton parallels
the above observations on early Palaeozoic acritarchs. However, while early Cambrian acritarchs
are generally abundant (Moczydlowska 1991) in standard palynological preparations, their late
Neoproterozoic lavishly ornamented counterparts are by comparison quite rare. Hence, acid-
resistant residues of organic-rich standard 50 g rock samples yield only modest numbers of large
acanthomorphic acritarchs accompanied by numerous sphaeromorphs and cyanobacterial sheaths,
in contrast to the hundreds of specimens commonly recovered from Cambrian rocks in comparable
facies associations. This circumstance was not previously noted but, in our experience, it certainly
applies also to the Ediacaran Pertatataka Formation in the Amadeus Basin of Australia. It is,
however, difficult to relate such observations to assemblages from cherts of the late Neoproterozoic
Doushantuo Formation in China, since microfossils from this unit were generally studied in
petrographic thin sections (Zhang 1984; Yin 1985a, 19856, 1987).

BIOSTRATIGRAPHY

Acritarchs attributed here to Appendisphaera grandis sp. nov., Cavaspina acuminata (Kolosova,
1991) comb, nov., C. basiconica sp. nov., Tanarium conoideum Kolosova, 1991 emend., T. irregulare
sp. nov., and T. tuberosum sp. nov. were formerly assigned to the generally Early Palaeozoic genus
Baltisphaeridium (Pyatiletov and Rudavskaya 1985; Rudavskaya and Vasileva 1989; Kolosova
1991). They were considered as evidence of an Early Cambrian age by Rudavskaya and Vasileva
(1989), but Kolosova (1991) regarded them as indicating a late Riphean age. Acritarchs attributed
here to T. irregulare sp. nov. were recently reported from the undoubtedly Neoproterozoic
(Ediacaran) Pertatataka Formation in the Amadeus Basin of Australia (Zang 1988, p. 282). Other
acritarchs reported by Zang (1988) are described here as A. grandis sp. nov. and A. tenuis sp. nov.
(see below).

Recently, Knoll (1992) reported large acanthomorph acritarchs from a metamorphic rock
succession in Prins Karls Foreland (Svalbard), including a taxon that he attributed to Briareus
borealis. He also pointed out the resemblance between this taxon and acritarchs attributed by
Rudavskaya and Vasileva (1989) to Baltisphaeridium varium (= T. irregulare sp. nov.; see below),
whereas acritarchs described as lAsterocapsoides sinensis were considered to resemble B. primarium

500

PALAEONTOLOGY, VOLUME 36

(Rudavskaya and Vasileva 1989 ; = T. conoideum (Kolosova) comb. nov. ; see below). Furthermore,
he noted that B. strigosum (in Rudavskaya and Vasileva 1989; see below) resembles C. magnum
from the Neoproterozoic Doushantuo Formation in China (Zhang 1984). Unfortunately the
markedly different state of preservation of the material from the Prins Karls Foreland adds some
uncertainty to their identification.

The lower age limit of the present acritarch assemblage from the Khamaka Formation and time-
equivalent beds cannot be established independently with absolute certainty, due to the absence of
underlying fossiliferous beds. However, an upper limit is clearly provided by the Tommotian to
early Atdabanian shelly fossils in the Bilir Formation (see above). Morphologically complex, age-
diagnostic phytoplanktic microfossils occur consistently within an interval of shales and thin-
bedded mudstones in the upper part of the siliciclastic unit, and dolostones of the lower part of the
carbonate unit. This diagnostic assemblage was recorded in twenty boreholes in the region
(Rudavskaya and Vasileva 1989; Kolosova 1991; and this paper).

Comparable associations of microfossils and isolated occurrences of individual species of
microfossils are known from a number of localities in other regions of the Siberian Platform
(Pyatiletov and Rudavskaya 1985; Kolosova 1991). However, these associations have not been
described in detail. Kolosova (1991) concluded that the presence of acritarchs identified as
Trachyhystrichosphaera aff. aimica Hermann in the rock interval 53-8 to 103-7 m of Borehole Torgo
G-2 is in agreement with a late Riphean age. Trachyhystrichosphaera is indeed typically late Riphean
(Hermann, 1990; Vidal et al. in press), but the identification of T. aff. aimica by Kolosova (1991,
fig. 3, 1-3) appears extremely uncertain. They are associated with acritarchs attributed to Cavaspina
acuminata (Kolosova) comb, nov., which also occurs abundantly in Borehole Talakan 806 at 1467-0
to 1473-9 m, between the Khamaka Formation and the clearly early Cambrian Bilir Formation
(Grausman and Zhernovskij 1989; Rudavskaya and Vasileva 1989). These observations could
suggest a post-late Riphean age for the material described by Kolosova (1991).

In summary, the acritarch evidence is here taken to indicate that, by comparison with acritarchs
from the Ediacaran Pertatataka Formation from Australia (Zang, 1988; Zang and Walter, 1992),
the age of acritarchs in the Khamaka Formation is late Vendian (Ediacaran). This conclusion also
accords with evidence presented by Knoll (1992) for comparable microfossils from Svalbard, which
he also interpreted as Vendian.

SYSTEMATIC PALAEONTOLOGY

Microscopic slides containing all figured specimens are kept in the collections of the Institute of Palaeontology,
Uppsala University, Uppsala (with prefix PMU-Sib.) and at the All-Union Scientific Research Geological
Prospecting Institute, St Petersburg (VNIGRI). The location of specimens on the microscopic slides is given
by England Finder coordinates.

The problems resulting from incorrect taxonomic assignation of certain Neoproterozoic taxa were
extensively discussed by Vidal and Nystuen (19906). Some acritarch taxa reported in this paper were initially
attributed to primarily Palaeozoic genera such as Baltisphaeridium , but are transferred here to new genera.
Diagnostic morphological features and dimensional parameters are shown in Text-figs 3-4, 8-9, 12-13.

Group acritarcha Evitt, 1963
Genus appendisphaera gen. nov.

Type species. Appendisphaera grandis sp. nov.

explanation of plate 1

Figs 1-2. Appendisphaera grandis sp. nov. PMU-Sib. l-L/27/1, paratype; Borehole Zapad 742; depth 1887
to 1894 m. 1, specimen partly covered by superposed tubular, probably cyanobacterial sheath, x 570.
2, enlarged part of the same specimen showing detail view of processes, x 2400.

PLATE 1

MOCZYDLOWSKA et al Appendisphaera

502

PALAEONTOLOGY, VOLUME 36

Appendisphaera gen. nov. simple, solid processes

4. grandis sp. nov.

A. fragilis sp. nov.

A. tenuis sp. nov.

ilALdli

short processes
slightly conical bases
sharp or blunt tips

A. ? tabifica sp. nov.

text-fig. 3. Generalized features of species of Appendisphaera gen. nov. showing diagnostic features of
A. grandis sp. nov., A. fragilis sp. nov., A. tenuis sp. nov. and A. Itabifica sp. nov.

MOCZYDLOWSKA ET AL.\ VENDIAN PHYTOPLANKTON

503

text-fig. 4. Distribution of dimensional
parameters of A. Itabifica sp. nov.,
A. tenuis sp. nov., A.fragilis sp. nov. and
A. grandis sp. nov. Min proc, lower range
of process length ; min dia, lower range of
vesicle diameter; max proc, upper range
of process length ; max dia, upper range
of vesicle diameter; mean dia , mean of
vesicle diameter; mean proc, mean of
process length; Sd dia, standard devi-
ation of vesicle diameter; Sd proc,
standard deviation of process length;
N spec, number of measured specimens.

Derivation of name. From Latin appendix - outgrowth, process, appendage; and Latin sphaera - sphere, ball.
The name refers to the spherical shape of the central vesicle and the processes that it bears.

Diagnosis. Organic-walled, acid-resistant microfossils consisting of medium to large circular to oval
vesicles (originally spherical) bearing relatively long processes evenly distributed on the vesicle wall.
The processes are simple and solid.

Remarks. Under transmitted-light microscopy, the processes appear to be solid, although they may
have slightly widened proximal attachment areas. The morphological and dimensional characters
of individual species are shown in Text-figs 3-4.

1985 Baltisphaeridium (?) strigoswn Jankauskas; Pyatiletov and Rudavskaya, p. 152, pi. 63, figs 7, 9.

1989 Baltisphaeridium strigosum Jankauskas; Rudavskaya and Vasileva, pi. 1, figs 2-4, 6; pi. 2, figs

Derivation of name. From Latin grandis - large, great. It refers to the large dimensions.

Types. Holotype specimen PMU-Sib. l-R/63/2 (Text-fig. 5a-d); paratype specimen PMU-Sib. l-L/27/1 (PI 1,
figs 1-2). Borehole Zapad 742, Nepa-Botuoba region, Yakutia. Thinly bedded mudstones at 1887 to 1894 m,
lowermost Khamaka Formation, Neoproterozoic, Upper Vendian.

Diagnosis. Vesicle circular in outline, originally spherical, bearing very abundant long processes
evenly distributed over its surface. The processes are homomorphic, simple and solid, proximally
slightly widened and distally tapering. Their tips are sharp-pointed. The processes are densely
distributed but clearly separated from each other and attached to the vesicle without clearly defined
basal structures.

Appendisphaera grandis sp. nov.
Plate 1, figs 1-2, Text-fig. 5

1-2.

504

PALAEONTOLOGY, VOLUME 36

text-fig. 5. Appendisphaera grandis sp. nov. PMV-Sib.l-R/63/2, holotype; Borehole Zapad 742; depth 1887
to 1894 m. a-c, enlarged part of the vesicle displaying details of processes, x 2000. d, full specimen, x 500.

Material. Thirty-one specimens, including twenty-four that are very well preserved.

Dimensions. N = 24. Diameter of central body 68-140 pm (holotype 105-108 pm), x = 106T pm, 5 = 100 pn i;
length of processes 9-33 //m (holotype 18-23 /mi), x = 17-4 pm, 3 = 6 0 /mi. Length of processes varies within
a range of 15-25% of the vesicle diameter.

MOCZYDLOWSKA ET AL.. VENDIAN PHYTOPLANKTON

505

Remarks. Kolosova (1990) proposed a new acritarch species T. perfectum. Despite attempts at
obtaining the original publication, there is substantial doubt as to whether it was validly published,
since only a preprint with submission number is available at the moment. In a later paper Kolosova
(1991) figured a specimen under this name, but without providing a description or indicating the
taxonomic status of the taxon. We suspect the taxonomic identity of T. perfection with the presently
erected A. grandis. However, the poor illustration of the equally poorly preserved holotype of
T. perfectum in Kolosova (1990, 1991) does not allow certain identification. For these reasons,
T. perfectum is not included in the formal synonymy.

In the present material, the vesicle wall is most probably smooth, but it is difficult to confirm this
due to the very dense arrangement of the processes. The processes, although thin, seem to be
relatively stiff and straight. As a result, they are commonly very well preserved and display a
considerable regularity in shape and distribution.

Occurrence. Yakutia, Siberian Platform, Nepa-Botuoba region: boreholes Zapad 742, depth 1887 to 1894 m,
Ozero 761, depth 1876 to 1884 m, and Talakan 823, depth 1534 to 1539 m; Khamaka Formation. Syugdzher
Saddle (NE of the Nepa-Botuoba region): boreholes Dyudan 291-0, depth 3414-3 to 3420 3 m and Nakyn
295-0, depth 3062 to 3068 m (this paper). Boreholes Ozero 750 (also known as Peleduj 750), depth 1835 to
1837 m and Byuk 715, depth 1964-8 m; Kursov Formation (Pyatiletov and Rudavskaya 1985).

Appendisphaera fragilis sp. nov.

Text-fig. 6a-b

Derivation of name. From Latin fragilis - fragile, thin, referring to the nature of the processes.

Holotype. Specimen PMU-Sib. l-Y/37/3 (Text-fig. 6a-b). Borehole Zapad 742, Nepa-Botuoba region,
Yakutia. Thinly bedded mudstones at 1887 0-1894 0 m, lowermost Khamaka Formation, Neoproterozoic,
Upper Vendian.

Diagnosis. Vesicle oval in outline, originally spherical, bearing long, slender and fragile processes.
The wall of the vesicle is smooth as shown by the vesicle outline. The processes are of approximately
equal length, thin and thread-like and have blunt tips. They are widely spaced.

Material. Three poorly preserved specimens.

Dimensions. N = 3. Diameter of central body is 57-121 pm, length of processes is 11-20 //m.

Remarks. The processes are only preserved on a small portion of the vesicle. Thus, their numbers
and distribution is uncertain. The part of the vesicle wall that bears processes does not differ
morphologically from the portions where processes are absent. The processes appear attached to the
wall without additional basal structures. It is possible that, if they were broken during deposition
and burial, they would not leave any traces on the surface of the vesicle. It is also possible that
processes may have been both numerous and distributed over the complete surface of the vesicle.
On the other hand, it is also conceivable that extremely poorly preserved specimens with a vesicle
totally lacking processes could be easily mistaken as spheromorphs. Thus, some of the thin-walled
spheromorph microfossils frequently found in Upper Proterozoic rocks could perhaps be in reality
poorly preserved specimens of process-bearing species.

Occurrence. Yakutia, Siberian Platform, Nepa-Botuoba region: Borehole Zapad 742, depth 1887 to 1894 m;
Khamaka Formation. Syugdzher Saddle (NE of the Nepa-Botuoba region). Borehole Dyudan 291-0, depth
3414-3 to 3420-3 m (this paper).

506

PALAEONTOLOGY, VOLUME 36

text-fig. 6a-b, Appendisphaera fragilis sp. nov. PMU-Sib.l-Y/37/3, holotype; Borehole Zapad 742; depth
1887 to 1894 m. a, complete specimen, x 600. B, enlarged portion of the vesicle and processes, x 1350.
C-D, Appendisphaera ‘habifica sp. nov. PMU-Sib.2-H/33/4, holotype; Borehole Zapad 742; depth 1887 to
1894 m. d, full specimen, x 400. c, magnified part of the vesicle showing details of processes and membrane,

x 850.

Appendisphaera tenuis sp. nov.

Text-fig. 7

Derivation of name. From Latin tenuis - thin, fine, delicate; with reference to the morphology of the processes.

Holotype. Specimen PMU-Sib. l-M/33 (Text-Fig. 7a-c). Borehole Zapad 742, Nepa-Botuoba region, Yakutia.
Thinly bedded mudstones at 1887 to 1894 m, lowermost Khamaka Formation, Neoproterozoic, Upper
Vendian.

MOCZYDLOWSKA ET A L.. VENDIAN PHYTOPLANKTON

507

Diagnosis. Vesicle circular in outline, with smooth or psilate wall surface bearing numerous, evenly
distributed, short spiny processes. The processes are solid, thin, and have sharp-pointed to blunt tips
and slightly expanded, conical bases.

Material. Three well preserved specimens.

Dimensions. N = 3. Diameter of central body is 1 15-147 //m and length of processes 7-12 pm.

Remarks. The present species differs from the Neoproterozoic (late Riphean) species Ericiasphaera

text-fig. 7. Appendisphaera tenuis sp. nov. a-c, PMU-Sib.l-M/33, holotype; Borehole Zapad 742; depth 1887
to 1894 m. a-b, enlarged part o‘f the vesicle with processes, x 1500; c, complete specimen, x 500. D,
VNIGRI.3758/2-U/55/1 ; Borehole Dyudan 291-0; depth 3413-3 to 3420-3 m, x 500.

508

PALAEONTOLOGY, VOLUME 36

spjeldnaesii Vidal, 1990 through the more dense distribution of processes and by the lack of ciliar
distal portions in the processes (Vidal 1990, p. 291).

Occurrence. Yakutia, Siberian Platform, Nepa-Botuoba region: Borehole Zapad 742, depth 1887 to 1894 m;
Khamaka Formation. Syugdzher Saddle (NE of the Nepa Botuoba region). Borehole Dyudan 291-0, depth
3414-3 to 3420-3 m (this paper).

A ppendisphaercP. tabifica sp. nov.

Text-fig. 6c-d

Derivation of name. From Latin tabificus - melting. It refers to distal portions of processes merging into the
membrane.

Holotype. Specimen PMU-Sib.2-H/33/4 (Text-fig. 6c, d). Borehole Zapad 742, Nepa-Botuoba region,
Yakutia. Thinly bedded mudstones at 1887 to 1894 m, lowermost Khamaka Formation, Neoproterozoic,
Upper Vendian.

Diagnosis. Vesicle circular in outline, originally spherical, having very abundant, extremely thin
processes that coalesce and acquire the appearance of a membrane in the equatorial zone. The
processes are simple, thin, solid and evenly distributed around the vesicle; they are supported by
intervening organic matter and are welded to it distally.

Material. A single very well-preserved specimen.

Dimensions. The vesicle diameter is 115 pm and the length of processes 4CM6 ^m.

Remarks. The present material consists of a single specimen bearing processes, and embedded in a
membrane that seems to be formed or supported by the processes. However, the membrane-like
material is restricted to the equatorial zone, despite the fact that processes are evenly distributed
around the vesicle. Since only one specimen is available, it is difficult to establish whether this is a
diagnostic feature or a preservational artefact (e.g. due to the accumulation of organic matter
trapped between very densely arranged processes). The species is assigned provisionally to
Appendisphaera. Future finds may provide evidence as to whether the equatorial membrane is a
diagnostic morphological character and thus require erection of a new genus.

Occurrence. As for the holotype.

Genus cavaspina gen. nov.

Type species. Cavaspina acuminata (Kolosova, 1991) comb. nov.

Derivation of name. From Latin cavus - hollow, and Latin spina- spine. The name refers to the hollow
processes, the cavities of which communicate with the vesicle cavity.

Diagnosis. Organic-walled, acid-resistant microfossils consisting of medium to large vesicles,
circular to oval in outline, originally spherical, bearing evenly distributed processes. The processes
are short and simple, cylindrical or conical, and hollow. The process cavities freely communicate
with the vesicle cavity.

Remarks. The general morphology and dimensions of the species attributed to this genus are shown
in Text-figs 8-9. Cavaspina superficially resembles Goniosphaeridium , but species attributed to
Cavaspina are generally much larger. The ratio of vesicle diameter to process length in species of
Cavaspina is substantially larger than in most species of Goniosphaeridium.

MOCZYDLOWSKA ET AL.. VENDIAN PHYTOPLANKTON

509

Cavaspina gen. nov. short, hollow processes

freely communicating with inner cavity

C. acuminata (Kolosova, 1991) comb. nov.

C. basiconica sp. nov.

conical processes with wide bases
and slender distal portions

text-fig. 8. Generalized features of species of Cavaspina gen. nov. showing diagnostic features of C. acuminata
(Kolosova 1991) comb, nov., and C. basiconica sp. nov.

Cavaspina acuminata (Kolosova, 1991) comb. nov.

Text-fig. 10a-b

1989 Rudavskaya and Vasileva, pi. 1, fig. 5.

1989 Baltisphaeridium pilosiusculum Jankauskas; Rudavskaya and Vasileva, pi. 2, figs 4—6.

1989 Baltisphaeridium sp.; Rudavskaya and Vasileva, pi. 2, fig. 7.

1991 Baltisphaeridium (?) acuminatum Kolosova, p. 57, fig. 4: 1-3.

Holotype. Specimen YIGS Nr 87-123 (Kolosova 1991, fig. 4: 1). Yakutia, Siberian Platform, Borehole Torgo
G-2 depth 70 to 74 m, Torgo Formation, Upper Proterozoic (Upper Riphean according to Kolosova 1991).

Translated original description. Microfossils (diameter 35-50 /an) more or less spherical, with slightly ribbed
surface, having abundant folds. Processes (length 3-6 //m) conical, with sharp tips, of unequal length, unevenly
distributed. Their number varies. Some processes have convex and the others straight surface. They differ in
size and shape within the same specimen. Some processes have thick basis and regular conical shape, the others
are sharply tapering in the distal portion. (Kolosova 1991, p. 57; translation by the authors.)

Remarks. The original description of the species does not provide information as to whether the
processes are hollow or solid. Moreover, the species was only doubtfully referred to the genus
Baltisphaeridium. However, the micrograph of the holotype (Kolosova 1991, fig. 4: 1), as well as the
other figured specimens, shows that the processes are hollow and have free communication with the
interior of the vesicle; diagnostically, Baltisphaeridium (a predominantly Ordovician genus) does
not include species in which the processes and vesicle cavities interconnect.

510

PALAEONTOLOGY, VOLUME 36

C. basiconica
C. acuminata

text-fig. 9. Distribution of dimensional
parameters of C. basiconica sp. nov. and
C. acuminata (Kolosova, 1991) comb,
nov. For abbreviations, see Text-fig. 4.

Emended diagnosis. Vesicle circular to oval in outline, spherical before compaction, bearing
numerous simple processes. The processes are short, conical, acuminate and hollow, their cavities
freely interconnecting with the vesicle cavity.

Material. Eight well-preserved specimens.

Dimensions. N = 8. The diameter of central body is 50-68 pm, x = 57-0 pm, S = 6-9 pm. Length of processes
is 3-5 /an, x = 4-5 /an, S = 0-7 /an.

Occurrence. Yakutia, Siberian Platform, Nepa-Botuoba region. Borehole Talakan 806, depth 1467.0 to
1473 9 m; Khamaka Formation. Lena-Anabar Depression (Mouth of Olenek River), Borehole Charchyk 1,
depth 2683-0 to 2712-3 m (this paper). Berezov Depression (western slope of Aldan Shield), Borehole Torgo
G-2, depth 70 to 74 m; Torgo Formation (Kolosova 1991).

Cavaspina basiconica sp. nov.

Text-fig. 1 1

1985 Baltisphaeridium (?) strigosum Jankauskas, 1976; Pyatiletov and Rudavskaya, p. 152, pi. 63,
fig. 8.

Derivation of name. From Latin basis - base; and Latin conicus - conical. The name refers to the quasi-conical
shape of the process base.

Types. Holotype specimen PMU-Sib.l-Y/55/2 (Text-fig. 11a-b, d); paratype specimen PMU-Sib. 1-0/56-2
(Text-fig. 11c). Borehole Zapad 742, Nepa-Botuoba region, Yakutia. Thinly bedded mudstones at 1887 to
1894 m, lowermost Khamaka Formation, Neoproterozoic, Upper Vendian.

Diagnosis. Vesicle circular to oval in outline, originally spherical, having numerous and evenly
distributed processes. The processes are approximately equal in length and have distinctive conical,
swollen bases that grade into thin, slender and twisting distal portions. The process bases form a
slightly wavy outline. The tips of the processes are tapering or blunt. The processes are hollow
proximally, their cavities communicating with the vesicle cavity.

MOCZYDLOWSKA ET AL.: VENDIAN PHYTOPLANKTON

511

text-fig. 10. a-b, Cavaspina acuminata (Kolosova, 1991) comb, nov.; Borehole Talakan 806; depth 1467 0 to
1473-9 m. a, VNIGRI. 1091/1-N/32, x950; B, VNIGRI.1091/1-M/32/3, x 950. c-D, Tanarium conoideum
Kolosova, 1991 emend.; Borehole Dyudan 291-0; depth 3414-3 to 3420-3 m. c, VNIGRI. 3758/3-N/27/3,

x 420; D, VNIGRI. 3758/3-J/20/1, x480.

Material. Seven well-preserved specimens.

Dimensions. N = 6. Diameter of central body is 83-133 pm (holotype 115-133 /mi), x = 96-9 /mi, S = 26-5 pm.
Length of processes 7-16 pm (holotype 11 /mi), x = 10-6 pm, 6 = 3-9 pm. One additional specimen displays
much smaller dimensions, i.e. 32^13 pm in diameter and the length of processes is 3-5 pm.

512

PALAEONTOLOGY, VOLUME 36

text-fig. 11. Ccivaspina basiconica sp. nov. a-b and d, PMU-Sib.l-Y/55/2, holotype; Borehole Zapad 742;
depth 1887 to 1894 m. a, complete specimen, x 480; b, d, magnified part of the same specimen showing conical
bases of the processes (b) and slender distal portions of processes (d), x 1350. c, PMU-Sib. 1-0/56/2, paratype;

Borehole Zapad 742; depth 1887 to 1894 m, x 500.

Occurrence. Yakutia, Siberian Platform, Nepa-Botuoba region: boreholes Zapad 742, depth 1887 to 1894 m
and Talakan 823, depth 1534 to 1539 m; Khamaka Formation (this paper). Borehole Byuk 715, depth
1964-8 m; Kursov Formation (Pyatiletov and Rudavskaya 1985).

Genus tanarium Kolosova, 1991 emend.

Type species. Tanarium conoideum Kolosova, 1991.

Translated original diagnosis. Microfossils dominantly of spheroidal form, diameter 80-200 /mi. They possess
processes of more or less equal size (in a single specimen), differing in number between specimens and not very

MOCZYDLOWSKA ET AL.\ VENDIAN PHYTOPLANKTON

513

Tanarium Kolosova, 1991 emend.

long, hollow processes or protrusions
freely communicating with inner cavity

T. conoideum Kolosova, 1991 emend.

T. tuberosum sp. nov.

text-fig. 12. Generalized features of species of Tanarium Kolosova, 1991 emend., showing diagnostic features
of T. conoideum Kolosova, 1991 emend., T. irregulare sp. nov., and T. tuberosum sp. nov.

evenly distributed ; processes are unbranching, conical, solid and have straight or indented sides or they are
needle-shaped thorns with greatly thickened bases. (Kolosova 1991, p. 56; translation by the authors).

Emended diagnosis. Organic-walled, acid-resistant microfossils consisting of a medium to large

514

PALAEONTOLOGY, VOLUME 36

T. conoideum
T. tuberosum
T. irregulare

text-fig. 13. Distribution of dimensional
parameters of T. conoideum Kolosova,
1991 emend., T. tuberosum sp. nov.,
T. irregulare sp. nov. For abbreviations,
see Text-fig. 4.

vesicle; the vesicle is circular, oval or irregular in outline, originally spherical or sub-spherical.
Processes and protrusions of various shapes arise from the vesicle. The processes and protrusions
are hollow and communicate with the vesicle cavity. They are conical or cylindrical, and tapering
or rounded distally. Simple or branching processes may occur in the same specimen.

Remarks. According to the original diagnosis, Tanarium possesses solid processes. This feature is
inconsistent with photo-micrographs of the type species, T. conoideum (Kolosova 1991, fig. 5: 1-3),
which clearly show that the processes have inner cavities communicating with the vesicle cavity.
Morphological criteria and dimensional parameters of species of Tanarium are shown in Text-
figures 12 and 13.

The most salient feature distinguishing species of Tanarium from acritarchs attributed to
Goniosphaeridium is the heteromorphic nature of their processes, a feature absent in
Goniosphaeridium. Moreover, as with Cavaspina (see above), the observed ratio between vesicle
diameter and process length seems generally larger in species of Tanarium than in the
morphologically related Goniosphaeridium.

Tanarium conoideum Kolosova, 1991 emend.

Text-fig. 10c-d

1985 Baltisphaeridium primarium Jankauskas; Pyatiletov and Rudavskaya, p. 152, pi. 63, figs 1-4.
1989 Baltisphaeridium primarium Jankauskas; Rudavskaya and Vasileva, pi. 1, fig. 7.

1991 Tanarium conoideum Kolosova, p. 57, fig. 5: 1-3.

Holotype. Specimen YIGS Nr 87-115 (Kolosova, 1991, fig. 5: 1-2). Yakutia, Siberian Platform, Borehole
Byuk-Tanar 715 (= Byuk 715), depth 1964 0 to 1970-6 m, Kursov Formation, Upper Proterozoic, Vendian.

Translated original description. Diameter of spheres 109-120 pm. Processes conical with straight sides. Their
length is up to 40 pm and width at the base is 16 pm. Their number varies, relatively few in holotype and more
numerous in other specimens. Due to their number, the distance between processes is variable. It varies between
7-2 and 90 pm (as stated in original publication) between different specimens. Some processes are broken.
Microfossils are light yellow and darker in colour around the outline and, thus, they seem to be spheroidal,
having narrow peripheral dark zone (Kolosova 1991, p. 57; translation by the authors).

MOCZYDLOWSKA ET AL.: VENDIAN PHYTOPLANKTON

515

text-fig. 14. Tanarium irregulare sp. nov. A, PMU-Sib.2-J/57/l, holotype; Borehole Zapad 742; depth 1887
to 1894 m, x 300. b, magnified part of the same specimen showing tubular processes with widened bases,
x 1650. c, PMU-Sib.l-Q/51/3-4, paratype. Borehole Zapad 742; depth 1887 to 1894 m, x 280. D, enlarged
part of the same specimen showing the vesicle with tubular process having free communication with the inner

cavity, x 1000.

516

PALAEONTOLOGY, VOLUME 36

Emended diagnosis. Vesicle circular to oval in outline, originally spherical or spheroidal, possessing
randomly distributed heteromorphic processes. The processes are conical or cylindrical with
tapering or rounded tips. Occasionally the processes are bifurcated distally. The bases of the
processes are often conspicuously widened. The processes are hollow and their cavities communicate
with interior of vesicle.

Remarks. The holotype of T. conoideum was selected by Kolosova (1991) from the same locality and
stratum as Pyatiletov and Rudavskaya (1985) recorded Baltisphaeridium primarium Jankauskas.
The latter species is here considered conspecific with T. conoideum.

Material. Six very well-preserved specimens.

Dimensions. N = 6. Diameter of central body is 1 10-176 /an, x = 130-2 pm, S = 22-4 pm. Length of processes
is 22-55 pm, x = 38-5 /an, 5 = 10-5 /an. Width of process bases is 7-22 pm.

Occurrence. Yakutia, the Siberian Platform, Syugdzher Saddle (NE of the Nepa-Botuoba region): Borehole
Dyudan 291-0, depth 3414-3 to 3420 3 m; Lena-Anabar Depression (Mouth of Olenek River), Borehole
Charchyk 1, depth 2703 0 to 2712-3 m (this paper). Nepa-Botuoba region : boreholes Byuk 715, depth 1968-8 m,
Kursov Formation (Pyatiletov and Rudavskaya 1985) and Ozero 761, depth 1876-3 to 1884-6 m (Kolosova
1991).

Tanarium irregulare sp. nov.

Text-fig. 14

1989 Baltisphaeridium varium Volkova; Rudavskaya and Vasileva, pi. 2, fig. 8.

Derivation of name. From Latin irregularis - irregular, referring to the shape of processes.

Types. Holotype specimen PMU-Sib.2-J/57/l (Text-fig. 14a-b); paratype specimen PMU-Sib.l-Q/5 1/3-4
(Text-fig. 14c-d). Borehole Zapad 742, Nepa-Botuoba region, Yakutia. Thinly bedded mudstones at 1887 0
to 1894 0 m, lowermost Khamaka Formation, Neoproterozoic, Upper Vendian.

Diagnosis. Vesicle irregular in outline with a general oval shape, originally probably sub-spherical.
The vesicle extends into long processes that are hollow, their cavities communicating freely with the
vesicle cavity. The wall of the vesicle and processes is smooth or psilate. The processes are
heteromorphic, simple or branched. They are tubular and of equal diameter along their complete
length. Alternatively, they may gradually taper towards the distal portion or be conical. Their tips
are sharply pointed to rounded. Some processes are distally bifurcated. Proximally the processes
arise abruptly from the vesicle wall, or have widened conical bases.

Material. Three well-preserved specimens.

Dimensions. N = 3. Diameter of central part of vesicle is 75-115 pm. Length of processes is 23^16 /mi.

Remarks. Microfossils of comparable shape and symmetry of the vesicle were reported by Zang
(1988, p. 282) from the Neoproterozoic Pertatataka Formation in the Amadeus Basin, Australia.

Occurrence. As for the holotype.

Tanarium tuberosum sp. nov.

Text-fig. 15

1989 Baltisphaeridium primarium Jankauskas; Rudavskaya and Vasileva, pi. 2, fig. 3.

Derivation of name. From Latin tuberosus - full of protuberances or lumps. It refers to the shape of the
morphological elements.

MOCZYDLOWSKA ET AL.\ VENDIAN PHYTOPLANKTON 517

text-fig. 15. Tanarium tuberosum sp. nov. a-b and D, PMU-Sib.4-J.30/3, holotype; Borehole Ozero 761;
depth 1876 to 1884 m. a, d, enlarged part of the vesicle with hollow protuberances freely communicating with
the inner cavity of vesicle, x 560. b, the same specimen in full, x 280. c, VNIGRI.3 142/2-W/56/3 ; Borehole

Nakyn 295-0; depth 3062 to 3068 m, x 550.

Holotype. Specimen PMU-Sib.4-J/30/3 (Text-fig. 15a-b, d). Borehole Ozero 761, Nepa-Botuoba region,
Yakutia. Thinly bedded mudstones at 1876 to 1884 m, lowermost Khamaka Formation, Neoproterozoic,
Upper Vendian.

518

PALAEONTOLOGY, VOLUME 36

Diagnosis. Vesicle circular, oval or irregular in outline, originally spherical to sub-spherical, and
possessing wide or conical protrusions. The protrusions are hollow and communicate with the
vesicle cavity.

Material. Eight well-preserved specimens.

Dimensions. N = 8. Diameter of central body is 66-207 p m, x = 1 24-3 pm, 6 = 494 //m. Length of protrusions
is 5-99 /mi, x = 45 6 /an, d = 25 4 /an. Width of protrusions is 6-46 /mi.

Remarks. Chen and Liu (1986) described structurally preserved microfossils from the Neo-
proterozoic (Sinian) Doushantuo Formation in China that they attributed to Megasphaera
inornata , Meghystrichosphaeridium wengaensis and M. chadianensis. The taxa in question appear to
represent three-dimensionally preserved large acritarchs, 200-800 pm in diameter. In particular,
illustrated specimens of Meghystrichosphaeridium wengaensis and M. chadianensis (Chen and Liu
1986, pi. 2, figs 1-4) undoubtedly resemble some acritarch specimens attributed here to T. tuberosum
sp. nov. However, the different preservation of these phosphatized specimens renders identification
difficult. Furthermore, specimens of T. tuberosum are substantially smaller, with overall dimensions
ranging from c. 70-300 pm (see above). Safe identification demands direct examination of the
Doushantuo microfossils.

Occurrence. Yakutia, Siberian Platform, Nepa-Botuoba region: boreholes Ozero 761, depth 1876 to 1884 m
and Zapad 844, depth 1700 0 to 1715-6 m; Khamaka Formation. Syugdzher Saddle (NE of the Nepa-Botuoba
region), Borehole Nakyn 295-0, depth 3062 to 3068 m. Lena-Anabar Depression (Mouth of River Olenek),
Borehole Charchyk 1, depth 2703-0 to 2712-3 m (this paper).

Spheromorphs
Text-fig. 16

Description. Acritarchs with circular to oval outline, spherical before compaction. The thickness of the vesicle
wall varies from thin to thick, a feature not related to the vesicle diameter. The surface of the wall is smooth,
psilate or chagrinate, and is often deformed into irregular or arcuate compression wrinkles.

Material. Abundant specimens in variable states of preservation.

Dimensions. N = 20. Diameter of vesicle varies between 46-118 //nr and 250-424 /mi.

Remarks. The taxonomy of spheromorphic organic-walled microfossils has been the subject of
several taxonomic reviews (Volkova 1964; Vidal 1976; Lindgren 1982; Vidal and Siedlecka, 1983;
Jankauskas 1989; Knoll et al. 1991 ; Moczydlowska 1991 ; Knoll 1992) that reflect the small number
of available morphological features. Diagnostic features used to subdivide the group into genera
and species have been essentially arbitrarily and inconsistently used. Features such as the diameter
of the vesicle, and thickness and surface sculpture of the wall, seem open to subjective judgement.
In the case of dimensional limits chosen to distinguish various taxa (e.g. within the genus
Leiosphaeridia Jankauskas, 1989) the criteria are purely arbitrary. In many taxa, the apparent
ornamentation of the wall seems to be a preservational feature introduced by corrosion,
biodegradation or mineral growth (Vidal 1974, 1976). Different states of preservation of the vesicle
wall have been used to recognize species. With the exception of a few diagnostically ornamented
taxa, the taxonomic attribution of unornamented spheromorphs remains problematic. Possibly,
more ‘objective’ characters could be provided by studies on the chemical composition and
ultrastructure. However, few such studies have been undertaken and in view of the present lack of
other evidence we refrain from attempting a more precise taxonomic attribution of these
microfossils.

MOCZYDLOWSKA ET AL.: VEND1AN PHYTOPLANKTON

519

text-fig. 16. Spheromorphic, organic-walled microfossils displaying corroded vesicle walls and compaction
wrinkles, a, PMU-Sib.4-V/32; Borehole Ozero 761; depth 1876 to 1884 m, x 240. b, PMU-Sib.l-J/49/4;

Zapad 742; depth 1887 to 1894 m, x 800.

Occurrence. Common in all investigated successions.

Acknowledgements. For support connected with logistics, field work and collecting of material we are most
grateful to Mrs Valentina V. Grausman and Dr Valeri E. Bakin (Geological Production Corporation, Lena
Gas and Oil Geology, Yakutsk, P. G. O., Lena, Yakutsk), Dr Genady Novikov (V.N.G.R.E., Kysyl-Syr,
Yakutia) and the Swedish Arctic Research Programme, Stockholm. G. Vidal and M. Moczydlowska wish to
acknowledge generous grant support from the Natural Science Research Council (NFR) that defrayed field
work in Yakutia, laboratory research and the work of V. A. Rudavskaya at the Micropalaeontological
Laboratory (Uppsala University). Our initially submitted manuscript was substantially improved through
constructive reviews by Drs Robert A. Fensome and Andrew H. Knoll.

REFERENCES

aitken, J. d. 1978. Revised models for depositional Grand Cycles, Cambrian of the southern Rocky
Mountains, Canada. Bulletin of Canadian Petroleum Geology, 26, 515-542.

AWRAMIK, S. M., McMENAMIN, D. S., YIN CHONGYU, ZHAO ZIQIANG, DING QIXIU and ZHANG SHUSEN 1985.
Prokaryotic and eukaryotic microfossils from a Proterozoic/Phanerozoic transition in China. Nature, 315,
655-658.

brasier, m. d. 1990. Nutrients in the early Cambrian. Nature, 347, 521-522.

1992. Global ocean-atmosphere change across the Precambrian-Cambrian transition. Geological
Magazine, 129, 161-168.

chen menge and Liu kuiwu 1986. [The geological significance of newly discovered microfossils from the Upper
Sinian (Doushantuo Age) phosphorites.] Scientia Geologica Sinica, 1986, 46-53. [In Chinese],

520

PALAEONTOLOGY, VOLUME 36

cowie, J. w. and brasier, m. d. (eds) 1989. The Precambrian- Cambrian boundary. Oxford Monographs on
Geology and Geophysics, Oxford University Press, 12, 213 p.
evitt, w. R. 1963. A discussion and proposals concerning fossil dinoflagellates, hystrichospheres, and
acritarchs. II. Proceedings of the National Academy of Sciences, U.S.A., 49, 298-302.
grausman, v. v. and zhernovskij, v. p. 1989. [Late Precambrian and Cambrian transitional beds in the deep-
drilling sections of the Western Yakutia.] 75-92. In khomentovski, b. b. and sovetov, y. k. (eds). [ Late
Precambrian and early Palaeozoic in Siberia.) Institute of Geology and Geophysics, Novosibirsk. [In
Russian].

hayes, J. M., kaplan, i. r. and wedeking, w. 1983. Precambrian organic geochemistry, preservation of the
record. 93-134. In schopf, j. w. (ed.). 93-134. Earth's earliest biosphere. Princeton University Press,
Princeton, New Jersey, 543 pp.

Hermann, T. N. 1990. [ Organic world billion years ago). Nauka, Leningrad, 50 pp. [In Russian].
jankauskas, T. v. (ed.) 1989. [Precambrian microfossils of the USSR). Nauka, Leningrad, 190 pp. [In Russian].
kaufman, a. j., knoll, a. h. and awramik, s. m. 1992. Biostratigraphic and chemostratigraphic correlation of
Neoproterozoic sedimentary successions: Upper Tindir Group, northwestern Canada, as a test case.
Geology, 20, 181-185.

khomentovski, B. B. 1985. [Vendian of the Siberian Platform], 83—161 In sokolov, b. s. and fedonkin, m. a.

(eds). [Vendian system. 2.) Nauka, Moscow. [In Russian].
knoll, a. h. 1984. Microbiotas of the Precambrian Hunnberg Formation, Nordaustlandet, Svalbard. Journal
of Paleontology , 58, 131-162.

— 1992. Vendian microfossils in metasedimentary cherts of the Scotia Group, Prins Karls Forland,
Svalbard. Palaeontology , 35, 751-774.

— and ohta, y. 1988. Microfossils in metasediments from Prins Karls Forland, western Svalbard. Polar
Research, 6, 59-67 .

— swett, K. and mark, j. 1991. Paleobiology of a Neoproterozoic tidal flat/lagoonal complex: the Draken
Conglomerate Formation, Spitsbergen. Journal of Paleontology , 65, 531-570.

kolosova, s. p. 1990. [Ancient acanthomorphs of the eastern part of the Siberian Platform.] 2-45. In [Organic
world and stratigraphy of deposits in the oil-gas and ore-bearing regions of Yakutia.) VINITI, Moscow [In
Russian; Preprint],

— 1991. [Late Precambrian spiny microfossils from the eastern part of the Siberian Platform]. Algologia, 1,
53-59. [In Russian].

lindgren, s. 1982. Taxonomic review of Leiosphaeridia from the Mesozoic and Tertiary. Stockholm
Contributions in Geology , 38, 21-33.

moczydlowska, m. 1991. Acritarch biostratigraphy of the Lower Cambrian and the Precambrian-Cambrian
boundary in southeastern Poland. Fossils and Strata, 29, 1-127.

— and vidal, G. 1992. Phytoplankton from the Lower Cambrian Laesa Formation on Bornholm,
Denmark: biostratigraphy and palaeoenvironmental constraints. Geological Magazine, 129, 17-40.

pyatiletov, v. G. and rudavskaya, v. a. 1985. [Acritarchs from the Yudomian Complex]. 151-158. In
sokolov, b. s. and ivanovskij, a. b. (eds). [Vendian system, 1.) Nauka, Moscow. [In Russian].
rovnina, L. v. 1981. Palynological method to determine the level of katagenesis of organic matter by using
Jurassic deposits of western Siberia. 427-432. In brooks, j. (ed.). Organic maturation studies and fossil fuel
exploration. Academic Press, London, 441 pp.

rozanov, a. y. and sokolov, b. s. (eds) 1984. [Lower Cambrian stage subdivision. Stratigraphy). Nauka,
Moscow, 184 pp. [In Russian].

— and ten others 1969. [Tommotian stage and the Cambrian lower boundary problem). Nauka, Moscow,
380 pp. [In Russian],

rudavskaya, v. A. and vasileva, n. j. 1989. [Talsy assemblage of acritarchs from the Nepa-Botuoba anticline],
5-11. In TIMOSHINA, N. A. (ed.). [Phytostratigraphy and spore morphology of the ancient plants in the oil-gas
provinces in the U.S.S.R.) VNIGRI, Leningrad. [In Russian].
rundquist, d. v. 1984. Metallogenic belts, provinces and epochs of the U.S.S.R. Episodes, 7, 28-32.
sokolov, b. s. and ivanovskij, a. b. (eds) 1985. [Vendian system , 1.) Nauka, Moscow, 220 pp. [In Russian],
vidal, G. 1974. Late Precambrian microfossils from the basal sandstone unit of the Visingso beds, South
Sweden. Geologica et Palaeontologica, 8, 1-14.

— 1976. Late Precambrian microfossils from the Visingso Beds in southern Sweden. Fossils and Strata , 9,
1-57.

— 1990. Giant acanthomorph acritarchs from the Upper Proterozoic in southern Norway. Palaeontology,
33, 287-298.

MOCZYDLOWSKA ET AL.: VENDIAN PHYTOPLANKTON

521

— and moczydlowska, m. 1992. Patterns of phytoplankton radiation across the Precambrian-Cambrian
boundary. Journal of the Geological Society of London , 149, 647-654.

and rudavskaya, v. a. in press. Biostratigraphical implications of a Chuaria-Tawuia assemblage

and associated acritarchs from the Neoproterozoic of Yakutia. Palaeontology , 36.

— and nystuen, j. p. 1990a. Micropalaeontology, depositional environment and biostratigraphy of the
Upper Proterozoic Hedmark Group, southern Norway. American Journal of Science , 290-A, 170-211.

— 1 9906. Lower Cambrian acritarchs and Proterozoic-Cambrian boundary in southern Norway.
Norsk Geologisk Tidsskrift , 70. 191-222.

— and siedlecka, a. 1983. Planktonic, acid-resistant microfossils from the Upper Proterozoic strata of the
Barents Sea region of Varanger Peninsula, East Finnmark, Northern Norway. Norges Geologiske
Undersokelse Bulletin , 382, 45-79.

volkova, n. a. 1964. [Phytoplankton of the oldest deposits from the north-western Moscow area and its
significance for the stratigraphy.] Izvestiya AN SSSR, Seriya geologicheskaya , 4, 74-84. [In Russian].
yin, l. 1985a. Microfossils of the Doushantuo Formation in the Yangtze Gorge District, western Hubei.
Palaeontologia Cathay ana, 2, 229-249.

— 1985b. Microfossils of the Doushantuo Formation in the Yangtze Gorge district, and Western Australia.
Proceedings of the Royal Society of Victoria , 83, 211-234.

— 1987. New data of microfossils from Precambrian-Cambrian cherts in Ningqiang, southern Shaanxi.
Acta Palaeontologica Sinica. 26, 187-195.

zang, w. 1988. Late Proterozoic-early Cambrian microfossils and biostratigraphy in China and Australia.
Unpublished Ph.D. thesis, Australian National University, Canberra.

— 1992. Sinian and Early Cambrian floras and biostratigraphy on the South China Platform.
Palaeontographica , Abteilung B. 224, 75-119.

— and Walter, M. R. 1989. Latest Proterozoic plankton from the Amadeus Basin in central Australia.
Nature , 337, 642-645.

— 1992. Late Proterozoic and Cambrian microfossils and biostratigraphy, Amadeus Basin, Central
Australia. Memoir of the Association of Australasian Palaeontologists, 12, 1-132.

zhang, z. 1984. Microflora of the late Sinian Doushantuo Formation Hubei Province, China. Scientific papers
for exchange, prepared for the 27 th International Geological Congress. Geological Publishing House, Beijing,
138-142.

MALGORZATA MOCZYDLOWSKA
GONZALO VIDAL

Uppsala University,
Micropalaeontological Laboratory,
Institute of Earth Sciences,
Norbyvagen 22, S-75236 Uppsala, Sweden

VALERIA A. RUDAVSKAYA

Typescript received 9 June 1992
Revised typescript received 21 December 1992

VNIGRI, Cejzdowskaya 27a,
St Petersburg, B-53, Russia

Note added in proof. In a recent monograph, Zang and Walter (1992) described similar acritarchs.
Solisphaeridiuml sp. in Zang and Walter (1992, p. 100, fig. 32H,I) is conspecific with Tanarium irregulare sp.
nov. Appendisphaera grandis sp. nov. is similar to Comasphaeridium sp. B (Zang and Walter 1992, fig. 28 F,
G). A. tenuis sp. nov. differs from Baltisphaeridium plerusente Zang, 1992 and B. rarusente Zang, 1992 by
having solid processes. The latter two species are considered conspecific and the generic attribution incorrect.
A.? tabifica differs from C. dilutopi/um. A.? tabifica processes are more even in length and longer than
C. dilutopilum. The generic attribution of the latter species is incorrect given that Cymatiosphaeroides possesses
an outer membrane enclosing the vesicle and processes (Knoll 1984; Knoll et al. 1991).

NEW MATERIAL OF AN EARLY CRETACEOUS
TITANOSAURID SAUROPOD DINOSAUR FROM

MALAWI

by LOUIS L. JACOBS, DALE A. WINKLER, WILLIAM R. DOWNS
and ELIZABETH M. GOMANI

Abstract. Compared to their Late Jurassic record, sauropod dinosaurs are poorly known in the Cretaceous
Period between 144 Ma and the terminal Cretaceous extinction event at 66 Ma. The Titanosauridae are the most
widespread and common of Cretaceous sauropods. The titanosaurid species from the Dinosaur Beds of
Malawi, Africa, here referred to Malawisaurus dixeyi comb, nov., has procoelous anterior caudal vertebrae, a
characteristic of the family, but middle and distal caudals with gently biconcave ends. Caudal neural spines are
low, a feature that is shared with South American Saltasaurus and North American Alamosaurus. A premaxilla
of Malawisaurus, the first known for the family, is primitive in having the external nares placed far anterior,
demonstrating that this titanosaurid has a blunter snout than other sauropods. Flattened teeth in Malawisaurus
suggest that pencil-shaped teeth may have evolved more than once within the Sauropoda. Titanosaurids
probably originated at a time when other sauropod families were differentiating in the Late Jurassic. The
Titanosauridae is the longest lived group of sauropods.

In a comprehensive review of sauropod dinosaurs, McIntosh (1990o) recognized six families.
Among these are the Diplodocidae, containing the familiar genera Diplodocus , Apatosaurus and
some others, the Brachiosauridae, or giraffe-necked sauropods, including the gigantic Brachiosaurus,
the Camarasauridae, the two primitive families Vulcanodontidae and Cetiosauridae, and the
Titanosauridae. All of the sauropod families were in existence by the Late Jurassic, but their origins
and interrelationships are poorly understood. Among the most enigmatic and least understood of
these groups is the latest surviving Titanosauridae (McIntosh 1 9906), which is the most widely
distributed, and last surviving of Cretaceous sauropods. It is unique among sauropod families in
that its members possess dermal ossifications or scutes.

Titanosaurids were first recognized from the Late Cretaceous of India, and the oldest known
generally accepted titanosaurid record is from the Late Jurassic of Africa. They are known from
Madagascar and Europe, and titanosaurid remains are numerous in the Late Cretaceous of South
America. The family includes the last recorded sauropod in North America, Alamosaurus , from
Texas, New Mexico, Utah, and Wyoming.

Titanosaurid dinosaurs from Malawi (previously Nyasaland), Africa, were first described by
Haughton (1928), who named Gigantosaurus dixeyi. No significant new collections were made in
Malawi until the initiation of a joint Southern Methodist University (SMU) and Malawi
Department of Antiquities programme in 1984 (Clark et al. 1989; Jacobs 1990; Jacobs et al. 1990,
1992). Since the inception of this project, approximately 150 dinosaur specimens, including both
isolated bones and articulated sets, many of which pertain to sauropods, have been removed from
nine localities in the Early Cretaceous Dinosaur Beds (Lupata Group) of the Sitwe Valley, near
Mwakasyunguti, northern Malawi (Colin and Jacobs 1990; Jacobs et al. 1990). The purpose of this
paper is to evaluate Gigantosaurus dixeyi using previously unknown skeletal elements, the
premaxilla, dentary, ischium, middle and distal caudal vertebrae, and cervical vertebrae.

All the material described here is from one quarry (designated CD-9) and occurred within 30
square metres. Some of the bones, notably sternal plates, middle and distal caudal vertebrae, and
some cervical vertebrae were found articulated, but other bones were scattered and isolated to some

(Palaeontology, Vol. 36, Part 3, 1993, pp. 523-534.]

© The Palaeontological Association

524

PALAEONTOLOGY, VOLUME 36

degree. The quarries this project is working occur within the same rock units and in the same areas
as the original collections made by Dixey (1928) and Migeod (1931a, 19316), from matching Dixey’s
photograph (1926, unnumbered plate between pp. 120 and 121, labelled 'Looking across the
Mwakasyunguti Valley from the southwest’). A faunal list from the Dinosaur Beds is given in
Jacobs et al. (1990), but only titanosaurid material is discussed here. All the material represents a
relatively small sauropod.

All specimens collected by our project remain the property of the government of Malawi;
specimens with numbers preceded by MAL have been taken to Southern Methodist University,
Dallas, Texas for study. Other designated specimens are housed in the Department of Antiquities
in Malawi.

SYSTEMATIC PALAEONTOLOGY

Subclass archosauria Cope, 1869
Order saurischia Seeley, 1888
Suborder sauropodomorpha Huene, 1932
Infraorder sauropoda Marsh, 1878
Family titanosauridae Lydekker, 1885

Emended diagnosis. The most diagnostic derived characters of the Titanosauridae include a
transversely expanded ischium and strongly procoelous anterior caudal vertebrae. (Those vertebrae
in the tail having caudal ribs or transverse processes are considered here to be anterior caudal
vertebrae.) In more derived titanosaurids the middle and posterior caudals are also procoelous.
Teeth are narrow, flattened in more primitive forms, but may be more pencil-like in derived forms.
Sternal plates are robust. The neural spines of cervical vertebrae are undivided and the cervical ribs
extend beyond the centrum to overlap the following vertebra, both of which are primitive
characters. Teeth are not limited to the anterior portion of the jaw, at least primitively. External
nares are far anterior.

Distribution. Late Jurassic of Africa; Early Cretaceous of Africa, Europe, and questionably South America;
Late Cretaceous of Africa, Madagascar, South America, North America, Europe, and Asia.

Taxonomic note. All of the material described below pertains to the Titanosauridae, and there is no
indication that more than one titanosaurid species is present in the Dinosaur Beds. The anterior
caudals collected by this project appear to be specifically identical with that originally described and
illustrated (Haughton 1928, pi. 3) as Gigantosaurus dixeyi. Haughton (1928) correctly considered
the specimens he included in G. dixeyi to be most closely related to some of those called
Gigantosaurus robustus from the Late Jurassic of Tendaguru, Tanzania.

Two species of Gigantosaurus had been named from Tendaguru, G. africanus and G. robustus ,
both by Fraas (1908). Unfortunately, the generic name was preoccupied so Tornieria was proposed
to take its place (Sternfeld 1911); thus the species named from Tendaguru became T. africana and
T. robusta. G. dixeyi from Malawi became known as Tornieria dixeyi without further justification.
However, the species named G. africanus by Fraas (1908), the type species, was subsequently
reassigned to Barosaurus africanus (Janensch 1922), leaving only T. robusta in the genus, which is
nomenclaturally untenable. Therefore, the name Tornieria was changed to Janenschia by Wild
(1991), so T. robusta is now Janenschia robusta. The taxon from Malawi is closely related to but
distinct from the titanosaurid genus Janenschia. It is less closely related to Barosaurus, a diplodocid.
For those reasons, a new genus is erected to accommodate the titanosaurid species from Malawi.

Genus malawisaurus gen. nov.

Derivation of name. Named for the country of Malawi plus - saurus , Greek, for lizard.

Diagnosis. As for the only known species, Malawisaurus dixeyi.

JACOBS ET A L.: TITANOSAURID FROM MALAWI

525

Malawisaurus dixeyi (Haughton, 1928) comb. nov.

Text-figs 1-2

v*1928 Gigantosaurus dixeyi sp. nov., Haughton, p. 70, pi. 2, figs 1-3, pi. 3, pi. 4 fig. 1.

1932 Tornieria ( Gigantosaurus ) dixeyi Haughton; Stromer, p. 7.

1954 Tornieria dixei ( sic); Lavocat, p. 67.

1987 ‘ Gigantosaurus ’ (? = Tornieria ) dixeyi ; Raath and McIntosh, p. 117.

1 990a Tornieria dixeyi ; McIntosh, pp. 352, 398.

1990 ‘ Gigantosaurus ’ dixeyi', Jacobs et al., p. 200.

Holotype. The type is taken to be the anterior caudal vertebra illustrated by Haughton (1928) and catalogued
as South African Museum no. 7405. The remainder of the material included by Haughton with the type (a right
pubis, an incomplete scapula, and sternal plates) is considered topotypic.

Type horizon and locality. Upper member of the Early Cretaceous Dinosaur Beds (Lupata Group);
Mwakasyunguti area, northern Malawi, Africa.

Additional material. Premaxilla, dentary, teeth, cervical, dorsal, and caudal vertebrae, sternal plates, ischium.

Diagnosis. A sauropod dinosaur having strongly procoelous anterior caudal vertebrae with short,
vertical neural spines. Additional characters from topotypic specimens indicate middle and distal
caudal vertebrae are not procoelous, the ischium is transversely expanded, cervical and dorsal
vertebrae have undivided neural spines, haemal arches do not bifurcate, cervical ribs extend
posteriorly beyond the centrum, the premaxilla is blunt with the external naris relatively anterior
in position, teeth are not restricted to the anterior portion of the lower jaw, and there are at least
15 tooth positions in the dentary.

Description. An apparently nearly complete premaxilla (MAL-6; Text-fig. 1a) is robust. The alveolar margin of
the premaxilla is slightly damaged, obscuring an accurate count of tooth positions. Three unerupted teeth are
present. The premaxilla exhibits less anterior elongation than seen in Camarasaurus and much less than the
projecting premaxilla of Brachiosaurus. The external naris is large and bordered anteriorly by a narrow,
dorsally oriented ascending process. The maxillary suture is below the external naris, rather than extending
anterior to it as in other sauropod taxa in which this feature is known. This premaxilla is therefore unique
among sauropods, so far as can be determined. It demonstrates that the species it represents had a blunt snout
and domed skull. The external nares were lateral on the skull, far forward, and less retracted than in
Camarasaurus. The placement of the external nares was very different from the dorsal position of diplodocids
and the dorsolateral condition of Brachiosaurus.

One dentary (MAL-174; Text-fig. 1b-c) has 15 tooth positions (minimum). Eight unerupted teeth can be
seen (tooth row length =175 mm; length from symphysis to surangular notch = 201 mm; height of dentary
at mid length = 51 mm. The posterior tooth position is 23 mm anterior to the notch for the surangular.
Anteriorly the dentary is gently curved towards the symphysis. The splenial groove on the medial surface is
broadly open posteriorly, tapering forward to a level approximately below the eighth tooth position.

Teeth, as exemplified by MAL-176 (height = 27 mm, length = 8 mm, width = 5 mm; Text-figs Id, 2a) are
not broadly spatulate as in Brachiosaurus or Camarasaurus. The surface of the teeth is rugose and the lingual
side is less convex than the buccal side. The distal and medial edges are well defined and sharp, extending from
the apex of the tooth. A faint furrow on the buccal surface parallels the medial and distal edges as in
Camarasaurus and Brachiosaurus. The enamel is thin at the base of the crown. The maximum width of the
crown is closer to the tip than the base. The crown is straight. While these teeth are not spoon-shaped, neither
can they appropriately be termed pencil-like.

One cervical vertebra (No. 89-78; centrum length = 386 mm, centrum height = 1 1 0 mm, maximum
preserved height of centrum and spine = 410 mm; Text-fig. 1e) has been prepared sufficiently for description.
The centrum is opisthocoelous. Anterior and posterior zygapophyses do not extend much beyond the level of
the end of the centrum. The neural spine is low and not bifurcated. There are no pleurocoels. The lamina
supporting the diapophysis is oriented parallel to the long axis of the centrum. The cervical ribs are long and
extend posteriorly well behind their associated centrum as thin rods. Isolated fragments of cervical ribs were
previously identified erroneously as ornithischian ossified tendons (Jacobs et al. 1990).

The most complete and best preserved of the dorsal vertebrae (No. 89-137) is distinctly opisthocoelous

PALAEONTOLOGY, VOLUME 36

526

text-fig. 1 . Malawisaurus dixeyi ; Malawi ; Lupata Group, Early Cretaceous ; referred material, a, MAL-6 ; right
premaxilla, lateral view, b-c, MAL-174; right dentary, lingual and lateral views, d, No. 90-69; isolated tooth,
posterior view. E, No. 89-78; cervical vertebra, right lateral view. F, Nos. 89-123, 89-124; articulated sternal
plates, ventral view. G, MAL-142; left ischium, lateral view. Scale bars, for D = 5 mm, for all others = 50 mm.

JACOBS ET AL. : TITANOSAURID FROM MALAWI

527

(centrum length = 200 mm, posterior centrum width = 170 cm, centrum height = 135 mm, height of centrum
and spine = 370, width at diapophyses = 530 mm). Wide transverse processes originate fairly low on the
centrum and project horizontally. The neural spine is short, simple, transversely broad, and not bifurcated. It
has a small central posterior lamina with larger, dorsally flaring lateral laminae. Its position in the dorsal series
is unclear.

Anterior caudal vertebrae are distinctly procoelous, thus conforming to the titanosaurid diagnosis. Based on
comparisons with Alamosaurus (Gilmore 1946, pis 5-8), of the two anterior caudals from Malawi, the more
anterior (No. 90-128; centrum length = 128 mm, posterior centrum width = 129 mm, centrum height =
119 mm, height of centrum and spine = 270 mm, width across transverse processes = 190 mm; Text-fig. 2b) is
probably from the second or third position. The other anterior caudal (No. 89-79; centrum length =131 mm,
anterior centrum width = 122 mm, centrum height = 106 mm, height centrum and spine = 215 mm, width at
transverse processes = 210 mm; Text-fig. 2c) follows closely in the series. It is likely to derive from a position
anterior to the seventh caudal. The posterior ball of the centrum is well developed in both, with the greatest
bulge in the dorsal half. The more anterior of the two anterior caudal vertebrae has the transverse processes
joined to the neural spine by laminae of bone, producing triangular wing-like projections in anterior view. In
both anterior caudals, the neural arch is positioned anteriorly on the centrum and the neural spine is low and
vertically oriented. The centrum appears asymmetrical in lateral view with the posterior margin deeper than
the anterior. In the more posterior of the two vertebrae (No. 89-79), the prezygapophyses extend beyond the
cranial limit of the centrum and facets for articulation with the haemal arch at the posterior margin of the
centrum are large. The transverse processes, while distinct and well developed, are isolated rods not connected
by winglike laminae to the neural spine. Some diplodocids possess mildly procoelous anterior caudals with
well-developed wings, but the anterior caudals from Malawi are clearly titanosaurid based on the extreme
procoely and the substantial length of the centra. None of the caudals from Malawi possesses pleurocoels.

An articulated tail section comprises 22 middle and posterior caudal vertebrae. It shows features consistent
with the procoelous anterior caudals, most notably, the neural arches are attached toward the cranial half of
the centra, neural spines are low, vertical, and do not extend beyond the posterior margin of the centrum,
haemal facets are large, and the prezygapophyses extend well beyond the anterior border of the centrum.
Haemal arches are articulated with the centra. They are simple, not bifurcated, and they are longer than the
neural spines.

An isolated caudal vertebra (MAL-1 ; centrum length = 86 mm, posterior centrum width = 57 mm, centrum
height = 64 mm, height centrum and spine = 138 mm; Text-fig. 2d) is indistinguishable from that in the 7th
position (from the anterior) of the articulated tail section in all essential observable features. Moreover, the
centrum of MAL-1 is gently amphicoelous. We consider that MAL-1 represents the same taxon as the
articulated tail section. Based on the possession of low vertical spines, strong haemal facets, anteriorly
positioned neural arches, and long, anteriorly projecting prezygapophyses, we suggest that the middle and
posterior caudal vertebrae belong to the same taxon as the procoelous anterior caudals. This leads to the
conclusion that there is a transition from strongly procoelous to gently amphicoelous centra progressing
posteriorly along the tail, as will be discussed in more detail below. The short and low neural spines, long
prezygapophyses, and long haemal arches that do not bifurcate preclude referral of the articulated tail section
to the Diplodocidae.

Sternal plates were found with their anteromedial margins articulated as in life. Each sternal plate flares
posterolaterally (Nos 89-123 and 89-124; combined distance across both = 770 mm; Text-fig. If). They appear
identical in all essential respects to the articulated sternal plates of* Alamosaurus illustrated by Gilmore (1946,
pi. 4b). Haughton (1928, pi. 4 fig. 1) illustrated a sternal plate from Malawi that he assigned to Gigantosaurus
dixeyi. It is difficult to interpret his illustration or to evaluate the reconstructive preparation, but it appears that
if the specimen were rotated clockwise such that the top edge became the medial articulation, it would resemble
closely the sternal plates reported here.

A left ischium (MAL-42) has a robust peduncle for articulation with the ilium (Text-fig. 1g). The medial
flange for articulation with the pubis is long (157 mm). The ischium distal to the flange is transversely broad
(132 mm) relative to total length (390 mm).

Comparisons. Based on the possession of strongly procoelous anterior caudal vertebrae, M. dixeyi
is clearly a titanosaurid. Some other sauropods (notably Diplodocus and to a lesser extent
Apatosaurus) have anterior caudals that are weakly procoelous (Osborn 1899; Gilmore 1936); but
the centra are short relative to those in Malawisaurus and other titanosaurids, and the posterior ball
is not nearly so well developed. Moreover, the chevrons of diplodocids are sledge-shaped.

528

PALAEONTOLOGY, VOLUME 36

text-fig. 2. Malawisaurus clixeyi ; Malawi; Lupata Group, Early Cretaceous; referred material. A, MAL-176,
isolated tooth, lingual view, x 1-28. B, No. 90-128; anterior caudal vertebra, posterior view, x0-28. c, No.
89-79; anterior caudal vertebra, left lateral view, x 046. d, MAL-1 ; middle caudal vertebra, left lateral view,
x0-56. e-f, calcite pseudomorphs presumed to represent dermal ossicles, lateral and dorsal views, x 1-3.

Bellusaurus sui from the Middle Jurassic of China is reported to have procoelous anterior caudals
and amphicoelous middle caudals (Dong 1990). We have not examined the material so comments
must be based on the published English summary and illustrations. Bellusaurus appears to be
primitive in having undivided neural spines on anterior dorsal and cervical vertebrae, in the narrow

JACOBS ET A L.: TITANOSAURID FROM MALAWI

529

outline of the ischium, and perhaps in the width of the teeth. While it was assigned to the
Brachiosauridae, the possibility that Bellusaurus may be the oldest known titanosaurid cannot be
ruled out.

M. dixeyi is similar to Janenschia robusta from the Late Jurassic of Tendaguru, Tanzania.
McIntosh (1990a, 19906) considered it to be unsettled whether Janenschia is a titanosaurid. The
problem stems from the lumping of a series of caudal vertebrae, which are strongly procoelous in
the anterior portion of the tail, with type material from another locality that includes no caudal
vertebrae (Janensch 1922, 1961). We assume the association of the Tendaguru caudal series having
procoelous anterior centra with the name Janenschia robusta to be valid. In any event, meaningful
comparisons with the caudal series from Tendaguru can be made, regardless of its binomial, and
that tail specimen, at least, in our opinion pertains to the Titanosauridae.

While the caudal series assigned to Janenschia is strongly procoelous anteriorly (Janensch 1929,
fig. 16), by approximately the 12th to 15th caudal position the posterior ball of the centrum is
much less pronounced that in the vertebrae nearer the cranial end, and the ball is lost completely
towards the end of the tail. The posterior caudal vertebra illustrated by Janensch (1929, fig. 19a) is
not procoelous. These vertebrae, if correctly identified, demonstrate the transition from strong
procoely to gentle amphicoely in the tail of Janenschia. Malawisaurus is similar to Janenschia in this
feature. The transition from procoely also appears in Bellusaurus and in some diplodocids as
mentioned above, although diplodocid anterior caudals are not strongly procoelous.

Malawisaurus is distinct from Janenschia in having lower erect caudal neural spines that do not
project beyond the posterior margin of the centrum and elongate prezygapophyses that extend well
beyond the anterior margin. The cervical vertebra referred to Janenschia by Janensch (1929, fig. 9)
differs from Malawisaurus in having irregular pleurocoels. The ischium of Janenschia (Janensch
1961, pi. 19, fig. 7) is less transversely expanded than that of Malawisaurus. In the transition from
procoely to amphicoely along the tail, the two genera are primitive among titanosaurids. In the
features in which the two differ, Malawisaurus appears derived.

Raath and McIntosh (1987) assigned a series of caudal vertebrae, a humerus, and an incomplete
femur from the Kadzi Formation, Zimbabwe, to Tornieria (now Janenschia). The caudal vertebra
that they illustrated [figs 5a(ii), 5a(iii)] is definitely titanosaurid based on its procoely. The humerus
is more robust than undescribed specimens from Malawi and more like some of those from
Tendaguru. If the associations at Kadzi, Tendaguru, and northern Malawi are correct, the Kadzi
titanosaurid is more like Janenschia than Malawisaurus.

The age of M. dixeyi makes it the only known African Early Cretaceous titanosaurid, and it is
among the most completely known titanosaurids of that age. The report of titanosaurids from the
Early Cretaceous of Gadoufaoua, Niger, while based on apparently excellent material, utilizes a
two-fold familial division of the sauropods, which includes diplodocids (sensu McIntosh 1990a) in
the Titanosauridae. The description of the Gadoufaoua material (Taquet 1976) clearly indicates
diplodocid, rather than titanosaurid, affinities in the sense that these terms are used here.

Titanosaurid fossils are known from the Late Cretaceous of Africa. Aegyptosaurus baharijensis
was described by Stromer (1932). Stromer’s illustrations (figs 4 a-c) indicate that the caudal
vertebrae were procoelous at least through the middle portion of the tail. Other specimens from
North Africa referred to Aegyptosaurus (Lapparent 1960, pi. 6, fig. 8) are fragmentary but also show
that procoely extend at least to the middle caudal vertebrae. Two titanosaurid vertebrae were
reported from the Late Cretaceous of False Bay, South Africa (Kennedy et al. 1987).

Malawisaurus is distinct from all other Cretaceous dinosaurs of Africa, so far as can be
determined. The Late Cretaceous Rebbachisaurus comprises two species. The type, R. garasbae from
Morocco, includes a dorsal vertebra with large pleurocoels and a long neural spine. The second
species, R. tamesnensis from Niger, has broad Camarasaurus- like teeth anterior dorsal vertebrae
with a bifurcated neural spine, and an ischium without transverse expansion, if all the specimens are
correctly allocated (Lapparent 1960, pi. 10, fig. 2, pi. 11, fig. 3). These species are not titanosaurid.
McIntosh (1990a) considered it plausible that Rebbachisaurus garasbae is a diplodocid.

Algoasaurus bauri from South America is a nomen dubium (McIntosh 1990a). Its age is uncertain.

530

PALAEONTOLOGY, VOLUME 36

but is either Late Jurassic or Early Cretaceous (Mateer et al. 1992). The type material is
fragmentary. The original description (Broom 1904, p. 447) mentions that it has the ‘peculiar
excavations seen in the centra of the American types’ (presumably = pleurocoels). Broom’s line
drawing (fig. 2) shows a fragmentary vertebra, identified as a posterior dorsal, with an undivided
neural spine. The teeth of three sauropod taxa were identified from the same formation that yielded
Algoasaurus (Rich et al. 1983): aff. Astrodon sp., aff. Pleurocoelus sp., and aff. Camarasauridae.
Judging from the variation that has been documented in the teeth of Brachiosaurus (Janensch 1935),
for the sake of heuristic argument, we suggest that these teeth exhibit no more variation than might
reasonably be expected in a single sauropod taxon. Further, affinities of Algoasaurus , and the South
African teeth might lie with Rebbachisaurus.

Titanosaurid material from the Late Cretaceous of Europe (Magvarosaurus, Hypselosaurus),
India ( Titanosaurus ), and Madagascar ( Titanosaurus or Laplatasaurus) is incomplete. Middle and
distal caudals are procoelous where representative bones are known, thus distinguishing the Late
Cretaceous forms from Malawisaurus. Macrurosaurus from the Early Cretaceous of Europe is
poorly known, but the caudal centra show a transition from strongly procoelous to gently
amphicoelous or flattened (Seeley 1876).

The oldest known and most primitive New World titanosaurid is Andesaurus from the Albian or
Cenomanian of Argentina. Based on the pelvis, it is a titanosaurid, but no anterior caudal centra
are known. The more posterior caudals are amphiplatyan (Calvo and Bonaparte 1991, fig. 4a-c).
It is similar to Malawisaurus in this feature, but the ischium (Calvo and Bonaparte 1991, fig. 5a) is
not so transversely expanded as in the African form. If Andesaurus is truly a titanosaurid, it is
significant in that it is the only known South American genus lacking procoelous posterior caudal
centra, and it is as primitive as African Janenschia and Malawisaurus in that character.

Comparisons with Late Cretaceous titanosaurids from the Americas are facilitated by higher
quality samples, especially of Saltasaurus in South America (Bonaparte and Powell 1980) and
Alamosaurus in North America (Gilmore 1922, 1946). Malawisaurus is distinct from, and primitive
relative to Saltasaurus, Laplatasaurus, Argyrosaurus and Alamosaurus in having gently
amphicoelous, rather than procoelous, middle caudal centra. It is similar in having low caudal
neural spines, transversely expanded ischia, and large sternal plates. One character that may
eventually prove to be a synapomorphy among American titanosaurids is the possession of a
biconvex centrum in the first caudal vertebra, a character known only in Alamosaurus and at least
some South American titanosaurids so far as it can be evaluated.

Special mention must be made of the South American genus Antaretosaurus, which has been
considered a titanosaurid, because Fluene (1929) assigned a lower jaw fragment (pi. 29, figs 4-5) and
some teeth (pi. 29, fig. 3) to this taxon. The teeth are pencil-like, although they are slightly flattened,
and the jaw fragment has an anteriorly restricted tooth row with an abrupt angle towards the
symphysis. These characters are certainly similar to Diplodocus, thus influencing reconstructions of
titanosaurids and resulting in the lumping of titanosaurids with diplodocids, even to the extent that
in Romer’s (1956) classification, the sauropods comprises only two families, the Brachiosauridae
and the Titanosauridae.

The features of Malawisaurus contradict an especially close relationship of titanosaurids with
diplodocids. The teeth of Malawisaurus are more narrow than in Brachiosaurus, but they are not
pencil-like. Broad, leaf-shaped teeth are primitive for sauropods. Teeth referred to Alamosaurus are
transversely rounded and elongate (Kues et al. 1980, fig. 2), the morphology referred to as pencil-
like in sauropods, as are others illustrated by Huene (1929, pi. 1, figs 12-13), which possibly pertain
to Saltasaurus or Laplatasaurus. On the other hand, the teeth of Camplodoniscus ( = Campylodon)
illustrated by Huene (1929, pi. 40, figs 1-2) are Late Cretaceous South American sauropod teeth
that are clearly not pencil-like. McIntosh (1990n) considered Camplodoniscus to be incertae sedis.

Pencil-shaped teeth are derived, but the morphological change from broad to narrow teeth is so
simple that convergence would be hard to document based on a comparison of end members.
Considering the morphology of Malawisaurus teeth, we conclude that ‘pencil-like’ teeth evolved at
least once in the Titanosauridae and in parallel at least once in the Diplodocidae.

JACOBS ET AL.. TITANOSAURID FROM MALAWI

531

The jaw fragment assigned to Antarctosaurus , with its pencil-like teeth, resembles diplodocids
because of its angulation toward the symphysis and its restricted tooth row. We consider it less likely
that titanosaurids and diplodocids would evolve those characters in parallel. Caudal vertebrae of
Antarctosaurus are unknown. There are no features on any of the bones illustrated by Huene that
exhibit any derived characters of titanosaurids, including the distal ischium (Huene 1929, pi. 32,
fig. 3), which would be expected to be more expanded. Therefore, based on the morphology of the
Malawisaurus jaw, which is dissimilar albeit primitive relative to diplodocids, and the lack of derived
titanosaurid features on other known parts of Antarctosaurus , the referral of Antarctosaurus to the
Titanosauridae can be questioned, even if not fully refuted.

Antarctosaurus is possibly a diplodocid, a family which is not otherwise known from the Late
Cretaceous of South America. A derived diplodocid, Amargasaurus , is known from the Early
Cretaceous of Argentina (Salgado and Bonaparte 1991). No bones in South America from the Late
Cretaceous other than the lower jaw of Antarctosaurus exhibit definite diplodocid synapomorphies
(e.g. sledge-shaped chevrons, bifurcated cervical neural spines). In fact, their only other Late
Cretaceous records are in Asia (McIntosh 1990a), those being the highly derived genera
Nemegtosaurus and Quaesitosaurus.

The referral of Antarctosaurus , with its diplodocid jaw structure, to the Titanosauridae led to the
persistent misconception that titanosaurid skulls should be similar in structure to diplodocid skulls.
This caused further confusion when Antarctosaurus was identified in India by Huene and Matley
(1933). In that same paper a sauropod maxilla was described, but it was considered of uncertain
genus and not identified to any lower level. The salient features of the Indian maxilla, as shown by
the figure and description in Huene and Matley (1933, p. 23, fig. 19), are the vertical orientation of
the ascending process separating the external nans from the antorbital fossa, and the position of the
maxillary-premaxillary suture. They clearly stated (p. 24) ‘very little of the anterior margin is
missing’. The entire length of the suture appears to lie beneath the external naris, not anterior to
it.

We agree with Huene and Matley (1933, p.24), based on their figure and description, that the
Indian maxilla represents a sauropod in which ‘the snout was short and very high’, and we further
suggest that it is the maxilla of a titanosaurid. The suggestion is made because the only other
sauropod specimen known that exhibits a maxillary-premaxillary suture entirely beneath the
external naris is the premaxilla of Malawisaurus from Africa.

Dermal armour

One of the most intriguing features of titanosaurids is the presence of dermal ossicles, known from
Madagascar (Deperet 1896) and South America (Bonaparte and Powell 1980). No bones have been
found in Malawi that can be said with certainty to be dermal ossicles. However, in the same quarry
with titanosaurid bones are calcite pseudomorphs with a shape remarkably like dermal armour
(Text-fig. 2e-f). In addition, calcite fills the marrow cavities of some of the long bones. The
pseudomorphs have a subcircular base, concave on one side, but with a tall keel on the other. The
base is flattened on one edge and the perimeter is ornamented with 12 to 13 small projections.
The pseudomorphs are bilaterally symmetrical with the spine being compressed and inclined
towards the flattened edge of the base, which is taken to be posterior. These pseudomorphs appear
to be biological in origin, and it is worth speculating that they may represent titanosaurid dermal
ossicles. The description and illustrations of the osteoderms of Saltasaurus (Bonaparte and Powell
1980, fig. 6) indicate a ventrally cupped base with a keeled upper surface in some specimens.

DISCUSSION

Janenschia robusta from the Late Jurassic of Tendaguru, Tanzania, has traditionally been
recognized as the oldest titanosaurid, and therefore, the family has been considered to be of African
origin. It may have had such an origin, but the description of Bellusaurus from the Middle Jurassic

532

PALAEONTOLOGY, VOLUME 36

of China, at a minimum, suggests alternatives. Until the Jurassic record is more fully known, the
question of where titanosaurids originated remains open.

By the Early Cretaceous, titanosaurids had spread to Europe. The terrestrial record for the Early
Cretaceous is poor in South America, but Andesaurus is possibly of that age and it lacks procoelous
posterior caudal centra, as does Malawisaurus. After South America and Africa split apart, South
America’s titanosaurid fauna entered a period of approximately 30 million years of evolution in
isolation from the influence of other continents.

Late Cretaceous titanosaurids from Madagascar and India may represent remnants from before
the separation of these land masses from Gondwana, or they may indicate a more complicated
geographical pattern. Currently, titanosaurids from those places are not sufficiently well known for
their geographical history to be understood. The geographical origin of Alamosaurus, the only
known North American titanosaurid, is clearer. Titanosaurids, in all probability, migrated up from
South America near the end of the Cretaceous at the same time ceratopsians, hadrosaurs, and
marsupials were invading South America from the north (Bonaparte 1984, Rage 1986). Then, at the
end of the Cretaceous, just a few million years after it reached North America, Alamosaurus,
possibly the last of the titanosaurids (and if so, the last of the sauropods) became extinct.

Acknowledgements . We gratefully acknowledge the support of Dr Yusuf Juwayeyi, the Malawi Department of
Antiquities, and our Malawian colleagues in the field, especially Fidelis Morocco, James Khomu, and the
villagers of Mwakasyunguti. We also thank Kent Newman, Alisa Winkler, Zefe Kaufulu, Laura MacLatchy,
and John Congleton. Text-figure 1 is by Mary Ann Zapalac. The following people allowed access to collections
in their charge: Hermann Jaeger and Wolf-Dieter Heinrich (Museum fur Naturkunde, Berlin); Jose Bonaparte
and Guillermo Rougier (Museo Argentino de Ciencias Naturales, Buenos Aires); Angela C. Milner (The
Natural History Museum, London); Jennifer Clack (University Museum of Zoology, Cambridge); and Gillian
King (South African Museum). Dr John S. McIntosh provided freely of his knowledge of sauropods. Support
for this project has been provided by the National Geographic Society, the Institute for the Study of Earth and
Man, Caltex Oil (Malawi) Limited, the Huntmgfield Corporation, the Carl B. and Florence E. King
Foundation, and Johnson and Johnson Orthopaedics. Special thanks are due to the Government and the
people of Malawi.

REFERENCES

bonaparte, J. F. 1984. Late Cretaceous faunal interchange of terrestrial vertebrates between the Americas.
19-24. In reif, w.-E. and westphal, f. (eds). Third symposium on Mesozoic terrestrial ecosystems. Attempto
Verlag, Tubingen, 259 pp.

-and powell, j. E. 1980. A continental assemblage of tetrapods from the Upper Cretaceous beds of El
Brete, northwestern Argentina (Sauropoda-Coelurosauria-Carnosauria-Aves). Memoires de la Societe
Geologique de France, Nouvelle Serie, 139, 19-28.

broom, r. 1904. On the occurrence of an opisthocoelian dinosaur ( Algoasaurus bauri) in the Cretaceous Beds
of South Africa. Geological Magazine , Series 5, 1, 445^147.
calvo, J. o. and bonaparte, j. f. 1991. Andesaurus delgadoi gen. et sp. nov. (Saurischia-Sauropoda),
dinosaurio Titanosauridae de la Formation Rio Limay (Albiano-Cenomaniano), Neuquen, Argentina.
Ameghiniana, 28, 303-310

clark, J. m., Jacobs, L. L. and downs, w. r. 1989. Mammal-like dentition in a Mesozoic crocodylian. Science,
244, 1064-1066.

colin, j.-p. and Jacobs, l. l. 1990. On the age of the Malawi Dinosaur Beds: evidence from ostracodes.

Comptes Rendus de I'Academie des Sciences, Paris , 311, 1025-1029.
cope, e. d. 1869. Synopsis of the extinct Batrachia, Reptilia and Aves of North America. Transactions of the
American Philosophical Society, 14, 1-252.

deperet, c. 1896. Note sur les dinosauriens sauropodes & theropodes du Cretace superieur de Madagascar.

Bulletin de la Societe Geologique de France, 24, 176—196.
dixey, F. 1926. The Nyasaland section of the Great Rift Valley. Geographical Journal , 68, 1 17-137.

— 1928. The Dinosaur Beds of Lake Nyasa. Transactions of the Royal Society of South Africa, 16, 55-66.
dong zhi ming 1990. [On remains of the sauropods from Kelamaili Region, Junggar Basin, Xinjiang, China].
Vertebra ta Palasiatica, 28, 43-58. [In Chinese with English summary].

JACOBS ET AL.\ TITANOSAURID FROM MALAWI

533

fraas, E. 1908. Ostafrikanische Dinosaurier. Palaeontographica , 55, 105-144.

gilmore, c. w. 1922. A new sauropod dinosaur from the Ojo Alamo Formation of New Mexico. Smithsonian
Miscellaneous Collections , 72, 1-9.

— 1936. Osteology of Apatosaurus with special reference to specimens in the Carnegie Museum. Memoirs
of the Carnegie Museum , 11, 175-300.

— 1946. Reptilian fauna of the North Horn Formation. U.S. Geological Survey , Professional Paper , 210C,
1-52.

haughton, s. H. 1928. On some reptilian remains from the Dinosaur Beds of Nyasaland. Transactions of the
Royal Society of South Africa , 16, 69-83.

huene, F. von. 1929. Los Saurisquios y Ornithisquios de Cretaceo Argentino. Anales del Museo de La Plata ,
3, 1-196.

— 1932. Die fossile Reptil-Ordnung Saurischia, ihre Entwicklung und Geschichte. Monographien zur
Geologie und Palaeontologie , 4, 1-361.

— and matley, c. a. 1933. The Cretaceous Saurischia and Ornithischia of the central provinecs [s/c] of India.
Memoirs of the Geological Survey of India , Palaeontologia Indica , 26, 1-74.

Jacobs, b. f. 1990. Lower Cretaceous diaspores from Malawi. National Geographic Research , 6, 516-518.
Jacobs, L. L., winkler, d. a. and downs, w. R. 1992. Malawi’s paleontological heritage. Occasional Papers of
the Malawi Department of Antiquities, 1, 5-22.

— kaufulu, z. m. and downs, w. R. 1990. The Dinosaur Beds of northern Malawi, Africa. National
Geographic Research , 6, 196-204.

janensch, w. 1922. Das Handskelett von Gigantosaurus robustus und Brachiosaurus brancai aus den
Tendaguru-Schichten Deutsch-Ostafrikas. Centralblatt fur Mineralogie, Geologie und Palaeontologie , 1922,
464-480.

— 1929. Material und Formengehalt der Sauropoden in der Ausbeute der Tendaguru-Expedition.
Palaeontographica, Supplement 7, 2, 1-34.

— 1935. Die Schadel der Sauropoden Brachiosaurus, Barosaurus und Dicraeosaurus aus den Tendaguru-
schichten Deutsch-Ostafrikas. Palaeontographica, Supplement 7, 2, 147-298.

— 1961. Die Gliedmaszen und Gliedmaszengiirtel der Sauropoden der Tendaguruschichten. Palaeonto-
graphica, Supplement 7, 3, 177-235.

Kennedy, w. J., klinger, h. c. and mateer, N. J. 1987. First record of an Upper Cretaceous sauropod dinosaur
from Zululand, South Africa. South African Journal of Science, 83, 173-174.
kues, b. s., Lehman, t. m. and rigby, J. K. jr 1980. The teeth of Alamosaurus sanjuanensis , a Late Cretaceous
sauropod. Journal of Paleontology, 54, 864-869.

lapparent, a. f. de 1960. Les dinosauriens du 'Continental Intercalaire ’ du Sahara central. Memoires de la
Societe Geologique de France, Nouvelle Serie, 88A, 1-57.
lavocat, r. j. m. 1954. Sur les dinosauriens du Continental Intercalaire des Kem-Kem de la Daoura. Congres
Geologique International, Comptes Rendus de la Dix-Neuvieme Session, Alger 1952, 15, 65-68.
lydekker, r. 1885. Introductory observations. Palaeontologia Indica, Series 4, 1, p. V.
marsh, o. c. 1878. Principal characters of American Jurassic dinosaurs. Part 1. American Journal of Science,
16, 411-416.

mateer, n. j. and twelve others 1992. Correlation of nonmarine Cretaceous strata of Africa and the Middle
East. Cretaceous Research , 13, 273-318.

McIntosh, j. s. 1990a. Sauropoda. 345^101. In weishampel, d. b., dodson, p. and osmolska, h. (eds). The
Dinosauria. University of California Press, Berkeley, xvi + 733 pp.

— 19906. Species determination in sauropod dinosaurs with tentative suggestions for their classification.
53-69. In carpenter, k. and currie, p. j. (eds). Dinosaur systematic^ : approaches and perspectives.
Cambridge University Press, Cambridge, xvi + 318 pp.

migeod, F. w. H. 1931a. British East Africa expedition. Account of work done in 1930. Natural History
Magazine, 1931, 86-103.

19316. Digging dinosaurs. Discovery, 12, 142-145.

osborn, H. F. 1899. A skeleton of Diplodocus. Memoirs of the American Museum of Natural History, 1,
189-214.

raath, m. a. and McIntosh, j. s. 1987. Sauropod dinosaurs from the central Zambezi Valley, Zimbabwe, and
the age of the Kadzi Formation. South African Journal of Geology, 90, 107-119.
rage, j.-c. 1986. South American/North American terrestrial interchanges in the latest Cretaceous: short
comments on Brett-Surman and Paul (1985), with additional data. Journal of Vertebrate Paleontology, 6,
382-383.

534

PALAEONTOLOGY, VOLUME 36

rich, T. h. v., molnar, R. E. and rich, p. v. 1983. Fossil vertebrates from the Late Jurassic or Early Cretaceous
Kirkwood Formation, Algoa Basin, southern Africa. Transactions of the Geological Society of South Africa ,
86, 281-291.

romer, a. s. 1956. Osteology of the reptiles. University of Chicago Press, Chicago, xxi + 772 pp.
salgado, L. and Bonaparte, J. F. 1991. Un nuevo sauropodo Dicraeosauridae, Amargasaurus cazaui gen. et sp.
nov., de la Formacion La Marga, Neocomiano de la provincia del Neuquen, Argentina. Ameghiniana, 28,
333-346.

seeley, H. G. 1876. On Macrurosaurus semnus (Seeley), a long-tailed animal with procoelous vertebrae from the
Cambridge Upper Greensand, preserved in the Woodwardian Museum of the University of Cambridge.
Quarterly Journal of the Geological Society , London, 32, 440-444.

— 1888. The classification of the Dinosauria. British Association for the Advancement of Science, Report,
1887, 698-699.

sternfeld, r. 1911. Zur Nomenklatur der Gattung Gigantosaurus Fraas. Sitzungsberichte der Gesellschaft
Naturforschender Freunde zu Berlin, 1911, 398.

stromer, E. 1932. Ergebnisse der Forschungsreisen Prof. E. Stromers in den Wusten Agyptens. II.
Wirbeltierreste der Baharije-Stufe (unterstes Cenoman). 11. Sauropoda. Abhandlungen der Bayerischen
Akademie der Wissenschaften Mathematisch-naturwissenschaftliche, Abteilung, Neue Folge, 10, 1-21.
taquet, p. 1976. Geologie et Paleontologie du gisement de Gadoufaoua (Aptien du Niger). Cahiers de
Paleontologie, Editions du Centre National de la Recherche scientifique, 15, 1-191.
wild, r. 1991. Janenschia n.g. robusta (E. Fraas, 1908) pro Tornieria robusta (E. Fraas, 1908) (Reptilia,
Saurischia, Sauropodomorpha). Stuttgart Beitrdge zur Naturkunde ( Geologie und Palaontologie) , 173, 1^4.

LOUIS L. JACOBS

Shuler Museum of Paleontology
and Department of Geological Sciences
Southern Methodist University
Dallas

Texas 75275, USA

DALE A. WINKLER
Shuler Museum of Paleontology
Southern Methodist University
Dallas

Texas 75275, USA

WILLIAM R. DOWNS
Bilby Research Center
Northern Arizona University
Flagstaff

Arizona 86011, USA

ELIZABETH M. GOMANI
Department of Geological Sciences
Southern Methodist University
Dallas

Texas 75275, USA
and

Department of Antiquities
P.O. Box 264
Lilongwe, Malawi

Typescript received 27 July 1992

Revised typescript received 4 December 1992

NEW ANATOMICAL CHARACTERS IN FOSSIL
CORALLINE ALGAE AND THEIR TAXONOMIC

IMPLICATIONS

by JUAN C. BRAGA, DAN W. J. BOSENCE and ROBERT S. STENECK

Abstract. Interfilamental cell-connections are considered important characters in the suprageneric and generic
taxonomy of present-day nongemculate coralline algae but to date they have not been used in the taxonomy
of fossil corallines. SEM observations of polished and etched specimens allow recognition of interfilamental
cell connections in fossils and therefore these characters can also be applied to the taxonomy of ancient
coralline algae. The shape and number of epithallial cells, which are important diagnostic features in delimiting
genera in the subfamily Melobesioideae, can also be recognized. The implications of this work are that many
fossil corallines have to be reassigned to different genera and can also be assigned to the subfamily classification
used for present-day corallines. An identification key is given, which permits identification of the known fossil
Cenozoic coralline algae using similar criteria to those used for modern corallines.

Traditionally, generic and suprageneric taxonomy of present-day nongeniculate corallines
was based on the characteristics of tissues and reproductive structures, which could be easily
recognized in fossil material with normal microscope procedures. As summarized by Wray (1977),
diagnostic supraspecific criteria for present-day and fossil material included: (a) type and location
of conceptacles; (b) character of hypothallium; (c) character of perithallium, and (d) presence or
absence and arrangement of heterocysts (trichocytes).

More recently, Johansen (1969) discovered that interfilamental cell-connections (Text-fig. 1a-b)
may be used as important characters to delimit coralline algal subfamilies. Since, Cabioch (1971,
1972), Adey and Johansen (1972), Johansen (1976, 1981) and Woelkerling (1987, 1988), employed
the type of cell-connections in both the suprageneric (Table 1), and generic taxonomy of corallines,
and this point of view seems now to be widely accepted by botanists working with modern
corallines.

In addition to the above characters, Adey (1970) employed the shape of epithallial cells as
diagnostic features in delimiting some genera. The number of epithallial cells was also considered
by Johansen (1976) to be of diagnostic significance at the generic level and even the pattern of cell
elongation was used in distinguishing supraspecific taxa (Adey 1964; Adey and Johansen 1972;
Woelkerling and Irvine 1986). Other anatomical features hardly observable or preservable in fossil
specimens, such as the emplacement of conceptacle primordia (Adey 1964; Adey and Johansen
1972), the pattern of spore germination (Chamberlain 1983), the presence of haustoria (Woelkerling
1988) and the presence of plugs in conceptacle pores (Woelkerling 1987, 1988), have also been
introduced as diagnostic criteria in generic and suprageneric coralline taxonomy.

Many of these characters have not been recognized in fossil material and therefore the taxonomy
of present-day corallines has become separate from the taxonomy of fossil corallines. These
differences between the palaeontological and neontological classifications led Wray (1977, p. 58) to
assume the impossibility of using these new criteria in dealing with fossil algae and led Poignant
(1984, p. 603) to conclude that the generic taxonomy of fossil corallines must be independent of that
of modern algae.

The aim of this paper is to show that key features such as cell-connections are commonly
preserved and can be recognized in fossil corallines. Epithallial cells and meristems are also

[Palaeontology, Vol. 36, Part 3, 1993, pp. 535-547, 2 pls.| © The Palaeontological Association

536

PALAEONTOLOGY, VOLUME 36

mineralised

cell wall liv,n9 ,lssue

secondary
pits

cell

fusions

text-fig. 1 . Scheme of the two types of interfilamental cell connections. A, secondary pit connections, b, cell
fusions.

table 1. Subfamily classification of the nongeniculate coralline algae (from Johansen 1981, Woelkerling 1987,
1988).

Subfamily

Cell fusions
joining contiguous
filaments

Secondary pit
connections joining
contiguous cells

Sporangial conceptacles ;
tetraspores

Lithophylloideae

No

Yes

Uniporate;

tetraspores lacking plugs

Melobesioideae

Yes

No

Uniporate;

tetraspores lacking plugs

Mastophoroideae

Yes

No

Multiporate;
tetraspores with plugs

Choreonematoideae

No

No

Uniporate;

tetraspores with plugs

sometimes calcified and can occasionally be observed in fossil material. Using appropriate SEM
techniques we show that it is possible to use the same taxonomic criteria, whether the material is
fresh, dried or fossilized, and that fossil material can be identified at generic and subfamily level
using the taxonomic systems currently used by botanists (Table 1 ; Woelkerling 1988). It is crucial
for understanding the history of the group and to developing the potential of these algae for
palaeoenvironmental interpretations that the taxonomy of fossil and present-day corallines is as
close as possible.

METHODS

Thin sections

Fossil corallines are commonly studied in thin sections using petrographic optical microscopes. It
is important to emphasize that the algal thalli must be properly orientated in order to visualize the
tissue organization and the conceptacle morphology. Usually two types of sections are needed for

BRAGA ET A L.: CORALLINE ALGAE

537

this purpose: one parallel to the direction of filament growth and perpendicular to the thallus
surface, and the other perpendicular to the direction of filament growth for measurement of cell
diameters. Observations made in sections with other orientations may help in recognizing
anatomical features but data obtained exclusively in sections other than the two basic orientations
can be misleading. There are many examples in the literature of incorrect interpretations of tissue
organization of fossil coralline species, which are based on inadequate or insufficient sections. Thin
sections should be thinly ground and ideally should be not much thicker than the cells are wide
(usually less than 20 /mi). Thick sections preserve a number of superimposed filaments and cells
which are difficult to interpret.

SEM techniques

The procedure which we have found to produce the best results is to prepare samples of fossil
corallines for scanning electron microscopy following the steps below.

(a) Thalli are split with the aid of a small chisel or wire cutters into 5 to 10 mm pieces which are
the most convenient size for handling.

(b) These pieces are oriented and embedded in a resin and the surface is cut and polished. Two
sections should be made; one parallel to, and the other perpendicular to the direction of filament
growth. The specimen is oriented best with a dissecting microscope.

(c) The polished surface is etched with EDTA (7% vol.) or HC1 (2% vol.). The appropriate
etching time largely depends on the diagenesis of the sample but, for most specimens, 3 minutes
using EDTA and one minute using HC1 yields good results. Cell size in the coralline thallus also
influences the appropriate etching times.

(d) The samples must then be mounted on stubs and coated with carbon and/or gold following
standard scanning microscopy procedures.

The procedure used in preparing modern corallines, by simple fracturing of thalli prior to
mounting and coating of the sample, does not give good results in fossil material as diagenetic
features obscure the original tissue characteristics. After the polishing and etching of fossil coralline
thalli, however, original cell walls and successive cements are clearly distinguishable (PI. 1, figs 1-2)
and the structure of calcified tissues is observable at magnifications similar to that used by botanists.
We have found that sometimes cell walls etch positively and sometimes negatively with respect to
the cements infilling the cell cavity. This occurs in material from the same horizon which has been
etched in the same way. We have no explanation as to why this should occur, but would expect the
smaller-celled micritic walls (with larger surface areas) to react faster to etching and preferentially
to dissolve out.

RESULTS

Cell connections

The types of interfilamental cell-connections are considered a diagnostic feature in delimiting
subfamilies, and hence genera, of present-day corallines. With the SEM or optical thin-section
procedure described above, interfilamental cell fusions are easily recognizable in fossil corallines (PI.
1, figs 1-2; Bosence, in press) and therefore we are able to use this character for the first time in
taxonomy. Cell fusions can be seen as voids extending from a cell of one filament to a cell or cells
of contiguous filaments. Cement(s) lining these voids delineate the original shape of fused cells and
clearly distinguish these original tissue features from fractures or later recrystallization of coralline
skeletons. In addition, diagenetic recrystallization of cell walls occurs in patches and not in discrete,
lateral intercell connections. Secondary pit-connections, although observable in some cases (PI. 1,
fig. 3), are more difficult to recognize as they are too small to allow cements to penetrate and fill
them. When cell fusions are absent, and the interfilamental cell-connections are secondary pits,
filaments appear as continuous and aligned rows of cells which are clearly delimited from adjacent
ones by continuous cell walls (PI. 1, fig. 4).

When the sample preparation for SEM work described above is applied to present-day coralline

538

PALAEONTOLOGY, VOLUME 36

thalli, the results are similar to those obtained in fossil corallines. Cell fusions are infilled by
impregnating resin or by cements in the older parts of the thallus. The response of these infillings
to etching is different from that of the calcified cell walls and the different final relief shows the
structure of tissues (PI. 1, fig. 5). Secondary pit-connections usually remain as empty, small holes
in the cell walls and are only visible occasionally as cement moulds. When cell fusions are absent
however, filaments are continuous and aligned, and clearly delimited from contiguous ones (PI. 1.
fig. 6).

In thin section the interlilamental cell-connection type is also recognizable, although it is initially
less clear and unequivocal. Coralline tissues with interlilamental cell fusions have a spongy
appearance in thin section, i.e. they are made up of irregularly sized and shaped cells, and cell
fusions are visible as discontinuities in lateral cell walls (PI. 2, fig. 1). Fracturing and recrystallization
of tissues, however, may be misleading and hardly distinguishable from original features of tissue
at the low magnifications usual in optical microscopy. When cell fusions are absent the tissue in thin
section shows clear and continuous filaments which give a tissue a characteristic structure (PI. 2,
fig. 2).

As an example of the taxonomic implications of these microstructural studies in fossil corallines
we have re-examined two late Miocene coralline species from southern Spain and Malta (see
Systematic Palaeontology section below). The evidence obtained from SEM analysis substantially
changes the subfamilial and generic ascription which would have been previously applied to these
species following traditional palaeontological taxonomy. In both examples, SEM analysis clearly
shows the presence of cell fusions (PI. 1, figs 1-2). These may also be seen in thin sections (PI. 2, figs
1, 3) and can be confirmed by SEM. Together with the tetra-bisporangial uniporate conceptacles
and non-geniculate thallus, these morphological characters undoubtedly place these species in the
subfamily Mastophoroideae (Table 1 ; Woelkerling 1988).

These examples indicate major changes in the taxonomic status of some, or even many, fossil
Cenozoic coralline species. The recognition of cell fusions will require species with uniporate
tetra/bisporangial conceptacles, and considered in the past to be members of the subfamily
Lithophylloideae, to be transferred to the subfamily Mastophoroideae. In our experience, many
fossils previously ascribed to Lithophyllum in the Mediterranean Tertiary are probably either
Spongites or Neogoniolithon. There are no previous references to Tertiary Spongites apart from our
preliminary work on this project (Bosence in press; Braga et al. 1990; Braga et al. 1991), as
Spongites has been a long-overlooked genus until its reassessment by Woelkerling (1985). Segonzac
(1972, 1990) referred several coralline species from the Neogene of eastern Spain to Neogoniolithon

EXPLANATION OF PLATE 1

Figs 1-2. Spongites sp. 1. Museo de Paleontologia Universidad de Granada, Sample TJ-21; Almanzora
Corridor, Almeria, Spain; upper Tortonian. 1, longitudinal section of perithallial filaments showing lateral
cell fusions (arrows) preserved by infilling cement; direction of growth towards top of figure, x 450. 2, detail
of Fig. 1 showing lateral cell fusions (arrows), x 900.

Fig. 3. Lithophyllum sp. BMNH, V.63740; secondary pit connection (arrow) between adjacent perithallial
cells; cell walls have been etched by HC1; Malta; upper Tortonian, Upper Coralline Limestone Formation
(location 39 of Bosence 1983), x 2000.

Fig. 4. Lithophyllum sp. Museo de Paleontologia Universidad de Granada, Sample TJ-25 ; longitudinal section
showing continuous cell and filament walls; direction of growth towards top right; Almanzora Corridor,
Almeria, Spain; upper Tortonian, x 500.

Fig. 5. Spongites sp. Museo de Paleontologia Universidad de Granada, Sample PR-8; Recent specimen in
longitudinal section with resin infill to cell cavities and etched cell walls showing lateral cell fusions;
La Caleta, Cadiz, Spain, x 500.

Fig. 6. Lithophyllum sp. Museo de Paleontologia Universidad de Granada, Sample PR-10; longitudinal view
of Recent specimen showing filaments with continuous filament walls and secondary pit connections
(arrows); La Caleta, Cadiz, Spain, x 500.

PLATE 1

BRAGA et al., Spongites, Lithophyllum

540

PALAEONTOLOGY, VOLUME 36

but the diagnostic criteria used in this assignment were not stated. Neogoniolithon has been
described from the Oligocene of Italy (Mastrorilli 1967) but the distinctive hypothallium was not
illustrated. Poignant (1977) described a coralline from the Palaeocene of the Paris Basin, France,
which he considered to be the earliest record of Neogoniolithon. However, the illustrated
‘megacytes’ or ‘cellules geantes’, that are diagnostic for Neogoniolithon, are very large, have
irregular shapes and thick micritized walls, and may be borings into the coralline tissue, and not
trichocytes. His detailed micrograph of the perithallial tissue does not indicate cell fusions; we
consider this to be an incorrect assignment.

Using microstructural analysis we can utilize the same subfamily-level taxonomy for fossil
nongeniculate corallines as is currently being employed for present-day corallines, with the
exception of the monogeneric subfamily of small epiphytic plants of the Choreonematoideae
(Woelkerling 1987) distinguished on the basis of the occurrence of plugs in uniporate sporangial
conceptacles. This subfamily, however, is also apparently characterized by the absence of any
interfilamental cell-connection (Woelkerling 1988), a feature that presumably can be used for
recognition in fossil material. The other three subfamilies are delimited by the type of cell-
connections and the number of pores in tetra/bisporangial conceptacles, which are preserved (see
identification key. Table 2).

Epithallial cells

Epithallial cells and initials have rarely been described in fossil material but they are preservable,
at least in some cases. They are mineralized (Steneck and Paine 1986) and may be preserved,
particularly when surrounded by micrite. In addition, coralline thalli can be overgrown by other
crusts or other organisms and Voight (1981 ) described the surface view of epithallial cells of Foslie/la
preserved by overgrowth (‘bioimmuration’) of encrusters.

Detailed SEM research has allowed us to recognize epithallial cells in vertical section and to
employ their characters in delimiting genera, which are currently indistinguishable in thin section.
This appears to be the only way of understanding the taxonomy of genera within the Melobesioideae
(e.g. Lithothamnion , Phymatolithon and Clathromorphum). Text-figure 2a-b shows the characteristic
flat epithallial cells of Lithothamnion in a fossil example from the upper Miocene of NE Spain.
Epithallial cells are recognizable in fossil Melobesioideae and therefore modern taxonomy can be
applied to corallines which currently are all assigned to Lithothamnion.

EXPLANATION OF PLATE 2

Fig. 1. Spongites albanensis (Lemoine) comb. nov. BMNH V. 60928; noncoaxial hypothallial tissue (below)
and perithallial tissue above in thin section; note the irregular grid of perithallial tissue and cell fusions
(arrowed); Malta; upper Tortonian, Upper Coralline Limestone Formation, x 125.

Fig. 2. Lithophyllum sp. Museo de Paleontologia Universidad de Granada, Sample TJ-31 ; detail of perithallial
tissue in longitudinal section, showing continuous filament walls; Almanzora Corridor, Almeria, Spain;
upper Tortonian, x 150.

Figs 3-4. Spongites albanensis (Lemoine) comb. nov. Museo de Paleontologia Universidad de Granada, Sample
PUR-63; Almanzora Corridor, Almeria, Spain; upper Tortonian. 3, perithallial tissue in thin section
showing bean-shaped conceptacles, irregular grid or spongy aspect and cell fusions (arrows), x 100. 4, SEM
view of polished and etched perithallial tissue showing cell fusions (arrowed); direction of growth towards
top left, x 200.

Figs 5-6. Spongites sp. 1. Museo de Paleontologia Universidad de Granada, Sample PUR-28; Almanzora
Corridor, Almeria, Spain; upper Tortonian. 5, longitudinal section of hypothallial filaments illustrating
noncoaxial arrangement of cell divisions and overlying perithallial tissue with cell fusions, x 150. 6, irregular
grid or spongy perithallial tissue and uniporate flask-shaped conceptacles; direction of growth towards top,
x 40.

PLATE 2

BRAGA et al Spongites , Lithophyllum

542

PALAEONTOLOGY, VOLUME 36

text-fig. 2. Lithothamnion sp. 2. Museo de Paleontologi'a Universidad de Granada, Sample ELX-13; SEM
view of epithallial cells (arrowed) in polished and etched section; Elche, Alicante, Spain; upper Tortonian.
a, general view of thallus embedded in fine-grained matrix, x 200. b, details of flattened epithallial cells, above
rectangular perithallial cells with deeply etched cell walls, x 1000.

IDENTIFICATION KEY

There have been several identification keys published for both present-day (e.g. Adey and
MacIntyre 1973) and fossil (e.g. Poignant 1979) coralline algae. The former has the drawback of
including characters which are not recognizable in fossils and the latter includes non-diagnostic
criteria (e.g. alignment of cells) within the key; both are outdated by the information presented in
this paper. A key published by Woelkerling (1988) for the identification of present-day corallines
can be followed, but uses many terms which are unfamiliar to palaeontologists. We present a new
key (Table 2) for the identification of fossil coralline algae which follows both the revised
subfamilial classification presented in Table 1, and the generic classification summarized in
Woelkerling (1988) with some additional details from Chamberlain et al. (1991) and Penrose and
Woelkerling (1992). It includes diagnostic criteria for all non-geniculate corallines which we know
have been described from fossil material. For those unfamiliar with coralline algal morphology we
recommend that the key should be used alongside Woelkerling’s (1988) monograph.

SYSTEMATIC PALAEONTOLOGY

Division rhodophyta Wettstein, 1901
Class rhodophyceae Rabenhorst, 1863
Order corallinales Silva and Johansen, 1986
Family corallinaceae Lamouroux, 1812
Subfamily mastophoroideae Setchell, 1943
Genus spongites Kutzing, 1841

Lectotype species'. Spongites fructiculosa Kutzing, 1841; designated by Woelkerling (1985).

Diagnosis'. Nonendophytic Mastophoroideae lacking a basal layer of palisade cells and coaxial
hypothallium. Pore canals of tetrasporangial conceptacles bordered by cells that arise from
peripheral filaments, protrude into the canals and are subparallel to the conceptacle roof (Penrose

BRAGA ET AL. \ CORALLINE ALGAE

543

table 2. Key to Cenozoic coralline algal genera using characters preservable in the fossil record. We do not
include here (a) the small, epiphytic and taeniform and weakly calcified corallines which are unknown as
fossils (Woelkerling 1988) or (b) Aethesolithon (Johnson 1964) which has distinctive vegetative tissue but
unknown conceptacles. Monomerous - a type of thallus construction involving a single system of repeatedly
branched filaments. Dimerous - a type of thallus construction involving two distinct groups of filaments
orientated more or less at right angles to one another (Woelkerling 1988).

1. Tetra/bisporangial conceptacles uniporate, cell fusions absent (secondary pits present, but
probably not preserved); trichocytes absent (LITHOPHYLLOIDEAE).

I. Thallus dorsiventral

i. Hypothallium palisade, single layered; margins bistratose (hypothallium and

epithallium, if preserved, only) - Titanoderma

(formerly Dermatolithon and most Tenarea)

ii. Elypothallium mainly non-palisade, single or multistratose, coaxial

or non-coaxial ; margins multistratose - Lithophyllum

IE Thallus isobilateral (back to back); crusts with medulla of two layers of palisade

cells (unknown as fossil) - Tenarea

2. Tetra/bisporangial conceptacles uniporate; cell fusions present (secondary pits absent);
trichocytes present or absent (MASTOPHOROIDEAE).

I. Thallus thin, 2-3 (5) cells thick.

i. Hypothallium unistratose, small celled ; perithallium absent, thin, or only

around conceptacles - Fosliella

(currently indistinguishable from Pneophyllum in fossil material).

ii. Multiple overgrowths of large (> 10-15 diameter) hypothallial cells -
II. Thallus composed of numerous layers of cells.

i. Pore canal of tetrasporangial conceptacles bordered by a ring of elongate and

conspicuous cells subperpendicular to the roof surface ; cell filaments
subperpendicular to the roof surface -

ii. Pore canal of tetrasporangial conceptacles bordered by cell filaments

subparallel to the roof surface and protruding into the canal

a. Hypothallium coaxial ; trichocytes absent, single or in vertical or

horizontal stacks -

b. Hypothallium non-coaxial; trichocytes absent, single, or in vertical or

horizontal stacks -

Lithoporella

Hydrolithom

Neogoniolifhoii

Spongites

3. Tetra/bisporangial conceptacles multiporate or sporangial sori; cell fusions present
(secondary pits absent) (MELOBESIOIDEAE).

I. Tetra/bisporangial conceptacles multiporate.

i. Thallus dimerous. Hypothallium unistratose, small celled; perithallium

reduced - Melobesia

ii. Thallus monomerous, composed of numerous layers of cells

a. Hypothallium coaxial; epithallium thin and unknown in fossil- Mesophyllum

b. Hypothallium non-coaxial

1. Epithallial cells flat; epithallium one-cell thick - Lithothammion

2. Epithallial cells not flat

a. Epithallium several-cell thick; (unknown as fossil)- Clathromorphum

b. Perithallial cells show elongation down from meristem;

epithallium thin and difficult to recognize in fossils - Phymatolithon

(currently indistinguishable from Leptophytum in fossil material)

II. Tetrasporangia in sori; hypothallium non-coaxial. Epithallium thin and unknown

in fossil - Sporolithon

(formerly Archaeolithothamnium)

544

PALAEONTOLOGY, VOLUME 36

and Woelkerling 1992). In fossil material cell filaments more or less parallel to the conceptacle roof,
sometimes protruding into pore canals, can be easily recognized.

Spongites albanensis (Lemoine) comb. nov.

Plate 2, figs 1, 3^f

1924 Lithophyllum (?) albanense Lemoine, p. 281, text-figs 8-9.

1939 Lithophyllum (?) albanense Lemoine; Lemoine, p. 105, text-figs 75-77.

1983 Lithophyllum albanense Lemoine; Bosence, p. 160, pi. 17, figs 1-4; text-fig. 7.

1988 Lithophyllum albanense Lemoine; Braga and Martin, p. 295, fig. 9.

Type specimens. The type material of this species, originated from the Burdigalian of Koritza (Albania), has
not been located and seems to be lost. It was illustrated by Lemoine (1924, text-figs 8-9).

Description. Thalli monomerous, forming crusts up to 2 mm thick, which develop into branching protuberances
2-6 mm in diameter and up to 40 mm long. Plumose hypothallium up to 300 /mi thick; the cells of which are
irregular in size and shape, 15-30 //m (mean 20 /an, s.d. 4) long and 9-15 pm (mean 13 pm, s.D. 2) in diameter
(PI. 2, fig. 1). The perithallial filaments have the aspect of a zoned, spongy grid, both in crusts and in
protuberances, where they are radially arranged (PI. 2, fig. 3). The cells in this part of the thallus are also
irregular, 12-20 //m (mean 15 /an, s.d. 2-2) long and 8-13 pm (mean 10 pm s.d. 1-2) in diameter. Fusions
between cells of contiguous filaments are common in the perithallium (PI. 2, figs 1, 3-4) and occasional in the
hypothallium.

All the fertile plants bear uniporate conceptacles. They are bean-shaped in section, 400-640 pm (mean
540 pm, s.d. 60) in diameter and 160-220 pm (mean 190 pm, s.d. 25) high (PI. 2, fig. 3). Pores are surrounded
by fans of cell filaments (PI. 2, fig. 3).

Remarks. This species ascribed to Lithophyllum has been reported in the Miocene sediments from
many Mediterranean localities. Our material comes from the Miocene of Malta (Bosense 1983) and
Almeria, Spain (Braga and Martin 1988) and in both cases was assigned to Lithiophyllum albanense.
The tissue of this species, however, shows clear and abundant cell fusions when observed with the
above described SEM method (PI. 2, fig. 4) and in thin section (PL 2, figs 1, 3) and therefore
belongs to the subfamily Mastophoroideae and not to the Lithophylloideae. This non-geniculate
coralline with cell fusions, a non-coaxial hypothallium and uniporate conceptacles lacking a ring of
cell filaments perpendicular to the roof, must be included in Spongites.

Spongites sp. 1

Plate 1, figs 1-2; Plate 2, figs 5-6

1988 Lithophyllum sp. 1, Braga and Martin, pp. 290, 295, 297.

Description. Thalli of monomerous crusts up to 2 mm thick, that give rise to branching protuberances 2-4 mm
in diameter and up to 10 mm in height. Hypothallium multistratose and plumose, 100-200 pm thick (PI. 2,
fig. 5). Hypothallial cells irregularly shaped, 18-32 pm (mean 24 /<m, s.d. 3 5) long and 10-18 pm (mean 14 /an,
s.d. 1-7) in diameter. Perithallial cells form a slightly zoned, irregular grid with a spongy aspect (PI. 2, fig. 6) due
to cell fusions (PI. 1, figs 1-2). Cells in this part of the thallus have variable sizes and shapes. Cells are 10-22 //m
(mean 16 //m, s.d. 3-5) long and 10-20 pm (mean 14 pm, s.d. 2-2) in diameter. Some isolated cells or vertically
arranged groups of cells, which are slightly larger than the average cell-size, may be interpreted with doubt
as trichocytes.

Only uniporate conceptacles are present (PI. 2, fig. 6). They are flask-shaped and are extraordinarily large,
measuring 660-1300 pm (mean 980 pm, s.d. 1 10) in diameter and 200-700 pm (mean 520 //m, s.d. 60) in height
from the floor to the roof (not including the pore) of the conceptacle. The conceptacle pore is bordered by fans
of cell filaments (PI. 2, fig. 6).

Remarks. This species would have been ascribed in a traditional palaeontological taxonomy either
to Lithophyllum or Neogoniolithon (see Wray 1977; Poignant 1979) if the larger cells mentioned

BRAGA ET A L. \ CORALLINE ALGAE

545

above were eventually interpreted as trichocytes. The presence of cell fusions in its tissue, and of
uniporate tetra/bisporangial conceptacles, clearly place this alga in the Mastophoroideae. The
characters of cells around the pore canal of conceptacles, hypothallium and perithallium are typical
of Spongites Kiitzing, 1841 within this subfamily (Woelkerling 1985; Penrose and Woelkerling 1988,
1992).

We prefer not to provide a new name for this alga because of the current revision of the type
material of the hundreds of species already established for Cenozoic coralline algae, many of which
are poorly described and illustrated. A revision of coralline species taxonomy is beyond the scope
of this paper.

CONCLUSIONS

Interfilamental cell-connections can be recognized in fossil corallines using SEM examination of
polished and etched sections. The recognition of these anatomical characteristics allows us to
employ the same diagnostic criteria in delimiting genera and subfamilies for fossil and present-day
corallines. The suprageneric and generic status of some or many Cenozoic species will change after
SEM analyses of their tissues. This is exemplified here by Spongites albanensis (Lemoine) comb. nov.
The recognition of cell fusions in tissues of this coralline permits it to be transferred from the
Lithophylloideae to the subfamily Mastophoroideae.

The presence of calcified epithallial cells and initials in some recent corallines suggests the
potential preservation of them in fossil material. The recognition of such features in fossil
Melobesioideae has allowed us to apply to them the same taxonomic criteria which delimit genera
of this subfamily in present-day corallines.

Acknowledgements. We are grateful to Dr W. J. Woelkerling for discussing problems of identifying ancient
coralline genera, for reading through our identification key and suggesting some minor improvements.
Funding for this joint work has been provided by the Acciones Integradas (J.C.B.), Central Research Fund of
the University of London (D. W.J.B.) and the National Science Foundation (NSF OCE 91 1696) for R.S.

REFERENCES

adey, w. h. 1964 The genus Phymatolithon in the Gulf of Maine. Hydrobiologia , 24, 377-420.

— 1970. A revision of the Foslie crustose coralline herbarium. Kongelige Norske Videnskabernes Selskabs
Skrifter , 1970. 1-46.

— and johansen, h. w. 1972. Morphology and taxonomy of Corallinaceae with special reference to
Clathromorphum , Mesophyllum , and Neopolyporolithon gen. nov. (Rhodophyceae, Cryptonemiales).
Phycologia , 11, 159-180.

— and macintyre, i. G. 1973. Crustose coralline algae. A re-evaluation in the geological sciences. Bulletin
of the Geological Society of America , 84, 883-904.

bosense, d. w. j. 1983. Coralline algae from the Miocene of Malta. Palaeontology , 26, 147-173.

— in press. Miocene coralline algae and coralline algal facies. In Jordan, c., colgan, m. and esteban, m.
(eds). Miocene reefs: a globed comparison , Springer, Berlin.

braga, J. c., bosence, d. w. J. and steneck, r. s. 1991. A re-evaluation of the microstructure and affinities of
some Cenozoic coralline algal genera. 5th International Symposium on Fossil Algae. Capri (Italy) April 1991.
Abstracts , 8-9.

— and martin, j. m. 1988. Neogene coralline-algal growth forms and their palaeonvironments in the
Almanzora river valley (Almeria, SE Spain). Palaeogeography, Palaeoclimatology, Palaeoecology, 67,
285-303.

— and alcala, b. 1990. Coral reefs in coarse-terrigenous sedimentary environments (Upper Tortonian,
Granada Basin, south Spain). Sedimentary Geology , 66, 135-150.
cabioch, j. 1971. Essai d’une nouvelle classification des Corallinacees actuelles. Comptes Rendues
Hebdomadaires des seances de I Academie des Sciences de Paris , 272 D, 1616-1619.

546

PALAEONTOLOGY, VOLUME 36

cabioch, j. 1972. Etude sur les Corallinacees. II. La morphogenese; consequences systematiques et
phylogenetiques. Cahiers de Biologie Marine , 13, 137-288.
chamberlain, y. m. 1983. Studies in the Corallinaceae with special reference to Fosliella and Pneophyllum in
the British Isles. Bulletin of the British Museum ( Natural History ), Botany Series , 11, 291^163.

irvine, l. m. and walker, r. 1991. A redescription of Lithophyllum orbiculatum (Rhodophyta,
Corallinales) in the British Isles and a reassessment of generic delimitation in the Lithophylloideae. British
Phycological Journal , 26, 149-167.

johansen, h. w. 1969. Morphology and systematics of coralline algae with special reference to Calliarthron.
University of California Publications in Botany , 49, 1-78.

— 1976. Current status of generic concepts in coralline algae (Rhodophyta). Phycologia, 15, 221-244.

— 1981. Coralline algae , a first synthesis. CRC Press, Boca Raton, Florida, 239 pp.

Johnson, J. H. 1964. Fossil and Recent calcareous algae from Guam. Professional Paper of the United States
Geological Survey. 403-G, 1 — 40.

kutzing, F. T. 1841. Ueber die 'polypiers calciferes' des Lamouroux. F. Thiele, Nordhausen, 34 pp.
lamouroux, j. v. f. 1812. Extrait d’une memoire sur la classification des polypiers coralligenes non entierement
pierreux. Nouveau Bulletin des Sciences par la Societe Philomatique de Paris, 3, 181-188.
lemoine, p. 1924. Contribution a l’etude des Corallinacees fossiles. VII. Melobesiees miocenes recueilles par M.
Bourcart en Albanie. Bulletin de la Societe Geologique de France, 23, 275-283.

— 1939. Les algues calcaires fossiles de l'Algerie. Materiaux pour la Carte Geologique de I'Algerie, ler Serie,
Paleontologie , 9, 1-128.

mastrorilli, v. i. 1967. Nuovo contributo alio studio delle Corallinacee deH’Oligocene Ligure-Piemontese :
I reperti della Tavoletta Ponzone. Atti dell ’ Istituto di Geologia della Universitd di Genova, 5,
153-406.

penrose, d. and woelkerling, w j. 1988. A taxonomic reassessment of Hydrolithon Foslie, Porolithon Foslie
and Pseudolithophyllum Lemoine (Corallinaceae, Rhodophyta) and their relationships to Spongites Kutzing.
Phycologia. 27, 159-176.

— 1992. A reappraisal of Hydrolithon and its relationship to Spongites (Corallinaceae, Rhodophyta).
Phycologia , 31, 8 1—88.

poignant, a. f. 1977. Une nouvelle algue Corallinacee Neogoniolithon montainvillense n. sp., dans le Paleocene
du Bassin Parisien. Geobios. 10, 129-134.

1979. Determination generique des Corallinacees mesozoiques et cenozoiques. Bulletin du Centre de
Recherche, Exploration-Production Elf- Aquitaine, 3, 757-765.

1984. La notion de genre chez les algues fossiles. A. Les Corallinacees. Bulletin de la Societe Geologique
de France. 26, 603-604.

rabenhorst, l. 1863. Kyptogamen-Flora von Sachsen, der Ober-Lausitz, Thuringen und Nordbohmen Abteilung
I. E. Krummer, Leipzig, 653 pp.

segonzac, G. 1972. Nouvelles especes de Corallinacees (algues calcaires) du Neogene d’Espagne. Bulletin de la
Societe dHistoire Naturelle de Toulouse , 108, 280—286.

1990. Algues neogenes betiques. Documents et Travaux. Institut Geologique Albert-de-Lapparent, 12-13,
35-41.

setchell, w. A. 1943. Mastophora and the Mastophorsae : genus and subfamily of Corallinaceae. Proceedings
of the National Academy of Sciences of the United State of America, 29, 127-135.
silva, p. c. and johansen, h. w. 1986. A reappraisal of the order Corallinales (Rhodophyceae). British
Phycological Journal, 21, 245-254.

steneck, r. s. and paine, r. T. 1986. Ecological and taxonomic studies of shallow-water encrusting
Corallinaceae (Rhodophyta) of the boreal northeastern Pacific. Phycologia, 25, 221-240.
voigt, e. 1981. Erster fossiler Nachweis des Algen-Genus Fosliella Howe, 1920 (Corallinaceae; Rhodophyceae)
in der Maastrichter und Kunrader Kreide (Maastrichtium, Oberkreide). Facies, 5, 265-281.
wettstein, r. r. 1901. Handbuch der systematischen Botanik. Vol. 1. Deuticke, Leipzig, 201 pp.
woelkerling, w. j. 1985. A taxonomic reassessment of Spongites (Corallinaceae, Rhodophyta) based on
studies of Kutzing’ s original collections. British Phycological Journal, 20, 123-153.

1987. The genus Choreonema in southern Australia and its subfamilial classification within the
Corallinaceae (Rhodophyta). Phycologia, 26, 111-127.

1988. The coralline Red Algae: an analysis of the genera and subfamilies of nongeniculate Corallinaceae.
British Museum (Natural History), London, and Oxford University Press, Oxford, 268 pp.

BRAGA ET A L. \ CORALLINE ALGAE

547

— and irvine, L. M. 1986. The typification and status of Phymatolithon (Corallinaceae, Rhodophyta). British
Phycological Journal , 21, 55-80.

wray, j. L. 1977. Calcareous algae. Elsevier, Amsterdam, 185 pp.

J. C. BRAGA

Departamento de Estratigrafia y Paleontologia
Universidad de Granada
Campus Fuentenueva
18002 Granada, Spain

D. W. J. BOSENSE

Department of Geology
Royal Holloway and
Bedford New College
University of London
Egham, Surrey TW20 OEX, UK

R. S. STENECK

Typescript received 29 April 1992

Revised typescript received 23 December 1992

Department of Oceanography
University of Maine
Darling Marine Centre
Walpole, Maine 04573, USA

PALAEOSCOLECID WORMS FROM THE MIDDLE
CAMBRIAN OF AUSTRALIA

by KLAUS J. MULLER and INGELORE HINZ-SCH ALLREUTER

Abstract. The Middle Cambrian of the Georgina Basin, Queensland, Australia, has yielded a large
association of secondarily phosphatized Palaeoscolecida in three-dimensional preservation. Many hitherto
unknown details are described, such as the differentiated aboral end, irregularly distributed tubules on the
outer surface, and a wide range of sclerites even on the same animal. Internal structure, functional morphology,
ecological aspects and systematic relationships are also discussed. Nineteen species belonging to nine genera
are described, of which the following are new: Austroscolex primitivus gen. et sp. nov., A. spatiolatus sp. nov.,
Corallioscolex gravius gen. et sp. nov., Euryscolex paternarius gen. et. sp. nov., Kaloscolex granulatus gen. et
sp. nov., Milaculum elongatum sp. nov., Murrayscolex inaequalis gen. et sp. nov., M. serratus sp. nov.,
Pantoioscolex oleschinskii gen. et sp. nov., Rhomboscolex chaoticus gen. et sp. nov., Schistoscolex
angustosquamatus gen. et sp. nov., S. mucronatus sp. nov., S. umbilicatus sp. nov., Shergoldiscolex nodosus gen.
et sp. nov., S. polygonatus sp nov., and Thoracoscolex armatus gen. et sp. nov.

Palaeoscolecida are a group of Early Palaeozoic worms that are characterized by annulation
and a striking surface ornamentation. Minute plates of variable shape are arranged in rows
perpendicular to the long axis over the armoured part of the organism. Such fossils have been
known for more than 110 years ( Protoscolex was described by Ulrich in 1878), but until recently
received little attention. They are proving to be widespread in the early Phanerozoic, and may be
of interest with regard to their possible stratigraphical applicability. As more-or-less entire fossils,
they are preserved only under special conditions, mainly in shales. Much more widespread are the
sclerotized plates from their surfaces. These are quite diverse, and have been described under
different generic names, such as Hadimopanella , Kciimenella , and Milaculum. Such isolated plates
have been recorded from all over the world; they range from the Early Cambrian to the Late
Silurian.

The material presented herein is the most comprehensive association yet known. It is secondarily
phosphatized and was etched from limestones. In this special type of preservation the specimens
have survived in three dimensions, and for the first time the aboral end of the organisms can be
documented. The end is usually broken. The worms are expected to have had a proboscis that
underwent rapid decay because of its softer integument. Different types of sclerites may occur on
the same animal, and variability in sclerite size and morphology can be studied in detail. Our
material shows that generic assignment of isolated plates may be extremely difficult in some cases.

HISTORICAL REVIEW

In the study of Palaeoscolecida we have to distinguish between descriptions of more-or-less entire
organisms embedded in shale and isolated sclerites. The source of the sclerites was discovered only
recently (Hinz et al. 1990).

The first palaeoscolecidan worms were described by Ulrich (1878) under the generic name
Protoscolex. This author introduced five species but the holotype of the type species ( P .
covingtonenesis ), in particular, lacks distinctive features (Conway Morris et al. 1982, p. 2154).
Therefore, the other four species of Ulrich (1878) and all subsequent assignments to this genus
should be revised. Miller and Faber (1892) established Protoscolex magnus from the Upper

(Palaeontology, Vol. 36, Part 3, 1993, pp. 549-592.|

© The Palaeontological Association

550

PALAEONTOLOGY, VOLUME 36

Ordovician Fulton Formation in Cincinnati, Ohio. They assigned the genus tentatively to the
Annelida. Almost 30 years later. Bather (1920) described Protoscolex latus from the Lower Ludlow
of Herefordshire. His drawing shows a double row of plates on each annulus. One row is smaller
and sometimes overlain by the adjacent ‘segment’ so that the annulus appears to have only one row
of plates. Details of spacing, outer shape and ornamentation of plates are not recorded. Bather
considered the papillae as initial stages of setae (see below). On the basis of the cuticle, missing
appendages and the presence of a gut, he stated a strong resemblance between Protoscolex and
Oligochaeta, comparing the genus also with Nematoda and Millipedia. Further, he regarded the
Upper Carboniferous Pronaidites carbonarius Kusta, 1888 as Protoscolex , but this genus was
included in the oligochaetid family Tubificidae Vejdovsky, 1884 by Howell (1962). Ruedemann
(1925a) published Protoscolex cf. covingtonensis Ulrich from argillaceous black shales from a bed
that is transitional between the Upper Ordovician Utica and Frankfort beds in upper New York
State. Another, but questionable representative, P. giganteus Ruedemann, 1925a, is extremely large,
being about 100 mm in length. The central canal of the specimen is apparently filled with mud. In
the same year Ruedemann (1925/?) described Protoscolex batheri from the Silurian Lockport
Limestone at Gasport, New York and supported a relationship of Protoscolex to Oligochaeta.

Whittard (1953) introduced Palaeoscolex and the Palaeoscolecidae for worms that had a papillate
surface but were more advanced than Protoscolex. The suggested separation between Palaeoscolex
and Protoscolex was supported by Howell (1962) who maintained the Palaeoscolecidae but
classified Protoscolex under phylum, class, order and family uncertain. The next palaeoscolecid
was described by Robison (1969) as Palaeoscolex ratcliffei from the Middle Cambrian Spence Shale
of Utah. The single, carbonized specimen is split longitudinally in the centre and therefore does not
show the surface of the sclerites; Conway Morris and Robison (1986) published additional material
from Spain, and established the new Class Palaeoscolecida. Due to the above-mentioned
preservational differences, P. cf. ratcliffei could not properly be identified with Robison’s (1969)
species. The new material still left the problem of both ends, annulus borders, metameric
segmentation and ornamentation of the individual plates unsolved. In connection with the
description of a possible annelid from the Trenton Limestone (Ordovician) of Quebec, Conway
Morris et al. (1982) discussed Protoscolex and Palaeoscolex , concluding that their interpretation as
oligochaetes is not credible. They preferred a general assignment to the Annelida. Glaessner (1979)
described Palaeoscolex antiquus from the Lower Cambrian Emu Bay silty shale or siltstone of
Kangaroo Island, Australia. These fossils are preserved either as moulds or as partial replacements
by calcite. They were rejected as palaeoscolecidan by some authors (e.g. Huo and Chen 1989). Xian
and Huo (1988) recorded Palaeoscolex sinensis from South China.

A more comprehensive study was provided by Kraft and Mergl (1989) who described various new
genera from the Lower Ordovician of Bohemia. Their well-preserved material came from clayey or
micaceous shales and siliceous nodules. They tentatively assigned the fossils to the Annelida.

Isolated sclerites of Palaeoscolecida have been established under various generic names as hard
parts of problematic affinity. Ethington and Clark (1965) were the first to report material discovered
in the course of routine etching for conodonts from the Lower Ordovician of Alberta, Canada, and
described the specimens as ‘plate forms A and B’ (non form C, which is enigmatic). Muller ( 1973)
introduced Milaculum for four species, including Ethington and Clark’s material, from Upper
Cambrian and Ordovician strata of Iran, Sweden and North America. Muller and Miller (1976)
described possible cysts as Utahphospha from limestones of the Upper Cambrian Orr Formation,
Utah. Repetski (1981) followed with another species of these cysts from the Ordovician Padre
Formation, Texas. He suggested that Hadimopanella and Lenargyrion might be junior synonyms of
Utahphospha if they agreed in their internal structure. Further taxa were almost simultaneously
published as Hadimopanella Gedik, 1977 (from the Upper Cambrian of Turkey) and Lenargyrion
Bengtson, 1977 (from the Lower Cambrian of Siberia). Until now, Lenargyrion has been considered
as a junior synonym of Hadimopanella. During the 1980s, many new, mostly hadimopanellid,
species were described. Gedik (1981, 1989) established further taxa and proposed a hadimopanellid-
based zonation for the entire Cambrian. Wrona ( 1982) and Gazdzicki and Wrona (1986) recognized

MULLER AND HINZ-SCH ALLREUTER: CAMBRIAN WORMS 551

HadimopaneUa apicata from the Lower Cambrian of Antarctica. Van den Boogaard (1983) reported
hadimopanellids in rock-forming quantities from the Middle Cambrian Lancara Formation,
northwestern Spain. Peel and Larsen (1984) described material from the Lower Cambrian of
Greenland. Dzik (1986) erected a Hadimopanella-Utahphospha group. Wrona (1987) illustrated a
fragment of a natural assemblage of hadimopanellid sclerites in quite an irregular and loose
arrangement. He suggested a tube-like structure for the animal. Further, he established a family
Utahphosphidae comprising Utahphospha and HadimopaneUa , and considered Milaculum to be
analogous but not homologous to Utahphosphidae. Marss (1988) introduced another palaeo-
scolecid under the name Kaimenella which is similar to Milaculum. Having been unaware of
Wrona’s (1987) paper he erected the Hadimopanellidae for HadimopaneUa and Kaimenella.
Boogaard (1988) described two new Milaculum species from the Upper Ordovician of Estonia and
the Lower Silurian of the Welsh Borderland. He was the first to illustrate sclerites of different sizes
in their original configuration, embedded in fragments of the preserved cuticle (Boogaard 1989a).
In a second paper (Boogaard 19896), he described another possible hadimopanellid species from the
Ordovician of Estonia. Bendix-Almgreen and Peel (1988) agreed with a relationship between
HadimopaneUa and Utahphospha , but contrary to Repetski (1981), they kept them apart. In their
study of material from the Lower Cambrian of North Greenland they were largely engaged in the
question of internal structure and hadimopanellid affinity. Wang (1990) recorded a further
HadimopaneUa from the Silurian (Llandovery) Wangjiawan Formation of South China. Esakova
{in Koneva et al. 1990) described a new species of Utahphospha from the Cambrian of Kazakhstan.

Contrary to the relatively intact worms which have been tentatively assigned to Annelida or even
Oligochaeta, opinions about the nature of the sclerites have been quite diverse. Originally, they were
described as problematica without any assignment. Bengtson (1977, p. 758) recognized Lenargyrion
as external dermal sclerites, but pointed out that it ‘ . . . shows no conclusive vertebrate characteristics.
Any homologisation would be highly speculative and there remained a strong possibility of an
unknown invertebrate origin that is unrelated to the vertebrate stock’. Wrona (1982, 1987)
compared HadimopaneUa with placoid scales of early vertebrates (heterostracans) but refrained
from any systematic assignment. Dzik (1986) suggested that his Hadimopanella-Utahphospha group
together with Milaculum Muller, 1973, could represent an evolutionary line of dermal skeleton in
primitive chordates. Marss (1988) noted some similarity to Agnatha but did not commit herself
with regard to the origin of the sclerites. Boogaard (1988, p. 8) considered the characteristics of
Milaculum as being ‘suggestive of an agnathan affinity’. Based on a written communication by
M. Mergl (June 1989), he shared the opinion that HadimopaneUa and Milaculum belong to
Palaeoscolecida (Boogaard 1989a). In his 19896 paper Boogaard (p. 12) noted a resemblance
between palaeoscolecidan sclerites and dermal plates of primitive chordates. He concluded that they
were ‘not ancestral to chordates but perhaps have derived from the same stock’. Bendix-Almgreen
and Peel (1988) recognized morphological similarities between HadimopaneUa and Utahphospha on
the one hand and ascidian spicules of Cystodites Drasche, 1884 on the other. Despite compositional
differences, they tentatively proposed a relationship between these forms. Wrona (1989) followed
the suggestions of previous authors about a relationship of HadimopaneUa and dermal sclerites of
primitive chordates.

MATERIAL AND METHODS

In the Middle Cambrian of the Georgina Basin, Queensland, Australia, seventy-four samples from nine
localities proved to be productive for palaeoscolecid worms (Table 1). The rocks are deeply weathered down
to about 200 metres. A total of 150 kg was etched with 15 per cent acetic acid on screens with a mesh size of
150 and 250 microns in order to differentiate the various fractions from the beginning. Subsequently, the finest
residues from the bottom of the buckets were washed on a 0-063 mm screen. Without further treatment these
residues were sorted under a binocular microscope. The majority of the approximately 3500 sorted specimens
were originally fragmented, which prevents statements about the real frequency of the individuals. Also the
assignment of the specimens to particular taxa proved to be very difficult without SEM photographs because
ornamentation of plates and platelets are mostly indistinguishable under a binocular microscope. Even

552

PALAEONTOLOGY, VOLUME 36

Uh a

r r-j

r~ r-- 2

.. Q<

■ NO

1 <N ' — • ai) fe 1

t rn . <- rr-t O

■ t~~ /-vIS 43 i
- . '•9 oi H '

■ ci 4- „ _ :

W3

<D

05 <4_|

id

^ 00 ^ 00" ^
; £ ^ ^ ^
jjo L, ^ H

- - 00 £ &
3 r t 1 ^ f — tuO

H V „ rT3

i) 2 2

H S S3 2

, 3 o 3, 5
g g>
3 a ~ 2? d

5 ^ w„ tr ^
; *s ^ s °

., t? fl >_ .0

U" S

m

« Q

u -3

N->

Cj S

00

o

X

o

X

Oh

W

Oh

<

JL>

'Oh
£ co
03 O
C/5 O

X .

M fC) 10 o (N m t (Nm^NO
'sDNor^oooooooo oooo

M CN (N (N (N rj M rn m m m

r^- t^~ t"- r~~ r-- r^- r~~ r- r- r-- r-"

5 ^°, V

2t-t
(N o
Oh - *0

2 00 -
^ OO
y o ^
Xh On o

^ W ^

£

O

g

‘o

Oh
Oh v

H1 -

I’d

2

O ''l

£ q) o

; 00 v

^ «TN

0>

m

t>X) ’“ ' (N

3 W73

&

£

u

1)

00 NO -
d vo Tf

2 00 o

^ v cr>
CO 00 v

J-; ^ d-

M° *

©St,

Ph ^ Ol

W C/!

7307

MULLER AND H I NZ-SCH ALLREUTER: CAMBRIAN WORMS

553

x

. . x . x .

XX .xxx

XX .XX .

. X . X . .

X X X X . X

X X X X X X

xxx .xxx

X

. . X X X X . X

. .xxx . .

X .X . . .. . X X X X X .

. .X .XX .

. . .XX...

. . X . . X .

X .

X ...x.x

. .XX .XX .X . . .X

X

X

oo o < m r-~ os

n n fO m ro
r-- r- r- r- r- r-

o

CN

rn

r-

— • <N
(N (N
m m
r- r-

Tf (n VO OO
tN CN <N <N
rn m m m
r~- r- r"- r-

rN
m m
m m f*~i

r- r- r-

ro

co

r-

*/~» vo r"~
r<) m m
r*~> rr i m
r-- r- r-

oo

m

cn

r-

os o

Tj-

r^) m m

r- r- r-

<N

m

03 't Vi

'Tt-

rn m m

r- r^- r-

r->

"Tf

m

r-

oo

rn

r-~

os o

Tf IT)

m m
r- r-

<D

s

u

>>

|h
I— i

o

s

VO

s

oo

<rs

z

6

M

o

Os

rn

w

©

lo

oo

T1-

o

<N

C/D

table 1. Palaeoscolecid localities, sample numbers and species distributions. Unidentified palacoscolecids came from: (1) About 1 km north of Mt
Murray, E 139° 58' 6", S 21° 48' 50,4", Cme (sample no. 6958); (2) Rogers Ridge. Trig. P., E 139° 58' 50,4", S 21° 45' 41,4", Cmc (sample nos 7262.

7263, 7275, 7280, 7282-4); (3) North of Rogers Ridge, E 139° 36.6', S 21° 44' 50.4 ", Cmi (sample nos 7302-4, 7306-8, 7310, 731 1, 7313, 7317, 7319. ^

7320); (4) About 1 km north of Mt Murray, E 139° 58' 27,6". S 21° 48' 50,4", Cme (sample nos 7321, 7322, 7324-6, 7328, 7331-45, 7347-50), (5) >

Phosphate Hill, excavation T15, E 139° 58' 2", S 21° 53' 26", Cme (sample nos 7351-9, 7362-7, 7371. 7374-6); (6) Phosphate Hill, excavation V9, £

E 139° 58' 2 ", S 21° 56' 36", Cme (sample nos 7377, 7383, 7390, 7393, 7394, 7400. 7401); (7) North of Rogers Ridge, E 139° 59,6', S 21c 44 21", Cmd m

(sample no. 7407). (8) North of Rogers Ridge, E 139° 59,6', S2T44' 19", Cme (sample no. 7414); (9) Thorntonia, section 418, E 138° 53,8', ®

S 19° 7,4'. Cmc (sample no. 7464); (10) BMR Duchess 18. E 139° 59' 16,4 S 21° 45', Cme (sample nos 7502, 7503). H

Explanation of abbreviations Cme, Monastery Creek Formation (late Templetonian); Cmi. Inca Shale (late Templetonian - 9 early Floran; Cmc. O

Currant Bush Limestone (Undillan); Cmd, Devoncourt Limestone (Undillan) Ap, Atistroscolex primilivus ; As, Austroscolex spatiolatus\ Cf, o
Corallioscolex formosus, Ep, Euryscolex paternarius; Hoi, aff. H. oezgitli form species I ; HoII, all' H oezguli form species II; Kg, Kaloscolex gravius , O
Me, Milacnltim elongation , Mi, Murrayscolex inaequalis; Ms, Murrayscolex serratus\ Po, Pantoioscolex oleschinskii ; Rc, Rhomboscolex cliao tints,

Sa, Scliistoscolex angnstosijiiamatas ; Sm, Scliistoscolex mucronatus; Su, Scliistoscolex umbilicatus ; Si, Scliistoscolex sp. indet ; Sn. Shergoldiscolex 0
nodosus, Sp, Shergoldiscolex polygonatus; Ta, Tlioracoscolex armalus ; Ha, cf. Hadimopanello apicata, Pi, Pal. gen. indet. sp. A.

Sample tn

Locality Age nos. Ap As Cf Ep Hoi HoII Kg Me Mi Ms Po Rc Sa Sm Su Si Sn Sp Ta Ha Pi

I km N Mt Murray
E 139° 58' 27,6";

S 21° 48' 50,4"

Rogers Ridge. TP point
E 139° 58' 50,4";

S 21° 45' 41,4"

N' Rogers Ridge
E 139° 58' 36,6".
S 21° 44' 50,4"

Cme 6958

Cme 7262
7263
7275
7280

7282

7283

7284
Cmi 7302

7303

7304
7306

I km N' Mt Murray
E 139° 58' 27,6",

S 21° 48' 50,4'

7307

7308

7310

7311
7313
7317

7319

7320

7321

7322

7324

7325

7326
7328

7331

7332

7333

7334

7335

7336

7337

7338

7339

7340

7341

7342

7343

7344

7345

7347

7348

7349

7350

554

PALAEONTOLOGY, VOLUME 36

text-fig. 1. Generalized geology of the northwestern Queensland Cambrian phosphogenic province (from

Shergold and Southgate 1986).

photography of the entire material would not have led to precise identifications because on many specimens
the critical details are concealed by phosphatic precipitation. The number of specimens listed for each taxon
is mainly based on photographed material only.

Outer morphological features have been documented using CamScanll photography. Internal structure was
studied by thin-sections and observations of exfoliated fragments in transmitted light under a microscope.
Measurements were obtained from the scanning photographs. The ratio between plates and platelets covering
the surface has been calculated using the videoplan program by Contron on much-enlarged photographs.

Illustrated material is housed in the Commonwealth Palaeontological Collections of the Bureau of Mineral
Resources, Canberra, under the numbers CPC 23001-23064.

STRATIGRAPHY

Middle Cambrian rocks extend virtually throughout the entire 325,000 km2 of the Georgina Basin (Text-fig.
1). The succession is most complete in the southern and eastern parts. The sediments are rich in phosphate and,
due to their economic importance, their investigation started in the 1960s (Shergold and Southgate 1986).

MULLER AND HINZ-SCH ALLREUTER: CAMBRIAN WORMS

555

UPPER CAMBRIAN -
ORDOVICIAN

MIDDLE

CAMBRIAN

LATEST

PROTEROZOIC -
L. CAMBRIAN

PROTEROZOIC

Undifferentiated

Inca Formation
Beetle Creek Formation
Thorntonia Limestone

Mt Birme Beds

Undifferentiated

— — — Fault

text-fig. 2. Phosphorite fields of the Duchess Embayment (from Shergold and Southgate 1986).

The present paper is based on material from the Thorntonia Region, the Duchess Embayment and the
Ardmore Outlier (Text-fig. 2). Stratigraphically it can be assigned to the Templetonian and Floran stages.

In general Middle Cambrian sedimentation starts with arkose, conglomerates and detrital sands which may
be fluvial. A dolomitic mudstone lithofacies passes into mudstone carbonates and coarser carbonates that are
often associated with evaporites, phosphorites and siliceous sediments. These are followed by black shale and
thin-layered carbonates or banked carbonates. The series terminates with more banked carbonates (for further
details see Southgate and Shergold 1991).

The basal conglomerate of the Ordian is 2-75 m thick and is overlain by dolomitic lithofacies of the
Thorntonia Limestone. This consists of recrystallized dolostones with wackestone and grainstone textures,
followed by trilobite and echinoderm coquinites. At the end of the Ordian, the region faced a short period of

556

PALAEONTOLOGY, VOLUME 36

emergence and subaerial erosion before rapid subsidence in the Early Templetonian led to the sedimentation
of a chert-siltstone-phosphorite-limestone lithosome.

In the Thorntonia region the Bronco Stromatolithe marks the boundary between Ordian and Templetonian.
Here, subsidence did not start before the late Templetonian. The type section of the Gowers Formation is
6 5 m thick and located at Section 418, southwest of Thorntonia Station. The basal part is formed by
dolomitized phosphatic packstones. Subsequent lime-mudstones are followed by phosphatic wackestone and
mudstone, capped by hardgrounds and phoscrete crusts.

The Ardmore Outlier is about 17 km west of Dajarra. The profile commences with the Riversdale Formation
and the overlying Thorntonia Limestone that terminates with the Ardmore Chert. The Monastery Creek
Formation follows disconformably in the early Templetonian. Its basal part, known as the Siltstone Member,
is composed of thin-layered, siliceous phosphorites that are only poorly exposed in this area. The upper part
of the Beetle Creek Formation, the Simpson Creek Phosphorite, is pelletal and well bedded with interbeds of
collophane mudstone.

The Duchess Region is about 140 km southeast of Mt Isa and consists of several isolated phosphorite
occurrences. They extend from Mt Birnie in the north to Phosphate Hill in the south. The outcrop at Rogers
Ridge is situated within a series of en echelon faults at the northeast of the Duchess Embayment. The section
exposes the Thorntonia Limestone, disconformably overlain by the Siltstone Member of the Monastery Creek
Phosphorite Formation. It is differentiated into siliceous, calcareous and non-phosphatic facies. The
calcareous facies consists of fetid, weakly dolomitic phosphatic limestone with micritic structure; chert and
shale are minor components. The profile at the Trigonometric Point has a thickness of about 15 m and is
formed by alternating layers of fossiliferous phosphatic carbonates, dolomite and cherts. They have been
assigned to parasequence 2 of sequence 1 of the Monastery Creek Phosphorite Formation (Shergold and
Southgate 1991) which is of late Templetonian age. Outcrops of the basal Inca Shale occur north of Rogers
Ridge. It is composed of bleached bituminous shales with intercalated fossiliferous carbonate nodules. The
presence there of Triplagnostus gibbus indicates a late Templetonian age. To the northwest, the shales pass into
a laminated carbonate facies.

In the Mt Murray area, samples were collected as isolated, flat carbonate nodules that proved to be extremely
fossiliferous. These also are of late Templetonian age. In contrast to the micrites at Rogers Ridge, the
sediments here are sparry and bioclastic carbonates.

The Phosphate Hill deposits crop out about 15 km SSW of Rogers Ridge. Excavation T15 (Text-fig. 2;
Localities 5 and 6) exposes a 11 m section showing ten lithological units (Shergold and Southgate 1986). The
deposits of excavation V9 (Text-fig. 2; Locality 7) at the southernmost end of Phosphate Hill do not correlate
easily with the recognized units of T15. Rogers and Crase (1980) therefore suggested an alphabetic division into
units A-E. Mudstone phosphorite and chert are less abundant than at T15 and the series is heavily faulted.

A drill core (BMR Duchess 18) made northeast of Rogers Ridge provided samples from the Inca Formation
which begins here prior to the late Floran Euagnostus opimus Zone. It thus can be regarded as a lateral
equivalent of the Gowers Formation of the Thorntonia Region.

PRESERVATION

Notwithstanding their plated armour, Palaeoscolecida were originally soft and pliable, since
occasional specimens are folded and some specimens show in part a narrow, high-relief annulation;
it was the cuticle between the sclerotized parts that provided the elasticity. The material is preserved
by secondary phosphatization, and traces of abrasion on the phosphatic coating indicate repeated
transport. Most of the specimens were fragmented prior to final deposition and they give little
evidence of the life position and ecology of the animals.

Between the individual samples there are considerable differences in completeness and mode of
phosphatization. In some cases even the finest details are preserved, while in others a coarse coating
may conceal otherwise distinct structures. Large worms are usually compressed, and due to
compaction of the sediment, some specimens show interference between adjacent fossils. These
differences in preservation are partly due to differences in sedimentation, and to dissolution and
redeposition of phosphate in the deeply weathered sequences; the porous rock is an aquifer in some
cases (J. H. Shergold, pers. comm.).

Almost complete specimens are rare and belong only to a few species. However, these specimens

MULLER AND H INZ-SCH ALLREUTER: CAMBRIAN WORMS 557

are usually thickly coated in phosphate; this cover prevented disintegration of the armour during
entombment and during the etching process, but conceals details of the surface structure.

Our material comes from beds that contain many primary phosphatic sclerites, such as conodonts
and horny brachiopods, together with other obviously secondarily phosphatized fossils, such as
molluscs, ‘ostracodes’ and trilobites. As has been documented (e.g. Bengtson 1977; Bendix-
Almgreen and Peel 1988), the sclerites are composed of two structurally different units: a fibrous
core and brim, with organic fibrils between phosphatic material, and a dense phosphatic capping
forming the surface ornamentation. According to the diagnosis given for Lenargyrion by Bengtson
(1977), the sclerites were regarded as primarily phosphatic; we consider this quite likely. However,
it seems that palaeoscolecidan sclerites have been discovered so far only from sediments in which
secondary phosphatization occurred (e.g. Bengtson 1977, text-fig. 1a; Wrona 1982, pis 1-2; Peel
and Larsen, fig. 2a-c; Hinz 1987, pi. 4, figs 3, 6; Bendix-Almgreen and Peel 1988, fig. 4a-c). The
originally flexible cuticle of our palaeoscolecidan material is without doubt secondarily
phosphatized. Despite the diagenetic origin of the phosphate, the lamellar wall structure is still
observable in thin-sections (Text-fig. 17a-c). Therefore, the distinct structures of the sclerites
cannot contribute to the question of whether the phosphatic nature is primary or diagenetic. This
is in accordance with the opinion of Bengtson and Conway Morris (1984), who studied halkierid
sclerites with regard to their original shell substance, and who stated that ‘Although secondary
phosphate itself normally shows amorphous structure, phosphatized specimens may preserve fine
surface sculpture and internal structure of the wall'.

As isolated palaeoscolecidan remains have been studied in the past exclusively from residues
prepared for conodont studies, a possible bias is suggested. If the palaeoscolecids had been
calcareous they would have been dissolved during the preparation process. As far as we are aware,
such remains have never been recovered by washing from soft clays. However, residues from such
samples commonly are rather bulky, and conodonts concentrated by heavy liquid techniques before
sorting. If not being actively sought in the smaller size-range, palaeoscolecidan remains may easily
be overlooked in routine processing for conodonts by this method.

MORPHOLOGICAL CHARACTERS
Our orientation of the worms has been based on the following.

1. In curved specimens, e.g. Schistoscolex umbilicatus (Text-fig. 1 Id), the convex side is considered
as dorsal.

2. The distinction between oral and aboral end is based on specimens as shown in Text-figure
12b, d, g. The broken end on the inner side of the spiral (Text-fig. 1 Id) shows termination of the
regular annulation of a broader portion; this suggests that the break would continue into the
nippled aboral end (as shown, for example, in Text-figure 11b).

3. On fragments lacking both ends, the orientation was deduced from the inclination of the annuli
according to principles of streamlining; the leading (stoss) side is considered to be oral, and the
opposite (lee) side aboral. The sclerotized plates have a higher preservation potential than the
whole animal. Accordingly they are common, and in many cases occur in great abundance.
However, they show only a few of the relevant taxonomic characters of this group.

The platelets exhibit a wide range of shape. However, as isolated elements, they are generally
small, and unidentifiable in the etched residues.

Characters of systematic value within the Palaeoscolecida (see Text-fig. 3) may be the following.

1. Number of rows of plates per annulus.

2. Rows that are distinguished by the size of the plates but not by their ornamentation.

3. Outline and ornament of individual plates.

4. Type of platelets (major or minor; with or without sculpture, rounded, polygonal or elongate;
‘crumpled cuticle’).

5. Arrangement of platelets (in rows, irregular etc.)

558

PALAEONTOLOGY, VOLUME 36

s o

text-fig. 3. Morphological terms applied to palaeoscolecids. Abbreviations: a, annulus; an, aboral nipple;
ba, bifurcated annulus; in, intercalation; iv, invagination; mp, microplate; n, nipple; p, plate; pit, platelet;

so, indicated surface ornamentation; t, tubule.

TAXONOMIC CHARACTERS

The above-mentioned characters have been used for the generic and specific classification of the
palaeoscolecids. However, a suprageneric taxonomy cannot be proposed because the characters
used occur too variably in members of the group. Thus, the number of rows per annulus, the
presence of intercalations, and the development of marginal rims of plates and platelets cannot be
used to recognize families.

The study of ontogenetic stages was largely prevented by the fragmentary preservation;
evolutionary trends could not be observed.

Worms having annuli composed of a double row of equally sized plates have ornamentation
showing a mirror-image between the rows. Those with rows of differently-sized plates have the rows
always placed in the same position (i.e. anterior or posterior) on the annulus. The position of
platelets on annuli may be an additional specific character. Furcation of the annuli seems to be a
taxonomically less important feature, because it is not essential for coiling, and in addition on a
single specimen non-, bi-, and trifurcated ribs have been observed. The orientation of specimens
broken both aborally and orally is sometimes recognizably the inclination of annuli towards the
aboral end. Characters of isolated plates may be insufficient for specific identification because plates
of the same form are found in different configurations and associations on different worms. For
convenience it is suggested that new finds of such isolated sclerites should be published under open
nomenclature.

SYSTEMATIC PALAEONTOLOGY
Phylum UNCERTAIN

Class palaeoscolecida Conway Morris and Robison, 1986
Family palaeoscolecidae Whittard, 1953

Discussion. The family comprises solitary worms with more-or-less rounded cross-sections but
mostly with differentiated dorsal and ventral surfaces. Diameter of individual increases from
anterior, may remain constant over quite a long distance distally but may slightly decrease again
in that direction. Somewhat flexible multilamellar cuticle is armoured by skeletal plates. The
capability of limited movements, such as contraction and relaxation, is documented in the different

MULLER AND HINZ-SCHALLREUTER: CAMBRIAN WORMS

559

profiles of the annuli. The structure of the tissue is net-like and forms some sort of matrix for the
insertion of the sclerites. Annulations are distinct and may be partly bi- or trifurcated on the dorsal
side. On the same specimen, single rows may alternate irregularly with bi- or trifurcated rows (Text-
fig. lie). The outer skin of the annuli bear one to four rows of more-or-less regularly arranged
plates. In between are much smaller platelets. Both are mineralized components but their primary
substance is unknown; they were probably phosphatic. There is a certain variability in shape and
adornment of the plates, particularly in respect to their position either on the dorsal or the ventral
side. In some taxa, different morphotypes of plates have been observed which are generally arranged
in alternating rows.

Ornamentation of plates and organization of the armour are considered as the main criteria for
the recognition of species. A number of taxa have intercalations between the annuli that may have
facilitated bending and other limited movements. These zones may form a zigzag band and have
platelets that differ from those of the annuli in shape and arrangement.

The aboral end bears four nipples arranged in two lateral pairs. A longitudinal, slit-like opening
between these pairs extends over the whole diameter of the worm. The slit opens towards an
invagination.

Nipple-like protuberances similar to those at the aboral end may appear also on the ventral side,
and are sometimes paired. Platelets are arranged concentrically around their bases, evidence that
they are an original structure of the cuticle. Some specimens expose tubules (e.g. Text-fig. 5e-f)
which may be also spread irregularly over the whole surface (Text-fig. 13a-b, f) and in this are
similar to priapulids (Oeschger and Janssen 1991).

The oral end is usually broken. Even a well-preserved annulus cannot be regarded as the
termination of the animal itself; it might represent the end of heavy armour passing into a proboscis
similar to the ‘unidentified worms’ illustrated by Chen et al. (1991, fig. 7).

The soft body itself is not preserved on our material. Specimens flattened on shale frequently
show a dark, narrow, longitudinal band extending over the entire length of the individual that may
be interpreted as the gut. Glaessner (1979) even observed muscle fibres beneath the cuticle.

Size-range. Large, almost complete specimens of 01 0~015 m length have been observed only in shales. The
length : width ratio of almost complete small specimens that were etched from limestones suggests a length of
about 0 06 m for the largest, fragmentary individuals.

Occurrence and age. Lower Cambrian to Upper Silurian, Antarctica, Australia, China, Czechoslovakia,
England, Estonia, Greenland, Siberia, Spain, Turkey, USA.

Genus austroscolex gen. nov.

Type species. Austroscolex spatiolatus sp. nov.

Derivation of name. Referring to its occurrence in Australia and scolex , Greek (worm).

Diagnosis. Annulation fairly broad, no intercalations. Annuli with widely spaced circular plates
arranged in one or two rows. Median annular zone broad. Ornamentation of plates with central
elevation that may differentiate into four or five nodes. Margins of plates and platelets jagged.

560

PALAEONTOLOGY, VOLUME 36

MULLER AND HINZ-SCHALLREUTER: CAMBRIAN WORMS

561

Austroscolex primitivus sp. nov.

Text-fig. 4e, i

Derivation of name. From primitivus , Latin, referring to the simple ornamentation.

Holotype. CPC 23006; from 1 km north of Mt Murray, Duchess, Queensland; Triplagnostus gibbus Zone,
Middle Cambrian.

Illustrated material. CPC 23006.

Other material. 14 specimens.

Diagnosis. Broad annulation with predominantly a single row of plates close to an annulus border.

Description. Isolated, laterally-compressed fragments of large worms. Annuli fairly broad and mainly with a
single row of small plates that is located close to ?oral annulus border. Opposite side only occasionally with
some plates. Second row only incompletely developed. Widely spaced plates with relatively large central
elevation, and jagged margins of both plates and platelets similar to A. spatiolatus. Plates: platelets ratio about
1:2.

Austroscolex spatiolatus sp. nov.

Text-fig. 4c-d, f-h, k-l

Derivation of name. From spatium, Latin (space) and latus, Latin (broad) referring to the broad interspace
between plates.

Holotype. CPC 23003; from 1 km north of Mt Murray, Duchess, Queensland; Triplagnostus gibbus Zone,
Middle Cambrian.

Illustrated material. CPC 23003-23005.

Other material. 67 specimens.

Diagnosis. Two rows of plates within an annulus; one row with smaller plates than the opposite row.
Other characters as for the genus.

Description. Fragments of fairly large worms. Annulation quite regular with distinct apical inclination.
Ornamentation of annuli with double row of widely spaced circular plates that differ considerably in size.
Orally directed row bears comparatively smaller plates than opposite one. Distribution of plates within rows
irregular and independent from each other. Plates with elevated centres with four small cones. Diameter of

text-fig. 4. a-b, Palaeoscolecida gen. et sp. indet. A, CPC 23001 (sample 7304); north of Rogers Ridge, late
Templetonian, Triplagnostus gibbus Zone; fragmentary specimen showing postmortal hypha-hke threads on
the inner surface, x 85. b, CPC 23002 (sample 7336); 1km north of Mt Murray, late Templetonian,
Triplagnostus gibbus Zone; fragmentary specimen with distinct, longitudinal structure in the median plane,
probably secondary, x 50. C-D, f-h, k-l, Austroscolex spatiolatus gen. et sp. nov. c-d, f, CPC 23003 (sample
7340), holotype; 1 km north of Mt Murray (= D640), late Templetonian, Triplagnostus gibbus Zone, c, general
view, x 30. d, detail of surface, x 150. F, detail of surface, x 900. G-H, CPC 23004 (sample 7262); Rogers
Ridge, late Templetonian, Triplagnostus gibbus Zone. G, general view, x40. h, detail of surface; note matrix
of exfoliated portion, x 215. k-l, CPC 23005 (sample 7262); Rogers Ridge, late Templetonian, Triplagnostus
gibbus Zone. K, detail of surface, x 235. l, general view, x 60. E, I, Austroscolex primitivus gen. et sp. nov., CPC
23006 (sample 7336), holotype; 1 km north of Mt Murray (= D640), late Templetonian, Triplagnostus gibbus
Zone, e, general view, x 50. i, detail of surface, x 500.

562

PALAEONTOLOGY, VOLUME 36

text-fig. 5. Corallioscolex gravius gen. et sp. nov. All specimens from 1 km north of Mt Murray (= D640),
late Templetonian, Triplagnostus gibbus Zone, except where otherwise stated, a-b, d, CPC 23007 (sample
7336), holotype. a, general view of fragment, x 110. b, detail of coralliomorph outer surface with radially
arranged platelets, x 1900. d, detail of armour with irregularly distributed plates; note folds at annulus border.

MULLER AND HINZ-SCH ALLREUTER: CAMBRIAN WORMS

563

entire plate is about twice as wide as the centre. Marginal rim jagged. Platelets with irregularly polygonal
outline and jagged margins. Upper surface topped by minute tubercles. Plates: platelets ratio about 1 :4. On
one specimen the uppermost layers of the cuticle are exfoliated and show a matrix in which the sclerites were
anchored (Text-fig. 4h).

Comparison. The main differences between Aaustroscolex primitivus and A. spatiolatus are the
incompletely developed second row of plates and the generally smaller size of the plates in
A. primitivus , as well as the different upper surface of the plates.

Genus corallioscolex gen. nov.

Type species. Corallioscolex gravius sp. nov.

Derivation of name. From corallium, Latin (coral), referring to the coralliomorph surface sculpture.

Diagnosis. Annuli with two to three rows of more-or-less equal plates. Intercalations not developed.
Plates with distinct central depression, surrounded by several nodes. Platelets forming coralliomorph
sculpture.

Corallioscolex gravius sp. nov.

Text-fig. 5

Derivation of name. From gravius, Latin (beautiful) referring to the surface ornamentation.

Holotype. CPC 23007; from 1 km north of Mt Murray, Duchess, Queensland; Triplagnostus gibbus Zone,
Middle Cambrian.

Illustrated material. CPC 23007-23011.

Other material. 5 specimens.

Diagnosis. As for the genus.

Description. Fragments of large worms with broad annulation. Intercalations are lacking; instead, stretching
of the worm exposes laminae between the annuli (Text-fig. 5h). Annuli ornamented by two to four rows of
widely- and irregularly-shaped plates.

The plates vary in size and have a depressed centre surrounded by comarginal nodes. On a number of plates
the surface seems to be almost completely abraded (Text-fig. 5k). Basal portion of plates appears as a girdle
and consists of densely-spaced, irregularly polygonal platelets. Diameter of plate at its base is about one-and-
a-half times larger than at the surface (Text-fig. 5i). The plate is embedded in a reticulate or cellular structure
that appears as matrix for the overlying coralliomorph platelets (Text-fig. 5l-m).

The platelets form rosette-like, coralliomorph structures (Text-fig. 5b). They are composed of irregular,
radially orientated blades, topped by larger, elongate sclerites with distinctly convex upper surface; four to six

x 400. c, e-f, CPC 23008 (sample 7338). c, general view of large, fragmentary specimen, x 40. e, annulated
tubule on outer surface, x 1150. F, detail showing plates, platelets and proximally broken tubules, x 800.
G, K, CPC 23009 (sample 7336). G, general view, x 55. K, detail of outer surface with strongly abraded plates,
x 465. H, CPC 23010 (sample 7324), Mt Murray, late Templetonian; surface detail showing stretched annulus
border, x950. i, l-n, CPC 23011 (sample 7338). i, surface detail of single plate; the broad basal portion and
part of the girdle were originally covered by secretory tissue, x 950. l, detail of inner layers with both
coralliomorph and labyrinthic structure; an exfoliated portion has been accidentally attached to the
coralliomorph surface, x 950. m, detail of surface showing structure of an underlying layer with labyrinthic

structure, x 500. N, general view of fragment, x 50.

564

PALAEONTOLOGY, VOLUME 36

of these sclerites are positioned on each rosette as radiating septa-like structures. Margins of platelets are partly
jagged. Plates : platelets ratio about 1 :4 or 1:3.

Another characteristic of the cuticle is the presence of tubules that seem to be annulated proximally (Text-
fig. 5e).

Genus euryscolex gen. nov.

Type species. Euryscolex paternarius sp. nov.

Derivation of name. From eurys , Greek (broad) referring to the broad median annular zone.

Diagnosis. Broad annulation, no intercalations. Annuli with three to four rows of widely-spaced
plates. Concave upper side of plates with radial sculpture. Platelets with jagged outer margin.

Euryscolex paternarius sp. nov.

Text-fig. 15a-d

Derivation of name. From patera , Latin (bowl), referring to the shape of the sclerites.

Holotype. CPC 23053; from 1 km north of Mt Murray, Duchess, Queensland; Triplagnostus gibbus Zone,
Middle Cambrian.

Illustrated material. CPC 23052-23053.

Other material. 4 specimens.

Diagnosis. As for the genus.

Description. Wall fragments of large worms, with annulus borders indicated only by narrow furrows.
Intercalations not observable. Annuli with flat relief and widely- but irregularly-spaced plates of different size.
Proper rows not distinguishable. Plates circular, depressed centre with faint radial ribbing. Margin elevated
and smoothly scalloped, topping a steep girdle with comparatively coarse vertical ribs. Platelets tiny, with
jagged margins, surface differentiation not recognizable; arranged in a distinct mosaic pattern of larger
polygonal units that are, however, still much smaller than the plates (Text-fig. 1 5d). Plates : platelets ratio about
1 :4.

Genus hadimopanella Gedik, 1977
Type species. Hadimopanella oezguli Gedik, 1977.

Diagnosis (after Gedik 1977, 1989). A phosphatic circular unit with strongly convex upper surface
decorated by tubercles in its central part and a slightly convex to plane and smooth lower surface.
The tubercles are coarse, i.e. greater than 10 //m and regularly distributed. Most tubercles are on
the outer circular row and a few on the central part. The tuberculated area forms about half of the
full diameter. The non-tuberculated marginal area, called marginal brim, is broad. The
height: diameter ratio is about 1:3.

Remarks. Sclerites similar to Hadimopanella oezguli were discovered in two different configurations
which, from what is known of other more complete worms, probably characterize different taxa. As
they are represented only by one or two fragments, they are described herein as form species I and
II.

MULLER AND HINZ-SCHALLREUTER: CAMBRIAN WORMS 565

text-fig. 6. Kaloscolex granulatus gen. et sp. nov. All specimens from 1 km north of Mt Murray (= D640),
late Templetonian, Triplagnostus gibbus Zone, a-b, CPC 23012 (sample 7338). a, general view of terminally
broken and wrinkled specimen, x 30. b, detail of surface; note irregular distribution of plates, x 215. c-d, CPC
23013 (sample 7336). c, surface detail; note ornamentation of platelets, x 300. d, general view of specimen,
x40. E, CPC 23014 (sample 7339); surface detail showing underlying layer with less distinct plates and more
closely-spaced platelets, x 300. f, i, CPC 23015 (sample 7338), holotype. F, lateral view, x 40. 1, detail showing
strong annular relief and furcation of ribs on dorsal side; ventral side flattened; ratio of platelets: plates
increased in this portion, x 235. g-h, CPC 23016 (sample 7338). G, surface detail of armour, x 265. H, general

view of large, deformed specimen, x 50.

566

PALAEONTOLOGY, VOLUME 36

lOOpm

100|jm

text-fig. 7. a-c, e-f, Milaculum elongatum sp. nov. a-c, CPC 23017 (sample 7336); 1 km north of Mt Murray,
late Templetonian, Triplagnostus gibbus Zone, a, general view of fragment representing about the width of an
annulus; note elongate platelets between plates, x 1 10. B, profile of single plate with fine radial ribs and jagged
outer, basal margin, x 800. c, surface detail with broken tubule, x 1550. e-f, CPC 23018 (sample 7336),
holotype; 1 km north of Mt Murray, late Templetonian, Triplagnostus gibbus Zone. E, piece of large specimen
with broad, flat annuli, x 100. f, detail showing ornamentation of plates and mosaic pattern of platelets, x 285.
d, G, aff. Hadimopanella oezguli form species 1; CPC 23019 (sample 7324); Mt Murray, late Templetonian,
Triplagnostus gibbus Zone, d, fragmentary, deformed specimen, x 95. G, detail of surface with plate and
adjacent nipple-like protuberance, x 365. h-k, aff. Hadimopanella oezguli form species II. H, CPC 23020

MULLER AND HINZ-SCH ALLREUTER: CAMBRIAN WORMS

567

aff. Hadimopanella oezguli Gedik, 1977

1977 Hadimopanella oezguli Gedik, p. 46, pi. 5, figs 1-5.

1988 Hadimopanella oezguli Gedik; Marss, p. 14, pi. 1, figs 1-8.

1989 Hadimopanella oezguli Gedik; Gedik, p. 69, pi. 1, figs 1-2.

1990 Hadimopanella oezguli Gedik; Hinz et al., fig. Id.

Form species I
Text-fig. 7d, g

Material. Figured specimen, CPC 23019.

Description. Piece of armoured cuticle with terminations and cross-section indeterminable. Annulation fairly
broad, intercalations not recognizable. Each annulus with a double row of equal plates. They have nearly the
same size and are approximately regularly arranged with distinct interspaces between them. Plates rounded,
with a circle of 6 or 7 marginal nodes and a central node that may be somewhat larger. Platelets polygonal in
the middle of the annuli; between plates, platelets are elongate and extending parallel to long axis of the worm.
All platelets have jagged margins that are particularly obvious close to the underside. The surface is further
characterized by tubules.

Remarks. This form seems to be similar to Palaeoscolex sinensis Hou and Sun, 1988, from the Lower
Cambrian Cheng-jiang fauna with Eoredlichia , in which the plates are similarly arranged and
ornamented, but the illustrations do not show the critical details. Furthermore, the authors state a
double row of ribs in the median zone of each annulus; by contrast Hadimopanella oezguli has
polygonal platelets. The latter feature, however, may be preservational; this can be confirmed only
by having suitable material at hand.

Form species II
Text-fig. 7h-k

Material. 2 specimens, CPC 23020-23021.

Diagnosis. Cuticle with densely-spaced plates. Border of annuli indistinct.

Description. Piece of wall of a large specimen. Annulation indistinct. Surface is covered by densely-spaced,
rounded plates. Surface of plates distinctly convex with central node encircled by further nodes of the same size.
Narrow interspace between plates filled with rounded, nodular platelets.

Genus kaloscolex gen. nov.

Type species. Kaloscolex granulatus sp. nov.

Derivation of name. From halos, Greek (beautiful) referring to the surface ornamentation.

Diagnosis. Annulation dense, partly furcated. Intercalations lacking. Annuli with one to two rows
of plates; upper surface of plates with four to five nodes. Platelets with median elevation leading to
a granular appearance of the cuticle.

(sample 7337); 1 km north of Mt Murray, late Templetonian, Triplagnostus gibbus Zone; piece of armoured
cuticle with indistinct annulation, x200. i, k, CPC 23021 (sample 7336); I km north of Mt Murray, late
Templetonian, Triplagnostus gibbus Zone. I, piece of densely armoured cuticle, x 90. K, detail with

ornamentation of plates and microplates, x 490.

568

PALAEONTOLOGY, VOLUME 36

Kaloscolex granulatus sp. nov.

Text-fig. 6

Derivation of name. From granulatus , Latin, referring to the granular ornamentation of platelets.

Holotvpe. CPC 23015; from 1 km north of Mt Murray, Duchess, Queensland; Triplagnostus gibbus Zone,
Middle Cambrian.

Illustrated material. CPC 23012-23016.

Other material. 3 specimens.

Diagnosis. As for the genus.

Description. Curved specimens with slight increase in diameter towards oral end; cross-section undeterminable,
intercalations lacking. Annuli partly with furcation on the convex side; accordingly, the ornamentation differs
between convex and concave sides: the convex face displays plates, the concave side exposes only irregular
platelets and erratic pillars (Text-fig. 6i). According to the curvature, the convex side is regarded as dorsal, and
the concave side as ventral.

Annuli ornamented by an irregular arrangement of widely spaced plates in one or two approximate lines.
Sometimes, smaller plates are randomly added. Size of plates highly variable, outline crudely rounded with
irregular margin. Plates with four to five nodes around the centre.

Platelets of two types: (a) oval sclerites, with distinct median elevation, more-or-less concentrically arranged
around plates; (b) smaller, polygonal to elongate platelets with irregular margin, surrounding the first type.

On exfoliated specimens, i.e. on a lower level of the cuticle, the convex platelets (type a) are more densely-
spaced. Plates are recognizable only in an incipient stage (Text-fig. 6e). Plates: platelets ratio about 1:1.
Tubules situated on the lateral face appear segmented.

Genus milaculum Muller, 1973
1989 Plasmuscolex Kraft and Mergl, p. 25.

Type species. Milaculum ruttneri Muller, 1973.

Milaculum elongation sp. nov.

Text-fig. 7a-c, e-f

Derivation of name. From elongatus, Latin, referring to the shape of the plates.

Holotype. CPC 23018; from 1 km north of Mt Murray, Duchess, Queensland; Triplagnostus gibbus Zone,
Middle Cambrian.

Illustrated material. CPC 23017-23018.

Other material. 43 specimens.

Diagnosis. Annulation very broad and flat; intercalations lacking. Annuli with two rows of plates
ornamented by vertical rows of nodes. Elongate platelets between individual plates of one row.
Median annular zone and border of annuli with polygonal platelets.

Description. Fragmentary large worms with fairly wide annuli; intercalations lacking. Ornamentation of annuli
with two rows of elongate plates arranged along the borders of each annulus. Spacing of plates distinct and
regular.

Plates subsymmetrical with elongate to oval outline and tapering very slightly towards the central area of
each annulus. Outer margin of plates smooth. Sculpture with comarginal, outwardly directed nodes or cones.

MULLER AND HINZ-SCH ALLREUTER: CAMBRIAN WORMS

569

One to three median cones may be present even within the same row of an annulus. A faint radial ribbing is
observable (Text-fig. 7b).

Platelets rather small in comparison with the large plates. Regarding their outer shape, two types of platelets
are developed: polygons, and narrow, elongate platelets. Towards the borders of each annulus, relatively large
polygonal platelets are present; by contrast, in the central area there are only small polygons. Between the
plates there are elongate platelets reflecting the outline of the plates. Up to about ten longitudinal rows of
platelets have been observed between two plates (Text-fig. 7a). All sclerites have minutely jagged outer margins
(Text-fig. 7b). Plates: platelets ratio about 2:1.

Remarks. There is some similarity to Hadimopanella oezguli in the presence of elongate platelets
between the plates, but shape and ornamentation of plates are distinctly different.

Genus murrayscolex gen. nov.

Type species. Murrayscolex serratus sp. nov.

Derivation of name. From its occurrence at Mt Murray, Queensland.

Diagnosis. Annulation moderately broad, intercalations lacking. Annuli with two rows of
alternating plates, one row with larger plates than the opposite row. Upper surface of plates
nodular, outer margin jagged. Microplates between plates in one row.

Murrayscolex inaequalis sp. nov.

Text-fig. 8d-f

Derivation of name. From Latin, referring to the unequal development of plates in both rows of an annulus.

Holotype. CPC 23026; from 1 km north of Mt Murray, Duchess, Queensland; Triplagnostus gibbus Zone,
Middle Cambrian.

Illustrated material. CPC 23026.

Other material. 2 specimens.

Diagnosis. Shape of plates tapering towards respective annulus borders. Upper surface of plates and
microplates highly ornamented. Other characters as for the genus.

Description. Parts of very large worms; annulation fairly broad, intercalations lacking. Ornamentation of
annuli with two rows of plates, one row with larger and more elongate plates than the opposite row. Elongate
plates quite closely spaced; towards centre of annulus they are broadly rounded, tapering slightly towards its
border. Broader portion with up to six fairly large nodes, smaller area with numerous tubercles that may also
be comarginal. Between plates, smaller, subtriangular microplates with tuberculate surfaces are intercalated.
Opposite rows with smaller, rounded plates of similar ornament and with microplates between. Interspace filled
by irregular platelets with jagged margins. Plates : platelets ratio about 3:1. On one specimen (Text-fig. 8e-f)
a distinct, segmented protuberance is visible.

Remarks. This form differs from Murrayscolex serratus mainly in the development of the larger
plates with their differentiated outer shapes and a tuberculated terminal zone, and in the narrower
central area of the annuli.

570

PALAEONTOLOGY, VOLUME 36

TEXT-FIG. 8. Murrayscolex gen. nov. a-c, G-k, Murrayscolex serratus sp. nov. A-B, CPC 23022 (sample 7390),
holotype; Phosphate Hill, excavation no. 9, late Templetonian, Triplagnostus gibbus Zone, a, general view,
x 1 15. b, detail of armour, x 535. c, CPC 23024 (sample 7337); 1 km north of Mt Murray (= D640), late
Templetonian, Triplagnostus gibbus Zone; lateral view of curved specimen with broken aboral and oral ends.

MULLER AND HINZ-SCH ALLREUTER: CAMBRIAN WORMS

571

Murrayscolex serratus sp. nov.

Text-fig. 8a-c, g-k

Derivation of name. From serratus , Latin, referring to the jagged margin of sclerites.

Holotype. CPC 23022; from Phosphate Hill, excavation no. 9, Duchess, Queensland; Triplagnostus gibbus
Zone, Middle Cambrian.

Illustrated material. CPC 23022-23025.

Other material. 2 specimens.

Diagnosis. Plates close to annulus borders. Ornamentation of plates with elevated centre, encircled
by tubercles. Median annular zone variable in width.

Description. Large fragments without preserved oral and aboral ends; cross-section nearly circular. Annulation
broad and fairly regular with convex dorsal and flattened ventral relief. Ornamentation of annuli with two rows
of irregularly alternating plates positioned close to borders of annuli. Rows are distinguished only by the size
of the plates. Plates rounded to elongate with three to four larger, partly fused central nodes, encircled by
minute, irregularly-spaced nodes. Towards annulus border, the nodular row may be doubled. Plates generally
highly variable in size and ornament. Microplates, with sculpture similar but not identical to plates, are
positioned between the plates and along borders of annulus. Platelets developed as small polygons with convex,
hardly differentiated upper surface. They are positioned in central zone of annulus. Platelets within
intercalations are oval-shaped with multinodular surface. Both plates and platelets have jagged margins.
Plates : platelets ratio about 1: 1 (Text-fig. 8a) or 1:2 (Text-fig. 8g-k).

Genus pantoioscolex gen. nov.

Type species. Pantoioscolex oleschinskii sp. nov.

Derivation of name. From pantoios, Greek (of all sorts) referring to the high variability in size of the sclerites.

Diagnosis. Ornamentation consisting of narrow rows of small plates alternating with zones of more
densely-spaced microplates. Ornamentation of sclerites with four to six nodes. Annulus borders
indistinct.

Pantoioscolex oleschinskii sp. nov.

Text-fig. 16a, c

Derivation of name. In honour of Mr Georg Oleschinski, Bonn.

Holotype. CPC 23057; from 1 km north of Mt Murray, Duchess, Queensland; Triplagnostus gibbus Zone,
Middle Cambrian.

x 40. G, i, K, CPC 23023 (sample 7350); 1 km north of Mt Murray (= D640), late Templetonian, Triplagnostus
gibbus Zone. G, oblique view from above, x 75. i, details of armour, x 265. k, detail of armour, x 490. H, CPC
23025 (sample 7335); 1 km north of Mt Murray (= D640), late Templetonian, Triplagnostus gibbus Zone;
surface detail of specimen with strongly abraded plates, almost down to planation, x 750. d-f, Murrayscolex
inaequalis sp. nov. CPC 23026 (sample 7336), holotype; 1km north of Mt Murray (=D640), late
Templetonian, Triplagnostus gibbus Zone, d, general view of large, compressed fragment, x 100. E, detail of
ornamentation with more or less triangular microplates between plates, x 550. F, surface detail with annulated

tubules, distally broken-off, x 3000.

572

PALAEONTOLOGY, VOLUME 36

Material. Figured specimen, CPC 23057.

Diagnosis. As for the genus.

Description. Large fragment of a single specimen that differs from all other taxa in its surface pattern. Annulus
borders indistinct, relief flat. Ornamentation with narrow rows of small plates in somewhat irregular
arrangement alternating with broad zones of more densely-spaced microplates. Sclerites highly variable in size,
and surface topped with circle of four to six nodes. Interspaces smooth, platelets not recognizable.
Plates : interspace ratio about 1 : 2.

Remarks. With regard to the surface pattern, there is superficial similarity between Pantoioscolex
oleschinskii and Palaeoscolex antiquus Glaessner, 1979. Both taxa show an alternation between rows
of larger plates and comparatively broad zones with smaller, densely spaced plates.

Conway Morris and Robison (1986) presented drawings of the ornamentation patterns of
Palaeoscolex piscatorum Whittard, 1953, P. ratcliffei Robison, 1969 and P. cf. ratcliffei Conway
Morris and Robison, 1986, as well as P. antiquus Glaessner, 1979. A paratype of P. piscatorum (RU
4200 and RU 4200 counterpart) in the collections of the British Geological Survey, shows a net-like
sculpture comparable to Whittard’s illustration (Whittard 1953, pi. 5, fig. 2). It is similar to the
matrix illustrated on exfoliated specimens (e.g. Text-fig. 4h) and suggests that not the outer surface
but an inner layer is exposed. As has been shown for Kaloscolex granulatus (Text-fig. 6e), the
sculpture may differ considerably between outermost and inner layers. Therefore, P. piscatorum is
unsuitable for direct comparison.

Genus rhomboscolex gen. nov.

Type species. Rhomboscolex chaoticus sp. nov.

Derivation of name. From rhombus , Latin, referring to the surface ornamentation by platelets.

Diagnosis. Broad, flat annulation with two rows of widely spaced plates. Intercalations lacking.
Ornamentation of plates with circle of tear-shaped nodes around central depression. Platelets
forming a superficial rhombic pattern, particularly in broad, central area of annuli.

Rhomboscolex chaoticus sp. nov.

Text-fig. 9

Derivation of name. From chaos , Latin, referring to the chaos theory and the repetition of non-identical
structures.

Holotype. CPC 23028; from 1 km north of Mt Murray, Duchess, Queensland; Triplagnostus gibbus Zone,
Middle Cambrian.

Illustrated material. CPC 23027-23030.

Other material. 3 specimens.

Diagnosis. As for the genus.

Description. Laterally compressed fragments of large individuals without preserved oral and aboral ends.
Annuli somewhat variable in width but generally fairly broad and without pronounced relief. Annuli with two
widely spaced rows of unequal plates, each row close to either border. Although highly variable in size, plates
are generally larger in one row than in the opposite row.

MULLER AND HINZ-SCHALLREUTER: CAMBRIAN WORMS

573

text-fig. 9. Rhomboscolex chaoticus gen. et sp. nov. a, c, CPC 23027 (sample 7340); Rogers Ridge, late
Templetonian, Triplagnostus gibbus Zone, a, general view of large, compressed fragment, x 40. c, detail of
annulus ornamentation; note different size of plates, x 365. b, d, CPC 23028 (sample 7338), holotype; 1 km
north of Mt Murray (= D640), late Templetonian, Triplagnostus gibbus Zone, b, surface detail with platelets
forming a superficial rhombic pattern, x250. d, general view of specimen, x 35. e-g, CPC 23029 (sample
7324); Mt Murray, late Templetonian, Triplagnostus gibbus Zone, e, surface detail of exfoliated portion
exposing crumpled cuticular structure at the base of the sclerites; plates ornamented with four high cones,
x 700. F, general view of specimen, x 55. G, surface detail; here the rhombic pattern is stressed by larger nodes
on each corner, x 215. h, CPC 23030 (sample 7339); 1 km north of Mt Murray (= D640), late Templetonian,
Triplagnostus gibbus Zone; surface detail of specimen in which the plate has its girdle exfoliated, x 950.

574

PALAEONTOLOGY, VOLUME 36

text-fig. 10. Schistoscolex gen. nov. a-c, e-g, i, l-m, Schistoscolex angustosquamatus sp. nov. a -b, CPC 23031
(sample 7336), holotype; I km north of Mt Murray (= D640), late Templetonian, Triplagnostus gibbus Zone.
a, ventrolateral view of specimen with preserved aboral portion, x 95. b, detail of surface ornamentation.

MULLER AND HINZ-SCHALLREUTER: CAMBRIAN WORMS

575

Plates rounded to oval-shaped with depressed centre encircled by tear-shaped nodes, which are arranged in
rosettes with inwardly directed tips. Girdle steep and passing into a flattened brim with jagged margin. Plates
are encircled by several rows of platelets (Text-fig. 9b).

Platelets are generally quite uniform, rounded to polygonal, partly with jagged margins, strongly convex and
developed in two size-orders. Larger platelets appear to build a trellis ornament. Superficial rhombic
arrangement sometimes stressed by larger, almost globular platelets topping the corners of each rhomb (Text-
fig. 9g). On a small portion of a single specimen, plates are extremely densely-spaced and ornamented by four
to five strikingly high cones (Text-fig. 9e-f). The significance of this feature is uncertain.

In general the outer layer consists of convex platelets beneath which a crumbled cuticular structure is
observable on a partly exfoliated specimen (Text-fig. 9e). Plates : platelets ratio about 1 :4.

Remarks. Although the rhombic structure appears consistent, closer inspection reveals that every
rhomb differs from the adjacent rhombs, which is not due to deformation, but is a primary feature.
In chaos theory the term fractal has been used for such a pattern which does not repeat itself.

Most of the specimens have a considerable phosphatic coating that conceals the fine original
structures.

Genus schistoscolex gen. nov.

Type species. Schistoscolex umbilicatus sp. nov.

Derivation of name. From schistos, Greek (separate), referring to the furcation of annuli.

Diagnosis. Annulation narrow, partly furcated. Intercalations small. Annuli with two rows of
relatively large plates in contact with each other. Surface of plates with one to four larger nodes.
Marginal tubercles may be developed towards annulus borders.

Schistoscolex angustosquamatus sp. nov.

Text-fig. 10a-c, e-g, i, l-m

Derivation of name. From angustus, Latin (narrow), and squamatus, Latin (adorned with plates), referring to
the dense arrangement of plates.

Holotype. CPC 23031; from 1 km north of Mt Murray, Duchess, Queensland; Triplagnostus gibbus Zone,
Middle Cambrian.

Illustrated material. CPC 23031-23035.

Other material. 8 specimens.

x 300. c, E, G, CPC 23032 (sample 7310), holotype; north of Rogers Ridge, late Templetonian, Triplagnostus
gibbus Zone, c, general view of fragmentary specimen, x 1 15. e, surface detail showing exfoliated portion with
matrix for the insertion of plates, x 400. G, surface detail with plates along annular border, x 400. F, CPC
23033 (sample 7341); 1 km north of Mt Murray (= D640), late Templetonian, Triplagnostus gibbus Zone;
lateral view of specimen with rather irregular annulation, x40. i, CPC 23034 (sample 7262); Rogers Ridge,
late Templetonian, Triplagnostus gibbus Zone; detail of specimen showing dorsal bifurcation of ribs, x 350.
l-m, CPC 23035 (sample 7337); I km north of Mt Murray (= D640), late Templetonian, Triplagnostus gibbus
Zone. L, general view of fragmentary specimen, x 100. M, detail showing inner surface and strongly folded
(contracted) profile of ribs, x 285. d, h, k, Schistoscolex mucronatus sp. nov., CPC 23036 (sample 7332),
holotype; 1 km north of Mt Murray (= D640), late Templetonian, Triplagnostus gibbus Zone, d, general view
of fragmentary and deformed specimen, x 125. H, detail showing strongly aborally inclined spines, x 365.

K, detail of armourment with alternating spiny plates, x 900.

576

PALAEONTOLOGY, VOLUME 36

Diagnosis. Annulation narrow, partly furcated; intercalations indistinct. Plates densely arranged,
ornamentation with one or two, large, inclined central nodes. Smaller nodes may be present towards
annulus border.

Description. Curved individuals with circular cross-section. Annulation narrow, partly bifurcated. Inter-
calations indistinct. Each annulus with two rows of large plates forming an acute angle along the median plane
in the contracted condition of the worm (Text-fig. 10e). Plates in close contact with each other. Elevated central
part of plates with four to five irregular nodes or cones, differentiating into one larger node positioned towards
the midline of the annulus and pointing to the annulus border. Several smaller and irregular nodes at opposite
side. The smaller nodes of the sclerites are always positioned along the annulus borders, the larger nodes are
developed in the middle of the annuli. Platelets are only developed as fillings of the small interspaces. Their
shape is thus dependent on the space available and appears irregularly elongate. Between the plates, towards
the annulus borders, there is a roughly oval to triangular platelet with a distinct median node (Text-fig. 10g).
Plates : platelets ratio about 9:1.

Remarks. There are similarities in the furcation of ribs, and arrangement and density of plates
between Schistoscolex angustosquamatus and S. umbilicatus. The main difference between the two
species is the inclined subcentral node in S. angustosquamatus.

Schistoscolex mucronatus sp. nov.

Text-fig. 10d, h, k

Derivation of name. From mucronatus , Latin (adorned with a spine or thorn), with reference to the spine-like
subcentral node.

Holotype. CPC 23036; from 1 km north of Mt Murray, Duchess, Queensland; Triplagnostus gihhus Zone,
Middle Cambrian.

Illustrated material. CPC 23036.

Other material. 1 specimen.

Diagnosis. Plates smaller than in other species; microplates irregularly distributed in central annular
zone. Ornamentation of plates with large, spine-like central elevation. Other characters as for the
genus.

Description. Fragmentary worms without preserved aboral and oral ends. Width of annuli somewhat variable,
convexity depending on degree of contraction. Intercalations not recognizable. Ornament of annuli with two
rows of alternating, relatively large, closely-spaced plates. Towards the border, a submedian, spine-like cone
is surrounded by a few nodes. The central zone of each annulus may be filled by polygonal to irregular platelets
or microplates that share the ornament with the plates. Plates: platelets ratio about 9:1.

Remarks. Oral and aboral ends of fragmentary worms may be deduced from the annular
inclination: the steep ‘stoss’ side is superimposed by the ‘lee’ side of the preceeding annulus. In this
respect, the visible spines are assumed to point towards the aboral end.

This form differs from Schistoscolex angustosquamatus in the spine-like development of the
median cone and in the convexity and inclination of annuli. Intercalations cannot be observed but
the general pattern of ornamentation indicates that the species belongs to the same genus.

Schistoscolex umbil ictus sp. nov.

Text-figs 1 1-12

Derivation of name. From umbilicus , Latin (navel) referring to the ornamentation of plates with elevated centre.

MULLER AND HINZ-SCH ALLREUTER: CAMBRIAN WORMS 577

text-fig. 11. Schistoscolex umbilicatus gen. et sp. nov. a-b, CPC 23037 (sample 7335); 1 km north of Mt
Murray (= D640), late Templetonian, Triplagnostus gibbus Zone. A, dorsolateral view, x85. b, aboral view
showing nipple-like appendages, x 265. c, E, CPC 23038 (sample 7335); I km north of Mt Murray ( = D640),
late Templetonian, Triplagnostus gibbus Zone, c, ventrolateral view documenting ventral flattening of annulus
profile, x 135. E, detail exposing narrow intercalations and dorsal furcation of ribs, x 265. d, f, CPC 23039
(sample 7324); Mt Murray, late Templetonian, Triplagnostus gibbus Zone, d, planispirally enrolled specimen,
x 80. f, detail of strongly abraded surface ornamentation, x 1550. g-h, CPC 23040 (sample 7340); 1 km north
of Mt Murray (= D640), late Templetonian, Triplagnostus gibbus Zone. G, aboral view showing dorsoventral
opening and two lateral pairs of nipples; the ventrally positioned nipples are larger than the dorsal ones, x 365.

h, lateral view of slender, curved specimen, x 25.

Holotype. CPC 23041; from 1 km north of Mt Murray, Duchess, Queensland; Triplagnostus gibbus Zone,
Middle Cambrian.

Illustrated material. CPC 23037-23042.

Other material. 10 specimens.

578

PALAEONTOLOGY, VOLUME 36

text-fig. 12. Schistoscolex umbilicatus gen. et sp. nov. All specimens from 1 km north of Mt Murray (= D640),
late Templetonian, Triplagnostus gibbus Zone, a-d, CPC 23041 (sample 7335), holotype. a, oblique side view
of terminally broken specimen, x 45. b, surface detail exposing underlying layer, x 100. c, detail with dorsal
furcation of ribs, x 200. D, detail showing circular structure; this is regarded as a primary feature as the
platelets are arranged concentrically around it, x 500. e-g, CPC 23042 (sample 7337). E, surface detail with
irregular annulation, x 465. F, surface detail with furcation, x 550. G, general view of deformed specimen,

x 40.

Diagnosis. Annulation narrow, partly furcated. Intercalations with elongate microplates. Annuli
with two rows of closely spaced plates. Ornamentation of plates with one or two median nodes and
smaller tubercles towards annulus border.

Description. Worms curved to planispirally enrolled; diameter of worm equal over entire length except for
initial part and a slight increase towards the distal termination. Cross section circular. Aboral end with four,
slightly outwardly-directed nipples. According to the curvature of the worm they are arranged as a smaller
dorsal and a larger ventral pair (Text-fig. 12g). An elongate opening between these pairs results from the
inwardly folded cuticle (Text-fig. 17a-c).

Annulation irregular in width. Dorsal bi- or trifurcation with an increased number of rows of plates observed
(e.g. Text-fig. 12c, f). Furcated ribs highly convex and narrowly folded dorsally, the ventral side being more

MULLER AND HINZ-SCHALLREUTER: CAMBRIAN WORMS

579

flattened. Intercalations very narrow and consisting of elongate platelets that are irregular in shape and
arrangement.

Ornamentation of annuli with two rows of comparatively large plates. Both rows consist of similar plates
that are in direct contact with each other or leave only a narrow interspace. They are rounded with irregular
margins and characterized by one or two median cones. Some comarginal tubercles may be developed towards
the annulus borders. Both rows are separated by a mosaic pattern of platelets. On the ventral face the plates
are more irregularly arranged than on the dorsal face (Text-fig. 12f).

Platelets irregular in shape, varying from almost hexagonal to elongate or completely irregular. A nodular
sculpture may be present. Elongate platelets occur predominantly between the plates of a single row.
Plates : platelets ratio about 7:3.

Schistoscolex sp. indet.

Text-fig. 13

Illustrated material. CPC 23043-23045; from 1 km north of Mt Murray, Duchess, Queensland; Triplagnostus
gibbus Zone, Middle Cambrian.

Other material. 1 specimen.

Description. Gently curved worms with circular cross-section. Annuli slightly variable in width on the dorsal
side; ventral width much more irregular and dependent on the degree of curvature. Intercalations sometimes
overlapped on the ventral side (Text-fig. 13e). Ornamentation of annuli with irregularly arranged plates. Plates
with irregularly rounded outline and central elevation. Interspaces filled with small, irregularly polygonal
platelets.

Ventral side of worms characterized by more-or-less paired nipples on each third or fourth annulus (Text-
fig. 13c-e). The nipples are outwardly directed. Smaller nipples indicated close to larger nipples on the same
or on an adjacent annulus. Around the nipples, the cuticle is folded and the annulus is widened.

Remarks. Structures which may represent the broken bases of tubules (compare also Text-figs 5e-f
and 8e-f) appear to be randomly distributed over the entire surface. They are surrounded by
concentrically arranged tiny plates. These tubules are similar to those described in the praipulid
Halicryptus spinutosus by Oeschger and Janssen (1991). Plates: platelets ratio about 1 :2.

Genus shergoldiscolex gen. nov.

Type species. Shergoldiscolex nodosus sp. nov.

Derivation of name. In honour of Dr John H. Shergold, Bureau of Mineral Resources, Canberra, Australia.

Diagnosis. Annulation moderately broad, intercalations narrow with oval, nodular microplates.
Annuli with two rows of highly ornamented plates. Central annular zone narrow to medium broad.

Shergoldiscolex nodosus sp. nov.

Text-fig. 14g-m

Derivation of name. From nodosus , Latin (full of nodes), referring to the extremely rich ornamentation of
plates.

Holotype. CPC 23051; from 1 km north of Mt Murray, Duchess, Queensland; Triplagnostus gibbus Zone,
Middle Cambrian.

Illustrated material. CPC 23049-23051.

Other material. 4 specimens.

580

PALAEONTOLOGY, VOLUME 36

text-fig. 13. Schistoscolex sp. indet. All specimens from 1 km north of Mt Murray (= D640), late
Templetonian, Triplagnostus gibbus Zone, a-b, CPC 23043 (sample 7338). A, ventrolateral view of terminally
broken specimen; annuli with irregularly-distributed openings; width of annuli increased at these openings,
x 45. b, detail of same specimen, x 1 10. c-G, CPC 23044 (sample 7338). c, ventrolateral view with parallel rows
of nipple-like appendages; nipples are positioned on every third to fourth rib; width of rib increased around
nipple, x 185. d, general view of specimen, x 75. e, detail of ventral side with irregular annulation and insertion
of nipples, x 335. F, surface detail with proximal part of tubule; note circular arrangement of platelets around
this structure, x 650. G, surface detail of dorsal side with irregularly distributed openings, x 120. h-i, CPC
23045 (sample 7336). H, general view of fragmentary specimen, x 70. I, detail of surface with basis of tubule,

x 700.

MULLER AND HINZ-SCH ALLREUTER: CAMBRIAN WORMS 581

text-fig. 14. Shergoldiscolex gen. nov. All specimens from 1km north of Mt Murray (=D640), late
Templetonian, Triplagnostus gibbus Zone, a-f, Shergoldiscolex polygonatus sp. nov. A, c, CPC 23046 (sample
7339). a, general view of straight specimen, x 40. c, view onto oral side with another, underlying, completely
armoured cuticle that could represent a stage shortly before moulting, x 470. B, D, CPC 23047 (sample 7338).
b, surface detail, x 265. d, lateral view of curved specimen; the variable depressions on the outer surface are
of a secondary nature, x 25. e-f, CPC 23048 (sample 7335), holotype. E, lateral view of compressed,
fragmentary specimen, x 85. f, surface detail showing intercalations between the annuli, x 335. G— M,
Shergoldiscolex nodosus sp. nov. G, I, CPC 23049 (sample 7336). G, oblique lateral view of inwardly-folded
specimen, x 45. i, surface detail showing inclination of annuli towards aboral end, x 200. h, k, CPC 23050
(sample 7335). H, surface detail with rich ornamentation, x 700. K, general view of fragmentary specimen; the
inner side reflects the outer ribbing, x35. l-m, CPC 23051 (sample 7336), holotype. L, large, compressed
fragment, x 50. M, detail of surface with intercalations, x 200.

582

PALAEONTOLOGY, VOLUME 36

Diagnosis. Annulation variably broad, with two rows of large, highly ornamented plates. Within a
single annulus, one row has larger plates than the opposite row. Ornamentation of plates with
elevated centre surrounded by smaller tubercles. Triangular microplates between plates in one row,
polygonal platelets in central annular zone.

Description. Fragmentary specimens without oral and aboral ends. Annulation may be somewhat variable in
width. Intercalations narrow with a single row of irregularly oval, nodular microplates.

Contracted annuli with asymmetrical, roof-like relief (i.e. with a triangular cross-section). One face consists
of a row of plates, the other face comprises the plates plus the central annular zone. Annuli ornamented by
two, non-identical rows of plates that are in close contact with the microplates of the intercalations. Outline
of plates circular to irregularly polygonal. Centre elevated, and with four to six nodes surrounded by tubercles
that form a double or triple row at the short axis. In the opposite row the plates may lack a large central
elevation; instead, many small nodes are developed. In general, the plates are not bilaterally symmetrical.
There is a high variability in fine details. Between plates, subtriangular microplates with nodular surface are
intercalated.

Platelets with polygonal outline and smooth margins, highly variable in size. Platelets more elongate between
plates than in the centre of annulus. Surface partly nodular, in particular between the plates of a single row.
Plates : platelets ratio about 2:1.

Shergoldiscolex polygonatus sp. nov.

Text-fig. 14a-f

Derivation of name. From polygonatus , Latin, referring to the outer shape of sclerites.

Holotype. CPC 23048; from 1 km north of Mt Murray, Duchess, Queensland; Triplagnostus gibbus Zone,
Middle Cambrian.

Material. CPC 23046-23048.

Diagnosis. General characters as for the genus. Plates irregularly polygonal with central elevation
of partly fused nodes. Smaller tubercles developed towards annulus borders. Polygonal platelets of
central annular zone fairly large.

Description. Worm fragments with flat annular relief. Intercalations present. Plates within an annulus crudely
rounded to irregularly polygonal and rather variable in size. Centre of plates highly convex with one to several,
partly fused nodes. Outer circle of tubercles mostly incomplete and best developed towards the annulus border.
Platelets forming a mosaic pattern of polygons; size of platelets relatively large, sometimes approaching the
size of a small plate that is distinct only by its ornamentation. Plates: platelets ratio about 2:3.

Remarks. The specimen illustrated in Text-figure 14c is composed of two overlying armoured
cuticles. The upper armour shows traces of abrasion, whereas the armour beneath has plates with
still pointed nodes or cones. The particular specimen is regarded as possible evidence for moulting
of palaeoscolecids.

Genus thoracoscolex gen. nov.

Type species. Thoracoscolex armatus sp. nov.

Derivation of name. From thorax , Greek (armour), referring to the armour of the cuticle.

Diagnosis. Annulation narrow, intercalations small. Annuli with two rows of densely-packed,
circular plates. Ornamentation of plates with two to five central nodes.

MULLER AND HINZ-SCHALLREUTER: CAM BRIAN WORMS

583

text-fig. 15. a-d, Euryscolex paternarius gen. et sp. nov.; 1km north of Mt Murray (=D640), late
Templetonian, Triplagnostus gibbus Zone, a, c, CPC 23052 (sample 7339). a, general view of fragment with
extremely flat relief, x 30. c, detail showing plates with exfoliated, ribbed girdle and concave, radially ribbed
upper sides, x 175. b, d, CPC 23053 (sample 7336), holotype. b, general view of large fragment, x 85. d, surface
detail showing mosaic pattern of platelets, x 315. e-i, Thoracoscolex armatus gen. et sp. nov. e-f, CPC 23054
(sample 7324), holotype; Mt Murray, late Templetonian, Triplagnostus gibbus Zone. E, general view of
deformed specimen, x 65. f, surface detail of tight armour with small, elongate platelets between rows of ribs,
x 1300. g-h, CPC 23055 (sample 6958). G, general view of fragment, x 70. h, detail of surface, x 285. i, CPC
23056 (sample 7339); detail of surface showing ornamentation of platelets, x 365.

584

PALAEONTOLOGY, VOLUME 36

Thoracoscolex armatus sp. nov.

Text-fig. 15e-i

Derivation of name. From arma, Latin (armour).

Holotype. CPC 23054; from 1 km north of Mt Murray, Duchess, Queensland; Triplagnostus gibbus Zone,
Middle Cambrian.

Illustrated material. 23054-23056.

Other material. 2 specimens.

Diagnosis. As for the genus.

Description. Large, fragmentary worms; oral and aboral ends not preserved. Intercalations between annuli
rather narrow and formed by a single row of irregular platelets (Text-fig. 15i). Sometimes intercalations are in
part widened slightly by a second row of even smaller platelets.

Annuli rather narrow and moderately convex. Ornamentation consisting of two rows of irregularly
alternating plates. Plates with subcircular outline and undulating or jagged marginal rims. Surface moderately
convex with two to five nodes; the most frequent pattern displays four nodes. Outline and ornament highly
variable with regard to number and position of nodes. Platelets rather variable from rope-like to irregularly
rounded, the latter with highly convex upper side and partly jagged margins. Size difference to plates
considerable. Plates ; platelets ratio about 9:1.

Palaeoscolecida gen. indet. cf. Hadimopanella apicata Wrona, 1982

Text-fig. 16g-h

cf. 1982 Hadimopanella apicata Wrona, p. 11, pis 1-4.

cf. 1984 Hadimopanella apicata Wrona; Peel and Larsen, p. 93, figs 2-5.

cf. 1987 Hadimopanella apicata Wrona; Hinz, p. 80, pi. 4, fig. 6.

cf. 1988 Hadimopanella apicata Wrona; Bendix-Almgreen and Peel, pp. 85-91, figs 3-7.

Material. Figured specimen, CPC 23061 ; from Rogers Ridge, Queensland; Triplagnostus gibbus Zone, Middle
Cambrian.

Description. Isolated, thick, subcircular plate with pronounced central cone that is steeply emerging. Outer rim
of plate slightly scalloped ; the broad periphery is marked by comparably coarse vertical structures (Text-fig.
16h).

Remarks. This specimen is considered to be related to Hadimopanella apicata based on the very
distinct feature of a strongly developed central cone. In our specimen the steep cone has almost
parallel sides, in contrast to the approximately triangular profile of the taxa referred to in the
synonymy.

Palaeoscolecida gen. indet. sp. A
Text-fig. 16b, d-f

Illustrated material. CPC 23058-23060; Rogers Ridge, Queensland; Triplagnostus gibbus Zone, Middle
Cambrian.

Other material. 147 specimens.

Description. Sclerites of variably oval outline. Upper surface smooth with marginally distributed, relatively
large cones. Cones with outwardly-directed tips and rather flat profiles. Upper surface of cones frequently with

MULLER AND HINZ-SCH ALLREUTER: CAMBRIAN WORMS

585

text-fig. 16 a, c, Pantoioscolex oleschinskii gen. et sp. nov., CPC 23057 (sample 7331), holotype; 1 km north
of Mt Murray (= D640), late Templetonian, Triplagnostus gibbus Zone, a, general view of fragment, x25.
c, detail of surface, x 95. b, d-f, Gen. indet. sp. A, Rogers Ridge, late Templetonian, Triplagnostus gibbus Zone.
b, e, CPC 23058 (sample 7283). b, view from above, x 215. e, oblique lateral view; note enlarged basal portion
with minute cones, x 250. D, CPC 23059 (sample 7282); oblique view from above, x 285. F, CPC 23060 (sample
7283); oblique lateral view; note tiny marginal cones on basal portion, x 315. g-h, Palaeoscolecida gen indet.
cf. Hadimopanella apicata Wrona, CPC 23061 (sample 7283); Rogers Ridge, late Templetonian, Triplagnostus
gibbus Zone. G, oblique lateral view, x450. H, surface detail of marginal rim, x 235.

shallow depression (Text-fig. 16e). Cones continue towards centre of sclerite in faint, alternating ridges (Text-
fig. 16f). Lower surface probably fibrous, but preservation prevents precise statements. Some specimens show
additional, much smaller cones along with the lower marginal rim (Text-fig. 16f). A single sclerite even has a
much expanded lower side with several rows of minute cones interbedded (Text-fig. 16b, e).

586

PALAEONTOLOGY, VOLUME 36

text-fig. 17. Palaeoscolecida gen. et sp. indet. a-b, CPC 23062 (sample 7503); drilling Duchess 18, late
Templetonian, Triplagnostus gibbus Zone, a, longitudinal section showing multilamellar wall structure and
aboral invagination, x45. b, detail of aboral portion, x210. c, CPC 23063 (sample 7340); 1 km north of
Mt Murray (= D640), late Templetonian, Triplagnostus gibbus Zone; longitudinal section; detail of aboral
portion with invagination; note different annular relief of dorsal and ventral sides, x 160. d, CPC 23064
(sample 7331); 1 km north of Mt Murray (= D640), late Templetonian, Triplagnostus gibbus Zone; section
through cuticle parallel to surface, showing plates with broad marginal brim which is usually embedded in

tissue, x 200.

INTERNAL STRUCTURE AND GROWTH

Palaeoscolecida are soft-bodied fossils that have their cuticle reinforced by an armour of sclerotized
plates and platelets. The cuticle itself is stratified (Text-fig. 17a-c), up to seven layers having been
observed. Traces of an underlying secretory epidermis and muscles have not been identified.
However, a variable degree of bodily contraction (i.e. steep to flat annuli) points to muscular
activity. Depending on the degree of contraction, the annulation is visible on both outer and inner
sides of the cuticle, but it is not considered to reflect segmentation or metamerism.

The plates are most probably primarily mineralized; previous detailed investigations by Bengtson
(1977), Wrona (1982, 1987), Dzik (1986) and Bendix-Almgreen and Peel (1988) have revealed a
dense, apparently non-lamellar capping and a fibrous core. The fibrils probably facilitated
attachment to the cuticle. The upper surface of plates visible on a preserved cuticle is much smaller
than their basal diameter. The plates are anchored along a broad rim marked by radial structures.
On isolated specimens of Hadimopanella collaris Marss, 1988 a similar marginal structure can be
observed, as can be seen in a thin section of a specimen parallel to the outer surface (Text-fig. 17d).
Partly exfoliated surfaces of different taxa reveal a matrix for anchoring plates and adjacent
platelets.

The plates achieve their definite ornament only in the uppermost layer. Further inward, the
ornament simplifies and the plates decrease in size until they cannot be distinguished from platelets.
The armoured cuticle is more-or-less equally thick over the entire length of a worm; thickness only
decreases aborally at the paired nipples. From a slit-like aboral opening the cuticle is invaginated

MULLER AND H INZ-SCH ALLREUTER: CAMBRIAN WORMS

587

to form a blind sac (Text-fig. 17b-c). The outer surface of the aboral region appears either smooth
or ornamented with relatively small, more-or-less equally developed platelets, independent of the
general annular ornament of the worm.

Palaeoscolecida are assumed to have grown by moulting. This assumption is strengthened by the
armour of their cuticle which permitted only limited elasticity. Furthermore, within a single taxon,
there are remarkable differences in width with specimens ranging from less than 400 //m to more
than 800 /mi, or from about 700 /an to 1500 /urn. The specimen illustrated in Text-figure 15a-c is
assumed to document the development of a young stage. The young stage has the same ornament
and arrangement of plates as the older stage, but the interspace between the single plates is relatively
smaller. The new cuticle is even a little folded because of the increased diameter. Obviously the
platelets are not fully developed. Length is considered to have increased proportionally by the same
procedure: plates of an annulus are more closely-spaced until the development of platelets sets them
apart.

Due to lack of suitable material, the number of moult stages remains unclear. Furthermore, it is
unknown whether this number varies between individual taxa. It is possible that forms with large
central annular areas would have needed fewer stages than those with closely-spaced plates because
in the former the annuli had little opportunity to expand with each moult.

VARIABILITY

First of all, we have to distinguish between variability within a single taxon and variable characters
on the same individual. Intraspecific variation affects almost all available characters, such as the
widths of the annuli and central annular zone (e.g. Text-fig. 8a, g-k), and size and ornamentation
of plates (e.g. Text-fig. 15b, h).

In the species descriptions the plates: platelets ratio is added as a further aid for the
characterization of a taxon. However, with regard to variability it is not systematically significant,
even if the degree of variation differs distinctly between the various taxa.

More important for the present material is the variability of features on the same animal. This
is shown by: (1) an irregular annulation (Text-fig. 10f); (2) an irregular mode of furcation (Text-
fig. 12c) which results in differentiation of ornament between dorsal and ventral sides (Text-fig.
1 1e); (3) the size difference between plates of a single row (Text-fig. 9b); and (4) a generally variable
surface ornamentation within certain limits. Austroscolex (Text-fig. 5) and Rhomboscolex (Text-fig.
9) are distinguished by their overall appearance, but their individual plates show similarities. The
same applies to isolated plates of Murrayscolex and Shergoldiscolex which are separated on the
basis of presence or absence of intercalations. Eventually, different forms could prove to be only
morphotypes of a single taxon, or the same types of plates could have developed convergently.
These observations indicate that some isolated plates may be taxonomically unassignable.

FUNCTIONAL MORPHOLOGY

The interpretation of isolated plates as dermal sclerites first given by Bengtson (1977) is confirmed
by findings of integuments of Palaeoscolecida.

The armoured cuticle was not as stiff as in other tubular fossils such as serpulids, styliolinids,
tentaculites etc. It consists of individual rows of plates with the interspace between filled with tiny
polygonal platelets. They do not form a complete mineralized cover but rather lie in soft tissue that
permits a certain elasticity. The plates and platelets are completely mineralized components. The
organic matter that previously filled the interspaces between these components is now left as distinct
gaps (Text-fig. 8k).

The worms are not assumed to have had great mobility. In particular, enrolled specimens with
bifurcated ribs at the dorsal side and insertion of rows of platelets point to quite a stable life attitude
with limited ability for dilatation and shrinkage only. Therefore, locomotion by contraction and
dilatation as well as snake-like, lateral movements are unlikely to have occurred. The function of

588

PALAEONTOLOGY, VOLUME 36

the aboral invagination is unclear. It might have enabled the four appendages to perform
movements such as straddling.

Wear of nodes on the upper surface of plates has been previously observed on isolated plates
(Bengtson 1977; van den Boogaard 1983). Functional wear can be distinguished from secondary
abrasion during sedimentation by studying the outer rim of the plates which was originally
embedded in integument and thus sheltered. Plates with eroded margins point to transportation
which has been proved for the high energy environments of the Duchess area.

ECOLOGY

Little is known about the mode of life of Palaeoscolecida. Gedik (1981) assumed that they were
nektonic, whereas Runnegar (1982) presumed a burrowing habit. Our material, which mostly comes
from Mt Murray in the Duchess Embayment, is apparently current-washed so that the original
ecological context has been lost. The material presents such a wide range of shapes and
ornamentations that the animals are assumed to have occupied different ecological niches rather
than lived in one place.

Some of the plates show abrasion which points to living on or in a much coarser sediment than
the calcareous mud in which they were collected. Other taxa (e.g. Schistoscolex) with long, non-
abraded spines probably preferred a different environment. Also the fact that the worms have been
discovered in various attitudes (from slightly curved to planispirally enrolled) supports the
assumption of different environments.

Boogaard (1983, p. 334) demonstrated the wide variation of sample productivity in successive
beds of various stratigraphical sections and concluded that it is not due to diagenetic or later
processes, but the result of varying biological and/or sedimentological conditions.

PHYLETIC RELATIONSHIPS

Palaeoscolecida have been tentatively referred to the phylum Annelida because of their annular
outline that was interpreted as segmentation, and evidence of bodily contraction and relaxation.
However, the preserved annular cuticle seems to us insufficient to indicate a relationship with
Annelida. The main characteristic of this phylum is metamerism, i.e. the repetition of identical
internal structures such as coelomic spaces, nerves, etc. The probable occurrence of chetae
(Whittard 1953) in Palaeoscolecida was not confirmed in subsequent studies (see the discussions of
Whittard’s paper by Thomas (in Whittard 1953) and Conway Morris and Robison (1986)) or in the
well-preserved material studied herein. This clearly demonstrates that the ‘papillae’ are in fact
nodes. In some cases (e.g. Hadimopanella coronata Boogaard, 1989) they may even be developed as
spines. In our opinion, an assignment to annelids is unlikely.

Another group of worms which are known back to the Early Cambrian are the Priapulida. They
are characterized by an eversible proboscis.

The occurrence of tubules, which are irregularly distributed over the whole surface, and the
presence of plates may suggest a relationship with the Aschelminthes at least in a wider sense
(R. M. Kristensen pers. comm.). Similar tubules are, for example, present in all developmental
stages of priapulids (D. Walossek pers. comm.).

COMPARISONS

From Lower Palaeozoic sediments, mainly Cambrian, a number of different genera are known that
have been assigned to Onychophora, Priapulida and Annelida. These forms are briefly discussed
and compared with our palaeoscolecidan material.

The type species of Palaeoscolex, P. piscatorum Whittard, 1953, comprises relatively large
specimens which are usually flattened on shale. The body shows distinct annulation and most of it

MULLER AND HINZ-SCH ALLREUTER: CAMBRIAN WORMS

589

is also decorated with plates, but the precise ornament of the plates is indeterminable. This is due
particularly to the fact that one of the inner layers is usually exposed. In splitting the shales, at least
part of the outer surface adheres to the rock and, as is seen in our material, the ornament becomes
increasingly indistinct towards the inner side. The ornamentation, however, is a diagnostic
character, together with the arrangement of plates and the development of platelets. Therefore,
none of the new Australian taxa could be assigned to Palaeoscolex. Palaeoscolex piscatorum
displays a structure similar to an inner layer of Corallioscolex.

Protoscolex Ulrich, 1878, from the Upper Ordovician, probably belongs to the Palaeoscolecida.
As the form is not represented in our material, we have not studied it in detail.

Cricocosmia Hou and Sun, 1988, from the Lower Cambrian of South China, is an annulated
worm with one to two pairs of highly convex plates on each annulus. This is different to the
ornamentation of Palaeoscolecida and also different to Priapulida which have a narrowly annulated
but unornamented middle part. Further, a proboscis is not recognizable on Cricocosmia and its
reference to Priapulida is doubtful.

Maotianshania Sun and Hou, 1987, from the Lower Cambrian of Yunnan Province is likely to
belong to the priapulids. Its blunt aboral end bears a spine, and orally it terminates in a spiny
proboscis. The specimens are compressed and preserved on the bedding surfaces of shale and mud.

VALUE FOR STRATIGRAPHY

Palaeoscolecidan remains obviously are much more widespread and common than is evident from
the literature. This is shown by the fact that hadimopanellids since their first description some 14
years ago have been reported from many localities by various authors. They may form a large
component of the sediment. Whittard (1953) recorded up to 30,000 plates on a single individual.
Their assumed benthic mode of life suggests restriction to certain facies.

Marss (1988) believed that the established species of Hadimopanella may be of some help for
stratigraphical purposes, although Palaeoscolecida, as a Class, are long-ranging. Gedik (1989)
described the stratigraphical distributions of Hadimopanella species in the western Taurids, Turkey,
and suggested a biostratigraphical zonation within the Lower, Middle and Upper Cambrian.
However, his diagram shows considerable ambiguity about the lower and upper ranges of the
individual species. Furthermore, as the limestones and dolomites of his sections are quite variable
in lithology, the occurrence and disappearance of Palaeoscolecida may be facies controlled.

The stratigraphical application of isolated plates is hampered because the same types of plates
may occur in different genera. What would have been described as different species if not genera,
may turn out to be mere morphotypes.

It is expected that the various taxa will have different potentials as index fossils; some may be
long-ranging, while others may be restricted to short-time intervals. This can be determined only
empirically, and much more information is needed before Palaeoscolecida can be confidentially
applied in correlation. Even 'species’ may not be restricted to zones. Common form-species have
already turned out to be long-ranging. Hadimopanella oezguli Gedik, 1977, was originally described
from the Upper Cambrian of the Middle Taurus, Turkey but is also present in the Middle Cambrian
of Australia (herein) and South Kirgizia (Marss 1988), as well as in Cambro-Ordovician boundary
beds in Estonia (Hadimopanella collaris Marss, 1988, in part). Again, Milaculum ruttneri Muller was
originally described from the Upper Cambrian of Iran and has been subsequently described as
Plasmuscolex nero and Plasmuscolex klabavensis Kraft and Mergl, 1989 from the Arenig of
Bohemia.

Until the present study, only flattened specimens had been reported from shales. In this type of
preservation the plates and platelets of the surface commonly are not well-preserved. Accordingly,
critical details in ornamentation could not be observed. The fair variety in morphology of these
plates permits the possibility of investigating their potential as index fossils. But the standard has
to be established at a section where the specimens are not current-washed, unlike our material.

590

PALAEONTOLOGY, VOLUME 36

Acknowledgements . The material was collected during a joint project between the Bureau of Mineral Resources,
Canberra and the University of Bonn in order to investigate the microfauna of the Middle Cambrian Georgina
Basin phosphorites. We would like to thank John Shergold (Canberra) for logistics and permission to
reproduce Text-figures 1 and 2, R. Below (Utrecht) and D. Walossek (Bonn) for sampling, and Mrs A.
Gossmann, Mrs D. Kranz and Mr G. Oleschinski for valuable technical help. Critical reading by S. Conway
Morris (Cambridge) is appreciated. The project has been kindly supported by the Deutsche Forschungs-
gemeinschaft.

REFERENCES

bather, F. a. 1920. Protoscolex latus, a new “worm” from Ludlow Beds. Annals and Magazine of Natural
History , (9), 5, 124-132.

bendix-almgreen, s. e. and peel, J. s. 1988. Hadimopanella from the Lower Cambrian of north Greenland:
structure and affinities. Bulletin of the Geological Society of Denmark, 37, 83-103.
bengtson, s. 1977. Early Cambrian button-shaped phosphatic microfossils from the Siberian Platform.
Palaeontology , 20, 751-762.

— and conway morris, s. 1984. A comparative study of Lower Cambrian Halkieria and Middle Cambrian
Wiwaxia. Lethaia , 17, 307-329.

boogaare), M. van den 1983. The occurrence of Hadimopanella oezgueli Gedik in the Lancara Formation in
NW Spain. Proceedings of the Koninklijke Akademie van Wetenschappen, Series B, 86, 331-341.

1988. Some data on Milaculum Muller, 1973. Scripta Geologica , 88, 1-25.

— 1989a. A problematic microfossil, Hadimopanella ? coronata sp. nov., from the Ordovician of Estonia.
Rijksmuseum van Geologie en Miner alogie. Series B, 92, 179-190.

1989A Isolated tubercles of some Palaeoscolecida. Scripta Geologica, 90, 1-12.

chen jun-yuan, hou xian-guang and lu hao-zhi 1989. [Early Cambrian netted scale-bearing worm-like sea
animal]. Acta Palaeontologica Sinica, 28, 1-16, pis 1-4. [In Chinese with English summary].

— , bergstrom, j., undstrom, m. and hou xian-guang 1991. Fossilized soft-bodied fauna. National
Geographic Research and Exploration 7, 8-19.

conway morris, s., pickerill, r. k. and harland, t. l. 1982. A possible annelid from the Trenton Limestone
(Ordovician) of Quebec, with a review of fossil oligochaetes and other annulate worms. Canadian Journal
of Earth Sciences , 19, 2150-2157.

and robison, r. a. 1986. Middle Cambrian priapulids and other soft-bodied fossils from Utah and Spain.

University of Kansas Paleontological Contributions, 9, (Paper 117), 1-22.

COOK, p. J. and shergold, J. H. 1980. Proterozoic-Cambrian phosphorites. Project 156. Canberra Publishing and
Printing Co., Canberra, xi+ 106 pp.

dzik, J. 1986. Chordate affinity of the conodonts. 240-254. In hoffmann, a. and nitecki, m. h. (eds).

Problematic fossil taxa. Oxford Monographs on Geology and Geophysics, 5, 267 pp.
ethington, R. and Clark, D. L. 1965. Lower Ordovician conodonts and other microfossils from the Columbia
Ice Fields Section, Alberta, Canada. Brigham Young University Geological Studies, 12, 185-205.
gazdzicki, a. and wrona, r. 1986. Polskie Badania Paleontologiczne W Antarkyce Zachodnij. Przeglad
Geologiczny, 11, 609-617.

gedik, i. 1977. Conodont stratigraphy in the Middle Taurus. Bulletin of the Geological Society of Turkey , 20,
35-18.

— 1981. Hadimopanella Gedik, 1977 nin stratigrafik dagilimi ve mikroyapisi konusunda bazi gozlemler.
Karadeniz Teknik Universitesi Yer Bilimleri Dergisi, Jeoloji, 1, 159-163.

— 1989. Hadimopanellid biostratigraphy in the Cambrian of the Western Taurids: A new biostratigraphic
tool in the subdivision of Cambrian System. Geological Bulletin of Turkey, 32, 65-77 .

glaessner, m. f. 1979. Lower Cambrian Crustacea and annelid worms from Kangaroo Island, South
Australia. Alcheringa, 3, 21-31.

hinz, i. 1987. The Lower Cambrian microfauna of Comley and Rushton, Shropshire/England. Palaeonto-
graphica, Abteilung A, 198, 41-100.

— kraft, p., mergl, m. and muller, k. j. 1990. The problematic Hadimopanella , Kaimenella, Milaculum
and Utahphospha identified as sclerites of Palaeoscolecida. Lethaia , 23, 217-221.

hou xian-guang and sun wei-guo 1988. [Discovery of Chengjiang fauna at Meishucun, Jinning, Yunnan].
Acta Palaeontologica Sinica, 27, 1-12, pis 1-3. [In Chinese with English summary],

— and chen jun-yuan 1989. [Early arthropod-annelid intermediate sea animal, Luolishania gen. nov. from
Chengjiang, Yunnan], Acta Palaeontologica Sinica, 28, 1-15, pis 1-4. [In Chinese with English summary].

howell, b. f. 1962. Worms. 144-177. In moore, r. c. (ed.). Treatise on invertebrate paleontology. Part W.

MULLER AND HINZ-SCH ALLREUTER: CAMBRIAN WORMS

591

Miscellanea. Geological Society of America and University of Kansas Press, Lawrence. Kansas and Boulder,
Colorado. 259 pp.

koneva, s. p., popov, l. E., uschatinskaja, G. T. and esakova, n. w. 1990. Biostratigrafija i Paleontologija
Kembrija Severnoi Asii. Novosibirsk ‘ NAUKA Sibirskoe Otdelenie, 1-222. [In Russian],

Kraft, p. and mergl, m. 1989. Worm-like fossils (Palaeoscolecida ; ?Chaetognata) from the Lower Ordovician
of Bohemia. Sbornik Geologickych Ved Paleontologie, 30, 9-36.
kusta, j. 1888. Beitrag zur Kenntnis der Steinkohlenfauna bei Rakonitz. Bohmische Gesellschaft der
Wissenschaften. Mathematisch-Naturwissenschaftliche Classe, Sizungsbericht , Jahrgang 1887 , 561-564,

1 Tafel.

marss, T. 1988. Early Palaeozoic hadimopanellids of Estonia and Kirgizia (USSR). Proceedings of the Academy
of Sciences of the Estonian SSR, Geology , 37, 10-17.

miller, s. a. and faber, c. 1892. Some new species of new structural parts of fossils. Journal of the Cincinnati
Society of Natural History , 15, 75-100.

missarzhevsky, v. v. and mambetov, a. m. 1981. Stratigrafiya i fauna pogranichnykh sloev kembriya
i dokembriya Malogo Karatau. Trudy Geologii Instituti Leningrad , 326, 1-92. [In Russian].
muller, K. J. 1973. Milaculum n. g. ein phosphatisches Mikrofossil aus dem Altpalaozoikum. Paldontologische
Ze it sc hr if. 47, 217-228.

— and hinz, i. 1992. Cambrogeorginidae fam. nov., soft-integumented Problematica from the Middle
Cambrian of Australia. Alcheringa , 16, 333-353.

— and miller, j. f. 1976. The problematic microfossil Utahphospha from the Upper Cambrian of the western
United States. Letliaia, 9, 391-395.

oeschger, r. and janssen, H. h. 1991. Histological studies on Halicryptus spinulosus (Priapulida) with regard
to environmental hydrogen sulfide resistance. Hydrobiologia 222, 1-12.
peel, J. s. and larsen, N. h. 1984. Hadimopanella apicata from the Lower Cambrian of western North
Greenland. Gronlands Geologiska Under sokning Bulletin, 121, 89-96.
repetski, J. E. 1981. An Ordovician occurrence of Utahphospha Muller and Miller. Journal of Paleontology,
55, 395-400.

robison, R. A. 1969. Annelids from the Middle Cambrian Spence Shale of Utah. Journal of Paleontology, 43,
1169-1173.

Rogers, J. K. and crase, n. j. 1980. The Phosphate Hill rock phosphate deposit, northwest Queensland,
Australia - an outline of geological development. 307-328. In sheldon, r. p. and burnett, w. c. (eds).
Fertilizer mineral potential in Asia and the Pacific. Proceedings of the Fertilizer Raw Materials Resources
Workshop, August 20-24, 1979, East-West Resources Systems Institute.
ruedemann, r. 1925a. The Utica and Lorraine Formations of New York, Part 2, Systematic Paleontology.
New York State Museum Bulletin , 262, 5-140.

— 1925A Some Silurian (Ontarian) faunas of New York. New York State Museum Bulletin 265, 5-83.
runnegar, b. 1982. Oxygen requirements, biology and phylogenetic significance of the late Precambrian worm

Dickinsonia, and the evolution of the burrowing habit. Alcheringa, 6, 223-238.
shergold, j. h. and Southgate, p. n. 1986. Middle Cambrian phosphatic and calcareous lithofacies along the
eastern margin of the Georgina Basin, western Queensland. Australasian Sedimentologists Group, Field Guide
Series, 2, 89 pp.

Southgate, p. n. and shergold, j. H. 1991. Application of sequence stratigraphic concepts to Middle
Cambrian phosphogenesis, Georgina Basin, Australia. Bureau of Mineral Resources Journal of Australian
Geology and Geophysics, 12, 119-144.

sun wei-guo and hou xian-guang 1987. [Early Cambrian worms from Chengjiang, Yunnan,
China: Maotianshania gen. nov.]. Acta Palaeontologica Sinica , 26, 303-305. [In Chinese with English
summary].

ulrich, e. o. 1878. Observations on fossil annelids and descriptions of some new forms. Journal of the
Cincinnati Society of Natural History, 1, 87-91.

walcott, c. D. 1911. Cambrian geology and paleontology II; Middle Cambrian annelids. Smithsonian
Miscellaneous Collections, 57, 109-144, pis 18-23.

wang chen-yuan 1990. [Some Llandovery phosphatic microfossils from south China], Acta Palaeontologica
Sinica, 9, 548-556, pis 1-2. [In Chinese with English summary],
whittard, w. F. 1953. Palaeoscolex piscatorum gen. et sp. nov., a worm from the Tremadocian of Shropshire.

Quarterly Journal of the Geological Society of London, 109, 125-135.
wrona, r. 1982. Early Cambrian phosphatic microfossils from southern Spitsbergen (Hornsund Region).
Palaeontologia Polonica, 43, 9-16.

592 PALAEONTOLOGY, VOLUME 36

wrona, R. 1987. Cambrian microfossil Hadimopanella from glacial erratics in west Antarctica. Palaeontologia
Polonica , 49, 37-48.

— 1989. Cambrian limestone erratics in the Tertiary glacio-marine sediments of King George Island, west
Antarctica. Polish Polar Research, 10, 533-553.

K. J. MULLER
I. HINZ-SCHALLREUTER

Institut fiir Palaontologie
Universitat Bonn
Nussallee 8

5300 Bonn 1, Germany

Present address of i. hinz-schallreuter

Geologisch-Palaontologisches Institut und Museum
Universitat Hamburg

Typescript received 24 February 1992 Bundesstrasse 55

Revised typescript received 30 July 1992 2000 Hamburg 13, Germany

EDIACARAN-LIKE FOSSILS IN CAMBRIAN
BURGESS SHALE-TYPE FAUNAS OF NORTH

AMERICA

by SIMON CONWAY MORRIS

Abstract. A number of fossils from the Stephen Formation (Burgess Shale, Middle Cambrian, British
Columbia) and Parker Slate (Lower Cambrian, Vermont) resemble Ediacaran taxa. Thaumaptilon walcotti
gen. et sp. nov., known from three specimens, consists of a broad frond bearing a central rachis from which
arise branches bearing possible zooids. The holdfast is relatively elongate and slightly swollen. T. walcotti
approaches closely a number of Ediacaran frond-like fossils, especially Charniodiscus , Vaizitsinia and
Khatyspytia. All these taxa appear to be pennatulacean anthozoans (Cnidaria). Mackenzia costalis Walcott,
known from about seventy specimens, is an elongate bag-like organism. The exterior was probably thrown into
longitudinal folds, and the interior may have borne septa. Attachment to the sea-bed is indicated by holdfasts
of eocrinoid stems and brachiopod/sponge spicule aggregations. M. costalis may be compared to several
Ediacaran taxa including Inaria, Protechiurus and possibly Platypholinia. The affinities of M. costalis are
uncertain, but a place within the actinarian anthozoans seems possible and this animal probably had a
cnidarian-grade of organization. Emmonsaspis cambrensis (Walcott), long interpreted as a possible chordate,
is shown to be a frond-like fossil with angled branches arising from the mid-line. No trace of a holdfast exists
in any of the three specimens. E. cambrensis resembles a number of Ediacaran frond-like fossils, but similarities
to taxa such as Pteridinium may be superficial. The two remaining taxa are known from only single specimens.
Gelenoptron tantaculatum gen. et sp. nov., originally described by Walcott as Redoubtia polypodia (pars), is
tentatively interpreted as a chondrophorine, with evidence for a float and tentacular margin. A unique
specimen consisting of a disc with annuli and tentacles is regarded as another type of chondrophorine. It occurs
in association with M. costalis. The description of these animals as hold-overs from the Ediacaran assemblages
casts some doubt on the general validity of Seilacher’s concept of Ediacaran taxa representing a distinctive
body-plan, known as the Vendobionta, separate from the metazoans.

The identity and coherence of the Vendian (latest Proterozoic) soft-bodied Ediacaran faunas is now
well-established (e.g. Glaessner 1984; Conway Morris 1985, 1990; Fedonkin 1987). With one
possible exception (Hofmann et al. 1990), where tillites are present in the same sections (and they
often are), the Ediacaran faunas lie above these glacial deposits. The age of the Ediacaran faunas
remains somewhat uncertain. Glaessner (1984) proposed an approximate range of 550/570—
650/660 Myr, although the apparently reliable date of 565 + 3 Myr obtained from zircons in an ash-
fall that smothered an Ediacaran fauna in southeast Newfoundland (Benus 1988; see also Jenkins
1989) suggests that as a whole Ediacaran faunas may fall towards the younger end of Glaessner’s
(1984) spectrum of ages. Such dates are also consistent with new evidence indicating that the
Precambrian-Cambrian boundary is unlikely to be older than about 540 Myr (e.g. Compston et al.
1992).

Ediacaran assemblages consistently underlie strata with Cambrian shelly fossils and a diversity
of trace fossils that includes Phycodes pedum (see Narbonne and Myrow 1988). In many areas of
the world the Vendian-Cambrian sections have unconformities and/or distinctive facies (e.g. fluvial
or peritidal) that separate the last appearances of Ediacaran taxa and the first appearances of shelly
taxa and/or Cambrian-type trace fossils. In a few sections, however, there appears to be a greater
degree of continuity in rock record and facies suitable for fossil preservation. Most notable,
perhaps, is a section in the Wernecke Mountains of northwest Canada where an interval above the

(Palaeontology, Vol. 36, Part 3, 1993, pp. 593-635, 8 pis.]

© The Palaeontological Association

594

PALAEONTOLOGY, VOLUME 36

last appearance of Ediacaran taxa is devoid of all except millimetric trace fossils despite what appear
to be facies that were appropriate for Ediacaran style preservation (see Narbonne and Hofmann
1987). What may represent a similar state of affairs was documented by Sokolov and Fedonkin
(1984, 1985) in the clastic sequences of the east European platform and the carbonate sections of
northern Siberia, where the Ediacaran assemblages are separated from the onset of the Cambrian
faunas by a poorly fossiliferous interval. It is clear also that with certain key exceptions, many of
which are discussed below, Ediacaran components are effectively absent from Cambrian
assemblages.

There are broadly two explanations for the disappearance of Ediacaran faunas. The first proposes
closure of a taphonomic window, thereby precluding the widespread preservation of soft-parts that
typifies Ediacaran life. Most often invoked in this closure are the agents of predators, scavengers
and bioturbators, whose widespread appearance was a harbinger for the Cambrian explosion.
While there is little doubt that certain taphonomic factors changed between the Vendian and
Cambrian, especially the onset of deeper and more extensive bioturbation, this alone seems unable
to explain the absence of Ediacaran faunas in very latest Vendian sections such as those in the
Wernecke Mountains and possibly elsewhere. An alternative explanation for the disappearance of
Ediacaran faunas is some type of mass extinction, albeit one at present unspecified in terms of
intensity, duration or cause (Seilacher 1984; Brasier 1989; Conway Morris 1989a; but see Jenkins
1989). There seems to be no compelling evidence for extra-terrestrial causes such as bolide impact,
but some parallels might be drawn with the end Permian debacle (see Schopf 1974, 1979; Maxwell
1989). Such factors might include withdrawal of shelf seas as part of a major regression, changes
in oceanic salinity, or arguably most significant of all, a major drop in levels of atmospheric oxygen
(Wignall and Hallam 1992).

A relatively neglected aspect of mass extinctions is the evolutionary and ecological behaviour of
post-catastrophe taxa, be they ‘Lazarus’ taxa that enter unspecified refuges during time of stress
or newly diversifying clades arising from surviving taxa. Here I document what appear to be hold-
overs from the once-flourishing Ediacaran assemblages, present as rare specimens in soft-bodied
faunas of Burgess Shale-type.

SUPPOSED EDIACARAN SURVIVORS: PREVIOUS REPORTS

If some Ediacaran taxa belong to the same clades as survive to the present day and are represented
by Recent cnidarians or annelids, then in one sense their identification as post-Ediacaran survivors
is trivial. The term ‘ Ediacaran-type survivors’, however, is taken here to presuppose a relatively
close phylogenetic relationship, perhaps at the taxonomic level of family. Moreover, while the
assignments of some Ediacaran taxa to groups such as the Cnidaria now appear to be reasonably
secure, in general comparisons are easier between Ediacaran and Cambrian forms than they are
with extant representatives. Ideally, these relationships would be established on the basis of cladistic
analysis, but these are in their infancy in terms of rigorous analysis (but see Jenkins 1985 for a
preliminary attempt).

To date there are several Cambrian fossils that have been claimed to be closely related to
Ediacaran taxa (see also Runnegar and Fedonkin 1992). Borovikov (1976) described a putative
specimen of the flat segmented worm Dickinsonia (see Wade 1972a; Runnegar 1982), from the
Lower Cambrian (Shabakty suite) of Maly Karatau, Kazakhstan. Its age is not in doubt because
the specimen is associated with trilobites and brachiopods. Glaessner (1984, p. 144) remarked that
the specimen might be correctly assigned, but ‘some essential diagnostic characters are obscured by
its preservation’. Rozanov and Zhuravlev (1992, fig. 24) re-illustrated this specimen and suggested
it be better considered as a trace fossil.

Durham (1971a; see also Firby and Durham 1974, text-fig. 2 for the only other mention in the
literature) commented in an abstract that ‘ poor specimens of Dickinsonia (?) ’ had been recovered
from the Lower Cambrian Poleta Formation in the White Inyo Mountains, California. A single
specimen (Text-fig. 1) comes from a locality on the west edge of Cedar Flat, White Inyo Mountains,

CONWAY MORRIS: CAMBRIAN EDI ACARAN-LIKE FOSSILS

595

text-fig. 1. Specimen of supposed Dickinsonia (UCMP 37450) from the Poleta Formation (Lower Cambrian),

Cedar Flat, California, x F5.

eastern Califormia (small knoll, centre of NWf of SEj, section 5). The outcrop also yields trilobites
and the echinoderm Helicoplacus. The specimen, which occurs in a pale-green sandstone, is
incomplete, but possesses prominent transverse markings. An attribution to Dickinsonia is possible,
but the apparent convergence of the "segments’ on the upper side of the specimen is not a feature
of Dickinsonia. An alternative suggestion to consider is that, as with the Borovikov (1976)
? Dickinsonia, this specimen represents a trace fossil, perhaps arthropod scratch marks such as
Monomorphichnus. Typically in this ichnogenus the scratches are not as densely arrayed, form well-
defined groups, and may taper in one direction (e.g. Crimes et at. 1977 ; Fritz and Crimes 1985; Peel
1990). Closely-spaced scratches in M. multilineatus were described by Alpert (1976, p. 234, pi. 1, figs
1-2) from the Lower Cambrian Harkless and Campito Formations (these straddle the Poleta
Formation) in the White Inyo Mountains, California. In Alpert’s material, however, the size of the
scratches in an array varies, whereas those of the supposed Dickinsonia are more even. Nevertheless,
a trace fossil origin is thought to be likely for this specimen. Johnson and Fox (1968) described
rosette-like structures from the Silurian of Pennsylvania as being related to Dickinsonia , but Cloud’s
(1973) view that these structures are pseudofossils is accepted.

Skania fragilis was described by Walcott (1931) from the Burgess Shale. The specimens are
typically less than 10 mm in size, and although Walcott (1931, p. 26) referred to a suite of 29

596

PALAEONTOLOGY, VOLUME 36

specimens, most of these are unidentifiable. Walcott’s (1931) account was placed in the framework
of an arthropodan anatomy, and this was accepted by some subsequent workers (e.g. Stormer 1944,
pp. 34-35). Cave and Simonetta (1975), however, rejected this interpretation, and were unable to
recognize either segmentation or appendages. They drew attention to an anchor-like structure,
defined by what appears to be slight thickening on the leading edges of the body that links to a
median strand. They interpreted these structures as intestinal caeca. As Cave and Simonetta (1975)
stressed, this arrangement is similar to the Ediacaran taxon Parvancorina minchami. This taxon
occurs in South Australia (Glaessner 1979n, 1980) and the White Sea area of Russia (Fedonkin
1985, 1987), although it appears not to have been illustrated from the latter region. Glaessner (1980)
described two principal sets of elongate structures that he interpreted as appendages, this being one
line of support for his tentative assignation of Parvancorina to the arthropods. Whether these
structures genuinely represent appendages is moot. Glaessner (1980) was evidently sceptical about
the relationships between Skania and Parvancorina , and Hou et al. (1991, fig. 5) described juvenile
specimens of Naraoia, from the Lower Cambrian Chengjiang fauna of south China, which are
similar to Skania. These authors noted the similarity to Parvancorina , but this they regarded as
superficial. Their arguments, however, are largely based on the supposition that Parvancorina is a
typical vendozoan or vendobiontan. In contrast, Gehling (1991) reiterated the case for a
phylogenetic connexion between Skania and Parvancorina.

Given the diversity of medusiform elements in Ediacaran faunas, the paucity of Cambrian
examples is noteworthy. Some confusion has arisen with supposed medusoids that transpire to be
trace fossils (e.g. Meer Mohr 1969), while this interpretation may apply also to such examples as
IBrooksella from the Middle Cambrian Spence Tongue of the Lead Bell Shale, Utah (Willoughby
and Robison 1979; see also Milashev 1958). In the same paper these authors identified another
putative medusoid {Cambromedusa fur cula) from the somewhat younger Wheeler Shale, and argued
that it was ‘most similar to the late Precambrian genus Cyclomedusa' (Willoughby and Robison
1979, p. 498). This interpretation is also disputed and, while the affinities of Cambromedusa remain
problematical, an affinity to the sponges seems conceivable with the supposed thin radial canals
representing spicules. Pickerill (1982) reported abundant specimens of medusiform fossils from the
Upper Cambrian (Agnostus Cove Formation) of New Brunswick, but here too the author was
unable to draw any convincing similarities to Ediacaran forms. In the case of the Nemakit-Daldyn
horizon of north Siberia, which has received wide attention on account of its early skeletal faunas
that appear at the Precambrian-Cambrian boundary, Khomentovsky (1986, p. 340) noted the
presence of medusoids that survived from the Vendian, but no details of these fossils appear to have
been published.

Another characteristic component of the Ediacaran assemblages is annulated discs of
Ovatoscutum, which have been widely interpreted as the remains of the chambered floats of
chondrophorine hydrozoans (e.g. Glaessner and Wade 1966; Fedonkin 1984, 1985; Gehling 1991),
although proponents of the vendobiontan hypothesis (Seilacher 1989, 1992) have reinterpreted
these fossils in the light of this new model.

The subsequent fossil record of Phanerozoic chondrophorines has been reviewed by Stanley
(1986; but see Conway Morris et al. 1991). None of these purported examples appears to be
significantly similar to Ovatoscutum (but see Runnegar and Fedonkin 1992, p. 372). In the context
of Ediacaran faunas, however, attention is drawn to Rotadiscus grandis from the Chengjiang fauna
of China (Sun and Hou 1987) and a series of related Palaeozoic fossils. The relevance of the
discussion here will be apparent from a review of Ediacaran biotas by Runnegar and Fedonkin
(1992, p. 372), where they write of ‘The previously unemphasized but possibly significant similarities

among various kinds of Vendian and Cambrian fossils such as Eldonia Walcott, Eomedusa

Popov, Rotadiscus Sun & Hou Stellostomites Sun & Hou, Velumbrella Stasinska,

Yunnanomedusa Sun & Hou’. The circular fossils of Rotadiscus have been interpreted as
chondrophorines, with the annulated disc interpreted as the float. During my re-examination of
specimens in the Nanjing collections it became clear that the chondrophorine interpretation is less
likely. Important information is available in undescribed specimens presently being studied by Chen

CONWAY MORRIS: CAMBRIAN EDI ACARAN-LIKE FOSSILS

597

Junyuan and Sun Weiguo, so I restrict my comments to specimens that have already been illustrated
(Sun and Hou 1987, pi. 3, figs In, 2 a-b). The cardinal observation is that the specimens are
bilayered, and consist of not only the annulated disc but a separate discoidal unit that occupies the
central region of the opposite side. This unit was misidentified by Sun and Hou (1987, pi. 3, fig. 1)
as a superimposed specimen of the medusiform taxon Stellostomites eumorphus which is either very
closely related or synonymous with the Middle Cambrian genus Eldonici (see Conway Morris and
Robison 1988). This discoidal unit consists of a series of radiating plates, apparently separated by
zones of softer integument, which along either side bear a distinctive series of pustules (just visible
in pi. 3, fig. 1 of Sun and Hou 1987). The proposal that a trifid structure in the centre of the
annulated disc represents the mouth is considered implausible, and unpublished specimens show
structures that were probably involved with feeding. The trifid structures may represent a split that
formed as the tough material of the disc was compacted. Adjacent to this trifid structure are two
small conical structures. These were interpreted as ‘ribbon-like appendages’, possibly gonozooids
by Sun and Hou (1987), but here are interpreted as probably epizoic tubes, comparable to
Cambrorhytium (see Conway Morris and Robison 1988; Jin et al. 1991).

Comparisons can be drawn between Rotadiscus and other Palaeozoic taxa, notably the Middle
Cambrian Velumbrella and the paropsonemids. Velumbrella was described by Stasinska (1960; see
Bednarczyk 1970 for revision of its age to Middle Cambrian) from the Holy Cross Mountains of
Poland. The large discs, which have diameters of up to 80 mm, occur in coarse sandstones. Stasinska
(1960; see also Brasier 1979; Scrutton 1979) considered the fossils to represent medusoids, but Dzik
(1991) considered this unlikely. Although unable to propose a systematic position for Velumbrella ,
Dzik noted (p. 50) that ‘they definitely were not scyphozoans’. Dzik (1991) also suggested that
Velumbrella was most closely comparable to the Chengjiang genera Yunnanomedusa and
Stellostomites , while ‘At the end of a morphocline of these Chinese discoidal fossils can be placed
Rotadiscus' (p. 50). Yunnanomedusa and Stellostomites are closely related to Eldonia , if not
synonymous. A relationship between these taxa and Rotadiscus is a possibility (see below), but it
seems likely that Velumbrella and Rotadiscus are even more closely related, and possibly
synonymous. In this context, I propose that in Velumbrella the discs with prominent radial grooves
are homologous with the discoidal unit composed of radiating plates in Rotadiscus. The fragments
of a discoidal structure, that have been referred to informally as ‘Brzechowia’ (see Dzik 1991, fig.
3a) are not a related species as Dzik (1991) suggested, but are regarded here as an integral part of
Velumbrella and correspond to the discoidal unit with concentric markings of Rotadiscus.
Supporting evidence for this proposal comes from the holotype (Stasinska 1960, pi. 1, figs 1-2)
where a portion of the concentric disc is in immediate association with the plated discoidal unit.

Specimens of Rotadiscus from Chengjiang may occur in close proximity (Sun and Hou 1987, pi.
3, fig. 2), and the overlap of discs of Velumbrella (Stasinska 1960, pi. 2, figs 1-2; see also Dzik 1991,
fig. 3a) also suggests a gregarious habit. In the case of another specimen illustrated by Dzik (1991,
fig. 3b), however, the superimposed discs are concordant and are interpreted here as being like the
holotype, i.e. equivalent to the two discoidal units that occur in Rotadiscus. In his comparison of
Upper Cambrian medusoids from New Brunswick, Pickerill (1982) drew comparisons to
Velumbrella , but considered them to be slight. This may have been premature, and while these
Canadian medusiforms show a number of obvious differences with Velumbrella , it remains possible
that the supposed radial canals (Pickerill 1982, fig. 10.4) are actually comparable to the plate-like
structures of the rotadiscids.

Dzik (1991) also commented on the possible relationships of these discoidal fossils to
Paropsonema , while Rozanov and Zhuravlev (1992, fig. 25) drew attention to similarities between
Velumbrella and paropsonemids (in the form of Eomedusa (Popov 1968)). This medusiform animal
was described from the Devonian of New York (Clarke 1900; Ruedemann 1916), and has
subsequently been noted from the Silurian of South Australia (Chapman 1926; Harrington and
Moore 1956) and the Cambrian of Siberia (Popov 1967, 1968; see also Conway Morris and Robison
1982; Dzik 1991; Rozanov and Zhuravlev 1992).

The Devonian paropsonemids were described briefly by Clarke (1900, see also Ruedeman 1916,

598

PALAEONTOLOGY, VOLUME 36

1934). The fossils require rescrutiny, and here only a few additional remarks are made. As noted by
Clarke (1900) the prominent radiating structures form two sets of intercalated rays that together
extend across much of the umbrella. On occasion the two series are divided by a series of prominent
nodules (NYSM 6818), but these are not invariably present. In addition, the outer series may
terminate by a major bifurcation, each branch of which subdivides into two or three narrow,
tapering branches. The rays may show lobate margins, and the transverse rows of pores often
converge as adjacent pairs so as to impart a chevron-like appearance along each ray.

Clarke (1900) proposed that the paropsonemids were echinoderms, interpreting the rays as
ambulacra. He noted the absence of calcareous spicules, let alone a skeleton, and commented that
assignment to the echinoderms had not received wide support in pre-publication discussion.
Thereafter, the emphasis changed with workers assigning Paropsonema to the chondrophores
(Fuchs 1905; Ruedemann 1916, 1934; Harrington and Moore 1956; Scrutton 1979; Stanley 1986).
Chapman (1926) was evidently unaware of the earlier work on paropsonemids and referred his
material (as Discophyllum mirabile ) to the scyphozoans, as did Popov (1967, 1968; as Eomedusa
datsenkoi). Rozanov and Zhuravlev (1992, p. 258) retained this fossil (and Velumbrella) in the
Cnidaria, but reverted to the notion of their being chondrophorines.

Pending a thorough redescription of the paropsonemids (D. Friend, pers. comm.), only
preliminary remarks are necessary. Clarke’s (1900) original proposal is regarded as broadly correct,
and their affinities are believed to lie close to, if not within the Echinodermata. In addition, there
seems reason to ally the paropsonemids with Rotadiscus and Velumbrella. This group, in turn, may
be compared with Eldonia. In any event neither Paropsonema nor Rotadiscus are accepted as
chondrophorines.

The status of Discophyllum in this context remains uncertain. These discs were originally
described by Hall (1847, pi. 75, fig. 3) from the Middle Ordovician of New York (see also Walcott
1898, pi. 47, figs 1-2). Clarke (1900, p. 178) drew attention to similarities between Discophyllum and
Paropsonema , but also noted that they were not ‘identical in all structural features’. Ruedemann
(1916, p. 26) restressed the similarities and considered ‘that the two are closely related organisms’,
while Chapman (1926) placed the Australian paropsonemid in Discophyllum , unaware of Clarke’s
(1900) work. Evidence, however, for a close relationship between Discophyllum and Paropsonema
remains wanting, and the arrangement of radial ridges and concentric markings in the former genus
that Ruedemann (1916; see also Ruedemann 1934) emphasized as points of significant similarity are
believed to be of superficial significance. Harrington and Moore (1956) reaffirmed Ruedemann’s
opinion that Discophyllum is the float of a chondrophorine.

Another characteristic component of Ediacaran faunas, that of the frond-like animals, until now
seems to have been effectively unrepresented in the Cambrian. Tarlo (1967) proposed that the
Lower Cambrian animal Xenusion auerswaldae was related to the Ediacaran taxa Rangea and
Charnia , but the alternative hypothesis of a relationship with the onychophores received renewed
support by Dzik and Krumbiegel (1989). A supposed specimen of Pteridinium from the Deep Spring
Formation (Lower Cambrian) of California (Cloud and Nelson 1966; see also Cloud and Nelson
1967) transpired to be an example of the Cambrian trace fossil Plagiogmus (Glaessner 1968; Cloud
and Bever 1973; note that Dzik and Krumbiegel (1989) proposed that one specimen might be a
poorly preserved xenusionid). With the exception of new frond-like fossils described below, there
appear to be no other reports from the Cambrian.

STRATIGRAPHY

The two occurrences described here are from Burgess Shale-type faunas (Conway Morris 19896), that occur
in the Lower Cambrian Parker Slate of north Vermont and the Middle Cambrian Burgess Shale (Stephen
Formation) of east British Columbia.

The Noah Parker quarry (USNM locality 319m, almost certainly the same as USNM locality 319g, but not
to be confused with USNM locality 25, see Shaw 1954, p. 1040) exposes the lower Parker Slate (Keith 1932;
Shaw 1954; originally known as the Colchester Formation, see Keith 1923) on the south-west flank of a small
hill (Parker's Cobble), 2-4 km north of Georgia Plains, Vermont. According to Shaw (1954, p. 1041) ‘the Noah

CONWAY MORRIS: CAMBRIAN EDI ACAR AN-LIKE FOSSILS

599

Parker quarry has been entirely quarried away’. The locality is well-known on account of a rich Lower
Cambrian fauna that includes trilobites (e.g. Resser and Howell 1938; Shaw 1955; Whittington 1990), other
arthropods (Resser and Howell 1938; Briggs 1976, 1979; Conway Morris 19896), sponges (Rigby 1987), other
shelly fossils (see Shaw 1955), trace fossils (Resser and Howell 1938), and soft-bodied fossils including the
putative Ediacaran survivor discussed here. It may be significant that Keith (1923, p. Ill) reported boulders
in the Parker Slate. A number of the other Burgess Shale-type localities around the Laurentian craton are
located adjacent to the proximal edge of the outer detrital belt, in a slope zone where slumping and boulder
beds often occur.

The Burgess Shale fauna has received extensive attention (e.g. Whittington 1985; Conway Morris 1989c,
1990). The majority of soft-bodied specimens come from an informal unit known as the Phyllopod bed,
exposed in the Walcott Quarry which is located about 5 km north of Field, British Columbia. Stratigraphically
this unit falls within the Pagetia bootes faunule of the Bathyuriscus-Elrathina Zone of Middle Cambrian age.
This unit is an integral part of the ‘thick’ Stephen Formation, which is composed predominantly of siltstones
and mudstones. The relationship between the ‘thick’ Stephen Formation and the ‘thick’ Cathedral Formation,
against which it is juxtaposed, has recently become a topic of controversy. A widely accepted interpretation
has been that the carbonates (now dolomitized) of the Cathedral formation formed a reef with a precipitous
escarpment against which the deep-water shales, including those of the Phyllopod bed, were deposited
(Mcllreath 1977; Conway Morris and Whittington 1979). This view has been challenged by Fudvigsen (1989),
who argued that the abruptness of the carbonate-clastic transition is exaggerated, and that rather than a reef
the original setting was a ramp with the Phyllopod bed biota owing its preservation to burial by tempestites.
Evidence against this reinterpretation, in support of the existing model, is presented by the original exponents
(Aitken and Mcllreath 1990; Fritz 1990; see also Fudvigsen 1990).

Institutional abbreviations. GSC, Geological Survey of Canada, Ottawa; NYSM, New York State Museum,
Geological Survey, Albany; UCMP, University of California Museum of Paleontology, Berkeley; USNM,
National Museum of Natural History (formerly United States National Museum of Natural History),
Smithsonian Institution, Washington, D.C.

SYSTEMATIC PAFAEONTOFOGY
Phylum, class, family uncertain
Emmonsaspis cambrensis (Walcott, 1890)

Plate 1, figs 1-2

1886 DiplograptusI simplex Walcott, pp. 15, 46, 51, 92-93, pi. 11, fig. 4 a [non fig. 4],

1889 Phyllograptus ? simplex ; Walcott, p. 388.

1890 PhyllograptuslI cambrensis Walcott, p. 604, pi. 59, fig. 3 [non fig. 3a],

1938 Emmonsaspis cambriensis [s/c] (Walcott); Resser and Howell, p. 233, pi. 9, fig. 1 [non pi. 9,
figs 2-4].

1952 Emmonsaspis ; Termier and Termier, p. 351.

1954 Emmonsaspis cambrensis (Walcott); Shaw, p. 1040.

1955 Emmonsaspis ? cambrensis (Walcott); Shaw, p. 775.

1955 Emmonsaspis cambrensis (Walcott); Shaw, pp. 785, 797.

1958 Emmonsaspis cambrensis (Walcott); Shaw, p. 531.

1968a Emmonsaspis', Termier and Termier, pp. 88, 91.

19716 Emmonsaspis', Durham, pp. 1105, 1121.

1972 Emmonsaspis cambriensis [j/c] (Walcott); Firby, p. 504.

1979 Emmonsaspis', Brasier, p. 125.

19896 Emmonsaspis cambriensis [x/c] (Walcott); Conway Morris, p. 278.

1990 Emmonsaspis', Bergstrom, p. 4.

1992 Emmensaspis [szc] ; Bengtson, p. 1033.

Revised diagnosis. Foliate body, tapering towards either end. Presumed distal termination simple;
proximal termination not known. Regularly spaced branches on either side inclined proximally to
impart chevron-like pattern.

600

PALAEONTOLOGY, VOLUME 36

Type and locality. The holotype (USNM 15314a) is associated on the same slab with two other specimens
(USNM 153146^c). Lower Cambrian, Parker Slate from USNM locality 3 1 9g at Parker’s Quarry, Parker’s
Cobble, Vermont.

History of research. When first described, Walcott (1886) compared the material to younger graptolites that
had been discussed by Emmons (1855), choosing as a name one of the junior synonyms for what Emmons
(1855) referred to as Diplograptus secalinus. The subsequent publications by Walcott (1889, 1890) suggest that
he may have become more sceptical about an assignation to the graptolites. Presumably aware that a place in
this group was inappropriate, Resser and Howell (1938, p. 233) erected Emmonsaspis. They noted that ‘The
central rod, the ribbing, and the general shape of this animal argue strongly for its reference to the chordates’,
a proposal then echoed by subsequent workers (e.g. Termier and Termier 1952, 1968a; Durham 19716; Brasier
1979). In contrast Shaw (1955) noted that the affinities of this animal were unknown. Further comments were
precluded because the specimens appeared to have been mislaid, and the line drawings of Walcott (1886, 1890)
and photographs published by Resser and Howell (1938) were inadequate for critical review. When the
specimens were finally relocated, it became clear that most were arthropodan and seemingly similar to the
Burgess Shale genus Perspicaris that had been described by Briggs (1977). What had been interpreted as the
central rod (and so a chordate notochord) was a gut-filled alimentary canal, while the apparently smooth
outlines revealed on closer inspection carapace and thorax, sometimes with traces of appendages. Of the
illustrated specimens only one (Walcott 1886, pi. 11, fig. 4a; 1890, pi. 59, fig. 3; Resser and Howell 1938,
pi. 9, fig. 1; herein PI. 1, fig. 1) is now attributable to Emmonsaspis , although it occurs with two more
poorly preserved specimens.

Description. Three specimens are on the same slab. Two are adjacent (PI. 1, fig. 2), one of which is very poorly
preserved. Both of these specimens are incomplete, although originally one of these was probably about 30 mm
long. The third specimen (the holotype; PI. 1, fig. 1) is better preserved, and although one end is incomplete
owing to rock breakage, the original length is estimated to have been approximately 41 mm (maximum width
is 10 mm).

Despite the vagaries of preservation and differences in size all three specimens show similar features (PI. 1,
figs 1-2). Overall the body tapers in either direction, but its orientation is conjectural and depends to some
extent on comparisons made with other organisms. Assuming that Emmonsaspis bears some relationship to
Ediacaran frond-like fossils (see below), then it is likely that the chevron-like arrangement of the branches was
directed abapically. The apex, therefore, appears to have been simple and obtuse. In neither of the better-
preserved specimens is the proximal region present, while in the third specimen little can be made in this region,
but no evidence exists for a holdfast or stalk. The margins of the body were smooth, lacking appendages or
other extensions.

Most prominent are the inclined branches of either side, that, meeting along the midline, impart the
distinctive chevron shape (PI. 1, fig. 1). So far as can be judged the branches arose opposite each other, and
although a midline is defined by the convergence of the branches there is no evidence for a discrete central stem
or rachis. In the holotype the branches total c. 14 per cm, indicating that originally this individual possessed
about 55 branches on each side. The branches are simple, lack subsidiary detail, but do show slight relief
indicating that in life they stood proud of the rest of the frond.

Mode of life. Incomplete preservation and uncertainty of taxonomic comparisons make
palaeoecological pronouncement on Emmonsaspis difficult. The animal is tentatively interpreted as
benthic, with the frond extending into the overlying water. The method of attachment is conjectural.
If the branches bore zooids, for which there is no direct evidence, then suspension feeding or
microcarnivory is plausible.

EXPLANATION OF PLATE 1

Figs 1-2. Emmonsaspis cambrensis (Walcott), Noah Parker Quarry, Georgia Plains, Vermont; Parker Slate,
Lower Cambrian. 1, holotype USNM 15314a, note inclined branches that impart chevron-like appearance
to the frond, x4-3. 2, USNM 153146 (left) and 15314c (right); the latter specimen is poorly preserved but
USNM 155146 shows features similar to the holotype, x 4-3. All three specimens co-occur on a single slab.
Photographed in white light.

PLATE 1

CONWAY MORRIS, Emmonsaspis

602

PALAEONTOLOGY, VOLUME 36

Rachis

Folded margin

Folded

margin

?Zooids

2 cm

Holdfast

Arcuate band

?Canal

Branch

text-fig. 2. Camera-lucida drawing of
the holotype of Thaumaptilon walcotti
gen. et sp. nov. (USNM 468028, Walcott
quarry, near Field, British Columbia;
Stephen Formation (Burgess Shale),
Middle Cambrian) to show interpreta-
tion of features. Drawing is of counter-
part (PI. 2, fig. 2), and some features
of the part have been combined with this
drawing by reversal.

EXPLANATION OF PLATE 2

Figs 1-2. Thaumaptilon walcotti gen. et sp. nov. Walcott quarry, near Field, British Columbia; Stephen
Formation (Burgess Shale), Middle Cambrian. Holotype, USNM 468028, entire specimens. 1, part;
2, counterpart, x 1 . Photographed in white light.

PLATE 2

CONWAY MORRIS, Thaumaptilon

604

PALAEONTOLOGY, VOLUME 36

Affinities. The graptolite affinities of Emmonsaspis may be rejected without further discussion. So
too may the notion that this creature is a chordate. As explained above the putative notochord
appears to represent the alimentary canal of Perspicaris- like animals filled with sediment, and their
preservation is sufficient to reveal unequivocal arthropodan features (Conway Morris 198%). While
the chevron-like arrangement of the branches vaguely recalls the disposition of chordate myotomes,
the absence of other features such as a notochord and alimentary canal render this interpretation
suspect. Emmonsaspis appears to approach no other Cambrian animal, including those from other
Burgess Shale-type faunas, and the preferred interpretation here is comparison with the Ediacaran
assemblage of frond-like fossils. In the context of Ediacaran-type organisms, closest comparison of
Emmonsaspis appears to exist with forms such as Pteridinium , but any similarities may be superficial
(see below).

Emmonsaspis from the Lower Cambrian of California? In an abstract Firby (1972) drew attention
to two types of Lower Cambrian fossils from the White Inyo Mountains, one lanceolate with a
‘central rod or tube’, the other ‘with a central rod, a broadly rounded “anterior” end tapering
toward a narrow “posterior” and having chevron-shaped impressions extending toward the edge
of the body’. Quoted as coming from the Campito Formation (Firby 1972), subsequent mention
(Firby and Durham 1974, text-fig. 2; Onken and Signor 1988, p. 142; see also Durham 1971a) places
the finds in the overlying formation, near the top of the middle of the Poleta Formation. Firby
(1972) proposed that these fossils were cephalochordates, comparable to Emmonsaspis. No formal
description has been published, and Durham (in litt. 2 October 1978) told me that in an attempt to
try a new casting technique the putative notochord of one specimen was destroyed. Thanks to the
kindness of Dr Firby I was able to examine this specimen, and concluded that at least this example
might be inorganic. In any event comparison of this specimen to Emmonsaspis or fossil
cephalochordates appears to be without foundation.

?Phylum cnidaria Flatschek, 1888

?Class anthozoa Ehrenberg, 1834

?Order pennatulacea Verrill, 1865
Family charniidae Glaessner, 1979a
Genus thaumaptilon gen. nov.

Type species. Thaumaptilon walcotti sp. nov.

Derivation of name. The generic name is a construct of the Greek words thauma (wonderful) and ptilon (soft
feather).

Diagnosis. Bilaterally symmetrical foliate animal with blunt holdfast. Leaf elongate and flattened,
with central broad axis. On one side branches arise on either side of axis, connected by narrow
strands, possibly canals. Proximal branches elongate, recurved adapically towards leaf margins;
distal branches more quadrate. Branches and other areas adjacent to axis papillate, possibly zooids.
Opposite side of leaf bears longitudinal ridges.

Thaumaptilon walcotti sp. nov.

Plates 2-3; Text-fig. 2

Derivation of name. The trivial name walcotti honours Charles Walcott, discoverer and first exponent of the
Burgess Shale fauna.

Diagnosis. See genus.

Types and locality. Holotype USNM 468028 (part and counterpart) is the largest and best preserved of the three
known specimens. Paratypes USNM 468029 and 468030 (parts only) are substantially smaller, and here are

CONWAY MORRIS: CAMBRIAN EDI ACARAN-LIKE FOSSILS

605

interpreted as juveniles. Middle Cambrian, Phyllopod bed (Burgess Shale), from USNM locality 35Ar at
Walcott Quarry near Field, British Columbia.

History of research. There is little doubt that Walcott had studied these specimens, especially as retouched
photographs of the holotype (A. Simonetta, pers. comm.) and one of the juveniles (USNM 468030) were found
with the specimens. During one of my searches through the collections the three specimens were set aside for
further study. Brief allusion was made to T. walcotti as having ‘a strong resemblance to a pennatulacean or
sea pen’ (Conway Morris 1979, p. 337), while the holotype was first illustrated in a recent review paper
(Conway Morris 1989c, fig. 5a) where its affinities with Ediacaran taxa were made explicit (see also Conway
Morris, 1990, p. 1 16).

Description. The holotype (PI. 2, figs 1-2; Text-fig. 2) is large, assumed to be mature, and with a total length
of 201 mm (straightened) and a maximum width of 55 mm. The other two specimens (PI. 3, figs 7-9) are
substantially smaller. The respective length and maximum width of USNM 468029 are 32 and 1 1 mm, while
those of USNM 468030 are 20 and 8 mm respectively. The basic form of the animal is a foliate frond and
holdfast. In USNM 468030 the frond is lanceolate (PI. 3, fig. 9), although its proximal taper may be
exaggerated by the non-preservation of the holdfast. In the holotype (PI. 2, figs 1-2; Text-fig. 2) and USNM
468029 (PI. 3, figs 7-8) the frond is more oblong, with the margins sub-parallel along much of their length but
tapering towards the apex and also progressively narrowing especially towards the holdfast. In the two
juveniles the apex is blunt, whereas in the adult it appears to be more rostrate, although this may have been
accentuated by folding of the distal end. The frond is assumed to have been relatively thin, and towards its
lateral margins there is evidence in the adult for overfolding (see below).

The frond consists of a central rachis and flanking areas. The rachis is most obvious in the holotype (PI. 2,
figs 1-2; Text-fig. 2). It appears not to extend into the holdfast, while its extension into the distal region is
obscure. Where obvious the rachis is relatively broad (9 mm), the margins are somewhat irregular, and local
deflections along the main axis occur. In USNM 468029 (PI. 3, figs 7-8) a median zone on the frond is taken
to mark the position of the rachis, but few details are evident. In USNM 468030 (PI. 3, fig. 9) the putative axis
is also marked by a dark strip, its narrowness consistent with the smaller size of the specimen. Apart from a
dark central strand along the part of the midline of the rachis in the holotype, the rachis itself is devoid of
specific features. On either side, however, dark strands extend from the margin of rachis adaxially to the ends
of the branches. These are clearest in the mid-region of the holotype (PI. 2, figs 1-2; Text-fig. 2), although their
configuration is rather variable, perhaps due to vagaries of preservation. Their basic arrangement, however,
appears to consist of two strands that arise from the proximal point of each branch and then diverge as they
extend to the rachis. This divergence could indicate that on reaching the rachis the strands were inserted into
upper and lower positions. There are instances, however, where instead of diverging, the strands appear to
converge. Such configurations may arise by local displacement of tissues, perhaps during decay.

The branches are the most prominent feature of the frond. They arise as pairs, opposite each other on either
side of the rachis. In no specimen can an exact total be counted, on account of either missing areas or
incomplete preservation. In the holotype (PI. 2, figs 1-2; Text-fig. 2) it is estimated that on each side there were
about 35^10 branches, a figure close to that estimated for the much smaller USNM 468029 (PI. 3, figs 7-8) with
c. 35 branches. In USNM 468030 (PI. 3, fig. 9), however, the total visible is lower (c. 15), but this might be an
underestimate on account of incomplete preservation, especially at the proximal end. The similarity in branch
totals between the holotype and USNM 468029, which is less than one-sixth of its size, indicates that branch
addition was limited to restricted budding at the apex, and that growth was largely accommodated by increase
in the dimensions of the individual branches. A comparison of branch width in the proximal region of the
holotype and USNM 468029 is also in the order of six times.

In the holotype the smallest branches are located adapically, where they form a closely-packed zone. The
branches, however, cannot be traced to the apex itself. In the proximal region the branches arise at an angle
of about 45° to the rachis. Adjacent to the margins of the body the branches not only narrow, but swing to
become subparallel to the edges. More distally the branches arise at a lower angle to the rachis, although the
differences on either side may be due to slightly oblique burial of this region. As seen the branches have simple
abrachial terminations, but this may be misleading because on both sides of the frond there appears to have
been overfolding (PI. 3, figs 4—5; Text-fig. 2), apparently truncating the branches. It may be significant that the
equivalent area of branches in USNM 468029 (PI. 3, fig. 7) shows an abrupt geniculation adjacent to the
margins; this is especially pronounced on the left-hand side. In USNM 468030 (PI. 3, fig. 9) the branches are
somewhat arcuate, with a more uninterrupted swing from rachis to margin, except in the apical zone where the
branches are straighter. This is also the case in USNM 468029, whereas in the holotype the apical branches
are quite strongly curved.

606

PALAEONTOLOGY, VOLUME 36

The branches are assumed to have formed cushion-like structures on the surface of the frond. In the holotype
they are connected by a series of low scarps (PI. 3, figs 2-3), and this gives the impression that the adapical edge
may have been free and was connected to the rest of the frond by a slight recess.

A notable feature of the branches are pustule-like structures (PI. 2, figs 1-2; PI. 3, figs 1,4; Text-fig. 2). These
structures have no precise counterpart in any other Burgess Shale fossil, but are tentatively interpreted as
retracted zooids. Although visible in both paratypes (PI. 3, figs 7, 9), they are clearest in the holotype. Locally
the pustules appear to form linear arrays, but overall there is no clear evidence for a particular disposition or
arrangement. As preserved the pustules form more or less homogenous blobs and in none of the specimens
is there evidence for tentacles or other substructure. Although the pustules are most obvious on the branches,
they also occupy quadrate areas adjacent to the rachis, most obviously in the proximal region (PI. 1, figs 1-2;
Text-fig. 2). The overlap with the rachis indicates that the latter structure may have been effectively internal.

At various points along the margin of the holotype the branches are truncated, and a change in level marked
by a scarp reveals marginal areas with prominent longitudinal striations (PI. 2, figs 1-2; PI. 3, figs 2, 4-5; Text-
fig. 2). These regions are interpreted as the overfolded edges of the opposite side of the frond, with the striations
believed to represent a system of somewhat irregular longitudinal ridges. These structures appear to be also
visible in USNM 468030 (PI. 3, fig. 9), especially on the proximal region of the right-hand side where a series
of prominent lineations run at a steep angle to the branches. In the holotype curving sets of lineations are
faintly visible on the frond, cross-cutting the trend of the branches (PI. 2, fig. 2; Text-fig. 2). They may represent
the ridges of the opposite side being impressed by compaction.

The holdfast is only well preserved in the holotype (PI. 2, figs 1-2; PI. 3, fig. 6). It is attached to the frond
by a fairly prominent constriction in body width, and thereafter the holdfast tapers to a blunt termination. The
holdfast is devoid of anatomical detail, apart from an arcuate reflective band (PI. 3, fig. 6) that is convex
towards the proximal end.

Mode of life. There is little doubt that T walcotti (Text-fig. 3) was benthic, with much or all of the
holdfast embedded in unconsolidated sediment and the frond extending into the overlying water.
Position relative to the sediment-water interface presumably was adjusted by muscular contractions
of the holdfast. The mode of feeding is rather conjectural, but if the putative zooids have been
identified correctly then food particles were presumably trapped by small tentacles.

?Class anthozoa Ehrenberg, 1834
?Order actiniaria Hertwig, 1882
Family mackenziidae fam. nov.

Diagnosis. Elongate polyps, atentaculate, with longitudinally-folded exterior, internal septa lining
gastric cavity.

explanation of plate 3

Figs 1-9. Thaumaptilon walcotti gen. et sp. nov. Walcott quarry, near Field, British Columbia; Stephen
Formation (Burgess Shale), Middle Cambrian. 1-6, holotype, USNM 468028. 1, part, detail of branches and
possible zooids, x 2-5. 2, counterpart, detail of branches showing evidence for imbrication and folded-over
edge of frond, x 2-0. 3, counterpart, same area as in fig. 2, x 2 0. 4, part, detail of branches and folded-over
edge of frond, x F2. 5, part, folded-over edge of frond and branches, x 1-2. 6, counterpart, holdfast, x IT.
7-8, USNM 468029, entire specimen, fragment in centre of specimen is a portion of eodiscid trilobite Pagetia
bootes (note other specimen at base), x 2-4 and x 2T respectively. 9, USNM 468030, entire specimen, x 3 3.
All photographs taken in ultra-violet radiation; 1, 2, 4-7, 9 under high angle radiation; 3 and 8 under low
angle radiation.

PLATE 3

CONWAY MORRIS, Thaumaptilon

608

PALAEONTOLOGY, VOLUME 36

Genus mackenzia Walcott, 1911
Type species. Mackenzia costalis Walcott, 1911.

Revised diagnosis (emended from Walcott 191 1, p. 55). Bag-like animal, atentaculate, with exterior
folded longitudinally in about 8-10 ridges. Probable internal septa lining extensive gastric cavity.
Holdfast or other proximal modification present.

Mackenzia costalis Walcott, 1911
Plates 4-7

1911 Mackenzia costalis Walcott, pp. 45, 54-55, pi. 13, figs 2-3.

1912a Mackenzia costalis Walcott; Walcott, p. 153.

19126 Mackenzia costalis Walcott; Walcott, p. 190.

1912 Mackenzia costalis Walcott; H. L. Clark, pp. 276-278.

1913 Mackenzia costalis Walcott; A. H. Clark, pp. 488-489, 503-504, 507.

1918 Mackenzia costalis Walcott; Walcott, p. 10, figs 6-7.

1921 Mackenzia costalis Walcott; Raymond, pp. 343-345, fig. 3.

1931 Miskoia Walcott, p. 38, pi. 3, figs 2-3.

1932 Mackenzia costalis Walcott; Croneis and McCormack, p. 126.

1934 Mackenzia costalis Walcott; Ruedemann, p. 76.

1952 Mackenzia ; Termier and Termier, p. 357.

1956 Mackenzia costalis Walcott; Wells and Hill, p. F233.

1957 Mackenzia costalis Walcott; Madsen, p. 281.

1960 Mackenzia ; Termier and Termier, p. 36.

1968a Mackenzia costalis Walcott; Termier and Termier, fig. 142.

19686 Mackenzia ', Termier and Termier, p. 92.

1968 Mackenzia costatis [.sic] Walcott; Arai and McGugan, p. 208.

1969 Mackenzia costalis Walcott; Rolfe, p. R331.

19716 Mackenzia costalis Walcott; Durham, pp. 1106-1107.

1973 Mackenzia', Alpert, p. 919.

1978 Mackenzia costalis Walcott; Conway Morris, p. 126.

1979 Mackenzia costalis Walcott; Conway Morris, p. 337.

1979 Mackenzia', Conway Morris and Whittington, p. 126.

1979 Mackenzia costalis Walcott; Scrutton, pp. 170, 177.

1982 Mackenzia', Sepkoski, p. 23.

1985 Mackenzia costalis Walcott; Whittington, pp. 53, 124, 127, fig. 4.16a-b.

1989c Mackenzia costalis Walcott; Conway Morris, pp. 343-344, fig. 5b-c.

1990 Mackenzia costalis Walcott; Conway Morris, p. 116, fig. 4a.

1991 Mackenzia', Gehling, p. 190.

1992a Mackenzia', Sepkoski, p. 35.

19926 Mackenzia ; Sepkoski, p. 1173.

1992 Mackenzia costalis Walcott; Conway Morris, p. 632.

1992 Mackenzia costalis Walcott; Bengtson et at., p. 434.

History of research. Walcott’s (1911) placement of M. costalis in the holothurians, with similarities being drawn
with the extant Synaptula , was soon questioned by H. L. Clark (1912) who drew comparisons with certain
actinians (Cnidaria). A. H. Clark (1913) complained about the tentativeness of H. L. Clark’s (1912) discussion,
and had no hesitation in regarding M. costalis as ‘undoubtedly mud-living actinians of the family
Edwardsiidae, closely related to Edwardsia' . Within a few years Walcott (1918) had conceded to A. H. Clark’s
(1913) proposal, writing that this species ‘closely resembles Edwardsia'. This view came to be widely accepted
(e.g. Croneis and McCormack 1932; Wells and Hill 1956; Madsen 1957; Termier and Termier 1960; Durham
19716; Sepkoski 1982, 1992a, 19926; Whittington 1985), although Scrutton (1979, p. 177) noted the need for
reappraisal. An interesting parallel to this debate may be found in the discussion of Pseudocaudina brachyura
from the Upper Jurassic Solnhofen Limestone of south Germany. The only known fossil was first described
as a holothurian (Broili 1926), but subsequently Heding (1932) suggested it was an actiniarian. This taxon was

CONWAY MORRIS: CAMBRIAN EDI ACARAN-LIKE FOSSILS

609

text-fig. 3. Reconstruction of Thau-
maptilon walcotti from the Middle Cam-
brian Stephen Formation (Burgess
Shale), British Columbia. Towards the
upper right the frond is depicted as being
folded over to reveal opposite side.

Approximately x 1 .

610

PALAEONTOLOGY, VOLUME 36

not mentioned by Wells and Hill (1956), and comparisons with Mackenzie! are drawn below. In recent review
papers (Conway Morris 1989c, 1990) I have introduced the notion of M. costalis being considered as an
Ediacaran-type survivor, although this need not prejudice its relationships to the Cnidaria (see below).

Material. Asterisk indicates that both part and counterpart exist. Lectotype (designated herein) USNM
57556*; other material includes USNM 57557*, 1 93929*-l 93937*, 1 93939*-l 93949*, 1 9395 1 *-193956*,
193959, 193961, 193964-193966, 193967*, 193968-193969, 193971-193972, 193974-193977, 193979-193982,
193984, 193988-193990, 193994, 193997, 193999, 194001, 194652, 196174, 202304; GSC 78489-78490,
78492-78500; ROM 38643, 48471-48475.

Stratigraphical levels in the Burgess Shale. Although about 60 specimens reside in the USNM, most must have
been collected after 1910 because at that time only two specimens had been obtained (Walcott 1911, p. 55).
Walcott (1912a, p. 153) noted that within the Phyllopod bed the specimens came from the richly fossiliferous
bed 10 (0 48 m above quarry floor), but apart from this information the stratigraphical range of Walcott’s suite
is uncertain. When the Geological Survey of Canada teams collected Burgess Shale material in 1966 and 1967
a more careful note was kept of heights of specimens within the Burgess Shale. Of the nine specimens from the
Phyllopod bed, the lowest horizon at which a specimen occurred was T 1 1"-2' 4" (0 58— 0-7 1 m) above the base
of the reopened Walcott Quarry (this may not coincide exactly with the original floor) and another specimen
was found at the 2' 4" (0-71 m) level. Above these three specimens came from the 2' 7"-3' 0" (0-78-0-91 m)
interval, and four specimens from the 3' 0"-3' 7" (0-78—1-09 m) horizon. From a horizon substantially above
the Phyllopod bed, at a level of 74-76' (22-56—23-1 6 m) above the base of the Walcott Quarry, two more
specimens were recovered.

Description. Overall M. costalis is elongate, cylindrical, without obvious appendages or other extensions
(PI. 4, figs 1-6; PI. 5, figs 1-2, 5, 7; PI. 6, figs 1, 4-5; PI. 7, figs 1-2, 4, 7). The body is more or less constant
in width, with blunt terminations. Mackenzia shows a wide range of size; mean length (n = 32; remaining
specimens are incomplete) is 86-5 mm (SD 384 mm), with recorded lengths between 25 and 158 mm (Text-fig.
4). To a first order widths are directly proportional to length, although in the more elongate specimens (i.e.
more than about 100 mm) greater length is achieved without noticeable increase in width (Text-fig. 4). There
is, moreover, a wide variation in length to width ratios, presumably reflecting the contractability of the body.

External features are few. Walcott (1911) drew attention to ‘eight to ten longitudinal bands ... outlined by
narrow, slightly elevated lines’. These observations are confirmed, in as much as there appears to be a direct
correspondence between the relatively prominent reflective strands (PI. 5, figs 1, 4; PI. 6, figs 4-5) and the
elongate lines that are delimited as low scarps (PI. 4, figs 1^4; PI. 5, figs 2-3, 5, 7; PI. 7, figs 1-3). In a number
of specimens the edges of the body are not entirely smooth, but consist of displaced margins that are also
separated from one another on different levels connected by scarps (e.g. PI. 5, figs 1-3, 7; PI. 7, figs 1-2). It is
conjectured that in life the circumference of the body was not simple but thrown into relatively deep folds and
intervening ridges, the expression of which is now seen in the elevated lines and displaced margins. Further
support for this comes from the distal end of some specimens which have a lobate appearance (PI. 4, figs 1-2,
6; PI. 5, fig. 4; PI. 7, figs 4, 6), interpreted as arising from the convergence of the external folds and ridges.

The proximal end of the body may be attached to a holdfast, but otherwise is normally blunt. An exception
to this is seen in GSC 78494, where the proximal area has an arcuate ornamentation and is connected to the
rest of the body by a narrow strand (PI. 6, figs 1-2). This unique feature is interpreted as an attachment disc,
anchoring the rest of the body via a stalk.

Although many specimens are not attached to a foreign holdfast, there are exceptions. In some individuals

EXPLANATION OF PLATE 4

Figs 1-6. Mackenzia costalis Walcott. Walcott quarry, near Field, British Columbia; Stephen Formation
(Burgess Shale), Middle Cambrian. 1-3, lectotype (designated herein), USNM 57556. 1-2, part, entire
specimen; 1, x 1-8; 2, x 2-0. 3, counterpart, entire specimen, proximal area missing owing to rock breakage,
x 24. 4, USNM 57557, entire specimen, x 1-2. 5, USNM 194652, entire specimen with a holdfast of an
eocrinoid stem (left), x 0-9. 6, USNM 83927a, entire specimen, associated with the enigmatic taxon Portalia
mira , x 1. Figs 1-4 photographed in ultra-violet radiation (1-2 at high angle; 2, 3, 6 at low angle); fig. 5
photographed in white light.

PLATE 4

CONWAY MORRIS, Mackenzia

612

PALAEONTOLOGY, VOLUME 36

at the proximal end there is a small cluster of skeletal debris, typically that of sponge spicules and brachiopods
(PI. 5, figs 5-6; PI. 6, fig. 8). This debris appears to have been located in a central area of the base. In other
specimens the proximal end is attached to the stem of an echinoderm (PI. 4, fig. 5 ; PI. 7, fig. 9). This association
is unlikely to be due to chance both because of its multiple occurrence in Mackenzia and the general scarcity
of echinoderm material in the Burgess Shale. The final example of a foreign holdfast is in GSC 78494 (PI. 5,
figs 1-2), where an indeterminate nodular mass is present in the basal region.

Features of internal anatomy are sparse. In his initial description Walcott (1911, p. 55) noted that at the
anterior end of the two specimens then discovered there was ‘a ring of what appear to be narrow plates
surrounding a central opening’. The presence of this feature is somewhat enigmatic because although
H. L. Clark (1912) ‘was unable to make out these points satisfactorily’, A. H. Clark (1913, p. 489) claimed to
see this ring of plates but reported that after immersion of the specimen in acid by Walcott ‘the ring seems to
have disappeared’. These discrepancies are resolved, however, when it is realized that the end of the specimen
in question (PI. 4, figs 1-2) undoubtedly represents the basal termination, and a ring-like structure is present
but represents the zone of attachment. In any event there is no evidence for calcareous plates or spicules in this
or any other specimen.

The most prominent of the unequivocal internal features are reflective strands (e.g. PI. 5, figs 1, 4; PI. 6, figs
4-5). They are tentatively interpreted as extensions of the body wall into an internal cavity, similar to the septa
of anthozoans. Apart from longitudinal fraying seen in a few specimens (e.g. GSC 78496; PI. 6, fig. 3), which
may represent partial decay, the strands show no substructure. Possibly separate from the above-mentioned
strands is a longitudinal reflective ribbon, visible in two specimens (GSC 78498, USNM 193947; PI. 6, figs 1,
6). It has a more organized structure than the longitudinal strands, with reflective margins enclosing an area
with reflective patches that range from elongate to more reticulate. The significance of this structure and its
original nature are obscure. The only other internal structure is a lenticular to more elongate mass that occurs
either towards the putative oral opening (PI. 7, figs 1, 3, 4—6) or in a more mid-portion (PI. 7, figs 7-8). They
have a fibrous to granular texture, but otherwise contain no identifiable structures. These masses are tentatively
identified as gastric residues.

Mode of life. Mackenzia was benthic, as inferred both from the absence of obvious flotant or
swimming structures and attachment to benthic debris, including echinoderm stems. It is these latter
examples that suggest that at least these specimens were not partly embedded in the sea-floor (cf.
the reconstruction in Conway Morris and Whittington 1979). Otherwise the mode of life of this
animal is largely enigmatic. The variation in length to width ratios (Text-fig. 4) indicates that the
animal was contractile, presumably achieving this by muscular contractions. The mode of feeding
is uncertain. Mackenzia lacks tentacles or other obvious grasping devices, and contains no
identifiable food items that might indicate its dietary preferences. Its sessile mode of life makes
either predatory activity or suspension feeding of particulate material more likely. In the former
it could be hypothesized that abrupt suctorial action could sweep macroscopic prey into the central
digestive cavity. Alternatively, the ciliated tracts could induce water circulation with particles being
trapped by mucous nets spun either into the surrounding water, suspended adjacent to the putative
oral opening, or collected over the surface of the body.

EXPLANATION OF PLATE 5

Figs 1-7. Mackenzia costalis Walcott. Walcott quarry, near Field, British Columbia; Stephen Formation
(Burgess Shale), Middle Cambrian. 1-2, GSC 78494, entire specimen; 1, x 1-2; 2, x 0 6. 3, GSC 78494, detail
of proximal region and holdfast composed of nodular material, x 1-3. 4, GSC 78490, entire specimen
associated with an echinoderm and also ?chondrophorine (see PI. 8, figs 1-2), x F3. 5, USNM 193967, entire
specimen, x 1-3. 6, USNM 193967, detail of proximal region with holdfast composed of a brachiopod and
sponge spicules, x 6. 7, GSC 78497, entire specimen, x 0 8. All specimens photographed in either high angle
(1, 4) or low angle (2-3, 5-7) ultraviolet radiation.

PLATE 5

CONWAY MORRIS, Mackenzia

614

PALAEONTOLOGY, VOLUME 36

?Class hydrozoa Owen, 1843

?Suborder chondrophorina Chamisso and Eysenhardt, 1821
Family uncertain
Genus gelenoptron gen. nov.

Type species. Gelenoptron tentaculatum sp. nov.

Derivation of name. The generic name is a construct of the Greek word enoptron (mirror) and Latin word gelatus
(gelatinous), an oblique reference to the strongly reflective preservation of the putative float and the inferred
gelatinous composition of the animal.

Diagnosis. Elongate structure, possibly a float, with marginal rim of densely spaced tentacles.
Separate lower unit, bearing larger, more stout tentacles with faint annulations.

Gelenoptron tentaculatum sp. nov.

Plate 8, figs 3-4

1931 Redoubtia polypodia Walcott, p. 3, pi. 2, fig. 2 (non fig. 3).

1989c ‘ Redoubtia polypodia' Walcott, Conway Morris, p. 343, fig. 5d.

Derivation of name. The trivial name tentaculum refers to the abundance of tentacles.

Diagnosis. See genus.

Types and locality. Holotype USNM 83925 (part and counterpart). Middle Cambrian, Phyllopod bed (Burgess
Shale) from USNM locality 35k at Walcott Quarry near Field, British Columbia

History of research. In 1918 Walcott (1918, fig. 5) illustrated a lobopod-like animal as Redoubtia polypodia.
Subsequently (Walcott 1931, pi. 2, fig. 3) this specimen was re-illustrated, with a second specimen (PI. 8,
fig. 3). Resser (in Walcott 1931, p. 3) noted that ‘Whether the second specimen really represents the same species
appears somewhat doubtful inasmuch as the tube feet [R. polypodia was regarded as a holothurian] are smaller
and more numerous. The larger appendages above the specimen .... are parts of another animal.’ Since then
R. polypodia has received sporadic mention, mostly concerned with its possible status as a holothurian and with
evident reference to the lobopod-like specimen (e.g. Croneis and McCormack 1932; Madsen 1957) rather than
the material illustrated here. Conway Morris (1989c) re-illustrated the holotype, and made passing reference
to its possible status as a chondrophorine.

Description. Although several specimens are known, only the holotype (PI. 8, figs 3-4) is particularly
informative and even here the overall anatomy is not entirely resolved. The bulk of the specimen consists of
a broad reflective area, tapering towards one end. Around the edge of this organ is a series of tentacle-like
structures, densely spaced and extending outwards for about 9 mm. It appears that these tentacles formed at
least two, and perhaps three, distinct layers. Each tentacle is narrow and tapers to a fine point. On the upper
side of the specimen (as illustrated) a separate set of structures (PI. 8, figs 3-4) lie at a distinctly different level

EXPLANATION OF PLATE 6

Figs 1-8. Mackenzia costalis Walcott. Walcott quarry, near Field, British Columbia; Stephen Formation
(Burgess Shale), Middle Cambrian. 1, GSC 78498, entire specimen, x 0 8. 2, GSC 78498, detail of proximal
area and holdfast, x L9. 3, GSC 78496, detail of strands of tissue, probably resulting from partial decay,
x 2-2. 4-5, USNM 193930, part and counterpart respectively, entire specimens, x 1. 6-7, USNM 193947,
counterpart and part respectively, entire specimens, x 1-4. 8, USNM 193968, entire specimen, note holdfast
with sponge spicules at proximal end, x 0-9. Fig. 1 photographed in white light, others in ultra-violet
radiation at either a high angle (2-6, 8) or low angle (7).

PLATE 6

CONWAY MORRIS, Mackenzia

616

PALAEONTOLOGY, VOLUME 36

(lower in the part). Adjacent to the reflectively preserved organ this unit is featureless, but distally it extends
into a series of stout tentacles with faint transverse annulations and pointed terminations. This structure is
regarded as an integral part of the organism, rather than a chance superposition (cf. Walcott 1931).

Discussion. The affinities of G. tentaculatum are uncertain. There is no support for an affinity with
the holothurians, and the recent tentative assignment to the chondrophorines (Conway Morris
1989c) is explained here. In this interpretation the broadly reflective area is interpreted as the float
or pneumatophore. In contrast to known chondrophorines, however, the float shows no indication
of chambers or other subdivisions. Comparing the tentacle-like structures in the fossil to the
complex series of zooids in living chondrophorines is not straightforward. It could be hypothesized
that the rim of tentacle-like structures is equivalent to the tentaculozooids or dactylozooids, while
the larger tentacles might conceivably form part of the set of the gonozooids. This interpretation,
however, is by no means secure. Attempts to compare G. tentaculatum with the siphonophores,
another group of complex hydrozoans (Harrington and Moore 1956), are no more satisfying. The
evidence for common possession of a float or pneumatophore cannot be matched by convincing
comparisons between the various zooids of siphonophores and the tentacular structures of the
fossil.

A number of other specimens (USNM 198613, 200621, 200607-200608, 200588 (part and
counterpart)) are comparable to the holotype in having margins that are either tentacular or hirsute,
but despite this similarity cannot be confidently assigned to G. tentaculatum. The affinities of these
specimens remain obscure, except that USNM 198613 appears to be some sort of worm with a pair
of prominent tentacles.

?Chondrophorine
Plate 8, figs 1-2; Text-fig. 5

Diagnosis. Disc (?pneumatophore) with closely spaced annulations and elongate tentacles
(?dactylozooids) extending beyond margin.

Material and locality. GSC 78491 (only known specimen, superimposed on GSC 78490, a specimen of
Mackenzia costalis Walcott). Walcott Quarry, 5 km north of Field, British Columbia.

Stratigraphical horizon. About 3' 0"-3' 7" (0-91-109m) (GSM 1967 collecting level 81210) above base of
Phyllopod bed, ‘thick’ Stephen Formation, Middle Cambrian.

Remarks. The specimen (PI. 8, figs 1-2; Text-fig. 5) is incomplete owing to rock breakage, and is
superimposed on a specimen of Mackenzia costalis Walcott. Also adjacent is an unidentified
echinoderm. Given the confusion that has arisen with specimens that were believed to show chance
superposition but proved to be integral (Conway Morris 1978; Whittington and Briggs 1985) and
continuing uncertainty about whether spinose extensions in some arthropods are in situ or chance

EXPLANATION OF PLATE 7

Figs 1-9. Mackenzia costalis Walcott. Walcott quarry, near Field, British Columbia; Stephen Formation
(Burgess Shale), Middle Cambrian. 1-2, USNM 193929, part and counterpart, entire specimens, x 1.
3, USNM 193929, part, detail of possible digestive contents, x 4-9. 4, USNM 193955, part, entire specimen,
x0-7. 5, USNM 193955, counterpart, detail of possible digestive contents, x 3-9. 6, USNM 193955,
counterpart, detail of distal region, x T3. 7, USNM 193961, entire specimen, x T8. 8, USNM 193961, detail
of possible digestive contents, x 3-3. 9, USNM 196174, two individuals (proximal portions only) attached
to stem of eocrinoid, x F5. Specimens photographed either in white light (1-2, 4, 6-7) or low angle ultra-
violet radiation (3, 5, 8-9).

PLATE 7

CONWAY MORRIS, Mackenzia

618

PALAEONTOLOGY, VOLUME 36

text-fig. 4. Plots of size frequency (upper histogram) and bivariate scattergram of length against maximum

width (lower graph) of Mackenzia costalis Walcott.

overlaps (Whittington 1981, p. 351 ; Conway Morris and Robison 1988, pp. 29-30), it is important
to determine whether this unique specimen was indeed an individual or integral part of Mackenzia.
Evidence for the former is principally the absence of anything similar in the remaining suite of
Mackenzia from the Burgess Shale (c. 70 specimens). An alternative possibility is the discoidal fossil
represents part of the life cycle of Mackenzia , specifically a swimming or floating stage that budded
from the sessile stage for dispersal. One point against this interpretation, perhaps, is the large size
of the disc relative to the specimen of Mackenzia.

EXPLANATION OF PLATE 8

Figs 1-2. Possible chondrophorine. GSC 78491, entire specimen photographed from two different angles of
radiation, superimposed on an individual of Mackenzia costalis Walcott (see PI. 5, fig. 4), x 31 and x 2-6
respectively. The branched structure on the lower side of the figures is an echinoderm.

Figs 3-4. Gelenoptron tentaculatum gen. et sp. nov. USNM 83925, part and counterpart respectively, x F3.

Walcott quarry, near Field, British Columbia; Stephen Formation (Burgess Shale), Middle Cambrian.

All photographs taken in ultra-violet radiation.

PLATE 8

CONWAY MORRIS, ?chondrophorine, Gelenoptron

620

PALAEONTOLOGY, VOLUME 36

Description. The specimen is poorly preserved, and difficult to resolve on account of its superposition with the
individual of Mackenzia. It is most obvious from the concentric annulations, which average about 1 1 per 5 mm.
The margins of the disc are not clear, but the original diameter is estimated to have been about 23 mm. On
one side four elongate structures arise; these are interpreted as the remains of tentacular organs.

A ffinities. This specimen is tentatively regarded as a chondrophorine hydrozoan, a group of colonial
cnidarians represented today by the pelagic Porpita and Velella. Under this interpretation the disc
with concentric annulations is taken to be the chitinous pneumatophore with annular gas-filled
chambers that imparted buoyancy. There is no evidence for a sail on the upper surface of the
putative pneumatophore, and in this sense the specimen resembles Porpita which in contrast to
Velella lacks a sail. Pursuing the comparisons with chondrophorines would suggest that the
tentacles be interpreted as some of the marginal dactylozooids. Incomplete preservation might
explain why they are not visible around the remainder of the disc margin, and why the other
more centrally located zooids (gonozooids and gastrozooids) are not visible.

DISCUSSION

Introduction

The interpretation of Ediacaran faunas is complicated considerably by the competing hypotheses
of whether the various species are more or less comparable to extant taxa, notably cnidarians,
annelids and arthropods (e.g. Glaessner 1984), or whether they are so distinct from known
metazoans as to justify erection of a separate category of Vendozoa or Vendobionta (Seilacher 1984,
1989, 1992; see also Bergstrom 1989, 1990). The merits of these proposals are returned to below, but
are of immediate relevance because organisms supposedly disparate according to one hypothesis are
closely related if the other is correct. For example, Spriggina has been widely interpreted as a
metamerically segmented metazoan of either annelid or arthropod grade (e.g. Birket-Smith 1981;
Glaessner 1984). However, the vendobiontan interpretation (e.g. Bergstrom 1989; Seilacher 1989,
1992) takes the supposed head-shield as an organ for benthic attachment, and the putative
segmented trunk as a frond-like body subdivided into transverse compartments.

The definition of taxonomic categories, therefore, is potentially elastic. This is especially true of
the frond-like organisms abundant in many Ediacaran assemblages. Some, especially Charniodiscus
and Vaizitsinia, are patently similar to Thaumaptilon , but other taxa differ to varying degrees. While
it is possible, therefore, to provide relatively detailed comments on the likely relationships between
a few Ediacaran taxa and younger forms, a comprehensive review is more difficult.

Possible relatives of Thaumaptilon

Among the Ediacaran taxa are those frond-like fossils with a broad blade, usually with a median
rachis, and often a stalk terminating in a more-or-less discoidal holdfast. Taxa include Charnia (e.g.
Ford 1958; Fedonkin 1981, 1985), Charniodiscus (e.g. Ford 1958, 1963; Glaessner and Daily 1959;
Fedonkin 1981, 1985, 1987; Jenkins et al. 1983; Sun 1986; Gehling 1991), Glaessnerina (e.g.
Glaessner and Wade 1966; Germs 1973; Jenkins and Gehling 1978), Paracharnia (e.g. Sun 1986),
Vaizitsinia (e.g. Fedonkin 1981, 1985), and possibly Ramellina (e.g. Fedonkin 1981, 1985) and
Valdainia (e.g. Fedonkin in Velikanov et al. 1983; Fedonkin 1985).

Among the Ediacaran frondose taxa, Thaumaptilon (Text-fig. 3) seems to approach most closely
Charniodiscus and Vaizitsinia , although with the latter genus comparisons cannot be as extensive on
account of more limited information. Charniodiscus is best known from South Australia (Jenkins
and Gehling 1978; Jenkins 1984, 1992; see also Glaessner and Daily 1959; Glaessner and Wade
1966; Jenkins et al. 1983; Gehling 1991; Runnegar 1992), although its type occurrence is from
Charnwood Forest, England (Ford 1958, 1963; Jenkins and Gehling 1978). Discoidal fossils
interpreted as the basal attachment organs of Charniodiscus have also been reported from the White
Sea, north-west Russia (Fedonkin 1981, pi. 3, fig. 8), the Wernecke Mountains of Yukon (Narbonne

CONWAY MORRIS: CAMBRIAN EDIACARAN-LIKE FOSSILS

621

text-fig. 5. Camera-lucida drawing of specimen of Mackenzia costalis Walcott; Walcott (GSC 78490; see
PI. 5, fig. 4) with superimposed circular fossil (GSC 78491; PI. 8, figs 1-2) tentatively interpreted as a
chondrophorine. Walcott quarry, near Field, British Columbia; Stephen Formation (Burgess Shale), Middle

Cambrian.

and Hofmann 1987, text-fig. 5 b-d), and the Mackenzie Mountains of north-west Canada
(Narbonne and Aitken 1990, pi. 1, fig. 7; this specimen shows an indistinct longitudinal trace to the
right of the disc as illustrated, evidently representing part of the stalk). The assignment of these discs
to holdfasts of frondose animals, including Charniodiscus , was strongly endorsed by Jenkins (1989).
In particular, he (Jenkins 1989, p. 313; see also Jenkins 1992, p. 157) noted ‘that Cvclomedusa
should best be considered as a form-genus representing variable holdfasts of sea-pen-like animals’,
the implication being that some are probably synonymous with Charniodiscus. Species of
Charniodiscus recognized on the basis of preserved fronds include C. concentricus (type species),
C. arboreus, C. longus and C. oppositus. Overall similarities between the taxa include a broad
lanceolate frond with central rachis, on the ventral side prominent primary branches that are fused
along their bases to the frond and extend into the rachis via canal-like structures, a more or less
smooth dorsal surface, and an elongate stalk and holdfast. Dimensions are also broadly
comparable. Among the species of Charniodiscus , Thaumaptilon seems to approach most closely
C. oppositus and perhaps C. concentricus.

There are, nevertheless, a number of differences, seemingly sufficient to denote generic separation,
but still inclusion in the same family, presumably the Charniidae Glaessner, 1979a, if Charnia is
accepted as being related to Charniodiscus and Thaumaptilon (see below). One difficulty in drawing
objective comparisons is the different style of preservation between the compression fossils of the
Burgess Shale in shale and the mouldic impressions of Ediacaran fossils in sandstones and siltstones.

The following are the most obvious differences: (a) Thaumaptilon appears to lack a membranous
edge to the frond, and the primary branches extend close to the margin; (b) the holdfast is relatively
elongate and, although showing some indication of expansion (PI. 3, fig. 6) does not match the disc-
like termination of Charniodiscus with its prominent radial ornamentation (?musculature; especially

622

PALAEONTOLOGY, VOLUME 36

in C. oppositus ) and concentric ridges (in C. concentricus) (see Jenkins 1992, p. 161). It remains
possible that the holdfast of Thaumaptilon was more contractile than present evidence would
suggest, and in life could expand into a discoidal structure. It should also be noted that Jenkins
(1992, p. 161) commenting on the holdfasts of Charniodiscus noted that ‘their apparent flat shape
resulted from sediment compaction and they were presumably subspherical in life’; (c) in
Thaumaptilon the branches seem to have been more recurved at the frond margin than in
Charniodiscus. In this latter genus, Jenkins and Gehling (1978; see also Glaessner and Daily 1959,
p. 386) presented evidence that although the primary branches were fused to the frond, the zone of
attachment was narrow so that overall the branches were relatively flexible. In contrast, the freedom
of movement in Thaumaptilon appears to have been confined to the distal margin so that the
branches were more cushion-like.

There are two other significant differences, but these may be effectively taphonomic. The first is
the absence of spicules in Thaumaptilon. In Charniodiscus putative spicules occur (e.g. Glaessner and
Daily 1959; Jenkins 1992) in the rachis. They also supposedly occur on the branches, where the
clarity of demarcation between the putative anthosteles along the primary branches (so defining the
secondary branches) is regarded as reflecting the presence of spicules (Jenkins and Gehling 1978).
No ultrastructural information on these spicules is available, and their status as spicules remains
questionable. It seems unlikely that spicules in Thaumaptilon were lost during diagenesis, because
a wide variety of spicular structures are evident in the Burgess Shale sponges (Rigby 1986). The
second difference is that the supposed anthosteles (i.e. secondary branches), in Charniodiscus find
no exact counterpart in Thaumaptilon. It is true that some of the branches in the latter taxon have
transverse striations, but in contrast to Charniodiscus these are less regular and more closely-spaced.
In any event it seems unlikely that the secondary branches in Charniodiscus are true anthosteles,
especially on account of their larger size. Jenkins (1984, p. 99) noted that if Charniodiscus had true
polyps, then these would have been difficult to preserve. This comment accords with the size and
preservation of the putative zooids (PI. 3, fig. 1) in Thaumaptilon. Narbonne and Hofmann (1987,
p. 655) reported ‘subcircular tubercles 0- 5-7-0 mm across and up to TO mm relief’, but considered
them to be ‘probably accidental’. The likelihood that these structures are zooids seems to be remote
given their irregular distribution and widely variable size. Examination of the exceptionally well-
preserved holotype of Charniodiscus oppositus with R. J. F. Jenkins (see also Jenkins 1992, p. 161)
in Adelaide (July, 1992) specifically to search for zooids did reveal very slight irregularities on the
sandstone surface of the branches. Nevertheless, as zooids they remained unconvincing and it
remains to be established what surfaces or regions of the frond are exposed in these Ediacaran
fossils.

The differences between C. oppositus and the other species of Charniodiscus also pertain to those
with Thaumaptilon. Most notable are the alternate insertion of the primary branches in C. arboreus
and the more elongate form of the eponymous C. longus. Despite being the type species relatively
little is known of C. concentricus because of its poor preservation. It may differ from Thaumaptilon
in having a more prominent discoidal holdfast (see Jenkins and Gehling 1978, fig. 4), but
C. concentricus has possibly significant similarities with Thaumaptilon in the disposition of the
primary branches, as well as possessing faint transverse striations that in spacing and arrangement
resemble those of the Burgess Shale taxon.

Comparisons with other Ediacaran frond-like fossils may be kept brief. Glaessnerina has been
illustrated on the basis of few specimens, none complete, and in the opinion of Jenkins (1992, p. 162)
remains an ‘enigmatic genus’. It is characterized by the primary branches meeting along the midline
as a zig-zag commissure that obscures the axis. Jenkins and Gehling (1978, p. 353) indicated,
however, that this juxtaposition of branches could be a post-mortem feature. Another difference
with Thaumaptilon are the very prominent secondary branches. Vaizitsinia, from the White Sea
region of north-west Russia, has not received extensive documentation. Existing descriptions
portray a frond with a more rhomboidal outline than Thaumaptilon, but the slightly inflated
holdfast and prominent primary branches with transverse striations (Fedonkin 1985, pi. 14, fig. 4;
see also Fedonkin 1983, fig. 39) are similar to the Burgess Shale example.

CONWAY MORRIS: CAMBRIAN EDI ACAR AN-LIKE FOSSILS

623

There are a number of other frond-like fossils from Ediacaran faunas that either require more
detailed documentation before extensive comparisons can be drawn with Thaumaptilon (e.g.
Khatyspytia grandis Fedonkin 1985; note that the only description of this new taxon occurs in the
plate description) or appear to differ more extensively with the Burgess Shale example. Foremost
in this respect is Charnia masoni (e.g. Ford 1958; Fedonkin 1981, 1985), which is characterized by
very well-developed secondary branches (at right angles to the primary branches) and also
discernible tertiary branches (Fedonkin 1981 ; Jenkins 1985). Note also that Runnegar (1992, p. 76;
see also Germs 1973, p. 5) proposes that Glaessnerina is synonymous with Charnia. Unpublished
observations from the Mistaken Point assemblage of the Avalon peninsula, Newfoundland also
demonstrate that Charnia possessed a circular holdfast (see also Jenkins 1984, 1985, 1992) similar
to that of Charniodiscus. Although Charniodiscus , Charnia and Rangea had been linked with a
number of other Ediacaran forms in a ‘Hypothetical phylogeny of Precambrian frond-like
octocorals’ by Jenkins (1985, fig. 8; see also Jenkins 1984, p. 100, text-fig. 4), subsequently this
worker (1989) emphasized differences by allying Rangea and Charnia in the Rangeomorpha, and
Charniodiscus and Glaessnerina in an unspecified higher taxon. Of the Rangeomorpha Jenkins
(1989, p. 313) noted ‘while they were evidently of cnidarian grade, they show little direct similarity
to living Cmdaria’, whereas the latter genera remained within the pennatulaceans. The differences
between the rangeomorphs and taxa such as Charniodiscus and Thaumaptilon are self-evident from
the literature (e.g. Fedonkin 1985; Jenkins 1985), but in my opinion the case for removing the
rangeomorphs from the pennatulaceans is not compelling. Apart from the general arrangement of
fronds, stalk and holdfast, the differences in branch arrangement would seem to be no more
remarkable than the analogous case of zooid configurations in extant scleractinians. Whether the
concept of the Charniidae Glaessner, 1979« is adequate to encompass all these forms is uncertain,
but with the higher taxonomy of Ediacaran fossils in a continuing state of flux this family is
retained.

Among other frond-like Ediacaran taxa attention should be drawn to Paracharnia , but the
specimens are difficult to interpret from the available photographs (Ding and Chen 1981 ; Sun 1986).
Paracharnia appears to be distinct because of the numerous leaf-like extensions that arise from the
main frond (see also Dorzhnamzhaa and Gibsher 1990 for a record of Paracharnia from Mongolia).
A number of other frond-like fossils (Archangelia valdaica , Vaveliksia veliknanovi and Zolotvtsia
biserialis , e.g. Fedonkin 1985) remain rather poorly known (see also Gehling 1988, fig. 7a).

As Thaumaptilon (Text-fig. 3) is evidently closely related to a number of Ediacaran frond-like
taxa, especially Charniodiscus , Vaizitsinia and Charnia , the next point to address is whether the
similarity to known pennatulaceans is significant. The evidence for the Charniodiscus-Thaumaptilon
complex being true pennatulaceans is based on three principal suppositions: overall shape, probable
presence of autozooids, and evidence for a system of internal canals similar to extant pennatulaceans
(e.g. Brafield 1969). In disputing the widely-held view that the Ediacaran fronds are pennatulaceans
(e.g. Glaessner and Wade 1966; Jenkins and Gehling 1978; Runnegar 1982; Jenkins 1989, 1992),
some emphasis has been given by critics of this scheme (e.g. Seilacher 1984, 1989) to the
arrangement of the branches: fused to a common membrane in Ediacaran taxa, but separate in taxa
such as Pennatula. While it remains true that no exact counterpart to extant pennatulaceans exists
among the Ediacaran fronds, it is also clear that the diversity of the former is very considerable and
includes forms far removed from taxa generally regarded as typical, such as Pennatula. In this
context one could draw attention to forms such as Renilla , where the polyps arise from a flattened
surface, and various club-like forms (e.g. Veretillum, Sclerobelemnon). Whether the putative polyps
in Thaumaptilon really are autozooids remains tentative, but their distribution and size (they are
assumed to be retracted) are consistent with this interpretation. The much smaller siphonozooids
are unlikely to be recognizable in fossils, but not withstanding the typical relatively coarse-grained
lithology of Ediacaran preservation (see above), it is possible that putative autozooids will be
identified.

The significance of the arcuate band, reflectively preserved on the holdfast of the holotype, is
uncertain. There is a suggestion of a pustule-like structure, but as this region was presumably

624

PALAEONTOLOGY, VOLUME 36

embedded in the sediment it may be thought unlikely that the pustules represent autozooids. In the
literature, however, there are reports of siphonozooids on the basal bulbous holdfast of the
pennatulacean Umbellula carpenteri (see Hickson 1916, p, 8).

Accepting the Charniodiscus-Thaumaptilon complex as pennatulaceans (see also Norris 1989), it
remains to be established what relationships might exist with other Ediacaran taxa as well as
Phanerozoic fossils assigned to this group. Apart from the Ediacaran taxa mentioned above, the
closest relations of this complex may he with Ranged. This is a distinctive form with three (or more)
fused fronds arising from a common base with a bulbous holdfast (see Jenkins 1985, 1989, 1992 for
further discussion).

The Phanerozoic fossil record of pennatulaceans is largely composed of isolated spicules, which
may extend to the Lower Cambrian (e.g. Bengtson et al. 1990) and more occasional calcified axes
(e.g. Kuz'micheva 1980; Malecki 1982; Kocurko 1988). While soft-bodied remains of other types
of octocoral are known (e.g. Ruedemann 1916; Glinski 1956; Hamilton 1958; Sass and Rock 1975;
Lindstrom 1978), comparable fossils of pennatulaceans seem to be exceedingly rare. Tremblay
(1941) described a possible example from Quebec, apparently from the Potsdam Sandstone (Upper
Cambrian). The supposed examples from the Tertiary of Trinidad and the south-eastern Moluccas
(Bayer 1955) were reinterpreted as trace fossils (Hantzschel 1958, 1975). Hantzschel (1958) also
discussed a number of other putative octocoral fossil remains and concluded they represented
various types of trace fossil.

Possible relatives of Emmonsaspis

Concerning the comparison of Emmonsaspis with Ediacaran taxa, there is more uncertainty. The
branching arrangement, imparting the characteristic chevron pattern, recalls approximately a
number of Ediacaran fronds. Nevertheless, there is an absence of evidence for either second
branches, a stalk or a holdfast. It is tentatively proposed that comparisons might be better drawn
to Pteridinium , a widespread taxon that consists of a more or less inflated body composed of
repeated transverse units (e.g. Glaessner 1963; Glaessner and Wade 1966; Pflug 1970; Fedonkin
1981, 1985; Gibson et al. 1984; Jenkins 1989; see also Narbonne and Aitken 1990). To my
knowledge there are no other fossils which resemble Emmonsaspis more closely. Nevertheless, the
comparisons are by no means precise. According to Pflug (1970) the body of Pteridinium consists
of three vanes (see also Jenkins 1992), whereas so far as can be told Emmonsaspis had a foliate body.
In addition, the ribs of Pteridinium are more siculate in shape and meet along a well-defined axial
groove.

Possible relatives of Mackenzia

Some Ediacaran faunas contain peculiar bag-like structures, especially the ernietiids, whose
systematic relationships remain uncertain. In seeking a possible relative to Mackenzia among the
Ediacaran assemblages, one possibility is a comparison with Platypholinia pholiata (e.g. Fedonkin
1985, pi. 19, figs 5-6). This is a rare sac-shaped organism, only known from the White Sea region
of north-east Russia. Platypholinia is rather featureless, but there is some evidence for an oral
opening. Fedonkin’s (1985) assignment of this taxon to the platyhelminths is considered unlikely.
It differs from Mackenzia in being smaller, less elongate, and in lacking obvious external folds.
Nevertheless, a relationship between the taxa is possible. Although superficially seemingly different,
comparisons might also be drawn between Mackenzia and the putative actinian Inaria karli ,
documented by Gehling (1988) from the Pound subgroup (Rawnsley Quartzite) of the Flinders
Ranges, South Australia. As reconstructed the most striking divergence is the proximal expansion
of Inaria , most noticeably in adults, into lobate expansions that give the animal a vague
resemblance to a garlic clove. From this region the polyp extends into a more cylindrical distal
portion, reconstructed as having a simple opening without tentacles, although Gehling (1988) does
not rule out their absence by non-preservation. In addition Inaria is inferred to have had internal
mesenteries with muscular strands. The similarities between Inaria and Mackenzia , including simple
oral opening, lobate walls, and internal mesenteries, suggest a fairly close relationship. Another

CONWAY MORRIS: CAMBRIAN EDIACARAN-LIKE FOSSILS

625

Ediacaran fossil that might be compared with Mackenzia is the only known specimen of
Protechiurus edmondsi from the Kuibis Quartzite, Namibia. Glaessner (19796) interpreted the
sandstone cast, which is approximately cigar-shaped and bears seven or eight longitudinal ridges,
as an early echiuroid worm. The evidence for this assignment is regarded as slender, but in shape
and dimensions the fossil accords with how a Mackenzia- like individual would appear if its central
cavity was infilled with fine sand. Mention should also be made of Ediacaran fossils such as
Beltanelliformis , which Narbonne and Hofmann (1987) synonymize with Nemiana and more
tentatively Beltanelloides. The specimens tend towards a bowl-like shape and are often found in
abundance and closely spaced. They have been generally interpreted as sedentary cnidarians (e.g.
Fedonkin 1985), although Narbonne and Hofmann (1987; see also Jenkins 1992) reconstructed
them as globular bodies of uncertain systematic position.

In passing Jenkins (1989, p. 311) noted that some specimens of what appear to be Eoporpita ,
widely interpreted as a chondrophorine (e.g. Wade 19726) showed structures that suggested they
could ‘be reasonably reconstructed as actinians’. In 1992, however, Jenkins (p. 161) returned to a
comparison with chondrophorines, although not abandoning his proposed link with the actinians.
This author (Jenkins 1989, fig. 2a-c) also illustrated a putative actinian (possibly Medusinites
asteroides; see also Jenkins 1992) with evidence of tentacles, noting that further study was required.
In addition, Jenkins (1989) proposed that other scyphozoan-like fossils (e.g. Ediacaria) had modes
of life similar to actinians. All these taxa require further study (see Jenkins 1992), and while they
differ from Mackenzia in the possession of what appear to be tentacles, they may belong to the same
clade of anemone-like anthozoans.

The wider relationships of Mackenzia and Inaria , and perhaps Platypholinia and Protechiurus , are
uncertain. They are regarded as being of a cnidarian grade, and may be most closely related to the
anthozoan actinians (see also Gehling 1988). Concerning the apparent absence of oral tentacles a
possibly significant comparison (see also Gehling 1988) might be drawn with the Recent actinians
Limnactinia nuda and L. laevis (Calgren 1921, 1927). Unlike other species of Limnactinia , these
species lack oral tentacles, but have nematocysts scattered over the exterior surface that form dense
batteries in the oral region. The resemblances between Limnactinia and Mackenzia are quite
striking, although in contrast to the Burgess Shale taxon, Limnactinia was evidently embedded m
the marine muds of its cold-water habitat. In the ceriantharian anemones a number of taxa are
known to lack tentacles. These include Anactinia , a small anemone that is interpreted as pelagic,
from the Bay of Bengal (Annandale 1909), although this genus and the related Par actinia are
probably larval forms (see Tiffon 1987). In a related context attention should also be drawn to the
scleractinian coral Leptoseris fragilis. This species is also remarkable in lacking oral tentacles but
here heterotrophic feeding is evidently facilitated by flagella (Schlichter 1991). Thus, if Mackenzia
were to be compared with atentaculate cnidarians, then L. fragilis might provide some revealing
similarities in terms of functional morphology. In other respects, however, this latter species differs
considerably from Mackenzia , but as discussed below its organization might provide analogies to
Ediacaran organisms.

Possible relatives of Gelenoptron and the ? chondrophorine

Concerning the fossil record, the study of Phanerozoic chondrophorines has undergone a minor
renaissance with a variety of fossils so assigned (see Stanley 1986 for a recent review) to complement
an existing roster (see Harrington and Moore 1956). While some of these assignations appear to be
reasonable, others are considerably more suspect (see Conway Morris 1989c, p. 346; Conway
Morris et al. 1991 ; Landing and Narbonne 1992) and discovery of an annulated disc alone may be
insufficient to place a given fossil in the chondrophorines.

In the context of this Burgess Shale fossil, however, particular interest devolves on putative
chondrophorines from the Cambrian, especially Burgess Shale-type faunas, and their descent from
taxa present in the Ediacaran assemblages. The status of the supposed chondrophorine Rotadiscus
grandis (Sun and Hou, 1987) from the Chengjiang fauna is discussed above, and although its
affinities are uncertain, a place within the Cnidaria, let alone the chondrophorines, seems unlikely.

626

PALAEONTOLOGY, VOLUME 36

Later, Chen et al. (1989, p. 271) mentioned another fossil which ‘bears a strong resemblance to
modern Porpita ’, but this is yet to be illustrated and discussed. Within the Ediacaran assemblages
reasonably convincing pneumatophores are present in the form of the annulated discs of
Ovatoscutum , known from both South Australia (Glaessner and Wade 1966; Jenkins 1984, 1989,
1992) and the White Sea area of Russia (Fedonkin 1984, 1985). Nothing is known of the soft-parts
which were associated with Ovatoscutum. Other putative chondrophorines from Ediacaran beds are
Eoporpita (Wade 19726; Jenkins 1984; Narbonne and Aitken 1990, but see Jenkins 1989, 1992) and
Kidlingia (Foyn and Glaessner 1979; see also Narbonne and Hofmann 1987 ; Narbonne and Aitken
1990). The former is associated with structures that are interpreted as zooids (but see Norris 1989),
while the latter genus has also been found in strata very close to the Precambrian-Cambrian
boundary in the Burin Peninsula, southeast Newfoundland (Narbonne et al. 1991). Chondroplon
was regarded by Wade (1971) as another chondrophorine, although Hofmann (1988) reinterpreted
the only known specimen as an example of Dickinsonia (but see Jenkins 1989, p. 31 1 ; 1992, p. 161),
a fossil which is widely interpreted as a putative annelid worm (Glaessner and Wade 1966; Wade
1972tf; Runnegar 1982). Of all these fossils, the ?chondrophorine appears to approach most closely
Kullingia , albeit differing in being considerably smaller (one Ediacaran specimen of 1 Kidlingia from
the Wernecke Mountains, Yukon is regarded as small with a diameter of 54 mm (Narbonne and
Hofmann 1987, p. 659)). Gelenoptron tentaculatum has no clear similarities to known Ediacaran
chondrophorines, but is included here because of its possible relationships to the hydrozoans.

The status of the Vendozoa

The hypothesis that the Ediacaran fossils represent a distinctive assemblage, perhaps entirely
separate from the Metazoa, whose basic body-plan has been compared to a series of chambered
units (Seilacher 1984, 1989, 1992; Bergstrom 1989, 1990) has attracted wide attention (e.g. Gould
1984, 1989). Nevertheless, it may be fundamentally incorrect, at least so far as the entire fauna is
concerned. Re-emphasis of the metazoan status of a number of the fossils has come from Gehling
(1991) and Runnegar (1992), while other workers (e.g. Jenkins, 1984, 1985, 1989, 1992) have
persisted with such assignments during the time of controversy.

In the context of this paper there are three items that bear on this controversy, and cast doubt
on the vendobiontan hypothesis, at least in its all-embracing form promulgated by Seilacher (1992).

First, there seems reasonable evidence that Thaumaptilon walcotti is related to Ediacaran taxa
such as Charniodiscus. There is also evidence that T. walcotti can be compared with pennatulaceans.
Most significant are the putative autozooids, whose distribution and size are consistent with the
pennatulaceans. In addition, although the foliate structure finds no exact parallel among extant
pennatulaceans, the overall arrangement of central axis and branches with connecting canals is
certainly similar. It is true that Recent pennatulaceans do not appear to have an exact counterpart
to the discoidal holdfasts of some Ediacaran types, such as Charniodiscus (see also Jenkins 1992).
However, the holdfasts of the former are quite often bulbous, and Hickson (1916) mentioned that
in one form the so-called end-bulb was thin-walled and showed a spherical to oval expansion. In
addition Kiikenthal and Broch (1911, pi. 15, fig. 11) illustrated a specimen of Kophohelemnon
heterospinosum with its holdfast inflated to form a ball-like structure, and similar structures are
depicted by Nutting (1912) in various pennatulaceans. There has, moreover, been an over-emphasis
in comparing the Ediacaran fronds with genera such as Pennatula , Pteroides and Scytalium, and a
reluctance to include the wide variety of other pennatulacean forms in the discussion. This is not
to imply that other Recent taxa approach more closely in form the putative Ediacaran sea-pens, but
rather to emphasize that the pennatulaceans adopt a very wide variety of shapes.

Even if T. walcotti and its putative Ediacaran relatives are admitted into the pennatulaceans, this
leaves important questions unresolved. These include (a) the status of various other Ediacaran taxa
such as Chamia and Rangea (see Germs 1973; Jenkins 1984, 1985); (b) the possible presence of
Cambrian octocorals as recorded from spicular remains (e.g. Bengtson et al. 1990) and their
connection with the soft-bodied record; and (c) the phylogenetic relationships between extant
pennatulaceans and the fossil examples discussed here.

CONWAY MORRIS: CAMBRIAN EDI ACARAN-LIKE FOSSILS

627

Mackenzia costalis Walcott is also interpreted, albeit more tentatively, as a cnidarian. The
comparisons with Limnactinia nuda which lacks tentacles, but possesses abundant nematocysts that
presumably trap prey, suggests that a comparison between Mackenzia and actinians is not far-
fetched. Overall, the body form of Mackenzia with evidence for a folded transverse section, possibly
reflected by internal partitions, and an attachment disc would be consistent with a cnidarian grade.
Even if the phyletic status of Mackenzia remains uncertain, there is little doubt it is a metazoan.
Paramount is the indirect evidence of a musculature, as inferred from the contractability of
specimens, while the presence of possible food boluses is consistent with a digestive cavity. While
placement of Mackenzia in the cnidarians is considered viable, there is little direct evidence to
support earlier suggestions that it is an actinarian (Wells and Hill 1956), perhaps even belonging to
the Edwardsiidae (Clark 1913)

In passing it is worth reviewing other putative fossil actinarians, as they are exceedingly
uncommon. Three taxa are candidates. Pa/aeactinia holli is known from a single specimen in the
Upper Ordovician of New York (Ruedemann 1934, pi. 10, figs 1-3). Ruedemann (1934) provided
a detailed description, and I re-examined the specimen in 1991. The fossil shows an apparent
holdfast and broad stalk, but it is impossible to discern details such as tentacles in the distal zone.
An actinarian affinity, while reasonable, remains unproven. The status of two other claimants as
fossil actinians is more suspect. Pseudocaudina brachyura has been variously identified as a
holothurian (e.g. Broili 1926; Frizzell and Exline 1966), an actinarian (Heding 1932), or a possible
ctenophore (Ziegler 1991, who proposed that this taxon is a junior synonym of Laffonia Helvetica ;
see also Hess 1973, p. 650). Fossils from the Cambrian of France were described by Dollfus (1875)
as an actinarian ( IPalaeactis vetula ), and this was accepted tentatively by Wells and Hill (1956).
Hantzschel's (1975) reassignment of these structures to trace fossils, probably comparable to
Berguaeria , is accepted. In the context of this paper it is worth noting that ichnotaxa such as
Bergaueria are generally considered to have been made by semi-infaunal anemone-like cnidarians
(e.g. Arai and McGugan 1968, 1969; Alpert 1973). Both Mackenzia and its putative relative Inaria
were epifaunal, but Gehling (1988) depicts a hypothetical path to an effectively infaunal descendant
capable of making Bergaueria- like traces.

Vendozoans : are there modern analogues?

Although the scleractinian Leptoseris fragilis is probably less relevant to a detailed understanding
of Mackenzia , its anatomy is of particular interest because it demonstrates how a cnidarian
bodyplan may be modified in a manner that would be analogous to many Ediacaran taxa. L. fragilis
is also remarkable because although it inhabits water on the edge of the photic zone (up to 145 m)
it possesses symbiotic zooxanthellae (Schlichter 1991). Various workers have speculated that
Ediacaran taxa obtained nutrition from chemosymbiosis or photosymbionts. Such hypotheses are
difficult to test, but one avenue of research might be the search for characteristic molecular
biomarkers in adjacent sediments.

Leptoseris fragilis, however, does not rely exclusively on zooxanthellae, and during months of
reduced sunlight it is heterotrophic, utilizing both dissolved organic matter (DOM) and capturing
food particles (Schlichter 1991). The role of DOM versus abundance of particulate food in
Ediacaran faunas is speculative, although one might hypothesize that in the apparent absence or
rarity of advanced triploblastic suspension feeders such as brachiopods and echinoderms concen-
trations of suspended material were higher. The use of flagellae in L. fragilis to capture food particles
suggests a possible parallel to how many Ediacaran species may have fed. It remains possible, of
course, that true cnidarians were equipped with nematocysts, that could entangle or poison small prey.

Two other related aspects of L. fragilis (Schlichter 1991) may throw light on Ediacaran
anatomies. First, this living coral has a series of specialized gastric ducts into which the food is
channelled. This arrangement is perhaps not entirely dissimilar to the hypothesized mattress-like
construction of some Ediacaran fossils where the body is composed of a series of canals. The second
curious aspect of L. fragilis is the presence of pores (evidently also present in Limnactinia nuda , see
Calgren 1927, p. 7) employed for exhalant flow of water from the gastric ducts.

628

PALAEONTOLOGY, VOLUME 36

L. fragilis may provide no more than a number of interesting analogies to the possible
construction and functional morphology of Ediacaran bodyplans. It does, however, suggest that
many of the Ediacaran species can be interpreted as diploblasts, of a grade comparable to cnidarians
and ctenophores. Indeed, some Ediacaran taxa may be true cnidarians, including Charniodiscus and
its relatives.

It is concluded that the Vendozoan hypothesis fails for many Ediacaran taxa, although this is not
to deny that there are a number of enigmatic forms that could represent experiments in multicellular
construction independent of the metazoans, being derived from other protistan groups. Restoring
the supposed vendobiontans to the metazoans leaves unanswered the problem of the peculiar
taphonomic circumstances that permitted widespread soft-part preservation in environments that
in the Phanerozoic are seldom sites of Fossil-Conservation-Lagerstatten. This problem remains
unsolved and understudied. However, proposals such as the role of bacterial sealing (Seilacher
et al. 1985), possible absence of degrading enzymes such as collagenases (Runnegar 1992), absence
of scavengers, and limited bioturbation are all possible. Yet another proposal could be that while
mobile infaunal worms, only known from trace fossils (e.g. Glaessner 1969; Fedonkin 1981),
required a thin flexible cuticle, sessile forms developed a conspicuously thick and taphonomically
resistant cuticle, perhaps as a response to elevated levels of oxygen (see Derry et al. 1992).

Acknowledgements. I thank F. J. Collier (National Museum of Natural History, Smithsonian Institution,
Washington D.C.) for making available specimens. In Nanjing, Chen Junyuan and Sun Weiguo generously
gave access to specimens of Rotadiscus grandis, while E. Landing (Albany, New York) facilitated examination
of the paropsonemids and possible actinian. Examination of possible specimens of Dickinsonia and a
?cephalochordate from California was facilitated by J. Firby, J. W. Durham and J. Lipps (Berkeley,
California). The visit to China was under the auspices of a Royal Society-Academia Sinica Exchange Scheme,
while work in the Smithsonian has been assisted by grants from the Royal Society and St John’s College,
Cambridge. Comments and assistance by R. J. F. Jenkins, J. G. Gehling, P. Cornelius and an anonymous
referee are all appreciated. B. K. Harvey photographed a number of specimens, H. Alberti assisted with
drafting and S. J. Last typed numerous versions of the manuscript. This is Cambridge Earth Sciences
Publication 2528.

REFERENCES

aitken, J. R. and MciLREATH, I. a. 1990. Comment [In defense of the Escarpment near the Burgess Shale fossil
locality]. Geoscience , Canada , 17, 111-116.

alpert, s. p. 1973. Bergaueria Prantl (Cambrian and Ordovician), a probable actinian trace fossil. Journal of
Paleontology, 47, 919-924.

— 1976. Trilobite and star-like trace fossils from the White-Inyo Mountains, California. Journal of
Paleontology, 50, 226-239.

annandale, n. 1909. A pelagic sea-anemone without tentacles. Records of the Indian Museum, 3, 157-162.
arai, M. N. and McGUGAN, a. 1968. A problematical coelenterate(?) from the Lower Cambrian, near Moraine
Lake, Banff area, Alberta. Journal of Paleontology, 42, 205-209.

— 1969. A problematical Cambrian coelenterate(7). Journal of Paleontology, 43, 93-94.
bayer, F-. m. 1955. Remarkably preserved fossil sea-pens and their Recent counterparts. Journal of the
Washington Academy of Sciences , 45, 294—300.

bednarczyk, w. 1970. Trilobite fauna of the Lower Paradoxides oelandicus stage from the Brzechow area in
the western part of the Swietokryskie Mts. Bulletin of the Academy of Polish Sciences, Scientific Series,
Geology and Geography, 18, 29-35.

bengtson, s. 1992. Proterozoic and earliest Cambrian skeletal metazoans. 1017-1033. In schopf, j. w. and
klein, c. (eds). The Proterozoic biosphere. A multidisciplinary study. Cambridge University Press, Cambridge,
xxiv-l- 1348 pp.

— , conway morris, s., cooper, B., jell, p. a. and runnegar, B. 1990. Early Cambrian fossils from South
Australia. Memoirs of the Association of Australasian Palaeontologists, 9, 1-364.

— , fedonkin, m. a. and lipps, j. h. 1992. The major biotas of Proterozoic to early Cambrian multicellular
organisms. 433 — 435. In schopf, j. w. and klein, c. (eds). The Proterozoic biosphere. A multidisciplinary study.
Cambridge University Press, Cambridge, xxiv + 1348 pp.

CONWAY MORRIS: CAMBRIAN EDIACARAN-LIKE FOSSILS

629

benus, A. 1988. Sedimentological context of a deep-water Ediacaran fauna (Mistaken Point Formation, Avalon
Zone, Eastern Newfoundland). Bulletin of the New York State Museum , 463, 8-9.
bergstrom, J. 1989. The origin of animal phyla and the new phylum Procoelomata. Lethaia , 22, 259-269.

— 1990. Precambrian trace fossils and the rise of bilateran animals. Ichnos , 1, 3-13.

birket-smith, s. J. R. 1981. A reconstruction of the Pre-Cambrian Spriggina. Zoologische Jahrbuch , Anatomie
unci Ontogenie der Tiere , 105, 237-258.

borovikov, l. i. 1976. First find of fossil Dickinsonias in Lower Cambrian sediments in the USSR. Doklady
Akademii nauk SSSR , 231, 53-55. [AGI English translation.]
brafield. a. E. 1969. Water movements in the pennatulid coelenterate Pteroides griseum. Journal of Zoology,
London , 158, 317-325.

brasier, m. d. 1979. The Cambrian radiation event. 103-159. In house, m. r. (ed.). The origin of major
invertebrate groups, Systematics Association Special Volume 12. Academic Press, London, x + 515 pp.

— 1989. On mass extinction and faunal turnover near the end of the Precambrian. 73-88. In donovan,
s. k. (ed.). Mass extinctions, processes and evidence. Belhaven, London, xiv + 266 pp.

briggs, d. e. G. 1976. The arthropod Branchiocaris n. gen., Middle Cambrian, Burgess Shale, British
Columbia. Bulletin of the Geological Survey of Canada, 264, 1-29.

— 1977. Bivalved arthropods from the Cambrian Burgess Shale of British Columbia. Palaeontology, 20,
595-621.

— 1979. Anomalocaris, the largest known Cambrian arthropod. Palaeontology, 22, 631-664.

broili, f. 1926. Eine Holothurie aus dem oberen Jura von Franken. Sitzungsberichte der mathematisch-
naturwissenschaftlichen Abteilung der Bayerischen Akademie der Wissenschaften zu Miinchen, 1926, 341-351.
calgren, o. 1921. Actiniaria. Part 1. The Danish Ingolf Expedition, 5, 1-241.

— 1927. Actiniaria and Zoantharia. Further Zoological Results of the Swedish Antarctic Expedition
1901-1903, 2 (3), 1-102.

cave, l. d. and simonetta, A. M. 1975. Notes on the morphology and taxonomic position of Aysheaia
(Onycophora?) and of Skania (undetermined phylum). Monitore zoologico italiano (N.S.), 9, 67-81.
chamisso, a. de and eysenhardt, c. G. 1821 De Animalibus quibusdam e classe vermium Linneana, in
circumnavigatione terrae, auspicante Comite N. Romanzoff, duce Ottone de Kotzebue, annis 1815-1818.
Fasciculus secundus, reliquos vermes continens. Nova Acta Physico-Medica Academiae Caesareae
Leopoldino-Carolinae , 10, 345-374.

chapman, F. 1926. New or little known fossils in the National Museum. Part XXX. A Silurian jelly-fish.

Proceedings of the Royal Society of Victoria, New Series 39, 13-17.
chen junyuan, hou xianguang and erdtmann, b. d. 1989. New soft-bodied fossil fauna near the base of the
Cambrian system at Chengjiang, eastern Yunnan, China. Chinese Academy of Sciences, Developments in
Geosciences, Contribution to the International Geological Congress, 1989, Washington, DC, USA, 265-277 .
Clark, a. h. 1913. Cambrian holothurians. American Naturalist , 47, 488-507.

Clark, h. l. 1912. Fossil holothurians. Science, 35, 274—278.

Clarke, j. m. 1900. Paropsonema cryptophya, a peculiar echinoderm from the intumescens zone (Portage beds)
of western New York. Bulletin of the New York State Museum, 39, 172-178.
cloud, p. 1973. Pseudofossils: a plea for caution. Geology , 1, 123-127.

— and bever, j. f. 1973. Trace fossils from the Flathead Sandstone, Fremont County, Wyoming, compared
with early Cambrian forms from California and Australia. Journal of Paleontology, 47, 883-885.

— and nelson, c. a. 1966. Phanerozoic-Cryptozoic and related transitions: new evidence. Science, 154,
766-770.

— 1967. Pteridinium and the Precambrian-Cambrian boundary. Science, 157, 958.
compston, w., williams, i. s., kirschvink, j. l., zhang zhichao and ma guoya. 1992. Zircon U-Pb ages for
the early Cambrian time-scale. Journal of the Geological Society of London , 149, 171-184.
conway morris, s. 1978. Laggania cambria Walcott, a composite fossil. Journal of Paleontology, 52, 126-131.

— 1979. The Burgess Shale (Middle Cambrian) fauna. Annual Review of Ecology and Systematics, 10,
327-349.

— 1985. The Ediacaran biota and early metazoan evolution. Geological Magazine, 122, 77-81.

— 1989a. Early metazoans. Science Progress, OxJ'ord , 73, 81-99.

— 19896. The persistence of Burgess Shale-type faunas: implications for the evolution of deeper-water
faunas. Philosophical Transactions of the Royal Society of Edinburgh. Earth Sciences, 80, 271-283.

— 1989c. Burgess Shale faunas and the Cambrian explosion. Science, 246, 339-346.

1990. Late Precambrian and Cambrian soft-bodied faunas. Annual Review of Earth and Planetary
Sciences, 18, 101-122.

630

PALAEONTOLOGY, VOLUME 36

Conway morris, s. 1992. Burgess Shale-type faunas in the context of the ‘Cambrian explosion’: a review.
Journal of the Geological Society of London, 149, 631-636.

and robison, r. a. 1982. The enigmatic medusoid Peytoia and a comparison of some Cambrian biotas.
Journal of Paleontology, 56, 116-122.

— 1988. More soft-bodied animals and algae from the Middle Cambrian of Utah and British
Columbia. University of Kansas Paleontological Contributions, Paper, 122, 1-48.

— , savoy, L. E. and Harris, a. g. 1991. An enigmatic organism from the ‘Exshaw’ Formation
(Devonian-Carboniferous), Alberta, Canada. Lethaia, 24, 139-152.

— and Whittington, h. b. 1979. The animals of the Burgess Shale. Scientific American, 241 (1), 122-133.
crimes, t. p., legg, i., marcos, a. and arboleya, m. 1977. ?Late Precambrian-low Lower Cambrian trace

fossils from Spain. 91-138. In crimes, t. p. and harper, j. c. (eds). Trace fossils 2, Geological Journal Special
Issue 9. Seel House Press, Liverpool, vii-|-351 pp.
croneis, c. and McCormack, j. 1932. Fossil Holothuroidea. Journal of Paleontology, 6, 111-148.
derry, L. a., kaufman, a. J. and Jacobsen, s. b. 1992. Sedimentary cycling and environmental change in the
late Proterozoic: evidence from stable and radiogenic isotopes. Geochimica et Cosmochimica Acta, 56,
1317-1329.

ding qixiu and chen yiyuan 1981. [Discovery of soft metazoan from the Sinian System along eastern Yangtze
Gorge, Hubei]. Journal of the Wuhan College of Geology, 1981 (2), 53-57. [In Chinese with English abstract].
dollfuss, G. 1875. Note sur des empreintes attribuables a une actinie (IPalaeactis vetula ) dans les schistes
cambriens des moitiers-d’Allonne. Memoires de la Societe Nationale des Sciences Naturelles et Mathematiques
de Cherbourg, 19, 224-232.

dorzhnamzhaa, d. and gibsher, a. s. 1990. [New information on the stratigraphic position of the
Vendian-Ediacaran fauna of Mongolia], 28. In repina, l. n. and zhuravelev, a. yu. (eds). Abstracts of the
Third International Symposium on the Cambrian System ( 1-9 August 1990, Novosibirsk, USSR). Institute of
Geology and Geophysics, Novosibirsk, 179 pp. [In Russian],

Durham, J. w. 1 97 1 <3. Some late Precambrian and early Cambrian fossils from the White-Inyo Mountains of
California. Geological Society of America, Abstracts with Program, 3, 114-115.

— 19716. The fossil record and the origin of the Deuterostomata. Proceedings of the North American
Paleontological Convention 1969, Part H, 1101-1132.

dzik, j. 1991. Is fossil evidence consistent with traditional views of early metazoan phylogeny? 47-56. In
simonetta, a. m. and conway morris, s. (eds). The early evolution of Metazoa and the significance of
problematic taxa. Cambridge University Press, Cambridge, ix + 296 pp.

— and krumbiegel, g. 1989. The oldest ‘onychophoran’ Xenusion: a link connecting phyla? Lethaia, 22,
169-182.

ehrenberg, h. 1834. Beitrage zur physiologischen kenntnis der Corollenthiere im allgemenen, und besonders
des rothen Meeres, nebst einem Versuche zur physiologischen Systematik derselben. Physikalische
Abhandlungen der Kdniglichen Akademie der Wissenschaften zu Berlin, 1832, 225 — 438.

Emmons, E. 1855. American geology, containing a statement of the principles of the science, with full illustrations
of the characteristic American fossils. With an atlas and a geological map of the United States. Volume 1, parts
I and II. Sprague, Albany, xvi + 194 + 232 pp.

fedonkin, M. a. 1981. [White Sea biota ofVendian (Precambrian non-skeletal fauna of the Russian Platform
North)]. Akademiya Nauk SSSR, Trudyi, 342, 1-100. [In Russian].

— 1983. [Vendian organisms], Itogi nauka i techniki, stratigrafiya, paleontologiya, 12, 1-128. [In Russian].

— 1984. [Promorphology of Vendian Radialia]. 30-58. In Ivanovskiy, a. b. and ivanov, i. b. (eds).
Stratigraphy and palaeontology of the earliest Phanerozoic. Izdatel’stvo Nauka, Moscow, 127 pp. [In
Russian].

1985. [Systematic description ofVendian Metazoa]. 70-106. In sokolov, b. s. and iwanowski, a. b. (eds).
The Vendian system. Volume 1. Paleontology. Nauka, Moscow, 221 pp. [In Russian, English translation
published by Springer-Verlag, Berlin, 1990, pp. 71-120],

1987. [Soft-bodied Vendian fauna and its place in metazoan evolution]. Akademiya nauk SSSR, Trudyi
Paleontologicheskogo Instituta, 226, 1-176. [In Russian].
firby, j. b. 1972. Possible cephalochordates from the Lower Cambrian. Geological Society of America,
Abstracts with Program, 4, 504.

and Durham, j. w. 1974. Molluscan radula from earliest Cambrian. Journal of Paleontology, 48,
1109-1119.

ford, t. d. 1958. Pre-Cambrian fossils from Charnwood Forest. Proceedings of the Yorkshire Geological
Society, 31, 211-217.

CONWAY MORRIS: CAMBRIAN EDIACARAN-LIKE FOSSILS

631

— 1963. The pre-Cambrian fossils of Charnwood Forest. Transactions of the Leicester Literary and
Philosophical Society , 57, 57-62.

foyn, s. and glaessner, M. F. 1979. Platysolenites, other animal fossils and the Precambrian-Cambrian
transition in Norway. Norsk Geologisk Tidsskrift, 59, 25 46.
fritz, w. h. 1990. Comment [In defense of the Escarpment near the Burgess Shale fossil locality]. Geoscience,
Canada , 17, 106-1 10.

— 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 , 83-12, 1-24.

frizzell, D. L. and exline, h. 1966. Holothuroidea - fossil record. U646-U672. In moore, r. c. (ed.). Treatise
on invertebrate paleontology. Part U. Echinodermata 3. Geological Society of America and University of
Kansas Press, Boulder, Colorado and Lawrence, Kansas, 695 pp.
fuchs, T. 1905. Ueber Parapsonema cryptophysa [.vie] Clarke und deren Stellung im System. Centralblatt fur
Mineralogie, Geologie und Palaontologie, Jahrgang , 1905, 357-359.
gehling, J. G. 1988. A cnidarian of actinian grade from the Ediacaran Pound Subgroup, South Australia.
Alcheringa , 12, 299-314.

— 1991. The case for Ediacaran fossil roots to the metazoan tree. Memoirs of the Geological Society of India,
20, 181-224.

germs, g. j. b. 1973. A reinterpretation of Rangea schneiderhoehni and the discovery of a related new fossil from
the Nama Group, South West Africa. Lethaia, 6, 1-10.

Gibson, G. G., teeter, s. a. and FEDONKIN, M. a. 1984. Ediacaran fossils from the Carolina slate belt, Stanly
County, North Carolina. Geology, 12, 387-390.

glaessner, M. F. 1963. Zur Kenntnis der Nama-Fossilien Siidwest-Afrikas. Annalen des Naturhistorischen
Museums in Wien, 66, 113-120.

— 1968. Biological events and the Precambrian time scale. Canadian Journal of Earth Sciences, 5, 585-590.
1969. Trace fossils from the Precambrian and basal Cambrian. Lethaia , 2, 369-393.

1979a. Precambrian. A79-A118. In robison, r. a. and teichert, c. (eds). Treatise on invertebrate
paleontology. Part A. Introduction. Geological Society of America and University of Kansas Press, Boulder,
Colorado and Lawrence, Kansas, 569 pp.

19796. An echiurid worm from the late Precambrian. Lethaia, 12, 121-124.

— 1980. Parvancorina - an arthropod from the Late Precambrian (Ediacaran) of South Australia. Annalen
des Naturhistorischen Museums in Wien, 83, 83-90.

— 1984. The dawn of animal life. A biohistorical study. Cambridge University Press, Cambridge, xi + 244 pp.

— and daily, b. 1959. The geology and late Precambrian fauna of the Ediacara fossil reserve. Records of the
South Australian Museum, 13, 369-401.

— and wade, m. 1966. The late Precambrian fossils from Ediacara, South Australia. Palaeontology, 9,
599-628.

glinski, a. 1956. Plumalina conservata n. sp. (Gorgonaria) aus dem Mittel-Devon der Eifel. Senckenbergiana
Lethaea, 37, 53-57.

gould, s. J. 1984. The Ediacaran experiment. Natural History , 93 (2), 14-22.

— 1989. Wonderful life. The Burgess Shale and the nature of history. Norton, New York, 347 pp.

hall, j. 1847. Palaeontology of New-York. Volume 1. Containing descriptions of the organic remains of the lower
division of the New- York system ( equivalent of the Lower Silurian rocks of Europe). C. van Benthuysen,
Albany, xxiii + 338 pp.

Hamilton, d. 1958. An Eocene sea-pen from Dunedin, New Zealand. Palaeontology, 1, 226-230.
hantzschel, w. 1958. Oktokoralle oder Lebensspur? Mitteilungen aus dem Geologischen Staatsinstitut in
Hamburg, 27, 77-87.

— 1975. Trace fossils and problematica. In teichert, c. (ed.). Treatise on invertebrate paleontology. Part W
(Revised). Miscellanea, 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. Siphonophorida. F145-F152. In moore, r. c. (ed.). Treatise on
invertebrate paleontology. Part F. Coelenterata. Geological Society of America and University of Kansas
Press, Boulder, Colorado and Lawrence, Kansas, 498 pp.
hatschek, b. 1888. Lehrbuch der Zoologie. Eine morphologische Ubersicht der Thierreiches zur Einfuhrung in das
Studium dieser Wissenschaft. Erste Liefrung. Jena, Fischer, iv + 304 pp.
heding, s. G. 1932. Paracaudina nom. nov. a correction, together with some remarks concerning the supposed
fossil holothurian Pseudocaudina brachyura Broili. Videnskabelige Meddelelser fra Dansk naturhistorisk
Forening i Kobenhavn , 92, 455-456.

632

PALAEONTOLOGY, VOLUME 36

hertwig, R. 1882. Report on the Actiniaria dredged by H.M.S. Challenger, during the years 1873-1876. Report
on the scientific results of the voyage of H.M.S. Challenger during the years 1873-76 under the command of
Captain George S. Nares, R.N., F.R.S. and Captain Frank Tourle Thomson , R.N., (Zoology), 7, 1-136.
Hess, h. 1973. Neue Echinodermenfunde aus dem mittleren Dogger des Aargauer Juras. Eclogae Geologicae
Helvetiae , 66, 625-656.

hickson, s. J. 1916. The Pennatulacea of the ‘Siboga’ Expedition. Siboga-Expeditie, 14, i-x, 1-135.
hofmann, J. J. 1988. An alternative interpretation of the Ediacaran (Precambrian) chondrophore Chondroplon
Wade. Alcheringa , 12, 315-318.

— , narbonne, G. m. and AITKEN, J. d. 1990. Ediacaran remains from intertillite beds in northwestern
Canada. Geology , 18, 1199-1202.

hou xianguang, ramskold, L. and Bergstrom, j. 1991. Composition and preservation of the Chengjiang
fauna -a Lower Cambrian soft-bodied fauna. Zoological Scripta , 20, 395-411.
jenkins, R. J. F. 1984. Interpreting the oldest fossil cnidarians. Palaeontographica Americana, 54, 95-104.
1985. The enigmatic Ediacaran (late Precambrian) genus Rangea and related forms. Paleobiology, 11,
336-355.

— 1989. The ‘supposed terminal Precambrian extinction event’ in relation to the Cnidaria. Memoirs of the
Association of Australasian Palaeontologists, 8, 307-317.

— 1992. Functional and ecological aspects of Ediacaran assemblages. 131-176. In lipps, j. h. and signor,
p. w. (eds). Origin and early evolution of the Metazoa. Plenum, New York, xiii + 570 pp.

— , ford, c. h. and gehling, j. g. 1983. The Ediacaran Member of the Rawnsley Quartzite: the context of
the Ediacara assemblage (late Precambrian, Flinders Ranges). Journal of the Geological Society of Australia,

30, 101-119.

— and gehling, j. g. r. 1978. A review of the frond-like fossils of the Ediacaran assemblage. Records of the
South Australian Museum, 17, 347-359.

jin yugan, wang HU a yu and wang wei 1991. Palaeoecological aspects of brachiopods from Chiungchussu
Formation of early Cambrian age, eastern Yunnan, China. 25-47. In jin yugan, wang jungeng and xu
shanhong (eds). Palaeoecology of China. Volume 1. Nanjing University Press, Nanjing, iii + 372 pp.
Johnson, H. and fox, s. k. 1968. Dipleurozoa from Lower Silurian of North America. Science, 162,
119-120.

keith, a. 1923. Cambrian succession of northwestern Vermont. American Journal of Science, (5th series), 5,
97-139.

— 1932. Stratigraphy and structure of northwestern Vermont. Journal of the Washington Academy of
Sciences, 22, 357 — 406.

khomentovsky, v. v. 1986. The Vendian System of Siberia and a standard stratigraphic scale. Geological
Magazine, 123, 333-348.

kocurko, M. J. 1988. Notes on fossil octocorals and comparisons of some modern and ancient octocoral
remains. Tulane Studies in Geology and Paleontology, 21, 105-1 14.
kukenthal, w. and broch, h. 1911. Pennatulacea. Wissenschaftliche Ergebnisse der Deutschen Tiefsee-
Expedition auf dem Dampfer ' Valdivia ’ 1898-1899, 13, i-vi, 113-576.
kuz'micheva, ye. i. 1980. [Fossil Gorgonida]. Paleontologichiskii Zhurnal , 1980 (4), 3-12. [In Russian; English
translation in Paleontological Journal, 14, 1-12].

landing, e. and narbonne, g. m. 1992. Scenella and ‘A chondrophorine (medusoid hydrozoan) from the basal
Cambrian (Placentian) of Newfoundland’. Journal of Paleontology, 66, 338.
lindstrom, m. 1978. An octocoral from the Lower Ordovician of Sweden. Geologica et Palaeontologica, 12,
541-552.

ludvigsen, r. 1989. The Burgess Shale: not in the shadow of the Cathedral Escarpment. Geoscience, Canada,
16, 51-59.

— 1990. Reply to comments by Fritz and Aitken and Mcllreath. Geoscience, Canada, 17, 116-118.
mcilreath, i. a. 1977. Accumulation of a Middle Cambrian, deep-water limestone debris apron adjacent to a

vertical, submarine carbonate escarpment, Southern Rocky Mountains, Canada. Society of Economic and
Petroleum Mineralogists, Special Publication, 25, 113-124.
madsen, F. J. 1957. On Walcott’s supposed Cambrian holothurians. Journal of Paleontology, 31, 281-282.
malecki, j. 1982. Bases of Upper Cretaceous octocorals from Poland. Acta Palaeontologica Polonica, 27,
65-75.

maxwell, w. d. 1989. The end Permian mass extinction. 152-173. In donovan, s. k. (ed.). Mass extinctions,
processes and evidence. Belhaven, London, xiv + 266 pp.
meer mohr, c. G. van der 1969. The stratigraphy of the Cantabrian Lancara Formation between the Luna

CONWAY MORRIS: CAMBRIAN EDI ACARAN-LIKE FOSSILS

633

River and the Esla River in the Cantabrian Mountain, Spain. Leidsche geologische mededelingen , 43,
233-316.

milashey, v. A. 1958. [Discovery of Upper Cambrian medusae in the Middle Olenek region of the Siberian
platform]. Sbomik Statei po paleontologii i biostratigrafii, 7, 19-21. [In Russian].
narbonne, G. M. and aitken, j. d. 1990. Ediacaran fossils from the Sekwi Brook area, Mackenzie Mountains,
northwestern Canada. Palaeontology , 33, 945-980.

and hofmann, h. j. 1987. Ediacaran biota of the Wernecke Mountains, Yukon, Canada. Palaeontology ,
30, 647-676.

and myrow, p. 1988. Trace fossil biostratigraphy in the Precambrian-Cambrian boundary interval.

Bulletin of the New York State Museum , 463, 12-16.

— , landing, E. and anderson, M. M. 1991. A chondrophorine (medusoid hydrozoan) from the basal
Cambrian (Placentian) of Newfoundland. Journal of Paleontology, 65, 186-191.
norris, R. D. 1989. Cnidarian taphonomy and affinities of the Ediacaran biota. Lethaia, 22, 381-393.
nutting, c. c. 1912. Descriptions of the Alcyonaria collected by theU.S. Fisheries Steamer ‘Albatross', mainly
in Japanese waters, during 1906. Proceedings of the United States National Museum , 43, 1-104.
onken, b. r. and signor, p. w. 1988. Lower Cambrian stratigraphic paleontology of the southwestern Great
Basin (White-Inyo Mountains of eastern California and Esmeralda County, Nevada). Bulletin of the
Southern California Paleontological Society , 20, 131-150.

Owen, r. 1843. Lectures on the comparative anatomy and physiology of the invertebrate animals delivered at the
Royal College of Surgeons in 1843. Longman, Brown, Green, and Longman, London, 392 pp.
peel, j. s. (ed.). 1990. Lower Cambrian trace fossils from Greenland. Gronlands Geologiske Under sogelse.
Rapport, 147, 1-62.

pflug, h. d. 1970. Zur Fauna der Nama-Schichten in Siidwest-Afrika. I. Pteridinia, Bau und systematische
Zugehorigkeit. Palaeontographica, Abteilung A , 134, 153-262.
pickerill, R. K. 1982. Cambrian medusoids from the St. John Group, southern New Brunswick. Current
Research Part B , Geological Survey of Canada, Paper , 82-1B, 71-76.
popov, yu. N. 1967. [New Cambrian scyphozoan], Paleontologichiskii Zhurnal, 1967 (2), 122-123. [In Russian;
English translation in Paleontological Journal, 1967 (2), 107-108].

— 1968. [Scyphomedusa from the type section of the Upper Cambrian of the Kulyumbe River]. 211-213.

In DARSENKO, V. A., ZHURAVLEVA, I. T., LAZARENKO, N. P., POPOV, YU. N. and CHERNYSHEVA, N. E.
[Biostratigraphy and fauna of the Cambrian deposits of the northwest of the Siberian platform}. Trudy Nauchno
issledovatelskogo Instituta Geologii Arktiki, 155, 221 pp. [In Russian],

Raymond, p. e. 1921. The history of corals and the Timeless’ oceans. American Journal of Science, 2, 343-347.
resser, c. E. and howell, B. F. 1938. Lower Cambrian Olenellus zone of the Appalachians. Bulletin of the
Geological Society of America, 49, 195-248.

rigby, J. K. 1986. Sponges of the Burgess Shale (Middle Cambrian), British Columbia. Palaeontographica
Canadiana, 2, 1-105.

— 1987. Early Cambrian sponges from Vermont and Pennsylvania, the only ones described from North
America. Journal of Paleontology, 61, 451^461.

rolfe, w. d. i. 1969. Phyllocarida. R296-R331. In moore, r. c. (ed.). Treatise on invertebrate paleontology. Part
R. Arthropoda 4. Geological Society of America and University of Kansas Press, Boulder, Colorado and
Lawrence, Kansas, 651 pp.

rozanov, a. yu. and zhuravlev, a. yu. 1992. The Lower Cambrian fossil record of the Soviet Union.
205-282. In lipps, j. h. and signor, p. w. (eds). Origin and early evolution of the Metazoa. Plenum, New
York, xiii + 570 pp.

ruedemann, r. 1916. Note on Paropsonema cryptophya Clarke and Discophyllum peltatum Hall. Bidletin of the
New York State Museum , 189, 22-27 .

— 1934. Paleozoic plankton of North America. Geological Society of America , Memoir, 2, i-vii, 1-141.
runnegar, b. 1982. Oxygen requirements, biology and phylogenetic significance of the late Precambrian worm

Dickinsonia, and the evolution of the burrowing habit. Alcheringa, 6, 223-239.

■— 1992. Oxygen and the early evolution of the Metazoa. 65-87. In bryant, c. (ed.). Metazoan life without
oxygen. Chapman and Hall, London, xii + 291 pp.

- - and fedonkin, M. A. 1992. Proterozoic metazoan body fossils. 369-388. In schopf, j. w. and klein, c.
(eds). The Proterozoic biosphere. A multidisciplinary study. Cambridge University Press, Cambridge,
xxiv+ 1348 pp.

sass, d. b. and rock, b. n. 1975. The genus Plumalina Hall, 1858 (Coelenterata) - re-examined. Bulletin of
American Paleontology, 67, 407-422.

634

PALAEONTOLOGY, VOLUME 36

schlichter, d. 1991. A perforated gastrovascular cavity in the symbiotic deep-water coral Leptoseris
fragilis: a new strategy to optimize heterotrophic nutrition. Helgo/dnder Meersuntersuchimgen, 45,
423-443.

schopf, t. j. m. 1974. Permo-Triassic extinctions: relation to sea-floor spreading. Journal of Geology, 82,
129-143.

1979. The role of biogeographic provinces in regulating marine faunal diversity through geologic time.
449-457. In gray, j and boucot, a. j. (eds). Historical biogeography, plate tectonics, and the changing
environment. Oregon State University Press, Corvallis, xii + 500 pp.
scrutton, c. t. 1979. Early fossil cnidarians. 161-207. In house, m. r. (ed.). The origin of major invertebrate
groups. Svstematics Association Special Volume 12. Academic Press, London, x-t-515 pp.
seilacher, a. 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, 450 pp.

1989. Vendozoa: organismic construction in the Proterozoic biosphere. Lethaia, 22, 229-239.

1992. Vendobionta and Psammocorallia: lost constructions of Precambrian evolution. Journal of the
Geological Society of London , 149, 607-613.

— , reif, w.-e. and westphal, f. 1985. Sedimentological, ecological and temporal patterns of fossil
Lagerstatten. Philosophical Transactions of the Royal Society of London, B311, 5-23.

SEPKOSKI, J. J. 1982. A compendium of fossil marine families. Milwaukee Public Museum Contributions in
Biology and Geology, 51, 1-125.

1992a. A compendium of fossil marine animal families. 2nd edition. Milwaukee Public Museum
Contributions in Biology and Geology, 83, 1-156.

19926. Taxonomy and stratigraphic ranges of animal families in the Vendian and Cambrian: data and
analytical results for section 11.4. 1171-1184. In schopf, j. w. and klein, c. (eds). The Proterozoic biosphere.
A multidisciplinary study. Cambridge University Press, Cambridge, xxiv+ 1348 pp.
shaw, a. b. 1954. Lower and lower Middle Cambrian faunal succession in northwestern Vermont. Bulletin of
the Geological Society of America, 65, 1033-1046.

1955. Paleontology of northwestern Vermont. V. The Lower Cambrian fauna. Journal of Paleontology,
29, 775-805.

1958. Stratigraphy and structure of the St Alban's area, northwestern Vermont. Bulletin of the Geological
Society of America , 49, 519-568.

sokolov, b. s. and fedonkin, m. a. 1984. The Vendian as the terminal system of the Precambrian. Episodes, 7,
12-19.

— 1985. [ Vendian system. Principles of historical geology and palaeontology. Volume 1, Palaeontology],
Nauka, Moscow, 222 pp. [In Russian, English translation published by Springer-Verlag, Berlin, 1990,
xiii + 383 pp.].

Stanley, G. d. 1986. Chondrophorine hydrozoans as problematic fossils. 68-86. In hoffman, a. and nitecki,
m. h. (eds). Problematic fossil taxa. Oxford University Press, New York, xi + 267 pp.
stasinska, a. 1960. Velumbrella czarnockii n. gen., n. sp - meduse du Cambrien inferieur des Monts de Sainte-
Croix. Acta Palaeontologica Polonica, 5, 337-344.

stormer, l. 1944. On the relationships and phylogeny of fossil and Recent Arachnomorpha. A comparative
study on Arachnida, Xiphosura, Eurypterida, Trilobita, and the fossil Arthropoda. Skrifter utgitt av det
Norske Videnskaps-Akademi i Oslo, 5, 1-158.

sun weiguo. 1986. Late Precambrian pennatulids (sea pens) from the eastern Yangtze Gorge, China:
Paracharnia gen. nov. Precambrian Research, 31, 361-375.

- and hou xianguang. 1987. [Early Cambrian Medusae from Chengjiang, Yunnan, China]. Acta
Palaeontologica Sinica, 26, 257-271. [In Chinese, with English abstract].

tarlo, L. B. H. 1967. Xenusion - onychophoran or coelenterate? Mercian Geologist, 2, 97-99.
termier, H. and termier, G. 1952. Histoire geologique de !a biosphere, la vie et les sediments dans les geographies
successives. Masson et Cie, Paris, 721 pp.

— - 1960. Paleontologie stratigraphique . Masson et Cie, Paris, 515 pp.

1968a. Evolution et biocinese. Masson et Cie, Paris, 237 pp.

- 19686. Biologie et ecologie des premiers fossiles. In grasse, p.-p. (ed.). Les grands problemes de la
biologie. Monographic 8. Masson et Cie, Paris, 213 pp.

tiffon, y. 1987. Ordre des Cerianthaires (Ceriantharia Perrier, 1883). 211-256. In doumenc, d. (ed.). Traite de
zoologie. Tome III, Fascicule 3, Cnidaires, Anthozoaires. Masson, Paris, 859 pp.
tremblay, p. 1941. Morphologie externe d’un fossile nouveau. Le Naturaliste Canadian, 68, 272.

CONWAY MORRIS: CAMBRIAN EDI ACARAN-LIKE FOSSILS

635

velikanov, v. A., assejeva, e. a. and fedonkin, m. a. 1983. [ Vendian of the Ukraine}. Naukova Dumka, Kiev,
164 pp. [In Russian].

verrill, a. e. 1865. Synopsis of the polyps and corals of the North Pacific Exploring Expedition, under
Commodore C. Ringgold and Captain John Rodgers, U.S.N., from 1853 to 1856. Collected by Dr Wm.
Stimpson, naturalist to the expedition. With descriptions of some additional species from the west coast of
North America. Proceedings of the Essex Institute , Salem , 4, 181-196.
wade, m. 1971. Bilateral Precambrian chondrophores from the Ediacara fauna. South Australia. Proceedings
of the Royal Society of Victoria, 84, 183-188.

— 1 972a. Dickinsonia: polychaete worms from the late Precambrian Ediacara fauna. South Australia.
Memoirs of the Queensland Museum, 16, 171-190.

— 19726. Hydrozoa and Scyphozoa and other medusoids from the Precambrian Ediacaran fauna. South
Australia. Palaeontology, 15, 197-275.

walcott, c. d. 1886. Second contribution to the studies on the Cambrian faunas. Bulletin of the United States
Geological Survey, 30, 1-369.

— 1889. Stratigraphic position of the Olenellus fauna in North America and Europe. American Journal of
Science , (3), 37, 374-392.

1890. The fauna of the Lower Cambrian or Olenellus Zone. United States Geological Survey, 10th Annual
Report, (1891), 515-760.

— 1898. Fossil Medusae. Monographs of the United States Geological Survey, 30, ix+ 1-201.

— 1911. Cambrian geology and paleontology. II. Middle Cambrian annelids. Smithsonian Miscellaneous
Collections , 57, 109-144.

1912n. Cambrian geology and paleontology. II. Middle Cambrian Branchiopoda, Malacostraca,
Trilobita, and Merostomata. Smithsonian Miscellaneous Collections, 57, 145-228.

— 19126. Cambrian of the Kicking Elorse Valley, B.C. Report of the Geological Survey of Canada, 26,
188-191.

— 1918. Geological explorations in the Canadian Rockies. In Explorations and field-work of the
Smithsonian Institution in 1917. Smithsonian Miscellaneous Collections, 68, 3-20.

— 1931. Addenda to descriptions of Burgess Shale fossils (with explanatory notes by C. E. Resser).
Smithsonian Miscellaneous Collections, 85, 1-46.

wells, J. w. and hill, d. 1956. Zoanthmiaria, Corallimorpharia, and Actiniara. F232-F233. In moore, r. c.
(ed.). Treatise on invertebrate paleontology. Part F. Coelenterata. Geological Society of America and
University of Kansas Press, Boulder, Colorado and Lawrence, Kansas, 498 pp.

Whittington, H. b. 1981. Rare arthropods from the Burgess Shale, Middle Cambrian, British Columbia.
Philosophical Transactions of the Royal Society of London, Series B, 292, 329-357.

— 1985. The Burgess Shale. Yale University Press, New Haven and London, xv+ 151 pp.

— 1990. Articulation and exuviation in Cambrian trilobites. Philosophical Transactions of the Royal Society
of London, Series B, 329, 27—46.

— and briggs, D. E. G. 1985. The largest Cambrian animal, Anomalocaris, Burgess Shale, British Columbia.
Philosophical Transactions of the Royal Society of London, Series B, 309, 569-609.

wignall, p. b. and hallam, A. 1992. Anoxia as a cause of the Permian/Triassic mass extinction: facies
evidence from northern Italy and the western United States. Palaeogeography, Palaeoclimatology,
Palaeoecology, 93, 21-46.

Willoughby, r. h. and robison, R. a. 1979. Medusoids from the Middle Cambrian of Utah. Journal of
Paleontology, 53, 494-500.

ziegler, g. 1991. Was ist Laffonia helvetica Heer? Stuttgarter Beitrdge zur Naturkunde, Serie B (Geologie und
Paldontologie ), 172, 1-10.

S. CONWAY MORRIS

Department of Earth Sciences
University of Cambridge
Downing Street
Cambridge CB2 3EQ, UK

Typescript received 26 June 1992

Revised typescript received 14 September 1992

DIPTERID FERNS FROM THE MESOZOIC OF
ANTARCTICA AND NEW ZEALAND AND THEIR
STRATIGRAPHICAL SIGNIFICANCE

by PETER MCA. REES

Abstract. Two genera of dipteridaceous ferns, Goeppertella and Hausmannia, are described for the first time
from the Mesozoic Elope Bay and Botany Bay assemblages of the northern Antarctic Peninsula, and
Goeppertella from the Clent Hills assemblage of New Zealand. These are the first gondwanan records outside
Argentina of Goeppertella. Two new species of the genus, G.jeffersonii and G. woodii, are described from Hope
Bay and Botany Bay. Based on the global distribution of Goeppertella , its occurrence in these gondwanan floras
indicates that they should be assigned an Early Jurassic or possibly earlier age, contrasting sharply with
recently published Late Jurassic or Early Cretaceous age assignments. A pre-Late Jurassic age for the Hope
Bay and Botany Bay assemblages is further supported by independent evidence from radiometric data. An
earliest Cretaceous age for these assemblages has been adopted in most recent interpretations of volcanic arc
evolution and palaeogeography in this region of Antarctica, the plant-bearing beds providing direct evidence
of terrestrial sedimentation: these interpretations are revised here, based upon the new evidence. The ages
assigned to a number of other late Mesozoic gondwanan floras, particularly from Argentina and India,
must be reconsidered since many of these were dated on the basis of comparison with the Hope Bay
assemblage.

Because of its great diversity and its early discovery and description (Halle 1913), the ‘classic’
fossil assemblage from Hope Bay, northern Graham Land, Antarctica (Text-fig. 1), has long been
considered a standard for floristic and biostratigraphical studies on other Mesozoic gondwanan
floras. Halle (1913) originally assigned a Middle Jurassic age to the assemblage, but it has
subsequently been variously dated as Early Jurassic (Orlando 1971) or Middle Jurassic (Rao 1953),
through to latest Jurassic or earliest Cretaceous (e.g. Stipanicic and Bonetti 1970b; Archangelsky
and Baldoni 1972). A latest Jurassic or earliest Cretaceous age has been adopted in recent
publications dealing with the palaeobotanical and geological history of the region (e.g. Baldoni
1981; Thomson et a/. 1983; Farquharson 1983, 1984; Farquharson et a/. 1984; Del Valle and
Fourcade 1986; Macdonald et al. 1988).

The assemblage of two hundred and twenty hand specimens from Hope Bay, as described
originally by Halle (1913), comprises fifty-nine species and two forms of unknown affinity, recently
revised to forty-three by Gee (1989) based on Halle’s specimens. The study of additional
undescribed material, totalling some two thousand hand specimens, from Hope Bay and a new
assemblage from nearby Botany Bay (Text-fig. Ib) has enabled an extensive revision of the Hope
Bay assemblage (Rees 1990). The Botany Bay assemblage comprises thirty-one species, twenty-five
of which also occur in the Hope Bay assemblage and they are so closely similar that they can be
considered as having essentially the same age. The previously unrecorded presence of the
dipteridaceous genus Goeppertella indicates an Early Jurassic or earlier age for these assemblages,
with an Early Jurassic age being most likely on present evidence from the assemblages as a whole
(Rees 1990). This is corroborated by recently published radiometric data (Millar et al. 1990) which
provide evidence of a Jurassic, rather than Cretaceous, age for the assemblages. Interpretations of
volcanic arc evolution and palaeogeography which utilized a latest Jurassic or earliest Cretaceous
age for these assemblages must be revised; volcanic arc uplift commenced prior to terrestrial

[Palaeontology, Vol. 36, Part 3, 1993, pp. 637-656, 3 pis.]

© The Palaeontological Association

638

PALAEONTOLOGY, VOLUME 36

text-fig. 1 . a, Antarctic Peninsula, showing location of main study area and the Cretaceous assemblage from
Alexander Island (♦). b, principal Jurassic terrestrial deposits in northern Graham Land, as well as Williams

Point (Cretaceous) in the South Shetland Islands.

deposition of the plant-bearing sediments in the Early Jurassic and not in the Early Cretaceous as
believed previously (e.g. Farquharson 1984).

Two new species, Goeppertella jeffersonii and G. woodii , are described here from Hope Bay and
Botany Bay, together with a specimen of Goeppertella from the Clent Hills assemblage of New
Zealand, considerably extending the previously known gondwanan distribution of the genus.
Another genus of the Dipteridaceae, Hausmannia , is also described for the first time from the Hope
Bay and Botany Bay assemblages; this genus was only known previously in Antarctica from
Alexander Island (Text-fig. 1a; Jefferson 1981).

MATERIAL AND METHODS

Fossil plant material was first collected from Hope Bay during the Swedish 1901-1903 expedition and has been
studied and described by Halle (1913) and Gee (1989). It would appear that no other worker has directly
studied the plants; certainly, no information from any subsequent collection has ever been published. British
expeditions collected extensive additional material from Hope Bay during Operation Tabarin in 1945 and the
Falkland Islands Dependencies Survey (FIDS) in 1946. Material was collected from the nearby Botany Bay
locality by W. N. Croft (FIDS, 1946) and G. W. Farquharson (1979/1980 British Antarctic Survey (BAS) field
programme); this has been supplemented by my own extensive collecting from Botany Bay, as part of the
1986/1987 BAS field programme. The material from these expeditions is housed in the Natural History
Museum (prefixed V.), London, and forms part of the palaeobotanical collections of the British Antarctic

REES: MESOZOIC FERNS FROM ANTARCTICA

639

Survey; it has provided the basis of the present revision. The Antarctic specimens studied for this paper are
numbered utilizing Natural History Museum registration numbers; they are listed, along with the equivalent
British Antarctic Survey station numbers, in the appendix.

The plant material from both localities occurs as impressions and coalified compressions. In addition to
diagenetic processes, it would appear that younger igneous intrusions have contact-metamorphosed the plant
beds (Farquharson 1984). Identifiable palynomorphs have not been recovered from either locality (T. H.
Jefferson in Farquharson 1984; D. Guy-Ohlson in Gee 1989; Rees 1990). The leaf cuticles have been converted
to high-rank coal and cannot be prepared by conventional methods, although their original structure is
occasionally preserved on impression surfaces, mainly on specimens from Botany Bay. Epidermal characters
of such specimens could be examined and photographed directly, utilizing scanning electron microscopy.
Macroscopic features of the specimens were enhanced by photographing them either under a thin coating of
ammonium chloride or under cross-polarized light. The specimen of Goeppertella from New Zealand (also at
the Natural History Museum) is similar to those from Hope Bay in preservation and was photographed only
at the macroscopic level; no epidermal details are preserved.

SYSTEMATIC PALAEONTOLOGY

Order filicales Engler and Prantl, 1898-1902
Family dipteridaceae Seward and Dale, 1901

The fern family Dipteridaceae comprises several genera, which are distinguishable on the basis of
their gross morphology and venation. Two of these, Hausmannia and Clathropteris , possess a
lamina which ranges from being entire to weakly (and often irregularly) segmented. In the other
genera segmentation is deeper and more consistent, producing more distinct frond-members which
are separate to their bases and are more or less pinnate. Of these, Thaumatopteris, Camptopteris and
Dictyophyllum possess frond-members which are characteristically once-pinnate (with pinnules
arising directly from each frond-member), whereas in Goeppertella they are twice-pinnate (with each
frond-member bearing pinnae, each pinna bearing pinnules). Additionally, vein meshes in
Hausmannia and Clathropteris are typically rectangular; these are typically polygonal in the other
dipteridaceous genera.

Problems can arise with fragmentary material which is not demonstrably bipinnate; for instance,
when it is not known whether a fragment is a pinna of Goeppertella or a frond-member of
Dictyophyllum. Such specimens from Hope Bay and Botany Bay can be assigned to Goeppertella on
the basis of their close association with, and similarity to, more complete material from these
localities which is assignable with confidence to this genus. The terminology used here in describing
specimens of Goeppertella is explained in Text-figure 2.

Genus goeppertella Oishi and Yamasita, 1936
Type species. Goeppertella microloba (Schenk) Oishi and Yamasita, 1936.

Previously recorded distribution. Late Triassic, possibly Early Jurassic, in the northern hemisphere; Early
Jurassic in Argentina.

Goeppertella jeffersonii sp. nov.

Plate 1, figs 1-3; Plate 3, fig. 4; Text-fig. 3b

Derivation of name. For the late T. H. Jefferson, in recognition of his contributions to Antarctic
palaeobotany.

Holotype. V. 63595 from frost-shattered debris derived from the Camp Hill Formation, Botany Bay, northern
Antarctic Peninsula (63° 4F S; 57° 53' W).

Material. From Botany Bay. V.63590-V. 63597.

640

PALAEONTOLOGY, VOLUME 36

text-fig. 2. Schematic diagram explaining the terminology used here for Goeppertella. A, complete frond and
frond rachis. b, portion of frond-member, c, portion of pinna, b and c are enlargements of the boxed areas

in a and B respectively.

Diagnosis. Frond-member bipinnate ; pinnae alternating regularly with well-defined rachial pinnules.
Shape and size of rachial pinnules comparable to that of pinnules on pinna rachis; pinnules wedge-
shaped, slightly falcate, 4-13 mm long x 2-8 mm wide, curving forward near their tips towards
pinna apex. Pinnule apices acute to subacute, margins typically entire. Venation reticulate, poorly
ranked, veins dividing to form polygonal vein meshes, first laterals branching uniformly from main
vein, pattern not disrupted by rachial veins.

EXPLANATION OF PLATE 1

Figs 1-3. Goeppertella jeffersonii sp. nov. ; Botany Bay. 1, V. 63595; fragment near apex of bipinnate frond-
member bearing pinnae and rachial pinnules, x 3. 2, V. 63593; two bipinnate frond-members with pinnae
bearing small pinnules, x 2-5. 3, V. 63595; detail of pinnule morphology and venation, with the pinnule on
the right showing marginal teeth on its basiscopic margin, x 10.

PLATE 1

REES, Goeppertellct

642

PALAEONTOLOGY, VOLUME 36

text-fig. 3. Pinnules of Goeppertella from Botany Bay, drawn from SEM montages, showing differing pinnule
shape, size and venation. A, Goeppertella woodii sp. nov. ; V. 63610. b, Goeppertella jeffersonii sp. nov. ; V.63595.

Scale bars = 1 mm.

Description. Main rachis and overall form of frond unknown. Frond-member bipinnate, known mainly from
fragments in its apical region. Member rachis stout, tapering in apical direction, up to at least 1-2 mm wide,
its epidermis composed of elongate cells with scale bases occurring at irregular intervals of about 40-400 pm.
Pinnae borne laterally in one plane, typically subopposite but occasionally alternate, at intervals of 12-15 mm
(measured between pinna rachises) and at angles of 40°-80° to the member rachis, becoming more acute
towards the apex of the frond-member. Pinnae ranging from an estimated 50 mm long (judged from taper of
a pinna in the holotype) to at least 105 mm. Pinna rachis 01-1 mm wide (mostly < 0-5 mm), typically straight
but sometimes curving forward, especially in the region of the frond-member apex; epidermis like that of
member rachis. Basalmost pinnule occurring on acroscopic side of pinna (based on three observations in two
sufficiently complete specimens). Basalmost pair of pinnules smaller than more distal ones, which are of fairly
uniform size until the pinna apical region where they become progressively smaller over a distance of
c. 6-20 mm. Pinnules characteristically subopposite, rarely opposite or alternate, fully confluent in their proxi-
mal regions to form a continuous lamina bordering the rachis; pinnule lamina raised slightly. Pinnule
length (measured along main vein from pinnule apex to rachis) ranging from 4 to 13 mm, mostly 5-1 1 mm, the
shortest ones (4—5 mm long) occurring in the apical regions of the pinnae. Pinnule width (between sinus points)
ranging from 2 to 8 mm, mostly 3-5-6 mm, the narrower pinnules (up to 4 mm wide) tending to occur near
pinna apices. Ratio of pinnule length to width ranging from 1-4:1 to 2-9: 1, mostly in range 1-6:1 to 2-2: 1, not
consistent with position of pinnules on pinna. Width of confluent lamina perpendicular between sinus point
and pinna rachis ranging from 1 to 3 mm. Pinnules ranging from squat to relatively narrow but basically
uniform, always forming wedge-shaped, slightly falcate segments which curve forward in the direction of the
pinna apex, especially at their tips which are characteristically sub-acutely to acutely pointed to almost
mucronate, rarely rounded. Pinnule margin normally appearing entire but occasionally very faintly and
shallowly lobed; where exceptionally well-preserved, very finely denticulate at intervals of about 0-2 mm and
showing distinct marginal vein.

Except at extreme apex of frond-member, where rachial lamina is absent, well-defined rachial pinnules
occurring uniformly, one such pinnule occupying the gap between each successive pair of pinnae, its main vein
departing from the member rachis at the same angle as the neighbouring pinna rachises. Shape and size of
rachial pinnules (5-7 mm long x 3-6 mm wide) comparable with that of pinnules on pinfia rachis, although
rachial pinnules are often a little squatter (ratio of length to width c. 1:1 to 1-6:1), with more rounded and less
falcate free portions. Rachial pinnules variously overlapping basal pinnules of the neighbouring pinnae, being
confluent in their proximal regions bordering the member rachis; distance from sinus points formed between
these two kinds of pinnules to the member rachis ranging from 1 to 2-5 mm.

REES: MESOZOIC FERNS FROM ANTARCTICA

643

Each pinnule on the pinna rachis supplied by a main vein up to 01 mm wide which runs to the pinnule apex
and defines its long axis; main vein departing from rachis at acute angles of 40-60° near pinna apex, 50-90°
(typically c. 60-70°) elsewhere, typically curving gently forward, especially in distal region of pinnule; following
a slightly sinuous course, reflecting its branching, becoming finer and more frequently sinuous towards the
pinnule apex. Coarse (first-order) lateral veins given off from main vein at intervals of about 1-3 mm and at
angles of about 60°, branching and thinning to produce a reticulate network of anastomosing finer veins
enclosing polygonal, 4-6-sided; islets of lamina about 0-5-1 mm in diameter but sometimes elongated up to
about 3 mm long, about twice as long as broad. Other orders of venation not clearly delimited, islets in
holotype often supplied by a blind-ending vein, either single or once-divided into a Y-shape. Epidermal cells
poorly preserved, with coarsely sinuous walls of irregular shape, size and orientation. Stomata obscure but
guard cells visible, about 30-50 pm long, apparently anomocytic, orientated in various directions. Cells over
veins elongate, about 50-200 /mi long x 10 /mi wide. Fertile material unknown.

Discussion. The bipinnate organization of the frond-member is diagnostic of Goeppertella within the
Dipteridaceae, it being the only member of the family which shows this feature (see Arrondo and
Petriella 1982 for further discussion). The type of ‘intercalary element’ (= rachial pinnule or
lamina) has been used by Arrondo and Petriella (1982) to distinguish species of the genus. Certain
other characters serve to separate these species; pinnule shape and orientation on the pinna rachis,
the style of pinnule apex (i.e. whether acute or obtuse), pinnule size and the type of margins (i.e.
whether entire or undulating).

Eight species of Goeppertella have been described previously (Arrondo and Petriella 1982 and
references therein). Of these, G. jeffersonii bears most resemblance to G.frenguelliana, G. microloba ,
G. macroloba and G. neuqueniana. The style of rachial lamina and pinnule size distinguishes it from
G. frenguelliana, which has a greatly reduced lamina, while G. microloba has differently-shaped
pinnules which have undulating margins. G. macroloba possesses pinnules which are considerably
larger than those of G. jeffersonii, whereas pinnules of G. neuqueniana are longer (with higher length
to width ratios) and have obtuse rounded apices. Additionally, pinnules of G. neuqueniana are
typically separate from one another on the pinna rachis, only becoming confluent in apical regions
of the pinna. The rachial pinnules of G. microloba , G. macroloba and G. neuqueniana are less
pronounced than those of G. jeffersonii, being little more than small lobes in the central regions of
the rachial laminae.

Goeppertella woodii sp. nov.

Plate 2, figs 1^1; Plate 3, figs 1-3, 5; Text-fig. 3a

1913 Dictyophyllum sp. ; Halle, text-fig. 2, pi. 1, figs 28, 28a.

1989 Dictyophyllum sp. ; Gee, pi. 2, fig. 13.

Derivation of name. For P. Wood, who was my companion in Antarctica and ensured the success and safety
of our field season.

Holotype. Y. 63602 from the Camp Hill Formation, Botany Bay, northern Antarctic Peninsula (63° 41' S;
57° 53' W).

Material. From Hope Bay -V. 63598, V. 63599; from Botany Bay - V. 63600 to V.63619.

Diagnosis. Frond-member bipinnate. Pinnae alternating with rachial lamina; lamina shape
irregular, but always broadening from proximal to distal sinus points, occasionally lobed distally.
Pinnules wedge-shaped, strongly falcate, relatively narrow in free portions, 6-25 mm
long x 5-13 mm wide, curving towards pinna apex. Pinnule apices acute to subacute, margins
typically entire. Venation reticulate; first-order lateral veins arising from main vein, joining with
veins arising from rachis between pinnules, dividing to form polygonal vein meshes. Fertile
segments with sori typically on rachial lamina and pinnule bases; sori 0-7-1 mm across, comprising
ten or more sporangia c. 01 mm in diameter.

644

PALAEONTOLOGY, VOLUME 36

Description. Main rachis and overall form of frond unknown. Frond-member bipinnate, large, known from
fragments both of the apical region and more proximally. Member rachis stout, tapering distally, up to at least
3 mm wide; longest length preserved, 120 mm. Epidermis of member rachis composed of both short and
elongate cells with no clear evidence of scale bases. Pinnae borne laterally in one plane, typically sub-opposite,
occasionally alternate or opposite, at intervals (measured between pinna rachises) of 1 1 mm to at least 40 mm
(mostly 12-20 mm) and at angles of 30-75° (mostly 40-60°) to the member rachis. Pinna separation and angle
of attachment to the member rachis decreasing towards the apical region of the frond-member ; greater pinna
separation and angle from the rachis associated with thicker member rachises (presumed to be from more
proximal regions of the frond-member). Measured pinnae up to 90 mm long, overall length ranging from an
estimated 80-100 mm up to more than 150 mm (judged from taper of pinna rachis). Pinna rachis gradually
tapering distally, from 0-2 to 1-3 mm wide (mostly 0-2-0-8 mm), typically straight but sometimes curving
forward in the apical region of the frond-member to become subparallel to the frond-member rachis. Epidermis
of pinna rachis like that of frond-member, occasionally with oval pits, presumably scale bases, 50-200 /im long.

Basal pinnules of pinna are usually smaller and less pronounced than more distal ones (with the exception
of those in the apical 20-30 mm of pinna). Pinnules characteristically subopposite, sometimes opposite or
alternate, fully confluent in their proximal regions to form a continuous lamina bordering the rachis ; pinnule
lamina raised slightly. Pinnule length (measured along main vein from pinnule apex to rachis) ranging from
6 to 25 mm, mostly 12-18 mm. Pinnule width (between sinus points) ranging from 5 to 13 mm, mostly
9-1 1 mm. Ratio of pinnule length to width ranging from 0-9: 1 to 2-3 : 1, mostly in range 11:1 to 1-4:1, not
consistent with position of pinnules on pinna. Width of confluent lamina perpendicular between sinus point
and pinna rachis ranging from 2 to 11 mm, mostly from 3-5 to 7 mm. Pinnules wedge-shaped, ranging from
squat (especially on more proximal pinnae of the frond-member) to, characteristically, relatively narrow in
their free part, strongly falcate, usually pointing strongly towards pinna apex, especially at their tips which are
typically acutely to subacutely pointed. Free regions of pinnules only rarely overlapping neighbouring ones but
often somewhat overlapping pinnules of neighbouring pinnae. Pinnule margin entire, occupied by a marginal
vein.

Interval between successive pinnae occupied by a rachial lamina except in extreme apical region of frond-
member, where rachial lamina is absent. Rachial lamina irregular in size and shape, but always broadening
from proximal pinna sinus point to just before the next more distal one, with the development in some cases
of a pinnule-like lobe (rarely, two, with the distal lobe being more pronounced). In other cases the lamina
broadens only slightly, occasionally being indistinguishable from the basal basiscopic pinnule of the next distal
pinna. Lamina ranging from 1 to 5 mm wide between proximal sinus points and member rachis, broadening
to 3-7 mm between next distal sinus points and rachis; distance between sinus points from 10 to 25 mm. Lobes,
where developed, from 6 to 13 mm long (measured from lobe apex to member rachis), symmetrical or curving
gently forward with rounded or pointed apices; lobe main veins arising from member rachis at 60-90°,
producing higher-order veins which divide in a similar manner to those of the pinnules on the pinnae.

Each pinnule supplied from pinna rachis by a main vein up to 0- 1 mm wide which runs to the pinnule apex
and defines its long axis ; main vein departing at angles of 40-90°, curving forwards to 0-40° near pinnule apex.
Main vein often slightly sinuous owing to lateral veins being given off. First order lateral veins arising at
intervals of about 1-5—3 mm, often at 90° or backwardly directed, joining with veins of similar strength given
off directly at 50-90° and at intervals of 2-5 mm from the pinna rachis to form a characteristic coarse mesh
(each c. 1-3 mm in diameter). First-order laterals producing second-order veins which anastomose to form a
fine mesh of polygonal, 4-6-sided islets of lamina about 0-25-0-5 mm in diameter. Veins with elongate cells.
Epidermal cells poorly preserved, with coarsely sinuous walls of irregular shape, size and orientation. Stomata
not clear.

EXPLANATION OF PLATE 2

Figs 1-4. Goeppertella woodii sp. nov. (from Botany Bay except Fig. 2, from Hope Bay). 1, V. 63603; proximal
region of fertile frond-member bearing two subopposite pinnae and characteristic rachial lamina, x F5. 2,
V.63598; proximal region of bipinnate frond-member with rachial lamina, x 1-5. 3, V.63616; frond-member
bearing alternate pinnae with sori present almost to pinnule apices, x 2. 4, V. 63619; partly fertile pinnae with
large broad pinnules, showing sori (dark spots) in their basal regions, x 2-25.

PLATE 2

REES, Goeppertella

646

PALAEONTOLOGY, VOLUME 36

Some frond-members and pinnae fertile, usually bearing sori within 0-5 mm of their rachises, with fewer
present further out on the pinnules, becoming rare or absent at extreme apex of pinnules. Occasionally,
neighbouring sori grouped together to produce appearance of one large or long sorus. Sori round to ovate,
c. 0-7-1 mm across, comprising at least ten sporangia c. 01 mm in diameter; sporangia coalihed.

Discussion. Goeppertella woodii differs from G. jeffersonii primarily in having an irregular rachial
lamina rather than rachial pinnules. Additionally, pinnules on pinnae of G. woodii are larger, more
falcate and generally narrower than those of G. jeffersonii. The rachial lamina of G. woodii, which
is irregular in size and shape, but which always broadens from the proximal sinus point to the distal
point, is unlike any seen in the eight species of Goeppertella described previously (Arrondo and
Petriella 1982 and references therein). Of these species, G. woodii is most similar to G.frenguelliana
and G. macroloba. However, the rachial lamina is greatly reduced in G.frenguelliana, being present
only as a narrow strip with margins which are parallel to the frond-member rachis, while the lamina
in G. macroloba is broadest in its central region, not distally as in G. woodii. The Dipteridaceae were
only known previously at Hope Bay from one poorly-preserved pinna fragment, comprising three
incomplete pinnules, which Halle (1913) and Gee (1989) assigned to Dictyophyllum sp. The
specimen which they described is closely similar in pinnule shape, size and venation pattern to more
complete material assigned here to G. woodii and can be included within this species.

Goeppertella cf. woodii sp. nov.

Text-fig. 4

Material. V. 157 19 from Haast Stream in the Clent Hills of South Island, New Zealand (collected in 1911 by
D. G. Lillie).

Description. Main rachis and overall form of frond unknown. Frond-member bipinnate, member rachis up to
1-2 mm wide, 47 mm long, bearing three pinnae. Pinnae subopposite, arising at angles of 45-60° to member
rachis, one pinna 40 mm long (complete to apex), another > 40 mm (incomplete) ; pinna rachises up to
0-3-0- 5 mm wide. Pinnules subopposite, falcate, main veins departing at 50-90° from pinna rachis, margins
entire, apices sub-acutely to acutely pointed. Pinnule length (measured along main vein from pinnule apex to
rachis) ranging from 7 to 13 mm (non-apical pinnules). Pinnule width (between sinus points) ranging from 4
to 10 mm. Width of confluent lamina perpendicular between sinus point and pinna rachis 3-5 mm. Rachial
lamina incompletely preserved, but broadens distally from proximal sinus point towards distal point, with a
distinct distal pinnule-like segment, 7-8 mm long, given off at 60-90° from the member rachis. Pinnule venation
similar to that seen in specimens of G. woodii from Antarctica. Epidermal and fertile details not known.

Discussion. This single specimen is most similar to G. woodii, agreeing in the style of its rachial
lamina and its pinnule size and shape as well as in the orientation of pinnules on the pinna rachis
and in venation pattern. As with G. woodii, it differs from the species described previously (Arrondo
and Petriella 1982) in the shape of its rachial lamina, which always broadens distally. Although

EXPLANATION OF PLATE 3

Figs 1-3, 5. Goeppertella woodii sp. nov.; Botany Bay. 1, V.63602; fragment near apex of bipinnate frond-
member bearing pinnae and rachial lamina, x 1. 2, V. 63614; fragment probably from proximal region of a
frond-member (the rachis either having been distorted during development of the frond or during
deposition), x 1. 3, V.63613; pinna fragments, the uppermost one being attached to a short length of frond-
member rachis at the extreme right of the block, with large pinnules bearing sori visible as dark patches on
their surfaces, x 1. 5, V. 63603, impressions of sori on a fertile frond-member showing annulus cells in a near-
vertical annulus, x 25. Fig. 4. Goeppertella jeffersonii sp. nov.; Botany Bay; V.63595; pinnule venation and
impressions of stomata, showing largely vertically aligned stomata and guard cells, x 90.

PLATE 3

REES, Goeppertella

648

PALAEONTOLOGY, VOLUME 36

text-fig. 4. Goeppertella cf. woodii. V. 15719; Clent Hills, New Zealand. A, bipinnate frond-member
photographed under cross-polarized light, showing the overall form of the pinnae and pinnules, x 1-25. b, the
same specimen coated with ammonium chloride, showing rachis and venation details, x 1-25.

pinnules on the pinnae of this specimen are similar in size to those of G. jeffersonii , the rachial
lamina is markedly different from the rachial pinnules seen on the frond-members of the Antarctic
species. The specimen is assigned here to G. cf. woodii until further material is available to confirm
its identity.

Genus hausmannia Dunker, 1846
Type species. Hausmannia dichotoma Dunker emend. Harris, 1961.

Hausmannia cf. nariwaensis
Text-fig. 5

1981 dictyophyl-phobos Jefferson, pi. 4.12, figs 1-2 (not figs 3-5).

Material. From Hope Bay-V.63420, V.63423, V.63620; from Botany Bay - V.63621-V.63623.

Description. Rachis not known, but region of rachial attachment seen in centre of complete lamina. Complete
lamina heart-shaped, up to 52 mm long x 43 mm wide, divided along its length into two main, almost identical,
lobes (incomplete lamina lobes seen, up to 95 mm long). Broadest point of lamina at about one third the
distance from the proximal to distal end, thereafter narrowing gradually until near the distal end. Proximal half
of lamina divided deeply to point of rachis attachment along lamina midline, the opposing margins of the two
lobes almost in contact or slightly overlapping along three fifths of the midline, then diverging to form the two
rounded proximal lobes of the lamina. Distal half of lamina divided from the distal end to half way to the point
of rachis attachment, divided distally into two rounded lobes which are narrower and less pronounced than
the proximal ones. Lamina margin entire within clefts between the two main lobes; elsewhere, markedly

REES: MESOZOIC FERNS FROM ANTARCTICA

649

text-fig. 5. Hausmannia cf. nariwaensis. A, NHM V.63620; near-complete fertile lamina and region of rachis
attachment, x 1-5. b, NHM V. 63420, fragment of large lamina with pronounced venation, showing vein orders
and areas with sori, x 5. c, NHM V. 63623, fragment of fertile lamina, x2. d, NHM V. 63623, detail of c
showing venation and sori, x 50. a-b from Hope Bay; c-D from Botany Bay.

650

PALAEONTOLOGY, VOLUME 36

crenulate, shallowly divided into convex lobes 0*5 — 1-5 mm deep, of variable length (from 4 to 12 mm) between
the main sinus points, several lobes with shallow medial sinus points indicating smaller lobes 2*5 — 5*5 mm long;
marginal vein evident, c. 01 mm wide.

In each half of the lamina, four main veins up to 0-2 mm wide radiate from the point of attachment to the
rachis, reaching to the margin and following a sinuous course as the various laterals are given off. Dichotomous
branching of main veins occurring at varying intervals, up to 10 mm apart, forming large-scale (first-order)
meshes; these are triangular through rectangular to polygonal (rarely more than pentagonal), often with sides
of variable length, shortest dimension about 3 mm, longest up to 15 mm. Within these first-order meshes,
thinner secondary vein branches arise from the main veins and anastomose to form secondary meshes of fairly
regular size, typically 1-5-4 mm long x 1-2-5 mm wide; these have 3 to 6 (rarely, 7) sides but are most
commonly rectangular. These meshes are traversed by finer veins which form third order meshes, 3 to 7 sided
but mostly rectangular, up to 1-5 mm x c. 0-5-1 mm. Occasionally, third-order veins dividing to form fourth-
order meshes which are mostly rectangular and from 0-5 to 1 mm long. A fifth order of venation may be present
in places, but divisions are unclear. Third-order meshes (or fourth-order, where seen) often occupied by a
rounded to oval sorus about 0-8 mm across, with a central receptacle and several sporangia (details obscured
by coarseness of matrix and/or coalification); epidermal details obscure, veins with elongate cells.

Discussion. Hausmannia is represented in the Hope Bay and Botany Bay assemblages by several
fragments of laminae and one near-complete lamina; it has not been described previously from
these localities. The specimens are similar to H. nariwaensis, described from Rhaetic floras of Japan
(Oishi 1932). They differ in the shape of the lamina, which is reniform (with the long axis
perpendicular to the median cleft) in the Japanese specimens and heart-shaped (the long axis being
parallel to the cleft) in the material from Hope Bay and Botany Bay. Further material is required
from these Antarctic localities in order to assess the significance of this difference, although the
specimens can be assigned to H. cf. nariwaensis on the basis of the close similarity in lamina division,
marginal lobing, venation pattern and soral details. Hausmannia ussuriensis, described by
Kryshtofovich (1923) from Rhaeto-Liassic rocks in Eastern Siberia is also similar to the material
described here but appears to have coarser and more numerous main veins, as well as a reniform
lamina. Hausmannia deferrariisii Feruglio (1937) from Argentina and Hausmannia sp. cf. H.
deferrariisii described by Herbst (1979) from Australia differ from the Antarctic material in having
a reniform lamina which, in addition, is less evenly incised. The only previous Antarctic record of
the genus is from the ?Aptian-Albian assemblage of Alexander Island (Text-fig. Ia), the material
being assigned to a biorecord, dictyophyl-phobos, by Jefferson (1981). The specimens from Hope
Bay and Botany Bay can be assigned to the same species as some of those described from Alexander
Island (Jefferson 1981, pi. 4.12, figs 1-2), since they are almost identical in lamina size, shape and
marginal lobing, as well as in venation pattern and soral details. However, the shape of the lamina,
as well as the marginal lobing and venation, differs in the other specimens figured by Jefferson (1981,
pi. 4.12, figs 3-5) and they possibly represent a different species of Hausmannia.

AGE OF THE HOPE BAY, BOTANY BAY AND CLENT HILLS ASSEMBLAGES
Age ranges of Goeppertella and Hausmannia

The only previous records of Goeppertella in the southern hemisphere are from beds in Argentina
dated as Early Jurassic (Herbst 1964, 1966, 1975; Arrondo and Petriella 1982 and references
therein; Baldoni 1987). Of the five Argentine species of the genus, two were assigned Early Jurassic
ages on the basis of the plants themselves, in the absence of independent age constraints such as
radiometric dating, palynology or marine faunas used to date the other three species (see Rees 1990
for details). It remains possible that these two Argentine species may be younger than Early
Jurassic, although they would be the first records of such an occurrence. Indeed, most species of
Goeppertella from the northern hemisphere are of Late Triassic age, with the possibility of some
ranging into the lower part of the Early Jurassic and none being known from younger floras (e.g.
Moller and Halle 1913; Oishi and Yamasita 1936 and references therein; Harris 1946). It is apparent

REES: MESOZOIC FERNS FROM ANTARCTICA

651

that Goeppertella is represented globally in strata which have a published age range of Late Triassic
to uppermost Early Jurassic and it has not been shown to occur in younger floras.

Unlike Goeppertella , Hausmannia is of limited stratigraphical value, being represented in floras
of latest Triassic and earliest Jurassic age (e.g. Kryshtofovich 1923; Harris 1931; Oishi 1932)
through to those of Early Cretaceous age (e.g. Seward 1913; Watson 1969; Jefferson 1981).

Age of the Hope Bay and Botany Bay assemblages

An earliest Cretaceous age for the Hope Bay and Botany Bay assemblages has been used in most
recent interpretations of Mesozoic volcanic arc evolution and palaeogeography in the northern
Antarctic Peninsula region (e.g. Farquharson 1984), with the palaeobotanical paper by Stipanicic
and Bonetti (19706) being the most frequently cited. Since the new evidence for an Early Jurassic
age contradicts previous arguments, the principal ones are reviewed here (see Rees 1990 for a
detailed account).

Stipanicic and Bonetti (1970a, 19706) reviewed the Argentine Jurassic floras and included (19706)
a discussion of the affinities and age of the Hope Bay plants. They concluded that they show an
equal degree of affinity with what they believed to be the Lower Cretaceous Rajmahal floras of India
as with those from the Middle Jurassic and Neocomian of Europe. For this reason, the authors
estimated that the Hope Bay assemblage was of latest Jurassic age, without discounting the
possibility that it could even be earliest Cretaceous. However, it would appear that their age
argument has two significant problems. Firstly, the Indian Rajmahal floras are imprecisely dated
and can only be reasonably assigned an age of ?Early Jurassic to ?Albian. An Albian age for the
Rajmahal plants is based upon the 100-105 Ma K-Ar dates for lavas which were believed to be of
the same age as the plant beds (McDougall and McElhinny 1970). Shah et al. (1973) considered that
the only criterion for determining the age of the Rajmahal Plant Beds was the plant remains and
concluded that they are of Early to Middle Jurassic age. Sengupta (1988, p. 154) discussed the
reasons for the contradictory radiometric (Early Cretaceous) and palaeobotanical (Jurassic) results
for the age of the Rajmahal flora. He argued that, although some samples of Rajmahal basalt (e.g.
those dated by McDougall and McElhinny 1970) indicated a Cretaceous age, their stratigraphical
and geographical location is poorly defined and cannot be used to assign a lower age limit to the
Rajmahal Formation. Sengupta (1988) concluded that the Rajmahal Formation may be considered
as Middle Jurassic to Cretaceous. Given the uncertainty concerning the age of this and other Indian
late Mesozoic plant-bearing sequences any age assignment based upon a correlation with them is
questionable. Secondly, it is difficult to accept that late Mesozoic floras from widely differing
palaeolatitudes (e.g. Antarctica and northern Europe) can be correlated and used with confidence
for stratigraphical purposes. When this type of correlation is carried out, it should be made clear
that further refinement, based upon local correlations, will be needed. For example, an
impression/coalified compression assemblage within the Fossil Bluff Formation on Alexander
Island, west of the Antarctic Peninsula (c. 71° S, 67° W ; Text-fig. 1a) has been independently dated
as ?Aptian-Albian on the basis of the presence of marine invertebrate fossils in the formation. The
assemblage contains twelve taxa which are morphologically similar to those from the Aptian of
Victoria, southern Australia, but it also has nine morphologically similar taxa in common with the
Middle Jurassic flora of Yorkshire, England (Jefferson 1981). If the floras from Victoria had not
been known to Jefferson, he may have concluded that the greatest affinity of the Alexander Island
flora was with that from Yorkshire. A Middle Jurassic age could then have been assigned to the
Cretaceous Alexander Island flora on palaeobotanical grounds. It appears that because Stipanicic
and Bonetti (19706) did not compare the Hope Bay plants with more local assemblages (particularly
those from Argentina), they assigned a latest Jurassic-earliest Cretaceous age to what is shown here
to be an Early Jurassic flora. It is interesting that Bonetti, both previously (1963) and subsequently
(1974), recognized the close similarity between the plants from Hope Bay and Argentina and
assigned ages to the latter based upon their close similarity with the assemblage from Hope
Bay.

652

PALAEONTOLOGY, VOLUME 36

Farquharson (1984) assigned the Hope Bay and Botany Bay plant-bearing beds to the Botany
Bay Group (BBG). The BBG was defined by Farquharson (1984, p. 28) as comprising ‘outcrops of
non-marine, mainly conglomeratic, sedimentary rocks derived from deformed metasedimentary
rocks ... [which] form a significant tectono- and litho-stratigraphic unit in the northern Antarctic
Peninsula’. Farquharson (1984) cited three lines of evidence for an earliest Cretaceous age for the
BBG.

Firstly, he cited the palaeobotanical arguments put forward by Stipanicic and Bonetti (1970fi) for
a latest Jurassic or earliest Cretaceous age for the Hope Bay assemblage. As demonstrated above,
however, these arguments cannot be used as evidence for this age.

Secondly was the presence of a marine intercalation within alluvial fan conglomerates in the
South Orkney Islands which Farquharson (1984) had included in the BBG; the marine sequence
contains ammonites indicative of an Early Cretaceous age (Thomson 1981). This age cannot be used
reliably to date what are merely lithologically-similar sequences from BBG localities elsewhere.

Thirdly, they used two radiometric ages of 130 + 7 Ma and 117 + 4 Ma obtained by Pankhurst
(1982) for rocks of the Antarctic Peninsula Volcanic Group (APVG); volcanic rocks of the APVG
overlie, or are commonly interbedded with, sedimentary sequences of the BBG. The Early
Cretaceous ages for rocks of the APVG from two localities in the region were used by Farquharson
(1984) to indicate a similar age for all of the BBG sequences. However, the relationship of these
dated volcanic rocks with those of the Botany Bay Group is uncertain, since they are not in contact
with them. Furthermore, Thomson and Pankhurst (1983, p. 328) remarked that ‘some caution is
necessary in accepting these ages since Rb-Sr whole-rock systems in acid volcanic rocks are widely
considered to be very easily reset without metamorphism’.

It can be seen that the evidence presented here for an Early Jurassic age for the Hope Bay and
Botany Bay assemblages outweighs that used previously to assign an earliest Cretaceous age to these
plants and to the Botany Bay Group as a whole. Significantly, new radiometric data indicate an
upper Middle or Late Jurassic age at youngest for the plants from Hope Bay and Botany Bay. At
Botany Bay, Sm-Nd dating of primary igneous garnets (from an andesitic sill within volcanic rocks
of the Antarctic Peninsula Volcanic Group) has yielded an age of 152 + 8 Ma; this age is believed
to indicate the time of intrusion of the sill (Millar et al. 1990). This corresponds to an age of lower
Callovian to lower Berriasian (Harland et al. 1982) or upper Bathonian to lower Kimmeridgian
(Haq et al. 1987). Thus, emplacement of the sill probably occurred sometime during the upper
Middle or Late Jurassic. The volcanic rocks (including the dated sill) overlie the plant-bearing
sedimentary sequence at Botany Bay (See Rees 1993 for further discussion). The close similarity
between the Botany Bay and Hope Bay assemblages indicates that the Hope Bay plants can be
assigned the same age as those from Botany Bay. The results of the radiometric dating (Millar et
al. 1990) are consistent with the revised age presented here for the Hope Bay and Botany Bay
assemblages and confirm that they should no longer be regarded as Cretaceous.

The new data cast considerable doubt upon the earliest Cretaceous age which was previously
assigned to all of the beds and formations included within the Botany Bay Group (e.g. Farquharson
1984). Indeed, it now seems probable that the BBG comprises sediments which were deposited in
discrete terrestrial basins during the Early Jurassic (Hope Bay and Botany Bay), with sedimentation
possibly continuing into the Early Cretaceous (South Orkney Islands). Although Farquharson
(1984) remarked that sediments from different outcrops of the Botany Bay Group may not have
been deposited contemporaneously, he was clearly not implying that their deposition spanned the
Early Jurassic to Early Cretaceous.

Age of the Clent Hills assemblage

Oliver et al. (1982) assigned a possible Middle to Late Jurassic age to the Clent Hills Group in the
Mount Somers area of South Island, New Zealand. They recognized two units within the group, one
non-marine and the other marine. The non-marine sequence is best represented at the Haast Stream
locality and contains the most abundant fossil plants, although no palynomorphs or macrofauna

REES: MESOZOIC FERNS FROM ANTARCTICA

653

have been found. Several workers have collected or identified plants from this locality and have
variously assigned Triassic or Jurassic ages (e.g. Haast 1877; Ettingshausen 1891; Arber 1917;
Edwards 1934). The plants were assigned Jurassic ages by Oliver et a I. (1982), who believed they
could not be assigned precise time-ranges. Microfloras have been discovered at two other non-
marine localities in the area and have been dated as Middle to Late Jurassic and Early Jurassic to
Early Cretaceous (Oliver et al. 1982 and references therein); their relationship to the macroflora in
Haast Stream is uncertain. The Middle to Late Jurassic age which Oliver et al. (1982) assigned to
the Clent Hills Group as a whole was based principally upon ages of invertebrate fossils from
marine sequences in the area. Significantly, the relationship between these sequences and those with
identifiable plant macrofossils is unknown, due to limited exposure and the absence of localities
showing gradations between the macrofloral and macrofaunal sequences. Goeppertella had not been
identified previously from any of the Clent Hills sequences. Its occurrence in the Haast Stream
assemblage indicates that the latter can be assigned a Late Triassic or Early Jurassic age.
Consequently, the stratigraphy of the Clent Hills Group should be reconsidered, since it now
comprises one non-marine sequence (and possibly more) of Late Triassic or Early Jurassic age as
well as Middle to Late Jurassic marine sequences.

CONCLUSIONS

The bipinnate specimens described here can be assigned with confidence to Goeppertella. The
previously recorded age range of the genus is from Late Triassic to uppermost Early Jurassic. It
remains possible that the new specimens (from Hope Bay, Botany Bay and Haast Stream) may
represent occurrences of the genus outside this range. However, the recent radiometric data of
Millar et al. (1990) indicate an upper Middle or Late Jurassic age for the volcanic rocks which
overlie the plant beds at Botany Bay. Also, previous arguments for a latest Jurassic or Early
Cretaceous age for the Hope Bay and Botany Bay assemblages (based upon palaeobotanical,
sedimentological and radiometric evidence) do not stand up to critical appraisal. It is concluded that
an Early Jurassic age assignment for the Hope Bay and Botany Bay assemblages is most likely on
present evidence. Certainly, they should no longer be regarded as latest Jurassic or Early
Cretaceous.

The Hope Bay and Botany Bay leaf fossils are now the oldest known from this area of Antarctica
since the assemblage from Williams Point, previously assigned a Triassic age (e.g. Lacey and Lucas
1981; Banerji and Lemoigne 1987), is now shown to be Cretaceous (Rees and Smellie 1989;
Chapman and Smellie 1992). Interpretations of palaeogeography and volcanic arc evolution in the
northern Antarctic Peninsula region have been revised in the light of these new age assignments (see
Rees 1993 for details). The new age assignment for the Hope Bay and Botany Bay assemblages
provides the first direct evidence that terrestrial sediments were deposited on a magmatic arc in at
least parts of the northern Antarctic Peninsula during the Early Jurassic. It is noteworthy that
marine beds of Jurassic or younger age are unknown from the central area of the northern Antarctic
Peninsula. It seems more probable that magmatic arc uplift occurred and an appreciable landmass
existed in this area from Early Jurassic times onwards, rather than from the Early Cretaceous as
suggested previously (e.g. by Farquharson 1984).

The present revision indicates that the ages assigned to a number of other Mesozoic gondwanan
floras must be reappraised, particularly those from Argentina which had been dated on the basis
of their close similarity to what had become regarded as the earliest Cretaceous assemblage from
Hope Bay. Further studies of the kind presented here are necessary in order to ensure that
reconstructions for instance, of palaeogeography and palaeoclimatic change are not severely
compromised by the use of inaccurate raw data.

Acknowledgements. I thank C. R. Hill for useful discussion and comments on the manuscript, as well as for
SEM work and detailed observations. Much of this work was carried out as part of my doctoral thesis; I thank

654

PALAEONTOLOGY, VOLUME 36

W. G. Chaloner and M. R. A. Thomson for their help and supervision. I am very grateful to P. Crabb for the
photographs and P. Wood for his help and good humour in Antarctica.

REFERENCES

arber, e. A. N. 1917. The earlier Mesozoic floras of New Zealand. Paleontological Bulletin , New Zealand
Geological Survey, 6, 1-80.

archangelsky, s. and baldoni, A. M. 1972. Revision de las Bennettitales de la Formacion Baquero (Cretacico
Inferior), Provincia de Santa Cruz. I. Hojas. Revista del Museo de La Plata, Seccion Paleontologia, 44,
195-265.

arrondo, o. G. and petriella, b. 1982. Revision del genero Goeppertella Oishi et Yamasita emend.

(Goeppertelloideae-Dipteridaceae). Ameghiniana, 19, 67-78.
baldoni, a. m. 1981. Tafofloras jurasicas y eocretacicas de America del Sur. 359-391. In volkheimer, w. and
musacchio, E. a. (eds). Cuencas sedimentarias del Jurasico y Cretacico de America del Sur, 2. Comite
Sudamericano del Jurasico y Cretacico, Buenos Aires, 713 pp.

— 1987. Megafloras del Jurasico inferior de la Patagonia extraandina, con algunas consideraciones sobre
Venezuela y Colombia. 159-199. In volkheimer, w. (ed. ). Bioestratigrafia de los sistemas regionales del
Jurasico y Cretacico de America del Sur, 1. Comite Sudamericano del Jurasico y Cretacico, Mendoza, 298 pp.

banerji, j. and lemoigne, y. 1987. Significant additions to the Upper Triassic flora of Williams Point,
Livingston Island, South Shetland Islands (Antarctica). Geobios, 20, 469-487.
bonetti, m. i. r. 1963. Florula mesojurasica de la Zona de Taquetren (Canadon del Zaino), Chubut. Revista
del Museo Argentino de Ciencias Naturales ‘ Bernardino Rivadavia Paleontologia, 1, 23-43.

— 1974. Florula Jurasica del Lago Argentino (Santa Cruz). Congreso Argentino de Paleontologia y
Bioestratigrafia, Tucumdn, Argentina (Resumenes), 1, 9-10.

chapman, j. l. and smellie, j. L. 1992. Cretaceous fossil wood and palynomorphs from Williams Point,
Livingston Island, Antarctic Peninsula. Review of Palaeobotany and Palynology, 74, 163-192.
del valle, R. A. and fourcade, N. H. 1986. La cuenca sedimentaria pos-triasica del extremo nororiental de la
peninsula Antartica. Contribuciones cientlficas del Instituto Antartico Argentino, 323, 1-24.

DUNKER, w. 1846. Monographic der norddeutschen Wealdenbildungen. Ein Beitrag zur Geognosie und
Naturgeschichte der Vorwelt. Publisher, Braunschweig, 83 pp.
edwards, w. n. 1934. Jurassic plants from New Zealand. Annals and Magazine of Natural History, 10, 81-109.
engler, a. and prantl, L. 1898-1902. Die naturlichen Pflanzenfamilien. 1 (4). Engelmann, Leipzig.
ettingshausen, c. von 1891. Contributions to the knowledge of the fossil flora of New Zealand. Transactions
of the New Zealand Institute, 23, 237-310.

farquharson, g. w. 1983. Evolution of late Mesozoic sedimentary basins in the northern Antarctic Peninsula.
323-327. In Oliver, r. l., James, p. r. and jago, j. b. (eds). Antarctic earth science. Australian Academy of
Science and Cambridge University Press, Canberra and Cambridge, 697 pp.

— 1984. Late Mesozoic, non-marine conglomeratic sequences of northern Antarctic Peninsula (the Botany
Bay Group). British Antarctic Survey Bulletin, 65, 1-32.

— hamer, R. D. and ineson, j. r. 1984. Proximal volcaniclastic sedimentation in a Cretaceous back-arc basin,
northern Antarctic Peninsula. In kokelaar, b. p. and howells, m. f. (eds). Marginal basin geology. Special
Publications of the Geological Society, London, 16, 219-229.

feruglio, E. 1937. Dos nuevas especies de Hausmannia de la Patagonia. Notas del Museo de La Plata
(Paleontologia), 2, 125-136.

gee, c. T. 1989. Revision of the Late Jurassic/Early Cretaceous flora from Hope Bay, Antarctica.
Palaeontographica, Abteilung B, 213, 149—214.

haast, j. von 1877. Notes on the geology of the Clent Hills and Mount Somers districts in the province of
Canterbury. New Zealand Geological Survey, Reports on Geological Exploration during 1873-4, 1-19.
halle, t. G. 1913. The Mesozoic flora of Graham Land. Wissenschaftliche Ergebnisse der Schwedischen
Siidpolar-Expedition 1901-1903, 3 (14), 123 pp.

haq, b. u., hardenbol, j. and vail, p. r. 1987. Chronology of fluctuating sea levels since the Triassic. Science,
235, 1156-1167.

HARLAND, W. B., COX, A. V., LLEWELLYN, P. G., PICKTON, C. A. G., SMITH, A. G. and WALTERS, R. 1982. A geologic
time scale. Cambridge University Press, Cambridge, 131 pp.

Harris, T. M. 1931. The fossil flora of Scoresby Sound, East Greenland. Part 1. Cryptogams (exclusive of
Lycopodiales). Meddelelser om Gronland, 85, 1-102.

REES: MESOZOIC FERNS FROM ANTARCTICA

655

— 1946. Liassic and Rhaetic plants collected in 1936-38 from East Greenland. Meddelelser om Gronland ,
114, 1-38.

— 1961. The Yorkshire Jurassic flora. I. Thallophyta-Pteridophyta. British Museum (Natural History),
London, 212 pp.

herbst, r. 1964. La flora liasica de C. Meschio, Provincia de Chubut, Argentina. Ameghiniana , 3, 227-234.
1966. Revision de la flora liasica de Piedra Pmtada, Provincia de Neuquen, Argentina. Revista del Museo
de La Plata, Seccion Paleontologia , 30, 27-53.

1975. Consideraciones sobre las Dipteridaceae (Filices) del Gondwana. Actas , I Congreso Argentino de

Paleontologia y Bioestratigrafla, 1, 525-535.

— 1979. Review of the Australian Dipteridaceae. Proceedings of the Linnean Society of New South Wales,
103, 7-21.

Jefferson, T. h. 1981. Palaeobotanical contributions to the geology of Alexander Island, Antarctica.
Unpublished Ph.D. thesis, University of Cambridge.

kryshtofovich, a. 1923. Pleuromeia and Hausmannia in Eastern Siberia, with a summary of recent
contributions to the palaeobotany of the region. American Journal of Science, 5, 200-208.
lacey, w. s. and LUCAS, R. c. 1981. The Triassic flora of Livingston Island, South Shetland Islands. British
Antarctic Survey Bulletin, 53, 157-173.

MACDONALD, D. I. M., BARKER, P. F., GARRETT, S. W., INESON, J. R., PIRRIE, D., STOREY, B. C., WHITHAM, A. G.,

kinghorn, R. R. f. and marshall, j. E. A. 1988. A preliminary assessment of the hydrocarbon potential of
the Larsen Basin, Antarctica. Marine and Petroleum Geology, 5, 34-53.
mcdougall, i. and mcelhinny, m. w. 1970. The Rajmahal Traps of India - K-Ar ages and palaeomagnetism.
Earth and Planetary Science Letters, 9, 371-378.

millar, i. L., milne, a. J. and whitham, a. G. 1990. Implications of Sm-Nd garnet ages for the stratigraphy of
northern Graham Land, Antarctic Peninsula. Zentralblatt fur Geologic und Paldontologie, 1, 97-104.
moller, H. J. and halle, T. G. 1913. The fossil flora of the coal-bearing deposits of south-eastern Scania. Arkiv
for Botanik, 13, 1-45.

oishi, s. 1932. The Rhaetic plants from the Nariwa district, Prov. Bitchu (Okayama Prefecture), Japan. Journal
of the Faculty of Science, Hokkaido University, Series 4, 1, 257—380.

— and yamasita, K. 1936. On the fossil Dipteridaceae. Journal of the Faculty of Science, Hokkaido
University, Series 4, 3, 135-184.

Oliver, p. j., Campbell, J. D. and speden, I. G. 1982. The stratigraphy of the Torlesse rocks of the Mt. Somers
area (S81) mid-Canterbury. Journal of the Royal Society of New Zealand, 12, 243-271.
orlando, h. A. 1971. Las floras fosiles de Antartida occidental y sus relaciones estratigraficas. Contribuciones
del Instituto Antdrtico Argentino, 140, 1-12.

pankhurst, r. j. 1982. Rb-Sr geochronology of Graham Land, Antarctica. Journal of the Geological Society,
London, 139, 701-711.

rao, a. r. 1953. Some observations on the Rajmahal flora. Palaeobotanist, 2, 25-28.

rees, p. M. 1990. Palaeobotanical contributions to the Mesozoic geology of the northern Antarctic Peninsula
region. Unpublished Ph.D. thesis. University of London.

— 1993. Revised interpretations of Mesozoic palaeogeography and volcanic arc evolution in the northern
Antarctic Peninsula region. Antarctic Science, 5, 77-85.

— and smellie, J. L. 1989. Cretaceous angiosperms from an allegedly Triassic flora at Williams Point,
Livingston Island, South Shetland Islands. Antarctic Science, 1, 239-248.

sengupta, s. 1988. Upper Gondwana stratigraphy and palaeobotany of Rajmahal Hills, Bihar, India. Memoirs
of the Geological Survey of India, Palaeontologia Indica, 68, 1-182.
seward, a. c. 1913. A contribution to our knowledge of the Wealden floras, with especial reference to a
collection of plants from Sussex. Quarterly Journal of the Geological Society, London, 69, 85-116.

— and dale, E. 1901. On the structure and affinities of Dipteris, with notes on the geological history of the
Dipteridinae. Philosophical Transactions of the Royal Society of London, Series B , 194, 487-513.

shah, s. c., singh, G. and gururaja, m. n. 1973. Observations on the Post-Triassic Gondwana sequence of
India. Palaeobotanist , 20, 221-237.

stipanicic, p. n. and bonetti, m. i. r. 1970a. Posiciones estratigraficas y edades de las principals floras
jurasicas argentinas. I. Floras liasicas. Ameghiniana, 7, 57-78.

1970A Posiciones estratigraficas y edades de las principales floras jurasicas argentinas. II. Floras
doggerianas y malmicas. Ameghiniana, 7, 101-118.

Thomson, m. r. A. 1981. Late Mesozoic stratigraphy and invertebrate palaeontology of the South Orkney
Islands. British Antarctic Survey Bulletin, 54, 65—83.

656

PALAEONTOLOGY, VOLUME 36

Thomson, m. r. a. and pankhurst, R. J. 1983. Age of post-Gondwanian calc-alkaline volcanism in the
Antarctic Peninsula region. 328-333. In Oliver, r. l., james, p. r. and jago, j. b. (eds). Antarctic earth
science. Australian Academy of Science and Cambridge University Press, Canberra and Cambridge, 697 pp.

— and clarkson, p. d. 1983. The Antarctic Peninsula - a late Mesozoic-Cenozoic arc (review).
289-294. In Oliver, r. l., james, p. r. and jago, j. b. (eds). Antarctic earth science. Australian Academy of
Science and Cambridge University Press, Canberra and Cambridge, 697 pp.
watson, J. 1969. A revision of the English Wealden flora. I. Charales-Ginkgoales. Bulletin of the British
Museum of Natural History , (Geology Series), 17, 209-254.

PETER MCA. REES

Department of Palaeontology
The Natural History Museum
Cromwell Road
London SW7 5BD, UK

Typescript received May 1991

Revised typescript received 6 November 1992

Present address
Department of Earth Sciences
University of Oxford
Parks Road
Oxford 0X1 3PR, UK

APPENDIX

Natural History Museum (NHM) registration numbers and corresponding original British Antarctic Survey
(BAS) station numbers assigned to specimens studied for this paper.

NHM number

BAS number

NHM number

BAS number

V. 63420

(no number)

V. 63606

D. 8951. 16

V. 63423

(no number)

V. 63607

D. 8953. 8

V. 63590

D. 8919. 1(A)

V. 63608

D. 9003.1

V. 63591

D. 8919. 1(B)

V.63609

D. 208. 1(1)

V. 63592

D.8919.1(X)

V. 63610

D. 208. 1(4)

V. 63593

D.8919.1(Y)

V.63611

D. 208. 1(5)

V. 63594

D. 8919. 2

V. 63612

D. 208. 1(6)

V. 63595

D. 8919. 3

V. 63613

D. 208. 1(A)

V. 63596

D. 8919. 4

V. 63614

D. 208. 1(B)

V. 63597

D. 8919. 5

V. 63615

D. 208. 1(C)

V. 63598

D.468.7

V. 63616

D.208.KK)

V. 63599

5826

V. 63617

D.208.KL)

V.63600

D.8868

V. 63618

D.208.1(M)

V. 63601

D. 8913. 2

V. 63619

D.208.1(P)

V. 63602

D. 8951. 5

V. 63620

D.1.1

V. 63603

D. 8951 .8

V. 63621

D. 8890.1

V. 63604

D. 8951. 9

V. 63622

D. 8890.2

V.63605

D. 8951. 12

V. 63623

D. 8890. 3

THE TEMNOSPONDYL AMPHIBIAN CAPETUS
FROM THE UPPER CARBONIFEROUS OF
THE CZECH REPUBLIC

by SANDRA E. K. SEQUEIRA and ANDREW R. MILNER

Abstract. The Carboniferous temnospondyl amphibian Capetus palustris is reassessed on the basis of new
and previously described specimens from the Gaskohle of Nyrany, Czech Republic. Capetus has frequently
been synonymized with Gaudrya, also from Nyrany, but the holotype of Gaudrya is a fragment of a large
specimen of Cochleosaurus. The phylogenetic position of Capetus within the primitive temnospondyls is
uncertain and there is no support for a cladistic relationship to the typical long-snouted edopoids. Although
a numerically rare element in the Nyrany assemblage, Capetus was one of the largest tetrapods present and was
probably a major component of the tetrapod biomass in the Nyrany fauna.

One of the richest sources of Upper Carboniferous amphibians has been the Gaskohle from the
Humboldt mine and other mines at Nyrany near Plzen in the Czech Republic. Fossil
vertebrates were collected from the Gaskohle from 1870 onwards and at least 700 amphibian
specimens have been recorded from museums in Europe (Milner 1987). Taxonomic revisions of this
fauna have reduced an initially large number of forms to about 25 monospecific genera, but some
are still poorly known or poorly understood because of the fragmentary nature of most large
specimens and the consequent difficulty of association. One such problematical series of specimens
in the Nyrany assemblage is that of the large, superficially crocodile-like temnospondyl amphibians
corresponding to the edopoid grade of organization. One form, Cochleosaurus bohemicus Fritsch,
1885, is usually readily recognizable from even fragmentary cranial material and its identity and
characteristics present few difficulties. The remaining non -Cochleosaurus material consists of
disparate large skull fragments of uncertain systematic position and this material is reviewed in the
following account.

In his original series of papers on the Permo-Carboniferous vertebrates of Bohemia (now part of
the Czech Republic), Fritsch (1885, 1901) described relatively few fragments of large temnospondyls,
and most of these were clearly referable to Cochleosaurus bohemicus. A few specimens were placed
in a second taxon Nyrania trachystoma , and one large anterior palatal fragment was made the
holotype of a third form Gaudrya latistoma. In the first half of the twentieth century, several large
temnospondyl skulls from Nyrany were described, all of which were clearly not Cochleosaurus.
Broili (1908) described and figured three Nyrany skulls (housed at Munich) referring them to
Sclerocephalus credneri which Fritsch (1901) had created for a specimen from the Permian of
Ruprechtice, Bohemia. Jaekel (1911, 1913) figured a reconstruction of a temnospondyl skull roof
based on a single specimen in the collection of Nyrany amphibians at the Museum fur Naturkunde
in Berlin, but did not describe or figure the original fossil. He referred this specimen to
Chelydosaurus vranyi, a taxon established by Fritsch (1885) for material from the Permian of
Olivetin and Ruprechtice in the Czech Republic. Steen (1938) described a further Nyrany skull
fragment at the British Museum (Natural History) as the new taxon Capetus palustris. She noted
that the largest of Broili’s three specimens was probably the same form, whilst one of the others was
a loxommatid. In their discussion of the relationships of Edops craigi from the Lower Permian of
Texas, Romer and Witter (1942) suggested that Capetus was a close relative of Edops and comprised
not only Steen’s holotype but also Broili’s largest specimen and Jaekel’s specimen. In his

(Palaeontology, Vol. 36, Part 3, 1993, pp. 657-680, 3 pis.)

© The Palaeontological Association

658

PALAEONTOLOGY, VOLUME 36

text-fig. 1. Capetus palustris Steen, 1938. BMNH R4706, latex cast of holotype specimen xO-66. For

interpretation, see Text-figure 2.

comprehensive revision of the ‘labyrinthodont’ amphibians, Romer (1947) ‘lumped’ together all
the large temnospondyl material from Nyrany which was not obviously assignable to Cochleosaurus.
This was again based on the specimens described by Broili, Jaekel and Steen but also included the
holotype palatal fragment of Fritsch’s Gaudrya latistoma and some of the material that Fritsch
referred to Nyrania trachystoma. The senior name for this combination of specimens thus became
Gaudrya latistoma and his binomen was used for the large non-Cochleosaurus temnospondyl
material from Nyrany by Romer (1947) and later authors, e.g. Langston (1953) and Milner (19806).
Romer (1966, 1968) briefly synonymized Capetus with a contemporary genus Macrerpeton from
Linton, Ohio, USA, but this synonymy was never explained or justified in print, and the two genera
are, in fact, distinct.

The discovery of an undescribed, largely complete skull of this type in Vienna in 1983, and its
subsequent preparation, has permitted the authors to associate most of the non-Cochleosaurus
specimens with greater confidence and also to reassign other specimens back to Cochleosaurus. As
a result of this revision, briefly reported by Milner and Sequeira (in press), the senior name for this
material reverts to Capetus palustris. In the following work this taxon is redescribed and its
relationships are reconsidered.

Location of specimens. BMNFL Department of Palaeontology, Natural History Museum, London, UK. BSP:
Bayerische Staatssammlung fur Palaontologie und Historische Geologie, Munich, Germany. MB: Museum fiir
Naturkunde, Humboldt Universitat, Berlin, Germany. NMW: Naturhistorisches Museum, Vienna, Austria.
UMZC : University Museum of Zoology, Cambridge, UK.

SEQUEIRA AND MILNER: CARBONIFEROUS AMPHIBIAN

659

text-fig. 2. Capetus palustris Steen, 1938. BMNH R4706, holotype. Abbreviations: a, angular;
bo, basioccipital ; eh, choana; cor, coronoid; d, dentary; ect, ectopterygoid; eo, exoccipital; f, frontal;
icl, interclavicle; it, intertemporal; j, jugal; 1, lacrimal; m, maxillary; n, nasal; p, parietal; pal, palatine;
pas, parasphenoid ; pf, postfrontal; pmx, premaxillary; po, postorbital; pp, postparietal ; pra, prearticular;
prf, prefrontal; pt, pterygoid; q, quadrate; qj, quadratojugal ; sa, surangular; se, sphenethmoid;
smx, septomaxillary ; spl, splenial; spp, postsplenial ; sq, squamosal; st, supratemporal; t, tabular; v, vomer.

Scale bar = 50 mm.

SYSTEMATIC PALAEONTOLOGY
Class AMPHIBIA

Order temnospondyli Zittel, 1888
Family incertae sedis
Genus capetus Steen, 1938

Type species. Capetus palustris Steen, 1938.

Diagnosis. As for the type (and only) species.

Capetus palustris Steen, 1938

1908
non 1908
1911
1913
1938

Plates 1-3; Text-figs 1-10

Sclerocephalus credneri Fritsch; Broili [partim] p. 55, pi. 1 figs 1, 3 [non Fritsch 1901].
Sclerocephalus credneri Fritsch; Broili, pi. 1 fig. 2.

Chelydosaurus vranyi Fritsch; Jaekel, fig. 124 [non Fritsch 1885],

Chelydosaurus vranyi Fritsch; Jaekel, fig. 5 [non Fritsch 1885].

Capetus palustris Steen, p. 241, text-fig. 27.

660

PALAEONTOLOGY, VOLUME 36

1947 Gaudrya latistoma Fritsch; Romer, p. 104, fig. 20 [ non Fritsch 1885],

1953 Gaudrya Fritsch; Langston, p. 374 [non Fritsch 1885].

1 9807? Gaudrya latistoma Fritsch; Milner, p. 453 [non Fritsch 1885].

1987 Gaudrya Fritsch; Milner, p. 506 [non Fritsch 1885].

1987 ‘Unnamed stem-eryopoid • ; Milner, p. 506.

Holotype. BMNH R4706, acid-etched mould of a skull table and cheek, in coal, figured by Steen (1938, text-
fig. 27; Text-figs 1-3).

Diagnosis. A primitive temnospondyl with a mixture of unique, derived and primitive characters
which does not readily permit it to be placed in a pre-existing family. The only possibly unique
character is: pterygoid-vomer sutures extending anterolaterally from the midline to the choanae at
about 45°. Derived characters shared with higher temnospondyls are: premaxillary antero-
posteriorly abbreviated with prominent alary process running mesial to large external naris (shared
with most Palaeozoic temnospondyls other than Edops , Cochleosaurus , Chenoprosopus and
Dendrerpeton ); jugal extending broadly anteriorly to level of anterior orbit margin (shared with
edopoid and ‘eryopoid ’-grade temnospondyls). Retained primitive characters include: lacrimal
entering orbit margin in large specimens; intertemporals present; elongate postorbital; pineal
foramen retained in large specimens; postparietal lappets absent; maxillary contacting quadrato-
jugal; vomers not elongate at level of choanae; interpterygoid vacuities narrow and terminating
anteriorly as points; pterygoids broadly suturing with vomers and probably excluding vomers from
margin of interpterygoid vacuity; large rhomboidal interclavicle.

Locality and horizon. Nyrany, 13 km southwest of Plzen, Czech Republic. Gaskohle, Nyrany Series of the Plzen
Basin, Westphalian D, Upper Carboniferous.

Referred Material

BSP. A partial skull and a fragment of snout tip, figured by Broili (1908 pi. 1, figs 1, 3). These two specimens
were probably destroyed in 1944. It is now uncertain whether they represented part of a single skull.

MB Am. 84. Acid-etched mould of a skull roof, mandible and interclavicle - the reconstructed skull roof figured
by Jaekel (1911, 1913) and Romer (1947) (PI. 1, fig. 1; Text-fig. 4).

MB Am. 92. Bitumen cast of a skull table and cheek. Location of original specimen unknown. Previously
undescribed (Text-fig. 5).

NMW 1893.32.53. Incomplete skull table and left cheek in counterpart and NMW 1895.2366, fragments of a
large snout and left mandible, all previously undescribed. These are unprepared fragments attributable to
Capetus and, despite the two catalogue numbers, represent a single specimen. Part of mandible figured here
(Text-fig. 8b).

NMW 1898.X.51. Mould of a large skull and mandibles in counterpart with some original palate material
remaining. Acid etched by S.E.K. Sequeira in 1989. Previously undescribed (Pis 2-3; Text-figs 6-7, 8c).
UMZC T.144. Previously part of the D. M.S. Watson collection as DMSW B.77. Acid-etched mould of outer
face of posterior right mandible and associated cheek. Previously undescribed (PI. 1, fig. 2; Text-fig. 8a).

Description and comparisons

General cranial features. The skull of Capetus is superficially alligator-like in shape. The most complete
specimens (MB Am. 84 and NMW 1898. X. 51) possess a moderately elongate snout accounting for about 59 per
cent of the total skull length. The skull broadens posteriorly to reach its maximum width at the level of the
anterior edge of the prominent otic notches. These demarcate the posterolateral edges of the skull table, which
is slightly wider than long, and which possesses a markedly concave posterior margin. Dermal bone surfaces
have typical temnospondyl pit and ridge ornament radiating outward from the centres of ossification. No
specimen shows any trace of lateral-line sulci or pits. The material comprises fragments or casts of incomplete
skulls, most of which were in the 230-300 mm range and had firmly interlocking sutures. The smallest skull
(MB Am. 84) is 180 mm long and may have belonged to a subadult individual as the sutures are relatively open
and some bones have slipped relative to one another (PI. 1, fig- 1 ; Text-fig. 4). Comparative cranial dimensions

SEQUEIRA AND MILNER: CARBONIFEROUS AMPHIBIAN

661

table 1. Capetus palustris. Cranial dimensions in mm. (e) = estimated dimension.

Dimension

MB Am. 84

NMW 1898.X. 51

MB Am.92

BMNH R4706

Midline skull length

180

235

Snout length-midline

110

138

—

—

snout tip to mid-orbit
Skull table length.

70

97

102(e)

114

midline mid-orbit to posterior

margin

Minimum interorbital width

53

69

74(e)

79

Orbit width

28

33

36(e)

37

Midline to tabular tip

40

54

64

65

are given in Table 1. Some fragmentary material (NMW 1893.32.53 and NMW 1895.2366) represents a
distinctly larger skull, perhaps 400 mm in length.

Skull roof. The dermal skull roof of Capetus comprises the full complement of dermal bones typical of primitive
temnospondyls and only elements of particular systematic significance are described.

The configuration of the premaxillaries is unusual among early temnospondyls. Each premaxillary forms a
narrow border along the anterior edge of the snout, and extends a pronounced wedge-like alary process
posteriorly over the anterior edge of the nasal. The alary process is situated slightly mesial to the medial edge
of a large external naris in MB Am. 84 (Text-fig. 4) and NMW 1898.X.51 (Text-fig. 6). Edopoids such as Edops ,
Cochleosaurus and Chenoprosopus possess much larger premaxillaries, which lack alary processes and which
border small inset external nares. Dendrerpeton appears to possess narrow premaxillaries with no alary
processes. In the more derived temnospondyls of the eryopoid grade ( sensu Milner 1990a), the premaxillaries
are similar in shape to those of Capetus but the alary processes may either border the nares or lie medially to
them (e.g. Sclerocephalus Boy, 1988; Onchiodon Boy, 1990).

The large nasals are about twice as long as wide. They occupy most of the surface area of the anterior snout
and extend anteriorly between the posterior ends of the premaxillaries. The external naris is bordered
posteriorly by a small septomaxillary which probably sutured with the anterior edge of the lacrimal. The
septomaxillary is visible as a disarticulated element in the left naris of MB Am. 84 (Text-fig. 4). The lacrimal
extends posteriorly to enter the anterior orbit margin and, in doing so, separates the prefrontal from the jugal.

text-fig. 3. Capetus palustris Steen, 1938. a, Steen’s reconstruction in which right postfrontal and frontal were
combined as a massive right postfrontal, giving an unusually large interorbital width and an apparent
resemblance to Edops. b, new reconstruction with frontals and postfrontals identified correctly, giving a skull

of more Sclerocephalus- like proportions.

662

PALAEONTOLOGY, VOLUME 36

text-fig. 4. Capetus palustris Steen, 1938. MB Am. 84, the specimen figured by Jaekel (1911, 1913) as
Chelydosaurus vranyi. For list of abbreviations, see Text-fig. 2. Scale bar = 50 mm.

This is visible in BMNH R4706 (Text-fig. 2), MB Am.84 (Text-fig. 4), NMW 1898.X.51. (Text-fig. 6) and one
of the lost BSP specimens figured by Broili (1908, pi. 1 fig. 1). This is the primitive tetrapod condition and is
an unusual symplesiomorphy to be found in a long-snouted temnospondyl. A thickened ridge extends from the
posterior end of the lacrimal anteriorly to a point partway along the lacrimal-maxillary suture. The two ridges

EXPLANATION OF PLATE 1

Capetus palustris Steen, 1938. 1. MB Am.84, silicone-rubber cast of specimen xO-66. For interpretation, see
Text-figure 4. 2. UMZCT.144, silicone-rubber cast of posterior region of right mandible xl. For
interpretation, see Text-figure 8a.

PLATE 1

SEQUEIRA and MILNER, Capet us

664

PALAEONTOLOGY, VOLUME 36

text-fig. 5. Capetus palustris Steen, 1938. MB Am. 92. This specimen is now represented only by a bitumen
cast, the location of the original being unknown. For list of abbreviations, see Text-fig. 2. Scale bar = 50 mm.

form a pair of parallel struts, which probably resisted part of the torsional stress generated between the anterior
dentition and the jaw hinge during feeding.

The frontals make the greatest contribution to the interorbital width, which is not great in any specimen
(. contra Steen 1938). Steen described and reconstructed the holotype specimen of Capetus as possessing an
interorbital width greater than the skull table length. This was based on a misinterpretation, in which the right
frontal and postfrontal bones were amalgamated as a postfrontal. That error was then compounded in a
mirror-image reconstruction with the midline suture misidentified (Text-fig. 3a). The resulting short skull table
and Edops- like widely spaced orbits were both artefacts of this misinterpretation, but may well have influenced
Romer and Witter (1942) when they associated Capetus with Edops. The skull of Capetus in fact bears a much
closer resemblance to that of Sclerocephalus in general shape and Text-fig. 3b depicts it reconstructed correctly.

The lateral edges of the orbits are bordered by a jugal, which is different in shape from those of most other
primitive temnospondyls. It intervenes broadly between the orbit and the maxillary and is as wide as, or wider
than, the orbit for most of its length. Anterior to the level of the orbits, the jugal narrows to a point along its
extensive common suture with the lacrimal. A similar condition pertains in Edops. The cheek margin is
bordered largely by the relatively slender maxillary, which narrows posteriorly to a point contact with the
quadratojugal, excluding the jugal from the skull margin (Text-figs 4, 6, 8). Cheek depth is never greater than
interorbital width and increases to a maximum towards the jaw articulation.

The circumorbital series is conservative for temnospondyls and is unremarkable apart from the right
postorbital in BMNH R4706 (Text-fig. 2). In this specimen, a pronounced lateral extension of this bone is
wedged into the medial face of the jugal just below the posterior orbital edge; postorbitals in other Capetus

SEQUEIRA AND MILNER: CARBONIFEROUS AMPHIBIAN

665

skulls are much less expanded at this point, so this may represent individual variation. The postorbitals of all
specimens are primitively anteroposteriorly elongate, and are substantially longer than wide, wedging
posteriorly between the supratemporal and squamosal.

The skull table structure is characteristic of primitive temnospondyls. A pair of intertemporals is present;
their shape and size varies within the material studied. In BMNH R4706 and MB Am. 92, they are about two-
thirds of the area of the supratemporals, but in MB Am. 84 they are relatively smaller. The parietals are large
and anteroposteriorly elongated; a small pineal foramen is located two-thirds of the way back along the
interparietal suture. The postparietals lack the lappets characteristic of Cochleosaurus and the unornamented
posterior face of each postparietal is a narrow backward-sloping surface with little evidence even of thickenings
associated with the underlying exoccipitals. The tabulars are anteroposteriorly narrow and grow
posterolaterally with size increase to enhance the concave curvature of the occiput. Each otic notch is deeply
ovoid and is bordered by the tabular, supratemporal and squamosal. A 2 mm wide margin of dense fine pitting
lines the squamosal border of the notch in BMNH R4706, but is not present in the smaller MB Am. 84.

Sclerotic ring. A partial sclerotic ring is preserved in the right orbit of NMW 1898.X.51 (Text-fig. 6). Eight
sclerotic plates appear to make up about one third of a ring, which suggests a normal temnospondyl sclerotic
ring of about twenty-five elements (Milner 1982).

Palate. Significant regions of the palate are visible only in NMW 1898.X.51 (Text-fig. 7) among the surviving
specimens, and even in this skull, much of the left side has been obscured by overlying jaw elements. General
palatal structure in Capetus is that of a primitive temnospondyl. The interpterygoid vacuities are small; in
comparison with the snout-quadrate length, they are a quarter as long and a sixth as wide. Pronounced anterior
narrowing of each vacuity to a point occurs as the pterygoid curves towards its median articulation with the
cultriform process of the parasphenoid.

The shape of the vomers is of considerable importance in determining the phylogenetic position of Capetus.
In NMW 1898.X. 51, the vomers are not well preserved but their general shape is clear. They are not very
elongate anteroposteriorly; most notably, they lack the extreme elongation anterior to the leading edge of the
choanae which is characteristic of cochleosaurs. This area is poorly preserved but there is no space for such
an anterior extension. One of the specimens figured by Broili (1908, pi. 1, fig. 3) is an anterior snout region in
palatal aspect and also appears to bear a pair of squarish vomers, each apparently with an anterior pit or
depression, probably for the reception of parasymphyseal fangs. No such pits are visible in NMW 1898.X.51
however, so their existence must be treated with caution. In NMW 1898.X.51, each vomer contacts the
pterygoid along an extensive common suture running at an angle of 45° from the anterior choanal edge towards
the midline. Because the vomer and the pterygoid are overlain by the displaced left mandible immediately
anterior to the interpterygoid vacuity, their configuration in this region can only be guessed at by extrapolating
the common suture back to the midline. This suggests that the vomer was completely excluded from the vacuity
and, if so, this represents a primitive character-state shared with Edops and Chenoprosopus. The portion of the
vomer level with the choanae is less elongated than that found in cochleosaurs. In NMW 1898. X. 51, there
appears to be a small and uninformative exposure of the right palatine between the right mandible and the
pterygoid (Text-fig. 7). No ectopterygoid is visible in any of the available specimens.

The broad pterygoid expands posterolaterally into a wide flange at the level of its articulation with the
braincase. In NMW 1898.X.51, elements of the single circular occipital condyle are preserved posterior to the
crushed braincase. In MB Am. 84, a series of crushed elements are superimposed on the left tabular and vicinity.
They are too damaged to merit description or to be figured in detail, but appear to represent a displaced part
of the braincase and occiput. Rod-like elements associated with this crushed material may represent the
paroccipital processes.

Mandible. Incomplete mandibles are preserved in NMW 1898. X. 51 (Text-figs 6-7), a portion of the left dentary
is present in MB Am. 84 (Text-fig. 4), the anterior of the left mandible is preserved in NMW 1895.2366 (Text-
fig. 8b) and the external face of the posterior region of the right mandible is represented in UMZC T.144 (Text-
fig. 8a). The mandibular rami are robust and deep, reaching a maximum vertical depth midway along the
extensive angular. In this region, mandible depth equals, and may exceed, cheek depth. The relationship of the
elements on the external face of the mandible (Text-fig. 9c) resembles that in Sclerocephalus (Boy 1988, fig. 6a)
and Eryops (Sawin 1941, pi. 5). The angular has a sharply curved ventral margin, like those in the above deep-
jawed genera but unlike many temnospondyls with flatter mandibular rami where the ventral margin is
relatively straight. Part of the medial face of the right mandible is exposed in NMW 1898.X.51 but little sutural
detail is visible.

666

PALAEONTOLOGY, VOLUME 36

text-fig. 6. Capetus pahistris Steen, 1938. NMW 1898. X. 51 part. Stippled area behind skull table represents
a patch of ossicle-bearing skin (see Text-fig. 8c). For list of abbreviations, see Text-fig. 2. Scale bar = 50 mm.

EXPLANATION OF PLATE 2

Capetus palustris Steen, 1938. NMW 1898. X. 51, silicone-rubber cast of main part x 0-66. For interpretation,
see Text-figure 6.

PLATE 2

SEQUEIRA and MILNER, Capetus

668

PALAEONTOLOGY, VOLUME 36

sa

text-fig. 7. Capetus palustris Steen, 1938. NMW 1898.X.51 counterpart. Stippled area is internal surface of
skull roof. For list of abbreviations, see Text-fig. 2. Scale bar = 50 mm.

EXPLANATION OF PLATE 3

Capetus palustris Steen, 1938. NMW 1898.X.51, silicone-rubber cast of counterpart x 0-66. For interpretation,
see Text-figure 7.

PLATE 3

SEQUEIRA and MILNER, Capetus

670

PALAEONTOLOGY, VOLUME 36

Dentition. Marginal teeth, palatal fangs and denticles are visible in MB Am.84, NMW 1895.2366,
NMW 1898.X. 51, UMZC T.144 and the two specimens figured by Broili. The teeth and fangs are simple sharp-
pointed cones of typical labyrinthodont appearance. They are dagger-like with no suggestion of curvature,
although the posterior teeth may be slightly backwardly-directed. In the large NMW 1895.2366, the anterior
dentary teeth appear to be genuinely flattened bilaterally and may have had a slight keel.

In MB Am.84 there is space for about fifty marginal teeth on the left upper jaw ramus. There appears to have
been space for about twelve premaxillary teeth and thirty-eight maxillary teeth, the latter growing to a height
of 15 mm in NMW 1898. X. 51. A precise dentary tooth count is not possible, and the presence or absence of
coronoid teeth cannot be ascertained. NMW 1898.X.51 shows some evidence of pseudocanine peaking with
two or three enlarged teeth around the premaxillary-maxillary junction. NMW 1895.2366 appears to have
some enlarged teeth at the very anterior end of the dentary (Text-fig. 8b), but these may be symphyseal fangs
crushed into the same plane as the marginal dentition.

In NMW 1898.X. 51, one or two vomerine fangs border the anteromedial edge of each choana. These fangs
are no larger than the marginal teeth. A poorly-defined structure which might be a much larger fang lies
displaced across the right palatine. The presence of relatively large palatine fangs is a feature of some primitive
temnospondyls. The presence or absence of palatine tooth-rows and ectopterygoid fangs or teeth cannot be
established.

In NMW 1898.X.51, a thin scatter of denticles covers the palatal surfaces of the vomer and pterygoid,
becoming most concentrated towards the medial edge of the latter. There are also patches of denticulate bone
in the interpterygoid region, but it is not clear whether these are part of the parasphenoid or isolated denticle-
bearing plates covering the interpterygoid region.

Stapes. A small bone fragment abutting the left otic notch in MB Am.84 (Text-fig. 4) may represent a portion
of the shaft of the left stapes.

Interclavicle. A large rhomboidal mterclavicle with a length: width ratio of 1-47: 1 is preserved in MB Am.84
(Text-fig. 4). It has been turned over relative to the skull, and also anteroposteriorly reversed by rotation
through 180°. The pattern of dermal ornament on its ventral face resembles that of postmetamorphic specimens
of Sclerocephalus ( Boy 1988, fig. 9) with pitting over the centre of ossification grading into radial grooves along
the margins. Large articular surfaces for the clavicles lie posterolateral to the strongly fimbriated anteromedial
edge of the interclavicle. Facet outlines suggest that the clavicles must have been broad-bladed structures.

Dermal ossicles. A flap of ossicle-bearing skin appears to have been preserved immediately behind the skull
table of NMW 1898. X. 51. This can be seen as a series of fine ridges and folds in the surface of the cast
immediately behind the postparietals (PI. 2). In places, these folds can be seen to incorporate small disc-like
structures which may be interpreted as osteoderms in the skin of the back (Text-fig. 8c). They are oval and bear
a pattern of fine concentric rings. They are only 4-5 mm in diameter but otherwise resemble the 10 mm
osteoderms described in a specimen of Eryops by Romer and Witter (1943).

Reconstruction. The reconstructions of the cranium and mandible in Text-figure 9 are composite and are based
on the general configuration and palate of NMW 1898.X.51, augmented with skull roof details from
BMNH R4706 and mandibular detail from UMZC T.144 and NMW 1895.2366. The reconstruction of the
smaller skull roof in Text-figure 10 is based solely on MB Am.84. These reconstructions differ in several details
from the provisional reconstructions previously produced by Milner and Sequeira (in press) and supersede
them. The general shape and depth of the skull is most similar to the slightly later Sclerocephalus from the
Lower Permian of Germany (Boy 1988).

DISCUSSION

Relationship of Capetus to other primitive temnospondyls

The following discussion of characters and character-states has the limited aim of comparing the

text-fig. 8. Capetus palustris Steen, 1938. A, UMZC T.144, posterior end of right mandible.
B, NMW 1895.2366, anterior end of left mandible, c, NMW 1898.X.51, dermal ossicles immediately behind
skull table. This figure represents the best-preserved part of the stippled area depicted in Text-figure 6. For list
of abbreviations, see Text-fig. 2. Scale bars = 10 mm.

SEQUEIRA AND MILNER: CARBONIFEROUS AMPHIBIAN

671

672

PALAEONTOLOGY, VOLUME 36

A B

text-fig. 9. Capetus palustris Steen, 1938. Whole skull reconstruction of large specimen, in a dorsal aspect,
b palatal aspect, and c right lateral aspect with mandible. Based on NMW 1898.X.51, BMNH R4706 and
UMZC T.144. For list of abbreviations, see Text-fig. 2. Scale bar = 50 mm.

characteristics of a small range of primitive temnospondyls, mostly superficially crocodile-like
forms, in order to establish their relationships. The genera involved (with sources in parenthesis)
are Capetus (redescribed here), Edops (Romer and Witter 1942 and authors’ personal obser-
vations), Cochleosaurus (Rieppel 1980, Godfrey and Holmes in press, and authors’ personal
observations), Chenoprosopus (Langston 1953), Trimerorhachis (Case 1935 and authors’ personal
observations), Selerocephalus (Boy 1988) and Onchiodon (Boy 1990). Outgroup comparison is made
with the loxommatids (Beaumont 1977). Comparison is not made with either Caerorhachis or
Dendrerpeton as the only skull of the former is both small and poorly preserved, while unpublished
work on the latter genus by the authors suggests that it is composed of at least two distinct taxa,
which precludes it from being a suitable outgroup at present. Unlike Boy (1990), we have not used
either trimerorhachoids or dissorophoids as outgroups, as both are believed to be more derived
temnospondyls than the edopoids, and dissorophoids are believed to be more derived than both

SEQUEIRA AND MILNER: CARBONIFEROUS AMPHIBIAN

673

q

qj

text-fig. 10. Capetus palustris Steen,
1938. Whole skull reconstruction of
medium specimen in dorsal aspect, based
on MB Am. 84. Scale bar = 50 mm.

edopoids and eryopoids (Milner 1990a, 19906). The character number prefixes relate to the
cladogram in Text-figure 11.

Edopoid characteristics

The Superfamily Edopoidea is a primitive clade of temnospondyls (Milner 19906) consisting of two
families: the Cochleosauridae, comprising Cochleosaurus and Chenoprosopus and defined below;
and the Edopidae, a monotypic family based on Edops craigi from the Lower Permian of Texas
(Romer and Witter 1942).

Edops shares the following with the Cochleosauridae.

ED.l. Enlarged premaxillaries with a long common median suture and extending well back along
the jaw margin behind the level of the common suture. Small naris set well back along snout
(Milner 19906). The small size of the naris is probably primitive, and its position is a
manifestation of the premaxillary shape. The pattern of enlargement of the snout by
expansion of the premaxillaries occurs elsewhere among primitive tetrapods only in some
urocordylid nectrideans where it is clearly a convergence (e.g. Sauropleura).

ED. 2. Jugal-prefrontal contact excluding lacrimal from the orbit margin (Milner 1980a, 19906).
This appears to be a derived character occurring in most but not all long-snouted
temnospondyls. It defines the Edopoidea against many other temnospondyls including
Dendrerpeton , Trimerorhachis , the East Kirkton temnospondyl, the Dissorophoidea and the
Zatrachydidae. However, it also occurs in the Eryopoidea and Parioxyidae, and in the
Actinodontidae and all its stereospondyl relatives.

One character used incorrectly to define the Edopoidea by Milner (1980a) followed by Godfrey
et al. (1987) was the presence of sculptured triangular septomaxillary sutured into the dermal roof
of the snout behind the external naris. This has proved to be more widely distributed among
temnospondyls, though seldom figured, and may characterize all except the eryopid-
dissorophoid-lissamphibian clade, where it is replaced by a free septomaxillary in the naris.

674 PALAEONTOLOGY, VOLUME 36

text-fig. 11. Cladogram of relationships of the temnospondyl genera discussed in the text. Numbers denote

characters as used in the text.

Cochleosaurid characteristics

Recent workers (Milner 19906; Godfrey and Holmes in press) have considered the Cochleosauridae

to be restricted to two genera and three species, namely Cochleosaurus bohemicus, C. florensis and

Chenoprosopus milleri. The following list of derived characters defining the family is modified from

that provided by Godfrey and Holmes (in press).

CO.l. Closure of the pineal foramen in adults (Steen 1938; Langston 1953; Godfrey and Holmes
Character 1).

CO. 2. Possession of depressed areas on the skull roof that exhibit subdued sculpturing (Rieppel
1980; Godfrey and Holmes Character 2).

CO. 3. Elongate anterior region of the palate produced by elongation of both the vomers and the
internal narial openings, resulting in the relatively posterior position of the anterior border
of the interpterygoid vacuities (Godfrey and Holmes Character 3 modified; Godfrey and
Holmes also cited the extremely large premaxillaries on the dorsal surface of the skull as part
of this character but these also occur in Edops - see below).

CO. 4. Ectopterygoid and maxillary excluded from the subtemporal fossa by a pterygoid-jugal
suture (Godfrey and Holmes Character 4). An inevitable correlate of this character is that
the jugal also separates the maxillary and the quadratojugal on the cheek. Milner (1980a)
attributed the latter character to all the long-snouted edopoids but there is a contact between
maxillary and quadratojugal in Edops , albeit a point contact.

CO. 5. Choanae wider anteriorly than posteriorly (Godfrey and Holmes Character 6).

Two other derived characters used by Godfrey and Holmes are of value only in defining the

Cochleosauridae within the Edopoidea, as they occur widely elsewhere within the Temnospondyli.

Transverse width of the skull through the mid-orbital region is less than the antorbital length
(Godfrey and Holmes Character 5).

Squamosal embayment ( = otic notch or temporal notch) with substantial participation of
supratemporal along its dorsal rim (Godfrey and Holmes Character 7).

SEQUEIRA AND MILNER: CARBONIFEROUS AMPHIBIAN

675

Post-edopoid characteristics

The following characteristics define the trimerorhachoid and eryopoid grades of temnospondyls,
and their descendent clades : the Brachyopoidea, Stereospondyli, Dissorophoidea and Lissamphibia.

PE. 1. Interpterygoid vacuities wide and anteriorly rounded.

PE. 2. Palatine rami of the pterygoids so reduced that the vomers enter the margin of the vacuities.
PE. 3. Prominent alary process (pars dorsalis) on the premaxillary which is a relatively narrow bone
bordering a large naris. Edopoids retain a simple massive premaxillary with a straight
posterior edge like that of loxommatids. Post-edopoid temnospondyls have relatively larger
nares and smaller premaxillaries. They include the East Kirkton temnospondyl, trimero-
rhachoids (not figured by Case 1935 but found in first-hand study by the authors) and all
eryopids, dissorophoids and stereospondyls. Milner (19906) suggested that this derived
character defined all temnospondyls with the situation in edopoids being a reversal, but it
appears to operate without contradiction at this higher node with the edopoid condition
regarded as primitive by outgroup comparison with loxommatids.

Eryopoid characteristics

The Eryopoidea is here taken as a grade of temnospondyl at the base of both the clade
Stereospondyli and the dissorophoid-lissamphibian clade (Milner 1990a, 19906). This grade
includes the stem-stereospondyl families Actinodontidae, Intasuchidae and Archegosauridae; and
the stem-dissorophoid families Zatrachydidae, Eryopidae and Parioxyidae. The defining characters
for the Eryopoidea define a node including all the descendent taxa.

ER. 1. Intertemporals absent (Milner 1990a). The loss of the intertemporals has demonstrably
occurred within several early tetrapod groups (e.g. Loxommatidae, Colosteidae) but is here
taken as a valid character in the absence of contradictory characters.

ER.2. Exoccipitals enlarged to give bilobed occipital condyle.

Edopoid-eryopoid relationships

Boy (1990) proposed a set of relationships for the edopoid and eryopoid temnospondyls in which
Edops is the sister-taxon to a clade of the Eryopidae, Parioxyidae and Zatrachydidae (hereafter the
E-P-Z group) while Chenoprosopus (and by implication Cochleosaurus) is the sister-taxon to the
Actinodontidae and Archegosauridae (hereafter A-A group). Boy’s E-P-Z grouping corresponds to
those ’eryopoid’ families which Milner (1990a, 19906) placed at the base of the dissorophoid-
lissamphibian clade, and their unity is not disputed. Likewise the A-A group were placed by Milner
(1990a, 19906) at the base of the Stereospondyli and their unity is also supported. However Boy's
suggestion that the Edopoidea and eryopoid-grade temnospondyls are both diphyletic is not
supported, for the following reasons.

1. It uses as outgroups, the Dissorophoidea, which Godfrey et al. (1987) and Milner (1990a,
19906) have argued to be more derived than the Edopoidea within the Temnospondyli, and the
Trimerorhachoidea which Godfrey et al. (1987) argued to be the sister-group of the Edopoidea, and
Milner (19906) argued to be more derived than the Edopoidea.

2. Boy (1990, p. 306, character 25) associated Chenoprosopus with the A-A group on the presence
of a nasal-maxillary contact excluding the lacrimal from the septomaxillary or naris margin.
However, the lacrimal does reach the septomaxillary in Cochleosaurus florensis (Godfrey and
Holmes in press), in Cochleosaurus bohemicus and in an undescribed cochleosaurid from Linton. As
the unity of these forms with Chenoprosopus in the Cochleosauridae is well demonstrated, we
conclude that the nasal-maxillary contact in Chenoprosopus must be convergent with that in the
Actinodontidae and the Archegosauridae, probably associated with snout elongation within each
group.

676

PALAEONTOLOGY, VOLUME 36

3. Boy (1990, p. 306, characters 4 and 26) associated Chenoprosopus with the A-A group on the
presence of an elongate prefrontal which is anteriorly constricted, whereas Edops was associated
with the E-P-Z group on the presence of an elongate prefrontal which is anteriorly expanded.
Clearly the elongate nature of the prefrontal cannot define either group and the constricted and
expanded alternatives cannot both be derived, so that at best only one of Boy’s two groupings might
be supported by this character. The character is correlated with general snout width and one can
argue on ontogenetic grounds that the narrow prefrontal found in juveniles is more likely to
represent the primitive condition, in which case only the character-state of anteriorly broad
prefrontal could be used to unite Edops and the E-P-Z group. It may be noted that other unrelated
broad-snouted temnospondyls also possess anteriorly broad prefrontals (e.g. Parotosuchus and
Cyclotosaurus (see Welles and Cosgriff 1965, figs 16, 20, 27) and this character could be argued to
be a convergent feature of broad-snouted temnospondyls.

4. Boy (1990, character 5) also united Edops and the E-P-Z group on the presence of a vertically
orientated ilium with an expanded dorsal region, but noted that this also occurs convergently in
some dissorophoids. It is thus a convergent feature of large terrestrial temnospondyls, and not a
particularly compelling character.

Thus the characters used by Boy to unite Chenoprosopus with the A-A group are argued not to
be valid, but the characters used to unite Edops with the E-P-Z group may be valid although one
occurs convergently elsewhere.

The two characters used by Boy to unite Edops with the E-P-Z group must be set against the five
characters listed here to unite post-edopoid temnospondyls, namely PE. 1-3 and ER.1-2.

The position of Capetus

Capetus palustris cannot be placed within the Cochleosauridae as it lacks characters CO. 1-5 listed
above as defining that family. The pineal foramen is retained, there are no areas of subdued
ornament, the vomers are not elongate, the jugal does not extend onto the cheek edge or palate and
the choanae are not of cochleosaurid shape. Capetus cannot be placed within the Edopoidea as it
lacks the characters ED. 1-2 defining that superfamily. The premaxilla is not enlarged and the
lacrimal is not excluded from the orbit margin by a jugal-prefrontal contact. Capetus does appear
to exhibit pseudocanine peaking, a character which it shares with Edops but this is contradicted by
the absence of other edopoid characters, and is a homeoplastic feature found in several large
temnospondyls. At most, Capetus could be a stem-edopoid (i.e. an earlier offshoot of the
superfamily than Edops) but we can find no derived character to support such a position. Capetus
cannot be placed in the Superfamily Edopoidea where it has previously resided.

Capetus certainly lacks PE.l and ER.l, probably lacks PE. 2 (ER.2 is unknown) and cannot be
associated with the Eryopoidea. However, it does possess PE. 3 - the narrow premaxillaries each
with a large alary process. This derived character suggests a position on the post-edopoid side of
the edopoid-post-edopoid dichotomy. It also occurs in the Trimerorhachidae and the undescribed
temnospondyls from East Kirkton, which would thus also be on the post-edopoid branch. In this
position, Capetus could not be placed readily in any pre-existing family, and to create a family for
Capetus alone would be to create a redundant taxon. Our solution is to treat Capetus as a plesion
sensu Patterson and Rosen 1977, i.e. an extinct holophyletic taxon of any size which is the sister-
taxon to a clade incorporating living taxa. Capetus would stand as a plesion on the stem of the
Lissamphibia, ranked after the Edopoidea but before the Stereospondyli.

However, one other distinctive feature of Capetus may contradict this position. Capetus is
unusual among long-snouted temnospondyls in retaining the character-state of the lacrimal entering
the orbit margin. In both edopoids and ‘eryopoid ’-grade temnospondyls, the jugal and prefrontal
meet to exclude the lacrimal. It might be argued that this is a shared derived character for the
Edopoidea and Eryopoidea, and that Capetus must therefore represent a more primitive side-
branch prior to the edopoid-eryopoid dichotomy. This, however, is an artefact of the restricted
range of taxa used in this comparison, in the interests of comparing homoeomorphic forms. The

SEQUEIRA AND MILNER: CARBONIFEROUS AMPHIBIAN

677

lacrimal enters the orbit margin in Trimerorhachis (Case 1935), the East Kirkton temnospondyl, the
zatrachydids and the dissorophoids, all of which are post-edopoid temnospondyls. Clearly some
convergence or reversal of this character has occurred. One possibility is that the edopoids and
eryopoids have separately excluded the lacrimal from the orbit margin, with the zatrachydids and
dissorophoids reversing the character. An alternative is that edopoids, stereospondyls and eryopids
have separately excluded the lacrimal from the orbit margin and that other groups retain the
primitive condition.

Conclusion

Capetus palustris is a primitive temnospondyl which cannot be placed readily in any pre-existing
family or superfamily. It is clearly not a member of the clade Edopoidea, but may be a very primitive
relative of the temnospondyls of the eryopoid grade of organization. Pending more comprehensive
character analysis of all primitive temnospondyls, which the authors plan to undertake, it is
considered to be a plesion within the ranked series of temnospondyl plesions leading to the
Lissamphibia.

SYSTEMATIC POSITION OF GAUDRYA LATISTOMA

Fritsch (1885, p. 31) created the taxon Gaudrya latistoma for a large anterior palate in the private
collection of Hr Cajetan Bayer of Plzen. Fritsch made plaster casts of the specimen; one is in his
original collection (NMP Fritsch Gypskopie 304) and others are widely distributed in museums.
The current location of the original specimen is not known to the authors; it does not appear to be
in the collection at the Narodni Muzeum, Prague or the Zapadoceske Muzeum at Plzen. The
systematic position of Gaudrya latistoma must therefore be determined from the cast and from
Fritsch’s illustrations (1885, pi. 61, fig. 1 (the specimen), figs 2-3 (sections of the teeth)).

The specimen comprises paired vomers in dorsal aspect bordered by the palatal component of the
premaxillaries including the premaxillary dentition in transverse section. In parts, the vomerine
bone is missing, revealing a natural mould of the ventral surface of the vomers. Each premaxillary
was anteroposteriorly elongate along the jaw margin and had space for 19 to 20 teeth. The small
elements figured by Fritsch as vomers appear to be palatal shelves (pars palatina) of the
premaxillaries. The large elements labelled palatines by Fritsch are the vomers (Text-fig. 12a). They
are unusually elongate from the level of the leading edge of the choana backwards, a condition
shared with Cochleosaurus and Chenoprosopus. The palatal surface is covered in tiny denticles. The
choana has a straight anterior margin perpendicular to the long axis of the skull. The anteromedial
margin curves sharply through about 120° to a straight, posterolaterally directed medial margin
extending back to the palatine. There is no evidence of intervomerine vacuities or depressions.
Scaled against the vomers of several specimens of Cochleosaurus , the Gaudrya vomers appear to have
belonged to a skull of between 185 and 210 mm total midline length.

Subsequently, the specimen was treated as indeterminate by Steen (1938, p. 261) and was ignored
by Romer and Witter (1942) in their discussion of Nyrany edopoids. However, Romer (1947, p. 104)
associated all large Nyrany temnospondyl specimens, not already in Cochleosaurus , under the
binomen Gaudrya latistoma as the senior binomen. Subsequent authors have accepted this
synonymy (e.g. Langston 1953; Milner 19806). No further species have been incorporated in the
genus.

Comparison of the cast of the holotype of Gaudrya with specimens of Cochleosaurus and the new
material of Capetus leads to the following three observations.

1. In the holotype of Gaudrya latistoma , the premaxillary ventral margin and tooth row extend
well back and imply the presence of anteroposteriorly elongate premaxillaries. This is character
ED. 1 described above and occurs in Cochleosaurus at Nyrany,. and also in edopoids such as Edops
and Chenoprosopus but does not occur in Capetus as can be seen in MB Am. 84 and
NMW 1898.X. 51.

2. In the same holotype, the vomers are very elongate behind the level of the anterior edge of the

678

PALAEONTOLOGY, VOLUME 36

A B C

text-fig. 12. Restored vomers of large long-snouted temnospondyls in the Nyrany assemblage. A, holotype of
Gaudrya latistoma Fritsch, 1885 (based on his plate 61, fig. 1). b, Cochleosaurus bohemicus Fritsch, 1885
(original figure based on NMW 1898.X.41). c, Capetus palustris Steen, 1938 (original figure based on

NMW 1898.X. 51).

choana (Text-fig. 12a). Such elongation in the posterior region of the vomer is unusual in
temnospondyls. It is character CO. 3 listed above for the Cochleosauridae and occurs in
Cochleosaurus (Text-fig. 12b) at Nyrany, and also in Chenoprosopus but does not occur in Capetus
as can be seen in NMW 1898. X. 51 (Text-figs 7-8, 12c).

3. If Edops- like palatal proportions are assumed, the Gaudrya snout-tip would have belonged to
a skull of at least 400-500 mm midline length and this large size may have influenced Romer to
associate it with the large Capetus , as none of the then-described Cochleosaurus skulls exceeded
250 mm in length. However, given that the elongate shape of the Gaudrya vomers corresponds to
those of Cochleosaurus in which they are proportionately large, the Gaudrya specimen could have
belonged to a Cochleosaurus skull of 185-210 mm length as noted above. The largest Cochleosaurus
specimen seen by the authors is MB Am. 85, a skull table with the diagnostic postparietal lappets.
This specimen, scaled against the skull tables of more complete skulls, appears to have belonged to
a skull of about 260 mm midline length. Thus the size of the Gaudrya holotype falls within the
known range for Cochleosaurus.

These features combine to suggest that the holotype of Gaudrya latistoma is simply a fragment of
a large specimen of Cochleosaurus bohemicus from the same locality. It is formally proposed that
Gaudrya latistoma be treated as a junior subjective synonym of Cochleosaurus bohemicus.

CAPETUS AND THE NYRANY ASSEMBLAGE

At least eight specimens of Capetus were collected, although the two at Munich were apparently
destroyed in 1944 and one in Berlin is represented only by a cast. These eight specimens form part
of an assemblage of at least 700 tetrapod specimens (Milner 1987). Capetus thus constitutes a
numerically small component (about one per cent) of the preserved tetrapod assemblage, although
it is more common than most other vertebrates of comparable size. The largest specimens had a
skull length of about 400 mm, suggesting an animal of about 1-5 m total length. Capetus would thus
have been one of the largest tetrapods in the Nyrany assemblage, and a few individuals would have
represented a significant fraction of the tetrapod biomass at any one time.

There are four tetrapods with superficially crocodile-like skulls in the Nyrany assemblage, namely
the temnospondyls Cochleosaurus and Capetus. and the loxommatids Baphetes and Megalocephalus.

SEQUEIRA AND MILNER: CARBONIFEROUS AMPHIBIAN

679

This diversity presents an interesting problem in understanding the niches which these forms might
have filled. Of the four genera, the two loxommatids are each represented by a single determinate
specimen (the Megalocephalus specimen is undescribed) and are clearly exotic elements in the
Nyrany assemblage. As suggested by Milner (1980fi, 1987) they were both small individuals in
comparison with representatives of these genera in the British Coal Measures, and it is reasonable
to assume that they were juveniles which had strayed from their typical habitat. The British
loxommatid material tends to be associated with fish-rich assemblages from large abandoned
channels, and it may be that the Nyrany loxommatids also typically lived in such a habitat. Baphetes
was broad-snouted and Megalocephalus had a longer narrower snout, and they may thus have been
the counterparts to alligators and crocodiles respectively in larger water bodies in the Late
Carboniferous.

Of the temnospondyls, Cochleosaurus is represented by nearly 100 out of a sample of 700 tetrapod
specimens, whereas Capetus is represented by eight. However, of the 100 or so specimens of
Cochleosaurus , only about ten are of individuals of comparable size to the Capetus remains, the rest
being smaller. Very small Cochleosaurus with 25 mm skulls may still be recognized by the large
premaxillaries and by the swellings on the postparietals marking the early ontogenetic stage of the
postparietal lappets which characterize the adults. This Cochleosaurus material is currently being
studied by the senior author, and no small Capetus specimens have been found in it. Consequently,
the assemblage contains a small number of presumed adults of each genus, together with a much
larger number of larvae and juveniles of Cochleosaurus but no juveniles of Capetus. Clearly
Cochleosaurus was reproducing in the Nyrany water-body and passing much of its early life-history
there, the varying sizes of juvenile suggesting that several age-classes were represented. The absence
of Capetus juveniles of any size indicates that it was not using this water-body as a breeding site.

Large individuals of Cochleosaurus and Capetus had skulls of different shape, the largest
Cochleosaurus having a narrow elongate snout, whereas that of Capetus was much broader and
slightly shorter. If, as with the loxommatids, the analogy with modern crocodilians is pursued, then
Capetus would have been a more alligator-like form feeding on aquatic and terrestrial slow-moving
tetrapods, while Cochleosaurus was more crocodile-like, perhaps specializing in predation on
swimming tetrapods.

Acknowledgements. For permission to study specimens in their care we thank Dr Angela C. Milner (Natural
History Museum, London), Dr Heinz A. Kollman (Naturhistorisches Museum, Vienna), Dr Jennifer A. Clack
(University Museum of Zoology, Cambridge) and the late Dr Hermann Jaeger (Museum fur Naturkunde,
Berlin). We also thank Angela Milner for preparing the cast of the Berlin specimen, Adrian Doyle (BMNH)
for helpful technical advice during preparation of other specimens by S.E.K. S., Dr Peter Wellnhofer for
information on the fate of the Munich specimens, and Dr Rob Holmes and Dr Stephen Godfrey for making
their unpublished typescript on Cochleosaurus florensis available to us. Photographic support was provided by
Birkbeck College Photographic Unit. A study trip to Vienna by A.R.M. was supported by the University of
London Central Research Fund.

REFERENCES

beaumont, E. H. 1977. Cranial morphology of the Loxommatidae (Amphibia : Labyrinthodontia). Philosophical
Transactions of the Royal Society of London, Series B, 280, 29-101.
boy, J. a. 1988. Uber einige Vertreter der Eryopoidea (Amphibia: Temnospondyli) aus dem europaischen
Rotliegend (?hochstes Karbon-Perm.). 1. Sclerocephalus. Paldontologisches Zeitschrift, 62, 107-132.

— 1990. fiber einige Vertreter der Eryopoidea (Amphibia: Temnospondyli) aus dem europaischen
Rotliegend (?hochstes Karbon-Perm). 3. Onchiodon. Paldontologisches Zeitschrift , 64, 287-312.
broili, F. 1908. fiber Sclerocephalus aus der Gaskohle von Niirschan und das Alter dieser Ablagerungen.

Jahrbuch der kaiserlich-koniglichen geologischen Reichsanstalt Wien, 58, 49-70.
case, E. c. 1935. Description of a collection of associated skeletons of Trimerorhachis. Contributions from the
Museum of Paleontology, University of Michigan, 4, 227-274.
fritsch, A. 1885. Fauna der Gaskohle und der Kalksteine der Permformation Bohmens , 2, 1-64. F. Rivnac,
Prague.

680

PALAEONTOLOGY, VOLUME 36

fritsch, A. 1901. Fauna der Gaskohle und der Kalksteine der Permformation Bohmens. 4, 65-101. F. Rivnac,
Prague.

Godfrey, s. J. and holmes, R. in press. The Pennsylvanian tetrapod Cochleosaurus florensis Rieppel, from the
lycopod stump fauna at Florence, Nova Scotia. Breviora.

— fiorillo, a. R. and carroll, R. L. 1987. A newly discovered skull of the temnospondyl amphibian
Dendrerpeton acadianum. Canadian Journal of Earth Sciences, 24, 796-805.

jaekel, o. 1911 Die Wirbeltiere. Borntrager, Berlin, 252 pp.

— 1913. Uber den Bau des Schadels. Verhandlungen der anatomischen Gesellschaft, 27, 77-96.
langston, w. jr. 1953. Permian amphibians from New Mexico. University of California Publications in

Geological Science, 29, 349 — 4 16.

milner, a. r. 1980a. The temnospondyl amphibian Dendrerpeton from the Upper Carboniferous of Ireland.
Palaeontology , 23, 125-141.

1980b. The tetrapod assemblage from Nyrany, Czechoslovakia. 439^496. In panchen, a. l. (ed.). The
terrestrial environment and the origin of land vertebrates. Academic Press, London and New York, 633 pp.

— 1982. Small temnospondyl amphibians from the Middle Pennsylvanian of Illinois. Palaeontology, 25,
635-664.

— 1987. The Westphalian tetrapod fauna; some aspects of its geography and ecology. Journal of the
Geological Society, London, 144, 495-506.

— 1990a. The relationships of the eryopoid-grade temnospondyl amphibians from the Permian of Europe.
Acta musei Reginae-hradecensis , 22, 131-137.

— 1990b. The radiations of temnospondyl amphibians. 321-349. In taylor, p. d. and larwood g. p. (eds).
Major Evolutionary Radiations. Systematics Association Special Volume 42, Clarendon Press, Oxford,
437 pp.

— and sequeira, s. e. k. in press. Capetus and the problems of primitive temnospondyl relationships.
Pollichia.

patterson, c. and rosen, d. e. 1977. Review of ichthyodectiform and other Mesozoic teleost fishes and the
theory and practice of classifying fossils. Bulletin of the American Museum of Natural History, 158, 81-172.
rieppel, o. 1980. The edopoid amphibian Cochleosaurus from the Middle Pennsylvanian of Nova Scotia.
Palaeontology , 23, 143-150.

romer, a. s. 1947. Review of the Labyrinthodontia. Bulletin of the Museum of Comparative Zoology, Harvard,
99, 1-368.

1966. Vertebrate paleontology, 3rd edition. Chicago University Press, Chicago, 468 pp.

— 1968. Notes and comments on vertebrate paleontology, Chicago University Press, Chicago, 304 pp.

— and witter, r. v. 1942. Edops, a primitive rhachitomous amphibian from the Texas red beds. Journal of
Geology, 50, 925-960.

— 1943. The skin of the rhachitomous amphibian. Eryops. American Journal of Science, 239, 822-824.
sawin, h. j. 1941. The cranial anatomy of Eryops megacephalus . Bulletin of the Museum of Comparative
Zoology, Harvard, 86, 407-463.

steen, M. c. 1938. On the fossil Amphibia from the Gas Coal of Nyrany and other deposits in Czechoslovakia.

Proceedings of the Zoological Society of London, Series B, 108, 205-283.
welles, s. P. 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 Science, 54, 1-148.
zittel, K. von, 1888. Handbuch der Palaontologie. Abteilung 1. Paldozoologie Band III. Vertebra ta ( Pisces ,
Amphibia , Reptilia, Aves). Munich and Leipzig, 900 pp.

SANDRA E. K. SEQUEIRA
ANDREW R. MILNER

Department of Biology
Birkbeck College

Typescript received 1 September 1992 Malet Street

Revised typescript received 1 November 1992 London WC1E 7HX, UK

EARLY CARADOC TRILOBITES OF EASTERN
IRELAND AND THEIR P AL AEOGEOGR APH IC AL

SIGNIFICANCE

by m. romano and a. w. owen

Abstract. Twenty-five trilobite species belonging to twenty-three genera are recorded from the Caradoc
Knockerk Formation of the Grangegeeth area, eastern Ireland. Arthrorhachis knockerkensis and Birmanites
salteri are new, while Barrandia sp. and Flexicalymene sp. possibly represent new species. The faunas occur
predominantly in the lower parts of the Knockerk House Sandstones Member and younger Brickwork’s
Quarry Shales Member. The older fauna is of early Caradoc age and shows affinities with species from the
Balclatchie and Lower Ardwell groups at Girvan. The younger trilobite fauna and associated graptolites are
indicative of the Climacograptus peltifer Biozone (possibly Harnagian). The Laurentian/Scoto-Appalachian
affinities of the Grangegeeth Caradoc faunas indicate that the Grangegeeth terrane was closest to Laurentia
during the early Caradoc and that the final lapetus Suture line must lie to the south of the Grangegeeth terrane.

The Ordovician shelly faunas of eastern Ireland have become increasingly important in determining
the position of the lapetus Suture in this complex sector of the Caledonide Orogen. In particular,
the area around Grangegeeth (Text-fig. 1) is sited near the proposed line of the suture, and an
understanding of the affinities of the faunas is vital to unravelling the tectonic history of this region.

The first detailed description of the Ordovician sequence and faunas of the area between Slane,
Co. Meath, and Collon, Co. Louth, in eastern central Ireland was by J. C. Harper (1952). In this
account. Harper listed eight trilobite species, and figured five, from his ‘Upper Tuffs and Shales
(including the brown shales of Mellifont)’. The area was remapped by Romano for an
undergraduate dissertation, during which a rich fauna with abundant trilobites was discovered
within Harper’s unit. These trilobites were recorded by Brenchley et al. (1967) who listed eleven
species and figured four. These authors ascribed the trilobite fauna to the Climacograptus peltifer
Biozone on account of the associated graptolites, and suggested that it may be of Harnagian age.
Harper and Romano (1967) described a new trinucleid trilobite, Decordinaspis bispinosa , from that
fauna. Romano (1970) proposed the name ‘Brickwork’s Quarry Shales’ for this richly fossiliferous
unit, and defined it as the third member of the Knockerk Formation; nineteen trilobite species were
described and figured from it. The two lower members also yield trilobites, though less abundantly.

In a re-appraisal of some Ordovician successions from eastern Ireland, Brenchley et al. (1977)
discussed their correlation with particular reference to dating the main episodes of volcanicity. The
affinity of the Brickwork’s Quarry Shales trilobite fauna to that from the Balclatchie Mudstones of
Scotland, was noted. Romano (1980) formally described the Ordovician lithostratigraphy of the
Slane-Collon area and listed eight trilobite species from the Knockerk House Sandstones Member,
three from the Knockerk House Shales Member and nineteen from the Brickwork’s Quarry Shales
Member.

The present paper is the first formal systematic work on the trilobite faunas and sets the trilobites
in a palaeobiogeographical context. In doing so, it places important constraints on the inter-
pretation of terrane provenance in eastern Ireland (Owen et al. 1992).

[Palaeontology, Vol. 36, Part 3, 1993, pp. 681-720, 4 pis.]

© The Palaeontological Association

682

PALAEONTOLOGY, VOLUME 36

text-fig. 1. Location map (inset) and summary
geological map of Grangegeeth and surrounding
area. Ordovician rocks are shaded black; Silurian
rocks with oblique lines. (For explanation of fault
nomenclature see Owen et al. 1992, fig. 1).

text-fig. 2. Generalised lithostratigraphy of the Grangegeeth area (after Romano 1980) showing ranges of
trilobites described in this paper. Question marks in range chart indicate uncertainty of: stratigraphical
position (Cybelinae indet.), identification (Birmanites salteri sp. nov.), or vertical range ( Arthrorhachis

knockerkensis sp. nov. and Tretaspis aff. reticulata ).

ROMANO AND OWEN: IRISH CARADOC TRILOBITES

683

LITHOSTRATIGRAPHY

The Caradoc and Ashgill rocks between Slane, Co. Meath and Collon, Co. Louth in eastern Ireland
(Text-fig. 1) were subdivided by Romano (1980) into two groups: the Grangegeeth Group below
and Mellifont Abbey Group above. The Grangegeeth Group is up to 1400 m thick and consists of
the Collon, Knockerk and Fieldstown Formations. The overlying Mellifont Abbey Group is at least
175 m thick and comprises the Broomfield and Oriel Brook Formations (Text-fig. 2). Rocks of
proven or inferred Caradoc age include all the formations of the Grangegeeth Group and the
Broomfield Formation of the Mellifont Abbey Group (Brenchley et al. 1977). Trilobites have only
been recovered from the Knockerk Formation (see below) and Oriel Brook Formation; those from
the latter unit require further sampling and will be described in a later publication.

The lithostratigraphy of the Knockerk Formation was described by Romano (1970, 1980). The
formation is approximately 400 m thick in the type area around Knockerk Flouse (Romano 1980,
p. 64, fig. 6) and is made up of four members (Text-fig. 2). The lowermost member, the Knockerk
Flouse Sandstones, consists of massive volcanic sandstones, with tuffaceous shales sporadically
present at the base of the unit but more common at the top. The unit is particularly fossiliferous
in the basal part with rich brachiopod faunas and rarer trilobites (see Harper 1952 and Romano
1980 for faunal lists and details of localities). The Knockerk House Sandstones grade up into the
Knockerk House Shales Member which is only sporadically tuffaceous. Brachiopods are rare in this
unit, with trilobites and graptolites dominating. The top of the Knockerk House Shales Member is
marked by the distinctive Tretaspis Bed which forms the base of the overlying Brickwork’s Quarry
Shales Member. The latter shale member is the most richly fossiliferous in the area, particularly near
the base where trilobites dominate assemblages which also contain brachiopods, bivalves,
gastropods, graptolites, ostracodes, bryozoans, orthocones and echinoderms. The bulk of the
material from this unit was collected from the Brickwork’s Quarry when it was superbly exposed
during its active life but now, due to flooding and disuse, collecting is extremely difficult.
Lithologically the member is most variable near the base where tuffaceous silty beds and
concretionary horizons are common. The upper part of the member frequently contains layers of
concretions and horizons with limonitic spots. It is these two features which serve to distinguish the
Brickwork’s Quarry Shales from the overlying Mullaghdillon Shales Member, where both
concretions and limonitic spots are absent. The junction between these two members is gradational
and no faunas have been found in the younger unit. The overlying Fieldstown Formation is not seen
in contact with the Mullaghdillon Shales and is distinguished by the presence of thin beds of lithic
and pumice tuffs. A single graptolite, Amplexograptus sp. indet., has been found in the Fieldstown
Formation (Romano 1980).

TRILOBITE DIVERSITY, DISTRIBUTION AND PRESERVATION

The composition and vertical ranges of the trilobite faunas are shown in Text-figure 2. The faunas
occur essentially in the lower part of the Knockerk House Sandstones Member and the base of the
Brickwork’s Quarry Shales Member. Twenty-five trilobite species are recorded from the Caradoc
sequences of the Grangegeeth area : eight from the lower part of the Knockerk House Sandstones
Member (excluding illaenid indet.) and one from the upper part; eighteen from the base of the
Brickwork’s Quarry Shales Member, of which three first appear in the middle of the underlying unit.
The compositions of the two major faunas are quite distinct and show a marked contrast in
diversity. Only one species, Sphaerocoryphe cf. pemphis, is common to both and the older fauna has
fewer skeletal elements (Table 1). Virtually all the specimens are disarticulated; only three occur as
complete exoskeletons. Cephala and cranidia dominate (64 per cent - but see caption to Table 1),
followed by pygidia (27 per cent), free cheeks (4 per cent), thoracic segments ( > 2 per cent) and
hypostomata ( > 1 per cent). The older fauna shows no particular species dominance from the thirty
specimens collected. The younger fauna is dominated by Tretaspis aff. reticulata (34 per cent of total

684

PALAEONTOLOGY, VOLUME 36

table 1. List of trilobite species from the Knockerk Formation, Grangegeeth area, showing the abundance of
their skeletal elements. The number of cephala, cranidia and lower lamellae of Tretaspis aff. reticulata and
Decordinaspis bispinosa is given as a single entry in each case. Ceph., cephala; Cranid., cranidia; Hypost.,
hypostomata; Thor, seg., thoracic segments; Pygid., pygidia; Artie, spec., articulated specimens.

Free

Thor.

Artie.

Ceph. Cranid.

cheek

Hypost.

seg.

Pygid.

spec.

Total

%

Arthrorhachis

9

rare

6

15

4-2

knockerkensis sp. nov
Telephina ( Telephops ) cf.

5

1

6

1-7

bicornis (Ulrich, 1930)
Remopleurides sp.

3

1

4

11

Birmanites salteri sp. nov.

2 3

6

5

4

+ 50

1

71

200

Barrandia sp. ?nov.

1

1

0-3

lllaenus sp.

1

2

3

0-8

illaenid indet.

1

1

2

0-6

Decoroproetus sp.

4

1

5

1-4

Tretaspis aff. reticulata

100

rare

20

121

33-9

Ruedemann, 1901
Decordinaspis bispinosa

10

10

2-8

Harper and Romano,
1967

Ampyx aff. repulsus Tripp,

23

2

1

7

33

9-2

1976

Ceraurinellal sp.

3

1

4

11

Sphaerocoryphe cf.

2 26

3

31

8-7

pemphis Lane, 1971
Sphaerexochus sp. indet.

1

1

0-3

Acanthoparyphyal sp.

1

1

0-3

Cybelinae indet. cf.

6

6

1-7

‘ Deacybele ' pauca
Whittington, 1963

Cybelinae indet.

1

1

0-3

A tractopyge ? sp.

1 3

1

1

6

1-7

Flexicalymene sp. ?nov.

4

2

6

1-7

Gravicalymene sp.

4

4

8

2-2

Achatella

2 1

1

4

1.1

truncatocaudatal
(Portlock, 1843)

Autoloxolichas cf. laxatus

2

2

0-6

(M‘Coy, 1846)
Autoloxolichas sp. indet.

1

1

0-3

Amphilichas sp.

1

1

1

3

0-8

Lichid indet.

1

1

0-3

Miraspis aff. solitaria

8

1

2

11

31

Reed, 1935

229

14

5

8

97

4

357

100-2

(64%)

(4%)

(1-4%)

(2-2%)

(27-2%) (1-1%)

trilobite remains) with Birmanites salteri (20 per cent), Atnpvx aff. repulsus (9 per cent) and
Sphaerocoryphe cf. pemphis (8 per cent) constituting the other major elements. All remaining species
each make up less than 5 per cent of the sample and six of these comprise less than 1 per cent
(Table 1).

ROMANO AND OWEN: IRISH CARADOC TRILOBITES

685

This Tretaspis dominance with abundant Ampyx is similar to that observed by Owen et al. (1986)
for the fauna from the late Caradoc Raheen Formation of County Waterford. Though there is an
absence of asaphids in the Raheen fauna, the nileid Homalopteon is the fourth most abundant taxon
and may have occupied a similar niche to that of Birmanites at Grangegeeth. Owen et al. (1986)
commented on other faunas dominated by trinucleids and raphiophorids, and concluded that a
fairly deep shelf environment was likely for the Raheen fauna. We propose a similar environment
for the upper Grangegeeth fauna.

Preservation in both the lower and upper faunas is mainly in the form of moulds. Abrasion is rare
although broken sclerites are relatively common. The almost invariably disarticulated exoskeletons
indicate reworking, but this is probably largely biogenic rather than current action since there is no
size or shape sorting. That some current action has occurred is apparent in the lower faunas where
bedding plane assemblages of brachiopods show some degree of orientation (Romano 1980, p. 66).

AGE OF THE TRILOBITE FAUNAS
Knockerk House Sandstones Member

This member contains at least eight species including 'lichid indet.’ which is clearly different from the co-
occurring Amphilichas sp., but excluding ‘lllaenid indet.’ which might belong in Illaenus sp. from this member.
Six are indeterminate and two have been left in open nomenclature. Achatella truncatocaudata is an Ashgill
form from the Cautleyan of Pomeroy, Ireland and closely related forms are of Hirnantian age from near
Girvan, Scotland (Owen 1986). Sphaerocoryphe pemphis occurs in the early Caradoc Balclatchie and Lower
Ardwell groups at Girvan. Among the specifically indeterminate forms, Amphilichas sp. is morphologically
close to Scottish Balclatchie and Lower Ardwell forms while Atractopvgel sp. resembles most closely middle
Ordovician forms from Estonia. Ceraurinellal sp. and Gravicalymene sp. are too incomplete to allow
comparisons.

Thus, although the fauna shows closest affinities with species of middle Ordovician to Hirnantian age, on
balance an early Mohawkian (early Caradoc) age is indicated by the affinities of species to Balclatchie and
Lower Ardwell group taxa. No diagnostic graptolites have been recovered from these beds but the
accompanying rich brachiopod faunas, showing Scoto-Appalachian affinities, indicate an early Caradoc age
(Harper and Parkes 1989; Owen et al. 1992).

Brickwork's Quarry Shales Member

Of the eighteen species recognized from this unit, seven are close to previously described species, two are new,
two are probably new, five have been identified only to generic level, one is endemic and one is indeterminate.
Among the seven forms left in open nomenclature, three are closest to Girvan species ( Sphaerocoryphe pemphis,
Ampyx repulsus, Miraspis solitaria) of mid Llandeilo to early Caradoc age, two show American affinities of
early Caradoc age ( Tretaspis reticulata) or slightly younger ( Telephina (Telephops) bicornis ), one is similar to
the mid Caradoc Welsh form ' Deacybele' pauca, while Autoloxolichas cf. laxatus is very close to a form
commonly occurring in the Caradoc of Britain, Ireland and Scandinavia. On balance an early Caradoc age is
indicated, as was suggested by Brenchley et al. (1977) who assigned these shales to the Harnagian mainly on
the basis of graptolites indicative of the Climacograptus peltifer Biozone (Brenchley et al. 1967).

No recent work has been published on the micropalaeontology of this unit but Downie (in Brenchley et al.
1967) concluded that the chitinozoans clearly pointed to a Caradoc age. Work on the brachiopods
(Dr D. A. T. Harper pers. comm.) confirms the correlations suggested by the trilobites.

BIOGEOGRAPHY

Determination of the biogeographical affinities of the Grangegeeth trilobites is crucial to an
understanding of the position of the area during the Caradoc relative to the major plates bordering
the Iapetus Ocean. As Cocks and Fortey (1990 and references therein) have summarized, the ocean
was wide in the early Ordovician and separated the shelf faunas of Laurentia at low latitudes from

686

PALAEONTOLOGY, VOLUME 36

those of Baltica in mid latitudes and Gondwana further south. The Tornquist Sea also provided a
barrier between Baltica and Gondwana. Marginal and deep water facies showed a considerably
lower level of endemicity, and pelagic faunas had a climatic (therefore latitudinal) zonation. By the
early Caradoc, however, the Tornquist Sea had ceased to prevent transmigration and Iapetus had
narrowed sufficiently to allow some mixing of the shelf faunas. However, Avalonia (including the
Anglo-Welsh area) had rifted from northern Gondwana with some concomitant divergence in their
faunas (see Cocks and Fortey 1990; Fortey and Cocks 1991; cf. Paris and Robardet 1990).

The early Caradoc shelly faunas at Grangegeeth were originally thought to have strong Baltic
affinities (Harper 1952; Williams 1956) and plate tectonic models subsequently positioned the
Iapetus suture to the north of the area (e.g. McKerrow and Soper 1989). More recent analyses of
the Grangegeeth faunas, however (Owen et al. 1992 and references therein), point to much stronger
links with Laurentian (largely Scoto-Appalachian) faunas during the Caradoc and thus indicate the
presence of a more southerly position of the suture with Avalonia.

As the distinction between the faunas of Laurentia (and its outboard terranes) and those of the
plates on the other side of Iapetus was breaking down during the early Caradoc, the
palaeobiogeographical analysis of early Caradoc faunas (such as those of Grangegeeth) must be
based on the critical assessment of the histories of each of the component taxa. It is now clear that
only a few taxa are diagnostic in determining the affinities of the Grangegeeth faunas and these
would probably be ‘swamped’ in a statistical analysis by the genera with a much more equivocal
palaeogeographical signal.

Only one genus, Decordinaspis, is endemic to Grangegeeth. Its ancestry is unclear, but Hughes et
al. (1975, fig. 120) suggested that it may lie in Paratrinucleus from the late Arenig-early Llanivirn
(cf. Hughes et al 1975) of New World Island in the oceanic Dunnage Zone of Newfoundland (see
Boyce 1987). Text-figure 3 summarizes the earlier distribution of the remaining twenty-one named
trilobite genera in the lower Caradoc Grangegeeth faunas. Many have a range which extends into
outer shelf and even slope environments and thus it is not surprising that ten had earlier trans-
Iapetus occurrences in Baltica and Laurentia ( Illaenus , Remopleurides , Telephina), the Anglo-Welsh
area and Laurentia (Decoroproetus, Flexicalymene), or all three areas ( Ampyx , Arthrorhachis,
Atractopyge, Gravicalymene, Miraspis ). Of all these genera, Telephina and possibly Remopleurides
were pelagic (Fortey 1985), and their earlier restrictions to Baltic and Laurentian sites may be
evidence against a high latitude (Gondwanan) location for the Grangegeeth terrane during the
Caradoc.

Achatella occurred earlier in Baltica and also in the Irish continuation of the Anglo-Welsh area
where it was represented by A. bailyi (Salter, 1864) described from the upper Llandeilo to lowest
Caradoc Tramore Limestone of Co. Waterford (see Morris 1988). However, it is also known in
Laurentian faunas from the early Rocklandian (Sloan 1991, table 2) which is only slightly younger
than the Grangegeeth occurrence. It is not therefore diagnostic palaeogeographically. The same
applies to the form described here as ‘cybeline indet. cf. Deacybele' . Although applied to a group
of Anglo-Welsh and Baltic species whose first appearance is in the Tramore Limestone, their origin
(and possibly generic placement) lies in Cybeloides — a Laurentian genus ranging from the early
Chazyan (late Llanvirn to early Llandeilo; see Sloan 1991, table 2). Autoloxolichas had an earlier
history in Baltica but appeared in Laurentia and the Anglo-Welsh area at about the same time as
at Grangegeeth.

Birmanites and Barrandia are only known from Gondwana (which included the Anglo-Welsh
area prior to the Llandeilo). As is noted in the systematic section, the former has a history extending
back to the early Ordovician of South East Asia. It has never been recorded from Laurentia but as
its earlier occurrences are in deep shelf and upper slope settings, it is not entirely surprising that it
might cross the Iapetus divide in the later part of the Ordovician. The same applies to Barrandia ,
hitherto known from the Anglo-Welsh area from the late Arenig to the early to mid Llandeilo. This
is also a deep shelf to upper slope trilobite which in the late Arenig of South Wales occurred in the
deep water cyclopygid association (Fortey and Owens 1987). A closely allied genus, Homalopteon ,
is known from various deeper water Llandeilo (and possibly Llanvirn) to Caradoc faunas in the

ROMANO AND OWEN: IRISH CARADOC TRILOBITES

687

text-fig. 3. Venn diagram showing the occurrences of selected trilobite genera from the Knockerk Formation
during the Caradoc, with reference to Laurentia, Baltica and Avalonia. The arrows show the ‘direction’
and timing of migratory pathways. The relative positions of the three continental plates are not intended

to be accurate.

Anglo-Welsh area. There is a single record of this genus in a similar bathymetric setting in the
Llandeilo at Girvan (Ingham and Tripp 1991, p. 27), further testimony to the unreliability of such
taxa for biogeographical reconstructions.

Six genera had an earlier history confined to Laurentian and Scoto-Appalachian faunas (Text-
fig. 3), but of these, Acanthoparypha , Amphilichas and Sphaerexochus appeared in the Anglo-Welsh
area during the early Caradoc, contemporaneous with or only slightly after the Grangegeeth faunas
(Owen et a/. 1992, fig. 2). The remaining three, Sphaerocoryphe , Tretaspis and Ceraurinella did not
extend their ranges into the Anglo-Welsh or Baltic areas until the mid Caradoc, late Caradoc and
mid Ashgill, respectively. All have environmental ranges from pure limestone facies to outer shelf
muds but they are the most reliable indicators of the provincial affinities of the Grangegeeth trilobite
genera.

The Laurentian/Scoto-Appalachian affinities of the Grangegeeth Caradoc faunas are even more
striking at species level. The species of Amphilichas, Ampyx, Flexicalymene, Miraspis, Sphaero-
coryphe, Telephina and Tretaspis are all closest to older or coeval Laurentian/Scoto-Appalachian
taxa and that of Achatella closest to a younger species from such faunas. In contrast, only the

688

PALAEONTOLOGY, VOLUME 36

species of Autoloxolichas and ‘cybeline indet. cf. Deacybele' have their strongest affinity to
(younger) Baltic and Anglo-Welsh forms.

The biogeographical affinity to Laurentian/Scoto-Appalachian faunas is evident at all the
fossiliferous levels of the Knockerk Formation (Text-fig. 2) and is also seen in the associated
brachiopods. The Grangegeeth area probably represents a discrete terrane within the Iapetus Suture
zone in eastern Ireland (Harper and Parkes 1989) but was closer to Laurentia than Baltica or
Gondwana during the early Caradoc. Limited graptolite evidence, however, suggests a position
closer to Gondwana during the Llanvirn and hence a northward migration of the Grangegeeth
terrane during the mid Ordovician (Owen et al. 1992).

LOCALITIES

Most of the trilobite faunas described were collected from the localities listed below within the Grangegeeth

area. To avoid repetition under the section heading ‘Horizon and locality’, details of each locality and

stratigraphical information are listed here. All material was collected from the Knockerk Formation, only the

member is listed below.

Locality 1 : (Hull et al. 1871 ; Harper 1952, p. 88; Romano 1980, p. 66); small quarry to-east of road, 436 m
north of Grangegeeth crossroads; base of Knockerk House Sandstones Member.

Locality 3: (Harper 1952, p. 87; Romano 1980, p. 66); small quarry to west of road, 419m 348° from
Grangegeeth crossroads; base of Knockerk House Sandstones Member.

Collon Quarry: (Romano 1980, p. 67); southern end of large quarry, 554 m 175° from Collon crossroads; base
of Knockerk House Sandstones Member.

Locality 210A: (Romano 1970); ditch section, 3226 m 330° from Slane crossroads; lower part of Knockerk
House Shales Member.

Locality 21 0B: (Romano 1970); ditch section, 2969 m 332° from Slane crossroads; approximately middle of
Knockerk House Sandstones Member.

Locality 38: (Harper 1952, p. 89; Romano 1980, p. 67); roadside exposure, 654m 210° from Collon
crossroads; ?upper part of Knockerk House Sandstones Member/lower Knockerk House Shales Member.

Locality 190: (Brenchley et al. 1967, p. 298); temporary well digging to west of road, 1007 m south of
Grangegeeth crossroads; exact horizon not known, but possibly near base of Brickwork’s Quarry Shales
Member.

Locality 208 and Locality 209: (Romano 1980, p. 65); temporary exposures, 2952 m 340° and 2902 m 337°
from Slane crossroads respectively; middle to upper part of Knockerk House Shales.

Locality 26: (Harper 1952, p. 89; Romano 1980, p. 70); ditch section and adjacent small quarry, 2214 m 102°
from Grangegeeth crossroads; ?Knockerk House Shales Member to basal Brickwork’s Quarry Shales
Member.

Brickwork’s Quarry: (Romano 1980, Fig. 6, p. 68); large quarry (flooded and partly overgrown in 1991),
2800 m 340° from Slane crossroads; from Tretaspis Bed through basal part of Brickwork’s Quarry Shales
Member.

Repositories

All figured specimens are housed in the National Museum of Ireland (NMI), Geological Survey of Ireland
(GSI) or Liverpool Museum (LM); other material is deposited in the above Institutions and the Natural
History Museum, London (BM It). Some incomplete or poorly preserved specimens are in the collections of
the Department of Earth Sciences, University of Sheffield.

SYSTEMATIC PALAEONTOLOGY

Family metagnostidae Jaekel, 1909
Genus arthrorhachis Hawle and Corda, 1847

Type species. Arthrorhachis tarda Hawle and Corda, 1847; from the Kraluv Dvur Formation of Ashgill age,
near Beroun, Czech Republic.

ROMANO AND OWEN: IRISH CARADOC TRILOBITES

689

Arthrorhachis knockerkensis sp. nov.

Plate 1, figs 1-7

71871 Agnostus trinodus ; Hull et a!., p. 29.

1952 ‘ Agnostus ' girvanensis Reed; Harper, p. 89.

71952 Agnostus trinodus' Harper, p. 90.

1967 Trinodus sp. ; Brenchley et al., pp. 298, 301, pi. 7.

1980 Trinodus sp.; Romano, pp. 68, 70, 771.

Derivation of name. After the townland of Knockerk in which Brickwork’s Quarry is situated.

Material. Holotype: NMI F20974. Paratypes: NMI F20971-3, F20975a-b; LM 1988.216.463;

BM It.25694-It.25695u-A

Horizon and locality. NMI F20971-F29075 and BM It. 25694 from concretions approximately 1 m above
the Tretaspis Bed; BM It.25695u-6 from the Tretaspis Bed, Brickwork's Quarry; L.M. 1988.216.463 from
Loc. 26.

Diagnosis. Glabella with constriction; well-marked posterolateral cephalic spines. Subquadrate
outline to pygidium, small triangular pygidial axis and very slightly divergent pygidial spines.

Description. Cephalic outline almost square, approximately as long as greatest width, latter being just anterior
to mid-length. Glabella occupies 55-65 per cent of cephalic length, narrowing forwards slightly in front of basal
lobes, smoothly rounded to blunt-ended anteriorly and constricted at about mid-length. Glabella strongly
convex transversely, gently convex longitudinally. Triangular basal lobes about twice as wide as long (exs.), not
meeting mesially thus the mesial glabella between them is bluntly pointed; delimited by furrows which are
broader abaxially. Axial furrows broad; fairly well incised laterally, less so in front of glabella. Cheeks of fairly
constant width, steeply declined laterally, gently declined anteriorly. Border widest anterolaterally; narrowing
markedly near genal angles where there is a small (coaptative) notch in the margin. Short, rearwardly directed
spines at posterolateral corners of cephalon. No sculpture except faint median glabellar node situated at
constriction on internal mould of some specimens (possibly elongated longitudinally on one specimen).

Pygidium sub-quadrate in outline, slightly wider than long; maximum transverse width towards rear end.
Axis occupies 50 per cent or less of pygidial length, sub-triangular in outline, narrowing evenly rearwards at
55-70° to rounded posterior margin; delimited by well-incised furrows. Short articulating half-ring separated
from anterior ring by deep furrow. Anterior ring only slightly shorter (sag.) than second ring and about two-
thirds as long as terminal piece. Posterior ring furrow complete, anterior furrow interrupted by longitudinal,
gently convex (tr.) ridge which crosses anterior and second axial rings and develops as a tubercle on the latter.
Axis strongly convex (sag. and tr.). Pleural field steeply declined laterally, less so posteriorly. Border generally
flat ; widest posterolaterally where extended into pair of short, slightly divergent spines which do not extend
beyond level of posterior margin of pygidium. No sculpture observed.

Discussion. Although there is clear intraspecific variation, the well-developed cephalic and pygidial
spines, small triangular pygidial axis and nearly parallel-sided pygidium are features not collectively
seen in any other species of this genus. Thus A. knockerkensis has longer, more posteriorly placed
pygidial spines, a more parallel-sided pygidium with a slightly narrower axis than A. tarda from the
Ashgill of Bohemia, Scandinavia, the British Isles and Kazakhstan (Whittington 1950; Kielan 1960;
Pek 1977 ; Owen 1981 ; Ahlberg 1989). It has a more quadrate outline to the pygidium and narrower
axis than A. doulargensis Tripp, 1965, from the Llandeilo of Girvan, Scotland, and the closely
related A. elspethi Hunt, 1967, from the Llandeilo-lower Caradoc of the southern Appalachians
and Sweden (see Ahlberg 1988). A. knockerkensis has a less well-rounded cephalon and shorter
pygidial axis than A. aspinosus Tripp, 1976, and a more constricted glabella and shorter pygidial
axis than A. comes Tripp, 1976, both from the Llandeilo^ Superstes Mudstone at Girvan.
A. girvanensis Reed, 1903, from the Balclatchie and Lower Ardwell groups is distinct in its shorter
glabella and broader, less tapered pygidial axis. T. agnostiformis M'Coy, from the Caradoc of

690

PALAEONTOLOGY, VOLUME 36

Enniscorthy, Co. Wexford, is the only species now ascribed to Trinodus (see Fortey 1980, p. 26) and
because of its poor preservation cannot be closely compared to the Grangegeeth species. The
position of the median glabellar node opposite the glabellar constriction precludes the assignment
of the Grangegeeth species to Galbagnostus Whittington, 19656, where this node is much more
forwardly placed.

Family telephinidae Marek, 1952
Genus telephina Marek, 1952

Type species. Telephus fractus Barrande, 1852, by original designation; from the Nucice Beds and Kraluv Dvur
Shales (late Caradoc to early Ashgill) of Bohemia.

Subgenus telephina (telephops) Nikolaisen, 1963

Type species. Telephus granulatus Angelin, 1854, by original designation; from the Elnes Formation
(‘ Ogygiocaris Shale’) (Llanvirn) of Hadeland, Norway.

Telephina ( Telephops ) cf. bicornis (Ulrich, 1930)

Plate 1, figs 8-11, 14

1967 Telephina ( Telephops ) sp. cf. T. (T.) bos Nikolaisen; Brenchley et al., pp. 298, 302, pi. 7, figs 7-8.
1980 Telephina ( Telephops ) cf. bicornis (Ulrich); Romano, p. 69.

1988 Telephina ( Telephops ) cf. bos Nikolaisen, 1963; Morris, p. 227 .

Material. NMI F20976-F20979; NMI G. 108/1965 (figured by Brenchley et al., 1967, pi. 7, figs 7-8).

Horizon and locality. All from concretions approximately 1 m above Tretaspis Bed, except NMI F20979 from
Tretaspis Bed, Brickwork’s Quarry and possibly from Loc. 26.

Discussion. The Irish specimens closely resemble those assigned to the American species T. (77)
bicornis (Ulrich, 1930, pi. 4, figs 1-14) from the Whitesburg Fimestone (late Whiterockian) of

EXPLANATION OF PLATE 1

Figs 1-7. Arthrorhachis knockerkensis sp. nov. 1, NMI F20971 ; internal mould of cephalon, dorsal view, x 7.
2, NMI F20972; internal mould of cephalon, dorsal view, xll. 3, 6, NMIF20973; internal mould of
cephalon, dorsal and lateral views, x 7, x 10, respectively. 4, NMI F20974 (holotype); internal mould of
pygidium, dorsal view, x 9. 5, 7, NMI F20975; internal mould of pygidium, dorsal and lateral views, both
xll. All from concretions approximately 1 m above Tretaspis Bed, Brickwork’s Quarry Shales Member;
Brickwork's Quarry.

Figs 8-11, 14. Telephina ( Telephops ) cf. bicornis (Ulrich, 1930). 8, 11, NMIF20976; internal mould of
cranidium, dorsal and lateral views, both x 4. 9, NMI F20977 ; internal mould of cranidium, dorsal view,
x 4. 10, NMI F20978; internal mould of cranidium, frontal view, x4. 14, NMI F20979; internal mould of
pygidium, dorsal view, x 7. All except 14 from concretions approximately 1 m above Tretaspis Bed (14 from
Tretaspis Bed), Brickwork’s Quarry Shales Member; Brickwork’s Quarry.

Figs 12-13, 15-16. Remopleurides sp. 12-13, NMI F20980; internal mould of cranidium, dorsal and frontal
views, both x 6. 15, NMI F20981 ; internal mould of cranidium, dorsal view, x 3. 16, NMI F20982; internal
mould of free cheek, x 4. All except 16 from Tretaspis Bed (16 from concretions approximately 1 m above
Tretaspis Bed), Brickwork’s Quarry Shales Member; Brickwork’s Quarry.

Fig. 17. Illaenus sp. NMI F14032/A (figured by Harper 1952, pi. 5, fig. 6 as NMI 1951/17); internal mould
of cephalon, dorsal view, x 1-5. Base of Knockerk House Sandstones Member; Loc. 1.

Fig. 18. Barrandia sp. ?nov. NMI F20983; internal mould of nearly complete specimen, dorsal view, x 1-5.
Probably from near base of Brickwork’s Quarry Shales; Brickwork’s Quarry.

PLATE 1

ROMANO and OWEN, Irish Caradoc trilobites

692

PALAEONTOLOGY, VOLUME 36

Virginia. Minor differences are the generally finer ornament on the Irish specimens, particularly on
the occipital ring, and the possibly smaller glabellar spines (although only one spine has so far been
seen on the Grangegeeth specimens). Also the American species does not appear to show the
shallow lateral occipital furrow. The pygidium described above differs from that of T. (T.) bicornis
in apparently lacking the fine tuberculation on the axial rings (however, the Irish specimen is an
incomplete, deformed internal mould). Ulrich compared his new species with T. ( T .) granulata
Angelin and figured copies of Swedish specimens from the Ogygiocaris shale. The proportions of
the cranidium (in Ulrich 1930, pi. 1, fig. 19) are quite unlike those of the present material although
another specimen (pi. 1, fig. 22) is more similar to the Irish cranidia. The pygidium of T. ( T .)
granulata however shows only single tubercles on the axial segments. This latter character was also
noted by Nikolaisen (1963) who figured two Norwegian cranidia of this species, the glabella of
which is relatively longer and more coarsely ornamented than in the Irish specimens.

Originally the single cranidium known from Grangegeeth was compared to T. (T.) bos Nikolaisen
(Brenchley et al. 1967), but additional and better preserved material from Grangegeeth differs from
the Norwegian species in having a relatively narrower cranidium and paired spines, as opposed to
single tubercles, on the pygidial axial rings.

Family remopleurididae Hawle and Corda, 1847
Genus remopleurides Portlock, 1843

Type species. Remopleurides colbii Portlock, 1843, by subsequent designation of Miller (1889); from the Killey
Bridge Beds, (Cautleyan), Desertcreat, Co. Tyrone, Northern Ireland.

Remopleurides sp.

Plate 1, figs 12-13, 15-16

71871 Remopleurides sp.; Hull et al ., p. 29.

71952 Remopleurides sp. ; Harper, p. 90.

1967 Remopleurides sp. indet. ; Brenchley et al., p. 298.

1980 7 Remopleurides sp.; Romano, p. 68.

1980 Telephina ( Telephina ) sp.; Romano, p. 68.

Material. NMI F20980, NMI F2098G-6, NMI F20982.

Horizon and locality. From Tretaspis Bed, Brickwork’s Quarry, except NMI F20982 from concretion
approximately 1 m above Tretaspis Bed in same quarry.

Discussion. In general glabellar outline and width of the tongue, Remopleurides sp. is very close to
Sculptella scripta Nikolaisen, 1983, pi. 9, figs 9-11, and 5. scriptoides Nikolaisen, 1983, pi. 10, figs.
1-8 from the Llandeilo of the Oslo Region. As no sculptural lines are evident in the internal or
external moulds of the Irish species, it is therefore provisionally placed in Remopleurides. The
relatively wide glabella, short tongue and rapidly widening palpebral rims posteriorly are, however,
not features collectively seen in any other remopleuridid species. Among the described Scottish
species of Remopleurides , the Irish form is closest to Remopleurides sp. from the Upper Balclatchie
Group (Tripp 1980, pi. 1, fig. 16), although the glabellar tongue is considerably narrower in the Irish
specimen. R. cf. granensis Stormer (Owen 1981, pi. 1, figs 19-23) from the Ashgill of the Oslo region
has a glabellar tongue which is closer to that of Remopleurides sp., but the latter does not show the
granulation or occipital tubercle of the Norwegian species. Middle Ordovician remopleuridids from
Norway include Remopleurides sp. G from the late Caradoc Solvang Formation in Ringerike
(Nikolaisen 1983, pi. 6, figs 9-11), which has a relatively narrow glabellar tongue but shows strong
sculpture on the occipital ring. Until more and better preserved material is available it is preferred
to leave the present species in open nomenclature.

ROMANO AND OWEN: IRISH CARADOC TRILOBITES

693

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

Type species. Ogygites birmanicus Reed, 1915; from the Hwe Mawng Beds (lower Ordovician) of Hwe Mawng
and Hpakhi, northern Shan States, Burma.

Discussion. The status of Birmanites Sheng, 1934 has been discussed by Chugaeva (1958), Zhou et
al. (1984), Zhou and Dean (1986), and Tripp et a/. (1989). Zhou et al. pointed out that a number
of species currently referred to Ogygites Tromelin and Lebesconte, 1876, Pseudobasilicus Reed,
1931, Birmanites , Opsimasaphus Kielan, 1960 and Nobiliasaphus Pribyl and Vanek, 1965, may be
assigned to Ogygites, based on the diagnosis given by Chugaeva (1958) and freely translated with
minor additions by Zhou et al. (1984, p. 17). Chugaeva considered Birmanites and Pseudobasilicus
as junior synonyms of Ogygites but Zhou et al. regarded Ogygites as being insecurely founded and,
pending an ICZN decision (Henningsmoen et al. 1980) preferred to use Birmanites. We follow Zhou
et al. in using Birmanites in preference to Ogygites (cf. Chugaeva 1958), and agree with Zhou and
Dean (1986, p. 754) in considering Opsimasaphus a junior subjective synonym of Birmanites.

Birmanites salteri sp. nov.

Plate 2, figs 1-10

1866 Asaphus radiatus', Salter, pi. 18, figs 4-5.

1952 Pseudobasilicus sp. ; Harper, pp. 90, 108, pi. 5, fig. 2.

1966 Opsimasaphus sp. ; Whittington, p. 78.

1967 Pseudobasilicus sp. indet.; Brenchley et al. p. 298.

1980 lOpsimasaphus sp. indet.; Romano, p. 67.

1980 Opsimasaphus sp.; Romano, pp. 68-70.

1988 Opsimasaphus radiatus (Salter); Morris, p. 253.

Derivation of name. After I. W. Salter who first figured a specimen of the species now described.

Material. Holotype: NMI F20984. Paratypes NMI F20985-F20991 ; BM It. 25696-It. 25703; Geological
Survey Museum (GSM) 12386, and counterparts 12387 and 12840 (the latter figured by Salter 1866, pi. 18, fig.
4 and Harper 1952, pi. 5, fig. 2).

Horizon and locality. Most material is from the Tretaspis Bed and overlying concretions, Brickwork’s Quarry.
The remainder from Locs 190, 209, 210A and from the ditch section at Loc. 26. GSM material is probably
from locality 8 of the Irish Survey Memoir (Hull et al., 1871, see Harper 1952, p. 108).

Diagnosis. Asaphinid with preglabellar field approximately 40 per cent of total cranidial length and
equal in width to almost 140 per cent of the width of the cranidium across the palpebral lobes. Well
marked posterior median glabellar lobe with small tubercle. Front of eye lobes one-third cranidial
length from posterior margin. Anterior margin of frontal glabellar lobe with blunt point; weakly
incised preglabellar furrow. Pygidial axis with up to 14 straight axial rings, pleural fields with 7 or
8 ribs.

Description. Cranidium nearly as wide as long, widest part (excluding posterior borders) at about two-thirds
distance from posterior margin. Glabella occupies just over 60 per cent of cranidial length. Occipital ring about
one-seventh glabellar length, transversely gently convex and more or less flat sagittally. Glabella (excluding
occipital ring) three-quarters as wide as long, gently convex (trs.); longitudinal profile fairly flat but sloping
down gently to anterior margin. Anteriorly glabella evenly rounded, frontal margin delimited by change in
slope rather than preglabellar furrow. Axial furrows absent opposite eye lobes where lateral margin of glabella
indistinct, elevated above the level of the weakly incised anterior and posterior portions of the axial furrows.
Wide, shallow basal glabellar furrows extend in slight curve from occipital furrow towards anterior end of
palpebral lobes but die out before reaching axial furrows. Longitudinal ridge-like median glabellar lobe

694

PALAEONTOLOGY, VOLUME 36

between glabellar furrows increases slightly in height posteriorly and abruptly ends before occipital furrow
where it carries a small median tubercle. Palpebral lobes long (exs.), equal to 20 per cent of maximum cranidial
length; anterior of lobe at 35 per cent of cranidial length from posterior end. Palpebral lobes semi-circular in
outline, flat, relatively narrow and separated from fixed cheeks by very shallow palpebral furrow. Top of
palpebral lobe lies just below level of highest part of glabella. Anterior facial sutures curve strongly outwards
and forwards to maximum cranidial width, then curve evenly forwards to meet in dorsally situated blunt point.
Preglabellar field long (40 per cent of total cranidial length) and wide (140 per cent the cranidial width at the
palpebral lobes), sloping very gently forwards in front of glabella then flattening to anterior margin. Lateral
portions of fixed cheeks slope gently abaxially. Posterior branches of facial sutures imperfectly preserved.
Posterior borders appear to be slightly shorter (exs.) than occipital ring; posterior border furrows deepen
abaxially. Free cheeks poorly preserved but wide (trs.) level with palpebral lobes. Long, wide genal spines with
broad base and possibly incurved posteriorly. Cheeks with doublure of unknown width. Sculpture on cephalon
(present on internal and external surfaces unless stated) consists of : (a) fine ridges on anterior half of glabella
directed subparallel to margins, (b) slightly coarser ridges on posterior half of glabella, concentrically arranged
around median tubercle, (c) ridges on preglabellar field and fixed cheeks running subparallel to each other and
cranidial margins, (d) ridges on palpebral lobes where subparallel to posterior margins and oblique to anterior
margins of lobes, (e) symmetrical, concentrically arranged ridges, anteriorly convex, on occipital ring, (f) faint
longitudinal (exs.) ridges on posterior border (not known if present on internal mould), (g) ridges on free cheeks
running subparallel to margins.

Excluding wings, hypostoma estimated to be approximately 75 per cent as wide as long; widest just posterior
to midlength. Lateral margins quite strongly curved, posterior margin deeply notched with posteriorly directed
prongs at least 25 per cent as long as hypostoma mesially. Suboval convex body slightly wider than long with
delimiting furrow more pronounced posterolaterally. Pair of prominent maculae situated posterolaterally of
convex body, area between maculae at higher level than lateral borders. Anterior wings separated from oval
body by inwardly curved border. Lateral borders covered with fine raised lines subparallel to margins.

Complete thorax unknown. Axis about 25 per cent of thoracic width and gently convex (trs.), axial furrows
shallow. Pleurae of uniform width (exs.), gently curved rearwards from about mid-length (trs.) and extending
into short posterolaterally directed, bluntly rounded spines. Oblique pleural furrows crossing from anterior end
adaxially to near posterior margin at distance of about 75 per cent pleural length (trs.) from axis. Abaxial half
of ventral surface of pleurae covered with fine longitudinal (exs.) ridges.

Pygidium approximately semi-circular in outline with gently curved anterior margin ; length to width ratio
lies between 1:1-4 and 1 : 2-0 ( N = 23). Gently tapering axis very narrow (varying from 10-20 per cent of pygidial
width anteriorly), and between 70-790 per cent of pygidial length. Axial furrows straight and faint, converging
evenly posteriorly at about 20°. Axis stands slightly above flat pleural regions. In lateral view axis highest
anteriorly and posteriorly. Seven (or 8) to 14 axial rings and rounded terminal piece, separated by straight axial
ring furrows; posterior axial rings being shorter (sag.). Seven (occasionally 8) pleural ribs present, curved
gently rearwards with abaxial end slightly wider (exs.) and terminating before margin. Posterior pleurae curve
rearwards more strongly. Interpleural furrows distinct, rarely faint pleural furrows present on anterior pleurae.
Doublure wide, occupying just over half width of pleural field, except posteriorly where constricted behind axis.
Doublure covered with up to twenty subparallel terrace lines lying oblique to pygidial margin. Between these

EXPLANATION OF PLATE 2

Figs 1-10. Birmanites salteri sp. nov. 1-2, 4, NMI F20984 (holotype); internal mould of cranidium, dorsal,
lateral and frontal views respectively, all x 2. 3, NMI F20985; internal mould of hypostoma, ventral view,
x2. 5, NMI F20986; internal mould of hypostoma, ventral view, x 4. 6, NMI F20987; cast of external
mould of hypostoma, ventral view, x 3. 7, NMI F20988; internal mould of pygidium, dorsal view, x 1.
8, NMI F20989; internal mould of pygidium, dorsal view, x 1. 9, 10, NMI F20990, 91 ; internal moulds of
pygidia showing doublure and terrace lines, dorsal views, x L5, x 3, respectively. 1-6, 9-10 from concretions
approximately 1 m above Tretaspis Bed, 7-8 from Tretaspis Bed, Brickwork’s Quarry Shales Member;
Brickwork’s Quarry.

Figs 1 1-13. Decoroproetus sp. 11, NMI FI 4026 (figured by Harper 1952, pi. 5, fig. 3 as NMI 1951/12); internal
mould of cranidium, dorsal view. 12, LM 1988.216.445; cast of external mould of cranidium, dorsal view.
13, LM 1988.216.444, cast of external mould of cranidium, lateral view. All from ?Knockerk House Shales
Member to basal Brickwork’s Quarry Shales Member; Loc. 26. All x 11.

PLATE 2

ROMANO and OWEN, Irish Caradoc trilobites

696

PALAEONTOLOGY, VOLUME 36

main terrace lines are finer striations aligned at angle to former. Postaxially and along downturned border the
terrace lines lie closer together.

The relative abundance of pygidia has allowed some quantitative analysis. Length (sag.) and width (trs.)
measurements on complete shields and axes indicate generally good linear relationships. All tend to show
isometric growth, at least in forms larger than 25 mm wide, and there is no indication of dimorphism or
clustering into instars.

Discussion. Birmanites salteri is fairly close to B. hupeiensis Yi, 1957, from the upper Llanvirn to
lower Llandeilo Miaopo Formation of Yichang, Hubei Province and the broadly contemporaneous
Shihtzupu Formation at Zunyi, Guizhou Province from where Yi’s species was described by Zhou
et al. (1984, p. 17, fig. 3c-/, i—j, m ). Zhou et al. noted that B. hupeiensis and the type species are the
only ones in which the width of the preglabellar area exceeds the palpebral width by 50 per cent (a
figure approached by B. salteri). However, in both of the Asian species the length of the preglabellar
field is approximately equal to that of the glabella (data for B. birmanicus taken from Moore 1959,
p. 0359, fig. 268.7), and the type species has well-marked cephalic sculpture and only 7 or 8 axial
rings on the pygidium. B. hupeiensis lacks cephalic sculpture but does not show the slightly pointed
mesial part of the anterior edge of the cranidium seen in B. salteri. The hypostoma of the Irish
species is broadly similar to that of B. hupeiensis, but is relatively longer and the inner edges of the
posterior fork are considerably less divergent. The pygidia of the Irish species differ from those of
B. hupeiensis in the slightly higher number of axial rings (maximum 14 as opposed to maximum 12)
and pleural ribs (7 or 8 rather than 6), and the generally longer axis (70-790 per cent rather than
55-70 per cent).

Birmanites sp. from the Shihtzupu Formation (Zhou et al. 1984, p. 18, fig. 4a-c, g) has the
pygidial axis of comparable length to that of B. salteri but its cranidium has a much shorter and
narrower preglabellar area. The hypostoma of B. salteri is proportionately longer, and the
hypostomal forks less divergent, than that of B. sp.

Birmanites aff. asiaticus (Petrunina in Repina et al. 1975) described by Zhou and Dean (1986,
p. 754, pi. 59, figs 7, 10-11, 13-15) from the Caradoc of Chedao, Gansu Province, China has a short
preglabellar area (approximately 30 per cent of the glabellar length) and, as noted by Zhou and
Dean, this is comparable to the condition seen in B. asiaticus from the Ashgill of southern Tian-
Shan and B. latus (Angelin, 1851; Kielan 1960) from the Ashgill of Vastergotland, Sweden.

Birmanites sp. of Tripp et al. (1989, p. 36, fig. 5 d-e, i, l ) from the Tangtou Formation (probably
lower Ashgill) of Jiangsu, China has a long preglabellar area but proportions are difficult to assess
from the illustrated specimen. The pygidia from the Tangtou Formation have the axis occupying
65 per cent of the sagittal length, 9 axial rings and 9 pleurae. Tripp et al. considered B. sp. to be
very close to B. juxianensis Ju, from the Huangnehkan Formation of Quxian, Zhejiang, differing
only in the latter’s much narrower pygidial border.

Family nileidae Angelin, 1854
Genus barrandia M'Coy, 1849

Type species. Barrandia cordai M‘Coy, 1849, by monotypy; probably from the Glyptograptus teretiusculus
Shales of early Llandeilo age (Hughes 1979), Powys, Wales.

Barrandia sp. ?nov.

Plate 1, fig. 18

1980 Barrandia sp.; Romano, p. 68.

Material. NMI F20983.

Horizon and locality. Not found in situ , probably from about 0-35-0-40 m above the Tretaspis Bed, Brickwork's
Quarry.

ROMANO AND OWEN: IRISH CARADOC TRILOBITES

697

Description. Specimen oval in outline, maximum width (excluding free cheeks) equal to approximately 75 per
cent of the length. Cranidium nearly flat but steeply downturned anteriorly. Very faint axial furrows which
diverge anteriorly in gentle arc, possibly just reaching anterior margin. Glabellar lobes and furrows absent.
Glabella longer than wide. Facial sutures very poorly preserved but fixed cheek about one-quarter cranidial
width. Though ill-defined, anterior part of ‘palpebral’ area approximately level with midlength of glabella.
Occipital ring, posterior border absent, free cheeks and hypostoma unknown.

Thorax just over twice as long as wide with axis occupying about 35 per cent of thoracic width anteriorly.
Axis narrows gradually posteriorly to just under 75 per cent its anterior width. Axial furrows not well-defined.
Eight thoracic segments ; pleurae parallel-sided and each bears a pleural furrow which runs obliquely from the
anteromesial corner terminating about 75 per cent of the distance (trs.) along pleura. Distal ends of pleurae
curved slightly rearwards.

Pygidium nearly flat; just over twice as wide as long, semicircular in outline, gently forwardly convex along
anterior margin where axis occupies just over 25 per cent of pygidial width. Axis very slightly convex (trs.)
narrowing quite rapidly backwards to terminate about midlength of pygidium. Axial furrows faint and weaken
posteriorly. Very faint axial ring furrow at just over 20 per cent axial length from anterior end. Pair of equally
faint pleural furrows extend posterolaterally from where axial ring furrow meets dorsal furrow. Suggestion of
a second furrow on right pleural field. Posterior border of pygidium declines steeply and shows indistinct
terrace lines.

Discussion. Hughes (1979, p. 164) summarized the important differences between Barrandia and
Homalopteon. Although the present specimen is poorly preserved it shows characteristics of the
former listed by Hughes, in particular the unfurrowed glabella and weakly furrowed pygidium.
Whittard (1961) described five species of Barrandia from the early Llanvirn of the Shelve Inlier,
Shropshire, England. One of these, B. cf. radians , is now included in Homalopteon (Hughes 1979,
p. 164). The Grangegeeth species is a relatively wide form; length to width ratio of the Irish
specimen is 1 : 1 -27 while the Shelve and Builth specimens range from 1 : 1 -36 to 1 : 1 .77 (measurements
taken from plates of Whittard and Hughes but excluding ‘narrow form’ of B. parabolica Whittard,
1961 (pi. 33, fig. 6). The pygidial outline of the Irish species is semicircular with rather sharp
anterolateral corners; this contrasts with the generally elliptical pygidial outlines of the Builth and
Shelve species. The presence of B. sp. ?nov. in beds of early Caradoc age extends the range of this
genus which was previously known to occur only up to the basal Caradoc ( N . gracilis Biozone).
Moreover, if our palaeogeographical interpretation is correct, it also extends the known
geographical distribution of this genus outside the Anglo-Welsh area (Whittard 1961 ; Hughes 1979;
Fortey and Owens 1987).

Family illaenidae Hawle and Corda, 1847
Subfamily illaeninae Hawle and Corda, 1847
Genus illaenus Dalman, 1827

Type species. Entomostracites crassicauda Wahlenberg, 1818, by subsequent designation; from the Crassicauda
Limestone (Llandeilo), Fjacka, Siljan, Sweden.

Illaenus sp.

Plate 1, fig. 17

1952 Illaenus richardsoni Reed; Harper, p. 107, pi. 5, fig. 6.

1980 Illaenus cf. richardsoni Reed; Romano, p. 66.

1980 Illaenus sp.; Romano, pp. 67-68.

Material. NMI F14032/A-B (the internal mould which was figured by Harper 1952, pi. 5, fig. 6 as 1951/17);
LM 1988.216.447 and BM It. 25704-It. 25705.

Horizon and locality. NMI FI 4032 from Loc. 1 ; LM 1988.216.447 from Loc. 3; BM It .25704 — It.25705 from
Collon Quarry.

698

PALAEONTOLOGY, VOLUME 36

Discussion. Whilst Harper (1952) considered that his material agreed with Reed’s (1914, p. 25)
description of I. richardsoni from the Balclatchie and Lower Ardwell groups (lower Caradoc) at
Girvan, he pointed out that it differed only from the Scottish specimens in having more clearly
defined axial furrows. To Harper’s observations may be added that the Irish cephalon is more
strongly convex posteriorly (sag.) and has wider (trs.) free cheeks than Reed’s material. We
therefore consider it best be left in open nomenclature. A cranidium housed in the Royal Museum
of Scotland (RMS-Geol 1870.12.950), was probably collected from the basal Knockerk House
Sandstones Member and probably belongs in Illaenus sp. It is more elongate than the present
figured specimen but this may be the result of deformation.

Fragmentary specimens of internal and external moulds of cranidia and free cheeks are here
assigned to ‘illaenid indet.’ from the Tretaspis Bed and basal part of the Knockerk House
Sandstones Member.

Family proetidae Salter, 1864
Subfamily tropidocoryphinae Pribyl, 1946
Genus decoroproetus Pribyl, 1946

Type species. Proetus decorus Barrande, 1846, by original designation; from the Liten Formation (Wenlock),
Lodenice, Prague, Czech Republic.

Decoroproetus sp.

Plate 2, figs 1 1-13

1952 Proetus ardmillanensis Begg; Harper, p. 107, pi. 5, fig. 3.

1980 ? Decoroproetus sp.; Romano, pp. 68, 70.

Material. NMI F14026 (figured by Harper 1952, pi. 5, fig. 3 as NMI 1951/12); LM 1988.216.443-446 (445
and 446 are part and counterpart).

Horizon and locality. From small quarry at Loc. 26.

EXPLANATION OF PLATE 3

Figs 1-6. Tretaspis aff. reticulata Ruedemann, 1901. 1, NMI F20992; internal mould of cephalon, frontal view,
x 3. 2, NMI F20993; internal mould of cephalon, dorsal view, x 2-5. 3, NMI F20994; internal mould of
incomplete fringe, oblique frontal view, x 2-5. 4, NMIF20995; cast of external mould of incomplete
cephalon, dorsal view, x 4. 5, NMI 20996/B; cast of external mould of pygidium, dorsal view, x 6.
6, NMI F20997; internal mould of pygidium, dorsal view, x 6. All from Tretaspis Bed, Brickwork’s Quarry
Shales Member; Brickwork’s Quarry.

Figs 7, 1 1. Decordinaspis bispinosa Harper and Romano, 1967. 7, NMI F20998; internal mould of cephalon,
dorsal view, x F5. 11, NMI F20999; internal mould of cephalon, oblique frontal view, x4. Both from
concretions approximately 1 m above Tretaspis Bed, Brickwork’s Quarry Shales Member; Brickwork’s
Quarry.

Figs 8-10, 12-13. Ampyx aff. repulsus Tripp, 1976. 8, NMI F21000; internal mould of cranidium, dorsal view,
x 2-5. 9, NMI F21001 ; internal mould of frontal part of glabella, lateral view, x 3. 10, 12, NMI F21002;
internal mould of cranidium, dorsal and lateral views, respectively, both x 3. 13, NMIF21003; internal
mould of damaged thorax and pygidium, dorsal view, x 2. 8-10, 12 from concretions approximately 1 m
above Tretaspis Bed, Brickwork’s Quarry Shales Member; Brickwork's Quarry. 13 from ?Knockerk House
Shales Member to basal Brickwork’s Quarry Shales Member, Locality 26.

Figs 14-16. Ceraurinellal sp. 14, NMI F21004; internal mould of damaged cranidium, dorsal view, x 3.
15, NMI F21005; internal mould of damaged pygidium, dorsal view, x 3. 16, NMI F21006; internal mould
of incomplete cranidium, dorsal view, x 2. All from basal Knockerk House Sandstones Member, Collon
Quarry.

PLATE 3

ROMANO and OWEN, Irish Caradoc trilobites

700

PALAEONTOLOGY, VOLUME 36

Discussion. Although the material is scant and poorly preserved, the specimens are best assigned to
Decoroproetus Pribyl rather than Astroproetus Begg (see Owens 19736) on account of the more
evenly rounded glabella and lack of occipital lobes. The absence of striations however, is typical of
Astroproetus. Harper (1952) identified the present form as Proetidella ardmillanensis (Begg, 1946),
now referred to Decoroproetus jamesoni (Reed, 1914; see Owens 1973a, p. 46, pi. 8, fig. 1 1) from the
Balclatchie Group, Girvan. As in the Grangegeeth cranidium the preglabellar area is long and
occupies about 30 per cent of the sagittal cephalic length. However, the preglabellar field of the Irish
cranidium is shorter than that of Begg’s specimen.

Family trinucleidae Hawle and Corda, 1847
Subfamily trinucleinae Hawle and Corda, 1847
Genus tretaspis IVTCoy, 1849

Type species. Asaphus seticornis Hisinger, 1840, by subsequent designation of Bassler (1915), from the Fjacka
Shale Formation (lower Ashgill) of Dalarna, Sweden.

Tretaspis aff. reticulata Ruedemann, 1901
Plate 3, figs 1-6; Text-fig. 4

1967 Tretaspis kiaeri (Stormer) subsp.; Brenchley et al., p. 298.

CO

c

CD

E

o

CD

Q.

CO

CD

-O

E

D3

4

3

2

1

0

Radius number of posterior Ei pit

Number of pits along posterior
margin of fringe

CO

c

CD

E

o

CD

CL

CO

CD

-O

E

D3

Radius number of posterior ln pit

text-fig. 4. Histograms showing the variation in selected fringe characteristics in Tretaspis aff. reticulata
Ruedemann, 1901. Specimens from Tretaspis Bed, Brickwork’s Quarry Shales Member; Brickwork’s Quarry.

ROMANO AND OWEN: IRISH CARADOC TRILOBITES

701

1977 Tretaspis ; Brenchley et al., p. 74.

1980 Tretaspis sp.; Romano, pp. 68, 70.

Material. NMI F20992-F20997 ; BM It.25706-It.257 16.

Horizon and locality. Tretaspis Bed and immediately overlying beds at Brickwork’s Quarry, Loc. 26 and
?Loc. 190.

Description. This species is sufficiently abundant for the fringe pit statistics to be described. Arcs E4_2, I4 3 and
In complete, with pits arranged in a single set of radii with lists bounding I2 and I3. Few adventitious pits
present on the genal prolongation. Arc E4 contains 22-5-28-5 pits (mean 25-5, N = 22; half-fringe counts).
E2 has the same number of pits as Ex in 2 (of 48) specimens, lacks the posterior pit in 41 specimens and the
posterior two in 5 specimens. Three specimens (of 48) contain an adventitious E pit posteriorly. I4 shares sulci
with the E arcs anteriorly but becomes discrete on the lateral or even anterolateral part of the fringe. In In,
20-25 pits are present (mean 22-5, N — 12). A fifth I arc (I4) is invariably present and is complete mesially in
most specimens (16 of 19). In the remaining specimens up to two pits (half-fringe) are missing mesially. Three
of 14 specimens show the arc extending to the posterior margin; the remainder containing 4-15-5 pits.
Specimens with I4 complete mesially have at least 7 pits in this arc. Seven to 13 pits present along the posterior
margin of the fringe (mean 9, N = 34).

Discussion. The Grangegeeth Tretaspis belongs in the T. sagenosa Group of Owen (1980, p. 718),
the typical representatives of which are Laurentian mid-Llandeilo to early Caradoc species having
a broad fringe and high peripheral pit count. The group gave rise to the European T. moeldenensis
Group in the late Caradoc (Owen 1980, 1987) and was discussed recently by Ingham and Tripp
(1991, p. 41).

The Irish species is closest to the broadly contemporaneous T. reticulata Ruedemann (see also
Whittington 1941, pi. 6, figs 26-27 , 31 ; Stauble 1953, figs 21-24) from the Rysedorph Conglomerate
(probably Blackriveran) of New York. Ruedemann's species is based on a limited amount of poor
material, but the fringe pitting of the Irish material seems to encompass that of T. reticulata , and
the only distinguishing features of the Grangegeeth specimens are the more pronounced cephalic
brim and possibly longer eye ridges. A specimen included in T. reticulata by Whittington (1941,
pi. 6) from probable Blackriveran strata near Tenth Legion, Virginia has a well-developed brim as
does one from the Lower Balclatchie Group at Girvan figured as T. cf. reticulata by Ingham and
Tripp (1991, fig. 8 d-e). The Grangegeeth form differs from both of these in its slightly lower E4 pit
count (up to 28-5 compared with about 32 and 29-5, respectively).

All the above forms have the pits in I4 distinct from those in In, even in those specimens of T. aff.
reticulata where the arc is short. In T. sagenosa Whittington (1959, p. 448, pis 24-27, pi. 28, figs 1-8)
from the lower Edinburg Formation (early Caradoc), this arc is restricted to a few pits situated very
close to those of In in front of the axial furrow. T. aff. reticulata also differs from Whittington’s
species in consistently lacking E3 pits (present in a few specimens of T. sagenosa ), in having the
lateral eye tubercles situated farther forward and closer to the glabella and in having fewer large
apodemes on the pygidial axis.

T. canadensis Stauble (1953, p. 203, figs 17-20, 22) from a debris flow in the early Caradoc Citadel
Formation in Quebec City, Canada and T. eximia Ingham and Tripp (1991) from the mid-
Llandeilo at Girvan, Scotland have arc I4 invariably complete and I5 developed anteriorly. The
latter also has a complete E3 arc and a very short eye ridge.

Genus decordinaspis Harper and Romano, 1967

Type species. Decordinaspis bispinosa Harper and Romano, 1967, by monotypy ; from the Brickwork’s Quarry
Shales Member, Knockerk Formation of early Caradoc age; Grangegeeth, Co. Meath, eastern Ireland.

702

PALAEONTOLOGY, VOLUME 36

Decor dinaspis bispinosa Harper and Romano, 1967
Plate 3, figs 7, 1 1

*1967 Decordinaspis bispinosa nov. sp.; Harper and Romano, p. 305, pi. 8, figs 1-2, pi. 9, figs 1^4
[described].

1967 Decordinaspis bispinosa Harper and Romano; Brenchley et a!., p. 298 [listed only].

1975 Decordinaspis bispinosa Harper and Romano, 1967; Hughes et al., p. 566, p. 5, figs 59-63.

1980 Decordinaspis bispinosa Harper and Romano; Romano, p. 68.

Material Holotype: NMI G. 104/1965 (Harper and Romano 1967, pi. 8, figs 1-2). Paratype: NMI G. 105/1965
(Harper and Romano 1967, pi. 9, figs 1-3) (see also Hughes et al. 1975, pi. 5, figs 59-62); NMI F20998,
F20999; BM It.257 1 7-It.257 1 9.

Horizon and locality. Concretions and enclosing shale just above the Tretaspis Bed, Brickwork’s Quarry.

Discussion. Following the discovery of further specimens, minor additional comments can be added
to the description given in Harper and Romano (1967). The largest specimen now known is
approximately c. 30 mm across the posterior border of the cephalon. The SI furrows are
sometimes slightly larger than S2 and in some specimens the former are continuous with the
occipital furrow. The shape of SI is variable but does not appear to be related to maturity; it also
varies either side of the glabella in a single specimen. On the internal moulds of three new specimens
a very faint eye ridge is present, directed slightly forwards from the lateral eye tubercle towards S3,
where it terminates in the axial furrow (this feature is known in Tretaspis and reedolithines but not
in Nankinolithus with which Decordinaspis has also been thought to have affinities — see Hughes
et al. 1975, p. 566). The fossulae situated in the axial furrows adjacent to the fringe are larger and
more pronounced than in the original material.

Decordinaspis has been recorded from the Upper Tramore Volcanic Formation at Kilbride in
southeast Ireland where the associated rich brachiopod fauna is closely comparable with that of the
Ballymoney Formation at Courtown, of Harnagian to Soudleyan age (Mitchell et al. 1972; Carlisle
1979). More recent extensive collecting by Dr M. Parkes has failed to confirm the presence of
Decordinaspis here (Parkes 1990) although another trinucleine is present and is currently under
investigation by one of us (A.W.O.) and Parkes. It may be this which was identified earlier as
Decordinaspis.

Family raphiophoridae Angelin, 1854
Genus ampyx Dalman, 1827

Type species. Ampyx nasutus Dalman, 1827, by monotypy; neotype described by Whittington (1950, p. 554)
from the Asaphus Limestone, upper Arenig of Ostergotland, Sweden.

Ampyx aff. repulsus Tripp, 1976
Plate 3. figs 8-10, 12-13

1967 Ampyx sp. cf. A. costatus Angelin; Brenchley et al., pp. 298, 302, pi. 7, figs 1-3.

1980 Ampyx sp.; Romano, pp. 68, 70.

Material NMI F21000-F21003; NMI G.109/1965 and NMI G.l 10/1965 (figured in Brenchley et al 1967,
pi. 7, figs 1-3); BM It.25720-It.25724.

Horizon and locality. Loc. 208 and ditch section at Loc. 26. Tretaspis Bed and concretions immediately above;
Brickwork’s Quarry.

Discussion. This material is close to a number of species listed by Tripp (1976, p. 393) as being
typified by A. mammillatus Sars, including: A. virginiensis Cooper, 1953, A. repulsus Tripp, 1976,

ROMANO AND OWEN: IRISH CARADOC TRILOBITES

703

A. americanus Safford and Vogdes, 1889 and A. incurvus Reed, 1906. To this list may be added
A. austinii Portlock, 1843 redescribed by Owen et al. (1986) from the upper Caradoc Raheen
Formation of Co. Waterford. Tripp listed the distinguishing features of this species-group as being
the short glabella with steep anterior slope, strongly developed second muscle scars and broad
posterior border furrows. The Grangegeeth form is closest to A. repulsus from the basal Superstes
Mudstones (middle Llandeilo) at Girvan, southwest Scotland, differing only in having a more
steeply descending glabella to the anterior border, a smooth rather than shallowly pitted cephalic
and pygidial surface and in the pygidial segmentation being a little more obvious. The more steeply
descending glabella in front of the glabellar spine, narrower anterior part of the fixed cheek, less
sigmoidal facial suture and longer pygidium distinguish it from A. virginiensis Cooper, 1953 (p. 15,
pi. 7, figs 1-11, 18), from the lower Edinburg Formation (lower Caradoc). The glabellar
proportions are similar to those of A. hornei Etheridge and Nicholson, (in Nicholson and Etheridge
1879; see Reed 1906, pi. 3, figs 8-9, and Tripp 1980, pi. 3, figs 8-9) from the lower Caradoc
Balclatchie Group at Girvan, but the Scottish species has a well-developed pair of genal ridges.
A. aff. repulsus is similar to A. austinii Portlock but its shorter glabella is again diagnostic.

Family cheiruridae Hawle and Corda, 1847
Subfamily cheirurinae Hawle and Corda, 1847
Genus ceraurinella Cooper, 1953

Type species. Ceraurinella typa Cooper, 1953, by original designation; from the Echinosphaerites beds of the
Edinburg Limestone of early Caradoc age, Shenandoah Valley, Virginia, USA.

Ceraurinella ? sp.

Plate 3, figs 14-16

1980 Ceraurinella sp; Romano, p. 67.

Material. NMI F2 1 004— F2 1006.

Horizon and locality. Loc. 1.

Discussion. Though incomplete, the material shows closest affinies with species placed in Ceraurinella
although the closely related upper Ordovician genus Xylabion Lane, 1971, also includes species with
some similar features to those found in the Grangegeeth form. The pygidium of the Irish form
differs from that of C. typa Cooper, 1953 (see Whittington and Evitt 1954) in that the second pair
of spines is directed more outwards and there is no third pair of spines. In some of the figured
pygidia of C. nahanniensis Chatterton and Ludvigsen, 1976 (p. 55, pi. 9, especially figs 13, 15, 19)
the third pair of spines are rudimentary and terminate well in front of the second pair. This
condition approaches that present in the single poorly preserved pygidium from Collon.

The two figured cranidia of Ceraurinella! sp. show differences in the length of the glabellar
furrows, shape of LI and shape of the frontal glabellar lobe. These might reflect basic
morphological differences but more probably are the result of deformation.

Subfamily deiphoninae Raymond, 1913
Genus sphaerocoryphe Angelin, 1854

Type species. Sphaerocoryphe dentata Angelin, 1854, by subsequent designation of Vogdes (1890); from the
upper Ordovician of Sweden.

704

PALAEONTOLOGY, VOLUME 36

Sphaerocoryphe cf. pemphis Lane, 1971
Plate 4, figs 1-3

1967 Sphaerocoryphe thomsoni (Reed); Brenchley et al. , pp. 298, 302, pi. 7, figs 5-6.

71980 Sphaerocoryphe sp. indet. ; Romano, p. 67.

1980 Sphaerocoryphe cf. thomsoni (Reed); Romano, p. 69.

Material. NMI F14064 (= NMI G. 106/1965 of Brenchley et al. 1967, pi. 7, figs 5-6), NMI F2 1007-
F21008; BM It.25725-It.25728.

Horizon and locality. Most specimens are from the Tretaspis Bed, Brickwork’s Quarry; some of the material
is possibly from the base of the Knockerk Flouse Sandstones Member, Collon Quarry.

Discussion. The Grangegeeth species possesses two pairs of secondary spines on the fixed cheeks, a
feature shared with Llandeilo to Ashgill species such as S. dentata Angelin, 1854 ( = S. thomsoni
Reed, 1906; see Kielan-Jaworowska et al. 1991), S. pemphis Lane, 1971, S. kingi Ingham, 1974,
S. punctata Angelin, 1854, ?S. atlantiodes Opik, 1937, S. ludvigseni Chatterton, 1980, and S. murphyi
Owen et al., 1986. The presence of such spines would appear to be a useful feature in identifying
species-groups but, as pointed out by Owen and Bruton (1980, p. 28), Shaw’s (1968) work showed
this to be a variable feature. However, of these species, the Grangegeeth form appears closest to
5. pemphis from the lower Caradoc Balclatchie and Lower Ardwell groups at Girvan, differing
mainly in details of the profixigenal spines. In particular those of the Grangegeeth species are not
separated by a short straight lateral margin and the posterior branch of the facial suture does not
cut the base of the anterior spine as it does in the Girvan species.

The specimens from the Knockerk House Sandstones Member are too incomplete to allow
specific identification, but the bulbous frontal glabellar lobe is similar in size and proportions to
S. cf. pemphis in the overlying Brickwork's Quarry Shales Member. The sculpture on the Knockerk
House Sandstones glabellae appears to be more reticulate than granular, but the specimens are
provisionally referred to the Brickwork’s Quarry Shales form.

EXPLANATION OF PLATE 4

Figs 1-3. Sphaerocoryphe cf. pemphis Lane, 1971. 1, NMI F21007; internal mould of cephalon, dorsal view.

2, NMI G106.1965 (figured by Brenchley et al. 1967, pi. 7, figs 5-6); internal mould of cephalon, dorsal view.

3, NMI F21008; cast of external mould of pygidium, dorsal view. All from Tretaspis Bed, Brickwork’s
Quarry Shales Member; Brickwork’s Quarry. All x 3.

Fig. 4. Sphaerexochus sp. indet. NMI F21009; internal mould of incomplete cranidium, lateral view; horizon
and locality as above, x 3.

Fig. 5. Acanthoparyphal sp. NMIF21010; internal mould of incomplete cranidium, oblique lateral view;
horizon and locality as above, x 3.

Fig. 6. Cybelinae indet. cf. ‘ Deacybele' pauca Whittington, 1965. NMI F21011 ; internal mould of incomplete
cranidium, dorsal view; Tretaspis Bed, Brickwork’s Quarry Shales; Brickwork’s Quarry, x4.

Fig. 7. Cybelinae indet. LM 1988.216.452; internal mould of incomplete cranidium, dorsal view; ?Knockerk
House Shales Member; Loc. 38, x 3.

Figs 8, 9. Atractopygel sp. 8, GSI F000879 (figured by Harper 1952, pi. 5, fig. 1 as GSI/JCH 1); internal mould
of incomplete individual, dorsal view, x F5. 9, NMI F14024/B; internal mould of incomplete cranidium,
dorsal view, x4. Basal Knockerk House Sandstones Member; Locs 1 and 3 respectively.

Figs 10-12, 14. Flexicalymene sp. ?nov. 10, NMI F21012; internal mould of pygidium, dorsal view. 11-12, 14,
NMI F21013; 11-12, internal mould of damaged cranidium, dorsal and lateral views; 14, cast of external
mould, dorsal view. All from Tretaspis Bed, Brickwork’s Quarry Shales Member; Brickwork’s Quarry. All
x 6.

Figs 13, 15. Gravicalymene sp. NMI F21014; internal mould and cast of external mould of incomplete
cranidium, dorsal views; Basal Knockerk House Sandstones Member; Collon Quarry, x 3.

sffiWPM

PLATE 4

ROMANO and OWEN, Irish Caradoc trilobites

706

PALAEONTOLOGY, VOLUME 36

Subfamily sphaerexochinae Opik, 1937 (emend. Lane and Owens 1982, p. 52)
Genus sphaerexochus Beyrich, 1845

Type species. Sphaerexochus mirus Beyrich, 1845, by monotypy; from the Wenlock of Bohemia.

Sphaerexochus sp. indet.

Plate 4, fig. 4

1980 Sphaerexochus sp. indet.; Romano, p. 69.

Material. NMI F21009.

Horizon and locality. Concretion approximately 1 m above the Tretaspis Bed, Brickwork’s Quarry.

Discussion. Sphaerexochus is common in the Girvan area (see Tripp in Williams 1962) from where
Tripp figured species from the confinis Flags (1962), Albany mudstones (1965), Stinchar Limestone
(1967, 1979), and Superstes Mudstones (1976). Tripp commented on the extreme variability within
some of these species, particularly in the length/width proportions of the glabella. The relatively
wide glabella of the Irish specimen is a feature of all of the above Scottish species and, in view of
Tripp’s comments, clearly cannot be used as a diagnostic characteristic. The three pairs of glabellar
furrows in the present specimen, is a feature mentioned by Tripp for his species S. eurys and
S. arcuatus , from the confinis Flags and Superstes Mudstones respectively. There is also a suggestion
of such structures in the Stinchar Limestone species S. filius Tripp, 1979 (pi. 38, fig. 18). As Tripp
remarked (1976, p. 400), these three Scottish species may only be reliably distinguished by their
pygidial characteristics and until more complete Irish material is recovered it would seem preferable
to leave the specimen in open nomenclature.

Genus acanthoparypha Whittington and Evitt, 1954

Type species. Acanthoparypha perforata Whittington and Evitt (1954, p. 72), by original designation; from the
Edinburg Limestone (early Caradoc) of Virginia, USA.

Acanthoparypha ? sp.

Plate 4, fig. 5

1980 Acanthoparypha sp.; Romano, p. 69.

Material. NMI F21010.

Horizon and locality. Concretion approximately 1 m above the Tretaspis Bed, Brickwork’s Quarry.

Discussion. Following Whittington and Evitt (1954, p. 71), the absence of an occipital spine would
preclude this cranidium from being assigned to Nieskowskia Schmidt, 1881, although Lane (1971,
p. 66) pointed out that this is a variable feature. The Grangegeeth form differs from the type species
of Acanthoparypha , A. perforata, in possessing smaller tubercles, a more parallel-sided glabella, and
less even longitudinal convexity to the glabella (cf. Whittington and Evitt 1954, pi. 14, fig. 17;
pi. 15, fig. 11). It is closer to A. chiropyga Whittington and Evitt, 1954, from the Lincolnshire
Limestone of Virginia, but is again distinguishable by its less tapered glabellar outline, more
triangular basal glabellar lobes and longitudinal convexity (cf. Whittington and Evitt, 1954, pi. 29,
fig. 6). Tripp (1967, p. 63) assigned a specimen to Acanthoparypha sp. from the upper Stinchar
Limestone at Girvan. Tripp did not figure the single incomplete cranidium, but the stated weak
convexity <~>f the glabella is unlike that seen in the Irish species. The specimen figured as
Acanthoparypha or Pandaspinapvga sp. by Tripp (1979, pi. 38, fig. 20) differs from the present

ROMANO AND OWEN: IRISH CARADOC TRILOBITES

707

species in having more curved SI. Lane and Owens (1982, p. 55) considered these two genera to be
possibly synonymous. A Shropshire species, A. stubblefieldi (Bancroft) from the Harnagian
Smeathen Wood Beds has a characteristic bend in SI and relatively longer S2 and S3; also the
glabella is only moderately convex longitudinally (Dean 1961, pi. 49, figs 3-6, 1 1).

Family encrinuridae Angelin, 1854
Subfamily cybelinae Holliday, 1942

Cybelinae indet. cf. ‘ Deacybele' pauca Whittington, 1965
Plate 4, fig. 6

1980 Deacybele cf. pauca Whittington; Romano, p. 69.

Material. NMI F2101 1 ; BM It.25729-It. 25730.

Horizon and locality. NMI F21011 and BM It. 25729 from the Tretaspis Bed; BM It. 25730 from concretions
approximately 1 m above the Tretaspis Bed, Brickwork’s Quarry.

Discussion. The status of Deacybele and its separation from Cybeloides Slocum, 1913 has been
questioned but remains unresolved (Owen 1981 ; Owen and Romano in Harper et at. 1985; Owen
et al. 1986). The following discussion is restricted to comparisons with other species which do not
show coalescence of the glabellar lobes. The Grangegeeth species is very close to ‘Z).’ pauca
Whittington, 1965, differing only in the relatively longer frontal glabellar lobe of the Bala specimens
(though the Irish specimen is imperfectly preserved), and more pedunculate eye of the Irish specimen
(presence of genal spines in the present material is unknown). " D.' pauca occurs in the Gelli-grin
Calcareous Ashes of early Longvillian age. Other similar species are ‘Z>.’ gracilis Nikolaisen, 1961
(see Owen and Bruton 1980, pi. 8, figs 14—16) from the late Caradoc Solvang Formation (‘Upper
Chasmops Limestone’) of Norway, and ‘Cybelinae cf. ‘C. ’ mchenryi Reed’ (Owen et ai 1986,
p. 108, figs 49-51) from the upper Caradoc Raheen Formation of southeast Ireland. In the
Norwegian and Raheen species, the eyes are more posteriorly placed, while in the Raheen form the
frontal lobe and posterior border are longer (sag. and exs.).

Cybelinae indet.

Plate 4, fig. 7

1952 Cybele cf. revaliensis Schmidt; Harper, p. 89.

1980 Deacybele sp. ; Romano, p. 67.

Material. LM 1988.216. 452^453 (part and counterpart).

Horizon and locality. Loc. 38.

Discussion. This form differs from cf. ‘ Deacybele ’ pauca in the following details: the cranidium is
proportionally wider, the lateral glabellar lobes are longer (trs.) and median lobe narrower, the
apodemes in the occipital furrow posterior to LI are less well marked, the palpebral lobes are level
with S2, the cheeks are pitted, and the glabella is probably smooth; the posterior borders widen
(exs.) considerably abaxially. There is also a ?short, slim genal spine and a small elongate (sag.)
indentation situated on the anterior slope of the frontal glabellar lobe. The large lateral glabellar
lobes, more anteriorly placed palpebral lobes, and presence of genal spines serve to distinguish this
form from Atractopyge revaliensis Schmidt, 1881 (pi. 13, figs 20 a-b) from the upper Llanvirn and
Llandeilo of the Baltic area.

The large lateral glabellar lobes, the possible presence of a median pit or depression on the frontal
lobe, and the suggestion ofian incipient branching of S3 (?caused by the presence of two apodemes
in this furrow) may indicate that this cranidium belongs in Stiktocybele Ingham and Tripp, 1991,
a genus founded on their new species S. bathytera from the Llandeilo Doularg Formation at Girvan.

708

PALAEONTOLOGY, VOLUME 36

It is differentiated from all other cybelines by the presence of 13 thoracic segments, the seventh of
which is macropleural. It shares with Cybelurus Levitskiy, 1962 and Lyrapyge Fortey, 1980 the
presence of a deep cleft on the frontal glabellar lobe. Ingham and Tripp (1991) also included
‘ Cybele' balclatchiensis Reed, 1914 from the Balclatchie Group at Girvan in Stiktocvbele (see Tripp
1980, p. 131 for full synonymy). The Grangegeeth form is provisionally distinguished from both
species by its considerably more slender genal spine.

Genus atractopyge Hawle and Corda, 1847

Type species. Calymene? verrucosa Dalman, 1827, by monotypy; probably from the Crug Limestone (Ashgill)
of South Wales (see Dean 1974; Price 1984).

A tractopyge ? sp.

Plate 4, figs 8-9

1952 Cybele ( Cybe/ella ) cf. rex (Nieszkowski); Harper, pp. 88, 106, pi. 5, fig. 1.

1980 Cybellela cf. rex (Nieszkowski); Romano, pp. 66-67 .

1988 Cybellela cf. rex (Nieszkowski, 1857); Morris, p. 64.

Material. GSLF00879 (figured as GSI/JCH by Harper 1952, pi. 5, fig. 1); NMI F14024/a-b;
LM 1988.216.403).

Horizon and locality. GSLF00879 from Loc. 1; F: 14024 and LM 1988.216.403 from Loc. 3; other material
from Collon Quarry.

Discussion. This material was originally referred to Cybellela Reed, 1928, a genus which was
maintained by Whittington (1965, p. 43) having earlier been placed in the synonymies of Cybele (by
Opik 1937) and Atractopyge (by Henningsmoen in Moore 1959). Dean (1971a, 1974) redescribed the
type species of the latter genus, A. verrucosa, and suggested that two of the species retained in
Cybellela by Whittington fC.’ aspera and ‘C. ’ dentata ) might best be reassigned to Atractopyge,
leaving only C. rex Nieszkowski, 1857 (the type species) and C. coronata Schmidt, 1881 in Reed’s
genus. Both species, from the middle Ordovician of Estonia, are in need of redescription. Ludvigsen
(1979, p. 47) tentatively ascribed a new species, C.? thor, from the middle Ordovician of western
Canada to Cybellela , but noted its similarity to other cybeline genera. Preliminary results of a
multivariate study of the Cybelinae by one of us (A.W.O.) and R. P. Tripp suggest that the Baltic
species of Cybellela fall well within the range of Atractopyge, and C.? thor is closest to Bevanopsis
Cooper, 1953. The Grangegeeth material is therefore provisionally placed in Atractopyge.

A.? sp. closely resembles the illustrations by Schmidt (1881) of" Cybele' rex and ' C. ' coronata in
gross cranidial and glabellar proportions and in possessing a long (sag., exs.) spinose preglabellar
area. Both Baltic species require redescription before detailed comparison can be made. The long,
spinose cranidial border of A.? sp. distinguishes it from A. condylosa Dean, 19716, from the
Llandeilo-lowest Caradoc of Newfoundland, A. michelli Reed, 1914 (see Morris and Tripp 1986;
Tripp 1980) from the Balclatchie and Lower Ardwell groups at Girvan, and A. killochanensis Tripp,
1954, from the Kiln Mudstones (Upper Ardwell Group). The frontal lobe is a little more expanded
(trs.) than that of A. michelli , but less so than those of A. condylosa and especially A. killochanensis.
The anterior cranidial border, absence of prominent paired glabellar tubercles and larger glabellar
lobes distinguish the Grangegeeth species from A. petiolulata Tripp, 1976, from the Llandeilo at
Girvan (see also Ingham and Tripp 1991, fig. 14/).

ROMANO AND OWEN: IRISH CARADOC TRILOBITES

709

Family calymenidae Milne Edwards, 1840
Subfamily flexicalymeninae Siveter, 1977
Genus flexicalymene Shirley, 1936

Type species. Calymene Blumenbachii var. Caractaci Salter, 1865, by original designation ; from the Woolstonian
and Marshbrookian (mid-Caradoc) of south Shropshire, England.

Flexicalymene sp. ?nov.

Plate 4, figs 10-12, 14

1967 calymenid; Brenchley et al., p. 298.

1980 lOnnicalymene sp.; Romano, p. 69.

Material. NMI F2 1 0 1 2-F2 1 0 1 3 ; BM It.2573 l-It.25733.

Horizon and locality. Tretaspis Bed; Brickwork’s Quarry.

Description. Cranidium about twice as wide as long. Preoccipital glabella about as wide as long, subparabolic
in outline with only very gently rounded or blunt-ended anterior margin and nearly straight lateral margins
anterior to LI. Glabella moderately convex transversely; in longitudinal profile it curves evenly forwards from
about level with SI to anterior part of frontal lobe, then nearly vertically to preglabellar furrow. Occipital ring
longest mesially, narrowing gradually abaxially behind LI ; near axial furrows there is slight development of
nodes. In lateral view occipital ring is gently convex. Behind median lobe occipital furrow transversely straight,
symmetrical in cross-section (sag.) and fairly shallow; behind LI it swings posteriorly a little and deepens. LI
large, subrectangular in outline slightly longer than wide and just over 25 per cent basal width of glabella. L2
suboval in outline, highest adaxially and sloping steeply downwards anteriorly ; just over half the length (exs.)
of LI. L3 small, elongate (trs.), about half the length (exs.) of L2. SI shallow from occipital furrow over adaxial
neck of LI; deepening and turning abaxially to anterolateral corners of LI where directed outwards and
forwards to meet axial furrow. Short anterior branch of SI crosses inner neck of L2. S2 shallower than SI,
abaxially more transverse, adaxially turn rearwards around L2. S3 short, narrower (exs.) than S2, barely reach
axial furrows. Frontal glabellar lobe with nearly straight anterior margin, three times as wide as long, level with
or just projecting beyond fixed cheeks. Axial furrows deep, curved around LI, then converge evenly forwards
at about 35° to join with transverse preglabellar furrow. Shallow anterior pits in axial furrows just anterior to
S3. In lateral view preglabellar furrow passes quite sharply into upturned, convex (sag.) ‘anterior border’ (see
Ingham 1977, p. 90 for discussion of the problem of terminology and homology of the preglabellar area in
calymenids). Anterior margin gently curved in dorsal view, quite strongly arched in frontal view. Posterior
border increases in width (exs.) abaxially; posterior border furrows of constant width, dying out before
reaching margins of fixed cheeks. Fixed cheeks gently convex transversely, slope more steeply anteriorly.
Posterior end of palpebral lobes opposite anterior of LI, anterior end level with middle to anterior end of L2.
Anterior branch of facial suture continues forwards in very broad curve to anterior margin. Posterior branch
directed very gently abaxially forwards then bends evenly posterolaterally to cranidial margin.

Free cheeks, hypostoma and thorax unknown. Pygidium just under twice as wide as long, and two and a half
times as wide as axis. Axis approximately 80 per cent length of pygidium; with five or six axial rings and terminal
piece. Axial rings of approximately constant width (sag. and exs.) separated by wide ring furrows one to four;
furrows five and six are indistinct. Axial furrows distinct except around terminal piece. Pleural regions with five
pairs of pleural furrows and four (?five) pairs of interpleural furrows. Pleural furrows widest (exs.) and deepest
adaxially, dying out at steeply downturned pygidial border. First pair extend onto smooth pleural facet.
Interpleural furrows fainter, especially proximally, and longer than pleural furrows. Anterior pleural bands
narrower than posterior bands, particularly on more anterior pleurae. Sculpture of cranidium and pygidium
granulate except in deepest parts of furrows.

Discussion. The short preglabellar area and posteriorly placed eyes of this form prompted Romano’s
(1970) generic placement; it was later only tentatively so assigned (Romano 1980). The status of
Onnicalymene has been questioned by Siveter (1977), who retained it as a subgenus of Flexicalymene.
Ingham (1977) regarded it as a junior synonym of Flexicalymene which has been followed by Owen

710

PALAEONTOLOGY, VOLUME 36

and Bruton (1980) who suggested that the informal term ‘ onniensis species group’ would be
sufficient to characterize those species previously assigned to Onnicalymene. The Grangegeeth form
clearly is best assigned to Flexicalymene ( sensu Siveter 1977, p. 353). The short preglabellar area,
fairly sharp change in slope from the 'preglabellar furrow’ to 'anterior border’, posteriorly placed
eyes, narrow anterior portion of fixed cheeks and straight sides to the glabella serve to distinguish
this species from other forms of this genus. Flexicalymene sp. ?nov. is fairly close to Flexicalymene
shirleyi Tripp, 1954, from the Kiln Mudstones, middle Caradoc, of Craighead Quarry, Girvan, but
the Grangegeeth form has a more gentle longitudinal curvature to the cranidium, a proportionately
longer glabella, more posteriorly placed eyes and fewer pygidial axial rings.

Genus gravicalymene Shirley, 1936

Type species. Gravicalymene convolva Shirley, 1936, by original designation; from the Birdshill Limestone
(Pusgillian) of Llandeilo, south central Wales.

Gravicalymene sp.

Plate 4, figs 13, 15

1980 ? Flexicalymene sp.; Romano, pp. 65, 67.

Material. NMI F21014; BM It.25734-It.25735.

Horizon and locality. BM It.25734 from Loc. 21 OB; others from Collon Quarry.

Discussion. The bell-shaped glabella, unconstricted axial furrows (Ingham 1977, p. 94) and
subquadrate form of the glabella in front of L2 (Siveter 1977, p. 383) indicate that the Irish form
should be assigned to Gravicalymene. The material differs from the type species in having a relatively
wider base to the glabella, straighter anterior margin to the frontal lobe and narrower (sag.)
‘anterior border’. It differs from G. jugifera Dean, 1962 (p. 116, pi. 14, figs 3^1, 8-9) from the
Pusgillian of Cross Fell, northern England, in its straighter anterior margin to the frontal lobe and
narrower preglabellar furrow. G. deani Ingham, 1977 (p. 96, pi. 20, figs 16-17; pi. 21, figs 1-6) from
the Cautleyan of Cautley, northern England, is similar to Gravicalymene sp. in having strongly
sigmoidal axial furrows and a very broadly rounded or nearly straight anterior margin to the frontal
lobe. The Irish form differs from G. deani in its much less thickened (sag.) ‘anterior border’. Of the
two Caradoc species of Gravicalymene from south Shropshire, Gravicalymene sp. contrasts with the
holotype of G. praecox (Bancroft, 1949) (see Dean 1963, p. 225, pi. 39, figs 1, 3, 9, 12-14), which
is an immature cranidium, in its longer frontal glabellar lobe. The mature specimens of G. praecox
are poorly preserved and difficult to compare with the Irish material. The other Shropshire species
G. inflata Dean, 1963 (p. 227, pi. 39, fig. 6) is only known from a single cranidium but the
considerable width (trs.) of the anterior part of the fixed cheek serves to distinguish it from
Gravicalymene sp. Gravicalymene capitovata Siveter, 1977 (p. 377, figs 11a-h, figs 12a, e-i, k-l)
from the uppermost Llanvirn Engervik Member of the Elnes Formation (‘ Ogygiocaris Shale
(4aa3)’; see Owen et al. 1990), of the Oslo Region has a relatively much larger LI and longer
‘anterior border’ than Gravicalymene sp.

Ross (1967, 1979) figured a number of species of Gravicalymene from the Caradoc of Kentucky
and Ohio, USA. The Irish form differs from these American species in its longer frontal glabellar
lobe and narrower (sag. and exs.) ‘anterior border’.

Family pterygometopidae Reed, 1905
Subfamily pterygometopinae Reed, 1905
Genus achatella Delo, 1935

Type species. Dalmanites achates Billings, 1860, by original designation; from the Cobourg beds (upper
Caradoc) of Ottawa, Ontario, Canada.

ROMANO AND OWEN: IRISH CARADOC TRILOBITES

711

Achatella truncatocaudata! (Portlock, 1843)

Text-fig. 5a, f

1980 1 Achatella cf. truncatocaudata (Portlock); Romano, p. 67.

Material. GSI F00880 (figured by Harper 1952, pi. 5, fig. 5); BM It. 25741.

Horizon and locality. GSI F00880 from Loc. 1 ; BM It. 25741 from Collon Quarry.

Discussion. The Grangegeeth specimens differ from A. truncatocaudata from the Killey Bridge
Formation (Cautleyan) of Pomeroy, Northern Ireland, in having more anteriorly placed eyes and a
less parallel-sided L2. A. cf. truncatocaudata (Owen 1986) from the High Mains Formation
(Hirnantian) at Girvan, southwest Scotland, also has more posteriorly placed eyes than the
Grangegeeth form. Two other Girvan species, A. retardata (Reed, 1914; Morris and Tripp 1986,
p. 172, pi. 4, fig. 2) and A. consobrina Tripp, 1954, (p. 683, pi. 4, figs 26-33) from the Rawtheyan, and
early Caradoc respectively, may be distinguished from A. truncatocaudata and the Grangegeeth
form by their longer (exs.) eyes, and more parallel-sided L2 in consobrina. The Grangegeeth form
differs from the type species (see Ludvigsen and Chatterton 1982) from the Shermanian and
Edenian stages (late Caradoc - early Ashgill) of Canada and the USA, in its relatively longer (sag.)
frontal glabellar lobe, more anteriorly placed eyes and larger number of lenses (twelve rather than
seven) in the vertical rows.

Family lichidae Hawle and Corda, 1847
Subfamily homolichinae Phleger, 1936
Genus autoloxolichas Phleger, 1936

Type species. Lichas st. mathiae Schmidt, 1885, by original designation; from the Caradoc of Estonia.

Autoloxolichas cf. la.xatus (M’Coy, 1846)

Text-fig. 5c

1967 Platylichas la.xatus (M’Coy); Brenchley et al ., p. 298.

1980 Platylichas cf. la.xatus (McCoy); Romano, p. 69.

Material. ?NMI F21015, NMI F21016 and fragmental specimens.

Horizon and locality. NMI F21016 from Loc. 209; other specimens from the Tretaspis Bed; Brickwork’s
Quarry.

Discussion. Thomas and Holloway (1988) assigned ‘ l ax at us' to Autoloxolichas. The Grangegeeth
material shows diverging bullar lobes and quite strongly expanding median glabellar lobe,
approaching the condition seen in Platylichas (see Thomas and Holloway 1988, p. 206). However
the apparent suppression of Lib, relatively close proximity of bullar lobe and LI a, and shallow
diffuse furrow joining these two lobes are characters regarded by Thomas and Holloway as typical
of Autoloxolichas.

A. laxatus is a common species in the Caradoc of Britain, Ireland and Scandinavia. It commonly
shows considerable variation (Dean 1963; Tripp 1958; Owen and Bruton 1980; Owen et al. 1986);
in particular, as pointed out by Owen et al. (p. 115), in the relative size and shape of the bullar
lobes, median glabellar lobe and anterior border. The present material differs from most of the
previously figured specimens of this species in possessing a less strongly rounded anterior margin
to the frontal glabellar lobe (cf. Tripp 1958, pi. 85, fig. 4; Dean 1963, pi. 43, figs 2, 10; Owen and

712

PALAEONTOLOGY, VOLUME 36

text-fig. 5. A, F, Achatella truncatocaudata ? (Portlock, 1843). GSI F00880 (figured by Harper 1952, pi. 5, fig.
5 as GSI/JCH 529); basal Knockerk House Sandstones Member, Loc. 1; internal mould of cephalon, dorsal
and lateral views respectively, both x 3. b, Autoloxolichas sp. indet. NMI F21015/B; Tretaspis Bed,
Brickwork’s Quarry Shales, Brickwork’s Quarry; latex of external mould of incomplete cranidium, x 3. c,
Autoloxolichas cf. laxatus (M'Coy, 1846). NMI F21016; Knockerk House Shales, Loc. 209; internal mould of
incomplete cephalon, dorsal view, x4. d-e Amphilichas sp.; d, NMI F21017; internal mould of incomplete
cephalon, dorsal view, x 2. e, NMI F21018/B; cast of external mould of pygidium, dorsal view, x 1.5. Both
from basal Knockerk House Sandstones Member, D from Collon Quarry, E from Loc. 3. g-l, Miraspis aff.
solitaria Reed, 1935; G, NMI F21019; internal mould of cranidium, dorsal view, x 4; H, NMI F21020; cast of
external mould of incomplete cranidium, dorsal view, x 7-5; I, NMI FI 4027 (figured by Harper 1952, pi. 5, fig.
4 as NMI 1951/13); cast of external mould of cranidium, dorsal view, x 7-5; J, NMI F21022; internal mould
of free cheek, dorsal view, x 8; K, NMI F21023; internal mould of incomplete cranidium, dorsal view, x 6;
L, NMI F21024; internal mould of part of thorax, dorsal view, x 3. G-H, j-k from concretions approximately
1 m above Tretaspis Bed, Brickwork’s Quarry Shales, Brickwork’s Quarry; i from ?Knockerk House Shales
Member to basal Brickwork’s Quarry Shales Member, Loc. 26. L from Tretaspis Bed.

ROMANO AND OWEN: IRISH CARADOC TRILOBITES

713

Bruton 1980, pi. 10, figs 9-10, 13) and more rounded posterior margin to the bullar lobe (cf. Dean
1963, pi. 43, fig. 11; Owen and Bruton 1980, pi. 10, fig. 6; Owen et al. 1986, p. 113, fig. 77). However,
Owen et al. (1986, figs 71-72) figured a cranidium assigned to Platylichas laxatus from the upper
Caradoc Raheen Formation in Co. Waterford which is very similar to the present species.

A fragmentary, large cranidium (NMI F21015; Text-fig. 5b) has an estimated width across the
front of the bullar lobes of 15 mm. The main features of this cranidium are the broadly rounded
frontal glabellar lobe, the wide (sag. and trs.) border, deep and narrow anterior border furrow with
small elongate lobe lying on inner part of border anterior to bullar lobe, wide and shallow lateral
furrow and fine tuberculate sculpture. It is assigned to Autoloxolichas on account of its general
characteristics and the presence of the small lobe lying anterior to the bullar lobe, a feature present
in A. nodulosus (see Thomas and Holloway 1988, pi. 9, figs 188-189). It differs from Autoloxolichas
cf. laxatus from the Knockerk House Shales Member in having a generally finer sculpture and it is
considerably larger. It is provisionally referred to Autoloxolichas sp. indet.

Subfamily tetralichinae Phleger, 1936
Genus amphilichas Raymond, 1905

Type species. Platymetopus lineatus Angelin, 1854, by monotypy; from the Boda Limestone (Ashgill) in
Dalarna, Sweden.

Amphilichas sp.

Text-fig. 5d-e

1952 Lichas ( Acrolichas ) hibernicus (Portlock); Harper, p. 88
1980 Amphilichas hibernicus (Portlock); Romano, p. 66.

1980 Amphilichas cf. ardmillanensis (Reed); Romano, p. 67.

Material. NMI F21017-F21018.

Horizon and locality. NMI F21017 from Collon Quarry; NMI F21018 from Loc. 3.

Discussion. At present, it is not known whether the material from Collon (cranidium) is conspecific
with that from Grangegeeth (segment and pygidium), even though they occur at a similar horizon.
As far as the authors are aware, no pygidium of A. ardmillanensis has been figured. The cranidium
differs from A. ardmillanensis (Reed, 1914; Tripp 1958, 1980) from the Upper Balclatchie and Lower
Ardwell groups (lower Caradoc), Girvan, southwest Scotland in the following respects: the
anterior margin of the central lobe is a smooth curve with no subconical projection in the middle;
the median lobe expands (trs.) posteriorly; the furrows posterior to the composite lateral lobes are
fainter; the anterior border is steeply declined. However, the Irish specimen is deformed, which may
account for some of these differences. Faint posterolateral glabellar furrows are also seen in other
Balclatchie Group species, A. panoplos and A. planus (Tripp 1980, pi. 4, fig. 17; Tripp 1954, pi. 1,
fig. 5, respectively), but the median glabellar lobe is wider in the two Scottish species. Discontinuous
lateral furrows are seen in A. sp. B (Tripp 1976, pi. 7, fig. 23) from the Superstes Mudstones at
Girvan but, this species has a more strongly rounded glabellar outline.

The Irish pygidium differs only slightly from that of A. hibernicus from the Lower Ardwell Group
at Girvan figured by Reed (1914) and Thomas and Holloway (1988, pi. 12, fig. 256). The axis of the
Irish specimen narrows less rapidly posteriorly and consists of a flatter portion reaching back to the
anterior part of the third segment where it then slopes down sharply to the near horizontal terminal
piece (compare Thomas and Holloway 1988, pi. 12, fig. 261). The oblique and incomplete second
axial ring furrows are larger in the Scottish specimen and the sculpture of the Irish pygidium is
slightly coarser. The doublure is broad in the Irish specimen and, as in A. hibernicus, reaches the
distal ends of the pleural furrows.

714

PALAEONTOLOGY, VOLUME 36

lichid indet.

1980 Lichid indet.; Romano, p. 67.

Material. BM It.25736n-6.

Horizon and locality. Collon Quarry.

Discussion. The material is fragmentary, but a comparison may be made with Metopolichasl (Dean
1963, pi. 43, fig. 4) from the Costonian of Shropshire, England. It differs from Dean’s species in
having a coarser sculpture on the frontal lobe. Thomas and Holloway (1988) pointed out the
uncertainty of the status of Metopolichas, but placed it in the Homolichinae on account of the form
of the hypostoma. On cranidial characteristics these authors noted that Metopolichas is very similar
to Lichas, thus casting further doubt on the generic assignment of the Knockerk Sandstone
specimen. A cranidium of Amphilichas sp. occurs at a similar horizon at Grangegeeth (see above),
but this species has a considerably more rounded anterior margin.

Family odontopleuridae Burmeister, 1843
Subfamily selenopeltinae Hawle and Corda, 1847
(= miraspidinae Richter and Richter, 1917; dicranuridae Prantl and Pribyl, 1949: see

Ramskold 1991)

Genus miraspis Richter and Richter, 1917

Type species. Odontopleura mira Barrande, 1846, by original designation; from the Liten Formation (Wenlock),
near Beroun, Bohemia.

Miraspis aff. solitaria Reed, 1935
Text-fig. 5g-l

1952 Acidaspis ( Onchaspis ) cf. lalage Wyville Thomson; Harper, pp. 89, 105, pi. 5, fig. 4.

1967 Diacanthaspis sp. cf. D. lalage (Wyville Thomson); Brenchley et ah, p. 298.

1980 Miraspis sp.; Romano, pp. 69-70.

Material. NMI F14027 (figured Harper 1952, pi. 5, fig. 4 as NMI 1951/13); NMI F21019, F21020, F21022-24;
BM It. 25737-It. 25740.

Horizon and locality. NMI F14027 from Loc. 26; others from Tretaspis Bed and concretions 1 m higher,
Brickwork’s Quarry.

Discussion. Miraspis solitaria (Reed 1935, p. 37, pi. 3, fig. 21) is from the basal Superstes Mudstones
(Llandeilo) at Aldons Quarry, Girvan, and was redescribed by Tripp (1976, p. 412, pi. 7, figs 27-32).
The Grangegeeth form shows minor differences on the free cheek, where its border and furrow are
more distinct, the granulation on the cranidium is finer, and the marginal spines are not directed as
obliquely. Until a pygidium is known from Grangegeeth some doubt must remain as to whether the
material does belong in M. solitaria.

Owen and Romano (in Harper et al. 1985, p. 306) distinguished the Grangegeeth Miraspis from
an unnamed Caradoc species from the Clashford House Formation near Herbertstown, Co. Meath,
eastern Ireland, on the basis of its more slender occipital spines, more circular LI and weaker
furrows at the side of the median glabellar lobe. The last two of these also distinguish the
Grangegeeth form from Miraspis sp. of Owen et al. (1986; see also Ramskold 1991, p. 171) from
the upper Caradoc Raheen Formation in Co. Wexford, which also lacks a median occipital
protuberance but, like the other Irish species, has a very weakly incised occipital furrow.

Acknowledgements. We thank Mr N. T. Monaghan (National Museum of Ireland), Mr P. W. Phillips
(Liverpool Museum) and Dr A. G. Sleeman (Geological Survey of Ireland) for loaning specimens in their care;

ROMANO AND OWEN: IRISH CARADOC TRILOBITES

715

Mr M. Cooper (University of Sheffield) for draughting the text-figures; Miss P. M. Mellor (University of
Sheffield) for producing Table 1 . A. W. O. thanks the Department of Geology and Applied Geology, University
of Glasgow, for funds to visit Sheffield.

REFERENCES

ahlberg, p. 1988. Agnostid trilobites from the Ordovician of Jamtland, Sweden. Geologiska Foreningens i
Stockholm Forhandlingar , 110, 267-278.

1989. Agnostid trilobites from the Upper Ordovician of Sweden and Bornholm, Denmark. Bulletin of the

Geological Society of Denmark , 37, 213-226.

angelin, N. p. 1851. Palaeontologia Suecica I. Iconographia Crustaceorum formationis transitionis, 1. Samson
and Wallin, Lund. 24 pp.

1 854. Palaeontologia Scandinavia I. Crustacea formationis transitionis, 2. Samson and Wallin, Lund, 2 1-92.

Bancroft, B. b. 1949. Upper Ordovician trilobites of zonal value in southeast Shropshire. Proceedings of the
Royal Society of London, Series B, 136, 291-315, pis 9-11.

barrande, j. 1846. Notice preliminaire sur le systeme Silurien et les trilobites de Boheme. C. L. Hirschfeld,
Leipsic, 97 pp.

1852. Systeme Silurien du centre de la Boheme. lere partie. Recherches paleontologiques. Prague and Paris,

xxx + 935 pp., 51 pis.

bassler, R. s. 1915. Bibliographic index of American Ordovician and Silurian fossils. Bulletin of the U.S.
National Museum , Washington, 92, 1-1521, pis 1-4.

begg, J. L. 1946. Some new fossils from the Girvan district. Transactions of the Geological Society of Glasgow,
21, 29-47.

beyrich, E. 1845. Ueber einige bohmische Trilobiten. G. Relmer, Berlin, 47 pp.

billings, E. 1860. Description of some new species of fossils from the Lower and Middle Silurian rocks of
Canada. Canadian Naturalist and Geologist, 5, 49-69.

boyce, w. D. 1987. Cambrian-Ordovician trilobite biostratigraphy in central Newfoundland. Current Research
(1987). Newfoundland Department of Mines and Energy, Mineral Development Division, Report 87-1, 335-341 .

brenchley, p. j., harper, j. c., mitchell, w. i. and romano, m. 1977. A reappraisal of some Ordovician
successions in eastern Ireland. Proceedings of the Royal Irish Academy, Series B, 77, 65-85.

—romano, m. and skevington, d. 1967. New Ordovician faunas from Grangegeeth, Co. Meath.

Proceedings of the Royal Irish Academy , Series B , 65, 297-303.

burmeister, h. 1843. Die Organisation der Trilobiten aus ihren lebenden Verwandten entwickelt ; nebst einer
systematischen Uebersicht aller zeither beschriebenen Arten. G. Relmer, Berlin, xii + 147 pp., 6 pis.

Carlisle, H. 1979. Ordovician stratigraphy of the Tramore area. County Waterford, with a revised Ordovician
correlation for south-east Ireland. 545-554 In Harris, a. l., Holland, c. h. and leake, b. e. (eds). The
Caledonides of the British Isles - reviewed. Special Publication of the Geological Society, London, 8,
xii + 768 pp.

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

— 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.

chugaeva, M. N. 1958. [Ordovician trilobites of the Chu-Ili Mountains] Trudy Geologicheskogo Instituta,
Leningrad, 9, 5-138, 11 pis. [In Russian],

cocks, L. R. M. and fortey, r. a. 1990. Biogeography of Ordovician and Silurian faunas. 97-104. In
mckerrow, w. s. and scotese, c. R. (eds). Palaeozoic palaeogeography and biogeography. Geological Society
Memoir, 12, 435 pp.

cooper, b. n. 1953. Trilobites from the Lower Champlainian Formations of the Appalachian Valley. Memoirs
of the Geological Society of America, 55, v + 69, 19 pis.

dalman, J. w. 1827. Om Palaeaderna eller de sa kallade Trilobiterna. Kungliga Svenska Vetenskapsacademiens
Handlingar, 1826 (2), 113-152, 226-294, pis 1-6.

dean, w. T. 1961. The Ordovician trilobite faunas of south Shropshire II. Bulletin of the British Museum
{Natural History ), Geology Series, 5, 313-358, 7 pis.

— 1962. The trilobites of the Caradoc Series in the Cross Fell Inker of northern England. Bulletin of the
British Museum {Natural History), Geology Series, 7, 67-134, 13 pis.

716

PALAEONTOLOGY, VOLUME 36

dean, w. t. 1963. The Ordovician trilobite faunas of south Shropshire. III. Bulletin of the British Museum
( Natural History ), Geology Series , 7, 213-254, pis 37-46.

1971m The trilobites of the Chair of Kildare Limestone (upper Ordovician) of eastern Ireland.
Monograph of the Palaeontographical Society , 125 (531), 1-60, pis 1-25.

19716. Ordovician trilobites from the Central Volcanic Mobile Belt at New World Island, northeastern
Newfoundland. Geological Survey of Canada , Bulletin, 210, 1-37, 7 pis.

1974. The trilobites of the Chair of Kildare Limestone (upper Ordovician) of eastern Ireland. Part 2.
Monograph of the Palaeontographical Society, 128 (539), 61-98, pis 1-19.
delo, D. m. 1935. A revision of the Phacopid trilobites. Journal of Paleontology, 9, 402-420, 1 pi.
fortey, r. a. 1980. The Ordovician trilobites of Spitzbergen. III. Remaining trilobites of the Valhallfonna
Formation. Norsk Polarinstitutt Skrifter, 171, 163, 25 pis.

— 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 cocks, l. r. m. 1991. The early Palaeozoic of the North Atlantic region as a test case for the use of
fossils in continental reconstructions. Tectonophysics, 206, 147-158.

— and Owens, r. m. 1987. The Arenig Series in South Wales. Bulletin of the British Museum (Natural
History), Geology Series, 41, 69-305.

harper, d. a. T., Mitchell, w. I., owen, a. w. and romano, m. 1985. Upper Ordovician brachiopods and
trilobites from the Clashford House Formation, near Herbertstown, Co. Meath, Ireland. Bulletin of the
British Museum ( Natural History), Geology Series , 38, 287-308.

and parkes, m. 1989. Palaeontological constraints on the definition and development of Irish Caledonide
terranes. Journal of the Geological Society of London, 146, 4 1 3 — 4 1 5 .
harper, j. c. 1952. The Ordovician rocks between Collon (Co. Louth) and Grangegeeth (Co. Meath). Scientific
Proceedings of the Royal Dublin Society, New Series, 26, 85-112, 3 pis.

and romano, m. 1967. Decordinaspis - a new Caradoc trinucleid trilobite from the Ordovician of Ireland.
Proceedings of the Royal Irish Academy, Series B, 65, 305-308, pis 8-9.
henningsmoen, G., jaanusson, v. and Stubblefield, c. J. 1980. Ogygiocaris (Trilobita): proposed conservation
under the Plenary Powers. Z.N. (S.) 439. Bulletin of Zoological Nomenclature, 36, 226-230.
hisinger, w. 1840. Lethaea Svecica seu petrificata Sveciae, supplementa secundi. Norstedt, Holmiae, 11 pp.,
3 pis.

holliday, s. 1942. Ordovician trilobites from Nevada. Journal of Paleontology, 16, 471-478.
hughes, c. p. 1979. The Ordovician trilobite faunas of the Builth-Llandrindod Inlier, Central Wales. Bulletin
of the British Museum (Natural History), Geology Series, 32, 109-181.

— 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, 537-604.

hull, E., cruise, r. j. and bailey, w. h. 1871. Explanatory Memoir to accompany Sheets 91 and 92 of the maps
of the Geological Survey of Ireland, illustrating parts of the Counties of Meath, Louth and Dublin. Memoir
of the Geological Survey of Ireland. 46 pp.

hunt, a. s. 1967. Growth, variation and instar development of an agnostid trilobite. Journal of Paleontology,
41, 203-208, 1 pi.

ingham, J. K. 1974. A monograph of the upper Ordovician trilobites from the Cautley and Dent districts of
Westmorland and Yorkshire. Monograph of the Palaeontographical Society , 128 (538), 59-87, pis 1-9.

— 1977. A monograph of the upper Ordovician trilobites from the Cautley and Dent districts of
Westmorland and Yorkshire. Monograph of the Palaeontographical Society, 130 (546), 89-121, pis 1-9.

— and tripp, r. p. 1991. The trilobite fauna of the Middle Ordovician Doularg Formation of the Girvan
District, Scotland and its palaeoenvironmental significance. Transactions of the Royal Society of Edinburgh,
Earth Sciences, 82, 27-54.

jaekel, o. 1909. Uber die Agnostiden. Zeitschrift der Deutschen Geologischen Gesellschaft, 61, 380-400.
kielan, z. 1960. Upper Ordovician trilobites from Poland and some related forms from Bohemia and
Scandinavia. Palaeontologica Polonica, 11, 1-198, 36 pis.
kielan-jaworowska, z., Bergstrom, J. and ahlberg, p. 1991. Cheirurina (Trilobita) from the Upper
Ordovician of Vastergotland and other regions of Sweden. Geologiska Foreningens i Stockholm
Forhandlingar, 113, 219-244.

lane, p. d. 1971. British Cheiruridae (Trilobita). Monograph of the Palaeontographical Society, 125 (530), 1-95,
pis 1-16.

— and Owens, R. m. 1982. Silurian trilobites from Kap Schuchert, Washington Land, western North
Greenland. Rapports Gronlands Geologiske Undersogelse, 108, 41-69.

ROMANO AND OWEN: IRISH CARADOC TRILOBITES

717

Levitskiy, E. s. 1962. O novom rode trilobitov- Cybelurus gen. nov. Izvestiya Vysshikh Uchebnykh Zavedenii.
Geologia i Razvedka, 7, 129-131, 1 pi.

ludvigsen, r. 1979. A trilobite zonation of the Middle Ordovician rocks, southwestern District of Mackenzie.
Geological Survey of Canada , Bulletin , 312, 1-99, 21 pis.

— and chatterton, B. D. e. 1982. Ordovician Pterygometopidae (Trilobita) of North America. Canadian
Journal of Earth Sciences , 19, 2179-2206.

marek, L. 1952. [Contribution to the stratigraphy and fauna of the uppermost part of the Kraluv Dvur Shales
(Ashgillian)]. Sbornik Ustfedniho Ustavu Geologickeho, 19, 429 — 455. [In Czech],
m'coy, f. 1846. A synopsis of the Silurian fossils of Ireland. Dublin University Press, Dublin, 72 pp.
McKERROW, w. s. and soper, n. j. 1989. The Iapetus Suture in the British Isles. Geological Magazine , 126, 1-8.
miller, s. a. 1889. North American geology and palaeontology. Western Methodist Book Concern, Cincinnati,
Ohio, 664 pp.

milne edwards, H. 1840. Histoire naturelle des Crust aces, comprenent F Anatomie, la physio/ogie et la
classification de ces animaux. 3, Roret, Paris, 638 pp.

mitchell, w. i., Carlisle, h., hiller, n. and addison, r. 1972. A correlation of the Ordovician rocks of
Courtown (Co. Wexford) and Tramore (Co. Waterford). Proceedings of the Royal Irish Academy, 72, 83-89.
moore, r. c. 1959. Treatise on invertebrate paleontology. Part O. Arthropoda I. Geological Society of America
and University of Kansas Press, Boulder, Colorado and Lawrence, Kansas, xix + 560 pp.
morris, s. F. 1988. A review of British trilobites, including a synoptic revision of Salter’s monograph.
Monograph of the Palaeontographical Society, 140 (574), 1-316.

— and tripp, r. p. 1986. Lectotype selections for Ordovician trilobites from the Girvan District, Strathclyde.
Bulletin of the British Museum ( Natural History), Geology Series, 40, 161-176.

nicholson, H. A. and etheridge, R. a. 1879. A monograph of the Silurian fossils of the Girvan district in Ayrshire ,
1, 1-135, pis 1-9.

nieszkowski, I. 1857. Versuch einer Monographic der in den silurischen Schichten der Ostseeprovinzen
vorkommenden Trilobiten. Archiv fur die Naturkunde Liv-, Ehst- und Kurlands , 1, 517-626, pis 1-3.
nikolaisen, F. 1961. The Middle Ordovician of the Oslo Region, Norway, 7. Trilobites of the suborder
Cheirurina. Norsk Geologisk Tidsskrift, 41, 279-309.

— 1963. The Middle Ordovician of the Oslo Region, Norway, 14. The trilobite family Telephinidae. Norsk
Geologisk Tidsskrift, 43, 345-398.

— 1983. The Middle Ordovician of the Oslo Region, Norway, 32. Trilobites of the family Remopleurididae.
Norsk Geologisk Tidsskrift, 62, 232-329.

opik, A. 1937. Trilobiten aus Estland. Acta et Commentationes Universitatis Tartuensis, Series A, 32, 1-163,
26 pis.

Owen, A. w. 1980. The trilobite Tretaspis from the Upper Ordovician of the Oslo region, Norway.
Palaeontology, 23, 715-747.

1981. The Ashgill trilobites of the Oslo Region. Palaeontographica, Abteilung A, 175, 1-88, 17 pis.

— 1986. The uppermost Ordovician (Hirnantian) trilobites of Girvan, SW Scotland with a review of coeval
trilobite faunas. Transactions of the Royal Society of Edinburgh, Earth Sciences, 77, 231-239.

— 1987. The trilobite Tretaspis at the Middle-Upper Ordovician boundary in Vastergotland. Geologiska
Foreningens i Stockholm Forhandlingar, 109, 259-266.

and bruton, d. l. 1980. Late Caradoc-early Ashgill trilobites from the central Oslo Region, Norway.
Palaeontological Contributions from the University of Oslo, 245, 1-62, 10 pis.

— bockelie, j. f. and bockelie, t. G. 1990. The Ordovician successions of the Oslo Region, Norway.
Norges Geologiske Undersokelse Special Publication, 4, 1-54.

— harper, d. a. t. and romano, m. 1992. The Ordovician biogeography of the Grangegeeth terrane and the
Iapetus suture zone in eastern Ireland. Journal of the Geological Society, London, 149, 3-6.

— tripp, r. p. and morris, s. f. 1986. The trilobite fauna of the Raheen Formation (upper Caradoc), Co.
Waterford, Ireland. Bulletin of the British Museum ( Natural History), Geology Series, 40, 91-122.

Owens, R. m. 1973«. British Ordovician and Silurian Proetidae (Trilobita). Monograph of the Palaeontographical
Society, 127 (535), 1-98, pis 1-15.

1973/). Ordovician Proetidae (Trilobita) from Scandinavia. Norsk Geologisk Tidsskrift, 53, 117-181.
Paris, F. and robardet, m. 1990. Early Palaeozoic palaeobiogeography of the Variscan regions. Tectonophysics,
177, 193-213.

parkes, M. a. 1990. The palaeontology of the Duncannon Group (middle-upper Ordovician) of southeast Ireland.
Unpublished Ph.D. thesis, National University of Ireland.

718

PALAEONTOLOGY, VOLUME 36

pek, i. 1977. Agnostid trilobites of the Central Bohemian Ordovician. Sbornik Geologickych Ved , Praha, 19,
7-44, 12 pis."

phleger, f. 1936. Lichadian trilobites. Journal of Paleontology, 10, 593-615.

portlock, j. E. 1843. Report on the geology of the County of Londonderry, and parts of Tyrone and Fermanagh.

Millikan and Longmans, Dublin and London, xxxi + 784 pp.
prantl, f. and pribyl, a. 1949. A study of the superfamily Odontopleuracea nov. superfam. (Trilobites).

Rozpravy Ustredniho Ustavu Geologickeho, 12, 1-221, pis 1-11.
pribyl, a. 1946. O nekolika novych trilobitovych rodech z ceskeho siluru a devonu. Priroda , Brno, 38, 89-95.

— and vanek, J. 1965. Neue Trilobiten des bohmischen Ordoviziums. Vestmk Ustredniho Ustavu
Geologickeho, 40, 277-282.

price, d. 1984. The Pusgillian Stage in Wales. Geological Magazine, 121, 99-105.

ramskold, l. 1991. Pattern and process in the evolution of the Odontopleuridae (Trilobita). The Selenopeltinae
and Ceratocephalinae. Transactions of the Royal Society of Edinburgh, Earth Sciences, 82, 143-181.
Raymond, p. e. 1905. Notes on the names Amphion, Harpina and Platymetopus. American Journal of Science,
(4) 19, 377-378.

— 1913. Revision of the species which have been referred to the genus Bathyurus. Bulletin of the Victoria
Memorial Museum, 1, 51-69.

reed, f. r. c. 1903. The Lower Palaeozoic trilobites of the Girvan district, Ayrshire. Monograph of the
Palaeontographical Society, 57 (270), 1^18, pis 1-6.

— 1905. The classification of the Phacopidae. Geological Magazine, 2, 172-178, 224-228.

— 1906. The Lower Palaeozoic trilobites of the Girvan district, Ayrshire. Monograph of the Palaeonto-
graphical Society, 60 (286), 97-186, pis 1-7.

1914. The Lower Palaeozoic trilobites of Girvan. Supplement. Monograph of the Palaeontographical
Society, 67 (329), 1-56, pis 1-8.

— 1915. Supplementary memoir on new Ordovician and Silurian fossils from the Northern Shan States.
Paleontologia Indica, 6, 1-98, 12 pis.

1928. Notes on the Family Encrinuridae. Geological Magazine, 65, 51-77.

— 1931. A review of the British species of the Asaphidae. Annals and Magazine of Natural History, 10,
441-72.

— 1935. The Lower Palaeozoic trilobites of Girvan. Supplement No. 3. Monograph of the Palaeontographical
Society, 88 (400), 1-64, pis 1-4.

REPINA, V. V., PETRUNINA, Z. E. and HAJRULLINA, T. I. 1975. [Trilobita], 100-351. In REPINA, L. N., YASKOVITCH,
B. V., AKSARINA, N. A., PETRUNINA, Z. E., PONIKLENKO, I. A., RUBANOV, D. A., BOLGOVA, G. V., GOLIKOV, A. N.,
hajrullina, t. i. and posokhova, M. M. (eds). [ Stratigraphy and fauna of the Lower Palaeozoic of the northern
submontane belt and Alai Ridges, southern Tyanshan.] Trudy Instituta Geologii i Geofiziki Akadmiya Nauk
SSR Sibirskoe Otdelenie [USSR], 278, 1-352. [In Russian].
richter, r. and richter, e. 1917. Uber die Einteilung der Familie Acidaspidae und fiber einige ihrer
devonischen Vertreter. Zentralblatt fur Mineralogie, Geologie und Palaontologie, 1917, 462^472.
romano, M. 1970. The stratigraphy and palaeontology of the Ordovician rocks of eastern central Ireland.
Unpublished Ph.D. thesis, University of Liverpool.

— 1980. The stratigraphy of the Ordovician rocks between Slane (County Meath) and Collon (County
Louth), eastern Ireland. Journal of Earth Science, Royal Dublin Society, 3, 53-79.

ross, R. J. 1967. Calymenid and other Ordovician trilobites from Kentucky and Ohio. United States Geological
Survey, Professional Paper, 583-B, 1-19.

— 1979. Additional trilobites from the Ordovician of Kentucky. United States Geological Survey,
Professional Paper , 1066-D, 1-13.

ruedemann, r. 1901. Trenton conglomerate of Rysedorph Hill Rensselaer Co. N.Y. and its fauna. Bulletin of
the New York State Museum, 49, 3-114.

safford, J. M. and vogdes, a. w. 1889. Descriptions of new species of Crustacea from the Lower Silurian of
Tennessee. Proceedings of the Philadelphia Academy of Natural Sciences, 39, 166-168.
salter, j. w. 1864-6. A monograph of the British trilobites from the Cambrian, Silurian and Devonian
formations. Monograph of the Palaeontographical Society, 16 (67), 1-80, pis 1-6; 17 (72), 81-129, pis 7-14;
18 (77), 129-176, plsl 5-25.

Schmidt, f. 1881. Revision der ostbaltischen Silurischen Trilobiten nebst geognostischer Ubersicht der
ostbaltischen Silurgebiets, Abtheilung I, Phacopiden, Cheiruriden und Encrinuriden. Memoires de
r Academie Imperiale des Sciences de St Petersbourg, (7) 30, (1), 1-237, pis 1-16.

ROMANO AND OWEN: IRISH CARADOC TRILOBITES

719

shaw, F. c. 1968. Early Middle Ordovician Chazy trilobites of New York. Memoir of the New York Museum
of Natural History, 17, 1-163.

sheng, s. f. 1934. Lower Ordovician trilobite faunas of Chekiang. Palaeontologica Sinica, Series B , 3, 1-19.
shirley, J. 1936. Some British trilobites of the family Calymenidae. Quarterly Journal of the Geological Society ,
London, 92, 384-421.

siveter, d. j. 1977. The Middle Ordovician of the Oslo Region, Norway, 27. Trilobites of the family
Calymenidae. Norsk Geologisk Tidsskrift, 56, 335-396.
sloan, r. e. 1991. A chronology of North American trilobite genera. In barnes, c. r. and williams, s. h. (eds).

Advances in Ordovician Geology. Geological Survey of Canada, Paper, 90-9, 165-177.

Slocum, a. w. 1913. New trilobites from the Maquoketa Beds of Fayette County, Iowa. Publications of the
Field Museum, Geology Series, 4, 43-83.

stauble, a. 1953. Two new species of the family Cryptolithidae. Le Naturaliste Canadien, 80, 83-119, 201-220.
thomas, A. T. 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.
tripp, r. a. 1954. Caradocian trilobites from mudstones at Craighead Quarry, near Girvan, Ayrshire.
Transactions of the Royal Society of Edinburgh, 62, 655-693.

— 1958. Stratigraphical and geographical distribution of the named species of the trilobite superfamily
Lichacea. Journal of Paleontology, 32, 574-582.

— 1962. Trilobites from the confinis Flags (Ordovician) of the Girvan District, Ayrshire. Transactions of the
Royal Society of Edinburgh, 62, 1 40.

— 1965. Trilobites from the Albany division (Ordovician) of the Girvan District, Ayrshire. Palaeontology,
8, 577-603.

'1967. Trilobites from the upper Stinchar Limestone (Ordovician) of the Girvan District, Ayrshire.

Transactions of the Royal Society of Edinburgh , 67, 43-93.

— 1976. Trilobites from the basal superstes Mudstones (Ordovician) at Aldons Quarry, near Girvan,
Ayrshire. Transactions of the Royal Society of Edinburgh, 69, 369-423.

— 1979. Trilobites from the Auchensoul and Stinchar Limestones of the Girvan District, Strathclyde.
Palaeontology, 22, 339-361.

— 1980. Trilobites from the Ordovician Balclatchie and lower Ardwell groups of the Girvan District,
Scotland. Transactions of the Royal Society of Edinburgh, Earth Sciences, 71, 123-145.

— zhou zhiyi and pan zhenqin. 1989. Trilobites from the Upper Ordovician Tangtou Formation, Jiangsu
Province, China. Transactions of the Royal Society of Edinburgh, Earth Sciences, 80, 25-68.

tromelin, G. de and lebesconte, p. 1876. Essai d’un catalogue raisonne des fossiles siluriens des departments
de Maine-et-Loire, de la Loire-Inferieur et du Morbihan, avec des observations sur les terrains paleozoiques
de l'Ouest de la France. Association Frangaise pour FAvancement de Science, 4e Session, Nantes (1875),
601-661.

ulrich, e. o. 1930. Ordovician trilobites of the family Telephidae and concerned stratigraphic correlations.

Proceedings of the United States National Museum, 76, 1—101.
vogdes, a. w. 1890. A bibliography of Palaeozoic Crustacea from 1698 to 1889, including a list of North
American species and a systematic arrangement of genera. Bulletin of the United States Geological Survey,
63, 1-177.

wahlenberg, g. 1818. Petrificata Telluris Svecanae. Nova Acta Regiae Societatis Scientiatum Upsaliensis, 8,
1-116, pis 1-4.

whittard, w. F. 1961. The Ordovician trilobites of the Shelve Inlier. Part 6. Monograph of the
Palaeontographical Society, 115 (494), 197-228, pis 26-33.

Whittington, h. b. 1941. The Trinucleidae - with special reference to North American genera and species.
Journal of Paleontology, 15, 21-41.

1950. Sixteen Ordovician genotype Trilobites. Journal of Paleontology, 24, 531-565.

— 1959. Silicified Middle Ordovician trilobites: Remopleurididae, Trinucleidae, Raphiophoridae, Endy-
mioniidae. Bulletin of the Museum of Comparative Zoology, Harvard , 121, 371-496, pis 1-36.

— 1965a. A monograph of the Ordovician trilobites of the Bala area, Merioneth. Part 2. Monograph of the
Palaeontographical Society, 118 (504), 33-62, pis 11-18.

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

— 1966. A monograph of the Ordovician trilobites of the Bala area, Merioneth. Part 3. Monograph of the
Palaeontographical Society, 120 (512), 63-92, pis 19-28.

720

PALAEONTOLOGY, VOLUME 36

Whittington, H. b. and evitt, w. R. 1954. Silicified Middle Ordovician trilobites. Memoir of the Geological
Society of America , 59, 1-137, pis 1-37.

williams, A. 1956. Productorthis in Ireland. Proceedings of the Royal Irish Academy, Series B, 57, 179-183.

— 1962. The Barr and Lower Ardmillan Series (Caradoc) of the Girvan District, south-west Ayrshire, with
descriptions of the Brachiopoda. Memoir of the Geological Society, London, 3, 1-267, pis 1-25.

yi, Y. 1957. [The Caradocian trilobite fauna from the Yangtze-Gorges], Acta Palaeontologica Sinica, 5, 23-27.
[In Chinese with English summary],

zhou zhiyi and dean, w. t. 1986. Ordovician trilobites from Chedao, Ganzu Province, north-west China.
Palaeontology, 29, 743-786.

— yin gongzheng and tripp, r. p. 1984. Trilobites from the Ordovician Shihtzupu Formation, Zunyi,
Guizhou Province, China. Transactions of the Royal Society of Edinburgh, Earth Sciences, 75, 13-36.

M. ROMANO

Department of Earth Sciences
University of Sheffield
Dainton Building, Brookhill
Sheffield S3 7HF, UK

A. W. OWEN

Department of Geology and Applied Geology
Lilybank Gardens
University of Glasgow
Glasgow G12 8QQ, UK

Typescript received 15 May 1992

Revised typescript received 26 September 1992

UPPER DEVONIAN TETRAPODS FROM
ANDREYEVKA, TULA REGION, RUSSIA

by o. a. lebedev and j. a. clack

Abstract. Devonian tetrapod remains have been recovered from the Famennian of Russia. They occurred in
a limestone with stromatolites, algae, numerous and diverse fish remains and the holotype of the tetrapod
Tulerpeton curtum. The conditions indicate a shallow-water basin with carbonate-rich water and perhaps an
estuarine or marine situation. The individual bones show some plesiomorphic similarities to those of other
Devonian tetrapods and osteolepiform fishes, such as a high premaxillary tooth count, large fangs plus a
marginal row of smaller teeth on the vomer, a low naris with unsutured premaxillary-maxillary junction, lack
of shagreen on the coronoids and lateral line canals in tubes through the bone, but also share derived characters
with some Carboniferous tetrapods, such as shagreen on the vomer, shape and suture pattern of supratemporal
and intertemporal, and possession of a tabular horn. The findings indicate that tetrapods were already diverse
by the Devonian, and that they may not have been confined to freshwater.

The earliest known tetrapods occur in Upper Devonian deposits, and have been described from
East Greenland (Save-Soderbergh 1932; Jarvik 1952, 1980; Clack 1988, 1989; Coates and Clack
1990, 1991), Australia (Warren and Wakefield 1972; Campbell and Bell 1977), Scotland (Ahlberg
1991) and Central Russia (Lebedev 1984, 1985, 1990). These fossils are of great importance in
enhancing our understanding of the transition from fish to tetrapod and of the acquisition of the
unique characters by which tetrapods are defined.

Tulerpeton curtum (Lebedev 1984, 1985, 1990) is one of only three Devonian tetrapods for which
articulated material has been described. It consists of complete, articulated, right fore and hind
limbs, in which the digits are preserved. Most of the left half of the shoulder girdle and part of the
left pelvic girdle, some centra, ribs and articulated ventral scalation are preserved in association.
The significance of Tulerpeton lies partly in its possession of six digits on the manus and probably
six on the pes. Other known Devonian tetrapods also have more than five digits on each limb.
Acanthostega (Coates and Clack 1990) has eight digits on the manus and Ichthyostega (Coates and
Clack 1990) has seven digits on the pes. Together, these genera have contributed greatly to our
understanding of the origin of tetrapod limbs (Coates and Clack 1990; Coates 1991 ; Gould 1991).
The postcranial material of Tulerpeton is being described in detail by one of the current authors
(O. A. L.) and M. I. Coates of the University of Cambridge, but this paper describes cranial material
associated with the holotype specimen, and other cranial remains from the same horizon.

The holotype of Tulerpeton curtum and two associated cranial elements derive from a single
block of limestone from the Andreyevka-2 locality in the Tula Region in Central Russia. The
sedimentology and associated biota are interpreted as deriving from an environment similar to the
Black Sea limans, essentially estuarine or brackish conditions with both freshwater sources and
occasional marine incursions (Lebedev in press). Most tetrapod bones were found as isolated
elements distributed through the horizon, as were those from placoderms (antiarchs), acanthodians,
sarcopterygians (new species of osteolepiforms (to be described elsewhere), porolepiforms,
struniiforms, dipnoans), and actinopterygians (palaeonisciforms), as well as ostracodes, worms,
stromatolites and charophytes. Apart from the articulated Tulerpeton limbs, the bones are
dissociated, but most can be readily identified as belonging to one or other of the fish groups. The
tetrapod elements clearly do not belong to any of the known fish groups, and are identified as
tetrapod on the basis of the shape and bone ornamentation. Initially these were all attributed to

[Palaeontology, Vol. 36, Part 3, 1993, pp. 721-734.)

© The Palaeontological Association

722

PALAEONTOLOGY, VOLUME 36

Tulerpeton curtum, as tetrapods are rare in the fauna and were assumed to be represented by a single
taxon. Subsequently, at least two types of tabular have been identified, suggesting that other taxa
may have been present. Therefore we must be cautious in attributing any material except that
associated with the holotype, to Tulerpeton , though we believe this to be the most likely possibility
for the majority.

Tulerpeton represents the first early tetrapod to be associated with an estuarine or occasionally
marine environment. The remains exhibit marked differences from those of the other two described
Devonian genera, and indicate that by the Famennian, diverse tetrapod morphologies and ecologies
already existed.

MATERIAL AND METHODS

The material was collected during three trips in 1982-1983 by one of the authors (O.A.L.), who
joined the field teams of the Palaeontological Institute of the Academy of Sciences. The locality was
discovered by M. F. Ivakhnenko. The material is stored in the Palaeontological Institute of the
Academy of Sciences, Moscow (PIN), collection number PIN 2921. The premaxilla (PIN 2921/8)
and vomer (PIN 2921/9), still in sutural attachment, were found with the holotype postcranial
material and are attributed to it. Other isolated elements cannot be safely attributed at this stage.
Other material discussed is in the Natural History Museum, London (BMNH), National Museum
of Scotland, Edinburgh (NMS), Geological Museum, Copenhagen (MGUH).

Most of the cranial elements were found following acid digestion of rock from the fossiliferous
layer, and are listed below: a premaxillary (PIN 2921/35); two incomplete jugals (PIN 2921/36, 37);
two postfrontals (PIN 2921/41, 457), one of the postfrontals fitting onto a parietal (PIN 2921/457),
two further parietals (PIN 2921/38, 3014); a postorbital (PIN 2921/3002); an intertemporal (PIN
2921/3003); two supratemporals (PIN 2921/39, 40); three tabulars (PIN 2921/42, 447, 458); a
dentary (PIN 2921/32); a coronoid (PIN 2921/33); two angulars (PIN 2921/31, 446). Other
fragments include part of a maxilla (PIN 2921/34), and part of a possible further tabular (PIN
2921/1000), but they are either too incomplete or too uncertainly identified to warrant description
here. Pencil specimen drawings were made by one of the authors (O.A.L.) using a binocular
microscope MBS-1 with a grid inserted into one of the eye-pieces. These were then redrawn in
Chinese ink.

STRATIGRAPHY, SEDIMENTOLOGY AND TAPHONOMY

The Andreyevka-2 locality is situated on the Tresna River, 300 m upstream from Andreyevka
village (Suvorov District, Tula Region, Russia). A small outcrop of Khovanshchina beds
(Zavolzhsky horizon, Famennian, Upper Devonian), surrounded by Carboniferous strata, has been
exposed by erosion on the right bank near water level. Dating was made on the basis of the presence
of Eusthenodon sp. nov., also found in the Khovanshchina beds of the Draguny locality on the Plava
River (South of the Tula Region). The ostracodes from Andreyevka-2 were determined as being of
Khovanshchina age (Fa 2d-Tn la of the French-Belgian Basin) (V. A. Chizhova, personal
communication) and include the following taxa: Aparchites globulus, Bykavites nativus, Evlanella
soholovi, Glyptolichwinella cf. G. spiralis, Healdianella punctata, Aparchitellina sp., Carbonita sp.
The lowermost bed is a limestone containing isolated bones and scales of Holoptychius cf. H.
nobilissimus and a new osteolepidid. It is overlain by an almost continuous stromatolite layer.
Above that lie limestones containing articulated Remigolepis armata and Bothriolepis carapaces,
isolated sarcopterygian bones, and the remains of Tulerpeton curtum.

The overlying layer is a bone bed about 100 mm thick, filled with bones, scales and teeth of many
taxa: Antiarchi: Remigolepis armata', Sarcoptergyii ; Eusthenodon sp. nov., Osteolepididae gen. et
sp. nov., Strunius sp.; Dipnoi; Andreyevichthys epitomus\ Chondrichthyi fam., gen. et sp. nov.;
Acanthodii; Devononchus concinnus, D. laevis, ‘ Cheir acanthus' sp. ; Palaeonisci; Moythomasia sp.

Invertebrates are represented by very thin-shelled, undeterminable bivalves, and tubes of the
sedentary worm Serpula vipera, abundant on the upper surfaces of stromatolites and penetrating

LEBEDEV AND CLACK: DEVONIAN TETRAPODS

723

them. Gyragonites and stem-cores of charophyte algae may belong to the genus Quasiumbella. The
upper part of the section consists of intercalated limestones and clays, containing a few detached
scales and bones of fishes (Lebedev 1986).

The sedimentary environment was probably a quiet shallow basin, of warm, possibly marine or
brackish water, containing a high percentage of dissolved carbonates and clay particles (Lebedev
in press).

Most of the bones are well preserved and unworn, but, with a few exceptions such as the material
of Tulerpeton, completely disarticulated. Coarse-grained material is almost absent, suggesting still
water conditions. The bones appear to show no current sorting nor preferred orientation, but as
most of the specimens have been recovered by acid digestion, this must be judged on a partial
sample.

The removal of most of the head and the left part of the body and tail, while the right side and
the scale cover remain in articulation, suggests postmortem disruption of the body by decay gases
rather than scavenging. The mass death of fishes and tetrapods seen in the upper fossiliferous layers
may result from the basin having dried up at some stage. There was almost no water transportation
and subaqueous maceration was fast and efficient.

DESCRIPTION

Cranial material of Tulerpeton curtum

Premaxilla. The premaxilla is sutured to the vomer (Text-fig. 1a-d), allowing both to be oriented with respect
to the midline. The whole unit (PIN 2921/8, 9) can usefully be compared with those of other early tetrapods
and sarcopterygian [osteolepiform] fishes, in particular, the contemporary Ichthyostega (Jarvik 1980) and
Acanthostega currently under study by one of us (J.A.C.).

The premaxilla bears characteristically tetrapod-like ornament consisting of irregular pits and ridges (Text-
fig. 1b-c). It is slightly wider than long and deeper medially than laterally, with a short symphysial region. Its
shape indicates an animal with a broad, low snout, which is more characteristic of early tetrapods than of any
of the contemporary fishes except Panderichthys and Elpistostege (Worobyeva 1973; Vorobyeva 1977, 1980;
Schultz and Arsenault 1985). A wide, short process meets the nasal; the contact is almost transverse to the
midline (Text-fig. Ib). Medially, there is an embayment probably for paired or a single internasal like those
found in loxommatids (Beaumont 1977) and Acanthostega (Clack 1989) or a fontanelle like that in
Crassigyrinus (Panchen 1985), chroniosuchids (Ivakhnenko and Tverdokhlebova 1980) and zatrachidids
(Langston 1953). There is no notch for the external naris nor a sutural surface for contact with the maxilla.
Thus there may only have been a ligamentous junction between these bones, as is probable in Proterogyrinus
(Holmes 1984) and Acanthostega.

The palatal lamina of the premaxilla forms the anterior margin of a narrow, bean-shaped anterior palatal
fossa, which tapers to a point towards the posterior part of the bone. The lamina expands here to meet the
lateral margin of the vomer. The edge is gently curved; the fossa is prolonged postero-laterally by a gradually
tapering fissure to the level of the middle of the palatal lamina of the premaxilla, as in Acanthostega.

Anterior palatal fossae (or fenestrae) are present in all known osteolepiform fishes such as Eusthenopteron
(Jarvik 1980), such porolepiforms as Glyptolepis (Jarvik 1980), and in several primitive tetrapods such as
Ichthyostega (Jarvik 1980), loxommatids (Beaumont 1977), Crassigyrinus (Panchen 1985) and Greererpeton
(Smithson 1982). In the latter and in Acanthostega, the fossae are paired, separated by a process from the
vomers. As in the latter, it is unclear whether the premaxilla contributed to the margin of the choana, though
if it did, the contribution can only have been minimal.

The sensory canal enters the bone at the mid-point of the suture with the nasal and passes anterolaterally,
branching to the surface with nine funnel-shaped foramina, only slightly larger than those found in the dermal
ornament. Its posterior outlet lies dorsolateral to the base of the posterior tooth. Several sensory canal
foramina are found on the lateral surface joining the main sensory canal. The sensory line also lies within the
bone in Acanthostega and Greererpeton, where a similar pattern of pores is seen. In contrast in Crassigyrinus
and other early tetrapods, the premaxillary portion of the infraorbital canal lies in an open sulcus.

PIN 2921/35 is a fragment from the posterior part of a premaxilla. It is similar in general outline to PIN
2921/8 except that the postero-lateral edge of the palatal lamina is much more strongly curved, perhaps
indicating a shorter snout and more transversely orientated apical fossa. In PIN 2921/8 the large posterior
teeth are bordered laterally by a ridge, while in PIN 2921/35 the ridge is absent and the teeth are situated

724

PALAEONTOLOGY, VOLUME 36

text-fig. 1. a-d, Tulerpetoii curium Lebedev; Andreyevka; Famennian; PIN 2921/8, 9, right premaxilla and
vomer in (a) ventral, (b) dorsal, (c) anterior and (d) posterior views; e, undetermined tetrapod, composite left
jugal based on PIN 2921/36, 37, in lateral view. Scale-bars represent 10 mm.

LEBEDEV AND CLACK: DEVONIAN TETRAPODS

725

immediately at the edge of the apical fossa. The most striking difference is the type of dermal ornament.
In PIN 2921/35, small pits he on a generally smooth surface, bearing occasional vascular pores, but in PIN
2921/8, the pits are funnel-shaped, separated by gentle ridges rather than flat surfaces.

There are fourteen teeth on premaxilla PIN 2921/8, their size gradually increasing caudally with the
exception of the last, which is much smaller. The teeth are long and conical, their apices being strongly curved
posteromedially. Longitudinal grooves at their base merge into fine striations apically, typical of labyrinthodont
teeth.

A transverse section of the first premaxillary tooth (PIN 2921 /8a) showed polyplocodont folding (Schultze
1969), in which the bone does not enter between dentine folds. The median line of the fold is straight, with
neither meanders nor branches. There are no dentine zones such as those in Panderichthys (Schultze 1969). The
pattern is most similar to that in Megalichthys. In Tulerpeton , the folds are closely appressed, with the bone
excluded from between them. It is characteristic of most early tetrapod tooth folding that the fold-line
meanders. It is possible that a section through a vomerine tusk (standardly used for cross-sections by Schultze
and others, for example Atthey (1876; Embleton and Atthey 1874)) rather than a premaxillary tooth would
show this more complex pattern. It is also possible that this tooth pattern is genuinely more primitive and fish-
like. ‘Dark dentine’ (Panchen 1985) is absent.

Vomer. The vomer (PIN 2921/9) (Text-fig. 1a) is diamond-shaped, almost flat and shagreen-covered, except
for a triangular area posteriorly. This region is pierced by several large vascular foramina and bordered
anteriorly and laterally by a row of denticles larger than those of the shagreen field. These lie on a curved ridge
bearing a series of teeth, including a fang and replacement pit, and three smaller teeth. The ridge borders the
choana anteromesially. The anteromedial corner of the vomer lacks shagreen but bears a network of large
vascular foramina. A rugose longitudinal projection lies along the medial suture, which may indicate the
presence of a cartilage-covered pad which may have acted as a shock-absorber during jaw closure, preventing
possible injury to the vomer caused by the tips of dentary tusks. In Ichthyostega , there is a boss in the same
position. Laterally a depression perhaps accommodated an adsymphysial tusk of the lower jaw. A slightly
smaller pit is situated at the base of a vertical ridge which runs parallel to the sutural area with premaxilla.

Most of the dorsal (internal) surface of the vomer is smooth and only the posterolateral portion, which slopes
gently down to the edge of choana, is rugose and pierced by numerous vascular foramina. This area is sharply
demarcated from the rest of the dorsal surface by a distinct angle. It marks the anterior limit of the nasal
capsule and corresponds to a similar tuberous, pore-bearing area on the dorsal side of the ventral lamina of
premaxilla. Anterolateral to the tooth-bearing ridge, the vomer is produced into a lamina which forms a tongue
and groove contact with the premaxilla.

Among tetrapods, the pattern of dentition on the premaxilla and vomer is closely matched by, but is clearly
different from, that of Acanthostega. In that genus, there are thirteen premaxillary teeth, with a similar size
distribution to that of this premaxilla. Greererpeton possesses a similar distribution and number, but there is
relatively less variation among all except the last three teeth. Greererpeton has a very small posteriormost tooth,
preceded by two which are enlarged into fangs comparable in size to those on the palate. Ichthyostega has fewer
premaxillary teeth (nine), with little size variation along the row. The tooth distribution of this premaxilla
resembles those of other early tetrapods more closely than it does those of osteolepiforms such as
Eusthenopteron and the osteolepidids ; Panderichthys rhombolepis has about twenty small teeth on either side
of the jaw; their size distribution varies in different individuals. In some respects, the premaxillae PIN 2921/8,
35 are similar to that of a newly recognized Devonian tetrapod (PIN 54/ 180c) from Latvia, previously
attributed to Panderichthys bystrowi Gross (Vorobyeva 1962; Ahlberg 1991), which is being described by Drs
P. Ahlberg, E. Luksevics and one of us (O.A.L.)

Like the premaxillary dentition, that of the vomer is most similar to Acanthostega among tetrapods, but
differs in two respects. It has an expanded lamina anteriorly, bearing shagreen. This character is typical of most
other early tetrapods; in lacking this lamina, Acanthostega resembles osteolepiform fishes. In most other
tetrapod vomers, however, a large part of the shagreen field lies level with or posterior to the vomerine teeth.
The palatal specimen attributed to Crassigyrinus (BMNH 30532) by Panchen (1985), appears to have been
misinterpreted. It is currently under study by one of us (J.A. C.), but preliminary investigations show the
vomer to lack a shagreen field and to have a tooth distribution similar to that of Acanthostega and PIN 2921/9.
Vomers are unknown in the early anthracosaurs Eoherpeton (Smithson 1985) and Proterogyrinus (Holmes
1984), however, broad vomers associated with broad flat heads seem to be characteristic of the majority of early
tetrapods.

The second difference between the vomer of PIN 2921/9 and that of Acanthostega lies in the position of the
fang pair. In Acanthostega, the fang pair lies mesial to a curved tooth-bearing ridge as in osteolepiforms.

726

PALAEONTOLOGY, VOLUME 36

text-fig. 2. Undetermined tetrapod, cranial elements; Andreyevka; Famennian; a, PIN 2921/457, left
postfrontal and parietal in dorsal view; B-c, PIN 2921/41, right postfrontal in dorsal and ventral views; d-e,
PIN 2921/38, left parietal in dorsal and ventral views; f-g, PIN 2921/458, right tabular in dorsal and ventral
views; H-l, PIN 2921/42, right tabular in dorsal and ventral views; j-l, PIN 2921/39, right supratemporal in
dorsal, ventral and posterior views; m-n, PIN 2921 /40, left supratemporal in dorsal and ventral views. Scale-

bar represents 10 mm.

LEBEDEV AND CLACK: DEVONIAN TETRAPODS

727

whereas in PIN 2921/9, the posterior fang lies more or less within the tooth-bearing ridge. The anterior
replacement pit however, lies mesial to the tooth bearing ridge. In Ichthyostega the vomerine tusk pair, or
rather slightly enlarged teeth are found at the beginning of the tooth row, followed by four or five smaller
teeth (personal observation, O.A.L., J.A.C.).

Most early tetrapods have only a few teeth on the vomers, usually not more than a pair or, in some cases,
clumps of small teeth. A fang pair plus a curving ridge bearing a row of smaller teeth and denticles is
characteristic of many sarcopterygian fishes, and the distribution of teeth on this premaxilla and that of
Acanthostega is the same as that found in Eusthenopteron and Panderichthys. Ichthyostega is intermediate
between Acanthostega and PIN 2921/9 in the number of vomerine teeth, but it lacks shagreen. Some advanced
temnospondyls, such as capitosaurs, also show a fang pair and a row of smaller teeth on the vomer (Bystrow
and Efremov 1940), superficially like that of Devonian tetrapods. However, the relationship of the tooth row
to the choana is different, and the row of teeth is continuous, rather than having the fang pair displaced from
the row of small teeth. The condition is presumably convergent.

Undetermined cranial material

Jugal. The jugal is represented by two partial specimens from the right side, which together give an almost
complete picture of the bone (PIN 2921/36, 37) (Text-fig. 1e). The suborbital process is low and long, and the
postorbital lamina high ; the orbit margin is a gentle curve, suggesting a relatively large orbit. The general shape
of the bone is similar to the pattern found in Proterogyrinus , with a long, low suborbital region, and a deep
notch for suture with the squamosal. The maxillary articulating surface is almost horizontal and slightly
roughened; there is a poorly developed processus alaris. Like the jugal of Acanthostega, this bone shows a
combination of lateral-line pores and a groove, with the jugal sensory line opening to the surface by a row of
ovoid pores, and the postorbital commissure running in an open, although deep groove.

Postfrontal. The postfrontal is a long crescentic element (PIN 2921/41, 457) (Text-fig. 2a-c). Ventrally, the
smooth surface is excavated so that while the lateral margin is thin, the bone thickens mesially to form a ridge
along the suture with the parietal and frontal. In this respect and in its general proportions and shape it most
closely resembles those of embolomeres such as Pholiderpeton (Clack 1987). Entry and exit foramina suggest
the possible presence of an internal sensory canal, but no pores can be observed on the surface of the bone.

Parietal. The parietal is known from three specimens, one in sutural attachment with its postfrontal (PIN
2921/457) (Text-fig. 2a), a second isolated and somewhat broken example (PIN 2921/38) (Text-fig 2d-e) and
a third very small specimen (PIN 2921 /3014). PIN 2921 /457 shows irregular pit and ridge ornament. The base
of each pit is pierced by 1-3 foramina for blood vessels. The ornament of PIN 2921/38 is similar, though the
pits are shallower, and there are no radiating grooves at the margins, which are almost smooth. The pineal
foramen is large, and somewhat anteriorly placed. The edge of the pineal foramen is significantly raised, and
lacks the usual pit and ridge ornament in PIN 2921 /38 and bears only tiny vascular foramina, or small round
pits. PIN 2921 /457 bears depressions laterally and anterolaterally to the pineal foramen, resembling Pteroplax
cornutus (Panchen 1970) in this respect. The parietal contacts (PIN 2921/38, 457 Text-fig. 2a, d-e) with
postfrontal, intertemporal, supratemporal and postparietal are almost equal in length and straight, with no
embayment for the supratemporal. The interparietal suture anterior to the pineal opening is complicated, with
an overlapping surface.

PIN 2921/3014 is only about 7 mm in length and the ornament is very poorly developed, consisting only of
small pits in the centre and radiating grooves laterally. It probably represents a very young individual.

Postorbital. The postorbital (PIN 2921/3002) (Text-fig. 3a-b), lacking only its posterior corner, is a plate-like
triangular bone of a rather simple construction. The anterior margin, representing the posterior edge of the
orbit, is slightly thickened dorsally at the area of a contact with the postfrontal. This contact is a simple smooth
surface, bearing several vascular pores posteriorly and two larger foramina anteriorly, but neither rugosity, nor
sutural sculpturing is evident. Ventrally the bone and corresponding orbital margin become thinner towards
the overlap area with the jugal. Where the postfrontal and the intertemporal meet, the surface is marked by
the opening of a rather large canal, perhaps representing the postorbital commissure of the lateral line canal.
The contact area with the intertemporal is a shelf, bearing two rows of vascular pores on its lateral surface;
the mesial surface is represented by a smooth narrow plate that appears to indicate kinetic attachment of skull
roof and cheek as found in anthracosaurs. In Crassigyrinus, a similar plate constitutes the dorsal margin of the
squamosal contacting the intertemporal posterior to the postorbital, seen in the holotype specimen NMS

728

PALAEONTOLOGY, VOLUME 36

text-fig. 3. Undetermined tetrapod, cranial elements; Andreyevka; Famennian; a-b, PIN 2921/3002, right
postorbital in dorsal and ventral views, x 4; c, PIN 2921/3003, left intertemporal in dorsal view, x 4; d, PIN

2921/447, right tabular in dorsal view, x 10.

G. 1859.33. 104. The dermal ornament consists of small pits concentrated in the antero-dorsal corner, with
grooves and ridges directed in a fan-shaped manner posteriorly and ventrally. The medial surface is smooth
except for a narrow groove running parallel to the orbit margin and to a groove-like depression on the lateral
surface parallel to the orbit margin.

LEBEDEV AND CLACK: DEVONIAN TETRAPODS

729

Intertemporal. The outline of the intertemporal (PIN 2921/3003) (Text-fig. 3c) is similar to that in
Crassigyrinus (Panchen 1985), ‘ Eogyrinus ’ (Panchen 1972a; = Pholiderpeton, Clack 1987), Proterogyrinus
(Holmes 1984) and Archeria (Holmes 1989). It is a roughly oval-trapezoid bone, its medial margin almost
straight except for a small angle at about the mid-point. Here the sutural surface changes from a dorsally
oriented anterior portion, to a more ventrally oriented posterior portion. The posterior margin bears an
overlap for the supratemporal. The anterior margin shows a sutural notch, probably for the posterior corner
of the postfrontal. The lateral margin, although partly broken off, is clearly sutureless; parallel to it runs a
groove, rugose posteriorly, and an acute ridge. The groove and ridge might indicate continuation of the kinetic
margin along the intertemporal as in Crassigyrinus. The ventral surface bears a central depression and several
vascular foramina, possibly marking the anterior part of the roof of the adductor chamber. The ornament of
most of the dorsal surface consists of a network of ridges with pits, more elongated towards the margins,
between them.

Supratemporal. The supratemporal is an approximately pentagonal bone and is represented by two specimens
of different sizes (PIN 2921 / 39, 40; Text-fig. 2j-n). The smaller specimen may represent a younger ontogenetic
stage; it is generally similar to the larger specimen, but with its features in a less well developed form, for
example, in dermal ornament. That of the smaller consists of numerous vascular pores; pits are present only
at the periphery of the bone and are not as conspicuous as those on the larger element. The lateral margin is
straight and lacks the interdigitations of a conventional sutural contact. This, like the matching margin of the
postorbital and intertemporal, may be evidence of a ‘kinetic’ junction between the cheek and skull table, as
in Pteroplax cornutus, which looks very similar (Clack 1987; O.A.L. personal observation). The contact for
the intertemporal is oblique and arch-shaped. The parietal suture is long and almost straight. The posterior
margin consists of two sutural facets : a lateral one for the tabular, occupying more than half of the total length,
and a relatively short mesial facet for the postparietal. The implication is that there was no tabular-parietal
contact as in anthracosaurs, but that the primitive condition of postparietal-supratemporal contact was
retained. The ventral surface bears two depressions medially, an anterior and a posterior one, separated by a
short ridge. The surface of the posterior depression is smooth and that of the anteromedial one bears slightly
developed radial ridges and rugosities. The posterior depression may represent part of the roof of a spiracular
chamber or its homologue, and the anteromedial one part of the roof of the adductor chamber.

Tabulars. Two elements are clearly identified as tabulars, PIN 2921/458 (Text-fig. 2f-g), PIN 2921/447, (Text-
fig. 3d). The tabular PIN 2921/458 bears a small ‘horn’, like those of many early tetrapods, such as
loxommatids, Crassigyrinus and Proterogyrinus. It is smooth and covered by vascular pores dorsally and
rugose ventrally. The rugosity suggests the attachment of ligaments, probably running between the tabular and
the shoulder girdle. The supratemporal suture is oblique and the contact area is wide; that for the postparietal
is somewhat shorter and a little embayed for a lateral process from the postparietal. The edge bordering the
temporal notch is gently curved. The mesial corner is produced farther posteriorly than the tabular horn. The
posterior edge is not thickened, and lacks the occipital flange characteristic of anthracosaurs, though it would
form a similar profile to their characteristic ‘widow’s peak’. Ventrally, a single large unfinished area, on a
raised boss, indicates the attachment facet for the paroccipital process. The condition is most similar to that
in loxommatids (Beaumont 1977 and personal observation O.A.L. , J.A.C.), where a single facet is also found
anteriorly placed on the bone. The dermal skull roof must have overhung the occipital face of the braincase
to a significant degree. Double paroccipital facets are found on the tabulars of Crassigyrinus , as in
anthracosaurs. In its relatively anterior position, the facet on this bone is most similar to the more anterior of
these, but it is not possible to be sure to which it is really homologous.

The tabular specimen PIN 2921/447 is generally similar to PIN 2921/458, but is much smaller. It probably
represents a younger individual, but it may derive from a different species. It differs from PIN 2921/458 in the
following respects. The bone is proportionately shorter and broader, and the horn is relatively larger and more
massive. The posterior margin of the ornamented surface is straight and the free lateral margin not embayed.
An oblique suture on the lateral margin may be for a narrow process of the supratemporal or for a process
of the squamosal, as, for example, in Loxomma acutirhinus (Beaumont 1977, fig. 2a). On the ventral surface,
there is no facet for the opisthotic.

PIN 2921/42 (Text-fig. 2h-i) is also identified as a tabular, though it is in some respects unusual. It has
similar features to PIN 2921/458, including the presence of the opisthotic facet on the ventral surface and a
small tabular horn. The margin of the temporal notch is much more strongly curved laterally, so that the
tabular would have contributed to the anterior as well as the dorsal margin of the temporal notch. The lateral
margin of this process bears an oblique suture, presumably for the squamosal, and in this feature, resembles

730

PALAEONTOLOGY, VOLUME 36

PIN 2921 /447. If correctly identified, this feature would indicate that the skull /cheek contact was a firm suture,
rather than a ‘kinetic’ one, and thus the element probably belongs to a separate taxon.

Lower jaw

Dentary. There are at least sixty marginal teeth on the dentary, their size gradually increasing caudally and
reaching a maximum at the beginning of the posterior third of the dentary, decreasing thereafter. The lateral
surface of the dentary (PIN 2921/32; Text-fig. 4a-b) is pierced by blood-vessel pores. These, situated at the

text-fig. 4. Undetermined tetrapod, cranial elements; Andreyevka; Famennian; a-b, PIN 2921/32, right
dentary in lateral and medial views; c, PIN 2921/33, left coronoid in dorsomedial view; d-e, PIN 2921/31,
right angular in lateral and medial views. Scale bar represents 10 mm.

bottoms of pits, are dispersed along the entire marginal tooth row at the uppermost margin of the bone. The
lower part of the lateral surface is smooth, but bears occasional longitudinal grooves. The height of the vertical
lamina decreases posteriorly to about a third of its maximum. The bone is enlarged anteriorly and forms a

LEBEDEV AND CLACK: DEVONIAN TETRAPODS

731

horizontal symphysial lamina which bears a pair of fangs, larger than the marginal teeth. Teeth in this position
are also found in Acanthostega, Ichthyostega (Jarvik 1980), and apparently in Proterogyrinus (Holmes 1984),
where they are also significantly larger than the marginal teeth. There is a rugose area on the lateral surface
of the vertical lamina ventral to the symphysial plate. This area could be of perichondral or ligamentous origin
and probably served as the attachment point for the lower jaw rami by a mentomandibular cartilage or short
ligaments.

Coronoid. One coronoid is represented in the collection (PIN 2921/33; Text-fig. 4c). It is unlikely to be a
posterior coronoid, since there is no adductor fossa notch posteriorly and each of the four sides is bounded
by a sutural surface. Two rows of coronoid teeth are present; a lateral row of small teeth on the low vertical
coronoid lamina as in osteolepiforms (for example, Eusthenopteron (Jarvik 1980), Chrysolepis (Lebedev, 1983),
Panderichthys rhombolepis (Gross 1941), and Holoptychius (Jarvik 1980), and a medial row of larger teeth of
which those in the centre are the largest. The structure and dentition of this bone is different from those of both
other tetrapods and of most sarcopterygian fishes (i.e. excluding dipnoans). In fishes, the medial row of teeth
consists of only a fang pair, to which the largest teeth in the row of PIN 2921/33 may be homologous. In
Ichthyostega and the tetrapod Doragnathus woodi (possibly a juvenile Spathicephalus ) (Smithson 1980a, 1980b)
th adsymphysial plate and the coronoids bear a vertical lamina with a single row of teeth; there are the gaps
in the tooth row between the coronoids in Doragnathus. In Crassigyrinus , fangs are situated within the main
tooth row. Most other tetrapods bear shagreen on the coronoids, but may also bear teeth, usually quite small
and irregularly arranged on the shagreen field. It is not clear to which row of teeth of Devonian tetrapods or
sarcopterygians those of later tetrapods may be homologous.

Angular. Isolated angulars show the posteroventral margin (PIN 2921 /31, 446) (Text-fig. 4d-e) to be a shallow
curve, the length being more than four times the height. The lateral lamina reaches its maximum curve in the
middle of the bone. The ornament consists of deep pits in the centre of the bone; dorsally and anteriorly they
turn into deep grooves that diverge and become shallow and numerous. The mandibular sensory canal was
housed in a deep groove in PIN 2921/31, like that in most other tetrapods, rather than in a canal as in
Acanthostega and Ichthyostega. In PIN 2921/446 however, the central part is enclosed within the bone, as it
is in a specimen from Celsius Bjerg, Greenland collected in 1947 (MGUH A88) (figured by Clack 1988, text-
fig. 8 as a ‘new taxon’). As in Crassigyrinus, there is a zig-zag suture with the postsplenial ventrally, where the
lateral line groove is carried forward on a process of the angular before it passes onto the postsplenial. The
mesial lamina is very narrow and bears a notch for a small Meckelian foramen, situated towards the middle
third of the bone.

DISCUSSION

This material demonstrates a number of characters in which it is most closely comparable with the
other Devonian tetrapods, Acanthostega , and Ichthyostega , and others in which it resembles post-
Devonian tetrapods more closely. Characters of the dentition provide some of the most useful and
illuminating contrasts between sarcopterygians, Devonian tetrapods, and post-Devonian tetrapods.

Premaxillary dentition. The configuration of the premaxillary dentition in the new material, as in the
other Devonian forms, and in Greererpeton, is apparently derived with respect to related
sarcopterygian fishes in the reduction in number of teeth, but primitive with respect to most later
tetrapods, in which many forms show a further reduction. Temnospondyls almost invariably retain
a large number of premaxillary teeth, usually more than ten, and sometimes as many as eighteen,
while reptiliomorphs, microsaurs, ai'stopods, lysorophids and nectrideans have fewer. Anthraco-
sauroids usually have fewer than six.

Vomerine dentition. The vomerine dentition of the new material resembles that of Acanthostega and
osteolepiform and porolepiform fishes rather than most tetrapods, but in possessing an expanded
lamina bearing shagreen it may share a derived character with other tetrapods.

Coronoid dentition. No Devonian tetrapod coronoid shows shagreen, and its possession may
constitute a derived character uniting post-Devonian tetrapods.

732

PALAEONTOLOGY, VOLUME 36

Naris. In the unsutured condition of the premaxillary-maxillary junction, the new material
resembles Acanthostega , Ichthyostega and Proterogyrinus. The naris must have been situated low on
the snout, and the weight of evidence now strongly suggests that this was the primitive condition,
rather than the high position found in Crassigyrinus. The low position is found in all three Devonian
tetrapods, and in a number of others judged to be primitive on independent grounds, for example
Proterogyrinus and Greererpeton. While the condition in Crassigyrinus superficially resembles
that of Eusthenopteron, it appears to be unusual in structure. It is currently being restudied by one
of the authors (J. A.C.). The polarity of this character was debated by Panchen (1985) with respect
to judging the validity of outgroup comparisons, but it becomes very hard to argue that a high
position as in osteolepiforms and Crassigyrinus is genuinely the primitive condition for tetrapods.
The question is analogous to that of the condition of the tetrapod stapes, in which strict outgroup
comparisons with osteolepiforms suggest that a double-headed stapes should be primitive (Bolt and
Lombard 1985). However, single-headed stapes have now been found in three very primitive and
unrelated tetrapods, and the weight of evidence (admittedly stratophenetic) strongly suggests that
this was really the plesiomorphic condition (Clack 1989, 1992; Bolt and Lombard 1992).

Skull table characters. Among characters of the skull roof bones, those of the tabular show
recognizable similarities to those of post-Devonian tetrapods, in particular loxommatids and
Crassigyrinus , in the presence of a tabular ‘horn’ or button, and in the form of the paroccipital
facet. The tabulars of the other two Devonian forms are each distinctive and apparently specialized.
Neither resembles those of sarcopterygian fishes nor other tetrapods. These characters of the new
tabulars may constitute synapomorphies with the later tetrapods, at least the so-called
‘labyrinthodont’ taxa. Similar arguments may apply to the supratemporal, in its proportions and
the form of its contacts with other bones.

The apparent occurrence of skulls both with and without a ‘kinetic line’ between skull table and
cheek is of some interest with respect to the polarity of this character. Presence of the ‘kinetic line’
is usually regarded as primitive (e.g. Watson 1926; Panchen 1970, 19726), and occurs in
Crassigyrinus and anthracosaurs, but it does not occur in the Devonian tetrapods Ichthyostega or
Acanthostega. A related question is the possession of an intertemporal, whose presence is almost
universally regarded as primitive but which is absent in Acanthostega and Ichthyostega. The
significance of these characters is being considered in studies of Acanthostega by one of the authors
(J.A.C.) and M. I. Coates (University of Cambridge).

Lateral line organs. In having the lateral line organs enclosed in canals through the bone, the new
material is fish-like. If the cranial material belonged to one taxon, a greater proportion of the lateral
line would have lain in open grooves than in Acanthostega, and we should see a condition more like
that found in Greererpeton (Smithson 1982).

Despite the fact that only two of these cranial elements can be safely attributed to Tulerpeton, they
supply an important insight into the evolution of certain tetrapod characters, in their divergence
from the corresponding sarcopterygian condition. In several characters, these elements most closely
resemble those of the contemporary and primitive Acanthostega and others more closely resemble
those seen in post-Devonian tetrapods. To summarize, in the following characters Tulerpeton
resembles the other known Devonian tetrapods; in the high number of premaxillary teeth, the
vomerine dentition, in lacking shagreen on the coronoid, in the unsutured premaxillary-maxillary
junction, and in the low position of the naris. In the following character it resembles post-Devonian
tetrapods; in having a shagreen field on the vomer. The postcranial material also shares characters
of the radius and ulna, tibia and fibula with post-Devonian tetrapods. The unattributed cranial
material shows characters of the tabular, supratemporal and intertemporal shared with post-
Devonian tetrapods. The apparent presence of both ‘kinetic’ and ‘unkinetic’ skull table forms
provides equivocal evidence about the polarity of this character. In the condition of the lateral line
canals, the material shows conditions intermediate between the Devonian genera and Lower
Carboniferous forms, but it may not all belong to one taxon.

LEBEDEV AND CLACK: DEVONIAN TETRAPODS

733

Acknowledgements. We thank V. D. Kolganov for redrawing the figures in ink, the University of Cambridge
Travel Fund and the Cambridge Philosophical Society for financial help for J. A. Clack to visit Moscow, and
the Trustees of the Curry Fund (Geological Association) for financial support of O. A. Lebedev during his visit
to Britain.

REFERENCES

ahlberg, p. E. 1991. Tetrapod or near-tetrapod fossils from the Upper Devonian of Scotland. Nature, 354,
298-301.

atthey, t. 1876. On Anthracosaurus russelli Eluxley. Annals and Magazine of Natural History, 18, 146-167.
beaumont, E. h. 1977. Cranial morphology of the Loxommatidae (Amphibia: Labyrinthodontia). Philo-
sophical Transactions of the Royal Society of London, Series B, 280, 29-101.
bolt, J. R. and lombard, r. e. 1985. Evolution of the amphibian tympanic ear and the origin of frogs.
Biological Journal of the Linnean Society of London, 24, 83-99.

1992. Nature and quality of the fossil evidence of otic evolution in early tetrapods. 377-404. In

webster, d., popper, A. N. and fay, r. R. (eds). Evolutionary biology of hearing. Springer-Verlag, New York,
859 pp.

bystrow, a. p. and efremov, J. A. 1940. Benthosuchus sushkini Efremov, a labyrinthodont from the Eotriassic
of Sharzhenga River. Travaux de I'institut paleozoologique de F Academic des sciences de FURSS, 10, 1-152.
Campbell, K. w. s. and bell, M. w. 1977. A primitive amphibian from the Late Devonian of New South Wales.
Alcheringa, 10, 369-381.

clack, J. A. 1987. Pholiderpeton scutigerum Huxley, an amphibian from the Yorkshire coal measures.
Philosophical Transactions of the Royal Society of London, Series B, 318, 1-107.

— 1988. New material of the early tetrapod Acanthostega from the Upper Devonian of East Greenland.
Palaeontology, 31, 699-724.

1989. Discovery of the earliest tetrapod stapes. Nature, 342, 425 — 430.

— 1992. The stapes of Acanthostega gunnari and the role of the stapes in early tetrapods. 405^120. In
webster, D. b., popper, A. N. and fay, R. R. (eds). Evolutionary biology of hearing. Springer-Verlag, New
York, 859 pp.

Coates, M. i. 1991. New palaeontological contributions to limb ontogeny and phylogeny. 325-337. In
hinchliffe, j. r. (ed.). Developmental patterning of the vertebrate limb. Plenum Press, New York, 452 pp.

and clack, j. a. 1990. Polydactyly in the earliest known tetrapod limbs. Nature, 347, 66-69.

— 1991. Fish-like gills and breathing in the earliest known tetrapod. Nature, 352, 234—236.
embleton, D. and atthey, T. 1874. On the skull and some other bones of Loxomma allmanni. Annals and
Magazine of Natural History, 14, 38-63.
gould, s. J. 1991. Eight (or fewer) little piggies. Natural History, 1, 22-29.

GROSS, w. 1941. Uber den Unterkiefer einiger devonische Crossopterygier. Abhandlungen der preussische
Akademie der Wissenschaften, Mathematisch-Naturwissenschaftliche Klasse, 1, 1-51.
holmes, r. 1984. The Carboniferous amphibian Proterogyrinus scheeli Romer and the early evolution of
tetrapods. Philosophical Transactions of the Royal Society of London, Series B, 306, 431-527.

— 1989. The skull and axial skeleton of the Lower Permian anthracosauroid amphibian Archeria crassidisca
Cope. Palaeontographica Abteilung A, 207, 161-206.

ivakhnenko, m. F. and tverdokhlebova, G. i. 1980. [Systematics, morphology and stratigraphic significance
of the Upper Permian chroniosuchids of the east of the european part of the USSR.] Saratov University
Publishers, Saratov, 68 pp. [In Russian].

jarvik, E. 1952. On the fish-like tail in the ichthyostegid stegocephalians. Meddelelser om Gronland , 114, 1-90.

1980. Basic structure and evolution of vertebrates, 1. Academic Press, London, 591 pp.

langston, w. 1953. Permian amphibians from New Mexico. University of California Publications in Geological
Science, 29, 349-416.

lebedev, o. A. 1983. A new crossopterygian fish from the territory of the central Devonian field.
Paleontologicheskiiy Zhurnal, 4, 68-75.

1984. [The first find of a Devonian tetrapod vertebrate in the USSR.] Doklady Akademii Nauk SSSR,
Palaeontology, 278, 1470-1473. [In Russian.]

1985. [The first tetrapods: searches and findings]. Priroda, 11, 26-36. [In Russian].

— 1986. [New ichthyofaunal finds and their significance for the resolution of the Devonian-Carboniferous
boundary problem]. 5-11. In zeisler, v. m. (ed.). [Geology and mineral resources of the central regions of the
East European Platform .] Moscow Naturalists’ Society, Nauka Publishers, Moscow, 154 pp. [In Russian],

1990. Tulerpeton, l'animal a six doigts. La Recherche, 225, 1274—1275.

734

PALAEONTOLOGY, VOLUME 36

panchen, a. l. 1970. Teil 5a, Anthracosauria. In kuhn, o. (ed.). Handbuch der Palaoherpetologie. Fischer,
Stuttgart, 84 pp.

— 1972a. The skull and skeleton of Eogyrinus attheyi Watson (Amphibia: Labyrinthodontia). Philosophical
Transactions of the Royal Society of London , Series B , 263, 279-326.

— 19726. The interrelationships of the earliest tetrapods. 65-88. In joysey, k. a. and kemp, t. s. (eds). Studies
in vertebrate evolution. Oliver and Boyd, London, 284 pp.

— 1985. On the amphibian Crassigyrinus scoticus Watson from the Carboniferous of Scotland. Philosophical
Transactions of the Royal Society of London, Series B , 309, 461-568.

save-soderbergh, g. 1932. Preliminary note on the Devonian stegocephalians from East Greenland.
Meddelelser om Gronland, 94, 1-107.

schultze, h.-p. 1969. Die Faltenzahne der rhipidistiiden Crossopterygier, der Tetrapoden und der
Actinopterygier-Gattung Lepisosteus, nebst einer Beschreibung der Zahnstruktur von Onychodus
(struniiformer Crossopterygier). Palaeontographica Italica, 65, 63-136.

— and arsenault, m. 1985. The panderichthyid fish Elpistostege : a close relative of tetrapods?
Palaeontology, 28, 293-310.

smithson, t. r. 1980a. An early tetrapod fauna from the Namurian of Scotland. 407 — 438 . In panchen, a. l.
(ed.). The terrestrial environment and the origin of land vertebrates. Academic Press, London, 633 pp.

— 19806. A new labyrinthodont amphibian from the Carboniferous of Scotland. Palaeontology, 23,
915-923.

— 1982. The cranial morphology of Greererpeton burkemorani Romer (Amphibia: Temnospondyli).
Zoological Journal of the Linnean Society of London, 76, 29-90.

1985. The morphology and relationships of the Carboniferous amphibian Eoherpeton watsoni Panchen.
Zoological Journal of the Linnean Society of London, 85, 317-410.
vorobyeva, E. i. 1962. [Rhizodont crossopterygian fishes of the main Devonian field of the USSR]. Trudy
Paleontologicheskogo Instituta, 94, 1-139. [In Russian],

— 1977. [Morphology and peculiarities of evolution of crossopterygian fishes]. Trudy Paleontologicheskogo
Instituta, 163, 1-239. [In Russian],

— 1980. Observations on two rhipidistian fishes from the Upper Devonian of Lode, Latvia. Zoological
Journal of the Linnean Society of London, 70, 191-201.

— and schultze, h. p. 1991. Description and systematics of panderichthyiid fishes, with comments on their
relationship to tetrapods. 68-109. In trueb, l. and schultze, h. p. (eds). Origins of the higher groups of
tetrapods. Cornell University Press, New York, 724 pp.

warren, j. w. and Wakefield, n. a. 1972. Trackways of tetrapod vertebrates from the Upper Devonian of
Victoria, Australia. Nature , 238, 469^170.

watson, d. m. s. 1926. The Croonian Lecture: The evolution and origin of the Amphibia. Philosophical
Transactions of the Royal Society of London, Series B, 214, 189-257.
worobyeva, E. i. 1973. Einige Besonderheiten in Schadelbau von Panderichthys rhombolepis (Gross) (Pisces,
Crossopterygii). Palaeontographica, Abteilung A, 143, 221-229.

O. A. LEBEDEV

Palaeontological Institute of the Academy of
Sciences

Profsoyuznaya 123
1173-21 Moscow B-321, Russia

Typescript received 26 June 1992

Revised typescript received 7 September 1992

I. A. CLACK

University Museum of Zoology
Downing Street
Cambridge CB2 3EJ, UK

IMPLICATIONS FOR THE GASTROPOD FOSSIL
RECORD OF MISTAKEN CRAB PREDATION ON
EMPTY MOLLUSC SHELLS

by SALLY E. WALKER and SYLVIA BEHRENS YAMAHA

Abstract. Durophagous crabs were found to make unusually high rates of predatory mistakes by attacking
empty gastropods and models of intact bivalves. This mistaken predation is attributed to the crypticity of the
shell : if a crab cannot readily determine whether a shell contains food, as is the case with gastropod shells, it
will crush it. In contrast, empty bivalve shells (represented by half-shells) are readily examined by crabs and
rejected. The taphonomic implications, and importance for the gastropod fossil record, are two-fold. First,
where predatory crabs are abundant, shells of gastropods are prone to detrimental biological destruction at
three levels: while alive, inhabited by hermit crabs and empty. Bivalves are subject to predation at only one
level: while alive. Second, because empty gastropods are preyed upon, peel marks on fossil gastropods are
therefore not a reliable indication of crab predation. Mistaken predation is a source of taphonomic bias that
needs to be considered in interpreting predation events in fossil gastropods.

Bias in the marine invertebrate fossil record stems from two main sources: (1) physical factors,
such as varying sedimentation rates or current sorting; and (2) biological factors, such as rate of
hard part production, secondary inhabitation of shells, or bioerosion (for reviews see Walker 1989,
1990, 1992; Parsons and Brett 1991 ; Kidwell and Bosence 1991). Our study addresses one potential
source of biological bias: the impact of durophagous crabs on molluscan assemblages. Vertebrate
predators, rather than invertebrate predators, have received the most attention in taphonomic
studies (e.g. Zapfe 1939a, 19396, 1939c; Davis 1959; Brain 1967, 1980; Sutcliffe 1970, 1973; Mellett
1974; Brothwell 1976; Hill 1976, 1980; Mayhew 1977; Dodson and Wexlar 1979; Behrensmeyer
and Dechant-Boaz 1980; Haynes 1980; Shipman and Walker 1980; Steadman 1986). Additionally,
vertebrate predators have been the focus, or suggested agents, of taphonomic anomalies in
molluscan assemblages (e.g. Teichert and Serventy 1947; Carter 1974; Boucot 1981 ; Lindberg and
Kellogg 1982; Cadee 1989). How have invertebrate predators, such as crabs, affected the potential
fossil record of shelled prey (e.g. bivalves and gastropods)?

Unsuccessful predation on gastropods and subsequent shell repair by the living snail provide a
record of predation events that can be tracked in the fossil record (Vermeij et al. 1981, 1982). Peeled
apertures on gastropod shells, indicating crab predation, may also be used to infer predation
intensity for some gastropods species (Vermeij 1983). However, crabs have been observed to attack
model oyster shells (LaBarbera 1981) and empty gastropod shells in soft-sediment environments
(Walker 1988). We have found that rocky shore crabs attack empty gastropod shells at relatively
high rates, peeling or crushing the shells as if the shells contained prey. We have termed this
phenomenon ‘mistaken predation’. If rates of mistaken predation are high in modern intertidal
communities, then predatory peel marks on fossil gastropods may not adequately represent the
amount of live shelled prey attacked.

Gastropod shells are also inhabited by many species of invertebrates (for reviews see Vermeij
1987; Walker 1990). The evolutionary effects of predation on hermit crabs in contrast to living
snails has been shown (LaBarbera and Merz 1992; see also Rossi and Parisi 1973; Vermeij 1977),
but many other invertebrate species remain to be studied. Gastropod shells, then, are susceptible to
destruction at three levels: while inhabited by the living snail, while occupied by secondary

[Palaeontology, Vol. 36, Part 3, 1993, pp. 735-741.]

© The Palaeontological Association

736

PALAEONTOLOGY, VOLUME 36

occupants, and while empty. Bivalves are preyed upon only once, while alive. After the bivalve is
dead, it gapes or disarticulates into two valves. A crab can readily discern if the bivalve shell is
empty, but it cannot determine if a gastropod shell is empty. This disparity in predation may lead
to differential preservation of bivalve shells over gastropod shells in habitats where durophagous
crabs are abundant.

The object of our study was two-fold : (1) to determine the extent of mistaken predation by crabs
in a modern environment; and (2) to determine whether bivalves and gastropods are equally
susceptible to this preservational bias. Crabs use predominantly chemical cues in prey selection, but
may also use tactile cues (Case 1964; Pearson et al. 1979; Vermeij 1983). We predict that live
gastropod and bivalve prey (with both cues) should be attacked at a higher rate than model prey
(only tactile cues). Empty bivalve shells, represented by single valves, will not be attacked by crabs,
whereas empty gastropod shells will be attacked. If model or empty 'prey’ are attacked at a rate that
is at least 10 per cent that of live prey, then we consider mistaken predation to be an important
factor in modification of gastropod death assemblages.

EXPERIMENTAL WORK

Field and laboratory experiments were conducted at Friday Harbor Marine Faboratory, San Juan
Islands, Washington, in August 1990, when red rock crabs ( Cancer productus) were most active.
Cancer productus is an opportunistic forager that routinely feeds on bivalves and gastropods
(Knudsen 1964; Boulding and Hay 1984; Robles et al. 1989). Of all the intertidal crabs, this species
attains the largest size (up to 180 mm carapace width) and is the most voracious predator in the
Pacific Northwest (Palmer 1985; Behrens Yamada and Boulding unpublished data). This species
may be an important agent in structuring intertidal communities in the Pacific Northwest (Casilla
and Paine 1987; Robles et al. 1989) but more studies are needed to determine the degree of
importance. Cancer productus first appeared in the Pliocene of North America (Nations 1975, 1979).
The propensity for biological modification of the gastropod fossil record, via shell destruction, is
greater, however, as an additional twelve species of Cancer originated in North America between
the Miocene and Pleistocene.

In the laboratory, live, empty and hermitted gastropod shells ( Littorina sitkana : shell height
range, 13-18 mm; mean, 14-7 mm) were given to fourteen red rock crabs, Cancer productus
(carapace width range, 84-115 mm; mean, 99-8 mm) in the following manner: all crabs were housed
in individual plastic containers (220 x 220 x 90 mm) with mesh holes to allow for water circulation.
All containers were submerged in aquaria with a constant flow of sea water. The crabs were offered
a series of three live L. sitkana and three empty L. sitkana for six trials (trial period of eight hours)
to test if the crabs selectively attacked live or empty L. sitkana. Next, the same crabs were fed a series
of three live L. sitkana and three hermitted (Pagurus hirsutuisculus) L. sitkana for seven trials (trial
period of eight hours) to test if the crabs attacked both types of shelled prey. At the end of each trial,
the shells were scored for predation. Only available shells were scored; snails or hermit crabs that
crawled out of reach of the crab were deemed unavailable and were not used in the analysis.

To determine if the laboratory results were artefacts of confined crabs, gastropod shells were
tethered in the field. Five and empty L. sitkana shells were tethered to fishing net lead line and placed
in the low intertidal zone (1-5 foot level) in front of Friday Harbor Faboratories. Empty gastropod
shells were plugged with plasticine clay to prevent hermit crab occupancy. A total of 141 live and
141 empty shells were used in iterations of 41 live: 41 empty; 40 live: 40 empty; 40 live: 40 empty;
20 live: 20 empty for four trials of 24 hours duration. At the end of each trial, each shell was scored
whether it was crushed or peeled (for the purpose of this study both crushed and peeled shells were
scored as 'crushed’).

Five mussels ( Mytilus edulis ) were chosen for the bivalve experiment. Five ( n = 62), model
{n = 76) and half-shell (n = 71) mussels were attached to lead lines and placed in the low intertidal
(1-5 foot level) for three trials (each trial was 24 hours). Model mussels were made from clean
M. edulis shells that had been boiled and dried, plugged with plasticine clay and sealed with Z-spar

WALKER AND BEHRENS YAMADA: MISTAKEN CRAB PREDATION

737

text-fig. 1 . a, crab predation on live and empty Littorina sitkana for six laboratory trials. Only crabs in trials
five and six attacked significantly more live snails than empty shells (trial five: /2 = 14-2, df = 1, P < 0.001 ;
trial six: /2 = 21T, df = 1, P < 0.001). B, frequency of crab attack on live and empty L. sitkana in field
experiments. Numbers above columns represent sample size. Significantly more live than empty L. sitkana were
crushed on trials one and three (trial 1 : %2 = 14-4, df = 1, P < 0.001); trial 3: /2 = 1 5-8, df = 1, P < 0.001).

(splash-zone) epoxy. Mussels were tethered to the lead line with monofilament (15 lb' line. The
tethered mussels were emplaced below the pier of the laboratory, so that observations on crab
feeding could be monitored. At the end of each trial, crab-damaged shells were recorded.

RESULTS

Cancer productus crushed over 50 per cent of live, empty, and hermitted L. sitkana shells in the
laboratory (Table 1). This represents a high attack rate on all shelled ‘prey’. Initially, the crabs did
not discriminate between empty and living Littorina sitkana. However, after five trials (forty hours)
three of the fourteen crabs significantly preferred live over empty shells (Text-fig. 1). This apparent
learning skewed the overall results (Table 1) towards a significant difference in attack rate between
live and empty shells.

table 1. Fate of live, empty and hermitted Littorina sitkana offered to Cancer productus. When only available
‘prey’ are considered, significantly more live L. sitkana were attacked then empty shells (j2 = 26-2, df = 1,
P< 0.001) or hermitted shells (j2 = 26 0, df = 1, P< 0.001). Frequency of crushed shells (number of
crushed/number of available prey) is given in parentheses.

Laboratory

experiments

Crushed

shells

Intact

shells

Unavailable

prey

Total

offered

Live L. sitkana
vs

162(0-86)

27

33

222

Empty shells

141 (0-64)

81

0

222

Live L. sitkana
vs

131 (0-78)

37

123

291

Hermitted shells

147 (0 54)

126

18

291

In the field, crab attacks on live L. sitkana varied from 35 to 60 per cent per day and on empty
L. sitkana from 5 to 25 per cent per day (Text-fig. 2a). Therefore, even though crabs will attack
empty gastropod shells in the field they exhibited a significant preference for live gastropods.

738

PALAEONTOLOGY, VOLUME 36

Crabs readily attacked live and empty (model) mussels in the field. Attack rates for the tethered
mussels ( Mytilus edulis) varied between 95 and 100 per cent for live and approximately 50 per cent
for model mussels (Text-fig. 2b). Again, crabs showed a significant preference for live prey. Very few
half-shell shells were chipped (n = 4), suggesting that the crabs could readily examine these shells
and determined that the shells were empty. At high tide, up to fourteen red rock crabs foraged along
the mussel lines at night, with most activity occurring between 11 p.m. and 2 a.m. Attacked model
mussels were clearly seen through the water because of the bright green and blue of the plasticine
clay.

text-fig. 2. Field experiments with live,
model and half-shells of the mussel
Mytilus edulis. Frequency of crushed live
mussels was higher, and the frequency of
crushed half-shells was lower, than
expected by chance alone (y2 = 103-7,
df=2, P < 0.001).

DISCUSSION

All three of our predictions were supported. First, Cancer productus crushed shells of L. sitkana
whether they were empty, hermitted or live. However, significantly more living Littorina were
selected over hermitted or empty shells of Littorina. Second, crabs attacked live mussels and model
mussels in the field, further indicating that crabs are tactile foragers. In Florida, model bivalves
made of resin also were not immune from attacks by the stone crab, Menippe mercenaria, and the
blue crab, Cal/inectes sapidus (LaBarbera 1981). However, chemical cues must aid crabs in
distinguishing between live and model mussels in our study. Crabs, as visual and tactile predators,
are fooled by empty and model molluscs, but not necessarily all the time, because the rates of
predation differed between live and model shells. The rate of predation on live mollusc shells in the
field was extremely high. Crabs attacked over 95 per cent of the living mussels and 50 per cent of
the model mussels. If crabs relied solely on tactile or visual cues, almost all the model shells would
have been attacked. Lastly, mussel half-shells were rarely attacked. Crabs could readily distinguish
that half shells had no meat in them and did not break them open.

Most individuals of Cancer productus in our laboratory experiments failed to distinguish between
empty and live prey. It appears that crabs cannot determine if a gastropod shell is empty, thus the
shells are cryptic in terms of their food resource. Shells that are not cryptic, like the mussel half-
shells, can readily be explored by the crab and not damaged. Perhaps crabs learn to attack empty
gastropod shells because they may contain a food resource, such as amphipods, hermit crabs, or a
living snail deeply retracted in the shell. Crabs can learn better prey handling techniques (Hughes
1979; Cunningham and Hughes 1984) and thus, some crabs may learn as we suggest, to distinguish
empty from food-rich, inhabited shells. Alternatively, not all behaviours are optimal from our
human-based perspective (Rothstein 1982; Pyke 1984; Blaustein and Porter 1990) yet may persist
as seemingly non-adaptive traits (Gould and Lewontin 1979). As Maynard-Smith (1978) suggested,
we need to know the relative importance of the various processes affecting what we might call
maladaptive or suboptimal behaviour.

In conclusion, from a taphonomic perspective, we propose that in habitats where crabs are
common, the gastropod shell may be selected against, and the death assemblage will be extremely

WALKER AND BEHRENS YAMADA: MISTAKEN CRAB PREDATION

739

biased towards peeled and fragmented shells. That is, crab predation not only occurs on the living
snail and hermitted shells, but also on empty gastropod shells, creating an additional and important
biological bias against gastropod representation in the fossil record. Additionally, peeled shells do
not directly indicate predation on the primary food source (i.e. the living snail). Peeling and
puncturing, characteristics of predatory crabs, can also occur in empty gastropod shells because of
mistaken predation. Therefore, our results indicate that predation rates (based on peeled/crushed
apertures) attributed to crabs in the fossil record may be anomalous.

Bivalves may not suffer from this bias for two reasons: first, empty (articulated) valves are not
occupied by secondary inhabitants in a similar manner as gastropod shells (e.g. by hermit crabs,
amphipods), and hence may not be attacked, and secondly, the gaping valves may readily be
explored by a predatory crab. In conclusion, mistaken predation by crabs is a potential source of
preservational bias between gastropods and bivalves, presently unaccounted for in the fossil record.

Acknowledgements. We would like to thank D. Willows, Director, and D. Duggins for access to Friday Harbor
Marine Laboratories, H. Campbell for her invaluable field assistance, A. Blaustein for non-optimality
argumentation, and A. J. Boucot, E. G. Boulding, M. Kowalewski, A. Olson, S. A. Navarrete, P. D. Taylor
and anonymous reviewers for improving this manuscript. This research was supported, in part, by A. Boucot
(S.E.W.), NSF grant EAR 900-4519 (S.E.W.) and the Oregon State University Research Council (S.B.Y.)

REFERENCES

behrensmeyer, a. k. and dechant-boaz, d. e. 1980. The recent bones of Amboseli Park, Kenya, in relation to
East Africa paleoecology. 79-93. In behrensmeyer, a. k. and hill, a. p. (eds). Fossils in the making.
University of Chicago Press, Chicago, Illinois, 338 pp.
blaustein, A. R. and porter, r. h. 1990. The ubiquitous concept of recognition with special reference to kin.
123-148. In bekoff, M. and Jamieson, d. (eds). Interpretation and explanation in the study of animal behaviour.
Volume 1 : Interpretation, intentionality, and communication. Westview Press, Boulder, Colorado, 505 pp.
boucot, a. J. 1981. Principles of benthic marine paleoecology. Academic Press, New York, 463 pp.
boulding, E. G. and hay, t. k. 1984. Crab response to prey density can result in density-dependent mortality
of clams. Canadian Journal of Fisheries and Aquatic Science, 41, 521-525.
brain, c. k. 1967. Hottentot food remains and their bearing on the interpretation of fossil bone assemblages.
Scientific Papers of the Namib Desert Research Station, 32, 1-11.

— 1980. Some criteria for the recognition of bone-collecting agencies in African caves. 107-130. In
behrensmeyer, a. K. and hill, A. p. (eds). Fossils in the making. University of Chicago Press, Chicago, Illinois,
338 pp.

brothwell, D. R. 1976. Further evidence of bone chewing by ungulates: the sheep of North Ronaldsay,
Orkney. Journal of Archaeological Science, 3, 179-182.
cadee, G. c. 1989. Size-selective transport of shells by birds and its palaeoecological implications.
Palaeontology, 32, 429-437.

carter, r. w. G. 1974. Feeding sea birds as a factor in lamellibranch valve sorting patterns. Journal of
Sedimentary Petrology. 44, 689-692.

case, J. 1964. Properties of the dactyl chemoreceptors of Cancer antennarius Stimpson and C. productus
Randall. Biological Bulletin , 127, 428^146.

casilla, j. c. and Paine, r. t. 1987. Predation and community organization on eastern Pacific, temperate zone,
rocky intertidal shores. Revista Chilena de Historia Natural, 60, 131-151.

Cunningham, p. n. and hughes, R. N. 1984. Learning of predaceous skills by shorecrabs Carcinus maenas
feeding on mussels and dogwhelks. Marine Ecological Progress Series, 16, 21-26.
davis, d. h. s. 1959. The barn owl’s contribution to ecology and palaeoecology. Ostrich Supplement , 3,
144-153.

dodson, p. and wexlar, d. 1979. Taphonomic investigations of owl pellets. Paleobiology , 5, 275-284.
gould, s. J. and lewontin, r. f. 1979. The spandrels of San Marco and the Panglossian paradigm: a critique
of the adaptationist programme. Proceedings of the Royal Society of London, Series B, 205, 581-598.
haynes, G. 1980. Evidence of carnivore gnawing on Pleistocene and Recent mammalian bones. Paleobiology,
6, 341-351.

hill, a. p. 1976. On carnivore and weathering damage to bones. Current Anthropology, 17, 335-336.

740

PALAEONTOLOGY, VOLUME 36

hill, a. p. 1980. Early postmortem damage to the remains of some contemporary East African mammals.
131-152. In behrensmeyer, a. K. and hill, A. p. (eds). Fossils in the making. University of Chicago Press,
Chicago, Illinois, 338 pp.

hughes, R. n. 1979. Optimal diets under the energy maximization premise: the effects of recognition time and
learning. American Naturalist , 113, 209-221.

kidwell, s. M. and bosence, d. w. j. 1991. Taphonomy and time-averaging of marine shelly faunas. 115-209.
In allison p. a. and briggs, d. e. (eds). Taphonomy : releasing the data locked in the fossil record. Plenum
Press, New York, 560 pp.

knudsen, J. w. 1964. Observations of the reproductive cycles and ecology of the common Brachyura and crab-
like Anomura of Puget Sound, Washington. Pacific Science , 18, 3-33.
labarbera, m. 1981. The ecology of Mesozoic Gryphaea, Exogyra , and Ilymatogyra (Bivalvia: Mollusca) in a
modern ocean. Paleobiology , 7, 510-526.

- and mertz, r. 1992. Postmortem changes in strength of gastropod shells: evolutionary implications for
hermit crabs, snails, and their mutual predators. Paleobiology , 18, 367-377.

lindberg, d. and Kellogg, m. 1982. Bathymetric anomalies in the Neogene fossil record: the role of diving
marine birds. Paleobiology , 8, 402-407.

mayhew, d. f. 1977. Avian predators as accumulators of fossil mammal material. Boreas , 6, 25-31.
maynard-smith, j. 1978. Optimization theory in evolution. Annual Review of Ecology and Systematics, 9,
31-56.

mellett, j. s. 1974. Scatological origins of microvertebrate fossil accumulations. Science, 185, 349-350.
nations, G. d. 1975. The genus Cancer (Crustacea: Brachyura): systematics, biogeography and fossil record.
Science Bulletin of the Natural History Museum of Los Angeles County , 23, 1-104.

- 1979. The genus Cancer and its distribution in time and space. Bulletin of the Biological Society of
Washington, 3, 153-187.

palmer, A. R. 1985. Adaptive value of shell variation in Thais lamellosa: effects of thick shells on vulnerability
to and preference by crabs. Veliger, 27, 349-356.

parsons, K. p. and brett, c. e. 1991. Taphonomic processes and biases in modern marine environments: an
actualistic perspective on fossil assemblage preservation. 22-65. In donovan, s. k. (ed.). The processes of
fossilization. Belhaven Press, London, 303 pp.

pearson, w. H., sugarman, p. c., woodruff, D. L. and olla, b. 1979. Thresholds for detection and feeding
behaviour in the Dungeness crab, Cancer magister. Journal of Experimental Marine Biology and Ecology, 39,
65-78.

pyke, G. H. 1984. Optimal foraging theory: a critical review. Annual Review of Ecology and Systematics, 15,
523-575.

robles, c., sweetham, d. a. and dittman, d. 1989. Diel variation of intertidal foraging by Cancer productus L.

in British Columbia. Journal of Natural History , 23, 1041-1049.
rossi, a. c. and parisi, v. 1973. Experimental studies of predation by the crab Eriphia verrucosa on both snail
and hermit crab occupants of conspecific gastropod shells. Bollettino di Zoologica, 40, 117-135.
rothstein, s. i. 1982. Successes and failure in avian egg and nestling recognition with comments on the utility
of optimality reasoning. American Zoologist, 22, 547-560.

Shipman, p. and walker, A. 1980. Bone-collecting by harvesting ants. Paleobiology, 6, 496-502.
steadman, d. w. 1986. Holocene vertebrate fossils from Isla Floreana, Galapagos. Smithsonian Contributions
to Zoology , 413, 1-103.

Sutcliffe, a. J. 1970. Spotted hyaena: crusher, gnawer, digester and collector of bones. Nature, 227,
1110-1113.

- 1973. Similarity of bones and antlers gnawed by deer to human artefacts. Nature, 246, 428-430.
teichert, c. and serventy, d. l. 1947. Deposits of shells transported by birds. American Journal of Science,

245, 322-328.

vermeij, G. J. 1977. The Mesozoic marine revolution: Gastropods, predators and grazers. Paleobiology, 3,
245-258.

- 1983. Traces and trends of predation, with special reference to bivalved animals. Palaeontology, 26,
455—465.

- 1987. Evolution and escalation. Princeton University Press, Princeton, New Jersey, 527 pp.
schindel, d. e. and zipser, e. 1981. Predation through geological time: evidence from gastropod shell

repair. Science, 214, 1024-1026.

— zipser, e. and zardini, r. 1982. Breakage-induced shell repair in some gastropods form the Upper
Triassic of Italy. Journal of Paleontology, 56, 233-235.

WALKER AND BEHRENS YAMADA: MISTAKEN CRAB PREDATION

741

walker, s. e. 1988. Taphonomic significance of hermit crabs (Anomura: Paguroidea): epifaunal hermit
crab-infaunal gastropod example. Palaeogeography , Palaeoclimatology, Palaeoecology, 63, 45-71.

1989. Hermit crabs as taphonomic agents. Palaios, 4, 439-452.

1990. Biological taphonomy and gastropod temporal dynamics. 391 — 412. In miller, iii, w. (ed.).
P ale o community temporal dynamics: the long-term development of multispecies assemblies. Paleontological
Society Special Publication No. 5. 421 pp.

1992. Criteria for recognizing marine hermit crabs in the fossil record using gastropod shells. Journal of
Paleontology. 66, 535-558.

zapfe, h. 1939a. Lebensspuren der eiszeitlichen Hoehlenhyane; die urgeschichtliche Bedeutung der
Lebensspuren knochenfressender Raubtiere. Palaeobiologica. 7, 111-146.

19396. Untersuchungen iiber die Lebensspuren knochenfressender Raubtiere, mit besonderer
Berucksichtigung der Hyane. Akademie der Wissenschaften, Wien. Mathematische Naturwissenschaftliche
Klasse, 76, 33-35.

1939c. Lebensspuren der eiszeitlichen Hoehlenhyane und deren urgeschichtliche Bedeutung. Forschungen

and Fortschritte, 15, 269-270.

SALLY E. WALKER

Department of Geosciences
University of Arizona
Tucson, AZ 85721, USA

SYLVIA BEHRENS YAMADA

Typescript received 5 August 1992
Revised typescript received 22 October 1992

Department of Zoology
Oregon State University
Corvallis, OR 97731, USA

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 R. M. Owens, Department of Geology, National
Museum of Wales, Cardiff CF1 3NP, UK, 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 1993 is £55 00 (U.S. $120) for Institutional
Members, and £20 00 (U.S. $36) 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 per cent below the
listed prices: contact the Marketing Manager. Non-members may obtain Nos 35-46 (at cover price) from Basil Blackwell Ltd.,
Journal Subscription Department, Marston Book Services, P.O. Box 87, Oxford OX2 0DT, UK, 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 2-35 are
still in print and are available together with those listed below:

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. $40).

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. $40).

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. $60).

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. $60).

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. $60).

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. $40).

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. $80).

43. (for 1990) : Benthic palaeoecology of the Late Jurassic Kimmeridge Clay of England, by p. r. wignall, 74 pp., 50 text-figs.
Price £30 (U.S. $60).

44. (for 1990): The Murchison Symposium. Edited by M. G. bassett, p. d. lane and d. Edwards. 397 pp., 137 text-figs,
3 plates. Price £60 (U.S. $120)

45. (for 1991): Coniacian ammonite faunas from the United States Western Interior, by w. J. Kennedy and w. A. cobban.
96 pp., 31 text-figs, 17 plates. Price £35 (U.S. $70).

46. (for 1991): Arundian (Lower Carboniferous) conodonts from South Wales, by J. J. STONE. 63 pp., 14 text-figs, 5 plates.
Price £30 (U.S. $60)

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. $16) (Members £6

or U.S. $12).

2. (1987): Fossils of the Chalk, compiled by E. owen; edited by a. b. smith. 306 pp., 59 plates. Price £11 50 (U.S. $23)

(Members £9 90 or U.S. $20).

3. (1988): Zechstein Reef fossils and their palaeoecology, by n. hollingworth and t. Pettigrew, iv + 75 pp. Price £4 95

(U.S. $10) (Members £3 75 or U.S. $7-50).

4. (1991): Fossils of the Oxford Clay, edited by D. M. martill and j. d. Hudson. 286 pp., 44 plates. Price £15 (U.S. $30)

(Members £12 or U.S. $24).

1982. Atlas of the Burgess Shale. Edited by s. conway morris. 31 pp., 24 plates. Price £20 (U.S. $40).

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, 1993

Palaeontology

VOLUME 36 -PART 3

CONTENTS

Neoproterozoic (Vendian) phytoplankton from the Siberian
Platform, Yakutia

M. MOCZYDtOWSKA, G. VIDAL and V. A. RUDAVSKAYA 495

New material of an Early Cretaceous titanosaurid sauropod
dinosaur from Malawi

L. L. JACOBS, D. A. WINKLER, W. R. DOWNS and E. M. GOMANI 523

New anatomical characters in fossil coralline algae and their
taxonomic implications

J. C. BRAGA, D. W. J. BOSENCE and R. S. STENECK 535

Palaeoscolecid worms from the Middle Cambrian of Australia

K. J. MULLER and I. HINZ-SCHALLREUTER 549

Ediacaran-like fossils in Cambrian Burgess Shale-type faunas of
North America

S. CONWAY MORRIS 593

Dipterid ferns from the Mesozoic of Antarctica and New
Zealand and their stratigraphical significance

P. McA. REES 637

The temnospondyl amphibian Capetus from the Upper
Carboniferous of Czechoslovakia

S. E. K. SEQUEIRA and A. R. MILNER 657

Early Caradoc trilobites of eastern Ireland and their
palaeogeographical significance

M. ROMANO and a. w. OWEN 681

Upper Devonian tetrapods from Andreyevka, Tula, Russia

O. A. LEBEDEV and J. A. CLACK 721

Implications for the gastropod fossil record of mistaken crab
predation on empty mollusc shells

S. E. WALKER and S. B. YAMADA 735

Printed in Great Britain at the University Press, Cambridge

ISSN 0031-0239

Published by

The Palaeontological Association • London
Price £40-50

THE PALAEONTOLOGICAL ASSOCIATION

(Registered Charity)

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

COUNCIL 1993-1994

President : Dr W. D. I. Rolfe, National Museums of Scotland, Chambers Street, Edinburgh EH1 1JF
Vice-Presidents: Dr A. W. Owen, Department of Geology and Applied Geology, The University, Glasgow G12 8QQ
Dr M. E. Collinson, Department of Geology, Royal Holloway and Bedford New College,

University of London, Egham, Surrey, TW20 OEX
Treasurer : Mr P. S. Clasby, 12 Haglane Copse, Lymington, Hants S041 8DT
Membership Treasurer : Dr M. J. Barker, Department of Geology, University of Portsmouth, Bamaby Road,

Portsmouth POl 3QL

Institutional Membership Treasurer: Dr J. E. Francis, Department of Earth Sciences, The University, Leeds LS2 9JJ
Secretary : Dr J. A. Crame, British Antarctic Survey, High Cross, Madingley Road, Cambridge CB3 OET
Newsletter Reporter : Dr D. Palmer, c/o Department of Earth Sciences, Downing Street, Cambridge CB2 3EQ
Marketing Manager (Sales) : Dr L. Cherns, Department of Geology, University of Wales College of Cardiff CF1 3YE
Marketing Manager (Publicity): Dr P. B Wignall, Department of Earth Sciences, The University, Leeds LS2 9JJ
Public Relations Officer: Dr P. A. Allison, P.R.I.S., The University, P.O. Box 227, Whiteknights, Reading RG6 2AB
Newsletter Editor: Dr R. B. Rickards, Department of Earth Sciences, University of Cambridge, Cambridge CB2 3EQ
Editor of Field Guides to Fossils: Dr E. A. Jarzembowski, Booth Museum of Natural History, 194 Dyke Road,

Brighton BN1 5AA

Editors

Dr J. E. Dalingwater, School of Biological Sciences, Williamson Building, The University, Manchester Ml 3 9PL (co-opted)
Dr P. D. Lane, Department of Geology, University of Keele, Keele, Staffordshire ST5 5BG
Dr P. Doyle, School of Earth Sciences, University of Greenwich, Brewhouse Lane, London El 9NU
Dr A. R. Milner, Department of Biology, Birkbeck College, Malet Street, London WC1E 7HX
Dr C. J. Cleal, 37 Lipscombe Close, Newbury, Berkshire, RG14 5JW
Dr R. M. Owens, Department of Geology, National Museum of Wales, Cardiff CF1 3NP

Other Members: Dr B. M. Cox, British Geological Survey, Keyworth, Nottingham NG12 5GG
Dr M. P. Smith, School of Earth Sciences, University of Birmingham, Birmingham B15 2TT

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 Palaeoanthropology,
Academia Sinica, P.O. Box 643, Beijing. Dr Rong Jia-yxj, 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: Professor F. Alvarez, Departamento de Geologia, Universidad de Oviedo, CL 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, Fredskowej 4, 2840 Holte, Denmark. USA : Professor A. J. Rowell, Department of Geology, University of
Kansas, Lawrence, Kansas 66044. Professor N. M. Savage, Department of Geology, University of Oregon, Eugene, Oregon
97403. Professor M. A. Wilson, Department of Geology, College of Wooster, Wooster, Ohio 44961. Germany: Professor
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 1993 are:
Institutional membership . . . £80 00 (U.S. $160) Student membership . . £1 T50 (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 J. E.
Francis, Department of Earth Sciences, The University, Leeds LS2 9JJ. 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 M. J. Barker, Department of Geology, University of Portsmouth,
Barnaby Road, Portsmouth POl 3QL. Subscriptions cover one calendar year and are due each January; they should be
sent to the Membership Treasurer. All members who join for 1993 will receive Palaeontology, Volume 36, 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, UK. For older issues contact
the Marketing Manager.

US Mailing: Second class postage paid at Rahway, New Jersey. Postmaster: send address corrections to Palaeontology,
c/o Mercury Airfreight International Ltd, 2323 EF Randolph Avenue, Avenel, NJ 01001, USA (US mailing agent).

Cover: Reconstruction of Paraostenia voultensis Secretan from the Middle Jurassic of the Ardeche, France, preying upon
a coleoid. This is a typical thylacocephalan, a recently recognized arthropod group of uncertain, but probably crustacean,
affinity, ranging from Silurian to Cretaceous, x 1.25. Reproduced by permission of the Royal Society of Edinburgh and
Dr W. D. I. Rolfe from Transactions of the Royal Society of Edinburgh, 76, 398, fig. 4.

PTERYGOMETOPINE TRILOBITES FROM THE
ORDOVICIAN OF BALTOSCANDIA

by VALDAR JAANUSSON and LARS RAMSKOLD

Abstract. A revision of the Pterygometopinae from Baltoscandia has revealed that the subfamily is more
diverse at the generic level than previously known. New genera are Ingriops, Oelandiops, Upplandiops, and
Keilapyge. The earliest species of Achatella are included in the new subgenus A. ( Vironiaspis ). The new species
Oelandiops mirificus, Upplandiops calvus, Estoniops maennili, and E. fjaeckensis are described. Upplandiops
calvus differs from most other dalmanitaceans in having only ten thoracic segments.

The Pterygometopinae, as defined by Ludvigsen and Chatterton (1982), has its centre of
distribution in the Baltoscandian region and the western slope of the central and southern Ural
Mountains. The subfamily began to spread to other regions, such as the British Isles and North
America, comparatively late in the Middle Ordovician. The Early Ordovician records of
Pterygometopus from Morocco (Destombes 1972) and Spain (Hammann 1972, 1974; Rabano 1989)
have recently been revised (Henry et a/. 1992) and shown to refer to early representatives of the
family Dalmanitidae.

Previous knowledge of Baltoscandian pterygometopines is largely confined to Schmidt’s (1881)
monograph on species from northern Estonia and Ingria (the district in Russia between Estonia and
Lake Ladoga; Text-fig. 1). The illustrations in the monograph are small-scale drawings with which
even Schmidt himself was not quite satisfied. Mannil (1958) established the genus Estoniops ,
provided adequate illustrations of the cephalon of E. exilis (Eichwald, 1858) and described a new
species, E. bekkeri, but for the remainder of the East Baltic pterygometopines Schmidt’s
monograph remained the only available published source of information. The Swedish material of
the subfamily has never been described except for the type species of Pterygometopus, redescribed
by Whittington (1950), and the very rare species P. sandbyensis (Olin, 1906) and P. schmidti
Warburg, 1925, each represented by a single, fragmentary cephalon. Wiman (1908) figured
specimens believed by him to represent Pterygometopus ( = Estoniops) exilis.

The intention of the present paper is a revision of the Baltoscandian pterygometopines at the
genus level. In addition to previously undescribed type species of new genera, only a few new species
are included. The available material contains some ten additional species which appear to be new,
but the description of this material is a separate task, as is a phylogenetic analysis of the group.

The figured specimens of Ingriops trigonocephalus (Schmidt, 1881; PI. 5, fig. 1) and Estoniops
panderi (Schmidt, 1881) (PI. 1, fig. 3) are from A. von Volborth’s collection and were presented to
the Swedish Museum of Natural History by F. Schmidt. They are identified by him on the labels,
and quite obviously represent syntypes because Volborth’s collection was one of F. Schmidt’s main
sources of material from Ingria. We do not attempt to designate lectotypes for these two species or
for other species described by Schmidt (1881) which are considered in this paper. Designation of
lectotypes should be done in the course of a substantial revision of the material available to F.
Schmidt. Such a revision was outside the scope of the present paper.

TERMINOLOGY

The terminology used in this paper mainly follows that of Harrington et al. (in Moore 1959, pp.
117-126). We prefer ‘dorsal furrow’ and ‘rachis’ to ‘axial furrow’ and ‘axis’ because these were the

(Palaeontology, Vol. 36, Part 4, 1993, pp. 743-769, 5 pis.]

© The Palaeontological Association

744

PALAEONTOLOGY, VOLUME 36

text-fig. 1 . Districts of Ordovician outcrop in Baltoscandia (shaded black), and the extent of subsurface and
submarine Ordovician on the Russian Platform (diagonal shading).

terms introduced by Dalman (1827) for these structures in the first systematic terminology of the
trilobite exoskeleton (see also Jaanusson 1956). An additional reason in favour of ‘rachis’ is that
it is difficult or impossible to use ‘axis’ in this sense in almost any western language other than
English. Following Jaanusson (1956), lateral glabellar lobes (L) and furrows (S) are numbered from
posterior to anterior

In many pterygometopines the frontal lobe of the glabella extends in lateral direction across the
facial suture. The portion of the lobe lateral to the facial suture is here termed ‘transsutural wing’
of the lobe. In several pterygometopines which have transsutural wings, a ridge is developed at the
cephalic margin. The ridge is most distinct anteriorly and fades gradually in posterolateral direction.
A comparable ridge is known from other phacopaceans, in particular some Devonian phacopines
(see, for example, Eldredge 1972). Ramskold and Werdelin (1991, p. 39) termed the structure
‘border ridge’. In this paper the term ‘marginal cephalic ridge’ is preferred because in
pterygometopines the former term could be confused with the lateral cephalic border, which in
many cases is ridge-like. The distinctness of the marginal ridge is in many species increased by a
furrow behind the ridge, in this paper termed the ‘admarginal cephalic furrow’.

Following Campbell’s (1977, p. 71) procedure for dalmanitids, we try to avoid ambiguous counts
of pleural ribs by counting the number of pleural furrows instead. We use the term rib only for
comparative purposes.

SYSTEMATIC PALAEONTOLOGY

Repositories. Figured specimens are housed in the Swedish Museum of Natural History, Stockholm (prefixed
RM Ar), the Institute of Geology, Estonian Academy of Sciences, Tallinn (ETAGI Tr), Museum of

JAANUSSON AND RAMSKOLD: ORDOVICIAN TRILOBITES

745

Palaeontology, Uppsala University (PMU B and PMU D), and Type Collection, Geological Survey of Sweden
(SGU).

Photography. Dorsal view of cephalon has been used as defined by Clarkson (1966) for phacopid trilobites.
Other orientations follow Whittington and Evitt (1954). Photographs are of external surfaces of the
exoskeleton, unless stated otherwise. AH specimens were painted with matt black opaque and coated lightly
with ammonium chloride prior to photography. All photographs are by the authors.

Family pterygometopidae Reed, 1905

Diagnosis. See Ludvigsen and Chatterton (1982). Attention should be paid to the fact that the
pygidium can have a semicircular outline, the number of rachial rings and pleural ribs can be as low
as three, and that in the adults of some forms no interpleural furrows can be discerned.

Subfamily pterygometopinae Reed, 1905

Diagnosis. Frontal lobe laterally strongly expanded, reaching beyond the anteromedian extent of
the visual surface of the eye. LI and L2 of about equal length. Eye bases normally surrounded
laterally by distinct subocular furrow. Pygidium with a semicircular to subparabolic outline.

Genera assigned. Pterygometopus Schmidt, 1881, Ingriops gen. nov., Oelandiops gen. nov., Estoniops Mannil,
1958, Upplandiops gen. nov., Keilapyge gen. nov., Achatella {Achatella) Delo, 1935, Achatella ( Vironiaspis )
subgen. nov.

Discussion. Schmidt (1881, p. 76) distinguished three groups of species in Pterygometopus'.

(1) Phacops ( Pterygometopus ) sclerops (Dalman) and P. (P.) trigonocephala Schmidt, each of
which was defined in a wide sense and is now known to comprise several separate species. In this
paper they are regarded as representing two genera, Pterygometopus and Ingriops gen. nov.

(2) P. (P.) panderi Schmidt, P. (P.) exilis (Eichwald) and P. ( P .) laevigata Schmidt. For this group
Mannil (1958) established the genus Estoniops. Although P. (P.) laevigatas was included, he noted
that this species belonged to a separate branch. We regard that branch as an independent, new
genus, Keilapyge.

(3) P. (P.) kuckersiana Schmidt, P. P. kegelensis Schmidt and P. (P.) nieszkowskii Schmidt. Mannil
(1958) considered these species to belong to the genus Achatella Delo, 1935.

In addition, from the Middle Ordovician of Sweden there are two new species, each belonging to
a new monotypic genus ( Oelandiops gen. nov. and Upplandiops gen. nov.).

Phacops jamesii Portlock, 1843, from the Tramore Limestone of the Republic of Ireland, does not
appear to fit into any of the above genera. Morris (1988) included the species in Estoniops but
Salter’s (1864, pi. 1, figs 39-41) figures show the frontal lobe of the glabella to be defined by a
distinct preglabellar furrow which posteriorly joins the dorsal furrow as in Pterygometopus. A very
long (exsag.) L3 appears to exclude it from the latter genus. We have not seen any material of this
species.

Pterygometopus huayinshanensis Lu, 1975, from the Neichiashan Series (stratigraphical unit
within the series not recorded) of Szechuan in central China is known only from a single,
fragmentary cephalon (Lu 1975, pi. 50, figs 6-10) in which details of several important
morphological features remain unclear. The comparatively long (exsag.) eyes recall Ingriops gen.
nov. but the genal angles are described as rounded and the development of the preglabellar furrow
is not comparable with either Ingriops or Pterygometopus. A unique feature for a pterygometopine
is the parallel-sided and posterolaterally inclined L3 which is only slightly longer (exsag.) than L2.
The species may belong to a new genus.

All these forms belong to a fairly well-defined unit within the family Pterygometopidae,
characterized by a laterally expanded frontal lobe of glabella, roughly equal length of the lateral

746

PALAEONTOLOGY, VOLUME 36

glabellar lobes LI and L2 and (with one exception) by the presence of a distinct subocular furrow.
The unit coincides with the subfamily Pterygometopinae as defined by Ludvigsen and Chatterton
(1982), but it has a much greater taxonomic diversity at the generic level than known before. The
known pterygometopines are all relatively small forms (maximum known cephalic length 16 mm).

Notes on pterygometopine morphology. The development of the anterior cephalic furrows and their topographic
relationship to the anterior branch of the facial suture vary within the subfamily.

In Pterygometopus the frontal lobe of the glabella is defined by distinct dorsal and preglabellar furrows. The
facial suture runs anteriorly and medially just in front of the preglabellar furrow (Text-fig. 2; faintly visible on
the left side in PI. 1, figs la, 2a; Whittington 1950, p. 539, fig. 3) so that the entire preglabellar furrow remains
situated on the cranidium.

text-fig. 2. Pterygometopus sclerops (Dalman, 1827). Reconstruction of cephalon in dorsal view showing course
of facial suture; based on lectotype RM Arl8074 (PI. 1, fig. 2); lower Holen Limestone, Kunda Stage, probably
lower Asaphus raniceps Zone; Husbyfjol (Vastana), Ostergotland, Sweden; x4.

In Ingriops gen. nov. (Text-fig. 3) the arrangement of the anterior cephalic furrows has the same appearance
as in Pterygometopus, but the facial suture runs in and not outside the preglabellar furrow. Thus the
preglabellar furrow is situated at the anterior margin of the cranidium.

In most other pterygometopines the original preglabellar furrow is effaced laterally. There is then no distinct
furrow to define the original lateral boundary of the frontal lobe of the glabella, although the location of the
boundary is indicated in some species. In Estoniops panderi , for example, the anterior branch of the facial
suture, in front of the dorsal furrow, is situated in a short, narrow and very shallow furrow (PI. 1, fig. 3c). In
E. maennili sp. nov. the original lateral extent of the frontal lobe is defined by a similar but still shorter furrow,
but especially by the marked difference in sculpture on either side of the facial suture (PI. 2, fig. 2). In several
other species of Estoniops, such as E. exilis (Mannil 1958, pi. 1, figs 1-6) and E. fjaeckensis sp. nov. (PI. 4, fig.
2a), a definable boundary between the original frontal lobe and the transsutural wing cannot be distinguished
in the relief of the dorsal exoskeletal surface. The transsutural wing tapers gradually posterolaterally until a
faint ridge remains which can be followed between the cephalic admarginal and border furrows almost to the
posterior branch of the facial suture (PI. 4, figs 2a, 2c). In the current terminology transsutural wings are
regarded as belonging to the frontal glabellar lobe, and in such species the lobe has an unusual extent. It should
be emphasized that the inclusion of the transsutural wings in the frontal lobe does not imply that the former
structures are homologous with any part of the frontal lobe as developed in forms in which the lobe is defined
both anteriorly and laterally by a preglabellar furrow. The use of the extended definition of the frontal lobe
is purely descriptive, necessitated by the lack of a discernible morphological boundary between the transsutural
wings and the remainder of the frontal lobe.

In several species of Estoniops, such as E. exilis (PI. 3, fig. 5; Manml 1958, pi. 1, figs 2-3, 5-6), E. fjaeckensis
sp. nov. (PI. 4, fig. 2) and E. panderi (PI. 1, fig. 3), the anterior margin of the frontal lobe is formed by an
admarginal cephalic furrow situated immediately behind the marginal cephalic ridge. In these forms the
anterior branch of the facial suture runs, at least for a short distance medially, in the admarginal cephalic

JAANUSSON AND RAMSKOLD: ORDOVICIAN TRILOBITES

747

furrow, suggesting that this portion (but not necessarily the remainder) of the furrow may be comparable to
the preglabellar furrow as developed in Pterygometopus and Ingriops.

In Keilapyge gen. nov., Upplandiops gen. nov. and several species of Estoniops (E. alifrons , E. maennili,
E. sp. nov. A) the marginal cephalic ridge and admarginal furrow are weak or absent medially, and the front of
the glabella merges into the anterior cephalic border. In some forms, such as Upplandiops calvus (PI. 3, figs lc,
4a), even the boundary between the dorsal cephalic surface and the doublure is poorly defined.

text-fig. 3. Ingriops trigonocephalus (Schmidt, 1881). Reconstruction of cephalon in dorsal view, based on
syntype RM Ar38514 (PI. 5, fig. 1); Voka Beds?, Kunda Stage (Didymograptus artus Zone); Pavlovsk, Ingria,

western Russia; x 4.

In Achatella the development of the preglabellar furrow varies from a condition comparable to
Pterygometopus to a laterally completely effaced furrow and short (tr.) transsutural wings. The variability is
discussed in the description of the genus.

Distribution. Removal of the Mediterranean forms from the pterygometopines (Henry et al. 1992) restricts the
occurrence of the early representatives of the subfamily to the Baltoscandian region and the central and
southern Ural Mountains (Antsygin 1970). They appear in the upper Arenig (Didymograptus hirundo Zone)
and show the greatest diversity in the Middle Ordovician. First from the lower Caradoc on, representatives of
the subfamily began to spread to other regions. The spread involved only a few genera: Estoniops occurs in
northern Wales and north-western England, and Achatella in south-western Scotland, eastern and central USA
(New York State, Illinois, Missouri, Ohio) and eastern Canada (Ontario and Quebec); for details see respective
genera below and for North American occurrences see Ludvigsen and Chatterton 1982. Possible new genera
are represented by Phacops jamesii Portlock, 1843 from south-eastern Ireland, and Pterygometopus
huayinshanensis Lu, 1975, from Szechuan Province of China. The latest pterygometopine has been recorded
from the Hirnantian of Scotland (Owen 1986).

Genus pterygometopus Schmidt, 1881

Type species. Calymene sclerops Dalman, 1827 (see discussion below); by subsequent designation of Bassler
(1915, p. 1065).

Other species. Phacops ( Pterygometopus ) sclerops var. angulata Schmidt, 1881; Pterygometopus bredensis
Weber, 1948. An additional species is here figured as Pterygometopus sp. nov. A (PI. 1, fig. 1).

748

PALAEONTOLOGY, VOLUME 36

Diagnosis. Preglabellar furrow distinct, joining dorsal furrow laterally. Anterior branch of facial
suture running just in front of preglabellar furrow; posterior branch situated in deep furrow. Eyes
of moderate size, anteriorly reaching the dorsal furrow. Genal angles rounded. Vincular furrow
distinct. Pygidial pleurae commonly faintly concave peripherally, with five to six pleural furrows.

Discussion. In Baltoscandia, Pterygometopus as defined in this paper has been identified routinely
as a single species, P. sclerops. Schmidt (1881) pointed out the variability of the species but preferred
to regard the variation as intraspecific. Only an especially distinctive form was distinguished as a
separate variety, var. angulatus. Subsequent to Schmidt’s (1881) monograph nobody is known to
have taken a close look at the fairly comprehensive material, including many articulated specimens.

Examination of the material of Pterygometopus for this paper disclosed that it includes several
well-defined, separate species. In the collections from the lower Holen Limestone of Ostergotland
(Sweden), the type horizon for P. sclerops , this species is not the commonest species of the genus.
The type species is especially characterized by the presence of a wide, rounded anterior continuation
of the palpebral lobe (PI. 1, fig. 2), which considerably restricts the extent of the visual surface of
the eye anteromedially. As a consequence, the anterior branch of the facial suture reaches the dorsal
furrow far anteriorly, just behind the lateral end of the frontal lobe. In the other examined species
of the genus the facial suture runs into the dorsal furrow at about the level of S3, and the visual
surface of the eye extends farther anteromedially than in P. sclerops. This, for example, is the case
with the species which is common in the Asaphus expansus Zone of Ostergotland (P. sp. nov. A;
PI. 1, fig. 1). P. sclerops differs from the other, still undescribed species by a set of additional features,
such as the relative height of the eyes, the configuration of the cephalic margin in anterior view and
the surface sculpture. All examined species of Pterygometopus have a vincular furrow (PI. 1, fig. 4;
for P. angulatus see Schmidt 1881, pi. 1, fig. 12). In P. sclerops the vincular furrow is long and
extends medially to the level of the lateral termination of the frontal lobe (PI. 1, fig. 4).

The specimen figured by Dalman (1827, pi. 2, figs 1 a-c) and refigured by Schmidt (1881, pi. 1,
figs 3 a-c, pi. 11, fig. 1), Whittington (1950, pi. 68, fig. 17, pi. 69, figs 1-3), Struve (in Moore 1959,
fig. 388: 2 a-c) and in this paper (PI. 1, fig. 2) was termed holotype by Whittington (1950). However,
when establishing the species, Dalman had several specimens at his disposal (Dalman 1827, p. 232)
and thus the correct term of the type specimen is lectotype. The other specimens figured by
Whittington (1950, pi. 68, figs 18-20) are not conspecific with the lectotype of P. sclerops.

Undoubted specimens of P. sclerops are known from the lower Holen Limestone of Ostergotland,
Narke and Vastergotland. The available material from the Kunda Stage of Oland and the Siljan
district is too fragmentary for a reliable identification at the species level. In the old collections from
Husbyfjol ( = Vastana) and several other localities in Ostergotland it is, because of similar lithology,
difficult to determine whether the specimens come from the Asaphus expansus Zone or the
lowermost Asaphus raniceps Zone. For this reason it is uncertain from which of these two units the
lectotype of P. sclerops and conspecific specimens is derived. The same uncertainty holds true also

EXPLANATION OF PLATE 1

Fig. 1. Pterygometopus sp. nov. A; lower Kunda Stage, Asaphus expansus Zone; Kungs Norrby, Ostergotland,
Sweden; 1 a-b, RM Ar 18047; dorsal and anterior views of slightly abraded cephalon with part of thorax,

x 4.

Figs 2, 4. Pterygometopus sclerops (Dalman, 1827); 2 a-b, RM Ar 18074, lectotype; lower Holen limestone,
Kunda Stage, probably lower Asaphus raniceps Zone; Husbyfjol (Vastana), Ostergotland, Sweden; dorsal
and anterior views of enrolled exoskeleton, x 3. 4, RM Ar55055 ; Sphaeronites Beds, Kunda Stage, lowermost
Asaphus raniceps Zone; Raback quarry, Kinnekulle, Vastergotland, Sweden; ventral view of part of
cephalon showing vincular furrow, x 5.

Fig. 3. Estoniops panderi (Schmidt, 1881); Aseri Stage (lower Didymograptus murchisoni Zone); Pavlovsk,
Ingria, western Russia; 3 a-c. RM Ar38516; anterior, lateral and dorsal cephalic views of a syntype, enrolled
exoskeleton, x 3.

PLATE 1

JAANUSSON and RAMSKOLD, Pterygometopus, Estoniops

750

PALAEONTOLOGY, VOLUME 36

with regard to specimens from Narke. However, on Kinnekulle in Vastergotland P. sclerops occurs
in the Sphaeronites Bed (RM Ar55055-Ar55056) which is within the lowermost A. raniceps Zone,
and specimens of the genus collected by J. W. Dalman in 1827 at Ulunda brook on Billingen (RM
Arl5507-Arl5510) are probably from beds of the same age. This indicates that the lectotype of
P. sclerops may be from the lower part of the Asaphus raniceps Zone rather than from the Asaphus
expansus Zone.

The Swedish material of Pterygometopus discussed above is from the lower and middle Kunda
Stage. In Ingria, Lamansky (1905) recorded P. sclerops only from the middle and upper Volkhov
Stage (corresponding to the Megistaspis simon and M. limbata Zones in the current Swedish
biostratigraphic terminology). Ingrian specimens of Pterygometopus from the Volkhov Stage which
were available for examination belong to P. angulatus Schmidt, 1881. This species differs clearly
from P. sclerops and other forms from the Kunda Stage by the characteristically arched anterior
cephalic margin and other features. On northernmost Oland Pterygometopus is not uncommon in
the Megistaspis limbata Zone but the material is fragmentary. However, from Halludden there is a
cephalon (RM Ar55010; 0-80-0-85 m below the upper boundary of the M. limbata Zone) which
shows the characteristic features of P. angulatus.

Occurrence. Middle and Upper Volkhov Stage (Didymograptus hirundo Zone) of Ingria and northern Estonia,
Upper Volkhov Stage of Sweden (Oland). Lower and Middle Kunda Stage ( Didymograptus artus Zone) of
Ingria, northern Estonia, Sweden and the Oslo Region of Norway (Brogger 1882). Arenig calcareous sandstone
of the Bredy region, eastern slope of southern Ural mountains (P. bredensis\ Weber 1948; Antsy gin 1970).

Genus ingriops gen. nov.

Derivation of name. From Ingria, the latinized form of the old name for the district in Russia between Estonia
and Lake Ladoga; Inkeri in Finnish, Ingermanland in Swedish. Gender masculine.

Type species. Phacops ( Pterygometopus ) trigonocephala Schmidt, 1881.

Other species. Phacops (Pterygometopus) trigonocephala var. intermedia Schmidt, 1881; Phacops (Pterygo-
metopus) trigonocephala var. estonica Schmidt, 1881; Phacops (Pterygometopus) trigonocephala var. genuina
Schmidt, 1881.

Diagnosis. Preglabellar furrow distinct, joining the dorsal furrow laterally. Anterior branch of facial
suture running in the preglabellar furrow; posterior branch situated in a deep furrow. Eyes fairly
long, 46 per cent of cephalic length in type species. Fixed cheeks with genal spines. Vincular furrow
distinct. Pygidium with a subparabolic outline, pleural areas with ten or eleven pleural furrows.

Discussion. In Schmidt’s concept, the array of forms which in this paper is included in Ingriops
gen. nov. constituted a single variable species. He recognized that this ‘species' is not taxonomically
homogeneous by distinguishing various forms as varieties. Most, if not all, of these varieties
represent separate species.

Lamansky (1905) recorded Pterygometopus trigonocephalus in the Ingrian sequence from the
Asaphus expansus and Asaphus raniceps Zones, and suggested that both var. estonicus and var.
genuinus from northern Estonia were younger, derived from the upper Kundan Asaphus eichwaldi
( = A. sulevi) Zone. Schmidt (1881) reported P. trigonocephalus also from Husbyfjol in Ostergotland.
In the Riksmuseum collections there is a fragmentary cephalon (RM Arl8016) of Ingriops from the
lower Holen Limestone of this locality, on the accompanying label identified by F. Schmidt as
' P. trigonocephala' , but in Ostergotland the genus appears to be very rare. A distinctive pygidium of
Ingriops has been found in the topmost beds of the Kunda Stage ( Megistaspidella gigas Zone) at
Ljung in Ostergotland (RM Arl8077).

JAANUSSON AND RAMSKOLD: ORDOVICIAN TRILOBITES

751

At first sight the cephalon of Ingriops appears to be fairly similar to that of Pterygometopus. The
conspicuous differences are the presence of genal spines and the larger eyes (in the eye of an adult
Ingriops trigonocephalus there are 32 files of up to 13 lenses; PI. 5, fig. 1</). A close examination
reveals that in Ingriops the facial suture runs in and not in front of the preglabellar furrow (see also
Schmidt 1881, pi. 1, fig. 15c). The vincular furrow in I. trigonocephalus was figured by Schmidt
(1881, pi. 1, fig. 12).

The pygidium of Ingriops has a distinctive appearance. It is comparatively longer than in
Pterygometopus , with a subparabolic outline and provided with far more numerous, flattened
pleural ribs.

Occurrence. Kunda Stage (Didymograptus artus Zone) of Ingria, northern Estonia and the province of
Ostergotland in Sweden.

Genus oelandiops gen. nov.

Derivation of name. From the Island of Oland, latinized Oelandia. Gender masculine.

Type species. Oelandiops mirificus gen. et sp. nov.

Other species. The genus is monotypic.

Diagnosis. Frontal glabellar lobe with short (tr.; transsutural wings. Marginal cephalic ridge and
admarginal cephalic furrow well-defined. Anterior branch of facial suture reaches anteromedially
admarginal furrow. Lateral glabellar lobes transverse. Posteromedian part of fixigena expanded
abaxially into long (exsag.), wing-like structure overhanging dorsal furrow from SO to S2. Genal
angles rounded. Pygidium strongly convex with steeply sloping outer portion of the pleural areas
and the postrachial area; pleural area concave along margin, with eight or nine deep pleural
furrows, interpleural furrows distinct. Dense tuberculation on thorax and pygidium.

Discussion. Oelandiops gen. nov. is in many respects so different from the other pterygometopines
that a detailed discussion of the distinguishing features is not necessary. It resembles Estoniops ,
especially E. panderi , in the extension of the frontal lobe laterally across the facial suture, but the
transsutural wings of the lobe are short. The transverse orientation of LI and L2 resembles that of
Upplcindiops gen. nov., but in other respects these two genera are very different. The pygidium is
similar to that of Pterygometopus or Estoniops panderi but the postrachial area is steeply sloping.

Occurrence. As for the type species.

Oelandiops mirificus sp. nov.

Plate 2, fig. 1 a-g

Derivation of name. Latin mirificus , amazing, astounding.

v.1960 Pterygometopinae n. gen. n. sp.; Jaanusson, pp. 225, 279.

Holotype. Complete enrolled exoskeleton (PI. 2, fig. ltf-g), RM Ar47921, Binnerback, northern Oland.
Segerstad Limestone, Zone of Illaenus planifrons (upper Aseri Stage; lower Didymograptus murchisoni Zone).
No other specimens are known.

Diagnosis. As for genus.

Description. Glabella fairly flat between the palpebral lobes, cephalon steeply sloping anteriorly and laterally.
Cephalic length (sag.) about 55 per cent of width. Lateral cephalic border highest along lateral margin of free

752

PALAEONTOLOGY, VOLUME 36

cheek, becoming gradually lower and less distinct anteromedially, fading posteriorly at level of posterior
border furrow. Admarginal furrow apparently shallow medially (specimen broken), deepening considerably
posterolaterally until merging with lateral border furrow somewhat in front of level of mid-length of eye.

Dorsal furrow deep, trench-like between eye and frontal lobe, widening along L3, tunnel-like between S2 and
SO, there overhung by both the lateral extremities of L2 and LI and the flat, wing-like medial projection of the
fixed cheek. Frontal lobe of glabella extends laterally beyond the facial suture, but transsutural wing of lobe
is fairly short and low, tapering in posterolateral direction until becoming gradually obsolete in front of level
of mid-length of eye. Axes of lateral glabellar lobes almost transversely directed. S3 meets dorsal furrow slightly
posterior to level of anterior margin of eye, fairly shallow, gently posteromedially curved. Maximum length of
L3 (exsag.) about twice that of L2. Lateral edge of L3 fairly strongly convex outward, surface of posterior
portion gently downwards sloping, posterolateral corner of lobe pointed (PI. 2, fig. lg). S2 deep, widening
adaxially. L2 and LI with flat dorsal surface which slopes slightly upward laterally (PI. 2, fig. lg), their rounded
outer margins almost reaching flat, exsagittally and adaxially expanded, wing-like extension of fixed cheek. SI
bifurcates adaxially, with small, posteromedian branch defining inner margin of lateral node of LI (PI. 2, fig.
1(>). Occipital ring laterally with an incipient, transversely directed furrow.

Palpebral lobes not quite reaching level of glabella in anterior view (PI. 2, fig. Id). Palpebral lobe extends
anteriorly to dorsal furrow slightly in front of SI: Palpebral furrow deep. Eyes set wide apart; distance between
inner margins of palpebral lobe anteriorly 1-6-F7 times width (tr.) of occipital ring. Eye length (exsag.) about
35 per cent of cephalic length (sag.). Visual surface with 25 files of up to 10 lenses; best preserved eye has lens
formula (from anterior) 456 788 989 9910 999 988 888 765 3. Subocular furrow deep, surrounded outwards by
prominent rim which slightly overhangs adjacent field of free cheek (PI. 2, fig. Id). Anterior branch of facial
suture running from eye exsagittally into dorsal furrow, ascends frontal lobe and continues in a wide arch,
reaching admarginal furrow medially. Adaxial portion of posterior branch of facial suture situated in a furrow,
lateral portion runs roughly parallel to posterior border furrow.

Only parts of doublure accessible for examination (PI. 2, fig. le); no vincular furrow. Hypostome unknown.

Thorax of eleven segments. Inner part of pleurae horizontal, outer part almost vertically sloping (PI. 2, fig. \e).
Rachial rings laterally with faint, transversely directed furrow (PI. 2, fig. lb, e). Pleural furrows deep, both
anterior and posterior pleural bands relatively narrow and high.

Pygidium 1-7 times wider than long (excluding articulating half ring); width about 64 per cent and length
about 67 per cent of respective cephalic dimensions. Width (estimated) of first rachial ring 37-39 per cent of
pygidial width. Rachis fairly strongly convex (tr.), with seven rings well defined by apodeme-carrying inter-ring
furrows, and a short posterior piece with a weakly discernible additional ring. Only inner pleural area visible
in dorsal view (PI. 2, fig. le), outer areas and postrachial area steeply to vertically sloping; border narrow but
distinct, outwardly flexed. Pleural areas with eight pairs of deep pleural furrows; an additional pair of furrows
on the postrachial area may be a ninth pair; furrows become obsolete at inner margin of border. Five distinct
interpleural furrows curving strongly backwards distally to reach the succeeding pleural furrow (PI. 2, fig. It).
Between pleural ribs of eighth pair is a median unpaired, bilaterally concave, raised band.

Sculpture which is not well preserved, consists of fairly coarse, apparently simple tubercles on all glabella,
adaxial parts of fixed cheeks, free cheeks, thoracic and pygidial rachial rings and most of pleurae (except
outermost parts). Cephalic marginal ridge and doublure with a fine granulation; such granulation may have
extended to other parts of the exoskeleton but is not preserved. Pygidial border smooth.

Occurrence. As for the holotype.

EXPLANATION OF PLATE 2

Fig. 1. Oelandiops mirificus gen. et sp. nov.; Segerstad Limestone, Zone of Illaenus planifrons (upper Aseri
Stage; lower Didymograptus murchisoni Zone); Binnerback, northern Oland, Sweden; 1 a-g, RM Ar47921,
holotype, enrolled exoskeleton; right lateral, dorsal cephalic, left lateral, anterior cephalic, dorsal pygidial,
and posterior pygidial views, and detail of cephalon in oblique anterodorsal view, 1 a-f x 3, lg x 5-3.

Figs 2-4. Estoniops maennili sp. nov. ; 2-3, Blidene Marl (lower Dicranograptus clingani Zone); western Latvia.
2, ETAGI Tr 2390, holotype; Adze boring (882-9 m); dorsal view, x 4. 3, ETAGI Tr 2368; Engure boring
(932-9 m); dorsal view of large cephalon, x 3. 4, uppermost Skagen Limestone (lower D. clingani Zone);
Kinnekulle, Mossen section, Vastergotland, Sweden; RM Ar55054; dorsal view of latex cast of internal
mould of cephalon, x 4.

PLATE 2

JAANUSSON and RAMSKOLD, Oelandiops, Estoniops

754

PALAEONTOLOGY, VOLUME 36

Genus estoniops Mannil, 1958

Type species. Acaste exilis Eichwald, 1858 from the Kukruse Stage ( Nemagraptus gracilis Zone) of northern
Estonia; by original designation.

Discussion. Antsygin (1970, p. 17) designated the cephalon ETAGI Tr. 1902 figured by Mannil
(1958, pi. 1, figs 1-3) as the lectotype of E. exilis. The designation is not valid because this cephalon
was not available to Eichwald (1858) when he described the species. The designation of a type
specimen must await a revision of Eichwald’s material.

Other species. Phacops ( Phacops ) alifrons M‘Coy, 1851; Phacops ( Pterygometopus ) panderi Schmidt, 1881;
Phacops sandbyensis Olin, 1906; Estoniops bekkeri Mannil, 1958; Estoniops oculeus Antsygin, 1970; Estoniops
fjaeckensis sp. nov.; Estoniops maennili sp. nov.

The examined material includes several additional species of Estoniops, such as the specimens figured as
Phacops exilis by Wiman (1908, pi. 7, figs 1-5) from the Dalby Limestone of the erratics in Uppland derived
from the South Bothnian submarine Cambro-Ordovician sequence (called E. sp. nov. B below), E. sp. nov. A
(PI. 3, figs 2, 7) from the Lasnamagi Stage of the File Haidar boring of Gotland, and a species represented
by a fragmentary cranidium (PMU OI. 1010) from the Folkeslunda Limestone of the same stage of northern
Oland (referred to as Estoniops sp. or E. aff. panderi by Jaanusson 1960, pp. 226, 278). A somewhat deformed
cephalon (Paleontologisk Museum Oslo, no. 69526) from the lower part of the Nakkholmen Mudstone of
Asker in the Oslo region (western side of the first tunnel at the Billingstad station) obviously belongs to an
additional new species of Estoniops.

Diagnosis. Frontal lobe of glabella laterally not defined by a continuous furrow; it either extends
across the facial suture into transsutural wings or is defined along the facial suture by a distinct
change in sculpture. L3 triangular, much longer (exsag.) than L2; axis of L2 distinctly
posterolaterally inclined. Eyes of medium size to fairly large, anterior end of palpebral lobe
normally reaching dorsal furrow. Genal angles rounded. Pygidium with roughly semicircular
outline, pleural area commonly evenly convex but may be faintly concave peripherally, with five to
eight pleural furrows.

Discussion. Definition of the generic characters of Estoniops is rendered difficult by the existence of
several species which show somewhat unusual features but which are poorly known either because
of a poor state of preservation or inadequate illustrations.

The cephalon of the type species was adequately figured by Mannil (1958, pi. 1, figs 1-6). Juvenile
specimens (PI. 3, fig. 5a-c) have a conspicuously shorter L3 than in adult cephala (Mannil 1958,
pi. 1, fig. 1) and approach in this respect the condition in Pterygometopus. The pygidium (PI. 3,
figs 8-9; PI. 4, fig. 7) has four well-defined rachial rings and a posterior portion with one or two faint
rings; pleural areas are evenly convex and provided with five pairs of faintly furrowed ribs of which
the most posterior pair is indistinct.

Opik (1937, p. 73) pointed out that in E. exilis the palpebral furrow unites directly with the dorsal
furrow, that is, the anterior end of the palpebral lobe extends to the dorsal furrow. In Estoniops this
is normally the case. The specimens from the upper Middle Ordovician Chergyn Stage of the
Nizhnie Sergi district, western slope of the central Ural mountains, which Antsygin (1970, pi. 4,
figs 7-16) identified as E. exilis, differ by having the anterior end of the distinctly smaller eyes situated
some distance from the dorsal furrow (Antsygin 1970, pi. 4, figs 7-8). Other differences include a
more strongly posterolaterally inclined L2, a shorter (sag.) occipital ring and, especially, the
continuation of the pleural furrows to the pygidial margin. In Estoniops there is normally a distinct,
narrow border on the pygidium. The absence of a border in combination with the other differences
casts some doubt on the generic affinity of the Uralian species, but the material is not well preserved
and the quality of the photographs not sufficient for safe conclusions. Estoniopsl angulatus
Antsygin, 1970 (pi. 5, figs 17-18) from the same stage of the central Ural mountains seems to be a
related species of the same doubtful generic affinity.

JAANUSSON AND RAMSKOLD: ORDOVICIAN TRILOBITES

755

The earliest species of Estoniops, E. panderi (Schmidt, 1881) from the Aseri Stage (lower
Didymograptus murchisoni Zone) of Ingria and Estonia (PI. 1, fig. 3 a-c), shows some
Pterygometopus-like features. It has a relatively short (exsag.) L3, the posterior branch of the facial
suture runs in a furrow, the cephalic dorsal furrows are very deep, and the pygidium is
Pterygometopus-Mke with peripherally distinctly concave pleural areas (PI. 1, fig. 3a). In the visual
surface there are 21 files of up to 9 lenses.

The specimens from the upper Middle Ordovician Chergyn Stage of the Nizhnie Sergi district,
western slope of the central Ural mountains, that Antsygin (1970, pi. 5, figs 1-5) identified as
E. panderi (a much earlier species) cannot possibly be conspecific with the latter. They have no
admarginal cephalic furrow visible in dorsal view, a long and triangular L3, a shallow S3, evenly
convex pygidial pleural areas, and a smaller number of both rachial rings and pleural ribs in the
pygidium.

An advanced type of Estoniops appeared early, represented in the material by E. sp. nov. A. The
somewhat fragmentary cephalon (PI. 3, fig. 2), from the Lasnamagi Stage (upper Didymograptus
murchisoni Zone) of the subsurface of Gotland, appears to lack a marginal cephalic ridge and an
admarginal cephalic furrow, L3 is relatively long, and the associated pygidium (PI. 3, fig. 7) has only
four pleural furrows (plus a hint of a fifth) and virtually effaced interpleural furrows.

In Estoniops as defined in this paper the development of the marginal cephalic ridge and the
admarginal cephalic furrow varies. The ridge is distinct in E. exilis (PI. 3, fig. 5c; Mannil 1958,
pi. 1, figs 2, 5), E. fjaeckensis sp. nov. (PI. 4, fig. 2b), and in E. panderi (PI. 1, fig. 3), somewhat less
distinct in the cephalon figured as E. exilis by Wiman (1908, pi. 7, fig. 1 ; E. sp. nov. B in this paper).
In several species, such as E. alifrons (Whittington 1962, p. 17), E. maennili sp. nov. and probably
also in E. sp. nov. A (PI. 3, fig. 2), the furrow is obsolete, and no distinct marginal cephalic ridge
is developed. Due to the pour quality of illustrations or insufficient state of preservation the
development of the marginal cephalic ridge and the admarginal furrow is unclear in most species
of Estoniops described by Antsygin (1970).

Occurrence. The known vertical range of Estoniops is from equivalents of the lower Didymograptus murchisoni
Zone (Aseri Stage) to the lower Dicranograptus clingani Zone (Blidene Marl, Skagen Limestone, Upper
Llongvillian). Baltoscandia (Ingria, Estonia, Latvia, Sweden, Oslo region in Norway); Wales, Bala district
(lower Bala series, uppermost Gelli-grin Group; E. alifrons in Whittington 1962); northern England, Cross
Fell Inlier (Upper Longvillian; E. alifrons in Dean 1962); western slope of the central Ural mountains (Chergyn
Stage: E. panderi in Antsygin 1970; Tupyl Stage: E. oculeus Antsygin, 1970).

Estoniops fjaeckensis sp. nov.

Plate 4, figs 2, 6

Holotype. A cephalon (PI. 4, fig. 2 a-c), RM Ar53596, Fjacka section, Siljan district, Skagen Limestone, L6 m
from the lower boundary.

Other material. Four pygidia from the Fjacka section (RM) which presumably belong to E. fjaeckensis, but are
all juvenile (PI. 4, fig. 6). No distinct differences from those of E. exilis can be observed.

Diagnosis. Admarginal cephalic furrow and transsutural wings of relatively weakly convex frontal
lobe of glabella extend posterolaterally almost to the posterior branch of the facial suture. Eyes
fairly long, about 47 per cent of the cephalic length; comparatively low. Occipital ring comprises
about 19-20 per cent of cephalic length. Outer portion of posterior branch of facial suture converges
laterally with posterior border furrow. Tubercles composite, with superimposed granulation.

Description. The species is similar to the much older E. exilis', emphasis is therefore put on distinguishing
characters.

Cephalon is represented only by the holotype. The configuration of the cephalon in front of the eyes is similar
to that of E. exilis except that the frontal lobe of the glabella is less convex. Transsutural wing of the frontal

756

PALAEONTOLOGY, VOLUME 36

lobe long, tapering posterolaterally into a weak ridge which reaches the posterior branch of the facial suture
(PI. 4, fig. 2c). The extent of the distinct admarginal furrow coincides with that of the frontal lobe. Eyes longer
(about 47 per cent of cephalic length) than in E. exilis (fairly consistently 44 per cent of cephalic length) and
comparatively lower. Visual surface of 19 files of up to 7 or 8 lenses, lens formula (from anterior) approximately
456 778 777 777 666 554 3 (some areas damaged). Anterior end of palpebral lobe extends to dorsal furrow
somewhat in front of level of S3 (PI. 4, fig. 2a), as in juvenile E. exilis (PI. 3, fig. 5 a) but farther anteriorly than
in adults of that species (Mannil 1958, pi. 1, figs 1, 4). This holds true also for the level at which the facial suture
reaches the dorsal furrow. Occipital ring fairly long (19-20 per cent of cephalic length) compared to E. exilis
(16 per cent of cephalic length) and rather wide (tr. ). The posterior branch of the facial suture turns first almost
straight anterolaterally, then curves fairly abruptly in posterolateral direction and continues straight to the
lateral border furrow, converging with the course of the posterior border furrow (PI. 4, fig. 2c). In E. exilis the
outer portion of the facial suture runs almost straight in lateral direction parallel to the posterior border furrow
(PI. 3, fig. 5b', Mannil 1958, pi. 1, figs 1, 3, 6). In both E.fjaeckensis and E. exilis the glabella carries fairly coarse
tubercles of various sizes, but in E. exilis the tubercles are simple whereas in E.fjaeckensis they are composite,
with a fine granulation encroaching onto or covering the tubercles.

Occurrence. Lower Skagen Limestone (probably equivalent to the topmost Diplograptus multidens Zone).
Sweden, Siljan district, Fjacka section.

Estoniops maennili sp. nov.

Plate 2, figs 2-4; Plate 4, fig. 4

v.1968 Estoniops cf. alifrons (M’Coy); Mannil et al., p. 89.

Derivation of name. After Dr Ralf Mannil who collected most of the Latvian material and recognized the
affinities of the species.

Holotvpe. Cephalon (PI. 2, fig. 2) ETAGI Tr. 2390, from the Blidene Marl, Adze boring (882-9 m), western
Latvia.

Other material. Four cephala and two pygidia from borings in western Latvia (ETAGI); two cephala, three
cranidia, and two pygidia from Vastergotland (RM) (see ‘Occurrence’ below).

Diagnosis. Cephalon in front of eyes moderately convex, without a distinct marginal cephalic ridge.
Frontal glabellar lobe laterally defined by distinct change in sculpture at facial suture. Eyes
comparatively low, about 39^40 per cent of cephalic length; palpebral lobe narrow anteriorly.
Occipital ring with a low median tubercle. Pygidial pleural areas with seven to eight pleural furrows,

EXPLANATION OF PLATE 3

Figs 1, 3—4, 6. Upplandiops calvus gen. et sp. nov.; Furudal Limestone, Uhaku Stage, Hustedograptus
teretiusculus Zone); Uppland province, Sweden; 1 a-c, RM Ar55040, holotype; erratic boulder, Estuna;
complete, slightly disarticulated exoskeleton, dorsal and anterior cephalic, and exterior pygidial views, x 6.
3, RM Ar55042; erratic boulder, Bergsbrunna 1; dorsal view of posterior part of articulated exoskeleton,
central part of pygidial rachis apparently not abraded, x 6. 4 a-c, RM Ar55041 ; erratic boulder, Bergsbrunna
1 ; anterior, lateral, and dorsal views of cephalon, x 6. 6 a-b, UM B580; erratic boulder no. 4; Borstil parish,
Hoganas, Torron; anterior, and ventral views of incomplete cephalon, x 6.

Figs 2, 7. Estoniops sp. nov. A; Lasnamagi Stage (Upper Didymograptus murchisoni Zone); File Haidar boring
(level 337-35—337-49 m), Gotland, Sweden; 2, SGU Type 8429a; dorsal view of cephalon, x 6. 7, SGU Type
8429 b \ dorsal view of pygidium, x 4.

Figs 5, 8-9. Estoniops exilis (Eichwald, 1858); Viivikonna Formation, Kivioli Member, Kukruse Stage
(Nemagraptus gracilis Zone); northeastern Estonia. 5 a-c, RM Ar51262; Kukruse; dorsal, lateral, and
anterior views of juvenile cephalon, x 10. 8, RM Ar55066; Kohtla-Jarve; dorsal view of incomplete
pygidium, x 5. 9, RM Ar55065; Kiittejou; exterior view of pygidium, x4.

PLATE 3

JAANUSSON and RAMSKOLD, Upplandiops , Estoniops

758

PALAEONTOLOGY, VOLUME 36

eighth furrow very weak. Surface sculpture of cranidium includes fairly closely spaced tubercles
with superimposed granules.

Description. E. maennili sp. nov. is in many respects similar to E. alifrons, described in detail by Whittington
(1962), and therefore emphasis is put on the features which differ between these two species or which were not
mentioned in the description of E. alifrons.

E. maennili attains a size which is large for Estoniops', a cranidium from Vastergotland reaches a length (sag.)
of 16 mm, and the largest known cephalon from Latvia (PI. 2, fig. 3) is 14 mm long. In material from both
Latvia and Vastergotland, frontal lobe of glabella is moderately and fairly evenly convex. No admarginal
cephalic furrow observed, nor any distinct marginal cephalic ridge. S3 curves posterolaterally just before
reaching the dorsal furrow. Posterolateral margin of frontal lobe between S3 and facial suture slightly but fairly
distinctly convex, especially in large specimens. Posterolateral boundary of frontal lobe also indicated by an
incision or short furrow at facial suture, apparently a remnant of the preglabellar furrow. Remainder of lateral
boundary of frontal lobe defined by clear difference in sculpture on either side of facial suture. Frontal lobe
between anteriorly directed branches of the facial suture with closely spaced tubercles; lateral to the suture the
sculpture consists solely of fine granulation. Lateral cephalic border rounded, anteromedian portion expanded
so that when reaching frontal glabellar lobe the border forms a lateral continuation of that lobe with regard
to width (exsag.) and convexity. Expanded portion of cephalic border resembles transsutural wings but there
is no clear boundary between that portion and remainder of border. L2 more strongly posterolaterally inclined
than in other species of Estoniops , excepting E. alifrons. Occipital ring with a low median tubercle. Posterior
branch of facial suture runs in a furrow lateral to the eyes, deepest medially, gradually shallowing laterally. All
these morphological details can be observed also in E. alifrons (Whittington 1962, pi. 3, fig. 6).

Distance (exsag.) between posterior margin of eye and posterior cephalic margin about 38-39 per cent of
length of eye. The eyes appear to be fairly low, although because of slight deformation caused by compaction,
their precise height is difficult to determine. Visual surface carries 22 files of up to 6 or possibly 7 lenses.
Anterior extension of the palpebral lobe distinct in anterior view but very narrow. Subocular furrow well
defined.

Sculpture of glabella consists of closely spaced tubercles of somewhat varying size, between and on which
is a fine granulation extending also to the borders. Genal field with irregular pits.

Anterior six inter-ring furrows of pygidial rachis deep, with apodemes. Posterior portion carries fairly
distinct seventh inter-ring furrow and indistinct eighth. Inner portion of pygidial pleural areas fairly flat, outer
portion comparatively steeply sloping, with a weak indication of a peripheral concavity (PI. 4, fig. 4b). Seven
deep pleural furrows define six flat ribs; eighth pleural furrow faintly discernible. Interpleural furrows faint,
observed on specimens both from Latvia and Vastergotland; not recognizable on internal moulds. Surface of
whole pygidium with fine granulation, densest along the outer margin (PI. 4, fig. 4b).

Discussion. Specimens, preserved as internal and external moulds, from mudstone intercalation of
the uppermost Skagen Limestone of Kinnekulle (PI. 2, fig. 4) and northern Mosseberg in

EXPLANATION OF PLATE 4

Figs 1, 3, 5. Keilapyge laevigata (Schmidt, 1881); Johvi and Keila Stages (uppermost Diplograptus multidens to
lowermost Dicranograptus clingani Zones); northern Estonia. 1 a-d, ETAGI Tr 1488; Ristna Beds of the
Keila Stage; Paaskiila; dorsal, anterior, lateral, and oblique posterior views of cephalon; la, x 5; Ib-d, x 4.
3 a-b, RM Ar54513; horizon and locality as fig. 1 ; dorsal, and anterior views of partly exfoliated cephalon,
x 5. 5 a-b, RM Ar55039; Johvi Stage; Aluvere; exterior, and posterior views of pygidium, x 4.

Figs 2, 6. Estoniops fjaeckensis sp. nov.; Fjacka section, Siljan district, Sweden. 2 a-c, RM Ar53596; Skagen
Limestone, 1-6 m above lower boundary; dorsal, anterior, and lateral views of holotype cephalon, x 5.
6, RM Ar53597; uppermost Dalby Limestone, bed 12 of the bentonite sequence; dorsal view of small
pygidium, x 13.

Fig. 4. Estoniops maennili sp. nov.; Blidene Marl (lower Dicranograptus clingani Zone); Blidene boring
(892-893 m), western Latvia. 4 a-c, ETAGI Tr 3610; lateral, posterior, and exterior views of pygidium, x 6.

Fig. 7. Estoniops exilis (Eichwald, 1858) ; Viivikonna Formation, Kivioli Member, Kukruse Stage ( Nemagraptus
gracilis Zone); Kukruse, northeast Estonia; RM Ar50492; exterior view of small pygidium, x 9.

PLATE 4

JAANUSSON and RAMSKOLD, Keilapyge , Estoniops

760

PALAEONTOLOGY, VOLUME 36

Vastergotland agree in all observable details with the specimens from the Blidene Marl of western
Latvia.

As recognized by Ralf Mannil, E. maennili is very similar to the roughly contemporary E. alifrons.
Both species are large not only for the genus but for the whole subfamily, the sculpture is practically
identical, and both have a low occipital tubercle. Neither species has a marginal cephalic ridge or
an admarginal cephalic furrow. In both species the posterior branch of the facial suture is associated
with a furrow, and both have a fairly steeply posterolaterally inclined L2. These two species belong
to a separate group within the genus, distinguished especially by the morphology of the frontal
glabellar lobe and the rounded lateral cephalic border.

E. alifrons has a well-defined, broad anterior extension of the palpebral lobe (Whittington 1962,
p. 17, pi. 3, figs 6, 8). In E. maennili this structure is very narrow, and the visual surface of the eye
extends farther medially. E. alifrons has higher eyes. In E. maennili there are 22 lens files of up to
6 or possibly 7 lenses, while in E. alifrons there are up to 11 or more lenses per file (exact
number of files unknown but, as judged from illustrations, apparently close to that in E. maennili).
Specimens of E. alifrons from both the Bala district and the Cross Fell Inlier (UK) are deformed to a
varying degree, and this makes a detailed comparison with the only slightly compressed specimens
from the Blidene Marl and uppermost Skagen Limestone difficult. The impression from the figured
specimens identified as E. alifrons is that the species has a distinctly more steeply sloping frontal lobe
and cheeks than in E. maennili.

E. sandbyensis Olin (1906, pi. 1, fig. 7) from the Sularp Shale (Diplograptus multidens Zone) of
the Fogelsang district in Scania is represented by a single, somewhat compressed, fragmentary
cephalon (Geological Institute, Lund Univ. LO 1901T), preserved mainly as internal mould.
Without access to further material it is difficult to define the species. The genal angles are not
preserved, and there is no evidence for the genal spines indicated by Olin (1906, pi. 1, fig. 7) on his
drawing. A point of difference from E. maennili is the sculpture which in E. sandbyensis appears to
consist of much more sparsely scattered large tubercles than in the former species.

The roughly contemporary E.fjaeckensis sp. nov. from the Skagen Limestone of the Siljan district
has both a distinct marginal cephalic ridge and an admarginal cephalic furrow and is also otherwise
clearly different.

Occurrence. Blidene Marl (lower Dicranograptus clingani Zone) of western Latvia : Adze boring (882-9 m),
Blidene boring (892-893 m, 893-2 m), Engure boring (932-9 m, 934 0 m), Remte boring (1036-6-1038 m).
Uppermost Skagen Limestone (lower D. clingani Zone), Vastergotland, Sweden: Kinnekulle, Mossen section;
Mosseberg, Jonstorp section.

Genus upplandiops gen. nov.

Derivation of name. After Uppland (latinized Upplandia), the province in central Sweden where the type species
is fairly common in erratic boulders of the Furudal Limestone, derived from the South Bothnian submarine
Cambro-Ordovician sequence. Gender masculine.

Type species. Upplandiops calvus gen. et sp. nov.

Other species. The genus is monotypic.

Diagnosis. Anterior cephalic margin broadly rounded in section, without marginal ridge. Axes of
glabellar lobes transversely directed, L3 only slightly longer (exsag.) at dorsal furrow than LI and
L2. Eyes fairly large. Genal angles rounded. No vincular furrow. Ten thoracic segments. Pygidium
small relative to size of cephalon, posterior margin broadly rounded. Three pairs of laterally well
defined but medially obsolete rachial rings and a posterior portion; pygidial pleural areas weakly
convex, with three pairs of short (tr.), shallow pleural furrows.

Discussion. The configuration of the lateral glabellar lobes of Upplandiops resembles that of
Oelandiops gen. nov., but otherwise these two genera are very different. In O. mirificus the dorsal

JAANUSSON AND RAMSKOLD: ORDOVICIAN TRILOBITES

761

furrow is wide and deep opposite L3, whereas in U. calvus the corresponding sector is constricted
and the dorsal furrow runs in a tunnel below a bridge formed by L3 and the anterior part of the
palpebral lobe. In anterior view the cephalon of Upplandiops is somewhat similar to that of some
species of Estoniops that lack a marginal cephalic ridge, but in a strictly dorsal view the frontal lobe
of the glabella in Upplandiops does not give the impression of being extended across the facial suture
into lateral wings, because at the facial suture the boundary between the frontal lobe and the lateral
cephalic border is marked by a change in convexity.

The genus, or al least the type species of the genus, is especially remarkable in having only ten
thoracic segments. In the suborder Phacopina the normal number of thoracic segments is eleven,
and Struve (in Moore 1959) stated this number of segments to be diagnostic for the whole suborder.
However, a few exceptions have been recorded. Maksimova (1957) reported ten thoracic segments
in a species of Isalaux, known only from a single specimen. Another species of Isalaux has eleven
segments (Frederickson and Pollack 1952). Ten segments were reported also in the single complete
exoskeleton of a species of Isalaux (Isalauxina) by Maksimova (1962), a number confirmed by
material figured by Semenova (1984). Isalaux is a pterygometopid genus that cannot at present be
assigned to a subfamily (Ludvigsen and Chatterton 1982). A further pterygometopid with ten
thoracic segments is the type species of the eomonorachine genus Liocnemis , L. recurvus
(Linnarsson, 1869; see Kielan 1960). Since the above genera are not closely related, the reduction
in the number of thoracic segments to ten must have taken place independently in several different
lineages.

Upplandiops has, relatively, the smallest pygidium among pterygometopines. In U. calvus its
length is only about 65 per cent that of the cephalon. In Estoniops the corresponding figure ranges
from some 78-79 per cent (E. exilis ; Schmidt 1881, pi. 1, fig. 20) to about 85 per cent ( E . panderi).
The difference in width is still more pronounced: in E. panderi the pygidial width is about 75-76
per cent of the cephalic width, in U. calvus only 52-53 per cent. Moreover, the pygidium of
Upplandiops has only a few, weakly marked pleural furrows.

Occurrence. As for the type species.

Upplandiops calvus sp. nov.

Plate 3, figs 1, 3-4, 6

v.1960 Estoniops n. sp.; Jaanusson, pp. 234, 279.
v.1963 Estoniops n. sp.; Jaanusson, pp. 21, 29, 37.
v.1966 Estoniops n. sp.; Mannil, fig. 12 (range Clc).
v.1976 Estoniops n. sp.; Jaanusson, text-fig. 9 (range).

Derivation of name. Latin calvus , bald, alluding to the impression of the glabella seen in anterior view.

Holotype. Complete but partly disarticulated specimen (PI. 3, fig. 1 a-c), RM Ar55040, from an erratic boulder
at Estuna, the province of Uppland. Furudal Limestone (Hustedograptus teretiusculus Zone).

Other material. Two exoskeletons (RM Ar55042, Ar50494), about 10 cephala and 10 pygidia (see ‘occurrence’
below).

Diagnosis. The genus is monotypic; see generic diagnosis.

Description. Cephalon fairly short (sag.), wide; length equals 40-42 per cent of width. Anterior cephalic margin
broadly rounded, boundary between dorsal surface and doublure poorly defined, not reflected even in
sculpture. In strictly dorsal view (PI. 3, figs 1 a, 4c) a slight change in width and convexity gives the impression
of a lateral termination of the frontal glabellar lobe at about the level of the most lateral extent of the facial
suture. No trace of a vincular furrow (PI. 3, fig. 6b).

Dorsal furrow deep, overhung by adaxial edge of fixed cheek, between S2 and S3 running in an extremely
narrow tunnel below a bridge formed by L3 and anterior of palpebral lobe. Fossula situated directly anterior to

762

PALAEONTOLOGY, VOLUME 36

first lens file of eye. L3 relatively short (exsag.); shape varying from almost parallel-sided (PI. 3, fig. 4c) to
somewhat triangular (PI. 3, fig. la). LI and L2 of about equal length (exsag.), with axes strictly transversely
directed. Occipital ring about as wide as glabella across L3.

Eyes 47-48 per cent of length (sag.) of cephalon. Visual surface of 15 files of up to 6 lenses; eye formula in
largest available cephalon (PI. 3, fig. 4) 345 656 666 555 543, from front backwards. Central portion of glabella
with low, composite tubercles of varying size, between and on which is a very fine granulation which extends
to the borders, doublure and central cheek portions. In specimens with best preserved surface very shallow pits
can be discerned on the genal field. Hypostoma unknown.

All three specimens with thorax preserved have ten thoracic segments.

Length of pygidium (excluding the articulating half-ring) equal to 64-66 per cent and its width to 52-53 per cent
of the respective dimensions of the cephalon. It is comparatively broad, 2-5-2-6 times wider than long,
weakly convex. Pygidial rachis weakly convex, relatively wide, maximum width at first rachial ring equal to
38-39 per cent of total pygidial width. Three rachial rings, defined laterally by deep furrows but medially
almost obsolete, in best preserved specimens marked by very faint furrows. Posterior portion short,
occasionally with one or two faint, transverse depressions medially. Pleural areas of pygidium weakly convex,
with three pairs of poorly discernible pleural furrows and, in best-preserved pygidia, with an indication of an
interpleural furrow on the anteriormost rib.

Occurrence. Uhaku Stage ( Hustedograptus teretiusculus Zone) of Sweden and southern Estonia. Siljan district,
Furudal Limestone, Fjacka section. Northern Oland, Persnas Limestone, Boda Hamn boring. Uppland, erratic
boulders of the Furudal Limestone derived from the South Bothman submarine Cambro-Ordovician
sequence; among the records of P. exilis by Wiman (1908) specimens (housed in PMU) from the following
boulders belong to U. calms : Ekeby 5 and 57, Harg 2, Kristineholm 2, Salsta 3, Sunnersta 2, and Torron 4
and 7. In addition, the species is common in the boulder Bergsbrunna 1 and occurs in a boulder from Estuna
(RM). Southern Estonia, Uhaku Stage, Karula boring (430 4 m, 430-9 m).

Genus keilapyge gen. nov.

Derivation of name. From Keila, a town in northwest Estonia, with a quarry in which F. Schmidt found much
of the material of the type species. Gender feminine.

Type species. Phacops ( Pterygometopus ) laevigatus Schmidt, 1881.

Other species. Estoniops latus Antsygin, 1970, from the upper Middle Ordovician Cherdyn Stage, western slope
of the central Ural mountains. A probable additional species, which differs by its relatively coarsely granulate
glabella (Schmidt 1881, p. 235, fig. 13) occurs in the Keila Stage of northern Estonia (see also Schmidt 1907,
p. 3). A poorly preserved specimen of Keilapyge was figured as Estoniops ( Pterygometopus ) sp. by Neben and
Krueger (1979, pi. 143, figs 25-27) from erratics on the island of Sylt in northern Germany.

Diagnosis. Cephalic margin anteriorly broadly rounded (sag.), without admarginal furrow.
Transsutural wing short (tr.). L3 long (exsag.). Eyes relatively small; anterior end of visual surface

EXPLANATION OF PLATE 5

Fig. 1. Ingriops trigonocephalus (Schmidt, 1881); Voka Beds?, Kunda Stage (Didymograptus artus Zone);
Pavlovsk, Ingria, western Russia. 1 a~d, RM Ar38514, syntype; dorsal, anterior cephalic, dorsal pygidial,
and lateral views of an enrolled exoskeleton, x 3.

Figs 2-3. Achatella ( Vironiaspis ) kuckersianus (Schmidt, 1881); Viivikonna Formation, Kivioli Member,
Kukruse Stage (Nemagraptus gracilis Zone); northeast Estonia. 2 a-c, ETAGI Tr 3612; Kohtla; dorsal,
lateral, and posterior views of pygidium (mainly internal mould), x 4. 3a-c, RM Ar50493 ; Kukruse; dorsal,
lateral, and anterior views of cephalon, x 6.

Fig. 4. Achatella ( s.l .) schmidti (Warburg, 1925); Boda Limestone, Ashgill; Osmundsberget, Dalarna, Sweden.
4 a-b, PMU D189, holotype; dorsal and lateral views, exoskeleton preserved on cheek area only; figured by
Warburg 1925, pi. 11, figs 27-28, x4.

PLATE 5

JAANUSSON and RAMSKOLD, Ingriops , Achatella

3b 4a

764

PALAEONTOLOGY, VOLUME 36

reaches about mid-length of L3. Fixed cheeks provided, at level of L2, with a characteristic,
medially pointed ridge. No vincular furrow. Genal angles rounded. Pygidium with semicircular
outline; four or five pleural furrows.

Discussion. Keilapyge is distinguished from other comparable pterygometopines, such as Estoniops
and Upplandiops gen. nov., by a set of distinctive cephalic characters. The eyes are comparatively
small and reach anteriorly only to about the level of mid-length (exsag.) of L3. In the type species
the length (exsag.) of the eyes is 32-34 per cent of cephalic length (sag.), and the visual surface is
composed of 15 files of up to 6 lenses (counts in material from the Ristna beds). K. lata (Antsygin
1970, pi. 5, fig. 7) appears to have a comparable relative length of the eyes. Transsutural wings of
the frontal lobe are distinct but short (tr.). The course of the anterior branch of the facial suture is
best visible on PI. 4, fig. 3 a (left side) where exfoliation of the exoskeleton follows the suture (it is
not visible on PI. 4, fig. 1). L3 is proportionally very long (exsag.) and with a slightly posteriorly
protruding posterolateral corner. A feature which is unique for Keilapyge is the development of a
medially pointed ridge on the fixed cheek at the level of LI, constricting the dorsal furrow. The ridge
is distinct in both K. laevigata (PI. 4, fig. 1 a, d) and K. lata (Antsygin 1970, pi. 5, figs 6-7), and also
in K. sp. indet. (Neben and Krueger 1979, pi. 143, fig. 25).

The pygidium appears to have approximately the same size relative to the cephalon as in E. exilis.
The type species has four pairs of distinct pleural furrows and a trace of a fifth pair (PI. 4, fig. 5;
Schmidt 1881, pi. 15, fig. 26). The pleural ribs are flat, with very faint traces of interpleural furrows.
The pygidia which Antsygin (1970, pi. 5, figs 8-9) attributed to K. lata have five pairs of pleural
furrows which extend almost to the pygidial margin and distinct interpleural furrows. They are very
similar to the pygidia which Antsygin (1970, pi. 4, figs 14—16) referred to Estoniops exilis, and it is
difficult to say how certain the attribution is of these pygidia to K. lata.

Occurrence. Johvi and Keila Stages (uppermost Diplograptus multidens to lowermost Dicranograptus clingani
Zones) of northern Estonia and upper Middle Ordovician of the western slope of the central Ural mountains.

Genus achatella Delo, 1935

Type species. Dalmanites achates Billings, 1860, by original designation (for distribution see Ludvigsen and
Chatterton 1982).

Diagnosis. Cephalon fairly flat. L3 triangular, with pointed anterolateral termination; S3 long,
anterolaterally directed, straight or faintly posteriorly curved; distance (exsag.) between adaxial
terminations of S2 and S3 less than half the length of L3 along dorsal furrow. Anterior branch of
facial suture runs just outside the preglabellar furrow where developed. Transsutural wings or
equivalent structures short (tr.). No vincular furrow. Pygidium with subparabolic outline, 10 to 14
pleural furrows.

Discussion. Achatella as currently defined appears to constitute a well-defined monophyletic group
of species, but it displays a morphological variation that is quite extraordinary for a dalmanitacean
genus. Genal spines can be long (Ludvigsen and Chatterton 1982, fig. 3) to absent (PI. 5, fig. 4). The
eyes can be comparatively short (exsag.), anteriorly not reaching the dorsal furrow (Ludvigsen and
Chatterton 1982, pi. 1, fig. 7), or fairly long and reaching the dorsal furrow as far forwards as in
front of L3 (PI. 5, fig. 3). A subocular furrow can be distinct (Ludvigsen and Chatterton 1982,
pi. 1, fig. 7) or absent (PI. 5, fig. 3), and the absence of the furrow is unique for the whole subfamily.
The development of the preglabellar furrow can be comparable to that of Pterygometopus (PI. 5, fig. 3)
or the furrow can be effaced laterally; in the latter case the frontal glabellar lobe is prolonged
laterally into short transsutural wings. The development of the preglabellar furrow varies even
medially because in some Ashgillian species from Scotland the furrow appears to be obsolete or
nearly so. A group of species which, also with respect to other characters, are close to the type

JAANUSSON AND RAMSKOLD: ORDOVICIAN TRILOBITES 765

species of Achatella have long (tr.) lateral glabellar furrows with S2 conspicuously shallowing close
to the dorsal furrow, whereas in some other species (PI. 5, fig. 3a) these furrows are relatively short
and S2 does not shallow laterally. Outer portion of the pygidial pleural area varies from convex
(Tripp and Morris 1986, pi. 4, fig. 2) to fairly strongly concave (PI. 5, fig. 2). On the pygidium the
interpleural furrows can be fairly distinctly marked (Ludvigsen and Chatterton 1982, pi. 1, fig. 1)
to obsolete (PI. 5, fig. 2).

It is clear that Achatella as currently defined requires revision but, as the distribution of the genus
is mainly outside the Baltoscandian basin, such a revision is beyond the scope of this paper. An early
representative of the genus from Baltoscandia differs from the other forms to such an extent that
it is here included in a new subgenus, A. ( Vironiaspis ).

A species with a doubtful position within Achatella is P t ery gome t opus schmidti Warburg, 1925,
known only from a single fragmentary cephalon (PI. 5, fig. 4; Warburg 1925, pi. 11, figs 27-28) from
the post -Pleurograptus linearis Zone Ashgill Boda Limestone of the Siljan district in Sweden. The
general relative flatness of the cephalon and the pronouncedly triangular shape of L3 of this species
is Achatella-hke. The frontal lobe is defined, both anteriorly and laterally, by a preglabellar furrow
which follows the facial suture as in A. ( Vironiaspis ) kuckersiana , but the furrow is narrower and
shallower and becomes obsolete posteriorly a short distance before reaching the dorsal furrow. The
anterior branch of the facial suture runs just outside the preglabellar furrow. A subocular furrow
is distinct. In contrast to other species of Achatella , A. (s. /.) schmidti lacks genal spines. Without
further information on the morphology the relationships of the species remain unclear.

Subgenus achatella (achatella) Delo, 1935

Type species. As for genus.

Other species. For North American species see Ludvigsen and Chatterton (1982). European species: Phacops
( Chasmops ) bailyi Salter, 1864 from the early Caradocian Tramore Limestone of Ireland; Phacops
( Pterygometopus ) nieszkowskii Schmidt, 1881, from the Rakvere Stage of northern Estonia; Achatella
consobrina Tripp, 1954, from the Kiln Mudstones at Craighead Quarry near Girvan, south-western Scotland.

A new species of Achatella ( Achatella ), similar to A. (A.) nieszkowskii , is represented by a cephalon
(Paleontologisk Museum, Oslo no. 131629) from the Mjosa Limestone at Bergsvika on Helgoya in the Mjosa
district of the Oslo Region, Norway. A fragmentary cephalon (Paleontologisk Museum, Oslo no. 21991) from
the Furuberget Formation ( Cyclocrinus beds) at Furuberget in the same district may be conspecific. The latter
specimen was recorded by Floltedahl (1910, p. 36) and Stormer (1953, p. 104) as Pterygometopus kuckersianus.

In addition, there is a group of late Ordovician species, included in Achatella by Morris (1988), which,
pending additional information on their cephalic and pygidial morphology, are only tentatively included in the
nominal subgenus. Such species are Phacops truncatocaudatus Portlock, 1843, from the Killey Bridge
Formation of the Pomeroy district in Northern Ireland, Phacops ( Pterygometopus ) retardatus Reed, 1914, from
the South Threave Formation, Starfish Bed, in the Girvan district, southwest Scotland (Morris and Tripp
1986, pi. 4, fig. 2), and Phacops ( Pterygometopus ) quarrelensis Reed, 1930, from the Quarrel Hill Formation,
Lower Drummuck Group in the Girvan district. Specimens, identified as Achatella cf. truncatocaudata, have
been recorded from the Hirnantian High Mains Formation of the Girvan district (Owen 1986, fig. 2 a-e).

Diagnosis. Glabellar furrows tend to be comparatively long, with S2 very shallow close
to the dorsal furrow but deepening medially. Eyes of moderate size, with the anterior end some
distance from the dorsal furrow. Subocular furrow distinct. Outer portion of the pygidial pleural
area convex to weakly concave.

Discussion. The type species was described in detail by Ludvigsen and Chatterton (1982, p. 2184,
pi. 1, figs 1-7).

Examination of cephala of A. (A.) nieszkowskii from Estonia revealed that the species is markedly
similar to A. (A.) achates in many important respects, such as the size and position of the eyes.

766

PALAEONTOLOGY, VOLUME 36

shallowing S2 laterally, relative length of glabellar furrows, and development of the subocular
furrow, and that it differs in these respects from the Estonian Middle Ordovician species which
Mannil (1958) included in Achatella.

Occurrence. North American Midcontinent region, Shermanian to Edenian Stages (approximately lower
Dicranograptus clingani to topmost Pleurograptus linearis Zones). Western Ireland, Tramore Limestone (upper
Nemagraptus gracilis Zone?). Scotland, Girvan district. Kiln Mudstone ( Dicranograptus clingani Zone).
Norway, Oslo Region, Furuberget and Mjosa Formations ( Dicranograptus clingani Zone). Estonia, Rakvere
Stage (uppermost Dicranograptus clingani Zone). For distribution of the species which are questionably
included, see above.

Subgenus achatella (vironiaspis) subgen. nov.

Derivation of name. From the name of the north-eastern Estonian province of Viru, latinized Vironia, the
district where the type species is found. Gender feminine.

Type species. Phacops ( Pterygometopus ) kuckersianus Schmidt, 1881 from the Kukruse Stage of northern
Estonia.

Other species. Tentatively assigned: Phacops (Pterygometopus) kegelensis Schmidt, 1881, from the Keila Stage
of northern Estonia.

Diagnosis. Glabellar furrows of moderate length, with S2 not at all, or only slightly shallowing
towards the dorsal furrow. Eyes fairly large, their anterior end reaching the dorsal furrow.
Subocular furrow not developed. Outer portion of pygidial pleural area distinctly concave.

Discussion. A. ( Vironiaspis ) differs from the nominal subgenus by the following main features:

(1) The eyes reach the dorsal furrow, slightly in front of L3 in the type species.

(2) The lack of both a subocular furrow and a distinct subocular ridge; there is only a faint,
narrow, groove-like constriction of the base of the eye as in many other dalmanitaceans.

(3) The lateral glabellar furrows are comparatively short (tr.) and S2 lacks the distinct abaxial
shallowing developed in all species unconditionally included in the nominal subgenus in this paper.

(4) The outer portion of the pygidial pleural area is conspicuously more concave than in any
other species referred to Achatella.

In A. ( Vironiaspis ) kuckersiana the frontal glabellar lobe is defined, both anteriorly and laterally,
by a narrow but distinct preglabellar furrow (PI. 5, fig. 3a), comparable to that in Pterygometopus,
and as in the latter genus the facial suture (not visible on our photograph) runs just outside the
furrow. An equivalent to the transsutural wing of the frontal lobe is situated lateral to the pre-
glabellar furrow and thus outside the glabella. Its anterolateral boundary is defined by the
cephalic border furrow which becomes effaced anteromedially in front of the preglabellar furrow.
A detailed comparison with the development of comparable structures in some species of
A. ( Achatella ) (e.g. Tripp 1954, pi. 4, figs 26 a-c, 27; Ludvigsen and Chatterton 1982, pi. 1, fig. 7) is
difficult without examining the specimens.

A. ( Vironiaspis^ ) kegelensis, tentatively included in the subgenus, is a rare species and the material
available in Estonian museums is fragmentary. It is not quite certain that the cephala referred to
this species from the erratic boulders of northern Germany (Neben and Krueger 1979, pi. 125, figs
16-17; pi. 142, figs 6-7) are conspecific.

Occurrence. Uhaku (Roomusoks 1970, p. 9; Hustedograptus teretiusculus Zone), Kukruse (Nemagraptus
gracilis Zone), Johvi (upper Diplograptus multidens Zone), and Keila (lower Dicranograptus clingani Zone)
Stages of northern Estonia.

Acknowledgements. For the loan of specimens we are indebted to Reet Mannil (Institute of Geology, Estonian
Academy of Sciences, Tallinn), Arvo Roomusoks (Tartu University), Gunnar Henningsmoen (Paleontologisk

JAANUSSON AND RAMSKOLD: ORDOVICIAN TRILOBITES

767

Museum, Oslo), Solveig Stuenes (Paleontologiskt Museum, Uppsala University), Anita Lofgren (Department
of Geology, Lund University), and the authorities of the Geological Survey of Sweden (Uppsala). Soren
Carlander kindly donated the holotype of Upplandiops calvus to the Riksmuseum. We thank also an anonymous
referee for a constructive review of the manuscript.

REFERENCES

antsygin, N. A. 1970. Trilobity semejstva Pterygometopidae iz ordovikskikh otlozhenij Urala. [Trilobites of the
family Pterygometopidae from the Ordovician deposits of the Urals.] 13-42. In Material y po Paleontologii
Urala. UraFskoe Territorial’noe Geologicheskoe Upravlenie, Sverdlovsk, 190 pp. [In Russian.]
bassler, r. s. 1915. Bibliographic index of American Ordovician and Silurian fossils. Bulletin of the United
States National Museum , 92, 1-1521, pis 1-4.

billings, e. 1860. Description of some new species of fossils from the Lower and Middle Silurian rocks of
Canada. Canadian Naturalist and Geologist , 5, 49-69.

brogger, w. c. 1882. Die silurischen Etagen 2 und 3 im Kristianiagebiet und auf Eker. Universitdtsprogramm
fur 2 Sem., 1882, i-viii+ 1-376. Kristiania (Oslo).

Campbell, K. s. w. 1977. Trilobites of the Haragan, Bois d’Arc and Frisco Formations (early Devonian),
Arbuckle Mountains region, Oklahoma. Oklahoma Geological Survey, Bulletin , 123, i-vi, 1-227, pis 1-40.
Clarkson, E. N. K. 1966. The life attitude of the Silurian trilobite Phacops musheni Salter, 1 864. Scottish Journal
of Geology, 2, 76-83.

dalman, J. w. 1827. Om palaeaderna, eller de sa kallade trilobiterna. Kungliga Svenska Vetenskaps-Akademiens
Handlingar, 1826, 113-152, 226-294.

dean, w. T. 1962. The trilobites of the Caradoc Series in the Cross Fell Inlier of northern England. Bulletin of
the British Museum ( Natural History), Geology Series, 7, 65-134.
delo, d. m. 1935. A revision of the phacopid trilobites. Journal of Paleontology , 9, 402^120.
destombes, J. 1972. Les trilobites du sous-ordre des Phacopina de FOrdovicien de FAnti-Atlas (Maroc). Notes
et Memoires du Service Geologique du Maroc, 240, 112 pp.
eichwald, c. E. 1858. Beitrag zur geographischen Verbreitung der fossilen Thiere Russlands. Alte Periode.

Bulletin de la Societe Imperiales des Naturalistes de Moscou, 30, 193-242.
eldredge, n. 1972. Systematics and evolution of Phacops rana (Green, 1832) and Phacops iowensis Delo, 1935
(Trilobita) from the Middle Devonian of North America. Bulletin of the American Museum of Natural
History, 147, 45-1 14.

frederickson, e. a. and pollack, j. m. 1952. Two trilobite genera from the Harding Formation (Ordovician)
of Colorado. Journal of Paleontology, 26, 641-644.

hammann, w. 1972. Neue propare Trilobiten aus dem Ordovizium Spaniens. Senckenbergiana lethaea, 53,
371-381.

— 1974. Phacopina und Cheirurina (Trilobita) aus dem Ordovizium von Spanien. Senckenbergiana lethaea,
55, 1-151.

henry, j.-l., vizcaino, d. and destombes, J. 1992. Evolution de Foeil et heterochronie chez les Trilobites
ordoviciens Ormathops Delo 1935 et Toletanaspis Rabano 1989 (Dalmanitidae, Zeliszkellinae). Palaont-
ologische Zeitschrift, 66, 277-290.

holtedahl, o. 1910. Studien iiber die Etage 4 des norwegischen Silursystems beim Mjosen. Videnskabs-
Selskabets Skrifter. I. Mathematisk Naturvidenskapelig Klasse, 1909 (7), 1-76.
jaanusson, v. 1956. On the trilobite genus Celmus Angelin, 1854. Bulletin of the Geological Institutions of the
University of Uppsala , 36, 35-49.

— 1960. The Viruan (Middle Ordovician) of Oland. Bulletin of the Geological Institutions of the University
of Uppsala, 38, 289-366.

1963. Lower and Middle Viruan (Middle Ordovician) of the Siljan District. Bulletin of the Geological
Institutions of the University of Uppsala, 43, 1-40.

— 1976. Faunal dynamics in the Middle Ordovician (Viruan) of Balto-Scandia. 301-326. In bassett, m. g.
(ed.). The Ordovician system: proceedings of a Palaeontological Association symposium. University of Wales
Press and National Museum of Wales, Cardiff, 696 pp.

kielan, z. 1960. Upper Ordovician trilobites from Poland and some related forms from Bohemia and
Scandinavia. Palaeontologia Polonica, 11 [for 1959], i-vi, 1-198.
lamansky, w. 1905. Drevnejshie sloi silurijskikh otlozhenij Rossii. [Die aeltesten silurischen Schichten
Russlands (Etage B).]. Memoires du Comite Geologique , N.S., 20 (1), i-vii, 1-203, pis. 1-2. [In Russian and
German].

768

PALAEONTOLOGY, VOLUME 36

linnarsson, j, G. o. 1869. Om Vestergotlands kambriska och siluriska aflagringar. Kimgliga Svenska
Vetenskaps-Akademiens Handlingar, 8 (2), 1-89.

lu, yen-hao 1975. Ordovician trilobite faunas of central and southwestern China. Palaeontologia Sinica , New
Series B, 11, 1^163. [In Chinese and English],

ludvigsen, r. and chatterton, b. D. e. 1982. Ordovician Pterygometopidae (Trilobita) of North America.
Canadian Journal of Earth Sciences , 19, 2179-2206.

maksimova, z. a. 1957. Nakhodka pantsyrya predstavitelya nadsemeystva Phacopoidea s 10 tulovishchnymi
segmentami. [A find of an exoskeleton of a representative of the superfamily Phacopoidea with 10 thoracic
segments]. Ezhegodnik Vsesoyuznogo Paleontologicheskogo Obshchestva , 16, 216-217. [In Russian].

1962. Trilobity ordovika i silura Sibirskoy platformy. [Ordovician and Silurian trilobites of the Siberian
platform.]. Trudy Vsesoyuznogo Nauchno-Issledovatelskogo Geologicheskogo lnstituta ( VS EG El), 76, 1-214.
[In Russian].

mannil, R. 1958. [Estoniops - a new genus of Phacopidae (Trilobita)]. Eesti NSV Teaduste Akadeemia
Toimetised , 6 [for 1957], 385-388. [In Russian with Estonian and English summaries],

1966. Istoria razvitiya Baltijskogo basseina v ordovike [ Evolution of the Baltic basin during the Ordovician ].
Eesti Teaduste Akadeemia Geoloogia Instituut, Tallinn, 200 pp.

— polma, l. and hints, l. 1968. Stratigrafia viruskikh i haryuskikh otlozheniy (ordovik) sredney Pribaltiki.
(Stratigraphy of the Viru and Harju Series (Ordovician) of the Central East Baltic area). 81-1 10. In grigelis,
a. (ed.). Stratigraphy of the Baltic lower Paleozoic and its correlation with other areas. The Geological
Institute, Vilnius, 318 pp. [In Russian with English summary].

m'coy, F. 1851. In Sedgwick, a. and m'coy, f. (eds). 1851-1855. A synopsis of the classification of the British
Palaeozoic rocks , with a systematic description of the British Palaeozoic fossils found in the geological museum
of the University of Cambridge. London and Cambridge, iv+ 184 pp.
moore, r. c. (ed.). 1959. Treatise on invertebrate paleontology. Part O. Arthropoda 1. Geological Society of
America and University of Kansas Press, Boulder, Colorado and Lawrence, Kansas, xix + 560 pp.
morris, s. f. 1988. A review of British trilobites, including a synoptic revision of Salter’s monograph.
Monograph of the Palaeontographical Society , 140 (574), 1—316.

— and tripp, r. p. 1986. Lectotype selections for Ordovician trilobites from the Girvan District, Strathclyde.
Bulletin of the British Museum ( Natural History ), Geology Series , 40, 161-176.

neben, w. and krueger, H. H. 1979. Fossilien kambrischer, ordovizischer und silurischer Geschiebe. Staringia,
5, 1-63.

olin, e. 1906. Om de Chasmopskalken och Trinucleusskiffern motsvarande bildningarna i Skane. Meddelanden
frdn Lunds Geologiska Faltklubb, B. 1, 1-79.

opik, A. 1937. Trilobiten aus Estland. Acta et Comment at iones Universitatis Tartuensis, 32 (3), 1-163.

Owen, a. w. 1986. The uppermost Ordovician (Hirnantian) trilobites of Girvan, SW Scotland with a review of
coeval trilobite faunas. Transactions of the Royal Society of Edinburgh , Earth Sciences, 77, 231-239.
portlock, J. E. 1843. Report on the geology of the County of Londonderry, and parts of Tyrone and Fermanagh.
Dublin and London, xxxi + 784 pp.

rabano, i. 1989. Trilobites del Ordovfcico Medio del sector meridional de la zona Centroiberica espanola. IV.

Phacopina, Scutelluina, Odontopleurida y Lichida. Boletin Geologico y Minero, 100, 971-1032.
ramskold, l. and werdelin, l. 1991. The phylogeny and evolution of some phacopid trilobites. Cladistics , 7,
29-74.

reed, f. R. c. 1905. The classification of the Phacopidae. Geological Magazine (5), 2, 172-178, 224-228.

— 1914. The Lower Palaeozoic trilobites of Girvan. Supplement. Monograph of the Palaeontographical
Society , 67 (329), 1-56, pis 1-8.

— 1930. Notes on some new Ordovician trilobites from Girvan. Annals and Magazine of Natural History
(10), 6, 193-201, pis 9-10.

roomusoks, A. 1970. Stratigrafiya viruskoy i kharyuskoy seriy ( ordovik ) severnoy Estonii. I. [ Stratigraphy of the
Viruan Series (Middle Ordovician ) in Northern Estonia] . Tartu Riiklik Ulikool, 345 pp. [In Russian, with
Estonian and English summaries].

salter, j. w. 1864. A monograph of the British trilobites from the Cambrian, Silurian and Devonian
formations. Part I. Monograph of the Palaeontographical Society, 16 (67), 1-80, pis 1-6.
schmidt, F. 1881. Revision der ostbaltischen silurischen Trilobiten, nebst geognostischer Ubersicht des
ostbaltischen Silurgebiets. I. Phacopiden, Cheiruriden und Encrinuriden. Memoires de f Academic Imperiale
des Sciences de St.-Petersbourg , Serie 7, 30 (1), 1-237.

— 1907. Revision der ostbaltischen silurischen Trilobiten. VI. Allgemeine Ubersicht mit Nachtragen und
Verbesserungen. Memoires de I'Academie Imperiale des Sciences de St.-Petersbourg, Serie 8, 20, i-xv, 1-104.

JAANUSSON AND RAMSKOLD: ORDOVICIAN TRILOBITES

769

semenova, v. s. 1984. Sredne- 1 verkhneordovikskie trilobity Sibirskoy platformy. [Middle and Upper
Ordovician trilobites from the Siberian Platform], 73-84. In Ordovik sibirskoj platformy. Paleontologicheskij
atlas. [Ordovician of the Siberian Platform. Palaeontological atlas.] Trudy Instituta Geologii i Geofiziki , no.
590. Akademiya Nauk SSSR, Sibirskoe otdelenie, Novosibirsk, 242 pp. [In Russian],
stormer, L. 1953. The Middle Ordovician of the Oslo Region, Norway. 1. Introduction to stratigraphy. Norsk
Geologisk Tidsskrift , 31, 37-141.

tripp, r. p. 1954. Caradocian trilobites from mudstones at Craighead Quarry, near Girvan, Ayrshire.

Transactions of the Royal Society of Edinburgh , 62, 655-693.
warburg, E. 1925. The trilobites of the Leptaena Limestone in Dalarne with a discussion of the zoological
position and classification of the Trilobita. Bulletin of the Geological Institutions of the University of Uppsala ,
17, 1-446.

weber, v. n. 1948. Trilobity silurijskikh otlozhenij SSSR. 1. Nizhnesilurijskie trilobity. [Trilobites of the
Silurian rocks of the USSR. 1. Lower Silurian trilobites], Monografii po Paleontologii SSSR , 69, 1-113. [In
Russian].

Whittington, h. b. 1950. Sixteen Ordovician genotype trilobites. Journal of Paleontology , 24, 531-565.

— 1962. The Ordovician trilobites of the Bala area, Merioneth. I. Monograph of the Palaeontographical
Society, 116 (497), 1-32, pis 1-8.

— and evitt, w. R. 1954. Silicified Middle Ordovician trilobites. Memoirs of the Geological Society of
America, 59, 1-137.

wiman, c. 1908. Studien fiber das Nordbaltische Silurgebiet. II. Bulletin of the Geological Institutions of the
University of Uppsala, 8 [for 1906-1907], 73-168.

VALDAR JAANUSSON
LARS RAMSKOLD*

Department of Palaeozoology
Swedish Museum of Natural History
Box 50007

S-104 05 Stockholm, Sweden

*Present address:
Palaeontological Museum
University of Uppsala

Typescript received 2 July 1992 Norbyvagen 22

Revised typescript received 21 March 1993 S-752 36 Uppsala, Sweden

PROBLEMATICAL MICROFOSSILS FROM THE
SILURIAN OF IRELAND AND SCOTLAND

by ANN ALISA FERRETTI, CHARLES HEPWORTH HOLLAND and ERIKA SYBA

Abstract. Problematical microfossils are described from the Silurian of the Dingle Peninsula, County Kerry,
Ireland and from a clast in the Old Red Sandstone Greywacke Conglomerate of the Midland Valley of
Scotland. They are assigned to Sandvikina under the new specific name conica. The Irish material includes also
single specimens of Sandvikina sp. and Regnellia camera. All these microfossils, which appear to be closely
related, are here assigned to the new family Regnellidae, which, on present evidence, is characteristic of the
Silurian.

In 1987 Syba found a single fossil in a thin-section of a clast from the Old Red Sandstone of the
Scottish Midland Valley. Its conical appearance with transverse partitions had suggested that it
might possibly be a cephalopod. However, its maximum dimension is only c. 2 mm, and there is no
indication of a siphuncle; clearly another attribution was needed. It resembled Salterella in which
shearing might have produced the marginal bending of the partitions, but E. Yochelson (pers.
comm.) indicated that this was not the case. N. H. Nitecki (pers. comm.) confirmed that our
specimen was not any of the plant or animal problematica known to him. S. Bengston (pers. comm.)
suggested that the fossil was possibly a spine or sclerite (he was looking at a photograph of the single
Scottish specimen only), though he would not call it a tommotiid on the present evidence.

In 1990 Ferretti was working on the microfacies of some Silurian limestones from the Dingle
Peninsula, Country Kerry. Her thin-sections revealed material evidently the same as the Scottish
microfossil already referred to. Subsequently she has collected more material. In a paper by
Lauritzen (1974) describing new microfossils from the Llandovery of the Oslo region one of the
forms referred to there appeared to be clearly related to our original enigmatic fossil; the other is
the same as one of those described by Regnell from the Swedish Silurian in 1947, and which occurs
in the Irish material.

STRATIGRAPHICAL SETTING OF THE SCOTTISH SPECIMEN

The basal Lower Old Red Sandstone Greywacke Conglomerate is exposed in a series of
discontinuous outcrops just north of, and parallel to, the Southern Uplands Fault. It has previously
been regarded as forming the base of the Devonian (House et al. 1977). Although fossil evidence
had been absent from the conglomerate and associated sandstones, Thirwall (1988) dated a major
felsite intruded into the Greywacke Conglomerate at approximately 412 Ma, which would suggest
its depositional age at no later than latest Silurian.

The composition of the Greywacke Conglomerate is markedly uniform, with fine to coarse
sandstones forming the majority of the clasts at any one locality. The sandstone clasts are
compositionally immature arenites and wackes, initially deposited as turbidites. These and other
clast lithologies are similar to those found in the Ordovician (upper Llandeilo to ?Ashgill) and
Silurian (Llandovery to Wenlock) of the Southern Uplands. This led many previous authors to
suggest that the source for the Greywacke Conglomerate lay there.

A re-appraisal of the sedimentology of the Greywacke Conglomerate has shown the sequence to
be characterized by small (2 km length) vertically stacked proximal alluvial fan sediments deposited

[Palaeontology, Vol. 36, Part 4, 1993, pp. 771-783, 3 pis.]

© The Palaeontological Association

772

78

PALAEONTOLOGY, VOLUME 36

81

T

82

633

77

79 80

0 100m

I I

text-fig. 1. Geological sketch map of part of the Midland Valley of Scotland to show locality from which the

clast was collected.

FERRETTI ET AL.: SILURIAN PROBLEMATICA

773

text-fig. 2. Geological sketch map of the Dingle Peninsula to show setting of the Annascaul inlier. Details of
the area at the north-eastern end of the inlier with Parkin’s (1976) localities 28 and 36 referred to in the text.

774

PALAEONTOLOGY, VOLUME 36

in a series of separate north-south trending basins, each with one dominant fault-controlled margin
towards the east (Syba 1989). This work also showed that the petrography and geochemistry of the
greywacke clasts could not easily be matched with rocks in the Southern Uplands. This led to the
suggestion that the source of the Greywacke Conglomerate was a cover of greywacke within the
Midland Valley. The petrographical and geochemical analysis showed that the tectonic setting for
the greywacke flysch deposits ranged from passive margin (near Girvan, approximately 60 km
southwest of the Hagshaw Hills) to active continental margin. The presence of passive margin-type
sands indicated that the source for the greywacke conglomerate may have been accreted to the
Midland Valley prior to or during the obduction phase of the Ballantrae ophiolite (Tremadoc-
Arenig). The age of the sedimentary sequence which gave rise to the Greywacke Conglomerate is
of clear importance in deciphering the relationship of the source to the Laurentian margin. The
single fossil referred to here was found within a sandstone clast from the Hagshaw Hills. The clast
was collected from c. 53 m above the base of the conglomerate at a small exposure (NS 773301) just
west of Monks Water (Text-fig. 1).

STRATIGRAPHICAL SETTING OF THE IRISH SPECIMENS

The Dingle Peninsula, County Kerry (Text-fig. 2) provides the richest Silurian shelly faunas to be
found in Ireland. This is especially so in the western Dunquin Inlier (Holland 1988). The shelly
fauna of the Annascaul Inlier, described in detail by Parkin (1976), is confined to the Ballynane
Member near the top of the late Wenlock Annascaul Formation, and to a younger formation in part
of the Ludlow sequence. At the mountainous eastern end of the inlier, on the western slopes of
Caherconree, there are local limestones within the tuffaceous Ballynane Member.

Si veter (1989) described a rich trilobite fauna from two small exposures, separated by a little over
one kilometre of unexposed ground. These were discovered in the nineteenth century by officers of
the Geological Survey, and described as ‘dove coloured limestone’ and ‘grey crystalline limestone’
(localities 28 and 36 of Parkin).

A considerable quantity of material from the latter (108 samples) has been examined by Ferretti
which has yielded the microfossils described below. The grey limestone is a crinoidal-trilobite silty
wackestone to packstone, redeposited in, and embedded by, a terrigenous siltstone. The matrix of
the limestone is lime mud with sporadic sparry cement among bioclasts.

Parkin (1976) regarded the Ballynane Member as likely to be upper Wenlock in age; its possible
extent into the Ludlow is constrained by the overlying Caherconree Formation which is of middle
Gorstian age. Aldridge (1980), on conodont evidence, suggested that Locality 28 might be of
Wenlock age and Locality 36 of Ludlow age. The best trilobite evidence suggested a ‘mid/late
Wenlock to earliest Ludlow’ age (Siveter 1989).

SYSTEMIC PALAEONTOLOGY
PROBLEMATICA

Family regnellidae fam. nov.

Diagnosis. Conical to cylindrical microfossils with relatively numerous transverse partitions.

explanation of plate 1

Figs 1-3. Sandvikina conica sp. nov. 1, Institute of Palaeontology, University of Modena no. 24211, holotype.
Dingle Peninsula; Ballynane Member; the basal cone and empty distal portion are clearly seen, x65.
2, Modena no. 24212. Ballynane Member, x 65. 3, University of Glasgow no. l-H-4. Midland Valley of
Scotland; Greywacke Conglomerate, x 75.

PLATE 1

FERRETTI et at Silurian problematica

776

PALAEONTOLOGY, VOLUME 36

Discussion. The various microfossils referred to below appear to be closely related. Regnell (1947)
was the first to give descriptions. One of the two forms to which he referred was subsequently given
the generic name Regnellia Lauritzen, 1974.

Genus sandvikina Lauritzen, 1974
Type species. Sandvikina brachiata Lauritzen, 1974.

Emended diagnosis. A microfossil up to about 5 mm in length, in section comprising a basal conical portion
(PI. 1, fig. 1) with distally concave transverse partitions and an empty distal portion which widens considerably
in comparison with the basal cone.

Discussion. Lauritzen (1974, p. 704) referred to a trapezoidal-shaped microfossil with complex wall
structure. Our additional material suggests that the former may be an accident of preservation in
the section, and the latter is not so characteristic of other forms we place in the same genus.

Sandvikina conica sp. nov.

Plate 1, figs 1-3; Plate 2, figs 1-3

Hoiotype. Institute of Palaeontology, University of Modena, no. 2421 1 (in thin section); limestone in Ballynane
Member, Annascaul Formation; Dingle Peninsula, County Kerry; Parkin (1976), locality 36; latest Wenlock
or possibly earliest Ludlow.

Other material. Institute of Palaeontology, University of Modena, nos 24212-24220, 24223; locality as for
hoiotype. Hunterian Museum, University of Glasgow, P 1380; clast from Greywacke Conglomerate; Hagshaw
Hills, Midland Valley of Scotland. All material is in thin sections.

Diagnosis. As for genus, but additionally concave partitions of basal cone bend back at outside, and
wall of distal portion is a network.

Description. Though oblique sections of a cylinder would produce a cone, there are sufficient specimens of
sufficient similarity to suggest that the basal portion really is conical. Its cross-section remains unknown.
Dimensions are given in Table 1. Some transverse partitions anastomose; sometimes the single partitions have
spine-like terminations. The wall of the distal portion appears to be composed of two distinct layers, as revealed
in all the thin sections. There are never single lines as would be expected in a ‘planar’ net but rather fragments
of a chain. In Modena 24213 (PI. 2, fig. 3) the inner wall seems to be continuous, even if this represents a
particular section of the organism. Modena 24220 (PI. 2, fig. 1) shows a wall in the distal portion with oblique
portions of network tangential to it. A single fragment of the net (letter c in Text-fig. 3 ; PI. 2, fig. 2) clearly shows
three different levels, two of which (a and b ) are calcareous, but with different crystal sizes, and a third ( c ) which
is silicified.

Discussion. Regnell (1947, fig. 2) illustrated what appears to be an identical form, though Lauritzen
referred only to his cylindrical material. Sandvikina brachiata Lauritzen, 1974, was described and
illustrated as having a layered wall structure in both portions. In this respect it differs from S. conica.
In the basal cone, referred to by Lauritzen (1974) as ‘the outgrowth’, the transverse elements are

EXPLANATION OF PLATE 2

Figs 1-3. Sandvikina conica sp. nov. 1, Institute of Palaeontology, University of Modena no. 24220. Ballynane
Member; fragment of distal portion, x 55. 2, Modena no. 24223. Ballynane Member; three different
levels are recognizable, two of which are calcareous (a and b) and one that has been silicified (c) x 55.
3, Modena no. 24213 and 24214. Ballynane Member; the distal portion of these specimens is very well
developed, x 35.

PLATE 2

FERRETT1 et al Silurian problematica

778

PALAEONTOLOGY, VOLUME 36

table 1. Measurements of specimens of Sandvikina conica sp. nov. from Ireland taken from thin sections and
therefore approximately orientated. In some cases accurate measurement is difficult and some specimens are
incomplete. Specimens numbers as indicated in the text.

Specimen

number

Total

length

(mm)

Length of
basal cone
(mm)

Number of

transverse

partitions

Angle of
increase of
basal cone

Angle of
increase of
distal part

2421 1

2-2

10

27

20

24212

5-2

20

26

21

—

24213

20

0-2

5

17

60

24214

20

0-5

—

19

90

24215

—

1-4

35

16

—

24216

—

0-5

14

55

—

24217

—

1-2

39

7-15

—

24218

1-8

0-2

7

50

80

24219

2-0

0-4

4

41

80

fewer and, except for one separating the two parts of the fossil, discontinuous; but this may well
be a question of the level of sectioning. The ‘main part' is shown as trapezoidal but two portions
of this structure are not connected and again it may be a fortuitous preservation of the fractured
distal part of the fossil, for which in this case there is no evidence of a network.

Sandvikina sp.

Plate 3, fig. 1

Material. Institute of Palaeontology, University of Modena, no. 24221 (in thin section); locality 36, Dingle
Peninsula (see above).

Description. Similar to S. conica, but basal cone with a definite thin outer wall and regularly placed longitudinal
as well as transverse elements internally. Total length 5-0 mm, length of basal cone 2-6 mm; number of
transverse partitions 40, approximate distance apart 0 06 mm; angle of increase of basal cone 10°, angle of
increase of distal part 38°. As there is only one section of this form a specific name is not suggested.

Genus regnellia Lauritzen, 1974
Type species. Regnellia camera Lauritzen, 1974.

Regnellia camera Lauritzen, 1974

Plate 3, figs 2-3

1974 Regnellia camera Lauritzen, p. 712, pi. 102, figs 2-3.

text-fig. 3. Sketch of thin section with specimens of Sandvikina conica sp. nov., whose distribution is indicated
by the grey colour. Institute of Palaeontology, University of Modena; A, holotype, no. 24211; B, no. 24212.
Ballynane Member; Dingle Peninsula, Ireland. Trilobite and crinoidal fragments are easily recognizable.
Direction of bedding is indicated in the section by the disposition of small levels of detrital quartz grains and
is parallel to specimen b. Specimen a is illustrated in Plate 1, figure 1 ; b, in Plate 1, figure 2; and c, in Plate 2,

figure 2.

FERRETTI ET AL.\ SILURIAN PROBLEM ATICA

779

780

PALAEONTOLOGY, VOLUME 36

Material. Institute of Palaeontology, University of Modena, no. 24222 (in thin section), locality 36, Dingle
Peninsula (see above).

Description. A cylindrical microfossil 3-6 mm in length, with c. 80 transverse concave partitions. These bend
more steeply towards the outside to give an imbricate impression. Both ends are incomplete. Cross-section
unknown. The specimen appears to be the same as the holotype figured by Lauritzen from Norway, and
Regnell’s material from Sweden, which the former suggested had a circular cross section.

COMPOSITION

The holotype of Sandvikina conica is partially silicified while the Scottish specimen is entirely
silicified. The former is calcareous in the basal cone and completely silicified in the distal part.
Silicification could easily be explained for the Irish fossil by the abundant tuff's in which the
limestones are intercalated. The other Irish material is preserved in a microcrystalline calcareous
matrix, which is sometimes slightly locally silicified. Regnell (1947, p. 1) commented that the fossils
‘seem to be so intimately connected with the surrounding rock-matrix that they cannot even be
detected in hand-specimens but in slides only’.

STRATIGRAPHICAL DISTRIBUTION

Lauritzen’s material is from Aeronian biomicritic limestones, from Sandvika, 12 km west-south-
west of the centre of Oslo, Norway. Regnell’s specimens came from a quarry at Kallholn in Dalarna,
central Sweden. The rock was described as ‘crack-filling in the Boda limestone’, suggesting that it
‘might not be older than the basal Silurian’. Lauritzen commented further that the fissures are filled
with graptolitic shales from the middle Llandovery and that these sometimes contain horizons of
dark concretionary limestone. The stratigraphical setting of our own Irish material has been
discussed above. The evidence so far suggests that the Regnellidae range from middle Llandovery
to Wenlock, or possibly basal Ludlow.

The lowermost exposed Silurian in the Hagshaw Hills (Midland Valley of Scotland) is represented
by the Hagshaw Group. The basal units of this are interpreted to be marine turbidites and have been
dated as upper Llandovery (Rolfe 1961). The Hagshaw Group is overlain by the Parisholm
Conglomerate (Igneous Conglomerate) and contains about 6 per cent of sedimentary clasts which
are indistinguishable from those found in the Igneous Conglomerate. Wenlock age fossils have been
found in the Igneous Conglomerate (Rolfe and Fritz 1966) and indicate that the age of the source
rock for the greywacke clasts is more likely to be Llandovery than Wenlock or younger.

The age of the fossil described above indicates a Silurian age source for the Greywacke
Conglomerate and not a pre-Arenig source terrane as previously suggested. The deposition of the
Greywacke Conglomerate into locally filled basins suggests that the greywacke source was either
autochthonous within the Midland Valley or accreted to the Midland Valley Laurentian margin. As
most of the strike-slip movement along the Laurentian margin is believed to have occurred before
the end of the Llandovery (Dewey and Shackleton 1984) the simplest model is the accumulation of
a relatively thick Llandovery age greywacke within the Midland Valley. The Silurian is poorly
exposed in the Midland Valley but occurs in a series of inliers from the southwest to the northeast
along the southern margin, where the Llandovery is represented as deep water turbidites.
Palaeocurrent data suggest that in the Hagshaw Hills the sediments were derived from the south,

EXPLANATION OF PLATE 3

Fig. 1. Sanvikina sp. Institute of Palaeontology, University of Modena no. 24221. Ballynane Member, x35.
Figs 2-3. Regnellia camera Lauritzen, 1974. 2, Modena no. 24222. Ballynane Member; close view of Plate 3,
fig. 3. x 135. 3, Modena no. 24222; Ballynane Member, x 25.

PLATE 3

FERRETTI et al., Silurian problematica

782

PALAEONTOLOGY, VOLUME 36

but equivalent formations in the most north-easterly inlier suggest sediment dispersal from the east-
north-east. Petrographic and geochemical analysis of the greywacke clasts, along with both a
southerly and northerly dispersal pattern in the Llandovery sediments, are consistent with
deposition in a mature Midland Valley inter-arc basin.

The source for the Greywacke Conglomerate is now believed to be the Llandovery age sediments
in the Midland Valley. These sediments were probably uplifted during Wenlock times and may
account for the disconforinity seen at the base of the Igneous Conglomerate in the Hagshaw Hills.
The deposition of the greywacke conglomerate into separate north-south trending basins, each with
a source direction from the east may represent east-west plate interaction between Laurentia and
Baltica.

AFFINITY AND FUNCTION

Neither Regnell nor Lauritzen could present any positive evidence concerning the nature of the
organism which produced the fossils described here. We must add that an affinity with the sponges
seems improbable as the scale and proportions of the skeletal elements are inappropriate. The fossils
remain for the present as problematica.

The Irish and Scandinavian specimens are associated with various groups of shelly fossils, notably
brachiopods, bryozoans, echinoderms, and trilobites. Lauritzen recorded the presence of Girvanella.
With regard to the Irish material, bivalves and gastropods, which generally dominate the near-shore
marine environment, are absent, as are typical Silurian pelagic organisms such as the graptolites and
nautiloid cephalopods. Similar shallow and well-ventilated water communities, dominated by
brachiopod-trilobite-crinoidal associations, are very common in volcaniclastic accumulations,
evidencing colonization during minimal supply of volcanic material by explosions or currents.
A relatively shallow water environment, rich in detrital material, seems likely, where the water was
agitated sufficiently to fragment the various fossils but not to destroy them.

The presence of R. camera in different environments led Lauritzen to suggest a pelagic mode of
life for the organism. The Irish specimens are associated only with benthic forms, and we are
inclined to suggest a benthic mode of life for the Regnellidae also.

Our suggestion is that the basal cone became attached to the substrate and grew apace with
sedimentation to provide an effective anchor, upon which foundation the delicate distal network was
fixed. This may have functioned in the manner of the skeleton of a sponge in a filter-like system or
possibly projections of the soft body emerged from the side. The presence of numerous partitions
in the basal cone and their disposition reveal a certain flexibility together with steadiness if subjected
to movement by currents. Regnellia camera may have grown at greater length obliquely through the
sediment as suggested by Hubbard (1970) for the superficially similar, but very much larger.
Carboniferous caninioid corals to which she referred. Its distal network, if ever present, remains
unknown.

Acknowledgements. We thank Professor Antio Rossi for his kind and helpful advice concerning
the composition of the fossils, and Professor Enrico Serpagli, for his support of work by A. F., and for reading
a draft of this paper. Drs S. Conway Morris and E. Flugel directed our attention to useful references. This is
a contribution to the EEC project ‘Silurian ecostratigraphy in Ireland and Sardinia’.

REFERENCES

aldridge, R. j. 1980. Notes on some Silurian conodonts from Ireland. Journal of Earth Sciences, Royal Dublin
Society, 3, 127-132.

dewey, J. f. and shackleton, R. M. 1984. A model for the Grampian tract in the early Caledonides and
Appalachians. Nature, 312, 115-121.

Holland, c. H. 1988. The fossiliferous Silurian rocks of the Dunquin inlier, Dingle Peninsula, County Kerry,
Ireland. Transactions of the Royal Society of Edinburgh, Earth Sciences, 79, 347-360.

FERRETTI ET AL.: SILURIAN PROBLEMATIC A

783

HOUSE, M. R., RICHARDSON, J. B., CHALONER, W. G., ALLEN, J. R. L., HOLLAND, C. H., and WESTOLL, T. S. 1977.
A correlation of the Devonian rocks in the British Isles. Geological Society of London , Special Report , No. 7,

1 10 pp.

hubbard, j. A. e. b. 1970 Sedimentological factors affecting the distribution and growth of Visean caninoid
corals in north-west Ireland. Palaeontology, 13, 191-209.
lauritzen, o. 1974. New microfossils from the Silurian (Llandovery Stage 6) of the Oslo Region, Norway.
Palaeontology , 17, 707-714.

parkin, j. 1976. Silurian rocks of the Bull’s Head, Annascaul and Derrymore Glen inliers, Co. Kerry.

Proceedings of the Royal Irish Academy , Series B, 76, 577-606.
regnell, g. 1947. Some problematic micro-fossils from the Silurian of Dalarna. Kungliga Fysiografiska
Sdllskapets i Lund , Forhandlingar, 17, 1-7.

rolfe, w. d. i. 1961. The geology of the Hagshaw Hills Silurian inlier, Lanarkshire. Transactions of the
Edinburgh Geological Society , 18, 240-269.

— and fritz, M. A. 1966 Recent evidence for the age of the Hagshaw Hills Silurian inlier, Lanarkshire.
Scottish Journal of Geology, 1, 159-164.

siveter, derek J. 1989 Silurian trilobites from the Annascaul inlier, Dingle Peninsula, Ireland. Palaeontology,
32, 109-161.

syba, e. (1989. The sedimentation and provenance of the Lower Old Red Sandstone Greywacke Conglomerate,
southern Midland Valley, Scotland. Unpublished PhD Thesis, Glasgow University.
thirlwall, M. F. 1988. Geochronology of Late Caledonian magmatism in northern Britain. Journal of the
Geological Society, London, 145, 951-967.

ANNALISA FERRETTI

Institute of Palaeontology
University of Modena
Via Universita, 4
41100 Modena, Italy

CHARLES HEPWORTH HOLLAND

Department of Geology
Trinity College
Dublin, 2, Ireland

ERIKA SYBA

Brigantian Exploration Ltd.
Thatched Cottage
The Green

Typescript received 27 July 1992 South Collingham

Revised typescript received 7 January 1993 Newark NG23 7LE, UK

ONTOGENY OF THE EODISCID TRILOBITE
SHIZHUDISCUS LONGQUANENSIS FROM THE
LOWER CAMBRIAN OF CHINA

by zhang xi-guang and EUAN N. K. CLARKSON

Abstract. The eodiscid Shizhudiscus longquanensis occurs abundantly in the Lower Cambrian Shuijintuo
Formation, Pengshui, Sichuan, China. Phosphatized specimens in excellent preservation have been isolated
from a single limestone lens, and include a few protaspides and disarticulated examples of all subsequent
growth stages. From this material the ontogeny of S. longquanensis is described and reconstructed.
Of particular note are two large spines on the axis of the transitory pygidium which eventually become those
of the second and third thoracic segments. The visual surface first appears in the earliest meraspid, with a single
lens flanked by two half lenses. An immature specimen of Shizhudiscus sp. from Shaanxi Province shows that
this trilobite genus was already capable of full enrollment in the degree 0 meraspid stage.

The small isopygous eodiscids are a specialized group of trilobites, occurring as a common
element in Cambrian faunas of world-wide distribution. Whereas their morphology is well
understood, their ontogeny has been poorly known until recently, when Rushton (1966) and Jell
(1970, 1975a) described growth series with degree 0 meraspides. A further study was made by Hu
(1971), though only one specimen which he described as a paraprotaspis is correctly assigned to an
eodiscid (Jell 1975a). Zhang (1989) and Shergold (1991) were able to elucidate largely complete
growth series of two related eodiscid trilobites from the Lower Cambrian of China, and the Middle
Cambrian of Australia respectively. Both of these descriptions included convincing protaspides.

Many eodiscid genera have been described from southwest China, and some of these, as Zhang
Wen-Tang (1987) suggested, may be synonymous and should be merged. Amongst these, two
genera that are probably synonymous are Shizhudiscus and Hupeidiscus , which were defined
principally on the characters of their pygidia. The differences in pygidial morphology, in the opinion
of Zhang Wen-Tang, however, are not such as to warrant generic status. If this view is accepted,
Shizhudiscus would have to be regarded as a junior synonym of Hupeidiscus. The cranidia of the two
genera are very similar, though there are some minor differences. Whether or not to synonymize
these two genera depends upon how to evaluate similarities and differences, and here ontogenetic
studies can provide additional sources of data. In general terms (Robison 1967; Ludvigsen and
Chatterton 1980), ontogeny has proved very useful in providing reliable criteria for the assignment
of some trilobite taxa and in determining relationships between groups.

In the present paper we show remarkable changes in the development of Shizhudiscus
longquanensis , especially in the pygidium with its distinct pleural furrows, fine tubercles, and the
strong axial spikes that appear first on the transitory pygidium, and are present in the holaspis on
the second and third thoracic segments. Such pronounced morphological changes should provide
many reliable indicators for assessing the relationship between S. longquanensis and other closely
related eodiscid groups.

Unfortunately, scarcely anything is known about the ontogenetic development of Hupeidiscus , in
which the adult pygidium does not have pleural furrows. We therefore retain the name Shizhudiscus
in the present study, rather than synonymizing it with Hupeidiscus , in view of the relatively limited
number of characters of taxonomic value common to both, as so far known. The proposal of Zhang

[Palaeontology, Vol. 36, Part 4, 1993, pp. 785-806, 2 pis.]

© The Palaeontological Association

786

PALAEONTOLOGY, VOLUME 36

Wen-Tang seems reasonable, and further investigations of eodiscid ontogeny should help to resolve
the problem.

The rich and specialized eodiscid fauna of southwest China is largely contemporaneous with the
Lower Cambrian high phosphate concentration episode described by Cook and Shergold (1984), with
which this local diversification is associated. Zhang & Clarkson (1990) gave a detailed description
of the structure and meraspid to holaspid development of the eyes of the Chinese eodiscids
S. longquanensis and Neocobboldia chinlinica , and their meraspid to holaspid development. We also
discussed the phosphatic preservation of the material and presented an account of the locality and
stratigraphy of the section from which the material came. All the Shizhudiscus specimens came from
a single lens within the upper part of the Lower Cambrian Shuijintuo Formation at Longquanxi in
Pengshui, Sichuan, and they were regarded as belonging to a single more-or-less contemporaneous
population. No material other than that from this single lens was included, thus avoiding confusion
with different morphs or shifts in modal size that may have developed within a time span (see
Sheldon (1988)). The same approach has been taken here, and all the phosphatized specimens of
S. longquanensis used in the present study were likewise retrieved from a single horizon. A few
cranidia (PI. 2, fig. 6) belonging to another eodiscid occur at this level, these are strikingly different
and easily distinguished from S. longquanensis.

Here a virtually complete growth series of S. longquanensis is described. Though all the specimens
are disarticulated, the strikingly well-preserved material provides a firm basis for assessing the
relationship of this genus to other eodiscids, and indeed the affinities of the eodiscids with other
trilobite groups. In addition we discuss moult stages and growth instars of Shizhudiscus , the
enrollment mechanism of the adult exoskeleton, and the early development of its eyes (as
complementary to our earlier paper, Zhang & Clarkson 1990), and its possible life-cycle pattern.

Material. All specimens figured or discussed in this paper are now housed in the Chengdu Institute
of Geology and Mineral Resources, Chengdu, People’s Republic of China.

INSTARS AND MOULTING

Trilobites grew during successive moults, and it is probably the case that the majority of trilobite
specimens found are exuviae. In some instances, and particularly in fine sediment, individuals of a

EXPLANATION OF PLATE 1

Shizhudiscus longquanensis S. G. Zhang and Zhu. Lower Cambrian; Pengshui, Sichuan. Figs 1-12; pygidia.
1, PSO 5610; (MOa) with three axial segments, posterior two bearing spines, x 167. 2, PSO 5616; lateral view
of (M06), showing a further ring added behind the two spinose rings, x 135. 3, PSO 5621 ; oblique ventral
view of (MOa), showing W-shaped outline of the posterior margin, where the doublure is virtually absent,
x 167. 4, PSO 5580; oblique dorsal view of (MOc), with two further rings behind the spinose rings; the first
thoracic segment appears to be nearly ready to separate from the transitory pygidium. 5, PSO 5561 ; lateral
view of (M2); the axis behind the two spinose rings bears about four segments and the pleurae have five ribs
with tubercles (arrowed) behind the interpleural furrow; a tubercle (arrowed) has appeared on the pleural
rib, x98. 6, PSO 5615; oblique ventral view of (M3), x68. 7, PSO 5605; posterodorsal view of (M06),
showing the W-shaped posterior border with fine ridges and an additional ring behind the spinose rings,
x 133. 8, PSO 5620; ventral view of (HO), with the third thoracic segment attached, x 65. 9, PSO 5617;
oblique-lateral view of (M2), with the first ring bearing a single spine, and about 5 rings behind it on the axis;
the paired tubercles on the axis are indicated by a white arrow, and the tubercles on the pleural ribs by a
black arrow, x 103. 10a-6, PSO 5614; 10a, oblique posterolateral view of (H2); seven axial rings can be
recognised by means of the paired tubercles, x 70; 106, detail of the tubercle on the pleural rib; an aperture
may be present in the centre of the tubercle, which is surrounded by fine granules, x 1 170. 1 la-6, PSO 5613;
1 la, right articulating facet of a holaspid, x 146; 116, details of tubercles and granules on the pleural rib and
furrow, x 228. 12, PSO 5619; detail of the two spines on the axis of (Ml), x 188.

PLATE 1

ZHANG and CLARKSON, Shizhudiscus

788

PALAEONTOLOGY, VOLUME 36

single species may be found together in all stages of growth. In such cases the ontogenetic
development of the species can usually be worked out by studying the specimens in gradational size
series. The terms protaspid, meraspid, and holaspid (Raw 1925) are commonly used to describe the
main stages of growth in trilobites. The terms anaprotaspid, metaprotaspid, and paraprotaspid have
also been used, and as the ontogeny of trilobites has become increasingly well known many new
descriptions, often with new terminology, have become available. The use of two sets of terms has
seemed to some authors a potential source of confusion and Edgecombe et al. (1988) and
Chatterton et al. (1990) felt that these terms have little value for comparing trilobite ontogenies
across the whole group. They preferred to describe ontogenies in terms of protaspid, meraspid, and
holaspid stages alone. This, it is agreed, would bring uniformity to the study of trilobite ontogenies
and would simplify comparisons of the development of different trilobite groups. These terms are
available for most, if not all trilobites, even if they are not necessarily homologous.

Whereas the use of such a standard terminology is highly desirable, it is not always possible to
put it into practice. In some trilobites, disarticulated elements cannot always be assigned to a specific
growth stage, especially where only parts of the ontogeny are known; likewise problems may arise
where the ontogeny of the trilobite is unusual. Where the growth stages are represented by complete
individuals it is possible to use the number of free thoracic segments present as a standard in
determining meraspid instars, as originally suggested by Barrande (1852). Thus Degrees 0, 1, 2, 3
etc. define the meraspid’s developmental stage by denoting the number of articulated thoracic
segments. Each degree, however, may include more than a single moulting episode, and in some
cases more than one segment may be released from the transitory pygidium at a single moult
(Whittington 1957, 1959; Feist 1970; Pabian & Fagerstrom 1972).

On the other hand, the increase in the number of axial rings appears to relate closely to growth
instars. It thus seems reasonable to subdivide the protaspid and meraspid periods into stages or
substages according to the number of rings in the rachis of the protopygidium or transitory
pygidium. This same idea has been used by Snajdr (1981) in elucidating the ontogenetic
development of Scharyia. Tripp and Evitt (1983) defined substages within the protaspid stages 2 and
3 of Dimeropyge virginiensis , and used clear and simple symbols to describe the changes in shield
outline and the increase in number of tubercles on the outer surfaces of the cranidia. Muller and
Walossek (1987) likewise adopted concise symbols for subdividing meraspides and holaspides of
Agnostus pisiformis into stages, using overall size and configuration of the tagmata, and the degree
of development of the excellently preserved appendages. This was particularly useful because
agnostids have only two thoracic segments.

Within a single stage there may be more than one moult, but between two connected moults, i.e.
from one ecdysis (E) to the next (as in Henningsmoen’s (1975, p. 180) intermoult cycle) only a single
stage need be defined morphologically. There could be, in addition, some bias in interpretation
depending upon whether the specimens examined were moulted exuviae or dead carcasses, and it
would be desirable, where possible, to be sure of this before assigning them to stages. Thus Tripp
and Evitt’s (1983) specimens were probably mainly exuviae, and were assigned to moults within a
stage. In Muller and Walossek’s (1987) study, on the other hand, the carcasses were preserved

EXPLANATION OF PLATE 2

Shizhudiscus longquanensis S. G. Zhang and Zhu; 1—4, PSO 5718; intact (M06) cephalon. 1. laterodorsal view
of right-hand side, xll7. 2, detail of 1 showing the right librigena with three small lenses, x 364.
3, laterodorsal view of left hand side, x 1 17. 4, detail of 3 showing the right librigena and small lenses, x 364.
5-7, Hypostomata of an unknown ?polymerid trilobite; 5-6; PSO 5991 ; 5, ventral view, x 135; 6, oblique
posteroventral view, x 165. 7, PSO 5990; laterodorsal view, x 94. 8, PSO 5194; oblique lateral view of a
cranidium of an unknown eodiscid trilobite, x 122. All from Lower Cambrian, Pengshui, Sichuan;
9, ZB9Ca001 ; Shizhudiscus sp. Lower Cambrian, Zhenba, Shaanxi. Lateroposterior view of an enrolled
(M0) exoskeleton, x 140.

PLATE

ZHANG and CLARKSON, Shizhudiscus

790

PALAEONTOLOGY, VOLUME 36

text-fig. 1. Plots of measurements for 6 cephala and 177 cranidia of Shizhudiscus longquanensis S. G. Zhang
and Zhu; Lower Cambrian, Pengshui, Sichuan; showing an approximately isometric growth pattern. Possible
instars occur within the three growth periods. The Arabic numerals and letters correspond to the specimens

illustrated in Text-figures 4 and 6.

complete with appendages and may have conserved the results of more than one moult, though it
is also likely that the differences between adjacent stages, as defined by Muller and Walossek merely
reflects the variation in morphology of an individual with growth, from one ecdysis to the next.

It is likely, as Henningsmoen (1975) suggested, that the majority of remains found are exuviae.
Most of the specimens used in the present study, occurring as they do in a disarticulated state and
in densely packed associations (Zhang and Clarkson 1990, Text-fig. 3a), are probably exuviae. Only
a few, possibly the five nearly complete protaspides, and more certainly specimens illustrated in
PI. 1, fig. 8 and PI. 2, figs 1-4, may have been the remains of dead animals (though they became
disarticulated after death). The much more common exuviae record the ‘instantaneous’
morphologies proper to each moulting event, i.e. stage E in Henningsmoen’s (1975, p. 180) division
of the intermoult cycle for trilobites. Our analysis has relied very largely on these.

ZHANG AND CLARKSON: LOWER CAMBRIAN EODISCID ONTOGENY

791

text-fig. 2. Plots of measurements for 203 pygidia of Shizhudiscus longquanensis S. G. Zhang and Zhu; Lower
Cambrian, Pengshui, Sichuan; showing nine possible instars (overlapping to some extent), within the three
growth periods. The Arabic numerals correspond to the specimens illustrated in Plate 1.

Quantitative definition of instars

In many previous studies of trilobite ontogeny the instars have been defined quantitatively (Palmer
1957, 1958, 1962; Hunt 1967; Robison 1967; Jell 1975a; Romano 1976; Busch and Swartz 1985;
Brezinski 1986).

In order to analyse statistically how Shizhudiscus longquanensis increases in linear dimensions
during ontogeny we measured two parameters. These were the length (sag.) from the anterior to the
posterior margin, and the width (trans.) between the palpebral lobes; this was effected in five almost
complete protaspides (Text-fig. 3a-h), onecephalon (PI. 2, figs l^f), and 1 78 cranidia (Text-fig. 4a-l)-
Because these two dimensions could be measured with great accuracy in the isolated
specimens, even some incomplete or damaged specimens could be included. In the scatter diagram
(Text-fig. 1), the 183 points representing all cephala and cranidia suggest an approximately
isometric increase, but there is no more than a hint of clusters corresponding to possible moult
stages. It is not possible therefore to define the growth stages in this way. The relatively even spread
of points denotes that changes in the morphology of the cranidium or cephalon were fairly gradual,
so these likewise give no help in defining instars.

Whereas ontogenetic changes in the 203 pygidia of S. longquanensis which we have measured are
gradual, there are particular characters that can be traced all the way from their earliest stages to
their final form. Of these the most striking are the two axial spines which form on the protopygidiuin
and move progressively forward, and are eventually released anteriorly to form the dorsal ‘thorns’

792

PALAEONTOLOGY, VOLUME 36

of the last two thoracic segments. These can be used as developmental markers in the same way that
Stubblefield (1926), and Fortey and Owens (1991) traced successive stages of the ontogeny of
Shumardia pusilla by means of the long macropleural spines. We have been able to recognize ten
kinds of pygidium (PI. 1, figs 1-12), excluding the protopygidium, and have assigned these to ten
corresponding growth stages (eleven with the protopygidium) (Text-fig. 2). There may in fact have
been more growth stages; individuals corresponding to younger instars seem not to have been
preserved. The scatter diagrams (Text-figs 1-2) show a fairly even spread of points. Thus the size
range for any one developmental stage (of the cranidium or pygidium) must have been quite
variable, for there is no clear pattern of instar peaks. It has not proved possible to identify distinctive
growth stages or instars, either using morphological features of the exoskeleton or statistical
analysis. This accords with Zhang’s (1989) equivalent investigation of the ontogeny of Neocobboldia
chinlinica , which likewise lacks instar peaks, especially in the adults. In both these eodiscid trilobites
therefore, the size ranges of each instar overlap, and this may be a feature common to eodiscids
generally. This lends support to Sheldon’s (1988) argument that it is not possible to define instars
on the basis of size alone, and we agree also with his view that in some previous studies too few
specimens have been used for meaningful statistical analysis of size-frequency distribution during
ontogeny.

As ontogeny progressed, the limited evidence of instar groupings seen in the early larvae is lost,
as clearly shown by the meraspid specimens in Text-fig. 2. The main reason for this is the shape and
size at any one time, of the transitory pygidium. The three thoracic segments were formed one after
the other during the meraspid period, being generated and then released forward from the transitory
pygidium. The transitory pygidium thus varied markedly in dimensions, showing in the scatter
diagram as a broad range of individuals for a single instar, and individual groups of different instars
would overlap. This accords with the explanation for overlap of trilobite instars in size-frequency
distributions previously given by Palmer (1958), Robison (1967), Sheldon (1988) and Zhang (1989),
all of which have proved equally suitable for eodiscid trilobites.

ONTOGENY OF SHIZHUDISCUS LONGQUANENSIS

Whereas it is possible to define at least nine instars for the pygidium of S. longquanensis on the basis
of axial rings and thoracic segments, no such discrimination is possible for the corresponding instars
of the cranidia. We have no complete meraspid or holaspid exoskeletons which might show directly
which developmental stage for the cranidium relates to a known transitory pygidium. Moreover,
morphological changes in the cranidia with ontogeny are very gradual, and size-overlapping makes
it impossible to distinguish potential instar peaks for the cranidium by quantitative measurement
alone. The succession changes from one ontogenetic stage to the next. Despite this limitation, the
enrollment mechanism of this eodiscid trilobite that we have been able to demonstrate may allow,
to some extent, the correspondence between some of these isolated cranidia and pygidia to be
traced. The disappearance of the bacculae and the genal spines may also be used in the subdivision
of growth series. Most disarticulated exoskeletons therefore can be referred, with reasonable
confidence, to meraspid and holaspid periods, and even to particular growth stages. The ontogenetic
series of S. longquanensis is summarized below.

Protaspid period

Stage 2 (P2). This stage is represented by five subovate, entire exoskeletons (Text-figs 3a-h, 7a-c);
these are all late protaspides. They are moderately convex transversely, but much more strongly
convex sagitally, since the protopygidium is strongly bent downwards. They range in length from
0-22 to 0-25 mm, and in width from 0-20 to 0-23 mm. The axis, which may consist of five(?) or six
segments is defined laterally by deep but relatively wide axial furrows. The glabella, tapering
forwards, has three glabellar lobes, weakly convex and separated by faint transverse furrows. The
fourth lobe on the glabella is the occipital ring, which forms a high, rounded node, often indented

ZHANG AND CLARKSON: LOWER CAMBRIAN EODISCID ONTOGENY

793

text-fig. 3. Shizhudiscus longquanensis S. G. Zhang and Zhu; Lower Cambrian, Pengshui, Sichuan, a-h,
protaspides (P2). a-b, PSO 5982; dorsal and lateral views of an almost complete exoskeleton with the genal
tubercles (arrowed) preserved, x 175. c-G, PSO 5981 ; c, dorsal view; the tubercle is indicated by the white
arrow, and the anterior border tubercle by the black arrow, x 175. D, oblique lateral view, x 175. E, detail of
D, showing fine granules arranged in rows on the outer surface of the right side of the cephalon, where as yet
the facial suture and librigena have not differentiated, x 482. F, lateral ventral view, x 175. G, postero-ventral
view, showing details of granules and fine ridges, x 350. H, PSO 5983; oblique lateral view of posterior part
of the exoskeleton, ornamented with fine granules, x 295.

by a short longitudinal furrow. Behind the occipital ring is a rather convex axial part, which curves
sharply downward to meet the posterior border. The rachis is more strongly curved than the
adjacent regions of the protopygidium (borders and pleural areas), and tucks into the posterior
sinus so that the rear border of the protopygidium is W-shaped. There are no transverse furrows
visible on the protopygidial axis, but its length would suggest that two rings were present. The
anterior margin is nearly straight, and the anterior border, defined by a rather shallow anterior
border furrow, is very narrow, bearing two anterior tubercles. The doublure is likewise very narrow.
The two elliptical fixigenal lobes are prominent, located on the anterior part of the pleural region.
Behind these lobes are a pair of rounded bacculae, positioned in the posteropleural areas. The
exoskeleton curves ventrally to form the doublure along the free border on the ventral side, so that

794

PALAEONTOLOGY, VOLUME 36

text-fig. 4. Shizhudiscus longquanensis S. G. Zhang and Zhu; Lower Cambrian, Pengshui. a-l; cranidia.
a, PSO 5210; oblique posterodorsal view of (MO b), x 122. b, PSO 5202; ventral view of (M06), x 122. c, PSO
5209; ventral view of (HO), showing the genal doublure, x235. d, PSO 5198; antero-dorsal view of (MOa),
showing the anterior tubercle (arrowed), x 1 18. e, PSO 5204; lateral dorsal view of (M06); arrow indicates the
anterior border tubercle, x 102. f, detail of E, showing granules arranged in a concentric ellipse surrounding
the fixigenal lobe, x 300. G, detail of E, showing the tubercles (arrowed) on the glabella, x 300. h, PSO 5165;
oblique-lateral view of (Ml)?, x 78. i, PSO 5211; dorsal view of (M3), x 56. j, PSO 5185; dorsal view of mature
cranidium, x 34. k, detail of j, showing right genal angle and fine granulation on the posterior fixigenal area,
x 95. l, PSO 5167; fine granules on the anterior border of (M3), x 245.

ZHANG AND CLARKSON: LOWER CAMBRIAN EODISCID ONTOGENY

795

the doublure along the posterior margin also has a W-shaped outline. In ventral view the doublure
is quite inflated and widest at the posterior margin, but tapers rapidly anteriorly to become very
narrow along the anterior margin. On the external surface of the cuticle is a fine reticulate sculpture
composed of fine ridges and granules. The fine ridges are found on the axis, the inner parts of the
fixigenal areas of the cephalic region and the doublure. The granules, however, are concentrated
chiefly on the outer parts of the fixigenal areas and on the protopygidium. A pair of small tubercles
is located on the posterolateral borders (Text-fig. 3a-c), which develop during subsequent growth
stages into short genal spines. The articulation between the cephalic region and the protopygidium
is not yet developed at this stage, but its position can be inferred from the location of the bacculae,
genal tubercles, and occipital ring.

Meraspid period

In all trilobites, meraspid degree 0 (MO) is defined upon the first appearance of articulation. In the
case of S. longquanensis the smallest isolated transitory pygidium has only three axial rings. More
axial rings are gained with growth but none is released as a thoracic segment until the axis consists
of five rings. This situation is the same as in the transitory pygidia of Neocobboldia chinlinica. This
emphasizes the fact that the MO transitory pygidia of eodiscids gain axial rings step by step before
the first thoracic tergite forms; the meraspid degree 0 is here divided into substages corresponding
to the number of axial rings. Thus the three metaprotaspid stages (MpO-2) of N. chinlinica should
be revised as three meraspid substages (MOa-c). By the same reckoning, the meraspid period of
S. longquanensis is divided into seven stages or substages (MOa, MOfi, MOc, Mia, Mlfi, M2, M3),
defined on the basis of rings added to the pygidial axis and the release of thoracic segments
anteriorly. Since the exoskeletons are incomplete, however, the corresponding meraspid cranidia
cannot be referred to particular growth stages. These range in size from 018 to 0-52 mm in length
and from 0-30 to 0-87 mm in width. There are some evident, though gradual changes in morphology
in progressively larger meraspid cranidia (Text-fig. 4h-i). Thus the anterior margin loses its
indentation and becomes outwardly curved ; this is accompanied by the broadening of the anterior
border. As the fixigenal areas continue to elevate, each of the fixigenal lobes and the adjacent
bacculae merge into a single high lateral lobe. The glabellar lobe becomes somewhat inflated yet
remains of relatively low convexity and lies well below the level of the fixigenal lobes. The axial
furrows become increasingly deep and narrow and the posterior furrow appears. On both sides of
the cranidium the sinus representing the last librigena has become relatively long and wide,
suggesting the addition of more lenses to the visual surface. The short genal spike underwent a
contraction and all that remains of it finally is a rounded genal angle (Text-fig. 4c).

Substage (M0a). It is at this growth stage that the first articulating hinge develops, so that the
transitory pygidium becomes free; all material is thus represented by disarticulated cranidia and
transitory pygidia. The cranidia (Text-figs 4b-d, 7d) are subtrapezoidal in outline, with lengths
ranging from 0T8 to 0-25 mm, and widths from 0-30 to 0-38 mm. The axis tapers forwards and is
defined by deep and narrow axial furrows. Both the paired, elliptical, fixigenal lobes and the
surrounding fixigenal areas are strongly elevated and peaked. Although the glabellar and occipital
furrows are very weak, the four glabellar lobes that they define can be faintly distinguished. Of these
the front glabellar lobe is elliptical in outline and moderately convex, the next two are progressively
wider and more convex, and the most posterior lobe forms the widest part of the axis. The occipital
ring is remarkably convex and bears a strong spine. The anterior border is somewhat curved
backwards centrally, and bears two distinct anterior tubercles (Text-fig 4d) in front of the large
fixigenal lobes.

These lobes by this stage are highly convex, and the rounded bacculae posterior to them have
become prominent. There is an ocular ridge extending laterally from each side of the frontal
glabellar lobe, fused with the palpebral lobe to form a forwardly curved broad ridge between the
anterior border and the fixigenal lobe. By this stage the facial suture has moved from the ventro-

796

PALAEONTOLOGY, VOLUME 36

lateral side to the anterolateral surface of the cephalon, so that there is an evident sinus where the
librigena is lost. The lateral border has broadened and is defined by a shallow lateral border furrow,
but the posterior border furrow is not as yet present. The genal spine is short and posterolaterally
directed. In ventral view the anterior part of the doublure is very narrow, as is the librigenal
doublure (Text-fig. 4b). The posterolateral corner of the doublure, however, carries a broad,
triangular doublure with tuberculate ornamentation (Text-fig. 4c). There is by now a substantial
change in surface sculpture; the fine reticulation of the outer surface has been replaced by a
granulose texture, concentrated particularly on the anterior and lateral borders, the fixigenal areas,
as concentric ellipses on the fixigenal lobes, and clustered towards the tip of the occipital spine. On
the axis and the posterior fixigenal areas some other, rather large granules are scattered.

The transitory pygidia (PI. 1, figs 1, 3) are, like the cranidia, subtrapezoidal in outline with a
pronounced ‘larval notch’. For this stage they range from 0T6 to 0-23 mm in length and 0-30 to
0-39 mm in width. The axis, consisting of three segments, is very strongly elevated, with the axial
furrows hardly defined. The axial rings likewise are so tightly fused that they can only be
distinguished by the two strong spines borne by the rear two rings and by the rather weak transverse
axial furrow separating the first ring from the second. The two axial spines rise abruptly and are
nearly straight, with only their apices pointing backwards. The lateral and posterior borders are
moderately convex, and separated from the pleural areas by shallow border furrows. At the
anterolateral corners of the pygidium is a pair of swellings. These develop into lateral pleural lobes
when the segment bearing them is released to form the first thoracic segments. Two further pairs
of swellings, more posteriorly located, define the positions of other pleural ribs. The outer surface
of the exoskeleton is covered with fine granules, most of which are concentrated upon the posterior
pleural regions, but fine ridges are only to be found upon the borders. At this stage in development,
the doublure beneath the posterior border is hardly in evidence (PI. 1, fig. 3).

Substage (MOfi). The cranidia (Text-fig. 4a, e-g) are morphologically similar to those of the
preceding substage, and are mainly distinguished by their size and their correspondence with
pygidia of equivalent dimensions on the scatter diagram (Text-fig. 1). Their lengths range from 0-25
to 0-34 mm, and widths from 0-38 to 0-47 mm. The only appreciable changes are that the fixigenal
lobes have become still higher, and the occipital spine is more elongated. The ventral morphology
is like that of preceding substage. Likewise the granulation on the surface of the exoskeleton is now
more clearly defined, and the concentric ellipses round the fixigenal lobes more pronounced and
regular (Text-fig. 4f).

The MOfi substage is much more clearly determined by the morphology of the pygidia (PI. 1, figs
2, 7). These range from 0-23 to 0 27 mm in length and 0-39 to 0-40 mm in width. Whereas they retain
many morphological characteristics from earlier stages they differ in that an additional axial ring
has appeared at the rear of the axis, a pair of pleural ribs has been added posteriorly, and the
transverse axial furrow between the first and second rings has now become distinct.

Substage (MOc). During this substage the fifth axial ring appears posteriorly and increases markedly
in size, while the first transverse axial furrow and the first pair of interpleural furrows join to form
a continuous faintly incised furrow, defining the rear edge of the first thoracic segment (which will
soon separate from the transitory pygidium). This first thoracic segment includes the first axial ring,
which lacks a spine (PI. 1, fig. 4; Text-fig. 7f). It may well have still been connected to the
articulating half-ring of the second segment (still part of the transitory pygidium (by an arthrodial
membrane.

Stage {Ml). As the first thoracic segment is released, the sixth axial ring and the fifth pair of pleural
ribs develop at the rear of the transitory pygidium. This now has five axial rings and four pleural
ribs, and such morphology defines substage (Mia; PI. 1, fig. 12). During subsequent growth the
pygidium gains a new axial ring (the seventh), and a new pair of pleural ribs but at this stage no
further thoracic segment is released. There is still only one free thoracic tergite. We may therefore

ZHANG AND CLARKSON: LOWER CAMBRIAN EODISCID ONTOGENY

797

assign pygidia which have reached this stage of development (PL 1, fig. 5, Text-fig. 7g) to substage
(Ml b).

Stage {M2). As development proceeds the second transverse axial furrow, together with its
associated interpleural furrows, form a new articulation separating off a second thoracic segment
from the transitory pygidium. The release of this second segment accompanies the appearance of
the eighth axial ring, defining stage (M2). The transitory pygidium now bears only a single spine
(PI. 1, fig. 9, Text-fig. 7h).

Stage {M3). Following upon this, the pygidium gains the ninth axial ring and a new pair of pleural
ribs, and with further development the final articulation forms, separating the third thoracic
segment from the pygidium (PI. 1, fig. 6). This stage (M3) is represented by spineless pygidia with
six axial rings and five pairs of pleural ribs. These indicate clearly that after the third thoracic
segment was released at the M3 stage, no further axial rings or pleural ribs appeared.

Besides changes in the axial rings, pleural ribs, and thoracic segments, to which we have referred,
there are other minor changes. The W-shaped posterior margin of the transitory pygidium becomes
semicircular, but with numerous small serrations along a free margin. The granules on the cuticle
do not seem to increase in size but a few somewhat larger granules (tubercles) appear in the early
meraspid (PI. 1, fig. 4) arranged in pairs on each axial ring and in rows on each pleural rib, just
behind the interpleural furrow (PI. 1, figs 5-9). As development proceeded the whole exoskeleton
became progressively more convex.

Holaspid {HO) period

This begins when the third thoracic segment is liberated from a pygidium bearing six axial rings and
five pairs of pleural ribs. Most holaspid cranidia and pygidia show comparatively few reliable
criteria which would enable them to be assigned to particular instars, and only the axial rings and
paired pleural ribs can be used to distinguish early holaspid pygidia (HO) from those of later
holaspides (HI, H2...).

The cranidia (Text-figs 4j-k, 7i) of the early holaspid period share many morphological features
with those of later meraspides. As the cranidia become progressively larger and more fully inflated
the anterior border becomes wide and moderately convex, while the anterior border furrow becomes
wide and deep. The fixigenal areas attained a considerable elevation, and while the palpebral lobe
became distinct, the ocular ridge all but disappeared. The glabella is by this stage very much tapered
(and less elevated) forwards, but the occipital ring and the rear part of the glabella are much
swollen. The three glabellar lobes are now only faintly defined by scarcely visible furrows. While the
occipital spine remains prominent, its growth rate seems to have reduced and it is relatively shorter
than in the earlier stages.

The pygidium of the early holaspis bears six axial rings and four pleural ribs. With further growth
it gains a tenth and final axial ring, and a further pair of pleural ribs. Thereafter, there are no rings
or ribs added. There are some supplementary changes in the holaspid pygidia though these cannot
be used to distinguish growth instars. Thus, as the rachis expanded the axial furrows became quite
faint, but the seven axial rings can still be distinguished by their paired axial tubercles, which
become very prominent. Likewise, the tubercles arranged in rows on the pleural ribs just behind the
interpleural furrows are distinct. With inflation of the pleural area the pleural furrows became
shallow, but still visible, while the axial furrows and the border furrows became more deeply incised.
The border, by the later holaspid stage is relatively narrow, and its free margin weakly serrate. The
fine granules on the outer surface of the pygidium seem to have remained much the same size and
are mainly concentrated on the border and on the posterolateral and rear parts of the pygidium.

Thoracic segments

The thoracic tergites were probably released one by one from the anterior part of the transitory

798

PALAEONTOLOGY, VOLUME 36

text-fig. 5. Shizhudiscus longquanensis S. G. Zhang and Zhu; Lower Cambrian, Pengshui, Sichuan, a-j;
thoracic segments, a, PSO 5962; dorsal view of the first thoracic segment of Ml, x 91. b, PSO 5960; ventral
view of the first thoracic segment of M3, x 81. c, PSO 5952; oblique dorsal view of a second (or third) thoracic
segment of M2, x 80. d, PSO 5959; ventral view of a second (or third) thoracic segment of a holaspid, x 1 17.
E. PSO 5954; lateral anterior view of the first thoracic segment of HO, x 81. F, PSO 5955; dorsal view of the
first thoracic segment of M2, x 157. G, PSO 5957; ventral view of the left side of the first thoracic segment of
H, x 91. H, PSO 5958; dorsal view of a first thoracic segment of holaspid, showing the articulating facet, x 142.
i, PSO 5938; ventral view of a second (or third) thoracic segment of holaspid, showing the doublure under the
axial ring, x 190. J, PSO 5969; dorsal view of a second (or third) thoracic segment of holaspid, showing the

left articulating facet, x 160.

pygidium, as we have described (though it is possible that more than one was liberated at one time).
These are mostly found isolated, and only a few segments were still attached to the transitory
pygidium (PI. 1, fig. 8), or ready to separate from it. (PI. 1, figs 4, 6). It is difficult, therefore,
confidently to assign any one segment to a particular growth stage, though simple comparisons of
the width of any thoracic segment with a pygidium of approximately the same width allows an
approximation.

ZHANG AND CLARKSON: LOWER CAMBRIAN EODISCID ONTOGENY

799

The first of the three thoracic segments does not bear an axial spine (Text-fig. 5a-b, e-h). Such
segments, though representing a broad, gradational size series in our material, are readily
distinguished. The second and third thoracic segments have a prominent axial spine (Text-fig.
5c-d, i— j ; Text-fig. 7i); it is not easy to distinguish the second from the third, for they are
morphologically very similar. Before articulating, the axial spine on the presumptive third segment
was the stronger and more highly curved, and it might be expected that this difference would be
retained after release. Yet there is no convincing indication in the isolated spinose segments that this
is so, and whether any one of these is a second or third segment remains unknown. In the

text-fig. 6. Shizhudiscus longquanensis S. G. Zhang and Zhu. Explanatory drawings of the cephalon illustrated
in PI. 2, figs 1-4. a, right librigena with visual surface present, showing one almost circular lens and two half-
lenses. b, similar initial small lenses on the left librigena. f, fixigena ; 1, librigena ; p, palpebral lobe thick line

indicates suture line.

reconstruction (Text-fig. 7i), they are shown as identical, though it is recognized that they may differ
in minor ways.

All the thoracic segments, whether meraspid or holaspid, have virtually straight pleural edges at
front and rear, which acted as articulating hinges, presumably connected by arthrodial membrane.
The outer parts of the pleura curved down ventrally. Here arthrodial membrane would be lacking,
but the free parts formed a continuous protective wall when the trilobite was enrolled, and the
articulating facets permitted full enrollment. The orientation of any isolated thoracic segment is
readily determined by the articulating facets and the backwardly pointed axial spine where present.
Each segment has an arched articulating half-ring, which would fit neatly under the larger axial ring
of the preceding segment (or the occipital ring). The axial ring has a narrow doublure with a straight
free edge (Text-fig. 5b, d, i). It is likely that arthrodial membrane connected this edge to the front
of the adjacent half ring. In our small thoracic segments, the articulating facets are present though
not especially pronounced, and their surfaces bear fine granules (Text-fig. 5a, f). In larger
specimens, however, the articulating facet has become more clearly developed, and the fine granules
have extended into a series of fine parallel ridges (Text-fig. 5h, j). These may have been adapted for
tight gripping of adjacent parts when the exoskeleton was enrolled.

Eyes

In our previous study (Zhang & Clarkson 1990, pi. 1, fig. 1 a-b), the smallest librigena of
Shizhudiscus longquanensis that we found bore nineteen lenses (including half-lenses). We estimated
that this came from a late meraspid or early holaspid, and now that the ontogeny of this species is
known in detail, we are able confidently to refer this librigena to the later meraspid (M3) stage, on
account of its dimensions.

800

PALAEONTOLOGY, VOLUME 36

text-fig. 7. Shizhudiscus longquanensis S. G. Zhang and Zhu. Reconstruction of successive stages in ontogeny.
a-c, protaspid in dorsal, ventral, and oblique-lateral views (cf. Text-fig. 3a, f, b). d, cranidium and pygidium
of (MOa) stage (cf. Text-fig. 4a-b, PI. 1, figs 1-2). e, cranidium of (M3) stage (cf. Text-fig. 4i). f-h, meraspid
pygidia in dorsal view: f, (MOc) stage (cf. PI. 1, fig. 4), G, (Ml) stage (cf. PI. 1, fig. 5), h (M2) stage (cf. PI. 1,
fig. 9). i, attempted reconstruction of adult exoskeleton (cf. PI. 1, fig. 10a, Text-figs 4j, 5a, c). This is tentative,
since exact matching of the developmental stages for the various tagmata is uncertain, as is the comparative
morphology of the second and third thoracic segments. Scale bars: a-c = 1 mm; d-h = 2 mm; i = 5 mm.

In the course of this present work we found a single specimen of S. longquanensis with both eyes
intact, representing a much earlier stage than any described hitherto. This adds valuable detail to
what is already known about eodiscid eyes. Such preservation is very rare, for even in some enrolled

ZHANG AND CLARKSON: LOWER CAMBRIAN EODISCID ONTOGENY

801

individuals (from other phosphatized material of Shizhudiscus sp. collected at Zhenba, Shaanxi)
the librigenae are absent (PI. 2, fig. 8). Our single specimen (PI. 2, figs 1 1-4; Text-fig. 6a-b) is
recognizable as a degree 0 meraspis (M0/>) and is probably the remains of a dead individual rather
than an exuvium (6a-d). Whereas we lack the intermediate stages between MO/) and M3, it is still
possible to determine the basic steps of the development of the eye.

Each of the librigenae bears a fully developed single lens with two half lenses on either side; these
become the central lenses of the uppermost horizontal row in later growth stages. These lenses are
widely separated from each other, and the lenses of the second horizontal rows are added below the
spaces between them. Thus from the earliest stage of development, the pattern of hexagonal close
packing, fundamental to trilobites (Clarkson 1975) is already in evidence. While the MO b eye has
one complete and two half lenses, the M3 eye has nineteen (including six half-lenses, arranged in
two horizontal rows of complete lenses with a further row of half-lenses below (Zhang & Clarkson
1990, p. 222). These are arranged in two horizontal rows of complete lenses with a further row of
half-lenses below. In all trilobites new rows are always added below existing ones and expand
laterally. In the eodiscids the newly formed rows consist of half-lenses which only become complete
lenses at the next moult. It is evident that the eyes must have expanded very rapidly at front and
rear after the M0/> stage, to form the elongated visual surface seen in M3.

Assuming only one moulting in an instar, at least four moults were passed through in the
transition from M06 to M3, so the increase of a single horizontal row required at least two ecdyses.
If the rate of increase was approximately constant, the first lens would be expected to have been
emplaced in the MOa stage, and possibly earlier.

The early emplacement of the lenses in eodiscids may be compared instructively with that in other
trilobites. For example, even in the schizochroal-eyed Phacopina the early stages of eye development
(Alberti 1972) are broadly the same. In both eodiscids and phacopids, in the earliest stages, a single
small lens appears which retains its position thereafter on the upper horizontal row, and a pattern
of hexagonal close packing develops as more lenses are emplaced. Furthermore, the initial lenses in
Shizhudiscus are widely spaced, as they are in the Carboniferous Paladin (Clarkson & Zhang 1991);
it is only later in development that the lenses become contiguous and the eye achieves a fully
holochroal condition. The initial stages of development of trilobite eyes generally, therefore, exhibit
patterns common to all.

Doublure and hypostomata

The doublure, as in Pagetia (Jell 1975; Whittington 1988) remains very narrow, but there may have
been, in addition a flat crescent-shaped rostral plate (possibly uncalcified), extending as far as the
anterior border furrow. This condition would be similar to that in olenelloids, and equivalent to the
construction postulated by Whittington for Pagetia. Both Whittington (1988) and Fortey (1990)
regard the eodiscid hypostome as natant, lying freely below the cephalon.

A few hypostomata (PI. 2, figs 5-7) are present in our material. They bear fine granules and ridges
and each has eleven radiating spines. These are figured here for completeness, but their morphology
is quite different from that of other eodiscid trilobites, such as those of Pagetia (Opik 1952; Jell
1975a), and Neocobboldia (Zhang 1989). The hypostomata figured here may therefore belong to an
unknown polymerid.

Cuticular sculpture

In many trilobites cuticular sculpture may change throughout ontogeny. Where preservation is
exceptionally good, a network of fine ridges may be seen covering the larval exoskeleton; each of
these cuticular polygons retains the shape of the epidermal cell below which secreted it. The cell
polygons may still be visible in the adult, as in the very thin cuticle of Homagnostus obesus described

802

PALAEONTOLOGY, VOLUME 36

by Wilmot (1990). More commonly the cell polygons fade and are replaced by tubercles or other
structures as in Paladin (Clarkson & Zhang 1990). This is also the case in the eodiscids Neocobboldia
and Shizhudiscus where reticulation is only present on the external surface of late (P2) protaspides.
The fine granules that appear later do not seem to bear any direct relationship to the epidermal cells.

Enrollment

Eodiscid trilobites undoubtedly possessed the ability to enroll, as shown by Jell’s (1975a, 19756)
illustrations of phosphatized Middle Cambrian Pagetia silicunda and Opsidiscus brevicaudatus. The
specimens were all tightly enrolled, with the cephalon and pygidium fitting snugly, and in some the
visual surface was still attached.

A single phosphatized specimen of Shizhudiscus sp., from the Lower Cambrian of Zhenba,
Shaanxi (PI. 2, fig. 9), reveals some details of the enrollment mechanism in this genus.

(a) The anterior border of the specimen, otherwise moderately rounded, is slightly indented
centrally, thus matching the W-shaped margin of the pygidium. The future first thoracic segment
is still welded to the transitory pygidium. Thus this enrolled individual can confidently be referred
to the late meraspid period (MOc), by direct comparison with S. longquanensis. More importantly
this specimen shows that full enrollment was possible even at this relatively early stage in ontogeny;
the anterior margin of the transitory pygidium acted as a hinge which would permit the two
exoskeletal parts to close up together. Agnostids (Jaekel 1909) likewise can enrol at meraspid degree
0 stage, as can trinucleids and raphiophorids, though must polymerids rely upon concerted
movements of many segments and transverse joints for enrollment.

(b) Although no suggestion of a future articulatory joint is yet evident in the five smallest
protaspides (P2) of S. longquanensis , its eventual position may readily be determined from the
location of the paired bacculae and the prominent occipital ring. At the P2 stage, the small
protopygidium is strongly curved downwards relative to the cephalon, but even if the protaspid
were flexible the protopygidium and the ventral side of the cephalon could never make contact, and
in any case would not match. The earliest stage at which true enrollment was possible is MO.

(c) Robison (1972) deduced a pelagic mode of life for the agnostids, on the basis of their
global distribution, and considered the unique articulatory system and mode of enrollment as
an adaptation to this life style. Muller and Walossek (1987) illustrated the enrollment mechanism
of Agnostus pisiformis from the first meraspid to holaspid stage, and clearly the style of enrollment
of the eodiscid S. longquanensis , in functional terms, has some characters in common with this.
Thus in both eodiscids and agnostids the first meraspides could fold up and close tightly by means of
a single articulation. As the thoracic segments were liberated from the transitory pygidium, so there
would be more space within the enrolled exoskeleton for soft parts and appendages, especially in the
three-segmented Shizhudiscus. In the opinion of Muller and Walossek ( 1987)^^05^ lived in a semi-
enrolled state when active, and the body could not be stretched out fully without damage. The
articulating facets in Agnostus are very narrow. In S. longquanensis , however, they are prominently
developed on the thoracic segments (Text-fig. 5f, h, j) and the pygidia (PI. 1, figs 10a, 1 la), as they
are in polymerid trilobites. This resemblance suggests that in contrast with the agnostids, the
eodiscids were able to stretch out in a fully extended attitude, as well as having the facility for
enrollment.

MODE OF LIFE

The morphology of eodiscid exoskeletons changes remarkably between protaspid and meraspid
periods, as seen particularly in the pygidia of Neocobboldia chinlinica (Zhang 1989) and Pagetia
ocellata (Shergold 1991). The two scatter diagrams (Text-figs 1-2) relate to size distribution of
cramdia and pygidia respectively. There is a distinct gap between the pygidia of protaspid degrees

ZHANG AND CLARKSON: LOWER CAM BRIAN. EODISCID ONTOGENY

803

F

- 50

- 40

- 20

- 10

0.5

T

1.0

5 protaspides
178 cranidia
203 pygidia

2.0 W(mm)

text-fig. 8. Histograms for 5 protaspides, 1 cephalon, 178 cranidia, and 203 pygidia of Shizhudiscus
longquanensis S. G. Zhang and Zhu; Lower Cambrian, Pengshui, Sichuan. F, frequency; W, width between the
palpebral lobes of the cranidium (or the cephalon), or width of the pygidium.

2 and 3, though no such gap is evident for equivalent cranidia. It is possible that the observed break
in pygidial size results from collection or preservational bias for protaspides of the size represented
by the gap and there is an evident contrast here with the pygidial size distribution of Neocobboldia
chinlinica (Zhang 1989, fig 2), where the scatter of points is continuous. While this might support
the idea of preservation failure for this stage in S. longquanensis , it could equally indicate a
substantial morphological saltation relating to an abrupt change in mode of life. Evitt (1961) noted
a striking difference in gross morphology between protaspides and meraspides of Isotelus , and
Chatterton (1980) emphasized how significant metamorphosis occurred during the protaspid period
or between protaspid and meraspid. A recent series of papers has highlighted this phenomenon
(Fortey and Chatterton 1988; Chatterton and Speyer 1989; Speyer and Chatterton 1989;
Chatterton et a/. 1990). Besides the changes in overall morphology in S. longquanensis across this
gap, two other striking modifications take place around the protaspid-meraspid transition; the
achievement of enrollment ability and the appearance of visual surface, with its first small lens. It
is likely that both these transformations were associated with a change in mode of life, or at least
represent a radical increase in functional capacity to adapt to the environment.

The protaspides lacked eyes, the exoskeleton was strongly curved, and the pygidium was sharply
bent down. Such morphology recalls that of the non-adult-like ovoid or inflated forms described by
Speyer and Chatterton (1989) as typical of planktonic larvae, and in our view the protaspides of
S', longquanensis were likewise planktonic.

Agnostid and eodiscid trilobites are regarded by many authors as sister-groups within the
Trilobita (Fortey 1990). Other workers, whose views are summarized by Shergold (1991), classify
eodiscids alongside the ptychopariids. Whereas the affinities of the eodiscids are still debated, one
possibility is that the agnostids derived from eodiscids by a neotenous retention of planktonic larval
habits. Such heterochrony would explain why agnostids are eyeless and have only two thoracic
segments. Perhaps even the absence of a calcified protaspis in agnostids may be part of the same

804

PALAEONTOLOGY, VOLUME 36

paedomorphic complex. If so, a close relationship between agnostids and eodiscids would be
favoured.

TAPHONOMY

No hypostomata in our material can be referred confidently to S. longquanensis. This may reflect
a poorly mineralized ventral exoskeleton, but more probably results from transportation and
sorting. Within the lens containing the trilobites, small-scale cross-lamination is apparent and
the densely packed exoskeletons must have been a consequence of current concentration. The
bell-shaped size-frequency diagram (Text-fig. 8) again suggests selective transportation and
preservation. It is not very likely, however, that the individuals lived very far away from where they
were buried, since delicate structures such as the eyes, fine granules on the cuticular surface and the
long spines are well-preserved. It is considered probable that the protaspides (P2), though few have
been preserved, likewise lived near their place of burial. Such rapid burial of many individuals tends
to confirm that most of the individuals used in this study belong to a single species.

Speyer and Brett (1986), working with Devonian trilobites, proposed a number of trilobite
taphofacies on the basis of preservation alone. Despite differences in age and hydrodynamic
properties, the life environment of 5. longquanensis may roughly be assigned to their Taphofacies 1.
Changes in life habit during ontogeny, such as we have proposed here, involve adaptation to new
ecological niches within the same taphofacies.

Acknowledgements. We thank the Sino-British Fellowship Trust, through the Royal Society of London, for
financial aid which enabled Zhang Xi-Guang to visit the University of Edinburgh, where this study was
initiated. Thereafter the project was supported by the National Natural Sciences Foundation of China (NSFC
49070073). We are grateful to Mr John Findlay (University of Edinburgh) and Ms Chen Ji-Yu (Chengdu
Institute of Geology and Mineral Resources) for assistance with SEM photography, and to Cecilia Taylor for
critically reviewing the manuscript. Comments from an anonymous referee much improved the quality of the
final text.

REFERENCES

alberti, h. 1972. Ontogenetische Studien von Phacopiden - Komplexaugen Trilobiten. Beitrdge zur
elektronenmikroskopischen Direktabbildung von Oberflachen, 5, 797-809.

barrande, j. 1852. Systeme silurien du centre de la Boheme. lere partie. Recherches paleontologiques, Vol. 1.
Crustaces : Trilobites. Prague & Paris. 935pp.

brezinski, d. k. 1986. An opportunistic Upper Ordovician trilobite assemblage from Missouri. Letbaia , 19,
315-325.

busch, r. m. and swartz, f. m. 1985. Molting and description of a new homalonotid trilobite from
Pennsylvania. Journal of Paleontology , 59, 1062-1074.

chatterton, b. d. e. 1980. Ontogenetic studies of Middle Ordovician trilobites from the Esbataottine
Formation, Mackenzie Mountains, Canada. Palaeontographica, Abteilung A, 171, 1-74.

— sr veter, d. j., edgecombe, G. d. and hunt, a. s. 1990. Larvae and relationships of Calymeninae
(Trilobita). Journal of Paleontology, 64, 255-277 .

- and speyer, s. e. 1989. Larval ecology, life history strategies, and patterns of extinction and survivorship
among Ordovician trilobites. Paleobiology, 15, 118-132.

clarkson, E. N. K. 1975. The evolution of the eye in trilobites. Fossils and Strata, 4, 7-31.

-and zhang, xi-guang, 1991. Ontogeny of the Carboniferous trilobite Paladin eichwaldi shunnerensis
(King, 1914). Transactions of the Royal Society of Edinburgh, Earth Sciences, 82, 277-295.

cook, p. and shergold, j. h. 1984. Phosphorus, phosphorites, and skeletal evolution at the Precambrian-
Cambrian boundary. Nature, 308, 231-236.

edgecombe, G. d., speyer, s. E. and chatterton, b. d. e. 1988. Protaspid larvae and phylogenetics of encrinurid
trilobites. Journal of Paleontology 62, 779-799.

evitt, w. R. 1961. Early ontogeny in the trilobite family Asaphidae. Journal of Paleontology, 35, 986-995.

ZHANG AND CLARKSON: LOWER CAMBRIAN EODISCID ONTOGENY

805

feist, R. 1970. Breviscutellum ( Meridioscutellum ) n. sg. (Trilobita) et son developpement larvaire. Geobios , 3,
41-73.

fortey, r. a. 1990. Ontogeny, hypostome attachment and trilobite classification. Palaeontology , 33, 529-576.

— and chatterton, b. d. e. 1988. Classification of the trilobite suborder Asaphma. Palaeontology , 31,
165-222.

— and owens, r. m. 1991. A trilobite fauna from the highest Shineton Shales in Shropshire, and the
correlation of the latest Tremadoc. Geological Magazine , 128, 437^464.

henningsmoen, G. 1975. Moulting in trilobites. Fossils and Strata , 4, 179-200.

hu, chung-hung, 1971. Ontogeny and sexual dimorphism of Lower Paleozoic Trilobita. Palaeontographica
Americana , 7, 1-155.

hunt, a. s. 1967. Growth variation and instar development in an agnostid trilobite. Journal of Paleontology,
41, 203-208.

jaekel, o. 1909. Uber die Agnostiden. Zeitschrift der Deutschen Geologischen Gesellschaft , 53, 380-401.
jell, p. a. 1970. Pagetia ocellata, a new Cambrian trilobite from northwestern Queensland. Memoirs of the
Queensland Museum , 15, 303-313.

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

19756. The abathochroal eye of Pagetia, a new type of trilobite eye. Fossils and Strata, 1, 33-43.
ludvigsen, R. and chatterton, B. D. E. 1980. The ontogeny of Failleana and the origin of the Bumastinae
(Trilobita). Geological Magazine 117, 467^178.

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.

OPIK, A. a. 1952. The hypostome of Pagetia. Journal of Paleontology , 26, 272-274.

pabian, r. k. and fagerstrom, J. A. 1972. Late Paleozoic trilobites from southeastern Nebraska. Journal of
Paleontology, 40, 789-816.

palmer, a. r. 1957. Ontogenetic development of two olenellid trilobites. Journal of Paleontology, 31, 105-128.

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

— 1962. Comparative ontogeny of some opisthoparian, gonatoparian, and proparian Upper Cambrian
trilobites. Journal of Paleontology, 36, 87-96.

raw, F. 1925. The development of Leptoplastus salteri (Calloway) and of other trilobites. Quarterly Journal of
the Geological Society, London, 81, 223-324.

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

— 1972. Mode of life of agnostid trilobites. International Geological Congress 24th Session, Montreal,
Section 7, 33-40.

romano, m. 1976. The trilobite genus Placoparia from the Ordovician of the Valongo area. North Portugal.
Geological Magazine, 113, 11-28.

rushton, a. w. a. 1966. The Cambrian trilobites from the Purley Shales of Warwickshire. Monograph of the
Palaeontographical Society, 120 (511), 1-55, pis 1-6.

Sheldon, p. r. 1988. Trilobite size-frequency distributions, recognition of instars, and phyletic size changes.
Lethaia, 21, 293-306.

shergold, j. h. 1991. Protaspid and early meraspid growth stages of the eodiscoid trilobite Pagetia ocellata
Jell. Alcheringa, 15, 65-86.

snajdr, m. 1981. Ontogeny of some representatives of the trilobite genus Scharyia. Sbormk geologickych ved,
Paleontologie, 24, 7-35.

Speyer, s. E. and brett, c. e. 1986. Trilobite taphonomy and Middle Devonian taphofacies. Palaios, 1, 312-327.

— and chatterton, b. d. e. 1989. Trilobite larvae and larval ecology. Historical Biology , 3, 27-60.
Stubblefield, c. J. 1926. Notes on the development of a trilobite, Shumardia pusilla (Sars). Journal of the

Linnean Society, Zoology Series, 36, 345-372.

tripp, r. p. and evitt, w. R. 1983. Silicified trilobites of the genus Dimeropyge from the Middle Ordovician of
Virginia. In briggs, d. e. g. and lane, p. d. (eds). Trilobites and other early arthropods ; papers in honour of
Professor H. B. Whittington, F.R.S. Special Papers in Palaeontology, 30, 229-240.

Whittington, h. b. 1957. The ontogeny of trilobites. Biological Reviews, 32, 421^469.

— 1959. Ontogeny of trilobites. 0127-0144. In moore, r. c. (ed.) Treatise on invertebrate paleontology. Part
O, Arthropoda 1. Geological Society of America and University of Kansas Press, Boulder, Colorado and
Lawrence, Kansas, xix + 560 pp.

806

PALAEONTOLOGY, VOLUME 36

Whittington, h. b.1988. Hypostomes and ventral cephalic sutures in Cambrian trilobites. Palaeontology, 31,
577-609.

wilmot, N. Cuticular structure of the agnostine trilobite Homagnostus obesus. Lethaia , 23, 87-92.
zhang, wen-tang. 1987. World’s oldest Cambrian trilobites from eastern Yunnan. 1-16. In Stratigraphy and
palaeontology of systemic boundaries in China , Precambrian-Cambrian boundary. Science Press, Beijing,
242 pp.

— LU YAN-HAO, ZHU ZHAO-LING, QIAN YI-YUAN, LIN HUANG-LING, HOU ZHI-YI, ZHANG SEN-GUI and YUAN JIN-
LING. 1980. Cambrian trilobite faunas of southwestern China. Palaeontologica Sinica, New Series C, 16,
1-497.

zhang, xi-guang. 1989. Ontogeny of an Early Cambrian eodiscoid trilobite from Henan, China. Lethaia, 22,
13-29.

and clarkson, E. n. k. 1990. The eyes of Lower Cambrian eodiscid trilobites. Palaeontology, 33, 91 1-932.

ZHANG XI-GUANG
Chengdu Institute of Geology
and Mineral Resources
North Renmin Road
Chengdu

Sichuan 610082, People’s Republic of China

Present address:

Department of Geology,
University of Saskatchewan,
Saskatoon, Canada S7N 0W0

EUAN N. K. CLARKSON

Department of Geology and Geophysics
University of Edinburgh
King’s Buildings
West Mains Road
Edinburgh EH9 3JW, UK

Typescript received 4 August 1992
Revised typescript received 16 February 1993

PHYLOGENY AND EVOLUTION OF
‘PENTAMERIDE' BRACHIOPODS

by SANDRA J. CARLSON

Abstract. Despite their importance in articulate brachiopod evolutionary history, relatively little is known in
detail about the phylogenetic relationships among ‘pentameride’ taxa, and of ‘pentamerides’ to other
articulates. Phylogenetic relationships among all named ‘pentameride’ families and rhynchonellide
superfamilies were reanalysed using outgroup methods of polarity determination. A detailed working
hypothesis of ‘ pentameride ’ phylogeny and the supporting evidence on character distribution is presented. As
currently diagnosed, Pentamerida and Syntrophiidina are paraphyletic, while Rhynchonellida and Penta-
meridina are monophyletic. Generally acknowledged patterns of morphological change may now be examined
in detail, as they are expressed in a comprehensive pattern of relationship. In the evolutionary history of these
taxa, strophic hinge line length decreased steadily and astrophic hinge lines evolved twice. Interlocking hinge
structures arose twice from the non-interlocking condition, and muscle platforms in the dorsal and ventral
valves evolved several times independently. These phylogenetic results have significant implications for several
issues relevant to the study of brachiopod systematics. Agreement between the stratigraphical first appearance
of ‘pentameride’ families and their cladistic rank is quite good, suggesting that both outgroup and
palaeontological methods indicate the same direction of character polarity in the evolution of ‘pentamerides’.
The paraphyletic ‘syntrophiidines’ suffer pseudoextinction in transforming to the monophyletic rhyn-
chonellides (extant) and the monophyletic pentameridines (extinct), which possess a combination of characters
(very strong biconvexity, large adult size, lack of pedicle, non-interlocking dentition) that apparently rendered
them less able to adapt over time to changes in their habitat. A highly corroborated phylogenetic hypothesis
provides an explicit framework within which causal hypotheses of macroevolutionary phenomena may be
generated and tested.

‘Pentameride’ brachiopods occupy a particularly important place, both temporally and
morphologically, in the evolution of articulate brachiopods. ‘Pentamerides’ are among the earliest
articulates in the fossil record, first appearing in the Lower Cambrian of Siberia (Andreeva 1987).
The Pentamerida is one of the first articulate brachiopod orders to become extinct (with the
Atrypida, at the end of the Devonian). Several morphological transformations of great significance
in articulate brachiopod evolution are manifest within the ‘pentamerides’. For example, the
‘pentamerides’ include the first cyrtomatodont articulates, making the transition from non-
interlocking to interlocking hinge structures (Jaanusson 1971). They also include the first astrophic
articulates, evolving curved hinge lines from those that were long and straight. Shell biconvexity
increases dramatically from the earliest to the latest ‘pentamerides’. A number of derived features
associated with extant brachiopods first appear quite early in ‘pentameride’ evolution. On the other
hand, various types of muscle platforms are developed in both the dorsal and ventral valves and are
a prominent feature of ‘pentameride’ internal shell morphology. Similar platforms are present in
several groups of Palaeozoic brachiopods, but are generally lacking in Recent forms (see Rudwick
1970). Considering their early appearance in the fossil record, ‘pentameride’ brachiopods thus
present interesting combinations of both primitive and unexpectedly derived morphological
features.

In the most general sense, ‘pentamerides’ are thought to have evolved from the orthides and
given rise to the rhynchonellides (see Text-fig. 1). Despite their considerable importance in our
understanding of brachiopod evolution and the origin of the modern brachiopod fauna, a detailed

[Palaeontology, Vol. 36, Part 4, 1993, pp. 807-837.)

© The Palaeontological Association

808

PALAEONTOLOGY, VOLUME 36

text-fig. 1. Stratigraphical ranges of articulate brachiopod superfamilies plotted according to their familial
diversity and pattern-coded by their ordinal classification; redrawn from Williams (1968). The hypothetical
phylogenetic relationships are as illustrated by Williams (1968).

reconstruction of phylogenetic relationships among the ‘pentamerides’, and of ‘pentamerides’ to
other articulates, is lacking. No comprehensive study of ‘pentameride’ phylogeny and character
evolution has been completed since G. A. Cooper’s pioneering work in the 1930s (e.g. Schuchert
and Cooper 1932; Ulrich and Cooper 1938). The number of ‘pentameride’ genera has more than
doubled in the last thirty years alone, since the publication of the brachiopod volumes of the
Treatise on invertebrate paleontology (Williams and Rowell 1965). It it time to re-examine
assumptions of character homology and polarity among all members of the order.

The primary goal of this study is to investigate phylogenetic relationships among ‘pentameride’
brachiopod families. Without a detailed and strongly supported phylogenetic hypothesis,
morphological transformations among the ‘pentamerides’ may be understood only in the most
general terms. The results of four experimental phylogenetic analyses using outgroup criteria for
polarity determination are compared and contrasted, and the implications of each to several issues
relevant to the study of brachiopod systematics are discussed: morphological character evolution
within the group; comparison of outgroup and stratigraphical methods of polarity determination;
past, present, and future interpretation of ‘pentameride’ classification; and the macroevolutionary
significance of the extinction of the paraphyletic ‘pentamerides’.

METHODS

Tax a

Two ‘pentameride’ suborders are recognized currently (Amsden 1965; Biernat 1965; see Table 1).
The Pentameridina includes the stereotypical ‘pentamerides’; large, highly biconvex brachiopods
with long, curved beaks in both dorsal and ventral valves. The Syntrophiidina includes a diverse
group of brachiopods with a number of morphological characteristics intermediate betweemthe

CARLSON: PENTAMERIDE BRACHIOPOD PHYLOGENY

809

table 1. Classification of the Order Pentamerida used in this study. Primary reference is the Treatise on
invertebrate paleontology (Biernat 1965; Amsden 1965), with additions from Nikiforova (1960), Amsden et al.
(1967), Gauri and Boucot (1968), Boucot and Johnson (1979). Syntrophiidine genera named since 1965 are
included in the families to which they were assigned by their authors. Three additional Chinese syntrophiidine
genera ( Disepta , Fengxiangella, and Limstrophina) were only recently brought to my attention by Dr Rong
Jia-Yu (personal communication, 1993), and thus are not included in these analyses.

PENTAMERIDA

SYNTROPHIIDINA

PORAMBONITACEA

Alimbellidae : Alimbella , Medessia , Mogoktella

Clarkellidae: Acanthoglypha , Calliglvpha , Clarkella , Diaphelasma , Stichotrophia ,
Syntrophina , Syntrophinella, ISyntrophioides, Thaumotrophia, Yangtzeella
Eostrophiidae : Cambrotrophia

Huenellidae: [Huenellinae] Huenella , Huenellina , Palaeostrophia , Plectotrophia;

[Mesonomiinae] Glyptotrophia, Mesonomia; [Rectotrophimae] Rectotrophia
Lycophoriidae : Lycophoria

Porambonitidae : Porambonites , Porambonitoides , Rosella , Talovia
Syntrophiidae : [Syntrophiinae] Rhyselasma, Syntrophia; [Xenelasmatinae]
Euorthisina, Xenelasma, Xenelasmella, Xenelasmopsis
Syntrophopsidae: Altunella , Bobinella, ICuparius , Hesperotrophia, Rhabdostrophia ,
Rhysostrophia, Syntrophopsis, Tcharella

Tetralobulidae: Doloresella, Imbricatia , Pseudoporambonites, Punctolira, Tetralobula

Karakulinidae: Karakulina

Uncertain : Triseptata

CAMERELLACEA

Brevicameridae : Brevicamera

Camerellidae : [Camerellinae] Bleshidimerus , Bleshidium , Camerella, Idiostrophia ,
Kokomerena , Liricamera, Llanoella , Neostrophia , Perimecocoelia,

Plectocamara , Plectosyntrophia, Psilocamerella, Tuloja, Xizangostrophia;
[Stenocamarinae], Boreadocamara, Stenocamara
Parastrophinidae: Anastrophia , Eoanastrophia , Grayina , Jolkinia , Liostrophia ,
Maydenella , Parastrophina , Parastrophinella
UNCERTAIN

Branconia , Schizostrophia , Swantonia

PENTAMERIDINA

PENTAMERACEA

Parallelelasmatidae : Didymelasma , 1 Metacamerella ( = Parallelelasma ), Salonia
Clorindidae: Antirhynchonella, Clorinda , Clorindella, Clorindina
Enantiosphenidae : Enantiosphen

Gypidulidae: Barrandina , ? Biseptum, Carinagypa, Devonogvpa , Gypidula ,

GypiduleUa , Gypidulina, Ivdelinia , Leviconchidiella, Levigatella,

Pentamerella , IProcerulina, Sieberella , Wyella , Zdimir
Pentameridae : Brooksina , Callipentamerus , Capelliniella , Harpidium, Jolvia,
Lissocoelina , IPentamerifera, Pentameroides , Pentamerus , IPleurodium,

Rhipidium

Stricklandiidae: Costistricklandia , Kulumbella , Microcar dinalia, Plicostricklandia,
Stricklandia

Subrianidae: Aliconchidium , Conchidium , Cymbidium , Lamelliconchidium ,
Plicocoelina , Severella, Spondylopyxis, Spondylostrophia,

St r ick Ian d is trophia, Subriana, Vagranella , Vosmiverstum
Virgianidae: Holorhynchus , Platymerella , Virgiana

810

PALAEONTOLOGY, VOLUME 36

strophic orthides and the astrophic rhynchonellides and terebratulides. The ‘syntrophiidines’ are
thought to serve as the ancestors of both the pentameridines, which represent a derived, short-lived
but highly successful ‘dead end’ of ‘pentameride’ evolution (see Johnson 1977; Boucot and
Johnson 1979), and the rhynchonellides, which have persisted to the present day remarkably
unchanged.

Choosing terminal taxa in a preliminary phylogenetic study of this sort presents a chicken-and-
egg dilemma. Which comes first, the phylogenetic analysis or the taxa? Terminal taxa should
represent systems of common ancestry (i.e. be monophyletic; see Wiley 1981). The monophyly of
named ‘pentameride' higher taxa has not been tested; it is highly likely that at least some are not
clades. A species level analysis of the ‘pentamerides’ would appear to be a necessary first step,
assuming that biological species represent ‘basic taxonomic units' (although see de Queiroz and
Donoghue 1988, 1990; Nixon and Wheeler 1990) and that fossil species are comparable in some
sense to biological species. However, a minimum estimate of several hundred named fossil
‘pentameride’ species exist. It is not reasonable to expect interpretable results from a phylogenetic
analysis, either by hand or by computer, that includes such a large number of terminal taxa. Even
at the genus level, which is commonly considered to represent a realistic operational taxonomic unit
in brachiopod palaeontology (Cooper 1970), well over one hundred taxa exist.

As a compromise between feasibility and taxonomic detail, phylogenetic relationships among
families of ‘pentamerides’ were chosen for analysis. Specimens were examined, when possible, in
addition to descriptive literature on all named genera assigned to each family. Character states per
family were coded as a consensus of character states present in each genus assigned to that family.
This strategy runs the risk of coding a taxon as a combination of characters that are not present
in that particular combination in any single individual. Nevertheless, if the taxon is monophyletic,
the character combination should represent the clade as a whole. Phylogenetic analyses of genera
and species within at least the ‘syntrophiidine' families are being conducted currently (Carlson in
preparation), and the results of these ongoing studies have the potential to affect the results
presented here.

Fifteen ‘syntrophiidine’ families (four of which are monogeneric), eight pentameridine familes or
subfamilies, and two rhynchonelhde superfamilies comprise the ingroup in these analyses (Table 1).
The diagnoses and generic composition of ‘syntrophiidine’ families, as listed in Table 1, largely
reflect the classification in Biernat (1965). Genera named since 1965 are included in the families to
which they were assigned by their authors. My knowledge of the Syntrophiidina is much greater
than for the Pentameridina; thus I have relied largely on the Amsden (1965) classification and
diagnoses to characterize the pentameridine families (but also consulted Amsden et al. 1967; Gauri
and Boucot 1968; Boucot and Johnson 1979). Many new pentameridine genera have been named
since then, but because I have had limited exposure to the specimens, I chose to exclude them from
this analysis.

The purpose of this study is to investigate phylogenetic relationships among ‘pentameride’
families. Revising ‘pentameride’ higher-level classification is a separate, subsequent endeavour, and
will not be accomplished here. Genus-level phylogenetic analyses of the ‘pentamerides’, particularly
the most primitive ‘pentamerides’ and the orthides, must be completed before final decisions of
classification can be reached. Decisions involving the redefinition of established higher taxa are
particularly delicate; they deserve the consideration of the full body of morphological evidence on
all articulates, which is currently under active investigation.

Throughout, informal taxon designations are used (e.g. orthides for Orthida, ‘syntrophiidines’
for Syntrophiidina, rhynchonellaceans for Rhynchonellacea, porambonitids for Porambonitidae)
that refer to groups of brachiopods currently classified in various higher taxa. The phylogenetic
status, relative taxonomic rank, and lower-level classification of these named taxa are all in the
process of being evaluated; thus, the informal name is used to convey some sense of the
brachiopods in question, without placing undue emphasis on the rank or current definition of the
taxon name itself. In the interest of consistency, suspected or identified paraphyletic taxa are always
referred to in quotation marks, following Gauthier (1986).

CARLSON: PENTAMERIDE BRACHIOPOD PHYLOGENY

811

Finally, it is very likely that components of the research of several Russian palaeontologists (e.g.
Nikiforova, Sapelmkov, Andreeva, Kulkov, Tcherkesova, among others) who have produced
extensive publications on the ‘pentamerides’ have been inadvertently overlooked. My knowledge
of the Russian literature is limited by the amount that has been made accessible through translation,
which represents only a small portion of the total body of research. Sapelnikov (1980, 1982, 1985)
in particular, has published phylogenetic reconstructions of the ‘pentamerides’. His assignment of
genera to families does not agree with the classification (Table 1) used as a starting point for these
analyses; thus, our conclusions about ‘pentameride’ relationships will necessarily differ.

Characters

The list of characters (Appendix A), data matrix (Appendix B), and list of apomorphies for
cladogram five, discussed below, (Appendix C) have been deposited with the British Library, Boston
Spa, Yorkshire, UK, as Supplementary Publication No. SUP 14043.

Because all but one of the terminal taxa in this study are extinct, all the characters used in the
analysis are necessarily those that fossilize, namely characters of skeletal anatomy. Homoplasy
(convergence or parallelism) in brachiopod skeletal anatomy is common (e.g. Buckman 1906;
Cooper 1930, 1972; Cloud 1941), and raises a legitimate concern that perhaps too few homologous
characters will be identified to generate a phylogenetic ‘signal’ above the homoplastic ‘noise’. Such
concerns may never be put to rest entirely in phylogenetic studies of extinct groups. However,
ignoring all extinct taxa is a highly unsatisfactory alternative to attempting an analysis with the
available morphological data and then critically evaluating the results. In this case, homology was
tested primarily by phylogenetic congruence (Patterson 1982) with other putative homologues.

Information was compiled on seventy-four morphological characters of skeletal anatomy,
including valve form and ornament, shell structure, the hinge region, and dorsal and ventral valve
interiors, especially the cardinalia (SUP 14043, Appendix A). Only characters that vary among two
or more terminal taxa were included. Autapomorphous characters are essential in identifying
individual taxa, but do not provide information on relationships among taxa. No attempt was made
to eliminate characters thought to be homoplastic prior to performing the analyses, under the
assumption that homoplasy would be revealed in the analysis itself by phylogenetic congruence
(Patterson 1982). Both binary and multistate characters were recognized. All characters were
initially unordered, allowing the outgroup (primitive) character states to polarize the direction of
character transformation. None were constrained to be irreversible. All were weighted equally in the
first analysis; all were reweighted according to their rescaled consistency indices (Farris 1989) in the
second analysis.

Many characters were coded as missing (SUP 14043, Appendix B), for one of three different
reasons: (1) the character is not applicable to the taxon (e.g. spondylium type in a taxon lacking a
spondylium); (2) it is not known for the taxon; or (3) the states are variable (polymorphic) among
genera in a family. Intentional ambiguity in coding polymorphic taxa as missing, enables the
polarity of the various character states to be reconstructed from the results of the phylogenetic
analysis. In other words, coding variability itself as a separate character state (e.g. shell ornament:
smooth [0], costate [1], both smooth and costate [2]) will tend to group polymorphic taxa together.
It is more likely that one of the two character states is primitive, as revealed in the analysis, and has
transformed within the polymorphic taxon. The cladogram topology is structured on the basis of
coded characters; missing characters do not play a role in cladogram construction (although see
Nixon and Davis 1991; Platnick et al. 1991; Novacek 1992).

Polarity determination

Traditional palaeontological methods of phylogenetic inference polarize the direction of character
transformation using (primarily) stratigraphical criteria (i.e. stratigraphically lowest fossils are most
primitive). Relying largely on stratigraphical polarity in phylogenetic reconstruction is problematic
for several reasons. Using this method, the ‘primitive condition’ is fundamentally empirical and

812

PALAEONTOLOGY, VOLUME 36

defined solely on the basis of characteristics observable in the oldest known fossils. As older and
older fossils are discovered, the concept of ' primitive ’ must necessarily change to accommodate them.
Also, stratigraphical resolution may be poor at times of critical evolutionary importance. For
example, diverse morphotypes appear nearly simultaneously in the Cambrian (e.g. Rowell 1977),
making it difficult to decide which of the conflicting characters is ‘the’ most primitive.

Outgroup criteria for polarity determination (Watrous and Wheeler 1981 ; Maddison et al. 1984)
were used to test the relationship between stratigraphical first appearance data and cladistic rank
among the 'pentamerides’ (see Gauthier et al. 1988; Norell and Novacek 1992). Using outgroup
criteria, character states present in outgroup taxa presumed to share most recent common ancestry
with the ingroup taxa function as the reference for the primitive state. Outgroup analyses generate
a phylogenetic framework that is not exclusively dependent upon the quality of preservation of the
fossil record, and allow predictions to be made about possible character combinations in fossils not
yet discovered. Ideally, both methods will indicate the same polarity, but use different criteria. If
they do not, new insights into the nature of character evolution or fossil preservation may be gamed.

Three orthide taxa were chosen as outgroups (Nisusiidae, Billingsellidae, and Orthacea), using
Williams’ (1968) phylogenetic tree as a working hypothesis of relationships among all articulates
(Text-fig. 1). The orthides function as outgroups, but they also happen to occur earlier in the fossil
record than most of the ingroup taxa. Because of logical problems with coding relative
stratigraphical position as a character in a morphological analysis, analyses were conducted
independently of stratigraphical position; stratigraphical and morphological results were then
compared. Stratocladistics methods (Fisher 1980, 1982, 1988, 1991, 1992; Maddison and Maddison
1992), in which stratigraphical data can be incorporated directly into a morphological analysis but
analysed in a manner necessarily different from morphological data, will soon be used and the
results compared with these.

Phylogenetic methods

A phylogenetic systematic methodology was employed to analyse genealogical relationships among
the ’pentamerides’ (see Hennig 1966; Eldridge and Cracraft 1980; Wiley 1981 ; Wiley et al. 1991).
The goal of this method of inference is to identify evolutionary patterns of common ancestry that
result from the process of descent with modification. The phylogenetic systematic approach can be
viewed 'not as an alternative to traditional evolutionary methods, but as a refinement of them’
(de Queiroz 1988, p. 244). Phylogenetic methods produce explicit, testable working hypotheses of
relationship.

All analyses were conducted using the microcomputer program PAUP 3.0 (Swofford 1990).
PAUP is a parsimony-based program that seeks to find the shortest (most parsimonious) branching
diagram compatible with the available data, as it is coded in a taxon-by-character data matrix
(SUP 14043, Appendix B). Despite the widespread use of parsimony-based methods, parsimony is
a contentious principle in phylogenetic inference (e.g. Felsenstein 1978, 1983; Sober 1983, 1985).
Much (but not all) of the controversy surrounds the use of parsimony methods when rates of
evolution vary considerably among the taxa being analysed. I have assumed that significantly
different rates of taxic evolution are not an issue in this study of 'pentameride’ brachiopods, but this
is admittedly difficult to estimate.

Any phylogenetic analysis is only as robust as the assumptions implicit in its methods, including
character homology, taxon monophyly, polarity determination, etc. Particularly given the ease with
which microcomputer programs designed to reconstruct phylogenetic patterns can be used, and
results (of greatly varying quality) obtained, it is important to experiment extensively with the data
matrix and the program, to avoid accepting uncritically the first (or most favoured) result (e.g. Cann
et al. 1987; Templeton 1992).

Four analyses and four of the cladograms that resulted from them form the basis of this study.
A heuristic search using global branch swapping methods with random addition of taxa in each of
ten replicate analyses was first employed. Branch and bound searches and bootstrap replications

CARLSON: PENTAMERIDE BRACHIOPOD PHYLOGENY

813

were prohibitively slow due to the size and structure of the data matrix and have not yet been
completed. One of the several equally most parsimonious cladograms that resulted was compared
with the fifty percent majority rule consensus cladogram of all the equally most parsimonious
cladograms. In the second analysis, each of the characters was re-weighted based on its rescaled
consistency index (the more stable characters receive proportionally more weight) from the first
analysis, and a heuristic search was then performed. The rhynchonellides were removed in the third
analysis, revealing their significance in structuring relationships among the ‘pentamerides’. Finally,
I experimented with selectively weighting four characters concerning the development of muscle
platforms, to investigate their effect in ‘pentameride’ classification.

RESULTS

Analysis I - Cladogram 5

The first analysis yielded twelve equally most parsimonious cladograms, each of length 286 and
consistency index of 0-389. Although the consistency index appears to be relatively low, it is still well
within the norm for analyses of twenty-eight taxa (see Sanderson and Donoghue 1989; Klassen
et al. 1991). The pattern of relationships that emerges (Text-fig. 2) is not at all unexpected, given
traditional concepts of ‘pentameride’ phylogeny. Moreover, it is quite consistent with the order of
first appearance of these taxa in the fossil record. At the family level, ‘pentamerides’ (currently
defined) are paraphyletic. Rhychonellides and pentameraceans each form clades, and each shares
most recent common ancestry with different ‘syntrophiidine’ groups.

More than merely confirming several sister-group pairs recognized in the traditional
‘pentameride’ phylogenies (e.g. Eostrophiidae ancestral to Syntrophiidae; Huenellidae ancestral to
Tetralobulidae; Virgianidae ancestral to the Pentameridae and Subrianidae), the analysis presents
a parsimonious and detailed working hypothesis of relationships among all the taxa included in the
analysis. To construct the pattern of relatedness among the ‘pentamerides’ and their relatives at this
level of detail, patterns of character homology and homoplasy must be evaluated more or less
simultaneously. In the past, certain selected characters (‘good’ or less variable characters) were used
to establish taxon diagnoses, while the distribution of highly variable or conflicting characters was
ignored. However, these ‘bad’ characters may be less problematic in other taxa and may even
define them. An explicit branching diagram, with apomorphies defining each node, serves as a basis
for discussion of homology and homoplasy in character evolution. Points of disagreement can be
established clearly when character distributions across the entire diagram are known.

Remarkable examples of brachiopod homeomorphy - specimens with identical external
morphologies and different internal morphologies (e.g. the orthide Platystrophia and the spiriferide
Spirifer) - are relatively common. Such striking convergence in whole suites of characters makes
one suspicious that less obvious examples of homoplasy are likely to be common among
brachiopods. The analysis bears out this prediction; most of the characters have a consistency index
of considerably less than 1-0 (the average is, of course, 0-389), indicating numerous reversals,
convergences, or parallelisms.

Given the great geological age of the ingroup and the extinction of all but one of its members,
it makes evolutionary sense to expect fairly low consistency among characters. Examining patterns
of relationship established over a period of two hundred million years among higher (presumably
monophyletic) taxa extinct for over three hundred and fifty million years, it is entirely reasonable
to expect relatively high levels of homoplasy. This is particularly true when the pool of characters
included in the analysis is limited to morphological characters as they are expressed in organisms
less morphologically complex than, for example, arthropods or vertebrates. This is not to say that
lower morphological complexity renders cladistic analyses ineffective, only that expectations of high
(‘statistically significant’) levels of congruence among characters is perhaps unrealistic from an
evolutionary perspective.

To facilitate discussion of the results, seven groups of taxa are recognized. Some correspond to
named higher taxa, others do not. Four of the seven together comprise the order Pentamerida

rt O c/) o

Billingsellidae

Nisusiidae

Matutellidae

Orthacea

Alimbellidae

Huenellidae

Tetralobulidae

Syntrophopsidae

Clarkellidae

Brevicameridae

Camerellidae

Eostrophiidae

Syntrophiidae

Lycophoriidae

Triseptata

Rhynchonellacea

Stenoscismatacea

Porambonitidae

Parastrophinidae

Karakulinidae

Parallelelasmat

Clorindinae

Gypidulinae

Enantiosphenidae

Stricklandiidae

Virgianidae

Pentamerinae

Subrianinae

□

□

□

iza

(ZZ

814

PALAEONTOLOGY, VOLUME 36

(D

<D

03

■o

03
c/3

03 (/)

E 3

CO

m z

03

CO

0

:= 03

03

0

0

0

0

0

0

0

C/3

0

0

0

0

3

0

0

0

0

o

0

0

0

0

0

o

0

0

3

xz

0

"0

c

3

-D

O

0

xz

CL

o

■0

3

zx

E

0

o

3

0

XZ

V—

E

0

Zi

t

0

H — '

c

>,

0

>

0

E

0

O

<

I

1-

C/3

o

CD

O

Q3

03

■o

03

CO

0

xz xz

Cl CL <-
O O

O

C/3

O

c

>»

CL CL
O 0
o c/3

>N ‘>-

03

03

O

_0

3

c

o

xz

o

c

>>

xz

o

JZl

E

C0

0

03

0

’c

0

_ 0
sz 0

§-1

ZD

0 JxC
0 0

0

CL

0

Z

0

E

0

_0

_0

_0

3

0

CL

0

13

C

0

03

0

-O

0 Zi .E

.E 2 c

E Q- 0

-E >> c

O O lu

Q- 0

£ c
O _co

zx
o

0

0

■g

’c=

0

0

0

C

L_

0

E

0

03 C

iz .h= 0
CO > CL

0

0

C

c

0

_Q

o

CO

text-fig. 2. Cladogram number five; one of twelve equally most parsimonious cladograms resulting from
Analysis I, using outgroup polarity. Solid dots under taxon names identify outgroup taxa. The cladogram is
rooted at an internal node with a basal polytomy. Note that the ingroup is not strictly monophyletic; Orthacea
appears to share more recent common ancestry with certain ‘syntrophiidine’ taxa than with other orthide taxa.
Nodes are identified by the letters beside them ; aponrorphies of each node are listed in the Appendix. The known
stratigraphical range of each taxon is plotted above the taxon names. Outgroup ranges are unshaded,
‘syntrophiidine’ families are shaded, rhynchonellides are black, and pentameridines are diagonally hatched.

CARLSON: PENTAMERIDE BRACHIOPOD PHYLOGENY

815

(Amsden 1965; Biernat 1965); three of the seven comprise the Syntrophiidina (Biernat 1965).
Characters that distinguish each group will be discussed and compared to current taxon diagnoses.

Pentamerida. Lower Palaeozoic biconvex brachiopods with impunctate shells, open delthyria,
distinctive dorsal cardinalia, and ventral spondylia are typically assigned to the order Pentamerida
(Amsden 1965; Biernat 1965; see Text-fig. 3). The spondylium, a spoon-shaped structure formed

text-fig. 3. Highly schematic reconstructions to facilitate comparison of relative biconvexity, hinge line shape
and length, and development of ventral muscle platforms. Upper row: dorsal view of interior of ventral valves;
lower row: lateral view of articulated valves, ventral valve on left, a, early orthide; b, generalized

syntrophiidine; c, derived pentameridine.

from the uniting of convergent dental plates and the ventral median septum, is particularly
diagnostic of the ‘pentamerides’, in concert with the other characters mentioned. The results of the
phylogenetic analysis (Text-fig. 2; Appendix; SUP 14043, Appendix C) are consistent with this
general characterization of the ‘pentamerides’, but the shared derived characters of the group span
several nodes (a-d) in the cladogram, rather than being clustered conveniently at a single node.

The cladogram in Text-figure 2 suggests three possibilities for the definition of Pentamerida.
Either the Pentamerida includes the matutellids, making the orthaceans (and alimbellids?) an early
offshoot from the ‘ pentameride ’ clade, or it includes only the alimbellids, or it excludes both
matutellids and alimbellids. Andreeva (1987) classified the matutellids in the Pentamerida and
considered them to be the ancestors of the alimbellids, while Williams and Bassett (1991 ) tentatively
suggest that the alimbellids may be more appropriately classified with the Orthida. Matutellids
appear to be morphologically intermediate between early orthides and ‘pentamerides’. They possess
a long, straight hinge line and an apical or supra-apical foramen like early orthides, but also possess
a very strong fold and sulcus and mantle canal markings similar to the ‘syntrophiidines’ (Andreeva
1987). Ultimately, the hierarchic pattern of acquisition of derived characters (Appendix) is more
informative phylogenetically than deciding how to define the taxon Pentamerida.

Syntrophiidina. The ‘syntrophiidines’ have long been considered to be a paraphyletic, early group
of ‘pentamerides’ (Schuchert and Cooper 1932; Williams 1968; see Text-fig. 2). They possess
derived characters of the Pentamerida (e.g. open delthyrium, spondylium, strong biconvexity, fold
and sulcus), but lack derived characters of the pentameraceans and rhynchonellides (Table 2).

‘ Early Syntrophiidina ’. Huenellidae and Tetralobulidae, Syntrophopsidae, and Clarkellidae (and

816

PALAEONTOLOGY, VOLUME 36

possibly Alimbellidae and/or Matutellidae) together comprise the earliest ‘ syntrophiidines a grade
of Cambro-Ordovician ‘pentamerides’ (Text-fig. 2). They possess derived ‘pentameride’ features,
while retaining certain primitive characters of the orthides (e.g. relatively long hinge lines and
extensive interareas, moderate to small adult size). It is the morphological combination of both
primitive and derived characters that gives this early paraphyletic group a certain morphological
‘integrity’.

‘ Camerellacea' . Nikiforova (1960) established the Camerellacea as a superfamily distinct from the
‘porambonitaceans’ largely by the combined possession of a spondylium duplex in the ventral valve
(a typical pentameracean character) and retention of a ‘syntrophiidine’ type of dorsal cardinalia
and muscle field. The Camerellidae, some parastrophinids, and the Stricklandiidae were originally
classified in this superfamily (Nikiforova 1960). Many ‘camerellaceans’ may possess a spondylium
simplex rather than duplex (Biernat 1965). The camerellids alone are a fairly morphologically and
taxonomically diverse group; they may not be monophyletic. In this analysis (Text-fig. 2), the
brevicamerids and camerellids are sister taxa characterized by a very reduced hinge line, a cruralium
(functionally comparable to a spondylium in the dorsal valve), and valve ornament and fold and
sulcus restricted to the anterior portions of the valves. Camerellacea redefined in this manner would
include only these two families; it is doubtful that these characters alone justify superfamily status
for the group. Nevertheless, genus-level analyses within these families may shed light on the
distribution and acquisition of characters in the clade and clarify its phylogenetic status. As
originally defined, however, the Camerellacea (Nikiforova 1960) does not represent a clade;
according to this analysis, its diagnostic characters are homoplastic.

‘ Late Syntrophiidina' . Eostrophiidae and Syntrophiidae, Porambonitidae, Lycophoriidae and
Triseptata , and Parastrophinidae comprise a poorly resolved grade of (largely) Ordovician
‘pentamerides’ (Text-fig. 2). Compared to the other ‘pentamerides’, these brachiopods together
possess unusual combinations of primitive and derived characters. They have short (but occasionally
long) hinge lines, a weak fold and sulcus, and a strong dorsal septalium. Spondylia are either lacking
entirely or are present as duplex spondylia; several have interlocking hinge structures, and many
reach large adult sizes. Cambrotrophia (the sole eostrophiid) is widely considered to be ancestral to
the syntrophiids (Biernat 1965), despite a considerable stratigraphical gap between them (Text-fig. 2).
The location of Cambrotrophia near the centre of the cladogram is somewhat anomalous
considering its early first appearance in the fossil record.

Despite the fact that Porambonites serves as the type genus for the superfamily, the porambonitids
represent a significant morphological departure from other ‘syntrophiidines’ because of their large
size, characteristic fenestrate ornament, robust hinge teeth (that interlock in some), and in lacking
a spondylium. The classification of Lycophoria has long been problematical as well. The genus has
been classified as a strophomenide (Lahusen 1886; St Joseph 1939), an orthide (Schuchert and
Cooper 1932), and a ‘pentameride’ (Biernat 1965). Although this analysis resolves the Lycophoria
plus Triseptata clade as sister-group to the rhynchonellides, future analysis may reveal that
Lycophoria (and perhaps Triseptata) is more closely related to nonsyntrophiidine brachiopods.
The parastrophinids share a number of derived characters with the pentameridines (e.g. astrophic
hinge line), but retain a ‘syntrophiidine’ dorsal cardinalia, including a strong septalium.

Rhynchonellida. Rhynchonellacea and Stenoscismatacea are sister taxa, consistent with their
classification in the order Rhynchonellida. Rhynchonellides are characterized by astrophic hinge
lines, extremely strong valve biconvexity with a deep fold and sulcus (primitively retained). Most
possess a characteristic costate ornament. The delthyrium is partly closed by deltidial plates. The
spondylium is absent in the rhynchonellaceans, but present and elaborated in the steno-
scismataceans. Brachiophores, bounding the socket, are longer than in the ‘syntrophiidines’ and
brachial processes (in the form of crura) are present. Many rhynchonellaceans possess a strong
dorsal septalium either retained primitively from their ‘syntrophiidine’ ancestry or (possibly)

table 2. Comparison of distinguishing characteristics of the four main groups in this study. Descriptions derive from taxon diagnoses, as referenced.
Descriptions in brackets not present explicitly in diagnosis, but characteristic of taxon nevertheless.

CARLSON: PENTAMERIDE BRACHIOPOD PHYLOGENY

817

03

0)

CJ

„

TJ oo , —

.§-2
CJ

g

£ bJ.
*0<

a3 <u
bU

cj ^
o oo ^

03 ^ '

S3 - c

C ^ <U
gOT)

2(J s
a x £

£cl«

£ 0
o3 y

u £

a3 rn

CJ ^
O OO
a3 i-h

g a x
o o

£ 22 g

2 £ S3

a

o

x

03

H

03

X

u

bD

a

o

ju
X "

.2

'C

o3

£,E

a

D. O

JO

03

>> U

</) 03

£

t; J2
«s a

bi Z> a, Z Prh

s -g

C/D ^

X> G*
o3 O

z£

bO

J-i

03

^ X
>» u

X‘ > CJ

'C o o «

« i: £ 8

> oo on

o3

a

CJ J-H

s °

2 c
'C «

03 Qh

> o

x

CJ

>

C

o

CJ u, ^

x .2 ^
aj >? s ^

X a
.2 O
d o- C

> £10

^ c

^ S s

\ QJ Ph

~£0

> °
G oj

8 c3

^CJ

3 X

2

X

3 3

Cj

G

03

CT O

CJ

P

<L> O

C/D

/

CJ

03

>

^ 6
X 00

o

CJ

CJ

Wh

Oh

C-

O

Co

-a

co

£

c 22
c.
P

a

QJ

£

C/d 1 / '

G QJ CJ

G 3

oj W) g

g O o

•2 o O

bf)

c

G

CJ

O

g

o

Q

D-'p
- CJ

G p

& 'g

G £
.2 ^
• • 75

Ph J-h
CJ ^

J

60

C 3c 03
O

X

C Co

P
22 £
^ o
1- M
<u O

s y
w

C

O

X

£

03

■— T O
G Oh

J J3
< 3

b bD
2 G
G X

G

O

X

£

03

Vh

O

Oh

<u

O

X

P<

o

Ph

S

.G Oh G tn CJ

x p< x 2 2

G G cj X x

tH C3 ^ ^

bi)

C

X

cj

as

X

XJ /-H J-H

CJ 1 ' (

_ - c3 1/5 -°

O X ct ~ Hr

_ 03 es

h 3 K o d

fl u u o u

jj c -8 s
& 8 .S tS

3 x c
73 < D

— Q,
<u

SJ C QJ

22 «3 c

o-OO

aj

03

"n.

C/D

‘x

^CJ

CJ

>

5 g

2

o ^
° g
> 1

■o 2

G u.

03

2 X

_ G
G

o3 s-h

O V

P^

> o

pH Q

C/3

C/J /— v

CJ rr, <Z) C/D

CJ g CJ UJ

o i! bij
b o -o 22
ax -c co
— a

2 o CJ G
•G IS X 'G

•g o y o

£ 2 ° 2

U cq ^pa

c a

o

x.,2

03 %

p 2

u ^

■3 = K w

X n O

3 a
X —

^ G

c ^

3 O

o 2
^ X

CJ

& s

3 S
° x

s ^
O o

^ X
CJ

o G
xLx

CJ

c

o

Z

o

G

O

Q

CJ

G

o

Z

CJ

l-H

x 2
&o

■a I
^ 3

818

PALAEONTOLOGY, VOLUME 36

developed independently from them. This phylogenetic hypothesis reveals the possible evolutionary
origins of characters that typify the rhynchonellides. In other words, the rhynchonellides embody
a mosaic of primitive and derived characters at various stages of transformation from their
‘syntrophiidine’ ancestry.

Pentameridina. Pentameridines typically attain large adult size; some pentameridines are among the
largest brachiopods known. They retain a deep spondylium and strong valve biconvexity. The
delthyrial (pedicle?) opening is frequently obscured or obstructed by the growth of the strongly
rostrate dorsal and ventral beaks. The most distinctive changes involve the dorsal cardinalia which
become elaborated (Text-fig. 4) from a single supporting plate (brachiophore plate) to a three-part

text-fig. 4. Highly schematic reconstructions of the dorsal cardinalia (left side only, as seen from the sagittal
plane) of generalized ‘pentameride’ taxa. Brachiophores (socket ridges) are black, brachiophore/brachial
plates are unshaded, brachial processes/crura are shaded. Sockets are represented as shaded ‘hooks’ on top
of brachiophores. Orthide morphology transforms to the generalized ‘syntrophiidine’ morphology, which then
transforms in one of three general ways: a, to the ‘camerellacean’ type; B, to the porambonitid and
rhynchonellide types; or c, to the parastrophinid and pentameridine types.

brachial plate (consisting of an inner plate, brachial process, and outer plate). Brachial processes
(not homologous with brachiophores) are likely to have served as lophophore supports.
Pentameridine brachial plates are typically longer than ‘syntrophiidine’ brachiophore plates. For
this reason, they often (but not always) enclose the adductor muscle field, unlike in the
‘syntrophiidines’.

Just as the position of the Pentamerida is debatable, so is the position of the Pentameridina.
Node Q, R, or T (Text-fig. 2) could serve to delimit the pentameridines, depending on the inclusion
of the parastrophinids, and/or the parallelelasmatids and Karakulina. Node R corresponds to the
point at which pentameridines are generally recognized today; twelve apomorphies characterize the
node, many of which are typical pentameridine features. Karakulina , originally classified as a
‘porambonitacean’ (Andreeva 1972), consistently clusters with the pentameraceans in this analysis.

CARLSON: PENTAMERIDE BRACHIOPOD PHYLOGENY

819

although many characters of this genus are still poorly known. Parallelelasmatidae, first classified
as a ‘syntrophiacean’ (Cooper 1956), was later considered to be a ‘porambonitacean’ family
(Williams 1962), and later still was placed in the Pentameracea (Amsden 1964, 1965); it also
consistently clusters with the pentameraceans in this analysis.

Consensus of results from Analysis I

The fifty percent majority rule consensus cladogram of the twelve equally most parsimonious
cladograms obtained in Analysis I is illustrated in Text-figure 5. This type of consensus diagram

CD

co
~o

m

CD

CD

CD

CD

JO

CO

CD

CO

JO

jD

CD

zz:

Q)

O

CD

c/5

— \

CO

_Q

Z3

)

JZ

F

<S)

"co

L-

\ —

Z

o

<

CD "D
CO

CD _0
c CO
0) -
^ CD

CD

CO

CD T5
CD \r)

Cl

O

CD

CO

CD

cO

"O

CD
CO
CD T3

_ -o E =

aE co 2

2 $ ,y cd
■F -Z > f

^ CO <D ^

CD

0 <D -n ®
CO CO •= CO
T3 TD ~ P

Z Z o '

Cl Cl -O

g 2 E

-<-• CO
CO C s-
O >, O

CO

CD

o

JO

"cD

c

o

o

<~ -LZ -i-
a ao

o CO ><
>.

co

CD

o

co

co

E

CO

o

CO

o

c

CD

CD

CO

1 CD

.E co
sz ~o

&I

-F =3

CO Zd

CO CO

co co

CO

E

CO

JO

CD

CD

CD

CO

"O

C

CD

CD

CO

T3

CD "O

= c o

CO -JZ

CO

E co
5 .2

co -9

■ F c
Cl co
>> c

Cl u
c
JO

o

CD
CO
C

CD

E

CO

CO C
CD

CD

CO

■D

C

CO

CD

CO

c

"c

CO

_Q

=5

6<I|-C/)OCQOLDWIl_lh[r(/)CLZQ.OOLiJW>a.(/)

text-fig. 5. Fifty percent majority rule consensus cladogram of twelve equally most parsimonious cladograms
from Analysis I. The numbers beside each node represent the percentage of the twelve cladograms that support
that node; unmarked nodes are supported by one hundred percent of the cladograms.

provides a good estimate of the nodes that are more or less consistently supported by the data. Only
three nodes are relatively poorly supported (by seventy-five percent or fewer of the cladograms);
the rest are quite robust. The topology of the consensus diagram is very similar to cladogram five
(Text-figs 2, 5). The Karakulina plus Parallelelasmatidae clade collapses to a polytomy with the
pentameridines at this level of consensus.

Billingsellidae

Nisusiidae

Matutellidae

Orthacea

Alimbellidae

Huenellidae

Tetralobulidae

Syntrophopsidae

Clarkellidae

Brevicameridae

Camerellidae

Eostrophiidae

Syntrophiidae

Porambonitidae

Lycophoriidae

Triseptata

Rhynchonellacea

Stenoscismatacea

Parastrophinidae

Karakulinidae

Parallelelasmat

Clorindinae

Gypidulinae

Enantiosphenidae

Stricklandiidae

Virgianidae

Pentamerinae

Subrianinae

CARLSON: PENTAMERIDE BRACHIOPOD PHYLOGENY

820

PALAEONTOLOGY, VOLUME 36

Analysis II

In the second analysis, characters were weighted by the value of their rescaled consistency indices
precisely according to their performance in the first analysis. The analysis produced three equally
most parsimonious cladograms of length 70 163 (weights are scaled to a base weight of 1000 = TO)
with a consistency index of 0-503 (Text-fig. 6). The consensus topology is generally similar to

CD

03 CD CD

CD

cc

TD

CD

0

cd

0

0

73

C/3

Q.

o

0

0

0

0

T3

0

0

0

■O

0

0

0

TD

0

0

T3

0

0

73

0

73

c

—

cd

22

cd

73

sz

73

3

73

E

—

"Jc

IE

o

CD

22

—

0

- z

CL

'■ —

-Q

■ —

0

0

CL

CL

S3

C/3

03

C.

C/3

Z3

C/3

0
-*— •
ZJ

"cd

o

0

SI

0

-Q

E

o

c

0

c

0

=3

O

0

0

0

0

o

>

0

0

E

0

O

"c/3

O

o

E

>

E

0

o

m

Z

6

<

CD

X

H

o

m

o

LU

CD

Q_

03

0

0

0

O

O

0

0

0

0

0

"0

E

' El

0

c:

C/3

o

sz

0

o

sz

"o

C/3

CL

Q.

o

o

o

0

c

CZ

o

>,

C/3

sz

0

C£

CD

0

0

73

0

"c

0

'sz

73

CL

C

o

j—.

3

C /3

ZsC

0

0

0

0

CL

Z

03

E

« a)

25 03

_0 C

CD 73
= C

oo -c
co

Cl O

CD

CO

~o

cz

Q- 03
>* cz

CD

03

73

3

c

0

JL

O

0 LU CO

CD

CO

cd g

03 *-

73 CD

"E E

CO co

CD

CO

C

"c

03

cd c n

E D 3
> CL CD

from Analysis II. Following Analysis I, characters were reweighted according to the value of their rescaled

consistency indices.

cladogram five (Text-fig. 2), although a few taxa change positions. Among the early
‘syntrophiidines’, the syntrophopsids become the sister group to the (Huenellidae plus
Tetralobulidae) clade. Porambonitids become the sister group to all the remaining derived
‘pentamerides’ and rhynchonellides. No major changes in the interpretation of character evolution
are required.

Analysis III

The rhynchonellides were removed from the analysis as an experiment, to see if the topology of
relationships among the ‘pentamerides’ would remain stable in their absence. The results are
different in several interesting respects. Forty-three cladograms of length two hundred and sixty-
eight, each with a C.I. of 0-404, were obtained. The consensus diagram of these forty-three
cladograms (Text-fig. 7) shows that resolution among the early ‘syntrophiidines’ collapses entirely;

CARLSON: PENTAMERIDE BRACHIOPOD PHYLOGENY

821

CD

CO

T3

3

0

CD

C

m

CD

CO

■O

CD

CO

-0

O m : — ' —

C / )
3
0

CD

CO

"O

CD

CO

"0

9r ft

o sz

£ -c

2 4j O

CD

CO

0

Q.

CD

0

« ~ 3
^ < H

CD

CO

■g

1/3

a. £
o

jz 72
Q. =
O CD
C=

C CD

>, 3

03 X

CD

CD

CO

CD

CO

"0 CD

= ft

2 1
O CD
CO -L

CD i?

CD

0

0

■0

0

0

TO

3

c

CD

T5

'k-

0

E

0

CD

T3

0

0

p

IE

0

0

-0

IE

0

td

"E

o

“0

3

3

Q.

o

0

0

0

3

03

E

0

0

0

0

0

C

0

0

c

3

0

sz

Cl

0

0

0

"O

0

0

C

"k_

0

0

0

c

CD

O

0

k

0

Q.

o

CL

O

.0

E

To

3

ZL

_0

3

c

3

■0

O

0

ZL

'c

0

E

0

3

0

>

E

To

H — >

c

0

0

i_

0

"0

Q.

d

0

O

CD

E

.0

kU

0

o

o

0

0

0

CJ

>

C

k-

0

3

CO

O

LU

C/3

CL

CL

X

CL

o

CD

LU

03

>

CL

03

text-fig. 7. Fifty percent majority rule consensus cladogram of forty-three equally most parsimonious
cladograms from Analysis III, in which rhynchonellides were removed. The numbers beside each node
represent the percentage of the cladograms that support that node.

most other nodes are very poorly supported, except among the pentameridines. Perhaps most
significantly, the Lycophoria and Triseptata clade breaks apart, and both taxa move down towards
the base of the cladogram. Lycophoria moves to the branch between the nisusiids and the
matutellids (in spite of a sizeable stratigraphical gap between them) consistent with their earlier
assignment to the orthides (Schuchert and Cooper 1932). Thus, without the influence of
rhynchonellide morphology, this problematic taxon shifts phylogenetic affinities significantly.
Stratocladistic methodology would seem to offer a potential resolution of this dilemma, by
considering both morphological and stratigraphical information simultaneously.

Analysis IV

In the final experiment, four characters (numbers 33, 37, 57, 60) were arbitrarily weighted four times
greater than the other characters. These particular characters, which all relate to the presence and
type of spondylium, cruralium, and septalium, were chosen because earlier investigations of
‘pentameride’ phylogeny (e.g. Schuchert and Cooper 1932; Ulrich and Cooper 1938) emphasized
the importance of characters related to the size and orientation of dental plates and brachial plates,
and the attachment of muscles relative to these plates. The results are quite different again from the
other topologies. Thirty-five cladograms of length three hundred and thirty-one, each with a C.I.
of 0-400, were obtained. In the consensus of these cladograms (Text-fig. 8), most of the nodes are

r

2,

t-

f

Billingsellidae

Nisusiidae

Matutellidae

Orthacea

Alimbellidae

Syntrophopsidae

Huenellidae

Tetralobulidae

Clarkellidae

Brevicameridae

Camerellidae

Eostrophiidae

Syntrophiidae

Porambonitidae

Lycophoriidae

Triseptata

Rhynchonellacea

Stenoscismatacea

Parastrophinidae

Karakulinidae

Parallelelasmat

Clorindinae

Gypidulinae

Enantiosphenidae

Stricklandiidae

Virgianidae

Pentamerinae

Subrianinae

§

2

£

Billingsellidae

Nisusiidae

Lycophoriidae

Orthacea

Matutellidae

Alimbellidae

Triseptata

Syntrophopsidae

Huenellidae

Tetralobulidae

Clarkellidae

Brevicameridae

Camerellidae

Eostrophiidae

Syntrophiidae

Porambonitidae

Parastrophinidae

Karakulinidae

Parallelelasmat

Clorindinae

Gypidulinae

Enantiosphenidae

Stricklandiidae

Virgianidae

Pentamerinae

Subrianinae

CARLSON: PENTAMERIDE BRACHIOPOD PHYLOGENY

Billingsellidae

Nisusiidae

Lycophoriidae

Porambonitidae

Eostrophiidae

Matutellidae

Orthacea

Alimbellidae

Huenellidae

Tetralobulidae

Syntrophopsidae

Clarkellidae

Camerellidae

Brevicameridae

Syntrophiidae

Triseptata

Parastrophinidae

Rhynchonellacea

Stenoscismatacea

Clorindinae

Gypidulinae

Karakulinidae

Parallelelasmat

Subrianinae

Pentamerinae

Enantiosphenidae

Stricklandiidae

Virgianidae

822

PALAEONTOLOGY, VOLUME 36

CD

03

XD

CD

03

"O

0

03

T3

O

0
0

03 'cO
O =J _
= 0 O

CQ Z Jj

Q_

O

0

0

H — ’

"c

o

JD

E

0

0

0

■0

CL

O

0
o o
CL LJJ

0

0

-g 0
= 0
0 0

-+— » -i—

0 03

^ o

0

0

T3

0

0 _
0-0 0
0 0 E

0

E £

0 CD

6 ■-
0 "O
o c

C 'C
0 O

0

0

T3

0

0 0

0 T3 0

0 0 0 T3

O 0 -q q_ CD CD -0

0 0 -.= q 0 "D 0

^^•SaECDmQ
3?0— O0(5oo
-O002=-^c'S-^
EOE:0mb0O0_

■— 00 >^TP0 cl :>, ;z 0-i— — -^roro d

CIl-WOOCDCOhCLCCWOO^CLW

0

0 0 0
0 0 O

“d g 0

E ^

In o

Q- E

o 9

0

0 0

--—00

j; j| 0 ,e

000E

"O ^ — —

'a. E E _q
>* 0 0

0

0

"D

g o

0 CD
g jz

CL

0 0

0

0

T3

.O CD

0

C

0

0

0

03

’c

0

• E 03

CL LU CO >

text-fig. 8. Fifty percent majority rule consensus cladogram of thirty-five equally most parsimonious
cladograms from Analysis IV, in which four characters were selectively weighted. The numbers beside each node
represent the percentage of the cladograms that support that node.

strongly supported, except within the pentameridines. Lycophoria, the porambonitids, and
Cambrotrophia , lacking muscle platforms, shifted towards the base of the cladogram, among the
orthides. A paraphyletic group in the middle of the cladogram, possessing well-developed
spondylia, septalia, and cruralia, bears some resemblance to the ‘Camerellacea’ ( sensu Nikiforova
1960), including (among others) the camerellids, brevicamerids, and parastrophinids. The
rhynchonellides, which lose the ventral platform, and pentameridines, which lose the dorsal
platforms, become sister taxa; relationships within the pentameridines are rearranged.

DISCUSSION

The evolution of character complexes

The evolutionary process of descent with modification necessarily results in character trans-
formation over time. Phylogenetic analysis can reveal the nature and direction of transformations,
and provide an evolutionary context for the identification and interpretation of evolutionary trends
(see Carlson 1992). Recall that characters derived with respect to one node are, at the same time,
primitive with respect to the next higher node in a cladogram. The everchanging perspective on the
relative state of primitive and derived, moving from the base of a cladogram to the final pair of sister
taxa, reveals the dynamic (mosaic) evolution of character complexes.

CARLSON: PENTAMERIDE BRACHIOPOD PHYLOGENY

823

The stereotypical ‘primitive’ articulate brachiopod, which derives largely from a composite of
early orthide fossils (see Williams and Wright 1965), includes the following elements: dorsal valves
of low convexity and ventral valves of generally low convexity; strong radial ornament common;
impunctate shell structure; long, straight hinge lines; weak, non-interlocking (deltidiodont) teeth
and sockets; wide interarea; large triangular delthyrial openings, often covered by a pseudo-
deltidium or deltidial plates, and notothyrial openings; a variably sized foramen at or near the apex
of the ventral valve; no mineralized support for the lophophore. As noted earlier, some of these
characters are retained by all ‘pentamerides’, while others are transformed slightly, and some
dramatically.

Relationship to substrate. The wide hinge lines and broad interareas in early brachiopods provide
a somewhat stable surface upon which the animal rested on soft, but presumably firm substrates
(Rudwick 1970; McGhee 1980/7). The earliest articulates possess delthyrial and notothyrial
openings in addition to a foramen at or near the apex of the ventral valve. The foramen is generally
interpreted as the pedicle opening, while the delthyrial opening may have accommodated only the
body of the adductor and diductor muscles extending from one valve to the other (see Rudwick
1970; Rowell and Caruso 1985; Carlson 1989). In brachiopods lacking an apical foramen (e.g.
‘pentamerides’), the pedicle, if present, is thought to have emerged from between the valves at the
delthyrial/notothyrial opening, although to my knowledge no direct fossil evidence exists to
support this assumption.

The hinge line is primitively wide in the ‘syntrophiidines’ with respect to both outgroup and
stratigraphical polarity. It first shortens and then lengthens somewhat in several groups
independently. Hinge line width can vary considerably among ‘syntrophiidines’, even within a
species; some specimens have essentially an astrophic (curved) hinge, while conspecifics may possess
distinct hinge lines. Astrophic hinges evolved twice from the strophic (straight) condition, first in
the rhynchonellides and again in the most recent common ancestor of the parastrophinids and
pentameridines. Most pentameridines are astrophic; stricklandiids and some gypidulids present
interesting exceptions. The delthyrial openings in pentameridines commonly become obstructed by
the growth of the beaks in very strongly biconvex ventral and dorsal valves, apparently preventing
the passage of a pedicle. In addition to preservational and palaeoenvironmental evidence, this has
lead to the interpretation that at least some pentameridines lacked pedicles entirely and became free-
living on soft substrates (i.e. Ziegler et al. 1966, 1968). Rhynchonellaceans, the only extant taxon
in these analyses, are astrophic, live on hard substrates attached by a pedicle, and possess well-
developed pedicle openings.

Hinge structures. The transition from deltidiodont to cyrtomatodont hinge structures has been
documented on a broad scale among all articulate brachiopods (Jaanusson 1971; Carlson 1989,
1992). The transition can be observed within the ‘pentamerides’ as well. In comparison with
inarticulate brachiopods, the lack of hinge structures in nisusiids and matutellids is primitive, small
teeth with non-interlocking articulation are derived, and large teeth with interlocking articulation
are derived relative to non-interlocking structures. Interlocking hinge structures appear to have
evolved twice independently among these taxa: in camerellids and in the porambonitid clade (at
Node M; Text-fig. 2). Taphonomic evidence (Sheehan 1978) also supports the morphological
evidence for cyrtomatodont dentitions in these taxa; they are only infrequently preserved in the
fossil record as disarticulated valves.

The possession of a cyrtomatodont dentition does not appear to be a shared derived character
of a single monophyletic subclade within the ‘pentamerides’, thus, cyrtomatodonts are not strictly
monophyletic. At a more universal level of analysis, however, cyrtomatodont dentitions in the most
recent common ancestor of the rhynchonellides and porambonitids may be a synapomorphy of the
cyrtomatodonts, with camerellids (possibly other brachiopods as well; see Jaanusson 1971)
convergent on the cyrtomatodont condition. More detailed phylogenetic analyses of all articulate
brachiopods will resolve this issue.

824

PALAEONTOLOGY, VOLUME 36

Mantle cavity volume. Overall body size increases considerably over the course of ‘ pentameride ’
evolution, as does valve convexity (Sapelnikov 1982). ‘Syntrophiidines’ are generally small, but
occasionally reach a fairly large size (e.g. Porambonites ); they commonly possess a strong fold and
sulcus as well. ‘Syntrophiidines’ first appear to increase convexity in lateral profile, by decreasing
their rate of whorl expansion during growth ( sensu McGhee 1980«). With shorter hinge lines, the
dorsal profile of more derived ‘syntrophiidines’ also becomes more rounded over time. Several taxa
approach a spherical shape (e.g. Lycophoria), which maximizes the ratio of body volume to surface
area.

As McGhee (1980u) has pointed out, several groups of biconvex articulate brachiopods appear
to maximize their mantle cavity volume, possibly as an adaptation for increasing the lophophore’s
size and food-gathering capabilities (see also Carlson 1992). Unfortunately, the ‘pentameride’
lophophore type is not known, but is presumed to be a spirolophe, which is the primitive condition
for brachiopods (by comparison with living inarticulates). Pentameridines achieve extremes in body
size and valve convexity, which is also consistent with the assumption that they were spirolophous
(La Barbera 1986). Rhynchonellides remain small to medium-sized and retain the near-spherical
mantle cavity shape. They are spirolophous, supporting only the proximal end of the lophophore on
short prong-like crura, which are probable homologues to pentameridine brachial processes.

Muscle platforms. Muscle platforms are structures in the posterior of either valve that serve to raise
the site of muscle attachments above the valve floor, to varying degrees. The elaboration of muscle
platforms and dorsal and ventral cardinalia are among the most significant evolutionary trends
within the ‘pentamerides’. Nearly all extant brachiopods possess tendinous muscles, in which
relatively short muscle bundles that attach to each valve are connected to one another by tendon
crossing the expanse of the mantle cavity (Rudwick 1970). Only thecideides, tiny brachiopods with
a small ventral platform, today possess columnar muscles in which the muscle bundle extends from
one valve to the other, with no tendon in between. Rudwick (1970) proposed that muscle platforms
evolve as an adaptive response in brachiopods that (1) possess columnar, rather than tendinous,
muscles and (2) are under selection for increasing valve convexity, presumably because of the
greater food-gathering capabilities a larger lophophore can attain.

Although ventral muscle platforms of various types have evolved independently several times in
articulate brachiopods (e.g. clitambonitaceans, gonambonitaceans, certain atrypides, etc.), the
spondylium appears to be a shared derived character of the ‘pentamerides’. A variety of types of
spondylia have been recognized (Kozlowski 1929; Schuchert and Cooper 1932). In a spondylium
simplex, the dental plates have converged and are supported by a single median septum.
A spondylium duplex is supported by a single structure that appears to have been formed by the
coalescence of two septae, possibly the distal extension of the dental plates themselves (Williams and
Rowell 1965, p. H153). The evolutionary transformation of spondylium type in this analysis
appears to change from pseudospondylia to sessile to simplex to duplex spondylia. Unfortunately,
neither the developmental origin nor the functional significance of the simplex versus duplex
spondylium is clear, making it difficult to evaluate the significance of this particular pattern of
character transformation. However, given the differences in the structure of these spondylia, the
multiple independent origins of simplex spondylia from the sessile condition and duplex spondylia
from parallel dental plates seem more likely (and just as consistent with the distribution of
characters in Text-fig. 2), although less parsimonious overall (Kozlowski 1929). Old (gerontic)
individuals of Porambonites commonly develop structures very similar to spondylia and cruralia
(Schuchert and Cooper 1932; Biernat 1965). If individuals can develop spondylium-like structures
within a life cycle, the multiple, peramorphic ( sensu Alberch et al. 1979) origins of spondylia over
evolutionary time are plausible.

Muscle platforms may also develop in the dorsal valve. The brachiophores (socket ridges) in
‘syntrophiidines’ are supported and united by short brachiophore plates (Text-fig. 4). When these
plates unite with and are supported by a median septum, the structure is called a septaliuin, which
may support the diductor muscles (functioning as a cardinal process), but never the adductor

CARLSON: PENTAMERIDE BRACHIOPOD PHYLOGENY

825

c

cc

o

m

m

O

Stratigraphies! rank

text-fig. 9. Plot of cladistic rank as it varies according to stratigraphical rank (following Gauthier et a 1 1988).
Stratigraphical rank was assigned at the series level (e.g. Lower Cambrian = 1, Middle Cambrian = 2, etc.)
according to the first appearance of the taxon in the fossil record. Cladistic rank was determined by counting
the number of nodes from the base of the cladogram to each taxon, and is scaled from 0 to 10.

muscles (Williams and Rowell 1965). A septalium is present in all ‘syntrophiidines’ except
Cambrotrophia, Lvcophoria , and the porambonitids, but appears to have evolved four times
independently. It is absent in the rhynchonellides and pentameridines. In some ‘syntrophiidines’
and pentameridines (and stenoscismataceans as a camarophorium), a cruralium develops anterior to
the septalium, which supports the adductor muscles and functions as a spondylium in the dorsal
valve. According to the results of Analysis I, a cruralium has evolved six times independently among
the ‘pentamerides’. Muscle platforms in ‘pentamerides’ appear to have been relatively easy to
construct and the selection pressure to construct them high.

In summary, morphological transformations previously known to occur within the
‘pentamerides’ can now be discussed with respect to a specific phylogenetic hypothesis of
relationships among all the named ‘pentamerides’ families. The pattern of transformations in all
characters can be compared simultaneously; the distribution of homologues, both primitive and
derived, as well as homoplastic characters is revealed. The order of acquisition of evolutionary
novelties (polarized by methods of outgroup comparison), including the loss or secondary
transformation of characters, may be traced on the cladogram and compared with the known
stratigraphical ranges of the analysed taxa.

826

PALAEONTOLOGY, VOLUME 36

Comparison of outgroup results and stratigraphical position

The fossil record provides series of morphologies in an ordered temporal sequence related to the
evolutionary time of origin and the direction of transformation of characters and character states.
Nevertheless, it is clear that the fossil record represents a more or less erratic sampling of the history
and diversity of life. Some critics have claimed that the record is too incomplete to record accurately
the true sequence of character transformation (Nelson 1978) and that ancestral taxa may either
never have been preserved as fossils or they may not appear in the record as early as their actual
time of origination (see Patterson 1981 ; Norell 1992). Despite these difficulties, agreement between
cladistic rank and stratigraphical rank is often quite good (Gauthier et al. 1988; Donoghue et at.
1989; Norell and Novacek 1992), as is the case in this analysis (Text-fig. 9).

Outgroup analyses provide criteria independent of relative first appearance in the fossil record for
evaluating the direction of character change in evolution. Studying the pattern of acquisition of
evolutionary novelties (apomorphies) in a cladogram obtained using outgroup methods, in
conjunction with the available geological evidence, may help to distinguish between alternative
explanations of the preserved stratigraphical record. One of the advantages of employing
stratocladistic methods (Fisher 1991, 1992) is that morphological and stratigraphical information
can be combined in a single analysis and the results compared with previous hypotheses.

Analysis I. A method for comparing stratigraphical first occurrence data with cladistic rank was
developed by Gauthier et al. (1988; further elaborated by Norell and Novacek 1992). If two sources
of information on polarity are congruent, they are positively correlated ; the greater the congruence,
the stronger the correlation. This method was used to compare the first occurrences of
‘pentameride’ families with their ranking in cladogram five (Text-fig. 2), although no attempt was
made to accommodate redundant ranks (Norell and Novacek 1992). A clear positive relationship
exists between the two variables (Text-fig. 9). Using ranked stratigraphical first occurrence as the
independent variable, a linear regression yields a correlation coefficient of 0602. Two outliers are
noticeable, the Enantiosphenidae and the Stenoscismatacea; eliminating them and recalculating the
correlation yields a coefficient of 0-748. Both first appear in the fossil record much later than would
be predicted on the basis of their cladistic ranking alone, indicating one of two things. Either their
true stratigraphical ranges extend further back in the record than currently known, or they share
common ancestry with a late-appearing and derived species in their sister taxon. Overall, the order
of appearance of ‘pentameride’ families in the fossil record agrees well with their ranking in a
phylogenetic diagram constructed independently of stratigraphical position.

An empirical example. An example from the ‘pentameride’ literature illustrates one of the difficulties
in relying exclusively on stratigraphical polarity in phylogenetic reconstruction. Prior to 1987, the
Eostrophiidae was the stratigraphically lowest ‘pentameride’ family, occurring in the Middle
Cambrian. The assignment of Syntrophioides to Clarkellidae (by Schuchert and Cooper 1931),
which extends the stratigraphical range of the clarkellids to the Middle Cambrian, is tentative and
may not survive revision of the group. As the stratigraphically lowest, Cambrotrophia has assumed
the role of ‘typical ancestral pentameride’ (e.g. Sapelnikov 1980). Therefore, I had predicted that
it would appear as a fairly primitive member of the ‘pentamerides’ in my analyses. Contrary to
expectations, eostrophiids appear near the middle of the cladograms, possessing an interesting
mixture of primitive and derived morphological features.

In 1987, Andreeva described a new genus, Tcharella , from the Lower Cambrian of Siberia. She
classified Tcharella in the Syntrophopsidae (and moved Cambrotrophia to this family as well,
although my analyses do not support her reassignment). With this new discovery, if the family
assignment is justified, Syntrophopsidae becomes the oldest ‘pentameride’ family known from the
fossil record. It also appears much closer to the base of the ‘pentameride’ clade in all cladograms
(Text-figs 2, 5-7) than does Eostrophiidae. In this instance, cladograms generated using outgroup
methods of polarity determination provided a basis for predicting where previous collecting biases

CARLSON: PENTAMERIDE BRACHIOPOD PHYLOGENY

827

had existed. In other words, recent collecting in the Cambrian of Siberia revealed that brachiopods
morphologically more primitive than Cambrotrophia (on the basis of this analysis) existed in the
fossil record earlier than previously thought, but were only recently collected and named. It is likely
that the stratigraphical ranges of some of the other primitive families (e.g. Huenellidae,
Tetralobulidae) will be extended further back in time, as collecting in Cambrian strata of less well-
known regions proceeds. This process of discovery cannot proceed indefinitely, but it is clear that
exploration and description of Cambrian fossils and strata in a number of remote areas is active and
ongoing, and yields new information regularly (e.g. Ushatinskaya 1986; Andreeva 1987 ; Popov and
Tikhonov 1990).

Phylogenetic analyses of the sort presented here establish morphological states and their
evolutionary transformations within systems of common ancestry without having to rely on the
collections of particular specimens from particular stratigraphical horizons that exhibit particular
combinations of characters. In other words, characteristics likely to have been present in the
common ancestor can be hypothesized based on the distribution of features in known specimens,
irrespective of whether a specimen has been collected that exhibits all those ancestral features
(de Queiroz and Gauthier 1990). As they are collected, new specimens will test existing hypotheses.

Previous views of ‘ pentameride' phytogeny and classification

A very brief review of the history of ‘pentameride’ classification reveals much controversy over the
status of ‘pentamerides’ as a unified group, and the identification of the group of brachiopods with
which ‘pentamerides’ share most recent common ancestry (see Muir-Wood 1955). The current view
(Williams 1968; Text-fig. 1) considers ‘pentamerides’ to have evolved from the orthides and given
rise to the rhynchonellides; each of these three groups is classified in a separate order of the
Articulata.

Schuchert and Cooper (1932) remains today the most detailed discussion of phylogenetic
relationships among ‘pentameride’ genera. In their tentative phylogenetic reconstructions, genera
that would today all be classified in the superfamily Porambomtacea (Biernat 1965) were placed in
four separate lineages and assigned to three different superfamilies (Text-fig. 10). Two lineages
derived from the Billingsellidae, one that gives rise to the pentameraceans via Syntrophiidae, and one
that leaves no descendant higher taxa. Two other lineages emerge from Orthidae, deep within the
Orthacea. Ulrich and Cooper (1938) briefly discuss possible phylogenetic relationships among these
brachiopods and are in general agreement with Schuchert and Cooper (1932). They tentatively
suggest that the rhynchonelloids may have evolved from the syntrophiids, which also gave rise to
the camerellids, and culminated in the parastrophinids. Comparing the phylogenetic tree in Text-
figure 10 to the cladogram in Text-figure 8 (where ventral and dorsal cardinalia were weighted
preferentially), certain similarities emerge, suggesting that these characters may have been given
particular weight in structuring relationships among the ‘pentamerides’.

In 1965, Biernat (in the Treatise on invertebrate paleontology ) proposed a classification that differs
considerably from that of Schuchert and Cooper (1932), but is comparable to the (partial)
classification in Cooper (1956). The Porambonitacea are united as a single (?monophyletic)
superfamily in the suborder Syntrophiidina (Table 1). Cooper (1956) placed both superfamilies
Pentameracea and Rhynchonellacea in the suborder Pentameroidea, perhaps as a reflection of the
close phylogenetic relationship envisioned between these taxa. Unfortunately, neither Cooper nor
Biernat presented much in the way of phylogenetic analysis of named ‘pentameride’ genera or
families; some sense of phylogeny must be inferred from the grouping of taxa into higher taxa
(although classifications may be arrangements of convenience and explicitly not phylogenetic in
structure; Cooper 1944).

Sapelnikov (1980) proposed a classification substantially different from previous sources. The
assignment of genera to families bears little resemblance to earlier schemes, and a number of genera
named prior to 1980 appear to have been omitted in the analysis. In addition to a new classification,
a hypothesis of phylogenetic relationships among the (newly reconstituted) families and subfamilies

828

PALAEONTOLOGY, VOLUME 36

PENTAMERACEA

ORTHACEA

text-fig. 1 0. Phylogenetic relationships among ‘ syntrophiacean ’ ( = ‘ porambonitacean ’), pentameracean, and
selected orthacean families (compiled and redrawn from Schuchert and Cooper 1932), illustrating the
apparently fragmented phylogenetic relationships among families (outlined in heavy black ovals) currently
classified in the Porambonitacea (Biernat, 1965). The Orthacea, as illustrated here, was reorganized by
Williams and Wright (1965) for the Treatise on invertebrate paleontology ; billingsellids and nisusiids were
removed and placed in a separate superfamily, the Billingsellacea. The Orthacea chosen as an outgroup in this
study reflects the more recent definition of the superfamily.

CARLSON: PENTAMERIDE BRACHIOPOD PHYLOGENY

829

was presented. Seven higher taxa emerge from the Huenellinae simultaneously at the base of the
Ordovician (Sapelnikov 1980, fig. 1); phylogenetic resolution is low.

It seems clear that classification of the ‘pentamerides’ has been fairly contentious for more than
a century. Major differences among systematists in both the naming and relative ranking of higher
taxa have resulted primarily from differences in the interpretation of ‘pentameride’ phylogeny,
specifically levels of character homology, relative to articulate brachiopod evolution.

Implications for ‘ pentameride' classification

As they are currently defined (Amsden and Biernat 1965), the Pentamerida, Syntrophiidina, and
Porambonitacea are each paraphyletic, the Camerellacea (Nikiforova 1960) is polyphyletic, and the
Pentameridina and Rhynchonellida are each monophyletic with respect to the result of the analyses
described here. Establishing the phylogenetic status of each of these named higher taxa with
reference to a working hypothesis of relationship (Text-fig. 2) is valuable and necessary if named
higher taxa are to play an interpretable role in macroevolutionary studies.

Controversy surrounds the field of classification, particularly in assessing the phylogenetic status
of existing higher taxa, and in naming new taxa and ranking those taxa in some kind of a hierarchic
scheme (see de Queiroz and Gauthier 1992). Newly named or rediagnosed higher taxa should be
monophyletic. Characters (synapomorphies) diagnose monophyletic taxa, and enable us to
determine whether a given organism is representative of the taxon or not. However, monophyletic
entities are systems of common ancestry that exist independent of our ability to recognize them (de
Queiroz and Gauthier 1990). Thus, we recognize snakes as tetrapods even though they lack limbs;
in this case, evolutionary character transformation is expressed as the loss of a character. It may also
be expressed as dramatic transformation (e.g. avian wing, mammalian inner ear).

While most systematists will admit that named taxa should have phylogenetic significance, there
is debate about whether or not paraphyletic groups (as only partial systems of common ancestry)
have phylogenetic significance. Many neontologists argue against the naming of paraphyletic
groups (e.g. de Queiroz and Gauthier 1990), while palaeontologists commonly argue in their favour
(e.g. Waller 1978). Particularly when dealing with fossils, a paraphyletic taxon has been named to
designate the group of plesions (sensu Wiley 1981) excluded from a derived clade; each taxon is
often given the same taxonomic rank (e.g. suborders Syntrophiidina and Pentameridina in the order
Pentamerida). If phylogenetic relationships among the organisms of interest are unknown, or very
poorly known, it is possible that paraphyletic taxa can be named by accident. However, when a
working hypothesis of phylogenetic relationship has been constructed (e.g. Text-fig. 2), paraphyletic
taxa can only be named on purpose (de Queiroz and Gauthier 1990), regardless of which taxonomic
philosophy one adopts.

In discussions of macroevolutionary phenomena, it is often useful to recognize groups of
organisms that share ecologically or functionally significant suites of primitive characters,
particularly if they also share similar extinction histories (Fisher 1985, 1991). Referring to these
known paraphyletic groups as, for example, non-avian dinosaurs or non-mammalian synapsids is
acceptable, but perhaps unnecessarily awkward. Informally, known paraphyletic groups are
denoted as such by enclosing their names in quotation marks (Gauthier 1986). To avoid confusion,
explicit reference should be made to an existing phylogenetic hypothesis (branching diagram) in
which membership in the paraphyletic group is clear. Without such a reference, dinosaurs could
represent either the paraphyletic group of fully terrestrial archosaurs or the monophyletic group
that also includes birds.

Taxonomic revisions necessarily invite confusion. Old taxon names rediagnosed have different
meaning; new taxon names are unfamiliar. The status of ‘pentameride’ higher taxa, as they are
currently diagnosed, with respect to these cladograms (Text-figs 2, 5-8) is clear. It is possible
formally to rediagnose the Pentamerida and Pentameridina on the basis of this phylogenetic
analysis. However, this paper focuses on characters rather than taxa - on the pattern of acquisition
of shared derived characters as represented in the branching diagrams themselves. It is debatable

830

PALAEONTOLOGY, VOLUME 36

whether Pentamerida is the most appropriate taxon name (irrespective of taxonomic rank) for the
clade of ‘pentamerides’, rhynchonellides, and their relatives. Such a decision must await the
completion of additional, complementary analyses by other brachiopod systematists, particulary
those working at lower taxonomic levels on the orthides and rhynchonellides.

Extinction of a paraphyletic taxon

Paraphyletic groups are characterized by the possession of derived characters and the retention of
primitive characters. The ‘syntrophiidines’ possess derived ‘pentameride’ features and can thus be
recognized as ‘pentamerides’, but also retain more primitive orthide characteristics than do either
of their descendants. The rhynchonellides and pentameridines retain primitive ‘ pentameride ’
features, but also possess unique sets of derived characters that distinguish them, as clades, from
their ancestors.

Of what evolutionary significance is the extinction of the paraphyletic "syntrophiidines’? Of the
monophyletic pentameridines? Or the survival of the monophyletic rhynchonellides? The
extinctions were apparently not clustered in time (at this level of resolution; Text-fig. 2); no mass
extinction eliminated all (and only) the primitive ‘pentameride’ families. The evolution of character
complexes over time, the acquisition of derived characters and the loss or transformation of
primitive characters, resulted in the recognition of successively younger groups of organisms as
distinct taxa. Thus, the process of evolution itself results in the pseudoextinction (Smith and
Patterson 1988; Fortey 1989) of paraphyletic taxa such as the ‘syntrophiidines’ or, more
universally, the ‘pentamerides’.

Comparing the features of the monophyletic survivors and victims, some interesting patterns
emerge. ‘Syntrophiidines’ are relatively small brachiopods with primarily strophic hinge lines and
non-interlocking hinge structures, strong biconvexity and fold and sulcus, no calcareous lophophore
supports, with variously developed muscle platforms in both the ventral and dorsal valves. They
lived on soft substrates, anchored by a pedicle in a manner presumably similar to their orthide
ancestors. Rhynchonellides retained the small ‘syntrophiidine’ adult size, strong biconvexity, and
fold and sulcus. However, they evolved astrophic hinge lines with interlocking teeth and sockets,
and live today attached to hard substrates by a strong pedicle. They also developed distinct supports
for at least the base of the lophophore. Some lost the ventral platforms; some elaborated platforms
in each valve. ‘Pentamerides’ retained the non-interlocking hinge structures and also, in general,
developed astrophic hinges, retained ventral but lost dorsal muscle platforms, retained a soft
substrate habitat but lost a functional pedicle as adults, and retained strong valve biconvexity but
grew to very large size. They also developed brachial processes thought to support the base of the
lophophore. The combination of astrophic hinge lines and non-interlocking teeth and sockets in
‘pentamerides’ was thought to be fundamentally unstable (Jaanusson 1971). Together with the lack
of a pedicle, these features may have contributed to the extinction of the group. Rhynchonellides
survived because of the particular combination of characters they possessed (interlocking hinge
structures and hard substrate habitat), given the environmental changes occurring at the end of the
Devonian (see Copper 1990).

Although this scenario makes logical sense, viewing the pattern of ‘pentameride’ extinction and
survival in these terms alone oversimplifies a more complex pattern. In other words, it successfully
accounts for a portion of the existing evidence, but ignores conflicting evidence. For example, many
strophomenides lacked a pedicle, lived on soft substrates, grew to large sizes, and had non-
interlocking hinge structures, but were very abundant and diverse through the entire Palaeozoic.
Was the additional combination of strong biconvexity and astrophic hinge lines sufficient to put
‘pentamerides’ at a strong selective disadvantage in the middle Palaeozoic? If so, why were
‘pentamerides’ themselves so abundant while they were extant (Ziegler et a/. 1968; Amsden 1969)?
They are often depicted as a classic specialized group, very well adapted for one mode of life
(apediculate, with their weighted posterior buried in the sediment) that they exploited in great
numbers, and were then eliminated by some change in their habitat to which they could not <jdapt

CARLSON: PENTAMERIDE BRACHIOPOD PHYLOGENY

831

(see Boucot and Johnson 1979). Detailed palaeoenvironmental investigations are more
appropriately suited to answer the how and why of patterns of extinction and survival. Nevertheless,
a phylogenetic context provides an explicit framework within which causal hypotheses of extinction
and adaptation may be generated and tested.

SUMMARY AND CONCLUSIONS

1. ‘Pentamerides’ played an important role in the early evolution of articulate brachiopods, yet
their phylogenetic relationships have not been studied in detail for the past several decades.
Phylogenetic relationships among named ‘pentameride’ families and rhynchonellide superfamilies
were reanalysed, using outgroup methods of polarity determination. A detailed working hypothesis
of ‘pentameride’ phylogeny and the supporting evidence on character distribution is presented. The
paraphyletic ‘syntrophiidine pentamerides’ gave rise, through the process of descent with
modification, to two clades: the pentameridines, a distinctive apediculate group extinct by the end
of the Devonian, and the rhynchonellides, which are represented little changed in the modern fauna.
This phylogenetic hypothesis will be tested in the future as genus-level analyses and stratocladistic
analyses (Fisher 1992; Maddison and Maddison 1992) are completed.

2. The phylogenetic hypothesis obtained can serve as a foundation for detailed studies of character
evolution within this important, early group of brachiopods. Strophic hinge line length decreases
steadily to the astrophic condition, interlocking hinge structures arise twice from the non-
interlocking condition and muscle platforms in the dorsal and ventral valves evolve several times
independently. These generally acknowledged patterns of morphological change, among others,
may now be examined in detail as they are expressed in a comprehensive pattern of relationship.
Levels of homology and patterns of convergence and parallelism are revealed ; primitive absence can
be distinguished from derived loss of characters. It is possible to study the dynamic mosaic of
character evolution among large numbers of characters simultaneously.

3. Outgroup methods can serve as an independent test of stratigraphical polarity methods. In this
study, agreement between the stratigraphical first appearance of ‘pentameride’ families and their
cladistic rank is quite good, suggesting that, at least in this example, both methods indicate the same
direction of character polarity in evolution. Outgroup methods can generate predictions about
ancestral character combinations independent of the empirical record, and provide a means of
generating phylogenetic hypotheses among taxa that appear more or less simultaneously in the
fossil record.

4. The phylogenetic status of named ‘pentameride’ higher taxa is established, at least with respect
to one hypothesis of relationship. Rhynchonellida and Pentameridina are monophyletic;
Pentamerida and Syntrophiidina are paraphyletic. More than simply categorizing the status of
named taxa, the nature of paraphyly is revealed explicitly. Comparisons between older classifications
and newer phylogenetic reconstructions reveal the macroevolutionary roles that named paraphyletic
taxa have played in the past.

5. What is the significance of the extinction of a paraphyletic taxon? In this instance, the extinction
of the ‘pentamerides’ signifies the end of an early stage in the evolutionary transformation of
articulate brachiopods. The ‘syntrophiidines’ suffer pseudoextinction in transforming to the
rhynchonellides (extant) and the pentameridines, which possess a combination of characters (strong
biconvexity, large adult size, lack of pedicle, non-interlocking dentition) that apparently rendered
them less able to adapt to changes in their habitat. There is value in recognizing, if only informally,
paraphyletic groups of taxa that possess unique and functionally or ecologically relevant
combinations of derived and primitive characters. This is particularly true when the derived state of
a character that defines a named subclade takes the form of loss or extreme transformation from
its primitive state. Paraphyletic groups should always be interpreted with respect to a specific
phylogenetic hypothesis so the nature of their paraphyletic status is clear. In this context, they may
be useful and informative in studies of character evolution.

832

PALAEONTOLOGY, VOLUME 36

Acknowledgements. I thank F. J. Collier, R. A. Doescher, R. E. Grant, R. B. Neuman and J. S. Thompson
for access to brachiopod collections at the U.S. National Museum of Natural History. J. Cooper very kindly
translated several reference articles from Russian. J. Fong is responsible for the illustrations. Finally, I thank
A. J. Boucot, C. H. C. Brunton, R. Cowen, D. C. Fisher, J. Gauthier, A. Rowell, S. Rudman and A. Williams
for discussions that relate, in one way or another, to many of the ideas expressed here. I particularly thank
A. J. Boucot, R. Cowen and A. Williams for reviewing a preliminary version of this paper. This research was
supported, in part, by grants from the National Science Foundation (BSR 8717424; BSR 9115589).

REFERENCES

alberch, p., gould, s. J., oster, g. F. and wake, D. b. 1979. Size and shape in ontogeny and phylogeny.
Paleobiology , 5, 296-317.

amsden, t. w. 1964 Brachial plate structure in the brachiopod family Pentameridae. Palaeontology , 7, 220-239.
1965. Suborder Pentameridina. 536-552. In moore, r. c. (ed.). Treatise on invertebrate paleontology. Part H.
Brachiopoda. Vol. 2. Geological Society of America and University of Kansas Press, Boulder, Colorado
and Lawrence, Kansas, 927 pp.

1969. A widespread zone of pentamerid brachiopods in subsurface Silurian strata of Oklahoma and the
Texas panhandle. Journal of Paleontology, 43, 961-975.

— boucot, a. J. and Johnson, j. g. 1967. Conchidium and its separation from the subfamily Pentamerinae.
Journal of Paleontology, 41, 861-867.

andreeva, o. n. 1972. Brachiopods of the Ordovician Kuragan Suite of the South Urals. Palaeontological
Journal. 1, 39 — 49.

1987. The Cambrian articulate brachiopods. Palaeontological Journal, 4, 27-38.

biernat, G. 1965. Suborder Syntrophiidina. 523-536. In moore, r. c. (ed.). Treatise on invertebrate paleontology.
Part H. Brachiopoda. Vol. 2. Geological Society of America and University of Kansas Press, Boulder,
Colorado and Lawrence, Kansas, 927 pp.

boucot, a. j. and Johnson, j. g. 1979. Pentamerinae (Silurian Brachiopoda). Palaeontographica, Abteilung A,
163, 87-129.

bryant, h. n. 1991. The polarization of character transformations in phylogenetic systematics: role of
axiomatic and auxiliary assumptions. Systematic Zoology, 40, 433-445.
buckman, s. s. 1906. Brachiopod homeomorphy: Pygope, Antinomia, Pygites. Quarterly Journal of the
Geological Society, London, 62, 433 — 455.

cann, r. l., stoneking, m. and wilson, a. c. 1987. Mitochondrial DNA and human evolution. Nature, 325,
31-36.

carlson, s. J. 1989. The articulate brachiopod hinge mechanism: morphological and functional variation.
Paleobiology, 15, 364—386.

— 1991. A phylogenetic perspective on articulate brachiopod diversity and the Permo-Triassic extinctions.
119-142. In Dudley, e. c. (ed.). The unity of evolutionary biology. Volume I. Dioscorides Press, Portland,
Oregon, 588 pp.

1992. Evolutionary trends in the articulate brachiopod hinge mechanism. Paleobiology, 18, 344-366.
cloud, p. E. 1941. Homeomorphy, and a remarkable illustration. American Journal of Science, 239, 899-904.
cooper, G. a. 1930. The brachiopod genus Pionodema and its homeomorphs. Journal of Paleontology, 4,
369-382.

1944. Phylum Brachiopoda. 277-365. In shimer, h. w. and shrock, r. r. (eds) Index fossils of North
America. M.I.T. Press, Cambridge, Massachusetts, 837 pp.

— 1956. Chazyan and related brachiopods. Smithsonian Miscellaneous Collections, 127, part 1, 1-1024,
part 2, 1025-1245.

— 1970. Generic characters of brachiopods. 194-263. In yochelson, e. l. (ed.). Proceedings of the North
American Palaeontological Convention, Field Museum of Natural History, Chicago, September 1969. Allen
Press, Lawrence, Kansas, 703 pp.

— 1972. Homeomorphy in Recent deep-sea brachiopods. Smithsonian Contributions to Paleobiology, 11,
1-25.

copper, p. 1990. Evolution of atrypid brachiopods. 35 — 40. In mackinnon, d. i., lee, d. e. and Campbell, j. d.

(eds). Brachiopods through time. A. A. Balkema, Rotterdam, Netherlands, 447 pp.
de queiroz, K. 1988. Systematics and the Darwinian revolution. Philosophy of Science, 55, 238-259.

— and donoghue, m. j. 1988. Phylogenetic systematics and the species problem. Cladistics, 4, 317-338.

— 1990. Phylogenetic systematics and species revisited. Cladistics, 6, 83-90.

CARLSON: PENTAMERIDE BRACHIOPOD PHYLOGENY

833

— and gauthier, j. 1990. Phylogeny as a central principle in taxonomy: phylogenetic definitions of taxon
names. Systematic Zoology, 39, 307-322.

1992. Phylogenetic taxonomy. Annual Review of Ecology and Systematics, 23, 449 480.
donoghue, M. J., doyle, J. a., gauther, j., kluge, a. and rowe, T. 1989. The importance of fossils in phylogeny
reconstruction. Annual Review of Ecology and Systematics , 20, 431-460.
eldredge, N. and cracraft, J. 1980. Phylogenetic patterns and the evolutionary process. Method and theory in
comparative biology. Columbia University Press, New York, 349 pp.
farris, J. s. 1989. The retention index and the rescaled consistency index. Cladistics, 5, 417-419.
felsenstein, J. 1978. Cases in which parsimony or compatibility methods will be positively misleading.
Systematic Zoology , 27, 401-410.

— 1983. Parsimony in systematics: biological and statistical issues. Annual Review of Ecology and
Systematics, 14, 313-333.

fisher, D. c. 1980. The role of stratigraphical data in phylogenetic inference. Geological Society of America,
Abstracts with Programs, 12, 426.

1981. The role of functional analysis in phylogenetic inference: examples from the history of Xiphosura.
American Zoologist, 21, 47-62.

— 1982. Phylogenetic and macroevolutionary pattern within the Xiphosurida. North American Paleont-
ological Convention III, Proceedings, 2, 175-180.

1985. Evolutionary morphology: beyond the analogous, the anecdotal, and the ad hoc. Paleobiology, 11,
120-138.

— 1988. Stratocladistics : integrating stratigraphical and morphological data in phylogenetic inference.
Geological Society of America, Abstracts with Programs, 20, A 186.

— 1991. Phylogenetic analysis and its application in evolutionary paleobiology. 103-122. In gilinsky, n. l.
and signor, p. w. (eds). Analytical paleobiology. Short Courses in Paleontology, Number 4. University of
Tennessee Press, Knoxville, Tennessee, 216 pp.

— 1992. Chapter 6. Stratigraphic Parsimony. 124-129. In maddison, w. p. and maddison, d. r. (eds).
MacClade : interactive analysis of phylogeny and character evolution. Version 3.0. Sinauer, Massachusetts,
398 pp.

fortey, r. A. 1989. There are extinctions and extinctions: examples from the Lower Palaeozoic. Philosophical
Transactions of the Royal Society of London, Series B, 325, 327-355.
gauri, k. l. and boucot, A. j. 1968. Shell structure and classification of Pentameracea M‘Coy, 1844.
Palaeontographica, Abteilung A, 131, 79-135.

gauthier, J. 1986. Saurischian monophyly and the origin of birds. 1-55. In padian, k. (ed.). The origin of birds
and the evolution of flight. Memoirs of the California Academy of Science, 8.

— kluge, a. G. and rowe, T. 1988. Amniote phylogeny and the importance of fossils. Cladistics, 4, 105-209.
hennig, w. 1966. Phylogenetic systematics. University of Illinois Press, Urbana, Illinois, 263 pp.
jaanusson, v. 1971. Evolution of the brachiopod hinge. Smithsonian Contributions to Paleobiology, 3, 33-

46.

klassen, G. J., mooi, r. d. and locke, a. 1991. Consistency indices and random data. Systematic Zoology, 40,
446-457.

kozlowski, r. 1929. Les brachiopodes gothlandiens de la Podolie Polonaise. Palaeontologica Polonica, 1,
1-254.

kulkov, N. p. 1968. [Structure of the brachial organ on brachiopods of the order Pentamerida]. Dokladv
Akademii Nauk SSSR, 180, 963-966 [In Russian].

la barbera, m. 1986. Brachiopod lophophores: functional diversity and scaling. 313-321. In racheboeuf, p. r.
and emig, c. c. (eds). Les Brachiopodes fossiles et actuels. Actes du ler Congres international sur les
Brachiopodes, Brest, 1986. Biostratigraphie du Paleozoi'que, No. 4, 500 pp.
lahusen, J. 1886. Notiz fiber die inneren Merkmale einer neuen Untergattung der Strophomeniden. Russische-
Kaiserliche Miner alogische Gesellschaft zu St Petersberg, Verhandlungen, 2, 219-222.
maddison, w. p. and maddison, D. R. 1992. MacClade : interactive analysis of phylogeny and character evolution,
version 3.0. Sinauer, Massachusetts, 350 pp.

-donoghue, m. j. and maddison, d. r. 1984. Outgroup analysis and parsimony. Systematic Zoology, 33,
83-103.

marshall, c. R. 1990. Confidence intervals on stratigraphical ranges. Paleobiology, 16, 1-10.

1991. Estimation of taxonomic ranges from the fossil record. 19-38. In gilinsky, n. l. and signor, p. w.
(eds). Analytical paleobiology. Short Courses in Paleontology, Number 4. University of Tennessee Press,
Knoxville, Tennessee, 216 pp.

834

PALAEONTOLOGY, VOLUME 36

McGhee, G. R. 1980a. Shell form in the biconvex articulate Brachiopoda: a geometric analysis. Paleobiology , 6,
57-76.

- 1980/7. Shell geometry and stability strategies in the biconvex Brachiopoda. Neues Jahrbuch fur Geologie
und Palaontologie, 3, 155-184.

muir-wood, h. m. 1955. A history of the classification of the phylum Brachiopoda. British Museum (Natural
History), London, 124 pp.

nelson, G. 1978. Ontogeny, phylogeny, paleontology, and the biogenetic law. Systematic Zoology , 27, 324-345.
Nikiforova, o. I. 1960. [Order Pentamerida.] 197-205. In sarycheva, t. g. (ed.). Osnovy paleontologii,
spravochnik paleontologov i geologov SSSR , Mshanki , Brakhiopody , 7. Akademiya Nauk SSSR, Moskva,
343 pp. [In Russian].

nixon, k. c. and davis, J. i. 1991. Polymorphic taxa, missing values and cladistic analysis. Cladistics , 7,
233-266.

— and wheeler, Q. d. 1990. An amplification of the phylogenetic species concept. Cladistics, 6, 211-223.
norell, m. a. 1992. Taxic origin and temporal diversity: the effect of phylogeny. 89-118. In novacek, m. j. and

wheeler, Q. D. (eds). Extinction and phytogeny. Columbia University Press, New York, 269 pp.

— and novacek, m. j. 1992. The fossil record and evolution: comparing cladistic and paleontologic evidence
for vertebrate history. Science , 255, 1690-1693.

novacek, m. j. 1992. Fossils, topologies, missing data, and the higher level phylogeny of eutherian mammals.
Systematic Biology , 41, 58—73.

patterson, c. 1981. Significance of fossils in determining evolutionary relationships. Annual Review of Ecology
and Svstematics, 12, 195-223.

- 1982. Morphological characters and homology. 21-74. In joysey, k. a. and Friday, a. e. (eds). Problems
of phylogenetic reconstruction. Systematics Association Special Volume No. 21. Academic Press, London
and New York, 442 pp.

platnick, n. I., Griswold, c. E. and coddington, J. a. 1991. On missing entries in cladistic analysis. Cladistics,
7, 337-344.

popov, l. Y. and tikhonov, y. a. 1990. [Early Cambrian brachiopods from South Kirgiziia]. Paleontologicheskii
Zhurnal, 1990, 33 — 45 . [In Russian],

rowell, a. j. 1977. Early Cambrian brachiopods from the southwestern Great Basin of California and
Nevada. Journal of Paleontology, 51, 68-85.

— and caruso, n. e. 1985. The evolutionary significance of Nisusia sulcata , an early articulate brachiopod.
Journal of Paleontology, 59, 1227-1242.

rudwick, m. j. s. 1970. Living and fossil brachiopods. Hutchinson University Library, London, 199 pp.
Sanderson, m. J. and donoghue, m. j. 1989. Patterns of variation in levels of homoplasy. Evolution, 43,
1781-1795.

sapelnikov, v. p. 1980. Sistema i filogeniia nadsemeistva Porambonitacea (Brachiopoda). [The taxonomy and
phylogeny of the Superfamily Porambonitacea (Brachiopoda)]. 3-22. In papulov, g. n. et al. (eds).
Palaeontology and stratigraphy of the Middle Palaeozoic Urals. Akademiya Nauk SSSR, Uralskii Nauchyi
Tsentr, Sverdlovsk, 93 pp. [In Russian],

- 1982. [Vascular and genital systems of the pentamerids]. 53-58. In khodalevich, a. n. et al. (eds). New
data on the systematics and morphogenesis of early Palaeozoic brachiopods. Akademiya Nauk SSSR, Uralskii
Nauchy Tsentr, Sverdlovsk, 263 pp. [In Russian],

— 1985. Morfologicheskaia i taksonomicheskaia evolutsiia brakhiopod. O triad pentamerid. [Morphologic and
taxonomic evolution of Brachiopoda, the Order Pentamerida.]. Akademiya Nauk SSSR, Uralskii Nauchyi
Tsentr, Sverdlovsk, 188 pp. [In Russian].

schuchert, c. and cooper, g. a. 1931. Synopsis of the brachiopod genera of the suborders Orthoidea and
Pentameroidea, with notes on the Telotremata. American Journal of Science, Series 5, 22, 241-251.

- 1932. Brachiopod genera of the suborders Orthoidea and Pentameroidea. Memoirs of the Peabody
Museum of Natural History, 4, 1-259.

sheehan, p. M. 1978. The hinging mechanisms of brachiopods - taphonomic considerations. Journal of
Paleontology, 52, 748.

smith, a. b. and patterson, c. 1988. The influence of taxonomic method on the perception of patterns of
evolution. Evolutionary Biology, 23, 127-216.

sober, E. 1983. Parsimony in systematics: philosophical issues. Annual Review of Ecology and Svstematics, 14,
335-357.

— 1985. A likelihood justification of parsimony. Cladistics, 1, 209-233.

ST Joseph, j. k. s. 1939. The brachiopod family Parastrophinidae. Geological Magazine, 78. 371^101.

CARLSON: PENTAMERIDE BRACHIOPOD PHYLOGENY

835

strauss, d. and sadler, p. m. 1989. Classical confidence intervals and Bayesian probability estimates for ends
of local taxon ranges. Mathematical Biology , 21, 411-427.
swofford, d. 1991. PAUP (Phylogenetic Analysis Using Parsimony ), version 3.0q. Distributed by the
author, Illinois Natural History Survey, 607 East Peabody Drive, Champaign, Illinois 61820.
tcherkesova, s. v. 1973. The vascular system in pentamerids. Paleontological Journal , 4, 464-472.
templeton, a. r. 1992. Human origins and analysis of mitochondrial DNA sequences. Science. 255, 737.
ulrich, e. o. and cooper, g. a. 1938. Ozarkian and Canadian Brachiopoda. Geological Society of America ,
Special Papers , 13, 1-323.

ushatinskaya, G. T. 1986. [Brachiopods of the Cambrian (review of the localities and some patterns of
geographic distribution)]. 8-34. In tamarinov, l. p. et. al. (eds). Problems in the paleobiogeography of Asia.
Joint Soviet-Mongolian paleontologic expedition. Trudy , 29. Nauka, Moscow, 93 pp. [In Russian],
waller, T. R. 1978. Morphology, morphoclines and a new classification of the Pteriomorphia (Mollusca:
Bivalvia). Philosophical Transactions of the Royal Society of London, Series B, 284, 345-365.
watrous, l. e. and wheeler, q. d. 1981. The out-group comparison method of character analysis. Systematic
Zoology, 33, 83-103.

wiley, e. o. 1981. Phylogenetics, the theory and practice of phylogenetic systematics. John Wiley and Sons, New
York, 439 pp.

— siegel-causey, d., brooks, d. r. and funk, v. a. 1991. The compleat cladist, a primer of phylogenetic
procedures. University of Kansas Museum of Natural History, Special Publication, 19, 1-158.
williams, a. 1962. The Barr and Lower Ardmillan Series (Caradoc) of the Girvan District, South-West
Ayrshire, with descriptions of the Brachiopoda. Geological Society of London, Memoir, 3, 1-267.

1968. Evolution of the shell structure of the articulate brachiopods. Special Papers in Palaeontology, 2,

1-55.

and bassett, M. 1991. [Compilation of Annual Reports (1990-1991). Treatise on invertebrate

paleontology. Part H, revision. Unpublished report],

and rowell, a. j. 1965. Morphology. 139-155. In moore, r. c. (ed.). Treatise on invertebrate paleontology.

Part H. Brachiopoda. Geological Society of America and University of Kansas Press, Boulder, Colorado and
Lawrence, Kansas, 927 pp.

and wright, A. 1965. Suborder Orthidina. 300-346. In moore, r. c. (ed.). Treatise on invertebrate

paleontology. Part H. Brachiopoda. Vol. 2. Geological Society of America and University of Kansas Press,
Boulder, Colorado and Lawrence, Kansas, 521 pp.

ziegler, a., boucot, a. j. and Sheldon, p. R. 1966. Silurian pentameroid brachiopods preserved in position of
growth. Journal of Paleontology, 40, 1032-1036.

cocks, l. r. m. and bambach, r. k. 1968. The composition and structure of Lower Silurian marine

communities. Lethaia, 1, 1-27.

SANDRA J. CARLSON

Department of Geology
University of California

Typescript received 1 September 1992 Davis,

Revised typescript received 20 January 1993 CA 95616, USA

APPENDIX

Descriptions of shared derived characters (and the primitive states from which they transformed) for all
internal nodes on cladogram number five.

Pentamerida

Node A: Ventral valve convexity greater than dorsal valve (from equal convexity); strong fold and sulcus
present (from absent); open delthyrium (from delthyrium covered by pseudodeltidium); pseudospondylium
present (from absent); mantle canal markings digitate (from saccate); brachiophore plates rudimentary (from
short); short dorsal median septum present (from absent).

Node B: Radial ribs present as costae (from costellae); pedicle opening as delthyrium (from foramen); dental
plates distinct (from absent).

Node C: More or less spherical lateral valve profile (from anteriorly compressed); strong valve biconvexity
(from moderate); dorsal valve convexity greater than ventral valve (from ventral valve greater); convergent

836

PALAEONTOLOGY, VOLUME 36

dental plates (from divergent); dental plates united with median septum (from not united); short ventral
median septum present (from absent); weak septalium present (from absent).

Node D: Valve outline subpentagonal (from subquadrate); short hinge line (from long); spondylium present
(from pseudospondylium); spondylium deep (from shallow); deep notothyrial cavity (from shallow);
brachiophore plates short (from rudimentary); convergent brachial plate orientation (from divergent).

‘ Early Syntrophiidina'

Node E: Posteriorly compressed lateral valve profile (from more or less spherical); dorsal and ventral valve
convexity approximately equal (from dorsal valve more convex); dorsal valve interarea orientation anacline
(from apsacline); hinge line of medium width (from short); four lateral septae in dorsal valve (from two);
dorsal adductor muscle attachment area quadrilobate (from other than quadrilobate).

Node F : Small adult valves (from moderate); fold and sulcus present on anterior two-thirds of valve only (from
entire valve); small hinge teeth (from large).

Node G: Ventral valve interarea reduced (from extensive); dorsal valve interarea reduced (from extensive);
spondylium simplex (from sessile); weak callosities around muscle attachment area in dorsal valve (from
absent).

Node H: Sulcus not extended into distinct tongue (fron tongue present); accessory septae supporting
spondylium not prolonged in front of spondylium (from prolonged); anterior dorsal adductor muscle scars
larger than posterior (from posterior larger); well-developed adductor muscle attachment area present anterior
to cruralium (if present) (from absent); dorsal adductor muscle attachment area quadrilobate (from other than
quadrilobate).

‘ Camerellacea

Node I: Lateral valve profile compressed posteriorly (from more or less spherical); bumpy or pustulose valve
ornament, if not radial ribs (from fenestrate); ornament present on anterior third of valve only (from entire
valve); fold and sulcus present on anterior third of valve only (from anterior two-thirds); extremely short hinge
line (from short); cruralium present (from absent).

‘ Late Syntrophiidina '

Node J : valve ornament smooth (from radial ribs); spondylium duplex (from simplex); ventral median septum
absent (from present and short); dorsal median septum long (from short); septalium absent (from present and
weak); callosities around muscle attachment area in dorsal valve absent (from present and weak).

Node K: Valve outline subquadrate (from subpentagonal); anteriorly compressed lateral valve profile (from
more or less spherical); weak fold and sulcus (from strong); fold and sulcus ornament smooth (from costate);
hinge line medium width (from short); cruralium simplex (from duplex).

Node L: Dorsal valve interarea obsolete (from reduced); spondylium wide (from narrow); brachial plates long
(from short); brachial processes blade-like (from absent); outer brachial plates present and short (from
absent); long septalium (from short); septalium on septum (from sessile); strong callosities around muscle
attachment area in dorsal valve (from absent).

Porambonitid clade

Node M: Large adult valves (from small); valve ornament as radial ribs (from smooth); interlocking hinge
structures (from non-interlocking); large hinge teeth (from small); dental plate orientation subparallel (from
convergent); spondylium absent (from present).

Node N: Valve length approximately equal to width (from width greater); subtriangular valve outline (from
subpentagonal); very strong valve biconvexity (from strong); well-developed cardinal process (from absent);
brachial plates short (from long); outer brachial plates absent (from present and short); brachial plates united
with cardinal process (from not united).

Node O: Valve length greater than width (from approximately equal); valve outline more or less spherical
(from subtriangular); radial ribs costellate (from costate); fold and sulcus lacking (from present and strong);
dorsal valve interarea reduced (from obsolete).

CARLSON: PENTAMERIDE BRACHIOPOD PHYLOGENY

837

Rhynchonellida

Node P: Small adult valves (from large); posteriorly compressed lateral valve profile (from more or less
spherical); moderately rostrate ventral valve beak (from not rostrate); growth lines absent (from present and
weak); ventral valve interarea orientation anacline (from apsacline); hinge line astrophic (from strophic);
delthyrium may be partly covered by deltidial plates (from open delthyrium); brachiophores long and curved
(from short and blunt).

Node Q: Moderately rostrate ventral beak (from not rostrate); ventral valve interarea obsolete (from reduced);
astrophic hinge line (from strophic); pedicle opening absent (from present as delthyrium); spondylium
supported on high septum (from low); ventral median septum long (from absent); ventral valve muscle
impressions confined to spondylium (from not confined).

Pentameracea

Node R: Valve length approximately equal to width (from width greater); ventral valve convexity greater than
dorsal valve (from dorsal valve greater); strongly rostrate ventral valve beak (from moderately rostrate); growth
lines absent (from present and weak); fold and sulcus extends the entire length of the valves (from anterior §
only); medium hinge width (from short); ventral median septum does not extend anterior to spondylium (from
extends beyond); mantle canal markings pinnate (from digitate); dorsal median septum absent (from long);
adductor muscle attachment area anterior to cruralium absent (from present and well-developed); adductor
muscle attachment area quadrilobate in shape (from other than quadrilobate); callosities around muscle
attachment area in dorsal valve absent (from present and strong).

Node S : Valve convexity equal (from ventral valve greater) ; valve ornament present only on anterior § of valve
(from present on entire valve); fold and sulcus absent (from present and strong); if present, fold and sulcus
only on anterior third of valve (from present on entire valve); ventral valve muscle impressions not confined to
spondylium (from confined); subparallel brachial plates (from convergent).

Node T: Very deep spondylium (from deep); brachiophores poorly developed (from short and blunt); long
outer brachial plates present (from short).

Node U : Large adult valves (from small); valve outline subtriangular (from subpentagonal); valve biconvexity
very strong (from strong); delthyrial cavity shallow (from deep); short ventral median septum (from long).
Node V: Very large adult valves (from large); fold and sulcus absent (from present and strong); shallow
notothyrial cavity (from deep); rod-like brachial processes (from blade-like); outer brachial plates absent (from
present and long); adductor muscle attachment area anterior to cruralium well-developed (from absent); outer
lamellar shell layer in dorsal and ventral valves not intersecting (from intersecting).

Node W : Anteriorly compressed lateral valve profile (from more or less spherical); moderate valve biconvexity
(from very strong); ventral valve beak not rostrate (from strongly rostrate); outer lamellar shell layer in dorsal
valve not continuous (from continuous); inner lamellar shell layer in dorsal valve present (from absent);
lamellar layer in ventral median septum absent (from present).

Node X: Valve length greater than width (from equal proportions); valve ornament as radial ribs (from
smooth); weak growth lines present (from absent); long spondylium (from short); deep spondylium (from very
deep); short outer brachial plates present (from absent); prismatic layer in ventral valve wedged (from not
wedged).

Node Y: Subparallel dental plate orientation (from convergent); long ventral median septum (from short);
cardinal process occasionally present as a longitudinally striated mound (from absent); subparallel brachial
plates (from convergent); adductor muscle attachment area anterior to cruralium absent (from well-
developed); outer lamellar shell layer in dorsal and ventral valves intersect (from not intersecting).

BIOLOGY AND EVOLUTION OF THE NASAL
REGION IN TREMATOPID AMPHIBIANS

by DAVID W. DILKES

Abstract. Postmetamorphic ontogeny of the nasal region in trematopid amphibians (Temnospondyli:
Dissorophoidea) from the Lower Permian of Texas is characterized by the early appearance of a narial flange
which is identical to that of an adult. Subsequent ontogenetic changes are the posterior expansion of the
external naris, positive allometric growth of rostral length, and the development of caniniform teeth on the
premaxillae and maxillae. It is hypothesized that the narial flange of dissorophoids, regardless of the details
of its morphology, was an adaptation for terrestriality, and probably acted in a manner analogous to the
primary concha of reptiles to enhance olfactory receptivity and improve the efficiency of moisture reclamation
from exhaled air. Evolution of caniniform teeth in the Trematopidae apparently localized and increased cranial
stresses in the rostrum during eating. Consequently, the plesiomorphic dissorophoid narial flange was modified
in trematopids to fulfil an additional role of cranial reinforcement. Expansion of the external naris was a
subsequent event, and probably a result, rather than a cause, of any alterations to cranial stresses. Existence
of a salt gland in the expanded external naris is equivocal and not necessary for an interpretation of
trematopids as terrestrial amphibians.

Among Palaeozoic tetrapods, the morphologically diverse temnospondylous amphibians of the
clade Dissorophoidea are a favourite topic for palaeobiological discussion (e.g. Carroll 1964;
DeMar 1968; Boy 1972; Bolt 1974 a; Milner 1982; Bolt and Lombard 1985). Dissorophoids are
limited traditionally only to taxa of the Carboniferous, Permian, and Early Triassic; Trueb and
Cloutier (1991) on cladistic analysis included a monophyletic Lissamphibia within the Dis-
sorophoidea. Considerable ontogenetic data are known for these fossil amphibians (e.g. Romer
1939; Boy 1974; Bolt 19746, 1977a, 19776, 1979; Olson 1985; Dilkes 1991) and have provided
insights into the biphasic life history of dissorophoids, and into the origin of the Lissamphibia.

Two families of dissorophoids, the Dissorophidae and Trematopidae, will be discussed in this
paper with the emphasis placed upon trematopids. In a re-evaluation of the Trematopidae (Dilkes
1990), a new genus, Phonerpeton , was erected on the basis of specimens first described by Olson
(1941) and other, undescribed, material collected since his publication. Three species of
temnospondyls are assigned currently to the Trematopidae (Berman et al. 1987; Dilkes and Reisz
1987; Dilkes 1990): Phonerpeton pricei , Acheloma cumminsi (senior synonym of Trematops (Dilkes
and Reisz 1987)), and Anconastes vesperus (Text-fig. 1). Until recently, all known specimens of
trematopids were restricted to the Early Permian of Texas, Oklahoma, New Mexico, and Ohio.
A report of a trematopid from the Upper Rotliegendes of Germany (Martens 1990) was the first
definitive occurrence outside North America. The Trematopidae is a monophyletic taxon diagnosed
by the presence of caniniform teeth on the premaxilla and maxilla, an elongated external naris with
a medially placed narial flange that contacts the antorbital bar, a median vomerine septum, an
inflected medial rim of the adductor fossa, and by the absence of dermal sculpturing along the
dorsal border of the otic notch (Dilkes 1990).

One of the most intriguing features of trematopids is the presence of an elongated external naris
that superficially resembles a key-hole (Text-fig. 1). First recognized by Williston (1909), this
character has been cited often as the defining feature of the family (Olson 1941 ; Vaughn 1969; Olson
1970; Eaton 1973). An elongated external naris has, however, evolved independently in the
dissorophoid genera Ecolsonia cutlerensis (Berman et al. 1985) and Mordex calliprepes (Milner

{Palaeontology, Vol. 36, Part 4, 1993, pp. 839-853]

© The Palaeontological Association

840

PALAEONTOLOGY, VOLUME 36

text-fig. 1 . Current members of the Trematopidae. a, Acheloma cumminsi (redrawn from Dilkes and Reisz
1987). b, Phonerpeton pricei (redrawn from Dilkes 1990). c, Anconastes vesperus (redrawn from Berman et al.

1987). Scale bars: a, c = 30 mm; b = 20 mm.

1986). Associated with the elongated external naris of trematopids is a distinctive, laterally concave
bone within the nasal chamber known as the narial flange (Text-figs 1-3). Its pronounced lateral
curvature permits an extensive contact with the antorbital bar. Superficially, the narial flange
appears to be a single lamina, but it is composed of three separate bones; anteriorly, the first is from
the nasal, the second from the prefrontal, and the third (and smallest) from the lacrimal (Bolt 1974a;
Dilkes 1990). Contact between successive bones that comprise the narial flange varies from a short
overlapping suture to a butt-like suture.

Taxa of the Dissorophidae also have a narial flange that is made of contributions from the same
three bones (Bolt 1974a, 1974c), but the resultant morphology is different. The exceptionally well-
preserved material of Doleserpeton, which according to Bolt (1977a, 1979) is a probable junior
synonym of Tersomius , from the early Permian fissure fills near Richard's Spur, Oklahoma,

DILKES: TREMATOPID NASAL REGION

841

text-fig. 2. Partial left half of skull of small trematopid (cf. Phonerpeton sp.; MCZ 2475) lacking braincase
and most of palate, a, exterior dorsolateral view, b, interior, ventromedial view, c, exterior of rostrum in
anterolateral view. D, interior of rostrum in posteromedial view. Concave narial flange is visible in B, c and d.

Scale bars: a-b = 20 mm, C-D = 10 mm.

provides the most information on the dissorophid narial flange. The contributions from the
prefrontal and lacrimal are separate laminae that are similar in size and lie against one another to
create a bilaminar structure which as a unit contacts the lamina of the nasal. The dissorophid narial
flange lacks the pronounced lateral concavity of the trematopid flange. Details of the narial flanges
of Ecolsonia and Mordex are unknown. Presence of a narial flange is restricted currently to
trematopids and dissorophids (Bolt 1974a; Dilkes 1990).

To date, the only discussion of ontogeny for the Trematopidae was by Olson (1985) who compared
a larval form with adult trematopids. He observed relative reductions in the sizes of the orbits
and interpterygoid vacuities, and a posterior elongation of the jaw suspensorium. However, this
larva is not a trematopid, though it can be identified confidently as a dissorophoid (Dilkes 1991).
New ontogenic data on the nasal region of trematopids was discovered during the course of the
revision by Dilkes (1986) and its discussion herein is the first for the family.

The most recent proposal for the evolution of the trematopid nasal region was by Bolt (1974a)
in which he hypothesized that the lateral expansion of a nasal salt gland led to the posterior
elongation of the external naris as the region occupied formerly by bone became filled with
glandular tissue. Bolt argued further that cranial stress patterns were modified in the rostrum by this
narial elongation, and the narial flange of trematopids acquired its unique morphology to provide
additional support for the rostrum. No further discussion on the evolution of the trematopid nasal
region has followed, and Bolt’s theory appears to have been generally accepted. However, no
ontogenetic data were then available to him.

842

PALAEONTOLOGY, VOLUME 36

text-fig. 3. Rostrum of Phonerpeton pricei. a, dorsal view, b, rostrum with roof over nasal chamber removed
to reveal dorsal surface of palate. Position of lower edge of the narial flange is shown by dashed lines. Note
relative locations of the palatine ridge, narial flange and choana. c, schematic cross section through rostrum
at level indicated by dashed line in A. Scale bar = 20 mm.

In this paper, the previously unpublished, new ontogenetic data from Dilkes (1986) are presented
and an alternative evolutionary hypothesis, derived in part from Bolt (1974a, 1974c), is proposed
for the trematopid narial flange and expanded external naris. In addition, a new physiological
interpretation of the dissorophoid narial flange based upon analogous structures in extant
ectotherms is presented.

MATERIAL

Discussion of the ontogenic growth of the trematopid skull is limited to the rostrum due to the
nature of available specimens. All but one of the trematopids discussed in this paper are from the
Lower Permian (Wichita Group) of north-central Texas. Terminology for their stratigraphical
positions follows Hentz (1988) and Hentz and Brown (1987), and precise locality data are given in
Dilkes (1990). MCZ (Museum of Comparative Zoology, Harvard University) 2531 was collected by
Dr L. I. Price in 1937 from the Cowan Ranch locality in the Petrolia Formation (formerly Belle
Plains Formation). Mr R. V. Witter collected MCZ 1417 and 1419 (holotype of Phonerpeton pricei)
in 1936 and AMNH (American Museum of Natural History) 7150 in 1945 from the Archer City
Bonebed in the Archer City Formation (formerly Putnam Formation). MCZ 2313 was collected by
Dr A. S. Romer in 1952 from the Archer City Bonebed. The only specimen not from the Wichita
Group is MCZ 2475 which was collected in the undivided Clear Fork Group (formerly Arroyo
Formation) near West Coffee Creek, Baylor County, Texas.

Abbreviations used in the Text-figures

c.nas.pr.

cavum nasi proprium

1

lacrimal

pf

postfrontal

ch

choana

l.a.

lamina ascendens

pm

premaxilla

col

intercoronoid

l.e.e.

lateral exposure of ectopterygoid

po

postorbital

coll

precoronoid

I.e.p.

lateral exposure of palatine

prf

prefrontal

ct

cultriform process

l.div.

lateral diverticulum

Pt

pterygoid

d

dentary

m

maxilla

q

quadrate

e.nar.

external naris

m.v.s.

median vomerine septum

qj

quadratojugal

e.nas.gl.

external nasal gland

mpal

maxillopalatine

s.n.

septum nasi

e.o.

eminentia olfactoria

n

nasal

sm

septomaxilla

ect

ectopterygoid

NF

narial flange

sp

splenial

eo

exoccipital

op

opisthotic

spp

postsplenial

f

frontal

org.J.

organ of Jacobson

sq

squamosal

i-gl-

intermaxillary gland

P

parietal

st

supratemporal

in.f.

internarial fenestra

P-C.

posterior concha

t

tabular

in.p.

internarial pit

p.r.

palatine ridge

V

vomer

J

jugal

pal

palatine

DILKES: TREMATOPID NASAL REGION

843

text-fig. 4. Antorbital view of juvenile trematopid (MCZ 2531). a, dorsal view, b, ventral view. 0, right lateral
view. Narial flange of this juvenile is similar in size and shape to that of an adult as shown in Text-figure 2.

Scale bar = 10 mm.

CRANIAL ONTOGENY OF THE TREMATOPIDAE

In his review of the Trematopidae, Olson (1941) assigned several skulls of small trematopids to two
new species of Acheloma. The types of these species were assigned to Phonerpeton pricei by Dilkes
(1990). Many of the skulls that were referred by Olson to his new species are incomplete and poorly
preserved, but probably belong to Phonerpeton.

In addition to the referred skulls, several partial skulls of younger juveniles were collected, but
not described by Olson (1941). A precise ontogenetic assessment of these skulls is difficult, but they
can be interpreted as postmetamorphic by comparison with descriptions of juvenile dissorophoids
(Milner 1982; Olson 1985; Dilkes 1991) judged to be larvae by the criteria of Boy (1974). The small
trematopids discussed in this paper are much larger than the known larval dissorophoids, have more
prominent dermal sculpturing, have palatal tusks on the vomer and palatine that are absent on
larvae, and their cranial sutures are more tightly knit than in larvae. As these skulls are different
sizes (and presumably different ontogenetic ages), it is possible to reconstruct a partial
postmetamorphic growth series of trematopids. Unfortunately, these juvenile trematopid specimens
are also fragmentary, and although found with adult skulls that can be identified as Phonerpeton ,
none of the diagnostic features of either Phonerpeton or Acheloma is preserved. Therefore, they will
not be assigned to any particular genus. Nonetheless, the statements given below apply equally to
both genera. Subadult and adult stages of Phonerpeton pricei will be used to complete the
ontogenetic sequence because specimens are readily available.

The smallest available postmetamorphic specimens of trematopids (MCZ 2531) are two
antorbital regions: one from an individual with an estimated midline skull length of 43 mm and the
second with an estimated skull length of 53 mm. Most of the information given below is derived
from the better preserved, larger specimen (Text-fig. 4). Both are gently convex anteriorly in dorsal
view. An oval internarial fenestra is situated on the rostrum at the midline junction between the
premaxillae and nasals. The internarial fenestra opens ventrally into a deep internarial pit that is
formed by the dorsal deflection and separation of the vomers. A median vomerine septum separates
the two nasal chambers, but does not contact the roof of the rostrum.

The narial flange of MCZ 2531 is a massive sheet of bone that projects deeply into the nasal
chamber and contacts the antorbital bar just in front of the orbit. No sutures are visible on the
narial flange; nonetheless, it is virtually identical to the concave flange of adult trematopids. The
posterior expansion of the external naris of MCZ 2531 is only slightly greater than the size of the
true narial opening and, therefore, does not meet the narial flange as it does in adult trematopids
(Text-fig. Ia-b). The prefrontal is excluded from the external naris as an apparent consequence of

844

PALAEONTOLOGY, VOLUME 36

text-fig. 5. Antorbital region of juvenile trematopid MCZ 1417. a, dorsal view, b, ventral view, c, left lateral

view. Scale bar = 10 mm.

text-fig. 6. Phonerpeton pricei. a, MCZ 1419. b, AMNH 7150. Both specimens are shown in left lateral view.
Differentiation of caniniform teeth is apparent in AMNH 7150 relative to MCZ 1419. Scale bar = 20 mm.

the small narial expansion. The marginal teeth are large, but lack the caniniform specialization that
is characteristic of adult trematopids.

A third antorbital region (MCZ 1417; Text-fig. 5) from a slightly larger individual (estimated
midline skull length = 56 mm) has a more elongate and narrower antorbital region than that of
MCZ 2531. The posterior division of the naris is larger and considerably closer to the contact of
the narial flange and antorbital bar. Marginal teeth are differentiated to the state of canine peaks
as defined by Chase (1963).

The holotypic skull of Phonerpeton pricei (MCZ 1419; Text-fig. 6a) with a skull length of 73 mm
has a laterally constricted antorbital region with slight expansions on the premaxilla and maxilla for
canine peaks. The posterior elongation of the naris is relatively greater than that of MCZ 1417 and
reaches the narial flange. Development of caniniform teeth has not, however, proceeded beyond
that of MCZ 1417 and it is herein considered to be a subadult.

AMNH 7150 (Text-fig. 6b; Dilkes 1990, fig. 3) is from an osteologically mature individual (skull
length = 77 mm) as shown by the advanced ossification, rugosity of dermal sculpturing, and highly
interdigitated cranial sutures. The shape of the antorbital region and the expansion of the external
naris are similar to that of MCZ 1419. Caniniform teeth are fully differentiated.

These specimens enable us to recognize several stages in the postmetamorphic ontogenetic growth
of the trematopid skull. One early stage is the formation of a narial flange which is identical in
position and shape to that of an adult. In fact, a narial flange may appear as early as the larval stage.
A specimen of a larval dissorophoid has the nasal portion of a narial flange clearly visible within
its left elongated external naris (Dilkes 1991). It is conceivable that dissorophids and trematopids

DILKES: TREMATOPID NASAL REGION

845

had similar premetamorphic growth patterns and the narial flange might have developed as early
as the late larval stage in both families. As postmetamorphic growth in trematopids progressed, the
external naris expanded posteriorly to contact the flange along the antorbital bar, and the antorbital
region changed from a broadly rounded outline in dorsal view to one that is more elongate and
narrower. Expansion of the external naris in adult specimens of trematopids does not occur
posterior to the contact of the narial flange with the antorbital bar. Hence, it can be inferred that
this contact between the narial flange and antorbital bar is the limit for the expansion of the external
naris. Since the expansion of the external naris is observed to occur together with the elongation of
the antorbital region, one may infer further that major allometric growth of the rostrum halts when
this limit is reached. Additional isometric growth of the entire skull would probably occur after the
major allometric changes were finished as suggested by a slightly larger skull (MCZ 2313, skull
length = 88 mm) of Phonerpeton that has similar proportions to that of AMNH 7150. The final
ontogenetic stage is the completion of premaxillary and maxillary caniniform teeth.

EVOLUTION OF THE TREMATOPID EXTERNAL NARIS
Previous research

The possible function of the elongate external naris of trematopids has been the subject of
speculation by Williston (1909), Olson (1941), and most recently Bolt (1974a). Williston (1909)
believed that the posterior division was an antorbital vacuity and the anterior division was the
functional naris. His interpretation was based apparently upon a superficial resemblance between
the posterior portion of the external naris and the characteristic antorbital fenestra of archosaurian
reptiles. Olson (1941) also referred to the posterior portion as an antorbital vacuity, but rejected the
notion of an unknown gland lodged within the vacuity. This statement was probably a reference to
Broom (1913) who first suggested that the archosaurian antorbital vacuity housed a gland. Olson
argued that the large medial chamber with the narial flange (his ‘descending septum of the nasal’)
communicated only with the anterior division of the external naris whereas the posterior division
opened only into the choana. The medial chamber was thought to contain a vomeronasal organ
and, thus, was concerned only with olfaction. The posterior division acted as the functional narial
opening. This interpretation is, however, refuted by the location of the septomaxilla which is found
primitively in tetrapods along the posterior margin of the external naris (Panchen 1967) to support
the posterior wall of the tubular vestibulum, the first of three major subdivisions of the tetrapod
nasal cavity (Parsons 1967). Prominent septomaxillae are present in Acheloma (Text-fig. 1a; Dilkes
and Reisz, 1987) and Phonerpeton (Text-fig. 1b; Dilkes 1990) at the junction between the anterior
and posterior portions of the external naris. Hence, the anterior division of the trematopid external
naris was the functional narial opening.

The term antorbital vacuity was finally abandoned by Romer (1947) who referred to the entire
opening as a posteriorly elongated external naris. All subsequent authors have followed Romer’s
terminology.

In the most recent consideration of the trematopid nasal chamber and the elongated external
naris. Bolt (1974a) thought it unlikely that the narial flange subdivided the nasal capsule to either
increase epithelium or separate functionally different chambers because among living amphibians
most flanges that protrude into the nasal sac are derived from the cartilaginous nasal capsule and
not the surrounding dermal bones. Instead, he proposed that the nasal capsule was situated lateral
to the narial flange. Placement of the nasal capsule lateral to the flange required a reinterpretation
of a raised ridge on the dorsal surface of the palatine that Bolt observed in the rostrum of a small
trematopid skull. This palatine ridge (Text-figs 3-5) has been reported in other temnospondyls (e.g.
Eryops\ Swain 1941; Parotosuchus : Warren 1980, fig. 3a) and is generally thought to brace the
posterior wall of the nasal capsule. Bolt argued that the palatine ridge was not a necessary indicator
of the shape of the capsule and perhaps marked the presence of some other cartilaginous structure.

Bolt (1974a) also examined the possible existence of a specialized narial gland in trematopids and

846

PALAEONTOLOGY, VOLUME 36

argued that a salt gland, possibly one that was homologous to the external nasal gland of reptiles,
was the most probable candidate for these terrestrial amphibians. He proposed that this salt gland
expanded laterally, presumably to enhance its efficiency as a salt-excreting organ. Expansion was
restricted to a lateral direction because medial expansion would force the nasal capsule against the
narial flange. As the salt gland expanded laterally, the external naris elongated posteriorly to
accommodate the larger gland. This narial elongation had profound consequences for the
mechanical bracing of the antorbital region. According to Bolt, forces generated during feeding
would be concentrated in the antorbital region and the reduction of this region to a bar by the
expansion of the external naris would weaken the skull. The narial flange was recruited to act as a
strut to brace the weakened antorbital region. Bolt pointed out that the dissorophid narial flange
probably also had a mechanical role in bracing sutural contacts between rostral bones and that his
hypothesis for the trematopid narial flange was based upon a change in the pattern of cranial
stresses.

There are several difficulties with the above scenario. Only meagre osseous evidence in the form
of impressions exists on the inner surfaces of dermal bones for a reconstruction of the cartilaginous
nasal capsule in trematopids. No details, such as size, shape, or even existence of possible fenestrae
can be determined. Thus, there is no a priori reason to assume that constraints on medial expansion
of an external nasal gland left lateral expansion as the sole option. Presumably, since the choana
occupies the entire space between the septomaxilla anteriorly and the contact of the narial flange
with the antorbital bar posteriorly, a lateral salt gland would have been above the choana. Any
expansion of this gland could occur in a number of different directions, including (but not
exclusively) laterally. To argue for a restriction of expansion to one specific direction would require
detailed knowledge of the soft anatomy of the nasal region that is not available.

Even if the existence of a laterally expanded external nasal gland is accepted, no clear causal
connection between its expansion and the elongation of the external naris can be demonstrated. All
reptilian salt glands share a common histology and embryology despite the independent origins of
these glands among different groups of reptiles (Parsons 1970; Dunson 1976). Reptilian salt glands
are branched or compound tubular glands and grow by repeated branching to reach their adult
shape late in embryology. All subsequent growth of the gland keeps pace with that of the skull. The
naso-labial gland of plethodontid salamanders is also a branched tubular gland that grows by
occupying any available cavity or foramen without affecting cartilaginous or osseous structures
(Whipple 1906). If the hypothesized salt gland of trematopids was also a tubular gland, then it
presumably grew in a fashion similar to the reptilian salt gland and the plethodontid naso-labial
gland. It is, therefore, unnecessary to hypothesize loss of bone as a consequence of salt gland
expansion. Nor is there strong evidence for a positive correlation between the presence of an
enlarged external nasal gland and an elongated external naris in an extant tetrapod. For example,
dermal bones around the external naris of Varanus are reduced so that the opening is greatly
enlarged and the nasal capsule is visible (Bellairs 1949). Skin and connective tissue cover the exposed
portion of the nasal capsule. The external nasal gland of Varanus is not enlarged and is located
within the conchal space behind the external naris.

Terrestriality and nasal anatomy

In a comparative study of the amphibian nasal region, Jurgens (1971) concluded that, contrary to
previous assertions (e.g. Jarvik 1942), there are numerous similarities between the nasal capsules of
the three recognized living orders. Jurgens identified several intermediate stages in some anurans
and urodeles which he believed demonstrated an especially close phylogenetic link between these
groups. Differences between representatives of each group could be explained by the formation and
elaboration of accessory nasal sacs. The degree of elaboration of accessory nasal sacs was correlated
positively with increased terrestriality, and Jurgens hypothesized that the adoption of a terrestrial
lifestyle was the selective agent responsible for the evolution of accessory nasal sacs. The simplest
accessory nasal sac of amphibians is the lateral diverticulum (lateral nasal sinus) of primitive
(aquatic) urodeles which forms a simple ventrolateral extension of the cavum nasi proprium and

DILKES: TREMATOPID NASAL REGION

847

c.nas.pr.
sm

org.J.

-mpal

text-fig. 7. Transverse sections of nasal chambers in representative extant ectothermic vertebrates. A, urodele
Salamandra maculata (redrawn from Seydel 1895). b, urodele Ambystoma maculatum (redrawn from Jurgens
1971). c, anuran Rana esculenta (redrawn from Seydel 1895). D, anuran Alytes obstetricans (redrawn from
Jurgens 1971). e-f, gymnophionan Ichthyophis glutinosus (drawn from photographs in Badenhorst 1978). G,
sphenodontid Sphenodon punctata (redrawn from Hoppe 1934). Black areas are bone and stippled areas are

cartilage. Not drawn to scale.

remains entirely continuous with the cavum (Text-fig. 7a). In contrast, the lateral diverticulum and
organ of Jacobson of terrestrial salamanders is subdivided from the cavum to varying degrees by
vertical or inclined cartilages (Text-fig. 7b). Separation of the lateral diverticulum from the cavum
and the isolation of the organ of Jacobson in a subchamber of the lateral diverticulum is developed
further in terrestrial anurans (Text-fig. 7c-d).

Trematopid amphibians are generally considered to be terrestrial carnivores (Olson 1970; Bolt
1974a; Dilkes 1990) largely on the basis of the presence of caniniform teeth, extensive cranial and
postcranial ossification, slender and relatively long limb elements, large otic notches, and the
absence of lateral line grooves or pits. Dissorophids share many of these general adaptations and
they form the other major group of terrestrial temnospondyls. If it is assumed that terrestriality was
an important selective agent in the evolution of dissorophids and trematopids and if one also
accepts the hypothesis of Jurgens (1971) that terrestriality among lissamphibians was the key
selective agent responsible for the evolution of accessory nasal sacs, then one may speculate that
dissorophids and trematopids had accessory nasal sacs. As the lateral diverticulum is the one
accessory nasal sac from which the others have apparently evolved, the search may be limited to
finding evidence for this nasal sac.

The narial flange would appear to be the only evidence for the presence of a lateral diverticulum
in Palaeozoic dissorophoids since it is the only ossified structure that projects into the nasal
chamber. However, the arguments by Bolt (1974a) on the relative contributions of dermal bone and
cartilage to internal nasal projections would seem to suggest otherwise. To determine whether or not
the dissorophoid narial flange does constitute evidence for a partly isolated lateral diverticulum, two
issues must first be addressed. These issues are: (1) the likelihood of an ossified as opposed to
cartilaginous internal nasal projection in temnospondyls; and (2) whether or not the dissorophoid
narial flange was actually inside the nasal capsule.

First, it is true that the cartilaginous chondrocranium of anurans and urodeles is more extensive
than the surrounding dermal bones and provides the principal support for the olfactory organ
(Text-fig. 7a-d). The chondrocranium in these amphibians would then be expected to produce any

848

PALAEONTOLOGY, VOLUME 36

internal projections, and, as observed by Jurgens (1971), cartilage is indeed the source for the varied
internal nasal flanges that appear in terrestrial amphibians. However, any assumption that extinct
terrestrial temnospondyls would necessarily follow the same pattern is questionable. Paedo-
morphosis was probably an important factor in the phylogeny of extant amphibians (Bolt 1977a,
1979; Milner 1988), so the rarity of dermal flanges in the nasal chamber of anurans and urodeles
may be a consequence of their inheritance of a larval level of ossification. Adult non-lissamphibian
dissorophoids with their highly ossified skulls may have had a different relationship between the
dermal skull and chondrocranium since they would represent a line of development divergent from
the one taken by extant amphibians. Support for this hypothesis is provided by the extant limbless,
burrowing amphibians known as caecilians. The extent of dermal cranial ossification is enhanced
secondarily relative to anurans and urodeles to produce the characteristic solidly ossified skull roof
(Wake and Hanken 1982). Consequently, the chondrocranium is largely resorbed during
development, and the olfactory organ is supported primarily by dermal bones of the roof and palate
(Text-fig 7e-f), in particular the maxillopalatine, septomaxilla, and vomer (Duellman and Trueb
1986). The prominent eminentia olfactoria on the floor of the olfactory sac is supported in anurans
and urodeles primarily by the floor of the nasal capsule (solum nasi) with a small contribution from
the vomer (Jurgens 1971). In caecilians, on the other hand, primary support for the eminentia
olfactoria is the vomer (Text-fig. 7e) with a lateral contribution from the maxillopalatine (Text-fig.
7f). Caution is advised though when applying these observations to temnospondyls because internal
impressions on the fossilized bone suggest a uniformly broad and plate-like ethmoid in addition to
the robustly ossified skull (Rocek 1990). Nonetheless, the projection of a dermal flange into the
nasal capsule of dissorophoids is consistent with the pattern of cranial development in extant
amphibians.

Second, contrary to the suggestion of Bolt (1974a) the dorsal ridge on the palatine does represent
the approximate position of the posterior wall of the nasal capsule. The hind wall of anuran and
urodelan nasal capsules is termed the postnatal wall and is connected to the palatoquadrate laterally
by the processus maxillaris posterior and medially by the commissura quadrato-cranialis anterior
(Jurgens 1971). In temnospondyls, the postnasal wall was separated from the palatoquadrate by the
palatine ridge (Rocek 1990). A postchoanal depression immediately anterior to this palatine ridge
in species of Branchiosaurus held the lateral portion of the postnasal wall (Watson 1940; Boy 1978;
Rocek 1990). The course of the palatoquadrate in branchiosaurids can be traced along a distinct
groove on the dorsal surface of the pterygoid to a groove on the palatine along the posterior side
of the palatine ridge. This groove on the palatine probably accommodated the processus maxillaris
posterior and the commissura quadrato-cranialis anterior. A similar pair of grooves on the
pterygoid and palatine of trematopids probably carried the same cartilaginous structures.
Trematopids lack a postchoanal depression on the palatine, but the anterior side of the palatine
ridge is concave and. as in branchiosaurids, probably held the postnasal wall. Since trematopids and
dissorophids have prominent palatine ridges and the narial flange of each family is located entirely
anterior to this ridge, the dissorophoid narial flange was clearly within the confines of the nasal
capsule. Thus, the narial flange subdivides partly the nasal chamber, and the slightly smaller region
above the choana was probably a lateral diverticulum.

Although the presence of a narial flange may be linked to terrestriality, differences in the shapes
of the trematopid and dissorophid flanges suggest that some additional agent, perhaps dietary,
directed the details of their construction. Bolt (1974c) argued that the magnitude of the stresses
imposed upon the skull of an early tetrapod during feeding are proportional to the size and position
of the palatal tusks and marginal teeth that engage the prey. According to Bolt, the orbits would
concentrate stresses within the antorbital and postorbital regions. Thus, the marked contrast in
dental morphology between trematopids and dissorophids may have been an important factor in
the evolution of the distinctive trematopid narial flange. Teeth of small dissorophids tend to be
small, uniform in size, and peg-like with the crowns varying from monocuspid to bicuspid (Bolt
1977a). Larger dissorophids have large, monocuspid teeth that remain relatively uniform in length.
Trematopids, however, are characterized by their caniniform teeth on the premaxillae and maxillae

DILKES: TREMATOPID NASAL REGION

849

(Dilkes 1990). Maxillary caniniform teeth are located anterior to the orbits, just in front of the
palatine tusks. As the caniniform teeth and palatal tusks would first contact the food item, they
would concentrate cranial stresses in the antorbital region. If the above assumption of a correlation
between stress concentration and the size and location of the teeth involved in prey capture is valid,
then the stresses imposed on the antorbital region may have been far greater in trematopids than
dissorophids. The establishment of a contact between the narial flange and narrow antorbital bar
would perhaps transfer these stresses away from the bar, as suggested by Bolt (1974a). The
functional replacement of that portion of the antorbital region behind the external naris by the
narial flange possibly led to the subsequent loss of this region by the gradual posterior expansion
of the narial opening.

Support for the phylogenetic origin of the trematopid narial flange and elongated external naris
by development of caniniform teeth can be found in the ontogenetic sequence for the rostrum
outlined in this paper. It is assumed that the phylogenetic and ontogenetic stages would be
comparable because in each instance there is a shift in habitat preference from aquatic to terrestrial.
Both the early ontogenetic appearance of a narial flange prior to elongation of the external naris
and the coincident development of caniniform teeth and the elongation of the external naris fit
major stages in the proposed scheme of trematopid phylogeny, but not the hypothesis of Bolt
(1974a) in which elongation of the external naris and development of the trematopid narial flange
are concurrent. Not all stages are, however, recapitulated. For example, the hypothesized
differentiation of a dissorophid-like narial flange to a trematopid narial flange does not occur during
the observed stages of ontogeny even though the proposed causal agent of caniniform teeth exhibits
a clear progressive development. However, it is possible that the relative ontogenetic timing of the
narial flange might have been modified in dissorophoid phylogeny. Presence of a narial flange in a
presumably aquatic larval dissorophoid (Dilkes 1991) is perplexing if a narial flange is an
adaptation for terrestriality. One possible explanation is a heterochronic (peramorphic) shift in the
formation of the narial flange to an earlier point in ontogeny. A peramorphic change could be
achieved either by a later termination or earlier beginning of growth or an acceleration of growth
(McKinney et al. 1990).

POSSIBLE PHYSIOLOGY OF THE TREMATOPID NARIS

Nasal systems of terrestrial vertebrates, in addition to their role of nasopharyngeal breathing, must
also detect often low concentrations of air-borne molecules that diffuse rapidly, and act as a site for
the conditioning of inspired air through filtering, warming, and humidification (Romer and Parsons
1986). Exchange of heat and water between inhaled and exhaled air and the nasal walls is found in
all terrestrial vertebrates (Schmidt-Nelson 1983). Significant quantities of heat and water that would
otherwise be lost to the environment through exhaled air can be recovered in the nasal sacs of
endotherms (Schmidt-Nelson et al. 1970) and ectotherms (Murrish and Schmidt-Nelson 1970).

The efficiency of the counter-current exchange mechanism in the nasal chamber of terrestrial
vertebrates is enhanced if the surface area for exchange is increased, the distance from the centre
of the air flow to the nasal wall is decreased, and the velocity of the passage of air is diminished.
Conchae, defined as any curved projection of the lateral wall of the nasal chamber (Romer and
Parsons 1986; Test-fig. 7e), are responsible for the facilitation of heat and water conservation in the
nasal chamber. All amniotes, with the exception of turtles, possess conchae in their nasal chambers
(Parsons 1967). Amphibians lack conchae (Jurgens 1971). Experimental data show that the
efficiency of the reclamation of heat and water is correlated with the sophistication of the conchae
(Murrish and Schmidt-Nielsen 1970; Schmidt-Nielsen et al. 1970).

Removal of heat from exhaled air results in important metabolic savings for endotherms
(Schmidt-Nielsen et al. 1970); however, since ectotherms rely upon external heat sources any
savings in their energy budget by this mechanism would probably be secondary to water
conservation (Murrish and Schmidt-Nielsen 1970). It is assumed that early tetrapods were
ectothermic, as the presence of growth rings in their bone would suggest (Enlow and Brown 1956).

850

PALAEONTOLOGY, VOLUME 36

Therefore, water rather than heat conservation could have been an important selective agent in the
evolution of the nasal region in Palaeozoic amphibians and reptiles.

Little detailed information can be gained directly from fossils on the complexity of a cartilaginous
nasal capsule, though in the case of temnospondyls it is possible to reconstruct the general aspects
of their ethmoidal cartilages (Rocek 1990). Despite these inherent difficulties, it is clear that the
presence of a bony flange which projected into the nasal capsule of dissorophoids would have
increased the surface area of the nasal epithelium and decreased the speed of air flow through the
nasal sac and, thus, aided the recovery of water. The rate of air flow is also an important variable
in determining the level of olfactory responses (Mozell et al. 1984). With the exception of mammals,
most of the air that passes through the nasal sac of amniotes and anamniotes crosses the olfactory
epithelium. An increase in the rate of air flow has a negative effect upon olfactory response because
the probability of an odorant molecule being absorbed by the mucus layer is decreased as the
velocity at which the molecules strike the mucus is increased. Thus, any obstructions to the air flow,
such as a narial flange, would enhance olfactory responses by increasing the proportion of odorant
molecules that penetrate the mucus and are detected by the receptor cells. Conservation of water
and the enhancement of olfaction may have been complementary factors in the evolution of the
dissorophoid narial flange which would then have fulfilled an analogous role to the concha of
reptiles.

No reference to a specific environment is implied in this hypothesis of physiological analogy
between reptilian conchae and the dissorophoid narial flange. Dissorophoids are a common
component of sites in the southwest of the United States that are interpreted as mixed load river
systems (Berman et al. 1987) and ponds (Sander 1989) whereas more northerly sites in North
America have yielded only a few dissorophoids from settings such as infilled abandoned river
channels (Hook and Ferm 1988) and deltas (Baird et al. 1985). The presence of a narial flange within
the nasal chamber of dissorophoids is interpreted as evidence for a more terrestrial lifestyle although
they apparently remained associated with water, in particular rivers and ponds.

None of the above arguments discounts the proposal by Bolt (1974a) of a salt gland in
trematopids derived from an external nasal gland. Amphibians and reptiles have homologous
external nasal glands (Parsons 1959), and the common presence of this gland in extant tetrapods
suggests that it is probably plesiomorphic for the group. Salt glands are an important extrarenal
route for the elimination of excess ions, especially potassium, taken in through the diet or
accumulated as a result of desiccation (Templeton 1964; Dunson 1976; Minnich 1982). However,
salt glands are absent or greatly reduced in size in some terrestrial reptiles that accumulate large
quantities of ions through their diet. Furthermore, terrestrial reptiles typically excrete nitrogenous
wastes in the form of urate which has the advantages of binding large quantities of the same ions
removed by salt glands and allowing significant amounts of water to be resorbed through the walls
of the bladder and cloaca (Minnich 1982). Clearly, there is no evidence in the fossils to suggest that
dissorophoids or any other extinct terrestrial vertebrate used urate excretion. Nonetheless, given the
reliance of any physiological hypothesis for an extinct vertebrate upon extant representatives, the
importance of urate excretion in terrestrial reptiles and at least two genera of arboreal anurans
(Shoemaker 1988) and reduction or absence of salt glands in some desert-dwelling lizards weakens
arguments for a strong positive correlation between terrestriality and the presence of a salt gland.

CONCLUSIONS

Development of the distinctive narial flange of trematopid amphibians occurred early in
postmetamorphic ontogeny. Posterior elongation of the external naris and the differentiation of
caniniform teeth on the premaxillae and maxillae are subsequent growth stages that reach their
fullest expression only in adults. Hence, at present, only the shape of the narial flange is a reliable
character for the identification of juvenile postmetamorphic trematopids. The narial flanges of
trematopids and dissorophids are reinterpreted as projections into the nasal sac that enhanced water
conservation and olfactory sensitivity in a manner analogous to the conchae of amniotes. Although

DILKES: TREMATOPID NASAL REGION

851

amphibians lack conchae, they show evidence of a positive correlation between terrestriality and the
presence of cartilaginous projections into the nasal sac. As the moist skin of extant amphibians is
the primary site for the loss and resorption of water (Duellman and Trueb 1986), the cartilaginous
projections probably improve only olfaction. Conchae, narial flanges, and cartilaginous projections
are independent evolutionary attempts to improve olfaction and, in the case of the first pair of
structures, conserve water among terrestrial tetrapods.

Acknowledgements. Sections of this paper were part of a thesis submitted as partial fulfilment of the
requirements for a degree of Master of Science in the Department of Zoology, University of Toronto, under
the supervision of Dr Robert Reisz. A postgraduate scholarship from the Natural Sciences and Engineering
Research Council of Canada supported this research. Thanks are extended to Dr Eugene Gaffney (American
Museum of Natural History, New York) and Mr Charles Schaff (Museum of Comparative Zoology, Harvard
University) for arranging loans of trematopid specimens. Drs Robert Reisz and Robert Holmes, Mr Michael
DeBragga, and Mr Michel Laurin read drafts of this paper, and their criticisms and suggestions are
appreciated. Comments from two anonymous reviewers and Dr Andrew Milner helped to clarify sections of
this paper.

REFERENCES

badenhorst, a. 1978. The development and phylogeny of the organ of Jacobson and the tentacular apparatus
of Ichthyophis glutinosus (Linne). Annale van die Universiteit van Stellenbosch. Serie A2 , 1, 1-26.
baird, G. c., shabica, c. w., anderson, J. L. and Richardson jr, e. s. 1985. Biota of a Pennsylvanian muddy
coast: habitats within the Mazonian delta complex, northeast Illinois. Journal of Paleontology , 59, 253-281.
bellairs, a. d’a. 1949. Observations on the snout of Varanus , and a comparison with that of other lizards and
snakes. Journal of Anatomy , 83, 116-146.

berman, d. s., reisz, r. r. and eberth, d. A. 1985. Ecolsonia cutlerensis, an Early Permian dissorophid
amphibian from the Cutler Formation of north-central New Mexico. New Mexico Bureau of Mines and
Mineral Resources, Circular , 191, 5-31.

— 1987. A new genus and species of trematopid amphibian from the Late Pennsylvanian of
north-central New Mexico. Journal of Vertebrate Paleontology, 7, 252-269.
bolt, J. R. 1974a. Osteology, function, and evolution of the trematopsid (Amphibia: Labyrinthodontia) nasal
region. Fieldiana: Geology , 33, 11-30.

— 19746. Armor of dissorophids (Amphibia: Labyrinthodontia): an examination of its taxonomic use and
report of a new occurrence. Journal of Paleontology, 48, 135-142.

— 1974c. Evolution and functional interpretation of some suture patterns in Paleozoic labyrinthodont
amphibians and other lower tetrapods. Journal of Paleontology , 48, 434-458.

— 1977a. Dissorophoid relationships and ontogeny, and the origin of the Lissamphibia. Journal of
Paleontology, 51, 235-249.

— 19776. Cacops (Amphibia: Labyrinthodontia) from the Fort Sill Locality, Lower Permian of Oklahoma.
Fieldiana: Geology, 37, 61-73.

— 1979. Amphibamus grandiceps as a juvenile dissorophid: evidence and implications. 529-563. In nitecici,
m. H. (ed.). Mazon Creek fossils. Academic Press, New York, 581 pp.

— and lombard, r. e. 1985. Evolution of the amphibian tympanic ear and the origin of frogs. Biological
Journal of the Finnean Society, 24, 83-99.

boy, j. a. 1972. Die Branchiosaurier (Amphibia) des saarpfalzischen Rotliegenden (Perm; SW-Deutschland).
Abhandlungen des Hessischen Landesamt fur Bodenforschung, 65, 5-137.

— 1974. Die Larven der rhachitomen Amphibien (Amphibia: Temnospondyli; Karbon-Trias). Palaonto-
logische Zeitschrift, 48, 235-268.

— 1978. Die Tetrapodenfauna (Amphibia, Reptilia) des saarpfalzischen Rotliegenden (Unter-Perm; SW-
Deutschland). 1. Branchiosaurus. Mainzer Geowissenschaftliche Mitteilungen, 7, 21-16.

broom, r. 1913. On the South African pseudosuchian Euparkeria and allied genera. Proceedings of the
Zoological Society of London, 1913, 619-633.

Carroll, R. L. 1964. Early evolution of the dissorophid amphibians. Bulletin of the Museum of Comparative
Zoology, Harvard University, 131, 161-250.
chase, J. n. 1963. The labyrinthodont dentition. Breviora, 187, 1-13.

852

PALAEONTOLOGY, VOLUME 36

demar, r. 1968. The Permian labyrinthodont amphibian Dissorophus multicinctus, and adaptations and
phylogeny of the family Dissorophidae. Journal of Paleontology , 42, 1210-1242.
dilkes, d. w. 1986. A revision of the family Trematopsidae (Temnospondyh : Dissorophoidea) with the
description of a new genus from the Lower Permian of Texas. Unpublished M.Sc. thesis, University of
Toronto.

— 1990. A new trematopsid amphibian (Temnospondyh : Dissorophoidea) from the Lower Permian of
Texas. Journal of Vertebrate Paleontology, 10, 222-243.

— 1991. Reinterpretation of a larval dissorophoid amphibian from the Lower Permian of Texas. Canadian
Journal of Earth Sciences, 28, 1488-1492.

— and reisz, r. r. 1987. Trematops milleri Williston, 1909 identified as a junior synonym of Acheloma
cumminsi Cope, 1882, with a revision of the genus. American Museum Novitates, 2902, 1-12.

duellman, w. e. and trueb, L. 1986. Biology of amphibians. McGraw-Hill, New York, xvii + 670 pp.
dunson, w. a. 1976. Salt glands in reptiles. 413-445. In gans, c. and dawson, w. r. (eds). Biology of the
Reptilia. Physiology A, 5. Academic Press, New York, xv + 556 pp.
eaton, T. h. 1973. A Pennsylvanian dissorophid amphibian from Kansas. Occasional Papers of the Museum of
Natural History, University of Kansas, 14, 1—8.

enlow, d. h. and brown, s. o. 1956. A comparative histological study of fossil and recent bone tissues. Part
I. Texas Journal of Science, 8, 405-443.

hentz, T. f. 1988. Lithostratigraphy and paleoenvironments of Upper Paleozoic continental red beds, north-
central Texas: Bowie (new) and Wichita (revised) groups. The University of Texas at Austin, Bureau of
Economic Geology Report of Investigations, 170, 1-55.

— and brown, L. F. 1987. Wichita Falls-Lawton Sheet. In barnes, V. E. (project director). The Geological
Atlas of Texas. The University of Texas at Austin, Bureau of Economic Geology.

hook, r. w. and ferm, j. c. 1988. Paleoenvironmental controls on vertebrate-bearing abandoned channels in
the Upper Carboniferous. Palaeogeography, Palaeoclimatology, Palaeoecology, 63, 159-181.
hoppe, G. 1934. Das Geruchsorgan von Hatteria punctata. Zeitschrift fur Anatomie und Entwicklungsgeschichte,
102, 434-461.

jarvik, e. 1942. On the structure of the snout of crossopterygians and lower gnathostomes in general.
Zoologiska Bidrag fran Uppsala, 21, 235-675.

jurgens, J. D. 1971. The morphology of the nasal region of Amphibia and its bearing on the phylogeny of the
group. Annale van die Universiteit van Stellenbosch, 46, 1-146.

McKinney, m. l., McNamara, k. j. and zachos, l. g. 1990. Heterochronic hierarchies: application and theory
in evolution. Historical Biology, 3, 269-287.

martens, T. 1990. First evidence of terrestrial tetrapods with North- American faunal elements in the red beds
of Upper Rotliegendes (Lower Permian, Tambach Beds) of the Thuringian Forest (G.D.R.) - First results.
Acta Musei Reginaehradecensis, Series A, 22, 99-104.

milner, A. R. 1982. Small temnospondyl amphibians from the Middle Pennsylvanian of Illinois. Palaeontology,
25, 635-664.

— 1986. Dissorophoid amphibians from the Upper Carboniferous of Nyrany. 671-674. In rocek, z. (ed.).
Studies in herpetology. Proceedings of the herpetological meeting ( 3rd ordinary general meeting of the societas
Europea Herpetologica) . Charles University Press, Prague, 754 pp.

— 1988. The relationships and origin of living amphibians. 59-102. In benton, m. j. (ed.). The phylogeny and
classification of the tetrapods , Volume 1 : amphibians, reptiles, birds. Systematics Association Special Volume,
35A. Clarendon Press, Oxford, x + 377 pp.

minnich, j. E. 1982. The use of water. 325-395. In gans, c. and pough, f. h. (eds). Biology of the Reptilia.

Physiology C, 12, Academic Press. New York, xv + 536 pp.
mozell, M. m., sheehe, p. r., swieck, jr, s. T., kurtz, D. b. and hornung, d. E. 1984. A parametric study of
the stimulation variables affecting the magnitude of the olfactory nerve response. Journal of General
Physiology , 83, 233-267 .

murrish, d. E. and schmidt-nielsen, k. 1970. Exhaled air temperature and water conservation in lizards.
Respiration Physiology, 10, 151-158.

olson, e. c. 1941. The family Trematopsidae. Journal of Geology, 49, 149-176.

1970. Trematops stonei sp. nov. (Temnospondyh: Amphibia) from the Washington Formation, Dunkard
Group, Ohio. Kirtlandia, 8, 1-12.

- 1985. A larval specimen of a trematopsid (Amphibia: Temnospondyh). Journal of Paleontology, 59,
1173-1180.

panchen, a. l. 1967. The nostrils of choanate fishes and early tetrapods. Biological Reviews, 42, 374-420.

DILKES: TREMATOPID NASAL REGION

853

parsons, t. s. 1959. Studies on the comparative embryology of the reptilian nose. Bulletin of the Museum of
Comparative Zoology, Harvard University , 120, 101-277.

— 1967. Evolution of the nasal structure in the lower tetrapods. American Zoologist , 7, 397 — 4 1 3.

1970. The nose and Jacobson’s organ. 99-191. In gans, c. and parsons, t. s. (eds). Biology of the Reptilia.
Morphology B , 2. Academic Press, New York, xiii + 374 pp.
rocek, z. 1990. Ethmoidal endocranial structures in primitive tetrapods: their bearing on the search for anuran
ancestry. Zoological Journal of the Linnean Society, 99, 389 — 407.
romer, a. s. 1939. Notes on branchiosaurs. American Journal of Science, 237, 748-761.

— 1947. Review of the Labyrinthodontia. Bulletin of the Museum of Comparative Zoology, Harvard College,
99, 1-368.

— and parsons, t. s. 1986. The vertebrate body. Saunders College Publishing, New York, vii + 679 pp.
Sander, p. m. 1989. Early Permian depositional environments and pond bonebeds in central Archer County,

Texas. Palaeogeography, Palaeoclimatology, Palaeoecology, 69, 1—21.
sawin, h. j. 1941. The cranial anatomy of Eryops megacephalus. Bulletin of the Museum of Comparative
Zoology, Harvard College, 88, 406-463.

schmidt-nielsen, k. 1983. Animal physiology: adaptation and environment. Cambridge University Press,
Cambridge, xiii + 619 pp.

— hainsworth, f. r. and murrish, d. e. 1970. Counter-current heat exchange in the respiratory passages:
effect on water and heat balance. Respiration Physiology, 9, 263-276.

seydel, o. 1895. Uber die Nasenhohle und das Jacobson’sche Organ der Amphibien. Morphologisches
Jahrbuch, 23, 453-543.

shoemaker, v. H. 1988. Physiological ecology of amphibians in arid environments. Journal of Arid
Environments, 14, 145-153.

templeton, j. r. 1964. Nasal salt excretion in terrestrial lizards. Comparative Biochemistry and Physiology, 11,
223-229.

trueb, L. and Cloutier, r. 1991. A phylogenetic investigation of the inter- and intrarelationships of the
Lissamphibia (Amphibia: Temnospondyli). 223-313. In schultze, h.-p. and trueb, l. (eds). Origins of the
higher groups of tetrapods: controversy and consensus. Comstock Publishing Associates, Ithaca, xii + 724 pp.
vaughn, p. p. 1969. Further evidence of close relationship of the trematopsid and dissorophid labyrinthodont
amphibians with a description of a new genus and species. Bulletin of the Southern California Academy of
Sciences, 68, 121-130.

wake, M. H. and hanken, j. 1982. Development of the skull of Dermophis mexicanus (Amphibia:
Gymnophiona), with comments on skull kinesis and amphibian relationships. Journal of Morphology, 173,
203-223.

warren, a. a. 1980. Parotosuchus from the Early Triassic of Queensland and Western Australia. Alcheringa,
4, 25-36.

watson, d. M. s. 1940. The origin of frogs. Transactions of the Royal Society of Edinburgh, 60, 195-231.
WHIPPLE, i. l. 1906. The naso-labial groove of lungless salamanders. Biological Bulletin, 11, 1-26.
williston, s. w. 1909. New or little-known Permian vertebrates: Trematops, a new genus. Journal of Geology,
17, 636-658.

DAVID W. DILKES

Bernard Price Institute for Palaeontological Research
University of the Witwatersrand

Typescript received 27 July 1992 Private Bag 3, Wits 2050

Revised typescript received 16 November 1992 Johannesburg, South Africa

THE LOWER CAMBRIAN TRILOBITE BIGOTINA
AND ALLIED GENERA

by G. L. PILLOLA

Abstract. The Lower Cambrian genus Bigotina is reinvestigated on the basis of new material (including larval
stages and pygidia) from the type locality of B. bivallata , from the upper third of the Saint-Jean-de-la-Riviere
Formation (Carteret, Normandy, France). A neotype for B. bivallata is proposed. The new material clarifies
the generic concept of this very early trilobite which exemplifies the primitive morphology of non-olenellid
trilobites. The taxonomic position of Bigotina is discussed in comparison with early redlichnds
(Pararedlichiinae, Wutingaspinae and Abadiellidae) and some ellipsocephalids; at present Bigotinidae are
placed in the superfamily Redlichiacea, and the relationship to ellipsocephalids is discussed. The occurrence
of B. bivallata and related taxa in Western Europe, Morocco and Siberia is of significance for correlation and
palaeogeography during Early to Middle Atdabanian time.

In 1925, Bigot discovered an Early Cambrian trilobite and archaeocyathan fauna, associated with
algae, in the Carteret region (Cotentin Peninsula, Normandy, France). These fossils came from
limestone beds cropping out at low tide, in front of the beaches of Saint-Jean-de-la-Riviere and
Saint-Georges-de-la- Riviere. Some of them were described and figured by Bigot in 1926. The sole
species of trilobite discovered, he assigned to Ptychoparia. The archaeocyathans were placed in
Spirocyathus and Protopharetra.

Cobbold (1935) described additional trilobite specimens from Carteret, which were provided by
Bigot, and erected the new genus and species Bigotina bivallata for their reception. Since then, the
algae, stromatolites, ichnofossils and sedimentology have been studied by Bigot (1929), Dore (1963,
1969), Derville (1933), Dangeard (1954), Dangeard and Dore (1957); archaeocyathan faunas have
been described by Debrenne (1958, 1964).

Flowever, the trilobite Bigotina has not been properly investigated and illustrated : for example,
the drawing by Cobbold (1935) reproduced in the Treatise (Harrington 1959, p. 51, fig. 36c;
Henningsmoen 1959, p. 210, fig. 151/1) is misleading in several respects. More recently, Suvorova
(1960), erected two new species and Repina (1960, 1966, 1981, in Repina and Luchinina 1981) four
others, all from Siberia. The genus was recorded by Sdzuy (1981) from Morocco, and Linan et al.
(1981) from the Cordoba area (southern Spain). A species assigned to Bigotina from Australia (Opik
1975) is now placed in Alanisia (Jell, in Bengtson et al. 1990). This paper was prepared with a view
to providing a description of the genus, which is an important one in the understanding of trilobite
phylogenetics; it is the earliest opisthoparian trilobite in several sections and may prove to be
pivotal in our understanding of early trilobite evolution. Bigotina exhibits much subtle
morphological variation and is of importance in considering Lower Cambrian biogeography.

STRATIGRAPHY AND GEOLOGICAL SETTING

Stratigraphy

Following the recognition of Lower Cambrian deposits in the Carteret region by Bigot (1925; Text-
fig. 1 ), the lithostratigraphy and sedimentology have been commented upon and described by Dore
(1963, 1969).

The base of the succession comprises the Carteret Formation (620 m of sandstones and siltstones)

| Palaeontology, Vol. 36, Part 4, 1993, pp. 855-881, 7 pls.j © The Palaeontological Association

856

PALAEONTOLOGY, VOLUME 36

text-fig. 1. Geographical and schematic geological setting of the study area. 1, Lower Cambrian (Tommotian)
Carteret Formation; 2, Lower Cambrian (Atdabanian) Saint-Jean-de-la-Riviere Formation; 3, Upper
Ordovician; 4, Devonian; 5, Recent deposits; 6, Faults; 7, municipality territorial divisions; 8, study section
(modified after Dore 1969 and Graindor et al. 1976).

PILLOLA: LOWER CAMBRIAN TRILOBITES

857

overlain by calcareous and terrigenous sediments of the Saint-Jean-de-la-Riviere Formation
(170-7200 m); this is overlain by red siltstones and sandstones (about 100 m).

The Carteret Formation is subdivided into : (a) sandstones and siltstones well exposed at the Cape
of Carteret, yielding Allonia ? and trace fossils; (b) silty sandstones with some calcareous nodules
containing Eotheca primitiva nom. nud. (hyolithid), Allonia tripodophora Dore and Reid 1965, and
Chancelloria sp. (coeloscleritophorans); (c) green to violet slates with some pink sandstones exposed
in the Dennemont quarry; and, finally, (d) green silty shales of Barneville, well exposed at the beach
of Carteret-Barneville (Dore 1963, 1969, 1984; Dore et al. 1986).

The Saint-Jean-de-la Riviere Formation is composed of dark blue-grey limestones, mostly oolitic
and bioclastic, and several levels containing planar stromatolites, relatively small stromatolitic
biconstructions with peri-reefal breccias, sometimes associated with rare archaeocyathans especially
in the upper part of the formation. The beds are generally thin (10-100 mm), exceptionally reaching
1 m thick. Some of the limestones contain a proportion of detrital material (large quartz grains and
black mica are visible to the naked eye). The coarse silt and shale intercalations are about two-thirds
of the sequence; trace fossils and rare, poorly preserved trilobites are present in these siliciclastic
beds.

Bigotina bivallata occurs from the basal oolitic limestone to the top of the Saint-Jean-de-la-
Riviere Formation in the section studied here (Text-fig. 2); archaeocyathans (Aldanocyathus
carteretensis, Protopharetra bigot i, Sibiricyathus sp., Coscinocyathus sp.) have been found only in
the upper part of the formation; trace fossils ( Skolithos , Planolites , Taphrelminthopsis , Phycodes
and Astropholichnus ), hyolithids and few algae or calcimicrobe taxa ( Epiphyton , Renalcis , Girvanella ,
Rosnaiella , Botomaella) occur at several levels; rare Aldanella (gastropod) and Indianites
(conchostraca) also occur. All the faunal and sedimentological indications are suggestive of
extremely shallow-water deposition.

Geological setting

The study area lies seaward of the Saint-Jean and Saint-Georges-de-la-Riviere beaches, between
Carteret-Barneville on the North and Portbail on the South. The rocky plateau, directed NW-SE,
is mostly occupied by the Saint-Jean-de-la-Riviere Formation ; the Cambrian sequence is tectonically
in contact with small outcrops of rocks of Upper Ordovician and Devonian age, as a result of the
NW-SE fault. The main structure between the Carteret Cape and Saint-Georges-de-la-Riviere
consists of a gentle syncline, with some less important faults and small changes in dip of the strata
(10°-30°). The axis of the syncline is directed NE-SW. A wide sandy channel separates the SE limb
of the main syncline from an antiform structure. The section is located on the southern limb of the
syncline, starting from the wide channel to the axis of the fold. Four main horizons containing
trilobites have been found in the section.

1 . The lowest one ( 1 b), located about 4 m above the first prominent limestone bed, contains a rich
trilobite accumulation layer and very rare hyolithids; trilobite exoskeleton is generally limonitized
with some tectonic deformation. Two small cranidia have also been found in the first oolitic bed ( la,
at 100 mm from the base of the Saint-Jean-de-la-Riviere Formation).

2. The second important fossiliferous level (2a) occurs at about 90 m from the base; it contains
well-preserved specimens of Bigotina bivallata , including larval stages.

3. The third level is about 128 m from the base of the formation, but has not yielded abundant
material.

4. The most fossiliferous trilobite horizon has been located at about 150 m from the base,
midway between the two main archaeocyathan and algal bioconstructions, which are 140-145 m
and 160-165 m from the base respectively. Trilobites are found in massive limestone, which is
locally oolitic.

The new material is deposited at the Museum National d’Histoire Naturelle, Paris (MNHN) and
at the British Museum (Natural History), London (BMNH).

858

PALAEONTOLOGY, VOLUME 36

text-fig. 2. Lithological column and
position of the most important trilobite
beds, a, siltstones and shales, locally
sandy siltstones with calcareous cement;
b, limestones, mostly oolitic and/or
bioclastic; c, stromatolitic dominated
hthofacies and algal-microbial archaeo-
cyathid build-ups; D, not exposed.

TERMINOLOGY

The nomenclature adopted here is mostly that of the Treatise (Moore 1959; Harrington et at. 1959) and Opik
(1958, 1961, 1967); however, some modifications must be emphasized, especially in the description ot the
cranidial features, because some terms are frequently used for Lower Cambrian trilobites and their definition
is sometimes ambiguous.

PILLOLA: LOWER CAMBRIAN TRILOBITES

859

1. Plectrum (Opik 1967 (= mesial ridge)). The plectrum is frequent in Lower Cambrian redlichiids, some
ellipsocephalids and olenellids. It consists of a more or less wide elevated structure varying to an extremely
narrow ridge running between the frontal lobe of the glabella and the anterior border furrow of the
cranidium or a rearward projection of the border in front of the glabella. In some cases, the plectrum is
confluent with the anterior part of the ocular ridges (= eye ridges, eye lines).

2. Facial lines (Opik 1958, 1961, p. 435). The facial lines are relatively thin ridges located in the distal portion
of the anterior limb, joining the ocular ridge and the anterior border. These ridges are usually divergent
anteriorly or, rarely, subparallel to the axis.

3. Parafrontal band (= segment X, of Hupe 1951). Relatively narrow, raised band encircling or flanking the
anterior part of the glabella and confluent with the anterior edge of eye ridges.

4. Interocular swelling (Opik 1967, p. 58; ‘bosse fixigenale’, Hupe 19536). Usually represented as two more
or less prominent humps, located at inner posterior part of interocular fixed cheeks. In some Lower
Cambrian trilobites two posterior and one anterior humps may be observed on each mterocular cheek. The
distal posterior and anterior humps are sometimes coalescent and separated, from the rest of the fixed cheek,
by shallow depressions reaching the posterior border furrow. A relationship between the interocular
swelling and the traces of the primary pleural segmentation, on the adult fixed cheeks, seems to be possible
because these features are more clearly seen during the larval development, and may be assumed a
primitive character.

5. Eye ridges and palpebral lobes: ocular furrow, epipalpebral furrow, ocular striga. The ocular furrow (sensu
Sdzuy 1978, p. 93), is a faint to well-impressed furrow extending from the anterior end of the palpebral lobe
up to the posterior end; this furrow separates the palpebral lobe in two parts, sometimes of about equal size
(cf. Lemdadella ) or, more frequently, into a narrow distal part and a more developed internal portion. The
internal portion of the palpebral lobe, when it is not interrupted, is in continuity with the posterior
‘segment’ of the ocular ridge; this latter is separated from the interocular cheek by the palpebral furrow.

The ocular furrow and the eye ridge furrow observed in Bigotina, Lemdadella and many other redlichiids,
is homologous with the ‘epipalpebral furrow’ (Cowie and McNamara 1978, p. 21) of early olenellids
(Fallotaspis, Eofallotaspis, Profallotaspis, Choubertella , Olenetlus). The palpebral furrow of these trilobites
extends into S3.

The ‘ocular striga’ (Opik 1961, p. 436) was originally designated as the faint furrow that divides the ocular
ridge into posterior and anterior ridges and extends into the palpebral lobe. In this restricted sense the term
‘ocular striga’ may be considered as synonym of ‘epipalpebral furrow’ of olenellids (cf. Opik 1961, fig. 7).
However, the structure of more advanced trilobites (e.g. some metaredlichiids and protolenids, ptychopariids,
zacanthoids) seems to be fundamentally different (Pillola 1991, p. 79). Thus, the homology between the
different elements constituting the ocular ridges and palpebral lobes of high taxonomic units is not completely
elucidated.

RELATIONS BETWEEN BIGOTINID AND SOME OTHER EARLY CAMBRIAN

REDLICHINA

The features seen on adult Bigotina are recognizable in larval stages of several redlichiid taxa; these
features occur progressively earlier during ontogeny of advanced Wutingaspiinae, Pararedlichiinae,
and Redlichiinae respectively. The same features are also present in Dolerolenidae and
Metadoxididae during larval development. This suggests that the morphology of Bigotina truly
represents a primitive one for the non-olenellid Trilobita.

It is relevant, therefore, to consider the similarities and differences between Bigotina and the other
trilobites from the lowest Cambrian which have been considered as representing primitive
morphology.

The ontogeny of Eoredlichia Chang, 1950, Redlichia (Pteroredlichia) Chang, 1966, Redlichia
(Redlichia) Cossmann, 1902 and Wutingaspis Kobayashi, 1944 allows an understanding of one of
the main trends in evolution of the Redlichiidae (Zhang and Lin, in Zhang et al. 1980). McNamara
(1986a) interpreted these genera as an example of a peramorphic evolutionary trend. Further
comparisons between the larval stages and/or adults of Bigotina , Bigotinops , Chaoaspis Chang,

860

PALAEONTOLOGY, VOLUME 36

1966, Dolerolemts Leanza, 1949, Lemdadel/a Sdzuy, 1978, Metadoxides and Parabadiellci Chang,
1966, have been given by Pillola (1991).

These comparisons show that features resembling the adult Bigotina are short-lived during the
ontogeny of Redlichiinae and generally longer-lasting in Wutingaspiinae (Text-fig. 3).

CRANIDIAL LENGTH (SAG.)

text-fig. 3. Occurrence of Bigotina- like adult cranidial features (indicated by thicker line) during the ontogeny
of some redlichiids. Circles indicate the measurement of only one specimen; meraspis to holaspis transition

occurs usually at 3-4 mm sagittal cranidial length).

A particular example is represented by Metadoxides armatus , in which the features of Bigotina
occur in large meraspides. Adult specimens of this species may have from 18 to 22 thoracic segments
and final length can reach 200 mm. Relatively large cranidia which have some immature features
suggest that the entire developmental stage has been ‘stretched out’. Metadoxides may represent an
example of hypermorphosis ( sensu McNamara 1986a).

The genera included in the Abadiellidae Hupe, 1953, have been claimed to represent an
homogeneous group of trilobites; however, the diagnostic characters given by Hupe (1953a, p. 152)
are shared with many other redlichiids. The occurrence together of a slender glabella with very
shallow, medially and backwardly directed glabellar furrows, strong occipital spine, and advanced
genal spine, may represent synapomorphies between the following genera: Abadie/la Hupe, 1953,
Granolenus Jago, 1980, and Lunolenus Sdzuy, 1961. This could be used to define Abadiellidae
in a restricted sense. Inclusion of Guangyuanaspis Zhang & Chien, 1974, Malongocephahts Zhang
and Lin, 1980, Parabadiella Chang, 1966, Shaanxia Zhang and Lin, 1980, and Sibiriaspis Repina,
1960, would serve to blur the distinction between Abadiellidae and Wutingaspiinae and/or
Pararedlichiinae. The existence of intermediate morphologies suggests that abadiellids sensu stricto
are perhaps to be regarded as a subfamily of Redlichiidae.

A few major innovations occur in the most representative ellipsocephalids: lateral glabellar
furrows and dorsal furrow are well impressed; more or less interrupted proximal end of ocular
ridge, and interrupted palpebro-ocular ridge; L4 well developed, and L5 reduced; plectrum
occasionally present. The exoskeleton of ellipsocephalids is smooth or finely punctate, rarely
granulose; general convexity of the exoskeleton strong.

The posterior end of the palpebral lobe, commonly close to the posterior furrow of the fixed cheek
(except in palaeolenids and Hamatolenus (Myopsolenus) Hupe, 1953), and the short, nearly straight,
posterior branch of the facial suture, the diminutive posterolateral limb and long eye ridges, are

PILLOLA: LOWER CAMBRIAN TRILOBITES

861

characters usually invoked for a high level distinction between Redlichioidea and Ellipsocephaloidea
(Hupe 1953a, 19536; Henningsmoen 1959; Moore 1959; Poulsen 1959; Opik 1975; Geyer 1990).
The ‘distinctive’ characters mentioned above occur in some adults of early redlichiids, and in larval
development of several taxa belonging to Redlichiacea.

According to Stubblefield (1936) and Whittington (1981), the axial characters (including ventral
ones), and especially glabellar features, are of significance in trilobite classification. The axial
characters of the cranidia of bigotinids, Abadiellidae, Pararedlichiinae and Wutingaspiinae, are
generally similar. The glabellar segmentation pattern and the shape of the glabella is relatively
constant; most of the variation can be explained by preservation and/or the stage of growth.

Few changes occur in the central region of the cranidium of more advanced redlichiids: forwardly
expanded glabella with more or less impressed S4 (some metaredlichiids); glabellar furrows sub-
perpendicular to the axis and sometimes anteriorly directed; shallowing and uniformly impressed
glabellar furrows (Oolerolenidae, some Gigantopygidae and some Yinitidae); connection between
ocular ridge and anterior part of the glabella, partly interrupted and with some changes in shape,
in morphology and relative topographic position from the anterior end of frontal glabellar lobe (late
Pararedlichiinae and Wutingaspiinae, Redlichiinae, some abadiellids).

By contrast, modifications of the morphological features are more marked on the peripheral
‘characters’ (Pillola 1991, p. 75); a relative large spectrum of variability is shown by: relative length
(sag.) of preglabellar field and anterior border; presence and shape of plectrum; morphology,
position and size of the palpebral lobe; facial suture pattern and relative proportion between fixed
and free cheeks and shape of the occipital segment. The combination of these characters, mainly the
condition of retraction, stasis or protraction of the glabella, the sutural pattern and position and
size of palpebral lobes, comprise a nearly complete view of described Redlichiidae.

Bigotinids lack important derived characters allowing them to be classified either with
Elhpsocephalacea or with Redlichiacea. Early representatives of both high taxa share similarities
which suggest that bigotinids may represent a morphology close to the common ancestor of the two
superfamilies of opisthoparian trilobites at the beginning of Cambrian times.

The provisional assignment of the Bigotinidae to the Redlichioidea seems to be preferable
because of their early stratigraphical position and because the ontogeny of redlichiids includes a
Bigot in a- like phase.

PALAEOBIOGEOGRAPHY

Bigotinids are the earliest opisthoparian trilobites in Lower Atdabanian sequences of northern
France, Spain, Morocco and Siberia.

The most ancient trilobite recorded in Morocco is Hupetina antiqua , from level T1 of Tiout
(Sdzuy 1978). In the same section Bigotina bivallata occurs in the T2 level, about 10 m above the
Hupetina bed. A slightly younger trilobite fauna includes Bigot inops, Lemdadella and Pararedlichia
Hupe, 1953 (= Eoredlichia Chang, 1950). The first occurrence of Bigotina in the Cordoba area
(southern Spain) is located at the base of Level 9 of the Pedroche Formation and it is replaced
higher in the section by Lemdadella (Pararedlichiinae) with which it may coexist (Linan et al. 1981).
Bigotinops is recorded in northern Spain associated with Pararedlichia. In the Carteret region,
Bigotina bivallata ranges almost entirely through the Saint-Jean-de-la-Riviere Formation, but only
in Level 4 does it coexist with Bigotina sp. 1.

Bigotina is also the oldest opisthoparian trilobite from Siberia, its first occurrence being at the
base of the Fallotaspis zone (Zhuravleva 1984); Bigotina is also present in the Atdabanian
Tolbachan Formation in Yakutia (Jegorova 1983).

The ‘bigotinid bioprovince’ (Pillola 1990, 1991) was located on the overlapping area between the
‘Olenellid Realm’ and the ‘Redlichiid Realm’ (Cowie 1971); this bioprovince was well defined
during Early to Middle Atdabanian time (Text-fig. 4), at which time bigotinids coexisted and
gradually were replaced by more advanced redlichiids and ellipsocephalids.

The South Iberian peninsula, northern France, and Tiout-Lemdad areas in Morocco provide a

862

PALAEONTOLOGY, VOLUME 36

^ Bigotina bivallata
Bigotina spp.
Bigotinops

H Hupetina antiqua
S Serrania verae
P Pruvostina
0 Ouijjania

v Lemdadella
■ Eoredlichia

( =Pararedlichia )
□ Wutingaspis

S AMERICA

text-fig. 4. a, land mass distribution during Atdabanian time, and occurrence of selected trilobite genera;
1, ‘olenellid Realm’; 2, 'redlichiid Realm’; 3, area with bigotinids, but without olenellids; 4, area with
olenellids, redlichiids and bigotinids. After Cowie (1971), Rozanov (1984), McMenamin (1987), Burrett et al.
(1990); modified from Pillola 1990, 1991. B, palaeogeographical distribution of selected bigotinids and early

Redlichiidae in Western Europe and Morocco.

PILLOLA: LOWER CAMBRIAN TRILOBITES

863

good example of synchronous similarities of sedimentary environments and fossil composition at
the beginning of Atdabanian (Text-fig. 4); a large proportion of archaeocyatha, trilobites and algae
are in common in these areas (Dore 1963, 1969; Debrenne and Debrenne 1978; Linan and Sdzuy
1978; Sdzuy 1978, 1981; Linan et al. 1981) the lack of early olenellids in Normandy and in the
Cordoba area seems unrelated to environmental setting, but no persuasive reasons have been
suggested.

Faunal exchange between a circum-Mediterranean area and the Far East (particularly with
southwest China) during Atdabanian and Botomian time was well established, as has been recently
pointed out on the basis of new data by Pillola (1990, 1991); however, there are remaining problems
in understanding paleobiogeographical relationships between peri-Gondwanan terranes.

Early Cambrian palaeomagnetic data indicate that south China and Australia were placed close
to one another, and together with Antarctica constituted a large part of Gondwana, located in an
intertropical to sub-tropical position. The disposition of facies belts and sedimentary environments
(Lin et al. 1985; Burrett et al. 1990) support this. Circum-Mediterranean regions were located in
apparently similar climatic conditions in Early Cambrian times.

The close affinities between Chino-Austral-Antarctic faunas and those of Siberia, Mongolia and
Kazakhstan, based on small skeletal fossils and trilobites (Bengtson et al. 1990), and the occurrence
of strong affinities also with circum-Mediterranean areas allow us to suppose a gradient of faunal
exchange all around the Gondwanan and Siberian regions. But lack of fossils from areas between
North East Africa (Morocco and Mauritania) and South America-South Africa (only partly filled
by Avalonian terranes, and Namibia) is a major limitation to the contribution that palaeontological
evidence can make to the model of a Precambrian to Early Cambrian supercontinent, which was
dismembered in Cambrian time.

SYSTEMATIC PALAEONTOLOGY

Order redlichiida Richter, 1933
Suborder redlichina Harrington, 1959
Superfamily redlichioidea Poulsen, 1927
Family bigotinidae Hupe, 1953

[nom. transl. bigotininae Hupe, 1953, emend. Geyer 1990]

Diagnosis. Redlichiacean trilobites typically retaining many primitive features: concave to flat
preglabellar field relatively narrow (sag.) or absent; very prominent and wide undifferentiated
palpebro-ocular ridge, not interrupted by dorsal furrows, clearly bipartite (especially near the
glabella), and parafrontal band; wide plectrum and anterior preocular facial lines present; short
(tr.) postocular cheek with wide posterior border furrow. Eyes reaching, or very close to the
posterior border furrow, and distant from the glabella. Pygidium small with 4 segments and
terminal piece; pleural field narrow and furrowed.

Discussion. The Bigotinidae Hupe, 1953, has been recently revised by Geyer (1990) to accommodate
Bigotina (the sole genus included by Hupe (1953, p. 212) in his Bigotininae) together with
Bigotinops, Ouijjania and Pruvostina (all Hupe, 1953), Hupetina Sdzuy, 1978, from Morocco,
Serrania Linan, 1978, from Spain, and the Siberian Bigotinella Suvorova, 1960, originally described
as a subgenus of Bigotina.

Hupe (1953a) assigned the Bigotininae to the Protolenidae (superfamily Ellipsocephaloidea);
later, Suvorova (1960, p. 21) placed Bigotina in the Neoredlichiidae Hupe, 1953, while Repina
(1960) placed it in the Aldonaiidae Hupe, 1953; Henningsmoen (1959, p. 212) and Repina (1966,
p. 136, 1969, p. 43), following Hupe, assigned the Bigotininae to the Protolenidae Richter and
Richter, 1948. Opik (1967, p. 151) indicated that Bigotina is close to ellipsocephalids and
protolenids, and later placed the Bigotininae in the Ellipsocephaloidea (Opik 1975, p. 38). The
affinities between Bigotina , Bigotinops , Pruvostina and Ouijjania were discussed by Sdzuy (1959,

864

PALAEONTOLOGY, VOLUME 36

1961), who assigned them to the Bigotininae, in the Dolerolenidae Kobayashi, 1951. A direct
relationship between Bigotinidae, Dolerolenidae and Metadoxididae has been questioned (Rasetti
1972).

Genus bigotina Cobbold, 1935

Type species. Bigotina bivallata Cobbold, 1935, by original designation, from the upper part of the Lower
Cambrian Saint-Jean-de-la-Riviere Formation; Saint Jean-de-la Riviere rock foreshore, Cotentin peninsula,
Normandy, France.

Diagnosis. Relatively wide anterior border, regularly arcuate about mid-line (i.e. neither constricted
nor angular in front of the glabella). Glabella occupying 60 to 75 per cent of the total cranidial
length. Prominent occipital ring with a small posterior spine. Short stout genal spine. Exoskeleton
sculpture granulose.

Geographical distribution and stratigraphical range. Cotentin Peninsula, Normandy, France; Cordoba area.
Ossa Morena, southern Spain; northern Spain; Morocco; Siberia. Early Cambrian, Atdabanian.

Discussion. There have been various opinions about the systematic position of Bigotina. Suvorova
(1960, pp. 37 and 40) erected two subgenera of Bigotina , diagnosed as follows:

Bigotina Cobbold, 1935; type species Bigotina bivallata Cobbold, 1935, palpebral lobes and
ocular ridge narrow, bipartite, short frontal area, preglabellar convexity is well expressed.

Bigot inella Suvorova, 1960; type species Bigotinella malykanica, glabella cylindrical or moderately
wider. Palpebral lobe and ocular ridge wide, bipleurality not well expressed, frontal area very short,
preglabellar field may be absent. Preglabellar convexity weak or absent.

The type material of Bigotinella described by Suvorova (PI. 1, figs 1-4) and her other figured
specimens have been studied; additional material is from the right bank of the Lena River (opposite
Malykan, 300 km southwest of Yakutsk, Siberia), Pestrotsvet Formation, bed 10 of the local
succession (British Geological Survey collections).

The type specimen of B. ( Bigotinella ) malykanica is a cranidium about 3 mm long (sag.) The
glabella is slightly tapered forwards and weakly furrowed; the occipital furrow is subperpendicular
to the axis and has a uniform width. The occipital ring is a nearly perfect half-ring shape. The
segmentation of the ocular ridge is not well expressed, but this condition is common in many other
small specimens, e.g. in Bigotina bivallata (PI. 7, figs 2-3), and in some specimens attributed to

EXPLANATION OF PLATE 1

Figs 1-4. Bigotinids from Siberia; all figured specimens are plaster replicas obtained from latex casts of the type
specimens. 1, ‘ Bigotina ' ( Bigotinella ) malykanica Suvorova, 1960, c. x 6. 2 a-c 'Bigotina (Bigotina)' angulata
Suvorova, 1960, (a) dorsal view, x4; ( b ) lateral oblique view, x4; (c) frontal view, x7; note the finely
punctate sculpture on the whole cranidium and the ‘angular’ shape of the anterior border due, mostly, to
the incomplete preparation of the specimen. 3, ‘ Bigotinops ' privus Suvorova, 1960, x 6. 4, ' Bigotinops'
patrius Suvorova, 1960, x 6. Horizon and locality information is in the text.

Figs 5-9. Bigotinids and pararedlichiids from Morocco and Spain. 5, Pruvostina nicklesi Hupe, 1953; internal
mould of the type G.147 (MNHN R. 50887); Lower Cambrian, Zone III; Mektar bou l’Baroud track, near
Amouslek, Morocco; c. x 2. 6, Bigotinops dangeardi Hupe, 1953; internal mould of the type G.154 (MNHN
R. 50838); Lower Cambrian, Zone 1; Tazemmourt, Morocco; x 5. 7, Gen. and sp. ind., internal mould.
D. Vizcaino collection; Lower Cambrian, Zone 1; Tazemmourt, Morocco, x2. 8, Pararedlichia pulchella
Hupe, 1953 (= Eoredlichia) ; internal mould of the type G.153 (MNHN R. 50899); Lower Cambrian, Zone
1; Tazemmourt, Morocco, x 2-6. 9, Serrania verae Linan, 1978; plaster cast of the specimen LPH/O/21.
Linan collection. Lower Cambrian, Tramo III of the Pedroche Formation; Puente de Hierro, Cordoba,
Spain, x 5.

PLATE 1

PILLOLA, bigotinids, pararedlichiids

866

PALAEONTOLOGY, VOLUME 36

Bigotina angulata. In contrast, the specimens of malykanica figured by Repina (1966, pi. 25, figs 1-2)
have quite well expressed segmentation and a cranidium 5 mm long has also a well furrowed conical
glabella. Other specimens of B. malykanica figured by Suvorova (1960, pi. 2, figs 23-25) are small
(4-5, 4-0 and 2-6 mm respectively). Repina's (1966, pi. 25, figs 1-2) specimens of malykanica , are 2-8
and 5 mm long (sag.). A specimen figured by Sdzuy (1981, fig. 20), attributed to B. cf. angulata , has
faint interocular swellings on the posterior intraocular cheek; the same cranidium shows clearly that
the dorsal furrows turn at the level of the anterior corner of the intraocular cheek and reach the
palpebral furrow at mid length of L3. This feature has never been observed in Bigotina. Specimens
of Bigotinella malykanica at lengths of 5 to 8 mm do not show any morphological difference from
‘ Bigotina ’ angulata Suvorova, 1960. The BGS material (FOR3207-FOR3210), six cranidia with
sagittal lengths of 2-5 to 8-0 mm, confirm that angulata and malykanica are synonyms based on
ontogenetic stages.

text-fig. 5. Bigotina bivallata Cobbold, 1935; reconstruction of the cephalon, x8.

I consider that Bigotinella Suvorova, 1960 is a valid genus, provisionally a member of the
ellipsocephalids, characterized by facial lines, plectrum, with anterior branch of facial sutures very
short and subparallel to the axis, and a sculpture of fine punctae of equal size distributed on all parts
of the cranidium.

Other species from Siberia have been assigned to Bigotina : B. ( Bigotina ) conij erica (considered a
synonym of egregica \ Repina 1966, p. 138), B. ( Bigotina ) egregica Repina, 1960, B. ( Bigotinella )
botomica Repina, 1966 and B. (Bigotinella) rara Repina in Repina and Luchinina, 1981. B. egregica
has a wide cranidium with slender and relatively short glabella, the preglabellar field is about three
times wider than the anterior border; this species seems to be related to some members of the
Aldonaiidae, for example Planaspis gelasinica Repina, 1960. Bigotinelia botomica is closely related
to B. malykanica , has a glabellar segmentation which is faint, except the occipital furrow, and

EXPLANATION OF PLATE 2

Figs 1-6. Bigotina bivallata Cobbold, 1935. Saint-Jean-de-la-Riviere Formation, level 4, Lower Cambrian
(Atdabanian); Saint-Jean-de-la-Riviere beach, Carteret region, Normandy, France. 1, holotype; reproduced
from Bigot 1926, pi. 4, fig. 1, x 6. 2, holotype; reproduced from Cobbold 1935, pi. 17, fig. I,x6. 3 a-c
MNHN B. 48948a, neotype; 3a, incomplete locally exfoliated cranidium; dorsal view, x 6; 36, latex cast of
the external mould, showing some parts not preserved in 3a, x6; 3c, frontal view of the latex cast, x 6.
4, MNHN B. 48968; nearly complete cranidium, x 5. 5, MNHN B. 48970; incomplete cranidium, x6.
6, MNHN B. 48947; incomplete sagittally compressed cranidium, x 5.

PLATE 2

PILLOLA, Bigotina

868

PALAEONTOLOGY, VOLUME 36

relatively narrow palpebral lobes and ocular ridges. Bigotinella rara (compare the specimen figured
by Egorova et al. 1983, pi. 45, fig. 5), like Mimdocephalinci bidjaensis Repina, 1964
(Ellipsocephaloidea), has a subtriangular occipital ring with a strong occipital spine; all these
Siberian taxa are closely similar to Australian and Moroccan ellipsocephalids like Elicicolci Jell,
1990, Ornamentaspis Geyer, 1990, and Kingaspoides Hupe, 1953.

Bigotina tina Opik, 1975 (p. 40, pi. 7, fig. 1) is now placed in Alanisia guillermoi (Richter and
Richter, 1940; see Jell 1990, and discussion in Bengtson et al. 1990).

Bigotina differs from Bigotinops Hupe, 1953, in having relatively shorter postocular fixed cheek
and posterior branch of facial suture; the posterior margin of the occipital ring is not expanded in
a cuspid or strong spine. The differences between the two taxa are small for a generic separation.
However, the lack of any other specimen assigned to the type species ( Bigotinops dangeardi Hupe,
1953, from Zone 1, Anti Atlas, Morocco, a cranidium and probably the adjacent librigena, internal
mould; PI. 1, fig. 6), does not allow more accurate comparison. At present, some other specimens
originating from Morocco (Sdzuy 1978, fig. 2b, horizon Til, should probably be assigned to
Bigotinops (Sdzuy 1981, p. 397, fig. 17; Linan et al. 1981, p. 284). This genus is also recorded from
Spain, but without illustrations (Linan and Sdzuy 1978; Sdzuy 1978).

Bigotinops from Siberia is represented by three species: B. patrius Suvorova, 1960, B. privus
Suvorova, 1960 and B. copiosus Jegorova, 1983. Bigotinops privus (PI. 1, fig. 3) has the glabella
expanded forward, occupying about 80 per cent of total cranidial length, the occipital furrow does
not cross the glabella and is interrupted before the dorsal furrow; SI to S3 weak, not impressed in
the distal portions and shallowing toward the middle; narrow (sag.) anterior border and
preglabellar field. Posterior border furrow narrow and shallow. The cuticle is finely punctate.

Bigotinops patrius is poorly known. Examination of the latex cast of the type specimen provided
by Suvorova (PI. 1, fig. 4), suggests that this species should be removed from Bigotinops. Because
no facial sutures have been discerned, this specimen seems to be more probably related to Fallotaspis
or Pr of allot aspis.

Bigotinops copiosus comes from 15 m above the base of the Tolbachan Formation on the Tolba
River section, Yakutia. The figured specimens (Jegorova 1983, pi. 6, figs 4—5) are closely similar to
Bigotina bivallata; Bigotinops copiosus shows both relatively short anterior branches of the facial
sutures and short (tr.) postocular cheeks; hence it is suggested that this species more properly
belongs in Bigotina.

Hupetina Sdzuy, 1978 (type species: H. antiqua Sdzuy, 1978, from the Tiout section, horizon TI,
Morocco) has a narrow, angular anterior border, widening distally, preglabellar field nearly absent
and librigena with a short advanced spine. The diagnosis given by Sdzuy includes also the parallel-
sided or slightly tapering glabella and the short, anteriorly directed S3. These characters are present
in the type specimen (Sdzuy 1978, pi. 1, fig. 6); however, the other figured specimens (Sdzuy 1978,
pi. 1, fig. 7; Sdzuy 1981, fig. 15) show a glabella that tapers forward, and the S3 furrow

EXPLANATION OF PLATE 3

Figs 1-9. Bigotina bivallata Cobbold, 1935. Saint-Jean-de-la-Riviere Formation, Lower Cambrian
(Atdabanian); Saint-Jean-de-la-Riviere beach, Carteret region, Normandy, France. 1 a-c, MNHN B. 48857;
nearly complete cranidium with well preserved exoskeleton, slightly sagittally compressed (anterior margin
incomplete); level 2a; In, dorsal view, x 7; 1 b, anterior oblique view, showing the clear subdivision of the
ocular ridge, x7; lc, lateral view, x 7. 2, MNHN B. 48864; small partly exfoliated cranidium, showing
terrace lines on the anterior border, facial lines, wide plectrum, ocular furrows and interocular swellings;
level 2 a; x 7. 3 MNHN B.48862; damaged pygidium; level 2a; x 20. 4, MNHN B. 48858; incomplete
cranidium, internal mould; level 2a; x 6. 5, MNHN B. 48861; small librigena; level 2a; x 12. 6, MNHN
B.48867n; small cranidium, internal mould; level 2a; x 10. 7, MNHN B. 48998; small exfoliated and
weathered cranidium; level 4; x 5. 8, MNHN B. 48895; incomplete pygidium; level 4; x 15. 9, MNHN
B.48865; small damaged cranidium; level 2a, x 5-5.

PLATE 3

PILLOLA, Bigotina

870

PALAEONTOLOGY, VOLUME 36

perpendicular to the sagittal line. Finally, an unclassified specimen (Sdzuy 1981, fig. 16), having a
cranidial length of 3 5 mm, is similar to some cranidia of Bigotina figured here (PI. 6, fig. 3).

Serrania Lilian, 1978, is also very close to Bigotina , but the former has a relatively longer glabella
(more than 85 per cent of the cranidial length; cf. Linan 1978 and PI. 1, fig. 9), sub-parallel or slightly
diverging anterior branches of the facial sutures and a different shape of anterior border. The
librigena may be relatively smaller than that of Bigotina.

Ouijjania Hupe, 1953a, is monospecific, and represented by only one cranidium, assigned to
O. meridionalis (Hupe, 1953a, fig. 52); the type specimen has never been figured, and the generic
characters are inevitably uncertain.

Pruvostina Hupe, 1953a (type species P. nicklesi Hupe, 1953a; cf. PI. 1, fig. 5) is similar to
Bigotina ; the type specimen, an internal mould of a slightly distorted, incomplete cranidium, shows
subparallel anterior branches of facial sutures, relatively wide (tr.) anterior border and occipital
ring. The preservation of this specimen (e.g. the partly abraded glabella) does not allow more critical
comparison with Bigotina.

Tolbinella Jegorova, 1983 (Pararedlichiinae) has been compared with Bigotina, from which it
differs in the relative length of frontal area and glabella and the long and more divergent anterior
branches of the facial sutures.

It is probable that more genera have been created among Bigotina- like trilobites than the sparse
material really warrants. The present description should allow a more objective assessment of the
synonymy in future.

Following the discussion of Siberian bigotinids, only the species described as Bigotinops copiosus
Jegorova, 1983, belongs in Bigotina. All other Siberian ‘ Bigotina ' should be placed in an
intermediate taxonomic unit between redlichiids and ellipsocephalids. Thus the only valid described
species are Bigotina bivallata, and Bigotina copiosa (Jegorova, 1983) which may prove to be a
synonym of bivallata. In addition, Bigotina cf. bivallata and Bigotina sp. 1 from Carteret are briefly
discussed below.

Bigotina bivallata Cobbold, 1935
Plates 2—4; Plate 5, figs 1-3, 5-7, 10; Plates 6-7; Text-fig. 5
1926 Ptychoparia, Bigot, p. 136, pi. 4, figs 1-4.

1935 Bigotina bivallata n. gen. n. sp., Cobbold, p. 384, pi. 17, figs 1-10.

1981 Bigotina sp.; Sdzuy, p. 396, fig. 18.

71981 Gen. and sp. ind., Sdzuy, p. 396, fig. 13.

1981 Gen. and sp. ind., Sdzuy, p. 396, fig. 14.

v.1991 Bigotina bivallata Cobbold; Pillola, pi. 31, figs 9-11.

Holotype. Specimen figured by Bigot (1926, pi. 4, fig. 1) and copied by Cobbold (1935, pi. 17, fig. 1) PI. 2, figs
1-2; missing and presumed destroyed during World War II; originally housed in Caen University (France).

EXPLANATION OF PLATE 4

Figs 1-7. Bigotina bivallata Cobbold, 1935. Saint-Jean-de-la-Riviere Formation, Lower Cambrian
(Atdabanian); Saint-Jean-de-la-Riviere beach, Carteret region, Normandy, France; all specimens except fig.
1 from level 1 b. la-6, Caen University collections; slightly damaged cephalon; a, dorsal view, x7;
b, anterior oblique view, x 7. 2, BMNH It. 25629a; cranidium, internal mould, x 6. 3, BMNH It. 25625;
small cranidium, x 10. 4, BMNH It. 25621 ; cranidium, internal mould, x 6. 5, BMNH It. 25622; cranidium,
internal mould, x6. 6 BMNH It. 25628; cranidium, internal mould, x 6. 7, BMNH It. 256996; cranidium,
internal mould, x 6.

Fig. 8, Bigotina cf. bivallata Cobbold, 1935; horizon and locality as for figs 2-7; BMNH It. 25630; cranidium,
internal mould ; with some trace of granulations, on the incompletely removed limonitic layer, x 4.

PLATE 4

PILLOLA, Bigotina

872

PALAEONTOLOGY, VOLUME 36

Neotype (proposed herein). Incomplete cranidium MNHN-IP B. 48948, internal and external mould preserved
in limestone; PI. 2, fig. 3 a-c.

Type locality. Saint-Jean-de-la Riviere rock foreshore, 4T5 km from the Carteret Cape lighthouse (Text-fig. 2);
map sheet 1/50000, Bricquebec-Surtainville (B. R. G. M. 1976, feuille XI-XII-11,); cropping out during low
tide. Dark grey limestones (bioclastic to oolitic, packstone and grainstone, locally weathered, located in the
upper part of the Sain-Jean-de-la-Riviere Formation (about 150 m from the base); Atdabanian, Lower
Cambrian.

Material. Level 1 a: two small cranida, one librigena (BMNH It. 25643); level 1 b, BMNH It. 25620— It. 25629,
It2563 1— H25639 : 22 cranidia and 2 meraspides, one librigena, several fragments including thoracic pleurae.
Level 2a, MNHN-IP B. 48856-B. 48932 : 96 cranidia (3 mm to 10 mm), 12 meraspides, 8 librigenae, 2 pygidia.
BMNH It. 25640— It. 25642: 4 cranidia and 4 meraspides, 1 librigena. Level 2b, MNHN-IP B.48933-B. 48942:
5 small cranidia, 7 meraspides, 1 ?protaspis. Level 2c, MNHN-IP B.48943-B. 48944: 2 cranidia. Level 2d,
MNHN-IP B.48945-B. 48946: 2 cranida. Level 4, MNHN-IP B. 48947-B. 49005 : 48 cranidia and 21 meraspides,
13 librigenae, 2 pygidia. BMNH It.25644 — It.25664 : 16 cranidia, 13 meraspides, 5 librigenae.

In Caen University collections: one nearly complete cephalon articulated with two poorly preserved thoracic
segments, more than 25 cranidia, several librigenae and pleurae, coming, in all likelihood, from level 4. A few
specimens of the Feist collection (Montpellier University, France) 9 incomplete or damaged cranida,
1 librigena, 2 meraspides.

All studied material: more than 230 cranidia and 60 larval stages, 40 librigenae and 4 pygidia. Several
specimens of Bigotina bivallata from Cordoba, Spain, have been studied in Zaragoza University (Linan
collections).

Diagnosis. Bigotina with diverging anterior branches of facial sutures; faint ocular furrow, and
coarsely granulose sculpture.

Description. Cranidium subrectangular with arcuate anterior border. Glabella convex, elevated, well rounded
anteriorly, occupying about 70-75 per cent of the cranidial length (sag.) defined by weak furrows and mostly by a
change of slope between the glabella itself and fixed cheeks. Occipital ring longer (sag.) medially, about 3 times
as wide (tr.) as long, showing a short posterior median spine. Occipital furrow deepest laterally, shallowing
and widening towards the middle, which it crosses transversely. Lateral deep parts of occipital furrow parallel
to SI. Three pairs of evenly spaced glabellar furrows, deeply impressed laterally, shallowing progressively
toward the middle. SI particularly bifurcates at about midway between the axial furrow and the sagittal line.

Frontal area composed of the concave to flat preglabellar field, almost entirely occupied by a more or less
prominent boss, the plectrum, and by anterior border; the last convex and about the same width (sag.) as the
preglabellar field, with terrace lines on the anterior edge. Front margin of the cranidium rather gently and
uniformly arcuate in outline, anterior border furrow deeply impressed on external side of the plectrum and

EXPLANATION OF PLATE 5

Figs 1-3, 5-7, 10. Bigotina bivallata Cobbold, 1935; Saint-Jean-de-la-Riviere Formation, Lower Cambrian
(Atdabanian); Saint-Jean-de-la-Riviere beach, Carteret region, Normandy, France. All specimens come
from level 4, except fig. 6 from level 2a. \a-b, MNHN B. 48996; incomplete weathered cranidium;
la, whitened with ammonium chloride, c. x 7; 16, showing caecal network and ocular ridge structure, x 8.
2, MNHN B.48953; damaged cranidium with well-marked ocular furrow and trace of sculpture on the
internal mould, x 7. 3, MNHN B. 48973; weathered largely exfoliated cranidium, x 6. 5, MNHN B.48955;
small incomplete cranidium; the angular shape of the anterior border is due to overlap of matrix, x 10.
6, MNHN B.48870; incomplete cranidium, internal mould, x 3-5. 7, MNHN B.48999; small incomplete
weathered cranidium with clear interocular swellings, x 6. 10, MNHN B. 48960; librigena, external mould, x 7.

Figs 4, 8-9 Bigotina sp. 1 ; Saint-Jean-de-la-Riviere Formation, Lower Cambrian (Atdabanian); Saint-Jean-de-
la-Riviere beach, Carteret region, Normandy, France. All specimens from level 4. 4, BMNH It.25644; small
cranidium, internal mould, x 10. 8, MNHN B. 48958; incomplete cranidium, x 12. 9, MNHN B. 48971;
nearly complete cranidium, x 8.

PLATE 5

PILLOLA, Bigot ina

874

PALAEONTOLOGY, VOLUME 36

shallow m front of the glabella. Concave to flat downsloping anterolateral limb, widens adaxially slightly, some
specimens showing the preocular facial lines (PI. 2, fig. 3; PI. 3, fig. 1), which are sometimes faint, forming an
angle of 35° with the axial line. Palpebral lobe and ocular ridge continuous, confluent with the anterior part
of the glabella, palpebral furrow begins at same level as S3. Crescentic palpebral lobe about one third of the
glabellar length, opposite middle third of the glabella. Intraocular cheek with outline of a quarter circle,
showing, in some cases, weakly developed interocular swellings posteriorly. Postocular fixed cheek not
extended laterally far beyond the palpebral lobe. Exceptionally broad, deep posterior border furrow. Posterior
border widens laterally, distally curved downwards.

Anterior branch of facial suture straight and diverging forward at 20-25° to sagittal line, turning inwards
across the border. Posterior branch short, sigmoid, well rounded, backwardly directed, cutting the posterior
border perpendicularly.

Free cheek when in place, inclined at an angle of about 30° to horizontal. Lateral border wide and gently
convex, with abaxial terrace ridges and carrying rows of small granules (PI. 3, fig. 5); librigenal field flat to
slightly concave, relatively narrow, rising towards the upturned circum-ocular region. Short, stout genal spine,
turning downwards.

Thorax of unknown number of segments, with a strong relief, each pleural region is slightly wider than the
rachis; axial ring relatively wide (sag.), with a small postero-median spine; half ring well developed. Pleurae
laterally directed, with distal fulcrum and wide (exsag.) pleural furrow; ending with a short backwardly
directed spine. Articulating facet observed in several isolated pleurae (e.g. PI. 7, fig. 7).

Pygidium small; prominent truncato-conical to oval rachis, with four segments and tiny terminal piece; sub-
triangular, furrowed pleural field, not extended beyond the rachis, anterior margin forming an angle of 30° with
the axis.

Sculpture consists of fine to coarse granulations distributed on all parts of the relatively thick exoskeleton ;
the internal mould is usually smooth, but traces of granulation are occasionally observed. Caecal network
sometimes revealed by selective weathering in the frontal area (PI. 4, fig. 1 6; PI. 6, fig. 9).

Discussion. Intraspecific variation in Bigotina bivallata shows every intergradation, as illustrated
here. I originally considered the specimens on PI. 4, figs 2-7 as a subspecies of bivallata. However,
no clear-cut differences can be proved, except the slightly more arcuate anterior margin of the
cranidium, and slightly wider posterior furrow of the fixed cheek. The differences in relative sagittal
length of preglabellar field and anterior border are due to these specimens being internal moulds.
The specimen figured by Sdzuy (1981, fig. 14) is conspecific with Bigotina bivallata , and particularly
resembles these specimens.

Bigotina copiosa (Jegorova, 1983) is known from only two specimens: one Jegorova (1983, pi. 6,
fig. 5) differs from B. bivallata in having a relatively more developed LI, but the other, the type
specimen has only one feature allowing specific separation from B. bivallata : the ‘ornament on the
cranidium consists of indistinct shallow pits’ (Jegorova 1983, p. 60).

Occurrence. Earliest Cambrian (Atdabanian), Saint-Jean-de-la- Riviere Formation; Carteret region
(Normandy, France). Pedroche Formation; Cordoba area, Sierra Morena, Spain. Lemdad and Amouslek
area, Anti Atlas, Morocco. ?Tolbachan Formation; Tolba River, Southwest Siberian platform.

EXPLANATION OF PLATE 6

Figs 1-8. Bigotina bivallata Cobbold, 1935. Saint-Jean-de-la-Riviere Formation, Lower Cambrian
(Atdabanian); Saint-Jean-de-la-Riviere beach, Carteret region, Normandy, France. 1, MNHN B.48925;
incomplete early meraspis, x 50. 2, MNHN B. 4892 la; incomplete early meraspis, x 50. 3, MNHN
B. 489266; incomplete early meraspis, x 50. 4 a-6, MNHN B. 48926; meraspis; a, dorsal view, x40;
6, anterior oblique view, x 50. 5, MNHN B. 48981 ; meraspis, x 25. 6 a-b, BMNH It. 25654; meraspis; a , SEM
photograph showing faint interocular swellings on intraocular cheek, x30; b, with clearly expressed
segmentation of proximal end of the ocular ridge, x 25. 7, BMNH It. 256556; meraspis with parallel-sided
glabella and transglabellar furrows, x 25. 8«-6, MNHN B. 48866; meraspis; a , SEM, x 30; 6, with clearly
expressed segmentation of proximal end of the ocular ridge, well-outlined frontal lobe, and subangular
narrow anterior border, x 20.

PLATE 6

PILLOLA, Bigot ina

876

PALAEONTOLOGY, VOLUME 36

Bigotina cf. bivallata Cobbold, 1935
Plate 4, fig. 8

Material. BMNH It. 25630; cranidium.

Discussion. This unique specimen has coarse granulation also on the concave part of the cranidium
(i.e. on the posterior border furrow). No other differences from bivallata have been discerned, and
it may be a variant.

Occurrence. Earliest Cambrian (Atdabanian), Saint-Jean-de-la-Riviere Formation, level 1 6; Carteret region,
Normandy, France.

Bigotina sp. 1.

Plate 5, figs 4, 8-9

Material. Three cranidia: BMNH It.25644; MNHN B.48958, B.48971.

Discussion. Backwardly directed ocular ridge forms, together with palpebral lobe, a nearly perfect
half circle; palpebral lobe with well-impressed ocular furrow; the slightly more developed and
arched frontal area, and less conical glabella are the only appreciable differences separating these
specimens from bivallata. A conspecific incomplete cranidium from Carteret, was figured by Sdzuy
(1981, fig. 19).

Occurrence. Earliest Cambrian (Atdabanian), Saint-Jean-de-la-Riviere Formation, level 4; Carteret region,
Normandy, France.

ONTOGENY

Because heterochrony has been invoked as an important source of evolutionary novelty in early
Cambrian trilobites (McNamara 19866), the ontogeny of Bigotina is important.

The smallest cranidium of Bigotina bivallata (PI. 6, fig. 1) represents an early meraspid stage. The
sagittal length is about 0-9 mm; it is subcircular in shape, overall slightly convex, and has weak
segmentation. The glabella is relatively narrow (tr.), reaching, with the anterior part of the ocular
ridge, the anterior border furrow; the glabella is defined by wide dorsal furrows shallowing
anteriorly. The occipital ring is semicircular, with wide (sag.) occipital furrow. The glabellar furrows
(S1-S3) are poorly impressed, transglabellar, and perpendicular to the axis. The frontal lobe is
semicircular anteriorly. The preglabellar field is absent, the anterior border furrow weak, and the
anterior border narrow. The palpebro-ocular ridge is broad, particularly near the frontal lobe.

EXPLANATION OF PLATE 7

Figs 1-9. Bigotina bivallata Cobbold, 1935; Saint-Jean-de-la-Riviere Formation, Fower Cambrian
(Atdabanian); Saint-Jean-de-la-Riviere beach, Carteret region, Normandy, France. 1, MNHN B. 48924a;
meraspis, x 30. 2, MNHN B.48923; meraspis, with conical glabella, x22. 3, MNHN B.48989; meraspis,
c. x 20. 4, MNHN B. 489246; meraspis or holaspis, with faint facial line, ocular furrow and fine granulated
sculpture, c. x 20. 5, BMNH It. 25655a; probable early holaspis, with all adult features, x 17. 6, MNHN
B. 48951 ; holaspid cranidium with well-impressed ocular furrow, x 10. 7, BMNH It. 25662; isolated pleurae,
with tiny spines, x 5. 8, MNHN B. 488576; small cranidium, x 7. 9, BMNH It.25660; anterior fragment of
a cranidium, showing differential weathering of the test; note the two cuticular layers, the caecal network,
and probable internal trace of the anterior trunk of the ocular ridge, a, anterior border; 6, boss-plectrum or
ocular ridge, x 7.

PLATE 7

PILLOLA, Bigotina

878

PALAEONTOLOGY, VOLUME 36

Slightly larger specimens ( c . 11 mm sagittal length) show anteriorly convergent dorsal furrows up
to S3, and an expanded frontal lobe (PI. 5, figs 2-3); they also have a very broad, poorly defined
proximal end of the ocular ridge, and short anterior and posterior branches of the facial suture. A
short intergenal spine occurs (PI. 6, figs 3-4). Cranidial segmentation is better expressed in the larger
specimen; the glabella is 80 per cent of the cranidial length, and it is preceded anteriorly by the
band. The preglabellar field is absent and the border furrow and anterior rim are better outlined.
The occipital and glabellar furrows are broad, and perpendicular to the axis. The occipital ring
expands posteriorly, and has a posteromedian spine.

Meraspides of T2 mm sagittal cranidial length (PI. 6, fig. 5) have a well-outlined glabella; SI is
arcuate backward. The posterior end of the palpebral lobe is placed transversely opposite SO. The
wide posterior border furrow is partly confluent with the ocular furrow, and the palpebral lobe
shows some continuity across the posterior border furrow; with the posterior border, this
connection represents a shallow ridge close to the posterior branch of the facial suture. This
character persists, with a certain attenuation, in the adult stages.

Some features are well expressed in cranidia of length T3 mm or a little more; they are
subquadrangular, with a conical to subcylindrical glabella which is well defined by dorsal furrows;
the occipital furrow is deep distally; SI is of same shape as SO, but shallower; S2 and S3 are
shallower still and subperpendicular to the axis. The occipital ring is semielliptical, wide medially
(tr.), with a well developed spine (or node). The glabellar frontal lobe is rounded anteriorly; the
preglabellar field is about twice as wide as the anterior border, and the latter becomes subangular
in front of the glabella (PI. 6, figs 6, 8). The posterior border furrow and ocular furrow are deep and
confluent. The ocular ridge is wide, subdivided into three branches in the vicinity of the frontal lobe
(PI. 6, figs 6, 8). The posterior area of the interocular cheek possesses two bosses (interocular
swellings), and has another between the ocular furrow and the dorsal furrow close to L3. These faint
structures are sometimes observed in adult specimens and may represent the traces of primary
furrowed interocular cheeks.

An incomplete cranidium (PI. 7, fig. 1) c. 1-5 mm long, seems to have a relatively short glabella
(75 per cent of cranidial length), due mostly to the small frontal lobe. The latter is, as in previously
described specimens, completely surrounded by the ocular ridge system. An incomplete internal
mould of a cranidium 2-5 mm long (PI. 7, fig. 3) clearly shows the same pattern in the frontal lobe-
ocular ridge connection, and a wide (tr.) plectrum in the preglabellar field. The anterior border is
wide and the anterior furrow deeper.

The external features of specimens sagittally 2-5 to 3-5 mm long (PI. 7, figs 5-6, 8) are quite similar
to those of adult individuals. A faint preocular line connects the anterior border and the ocular
ridge; the palpebral lobe is sometimes subdivided by the ocular furrow; a reduction in size of the
occipital ring and the occipital spine can be observed. Fine to coarse granulose sculpture and terrace
lines are also present.

Acknowledgements. Financial support for research was provided by a EEC Postgraduate Fellowship for which
I am very grateful. I am indebted to R. A. Fortey and C. Mellish for discussions and criticism. Thanks are due
to A. W. A. Rushton (BGS, Nottingham), N. P. Suvorova (Moscow), E. Linan (Zaragoza), J. J. Chauvel and
M. Robardet (Rennes), F. Dore and L. Dupre (Caen) and D. Vizcaino (Maquens), for the friendly
collaboration and fruitful discussions.

REFERENCES

bengtson, s., conway morris, s., cooper, b. J., jell, p. a. and runnegar, b. n. (eds). 1990. Early Cambrian
fossils from South Australia. Memoir of the Association of Australasian Palaeontologists, 9, 364 pp.
bigot, a. 1925. Sur la presence de Trilobites et d’Archaeocyanthides dans les couches cambriennes des environs
de Carteret (Manche). Compte Rendue de PAcademie des Sciences, Paris, 180, 1237.

— 1926. Sur les calcaires cambriens de la region de Carteret et leur faune. Note preliminaire. Bulletin de la
Societe Linneenne de Normandie, 7 (for 1925), 130-144.

PILLOLA: LOWER CAMBRIAN TRILOBITES

879

— 1929. Les recifs en coupole du Cambrien de Carteret et les recifs de Chlorellopsis. Bulletin de la Societe
Linneenne de Normandie , 8, 68-71.

burrett, c., long, J. and st ait, b. 1990. Early-Middle Palaeozoic biogeography of Asian terranes derived
from Gondwana. 163-174. In mckerrow, w. s. and scotese, c. r. (eds). Palaeozoic palaeogeography and
biogeography. Geological Society, London, Memoir , 12, 435 pp.
cobbold, e. s. 1935. A new genus of the Trilobita and a new species of the Conchostraca from the Cambrian
of the Carteret Region, N.W. France. Annals and Magazine of Natural History, Series 10, 15, 381-392.
cowie, j. w. 1971. Lower Cambrian faunal provinces 31—46. In middlemiss, f. a., rawson, p. f. and newall,
g. (eds). Faunal provinces in space and times. Geological Journal, Special Issue, 4, 236 pp.
dangeard, l. 1954. Algues microscopiques a structure conservee dans le Cambrien de Carteret, Manche.
Bulletin de la Societe Linneenne de Normandie, 9, 54—55.

— and dore, F. 1957. Observations nouvelles sur les Algues et les stromatolithes du Cambrien de Carteret
(Manche). Bulletin de la Societe Geologique de France, 6 (7), 1069-1075.

debrenne, f. 1958. Sur les Archaeocyatha du Cambrien de Carteret (Manche). Bulletin de la Societe Geologique
de France, 6(8), 615-620.

— 1964. Archaeocyatha. Contribution a l’etude des faunes cambriennes du Maroc, de Sardaigne et de France.
Notes et Memoires du Service Geologique de Maroc, 179, 1-371.

— and debrenne, M. 1978. Archaeocyathid fauna of lowest fossiliferous levels of Tiout (Lower Cambrian,
Southern Morocco). Geological Magazine, 115, 101-119.

derville, h. 1933. Les recifs en coupole du Cambrien de Carteret. Bulletin de la Societe Geologique de France,
5(3), 603-618.

dore, f. 1963. Age et position des calcaires de Saint-Jean-de-la-Riviere dans le Cambrien des Moitier d’Allonne
(Manche). Bulletin de la Societe Geologique de France, 7(5), 272-279.

— 1969. Les formations cambriennes de Normandie. Unpublished Ph.D. thesis. University of Caen.

— 1984. The problem of the Precambrian-Cambrian boundary in the Armorican Massif. Bulletin de Liaison
et Information, International Geological Correlation Programme, Project 196, 2, 39-43.

— dupret, L. and le gall, J. 1986. Les premiers temps de la transgression cambrienne sur le socle dans le
Cotentin. Bulletin de la Societe Geologique et Mineralogique de Bretagne, 16, 49-59.

— and reid, R. E. 1965. Allonnia tripodophora nov. gen., nov. sp., nouvelle eponge du Cambrien inferieur de
Carteret (Manche). Compte Rendu Sommaire des Seances de la Societe Geologique de France , 1965, 20-21.

egorova, L. p., repina, L. N. and Suvorova, N. p. 1983. [Class Trilobita], 102-145. In sokolov, b. s. and
Zhuravleva, i. t. (eds). [Lower Cambrian stage subdivision of Siberia. Atlas of fossils]. Trudy Instituta
Geologii i Geofiziki. IzdateT stvo Nauka, Moscow, 558, 216 pp. [In Russian].
geyer, g. 1990. Die marokkanischen Ellipsocephalidae (Trilobita: Redlichiida). Beringeria, 3, 3-363.
GRAINDOR, M. J., ROBLOT, M. M., ROBARDET, M., DORE, F., PONCET, J., LORENZ, C. and HOMMERIL, P. 1976. Carte
geologique de la France 1/50000, feuille XI- XII- 11 Briquebec-Surtainville. Bureau de Recherches
Geologiques et Minieres.

Harrington, H. J. 1959. General description of Trilobita. 38-126. In moore, r. c. (ed.). Treatise on invertebrate
paleontology. Part O. Arthropoda 1. Geological Society of America and University of Kansas Press, Boulder,
Colorado and Lawrence, Kansas, xix + 560 pp.

— moore, r. c. and Stubblefield, c. j. 1959. Morphological terms applied to Trilobita. 117-126. In moore,
r. c. (ed.). Treatise on invertebrate paleontology. Part O. Arthropoda 1. Geological Society of America and
University of Kansas Press, Boulder, Colorado and Lawrence, Kansas, xix + 560 pp.

henningsmoen, G. 1959. Ellipsocephalacea. 207-212. In moore, r. c. (ed.). Treatise on invertebrate paleontology.
Part O. Arthropoda 1. Geological Society of America and University of Kansas Press, Boulder, Colorado
and Lawrence, Kansas, xix + 560 pp.

hupe, p. 1951. Sur les affmites des Trilobites. Bulletin de la Societe Geologique de France, 6, 169-186.

— 1953a. Contribution a l’etude du Cambrien inferieur et du Precambrien III de FAnti-Atlas marocain.
Contribution a l’etude des Trilobites du Maroc. Faunes de trilobites et zones paleontologiques du Cambrien
inferieur de FAnti-Atlas. Notes et Memoires du Service Geologique du Maroc, 103 (for 1952), 41-296,
335-357, 385-402.

19536. Classification des Trilobites. Annales de Paleontologie, 39, 59-168.

jegorova, l. i. 1983. [New Lower Cambrian trilobites from the southeastern part of the Siberian platform].

Paleontologicheskii Zhurnal, 1983, 59-64. [In Russian].
jell, p. a. 1990. Trilobita. 257-322. In bengtson, s., conway morris, s., cooper, b. j., jell, p. a and
runnegar, b. n. (eds). Early Cambrian fossils from South Australia. Memoir of the Association of
Australasian Palaeontologists, 9, 364 pp.

880

PALAEONTOLOGY, VOLUME 36

lin jin-lu, fuller, M. and zhang wen- you 1985. Paleogeography of the North and South China blocks during
the Cambrian. Journal of Geodynamics, 2, 91-114.

linan, e. 1978. Bioestratigrafia de la Sierra de Cordoba. Boletin de la Universidad de Granada , 191, 1-212.

— moreno-eiris, e., perejon, a. and schmitt, m. 1981. Fossils from the basal levels of the Pedroche
Formation, Lower Cambrian (Sierra Morena, Cordoba, Spain). Boletin de la Real Sociedad Espahola de
Historia Natural , (Seccion Geologica), 79, 277-286.

— and sdzuy, K. 1978. A trilobite from Lower Cambrian of Cordoba (Spain) and its stratigraphical
significance. Senckenbergiana Lethaea, 59, 387-399.

McMENAMiN, m. a. s. 1987. The emergence of animals. Scientific American, 255, 94-102.

McNamara, k. J. 1986c/. A guide to the nomenclature of heterochrony. Journal of Paleontology , 60, 4-13.

— 19866. The role of heterochrony in the evolution of Cambrian trilobites. Biological Reviews , 61, 121-156.
moore, r. c 1959. Superfamily Redlichiacea. 198. In moore, r. c. (ed.). Treatise on invertebrate paleontology.

Part O. Arthropoda 1. Geological Society of America and University of Kansas Press, Boulder, Colorado
and Lawrence, Kansas, 560 pp.

opik, a. a. 1958. The Cambrian trilobite Redlichia : organization and generic concept. Bulletin of the Bureau of
Mineral Resources, Geology and Geophysics , 42, 1—51.

1961. Alimentary caeca of agnostids and other trilobites. Palaeontology, 3, 410-438.

— 1967. The Mindyallan fauna of north-western Queensland. Bidletin of the Bureau of Mineral Resources,
Geology and Geophysics, 74, 1 — 404.

1975. Cymbric Vale fauna of New South Wales and Early Cambrian biostratigraphy. Bulletin of the
Bureau of Mineral Resources, Geology and Geophysics, 159, 1-78.
pillola, G. L. 1990. Lithologie et Trilobites du Cambrien inferieur du SW de la Sardaigne (Italie): implications
paleobiogeographiques. Comte Rendu de TAcademie des Sciences, Paris, 310, 321-328.

1991. Trilobites du Cambrien inferieur du SW de la Sardaigne, Italie. Palaeontographia Italica, 78, 1-174.
poulsen, c. 1927. The Cambrian, Ozarkian and Canadian faunas of northwest Greenland. Meddelelser om
Gronland , Series 2, 40, 1-83.

— 1959. Redlichiidae, Redlichiinae, Pararedlichiinae. 199-201. In moore, r. c. (ed.). Treatise on invertebrate
paleontology. Part O. Arthropoda 1. Geological Society of America and University of Kansas Press, Boulder,
Colorado and Lawrence, Kansas, xix + 560 pp.

rasetti, f. 1972. Cambrian trilobite faunas of Sardinia. Atti dell ' Accademia Nazionale dei Lincei, Memorie,
Classe di Scienze fisiche, matematiche e naturali. Serie 8, 11, 1-100.
repina, L. N. 1960. Komplesky trilobitov nizhnego i srednego kembriya zapdanoi chasti vostochnogo Sayana.
[Complexes of trilobites of the Lower and Middle Cambrian of the western part of the Eastern Sayan].
Regional' naya Stratigraphiya SSSR, Akademiya Nauk, 4, 171-232. [In Russian].

— 1966. Trilobity nizhnego kembriya yuga Sibiri (nadsemeistvo Redlichioidea), 1. [ Trilobites of the Lower
Cambrian of southern Siberia. Superfamily Redlichioidea, 7], Trudy Instituta Geologii i Geofiziki. Sibiriskoe
Otdelenie, 204 pp. [In Russian].

— 1969. Trilobity nizhnego i srednego kembriya yuga Sibiri. (nadsemeistvo Redlichioidea), 2. [Trilobites of
the Lower and Middle Cambrian of southern Siberia. Superfamily Redlichioidea, 2], Trudy Instituta Geologii
i Geofiziki. Sibiriskoe Otdelenie , 67, 1-109. [In Russian].

— and luchinina, v. a. 1981. [On the biostratigraphy of the lower part of the Lower Cambrian of north-
west Anabarya], Trudy Instituta Geologii i Geofiziki. Sibiriskoe Otdelenie, 475, 3-19. [In Russian].

richter, R. 1933. Crustacea (Palaontologie). In Handworterbuch der Naturwissenschaften, 2, 860-864 pp.
rozanov, A. yu 1984. Some aspects on bio- and palaeogeography of the Cambrian. 27th International Geological
Congress, Moscow 1984, (2) Paleontology. Science Press, Utrecht, 143-157.
sdzuy, K. 1959. Die unterkambrische Trilobite-Familie Dolerolenidae. Senckenbergiana Lethaea , 40, 389 — 407.
1961. Das Kambrium Spaniens. Teil II: Trilobiten. Akademie der Wissenschaften und der Literatim
Abhandlungen der Mathematisch-Naturwissenschaftlichen Klasse , 8, 217-408.

— 1978. The Precambrian-Cambrian boundary beds in Morocco (preliminary report). Geological Magazine,
115, 83-94.

1981. Der Beginn des Phanerozoikums - Paliiobiologische und stratigraphische Probleme. Natur und
Museum , 111, 390-398.

Stubblefield, c. J. 1936. Cephalic sutures and their bearing on current classification of trilobites. Biological
Reviews, 2, 407^140.

Suvorova, n. p. 1960. Trilobity kembriya vostoka sibiriskoi platformy. 2: Olenellidy-Granulariidy. [Trilobites
of the Cambrian of the eastern Siberian Platform. 2: Olenellidae-Granularhdae]. Trudy Palaeontolo-
gicheskogo Institua. Akademiya Nauk SSSR. Moskva , 84, 1-238. [In Russian],

PILLOLA: LOWER CAMBRIAN TR1 LOBITES

881

Whittington, H. B. 1981. Paedomorphosis and cryptogenesis in trilobites. Geological Magazine , 118, 591-602.
zhang wen-tang and lin huang-ling 1980. Superfamily Redlichiacea. 61-229. In zhang wen tang, lu yan-

HAO, ZHU ZHAO-LING, QIAN YI-YUAN, LIN HUAN-LING, ZHOU ZHI-YI, ZHANG SHENG-QUI and YIUAN JIN-LING,

1980. [Cambrian trilobite faunas of south-western China]. Palaeontologia Sinica, Series B, 16, 497 pp. [In
Chinese with English summary].

Zhuravleva, i. t. 1984. The Atdabanian Stage. 96-100. In rozanov, a. yu and sokolov, b. s. (eds). Lower
Cambrian stage subdivision, stratigraphy. Trudy lnstituta Geologii i Geofiziki , Moscow , 184 pp. [In Russian].

G. L. PILLOLA

Typescript received 5 May 1992
Revised typescript received 15 March 1993

Dipartimento di Scienze della Terra
Universita di Cagliari
Via Trentino 51
09127 Cagliari, Italy

*

IMPLICATIONS OF THE AMINO ACID
COMPOSITION OF RECENT NEW ZEALAND

BRACHIOPODS

by DEREK WALTON, MAGGIE CUSACK and GORDON B. CURRY

Abstract. Intracrystalline proteins and amino acids were extracted from within the shells of several species
of Recent brachiopods from New Zealand. These molecules were characterized both by amino acid analysis
and by SDS-PAGE and the data derived from this summarized by multivariate statistical analysis. Results
indicate that the amino acids from within the shell may be used as objective taxonomic discriminators, and that
the groupings recognized by this method are generally in confirmation with those from the Treatise on
invertebrate paleontology. The presence of amino acids containing taxonomic information within the shells of
Recent brachiopods has implications for the use of amino acid analysis in the investigation of fossil species.

The intracrystalline organic molecules which are present as a mixture within the shells of Recent
brachiopods are a rich source of information regarding the taxonomy of the organisms (Collins et al.
1988, 1991 b\ Curry et al. 1991 b, 1991c; Cusack et al. 1992). The molecules consist mostly of
protein, although quantities of lipids (Curry et al. 1991 b), alcohols (H. Clegg pers. comm.) and
carbohydrate (Collins et al. 1991) are also present. Although the function of these molecules, and
their role, if any, in the process of biomineralization is unclear, they are of significance in the field
of molecular palaeontology.

Of the molecules present within the shells, lipids, alcohols and carbohydrates generally contain
little information of significant taxonomic value. Proteins, however, contain information derived
from the DNA of the organism, which may be utilized in studies of molecular taxonomy (e.g. Joysey
1988). Proteins are composed of relatively stable monomeric units, amino acids, which are present
in a defined order along the chain of the molecule. It is this information, the primary structure,
which is conventionally used for biochemical comparison of protein, as differences in the nucleotide
bases in the DNA of organisms are usually (but not always) reflected in differences in the nature and
order of amino acids (Crick et al. 1961). Analysis of the position of the changes in amino acids or
nucleotide bases can only be completed by sequencing either the proteins or the DNA of the
organism. The complete sequencing of proteins is a difficult and time-consuming process; even using
automated sequencers, it is rare to be able to sequence more than 40 amino acids along a protein
chain at any one time (counting from the A-terminus). Either enzymatic and/or chemical cleavage
of proteins, or the use of DNA probes, is necessary to elucidate further sequence information,
although there is no guarantee of success through either of these procedures.

Curry et al. (1991a) and Cusack et al. (1992) obtained A-terminal sequence information from
brachiopod intracrystalline proteins. These proteins are homologous between different brachiopod
species, although they do not correspond to the sequence of any other known proteins. The
sequence of a small (6-5 kDa) protein analysed from three New Zealand genera (Cusack et al. 1992)
shows few substitutions between the proteins. However, as yet, no study has completed the
sequencing of more than a quarter of the total residues contained within any particular brachiopod
shell protein.

Differences in the DNA between species will also be reflected by different relative proportions of
amino acids within a sample, which may be easily quantified by amino acid analysis. This will reduce
to some degree the level of information available from the proteins (as the primary structure is lost).

IPalaeontology, Vol. 36, Part 4, 1993, pp. 883-896.|

© The Palaeontological Association

884

PALAEONTOLOGY, VOLUME 36

but this is compensated for by the much shorter time and much smaller sample size since c. 10 pg
shell is sufficient for this type of analysis, while protein sequencing requires c. 100 g shell to provide
sufficient information. Amino acid quantification is rapid and is able to identify variation in the
relative proportions of amino acids, and therefore differences in the composition of the proteins, of
a sample, although the location of this variation cannot be determined. The relative abundance of
each of the amino acids within each sample is a representation of their relative frequency within the
polypeptides of the sample, and hence differences between the relative abundance will equate to
the minimum number of differences between the codons of the DNA of the organisms being
compared, although some changes in nucleotide base will not result in a changed amino acid residue
in the protein (Crick et al. 1961).

It was with the inherent difficulties of protein sequencing in mind that the intracrystalline proteins
from several Recent brachiopod species (listed in Table 1) were characterized by their amino acid
compositions. This information was examined by multivariate statistical techniques to extract
potential taxonomic data from the variation in concentration of the molecules. Values obtained by
the study of extant species can also be used as a baseline (or time = 0) for studies of molecules
contained within shells of the same species in the fossil record.

table 1 . Species and localities used in SDS-PAGE and amino acid analysis.

Brachiopod

Locality

Neothyris lenticularis (Deshayes, 1839)
Neothyris parva (Cooper, 1982)

Waltonia inconspicua (Sowerby, 1846)
Notosaria nigricans (Sowerby, 1846)
Terebratella sanguinea (Leach, 1814)
Terebratella haurakiensis (Allan, 1931)
Liothyrella neozealandica (Thomson, 1918)
Gyrothyris mawsoni (Thomson, 1918)

Paterson Inlet, Stewart Island

Continental shelf, 10 miles SW of Otago Peninsula

Otago Harbour

Tikoraki Point, Moeraki, Otago
Paterson Inlet, Stewart Island
Hauraki Gulf, North Island
Dusky Sound, Fiordland

Continental shelf, 10 miles SW of Otago Peninsula

Other studies have shown that the number of proteins contained within the shell also varies
taxonomically (Curry et al. 19916). To provide a method of secondary confirmation of the
taxonomic conclusions of the amino acid technique, the organic extract of several of the samples
was fractionated by sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE)
using the method of Schagger and Von Jagow (1987). As SDS-PAGE requires large amounts of shell
material, Terebratellci haurakiensis , Gyrothyris mawsoni and Neothyris parva , could not be analysed
in this way. SDS-PAGE reveals which species have a similar number and size of proteins present
within their shells, which provides a secondary method of confirmation of similarity (Curry et al.
19916) in case of spurious relationships derived from the amino acid data. Throughout this study,
the standard one- or three-letter codes are used to denote the amino acids.

METHODS AND MATERIALS

Preparation

Samples of Recent brachiopods were collected from the locations given in Text-figure 1 and Table [.Wherever
possible, samples which were excessively bored or fractured were omitted from further study.

Sediment was scrubbed from the sample and encrusting epifauna removed by scraping. Articulated shells
were disarticulated and any remaining body tissue removed, before soaking in an aqueous solution of bleach
(10% v/v) for 2 hours at room temperature to digest adhering particles of organic matter, washed extensively
with Milli RO® water, and air dried. Samples were ground using a ceramic pestle and mortar, and the
powdered samples incubated in an aqueous solution of bleach (10% v/v) under constant motion for 24 hours
at room temperature to destroy intercrystalline molecules by oxidation. Samples were washed by repeated

WALTON ET AL.: BRACHIOPOD AMINO ACIDS

885

text-fig. 1. Location of collection sites
for samples of extant species used in the
analyses.

agitation with MilliQ® water and centrifugation (typically ten washes). Powders were lyophilized, allowing
almost indefinite storage at room temperature.

To release the incarcerated molecules from within the calcium carbonate, it is necessary to dissolve the
inorganic phase. Two methods were employed :

1. Hydrochloric acid. For amino acid analysis, an aqueous solution of 2 m HC1 at a ratio of 1 1 /d/mg was
used to dissolve the shell powder. Once demineralization was complete, samples were centrifuged (20 |?.h.) to
remove any remaining insoluble particles. This method of preparation required no further concentration.

2. Ethylene diamino tetra acetic acid, disodium salt (Na2-EDTA). For samples to be analysed by
SDS-PAGE, an aqueous solution of Na2-EDTA (20% w/v) at a ratio of 23 ml/g shell was used to dissolve
the shell powder (Collins et al. 1988). Once demineralization was complete, samples were centrifuged (20 g.h.) to
remove any remaining insoluble particles. The sample was initially filtered and concentrated using the
Minitan® tangential flow system (Millipore; Cusack et al. 1992), which removed the EDTA/calcium
complexes. The intracrystalline extract was further concentrated using a Minicon® static concentrator ( Amicon).
Both filtration systems used 10 kDa cutoff filters.

SDS-PAGE

The polyacrylamide gels used in this study were prepared to a method based on that of Schagger and Von
Jagow (1987). An aqueous solution of the protein was mixed with an equal volume of buffer solution
containing final volumes of 0T5 m Tris/HCl, pH 6 8, 0 2 m 2-mercaptoethanol, 01 % (w/v) SDS, 30% (w/v)
glycerol and 0 0002% (w/v) of tracking dye (bromophenol blue) and incubated at 100 °C for 4 minutes to
denature the protein and to allow SDS to bind to the protein. Proteins were separated by the application of
a constant voltage of 50 V for 4 hours. Following electrophoresis, proteins were visualized by Coomassie
Brilliant Blue R-250 stain.

Amino acid analysis

Hydrolysis of the peptide bonds which link individual amino acids in proteins and peptides is a prerequisite
for amino acid analysis. Vapour phase hydrolysis, using 6 n HC1, was completed using automated hydrolysis
on the Applied Biosystems 420H Amino Acid Analyser. Standard proteins and peptides were used during every
analysis to ensure that hydrolysis proceeded to completion. Blank analyses were included to check for
background levels of contamination.

Following hydrolysis, individual amino acids may be derivatized using Phenylisothiocyanate (PITC;
Heinrikson and Meredith 1984) on the AB1 420H (Dupont et al. 1989), which absorbs UV light strongly at
254 nm. The derivatized amino acids were transferred to the dedicated ABI 130A narrowbore hplc system for
separation and quantification (Dupont et al. 1989).

886

PALAEONTOLOGY, VOLUME 36

49 kDa

Neothyris

lenticularis

Waltonia

inconspicua

Terebratella

sanguinea

Notosaria
nigricans
50 kDa h

20 kDa mm

Liothyrella

neozealandica

« 42 kDa
» 22 kDa

16 kDa

mm

17 kDa s

h 16 kDa

6.5 kDa

“

14 kDa

«■ 6.5 kDa

text-fig. 2. Representation of the size and distribution of the major protein bands from the > 10 kDa
fraction of the organic extract of Recent brachiopods, separated by SDS-PAGE. Note the differing numbers
and sizes of the proteins in the samples. (Original photographs of the SDS-PAGE separated samples are

available from the senior author on request.)

To assess the proportion of the total shell protein which is due to the intracrystalline fraction, samples of
Neothyris lenticularis , Waltonia inconspicua , and Liothyrella neozealandica were decalcified by HC1 without
prior bleaching. Concentrations of amino acids were converted to weight percentages (wt%) using the
EXCEL® spreadsheet and analysed by the statistical program DATADESK® on the Macintosh
microcomputer.

Direct confirmation of the presence of intracrystalline amino acids

Within their shells, brachiopods contain organic molecules with antigenic properties (Collins et al. 1988). To
confirm the presence of these molecules by direct analysis, and to confirm that the bleaching process effectively
removed the intercrystalline molecules, the following experiment was completed.

A sample of shell powder (0-70 g) of Terebratella sanguinea was directly weighed into newly pyrolysed
(500 °C/4 hours) glass universal bottles, 10 ml of pure water (MilliQ®) added and the sample incubated at
110°C for 24 hours. On removal, the sample was allowed to cool, and an aliquot (4-5 ml) removed and
concentrated on a rotary evaporator (Elowe Gyrovap) to 25 p\. An aliquot (20 pX) was added to the sample frit,
and the standard hydrolysis and derivatization cycles run.

RESULTS

Gel electrophoresis

A composite gel of the samples separated by SDS-PAGE is shown in Text-figure 2. In all cases, with
the exception of Liothyrella neozealandica , up to 20 replicates of each sample were completed. There
was only sufficient material of L. neozealandica to allow a single replicate. Samples are grouped
according to their taxonomic positions given by Williams et al. (1965).

Neothyris lenticularis , Waltonia inconspicua and Terebratella sanguinea (Order Terebratulida,
Suborder Terebratellida) all show several major bands of approximate molecular weights 49, 16 and
6-5 kDa. Liothyrella neozealandica (Order Terebratulida, Suborder Terebratulidina) has 4 main
bands of approximate molecular 42, 22, 16 and 6-5 kDa. Notosaria nigricans ( Order Rhynchonellida)
has 4 main bands of approximate molecular weights 50, 20, 17 and 14 kDa. The number and size
of the proteins contained within the shell varies taxonomically at the subordinal level, providing a
low sensitivity method of taxonomic discrimination.

Direct confirmation of the intracrystalline nature of the amino acids

Results indicated that for every milligram of powder analysed in this way, there are 0-7 nanograms
of amino acid. The shell powder was frozen and lyophilized again, and the inorganic phase
demineralized by 2 N HC1 as above. The concentration of amino acid yielded by this process was
169-4 ng/mg, slightly lower than the average for Recent Terebratella sanguinea (180 ng/mg), but
within the range of experimental error. The proportion of amino acid which may be attributed to

WALTON ET A L.\ BRACHIOPOD AMINO ACIDS

887

the intercrystalline fraction which remains undestroyed by the bleaching procedure is c. 04%. The
amino acid quantified upon demineralization must therefore have been enclosed within the shell of
the brachiopod (i.e. intracrystalline).

Absolute abundance of amino acids

Amino acid analysis of the samples that were not bleached shows that the molecules extracted from
the intracrystalline fraction account in all three cases for 3CM-0 % of the total amino acid present
within the shell. On average, less than 10 % of the total amino acid present within the shell of Recent
brachiopods is in the form of free amino acids (Table 2), indicating that the vast majority are
combined into proteins and peptides. A notable exception to this is tyrosine in Notosaria nigricans,
which is c. 75% free. As there are generally very few free amino acids in these Recent samples, die
indications are that the acid decalcification does not hydrolyse many (if any) of the peptide bonds
in the protein. The absolute abundance of all amino acids (free and combined) varies taxonomically,
ranging between 70 and 800 ng/mg (amino acid / shell), equivalent to 0-007 and 0-08% of the total
weight of the shell. Sample variability was between 5 and 8%. Concentrations which had a higher
variability than this were repeatedly re-analysed until a series of consistent results was produced.
Liothyrella neozealandica and Neothyris lenticular is both contain low concentrations of amino acid,
which contrasts with the high concentrations found in Notosaria nigricans , a feature comparable to
the values found for the total organic matter contained within the shell (Curry et al. 1989).

Decalcification of the Notosaria nigricans shell powder, in contrast to the remaining samples, left
a black insoluble residue. This is likely to be caused by the differing nature of the shell protein
contained in the shell matrix of this species, which consists partly of an acid-insoluble fraction,
which is more resistant to the oxidative effect of the exposure to sodium hypochlorite (Collins et al.
1991 b). The insoluble compounds were removed by centrifugation before amino acid analysis of
the soluble fraction.

Relative abundance of amino acids

To provide a basis for the comparison of the amino acids without the discrimination being solely
due to the concentration of the molecules, some form of standardization is necessary. In this study,
the concentrations were converted to weight percentages. These relative abundances are shown in
Table 2, which shows the variation in the amino acid content of the samples. These datasets are
based on between 2 and 10 replicates of each sample, with 2 to 4 plotted in each case, depending
upon the number of replicates. The most striking variations he in high Asp/Asn in Notosaria
nigricans (36-94 wt%) and Liothyrella neozealandica (11-51 wt%). N. nigricans also contains high
Tyr (7-44 wt%, much of which is present in the free state), but low Leu (0-68 wt%) and Glu/Gln
(2-47 wt%) in contrast to the other samples. Neothyris lenticularis and Neothyris parva contain high
Gly (51-51 and 54-76 wt% respectively). These proportions (i.e. high Gly, Ala and Asp/Asn) are
comparable to those found in intercrystalline molecules from the same species of brachiopods (Jope
1977; Kolesnikov and Prosorovskaya 1986). The actual values are somewhat different, which is not
surprising given that the molecules in this study are intracrystalline and are likely to be different
from the matrix proteins considered in these other studies.

The low proportion of free amino acids within the samples (Table 2) is important as it indicates
intact protein and peptide survival within the biocrystals of Recent shells, and also the absence of
contamination by sample handling (Walton and Curry 1991).

Glycine, the simplest of the amino acids, accounts for the highest proportion of the molecules
within the shells, in all cases being higher than 25 wt% and ranging up to more than 50 wt% of the
total. Glycine is also the most common of the amino acids found on human fingertips. However,
low concentrations of Gly are present in the free state, indicating that it is released by hydrolysis
of the proteins and peptides, and not from contamination by sample handling.

Direct comparisons of the relative abundance of amino acids between samples are hard to make,
as the overall change of so many variables is difficult to observe. The illustration of the scale and
direction of variation of the amino acid content of the samples is not possible, as this would require

table 2. Amino acid composition of the total organic extract from Recent brachiopods. Absolute concentrations in ng/mg. relative proportion in
weight percentage.

Amino acid

Sample

D/N

E/Q

S

G

R

T

A

P

Y

V

I

L

F

K

Waltonia inconspicua

absolute

8 23

9-82

694

76-76

6-70

3-74

11-66

12-85

2-94

10-74

4-31

3-96

2-22

3-58

relative

4-89

5-96

4 17

4503

3 86

2-27

7 14

7-93

1-72

663

3-52

3-42

1-25

2 41

%free

0

0

0

1 86

0

0

0

0

0

2-89

0

3396

0

0

Terebralella sanguined

absolute

1212

13-51

766

6970

6-28

6 58

1224

17-32

5-42

12-82

7-39

769

4-15

7-42

relative

637

7 10

403

3663

3-30

3-46

643

9 10

2-85

6-74

3-88

4-04

2 18

3-90

%frce

5-45

0

16-13

5-24

0

0

4-25

4-45

0

0

0

11 64

15 18

0

Terebratella haurakiensis

absolute

19 78

2206

20-60

132 04

15-79

8-73

3002

29-70

12-34

24-25

16-26

17-10

14-06

8 82

relative

5-32

5-94

5-54

35-54

4-25

2-35

8-08

7-99

3-32

6 53

4-38

4-60

3-78

2-37

%free

0

0

0

2-63

0

0

0

0

0

1-77

0

2-57

0

0

Neothyris lenticular is

absolute

6-76

8 35

354

56-06

3 15

2-69

5 18

7-74

1 39

6 16

1-78

2-92

1-70

1 47

relative

6 21

7 67

3 25

51 51

289

2-47

476

7 1 1

1 28

566

1 64

2-68

1 56

1 35

%free

9-76

0

0

5 05

0

0

5 21

0

0

0

0

16 10

0

21-77

Neothyris parva

absolute

4-23

5 34

246

34-76

1 63

1 58

2-22

3-47

034

3 43

1 41

1 48

0-93

043

relative

6-66

8 37

3 84

54-76

2-54

2-47

3-46

5-43

0-54

5 37

2-20

2-29

1 44

0-66

%free

ND

ND

ND

ND

ND

ND

ND

ND

ND

ND

ND

ND

ND

ND

Notosaria nigricans

absolute

214-52

14-35

1 181

21820

9-25

5-89

16 08

1496

43-22

7-90

2 81

396

13 45

4 39

relative

3694

2-47

2-03

37-57

1-59

101

2-77

2-58

7-44

1 36

0-48

0-68

2-32

0-76

%free

0 41

0

0

1 04

0

0

2-24

2-74

74-32

3 16

0

9 85

4 01

0

Gyrothyris mawsoni

absolute

10-29

14 16

6-86

71-15

8 81

4-86

12-37

13-37

3-92

966

5 31

5 53

3-10

2-99

relative

5-97

8 21

3 98

41 28

5 11

2-82

7 18

7-76

2-27

560

308

3 21

1-80

1-73

%free

0

0

0

205

0

0

2-26

0

0

0

0

6 33

0

0

Lioth vrella neoiealandica

absolute

706

4 91

2 1 1

27-89

1 36

1-93

3 12

3-62

0 41

4-88

1-37

1-50

0-54

0-65

relative

11 51

800

344

4546

2-22

3-15

509

5-90

067

7 95

2-23

2 44

0-88

1-06

%free

0

0

0

6-49

0

0

0

0

0

0

0

0

0

56-92

table 2. Amino acid composition of the total organic extract from Recent brachiopods. Absolute concentrations in ng/mg, relative proportion in
weight percentage.

PALAEONTOLOGY, VOLUME 36

00

04

O

<N

r-

o-

in

O-

co

X

Ov

X

Ov

CO

in

X

04

in

Of

of

p

OO

m

p

p

P

Of

X

Q

p

0-

p

p

X

O

OV

cb

04

O

O'

cb

0

do

04

O

04

0

0

Z

of

0

0

04

1

O

0

1

vb

in

04

in

in

OO

00

vo

00

0

X

CO

of

in

04

,

O

0

of

00

04

04

— -1

— •

p

p

p

p

in

Ov

p

Q

p

co

p

p

°p

in

OO

Ph

04

~

O

of

04

in

d"

cb

O

“

"

O

O

'__l

z

cb

04

of

cb

■ 1

O

O

O

O

X

oi

vo

Ov

of

of

O

O

04

00

O

OO

Ov

X

OO

m

co

f (

CO

O

of

p

Txf

p

X

O

vp

X

m

p

X

«— 1

p

04

n

qv

X

00

p

p

CO

p

of

cb

cb

cb

O'

Of

X

bf

04

04

04

vb

X

04

l— 1

cb

0

dv

in

X

vb

X

04

O

co

’ 1

’ 1

Z

04

ov

OO

vo

00

OO

Of

,

O

,

00

!

00

O'

co

co

in

co

OO

<N

pi

p

X

p

04

Q

of

CO

p

p

04

-

of

cb

0

O'

cb

O

vb

0

0

04

z

04

O

0

in

cb

0

r~~i

04

O

of

co

OV

04

of

in

m

X

X

CO

O-

O

X

X

X

O

00

m

O'

vp

OO

OO

c 1

>n

p

p

X

p

p

Q

qv

p

p

X

X

00

Ov

>

0

vb

04

dj

vb

O

04

VO

vb

in

0

cb

in

z

O'

cb

dv

in

0

of

O'

O

of

<N

04

in

r 1

OV

00

of

of

04

of

04

04

!

O'

OV

p

of

00

p

cp

CO

04

CO

m

Q

04

p

p

qv

04

of

X

>H

04

O

in

04

O

04,

cb

0

"

~

0

O

O

z

cb

of

O'

of

O'

cb

04

0

0

0

O

co

04

O

in

O

Ov

of

co

X

>0

of

X

04

0

00

ON

co

*— «

of

r-

p

p

p

n

p

p

r-

p

p

X

p

04

O'

O

X

dv

d-

dv

r-

0

O-

r-

0

cb

in

Z

of

04

04

cb

O'

0

cb

in

O

Cl,

04

,"H

VO

of

of

co

m

04

00

OO

X

,

04

X

00

of

O'

00

X

04

Ov

VO

p

04

of

04

O

p

p

p

p

04

p

n

0

O'

04

CO

—

04

- — 1

O

<

1—1

O'

O

04

vb

of

O

m

do

0

in

bf

in

04

cb

Z

vb

04

04

04

O'

04

cb

in

O

of

O'

OO

vo

m

in

Ov

o-

OO

Q

ov

,

X

04

co

m

p

04

in

p

p

cq

0

X

of

in

of

op

p

°q

00

p

H

cb

04

O

vb

cb

O

do

04

04

04

O

04

Z

in

O

of

04

O

cb

O

O

X

00

0

ov

in

in

OV

co

of

m

Ov

!

X

04

O'

cp

04

p

p

p

p

00

X

in

Q

04

p

°o

p

p

04

vb

cb

O

vb

cb

O

in

of

0

cb

04

O

04

z

dv

1—1

O

do

in

O

04

O

VO

co

vo

0

CO

of

of

CO

X

,

in

X

X

0

O'

of

in

00

in

Os

X

Ov

o>

p

OO

O'

VO

04

O

in

vp

p

in

p

p

p

ri

•n

O

’“H

04

0

00

of

of

vb

in

1— H

dv

vb

in

04

in

04

vb

X

in

of

of

Q

00

O'

X

X

X

04

O'

in

vb

0

r-

of

vo

co

m

co

in

in

CO

m

z

04

co

O'

of

04

of

of

r-

vo

co

co

0

of

of

m

X

of

,

CO

X

00

of

Ov

p

vo

p

vo

in

in

04

of

00

Q

°!°

O

00

p

- — 1

of

vb

of

O

O'

of

vb

0

wb

O

cb

cb

0

04

cb

•—r

X

04

O

vb

cb

O

04

cb

0

oo

04

Z

r— '

O'

04

vo

,

O

VO

Of

in

O'

Of

in

O'

X

O

OO

p

in

p

O

p

cq

X

cq

co

n

p

p

p

Cl

P

p

T3

w

dv

in

O

cb

d-

0

04

in

O

00

O'

O

in

OO

of

04

O

of

do

O

of

dD

0

O

a

04

z

o

co

Ov

04

O'

in

OO

04

X

,

X

rr>

X

04

of

_

OV

O'

X

!

c

7

04

OO

co

p

p

r-

04

r-

p

X

p

Ov

of

04

p

O

in

'g

Z- 1

00

of

O

04

vb

in

dv

wb

O

vb

vb

dv

Of

vb

Q

of

vb

O

O

in

O

O-

X

0

<

Q

.6Q

Z

04

CO

Q

3

Q

.g>

3

,SI

<>5

-S

.Vj

’5,

§D

3

■5

3

S

<>5

cu

05

5

5

3

.JJ

p»

!V.

S3

.go

N

0

<D

<D

-S

<v

0)

<D

.05

<u

1)

<D

<D

<D

.Jo

<D

.jo

o>

<u

<L>

<D

.3

0>

J3

o>

_>

<L>

p

>

o>

p

_>

<D

p

>

1)

£

>

<u ^

>

<D

n,

£

'o

C/5

JS

<D

*

»c>

O

C/5

C3

<L)

cfc|

53

-C)

0

C/5

p

<D

b

0

C/5

p

<u

43

'o

C/5

p

O

43

3

O

C/5

p

<v

43

-5‘

0

C/5

q

O

C/5

<D

£

s

c3

(Z)

1

53

J-t

£

£

X)

c3

13

u.

£

x.

X

53

i-H

v®

oN

3

X

ctf

13

J-H

v®

ox

3

X

d

13

l-H

0

1

X

a

13

0

6‘

X

cd

13

l-H

ox .O

^4

X

cd

13

£

WALTON ET AL BRACHIOPOD AMINO ACIDS

889

U1

<■

A Gvrothyris
■ N. lenticularis
v" N. parva
'k Liothvrella
-#■ Notosaria

• Waltonia

# T. sanguinea

x T. haurakiensis

x

x

♦

text-fig. 3. Graph of the first three principal components (denoted by Ul, U2 and U3) of the relative
proportion data for amino acids from Recent brachiopods (complete organic extract). Note close groupings
of related species (e.g. Neothyris lenticularis and N. parva) and the distance of more distantly related species

(e.g. Notosaria ).

multi-dimensional space which cannot be used. For this reason, multivariate statistical techniques
were applied to the dataset to summarize the variables and produce more meaningful information
regarding the variation of the samples. Principal component analysis (PCA) is a method of
analysing the dataset in a way which summarizes it into derived variables, which may then be
compared in graphical form. Text-figure 3 shows the plot of the first three principal components
extracted from the dataset, presenting a summary of the data in which most of the variation is
contained within fewer, derived variables. The new dataset is plotted into a rotating plot, allowing
three dimensional observation (Text-fig. 3), and onto a scatterplot (Text-fig. 4). In Text-figures
3 and 4, samples that are closely related plot close together.

Amino acid distribution in intracrystalline proteins allows separation of the samples solely in
terms of the variability in the dataset; i.e. an objective process where the difference between the
genomes of the organisms is reflected by different amino acid compositions. PCA scores each amino
acid in terms of how much of the total variation in the dataset is due to that variable (i.e. the size
of the eigenvalue). The first principal component contains 47-8% (Table 3) of the variability of the
dataset, and this variation is mainly due to differences between samples in the proportion of Ser
( — 0-328), Ala ( — 0-349), Pro ( — 0-351), lie ( — 0-361) and Leu ( — 0-372). The second principal
component contains 27-4% of the variability, mainly due to differences in the proportion of Gly
(0-332), Tyr ( — 0-500), Val (0-335) and Phe ( — 0-362). The third principal component contains
10-5 % of the variation, due mainly to variation in the abundance of Asp/Asn ( — 0-304), Gly (0-350),
Arg (0-542), Thr ( — 0-451) and Lys ( — 0-375). The signs of the eigenvalues are also important, as
these indicate the direction of the sample along the particular eigenvector. Both the first two
principal components (75-2%) and the first three principal components (85-7%) contain more than
75% of the total variability, and therefore conclusions reached from these data are valid (Sneath
and Sokal 1973).

Taxonomic information is difficult to derive from the principal component data alone, as this
method provides only discrete groupings of samples which are not linked. The method, at this scale
at least, is unable to separate the two species of the genus Neothyris. Cluster analysis is a method
of assaying the degree of ‘relatedness’ of the dataset, and the variables were analysed by single

890

PALAEONTOLOGY, VOLUME 36

*

2 ■■

1 ■■

0 --

-1 -

A Gvrothyris
■ N. lenticularis
■fr N. parva
t V Liothvrella
^ Notosaria

# Waltonia

# T. sanguinea

x T. haurakiensis

-2.50

x

x

I -

-1.25

4-

=f—

0.00

Second Principal Component

text-fig. 4. Scatterplot of the first two principal components of the relative proportions of amino acids from

Recent brachiopods.

table 3. Principal component analysis of the relative proportion data, calculated for samples of Recent
brachiopods. Only the first three eigenvectors and eigenvalues are given.

Eigenvalues

Eigenvectors

VI

V2

V3

value

proportion

D

0-267

-0-294

-0-304

el

6-694

47-8

E

-0091

0-437

0019

e2

3-839

27-4

S

-0-328

-0-032

0149

e3

1-470

10-5

G

0-201

0-332

0-350

R

-0-224

-0-081

0-542

T

-0-203

0-289

-0-451

A

-0-349

-0-067

0176

P

— 0-351

0-068

0029

Y

0-003

-0-500

-0-094

V

— 0180

0-335

-0-263

I

— 0-361

-0-058

0024

L

-0-372

-0021

-0-085

F

— 0-218

-0-362

0-083

K

-0-296

-0-122

-0-375

WALTON ET AL.: BRACHIOPOD AMINO ACIDS

891

N. parva
N. lenticularis
Liothvrella
Waltonia

Gvrothvris
T. sanguinea
T. haurakiensis
Notosaria

text-fig. 5. Single linkage cluster analysis showing apparent relationships of extant brachiopods from New
Zealand. Note the aberrant position of Liothvrella , caused by chance similar ratios of amino acids. Liothyrella
contains a different number and size of proteins (Text-fig. 2), indicating that the extracts should only be

compared with caution.

linkage analysis, which is equivalent to the nearest neighbour analysis of Sneath and Sokal (1973).
The dendrogram in Text-Figure 5 is derived from cluster analysis of the relative proportion data
shown in Table 2. Taxonomic relationships can be seen by the positions of the clusters and in the
links between clusters. The genus Notosaria plots away from the remaining data, whereas the two
species of the genus Neothyris form a discrete cluster. Waltonia inconspicua and Gyrothyris mawsoni
are closely related. The two species of the genus Terebratella , however, do not appear to be as
closely related as would be expected, while Liothyrella neozealandica appears to be related to the
genus Neothyris.

DISCUSSION

Application of this method to numerical taxonomy

Numerical taxonomy was defined by Sneath and Sokal (1973, p. 4) as ‘the grouping by numerical
methods of taxonomic units into taxa on the basis of their character states’, and implied
repeatability and objectivity. Any method which relies on objective data collection may therefore
be utilized in a study of numerical taxonomy. The work of Degens et al. (1967) provides a theoretical
and practical basis for the characterization of species by the amino acid composition of proteins
from their body tissue, and illustrates the use of multivariate statistical analysis of amino acid
differences for taxonomic purposes using the principles of numerical taxonomy, i.e. utilizing non-
subjective criteria.

The theory of the relationship of proteins based on their differences in amino acid composition
has important implications for the use of amino acid analysis of intracrystalline molecules as a
taxonomic tool. The technique has been criticized in the literature for not providing enough data
on the taxonomic difference between samples, and also for being difficult to interpret (e.g. Collins
et al. 1991; Logan et al. 1991). However, these arguments have rested almost entirely on the
assessment of differences of amino acids on a ratio and relative proportion basis, the limitations of
which were demonstrated by Walton and Curry (1991). In that study, the amino acid compositions
of finger tips and sediments showed a gross similarity with the relative proportions of amino acids
present in fossils, a relationship which did not stand up to reappraisal by more rigorous multivariate
statistical techniques which revealed a clear discrimination between the data. Comparisons of the

892

PALAEONTOLOGY, VOLUME 36

ratios of amino acids are misleading, and the use of more powerful multivariate statistical
techniques, such as cluster analysis and PCA, can lead to more certain discrimination between data
points.

Factor analysis, a multivariate statistical technique, was used by Degens et al. (1967) in an
analysis of the bulk amino acid composition of both calcified and non-calcified tissues from the
Mollusea. Degens et al. (1967) produced a phylogenetic tree from bulk amino acid analyses of
proteins, which is entirely consistent with the phylogeny produced from conventional morphological
techniques. However, Degens et al. (1967) used only numerical descriptions of the factor scores of
the data, and made no attempt to describe it graphically, rendering interpretation of the data
difficult. Q-mode factor analysis of shell proteins was used in a phylogenetic study of Recent and
fossil planktonic foraminifers (King and Hare 1972), where the data were expressed on a triangular
plot, representing the equivalents of the first three principal components (cf. Text-fig. 3). King and
Hare (1972) were also the first to conclude that chemotaxonomy of amino acids was viable in extinct
species. Haugen et al. (1989) completed a chemotaxonomic study of benthic foraminifers using
amino acid data and combined the use of ratios and PCA to infer relationships between the species.

All the above studies utilized analysis of the bulk extracts of the shell (i.e. both intercrystalline
and intracrystalline molecules) to examine taxonomy. Intercrystalline molecules are in a position
whereby they may be easily contaminated. This possibility is not present when the intracrystalline
molecules alone are analysed, as is the case in this study.

Taxonomic conclusions

The following taxonomic conclusions may be reached for the samples in this study. These
conclusions utilize the results from both SDS-PAGE analysis of samples as well as the amino acid
analyses. The amino acid concentrations are used entirely objectively. Samples of the genus
Neothyris are considered to be members of separate species, in terms of their position in the cluster
diagrams (Text-fig. 5). The plot of the first three principal components (Text-fig. 3) is unable to
separate the samples, indicating that they are closely related. Gyrothyris mawsoni and Waltonia
inconspicua are also closely related, although this relationship is obviously not as close as for the
Neothyris samples. This conclusion is in agreement with the data from the Treatise on invertebrate
paleontology (Williams et al. 1965), which shows that these two species are members of the same
subfamily.

The genus Terebratella is represented in this study by T. haurakiensis and T. sanguinea. As these
species are classified as being members of the same genus, it was expected that they would plot
together on the PCA plot (Text-fig. 3), and form a discrete cluster (Text-fig. 5), similar to the two
species of Neothyris. However, the bulk amino acid compositions of the two species of Terebratella
are markedly different (Table 2), with T. haurakiensis having a higher overall concentration of the
amino acids, and also a different relative proportion of the amino acids. This leads to the separation
of the data points for these samples in both the PCA plots and in cluster analysis. It therefore
appears likely, from these bulk amino acid data, that the classification of this genus may need
revising. Further work on the composition and size of the intracrystalline molecules from
Terebratella species, especially from an area where the two species coexist, is necessary in order to
confirm the need for revision of the classification. Unfortunately, only a limited amount of material
was available for the study of T. haurakiensis , rendering it impossible to employ SDS-PAGE to
assess whether the differences in the amino acid composition between the two samples are due to
changes in the actual composition of similar sized proteins in the species, or whether the samples
contain different numbers of proteins of different sizes and are therefore more distantly related than
currently thought (cf. the amino acid composition of Liothyrella neozealandica below). The
morphological difference between the two species is in terms of the ribbing pattern of the shell:
T. sanguinea is heavily ribbed, whereas T. haurakiensis is smooth. This could be due to ecophenotypic
variation, as seen by the ribbing of the related, normally smooth-shelled Waltonia inconspicua ,
which is ‘not uncommon’ (D. E. Lee pers. comm.), or it could be due to a more fundamental
difference, as the amino acid data would tend to suggest.

WALTON ET AL .: BRACHIOPOD AMINO ACIDS

893

The species Liothyrella neozealandica is problematical; although belonging to the same order as
the Terebratellida, it is a member of a different superfamily. However, in PCA (Text-fig. 3)
L. neozealandica plots close to the other data, and in the cluster analysis (Text-fig. 5) it clusters with
the genus Neothyris. SDS-PAGE shows that L. neozealandica contains a different number of
protein bands compared to the Terebratellida, and at different sizes, indicating that the close
relationship of the samples in the analysis is an artefact of the data, and that similar relative
concentrations of amino acids are present in different proteins. The amino acid composition of these
proteins of different size has produced apparent relationships which do not exist. This is the major
limitation of this technique of using comparisons of bulk amino acid compositions to relate
proteins, although, as this study has shown, the use of an additional technique such as SDS-PAGE
can assist in the recognition of false taxonomic interpretations.

The data from the amino acid analysis of the intracrystalline molecules from Notosaria nigricans
would be expected to be very much different from the remaining samples. This brachiopod differs
from the remaining samples at the ordinal level, and also contains shell proteins of different
molecular weights than those of the Terebratellida. The difference in the composition of the shell
proteins is shown by the distance between N. nigricans and the other species, both in cluster analysis
and PCA (Text-figs 5 and 3), which accurately record the differences between the amino acids
present within the species. This shows similarity with the immunological investigation of Endo
(1992), where antibodies produced in response to shell macromolecules from N. nigricans showed
no reaction with any other brachiopod shell protein, indicating large differences in the composition
and sequence of the proteins.

This study has shown that the separation of Recent samples is possible to the specific level in
brachiopods using only the proteins and amino acids extracted from within the shell, i.e. using a
technique with only objective methodology. The technique is rapid and requires very small
quantities of material, making it available for the study of material in short supply, which is not
available in quantities sufficient for protein sequencing. The relationships between the samples are
easily seen by the graphical presentation of the PCA data.

Anomalies between partial sequence and amino acid analyses

The high proportions of Gly found within the bulk amino acid composition of the samples contrast
with the partial sequence data so far derived for brachiopod intracrystalline proteins (Curry et al.
1991u; Cusack et al. 1992) which include only moderate amounts of Gly. This could be due to
different sample preparation methods: the proteins for sequencing undergo concentration by
filtration which removes most compounds with a molecular weight of less than 10,000 (10 kDa),
although this also depends on the conformation of the protein. Polypeptides with a molecular
weight of less than this may contain larger abundances of Gly.

As discussed previously, this method has significant advantages over methods of protein
sequencing in terms of rapidity and required sample size. An alternative approach would be to
separate proteins from Recent brachiopods using SDS-PAGE and to obtain the amino acid
composition of the purified protein. Taxonomic conclusions could then be reached by application
of the theories of Cornish-Bowden (1979, 1983). This approach would increase the level of
information gained from such molecules and has indeed been attempted by the authors. However,
consistent results have not yet been achieved due to the interaction of the support membrane with
the automated hydrolysis of the analysis system. This approach also has the same disadvantage as
protein sequencing in requiring large amounts of shell powder for analysis, in contrast to the
method described here.

Implications for the study of fossil molecules

The information contained within intracrystalline molecules was sealed by the mineralization of the
shell carbonate; hence both surviving molecules and their degradation products will be trapped in
a contained microenvironment. This contrasts with the previously published data for the amino acid

894

PALAEONTOLOGY, VOLUME 36

T. sanguinea
T, hairrakiensis
Waltonia
Gvrothvris
N, parva
N, lenticularis
Liothvrella
Notosaria

text-fig. 6. Dendrogram showing morphological relationships from the Treatise (after Williams et al. 1965).

composition of the intercrystalline shell matrix (e.g. Jope 1967a, 19876; Konesnikov and
Prosorovskaya 1986). These molecules are in sites where they may be easily degraded and leached
by percolating ground fluids or consumed by micro-organisms, which would alter the molecular
record. In a study of the Recent articulate brachiopod Terebratulina retusa, Collins (1986) found
that the intercrystalline protein between the secondary shell fibres of brachiopods is degraded in less
than one year, and that after this time the fibres behave as individual units, rather than being part
of a layer, and may separate from the remainder of the shell. This corresponds to the decay of
intercrystalline molecules and possible leaching of the decay products (Abelson 1955). These studies
indicate that intracrystalline molecules from fossils should be a much more reliable source of
historical information. In any study of fossil molecules, however, study of Recent samples is
required as a starting point. The study of the Recent samples has also shown that it is possible to
assess taxonomic relationships by the use of multivariate statistical analysis of amino acid
compositions, regardless of the numbers of different proteins present.

The degradation products of the fossil equivalents of these Recent intracrystalline molecules will
remain trapped within fossil shells until the inorganic phase is dissolved. The trapped molecules will
record molecular taxonomic differences in the fossil record. This is in contrast to the changes
between Recent and fossil molecules in the intercrystalline matrix of the shells, which may be
diagenetic rather than genetic. The study described here is an essential starting point for the study
of fossilized amino acids and proteins from brachiopod shells.

CONCLUSIONS

Taxonomic analysis of brachiopods using the weight percentage of intracrystalline shell amino acids
reinforces the morphological taxonomy. Using the classification of the brachiopods from the
Treatise (Williams et al. 1965), the data are divided into two orders, the Order Rhynchonellida,
represented by Notosaria , and the Order Terebratulida, which encompasses all the remaining
brachiopods in this study. Within the Order Terebratulida, Liothyrella is included within the
Suborder Terebratulidina, whereas the remainder are included in Suborder Terebratellida. At the
subfamily level, Neothyris is included in the Subfamily Neothyridinae, and the remainder in the
Subfamily Terebratellinae. From these data, an approximate dendrogram may be produced
(Text-fig 6) to show suggested relationships. It would be expected that Liothyrella and both species of
Neothyris would plot away from the Terebratellinae.

WALTON ET AL.\ BRACHIOPOD AMINO ACIDS

895

In Text-figure 5, however, Liothyrella plots closest to Neothyris, and the species of Terebratella
do not form a discrete cluster. As would be expected from the data in the brachiopod Treatise ,
Gyrothyris and Waltonia plot together as members of the same subfamily. SDS-PAGE indicates that
the relationship between the genus Neothyris and L. neozealandica is false, and is due to the occurrence
of similar ratios of amino acid rather than any true genetic similarity.

Acknowledgements. This work was carried out during the tenure of a studentship from the Natural
Environment Research Council (GT4/89/GS/42) to D.W. at the Department of Geology and Applied
Geology, University of Glasgow, which is gratefully acknowledged. M.C. and G. B.C. are supported by the
Royal Society. We wish to thank Sandra Miller for technical assistance given during sample preparation. We
acknowledge the gift of samples for destructive analysis from Drs D. E. Lee (University of Otago), C. W.
Thayer (University of Pennsylvania) and M. C. Rhodes (Academy of Natural Sciences, Philadelphia), and
assistance in collection from Chris Spiers (Portobello Marine Laboratory, University of Otago) and the New
Zealand Department of Conservation (Stewart Island).

REFERENCES

abelson, p. h. 1955. Organic constituents of fossils. Carnegie Institute of Washington Yearbook. 54, 107-109.
allan, R. s. 1931. On a new Recent Terebratella from northern New Zealand. New Zealand Journal of Science
and Technology, 12, 385-387.

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.

— kvick, a., leiserowitz, L., nelson, M. and weiner, s. 1990. Intercalation of sea urchin proteins in
calcite: study of a crystalline composite material. Science, 250. 664—667.
collins, M. J. 1986. Post mortality strength loss in shells of the Recent articulate brachiopod Terebratulina
retusa from the West Coast of Scotland. 209-219. In racheboeuf, p. r. and emig, c. c. (eds). Les
brachiopodes fossiles et actuels. Biostratigraphie du Paleozo'ique 4. Universite de Bretagne Occidentale, Brest,
500 pp.

curry, G. b., quinn, R., muyzer, G., ZOMERDIJK, t. and westbroek, p. 1988. Sero-taxonomy of skeletal

macromolecules in living terebratulid brachiopods. Historical Biology, 1, 207-224.

— muyzer, G., curry, G. b., Sandberg, p. and westbroek, p. 1991. Macromolecules in brachiopod shells:
characterization and diagenesis. Lethaia, 24, 387-397.

cooper, G. a. 1982. New brachiopods from the Southern Hemisphere and Cryptopora from Oregon (Recent).
Smithsonian Contributions to Paleobiology, 41, 1-43.

cornish-bowden, a. 1979. How reliably do amino acid composition comparisons predict sequence similarities
between proteins? Journal of Theoretical Biology, 76. 369-386.

— 1983. Relating proteins by amino acid composition. Methods in Enzymologv, 91, 60-75.

crick, F. H. c., barnet, h. l., brenner, s. and watts-tobin, r. j. 1961. General nature of the genetic code for
proteins. Nature, 192, 1227-1232.

curry, g. b., ansell, a. d., James, m. and peck, l. 1989. Physiological constraints on living and fossil
brachiopods. Transactions of the Royal Society of Edinburgh, Earth Sciences, 80, 255-262.

-Cusack, M., endo, k., walton, d. and quinn, R. 1991a. Intracrystalline molecules from brachiopod
shells. 35-40. In suga, s. and nakahara, h. (eds). Mechanisms and phytogeny of mineralization in biological
systems. Springer-Verlag, Tokyo, 517 pp.

- — walton, D., endo, K., clegg, h., Abbott, G. and Armstrong, h. 19916. Biogeochemistry of
brachiopod intracrystalline molecules. Philosophical Transactions of the Royal Society of London , Series B,
333, 359-366.

— QUINN, R., COLLINS, M. J., ENDO, K., EWING, S., MUYZER, G. and WESTBROEK, P. 1991c. Immunological
responses from brachiopod skeletal macromolecules; a new technique for assessing taxonomic relationships
using shells. Lethaia, 24, 399^407.

cusack, m., curry, G. B., clegg, h. and Abbott, G. 1992. An intracrystalline chromoprotein from red
brachiopod shells: implications for the process of biomineralisation. Comparative Biochemistry and
Physiology, 102B, 93-95.

degens, E. T., spencer, d. w. and Parker, R. H. 1967. Paleobiology of molluscan shell proteins. Comparative
Biochemistry and Physiology, 20, 553-579.

desha yes, g. p. 1839. Terebratules de la Nouvelle Zelande. Revue Zoologicpie, par la Societe Cuvierienne, 2, 359.

896

PALAEONTOLOGY, VOLUME 36

dupont, d. r., keim, p. s., chui, a., bello, R., bozzini, M., and Wilson, k, j. 1989. A comprehensive approach
to amino acid analysis. 284—294. In hugli, t. e. (ed. ). Techniques in protein chemistry. Academic Press, New
York, 612 pp.

endo, k. 1992. Molecular systematics of Recent and Pleistocene brachiopods. Unpublished Ph.D. thesis,
University of Glasgow.

haugen, j.-e., sejrup, h.-p. and vogt, n. b. 1989. Chemotaxonomy of Quaternary benthic foraminifera using
amino acids. Journal of Foraminifera! Research , 19, 38-51.

heinrikson, r. l. and Meredith, s. c. 1984. Amino acid analysis by reverse-phase high-performance liquid
chromatography: precolumn derivatization with phenylisothiocyanate. Analytical Biochemistry , 136, 65-74.

jope, m. 1967u. The protein of brachiopod shell. I. Amino acid composition and implied protein taxonomy.
Comparative Biochemistry and Physiology , 20, 593-600.

19676. The protein of brachiopod shell. II. Shell protein from fossil articulates: amino acid composition.
Comparative Biochemistry and Physiology , 20, 601-605.

1977. Brachiopod shell proteins: their functions and taxonomic significance. American Zoologist , 17,
133-140.

joysey, K. A. 1988. The use of amino acid sequences in phylogenetic analysis. 34 — 48. In runnegar, b. and
schopf, j. w. (eds). Molecular evolution and the fossil record. Short Courses in Paleontology , 1
Paleontological Society, Knoxville, 167 pp.

king, k. jr. and hare, p. e. 1972. Amino acid composition of the test as a taxonomic character for living and
fossil planktonic foraminifera. Micropalaeontology , 18, 285-293.

Kolesnikov, c. M. and prosorovskaya, e. l. 1986. Biochemical investigations of Jurassic and Recent
brachiopod shells. 113-119. In racheboeuf, p. r. and emig, c. c. (eds). Les brachiopodes fossiles et actuels.
Biostratigraphie du Paleozoi'que 4. Umversite de Bretagne Occidental, Brest, 500 pp.

LEACH, w. E. 1814. The Zoological Miscellany; being descriptions of new, or interesting animals. Volume I.
E. Nodder and Son, London, 144 pp.

logan, G. a., collins, m. J. and eglinton, g. 1991. Preservation of organic biomolecules. 1-24. In allison,
p. a. and briggs, D. E. G. (eds). Taphonomy: releasing the data locked in the fossil record. Plenum Press, New
York, 553 pp.

lowenstam, H. a. 1981. Minerals formed by organisms. Science, 211, 1126-1131.

schagger, h. and von jagow, G 1987. Tricine-sodium dodecyl sulfate-polyacrylamide gel electrophoresis for
the separation of proteins in the range from 1 to 100 kDa. Analytical Biochemistry, 166, 368-379.

sneath, p. h. a. and sokal, r. r. 1973. Numerical taxonomy. W. H. Freeman and Company, San Francisco,
573 pp.

sowerby, G. b. 1846. Descriptions of thirteen new species of brachiopods. Proceedings of the Zoological Society
of London , 14, 91-95.

Thomson, J. a. 1918. Brachiopoda. Scientific Reports of the Australian Antarctic Expedition, 1911-1914. Series
C - Zoology and Botany, 4, 1-76.

walton, d. and curry, g. b. 1991. Amino acids from fossils, facies and fingers. Palaeontology, 34, 851-858.

williams, a. and 18 others 1965. Brachiopoda. H1-H197. In moore, r. c. (ed). Treatise on invertebrate
paleontology. Geological Society of America and University of Kansas Press, Boulder, Colorado and
Lawrence, Kansas, 927 pp.

— 1968. Evolution of the shell structure of articulate brachiopods. Special Papers in Palaeontology, 2, 1-55.

DEREK WALTON

Department of Earth Sciences
University of Derby
Kedleston Road, Derby
DE22 1GB, UK

MAGGIE CUSACK
GORDON B. CURRY

Department of Geology and Applied Geology
University of Glasgow
Lilybank Gardens, Glasgow
G12 8QQ, UK

Typescript received 10 December 1992
Revised typescript received 11 February 1993

THE PROSAUROPOD DINOSAUR AZENDOHSAURUS
LAAROUSSII FROM THE UPPER TRIASSIC OF

MOROCCO

by FRAN^OIS-XAVIER GAUFFRE

Abstract. Azendohsaurus laaroussii was described by Dutuit as an ornithischian dinosaur on the basis of a
dentary fragment and two isolated teeth from the Upper Triassic Argana Formation of Morocco. It was
subsequently suggested to be a mixture of ornithischian and saurischian remains. Study of further material
demonstrates that all the A. laaroussii material represents a single valid taxon within the Prosauropoda.
However, there is insufficient taxonomic information to place it more precisely and it is considered as
Prosauropoda incertae sedis. A. laaroussii derives from the Paleorhinus biochron and is hence Carnian in age
and one of the earliest known dinosaurs. Some apomorphic characters attributed to the Ornithischia by Sereno
and to the Sauropodomorpha by Gauthier are shown to be invalid. A new synapomorphy of the Prosauropoda
is proposed, based on the construction of the maxillary.

A characteristic vertebrate fauna from the important transitional period of the Upper
Triassic has been collected from numerous outcrops of the Argana Formation, between Agadir and
Marrakech in the western High Atlas mountains of Morocco (Text-fig. 1). These outcrops were
discovered and excavated by Dutuit between 1962 and 1969. The material described in this work
originated in an outcrop discovered in 1965 and numbered XVI by Dutuit (1976), which is situated
in the northern part of the Argana Formation. Lithostratigraphic limits were established by
Tixeront (1973) and the outcrops are subdivided into eight levels from the Permian to the Upper
Triassic (Text-fig. 2). Outcrop XVI was found at the base of the t5 level, which consists of
sandstones and mudstones. Biostratigraphical dating was based on the faunal association of
metoposaurs and phytosaurs typical of this period. Outcrop XVI has yielded only one well-
preserved specimen, namely a nearly complete phytosaur skull, Paleorhinus magnoculus Dutuit,
1977; together with some fragmentary remains of dicynodonts: Moghreberia nmachouensis Dutuit,
1988, Azarifeneria barrati Dutuit, 1989a and Azarifeneria robustus Dutuit, 19896; some reptiles
including procolophonids (Dutuit, pers. comm.); temnospondyl remains and many tooth-bearing
bones and isolated teeth.

Azendohsaurus laaroussii is based upon fragmentary cranial material. Dutuit (1972) described
only a fragment of dentary and two isolated teeth. He compared this new taxon only with
ornithischians such as Fabrosaurus , Scelidosaurus and the Heterodontosauridae, and on the basis
of this comparison established a new genus and species of ornithischian. This phylogenetic position
was immediately challenged by Thulborn (1973, 1974) and Bonaparte (1976), who argued that
A. laaroussii was a prosauropod dinosaur. Dutuit later revised his position, considering A. laaroussii
to be firstly a ‘pre-ornithischian’ (Dutuit 1981), and later to be a prosauropod (Dutuit 1983).
Galton (1984, 1985a, 1990) and Weishampel (1990) expressed the opinion that the material was
from two distinct taxa; the fragment of dentary and one isolated tooth belonging to a prosauropod,
and the other isolated tooth belonging to an ornithischian. Additional material, comprising further
tooth-bearing bones and teeth, is used to establish the phylogenetic position of Azendohsaurus
proposed in this study.

The whole material described in this work is deposited in the collections of the Department of

IPalaeontology, Vol. 36, Part 4, 1993, pp. 897-908, 1 pl.|

© The Palaeontological Association

898

PALAEONTOLOGY, VOLUME 36

MARRAKECH

• Essaouira

• Chichaoua

Essaouira and Haouz plain
Argana Formation

Ourika

Mesozoic coastal
tablelands

• Imi n'Tanoute

Palaeozoic massifs

Argana

Western' High Atlas

(d '

Sous plain

Agadir

Taroudannt ^ Anti-Atlas

15 km

i 1

text-fig. 1 . Permian and Triassic outcrops from Argana and Ourika in the High Atlas mountains of Morocco

(modified from Dutuit 1976). +, Outcrop XVI.

Lower Lias

Discordance — —

Upper series

t8: saliferous silts and sands and basalts with doleritic structure
t7: Argana sandstones
t6: pink sandstones

Discordance — — - -

t5: sandstones and red clays

t4: silts and sands and sandstones with thin level of conglomerate
t3: conglomerate

- — Discordance - — —

t2: red sandstones and conglomerate at the top
tl: conglomerate

Discordance

Palaeozoic

text-fig. 2. Lithostratigraphic section of the Permian and Triassic Argana Formation (from Tixeront 1973).

GAUFFRE: PROSAUROPOD FROM MOROCCO

899

Palaeontology, Museum National d’Histoire Naturelle (MNHN), Paris. The acronym MTD
(Maroc Trias Dutuit) is a temporary museum reference pending definitive cataloguing.

STRATIGRAPHY

Biron and Dutuit (1981) ascribed a Carnian age to the t5 level on the basis of similarities observed
between t5 of the Argana Formation and F5 of the Ourika Formation (Text-fig. 1). This F5 level
was dated directly by palynostratigraphy (Cousminer and Manspeizer 1976; Biron and Courtinat
1982). However, the t5/F5 similarities are of uncertain significance for the following reasons.

(1) The lithology is not strictly identical between t5 and F5, and Biron (1982) admitted that the
formations were not necessarily synchronous.

(2) The osteological similarities were based only on very fragmentary material restricted to a 1 m
thick calcareous conglomerate, situated at the base of F5 (Biron 1982).

(3) The lower part of the F5 has produced trackways morphologically similar to those from t5,
but such ichnological data is of limited value in stratigraphical correlations.

However, one argument is more convincing for assigning a Carnian age to Azendohsaurus. Hunt
and Lucas (1991) have argued for the existence of a Paleorhinus biochron, which can be recognized
across much of Late Triassic Pangaea and which can be dated as Carnian in several localities. Since
Paleorhinus occurs in the outcrop XVI, Azendohsaurus laaroussii is considered to be Carnian in age.

SYSTEMATIC PALAEONTOLOGY

Superorder dinosauria Owen, 1842
Order saurischia Seeley, 1888
Suborder sauropodomorpha Huene, 1932
Infraorder prosauropoda Huene, 1920
Family Incertae sedis
Genus azendohsaurus Dutuit, 1972

Type species. Azendohsaurus laaroussii Dutuit, 1972 from the Upper Triassic of Morocco.

Azendohsaurus laaroussii Dutuit, 1972
Plate 1 ; Text-figs 3-7

Holotype. MTD XVI 1 : a fragment of mandible.

Paratypes. MNHN-ALM 424-5 (MTD XVI 2) and MNHN-ALM 424-4 (MTD XVI 3): two isolated teeth
(Text-fig. 7a-b).

Type locality. Outcrop XVI, 1-5 km east of Azendoh, ‘Argana corridor', Morocco, coordinates in Lambert
system: 161 .4/457.5.

Type horizon. Argana Formation, base of t5 level; Paleorhinus biochron (Hunt and Lucas 1991), Carnian,
Upper Triassic.

Principal referred material. MNHN-ALM 351, a nearly complete left dentary visible in external aspect only (PI.
1, fig. 1 ; Text-fig. 3). MNHN-ALM 365-20, a right dentary visible in lingual aspect only and missing the ends
(Text-figs 4, 6). Other dentary fragments include MNHN-ALM 353, ALM 365-17 and ALM 365-18. MNHN-
ALM 355-3, left maxillary visible in lingual aspect only, and with the posterior region missing and with a
damaged dorsal region of the dorsal process (PI. 1, fig. 2; Text-fig. 5). MNHN-ALM 365-21, isolated right
maxillary.

In total, there are fourteen fragments of dentary, including the above, but only four specimens can be
examined from both sides. Eight maxillary fragments have been recognized but only one is visible from both

900

PALAEONTOLOGY, VOLUME 36

sides. Although all jaw fragments bear teeth, the entire dentary row is not preserved on any one specimen. Nine
isolated teeth are available.

Diagnosis. A prosauropod dinosaur with the following autapomorphies: neck between crown and
root of tooth consistently present; antero-posterior expansion of crown always beginning at its
base; prominent longitudinal keel on medial face of maxillary; fossa posterior to dorsal process on
medial face of maxillary; entire anterior rim of dentary inclining ventrally, forming angle with
dorsal rim of dentary. Other significant characters are: a series of foramina on the medial face of
the maxillary, each situated at the base of a tooth; truncate cone-shaped anterior maxillary process
with a wide base.

Description

Dentary (PI. 1, fig. 1; Text-figs 3^f, 6). Although these elements are incomplete, the anterior end of the
coronoid protuberance (not a true coronoid process) can be seen in some fragments on the posterodorsal side.
The dentaries bear no indication of the external mandibular fenestra. In most fragments, only the outer face
is visible, except for four which show the inner face and, in particular, the mandibular symphysis (ALM 353,
ALM 365-17, ALM 365-18). The symphysial surface is very irregular and difficult to delimit, but there is no
indication of a predentary. The outer surfaces are entirely flat and smooth, with only some foramina situated
near the anterior (ALM 351, ALM 353) or along the median portion of the dorsal rim (ALM 365-20). All the
dentaries bear teeth, usually between six and eleven. However, teeth have clearly been lost in all specimens and
the total number of dentary teeth may have reached seventeen.

Maxillary (PI. 1, fig. 2; Text-fig. 5). A major portion of the posterior region of the maxillaries is always missing.
The available fragments represent two separate regions, the body of the bone and the dorsal process. An
anterior process also is present in the continuation of the main body of the bone. The external face of the
maxillary is flat and smooth, while the internal face displays a prominent longitudinal keel between the tooth
row and the dorsal process, along the entire length of the bone. It also bears a narrow, deep longitudinal fossa
situated above the keel, posterior to the dorsal process. A series of ovoid foramina are situated between the
tooth row and the longitudinal keel, each one at the base of a tooth. The total number of maxillary teeth is
estimated at fifteen to sixteen.

Dentition (Text-fig. 7a-b). The teeth share the following characteristics: a spatulate crown with conical
denticles which form an angle of 45° with the rim of the crown, the presence of a neck between the crown and
the root, the overlap of the teeth (the anterior rim of each tooth being overlapped laterally by the posterior rim
of the preceding tooth), and a thecodont dentition. Most of the teeth can be categorized as one of two types:

Type 1 : 8-12 lateral denticles, height/length ratio of 1-5-2, medio-longitudinal keel or ridge present on both
faces, edges asymmetrical with different denticle count on anterior and posterior edges, cingula present on both
faces. Most but not all dentary teeth are of this type.

Type 2: 4-6 lateral denticles, height/length ratio 1-T4, medio-longitudinal keels or ridges absent, edges
symmetrical with same denticle count on anterior and posterior edges, cingulum present only on outer face.
Most but not all maxillary teeth are of this type. There are other teeth which do not conform to either of these
types, but show a significant variability of these features with virtually all combinations between the two.

The existence of two different dental morphologies could mean that the fragments derive from two different
taxa, but three arguments contradict this. Firstly, on several fragments (ALM 365-20 (Text-figs 4, 6), ALM
352-1, ALM 356, ALM 361) it can be clearly seen that the two tooth types are associated, sometimes with
intermediate types. Secondly, several genera of prosauropod dinosaur show significant variation of the dental
morphology (e.g. Sellosaurus (Galton 19856), Plateosaurus (Galton 19856, 1985c) and Massospondylus (Gow,
et al. 1990)). Thirdly, the existence of replacement teeth demonstrates that an individual animal can have

EXPLANATION OF PLATE 1

Figs 1-2. Azendohsaarus laaroussii Dutuit; t5 level; Azendoh, Morocco. 1, MNHN-ALM 351 ; left dentary in
external view, x 1-5. 2, MNHN-ALM 355-3; left maxillary in internal view, x2-5.

PLATE 1

GAUFFRE, Azendohsciurus

902

PALAEONTOLOGY, VOLUME 36

text-fig. 3. Azendohsaurus laaroussii Dutuit; t5 level; Azendoh, Morocco. MNHN-ALM 351, left dentary in

external view, x 1. Scale bar = 10 mm.

foramina

text-fig. 4. Azendohsaurus laaroussii Dutuit; t5 level; Azendoh, Morocco. MNHN-ALM 365-20, right

dentary in external view, x 1-3. Scale bar = 10 mm.

text-fig. 5. Azendohsaurus laaroussii Dutuit; t5 level; Azendoh, Morocco. MNHN-ALM 355-3, left maxillary

in internal view, x 2. Scale bar = 10 mm.

GAUFFRE: PROSAUROPOD FROM MOROCCO

903

text-fig. 6. Azendohsaurus laaroussii Dutuit; t5 level; Azendoh, Morocco. MNHN-ALM 365-20, right

dentary in external view, x2. Scale bar = 10 mm.

different tooth morphologies present on the tooth-bearing bones (Gow 1981) and, for example, the number of
denticles changes between successive dental generations (Gow et al. 1990).

In conclusion, all the fragments attributed to A. laaroussii can be assigned to a single genus and species.

SYSTEMATIC POSITION

In order to place Azendohsaurus laaroussii systematically, it is necessary to search for derived
characters or apomorphies defining dinosaurian taxa, based on the teeth and tooth-bearing bones.
Benton (1984, 1990), Gauthier (1986), Sereno (1986), Novas (1992), Sereno, Forster, Rogers and
Monetta (1993) have provided cladistic analyses leading to definition of the Dinosauria and its
major subgroups. However, they found few fundamental diagnostic characters based on the skull.
Novas (1992) did not list any cranial apomorphy for the Dinosauria. Gauthier (1986) and Benton
(1990) mentioned only the elongate vomers reaching caudally at least to the level of the antorbital
fenestra, and Sereno et al. (1993) mentioned five cranial apomorphies, none of which concern the
maxillary and dentary. These characters thus cannot be used here, and consequently no character
can be used to assign A. laaroussii to the Dinosauria as defined at its basal node. However, a general
comparison with the non-dinosaurian groups of this period suggests that A. laaroussii belongs to
this group, and the morphology of the teeth closely resembles that of prosauropods and
ornithischians.

A limitation on further analysis is that the Upper Triassic and Early Jurassic Ornithischia are
known from only a few well-characterized taxa, namely Fabrosaurus (Lesothosaurus is regarded as
a junior synonym following Thulborn 1992), Scelidosaurus and Heterodontosaurus , most other
genera being based on more fragmentary remains.

Defining apomorphies utilizing the teeth and tooth-bearing bones are claimed for the Ornithischia
(six listed by Sereno 1984, 1986) and the Sauropodomorpha (two listed by Gauthier 1986), but none
are reported for the Prosauropoda (Sereno 1989; Galton 1990). These and other characters (Paul
1984, Galton 1985a) are listed and discussed below.

904

PALAEONTOLOGY, VOLUME 36

text-fig. 7. Azendohsaurus laaroussii Dutuit; t5 level; Azendoh, Morocco. Paratypes. (a) MNHN-ALM 424-
4, isolated tooth in external aspect, (b) MNHN-ALM 424-5, isolated tooth. Both x 6. Scale bars = 5 mm.

Dental characters

(1) Recurvature absent in maxillary and dentary teeth. An ornithischian apomorphy according to
Sereno (1986), but a similar condition occurs in prosauropods.

(2) Overlap of adjacent crowns in maxillary and dentary teeth. An ornithischian apomorphy
according to Sereno (1986), but a similar condition occurs in prosauropods.

(3) Well-defined neck separating crown from root. An ornithischian apomorphy according to
Sereno (1986), but a similar condition occurs in some prosauropods ( Sellosaurus Galton 19856,
Thecodontosaurus Kermack 1984, Massospondylus Gow et al. 1990).

(4) Low and triangular crowns in lateral view. An ornithischian apomorphy according to Sereno
(1986). A high crown is the widespread condition in prosauropods, but Lufengosaurus (Young 1947)
bears low triangular crowns, and other prosauropods show quite fundamental variation in crown
size on the tooth rows (Kermack 1984; Galton 1985a, 19856; Gow et al. 1990). In contrast, the

GAUFFRE: PROSAUROPOD FROM MOROCCO

905

ornithischians Heterodontosaurus (Charig and Crompton 1974) and Pisanosaurus (Bonaparte 1976)
have high crowns which are rather rectangular in shape.

(5) Maximum tooth size attained near the central or posterior region of the maxillary and
dentary tooth rows. An ornithischian apomorphy according to Sereno (1986), and not contradicted
here.

(6) Buccal emargination on maxillary (Genasauria apomorphy, Sereno 1986), slight emargin-
ation is visible in Fabrosawus (Sereno 1991).

(7) Lanceolate teeth with coarsely serrated crowns. A sauropodomorph apomorphy according to
Gauthier (1986), but invalid as most sauropods lack serrated crowns (see also character 4).

(8) Maximum tooth size at a quarter to a third of the length from the anterior of the maxillary
and dentary rows. A sauropodomorph apomomorphy according to Gauthier (1986), and not
contradicted here.

According to Galton (1985a) and Sereno (1986, 1991), the following dental characters are distinct
in ornithischians and prosauropods.

(9) Wear surfaces in most ornithischians (except Fabrosawus ), but not in most prosauropods
(except yunnanosaurids). All ornithischians show wear surfaces, including Fabrosawus (Sereno
1991), and contrary to Galton (1990), I consider that only yunnanosaurids show clear wear on
surfaces with a very typical dental morphology, namely an asymmetrical crown with an apex on the
inner side.

(10) Symmetry of the tooth edges in ornithischians but not in prosauropods. However, in
Fabrosawus there is slight asymmetry of the anterior and posterior edges, and a different number
of denticles on these edges in all teeth.

(11) Number of denticles, low in ornithischians and high in prosauropods. However, the number
of denticles in some prosauropods is small: five to seven on some Massospondylus teeth (Crompton
and Attridge 1986), and the number may vary within and between genera (Gow et al. 1990).

(12) A cingulum is present in ornithischians but is relatively uncommon in prosauropods. There
is little published information available on the cingulum, and it is not discussed further.

Osteological characters

(13) Ventral margin of antorbital fossa parallels maxillary tooth row. An ornithischian apomorphy
according to Sereno (1986), but a similar condition occurs in the prosauropods Lufengosaurus ,
Massospondylus , Plateosaurus and Anchisaurus (Galton 1990).

(14) A predentary is an ornithischian apomorphy (Sereno 1986).

(15) A spout-shaped mandibular symphysis is an apomorphy for Genasauria (Sereno 1986). It
is present in Fabrosawus (Sereno 1991).

(16) In prosauropods, the dentary is curved downwards (Paul 1984). However, ventral curvature
is absent in Anchisaurus (Galton 1976), juvenile Massospondylus and Tawasaurus (Sereno 1991) and
weakly developed in Co/oradisaurus, Thecodontosaurus juvenile and Sellosaurus.

(17) Numerous small foramina in prosauropods, in contrast to few large foramina in
ornithischians (Paul 1984). In the ornithischian Fabrosawus there are numerous small foramina
(eight to nine on the maxillary) as found in some prosauropods ( Massospondylus Gow et al. 1990).
Conversely, there can be low numbers of foramina in other prosauropod genera (e.g. Plateosaurus
Galton 19856, 1985c).

Personal observation permits the addition of a further character which I propose as an
apomorphy for the Prosauropoda, using respectively sauropods and ornithischians as outgroups.

(18) The dorsal process of the maxillary is fully individualized and its base is situated in the
anterior half of the bone, usually in the anterior third, but never in the posterior half.

Of the above eighteen characters, eleven (1-4, 7, 10-13, 16-17) are sufficiently ambiguous in their
distribution that they do not permit Azendohsaurus laaroussii to be placed in the dinosaur
phylogeny, and one (15) is not visible in Azendohsaurus. However, six characters appear to stand

906

PALAEONTOLOGY, VOLUME 36

as valid apomorphies defining major groups, namely character 8 defining the Sauropodomorpha,
character 18 defining the Prosauropoda, characters 5, 6 and 14 defining the Ornithischia, and
character 9 defining the Ornithischia but convergently present in the Yunnanosauridae.

As Azendohsaurus possesses characters 8 and 18, and lacks characters 5, 6, 9 and 14, it can be
argued to be a non-yunnanosaurid prosauropod dinosaur.

CONCLUSION

The material of Azendohsaurus laaroussii is consistent with attribution to a single genus and species
of dinosaur. It possesses the following taxonomically significant characters: the dorsal process of
the maxillary is well individualized with its base situated in the anterior half of the bone
(prosauropod); there are no wear surfaces on the teeth (non-yunnanosaurid prosauropod); the
maximum height of the dentition is reached between the anterior quarter and the anterior third of
the tooth row (sauropodomorph); there is no evidence of a predentary (non-ornithischian); there
is no buccal emargination on the maxillary (non-ornithischian). In conclusion, Azendohsaurus
laaroussii is a member of the Prosauropoda, but is not determinate within that infraorder, beyond
the fact that it is not a yunnanosaurid. It appears to be of Carnian age and is thus not only one of
the earliest prosauropods but also one of the earliest dinosaurs.

Acknowledgements. I thank Philippe Janvier for many helpful suggestions, Andrew Milner for reading the
manuscript and especially for correcting my spelling and punctuation, and Paulo Brito and Guido Dingerkus
for assisting with the translation.

REFERENCES

benton, M. j. 1984. Fossil reptiles of the German Late Triassic and the origin of dinosaurs. 13-18. In reif,
w. e. and westphal, F. (eds). Third symposium on Mesozoic terrestrial ecosystems. Attempto Verlag,
Tubingen, 259 pp.

— 1990. Origin and interrelationships of dinosaurs. 1 1-30. In weishampel, d. b., dodson, p. and osmolska,
h. (eds). The Dinosauria. University of California Press, Berkeley, Los Angeles, Oxford, 733 pp.

biron, p.-e. 1982. Le Permo-Trias de la region de FOurika (Haut- Atlas de Marrakech , Maroc). These de 3eme
cycle de geologie appliquee. Universite scientifique et medicale de Grenoble, 2 tomes. 170 pp.

— and courtinat, b. 1982. Contribution palynologique a la connaissance du Trias du Haut-Atlas de
Marrakech, Maroc. Geobios, 15, 231-235.

— and dutuit, J. M. 1981. Figurations sedimentaires et traces d’activite au sol dans le Trias de la formation
d’Argana et de FOurika (Maroc). Bulletin du Museum National d Histoire Naturelle , 4eme serie, 3 ( section
C, 4), 339^127.

bonaparte, j. F. 1976. Pisanosaurus mertii Casamiquela and the origin of the Ornithischia. Journal of
Paleontology , 50, 808-820.

charig, a. J. and crompton, a. w. 1974. The alleged synonymy of Lycorhinus and Heterodontosaurus. Annals
of the South African Museum , 64, 167-189.

chatterjee, s. 1984. A new ornithischian dinosaur from the Triassic of North America. Naturwissenschaften,
71, 630-631.

cousminer, h. l. and manspeizer, w. 1976. Triassic pollen date Moroccan High Atlas and the incipient rifting
of Pangea as Middle Carnian. Science, 191, 943-945.

crompton, a. w. and attridge, J. 1986. Masticatory apparatus of the larger herbivores during the Late
Triassic and early Jurassic times. 223-236. In padian, k.. (ed.). The beginning of the age of dinosaurs.
Cambridge University Press, 378 pp.

dutuit, J. M. 1972. Decouverte d'un dinosaure ornithischien dans le Trias superieur de l’Atlas occidental
marocain. Comptes Rendus de F Academie des Sciences, 275 (serie D), 2841-2844.

— 1976. Introduction a l’etude paleontologique du Trias continental marocain. Description des premiers
Stegocephales recueillis dans le couloir d'Argana. Memoire du Museum National d Histoire Naturelle, ( serie
C), 36, 253 pp.

GAUFFRE: PROSAUROPOD FROM MOROCCO

907

— 1977. Paleorhinus magnoculus , Phytosaure du Trias superieur de l’Atlas marocain. Geologie
mediterraneenne, Annales de l’Universite de Provence, tome IV, 3, 255-268.

1988. Osteologie cranienne et ses enseignements, apports geologique et paleoecologique, de Moghreberia
nmachouensis, Dicynodonte (Reptilia, Therapsida) du Trias superieur marocain. Bulletin du Museum
National dHistoire Naturelle, 4eme serie, 10 (3), 227-285.

— 1989u. Azarifeneria barrati , un deuxieme genre de Dicynodonte du Trias superieur marocain. Comptes
Rendus de I'Academie des Sciences, serie II, 309, 303-306.

— 1989 b. Confirmation des affinites entre Trias superieurs marocain et sud-americain: decouverte d'un
troisieme Dicynodonte (Reptilia, Therapsida), Azarifeneria robustus, n. sp., de la formation d'Argana (Atlas
occidental). Comptes Rendus de I'Academie des Sciences, serie II, 309, 1267-1270.

galton, p. M. 1976. Prosauropod dinosaurs (Reptilia: Saurischia) of North America. Postilla , 169, 98 pp.
1985u. Diet of prosauropod dinosaurs from the Late Triassic and Early Jurassic. Lethaia , 18, 105-123.
19856. Cranial anatomy of the prosauropod dinosaur Sellosaurus gracilis, from the Middle
Stubensandstein (Upper Triassic) of Nordwiirttemberg, West Germany. Stuttgarter Beitrdge zur Naturkunde,
118 (B), 1-39.

1985c. Cranial anatomy of the prosauropod dinosaur Plateosaurus from the Knollenmergel (Middle
Keuper, Upper Triassic) of Germany. Geologica et Palaeontologica, 19, 119-159.

1990. Basal Sauropodomorpha-Prosauropoda. 320-344. In weishampel, d. b., dodson, p. and osmolska,
h. (eds). The Dinosauria. University of California Press, Berkeley, Los Angeles, Oxford, 733 pp.

Gauthier, J. 1986. Saurischian monophyly and the origin of birds. 1-55. In padian, k. (ed). The origin of birds
and evolution of flight. California Academy of Sciences, Memoir 8, 98 pp.
gow, c. E. 1981. Taxonomy of the Fabrosauridae (Reptilia, Ornithischia) and the Lesothosaurus myth. South
African Journal of Science, 77, 43.

— kitching, J. w. and raath, m. a. 1990. Skulls of the prosauropod dinosaur Massospondylus carinatus
Owen in the collections of the Bernard Price Institute for palaeontological research. Palaeontologica
Africana, 27. 45-58.

huene, f. von. 1920. Bemerkungen zur systematik und stammesgeschichte einiger reptilien. Zeitschrift fur
Induktive Abstammungs und Vererbungslehre , 24, 162-166.

— 1932. Die fossile reptil-ordnung Saurischia, ihre entwicklung und geschichte. Monographien zur Geologie
und Paleontologie, 4 (ser. 1), 1-361.

kermack, d. 1984. New prosauropod material from South Wales. Zoological Journal of the Linnean Society,

82, 101-117.

novas, F. E. 1992. Phylogenetic relationships of the basal dinosaurs, the Herrerasauridae. Palaeontology, 35,
51-62.

owen, R. 1842. Report on British fossil reptiles. Report of the British Association for the Advancement of
Science, 2, 60-204.

padian, K. 1990. The ornithischian form Revueltosaurus from the petrified forest of Arizona (late Triassic:
Chinle Formation). Journal of Vertebrate Paleontology, 10, 268-269.

Paul, g. s. 1984. The stegosaurian dinosaurs: relics of the prosauropod-ornithischian transition? Journal of
Vertebrate Paleontology, 4, 507-515.

seeley, h. G. 1888. The classification of the Dinosauria. Report of the British Association for the Advancement
of Science, 57, 698-699.

sereno, p. c. 1984. The phylogeny of the Ornithischia, a reappraisal. 219-222. In reif, w. e. and westphal, f.
(eds). Third symposium on Mesozoic terrestrial ecosystems. Attempto Verlag Tubingen, 259 pp.

— 1986. Phylogeny of the bird-hipped dinosaurs (order Ornithischia). National Geographic Research, 2,
234-256.

— 1989. Prosauropod monophyly and basal sauropodomorph phylogeny. Journal of Vertebrate Paleont-
ology, 9 (supplement 3), 38a.

1991. Lesothosaurus, “Fabrosaurids”, and the early evolution of Ornithischia. Journal of Vertebrate
Paleontology, 11, 168-197.

— forster, c. a., Rogers, R. R. and monetta, a. m. 1993. Primitive dinosaur skeleton from Argentina and
the early evolution of Dinosauria. Nature, 361, 64-66.

thulborn, R. a. 1973. Teeth of ornithischian dinosaurs from the Upper Jurassic of Portugal. Servigos
Geologicos de Portugal, Memoria, 22, 89-134.

— 1974. A new heterodontosaurid dinosaur (Reptilia: Ornithischia) from the Upper Triassic Red Beds of
Lesotho. Zoological Journal of the Linnean Society, 55, 151-175.

908

PALAEONTOLOGY, VOLUME 36

thulborn, r. a. 1992, Taxonomic characters of Fabrosaurus australis, an ornithischian dinosaur from the
lower Jurassic of Southern Africa. Geobios , 25, 283-292.
tixeront, m. 1973. Lithostratigraphie et mineralisations cupriferes et uraniferes stratiformes syngenetiques et
familieres des formations detritiques Permo-triassiques du couloir d’Argana, Haut-Atlas occidental (Maroc).
Notes du Service Geologique du Maroc, 33 (249), 147-177.
weishampel, D. B. 1990. Dinosaurian distribution. 63-139. In weishampel, d. b., dodson, p. and osmolska, h.
(eds). The Dinosauria. University of California Press, Berkeley, Los Angeles, Oxford, 733 pp.

— dodson, p. and osmolska, h. 1990. The Dinosauria. University of California Press, Berkeley, Los Angeles,
Oxford, 733 pp.

young, c. c. 1947. On Lufengosaurus magnus Young (sp. nov.) and additional finds of Lufengosaurus huenei
Young. Palaeontologica Sinica, 132 (new series C, 12) 1-53.

FRAN^OIS-XAVIER GAUFFRE

Institut de Paleontologie

Typescript received 12 May 1992 8 rue Buffon

Revised typescript received 17 February 1993 75005, Paris, France

TAXONOMY, EVOLUTION AND
BIOSTRATIGRAPHICAL IMPORTANCE OF THE
LLANDOVERY GRAPTOLITE SPIROGRAPTUS

by DAVID K. LOYDELL, PETR STORCH and MICHAEL J. MELCHIN

Abstract. Four species are assigned to the Llandovery graptolite genus Spirograptus. S. andrewsi from
the upper Aeronian, is the earliest representative and is recorded here for the first time outside Australia.
S. turriculatus and S. guerichi sp. nov. occur in the lower Telychian. S. guerichi is distinguished from
S. turriculatus by its lesser dorso-ventral width, rhabdosome apical angle and whorl diameter. The subspecies
S. turriculatus minor is here considered a junior synonym of S. turriculatus. S. minimus is found only in the
middle Telychian of Russia and is not described here. Spirograptus probably evolved from Stimulograptus.
Evolution within Spirograptus involved initially increased curvature of the sicula, attainment of a greater
dorso-ventral width and modification of the thecae. A new Spirograptus guerichi Biozone is erected to replace
the Spirograptus turriculatus minor Biozone of some authors.

W hether there are one or two species of Spirograptus Gurich, 1908, in the lower Telychian (Upper
Llandovery) has been debated since Boucek (1932) erected a subspecies, considered ancestral to
Spirograptus turriculatus (Barrande, 1850), which he named Monograptus turriculatus mut. minor on
the basis of its smaller size and its fewer spirals.

Some authors consider them synonyms ( Rickards 1 970 ; Hutt 1975; Lenz 1 982 ; Melchin and Lenz
1986). This view was supported, but with the qualification that the greater number of small
specimens occur in older strata, by Bjerreskov (1975) and Loydell (1991). Others (Sennikov 1976;
Melchin 1989) recognized S. turriculatus and S. t. minor as distinct taxa and used S. t. minor as a
biozonal index subspecies for the earliest part of the Telychian; the S', t. minor Biozone was
succeeded by a S. /. turriculatus Biozone. Strachan (1971), Chen (1984), Fu and Song (1986), and
Wang (1978) also recognized minor as a separate taxon, but did not use it as a biozonal index.

Here we demonstrate that there are two early Telychian spirograptid taxa (one of which is a new
species), and that S. turriculatus and S. t. minor (as defined by Boucek 1932) are synonyms. The
differences between S. turriculatus and our new species are illustrated both with material chemically
isolated from limestone nodules, and with flattened specimens.

BIOSTRATIGRAPHICAL DIVISION OF THE EARLY TELYCHIAN

A variety of graptolite biozonal schemes has been used for the early Telychian. Some, however, use
the same index species to denote different intervals: for example, the turriculatus Biozone of Boucek
(1953) is equivalent to only the upper part of the turriculatus Biozone of Rickards (1976, 1989), and
thus some confusion has arisen. (See, for example, Khz 1991, text-fig. 1 who placed the base of the
Telychian at the base of the S. turriculatus Biozone in Bohemia, at a level which does not correlate
with the base of the Telychian elsewhere). Some of the biozonal schemes that have been employed
for the lower Telychian are correlated in Text-figure 1.

SYSTEMATIC PALAEONTOLOGY

The terminology used is mainly that of Bulman (1970). Thecal spacing is expressed in terms of a two
theca repeat distance (2TRD; Howe 1983) with thecae per 10 mm quoted in brackets. Due to

[Palaeontology, Vol. 36, Part 4, 1993, pp. 909-926, 2 pis.]

© The Palaeontological Association

910

PALAEONTOLOGY, VOLUME 36

LOYDELL

1992

LAPWORTH

1878

(Dob's Linn)

JONES

1921

RICKARDS

1976

BOUCEK

1953

MELCHIN

1989

Biozone

Subzone

Biozone

Biozone

Biozone

Subzone

Biozone

Subzone

Biozone

crispus

crispus

crispus

crispus

crispus

turriculatus

carnicus

turriculatus

turriculatus

turriculatus

runcinatus
(= storchi)

turriculatus

proteus

turriculatus

johnsonae

utilis

guerichi

Imnaei

hispanicus

minor

renaudi

maximus

maximus

gemmatus

palmeus

runcinatus

halli

maximus

halli

sedgwickii

sedgwickii

rastrum

sedgwickii

spinigerus

sedgwickii

text-fig. 1 . Correlation of graptolite biozonal and subzonal schemes for the upper Aeronian (St. sedgwickii
and St. halli Biozones) and lower Telychian (S. guerichi-' M crispus Biozones) stages of the Llandovery. The
carnicus subzone will be erected formally by J. Zalasiewicz (pers. comm.).

difficulties resulting from the rhabdosome’s spiral curvature in measuring distance from the sicula
in terms of number of thecae, measurements of dorso-ventral width are given for the mid-points of
successive whorls instead of at specified thecae as is the normal practice. Dorso-ventral widths
quoted are exclusive of thecal spines. The apical angle of the rhabdosome and the rhabdosome’s
diameter at specified whorls were measured only in specimens preserved with the axis of
rhabdosome coiling approximately parallel to bedding, as was the number of whorls in 10 mm from
the sicular aperture. Text-figure 2 illustrates the main features measured. The normal abbreviation
for theca (th.) is used throughout the descriptions.

As there are many problems in taking measurements from spiral rhabdosomes oriented at a
variety of different angles to bedding, figures quoted herein are approximate.

Synonymies include only the most important references; full synonymies of S. guerichi and
S. turriculatus are given in Loydell (1992). Synonymies are annotated with the symbols listed by
Matthews (1973).

Repositories of figured and cited material are abbreviated as follows: BGS, British Geological
Survey, Keyworth; ROM, Royal Ontario Museum, Toronto; NIGP, Nanjing Institute of Geology
and Palaeontology; NMP L, National Museum, Prague; PS, Storch Collection, in the Geological
Survey, Prague; UCWG, Institute of Earth Studies, University College of Wales, Aberystwyth.

Subfamily monograptinae Lapworth, 1873
Genus spirograptus Gurich, 1908
(= Tyrsograptus Obut, 1949)

Type species. By subsequent designation of Bulman (1929, p. 182); Graptolithus turriculatus Barrande, 1850.
p. 56, pi. 4, figs 7-11; from the Llandovery of Bohemia.

Diagnosis. Rhabdosome tightly trochispirally coiled; thecae hooked, in the manner of those of

LOYDELL ET A L : SPIROGRAPTUS

911

text-fig. 2. Measurements taken from
Spirograptus rhabdosomes. Dorso-
ventral width was measured at the mid-
point of each whorl. Whorl diameter was
measured also at whorl five and the
number of whorls in 5 mm from the
sicular aperture was also recorded.
2TRD = two theca repeat distance
(Howe 1983).

Stimulograptus Pribyl and Storch, 1983, and bearing one or two apertural spines. Thecal apertures
may be symmetrical, or asymmetrical resulting from the differential development of the apertural
margin, but without significant torsion. Thecal overlap is negligible. Sicula is small ( < 1-5 mm) and
dorsally curved, in one species at least also exhibiting dextral torsion.

Remarks. Gurich’s (1908, p. 34) original diagnosis for the genus (‘Rhabdosom spiralig eingerollt,
Zellen spitz endigen. Die Sicula ist an der Spitze der Spirale.’) was emended by Mu (1955, p. 10)
as follows: ‘ Monograph in which the rhabdosome is spiral, the thecae are hooked, usually with
apertural spines.’ Mu’s (1955) diagnosis does not differ significantly from ours.

Pribyl (1945) included a large number of species within Spirograptus. These possess a variety of
thecal forms, and belong to a number of different genera. Objections to Pribyl’s (1945) usage of the
genus (Mu 1955; Rickards et al. 1977) and incomplete knowledge of the thecal morphology of
S. turriculatus resulted in a lack of wide acceptance of Spirograptus. However, with the description of
chemically isolated specimens of S. turriculatus from the Canadian Arctic (Melchin and Lenz 1986),
Spirograptus may now be more rigorously defined, and is of value in uniting a small number of
closely related taxa.

Included within Spirograptus , as defined above, are S. turriculatus (Barrande, 1850), S. guerichi
sp. nov., S. andrewsi (Sherwin, 1974) and S. minimus Obut and Morozova, 1988 (in Obut et al.
1988). The generic status of 'Monograptus' woodae Haberfelner, 1931, is uncertain. The form of its
tiny rhabdosome (which has a dorso-ventral width of 0-2 mm throughout its length, according to
Haberfelner) suggests that it, too, may in the future be assigned to Spirograptus.

The thecal morphology of many species with spiral rhabdosomes previously assigned to
Spirograptus is similar to that of ‘ Monograptus ’ planus (Barrande, 1850) in that the apertural
portion of the metathecae exhibits torsion so that apertures open onto the reverse side of the
rhabdosome (see illustrations of chemically isolated specimens of ‘ M. ’ planus in Loydell and Zhao
1990). Loydell (1992) erected a genus for this ‘ planus group’ which includes ' M. ’ denticulatus

912

PALAEONTOLOGY, VOLUME 36

Tornquist, 1899, ‘M.’ involutus Lapworth, 1876, ‘ M.' proteus (Barrande, 1850), ‘M.’ spiraloides
(Pribyl, 1945) and probably ‘M.’ tullbergi Boucek, 1931.

Other spiral taxa possess thecae which exhibit torsion only in part (mesially and distally) of the
rhabdosome. In these cases, the whole of the theca is twisted, apertures are laterally expanded, and
may bear spines. Species possessing such thecae should be assigned to Oktavites Levina, 1928, the
type species of which, Graptolithus spiralis Geinitz, 1842, has recently been described from
chemically isolated material from Canadian limestone nodules by Lenz and Melchin (1989). Loydell
and Zhao (1990) illustrated the very similar thecal morphology of Oktavites contortus Perner, 1897,
(described by Loydell and Zhao as M. aflf. M. spiralis).

For several spirally curved species, thecal morphology is not yet known in sufficient detail for
confident generic assignment to be made.

Spirograptus turriculatus (Barrande, 1850)

v. * 1850
P 1897
v. 1932
. 1936
v. 1945
v. 1945

. 1971

non 1974
p 1975
p 1982

v. 1986
v. 1989
. 1990
vnon 1991

Plate 1, figs 2, 6; Text-figs 3a, 4a-g, 5
Grapt. turriculatus Barrande, p. 56, pi. 4, figs 7-11.

Monograptus turriculatus , Barr.; Perner, p. 15, pi. 12, figs 34, 36-38 ( non 35 = S. guerichi).
Monograptus turriculatus mut. minor,, n. mut., Boucek, pp. 153, 155, figs 1 c-d.

Monograptus turriculatus var. fimbriatus, Hundt, p. 29.

Spirograptus turriculatus turriculatus (Barrande, 1850); Pribyl, p. 211, pi. 10, figs 1-2.
Spirograptus turriculatus minor (Boucek, 1932); Pribyl, p. 213, text-fig. 3, pi. 20, figs 1-2 [copies
of Boucek’s 1932 figures].

Monograptus ( Spirogr .) turriculatus turriculatus (Barr., 1850); Schauer, p. 74, pi. 30, figs 1-5;
pi. 31, figs 11-13; pi. 45, figs 1-3.

Monograptus turriculatus (Barrande, 1850); Sherwin, p. 172, pi. 12, fig. 6 (= S. guerichi).
Monograptus turriculatus (Barrande, 1850); Bjerreskov, p. 70, pi. 10, fig. H.

Monograptus turriculatus (Barrande, 1850); Lenz, p. 118, figs 9 N, 32 E, 32 G, 33 B, 33 C ( non
32 D, 33C = S. guerichi).

Monograptus turriculatus (Barrande, 1850); Melchin and Lenz, p. 579, figs 1 a-i.
Monograptus turriculatus turriculatus (Barrande); Melchin, fig. 12o.

Spirograptus sinicus Geh; Ge, p. 97, pi. 18, figs 1-7.

Monograptus turriculatus (Barrande, 1850); Loydell, p. 243, figs Id, 8 j, lOg, 1 Id, 12/
(= S. guerichi).

Lectotype. Designated Pribyl 1945, p. 212; specimen NMP L 27597, figured Barrande 1850, pi. 4, fig. 10, from
the Llandovery of Litohlavy, Bohemia PI. 1, fig. 6; Text-fig. 4f. Pribyl (1945), when selecting the lectotype,
incorrectly stated its locality as Zelkovice.

Material. Several hundred specimens from Litohlavy, Bohemia, including the type and figured material of
Barrande (1850). Also Boucek’s (1932) specimens of Monograptus turriculatus minor from Ratinka, Bohemia.
The diagnosis and description below incorporate details (from Melchin and Lenz 1986) which are clearly visible

EXPLANATION OF PLATE 1

Figs 1, 3, 5. Spirograptus guerichi sp. nov. 1, PS 358/1; lower R. linnaei Biozone; Zelkovice, Bohemia.
3, PS 360/1; upper R. linnaei Biozone, Zelkovice Formation; Hlasna Treban, Bohemia. 5, PS 359/1,
holotype; upper R. linnaei Biozone, Zelkovice Formation; Hlasna Treban, Bohemia.

Figs 2, 6. Spirograptus turriculatus (Barrande). 2, NMP L 25043/245, lectotype of Monograptus turriculatus
minor Boucek; ‘ M." crispus Biozone, Litohlavy Formation; Ratinka, Bohemia. 6. NMP L 27597, lectotype
of S. turriculatus', S. turriculatus Biozone, Litohlavy Formation; Litohlavy, Bohemia.

Fig. 4. Spirograptus andrewsi (Sherwin). PS 513; upper St. sedgwickii Biozone, Zelkovice Formation; Zdice,
Bohemia.

All figures are x 5.

PLATE 1

LOYDELL et al Spirograptus

914

PALAEONTOLOGY, VOLUME 36

text-fig. 3. a, Spirograptus turriculatus (Barrande). PS 370/2; upper S. turriculatus Biozone, Litohlavy
Formation; Litohlavy, Bohemia, x 2-9. b, Spirograptus guerichi sp. nov. UCWG 267; isolated specimen,
lacking proximal end; S. guerichi Biozone; Osmundsberget, Dalarna, Sweden, x 25.

only in chemically isolated material. Other specimens examined are from western Wales and northwestern and
Arctic Canada.

Diagnosis. Rhabdosome trochispirally coiled. Thecae hooked, with differential development of the
apertural margin producing a tear-shaped aperture. Proximal thecae bear a single spine, directed to
the obverse side of the rhabdosome. Distal thecae are less retroverted, and bear two spines. Dorso-
ventral width increases rapidly, such that by the fifth whorl it is nearly always greater than 1 mm.
The diameter of the rhabdosome at the third whorl is usually within the range 4-7-5-8 mm
(exceptionally 4-0 mm) and at the fifth whorl is 6-9-8-9 mm (exceptionally 6-2 mm). There are three
to five whorls in 10 mm from the sicular aperture. The sicula is approximately 1 mm long, curved
dorsally and has dextural torsion near its apex.

Description. The rhabdosome is trochispirally coiled through up to fourteen whorls in our material. The apical
angle of the ‘cone’ formed by the rhabdosome generally lies between 40° and 50°. Thecae are without overlap,
retroverted throughout, although less so distally. As a result of differential development of the apertural
margin the thecal apertures are tear-shaped. Thecal axes are oriented parallel to that of rhabdosome growth.

TEXT-FIG. 4. Spirograptus turriculatus (Barrande). A, NMP L 25043/245, lectotype of Monograptus turriculatus
minor Boucek; " M.' crispus Biozone, Litohlavy Formation; Ratinka, Bohemia, b; chemically isolated proximal
end showing one spine per theca (after Melchin and Lenz 1986). c, BGSDKL 1637; Str . johnsonae Subzone
(5. turriculatus Biozone), Aberystwyth Grits Formation; locality A62 of Loydell (1992), western mid- Wales.
d, NMP L 25043/244; obliquely preserved specimen; originally figured by Boucek (1932, fig. lc) as
Monograptus turriculatus minor. E, BGS DKL 2176; upper St. utilis Subzone (S. turriculatus Biozone),
Aberystwyth Grits Formation; locality A50 of Loydell (1992), western mid-Wales, f, NMP L 27597, lectotype;
S. turriculatus Biozone, Litohlavy Formation; Litohlavy, Bohemia. G, PS 370/2; S. turriculatus Biozone,
Litohlavy Formation; Litohlavy, Bohemia. All figures x 5, except b, which is x 10.

LOYDELL ET AL: SPIROGRAPTUS

91

916

PALAEONTOLOGY, VOLUME 36

Proximal thecae bear a single proximo-ventrally directed apertural spine on the apertural margin on the
reverse side of the rhabdosome. Distal thecae bear two spines, that on the obverse side (inside of the spiral)
being shorter than that on the reverse margin. In Bohemian specimens the earliest theca seen to possess paired
spines is th5. Maximum lengths of spines in isolated material are L4 mm and 0-6 mm for those on the reverse
and obverse apertural margins respectively. Spines on specimens preserved within the rock have been observed
to be up to 2 mm in length. Measurements of dorso-ventral width at successive whorl mid-points in Bohemian
and Welsh specimens are given in Table 1. Melchin and Lenz (1986) quote a dorso-ventral width at thl of
05 mm, at th5 07 mm and, more distally, 1-3 mm. Maximum distal dorso-ventral width in flattened and low-
relief specimens from Bohemia is T55 mm. At th2 in this material dorso-ventral width is 05-07 mm. More
strongly diagenetically compressed Canadian material attains a maximum dorso-ventral width approaching
2 mm. 2TRD in isolated specimens is 1 mm proximally, and 1-5 mm distally. In Bohemian material, distal
2TRD is IT— 1-2 mm, 09-1-3 mm being the range observed in Welsh specimens. The sicula, 08-10 mm in
length, is dorsally curved and undergoes slight dextral torsion near its apex, which reaches to the top of th2.

table 1. Measurements of dorso-ventral width (DVW) at the mid-points of successive whorls of Welsh and
Bohemian specimens of Spirograptus turriculatus (Barrande, 1850).

Whorl . . .

2

3

4

5

6

7 8

DVW (Wales)
(n = 28)

0-7-10

0-85-1-25

10-1-3

1-15-14

1-2-13

DVW (Bohemia) 0-5-0-85
(« = 21)

0-75-11

0-85-1-2

0-95-1 25

0-95-1-35

10-1-4 1-0-15

Remarks. The differences between S. turriculatus and other spirograptids are summarized in Table
3. In fully developed rhabdosomes preserved with the rhabdosome coiling axis parallel to bedding
distinguishing specimens of S. turriculatus from other spirograptids is not difficult.

Schauer (1971) distinguished two ‘Formgruppen’, A and B, of S. turriculatus on the basis of
average number of whorls and the position of the longest thecal spines within the rhabdosome.
‘Formgruppe A’ is characterized by 7-5 to 8 whorls and particularly long thecal spines proximally
and mesially. ‘Formgruppe B’, confined stratigraphically to the crispus Biozone, consists of, on
average, 8-10 whorls, with spines particularly prominent distally.

Hutt (1975, p. 1 12) noted bifurcating spines on some specimens of S. turriculatus from the English
Lake District. This phenomenon might account for earlier suggestions (Bulman and Rickards 1970)
of each theca possessing more than two apertural spines. Some Bohemian specimens also appear to
possess more than two spines on distal thecae.

Hundt's (1936) subspecies fimbriatus was erected in the erroneous belief that the thecae of
S. turriculatus lack spines. The supposed distinguishing feature of fimbriatus was the presence of
‘Harchen’ (hairs). Pribyl (1945) placed fimbriatus in synonymy with turriculatus.

Boucek’s (1932) figures of the subspecies minor were copied by Pribyl (1945). These figures are
misleading in that they suggest that the specimens are narrower than they really are (compare
Boucek's figures with Text-figures 4a and 4d herein). The lectotype of S. t. minor , selected by Pribyl
(1945), is preserved somewhat obliquely to bedding (note the position on the virgula on final whorl;
Text-fig. 4a) and thus dorso-ventral width cannot be measured accurately. Comparison of Text-
figure 4a with 4f (lectotype of S. turriculatus) indicates that S. turriculatus and S. turriculatus minor
are conspecific; however, comparison of Text-figure 4a with Text-figures 6c-h illustrates the larger
size of minor as compared to guerichi. The other specimen figured by Boucek (1932, fig. 1 c\ Text-
fig. 4d) is preserved even more obliquely to bedding than the lectotype, and would be very difficult
in isolation to assign with confidence to either S. guerichi or 5. turriculatus. However, the torsion
of the sicula, characteristic of S. turriculatus , is visible in this specimen.

On the reverse side of the slab bearing the lectotype of S. t. minor is a specimen of " Monograptus'

LOYDELL ET AL: SPIROGRAPTUS

917

text-fig. 5. Reverse side of slab (NMP L 25043/245) bearing lectotype of Monograptus turriculatus minor on
obverse. ‘ Monograptus ' crispus Lapworth, 1876 is at top left; Spirograptus turriculatus (Barrande, 1850) is at
bottom right; Litohlavy Formation; Ratinka, Bohemia, x 5.

crispus Lapworth, 1876, (Text-fig. 5). This species not found in strata yielding S. guerichi, but
occurs with S. turriculatus in the ‘ M. ’ crispus Biozone. There is no longer an exposure at the type
locality for S. t. minor at Ratinka, Bohemia.

Spirograptus guerichi sp. nov.

Plate 1, figs 1, 3, 5; Plate 2, figs 1-6, 8-10; Text-figs 3b, 6c-h

p 1897 Monograptus turriculatus Barr.; Perner, p. 15, pi. 12, fig. 35 (non 34, 36-38 = S. turriculatus).

. 1971 Monograptus turriculatus minor Boucek, 1932; Schauer, p. 74, pi. 30, figs 7-8; pi. 31, figs 7-9.

. 1974 Monograptus turriculatus (Barrande, 1850); Sherwin, p. 172, pi. 12, fig. 6.

p 1975 Monograptus turriculatus (Barrande, 1850); Bjerreskov, p. 70, (non pi. 10, fig. FI =

S. turriculatus).

? 1976 Spirograptus minor (Boucek, 1932); Sennikov, p. 194, pi. 13, figs 5-7.

? 1978 Spirograptus minor (Boucek); Wang, p. 311, pi. 1, fig. 7.

p 1982 Monograptus turriculatus (Barrande, 1850); Lenz, p. 118, figs 32d, 33c (non figs 9n, 32e, 32g,

33a-b = S. turriculatus).

v. 1984 Spirograptus turriculatus (Barrande); Chen, p. 70, pi. 15, figs 9-10; pi. 16, figs 1-2.

v. 1984 Spirograptus minor (Boucek); Chen, p. 71, pi. 16, figs 3-4, 7 (? 12).

1986 Spirograptus turriculatus minor (Boucek); Fu and Song, p. 121, pi. 12, figs 11-14.

v. 1989 Monograptus turriculatus minor Boucek; Melchin, fig. 11m.

v. 1991 Monograptus turriculatus (Barrande, 1850); Loydell, figs Id, 8 /, lOg, llrf, 12 f

Derivation of name. After G. Giirich, author of the genus Spirograptus.

Holotype. Specimen no. PS 359/1; from the uppermost part of the Zelkovice Formation at Fllasna Treban,
Bohemia; Llandovery, upper Rastrites linnaei Biozone (PI. 1, fig. 5; Text-fig. 6d).

Material. Several hundred chemically isolated specimens from limestone nodules from Osmundsberget,

918

PALAEONTOLOGY, VOLUME 36

Dalama, Sweden (nodules from within 4 m above the unconformable contact between the Llandovery and the
Upper Ordovician Boda Limestone; from sections on both sides of quarry at northern end of main quarry,
4-5 km north of Boda Church, Siljan District [UTM Grid Reference WH 1095 6870]); these are almost
exclusively proximal fragments, representing the first one to three whorls. More than one hundred low relief
(pyrite) specimens from the early Telychian sequences of Hlasna Treban, and Zelkovice (Bohemia). Several
hundred specimens from western mid-Wales (see Loydell in press) and from northwestern and Arctic Canada.

Diagnosis. Tightly trochispirally coiled rhabdosome bears hooked thecae, each furnished with two
apertural spines. Maximum dorso-ventral width (excluding spines) averages 1 mm (maximum is
0-8 mm in Bohemian material). In specimens compressed with the axis of coiling approximately
parallel to bedding, rhabdosome diameter at whorl 3 usually falls within the range 3-4^-4-3 mm and
at whorl 5 is 5 0-5-7 mm. In 10 mm from the sicular aperture there are 6-7 whorls. Sicula dorsally
curved, 0-8-1 T 5 mm in length.

Description. The rhabdosome is trochispirally coiled. Specimens exhibiting more than seven whorls are rare,
eight whorls being the maximum recorded. The apical angle of the ‘cone’ formed by the rhabdosome usually
lies in the range 30-40°, but may be greater and, very rarely, exceeds 50°. In Bohemian material the diameter
of the rhabdosome at the third whorl usually falls within the range 34-4-3 mm (maximum 4-8 mm) and at the
fifth whorl is 5-0-5- 7 mm. There are usually 6-7 whorls in 10 mm from the sicular aperture. The thecae are
hooked, with simple symmetrical apertures. Distal to th2, the thecae show slight torsion so that the apertures
are directed to the reverse side of the rhabdosome (i.e. inwards, towards the centre of the spiral). Thl exhibits
no such torsion. The periderm of the dorsal wall of the metathecae is very thin and is rarely preserved (in
isolated specimens, Plate 2, figs 1 and 5). Thecal apertures are furnished at their lateral margins by a pair of
ventro-laterally directed spines, up to 0-8 mm in length (invariably broken on isolated specimens). In isolated
specimens dorso-ventral width is 0-42-0-47 mm at thl, with a maximum distal value of 0-7 mm. Flattened and
low relief specimens from Bohemia have a dorso-ventral width of c. 04 mm at th2 and a maximum distal
dorso-ventral width of 0-8 mm. Welsh specimens attain a maximum dorso-ventral width of 1-0 mm, whilst
Arctic Canadian specimens (the most diagenetically flattened in our collections) attain a maximum dorso-
ventral width of 1-2 mm. Measurements of dorso-ventral width at successive whorl mid-points for Welsh and
Bohemian specimens are given in Table 2. Thecal overlap is negligible. Thecae are very closely spaced. 2TRD
in Welsh specimens is 0-8-1-2 mm (25-17 thecae in 10 mm) at the mid-point of whorl 2 and increases to
1 0—1-3 mm (20-15 thecae in 10 mm) distally. The sicula is dorsally curved, 0-8-1T5 mm in length. Its apex
reaches to from just below the base of th2 to just below the base of th3.

Remarks. The differences between S. guerichi, S. turriculatus and the other spirograptid species are
summarized in Table 3. S. andrewsi (Sherwin, 1974) differs primarily from S. guerichi in its much
greater apical angle (giving the rhabdosome a very low trochispiral form) which results in specimens
preserved with the coiling axis perpendicular to bedding being by far the most commonly
encountered. S. minimus has a maximum dorso-ventral width of only 0-7 mm and has narrower
prothecal bases than S. guerichi. As the extreme of the range of variation in apical angle values in
S. guerichi overlaps with the commonest apical angle observed in S. turriculatus, this on its own is

EXPLANATION OF PLATE 2

Figs 1-6, 8-10. Spirograptus guerichi sp. nov.; chemically isolated specimens; S. guerichi Biozone;
Osmundsberget, Dalarna, Sweden. 1, UCWG 934C; part of first whorl; note poor preservation/absence of
dorsal wall; all but one spines have been broken off, x 50. 2, UCWG 934a; thecal aperture and one of the
apertural spines of th3, x 150. 3, UCWG 953d; looking down upon sicular aperture and aperture of thl,
x 50. 4, UCWG 9451; dorsally curved sicula and protheca of thl, x 50. 5, UCWG 953e; proximal end, note
paired spines on thl, x 50. 6, UCWG 945j; dorsally curved sicula and protheca of thl, x 50. 8, UCWG
934c; x 30. 9, UCWG 934d; proximal end, note paired spines on th2, x 50. 10, UCWG 934b; x 30.

Fig. 7. Stimulograptus halli (Barrande), UCWG 937f; chemically isolated proximal end; S. guerichi Biozone;
Osmundsberget, Dalarna, Sweden, x 25.

PLATE 2

LOYDELL et al., Spirograptus , Stimulograptus

920

PALAEONTOLOGY, VOLUME 36

text-fig. 6. a-b. Spirograptus andrewsi (Sherwin). Upper St. sedgwickii Biozone, Zelkovice Formation; Zdice,
Bohemia; a, PS 5 1 3a. B, PS 5136. c-h, Spirograptus guerichi sp. nov. c, BGS DKL 1981; Cwmsymlog
Formation, ‘M.’ gemmatus Subzone (S. guerichi Biozone); locality Cl of Loydell (1992), western mid-Wales.
d, PS 359/1, holotype; Zelkovice Formation, upper R. linnaei Biozone; Hlasna Treban, Bohemia, e, ROM
46011; Cape Phillips Formation; Fluff Ridge, Ellesmere Island, Canada, f, PS 360/1; Zelkovice Formation,
upper R. linnaei Biozone; Hlasna Treban, Bohemia. G, PS 358/1; Zelkovice Formation, lower R. linnaei
Biozone; Zelkovice, Bohemia. H, BGS DKL 819; Aberystwyth Grits Formation, lower St. utilis Subzone (S.
guerichi Biozone); locality A16 of Loydell (1992), western mid-Wales. All figures are x 5.

table 2. Measurements of dorso-ventral width (DVW) at the mid-points of successive whorls of Welsh and
Bohemian specimens of Spirograptus guerichi sp. nov.

Whorl...

2

3

4

5

6

7

DVW (Wales)
(n = 25)

0-5-0-7

0-58-0-85

06-0-9

0-85-0-9

0-75

DVW (Bohemia)
(n = 31)

0-45-0-6

0-45-0-7

0-55-0-7

0-6-0-75

0-65-0-8

0-7-08

not an infallible basis for distinguishing the two taxa. The greater dorso-ventral width, and whorl
diameter of S. turriculatus are the most useful of the features listed in Table 3 for separating this
species from S. guerichi.

LOYDELL ET AL : SPIROGRAPTUS

921

table 3. Comparison of Spirograptus species (w. = whorl).

species

feature^""^^^

andrewsi

guerichi

turriculatus

minimus

Sicula: shape

Slightly dorsally

Dorsally curved

Dorsally curved

Not known

curved

No torsion

Dextral torsion
near apex

length

1. 2-1.4

0.8-1.15

0.8-1. 0

Thecal

Symmetrical (in

Symmetrical

Asymmetrical

Not known

apertures

Sherwin’s 1974
reconstruction)

(tear -shaped)

Thecal

Probably single

Paired

Prox. single

Number not

spines

Dist. paired (and
bifurcating?)

known

length (max.)

1.3

0.8

2.0

1.5

Max. no. of
whorls

3

8

16

4

Apical angle

Higher than
turriculatus

30-40"

40-50’

c. 40°

Dorso- w.3

1.1 -1.2

0.45 r 0.85

0.75-1.25

c.0.5

ventral w.5

0.6-0. 9

0.95-1.4

width max.

0.95 1.1-1. 2

(Aust.) (Bohemia)

0.8-1. 2

1. 5-2.0

0.7

Whorl w.3

3.4 -4. 3

4.7-58

3.3-3. 5

diameter

(exceptionally

(exceptionally

w.5

4.8)

4.0)

5.0 -5.7

6.9-8. 9

(exceptionally

6.2)

No. of in 5mm

3-5

2-3

3

whorls 10mm

6-7

3-4

Other features

Prothecae are
narrower at base
than in guerichi

The differences between 5. guerichi and Boucek’s (1932) specimens of S. turriculatus minor have
been discussed above under S. turriculatus. In general, specimens referred to S. t. minor by non-
Bohemian authors are S. guerichi (see synonymy list above and Loydell 1992).

Pribyl (1945) and Schauer (1971) used the number of whorls comprising the rhabdosome to
distinguish between S. turriculatus and S. guerichi (which they identified as the subspecies minor).
Specimens of S. turriculatus may possess up to sixteen whorls (according to Obut et at. 1988), with
fourteen the maximum observed in Bohemian material. Eight is the maximum number recorded in
S. guerichi. This does not allow growth stages possessing fewer whorls to be distinguished.

922

PALAEONTOLOGY, VOLUME 36

Considerable caution should be exercised when attempting to identify to specific level specimens
which are preserved obliquely to bedding, short proximal fragments, and also tectonically distorted
specimens. Schauer (1971) illustrated the dramatic effects of tectonic distortion on S. turriculatus,
recognizing ‘ tektonische Lang for men Y tektonische Kurzformen ’ and ‘ tektonischen Schragformen' ,
depending upon the orientation of the rhabdosome to the direction of principal stress.

That Hutt et al. (1971) did not encounter specimens which now would be attributable to
S. guerichi in their study of the Osmundberget fauna is surprising as this species was common in the
nodules processed for this present work.

Biostratigraphically, S. guerichi is restricted to the earliest Telychian (Text-fig. 7), making its last
appearance in the lowest part of the S. turriculatus Biozone (upper Stimulograptus utilis Subzone
of Loydell 1991). Loydell (in press) noted that the species appears to be cosmopolitan, and not
geographically restricted in it distribution as was suggested by Melchin (1989).

Spirograptus andrewsi (Sherwin, 1974)

Plate 1, fig. 4; Text-fig. 6 A-B

* 1974 Monograptus andrewsi sp. nov., Sherwin, p. 170, pi. 10, figs 5-6; pi. 11, fig. 11; pi. 12, fig. 8; text-
figs 3 a-c.

Holotype. By original designation, the specimen illustrated by Sherwin 1974, pi. 11, fig. 11, from the
Llandovery of the Forbes District, New South Wales, Australia.

Material. Approximately thirty flattened, mostly incomplete specimens from the upper part of the
Stimulograptus sedgwickii Biozone of Zadni Treban, Zdice (‘Colonie Lapworth’), and Zelkovice, Bohemia.
This level in Bohemia possibly correlates with the Stimulograptus halli Biozone of Britain (see Loydell 1991).

Diagnosis. Rhabdosome forms a very low trochispiral. Thecae hooked and spinose, without overlap.
Proximal dorso-ventral width 0-5-06 mm, distally 0-55-0-95 mm in Australian specimens,
IT— 1-2 mm in Bohemian specimens. Sicula 1-2-1-4 mm long, very gently dorsally curved, its apex
reaching half way up th2.

Description. In the Bohemian material, the rhabdosome is in the form of a very low trochispiral, with a
maximum of three whorls developed. The maximum diameter of the last whorl is 8 mm. The sicula and earliest
thecae have not been seen. The thecae are triangular, without overlap. The ventral thecal wall is inclined at
c. 45° to the rhabdosome axis. The dorsal wall is nearly perpendicular to this axis. Stout thecal apertural spines
are up to 0-45 mm in length but many are broken. There appears to be only one spine per theca. Proximal
dorso-ventral width is 0-6-0-75 mm increasing to a maximum of 1 - 1—1 -2 mm distally. 2TRD is 1-4-1-95 mm at
the mid-point of whorl two.

Remarks. The form of the rhabdosome and the very gentle sicular curvature are useful in
distinguishing this species from S. turriculatus and S. guerichi (see Table 3). Short fragments could
prove impossible to distinguish, however.

Sherwin’s (1974) specimens of S. andrewsi were collected from strata older than those yielding
S. guerichi (described as Monograptus turriculatus by Sherwin), a situation which parallels that in
Bohemia. Sherwin was unsure as to whether the thecae of S', andrewsi bear one or two spines.

PHYLOGENY

The close similarity between the thecal morphology of S. turriculatus and that of Monograptus
sedgwickii (Portlock, 1843), a species now assigned to Stimulograptus Pribyl and Storch, 1983, was
noted by Elies and Wood (1913). Plate 2, figure 7 illustrates the proximal end of another

LOYDELL ET AL. SPIROGRAPTUS

923

stimulograptid, Stimulograptus halli (Barrande, 1850), for comparison with that of S. guerichi,
which is strikingly similar (PI. 2, fig. 5) except for its curvature. Melchin and Lenz (1986) proposed
that St. sedgwickii was directly ancestral to S. turriculatus and indeed it seems certain that the
spirograptids evolved from a stimulograptid ancestor by introduction of a spiral rhabdosome
morphology accompanied by curvature of the sicula.

Within Spirograptus, early evolution involved increasing curvature of the sicula, changes in thecal

924

PALAEONTOLOGY, VOLUME 36

morphology, more rapid attainment of a greater dorso-ventral width and also an increase in the
overall number of whorls (Text-fig. 7). The stratigraphically youngest representative of the genus,
Sp. minimus Obut and Morozova, 1988 (in Obut et al. 1988) is also the most diminutive, however.

A REVISED BIOZONAL SCHEME FOR THE EARLY TELYCHI AN

With Monograptus turriculatus minor proving to be a junior synonym of Spirograptus turriculatus ,
the minor Biozone of authors being based upon S. guerichi, and S. turriculatus sensu stricto absent
from strata currently assigned to the earliest Telychian (the lower part of the turriculatus Biozone
(Rickards 1989)) a revision of early Telychian biostratigraphy is necessary.

The S. guerichi Biozone is here erected for the interval from the first appearance of S. guerichi to
the first appearance of S. turriculatus , the end of the turriculatus Biozone being marked by the first
appearance of' Monograptus' crispus Lapworth, 1876. S. guerichi is virtually confined to its biozone
(Text-fig. 7), whilst S. turriculatus ranges high into the crispus Biozone.

With regard to the base of the Telychian Stage, the only identifiable graptolite recovered from the
Telychian of the Llandovery Series of the Llandovery area (Cocks et al. 1984) appears to be lost.
If this specimen is Paradiversograptus runcinatus (Lapworth, 1876), as identified by Cocks et al.
(1984), then this would probably mean that the base of the Telychian, as formally defined in Bassett
(1985), may coincide roughly with the base of the S. guerichi Biozone. Pa. runcinatus has its acme
in the lower part of the S', guerichi Biozone (Bjeireskov 1975; Loydell 1991).

Acknowledgements. M. G. Bassett and R. B. Rickards are thanked for providing the Osmundbeget limestone
nodules. D. K. L. acknowledges with thanks the financial support of the University of Wales. M. J. M gratefully
acknowledges the financial support of an NSERC Operating Grant and logistical support in the Arctic from
the Polar Continental Shelf Project, and thanks A. C. Lenz for providing access to collections from
northwestern Canada.

REFERENCES

barrande, J. 1850. Graptolites de Boheme. J. Barrande, Prague, vi + 74 pp.
bassett, m. G. 1985. Towards a 'Common Language’ in stratigraphy. Episodes , 8, 87-92.
bjerreskov, M. 1975. Llandoverian and Wenlockian graptolites from Bornholm. Fossils and Strata, 8, 1-94,
pis 1-13.

boucek, b. 1931. Predbezna zprava o nekterych novych druzlch graptolitu z ceskeho gotlandienu.
[Communication preliminaire sur quelques nouvelles especes de Graptolites provenant du Gothlandien de
la Boheme.] Vestm'k Statm'ho Geologickeho Ustavu, 7, 293-313.

- 1932. Predbezna zprava o nekterych novych druzlch graptolitu z ceskeho gotlandienu. (Cast II).
[Preliminary report on some new species of Graptolites from the Gothlandian of Bohemia.] Vestmk Statm'ho
Geologickeho Ustavu Republikv Ceskoslovenske, 8, 150-155.

1953., Biostratigraphy, development and correlation of the Zelkovice and Motol Beds of the Silurian of
Bohemia. Sbormk Ustredm'ho Ustavu Geologickeho, 20, 473 — 484.
bulman, o. M. B. 1929. The genotypes of the genera of graptolites. Annals and Magazine of Natural History,
(10), 4, 169-185.

- 1970. 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-l- 163 pp.

chen, x. 1984. [Silurian graptolites from southern Shaanxi and northern Sichuan with special reference to
classification of Monograptidae], Palaeontologica Sinica, New Series B, 166, 1-102. [In Chinese with English
summary],

cocks, l. r. m., woodcock, N. H., rickards, R. b., temple, j. t. and lane, p. d. 1984. The Llandovery Series
of the type area. Bulletin of the British Museum (Natural History), Geology Series, 38, 131-182.
elles, G. l. and wood, e. m. r. 1913. A monograph of British graptolites. Part 10. Palaeontographical Society
Monograph, 67 (327), 487-526, pis 50-52.

LOYDELL ET AL: SPIROGRAPTUS

925

fu, l. p. and song, L. s. 1986. [Stratigraphy and palaeontology of Silurian in Ziyang Region (Transitional Belt)]
Bulletin of the Xi'an Institute of Geology and Mineral Resources , Chinese Academy of Geological Sciences, 14,
1-198, pis 1-31. [In Chinese].

ge, m. y. 1990. [Silurian graptolites from Chengkou, Sichuan], Palaeontologica Sinica , New Series B , 179,
1-157. [In Chinese with English summary].
gurich, G. 1908. Leitfossilien (Kambrium bis Silur). Lief 1. Berlin, 95 pp.

haberfelner, e. 1931. Graptolithen aus dem Obsersilur der Karnischen Alpen. I. Teil: Hochwipfel, Nordseite.

Sitzungsberichte der Akademie der Wissenschaften in Wien (1), 140 (1-2), 89-168.
howe, m. p. a. 1983. Measurement of thecal spacing in graptolites. Geological Magazine , 120, 635-638.
hundt, R. 1936. Graptolithenleben und Graptolithensterben im Mitteldeutschen Silurmeer. Zeitschrift fur die
gesamte Naturwissenschaften , 1, [fide Pribyl 1945].

hutt, J. E. 1975. The Llandovery graptolites of the English Lake District. Part 2. Palaeontographical Society
Monograph, 129 (542), 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.

jones, o. t. 1921. The Valentian Series. Quarterly Journal of the Geological Society, London , 77, 144—174.
kriz, J. 1991. The Silurian of the Prague Basin (Bohemia) - tectonic, eustatic and volcanic control on facies
and faunal development. Special Papers in Palaeontology, 44, 179-203.
lapworth, c. 1873. Notes on the British graptolites and their allies. 1. On an improved classification of the
Rhabdophora. Geological Magazine , (1), 10, 500-504, 555-560.

- 1876. On Scottish Monograptidae. Geological Magazine , (2), 3, 308-321, 350-360, 499-507, 544—552.
— 1878. The Moffat Series. Quarterly Journal of the Geological Society, London, 34, 240-346.
lenz, a. c. 1982. Llandoverian graptolites of the Northern Canadian Cordillera: Petalograptus, Cephalo-
graptus, Rhaphidograptus, Dimorphograptus, Retiolitidae, and Monograptidae. Life Sciences Contributions,
Royal Ontario Museum, 130, 1-154.

— and melchin, m. j. 1989. Monograptus spiralis and its phylogenetic relationship to early cyrtograptids.
Journal of Paleontology, 63, 341-348.

levina, E. F. 1928. Graptolitit iz Aq-Tengi v Turkestanom chrebte. Trudy Sredie-Aziatskogo Gosydarstvennogo
Universiteta, Series 7a, Geologiya, Fascicule 5, 1-18.

loydell, d. k. 1991. The biostratigraphy and formational relationships of the upper Aeronian and lower
Telychian (Llandovery, Silurian) formations of western mid-Wales. Geological Journal, 26, 209-244.

— 1992. Upper Aeronian and lower Telychian (Llandovery) graptolites from western mid-Wales. Part 1.
Palaeontographical Society Monograph , 146 (589), 1-55, pi. 1.

— and zhao, y. h. 1990. Studies on some monograptids using the scanning electron microscope. Acta
Palaeontological Sinica, 29, 331-336.

Matthews, s. c. 1973. Notes on open nomenclature and on synonymy lists. Palaeontology, 16, 713-719.
melchin, m. J. 1989. Llandovery graptolite biostratigraphy and paleobiogeography, Cape Phillips Formation,
Canadian Arctic Islands. Canadian Journal of Earth Sciences, 26, 1726-1746.

— and lenz, a. c. 1986. Uncompressed specimens of Monograptus turriculatus (Barrande, 1850), from
Cornwallis Island, Arctic Canada. Canadian Journal of Earth Sciences, 23, 579-582.

mu, A. T. 1955. On Spirograptus Gurich. Acta Palaeontologica Sinica, 3, 1-10.

OBUT, A. M. 1949. Polevoi atlas rukovodyaaschchich graptolitov verchnego silura Kirizskoy SSR. Publishing
House of the Kirgiz Branch of the Academy of Sciences of the U.S.S.R., Frunze, 57 pp.

— morozova, f. I., Moskalenko, t. a. and chegodaev, L. D. 1988. Graptoliti, konodonti i stratigrafiya
silura, nizhnego devona Severnogo Kavkaza. Akedemiya Nauk SSSR Sibirskoe Otdelenie. Trudy Instituta
Geologii i Geofiziki im. 60-letiya Soyuza SSR, 698, 1-220.

perner, J. 1897. Etudes sur les Graptolites de Boheme. Illieme Partie. Monographic des Graptolites de L Etage
E. Section a. R. Gerhard, Prague. 25 pp., pis 9-13.

PORTLOCK, J. E. 1843. Report on the geology of the county of Londonderry, and of parts of Tyrone and Fermanagh.

Milliken; Hodges and Smith; Longman, Brown, Green and Longmans, Dublin and London, xxi + 784 pp.
pribyl, a. 1945. The Middle-European monograptids of the genus Spirograptus Gurich. Bulletin International
de 1' Academie Tcheque des Sciences, 45, 185-231.

— and storch, p. 1983. Monograptus ( Stimulograptus ) subgen. n. (Graptolites) from the Lower Silurian of
Bohemia. Vestnlk Ustredniho Ustavu Geologickeho, 58, 221-226.

rickards, R. B. 1970. The Llandovery (Silurian) graptolites of the Howgill Fells, Northern England.
Palaeontographical Society Monograph, 123 (524), 1-108, pis 1-8.

1976. The sequences of Silurian graptolite zones in the British Isles. Geological Journal, 11, 153-188.

926

PALAEONTOLOGY, VOLUME 36

rickards, r. b. 1989. Exploitation of graptoloid cladogenesis in Silurian stratigraphy. 267-274. In Holland,
c. H. and bassett, M. G. (eds). A Global Standard for the Silurian System. National Museum of Wales,
Geological Series No. 9, Cardiff, 325 pp.

— hutt, J. E. and berry, w. b. N. 1977. Evolution of the Silurian and Devonian Graptoloids. Bulletin of the
British Museum ( Natural History), Geology Series, 28, 1-120.
shauer, M. 1971. Biostratigraphie und Taxonomie der Graptolithen des tieferen Silurs unter besonderer
Berucksichtigung der tektonischen Deformation. Freiberger Forschungshefte (C), 373, Palaontologie, 1-185.
sennikov, n. v. 1976. Graptoliti i stratigrafiya nizhnego silura Gornogo Altaya. Nauka, Moscow, 274 pp.
sherwin, L. 1974. Llandovery graptolites from the Forbes District, New South Wales. 149-175. In rickards,
r. b., jackson, d. e. and hughes, c. p. (eds). Graptolite studies in honour of O. M. B. Bulman. Special
Papers in Palaeontology, 13, 1-261.

strachan, i. 1971 . A synoptic supplement to ‘A monograph of British graptolites by Miss G. L. Elies and Miss
E. M. R. Wood’. Palaeontographical Society Monograph , 125 (529), 1-130.
tornquist, s. l. 1899. Researches into the Monograptidae of the Scanian Rastrites Beds. Lunds Universitets
Arsskrifter, 35, 1-25.

wang, x. F. 1978. [A restudy on the graptolite and the age of the Wentoushan Formation from Liantan,
Kwangtung]. Acta Geologica Sinica, 4, 303-317. [In Chinese.]

DAVID K. LOYDELL

Institute of Earth Studies
University of Wales
Aberystwyth,

Dyfed SY23 3DB, UK

PETR storch

Czech Geological Survey
Malostranske namesti 19
118 21, Praha 1, Czech Republic

Typescript received 6 August 1992
Revised typescript received 25 November 1992

MICHAEL J. MELCHIN

Department of Geology
St Francis Xavier University
Antigonish,

Nova Scotia, Canada, B2G ICO

RETRODEFORMATION OF FOSSILS-A SIMPLE

TECHNIQUE

by A. W. A. RUSHTON Gild M. SMITH

Abstract. We describe a simple and quick method of restoring photographic or line-drawing images of
deformed fossils, using a Canon CLC 300 laser copier in place of computer hard- and software; it will help in
the identification of material from deformed rocks. As an example, retrodeformation of the trilobite Angelina
sedgwickii Salter rapidly gave results that compare favourably with previous restorations of the same species.

To the structural geologist, deformed fossils, in particular those with bilateral symmetry or known
angles and dimensions, provide valuable information on the state of strain in a rock. They are easily
measured and analysed using the strain techniques described by Ramsay (1967), and Ramsay and
Huber (1983). On the other hand, this deformation is troublesome to the palaeontologist. Features
used for specific discrimination, such as measurements and length-to-width ratios, become altered.
An assessment of strain is essential for the determination of deformed fossils, but because
palaeontologists rely so much on illustration, the restoration of images of deformed fossils is also
important to their interpretation. This was tellingly demonstrated by Hughes and Jell (1992) who,
by computer-based analysis of a population from deformed rocks, were able to assign seven earlier-
described taxa to a single species and to published restored illustrations. Cooper (1990) described
various techniques for restoring deformed fossils, and their limitations and underlying assumptions.

The purpose of this note is to draw attention to one of the techniques for restoration mentioned
by Cooper (1990, p. 331). It is so quick and simple that we believe it will be of value to anyone
identifying or describing deformed fossils. The technique is akin to that described by Hughes and
Jell (1992), but instead of using a computer it uses the ‘anamorphic zoom’ facility available on the
Canon CLC 300 (colour laser copier). This facility enables a pair of orthogonal axes (x and y) to
be independently enlarged or reduced, thereby allowing one to ‘squash, or ‘stretch’ a line-drawing
or half-tone photograph of the deformed fossil by a predetermined amount. This ‘retrodeformation’
(Williams 1990) takes only a few moments and comparatively high quality images can be produced.
The relative expense of operating the colour copier means that it is appropriate to calculate the
amount of deformation, rather than to use trial-and-error, to obtain good retrodeformed images.
The amount of deformation can be calculated by a variety of techniques (Ramsay and Huber 1983;
Cooper, 1990) which provide the shape and orientation of the strain ellipse in the deformed rock.
The ratio of the major to minor axes of this ellipse can then be used for the ratio of the enlargement
or reduction for the x and y axes on the copier.

The example given here is based on the trilobite Angelina sedgwickii Salter, 1859, from the
Tremadoc Series of North Wales. This species is known from numerous more or less complete
specimens, but all of them are distorted. Fortey and Owens (1992a) reviewed several attempts to
restore Angelina sedgwickii, and compared these restorations with an undeformed cranidium
described from Shropshire (Fortey and Owens 1992 b).

A photograph was used of a slab with two fairly complete specimens which he at different angles
to a tectonic lineation (Text-fig. 1). The fossils occur in a deformed but poorly-cleaved silty
mudstone; the bedding and regional (Sr) cleavage are nearly orthogonal to each other, and the
cleavage trace on the bedding surface is indicated by a weak intersection lineation. It is assumed that
this intersection lineation is parallel to the maximum extension direction in the rock. Lines were

IPalaeontology, Vol. 36, Part 4, 1993, pp. 927-930.)

© The Palaeontological Association

928

PALAEONTOLOGY, VOLUME 36

text-fig. 1: Angelina sedgwickii Salter, 1859. Upper Tremadoc, 820 m north of Minfordd British Rail
station (Grid Ref. SH 394 600). British Geological Survey GSM 102628 (A. J. J. Goode coll.), a, as preserved,
x 1 ; b, the same figure retrodeformed. The thick vertical line represents the trace of the cleavage on the bedding;
the short lines indicate the sagittal and transverse lines for each trilobite.

drawn on the photograph parallel to the tectonic lineation, and, for each trilobite, the sagittal line
and a transverse line at the base of the cephalon. These could be drawn to an accuracy of about one
degree.

RUSHTON AND SMITH: FOSSIL RETRODEFORM ATION

929

Three graphical methods, the Mohr Circle, Breddin curves, and Wellman’s method (Cooper
1990) were used to determine the ellipticity (R) of the strain ellipse. The results are summarized in
Table 1, and show close agreement between the different techniques.

table 1. Comparison of estimates of strain ratio for the fossils in Text-figure 1, determined by three different
techniques, y/ = angular shear strain, <p' = orientation angle.

Breddin

curve

Mohr

circle

Wellman

method

Upper specimen

ys = +37°

2-01

1-996

(!>' = 26

2-08

Lower specimen

y/ = + 18°

2-00

2-02

<t>' = 6

To prepare the retrodeformed image we aligned the photograph on the copier such that the
tectonic lineation was parallel to the y axis (this is made easy by trimming one edge of the
photograph, or its mount, parallel to the lineation), and dialled enlargement/reduction values of x
and y in the ratio 2-02: TO. In order to make the restored image (and the strain ellipse) equal in area
to the original, values of .v = 142 per cent, and y = 70 per cent were used (we are assuming here that
deformation is homogeneous, and are ignoring any extension or compression normal to the plane
of bedding). The copier was set to reproduce in black and white, but ‘full colour’ copies, even of
black and white prints, can give very good results.

The resulting image is, we believe, the best representation so far of the undeformed shape of
Welsh examples of A. sedgwickii. The sagittal and transverse lines on each are practically at right
angles, namely 90° on the upper example in Text-figure 1, and 89° on the lower. The trilobites, which
are of similar size and therefore about the same stage of growth, have the same proportions. Table
2 gives a comparison of our restoration with those given in Fortey and Owens (1992a, text-figs 2,

table 2. Proportions of restorations of Angelina sedgwickii as given by Fortey and Owens (1992a) compared
with Text-fig 1.

Length/width
of cephalon

Length/ width
of glabella

Salter’s ‘broad form’

0-40

110

Appleby and Jones

0-62

1-56

Ramsay and Huber

0-48

1 20

Fortey and Owens

0-47

1 08

Fig. 1, upper specimen

0-48

1-22

Fig. 1, lower specimen

0-49

1 -25

4 and 5): of these Salter's restoration is too broad and that of Appleby and Jones is too narrow;
Ramsay and Huber’s reconstruction is closest to ours. Although the morphological effects of
tectonic strain are removed by this technique, in the present instance the image of the cleavage trace,
captured on the photograph, remains, and can be seen striating the surfaces of the trilobites.

The chief advantages of this method are its rapidity, the fact that not only drawings but
photographs can be reshaped, and that no computer set-up is required (the facility being available

930

PALAEONTOLOGY, VOLUME 36

in many high-street copying services). The Analogue Video Reshaper ( Appleby and Jones 1976) and
the computer technique outlined by Boyce (1990) are more adaptable, but it is less easy to obtain
publication-quality figures from them. The microcomputer apparatus described by Williams (1990)
is very suitable for trial and error experimentation, but does not give good images of photographs
and half-tone plates; it could, however, be used in conjunction with the Canon copier to provide
publishable retrodeformed figures, in much the same way as Hughes and Jell (1992, p. 319)
described.

Acknowledgements. We thank Dr R. A. Fortey for discussion of Angelina. This paper is published by
permission of the Director, British Geological Survey (N.E.R.C.).

REFERENCES

appleby, r. m., and jones, G. L. 1976. The Analogue Video Reshaper -a new tool for palaeontologists.
Palaeontology , 19, 565-586.

boyce, w. d. 1990. Computer-aided restoration - reconstruction of trilobites (CARROT). Current Research
(1990) Newfoundland Department of Mines and Energy, Geological Survey Branch, Report , 90-1, 277-280.
cooper, r. a. 1990. Interpretation of tectonically deformed fossils. New Zealand Journal of Geology and
Geophysics, 33, 321-332.

fortey, r. A., and owens, R. M. 1992a. The Trilobite Angelina unstretched. Geology Today, 8, 219-221.

— 19926. The Habberly Formation: youngest Tremadoc in the Welsh Borderlands. Geological
Magazine, 129, 553-566.

hughes, N. c., and jell, p. a. 1992. A statistical/computer-graphic technique for assessing variation in
tectonically deformed fossils and its application to Cambrian trilobites from Kashmir. Lethaia, 25, 317-330.
ramsay, j. G. 1967. Folding and fracturing of rocks. McGraw-Hill, New York, 568 pp.

and huber, m. i. 1983. The techniques of modern structural geology. Volume 1 : Strain analysis. Academic

Press, London, xiii + 307 pp.

salter j. w. 1859. In murchison, r. i. Siluria (3rd edition). John Murray, London, xix + 592 pp.
williams, s. h. 1990. Computer-assisted graptolite studies. 46-55. In bruton, d. l., and harper, d. a. t. (eds).
Microcomputers in palaeontology. Contributions from the Palaeontological Museum, University of Oslo, no.
370, 105 pp.

A. W. A. RUSHTON
M. SMITH

Typescript received 16 November 1992
Revised typescript received 21 April 1993

British Geological Survey
Keyworth,

Nottingham NG12 5GG, UK

ROLE OF SHELL STRUCTURE IN THE
CLASSIFICATION OF THE ORTHOTETIDINE

BRACHIOPODS

by ALWYN WILLIAMS Cind C. H. C. BRUNTON

Abstract. The secondary shell of the spire-bearing Davidsonia is fibrous, whereas in all true orthotetidine
brachiopods it is laminar. For this reason, Davidsonia and related genera, which constitute the Davidsoniidae, are
transferred to the spire-bearing brachiopods, the Atrypidina. The oldest known orthotetidines are impunctate,
but the Ashgillian Fardenia scotica rarely bears incipient pseudopunctae, which apparently arise through spiral
perpetuation of screw dislocations. This origin seems appropriate for orthotetoid pseudopunctae as a whole,
which have not yet been found to contain undoubted taleolae. Among schuchertellids, inwardly projecting
pseudopunctae were replaced by outwardly pointing extropunctae which could have evolved by changes in the
rate of shell secretion relative to a different kind of organic holdfast. Koskinoid perforations also penetrate most
orthotetidine shells, but they do so without deflecting lamination and were probably drilled mechanically by
boring organisms.

Assuming shell structure and the loss of a functional pedicle foramen each to have the same taxonomic weight
as all the morphological features developed for articulation and muscle support, phylogenetic analysis confirms
that the orthotetidines belong to two superfamilies: an older paraphyletic Chilidiopsoidea, and a younger
monophyletic Orthotetoidea. Both groups were affected by homeomorphic trends resulting from cementation
and conical deepening of the ventral valves of many independent stocks. They can, however, be distinguished
by phylogenetic analysis which provides cladograms consistent with their stratigraphic distribution.

The orthotetidine brachiopods have always been the subject of taxonomic confusion. The
distinctiveness shared by core genera, like Orthotetes Fischer de Waldheim, 1850, Hipparionyx
Vanuxem, 1842, and Streptorhynchus King, 1850, has never been in doubt, but their precise affinities
with other brachiopod groups have repeatedly given free rein to taxonomic practices bordering on
the eccentric. These have been well documented by Manankov (1979a) and will not be reiterated
except where they touch upon amendments of recent classifications which have led to the one being
offered here.

The first authoritative grouping of the orthotetoids within the Brachiopoda as a whole was that
presented by Schuchert (in Schuchert and Le Vene 1929, p. 16), who accepted the Orthotetinae of
Waagen (1884) as a strophomenoid subfamilial repository, not only for all orthotetoid genera then
known but also for an orthoid ( Orthidium ) with a vaguely ‘orthotetoid ’ cardinal process, and for all
resupinate strophomenides!

In the 1950s, when the Superfamily Orthotetacea was first proposed (Williams 1953, p. 9), a
number of families were erected by various students of the group so that, by the end of the decade,
seven such taxa were recognized (Williams 1953; Stehli 1954; G. A. Thomas 1958; Boucot 1959).
These new taxa largely clarified the definitive orthotetoid character states, although further
complications arose with the assignment to the Superfamily of the Davidsoniidae, Gemmellaroiidae,
Scacchinellidae and the Thecospiridae by Williams ( 1953), for no better reason than that they were
cemented, strophic stocks allegedly without spines but with a pseudopunctate shell.

By 1965, when the brachiopod volumes of the Treatise on invertebrate paleontology were
published, the Superfamily had acquired the name of Davidsoniacea in place of Orthotetacea, in
accordance with nomenclatorial rules of priority. The Gemmellaroiidae and Scacchinellidae had

(Palaeontology, Vol. 36, Part 4, 1993, pp. 931-966, 5 pis.]

© The Palaeontological Association

932

PALAEONTOLOGY, VOLUME 36

both been found to be spinose and had been removed to the Productidina (Muir-Wood and Cooper
1960, p. 66); and, although the Tnassic spire-bearing Thecospiridae had been retained in the
Superfamily, the orthotetoids could be generally typified as: strophic, biconvex to resupinate,
articulate brachiopods; initially with a supra-apical foramen which was lost in all younger,
cementing species; and invariably with a secondary laminar shell, initially impunctate but later
becoming pseudopunctate.

Since the Treatise study of the davidsoniaceans in 1965 (Williams, pp. H405-412), the number of
genera assigned to the group has more than doubled to one hundred and four. However, only
Cooper and Grant (1974) and Manankov (1979a) have offered a comprehensive revision of the
classification to cope with this generic proliferation. The most important steps taken by Cooper and
Grant were: (1) to transfer to the Strophomenidina all impunctate genera which were assembled,
within an amended Davidsoniacea, into two Families, the Davidsoniidae and Fardemidae (a junior
synonym of the Chilidiopsidae); and (2) to erect a new Suborder, Orthotetidina, for all
pseudopunctate genera which were grouped into two Superfamilies, the Orthotetacea and
Derbyiacea, containing seven families and seven subfamilies. On the other hand, Manankov (1979a)
retained both impunctate and pseudopunctate taxa within an amended Davidsoniacea which
embraced four families and ten subfamilies, after the removal of the Thecospiridae and its
promotion to the rank of superfamily with the Strophomenida.

The differences between these two classifications are much more fundamental than differences in
the number of suprageneric taxa recognized by their authors. Manankov’s classification is
essentially phylogenetic in theme and prompted him to identify several recurrent trends, especially
in changes in shell shape and in the elaboration of dental plates and cardinalia. In contrast. Cooper
and Grant (1974) paid much less attention to taxonomic complications that could have arisen from
the recurrence of parallel trends, particularly in the development of the cardinalia. Indeed, their
approach is basically monothetic, as is exemplified by their exclusion of all impunctate species from
their Orthotetidina (1974, p. 256). (Their reassignment of such stocks to the Strophomenidina
disregards that that Suborder is pseudopunctate par excellence.)

At present then, three very different, flawed classifications are being used in systematic studies of
orthotetidines and/or ‘davidsoniaceans’, as the case may be. Their deficiencies have now been
brought into sharp focus by our attempts to update the taxonomy of the orthotetoids for the revised
edition of the Treatise on the Brachiopoda. Thus the oldest classification currently in use, which was
drawn up by one of us (A.W.) for the first edition of the Treatise , is not only incapable of
accommodating all valid genera erected since 1965, but is also flawed in its use of shell structure for
taxonomic purposes. In particular, the classification did not take into account the important
discovery by Thomas (1958, p. 36) that in Streptorhynchus and allied genera microscopic conical
flexuring of the laminar shell points outwardly, not inwardly as in the true pseudopunctate
condition. To compound a more general indictment, our own studies of shell structure and
morphology and our use of phylogenetic analysis to assess the merits of the two other commonly
used classifications, those of Cooper and Grant (1974) and Manankov (1979a), reveal that they are
also deficient in one way or another. In fact, we now know enough about this seemingly
monophyletic group to conclude that orthotetidine phylogeny is much more complicated than can
presently be deduced from the variability and range of known species; and, as the course of
orthotetidine evolution becomes unravelled by future studies of new and extant collections, it will
continue to prompt changes in taxonomy. Nonetheless, classifications, ephemeral though they may
be, have to be put to the test. We have, therefore, decided to publish our current findings, before
committing ourselves to a final version for the Treatise , in the hope that the model will be improved
by the critical appraisal of a wider audience.

MATERIALS AND METHODS

Apart from recourse to literature, supplemented by the vast collections of specimens readily
available to one of us, this study entailed the preparation of material for the computer as well as

WILLIAMS AND BRUNTON: ORTHOTETIDINE BRACHIOPODS

933

for the electron microscope. It, therefore, seems appropriate to outline the processing of data for
both lines of investigation in this section.

Shell structure

The study of shell structure was undertaken in two ways. Gross identifications, like the presence of
pseudopunctae or koskinoid perforations, could adequately be made under the binocular
microscope. Consequently, the shell structure of all genera of strategic taxonomic importance was
routinely checked by this means except, of course, that of species which are presently known only
as moulds or silicified replacements. The data obtained from such surveys have been used in this
paper without reference to their precise source, unless they have been incorporated into text-figures.

More detailed studies to determine the basic biomineral units and any other microscopic features
of the shell successions were carried out under the scanning electron microscope (SEM). Some
whole or fractured specimens were examined in an environmental chamber (WETSEM), using an
ISI ABT 55 machine, and required no special preparation. The shell structure of many genera,
however, including all those illustrated in this paper, were studied under a Cambridge Stereoscan
360. For this purpose, some specimens were embedded in London resin and, after polymerization
of the resin, were cut along preferred planes, which were then polished with alumina (Gamma 100)
and etched in 2 per cent EDTA for about 30 minutes. Other specimens were broken to provide
fracture sections through the shell or fragments of external and internal surfaces. These pieces were
sonicated for 10 to 15 minutes, first in a weak detergent and then, after washing, in acetone, to
remove adherent particles before mounting on stubs. All such surfaces and etched sections were
coated with gold before examination under the microscope.

In addition to obtaining information on the shell structure of representative orthotetidines, it was
necessary to ensure that the features being studied were not the result of changes during the
fossilization and subsequent diagenesis of entombing sediments. Control specimens chosen to
monitor this possibility were either contemporaneous but unrelated species from the same
lithofacies and preferably the same locality, or species with well studied shell structures which could
serve as standards for comparison. Details of the specimens used in the SEM studies for this paper
are as follows:

Cranioidea

Neocrania anomala (Muller), Recent, near Oban, Scotland. L14924.

Petrocrania scabiosa (Hall), Upper Ordovician, Maysville Formation, Cincinnatti, Ohio, USA.
LI 4920a.

Strophomenida

Leptagonia caledonica Brand, Lower Carboniferous, Great Limestone Shale, Cocklaw Quarry,
Scotland. L10106/1.

Rafinesquina alternata (Hall), Upper Ordovician, Maysville Formation, Cincinnatti, Ohio, USA.
LI 49206.

Strophomena planumbona (Hall), Upper Ordovician, Trenton Group, Cincinnatti, Ohio, USA.
BMNH 73834.

Orthotetidina

Apsocalymma shiellsi McIntosh, Lower Carboniferous, Lower Limestone Group, Trearne
Quarry, Beith, Scotland. L14922.

Brochocarina trearnensis McIntosh, Lower Carboniferous, Lower Limestone Group, Trearne
Quarry, Beith, Scotland. B42729.

Fardenia scotica Lamont, Upper Ordovician, Lower Drummuck Subgroup, Craighead Inlier,
Scotland. L4835/40.

Orthopleura sp., Upper Devonian, Cedar Valley Limestone, Washington Highway 11, 12 miles
north of Cedar Rapids, Iowa, USA. L 1492 1 .

934

PALAEONTOLOGY, VOLUME 36

Schuchertella lens (White), Mississippian, Louisiana Limestone, Louisiana, Missouri, USA.
L14923.

Streptorhynchus pelargonatus (Schlotheim), Upper Permian, Gera, Germany. B9329.

Streptorhynchus pelicanensis Fletcher, Upper Permian, Kazaman limestone. Pelican Greek,
Queensland, Australia. B1749.

Xystostrophia umbraculum (Schlotheim), Middle Devonian (Eifelian), Gerolstein, Eifel, Germany.
B39585.

Atrypidina

Davidsonia verneuili Bouchard, Middle Devonian (Eifelian), Gerolstein and Romersheim, Eifel,
Germany. B5484, B39660.

Spiriferidina

Spinocyrtia astiolata (Schlotheim), Middle Devonian (Eifelian), Germany. B2677.

Repository numbers prefaced by L and B or BMNH refer to specimens housed in the Hunterian
Museum, Glasgow and The Natural History Museum, London, respectively.

Phylogenetic analysis

There are many reasons for attempting a comprehensive reclassification of the Orthotetidina at the
present time. Certain basic assumptions, which play a crucial role in shaping the three extant
classifications, are no longer tenable. The shell structure of genera assigned to the Orthotetidina
(and/or Davidsoniacea) has proved not to be exclusively impunctate or pseudopunctate as current
taxonomic practices dictate. Furthermore, convergent dental plates, which characterize many of the
later Palaeozoic stocks, did not always function as ‘spondylia’, although all such morphological
features are generally given the same taxonomic weight within a classification, whereas relatively
minor changes in the cardinalia may be assigned widely differing values. In short, although
recurrent transformations of shell shape and, concomitantly, of articulatory and muscle-bearing
devices were widespread, each of the prevailing classifications had been proposed in the expressed
belief that homeomorphy affected only those characters which were not important to the erection
of the favoured hierarchy!

In the face of such conflicting taxonomic treatment of homeomorphy, it was decided to reclassify
the orthotetidines by phylogenetic analysis, deriving the cladogram(s) by parsimonious means and
rooting it (them) to outgroups chosen on inferred symplesiomorphies. The program used (PAUP,
Version 3. On) was created and updated by David L. Swofford (January 1991). Each search for the
optimal tree (or equally parsimonious trees) was carried out heuristically with ten branch-swapping
entrances into the data set; and information was also sought on consistency and homoplasy indices,
apomorphic homologues and consensus cladograms of rooted trees.

Taxonomic and diagnostic data. Notwithstanding the flexibility of the PAUP program, the
orthotetidine data at our disposal were not instantly amenable to phylogenetic resolution at the
generic level. By 1992, according to the databases being maintained at the Smithsonian Institution,
Washington D.C. and the University of Glasgow, one hundred and four genera had been assigned
to the orthotetidine (s.l.) group of brachiopods. Of these, eighty-seven (including twenty-five
classified as junior synonyms) are, in our opinion, true orthotetidines; eight, including Davidsonia ,
are best assigned to the Atrypidina; two to the Strophomenidina (s.l.); and seven to the Articulata.

The sixty-two ‘valid’ genera constituting the orthotetidines Q.^.) can be uniquely described by
thirty-seven characters in two to five transformational states. These states were our version of the
variability in diagnoses distinguishing genera from one another. In effect, they varied among two
or more genera constituting the orthotetidine set although, as the same assemblage of characters
was used in analyses of superfamilial subsets, a changing minority became ‘uninformative’. They
defined transformations in: (1) shell size, shape and ornamentation; (2) shell structure; (3) the
cardinal areas, particularly delthyrial and notothyrial features; and (4) all internal features which.

WILLIAMS AND BRUNTON: ORTHOTETIDINE BRACHIOPODS

935

in this suborder, were exclusively those of articulation and musculature. The characters were not
programmed as being in ordered states, but a minority were weighted as twice the standard default
unit. These included: seven defining the articulatory and muscle supports in the program analysing
the Orthotetoidea (transformations in these characters are linked to the conical elongation of
ventral valves which was common in this superfamily); and four defining changes in the pedicle
foramen, shell structure and dental plates in the primitive Chilidiopsoidea.

Notwithstanding these adjustments, analysing a 62 x 37 matrix (exclusive of outgroups) would
have been a prohibitively formidable exercise for the Apple Macintosh Ilex at our disposal,
especially as we wished to vary the weighting of characters and the ordering of their states to test
the effects of certain assumptions. In any event, a matrix where the number of taxa is almost double
that of the characters identifying them is not amenable to meaningful analysis. We, therefore,
decided to explore the prospects for assembling suprageneric units into a taxonomic framework,
within which genera could be segregated into several, small groups for analysis.

The operational units chosen were the sum total of mutually exclusive families and subfamilies
recognized by Cooper and Grant (1974) and Manankov (1979a) in their classifications of the
Orthotetidina and/or Davidsoniacea. The Adectorhynchidae, which had been later erected by
Henry and Gordon (1985, p. 36), was also included, but not the monotypic Dorsoscyphinae
( Roberts 1 97 1 , p. 49), which, possibly through oversight, had been omitted from both classifications.
(The subfamily is cited as a synonym of the Derbyiinae in the classification proposed in this paper.)
The Thecospiridae (retained by Cooper and Grant in their classification, but see Brunton and
MacKinnon 1972) were also excluded, as were all spire-bearing genera assigned to the Davidsoniidae,
except Davidsonia itself which was retained as an outgroup. The Triplesiidae were also chosen as an
outgroup. This family, which is typically a biconvex, laminar-shelled stock with a supra-apical
foramen restricted by a monticular pseudodeltidium and a bilobed cardinal process, is assumed by
us to be the sister group of the Orthotetidina, descended from a billingselloid ancestor.

In all, four families and ten subfamilies constituted our terminal taxa. They could all be diagnosed
uniquely by an assemblage of fifteen characters in two to four transformations (Table 1). Of course,
the precision with which taxa can be so defined depends upon the way they are represented. Ideally,
a family (or subfamily) should be categorized by the sum total holomorphologies of its constituent
species. But it is not feasible to retrieve data (much of it imperfect) on this scale. We have, therefore,
assumed that each family (or subfamily) involved in our analysis is monophyletic and can be
adequately represented by its type genus and/or well described, closely related species. Admittedly,
this assumption rules out any immediate phylogenetic appraisal of the variability inherent in
monophyletic clusters of genera. However, this deficiency has been partly mitigated by our
subsequent use of all well-founded genera to refine, and determine the contents of, those suprageneric
taxa which survived the first round of analysis.

The choice of characters for the first stage segregation of orthotetidine taxa was determined partly
by the extent to which they have been used in previous suprageneric classification; and partly by
the new information provided herein, especially on shell structure. Leaving aside classifications
which are strictly monothetic, like those of Likharev (1932) and He and Zhu (1986), a general
consensus has emerged on which characters are reliable for taxonomic discrimination at the
suprageneric level. The presence then loss of a supra-apical foramen, signalling pedicle atrophy, the
distinction between fibrous and laminar secondary shell, and the development of pseudopunctae (or
extropunctae as defined in this paper) in an impunctate stock, have usually been perceived as
synapomorphies, at subfamily level at least. However, there is a wide divergence of opinion on the
relative taxonomic weight of most of these characters. The presence of pseudopunctae, for example,
is accorded subordinal and subfamilial recognition by Cooper and Grant (1974) and Manankov
(1979(7), respectively, in their placing of the Chilidiopsidae within the orthotetidine ( s.l .) hierarchy.

It is, nonetheless, universally conceded that all such characters are of greater taxonomic
importance than, say, incremental changes in the articulatory devices or in the muscle supports of
the shell, which have hitherto largely determined the structure of the orthotetidine taxonomic
hierarchy. Clearly, a differential weighting would have to be introduced to strike a balance between.

936

PALAEONTOLOGY, VOLUME 36

table 1. Characters used in the suprageneric classification of the orthotetidine brachiopods showing: types
(I = irreversible; O = ordered; U = unordered); weights (13 or 1); and states (0-4).

Type

Weight

State

I

13

(1)

supra-apical foramen

present (0), usually present in young stages (1), absent (2)

U

13

(2)

shell structure

impunctate (0), pseudopunctate (1), extropunctate (2)

U

1

(3)

dental ridges

discrete (0), sporadically convergent (1), homeospondylium (2)

o

1

(4)

dental plates

absent (0), short, apical (1), parallel (2), parallel, long (3), convergent (4)

o

1

(5)

spondylium

absent (0), sessile (1), with median septum (2), free (3)

u

1

(6)

ventral median septum

absent (0), low to variable (1), high (2), ankylosed to pseudodeltidium (3)

I

1

(7)

cardinal process bases

discrete (0), becoming fused in later stocks (1), with single shaft (2)

I

1

(8)

cardinal process lobes

separate, associated with chilidium and hingeline (0), fused, myophore slots
postero- ventral of chilidium and hingeline (1)

o

1

(9)

socket plates

absent or vestigial (0), short, variably disposed (1), recurved (2), recurved to
divergent (3), divergent (4)

o

1

(10)

fusion of socket plates

distinguishable from cardinal process lobes (0), fused with lobes (1)

u

1

(ID

chilidium

large, discrete plates or single convex arch (0), large, grooved arch (1), narrow
convex arch (2), narrow, grooved arch (3), vestigial or residual boss (4)

o

1

(12)

brachiophores

absent to vestigial (0), developing in later stocks (1), present (2), with
promontorium (3)

o

1

(13)

ventral umbo

symmetrical, low interarea (0), variable (1), asymmetrical, high interarea (2)

o

1

(14)

dorsal median septum

absent or vestigial (0), low (1), high with raised muscle margin (2)

u

1

(15)

radial ornamentation

coarsely costate or costellate (0), costellate (1), finely costellate (2), secondarily
plicate to costate (3), smooth to costellate (4)

for example, the five characters delineating the detailed morphology of the dorsal cardinalia and the
single character defining the microtexture of the shell. In the circumstances, we decided that
personal judgement could be at least as telling as reweighting characters commensurate with their
rescaled consistency indices. We, therefore, decided to give each of the two basic characters in our
analysis - the loss of pedicle and the development of pseudopunctae and extropunctae - a weighting
equal to the total scored by the other thirteen characters of the assemblage (Table 1).

The states of the majority of characters used for suprageneric analysis could also be placed in an
ordered transformation; in our estimation, some even irreversibly so. Irreversible transformations
were the loss of the pedicle and the proximal development of a shaft for an elevated pair of fused
cardinal process lobes. The recurrence of many of these trends during orthotetidine evolution was,
of course, responsible for the extraordinary amount of homeomorphy affecting the Suborder, which
is well shown by identifying ordered characters used in the analysis (Table 1).

WILLIAMS AND BRUNTON: ORTHOTETIDINE BRACHIOPODS

937

table 2. A matrix of sixteen suprageneric units (including the Triplesiidae and Davidsoniidae as outgroups)
x fifteen characters. It is based on taxa recognized by Cooper and Grant (1974), Manankov (1979) and Henry
and Gordon (1985), but diagnosed according to the interpretation of shell structure and morphology given in
this paper.

Before analysing the assembled matrix of sixteen suprageneric units x fifteen characters (Table 2),
the classifications of Cooper and Grant (1974) and Manankov (1979a) were each subjected to a
PAUP run involving only those families or subfamilies which were orthotetidine in our sense and
distinguished by thirteen sets of character states which had been used by all three authors to define
their suprageneric units. The triplesiid outgroup was used for both exercises. The operational
taxonomic units of the Cooper and Grant classification consisted of twelve families and subfamilies
which formed five equally parsimonious trees with a consistency index of 0-596 and a homoplasy
index of 0-404. Of the three superfamilies recognized by them, the Orthotetacea was polyphyletic
and the Derbyiacea and ‘ Fardeniacea’ (in lieu of Davidsoniacea) were paraphyletic in the sense of
Farris (1974). Ten orthotetidine families and subfamilies recognized by Manankov were used to test
his classification which was based on one superfamily. Two equally parsimonious trees were derived
with a consistency index of 0-788 and a homoplasy index of 0-212. Yet all three orthotetidine
families featured in the classification, the Orthotetidae, the Schuchertellidae and the Meekellidae,
were paraphyletic (Farris 1974).

The phylogenetic analysis of the suprageneric matrix shown in Table 2, identified nine equally
parsimonious trees of 132 steps. A strict consensus of the nine trees was then obtained. It revealed
that differences among the trees arose from transpositions in the two outgroups relative to the
Chilidiopsidae and from some variation in the phylogenetic distance between the Derbyoidinae and
the Hypopsinae and Orthotetellidae branch. One of the trees, which was comparable with the
consensus cladogram except for the placing of the outgroups relative to the Chilidiopsidae, was
chosen to provide further information on reconstructed states for internal nodes and the apomorphic
relationships between them and the terminal taxa.

The taxa composing the chosen tree (Text-fig. 6), which had consistency and homoplasy indices
of 0-652 and 0-348 respectively, segregated into three groups on the microtexture of the shell. They
were: (1) the impunctate Chilidiopsidae and Adectorhynchidae (including the Areostrophiinae);
(2) the pseudopunctate Orthotetidae (including the Pulsiinae), Orthotetellidae (including the
Hypopsiinae and Derbyoidinae), Meekellidae (including the Omboniinae) and Derbyiidae
(including the Diplaninae); and (3) the extropunctate Schuchertellidae and Streptorhynchidae.

111111

Adectorhynchinae

Pulsiinae

Hypopsinae

Derbyiidae

Orthotetidae

Areostrophinae

Chilidiopsinae

Derbyoidinae

Diplaninae

Schuchertellidae

Omboniinae

Meekellinae

Triplesiidae

Orthotetellinae

Streptorhynchidae

Davidsoniidae

123456789012345

200000214137212

210200002000071

210420002002211

211003214132212

212002002000012

200001102040102

100100001000011

210001002002112

210000212131200

220000002042111

210400214132212

210300214143213

000100210040104

210430004122112

220000214132221

000002000140224

938

PALAEONTOLOGY, VOLUME 36

The taxonomic validity of this microtextural segregation was then tested by phylogenetic analyses
of the genera that could be unhesitatingly assigned to one or other of the groups, which will be
conveniently referred to as the chilidiopsoid, orthotetoid and schuchertelloid groups. Not all genera
accepted as orthotetidines were involved in the exercise. Some of the more obvious junior synonyms
were withheld, such as the many genera erected by Likharev (1934) as variants of Derbvia Waagen,
1884. Others, however, were explicitly included to test the validity of synonymy as in the case of
Chilidiopsis Boucot, 1959, which is currently suppressed in favour of Coolinia Bancroft, 1949.
Poorly known genera which had been founded on inadequate diagnoses and/or material were also
excluded from the initial analyses. Thus, only twenty-four of the thirty-seven characters used to define
orthotetidine genera could be ascertained from the description and illustrations of Magicostrophia
Zhu (1985, p. 51). This lack of data increased the number of cladograms retained at the end of a
program involving Magicostrophia , without giving any indication which generic combinations were
attributable to the deficiency. However, when the genus was fed into the chilidiopsoid program
after the preferred cladogram had been derived, it was found to be synonymous with Iridiostrophia
Havlicek, 1965, although the nature of features presently unknown, like the pseudodeltidium and
chilidium, may eventually determine otherwise.

The same set of thirty-seven orthotetidine characters were used to build up a matrix for each
group. The numbers of orthotetidine genera involved were nineteen, twenty and ten for the
chilidiopsoid, orthotetoid and schuchertelloid matrices, respectively; while the Triplesiidae and the
pseudopunctate laminar-shelled Stropheodontidae (in place of the unrelated Davidsoniidae) served
as outgroups. Much of our information on genera was obtained from diagnoses and illustrations;
their variable quality is reflected in the low consistency indices of 0-474, 0-534 and 0-678 for three
chilidiopsoid, eight orthotetoid and four schuchertelloid equally parsimonious trees, respectively.
The generic tree chosen to typify each group was that nearest to the consensus cladogram. Finally,
the clustering of genera within each chosen tree was compared with the contents of currently
recognized subfamilies and families and attempts were made to reconcile or rationalize the many
differences between the cladograms and published taxonomic hierarchies of the orthotetidines as a
whole, although some genera required transfer from one group to another or even removal from the
Suborder.

Terminology

Morphology. The terminology used to describe orthotetidine morphology is essentially that
compiled for the brachiopod volumes of the Treatise on invertebrate paleontology (Williams and
Rowell 1965, pp. H139-H155). The glossary has been widely accepted and applied with little
emendation. However, Cooper and Grant (1974, pp. 255-256) claimed that a number of terms were
inconsistently defined; they coined new ones in their place or to describe features, especially of the
cardinalia, which in their opinion were important enough to warrant formal recognition. The
application of these new terms has caused difficulty. They have been ambiguously defined, especially
in relating the newly named structures to one another and to the cardinal process (Cooper and
Grant 1974, p. 352); also the labelling of ‘gusset’ in figure 40 (Cooper and Grant p. 351) is at
variance with the text and with relevant plate figures (compare Cooper and Grant, pi. 110, figs
18-22 with pi. 104, figs 13-17). Our understanding of the terms as used by Cooper and Grant is
based on their more succinct definitions (Cooper and Grant, pp. 257-260), insofar as they are
consistent with the plate figures.

We see no advantage in accepting the radical terminological changes proposed by Cooper and
Grant. Indeed, we share Manankov’s concern (1979, pp. 28-29) over their introduction. Their
terms, as well as those others used to define features that previously had not been formally named,
have been applied without regard for the dynamic relationship between shell and secreting
epithelium. Thus, the so-called ‘erismata’, introduced to distinguish divergent plates associated with
the cardinal process from those labelled ‘socket plates’ by Thomas (1958, p. 19, fig. 6), must have
been secreted in exactly the same way as the socket plates, irrespective of their early ontogenetic

WILLIAMS AND BRUNTON: ORTHOTETI DINE BRACHIOPODS

939

appearance. There are, of course, differences in the size of these plates and in the disposition of their
constituent facets, but these variations do not warrant a new terminology. We have, therefore,
continued to describe orthotetidine cardinalia in terms defined in the Treatise , which incidentally
have the same meaning as those of Thomas (1958, p. 9). Our correlation of such terms with those
employed by Cooper and Grant (1974, p. 351, Fig. 40) is given in Text-figure 1.

cardinal process
/ \
lobe shaft

cardinal process

x \

lobe shaft

text-fig. 1 . Stylized representations of the cardinalia of Meekella attenuata Cooper and Grant (1974, pi. 104,
fig. 16), identifying the terms used by these authors for the various parts of the structure in the right-hand
diagram and those used in this paper in the left-hand diagram.

We also share Manankov’s (1979) reservations about the interpretations offered by Cooper and
Grant (1974) on other orthotetidine morphological features, especially those involving ‘dental
plates’, ‘ridges’, ‘septa’ and ‘spondylia’. The definitions of these terms have been amended to take
into account whether the features, to which they refer, are of ‘primary’ or ‘secondary’ origin. This
distinction appears to be based solely on the size of silicified specimens, in which they were first
observed, as no shell sections have been described or figured. We do not, therefore, know the nature
of the secondary shell accretion or resorption which various features are alleged to have undergone.
In these circumstances, we have stuck to the more traditional definitions of the terms in question.

Taxonomy. In classifying the Brachiopoda, it has been the practice to use the suffix ‘-acea’ for
superfamilies. However, the International Code of Zoological Nomenclature has recently
recommended the general adoption of ‘-oidea’. This recommendation has been accepted by all
contributors to the revision of the brachiopod volumes of the Treatise on invertebrate paleontology.
It is implemented in this paper except when referring to superfamilies in the way they had been
taxonomically defined in published works.

ORTHOTETIDINE SHELL STRUCTURE

Ultrastructural studies of the shell provide information on its microtexture and its micromorph-
ology. The microtexture is the basic pattern resulting from the periodic secretion of biomineral
constituents by the outer epithelium of the mantle. This pattern can be modified by regularly
occurring micromorphological features which usually result from microscopic extensions or

940

PALAEONTOLOGY, VOLUME 36

invaginations of the mantle into the shell. Both types of microstructure play a crucial role in
brachiopod classification, and current investigations of them have also prompted a reappraisal of
orthotetidine phylogeny.

Microtexture of the orthotetidine shell

Our preliminary survey showed that the microtexture of Davidsonia verneuilli Bouchard was not
laminar in the manner of strophomenides in general and other orthotetidines in particular (Williams
1968, 1970, 1973). The taxonomic position of Davidsonia has been controversial since the discovery
of calcareous spiralia in Davidsonia (Garcia-Alcalde 1973). The two most authoritative
consequential reviews of Davidsonia itself have been contradictory, with Copper (1979) advocating
its transfer to the atrypidines and Johnson (1982) its retention within the orthotetidines. It was,
therefore, decided to compare the skeletal ultrastructure of Davidsonia with those of the spiriferide
Spinocyrtia astiolata (Schlotheim) and the orthotetide Xystostrophia umbraculum (Schlotheim)
from, the same Middle Devonian (Eifelian) successions of Gerolstein and Romersheim, to check the
effects of any diagenetic changes on shell microtextures.

The microtexture of the three specimens of Davidsonia available for study under the SEM was
fibrous. The specimens had been recrystallized so that the calcitic internal matrix formed a sharp
micritic boundary with the floors of the dorsal and ventral valves, which were 0-5 mm and 2 mm
thick respectively in the best preserved shell. Recrystallization, however, had not obscured details
of individual fibres, which were orthodoxly stacked and more or less radially disposed with some
flexuring (PI. 1, fig. 1). The fibres were up to 20 /tm wide and 7 /an thick and the externally facing
saddles were gently concave and about 6 pm wide.

This fibrous aggregation was characteristic of all fracture surfaces examined. Here and there,
however, fibres were interleaved with lenses of more vertically disposed components (PI. 1, fig. 2),
which were up to 35 pm thick and first appeared about 200 /mi internally of the outer surface of the
dorsal valve. These have been interpreted as impersistent lenses of prismatic calcite. Further study
of Davidsonia is likely to confirm a first impression that the margin of the ventral valve is thickened
by interleaves of prismatic calcite.

The microtexture of Davidsonia was identical with that of Spinocyrtia, except for the smaller size
of the fibres of the latter, seldom more than 10 pm wide. It was fundamentally different from the
microtexture of Xystrostrophia , which has been classified by Cooper and Grant (1974, p. 256) as a
davidsoniacean within the Strophomenidina and by Manankov (1979a, p. 30) as a meekellid within
the Davidsoniacea. The shell succession of Xystostrophia was laminar with individual laminae
thinner than 100 nm although usually aggregated into sets, up to 30 //m thick. Each lamina was
composed of an amalgamated array of parallel-sided, platy laths 2-3 //m wide (PI. 1, fig. 3). The only

EXPLANATION OF PLATE 1

Figs 1-2. Davidsonia verneuili Bouchard. Middle Devonian (Eifelian); Gerolstein, Eifel, Germany. 1, B39660;
external fracture surface of dorsal valve, showing orthodoxly stacked fibres of secondary shell with well
developed saddles directed externally, x 380. 2, B5484; polished and etched subradial section of dorsal valve,
with a lens of prismatic shell intercalated (submedially) within the fibrous succession of the secondary layer,
x 1470.

Figs 3—4. Xystostrophia umbraculum (Schlotheim). Middle Devonian (Eifelian); Gerolstein, Eifel, Germany;
B39585; external fracture surface and polished and etched subradial section of dorsal valve, showing
disposition and parallel-sided successions of impunctate cross-bladed laminae with some crested laths in the
section (fig. 4) especially towards the bottom right-hand corner, x 1200, x 1000.

Figs 5-6. Neocrania anomala (Muller). Recent; near Oban, Scotland; L14924; bleached interior of dorsal valve
showing, in general view and detail, concentrically packed laminae forming the walls of punctae, x 720,
x 1350.

All scanning electron micrographs.

PLATE 1

WILLIAMS and BRUNTON, Davidsonia , Xystostrophia , Neocrania

942

PALAEONTOLOGY, VOLUME 36

variations found were sporadic lenses of crested laths with gently convex outer surfaces and of
highly inclined laths which are being studied further (PI. 1, fig. 4). The laths within a set of laminae
were aligned in the same direction, which usually changed at acute angles from one contiguous set
to the next.

The microtexture of the Xystostrophici shell is identical with the standard cross-bladed laminar
successions of all strophomenides and productides (except for the fibrous but pseudopunctate
plectambonitoids and some early chonetidines). It is certainly typical of all orthotetoid shells which
have been studied ultrastructurally to date. For this paper, detailed microtextural surveys were
restricted to a few representative genera, although these are sufficiently distant from one another
phylogenetically to suggest that cross-bladed lamination is the hallmark of the Orthotetidina as
amended herein. There was some variation in lath width with ranges of : 2-5-5 //m for Apsoccilymma
shiellsi McIntosh; 3-4 pm for Fardenia scotica Lamont; 4-6 pm for Streptorhynchus pelargonatus
(Schlotheim); and 4—7-5 //m for Schuchertella lens (White); crested laths were also found in
Schuchertella. In general, however, one micrograph of the shell structure of these species was
indistinguishable from another so far as microtexture was concerned.

Micromorphology of the orthotetidine shell

Terminology. The mam micromorphological features of the orthotetidines are conical deflections of
the shell successions, which may point externally or internally; but before describing them, it seems
appropriate to outline our interpretations of the terms currently used for such features.

In general, our usage conforms to that of the Treatise (Williams and Rowell 1965, H139-H155),
except that the terms have been more precisely defined to take into account the new information
obtained since 1965. Thus conical deflections of shell successions which point externally are usually
referred to as punctae, on the assumption that they trace the paths of canals, accommodating
extensions of the mantle. There are, however, several kinds of extensions. Papillose outgrowths
(caeca) of the outer epithelium itself may either terminate at the periostracum, as in the Cranioidea,
or be separated from it by a canopy of shell perforated by microvillous canals, as in the
Terebratulida and Thecideidina. This difference in the termination of the canals accommodating
caeca warrants the restriction of the terms ‘puncta’ and ‘endopuncta’ to the cranioid and
terebratulide types respectively. A system of canals permeating the brachiopod shell can also result
from the secretion by outer epithelium of persistent proteinaceous strands. This system is especially
characteristic of the organo-phosphatic brachiopods; and, although they have been described as
‘punctae’, it is more informative to refer to them simply as ‘canals’ (Williams et al. 1992, p. 87).
Gaspard (1990, p. 54) has designated similar micro-morphological features, found in terebratulides.

EXPLANATION OF PLATE 2

Fig. 1. Petrocrania scabiosa (Hall). Upper Ordovician (Maysville Formation); Cincinnatti, USA; L14920n;
external fracture surface of dorsal valve, showing calcitic infill of puncta within secondary laminar layer,
x 1750.

Figs 2-5. Schuchertella lens (White). Mississippian (Louisiana Limestone); Missouri, USA; L14923.
2, polished and etched lateral subradial section of dorsal valve, showing part of an extropuncta with conical
deflections of secondary laminae directed towards the exterior beyond the lower edge of micrograph, x 550.
3^1, external surface of fragment of laminar secondary shell, with general view and detail of radially
arranged, externally directed tubercular structures of extropunctae, x275, x 2500. 5, internal surface of
fragment of laminar secondary shell showing conical depression of extropuncta delineated by spirally
disposed laminae, xll50.

Fig. 6. Apsocalymma shiellsi McIntosh. Lower Carboniferous (Lower Limestone Group); Beith, Scotland;
LI 4922; external view of fragment of secondary shell of ventral valve, showing disposition of laminae
around pseudopunctate depression filled with obliquely and spirally arranged laminae, x 1500.

All scanning electron micrographs.

PLATE 2

WILLIAMS and BRUNTON, Petrocrania, Schuchertella, Apsocalymma

944

PALAEONTOLOGY, VOLUME 36

as ‘micropuncta’. However, further study may show that, as in lingulides, they accommodate
secretory products rather than membranous extensions of the outer epithelium.

Inwardly directed conical deflections of shell successions, which form tubercles on the valve floor,
are pre-eminently characteristic of the strophomenides, productides and certain orthides. They are
unknown in living species, the tubercles of thecideidines and terebratulides like Megerlina being
unrelated, superficial outgrowths. Consequently, their inferred relationship with outer epithelium
has always been a source of controversy, as has been well described by Manankov (19796). Thus,
the cores of pseudopunctae may be occupied by calcite rods (taleolae), which were probably a
distinctive components of the shell in vivo. Pseudopunctae consisting exclusively of superimposed
cones are also found, reputedly interspersed with those with taleolae in many species, and have been
renamed ‘propunctae’ by Afaneseve (1980). However, we would not advocate the adoption of this
term until a comprehensive study has established the true relationship between pseudopunctae with
and without taleolae, as both kinds could have served as bases for fibrillar holdfasts of the mantle
(Williams 1968, p. 41). In that respect, we do not subscribe to the idea that pseudopunctate tubercles
acted as seats for ‘setae’ facilitating water flow within the mantle cavity (Grant 1968, p. 15). The
inner epithelium of the strophomenide mantle would have been densely ciliated in the manner of
living brachiopods and would have adequately performed all the functions envisaged for fimbriae.

In his survey of orthotetoid shell structure, Thomas (1958, p. 34) drew attention to the fact that
the ‘pseudopunctae’ of Streptorhynchus , which are arranged radially along the axes of costellae, are
deflected outwardly; he concluded that they were the sites of canals. In 1971 (p. 34), he recorded the
same type of structures in Schuchertella lens (White) from the type locality; and in a personal
communication (August 1992) he generously commented on his unpublished researches and listed
the genera in which he had found these outwardly deflecting structures (now described by him as
‘endopunctae’). They included Arctitreta and Kiangsiella as well as Schuchertella and Strepto-
rhynchus. Manankov (19796, p. 33) had already confirmed the existence of these microstructures in
Arctitreta , Kiangsiella and Streptorhynchus , but was content to continue referring to them as
pseudopunctae. Whether these outwardly deflecting features should be identified as ‘endopunctae’
or ‘pseudopunctae’ or should be given a new name, is dealt with later during discussion of our own
findings.

Micromorphology of representative orthotetidines. Well preserved specimens of Apsocalymma shiellsi
McIntosh and Brochocarina trearnensis McIntosh have been studied in detail to ascertain the
micromorphology of the orthotetid G.s.) shell. Pseudopunctae were openly distributed at 25-30 mm2
and were uniformly asymmetrical in profile (PI. 3, figs 2-3) in relation to the inferred stress couples
set up between the mantle and the thickening shell (Williams 1968, p. 39). Where seen in transverse
fracture sections on exfoliated surfaces, mature pseudopunctae formed rosettes, up to about 50 pm

EXPLANATION OF PLATE 3

Figs 1-3. Apsocalymma shiellsi McIntosh. Lower Carboniferous (Lower Limestone Group); Beith, Scotland;
LI 4922; 1, external view of fracture surface of secondary shell of ventral valve, showing disposition of
laminae around pseudopunctate infill of inclined laminar fragments, x 1800. 2-3, views of subradial fracture
section of ventral valve, showing laminar structure of inwardly projecting tubercles with externally facing
pseudopunctate depression in top left-hand corner of figure 3, x 370, x 570.

Fig. 4. Rafinesquina alternata (Hall). Upper Ordovician (Maysville Formation); Cincinnatti, USA; L149206;
external view of fracture surface, showing core of large pseudopuncta made up of recrystallized laminar
fragments, parts of which are still identifiable along left-hand margin, x 900.

Figs 5-6. Strophomena planumbona (Hall). Upper Ordovician (Trenton Group); Cincinnatti, USA;
BMNH 73834; weathered and partly exfoliated exterior of ventral valve, with general view of pseudopunctate
base on crest of costella flanked by granular interspatial depression (to left and right respectively of figure
5) and detail of pseudopunctate core composed of spirally inclined laminae (fig. 6), x 650, x 3000.

All scanning electron micrographs.

PLATE 3

WILLIAMS and BRUNTON, Apsocalymma , Rafinesquina , Strophomena

946

PALAEONTOLOGY, VOLUME 36

in diameter, of coriically disposed laminae which became increasingly inclined towards a core,
22-25 /mi across, consisting of a variety of calcitic structures. Up to twenty or so laminae and
laminar sets made up the concentric layering around the core which formed a horizontal floor of
solid calcite within a shallow hollow in some pseudopunctae seen from the exterior (PL 2, fig. 6).
In others, the floor was tilted into discrete laminar sets, an arrangement which was well seen in some
internal tubercles where tilted sets (PI. 3, fig. 1) were enclosed within successive laminar cones with
gently convex tops. Indeed, some tubercles were completely covered by dome-like laminae which,
although affected by some diagenetic changes, seemed to have been unbroken in the original state
(PI. 3, fig. 2). We, therefore, conclude that orthotetid pseudopunctae typically consisted of a
succession of superimposed laminar cones with gently convex peaks. Arrays of these cones have
been traced for almost 0-5 mm, throughout a shell etched by EDTA, along a sinuous path about
40 //m wide. There was no evidence of a core composed of anything other than the amalgamated
peaks of successive laminar cones.

Schuchertellid micromorphology differs from the pattern of other orthotetoids in several respects.
The immediately obvious difference is that although it also consists of arrays of asymmetrical
conical deflections, they invariably point externally not internally. In recognition of this and other
differences, which preclude any homology with punctation, we propose that these structures be
called ‘extropunctae’.

In specimens of Schuchertella lens (White) and Streptorhynchus pelicanensis Fletcher investigated
by us, the extropunctae were densely arranged, more or less radially (PI. 2, fig. 3), at about 150 per
mm2. On internal exfoliated surfaces, mature extropunctae formed shallow craters (PI. 2, fig. 5),
about 50 /an in diameter, bounded by up to ten laminar sets arranged concentrically about elliptical
cores, 4-5 pm in maximum diameter, and usually with a medial slot. Exceptionally, a single lamina
lined part of the crater sides and merged with the core as a spirally twisted band with a medial slot.
On external exfoliated surfaces, extropunctae occurred as low domes, up to 30 //m across (PI. 2,
fig. 4), which consisted of successions of curved laminae disposed around cores made up of oblique
or twisted plates, some with medial slots. Extropunctate trails were also revealed by etching a
transverse section of a shell with EDTA. These externally directed conical deflections, which were
about 50 /mi wide, were seldom more than 200 /mi long (PI. 2, fig. 2), although one could be traced
for over 0-6 mm. The impersistence of extropunctae in such sections is probably due more to their
sinuosity than to periodic lapses in their development.

The shell of Xystostrophia umbraculum has already been described as having a cross-bladed
laminar microtexture. It is further characterized by an absence of micromorphological deflections
of any kind; and this impunctate condition is typical of other genera assigned to the Chilidiopsidae
(or its junior homonym Fardeniidae) by Cooper and Grant (1974) and Manankov (1979o). This
impunctate condition had already been confirmed in Floweria prava (Hall) from the Upper
Devonian of Iowa, and Fardenia scalena Williams from the Caradocian of Scotland. However, the
present review also afforded an opportunity to ascertain the structure of the shell of one of the
earliest orthotetidines, Fardenia scotica Lamont from the Ashgillian of Scotland, with intriguing
results.

Fardenia is mostly represented in collections by moulds in a weathered siltstone, but a few
specimens with adherent shell occur among topotypes in the Hunterian Museum, Glasgow. A
fragment, about 2-4 mm long, adhering to the postero-median area of a dorsal valve, was prised
away from a weathered shell, and four microscopic slivers were dislodged at the same time. All were
mounted on the one stub for examination with the SEM. The large fragment and three of the four
slivers were impunctate; the fourth was pseudopunctate !

There can be no doubt that the sliver in question (PI. 4, fig. 5) came from the dorsal valve along
with the other pieces. All five fragments have a cross-bladed laminar microtexture with laminae
made up of monolayers of parallel-sided, amalgamated laths between 2 and 2-5 //m wide and
commonly aggregated into thick sets up to 3 pm or so. However, two other features confirm that
the pseudopunctate sliver was an integral part of the Fardenia shell. First, all fragments, and the
adherent shell from which they were dislodged, were relatively coarsely recrystallized so that the

WILLIAMS AND BRUNTON: ORTHOTETIDINE BRACHIOPODS

947

edge

text-fig. 2. Diagrammatic reconstruction of the origin and essential structure of a pseudopuncta, based on
those found on internal laminae of Fardenia scotica and illustrated in Plate 4, figure 6.

laminar surfaces were distinctly roughened by granules up to 300 nm in diameter (cf. PI. 4, fig. 6).
Secondly, the pseudopunctae are scattered along a gently arched feature about 250 jum wide, which
formed the long axis of the pseudopunctate sliver. This structure is an internally facing interspace,
comparable in attitude and wavelength with those underlying the sharply crested costellae of the
large fragment and of its counterpart mould on the dorsal valve.

The pseudopunctate sliver had an area of about 0-3 min2 and was less than 100 /;m thick. The
surface studded with pseudopunctae (PI. 4, tig. 5) was not part of the dorsal valve floor but an
assemblage of freshly exfoliated facets of about ten laminar sets. The sliver was, therefore, a piece
of the internal succession of the Fardenia shell. All the pseudopunctae were shallow and, at most,
immature in development because the dome-like laminae accommodating them were seldom more

948

PALAEONTOLOGY, VOLUME 36

than 40 pm across (PI. 5, fig. 1) while their cores, with diameters of about 10 //m, were usually
encircled by fewer than seven laminar sets. Indeed, the most immature one was less than 13 pm
across and consists of only four or five laminae around a core with a diameter of 5 pm. The most
interesting feature of this incipient pseudopuncta is that the core was really made up of two spirally
continuous laminae inclined towards a central slit which appeared to divide it into two halves
(PI. 4, fig. 6).

The discovery of pseudopunctae on a sliver of shell of Fardenia scotica prompted us to check
shell-bearing specimens of this and other Fardenia in Ordovician collections from Scotland and
Anticosti Island in Canada. Yet only one other of the six shells systematically examined for
micromorphological features under the SEM revealed any corroborative evidence: a solitary
pseudopuncta in the postero-median area of a ventral valve of F. scotica. We have, therefore,
concluded that Fardenia could be regarded as impunctate for classificatory purposes, but had a
genetic propensity for pseudopunctation, albeit sporadically in impersistent patches in that part of
the shell supporting the musculature.

The fragment has also presented a composite picture of the origin and development of at least
one type of pseudopuncta. Starting with the spirally arranged laminae at the core of the incipient
pseudopuncta, the most likely way for this arrangement to have originated would have been for a
cell with a diameter of about 5 //m to have started secreting, on an interlaminar membrane, not
calcite but fibrillar proteins or filaments connected by hemidesmosomes, and to have continued to
do so at a faster rate than the deposition of laminae by surrounding cells (Text-fig. 2). These rapidly
lengthening proteinaceous strands (or filaments) would, in turn, have caused adjacent cells to secrete
laminae in a steeply inclined coil around the organic strands to form the biomineralized core to the
growing pseudopuncta. From time to time the process would have been stopped by the cessation
of proteinaceous secretion which could have been selective or universal. This would account for
those pseudopunctate tubercles, outcropping on the floors or on internal exfoliated surfaces of
strophomenide valves, which are capped by entire as well as perforate laminae. This would not have
precluded the growth of succeeding pseudopunctae on the same sites and at new loci, a concurrence
which occurs during the maintenance of the micromorphological canal system pervading the shell
of Discina (Williams et a/. 1992, p. 98).

The postulated development of pseudopunctae in Fardenia is compatible with the micromorph-
ology of the extropunctae of Schuchertella as well as the pseudopunctae of Apsocalymma. The conical
deflections of both genera have cores consisting of tilted, discrete blocks of lamina; and spirally
disposed laminae have been found lining extropunctate craters of Schuchertella (cf. PI. 2, fig. 5).
Both types may also be capped by entire laminae which could only have been secreted during inter-
ruptions of the processes responsible for the differentiation of the cores.

EXPLANATION OF PLATE 4

Figs 1^4. Leptagonia caledonica Brand. Lower Carboniferous (Great Limestone Shale); Cocklaw, Scotland;
L10106/1. 1, general dorsal view of tubercles with taleolar cores on internal surface of ventral valve, x 100.
2, external view of transverse fracture section of pseudopuncta with roughened, pock-marked surface to
taleolar core, x 470. 3^1, general view and detail of polished and etched subradial section of ventral valve
showing disposition of laminae around taleolae (exterior towards the top); fully developed taleolar base
secreted uncomfortably on horizontal laminae seen in bottom right-hand corner of figure 3, x410; and
taleola occupying much of figure 4, separated from laminae of top-left hand and bottom right-hand corners
by patina and seamed with canals, x 1700.

Figs 5-6. Fardenia scotica Lamont. Upper Ordovician (Lower Drummock Subgroup); Craighead, Scotland;
L4835/40; general view and detail of granular internal surface of fragment of secondary shell of dorsal valve,
showing incipient pseudopunctae breaking through cross-bladed laminae in spirally disposed arrangements
perpetuating screw dislocations as in figure 6, x 200, x 2750.

All scanning electron micrographs.

PLATE 4

WILLIAMS and BRUNTON, Leptagonia , Fardenia

950

PALAEONTOLOGY, VOLUME 36

internal lamina

text-fig. 3. Diagrammatic reconstruction of the origin and essential structure of an extropuncta shown as a
variant of the pseudopuncta illustrated in Text-figure 2 (note differences of the extropunctae of Schuchertella

lens shown in Plate 2, figures 2-5).

There is, of course, a basic difference in the orientation of laminae around the cores. Thomas
(pers. comm. 1992) has contended that Schuchertella was ‘endopunctate’ not pseudopunctate. In
that respect, a comparative study of the punctate shell of Neocrania, which is laminar not fibrous,
is instructive. At the internal surface of a bleached valve, punctae are delineated by concentric bands
of laminae (PI. 1, fig. 5). But aggregations of these form cylinders not cones (PI. 1, fig. 6); and,
although they can be dislodged by the chemical degradation of interlaminar membranes, they do
not collapse into discrete blocks of laminae filling the canals which remain open to be filled by
sediment or diagenetic precipitates during fossilization, as can be seen in Petrocrania scabiosa (Hall)
(PI. 2, fig. 1). If, therefore, the fine structure of the extropunctate core is the same as that of a
pseudopuncta, the reversed orientation of deflection of the surrounding laminae must be due to
different rates of secretion of the organic components of the cores. Thus, more slowly growing

WILLIAMS AND BRUNTON: ORTHOTETIDINE BRACHIOPODS

951

taleola

text-fig. 4. Diagrammatic reconstruction of the origin and essential structure of a pseudopuncta containing
a taleola, shown as a variant of the pseudopuncta illustrated in Text-figure 2 (note differences of the taleolae
of Leptagonia caledonica shown in Plate 4, figures 1-4).

keratin filaments (with desmosomal attachments) may have been the dominant constituent in the
extropuncta compared with rapidly secreted strands of membraneous proteins in the pseudopuncta
(Text-fig. 3.).

The Fardenia model of pseudopunctate development can also be used to explain the growth of
the pseudopunctae of the Ordovician strophomenoids Rafinesquina alternata (Hall) and

952

PALAEONTOLOGY, VOLUME 36

Strophomena planumbona (Hall), which have been studied under the SEM for comparative
purposes. In the former species, rosettes may be as much as 150 pm across with a core about one-
third of that diameter (PI. 1, fig. 4). Yet there can be little doubt that the cores, even of these large
structures, are mainly composed of tilted blocks of laminae. The pseudopunctae of Strophomena are
smaller, with rosettes about 40 pm in diameter and cores one-quarter or so of that length (PI. 3, fig. 6).
However, the specimen studied was well-enough preserved externally to provide information on
the first-formed parts of Strophomena pseudopunctae. The cores and surrounds of those
pseudopunctae originating in the interspaces tend to be coarsely granular, which we have taken to
indicate recrystallized primary shell, at least in part (PI. 3, fig. 5). In contrast, pseudopunctae
exposed on exfoliated surfaces on the crests of costellae and beneath the ornamented superficial
layer of the valve, have cores composed of obliquely stacked laminar blocks (PI. 3, fig. 6) which,
apart from size, are closely comparable with those of Fardenia.

The pseudopunctae of the leptaenid Leptagonia caledonica Brand are quite different (PI. 4, fig. 1).
Rosettes of inwardly inclined laminae, which can exceed 75 //m in diameter, are grouped around
taleolae, up to 30 pm or so in diameter (PI. 4. fig. 2), which can frequently be traced throughout the
shell successions for T5 mm or more (cf. PI. 4, fig. 3). Taleolae are demonstrably different from
other pseudopunctate cores of laminar blocks or the matrix infill of punctae. A taleola is fully
developed and differentiated from the microtexture of the host shell when first formed and its
distinctiveness is further emphasized by the way its surface forms a calcified patina, which is sharply
separated from the surrounding laminae even when traces of interlaminar boundaries are preserved
upon it (PI. 4, figs 3-4).

The most startling difference, however, was brought out by etching polished sections with EDTA.
The bedded nature of the laminae was enhanced by etching, whereas a taleola became porous and
remained free of any laminar traces. The etched pits within the taleola were commonly delineated
by rhombohedral planes, but were clearly part of an interlacing series of canals, up to 300 nm in
diameter, permeating the entire feature (PI. 4, fig. 4). In the face of this evidence of heterogeneity
in its original composition, we concluded that a taleola, in vivo, consisted of a calcitic mesh
permeated by interconnected tunnels which were filled with organic materials; and was probably
bounded by a membrane continuous with those between the calcitic components of the laminar
succession (Text-fig. 4).

Tubercles with or without taleolae can also be partly or entirely capped by laminae, which,
temporarily at least, terminated taleolar growth. This, and the fact that our interpretation of a

EXPLANATION OF PLATE 5

Fig. 1. Fardenia scotica Lamont. Upper Ordovician (Lower Drummock Subgroup); Craighead, Scotland;
L4835/40; detail of internal surface of fragment of secondary shell (PI. 4, fig. 5), showing two
pseudopunctae, with spirally arranged laminae well seen in lower one, x 1880.

Figs 2-6. Koskinoid perforations. 2-4, Brochocarina trearnensis McIntosh; Lower Carboniferous (Lower
Limestone Group) ; Beith, Scotland ; B42729 ; fracture surface and section near ventral umbo ; 2, general view
of three perforations in transverse section with bulbous surface to lower infill and circumferential rubbly
surround to upper left, x 340; 3, canal with cleaved infill and boundary patina perforating undeflected
laminar succession, x 600; 4, circular infill representing tunnel in bottom left-hand corner connecting with
rubbly infill representing two lateral galleries separated by shelf of laminae which has been penetrated by the
perforation in top left-hand corner, x 475. 5, Streptorhynchus pelicanensis Fletcher; Upper Permian
(Kazanian Limestone); Pelican Creek, Queensland, Australia; B1749; oblique view of part of perforation
penetrating fracture section of secondary laminae, showing recrystallized core infill and the circumferential
rubbly zone, x 680. 6, Orthopleura sp.; Upper Devonian (Cedar Valley Limestone); Cedar Rapids, USA;
LI 4921 ; internal fracture surface (external view), showing transverse section of koskinoid perforation with
micritic interface between boundary laminae and recrystallized infill, x 1600.

All scanning electron micrographs.

PLATE 5

WILLIAMS and BRUNTON, Fardenia , Brochocarina, Streptorhynchus , Orthopleura

954

PALAEONTOLOGY, VOLUME 36

‘living’ taleola continues to involve the existence of a calcitic framework, affirms that the main
function of tubercles was to provide holdfasts for mantle filaments.

Koskinoid perforations

Microscopic perforations penetrating the ventral valves of the atrypidine Uncites and many
orthotetidine genera have been recognized for well over a century. The perforations tend to be
concentrated in the umbonal region; and, since perforated species invariably lacked a functional
pedicle opening, they have been variously interpreted as accommodating: (1) byssus-like threads
(Jux and Strauch 1966); (2) finely divided distal branches of mature or juvenile internal pedicles
(Schumann 1969; Martinez-Chacon and Garcia-Alcalde 1978); or (3) attachment fibrils secreted by
papillae of outer epithelium, which first made the perforations by shell resorption (Grant 1980).

Within the context of this paper, the origin of these perforations has to be explored, as all three
interpretations envisage features which could be critically important to orthotetidine classification.
Indeed, Grant (1980, p. 314) has gone so far as to transfer the impunctate Morinorhynchus to the
Orthotetoidea solely on the grounds that it is the only chilidiopsoid, known to him, which has
koskinoid perforations. In so doing, he has accorded these perforations greater taxonomic weight
than the combined morphological and other structural features of Morinorhynchus.

During our own studies of specimens representing thirty or so orthotetidine genera, we were able
to confirm a general but not a complete absence of koskinoid perforations from chilidiopsoids and
their presence in orthotetoids. We further confirmed Grant’s observations (1980, p. 315) that,
although the perforations were concentrated in umbonal regions, they also occurred on cardinal
areas and elsewhere on ventral valves (especially the flatter ones) but were absent from the dorsal
valves. However, ultrastructural studies on the perforations in the orthotetid Brochocarina
trearnensis McIntosh, the schuchertellid Streptorhynchus pelicanensis Fletcher and the chilidiopsid
Orthopleura sp. suggest that they may not have been a growth feature of the brachiopods bearing
them.

External and exfoliated surfaces, as well as fracture sections, show that the perforations are
normally orthogonal to the shell and occur as close clusters of near perfectly circular transverse
sections on laminar surfaces (PI. 5, fig. 2). In Brochocarina, twelve perforations were counted in
0-25 mm2, with an average diameter of 69 //m (for fourteen sections with a range of 52-78 pm). They
seldom overlapped and were normally dispersed at distances of 70-80 pm from one another,
although not in any discernible pattern. The most noteworthy aspect of the perforations is that they
had been neatly drilled through the laminar successions of the shell without any deflection or
general disturbance of the laminae themselves (PI. 5, fig. 3), other than rare fracture cleavage in the
vicinity of the perforations. In effect, the perforations are cylindrical tunnels seldom deviating from
the vertical. Except for their chimney-like openings at the external shell surface, which were 30 pm
or so deep, they were filled with recrystallized, cleaved calcite disposed irregularly as foliated
rhombs (PI. 5, fig. 4) or as more regular arrays of cleaved plates more or less parallel to the long
axes of the tunnels (PI. 5, fig. 3); only occasionally were the medial regions of the infill occupied by
irregular cavities, a few micrometres in size.

The sides of the tunnels, as revealed by oblique fractures, were relatively smooth and sharply
distinguishable from the infill (cf. PI. 5, fig. 6) which was bounded by either a micritic interface or
a circumferential rubbly wall about 10 pm thick consisting of smaller fragments of calcite. Some of
these fragments were joined with the laminae defining the tunnel walls by thin isthmuses of calcite,
but it was not possible to determine whether these junctions resulted from a post-depositional
recrystallization of wall and matrix or were the residues of chemical solution occurring during
tunnel formation.

In the samples at our disposal, the tunnels were almost always discrete. Even when overlap
occurred, it is evident that one tunnel had been superimposed on another at a later time and there
were few signs of branching. However, one vertical fracture section showed a horizontal gallery, up
to 175 /mi high and extending for more than 760 pm. The gallery, which contained a shelf of

WILLIAMS AND BRUNTON: ORTHOTETIDINE BRACHIOPODS

955

laminae nearly 20 pm thick, was continuous with at least two tunnels about 200 pm apart so that
all three structures were filled by an uninterrupted matrix of recrystallized calcite (PI. 5, tig. 4).

The koskinoid perforations found in Orthopleura and Streptorhynchus differ from those piercing
Brochocarina shells only in the diameters of their almost perfectly circular outlines in transverse
sections (PI. 5, figs 5-6). Ten of the more densely distributed perforations in Streptorhynchus
averaged 1 1-8 //m (range 6-8-13-0 //m), compared with 30-5 pm (range 26-34 /un) for seven such
structures in the antero-medial region of Orthopleura. In both genera, the recrystallized blocky
matrix filling the perforations do not deflect the laminae forming koskinoid walls. In effect, all
koskinoid structures studied by us are consistent with the features represented in Text-figure 5.

canal wall

exterior

gallery infill

canal infill

membrane

lamina

cross bladed lath

infill

interior

text-fig. 5. Diagrammatic representation of koskinoid perforations, based on those found in Brochocarina

trearnensis and illustrated in Plate 5, figures 2-4.

Assuming that the micromorphology of the koskinoid tunnels of Brochocarina , Streptorhynchus
and Orthopleura is typical of other perforate orthotetoids (and Uncites ), a number of constraints
now have to be observed in forming any view on their origin.

956

PALAEONTOLOGY, VOLUME 36

First and foremost, the absence of any ordered deflection of the laminae forming the walls of
koskinoid tunnels precludes the development of the perforations during the growth and thickening
of the shell. Had the tunnels accommodated byssus-like threads or branching pedicles, they would
have been lined with a membrane in continuity with the periostracum; and the differential secretion
of the thickening shell around each byssus thread or pedicle branch would have resulted in outward
conical deflections of the surrounding laminae and their interleaved membranes. In any event, the
orthotetidines belonged to an order characterized by a general atrophy of the pedicle, which must
have taken place before the emergence of cementing orthotetoids, so that, by the time koskinoid
perforations began appearing, even the pedicle epithelium would have become modified to secrete
an adhesive pad rather than any byssus-like structure (compare the development of Neocrania
(Nielsen 1991, p. 12)).

The assumption by Grant (1980, p. 317) that papillae of outer epithelium could have resorbed
koskinoid tunnels and then secreted fibrils in them is also untenable. Shell resorption in brachiopods
cannot be so finely focused as to drill neat holes through an alternating succession of calcitic laminae
and proteinaceous membranes as well as the external cover of tanned periostracum. In living
brachiopods, resorption patches, associated with the advance of muscle bases or the growth of
loops, are invariably surrounded by transitional zones of partially digested carbonate and
proteinaceous membranes, which are many microns wide. No surface as cleanly cut as the typical
interface between shell and koskinoid perforation has yet been attributable to resorption in living
species. Moreover, had the same, randomly distributed patches of outer epithelium later secreted
and sustained fibrils protruding through the koskinoid tunnels, the inner laminar successions
bordering such perforations would have been deflected outwards.

The rejection of any role for the mantle and pedicle of the host brachiopod in the formation of
its koskinoid perforations inevitably leads to the assumption that they were excavated by other
types of organism, a conclusion shared with Thomas (1958, p. 37). The lack of any pattern to their
distribution and of any regular interconnections suggests that the vertical tunnels were occupied by
solitary organisms seldom more than 1 mm or so long and 130 //m in diameter. Such an organism
would have been capable of grinding through calcite as well as proteinaceous membranes, to
account for the mechanically drilled nature of the perforations. Even so, the rarity of galleries
joining the vertical tunnels suggests that the organism did not live by digesting the shell itself but
was probably parasitic on the soft parts of its host.

This interpretation of koskinoid perforations does, of course, raise important questions which are
not easily answered. In particular: why such structures should be restricted to ventral valves and
concentrated in their umbonal regions; and why burrowing parasites, even if in symbiotic
association, should be so selective of their hosts as to be known only in the later orthotetidines and
Uncites.

We suggest that the ventral valve with its umbo cemented to, or buried within, substrates, would
always have been susceptible to invasion by infaunal infestations. This would account for the
relatively widespread distribution of perforations on flatter ventral valves and their absence from
the upper dorsal valves. (It might also account for the absence of perforations from
contemporaneous stropheodontoids which, although of comparable shape, were never cemented to
the substrate and were probably capable of repeated movement of the entire shell (Williams 1953,
P- 34).)

The apparent restriction of koskinoid perforations to later orthotetidines and Uncites is more
difficult to explain. Both stocks differ from pedicle-bearing brachiopods in their attachment to
substrates by cementation which, as already noted, would have facilitated infestation. On the other
hand, contemporaneous productidines were also anchored and immobile and yet escaped koskinoid
depredations. This could have been due to the relatively elevated habit of spinose productidine
shells. Surface settlement might also have been repulsed by the nature of the productidine
periostracum. In fact the periostraca of many extinct brachiopod groups might have been
sufficiently robust and antibiotic to have deterred shell entry by boring organisms (a point
overlooked by Owen and Williams (1969, p. 200) in their comparison of the distribution of

WILLIAMS AND BRUNTON: ORTHOTETIDINE BRACHIOPODS

957

burrowing polychaetes and sponges in the shells of Waltonia and Hemithiris). Moreover, if
burrowing parasites had been responsible for koskinoid perforations, they could also have infested
many late Palaeozoic groups of pedicle-bearing brachiopods without leaving any trace on the shell
by effecting entry at the junction between the pedicle and outer epithelium.

Tentative as our interpretation of koskinoid perforations may be, we feel that there is good reason
for excluding this feature from the lists of characters used to classify the orthotetidines.

ORTHOTETIDINE CLASSIFICATION

The phylogenetic tree of suprageneric taxa, shown in Text-figure 6, was constructed in accordance
with the criteria outlined in Materials and Methods and our interpretation of orthotetidine
morphology and shell structure. The terminal taxa of the cladogram include all established
orthotetidine subfamilies and families except for the Dorsoscyphinae Roberts, 1971, and the
Tropidelasminae Waterhouse, 1983, which are judged to be synonyms of the Derbyiidae and the
Streptorhynchidae respectively. The taxa were redefined in conformity with the character states
listed in Table 1. Their relationships within the tree clarify some issues but raise others.

The fibrous-shelled Davidsoniidae are taxonomically distant from the Triplesiidae (the other
outgroup) and the remaining orthotetidines, all of which are laminar shelled. The davidsoniids,
therefore, can no longer be classified as orthotetidines.

The impunctate chilidiopsoid group, embracing the Chilidiopsidae and Areostrophiidae (with the
Adectorhynchinae), is paraphyletic. It includes the oldest known orthotetidines in which the pedicle
remained functional in adult shells, although atrophy of the organ took place within the group and
was signalled by the later emergence of free-lying chilidiopsids with no trace of a pedicle foramen
and, in turn, by cementing areostrophiids.

The pseudopunctate orthotetoid group consists of the Pulsiinae, Orthotetinae, Derbyoidinae,
Orthotetellidae (with the Hypopsiinae), Derbyiidae (with the Diplaninae) and the Meekellidae (with
the Omboniinae). The pseudopunctate condition, which immediately distinguishes the group from
other orthotetidines, is invariably characteristic of at least the entire postlarval shell, albeit with
varying density. Even so, the group is paraphyletic with the Derbyiidae and Meekellidae rooted
with the extropunctate Schuchertellidae and Streptorhynchidae.

At first sight, this aggregation appears to support the recognition of four superfamilial groups:
(1) impunctate chilidiopsoids normally with small cardinalia; (2) pseudopunctate orthotetoids with
moderately developed cardinalia; (3) extropunctate schuchertellids with variably developed
cardinalia; and (4) pseudopunctate derbyioids with elaborate cardinalia. Such a classification would
have some common ground with that of Cooper and Grant (1974). In particular, they elevated the
Derbyiidae to a Superfamily, the Derbyiacea, for orthotetidines with elaborately developed socket
plates (their ‘erismata’, p. 259). These structures, however, must have evolved as repeatedly as
convergent dental plates or elevated cardinal processes, which would explain why the ‘Derbyiacea’
{sensu Cooper and Grant) contains such dissimilar groups as the orthotetellids and the strepto-
rhynchids.

Moreover, there is another basic reason for questioning the need to proliferate orthotetidine
superfamilies. The pseudopunctate orthotetoids and the extropunctate schuchertellids together
constitute a monophyletic group, immediately distinguishable from their ancestral, impunctate
chilidiopsoids. Accordingly, we propose that superfamilial recognition be restricted to the
Chilidiopsoidea and the Orthotetoidea with the latter embracing the extropunctate schuchertellids
and streptorhynchids as well as all pseudopunctate orthotetidines. This would rationalize the
taxonomic position of the schuchertellids (including the streptorhynchids) and, simultaneously, put
micromorphological changes affecting the orthotetidine shell into perspective. Certainly, the
previous classifications of this group could not have been more at odds. Williams (1965, p. H409)
united both stocks as subfamilies of the Schuchertellidae which were characterized as lacking dental
plates. Cooper and Grant (1974, p. 256) assigned the Schuchertellidae and Streptorhynchidae to the
Orthotetacea and Derbyiacea respectively on differences in their cardinalia. Manankov (1979u, p. 31)

958

PALAEONTOLOGY, VOLUME 36

Davidsoniidae

Triplesiidae

Chilidiopsidae

Adectorhynchidae

Areostrophiinae

Pulsiinae

Orthotetidae

Derbyoidinae

Hypopsiimae

Orthotetellidae

Schuchertellidae

Streptorhynchidae

Diplaninae

Berbyiidae

r— — Omboniinae
Meekellidae

text-fig. 6. Cladogram of fourteen widely recognized orthotetidine subfamilies and families (with the
Triplesiidae and Davidsoniidae as outgroups), diagnosed according to the fifteen character sets listed in Table 1

and derived from the matrix of Table 2.

having confirmed the researches on shell structure by Thomas (1958, p. 34), recognized the
close affinity between the Schuchertellidae and the Streptorhynchidae. Yet he also included the
impunctate Areostrophiinae within his version of the Schuchertellidae.

WILLIAMS AND BRUNTON: ORTHOTETIDINE BRACHIOPODS

959

The use of generic programs to check the contents of the constituent subfamilies and families of
the phylogenetic tree just described, resulted in a cladogram (Text-figure 7) based on the matrix of
Table 3, which we now propose as the basis for orthotetidine classification. The character states
determining the taxa are shown in Table 1 . The cladogram is one of nine equally parsimonious trees
of 128 steps and has consistency and homoplasy indices of 0-664 and 0-336 respectively. The changes
are more profound than appears at first sight for, in its compilation, a number of genera were
transferred from one family to another or indeed removed from the Orthotetidina altogether
( Schuchertellopsis Maillieux, for example, is probably an atrypidine). Generic reclassification,
however, is not within the province of this paper. That part of our analysis will appear in the
Treatise in due course, by which time it will certainly have undergone further changes to
accommodate the new genera proposed in the intervening period.

In contrast, the suprageneric classification offered here may prove to be comprehensive enough
to incorporate new taxa without disintegrating. Comparison of Text-figures 6 and 7 shows that the
suprageneric groupings in the latter largely retain their initial phylogenetic relationship as
determined by PAUP, although the suppression of the Derbyoidinae and the assignment of its
constituent genera to the Orthotetinae has reduced the branching of the preferred tree. It has also
been necessary to erect a new monotypic subfamily based on the earliest known orthotetidine,
Gacella. These changes, however, have not affected our version of the chronology of the main events
in orthotetidine history, which are outlined below.

The loss of a pedicle occurred early in the evolution of the most primitive orthotetidines, the
Chilidiopsoidea. The presence of a supra-apical foramen in the adult ventral valve seems to have
been restricted to early Upper Ordovician species of Fardenia and Gaceda of Scotland and Virginia.
Little is known about the occurrence of a functional foramen in young chilidiopsoids, except that
supra-apical sheaths have been seen by one of us (A.W.) in immature shells of Coolinia from the
Middle Silurian Waldron Shale of North America. However, other chilidiopsoids are almost
invariably symmetrical in outline without any sign of distortion of the ventral umbo through
cementation, and it seems safe to assume that they were free-lying on the substrate. This unattached
habit was also probably the mode of life of many early orthotetids like Pulsia , Schellwienella,
Orthotetes and related genera. Our inference is that a universal atrophy of the pedicle led to the
widespread distribution of unattached stocks. Many of these became cemented to the substrate and
independently developed distorted subconical ventral valves, a characteristic feature of the
Areostrophiidae, Orthotetellidae, Derbyiidae, Meekellidae and the Streptorhynchidae.

Micromorphological transformations of shell structure serve to particularize the emergence of a
monophyletic family, the extropunctate schuchertellids, but the phylogenetic status of the
pseudopunctate orthotetoid stocks relative to the impunctate chilidiopsoids is less certain. The rare
occurrence of the impersistent pseudopunctae in a few specimens of Fardenia , which is otherwise
impunctate, may be taxonomically unimportant, but it does support the assumption that the
orthotetoids descended from the chilidiopsoids. Of course, pseudopunctae could also have
developed in impunctate chilidiopsoids other than Fardenia ; and we concede the possibility that the
pseudopunctate condition was polyphyletic in origin, which would not be surprising in view of its
development among other strophomenide stocks.

The development of extropunctae is of phylogenetic interest in several respects. On parsimonious
grounds, one would expect the extropunctate condition to have been an apomorphy of the
impunctate state. Indeed, Manankov (1979a, p. 31) showed that the extropunctate Schuchertella as
having evolved from impunctate areostrophiids, and the pseudopunctate Schelhvienella (his stem
stock for the orthotetoids) as having descended from the chilidiopsids. The PAUP program,
however, consistently showed extropunctae as homologues of pseudopunctae and obliged us to
consider such a route for the micromorphological evolution of the orthotetidine shell. We
subsequently found that the outwardly directed extropunctae could feasibly have been derived from
inwardly directed pseudopunctae by assuming that an evolutionary change took place in the organic
components of these structures. Our interpretation can be tested, because a number of Permian
orthotetoid genera have been founded exclusively on silicified material, so that their shell

960

PALAEONTOLOGY, VOLUME 36

Triplesiidae

2

Gacellinae

1

3

H-

4, 5, 6,7

10

9

H—

12,13

14

16

Chilidiopsinae

A — Adectorhynchinae

8

Areostrophiinae

-I — Pulsiinae

11

— Orlholetinae

Hypopsiinae

■{ — Orlhoietellinae

15

Schuchertellinae

\ — Streptorhynchinae

8

H—

20

Diplaninae

Derbyiinae

Omboniinae

Meekellinae

text-fig. 7. Cladogram of fourteen orthotetidine subfamilies (with the Triplesiidae as an outgroup) derived
from the matrix shown in Table 3, based on characters as ordered, weighted and categorized in Table 1. The
major character changes enumerated are: 1, development of short, variably disposed socket plates; 2, short
socket plates parallel with hinge-line; 3, loss of pedicle foramen in adult shells; 4, loss of pedicle foramen;
5, loss of dental plates; 6, socket plates becoming recurved and longer; 7, ventral umbo distorted by cementation;
8, high cardinal process lobes directed postero-ventrally and supported by proximal shaft ankylosed to socket
plates; 9, development of pseudopunctate shell; 10, ventral valves seldom distorted by cementation;

WILLIAMS AND BRUNTON: ORTHOTETIDINE BRACHIOPODS

961

table 3. The matrix of fifteen suprageneric units x fifteen characters used in the orthotetidine classification
proposed in this paper. The Triplesiidae served as an outgroup. The matrix was derived from that of Table 2
after the redistribution of some genera (see text), which did not affect any of the character states defining the
taxa but did result in the suppression of the Derbyoidinae with consequential effects on the cladogram
of Text-figure 7.

111111

123456789012345

Adectorhynchinae

200000214137212

Pulsiinae

210200002000071

Hypopsinae

210420002002211

Derbyiidae

211003214132212

Orthotetidae

212002002000012

Areostrophinae

200001102040102

Chilidiopsinae

100100001000011

Diplanmae

210000212131200

Schuchertellidae

220000002042111

Omboniinae

210400213132212

Meekellinae

210300214143213

Triplesiidae

000100210040104

Orthotetellinae

210430004122112

Streptorhynchidae

220000214132221

Gacellinae

000200001000012

micromorphology is presently unknown. Future studies of unsilicified specimens of these genera will
afford a check of both the shell structure and the merits of the classification now being proposed.
In particular, we anticipate that Diplanus , which is provisionally assigned to the Derbyiidae, will
prove to be extropunctate and more akin to the Schuchertellidae.

The other significant changes in orthotetidine morphology were essentially elaborations of the
articulatory and muscle-supporting devices attendant upon increases in shell volume, especially
through the conical deepening of the ventral valve. The features which most obviously underwent
interrelated changes were the teeth ridges and dental plates, the bilobed cardinal process and the
socket plates. Other structures also underwent compatible changes. Modifications of the
pseudodeltidium and chilidium, for example, were in phase with the disposition of teeth ridges, and
especially of the posterior face of the base of the cardinal process; but only the development of the
socket ridges needs to be taken into account here.

The teeth ridges, which trace the growth of the hinge-teeth on either side of the delthyrial cavity,
were supported by short dental plates in the comparatively shallow chilidiopsid ventral valves. The
subsequent evolution of these dental structures seems to have been a widespread atrophy followed
by independent recurrences of excessively developed plates. The plates were lost with the emergence
of the areostrophiids, the orthotetids, the schuchertellids and the derbyiids. Within these groups,
exaggerated teeth ridges became convergent onto a ventral median septum to form the so-called
homeospondylium of the orthotetids and a more sporadically developed apical chamber among
adult Derbyia.

Post-chilidiopsid dental plates, on the other hand, arose secondarily on at least three different
occasions in the orthotetidine history, and with subtly different manifestations. Among the pulsiids.

1 1, development of strong, parallel dental plates; 12, socket plates becoming ankylosed to cardinal process;
13, brachiophores well developed; 14, dental plates becoming convergent to form spondylium; 15, development
of free spondylium and divergent socket plates; 16, development of extropunctate shell; 17, development of
high cardinal process with fused lobes supported by proximal shaft; 18, socket plates becoming larger and
divergent; 19, long dental plates becoming convergent; 20, cardinalia becoming flanked by promontoria.

Number of genera(inclusive):

Omboniinae

962

PALAEONTOLOGY, VOLUME 36

1 to 4

1 5 to 9
1 10 or more

short socket plates and:

iiiiiiiiiiiiiiii: discrete cardinal process lobes

WILLIAMS AND BRUNTON: ORTHOTETIDINE BRACHIOPODS

963

the plates were of variable length and bounded the ventral muscle field. In our opinion, the
orthotetellid dental structure evolved independently of that of the pulsiids, with a convergence of
plates eventually to form a free spondylium accommodating the entire ventral muscle field. The
convergent dental plates of the meekellids are superficially similar; but they must have developed
independently of those giving rise to the orthotetellid spondylium, because the bases of the diductor
muscles were inserted on the floor of the ventral valve, on either side of the convergent dental plates
of Meeke/la or of the septum formed by the convergence of the plates in Ombonia.

The transformation of the chilidiopsid cardinal process, consisting of a pair of low, discrete lobes
with broad, posteriorly facing myophores, to a high shafted, distally bilobed structure with slit-like
myophores facing postero-ventrally, was clearly related to the conical deepening of the ventral
valve. The elaboration of the cardinal process was, therefore, polyphyletic, with the high shafted
version characteristic of the late chilidiopsoid areostrophiids and the orthotetoid streptorhynchins,
orthotetellids, derbyiids and meekellids. In these families, the elaboration of the cardinal process
was accompanied by an extraordinary development of the socket plates (the erismata of Cooper and
Grant 1974, p. 259) and other associated features, especially the oblique socket ridges with their
brachiophore-like prolongation (the dentifers and ancillary plates of Cooper and Grant 1974). As
a result, long, divergent socket plates were united with the shaft of the cardinal process into a single
structure. Well developed socket ridges with ventral prolongations were ankylosed to the lateral
faces of this device; and were also flared postero-laterally in the meekellids (the promontaria of
Cooper and Grant 1974).

The cumulative effects of these trends are illustrated in Text-figure 8 in relation to the
stratigraphic ranges of the main orthotetidine groups. The chronology of the Suborder is broadly
consistent with the cladograms derived by phylogenetic analysis. The diagram also illustrates the
extent of homeomorphy during orthotetidine evolution, with no fewer than four of the nine terminal
families independently featuring cardinalia of closely comparable complexity.

CONCLUSIONS

The orthotetidines constitute one of the few suborders of the Brachiopoda characterized by several
basic differences in the ultrastructure of their shells. All true orthotetids have a secondary shell of
cross-bladed laminae bearing closely distributed pseudopunctae, composed of microscopic conical
deflections of the laminae which are directed inwardly. They evolved from the laminar-shelled
chilidiopsoids, which are impunctate except for a few specimens of late Ordovician Fardenia which
bear sporadically occurring, impersistent pseudopunctae. The pseudopunctate orthotetoids were in
turn ancestral to the laminar-shelled schuchertellids, which are extropunctate with radially
distributed microscopic conical deflections of the laminae pointing outwardly.

The typical orthotetoid shell is ultrastructurally indistinguishable from that of the strophomenids,
although the pseudopunctae arose independently in both stocks. So far as is known, the extro-
punctate condition is unique to the orthotetidines; however, pseudopunctae with taleolae, so
characteristic of the leptaenids, stropheodontids, chonetidines, productidines and related aberrant
Permian forms, have yet to be positively identified in orthotetidines.

The orthotetidines were also closely related to the other strophomenidines in many basic
morphological features. The presence, in the older species of both groups, of a pseudodeltidium with
a supra-apical foramen is indicative of the existence of a ventral body wall in the living state
(Williams 1956, p. 258), which was absent from other articulate brachiopods except for some
primitive orthides. The sealing-off of the foramen in all later Palaeozoic strophomenides confirms
that a universal atrophy of the pedicle had taken place throughout the Order by Carboniferous
times. Subsequently, many strophomenides (including the orthotetidines) acquired a cementing

text-fig. 8. Chronostratigraphy of orthotetidine phylogeny, based on the cladogram of Text-figure 7 and
showing the main trends in the evolution of the pedicle foramen, shell structure and cardinalia; all taxes outside

the designated boxes are pseudopunctate.

964

PALAEONTOLOGY, VOLUME 36

habit; and, since the davidsoniids were also cemented to the substrate and appeared to have a
pseudodeltidium, they were widely accepted as orthotetidines and, indeed, gave their name to the
Suborder under the priority rules of the International Code for Zoological Nomenclature. Yet, as
Johnson (1982, pi. 1, figs 11, 14) illustrated, the so-called pseudodeltidium is a deltidium and, with
the discovery that the shell is fibrous not laminar, Davidsonia and other related, cementing Middle
Palaeozoic brachiopods with calcareous spiralia must now be transferred, without further demur,
to the atrypidines.

The widespread acquisition by many orthotetidines of a cementing habit led repeatedly to the
elevation of their shells above the substrate by excessive conical deepening of the attached ventral
valves. This conical deepening was accompanied by complementary extensions of skeletal
articulatory devices. The morphological effects were quite dramatic, especially with regard to
variations in the proportionate development of the ridges and plates supporting the teeth, the
bilobed cardinal process accommodating the dorsal diductor bases, and other associated parts of
the cardinalia defining the dental sockets.

Not surprisingly, these repeated trends gave rise to similar, spectacular structures in several
independent stocks. The trends were broadly synchronous within a readily identifiable phylogeny
that was evidently compatible with the stratigraphic ranges of constituent taxa (Text-fig. 8). As a
result, previous classifications have been dominated by the preferential weighting of one kind of
feature, for example dental plates or socket plates, at the expense of others. In effect, homeomorphy
has played a more important role than homology in determining the structure of previous
orthotetidine classifications.

These homeomorphic trends can be disentangled by paying due regard to morphology as a whole
through phylogenetic analysis, and especially to the more stable changes attending the evolution of
shell structure. The classification proposed herein is an attempt to meet these conditions. Even so,
it is provisional on getting further information not only on the many poorly described genera
currently in circulation but also on taxa like Diplanus and Hypopsia , whose exquisitely silicified
morphology could well be at variance with their original shell structure which is as yet unknown.

Acknowledgements. We wish to thank all those who have provided us with advice, specimens and facilities
during the preparation of this paper. In particular, we are indebted to: Dr George A. Thomas of Victoria,
Australia, for access to his notes on schuchertelhd punctation; Dr Felicita d’Escrivan of the Museum of
Comparative Zoology, Harvard University, for the loan of specimens of Davidsonian Dr Richard E. Grant of
the National Museum of Natural History, Washington D.C., for specimens of Schuchertella and Orthopleura
for sectioning; Dr Sandra J. Carlson of the University of California for her advice on various aspects of the
typescript; and to Mr Peter Aynsworth and Mr Douglas Maclean of the Department of Geology and Applied
Geology of Glasgow University for access to a Cambridge Stereoscan 360 and for printing the micrographs
figured in this paper. The work was undertaken when one of us (A. W.) was in receipt of grants from the Royal
Society, British Petroleum and NERC; we greatly appreciate the enhancement of facilities afforded by these
awards.

REFERENCES

afaneseva, G. a. 1980. Pseudopunctate brachiopods of the superfamily Chonetacea. Paleontological Journal ,
14, 46-53. [Paleontologicheskii Zhurnal, 2, 52-58].

boucot, a. j. 1959. A new family and genus of Silurian orthotetacid brachiopods. Journal of Paleontology, 33,
25-28.

brunton, c. h. c. and mackinnon, d. I. 1972. The systematic position of the Jurassic brachiopod Cadomella.
Palaeontology , 15, 405^4 1 1 .

cooper, g. a. and grant, r. e. 1974. Permian brachiopods of West Texas, II. Smithsonian Contributions to
Paleobiology , 15, 289-331.

copper, p. 1979 [1978]. Devonian atrypoids from western and northern Canada. Geological Association of
Canada , Special Paper , 18, 233-793.

farris, J. s. 1974. Formal definitions of paraphyly and polyphyly. Systematic Zoology , 25. 271-282.

WILLIAMS AND BRUNTON: ORTHOTETIDINE BRACHIOPODS

965

garcia- alcalde, J. L. 1973. Braquiopodos Devonicos de la Cordillera Cantabrica (5). El aparato braquial de
Davidsonia Bouchard-Chantereaux, 1849 (Strophomenida, Davidsoniacea). Breviora Geologica Asturica, 17,
1-5.

gaspard, d. 1990. Diagenetic modifications of shell microstructure in the Terebratulida (Brachiopoda,
Articulata). Pis 135-140. In carter, j. g. (ed. ). Skeletal biomineralization: patterns, processes and
evolutionary trends. Volume II, Atlas and Index. Van Nostrand Reinhold, New York, 101 pp., 200 pis.
grant, r. e. 1968. Structural adaptation in two Permian brachiopod genera. Salt Range, West Pakistan.
Journal of Paleontology , 42, 1-32.

1980. Koskinoid perforations in brachiopod shells: function and mode of formation. Lethaia , 13,
313-319.

he, x. and zhu, m. 1986. Evolutionary tendency and classification of the Superfannly Orthotetacea Williams.
95-101. In racheboeuf, p. r. and emig, c. (eds). Les brachiopodes fossiles et actuels. Biostratigraphie du
Paleozoique, 4. 500 pp.

henry, T. w. and Gordon, m. 1985. Chesterian davidsoniacean and orthotetacean brachiopods, Ozark region
of Arkansas and Oklahoma. Journal of Paleontology, 59, 32-59.

Johnson, J. G. 1982. Davidsonia and Davidsoniacea, Devonian strophomenid brachiopods. Journal of
Paleontology, 56, 973-976.

jux, u. and strauch, f. 1966. Die mittledevonische Brachiopoden-Gattung Uncites De France, 1825.
Palaeontographica , 125, 176-222.

likharev, b. k. 1932. Fauna of the Permian deposits of northern Caucasus, I. Brachiopod subfamily
Orthotetinae (Waagen). Trudy Vsesoyuznogo Geologo- Razvedochnogo Ob' edineniya NKTP SSSR, 215, 1-54.
[In Russian and English],

— 1934. On some new genera of Upper Palaeozoic Brachiopoda. Comptes Rendus de PAcademie des
Sciences de PURSS, N.S., 1, 210-213. [In Russian and English].

manankov, I. N. 1979a. Strophomenida of the late Palaeozoic of Siberia and the Arctic. Trudy
Paleontologisheskogo Instituta, Akademiia Nauk SSSR, 171, 1-101. [In Russian and English].

— 1979b. Pseudopunctae in Strophomenids. Paleontological Journal, 13, 332-338. [Paleontologicheskii
Zhurnal , 3, 72-78],

martinez-chacon, M. L. and garcia-alcalde, J. L. 1978. La genesis del koskinoide en braquiopodos
articulados. Revista de la Facultad de Ciencias, Universidad de Oviedo , 17-19, 261-279.
muir-wood, h. m. and cooper, G. a. 1960. Morphology, classification and life habits of the Productoidea
(Brachiopoda). Geological Society of America, Memoir, 81, 1-447.
nielsen, c. 1991. The development of the brachiopod Crania ( Neocrania ) anomala (O. F. Muller) and its
phylogenetic significance. Acta Zoologica, 72, 7-28.

owen, G. and williams, a. 1969. The caecum of articulate brachiopods. Proceedings of the Roved Society of
London , Series B, Ml, 187-201.

Roberts, J. 1971. Devonian and Carboniferous brachiopods from the Bonapart Gulf Basin, northwestern
Australia. Bureau of Mineral Resources, Geology and Geophysics, Bulletin, 122, 1-319.

SCHUCHERT, C. 1929. In schuchert, c. and le vene, c. Fossilium catalogus. 1. Animalia. Pars 42: Brachiopoda.
Junk, Berlin, 140 pp.

Schumann, D. 1969. ‘ Byssus '-artige Stielmuskel-Konvergenzen bei artikulaten Brachiopoden. Neues Jalubuch
fur Geologie und Paldontologie, Abhandhmgen, 133, 199-210.
stehli, F. G. 1954. Lower Leonardian brachiopods of the Sierra Diablo. Bulletin of the American Museum of
Natural History, 105, 257-358.

swofford, d. 1991. PAUP ( Phylogenetic Analysis Using Parsomony ) version 3. On. Distributed by the author,
Illinois Natural History Survey, 607 East Peabody Drive, Champaign, II. 61820, USA.
thomas, G. a. 1958. The Permian Orthotetacea of Western Australia. Bureau of Mineral Resources, Geology
and Geophysics, Bulletin, 39, 1-159.

— 1971. Carboniferous and early Permian brachiopods from Western and Northern Australia. Bureau of
Mineral Resources, Geology and Geophysics, Bulletin, 56, 1-274.

waagen, w. h. 1884. Salt Range fossils. Productus- Limestone fossils: Brachiopoda. Palaeontologica Indica,
4(3), 547-610.

williams, A. 1953. The classification of the strophomenoid brachiopods. Journal of the Washington Academy
of Sciences, 43, 1-13.

1956. The calcareous shell of the Brachiopoda and its importance to their classification. Biological
Reviews, 31, 243-287.

1968. A history of skeletal secretion among articulate brachiopods. Lethaia, 1, 268-287.

966

PALAEONTOLOGY, VOLUME 36

williams, a. 1970. Origin of laminar-shelled articulate brachiopods. Lethaia , 3, 329-342.

1973. The secretion and structural evolution of the shell of thecideidine brachiopods. Philosophical
Transactions of the Royal Society of London , Biological Sciences, 264, 439 — 478.

1965. Superfamily Davidsoniacea King, 1850. H403-H412. In moore, r. c. (ed.). Treatise on invertebrate
paleontology. Part H. Brachiopoda. Geological Society of America and University of Kansas Press, Boulder,
Colorado, and Lawrence, Kansas, 927 pp.

— mackay, s. and cusack, m. 1992. Structure of the organophosphatic shell of the brachiopod Discina.
Philosophical Transactions of the Royal Society of London , Series B, 337, 83-104.

— and rowell, a. j. 1965. Morphological terms applied to brachiopods. H139-H155. In moore, r. c. (ed.).
Treatise on invertebrate paleontology. Part H. Brachiopoda. Geological Society of America and University
of Kansas Press, Boulder, Colorado, and Lawrence, Kansas, 927 pp.

zhu, c. 1985. Late Ordovician brachiopods from Wulong-tun district of Eastern Da Hinggan Ling Region,
Northeast China. Bulletin of the Shenyang Institute of Geology and Mineral Resources, 12, 49-53. [In
Chinese].

ALWYN WILLIAMS
Palaeobiology Unit

Department of Geology and Applied Geology
University of Glasgow
8 Lilybank Gardens
Glasgow G12 8QQ, UK

C. H. C. BRUNTON

Typescript received 26 January, 1993
Revised typescript received 15 March 1993

Department of Palaeontology
The Natural History Museum
Cromwell Road
London SW7 5BD, UK

THE STATUS OF THE NOTHOSAURIAN REPTILE
ELMOSAURUS LELMENSIS , WITH COMMENTS ON
NOTHOSAURUS MIRA BILIS

by OLIVIER RIEPPEL

Abstract. The type and only known specimen of Elmosaurus lelmensis Huene was originally described as a
pachypleurosauroid and plesiomorphic sauropterygian, but is here redescribed and identified as skull fragment
of Nothosaurus cf. mirabilis. It shares with the Nothosauridae the presence of two caniniform teeth on the
maxilla. Elmosaurus is compared with a newly prepared and as yet undescribed skull of Nothosaurus mirabilis
Munster.

In 1957, Huene described a new genus and species of sauropterygian reptile, Elmosaurus lelmensis,
which had been found by H. Wehrmann in 1930 along a footpath west of Lelm, near Braunschweig,
Germany. The specimen comes from the upper Ceratites- layers of the Hauptmuschelkalk (Upper
Muschelkalk, Ladinian). It consists of a partial skull, broken anteriorly in a transverse plane just
in front of the external nares, and posteriorly along an oblique line of fracture passing through the
anterior corner of the left upper temporal fossa, continuing through the interorbital space and
in front of the right orbit. During preparation of the fossil, impressions of the anteromedial margin
of the left upper temporal fossa were noted in the surrounding matrix, as well as what Huene (1957,
p. 92) interpreted to be the left lateral margin of the parietal foramen (Text-fig. 2a). These
impressions were cast, and the cast subsequently attached to the fossil. The palate is exposed in
ventral view, but badly eroded.

Huene (1957) recognized the sauropterygian affinities of the fossil, but noted problems in the
analysis of its relationships within the group. He concluded (Huene 1957, p. 97) that the genus
represents a 'primitive pachypleurosaurid’, and 'the most primitive nothosaurian’ known to be
derived from a pelycosaurian ancestor. His conclusions notwithstanding, the systematic status of
Elmosaurus continued to be problematical. Carroll (1988) treated the taxon as a nothosaur incertae
sedis, while Storrs (1991, p. 135) considered it as a 'possible early offshoot of the Sauropterygia’.

In view of its allegedly plesiomorphic status within the Sauropterygia, a redescription of the
specimen seems justified in order to assess its significance for the analysis of sauroptergyian
interrelationships.

SYSTEMATIC PALAEONTOLOGY

Superorder sauropterygia Owen, 1860
Family nothosauridae Baur, 1889
Genus nothosaurus Munster, 1834

Type species: Nothosaurus mirabilis Munster, 1834, from the Upper Muschelkalk, (Middle Triassic), Germany

Nothosaurus cf. mirabilis
Text-figs 1-4

1957 Elmosaurus lelmensis, Huene, p. 97.

1988 Elmosaurus lelmensis, Carroll, p. 619.

1991 Elmosaurus lelmensis, Storrs, p. 135.

[Palaeontology, Vol. 36, Part 4, 1993, pp. 967-974|

© The Palaeontological Association

PALAEONTOLOGY, VOLUME 36

text-fig. 1. The holotype and only known specimen of Elmosaurus lelmensis Huene, 1957 (SMNS 59077);

dorsal (a) and ventral (b) views x 1.

Material. Staatliches Museum fur Naturkunde, Stuttgart (SMNS) 59077, a partial skull (Text-figs 1-3), the
holotype of Elmosaurus lelmensis Huene. The specimen formerly belonged to the geological collections of the
Technische Hochschule in Braunschweig. The original label attached to the specimen reads as follows: 1930

RIEPPEL: NOTHOSAURIAN REPTILE

969

text-fig. 2. The skull of Elmosaurus lelmensis Huene, 1957 ; dorsal (a) and ventral (b) views. The hatched part
represents the plaster case of the natural mould left on the matrix by the eroded skull table. Abbreviations:
can, caniniform teeth on maxilla; f, frontal; ju, jugal; m, maxilla; n, nasal; pf, postfrontal; ‘pin’, pineal
foramen as identified by Huene 1957; pi, palatine; pm, premaxilla; p, postorbital; prf, prefrontal; v, vomer.

Scale bar = 20 mm.

Mixosaurus. Ob. Muschelkalk, Ob. Ceratitenschichten, Lelm (Elm), Steinhaufen, Feldweg. Sammlung
Wehrmann, Braunschweig; SMNS 59074, a previously undescribed skull (Text-fig. 4) referred to Nothosaurus
mirabilis Munster.

Locality and horizon. SMNS 59077, upper Ceratites-\ayers, Upper Muschelkalk (Ladinian), Lelm near
Braunschweig, Germany. SMNS 59074, spinosus zone. Upper Muschelkalk (Ladinian), Hegnabrunn near
Kulmbach, Germany.

Description

The skull in dorsal view (Text-figs 1a, 2 a). The preorbital region of the skull shows the fully preserved nasal
bones, which meet along the dorsal midline of the snout in an extensive suture, thus broadly separating the
posterior (nasal) processes of the premaxillae from the frontal bone. Between the external nares, the nasal

970

PALAEONTOLOGY, VOLUME 36

bones are reduced to slender anterior processes which form most of the medial margins of the external nares,
terminating in their anteromedial corner. The premaxillae form pointed posterior (nasal) processes, entering
between the nasal bones and extending to the level of the posterior margin of the external nares. Posteriorly,
the nasal bones are drawn out into pointed tips reaching a level well behind the anterior margin of the orbits,
and embracing between them a small anteromedial projection of the frontal bone.

The premaxilla enters the anterior margin of the external naris and meets the maxillary bone at the
anterolateral edge of the external naris. The maxilla thus forms most of the lateral margin of the external naris,
as well as the latter’s posterior margin as it extends dorsally to meet the lateral edge of the nasal bone. More
posteriorly, the maxilla enters the anterior margin of the orbit and extends further posteriorly below it to form
its lower margin, meeting the postorbital bone in the posteroventral corner of the orbit. The snout on the whole
appears slightly constricted at the level of the external naris.

The frontal bone is unpaired. It enters the dorsal margin of the orbit between the prefrontal and the
postfrontal bones. Anterolaterally, the frontal forms an elongate and tapering process extending well beyond
the level of the anterior margin of the orbit, thus entering deeply between the nasal and maxillary bones,
separating the prefrontal from the nasal. The prefrontal bone lines the anterodorsal margin of the orbit. It
gains a very limited exposure only on the dorsal surface of the skull.

The postfrontal bone participates broadly in the formation of the posterodorsal margin of the orbit. More
posteriorly, however, it narrows abruptly to a slender posterior process which ends at the posterior oblique line
of fracture. The relation of the postfrontal to the anteromedial margin of the upper temporal fossa remains
unknown. Below the postfrontal, the postorbital participates in the formation of the posteroventral margin of
the orbit. It extends posteriorly, forming a broad postorbital arch and can be observed to define the
anterolateral margin of the upper temporal fossa. The ventral contact of the postorbital with the maxilla
excludes the jugal bone from the posterior margin of the orbit. The jugal is represented by its anteriormost part
only.

The supplementing cast, attached to the posterior end of the fossil along the posterior oblique line of
fracture, shows the anteromedial margin of the upper temporal fossa in continuation with the postorbital. As
preserved, the anterior corner of the upper temporal fossa appears narrow and distinctly smaller than the orbit.
Huene’s (1957) identification of the lateral margin of the pineal foramen in a very anterior position, at about
the level of the posterior margins of the orbits, cannot be corroborated. All that can be seen is a small nick
in the margin of the supplementing cast with no clear anatomical relation.

The skull in ventral view (Text-figs 1 b, 2b). The bones of the ventral surface of the skull are badly eroded, but
remains of both maxillae, of the vomers, and of the left palatine can be identified. The left internal naris and
its surrounding bones are well preserved. The narrow vomer forms most of its medial margin, and meets the
maxilla in a suture at the midline of its anterior margin. The maxilla lines the internal naris laterally, while the
palatine bone is seen to enter the posterior margin of the choana. The broken roots of two distinctly enlarged,
i.e. caniniform, teeth can be identified in the maxillary bone at a level of the posterior part of the internal naris.

table 1. Measurements of the holotype of Elmosaurus leimensis Huene (to the accuracy of 05 mm). Values in
parentheses refer to the right side of the skull.

Total maximal length of skull fragment 100

Width of skull between external nares 32

Longitudinal diameter of external naris 16-5 (17)

Transverse diameter of external naris 9-5 (10)

Longitudinal diameter of internal naris 19

Transverse diameter of internal naris 6-5

Longitudinal diameter of orbit 28

Transverse diameter of orbit 22

Distance between external naris and orbit 14-5

Distance between orbit and upper temporal fossa 18

Width of bony bridge between external nares 7

Width of interorbital space (at posterior tips of prefrontals) 12-5

RIEPPEL: NOTHOSAURIAN REPTILE

971

DISCUSSION

The description of SMNS 59077 given above, as well as the corresponding reconstruction (Text-fig. 3),
differs markedly from Huene’s (1957) description of Elmosaurus lelmensis and refutes the
hypothesis of pachypleurosauroid relationships, as well as the assumption of a plesiomorphic status
of the genus within the Sauropterygia.

SMNS 59077 differs from pachypleurosauroids (Carroll and Gaskill 1985; Rieppel 1989; Sander
1989) with respect to almost all aspects of its morphology, including the constriction of the snout,
shape and relations of the nasal bones (which in pachypleurosaurs do not extend along the medial
margin of the external naris), the small size of the prefrontal bones, the shape and relations of the
postfrontal bone and of the jugal bone, the latter forming a slender and curved element defining
most of the ventral and posterior border of the orbit in pachypleurosaurs. Indications of
heterodonty, and in particular the presence of paired maxillary fangs, is additional evidence against
pachypleurosauroid affinities and suggests nothosaur relationships instead.

Indeed, the Nothosauridae (sensu Tschanz 1989), and Nothosaurus in particular ( Paranothosaurus
may well be congeneric with Nothosaurus : Kuhn-Schnyder 1966; Storrs 1991; Rieppel and Wild
unpublished), are characterized by shared derived characters such as the exclusion of the jugal from
the posterior margin of the orbit, and the presence of two caniniform teeth in the maxilla. In
addition, SMNS 59077 shares with Nothosaurus the configuration of circumorbital bones, as well
as the arrangement of elements around the external naris and the choana. It is therefore concluded
that SMNS 59077 represents the partial skull of Nothosaurus sp. (Text-fig. 3).

text-fig. 3. Reconstruction of the skull of Elmosaurus lelmensis Huene, 1957, interpreted as a partly preserved

skull of Nothosaurus sp.

The diagnosis of the genus Nothosaurus , and its systematics at the species level, remain
problematical, which precludes the definitive assignment of SMNS 59077 to any particular
nothosaur species at the present time. The problem originates with the introduction of the genus
Nothosaurus by Munster (1834), who described the type species of the genus, Nothosaurus mirabilis ,
on the basis of a postcranial skeleton, associated with a fragment of the lower jaw. The only
nothosaur with a complete skull associated to a postcranial skeleton in Nothosaurus raabi Schroder
(1914; a junior synonym of N. venustus Munster, 1834; see Schultze 1970). The problems originating
from this situation concern the diagnosis of, Nothosaurus mirabilis on the basis of skull characters,
as well as the generic assignment of material known from skulls only. There is, however, a general
consensus that Nothosaurus is represented by medium to large sized eusauropterygians with a
longirostrine skull, constricted snout, large and elongated upper temporal fossae, posteriorly

972

PALAEONTOLOGY, VOLUME 36

displaced parietal foramen, and paired caniniform teeth in the maxilla (Tschanz 1989; Storrs 1991).
H. V. Meyer (1847-1855, PI. 1, fig. 1) figured a skull which he assigned to Nothosaurus mirabilis and
which, in comparison to other nothosaur species, shows a distinctly elongated snout (Schultze, 1970,
table 1). The relative length of the snout seems unique among the Nothosaurus species so far
described, and is matched by the hitherto undescribed specimen SMNS 59074 from the Upper
Muschelkalk (Text-fig. 4).

text-fig. 4. The skull of Nothosaurus cf. mirabilis-, dorsal (a) and lateral (b) views. Abbreviations: ec,
ectopterygoid; f, frontal; ju, jugal; m, maxilla; n, nasal; p, parietal; pf, postfrontal; pi. palatine; pm,
premaxilla; po, postorbital; prf, prefrontal; q, quadrate; v, vomer. Scale bar = 50 mm.

The total length of the skull of SMNS 59074, measured from the tip of the snout to the right
mandibular condyle (preserved in situ ) is 324 mm. The distance from the anterior margin of the
external naris to the tip of the snout is 72 mm; the longitudinal diameter of the external naris is
23-5 mm (left) and 24-5 mm (right); the longitudinal diameter of the orbit is 36 mm (right); the
internarial space is 10 mm; the interorbital space is 19 mm; the distance between the orbit and the
external naris is 23-5 mm (left) and 23 mm (right), the distance from the orbit to the upper temporal
fossa is 21 mm (right). The relation of the diameter of the external naris to the longitudinal diameter
of the orbit is 1 : 15; the relation of the distance of the orbit to the external naris to the distance from
the orbit to the upper temporal fossa is 11:1. The relative length of the snout (distance from
external naris to tip of snout as percentage of total skull length) in the species of Nothosaurus
recognized by Schultze (1970, his Table 1) is as follows: edingerae : 65-4; juvenilis: 54; venustus ; 67-2;
procerus : 59-5; andriani : 61 -6; chely drops : 64-5; mirabilis'. 47-7; SMNS 59074: 45. In conclusion,
SMNS 59074 is here compared to Nothosaurus mirabilis Munster, 1834, a species provisionally
diagnosed by the apomorphic elongation of the snout, pending future revision.

The last comprehensive review of Nothosaurus was conducted by Schultze (1970), who recognized
two species from the Lower Muschelkalk (N. venustus Munster, 1834; N. procerus Schroder, 19,14),

RIEPPEL: NOTHOSAURIAN REPTILE

973

and four possible species from the Upper Muschelkalk; N. mirabilis Munster, 1834; N. giganteus
Munster, 1834 (a skull which lacks diagnostic features according to Schultze, 1970); N. andriani
Meyer, 1839; and N. juvenilis Edinger, 1921 (another problematical species, possibly a juvenile of
N. mirabilis). Schultze (1970) identified features which consistently distinguish the species from the
Lower and Upper Muschelkalk, some of which are also observable in SMNS 59077. Nothosaurus
from the Lower Muschelkalk shows the separation of the prefrontal from the nasal bone by the
frontal, and the distance between the external naris and the orbit is somewhat larger than the
distance between the orbit and the upper temporal arch (11:1 to 1-4:1). Species from the Upper
Muschelkalk usually show a larger exposure of the prefrontal on the dorsal surface of the skull, the
element meeting the nasal bone; the distance between the external naris and the orbit is smaller than
the distance between the orbit and the upper temporal fossa (0-8: 1 to 0-9: 1).

Measurements and relations involving the orbit are prone to ontogenetic variation, since the orbit
usually grows negatively allometric (e.g. Sander 1989). This observation may bear on the
interpretation of SMNS 59077, which is a relatively small skull as compared to the other species
from the Upper Muschelkalk except for the even smaller Nothosaurus juvenilis (Edinger 1921).
Schultze (1970, p. 213) addressed the problem of ontogenetic variation in Nothosaurus skulls, and
found that the relative size of the orbit (as compared to the size of the external naris) is similar for
small and large species except in N. juvenilis , which shows relatively much larger orbits than the
other species. The ratio of the longitudinal diameter of the external naris to the longitudinal
diameter of the orbit in SMNS 59077 (1 : 1-67) falls into the range of variation of other nothosaur
species (Schultze 1970, p. 213, 1 : 1-5-1 : L94) and shows SMNS 59077 to have relatively much
smaller orbits that N. juvenilis (1:2-7; Schultze 1970, p. 213). The ratio of the distance from the
external naris to the orbit to the distance from the orbit to the upper temporal fossa is 0-8 : 1 in
SMNS 59077, and hence corresponds to the values obtained by Schultze (1970) for the Nothosaurus
species of the Upper Muschelkalk. Nothosaurus juvenilis (Edinger 1921 ; see also Haas 1963, pi. 12)
differs from SMNS 59077 by its smaller size, the narrow postorbital arch (perhaps correlated with
the large size of the orbit), and by the broad postfrontal bone.

In contrast to SMNS 59077, the prefrontals usually meet the nasals in the nothosaurs from the
Upper Muschelkalk. However, Schultze (1970, p. 223) discussed variability of this character, as is
further documented by the Nothosaurus mirabilis skull (SMNS 59074) from the Upper Muschelkalk,
which shows the prefrontal meeting the nasal on the left side, whereas the two bones remain
separated by the frontal on the right side of the skull (Text-fig. 4).

In his assessment of Elmosaurus as a pachypleurosauroid, Huene (1957) was intrigued by the
anterior position of the parietal foramen (not corroborated), and by what appeared to him to be
small and narrow upper temporal fossae. Only the anterior corner of the upper temporal fossa is
preserved in SMNS 59077, and it does appear relatively narrow. However, the size and shape of the
upper temporal fossa in Nothosaurus mirabilis (Text-fig. 4) and other species from the Upper
Muschelkalk (Schultze 1970, figs 9—12) is entirely compatible with the observations on SMNS
59077. In these species, the upper temporal fossa is large and elongate, but its anterior corner is
distinctly constricted due to a convex anterolateral margin of the parietal (this constriction of the
anterior corner of the upper temporal fossa is absent in the species from the Lower Muschelkalk:
Schultze 1970, figs 2-8). The narrow anterior corner of the upper temporal fossa therefore does not
preclude the existence of large temporal fossae in SMNS 59077.

On the basis of the evidence discussed above, Elmosaurus Huene, 1957, is considered a junior
synonym of Nothosaurus Munster, 1834. The skull as preserved differs from the Lower Muschelkalk
species by the relative width of the postorbital arch and by the shape of the upper temporal fossa,
but it is not diagnostic among the species of the Upper Muschelkalk. Elmosaurus lelmensis is
therefore referred to Nothosaurus cf. mirabilis pending future revision of the genus.

Acknowledgements. I thank Professor Dr P. Carls from the Institute of Geology and Palaeontology, Technische
Universitat Braunschweig, who made it possible for me to study the type material of Elmosaurus at the
Staatliches Museum fur Naturkunde in Stuttgart. Dr R. Wild and Dr R. Bottcher provided generous

974

PALAEONTOLOGY, VOLUME 36

hospitality, workspace and access to the collection at the latter institution. The photographs of Elmosaurus
were taken by H. Lumpe, also from the Staatliches Museum fur Naturkunde in Stuttgart. I thank Drs S. E.
Evans, A. R. Milner and an anonymous reviewer for comments on an earlier draft of the paper. This work was
supported, in part, by NSF grant DEB - 9220540.

REFERENCES

baur, G. 1889. Palaeohatteria Credner, and the Proganosauria. American Journal of Science, 37, 310-313.
carroll, R. L. 1988. Vertebrate paleontology and evolution. W. H. Freeman and Co., New York, 698 pp.

— and gaskill, p. 1985. The nothosaur Pachypleurosaurus and the origin of plesiosaurs. Philosophical
Transactions of the Royal Society of London , Series B, 309, 343-393.

edinger, t. 1921. Uber Nothosaurus. II. Zur Gaumenfrage. Senckenbergiana , 3, 193-205.
haas, G. 1963. Micronothosaurus stensioi , ein neuer Nothosauride aus dem Oberen Muschelkalk des Wadi
Ramon, Israel. Paldontologische Zeitschrift , 37, 161-178.
huene, f. v. 1957. Ein neuer primitiver Nothosauride aus Braunschweig. Paldontologische Zeitschrift , 31,
92-98.

kuhn-schnyder, E. 1966. Der Schadel von Paranothosaurus amsleri Peyer aus dem Grenzbitumenhorizont der
anisisch-ladinischen Stufe der Trias des Monte San Giorgio (Kt. Tessin, Schweiz). Eclogae geologicae
Helvetiae , 59, 517-540.

meyer, H. v. 1839. Uber Nothosaurus : mittel-tertiare Knochen von Weisenau; Hyalith bei Frankfurt. Neues
Jahrbuch fiir Mineralogie , Geognosie und Petrefaktenkunde , 1839, 559-560.

— 1847-55. Zur Fauna der Vorwelt. Die Saurier des Muschelkalkes mit Riicksicht auf die Saurier aus Buntem
Sandstein und Keuper. H. Keller, Frankfurt am Main, viii + 167 pp., 70 pis.

munster, G. zu. 1834. Vorlaufige Nachricht iiber einige neue Reptilien im Muschelkalke von Baiern. Neues
Jahrbuch fiir Mineralogie, Geognosie, Geologie und Petrefaktenkunde, 1834, 521-527.

Owen, r. 1860. Palaeontology. Adam and Charles Black, Edinburgh, xv + 420 pp.

rieppel, o. 1989. A new pachypleurosaur (Reptilia: Sauropterygia) from the Middle Triassic of Monte San
Giorgio, Switzerland. Philosophical Transactions of the Royal Society of London, Series B, 323, 1-73.
Sander, p. M. 1989. The pachypleurosaurids (Reptilia: Nothosauria) from the Middle Triassic of Monte San
Giorgio (Switzerland), with the description of a new species. Philosophical Transactions of the Royal Society
of London, Series B , 325, 561-670.

Schroder, h. 1914. Wirbeltiere der Riidersdorfer Trias. Abhandlungen der Koniglichen Preussischen
Geologischen Landesanstalt, N.E. , 65, 1-9.

schultze, h.-p. 1970. Uber Nothosaurus. Neubeschreibung eines Schadels aus dem Keuper. Senckenbergiana
lethaea, 51, 211-237.

storrs, G. w. 1991. Anatomy and relationships of Corosaurus alcovensis (Diapsida: Sauropterygia) and the
Triassic Alcova Limestone of Wyoming. Bulletin of the Peabody Museum of Natural History, 44, 1-151.
tschanz, K. 1989. Lariosaurus buzzii n.sp. from the Middle Triassic of Monte San Giorgio (Switzerland), with
comments on the classification of nothosaurs. Palaeontographica, Abteilung A, 208, 153-179.

OLIVER RIEPPEL

Department of Geology
Field Museum of Natural History
Roosevelt Road at Lakeshore Drive

Typescript received 1 December 1992 Chicago

Revised typescript received 26 February 1993 Illinois 60605-2496, USA

INDEX

Pages 1-248 are contained in Part I ; pages 249-493 in Part 2; pages 495-741 in Part 3; pages 743-980 in
Part 4.

A

Acanthoparyphal sp., 706
Achatella, 763; A. truncatocaudata? , 71 I
Acritarcha (Plant microfossils), 390, 500
Ambitisporites avitus, 173; A. dilutus , 174
Amino acids: brachiopods. Recent, 883
Ammonites: Cretaceous, Japan, 249
Amphibians: Carboniferous, Czechia, 657;

Devonian, Russia, 721; Permian, USA, 839
Amphilichas sp., 713
Ampyx aff. repulsus, 702

Ancyloceratidae (Mollusca, Cephalopoda), 251
Antarctica: Jurassic ferns, 637; Triassic gymno-
sperms, 337

Appendisphaera gen. nov., 500; A. fragilis sp. nov.,
505; A. grandis sp. nov., 503; A.? tabifica sp. nov.,
508; A. tenuis sp. nov., 506
Arctostroma, 214

Arthropods: Silurian, Australia, 319
Arthrorhachis knockerkensis sp. nov., 689
Asaphidae (Arthropoda, Trilobita), 693
Atopostroma, 220
Atractopygel sp., 708

Australia: Cambrian palaeoscolecid worms, 549;
Silurian arthropods, 319

Austroscolex gen. nov., 559; A. primitivus sp. nov.,
561; A. spatiolatus sp. nov., 561
Autoloxolichas cf. laxatus , 71 1
Azendohsaurus laaroussii , 899

B

Balatonospiral cf. B. lipoldi , 452

Bannhuanaspis gen. nov., 299; B. vukhuci sp. nov., 299

Barrandia sp., 696

Barthelopteris gen. nov., 75; B. germarii, 75
Basiliolidea (Brachiopoda, Articulata), 196
Bassett, M. G. See Jaanusson, V. and Bassett, M. G.
Bigotina bivallata, 870; B. sp., 876
Bigotinidae (Arthropoda, Trilobita), 863
Biostratigraphy: ferns, Triassic and Jurassic, 650;
fishes, Cretaceous, 246; graptolites, Silurian, 924;
palaeoscolecid worms, Cambrian, 589 ; plant micro-
fossils, Neoproterozoic, 387, 495; plant micro-
fossils, Silurian, 182; trilobites, Ordovician, 685
Birmanites salteri sp. nov., 693
Blake, D. B. A new asteroid genus from the Jurassic
of England and its functional significance, 147

Bosence, D. W. J. See Braga, J. C., Bosence, D. W. J.
and Steneck, R. S.

Brachiopods: classification, 481, 931; Jurassic, Eng-
land, 195; Ordovician, 21; phylogeny and evol-
ution, 807; Recent, New Zealand, 883; Triassic,
USA, 439

Brachisolaster gen. nov., 147; B. moretonis, 148
Braga, J. C., Bosence, D. W. J. and Steneck, R. S.
New anatomical characters in fossil coralline algae
and their taxonomic implications, 535
Brunton, C. El. C. See Williams, A. and Brunton,
C. H. C.

C

Calymenidae (Arthropoda, Trilobita), 709
Cambrian : Ediacaran-like fossils, USA and Canada,
593; palaeoscolecid worms, Australia, 549; rhab-
dopleurids, Siberia, 283; trilobites, China, 785;
trilobites, France, 855

Canada: Cambrian Ediacaran-like fossils, 593; Car-
boniferous gymnosperms, 65; Ordovician grap-
tolites, 267
Capetus palustris, 659

Cappetta, H. See Prasad, G. V. R. and Cappetta, H.
Carboniferous: amphibians, Czechia, 657; fishes,
Scotland, 123; gymnosperms, Canada, 65
Carlson, S. J. Phylogeny and evolution of ‘pen-
tameride’ brachiopods, 807
Cavaspina gen. nov., 508; C. acuminata, 509;

C. basiconica sp. nov., 510
Centroscymnus schmidi, 16
Ceraurinellal sp., 703
Charniidae (Anthozoa, Cnidaria), 604
Charniids: Cambrian, Canada, 604
Cheilotetras gen. nov., 161; C. caledonica sp. nov.,
162

Cheiruridae (Arthropoda, Trilobita), 703
China: Cambrian trilobites, 785; Permian ferns, 81
Chuaria circularis, 390

Clack, J. A. See Lebedev, O. A. and Clack, J. A.
Cladistics: amphibians, 670; brachiopods, 807, 931;

crinoids, 93; dinosaurs, 374; fishes, 137
Clarkson, E. N. K. See Zhang Xi-Guang and Clark-
son, E. N. K.

Cleal, C. J. See Zodrow, E. L. and Cleal, C. J.
Coates, M. I. New actinopterygian fish from the

976

INDEX

Namurian Manse Burn Formation of Bearsden,
Scotland, 123
Coenostroma , 2 1 1
Columnostroma, 224

Conularia sarae sp. nov., 412; C. wimani sp. nov., 416;

C. sp. a, 418

Conularnds: Silurian, Sweden, 403
Conway Morris, S. Ediacaran-like fossils in Cambrian
Burgess Shale-type faunas of North America, 593
Corallinaceae (Rhodophycaceae, Rhodophyta), 542
Coralline algae: taxonomy, 535
Corallioscolex gen. nov., 563; C. gravius sp. nov., 563
Corystospermaceae (Gymnospermophyta), 338
Cretaceous: ammonites, Japan, 249; dinosaurs, Eng-
land, 425; dinosaurs, Malawi, 523; dinosaurs,
Romania, 361; echinoids, France, 311; fishes,
India, 231; fishes, Sweden, 1
Crinoids: phylogeny, 91
Crioceratites ( Paracrioceras ) asiaticum, 251
Cruickshank, A. R. I. See Taylor, M. A., Norman,

D. B. and Cruickshank, A. R. I.

Curry, G. B. See Walton, D., Cusak, M. and Curry,
G. B.

Cusak, M. See Walton, D., Cusak, M. and Curry,
G. B.

Cuticles: gymnosperms, Carboniferous, 65
Czechia: Carboniferous amphibians, 657

D

Decordinaspis bispinosa, 702
Decoroproetus sp., 698
Deformed fossils, 927

Devonian: amphibians, Russia, 721 ; fishes, Vietnam,
297

Dielasmatidae (Brachiopoda, Articulata), 467
Dilkes, D. W. Biology and evolution of the nasal
region in trematopid amphibians, 839
Dinosaurs: Cretaceous, England, 425; Cretaceous,
Malawi, 523; Cretaceous, Romania, 361; Jurassic,
England, 357; Triassic, Morocco, 897
Dipteridaceae (Pteridophyta, Filicopsida), 639
Downs, W. R. See Jacobs, L. J., Winkler, D. A.,
Downs, W. R. and Gomani, E. M.

Durman, P. N. and Sennikov, N. V. A new rhab-
dopleurid hemichordate from the Middle Cam-
brian of Siberia, 283

Dyadospora murusattenuata, 169; D. murusdensa , 170
E

Echinoids: Jurassic, England, 147; sexual dimorph-
ism, 31 1

Ediacaran-like fossils: Cambrian, USA and Canada,
593

Emmonsaspis cambrensis , 599
Encrinuridae (Arthropoda, Trilobita), 707

England: Cretaceous dinosaurs, 425; Jurassic brach-
iopods, 195; Jurassic dinosaurs and pliosaurs, 357;
Jurassic echinoids, 147; Triassic trace fossils. 111

Eoetmopterus cf. E. supracretaceus, 1 5

Estonia: Ordovician trilobites, 743

Estoniops fjaeckensis sp. nov., 757; E. maennili sp.
nov., 757

Euryscolex gen. nov., 564; E. paternarius sp. nov.
564

Euthycarcinoidae (Uniramia, Euthycarcinoidea),
324

F

Ferestromatopora, 212

Ferns: Jurassic, Antarctica, 637; Permian, China, 81 ;
Triassic, New Zealand, 637

Ferretti, A., Holland, C. H. and Syba, E. Problematic
microfossils from the Silurian of Ireland and
Scotland, 771

Fishes: Carboniferous, Scotland, 123; Cretaceous,
India, 231; Cretaceous, Sweden, 1; Devonian,
Vietnam, 297

Flexicalymene sp., 709

France: Cambrian trilobites, 855; Cretaceous echin-
oids, 31 1

Frederichthvs gen. nov., 134; F. musadentatus sp.
nov., 135

Fructifications: ferns, Permian, 82

Functional morphology: palaeoscolecid worms,

Cambrian, 587; problematica, Silurian, 782

G

Gao Zhifeng and Thomas, B. A. A new fern from the
Lower Permian of China and its bearing on the
evolution of the marattialeans, 81

Gastropods: taphonomy, 735

Gauffre. F.-X. The prosauropod dinosaur Azen-
dohsaurus laaroussii from the Upper Triassic of
Morocco, 897

Gelenoptron gen. nov., 614; G. tentaculatum sp. nov.,
614

Germany: Triassic reptiles, 967

Glyptostromoides , 216

Goeppertella jeffersonii sp. nov., 639; G. woodii sp.
nov., 643

Gomani, E. M. See Jacobs, L. J., Winkler, D. A.,
Downs, W. R. and Gomani, E. M.

Graptolites: population and orientation analyses,
267; Silurian, 909

Gravicalymene sp., 710

Grigorescu, D. See Weishampel, D. B., Norman,
D. B. and Grigorescu, D.

Gymnosperms: Carboniferous, Canada, 65; Triassic,
Antarctica, 337

INDEX

977

H

Hadimopanella apicata, 584; H. oezguli , 567
Hadrosauridae (Chordata, Reptilia), 362
Habrostroma, 222
Hausmannia cf. nariwansis, 648
Hinz-Schallreuter, I. See Muller, K. J. and Hinz-
Schallreuter, I.

Holland. C. H. See Ferretti, A., Holland, C. H. and
Syba, E.

Howse, S. C. B. and Milner, A. R. Ornithodesmus - a
maniraptoran theropod dinosaur from the Lower
Cretaceous of the Isle of Wight, England, 425

I

Igdabalis indicus sp. nov., 239
Illaenidae (Arthropoda, Trilobita), 697
Illaenus sp., 697
India: Cretaceous fishes, 231
Ingriops gen. nov., 750

Ireland: Ordovician trilobites, 681; Silurian prob-
lematica, 771

J

Jaanusson, V. and Bassett, M. G. Orthambonites and
related Ordovician genera, 21
Jaanusson, V. and Ramskold, L. Pterygometopine
trilobites from the Ordovician of Baltoscandia, 743
Jacobs, L. J., Winkler, D. A., Downs, W. R. and
Gomani, E. M. New material of an Early Cret-
aceous titanosaurid sauropod dinosaur from
Malawi, 523

Janvier, P., Tong-Dzuy T. and Ta-Hoa, P. A new
Early Devonian galeaspid from Bac Thai Province,
Vietnam, 297

Japan: Cretaceous ammonites, 249
Jerre, F. Conularhd microfossils from the Silurian
Lower Visby Beds of Gotland, Sweden, 403
Jurassic: brachiopods, England, 195; ferns, Ant-
arctica, 637; dinosaurs and pliosaurs, England,
357; echinoids, England, 147

K

Kalbarria gen. nov., 324; K. brimmellae sp. nov., 324
Kaloscolex gen. nov., 567; K. gramdatus sp. nov., 568
Keilapyge gen. nov., 762
Krattorthis gen. nov., 56

Kykloxylon gen. nov., 338; K. fremouwensis sp. nov.,
'338 '

L

Laevolancis (Archaeozonotriletes) divellomedium , 172;
L. plica ta, 173

Laballidae (Brachiopoda, Articulata), 455
Latvia: Ordovician trilobites, 743
Lebedev, O. A. and Clack, J. A. Upper Devonian
tetrapods from Andreyevka, Tula Region, Russia,
721

Lichidae (Arthropoda, Trilobita), 711
Lineastroma , 213

Loydell, D. K., Storch, P. and Melchin, M. J. Tax-
onomy, evolution and biostratigraphical import-
ance of the Llandovery graptolite Spirograptus , 909

M

Mackenzia costalis, 608

Mackenziidae fam. nov. (Anthozoa, Cnidaria), 606
McNamara, K. J. and Trewin, N. H. A euthycar-
cinoid arthropod from the Silurian of Western
Australia, 319

Malawi: Cretaceous dinosaurs, 523
Malawisaurus gen. nov., 524; M. dixeyi , 525
Malta: Neogene coralline algae, 535
Marattiaceae (Pteridophyta, Filicopsida), 82
Matsukawa, M. and Obata, I. The ammonites
Crioceratites ( Paracrioceras ) and Shasticrioceras
from the Barremian of southwest Japan, 249
Melchin, M. J. See Loydell, D. K., Storch, P. and
Melchin, M. J.

Mesopoma carricki sp. nov., 126; M. pancheni sp.

nov., 132; M.? smithsoni sp. nov., 129
Metaconularia aspersa, 419
Metagnostidae (Arthropoda, Trilobita), 688
Meyer-Berthaud, B., Taylor, T. N. and Taylor, E. L.
Petrified stems bearing Dicroidium leaves from the
Triassic of Antarctica, 337
Microetmopterus gen. nov., 10; M. wardi sp. nov., 10
Milaculum elongatum sp. nov., 568
Milner, A. R. See Howse, S. C. B. and Milner, A. R.
Milner, A. R. See Sequeira, S. E. K. and Milner,
A. R.

Miraspis aff. solitaria , 714

Moczydlowska, M., Vidal, G. and Rudavskaya, V. A.
Neoproterozoic (Vendian) phytoplankton from the
Siberian Platform, Yakutia, 495
Moczydlowska, M. See Vidal, G., Moczydlowska, M.

and Rudavskaya, V. A.

Morocco: Triassic dinosaurs, 897
' Moyeria' cabottii , 175

Muller, K. J. and Hinz-Schallreuter, I. Palaeoscolecid
worms from the Middle Cambrian of Australia,
549

Murrayscolex gen. nov., 569; M. inaequalis sp. nov.,
569; M. serratus sp. nov., 571

N

Neogene: coralline algae, Malta and Spain, 535
Neoproterozoic: plant microfossils, Siberia, 387, 495

978

INDEX

Neraudeau, D. Sexual dimorphism in mid-Cret-
aceous hemiasterid echinoids, 311
New Zealand: Jurassic or Triassic ferns, 637; Recent
brachiopods, 883

Nileidae (Arthropoda, Trilobita), 696
Norman, D. B. See Taylor, M. A., Norman, D. B.
and Cruickshank, A. R. I.

Norman, D. B. See Weishampel, D. B., Norman,
D. B. and Grigorescu, D.

Norway: Ordovician trilobites, 743
Nothosauridae (Chordata, Reptilia), 967
Nothosaurus cf. mirabilis, 967

O

Obata, I. See Matsukawa, M. and Obata, I.
Odontopleuridae (Arthropoda, Trilobita), 714
Oelandiops gen. nov., 751; O. mirificus sp. nov., 751
Ontogeny: trilobites, 785, 876
Ordovician: brachiopods, 21; graptolites, Canada,
267; trilobites, Baltoscandia, 743; trilobites, Ire-
land, 681

Ornithodesmus cluniculus , 429
Orthidae (Brachiopoda, Articulata), 22
Orthambonites calligramma , 26
Orthis callactis , 54

Owen, A. W. See Romano, M. and Owen, A. W.

P

Palaeobiogeography : ammonites, Cretaceous, 262;
brachiopods, Triassic, 446; trilobites, Cambrian,
861; trilobites, Ordovician, 681
Palaeoecology : brachiopods, Triassic, 445; echino-
derms, Jurassic, 152; graptolites, Ordovician, 267;
gymnosperms. Carboniferous, 77; palaeoscolecid
worms, Cambrian, 588; sharks, Cretaceous, 17;
trace fossils, Triassic, 111
Palaeoscolecidae (Palaeoscolecida), 558
Palaeoscolecid worms : Cambrian, Australia, 549
Pantoioscolex gen. nov., 571 ; P. oleschinskii sp. nov.,
571

Paralenorthis , 33
Parallelopora, 225
Parallelostroma , 220

Permian: amphibians, USA, 839; ferns, China, 81
Phylogeny: ammonites, 262; brachiopods, 807, 931;
crinoids, 91 ; ferns, 86; galeaspids, 305; graptolites,
922; hexapoda, 327; stromatoporids, 209
Pillola, G. L. The Lower Cambrian trilobite Bigotina
and allied genera, 855

Plant microfossils: Neoproterozoic, Siberia, 387, 495;
Silurian, Scotland, 155

Plectoconcha aequiplicata , 460; P. newbyi sp. nov.,
465

Pliosaurs: Jurassic, England, 357
Pliosaurus brachyspondylus , 357

Prasad, G. V. R. and Cappetta H. Late Cretaceous
selachians from India and the age of the Deccan
Traps, 231

Problematica: Silurian, Ireland and Scotland, 771
Proetidae (Arthropoda, Trilobita), 698
Proetmopterus gen. nov., 12; P. hemmooriensis , 14
Prosser, C. D. The brachiopod Stolmorhynchia sto-
lidota from the Bajocian of Dorset, England. 195
Pseudoconularia aff. scalaris , 420
Pseudodyadospora petasus sp. nov., 168
Pseudotrupetostroma , 2 1 7

Pterygometopidae (Arthropoda, Trilobita), 710, 745
Pterygometopus , 747

R

Raja sudhakari sp. nov., 234
Rajidae (Chordata, Chondrichthyes), 234
Ramskold, L. See Jaanusson, V. and Ramskold, L.
Raphiophoridae (Arthropoda, Trilobita), 702
Recent: brachiopods. New Zealand, 883
Rees, P. M. Dipterid ferns from the Mesozoic of
Antarctica and New Zealand and their strati-
graphical significance, 637
Reguellia camera, 778
Regnellidae fam. nov. (problematica), 774
Remopleurides sp., 692

Remopleurididae (Arthropoda, Trilobita), 692
Reptiles: Triassic, Germany, 967
Reticulopteris muensteri, 74
Rhabdopleura obuti sp. nov., 284
Rhabdopleuridae (Hemichordata, Pterobranchia),
284

Rhabdopleurids: Cambrian, Siberia, 283
Rhaetina gregaria, 468
Rhombodus, 244

Rhomboscolex gen. nov., 572; R. chaoticus sp. nov.,
572

Richardson, J. B. See Wellman, C. H. and Richard-
son, J. B.

Rieppel, O. The status of the nothosaurian reptile
Elmosaurus lelmensis , with comments on Noth-
osaurus mirabilis, 967

Rigby, S. Population analysis and orientation studies
of graptoloids from the Middle Ordovician Utica
Shale, Quebec, 267
Rimosotetras problematica, 163
Romania: Cretaceous dinosaurs, 361
Romano, M. and Owen, A. W. Early Caradoc
trilobites of eastern Ireland and their palaeo-
geographical significance, 681
Rudavskaya, V. A. See Moczydlowska, M., Vidal,
G. and Rudavskaya, V. A.

Rudavskaya, V. A. See Vidal, G., Moczydlowska, M.
and Rudavskaya, V. A.

Rushton, A. W. A. and Smith, M. Retrodeformation
of fossils - a simple technique, 927

INDEX

979

Russia: Devonian amphibians, 721; Ordovician tri-
lobites, 743

S

Salairella , 219

Sandvikina conica sp. nov., 776; X. sp., 778
Sandy, M. R. and Stanley, G. D. Jr. Late Triassic
brachiopods from the Luning Formation, Nevada,
and their palaeobiogeographical significance, 439
Schistoscolex gen. nov., 575; .S', angustosquamatus sp.
nov., 575; S. mucronatus sp. nov., 576; S. umbilictus
sp. nov., 576; S. sp. mdet., 579
Scotland: Carboniferous fishes, 123; Silurian plant
microfossils, 155; Silurian problematica, 771
Sennikov, N. V. See Durman, P. N. and Sennikov,
N. V.

Sequeira, S. E. K. and Milner, A. R. The temno-
spondyl amphibian Capetus from the Upper
Carboniferous of the Czech Republic, 657
Sevastopulo, G. D. See Simms, M. J. and Sevasto-
pulo, G. D.

Sexual dimorphism: echinoids, 31 1
Sharks: Cretaceous, Sweden, 1
Shasticrioceras intermedium sp. nov., 258; S. nippon-
icum , 255; S. patricki, 256
Shell structure: brachiopods, 931
Shergoldiscolex gen. nov., 579; S. nodosus sp. nov.,
579; S. polygonatus sp. nov., 582
Shoshonorthis gen. nov., 51

Siberia: Cambrian rhabdopleurids, 283; Neo-

proterozoic plant microfossils, 387, 495
Silurian: arthropods, Australia, 319; conulariids,
Sweden, 403; graptolites, 909; plant microfossils,
Scotland, 155; problematica, Ireland and Scotland,
771

Simms, M. J. and Sevastopulo, G. D. The origin of
articulate crinoids, 91

Siverson, M. Maastrichtian squaloid sharks from
southern Sweden, I

Sivorthis gen. nov., 45; S.filistera sp. nov., 47
Smith, M. See Rushton, A. W. A. and Smith, M.
Solasteridae (Echinodermata, Asteroidea), 147
Spain: Neogene coralline algae, 535
Sphaerexochus sp., 706
Sphaerocoryphe cf. pemphis, 704
Spiriferinidae (Brachiopoda, Articulata), 450
Spirograptus andrewsi , 922; S. guerichi sp. nov., 917;

S. turriculatus, 912
Spondylospira lewesensis, 459
Spongites albanensis , 544; S. sp. 1, 544
Squalidae (Chordata, Chondrichthyes), 4
Squalus ballingsloevensis sp. nov., 6; S. balsvikensis
sp. nov., 6; S. gabrielsoni sp. nov., 8
Stanley, G. D. Jr. See Sandy, M. R. and Stanley,
G. D. Jr.

Stearn, C. W. Revision of the order Stromatoporida,
201

Steneck, R. S. See Braga, J. C., Bosence, D. W. J. and
Steneck, R. S.

Stolmorhynchia stolidota, 196

Storch, P. See Loydell, D. K., Storch, P. and Melchin,

M. J.

Stromatopora, 210

Stromatoporidae (Porifera, Stromatoporoidea), 210
Sulevorthis gen. nov., 37; X. lyckholmiensis, 40
Sweden : Cretaceous sharks, 1 ; Ordovician trilobites,
743; Silurian conulariids, 403
Syba, E. See Ferretti, A., Holland, C. H. and Syba, E.
Syr ingo stroma, 222

Syringostromatidae (Porifera, Stromatoporoidea),
220

Syringostromella, 218

Syringostromellidae (Porifera, Stromatoporoidea),
218

T

Ta-Hoa P. See Janvier, P., Tong-Dzuy T. and Ta-
Hoa, P.

Taiyuanitheca gen. nov., 82; T. tetralinea sp. nov., 82
Taleastroma, 215

Tanarium conoideum, 514; T. irregulare sp. nov., 516;

T. tuberosum sp. nov., 516
Taphonomy: gastropods, 735; trilobites, 803
Tawuia dalensis, 393

Taylor, E. L. See Meyer-Berthaud, B., Taylor, T. N.
and Taylor, E. L.

Taylor, M. A., Norman, D. B. and Cruickshank,
A. R. I. Remains of an ornithiscian dinosaur in a
pliosaur from the Kimmeridgian of England, 357
Taylor, T. N. See Meyer-Berthaud, B., Taylor, T. N.

and Taylor, E. L.

Teeth: fishes, Cretaceous, 1, 231
Telephina ( Telephops ) cf. bicornis, 690
Telephinidae (Arthropoda, Trilobita), 690
Telmatosaurus transsylvanicus , 362
Terebratulidae (Brachiopoda, Articulata), 460
Tetrahedraletes medinensis , 164
Thaumaptilon gen. nov., 604; T. walcotti sp. nov., 604
Thomas, B. A. See Gao Zhifeng and Thomas, B. A.
Thoracoscolex gen. nov., 582; T. armatus sp. nov.,
584

Titanosauridae (Chordata, Reptilia), 524
Tong-Dzuy T. See Janvier, P., Tong-Dzuy T. and
Ta-Hoa P.

Trace fossils: Triassic, England, 111
Tretaspis aff. reticulata , 700

Trewin, N. H. See McNamara, K. J. and Trewin,

N. H.

Triassic: brachiopods, USA, 439; dinosaurs, Mo-
rocco, 897 ; ferns. New Zealand, 637 ; gymnosperms.

980

INDEX

Antarctica, 337; reptiles, Germany, 967; trace
fossils, England, 1 1 1

Trigonocarpales (Gymnospermophyta, Cycadop-
sida), 74

Trilobites: Cambrian, China, 785; Cambrian, France,
855; Ordovician, Baltoscandia, 743; Ordovician,
Ireland, 681

Trinucleidae (Arthropoda, Trilobita), 700
Triplesiidae (Brachiopoda, Articulata), 481
Troodontidae (Chordata, Reptilia), 426
Tulerpeton curtum , 721

U

Upplandiops gen nov., 760; U. calvus sp. nov., 761
USA: Cambrian Ediacaran-like fossils, 593; Permian
amphibians, 839; Triassic brachiopods, 439

V

Vidal, G., Moczydlowska, M. and Rudavskaya, V. A.
Biostratigraphical implications of a Chuaria-
Tawuia assemblage and associated acritarchs from
the Neoproterozoic of Yakutia, 387
Vidal, G. See Moczydlowska, M., Vidal, G. and
Rudavskaya, V. A.

Vietnam: Devonian fishes, 297

W

Walker, S. E. and Yamada, S. B. Implications for the
gastropod fossil record of mistaken crab predation
on empty mollusc shells, 735
Walton, D., Cusak, M. and Curry, G. B. Implications

of the amino acid composition of Recent New
Zealand brachiopods, 883

Wang Guangzhong. Xiphosurid trace fossils from the
Westbury Formation (Rhaetian) of southwest
Britain, 1 1 1

Weishampel, D. B., Norman, D. B. and Grigorescu,
D. Tebnatosaurus trcmssylvanicus from the Late
Cretaceous of Romania: the most basal hadro-
saurid dinosaur, 361

Wellman, C. H. and Richardson, J. B. Terrestrial
plant microfossils from Silurian inkers of the
Midland Valley of Scotland, 155

Williams, A. and Brunton, C. H. C. Role of shell
structure in the classification of the orthotetidine
brachiopods, 931

Winkler, D. A. See Jacobs, L. J., Winkler, D. A.,
Downs, W. R. and Gonrani, E. M.

Wright, A. D. Subdivision of the Lower Palaeozoic
articulate brachiopod family Triplesiidae, 481

Y

Yamada, S. B. See Walker, S. E. and Yamada, S. B.

Z

Zeilleria cf. Z. elliptica , 470

Zeilleriidae (Brachiopoda, Articulata), 470

Zhang Xi-Guang and Clarkson, E. N. K. Ontogeny
of the eodiscid trilobite Shizhudiscus longquanensis
from the Lower Cambrian of China, 785

Zodrow, E. L. and Cleal, C. J. The epidermal struc-
ture of the Carboniferous gymnosperm frond
Reticulopteris , 65

Zugmayerella uncinata , 455; ?Z. sp., 457

VOLUME 36

Palaeontology

1993

PUBLISHED BY THE
PALAEONTOLOGICAL ASSOCIATION
LONDON

Dates of Publication of Parts of Volume 36

Part 1, pp. 1-248
Part 2, pp. 249-493
Part 3. pp. 495-741
Part 4, pp. 743-980

30 March 1993
21 July 1993
22 September 1993
22 December 1993

THIS VOLUME EDITED BY C. J. CLEAL, J. E. DALINGWATER, P. DOYLE, D. EDWARDS,
P. D. LANE, R. M. OWENS, P. A. SELDEN AND P. D. TAYLOR

© The Palaeontological Association, 1993

Printed in Great Britain
by Cambridge University Press

CONTENTS

Part

Bassett, M. G. See Jaanusson, V. and Bassett, M. G.

Blake, D. B. A new asteroid genus from the Jurassic of England and its functional significance 1
Bosence, D. W. J. See Braga, J, C., Bosence, D. W. J. and Steneck, R. S.

Braga, J. C., Bosence, D. W. J. and Steneck, R. S. New anatomical characters in fossil
coralline algae and their taxonomic implications 3

Cappetta, H. See Prasad, G. V. R. and Cappetta, H.

Carlson, S. J. Phylogeny and evolution of ‘pentameride’ brachiopods 4

Clack, J. A. See Lebedev, O. A. and Clack, J. A.

Clarkson, E. N. K. See Zhang Xi-Guang and Clarkson, E. N. K.

Cleal, C. J. See Zodrow, E. L. and Cleal, C. J.

Coates, M. E New actinopterygian fish from the Namurian Manse Burn Formation of

Bearsden, Scotland I

Conway Morris, S. Ediacaran-like fossils in Cambrian Burgess Shale-type faunas of North
America 3

Cruickshank, A. R. I. See Taylor, M. A., Norman, D. B. and Cruickshank, A. R. I.

Curry, G. B. See Walton, D., Cusak, M. and Curry, G. B.

Cusak, M. See Walton, D., Cusak, M. and Curry, G. B.

Dilkes, D. W. Biology and evolution of the nasal region in trematopid amphibians 4

Downs, W, R. See Jacobs, L. J., Winkler, D. A., Downs, W. R. and Gomani, E. M.

Durman, P. N. and Sennikov, N. V. A new rhabdopleurid hemichordate from the Middle
Cambrian of Siberia 2

Ferretti, A., Holland, C. H. and Syba, E. Problematical microfossils from the Silurian of
Ireland and Scotland 4

Gao, Zhifeng and Thomas, B. A. A new fern from the Lower Permian of China and its bearing
on the evolution of the marattialeans 1

Gauffre, F.-X. The prosauropod dinosaur Azendohsaurus laaroussii from the Upper Triassic of
Morocco 4

Gomani, E. M. See Jacobs, L. J., Winkler, D. A., Downs, W. R. and Gomani, E. M.
Grigorescu, D. See Weishampel, D. B., Norman, D. B. and Grigorescu, D.
Hinz-Schallreuter, I. See Muller, K, J. and Hinz-Schallreuter, I.

Holland, C. H. See Ferretti, A., Holland, C. H. and Syba, E.

Howse, S. C. B. and Milner, A. R. Ornithodesmus - a maniraptoran theropod dinosaur from
the Lower Cretaceous of the Isle of Wight, England

Jaanusson, V. and Bassett, M. G. Orthambonites and related Ordovician brachiopod genera 1
Jaanusson, V. and Ramskold, L. Pterygometopine trilobites from the Ordovician of
Baltoscandia 4

Jacobs, L. J., Winkler, D. A., Downs, W. R. and Gomani, E. M. New material of an Early
Cretaceous titanosaurid sauropod dinosaur from Malawi 3

Janvier, P., Tong-Dzuy T. and Ta-Hoa P. A new Early Devonian galeaspid from Bac Thai
Province, Vietnam 2

Jerre, F. Conulariid microfossils from the Silurian Lower Visby Beds of Gotland, Sweden 2
Lebedev, O. A. and Clack, J. A. Upper Devonian tetrapods from Andreyevka, Tula Region,
Russia 3

Loydell, D. K., Storch, P. and Melchin, M. J. Taxonomy, evolution and biostratigraphical
importance of the Llandovery graptolite Spirograptus 4

McNamara, K. J. and Trewin, N. H. A euthycarcinoid arthropod from the Silurian of Western
Australia

Matsukawa, M. and Obata, I. The ammonites Crioceratites (Paracrioceras) and Shasticrioceras
from the Barremian of southwest Japan
Melchin, M, J. See Loydell, D. K., Storch, P. and Melchin, M. J.

Meyer-Berthaud, B., Taylor, T. N. and Taylor, E. L. Petrified stems bearing Dicroidium
leaves from the Triassic of Antarctica

Page

147

535

807

123

593

839

283

771

81

897

425

21

743

523

297

403

721

909

319

249

337

IV

CONTENTS

Milner, A. R. See Howse, S. C. B. and Milner, A. R.

Milner, A. R. See Sequeira, S. E. K. and Milner, A. R.

Moczydlowska, M., Vidal, G. and Rudavskaya, V. A. Neoproterozoic (Vendian) phyto-
plankton from the Siberian Platform, Yakutia 3

Moczydlowska, M. See Vidal, G., Moczydlowska, M. and Rudavskaya, V. A.

Muller, K. J. and Hinz-Schallreuter, I. Palaeoscolecid worms from the Middle Cambrian of
Australia 3

Neraudeau, D. Sexual dimorphism in mid-Cretaceous hemiasterid echinoids 2

Norman, D. B. See Taylor, M. A., Norman, D. B. and Cruickshank, A. R. I.

Norman, D. B. See Weishampel, D. B., Norman, D. B. and Grigorescu, D.

Obata, I. See Matsukawa, M. and Obata, I.

Owen, A. W. See Romano, M. and Owen, A. W.

Pillola, G. L. The Lower Cambrian trilobite Bigotina and allied genera 4

Prasad, G. V. R. and Cappetta, H. Late Cretaceous selachians from India and the age of the
Deccan Traps 1

Prosser, C. D. The brachiopod Stolmorhynchia stolidota from the Bajocian of Dorset, England 1
Ramskold, L. See Jaanusson, V. and Ramskold, L.

Rees, P. M. Dipterid ferns from the Mesozoic of Antarctica and New Zealand and their
stratigraphical significance 3

Richardson, J. B. See Wellman, C. H. and Richardson, J. B.

Rieppel, O. The status of the nothosaurian reptile Elmosaurus lelmensis , with comments on
Nothosaurus mirabilis 4

Rigby, S. Population analysis and orientation studies of graptoloids from the Middle Ordovician
Utica Shale, Quebec 2

Romano, M. and Owen, A. W. Early Caradoc trilobites of eastern Ireland and their
palaeogeographical significance 3

Rudavskaya, V. A. See Moczydlowska, M., Vidal, G. and Rudavskaya, V. A.

Rudavskaya, V. A. See Vidal, G., Moczydlowska, M. and Rudavskaya, V. A.

Rushton, A. W. A. and Smith, M. Retrodeformation of fossils - a simple technique 4

Sandy, M. R. and Stanley, G. D. Jr. Late Triassic brachiopods from the Luning Formation,
Nevada, and their palaeobiogeographical significance 2

Sennikov, N. V. See Durman, P. N. and Sennikov, N. V.

Sequeira, S. E. K. and Milner, A. R. The temnospondyl amphibian Capetus from the Upper
Carboniferous of the Czech Republic 3

Sevastopulo, G. D. See Simms, M. J. and Sevastopulo, G. D.

Simms, M. J. and Sevastopulo, G. D. The origin of articulate crinoids 1

Siverson, M. Maastrichtian squaloid sharks from southern Sweden 1

Smith, M. See Rushton, A. W. A. and Smith, M.

Stanley, G. D. Jr. See Sandy, M. R. and Stanley, G. D. Jr

Stearn, C. W. Revision of the order Stromatoporida 1

Steneck, R. S. See Braga, J. C., Bosence, D. W. J. and Steneck, R. S.

Storch, P. See Loydell, D. K., Storch, P. and Melchin, M. J.

Syba, E. See Ferretti, A., Holland, C. H. and Syba, E.

Ta-Hoa P. See Janvier, P., Tong-Dzuy T. and Ta-Hoa P.

Taylor, E. L. See Meyer- Berthaud, B., Taylor, T. N. and Taylor, E. L.

Taylor, M. A., Norman, D. B. and Cruickshank, A. R. I. Remains of an ornithiscian
dinosaur in a pliosaur from the Kimmeridgian of England 2

Taylor, T. N. See Meyer- Berthaud, B., Taylor, T. N. and Taylor, E. L.

Thomas, B. A. See Gao Zhifeng and Thomas, B. A.

Tong-Dzuy T. See Janvier, P., Tong-Dzuy T. and Ta-Hoa P.

Trewin, N. H. See McNamara, K. J. and Trewin, N. H.

Vidal, G., Moczydlowska, M. and Rudavskaya, V. A. Biostratigraphical implications of a
Chuaria-Tawuia assemblage and associated acritarchs from the Neoproterozoic of Yakutia 2

Vidal, G. See Moczydlowska, M., Vidal, G. and Rudavskaya, V. A.

Walker, S. E. and Yamada, S. B. Implications for the gastropod fossil record of mistaken crab
predation on empty mollusc shells 3

Walton, D., Cusak, M. and Curry, G. B. Implications of the amino acid composition of
Recent New Zealand brachiopods 4

495

549

311

855

231

195

637

967

267

681

927

439

657

91

1

201

357

387

735

883

CONTENTS

Wang, Guangzhong. Xiphosurid trace fossils from the Westbury Formation (Rhaetian) of

southwest Britain 1 111

Weishampel, D. B., Norman, D. B. and Grigorescu, D. Telmatosaurus transsylvanicus from

the Late Cretaceous of Romania: the most basal hadrosaurid dinosaur 2 361

Wellman, C. H. and Richardson, J. B. Terrestrial plant microfossils from Silurian inlicrs of the

Midland Valley of Scotland 1 155

Williams, A. and Brunton, C. H. C. Role of shell structure in the classification of the

orthotetidine brachiopods 4 931

Winkler, D. A. See Jacobs, L. J., Winkler, D. A., Downs, W. R. and Gomani, E. M.

Wright, A. D. Subdivision of the Lower Palaeozoic articulate brachiopod family Triplesiidae 2 481

Yamada, S. B. See Walker, S. E. and Yamada, S, B.

Zhang Xi-Guang and Clarkson, E. N. K. Ontogeny of the eodiscid trilobite Shizhudiscus

longquanensis from the Lower Cambrian of China 4 785

Zodrow, E. L. and Cleal, C. J. The epidermal structure of the Carboniferous gymnosperm
frond Reticulopteris 1 65

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 R. M. Owens, Department of Geology, National
Museum of Wales, Cardiff CF1 3NP, UK, 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 1993 is £55 00 (U.S. $120) for Institutional
Members, and £20 00 (U.S. $36) 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 per cent below the
listed prices: contact the Marketing Manager. Non-members may obtain Nos 35^16 (at cover price) from Basil Blackwell Ltd.,
Journal Subscription Department, Marston Book Services, P.O. Box 87, Oxford OX2 0DT, UK, 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 2-35 are
still in print and are available together with those listed below:

36. (for 1986): Campanian and Maastrichtian ammonites from northern Aquitaine, France, by w. 1. Kennedy. 145 pp.,
43 text-figs, 23 plates. Price £20 (U.S. $40).

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. $40).

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. $60).

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. $60).

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. $60).

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. $40).

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. $80).

43. (for 1990): Benthic palaeoecology of the Late Jurassic Kimmeridge Clay of England, by p. r. wignall, 74 pp., 50 text-figs.
Price £30 (U.S. $60).

44. (for 1990): The Murchison Symposium. Edited by M. G. bassett, p. d. lane and D. Edwards. 397 pp.. 137 text-figs,
3 plates. Price £60 (U.S. $120)

45. (for 1991): Coniacian ammonite faunas from the United States Western Interior, by w. j. Kennedy and W. a. cobban.
96 pp., 31 text-figs, 17 plates. Price £35 (U.S. $70).

46. (for 1991): Arundian (Lower Carboniferous) conodonts from South Wales, by J. J. stone. 63 pp., 14 text-figs, 5 plates.
Price £30 (U.S. $60)

Field Guides to Fossils and Other Publications

These are available only from the Marketing Manager. Please add £1 00 (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. $16) (Members £6

or U.S. $12).

2. (1987): Fossils of the Chalk, compiled by e. owen; edited by a. b. smith. 306 pp., 59 plates. Price £11 50 (U.S. $23)

(Members £9 90 or U.S. $20).

3. (1988): Zechstein Reef fossils and their palaeoecology, by N. hollingworth and t. Pettigrew, iv + 75 pp. Price £4 95

(U.S. $10) (Members £3-75 or U.S. $7 50).

4. (1991): Fossils of the Oxford Clay, edited by d. m. martill and J. d. Hudson. 286 pp., 44 plates. Price £15 (U.S. $30)

(Members £12 or U.S. $24).

1982. Atlas of the Burgess Shale. Edited by s. conway morris. 31 pp., 24 plates. Price £20 (U.S. $40).

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, 1993

Palaeontology

VOLUME 36 - PART 4

CONTENTS

Pterygometopine trilobites from the Ordovician of Baltoscandia
v. jaanusson and l. ramskold 743

Problematical microfossils from the Silurian of Ireland and
Scotland

A. FERRETTI, C. H. HOLLAND and E. SYBA 771

Ontogeny of the eodiscid trilobite Shizhudiscus longquanensis
from the Lower Cambrian of China

ZHANG XI-GUANG and E. N. K. CLARKSON 785

Phylogeny and evolution of ‘pentameride’ brachiopods

s. J. CARLSON 807

Biology and evolution of the nasal region in trematopid
amphibians

d. w. dilkes 839

The Lower Cambrian trilobite Bigotina and allied genera

G. l. pillola 855

Implications of the amino acid composition of Recent New
Zealand brachiopods

D. WALTON, M. CUSACK and G. B. CURRY 883

The prosauropod dinosaur Azendohsaurus laaroussii from the
Upper Triassic of Morocco

F.-X. GAUFFRE 897

Taxonomy, evolution and biostratigraphical importance of the
Llandovery graptolite Spirograptus

D. K. LOYDELL, P. STORCH and M. J. MELCHIN 909

Retrodeformation of fossils -a simple technique

A. W. A. RUSHTON and M. SMITH 927

Role of the shell structure in the classification of the
orthotetidine brachiopods

A. WILLIAMS and C. H. C. BRUNTON 931

The status of the nothosaurian reptile Elmosaurus lelmensis, with
comments on Nothosaurus mirabilis

O. RIEPPEL 967

Printed in Great Britain at the University Press , Cambridge ISSN 0031-0239

>' 5

z co

INSTITUTION

_

NOIiDillSN! NVINOSHilWS^

saiavaan

x

Wj#-, /ft/:. CO ipc 2S51 CO

/ t xZKauk?/ s .

> s

LI BRAR I ES^SMITHSONIAN INSTITUTION ” N
CO = to

INSTITUTION

si

NOIlfUIiSNI

NviN0SHiiws"S3iavaan

CO

X
co
o
z
> "

LI BRAR I ES^SMITHSONIAN
co

INSTITUTION NC
co

vss

SMITHSONIAN

INSTITUTION NOliniUSNS NVIN0SH1IINS

saiavaan li

F=

nvinoshiiins Siiavaan

CO ;

<jnj, uaS>

LIBRARIES

sa i avaa n

X w

LIBRARIES SMITHSONIAN

CO

5 xv >

INSTITUTION NOIlfUIiSNI
co

co

nvinoshiiins S3 1 avaa n

w co

> S \\XV > _ . _

.SNI_NVIN0SH1IWSWS3 I bVU 8 ll^LI B RAR I ES^SMITHSONIAN^lNSTSTUTION ^NOIlfUllSNI^NVINOSHim!

— w — = V) =: in z

co

o

z

5 “

z _ _ _

IES SMITHSONIAN INSTITUTION NOIlfUllSNI NVINOSHIIINS SliaVHan LIBRARIES SMITHSONIAI
z C z r* . z •” ^

O — /S'WX O ><5^vaT\ jvjCA ■: "*) O

v : m g s "X £ m

— W — . {/) *>. 5 co «—

,SNI NVIN0SH1IINS S3 1 ava an LIBRARIES SMITHSONIAN INSTITUTION NOIlfUllSNI NVINOSHIIW!
z _ w z in -r in

CV* Z /IWfftTA -H ' Z H /H

Vfc.W. T /ovA^-™-v q

CO

SksX “ Vh’C sUkW^y ’ Vk>'» <<r- 'v//7'syy// ' ' v.vm

I > '«r-' 2 ^ g I i

I ES SMITHSONIAN_ INSTITUTION NOIlfUliSN^NVINOSHlIINS^SS I HVH 3 Ilf LI B RAR I ES^SMITHSONIAP

“ s | 2

b; ft- «= VSJrc,..^?/ Of lo^ <3*/ C P ViMK. 3?/ c

—I vv*

0 z 5 5 v >

SNI^NVIN0SH1I!AIS2S3 I aVB a n IS B RAR I ESSSMITHSONIAN"J|NSTITUTION2:NOIinillSNl“,NVINOSHll!NS

1 40th l 40th S 4% 1 40*k I A 5 "■»»

»

> »“ 8SS; > p|£ ^S$ P > p^ 29$ p °$£ SH >

75 P \?^.,i«o/ 33 life . . ,^1 r- m/{, ■ 35 — ilw* , . b»o / 33

r*1 s /\?T 2 N>siyAS^>/r |TJ X^VOSVA^^ ^ jfTJ X^OSV^^ ^ HI

^ — CO CO 5 CO

ES SMITHSONIAN INSTITUTION NOIlfUllSNI NVINOSHillNS SBiavaOll LIBRARIES SM1THS0NIAP
S , z \ in z in z v </>

5 ^§1|K 1 X § 1 - v s

ill as 1 34 i Wm I I » s s

Wf I #1

o

<S) v

| S It 5 ^.w

SNI_NVIN0SH1IINS^S3 I a Va a n^LI B RAR S £SOTSMITHSONIAN2|NSTJTUTlON 00 NOIlfUllSNI 2 NVIN0SH1IIN5
z \ y? - — . i __ y> 5 __ 5 *”

ES^ SMITHSONIAN INSTITUTION NOIlfUllSNI NVIN0SH1SWS SilHVHail LIBRARIES SMITHSONIAr

CO
33
>

33

m ^iL£S^ g ^ w m X^Tc

SNI NVIN0SH1IINS S3 I HVa 8 11 LI B RAR I ES^SMITHSONIAN”lNST]TUTlONS° N0linillSN!™NVIN0SHlUA|5

«t- .r/ .CO 2^ CO 21

ES SMITHSONIAN^lNSTiTUTION NOUfUllSN! NVINOSHIIIAIS S3iavaail LIBRARIES SMITHSONIAP

2 CO *=

^ m w ^ JK.: UJ S uj