u> -Z in •** z _

JTION NOIinillSNI_ NVINOSHIIWS S3iyvy8H LIBRARIES SMITHSONIAN INSTITUTION NOlin.

\ <o ~ m — tn

in

>

Z

in

o -^vjr ~ o 5

2 J 2 j z _i

M8n LIBRARIES SMITHSONIAN INSTITUTION NOlinillSNI NVINOSHIIWS S3IHVHan LIBRA

z r~ 2 r-

>

30 (- ™

' 2 m co m co m

co - co ± — w \ z tn

jtion NouniiiSNi NViNOSHims saiyvyan libraries Smithsonian institution Noun

Z _ CO Z , y to 2 co

^ E . . <■* ^ E ^

\

b.

y a n_ u b r ar i es Smithsonian institution NouniiiSNi nvinoshiiws sgiyvaaii libra

-?> tn c/> — . c/> —

v}S\ / « /^SOf^/Sv Ui / >s XlMSOV^S. UJ iJi\ • * l y^TifwyTS. ^

O xgN.pc^ ~ W" O 5 ^ 5 N^%j

JTION ~ NOlinillSNI^NVINOSHimS^Sa lyvyail^LIBRAR IESZSMITHSONIAN~'lNSTITUTIONZNOIin

z — ” > z r~ 2 f~ 2

m ' \x> -7- m v/\os*sz ^ m

— CO — to Z o _

y 8 IT LI B RAR I ES^SMITHSONIAN ^INSTITUTION NOlinillSNI _NVIN0SH1IWS _ S3 I y VH 8 |"| _LI B RA
<

v''W* - 'v!^fNW^9/ t Z ,

S s > s X^vos^ >

in c/> *'* 2! to — 2! CO

TION NOlinillSNI _ NVINQSH1IWS S3iyvy8IT LIBRARIES SMITHSONIAN INSTITUTION NOlin.

z \ y? r-. 5 CO ^ - co

< i "

O 5 2- ' O 2

Han LIBRARIES SMITHSONIAN INSTITUT!0N~N0liniJlSNrNVIN0SHlllNS~S3iyvy8n~LIBRA

f z y r- z r-

^ m 2 , . ,cd 2 m ^>v 2 ro

m x^vpc^ — m -^2i^ ~ xvo7iB2X m X v> n^u^/ m

co — co £ — to \ £ to

JTION NOlinillSNI NVIN0SH1IINS S3IHVyail LIBRARIES SMITHSONIAN INSTITUTION NOlin.

CO 2 to Z CO 2 CO

S •- t ^ 4KW E i s

z -I .y ,.:/ V?* Z «\> =? /&f?T>&v z

£ I (Ir SI) ° il6k x w ^

§ W £ i IW?# g *% § %

* Z CO Z CO * Z ^ V 2

y 8 n_ LI B RAR I ES SMITHSONIAN INSTITUTION NOlinillSNI NVIN0SH11WS S3iyvyon LIBRA

CO ^ CO — CO z

K

> '

^ X£V OS^iX > -C- ,W £" *2

^ — z to N z to

noshxiws samvaan libraries smithsonian institution NoixnxixsNi_NViNOSHxms sh

to _ = to zz _ <n _ . z

«4 *

m

o '<!$oZ£y o '<»' 3 o

ITHSON IAN ”* INSTITUTION*2 NO I inXIXSNI^NVINOSHiUAIS S 3 I a V 8 8 I 1^ L I B R A R I ES^ SMITHSONIAN^IN:

/ x fiX I /^£5\ ® § xgsx 5 S ,„/ - - 5

• ■“ / 0/!V ’m.-r\ -J ^

^ ' »_ lote» .ago

«/ 73

S' X^'S¥'‘Tr .

'cWn«i\ 73 ’-* [ges" r#ir\ ~ IzL . ,

h '#SKS

,<sy >- X'lZtorAkwj/ — "*•' "s&x •- \vxm35','xy — vc/w^xy wvsk

m -vjv. to xSqxn^z m E2 x^msgx rn xfo dcz z f11

CO « to . ^ co — to

INOSHXIWS SBIBVyen LIBRARIES SMITHSONIAN INSTITUTION^ NOIXnXIlSNI NVINOSHXIWS S3

il 8 JSs x ^Hk' I fJ^jl I 1 ^

5 s ^ 5 > 'WA 2

^ (/) * Z (/) 2 CO ^

ITHSONIAN INSTITUTION NOlinillSNI NVIN0SHX11MS S3 I a VH 8 ll_ LI B R AR I ES SMITHSONIAN _ IN
— CO ~ _ co -p ^ ^

/ Aj , ,

, Jpw c

5 5 ^ 5 U| 'W o

INOSHIIWS^S 3 I aVBSIl^UBRARI ES^ SMITHSONIAN^INSTITUTION ^NOIXnXIXSNI^NVINOSHXIWS^S)

O /TXtTtjTn. Z^o~aSv. O m zf^o^oX 5 m

m £ * m ^ X_;v^z rn

*— (f) ^

iithsonian'institution NouniiiSNi~NviNOSHims S3 1 a va 8 n li b r ar i e$ smithsonian ir

-r ■ rn Z CO Z •«. ~f ^

fii

' x Wv

S ' N^v > " S X^osv^y > ^ g 2

nNOSHlUNS^0 S3 I ava 8 nZLI B RAR I ES^SMITHSONIAN INSTITUTION NOIinillSNI_NVINOSHillMS S

to - to - ^ to — z \ “2

uj xS5Rn^s. ^ ui ,., ^

t\ _i A^ma^SK — mwAA

O

z

<A Hi

^ §

O ^osvanJX — O

_ Z -J Z -J

^ITHSONIAN INSTITUTION NQiXnj.IJ.SN! NVINOSHXUAIS S3iava91 ^LIBRARIES SMITHSONIAN^ I

z
co

73
>

73

rn Dt^z ^ m S2 XvgjiTs^y m ^^dcJX ^

C/) ~ CO ^ CO —

tflNOSHXIINS S3 lava 8 11 LIBRARIES SMITHSONIAN INSTITUTION NOIXnXIXSNl NVINOSHXIIMS S
to z .,. w z >v.

#0# o |Sg% 3 /i^l i 3 5

= CO

CO

4 £ 5 z

_ ' ^ ^ >

^ co ^ z co v z co

VIITHSONIAN INSTITUTION NOIXOXIXSNI NVINOSHXIWS SBiavaail LIBRARIES SMITHSONIAN _ I
— to — <o 5 00 z

<CkvaT\ ~ jXS. .. U* /Cvtsv,a.r\ ,n *^3 .CO

VOLUME 11 • PART 3

Palaeontology

JUNE 1968

PUBLISHED BY THE
PALAEONTOLOGICAL ASSOCIATION
LONDON

Price £3

THE PALAEONTOLOGICAL ASSOCIATION

The Association was founded in 1957 to further the study of palaeontology. It holds
meetings and demonstrations, and publishes the quarterly journal Palaeontology.
Membership is open to individuals, institutions, libraries, etc., on payment of the
appropriate annual subscription:

Institute membership £7. 0.y. (U.S. $20.00)

Ordinary membership £5. Os. (U.S. $13.00)

Student membership £3. 0s. (U.S. $8.00)

There is no admission fee. Student members are persons receiving full-time instruction
at educational institutions recognized by the Council; on first applying for member-
ship, they should obtain an application form from the Secretary or the Treasurer. All
subscriptions are due each January, and should be sent to the Membership Treasurer,
Dr. A. J. Lloyd, Department of Geology, University College, Gower Street,
London, W.C.l, England.

Palaeontology is devoted to the publication of papers (preferably illustrated) on all
aspects of palaeontology and stratigraphical palaeontology. Four parts at least are
published each year and are sent free to all members of the Association. Members
who join for 1968 will receive Volume 11, Parts 1 to 5.

All back numbers are still in print and may be ordered from B. H. Blackwell,
Broad Street, Oxford, England, at £3 per part (post free). A complete set, Volumes
1-10, consists of 38 parts and costs £114.

Special Papers in Palaeontology is a series of substantial separate works published
by the Association. The subscription rate is £6 (U.S. $16.00) for Institute Members
and £3 (U.S. $8.00) for Ordinary and Student Members. Subscriptions and orders
by members of the Association should be placed through the Membership Treasurer.

The following Special Papers are available. Members may obtain them at reduced
rates through the Membership Treasurer. Non-members may obtain them from
B. H. Blackwell, Broad Street, Oxford, England, at the prices indicated.

Special Paper Number One (for 1967): Miospores in the Coal Seams of the
Carboniferous of Great Britain, by A. H. V. Smith and M. A. Butterworth.
324 pp., 72 text-figs., 27 collotype plates. Price £8 (U.S. $22.00), post free.
Special Paper Number Two (for 1968): Evolution of the Shell Structure of
Articulate Brachiopods, by Alwyn Williams. 55 pp., 27 text-figs., 24 collotype
plates. Price £5 (U.S. $13.00).

Special Paper Number Three (for 1968): Upper Maestrichtian Radiolaria of
California, by Helen P. Foreman. 83 pp. with 8 collotype plates. Price £3 (U.S.
$8.00).

Typescripts on all aspects of palaeontology and stratigraphical palaeontology are
invited. They should conform in style to those already published in this journal, and
should be sent to Mr. N. F. Hughes, Department of Geology, Sedgwick Museum,
Downing Street, Cambridge, England, who will supply detailed instructions for
authors on request (these are published in Palaeontology, 10, pp. 707-12).

© The Palaeontological Association 1968

THE FEEDING MECHANISMS AND
AFFINITIES OF THE TRIASSIC
BRACHIOPODS THECOSPIRA ZUGMAYER
AND BACTRYNIUM EMMRICH

by M. J. S. RUDWICK

Abstract. The morphology of Thecospira and Bactrynium is analysed functionally. It is inferred that the spiral
brachidium of Thecospira, with its unique grooved lamellae, bore a simple spirolophe, and that the Tobate
apparatus’ of ridges and grooves in Bactrynium bore a complex ptycholophe. The morphology of both genera
is consistent with ciliary feeding mechanisms similar to those of living brachiopods: it is inferred that Thecospira
operated a spirolophous current system of ‘exhalant’ or ‘spirifer’ type, analogous to the living Discinisca ; and
that Bactrynium operated a ptycholophous system analogous to the living Megathiris. Both brachiopods were
cemented in early growth stages but later became free-lying; both had a strophic hinge, and normal articulation
and musculature. Bactrynium is pseudopunctate, Thecospira includes punctate and impunctate (or obscurely
pseudopunctate) species. Affinity with Permian Davidsoniacea is accepted for Thecospira', the nature of its
derivation is discussed in functional terms. Bactrynium is assigned to the Thecideacea; its resemblance to Permian
Lyttoniacea is interpreted as due to functional convergence. It is concluded that the Thecideacea were derived
from Davidsoniacea, probably during Permian time and probably by neoteny; affinity to Spiriferida or Tere-
bratulida is thus rejected, and the older view reaffirmed, that Thecideacea are surviving Strophomenida. The
phylogenetic history of the feeding mechanisms of Thecideacea and Thecospira is interpreted in terms of a
concept of ‘functional zones’, and is regarded as an example of ‘size-required allometry’.

The small and rare Triassic brachiopods Thecospira and Bactrynium are of special
interest in the study of the functional evolution of the phylum. Of all the changes in
brachiopod faunas between the late Palaeozoic and the early Mesozoic, the most
striking is the almost complete disappearance of the Strophomenida. Strophomenides
were abundant and diverse in the Permian period, but their representatives in the
Triassic are rare, and so changed in form that their affinities have remained contro-
versial.

One such relict strophomenide is the Middle to Upper Triassic Thecospira. Its stro-
phomenide affinities are now generally recognized, and it is regarded as the only known
post-Palaeozoic davidsoniacean (Williams 1953n, 1965). But it has undergone significant
morphological changes, of which the most important is its acquisition of a spiral brachi-
dium with grooved lamellae.

Another relict strophomenide, the late Triassic (Rhaetian) Bactrynium, is more
problematical. It has a concavo-convex shell resembling that of many Palaeozoic
strophomenides. The interior of the dorsal valve bears a series of multi-lobed ridges and
grooves which give it a striking resemblance to the Permian lyttoniacean Oldhamina.
On the other hand it was originally regarded as being related to the post-Palaeozoic
thecideaceans. The functional analysis of Bactrynium in this paper is designed to help
to distinguish more clearly those characters that are likely to be due to affinity from those
that result from functional parallelism or convergence.

Discussion of Thecospira and Bactrynium naturally bears on the controversial problem
of the affinities of the thecideaceans. Williams’s and Rowell’s recent contribution to this

[Palaeontology, Vol. 11, Part 3, 1968, pp. 329-60, pis. 65-68.]

C 5586 Z

330

PALAEONTOLOGY, VOLUME 11

problem (1965, pp. H187-9) seems to over-stress features indicating possible affinity to
spiriferides or terebratulides, at the expense of those that have previously suggested a
connexion with the strophomenides. I believe that a strophomenide affinity for the theci-
deaceans should be given further consideration.

Taxonomy and material. The genus Bactrynium was erected by Emmrich in 1855 for a
‘Problematicum’ from the Kossener Schichten (Rhaetian) of Austria. Pterophloios
Gtimbel 1861 is a junior synonym. The structure of Bactrynium was first clearly
described by Zugmayer (1880), who emphasized its similarities to ‘ Thecidea ’, i.e., in
modern terms, to the Thecideacea. At the same time Zugmayer erected the genus
Thecospira for small Triassic shells that had previously been assigned to ‘ Thecidea ’, but
in which he discovered a spiral brachidium of a unique kind.

For some time Thecospira was commonly regarded as a spiriferide, purely on account
of its brachidium, but Williams (1953n) showed that in all other respects it is clearly
related to the Strophomenida, and in the Treatise (Williams 1965) it is placed in a
monotypic family Thecospiridae within the Davidsoniacea.

The affinities of Bactrynium have remained more controversial. Waagen (1883) and
Bittner (1890) agreed that it differed from ‘ Thecidea ’ and either represented a transition
from Oldliamina and related genera (i.e. Lyttoniacea) or else belonged to that group.
In their studies of Permian Lyttoniacea from Timor, Broili (1916) and Wanner (1935)
concurred in this judgement; but, on the other hand, Noetling (1905) and Kozlowski
(1929) argued that its similarities to the lyttoniaceans were the result of convergence.
The former view has been maintained more recently by Stehli (1956); and the latter by
Sarycheva (1964), and also by Makridin (1960) in the Osnovy (where Bactrynium
appears under the name Pterophloios in the Thecideacea). In the Treatise, on the
other hand, it is given a monotypic family Bactryniidae, which is placed provisionally
among the Lyttoniacea (Williams 1965).

Bactrynium is known only from the Rhaetian of Austria. Its rarity is such that many
features of the morphology remain as obscure as they were in Zugmayer’s day. Never-
theless enough can be determined from the available specimens to make a functional
reconstruction feasible. I have re-examined Zugmayer’s original specimens of the type
species B. bicarinatum (Emmrich), which are now in the Palaontologisches Institut der
Universitat, Vienna (PIUW), together with others in the Naturhistorisches Museum,
Vienna (NHMW), the U.S. National Museum, Washington (USNM), and the British
Museum (Natural History), London (BMNH).

Thecospira is less rare, but is never a common fossil. It has a wider stratigraphical
range, from Karnian to Rhaetian, but is confined to the East Alpine region. The best
specimens are from the Raibl and St. Cassian beds (Karnian) of the Italian Alps, and
from the Kossener and Starhemberger Schichten (Rhaetian) of Austria. I have re-
examined Zugmayer’s original specimens of the type species T. haidingeri (Suess),
together with other specimens of that and other species, in the museums already listed.
I have also collected Thecospira from the St. Cassian beds near Cortina d’Ampezzo;
these specimens are now in the Sedgwick Museum, Cambridge (SM).

Acknowledgements. I am particularly grateful to Dr. R. Cowen for invaluable help with the preparation
of this paper. I am indebted to Dr. F. Steininger, Professor Dr. H. Zapfe, Dr. G. A. Cooper, and
Dr. C. H. C. Brunton, respectively, for access to the collections mentioned above, and for loans of

RUDWICK: TRIASSIC BRACHIOPODS THECOSPIRA AND BACTRYNIUM 331

some specimens. Signor R. Zardini kindly gave advice on field-work, and supplied photographs of
specimens in his private collection. My wife gave valuable assistance with collecting, and drew
some of the text-figures. Part of this work was done under a research grant from the Science Research
Council.

GENERAL FUNCTIONAL MORPHOLOGY

Relation to substratum. Both Thecospira and Bactrynium are small brachiopods, the
largest dimension of the shell being generally less than a centimetre. Both have a roughly
elliptical commissure plan, the major axis being transverse in Thecospira and longitudinal
in Bactrynium. In the latter, the commissure may also have a pair of postero-lateral
horizontal deflections (cf. Rudwick 1959, p. 10) which produced corresponding ear-like
extensions of the valves themselves. The ventral valve of both brachiopods is strongly
and fairly uniformly convex. The dorsal valve of Thecospira is weakly convex near
the umbo, becoming plane and then weakly concave towards the commissure: in other
words the convexity was reversed gradually during ontogeny. The dorsal valve of
Bactrynium is more strongly concave; available specimens do not show clearly whether
it too passed through a convex stage early in ontogeny.

The terms dorsal valve and ventral valve are used throughout this paper, because brachial valve and
pedicle valve are, in my opinion, misleading and inappropriate : in living articulates neither the brachia
nor the pedicle ‘belong’ to one valve more than the other, in embryonic origin or in adult anatomy;
and in many fossil articulates no pedicle existed and there is no direct evidence of the brachia.

There is no trace of a pedicle foramen in either brachiopod, but there is a prominent
cicatrix (‘scar’) of attachment at the apex of the ventral valve. Its appearance suggests
that the young shells were generally cemented to shell fragments. The size of the cicatrix
shows that when the ventral valve had grown to some 2-4 mm. in breadth its edge began
to grow away from the surface of attachment, with consequent increase in the convexity
of the valve. Concurrently, at least in Thecospira, the dorsal valve became plane and
then concave. The orientation of the cicatrix in larger specimens, and its size relative to
the whole shell, make it unlikely that the cementation attachment remained effective
throughout life. The enclosing argillaceous sediments suggest that apart from scattered
shell fragments the substratum was probably soft and muddy. As they grew in size,
both brachiopods probably outgrew the shell-fragments on which the spat had settled,
and gradually ‘heeled over’ until they lay freely on the ventral valve, resting on or partly
in the soft substratum (text-figs. 8, 10). The ever-increasing convexity of the ventral
valve could have ensured that the commissure remained above the surface throughout.
A similar relation to the substratum, with an attached stage gradually superseded by
a free-lying stage, has probably been common in several brachiopod groups with
concavo-convex shells (Rudwick 1965, p. H203; Grant 1966).

Hinge and musculature. There is a moderately broad strophic hinge-line in Thecospira.
Growth on the hinge-line sectors of the valves produced broad plane interareas on both
valves, that on the dorsal valve being very low (text-fig. 1a, PI. 65, figs. 1, 3; PI. 66, figs.
17, 18; PI. 68, figs. 12, 13). Contrary to the description in the Treatise (Williams 1965),
Bactrynium is also strophic, with a very broad hinge-line. This is most clearly seen in
a specimen of Zugmayer’s, where the dorsal valve is bounded posteriorly by a low
plane interarea similar to that of Thecospira (text-fig. 2a; PI. 68, fig. 2). Williams (1965)

332

PALAEONTOLOGY, VOLUME 11

notes correctly that the ventral valve shows no trace of the ‘posterior flap’ of shell
material that is characteristic of lyttoniaceans.

In Thecospira the pseudodeltidium and chilidium are often precisely flush with the
interareas; the growth-lines on the interareas thus cross the mid-line without deflection

text-fig. 1. Drawings of Thecospira, to illustrate morphology.

a. Postero-dorsal view of T. haidingeri (Suess), x4 (specimen as in PI. 65, figs. 1-4).

b. Ventral view of dorsal valve of T. tyrolensis (Loretz), x4 (specimen as in PI. 65, figs. 12, 13).

c. d. Oblique and ventral views of dorsal valve of T. haidingeri, X4-5 (specimen as in PI. 65, figs. 8,9.)
E. Dorsal view of ventral valve of T. haidingeri, X4-5 (specimen as in PI. 65, fig. 7).

(text-fig. 1a; PI. 65, fig. I? PI. 66, figs. 17, 18). In some shells the pseudodeltidium is
more clearly delimited (PI. 68, figs. 11, 12); and there are other Triassic shells originally
described as ‘ Thecidea ’, which may in fact be Thecospira or closely related to it, in which
the pseudodeltidium is slightly but distinctly convex (PI. 68, fig. 13).

EXPLANATION OF PLATE 65

Figs. 1-11. Thecospira haidingeri (Suess), Rhaetian, Austria. 1-4, Dorsal, ventral, posterior and left
lateral views of whole shell x3 (PIUW; figured Zugmayer 1880, pi. 2, fig. 33), from Kitzberg.
5-6, Details of ventral and dorsal valve surfaces of same specimen X 9. 7, Interior of ventral valve
x3 (PIUW; figured Zugmayer 1880, pi. 2, fig. 35), from Kitzberg. 8-9, Ventral and oblique pos-
terior views of interior of dorsal valve x3 (PIUW; figured Zugmayer 1880, pi. 2, fig. 36), from
Diirnbach. 10-11, Ventral and oblique posterior views of spiral brachidium in dorsal valve, exposed
by grinding and polishing x9 (PIUW; figured Zugmayer 1880, pi. 2, fig. 39), locality unknown.
J, jugum; C, crus. 12-13, Thecospira tyrolensis (Loretz), Raibler Schichten, Romerlo, near Cortina
d’Ampezzo, Italy. Ventral and oblique posterior views of interior of dorsal valve X 3 (NHMW).

Palaeontology, Vol. 11

PLATE 65

RUDWICK, Thecospira

73

RUDW1CK: TRIASSIC BRACHIOPODS THECOSPIRA AND BACTRYNIUM 333

The form of the pseudodeltidium and chilidium in Bactrynium are not known with
certainty. The fragmentary dorsal valve already referred to is not well enough preserved
to show the chilidium clearly, though it was probably convex and arched over the base
of the cardinal process. ‘ Thecidea bicar inata ’ Klipstein, which may be congeneric, has
a convex pseudodeltidium and chilidium.

Thecospira has a pair of strong teeth, unsupported by dental lamellae, projecting
forwards from beneath the pseudodeltidium (PI. 68, fig. 1 1). The corresponding sockets
are defined medially by the sides of the large rectangular cardinal process, and anteriorly
by ‘socket plates’, which are simply low ridges formed by the sides of the cardinal

muscle socket

hinge-line

text-fig. 2. Morphology of Bactrynium.

A. Postero-ventral view of fragment of dorsal valve, showing strophic hinge and low interarea, car-
dinalia, muscle scar, and posterior region of ‘lobate apparatus’, X 10 (specimen as in PI. 68, fig. 1, 2).
b. Longitudinal section through shell, showing concavo-convex form, with growth tracks of lateral
septa, X 5 (specimen as in PI. 68, fig. 9).

process curving antero-laterally to merge with the floor of the valve (text-fig. 1b-d;
PI. 65, figs. 8, 9, 12, 13; PI. 66, fig. 16). The teeth of Bactrynium are not yet known, but
the cardinalia in the dorsal valve are very similar to those of Thecospira , though the
cardinal process is relatively smaller (text -fig. 2a; PI. 68, fig. 1, 2). These hinge structures
are in no way abnormal.

The degree of rotation possible with this articulation cannot easily be inferred. The
hinges have some mechanical resemblance to that of the living Lacazella , in which the
angle of opening approaches 90° (Lacaze-Duthiers 1861). Possibly the shells of Theco-
spira and Bactrynium opened to a similar extent. The feeding mechanisms reconstructed
in this paper would be feasible with either a small or large angle of opening.

The cardinal process of Thecospira is indistinctly bilobed (text-fig. 1b-d; PJ. 65, figs.
8, 9, 12, 13; PI. 68, fig. 1 2), and the diductor attachments are visible distally. The cardinal
process projects into a deep cavity beneath the pseudodeltidium (PI. 68, figs. 11, 12). On
the floor of the ventral valve the diductor attachments are large and flabellate (text-fig.
1e; PI. 65, fig. 7). Between them are the narrow lanceolate attachments of the adductors.
I have not seen the median septum in the ventral valve mentioned in the Treatise
description. The adductor scars on the dorsal valve are elliptical, are not divided into
anterior and posterior pairs, and are separated in the mid-line by a long narrow ridge
(text-fig. 1b-d; PI. 65, figs. 8, 13; PI. 66, fig. 16).

334

PALAEONTOLOGY, VOLUME 11

The cardinal process of Bactrynium is also indistinctly bilobed. Lateral to its base is
a smooth transversely elliptical area (text-fig. 2a; PI. 68, fig. 2), which closely resembles
the scar of the ‘lateral adductor’ muscle in the living LacazeUa and the similar scars in
Jurassic thecideaceans such as EudeseUa and Davidsonella (PI. 68, figs. 3, 4). No other
scars are visible on the dorsal valve, but the preservation is not ideal; the scars on the
ventral valve are not known at all.

In Thecospira the muscular leverage involved in opening and closing the shell was in
no way unusual. The musculature is here reconstructed (text-fig. 8) on the conjectural
assumption that the muscles were columnar (as in living thecideaceans and inarticulates)
and not tendonous (as in living rhynchonellides and terebratulides). If the adductors were
differentiated into striated ‘quick’ and unstriated ‘catch’ portions, these portions were
not separated spatially like the anterior and posterior adductors of living rhynchonellides
and terebratulides. The orientation of the diductors shows that a wide degree of opening
of the shell would have been mechanically possible, if allowed by the articulation.

As far as available evidence goes, the muscular leverage of Bactrynium was also
‘normal’, apart from the lateral shift of the main adductors, which is also known in the
thecideaceans.

In both brachiopods the soft tissues would have been sealed extremely effectively and
strongly when the shell was closed, for the lid-like dorsal valve has a strongly thickened
submarginal rim (text-fig. 1b-d; PI. 65, figs. 8, 9, 12, 13).

‘ Ornament ’ and shell structure. Primary-layer shell covers the exterior of both valves in
the normal way. In Bactrynium it is marked only with finely wrinkled growth-lines (cf.
Bittner 1890, pi. 26, fig. 18). In Thecospira there are small tear-shaped pustules, oriented
radially and apparently scattered over the entire shell surface (PI. 65, figs. 5, 6). Where
best preserved, they appear to have been short slender spinelets (PI. 65, fig. 6, central
area of dorsal valve); and among them are faint traces of radial costellae (PI. 65, fig. 6,
umbonal region). Some shells from the Cortina area have fewer pustules and much more
conspicuous radial costellae, with new costellae added by intercalation (PI. 68, fig. 12).
There are no large tubular spines of productoid type on either brachiopod.

A detailed study of shell structure is beyond the scope of this paper, but the known
structures must be mentioned briefly in view of their prominence in discussions on
affinity. Bactrynium is said to be pseudopunctate (Williams 1965), and this seems to be

EXPLANATION OF PLATE 66

All figures X 3. Figs. 1-7, 13-18 are whitened with ammonium chloride.

Figs. 1-14. Bactrynium bicarinatum Emmrich, from the Kossener Schichten, Austria. 1-4, Ventral,
postero-ventral, posterior and left lateral views of interior of dorsal valve of USNM 108873, from
Baytal, Gumpoldskirchen, near Vienna (Williams 1965, fig. 397 shows same specimen before pos-
terior spike was broken off). 5-7, Ventral, oblique and postero-ventral views of interior of dorsal
valve of USNM 129877a, from the same locality. 8, Interior of fragment of dorsal valve from Diirn-
bach (PIUW; figured Zugmayer 1880, pi. 2, fig. 18). 9-12, Right oblique, ventral, left oblique and
right lateral views of interior of dorsal valve of specimen from the same locality (NHMW). 13, 14,
Ventral and posterior views of interior of dorsal valve of USNM 129877b, from Baytal. 15, Theci-
diopsis hieroglyphica (Goldfuss), from the Maestrichtian (late Cretaceous) of Limburg, Flolland.
Interior of dorsal valve of SM.F.31 1 1. 16-18, Thecospira sp., from the St. Cassianer Schichten, Alpe
di Specie (Seelandalpe), near Cortina d’Ampezzo, Italy. 16, Interior of dorsal valve of SM.G.1261.
17, 18, Postero-dorsal and posterior views of SM.G.1262.

Palaeontology, Vol. 11

PLATE 66

RUDWICK, Bactrynium, Thecospira, Thecidiopsis

RUDWICK: TRIASSIC BRACHIOPODS THECOSPIRA AND BACTRYNIUM 335

confirmed by a study of Zugmayer’s original sections, though these are crude by
modern standards. The structure of Thecospira is more problematical. Suess noted in
his original description of T. haidingeri that it was punctate (1854, p. 44), and this is very
clearly shown by Zugmayer’s specimens (PI. 65, figs. 5, 6). But Bittner (1890) com-
mented that although some species are punctate, others are apparently impunctate.
Cellulose peels taken from several specimens confirm that there is a genuine diversity of
shell structure within the genus. This must clearly be taken into account in any future
systematic revision; for the present, it is sufficient to outline the nature of the diversity.

text-fig. 3. Morphology of Thecospira.

A. Diagram based on a polished longitudinal section through a dorsal valve of Thecospira sp., showing
punctae coalescing as they traverse the shell; external surface above. More punctae are shown than
actually lie in the plane of the section; since they are picked out by matrix they are visible for a
further part of their course. St. Cassianer Schichten, Alpe di Specie, near Cortina d’Ampezzo, Italy;
SM.G.1260.

b. Reconstruction of brachium of Thecospira : cross-sections of two adjacent whorls, with filaments in
inferred natural orientation; for further explanation see text.

Specimens of Thecospira sp., from the St. Cassian beds of the Alpe di Specie [See-
landalpe], near Cortina, are punctate, but the punctation is most unusual. Closely
spaced, thick punctae of the exterior shell layers coalesce into comparatively few, more
widely spaced punctae as they cross the shell thickness (text-fig. 3a). Probably the struc-
ture is the same as that described loosely by Bittner in Thecospira davidsoni (1890, p. 287).
Similar types of punctation are known elsewhere in the phylum only in some inarticu-
late brachiopods (e.g. Crania) and in the Silurian enteletacean Dicoelosia (Wright 1966);
but all three occurrences of this type of punctation are likely to represent independent
developments of the structure.

T. tyrolensis (Loretz) is impunctate, but the shell structure appears complex because
of the presence of small pimples or pustules on the inner surface of the valves. As shell
growth proceeds, the pustules are incorporated in the secondary shell layer. T. guembeli
(Pichler), on the other hand, has a shell structure which can be regarded as obscurely
pseudopunctate; there is a ‘normal’ primary and secondary-layer structure with an
inner layer in which the orientation of the crystals is irregular.

Coelom and body wall. In living articulates the coelom, bounded by the body wall and
containing the muscles, gut, etc., is confined to a small postero-median part of the shell
cavity. The body- wall is ‘wrapped’ closely round the anterior and lateral surfaces of the

336

PALAEONTOLOGY, VOLUME 11

muscles, and laterally it incorporates the crura. A consistent reconstruction of the body
wall of Thecospira can be made by homology. The crura or ‘ apophyses ’ project forwards
from the ‘socket plates’ at the base of the cardinal process, in a position consistent
with a body-wall wrapped around the lateral surfaces of the muscles (text-figs. 1b-d,
8, 9).

A reconstruction of the gut, though given here for the sake of completeness, is neces-
sarily conjectural. Its position and form as shown (text-fig. 9) are, however, consistent
with what is known of living articulates. An enlarged stomach and lateral digestive
glands would have fitted neatly into the space (otherwise empty) behind the adductors

A B

text-fig. 4. Diagrams to show (a) morphology of ‘lobate apparatus’ on dorsal valve of Bactrynium,
and (b) its interpretation as a supporting structure for a complex ptycholophous lophophore.

and above the diductors, occupying also the concavity at the base of the cardinal process
(text-fig. 1b-d; PI. 65, figs. 9, 13; PI. 66, fig. 16).

When the shell of Bactrynium is preserved closed there is little space between the
ridged internal surface of the dorsal valve and the corresponding surface of the ventral
valve. Posteriorly, however, the dorsal valve is ‘excavated’ (strictly speaking, it is not
thickened like the rest of the valve) into a broad median cavity (text-fig. 4a; PI. 66,
figs. 2, 10). There is a similar cavity in thecideaceans (PI. 66, fig. 15), and in the living
Lacazel/a this is known to accommodate the dorsal part of the ‘body’ (coelom) with
viscera and muscles. It is, therefore, reasonable to infer that the ‘body’ of Bactrynium,
like that of Thecospira, was confined to a small median-posterior portion of the shell
cavity (text-fig. 4b).

THE LOPHOPHORE

The brachidium in Thecospira. The spiral brachidium of Thecospira consists of a pair of
low conical spiralia, attached to the dorsal valve by a pair of crura or ‘apophyses’, and
linked to each other proximally, across the median plane, by a simple jugum.

RUDWICK: TRIASSIC BRACHIOPODS THECOSPIRA AND BACTRYNIUM 337

The crura are short peg-like processes, circular in cross-section, which project for-
wards from the ‘socket plates’ near the base of the cardinal process. In most specimens
they are broken off, and their position is only shown by a pair of small residual tubercles
on the ‘socket plates’ (text-fig. 1b-d; PI. 65, figs. 9, 12). Their connexion with the
spiralia is, however, visible on Zugmayer’s ground and polished specimen (PI. 65,
fig. 11). The crura manifestly had the same function as the crura of other brachiopods,
and they project from the cardinalia in essentially the same way; therefore it seems
hardly necessary to distinguish them by another term (‘apophyses’, cf. Williams 1965).
The differences, which are part of the evidence for regarding the brachidium of Theco-
spira as a separate development, are in the cardinalia rather than the crura themselves.

The spiral axis of each spiralium is inclined, with the apex of the spiral pointing
latero-ventrally. The spiralia occupy almost the whole of the shell cavity, and when the
shell is closed the spiral lamellae lie very close to the internal surface of the ventral
valve (cf. Bittner 1890, text-fig., p. 310). The lamella is unique, as Zugmayer emphasized,
by being U-shaped in cross-section. No such gutter-like lamella is known in any of the
Spiriferida. The gutter faces outwards around the spiral (PI. 65, figs. 10, 11). The
ventral limb of the ‘ U ’ is narrower than the dorsal. The ventral side passes proximally
into the jugum, the dorsal into the crura. The structure of the jugum is not very clear in
Zugmayer’s specimens, and it has not been detected with certainty in serial cellulose
peels. Possibly it consists only of crural processes, projecting towards the mid-line
without meeting each other (PI. 65, fig. 11).

Lobate apparatus o/Bactrynium. On the internal surface of the dorsal valve of Bactry-
nium there is a series of lobed ridges and grooves. This ‘lobate apparatus’ is approxi-
mately, but not accurately, symmetrical about the median plane. It consists of a pair of
longitudinally elongated primary lobes separated by a long median indentation, and up to
ten pairs of laterally elongated secondary lobes projecting laterally from the primary
lobes and separated from one another by lateral indentations (text-fig. 4a). An occa-
sional tertiary lobe branches from a secondary lobe as a minor irregularity (text-fig. 5b,
lobe 5 left). The lobes have sub-parallel sides, and approximate to a uniform width of
about 0-9 mm. in all parts of all specimens (text-fig. 5). The lateral indentations like-
wise have sub-parallel sides, and approximate to a uniform width of about 0-3 mm.
The median indentation has a similar width posteriorly, but broadens anteriorly. (The
uniform width of the lobes and indentations is partly obscured in text-fig. 5 by the
foreshortening of the peripheral parts, due to the strong convexity of the internal surface
of the valve.)

The lobes and indentations are outlined by the sinuous course of a closely contiguous
groove and ridge (text-fig. 4a). These can be seen with variable clarity in different speci-
mens; this is partly due to differences of preservation, but there may be some genuine
variation. The groove is internal in position, the ridge external. The steep slope between
them is always clearest : the course of this outer edge of the groove (or inner edge of the
ridge) is shown by a continuous line in text-figs. 4a and 5, and is the basis of the measure-
ments given above.

The ridge is clearly visible, and clearly raised above the rest of the valve surface, in
the broad anterior part of the median indentation (PI. 66, figs. 1, 2, 5-7, 10). In the
narrower posterior part the ridges of the left and right sides generally coalesce into

338

PALAEONTOLOGY, VOLUME 11

a single ridge; but in the best-preserved specimens the median indentation is clearly
‘double’ right to its tip (PI. 67, fig. 5; PI. 68, fig. 1). The posterior tip of the median
indentation projects as a conspicuous spike overhanging the postero-median cavity
(PI. 66, figs. 1, 10): this spike is vulnerable to damage and has generally been broken
off at its base. As in the median indentation, the ridges on either side of each lateral

text-fig. 5. Drawings of the ‘lobate apparatus’ on dorsal valves of Bactrynium, to show uniformity of
dimensions of lobes and indentations, all X 5. Continuous line indicates slope between groove and
ridge: dashed line, inner edge of groove. (Details of specimens given in captions to plates: a, as PI.
66, figs. 1, 2; b, as PI. 66, figs. 9-12; c, as PL 66, figs. 5-7; d, as PI. 67, fig. 7; e, as PI. 67, fig. 4;
f, as PI. 66, fig. 8; g, as PI. 67, fig. 3.) Scale represents 5 mm.

EXPLANATION OF PLATE 67

Figs. 1-8. Bactrynium bicarinatum Emmrich, from the Kossener Schichten of Austria. All figures x6;
figs. 1-4, 7-8 are whitened with ammonium chloride.

1, 2, Internal surface of dorsal valves, enlarged from PI. 66, figs. 1 and 5 respectively, to show lobed
groove and pustular area within groove. 3, Internal surface of dorsal valve of small (young?)
specimen from Kossen, to show smaller number but similar dimensions of lobes (PIUW ; figured
Zugmayer 1880, pi. 2, fig. 17). 4, Internal surface of fragment of dorsal valve from Kitzberg, to
show pustular area within groove (PIUW ; figured Zugmayer 1880, pi. 2, fig. 20). 5, Internal surface
of dorsal valve, enlarged from PI. 66, fig. 10, to show postero-median cavity and minor irregularities
in lateral lobes. 6, Dorsal valve ground down to show lateral indentations, median spike, and pos-
terior bridge (PIUW ; Zugmayer coll. ; specimen coated with oil before photographing). 7, Internal
surface of dorsal valve of small (young?) specimen from Kossen; compare with fig. 3 (BMNH
B.2103). 8, Internal surface of dorsal valve, enlarged from PI. 66, fig. 13.

ummQiDDg ‘>i3[/\\an>I

L9 31Vld

H -/OA ‘AS0/O}U0dV/DJ

RUDWICK: TRIASSIC BRACHIOPODS THECOSPIRA AND BACTRYNIUM 339

indentation are often coalesced into a single ridge or ‘lateral septum’; but here too the
best specimens show that the ridge is double in origin (PI. 67, fig. 4; PI. 68, fig. 1).
Around the distal end of each lobe the ridge may be clearly marked or may merge
insensibly into the peripheral part of the valve surface.

The groove likewise is variably expressed; it may be clearly differentiated from the
central area of the lobes or may merge insensibly into it. Where the inner edge of the
groove is distinct, it is shown by a dashed line in text-fig. 5. The central area of the lobes
is differentiated, at least in the best specimens, by a distinctly pustular surface (PI. 67,
figs. 2-5).

The posterior extremity of the groove/ridge structure is not entirely clear. The most
posterior pair of secondary lobes are generally short, and lie antero-lateral to the
cardinal process. At first sight it might appear that the groove runs into the base of the
cardinal process (PL 66, figs. 3, 8, 11; cf. Williams 1965, p. H521). But this is an illusion.
A specimen of Zugmayer’s, ground down to expose the ridges, shows clearly that the
ridges of left and right sides fused with one another in the median plane to form a
forwardly projecting V-shaped structure (PI. 67, fig. 6). Zugmayer himself referred to
this as a bridge, and this interpretation appears to me to be correct. In other words, at
points antero-lateral to the cardinal process the ridge rises away from the floor of the
valve and crosses the postero-median cavity as a bridge-like lamella (text-fig. 4a). In
the best-preserved specimens (PI. 66, fig. 3; PI. 68, fig. 1) the ridges appear to end
abruptly, but these are evidently breakage surfaces at the points where the ridges
became a free lamella. It is clearly distinct from the ‘socket plate’ running into the base
of the cardinal process (text-fig. 2a).

Brachial axis o/Thecospira. The entire brachidium of Thecospira, like that of any other
articulate, is clearly an outgrowth of the secondary-layer material of the dorsal valve.
As such, it must have been sheathed by, and secreted by, an extension of the outer
mantle epithelium (cf. Williams 1956). As with the spiral brachidia of spiriferides, geo-
metrical corsiderations alone are sufficient to prove that this epithelium must have had
the power of resorption as well as secretion, or else the brachidium could not have
grown in size while retaining its form and relative position.

In their reconstruction of the lophophore of spiriferides, Williams and Wright (1961)
make the assumption that the growing tips of the brachia necessarily remained together
in the mid-line, on the jugum, throughout ontogeny. But this assumption is not fully
supported even by their cited analogy of plectolophes in living brachiopods; in Gryphus
vitreus, for example, the growing tips are far apart from one another at the distal end
of the median coil. Even supposing that a complete jugum was present in Thecospira,
therefore, I see no justification for reconstructing a ‘deuterolophe’ (i.e. doubled brachial
axes) on the spiral lamellae. The simpler reconstruction of an ordinary spirolophe
seems to me more plausible, in view of the close resemblance between the form of the
spiral brachidium and that of the spiral lophophores of living rhynchonellides.

On this reconstruction each whorl of the lamellae would have borne a corresponding
single whorl of the brachial axis, and the growing tips of the brachia were at the tips
of the lamellae. As suggested previously (Rudwick 1960, text-fig. 6d), the brachial axis
is shown, conjecturally, nestling within the ‘gutter’; if this is correct the filament-row, at
least near its base, would have projected outwards around each whorl. If the lophophore

340

PALAEONTOLOGY, VOLUME 11

was supported throughout its length by the brachidium, the great brachial canal, which
in most living articulates acts as a hydrostatic skeleton, may have been vestigial or
absent; and it is so shown, conjecturally, in the reconstruction given here (text-fig. 3b).

Proximally, by homology with living brachiopods, the axis would have continued
towards the mid-line, lying adjacent to the crural processes or jugum. The mouth would
have been situated in the mid-line. At this point, if the basic orientation of the lopho-
phore was like that of all living brachiopods, the filament-row must have been on the
posterior or ventral side of the mouth and food-groove, and the lip on the anterior or
dorsal side. Then, by tracing the axis laterally on to the spirals, it follows by simple
topology that the filament-row must have been on the dorsal side of the food-groove,
and the lip on the ventral, on each of the outwardly facing whorls of the spirolophe. The
brachial axis is similarly twisted in the proximal parts of the spirolophes and plecto-
lophes of living brachiopods (Rudwick 1962, text-figs. 7c, 9c, 12b).

Brachial axis of Bactrynium. The mode of growth of the lobate apparatus can be in-
ferred quite simply (text-fig. 6). A section of Zugmayer’s, cut longitudinally through the
dorsal valve parallel to the median plane, shows clearly that the lobate apparatus grew
by simple accretion. The ‘lateral septa’ (i.e. the coalesced doubled ridges in the lateral
indentations) remained in the same absolute position during the growth of the valve, for
their growth tracks can be seen projecting through the thickness of the valve perpendicu-
lar to its outer surface (text-fig. 2b; PI. 68, fig. 9). Since the valve grew not only in thick-
ness but also in area, any given lobe must therefore have changed in relative position
during ontogeny, becoming progressively more posterior. The distal ends of the lobes
are near the valve edge in specimens of all sizes: therefore the lobes must have grown in

EXPLANATION OF PLATE 68

Bactrynium, Thecospira and other brachiopods for comparison. All figures X 6. All except fig. 9 whitened
with ammonium chloride.

Figs. 1, 2. Bactrynium bicarinatum Emmrich; ventral and posterior views of fragment of dorsal valve;
see also text-fig. 2a. Kossener Schichten of Diirnbach (PIUW; figured Zugmayer 1880, pi. 2, fig. 18).

Fig. 3. Eudesella mayalis (Deslongchamps), from the ‘Couche a Leptaena’ (Upper Lias), of May,
Calvados, France; interior of dorsal valve (SM.F.20273).

Figs. 4, 5. Davidsonella sinuata (Deslongchamps), from the same locality and horizon; postero-ventral
and ventral views of dorsal valve (SM.F.20199).

Fig. 6. Moorellina leptaenoides (Deslongchamps), from the same locality and horizon; interior of
dorsal valve (SM.F.20258).

Figs. 7, 8. Elliottina deslongchampsi (Davidson), from the Middle Lias of May, Calvados, France;
ventral and postero-ventral views of dorsal valve (SM.F.20288).

Fig. 9. Bactrynium bicarinatum Emmrich, longitudinal section of dorsal valve (and part of ventral),
showing growth tracks of lateral septa; see also text-fig. 2b. (PIUW; figured Zugmayer 1880, pi. 2,
fig. 32.) Locality unknown.

Fig. 10. Lyttonia sp., internal surface of part of dorsal valve, for comparison with Bactrynium. Sandy
limestone beds above Middle Productus Limestone (Permian), Banschang, East of Chideru, Salt
Range, Pakistan (Yale University, Peabody Museum, 22954).

Figs. 11, 12. Thecospira sp. Dorsal views of ventral valve alone and of both valves as found in loose
association, to show pseudodeltidium, articulation and radial costellae. Raibler Schichten, Rumerlo,
near Cortina d’Ampezzo, Italy (R. Zardini coll.).

Fig. 13. ‘ Thecidea sp.’ (Thecospira?), dorsal view of shell, to show convex pseudodeltidium, from
Helenenthal, Baden (PIUW; figured Zugmayer 1880, pi. 2, fig. 41).

Palaeontology, Vo I. 11

PLATE 68

12

13

RUDWICK, Bactrynium, Theeospira and comparative species

RUDWICK: TRIASSIC BRACHIOPODS THECOSPIRA AND BACTRYNIUM 341

length during ontogeny, extending laterally as the valve increased in breadth. At the
same time the valve was also growing in length, and therefore the most anterior secondary
lobe must have become progressively further from the anterior edge of the valve. But
the most anterior secondary lobes, and the primary lobes themselves, are near the
anterior edge in specimens of all sizes: therefore new secondary lobes must have been

text-fig. 6. Diagrams to show inferred mode of growth of ‘lobate apparatus’ of Bactrynium (a, b),
compared with Thecidiopsis (c, d). Figs. A, c show course of lobed groove and ridge; figs, b, d show
corresponding growth tracks of axes of lobes ; lobes are numbered consecutively in order of derivation
from the persistent primary lobes P. As dorsal valve grows in size from AA to BB, lobes grow in
length but do not change in absolute position; new lobes (no. 8 in fig. a, no. 5 in fig. c) are budded off

from persistent primary lobes.

budded off at intervals during ontogeny from the pair of persistent primary lobes. Thus
in text-fig. 6 a, during growth of the valve margin from A A to BB, secondary lobes 1 to 7
extended laterally while remaining in the same absolute position on the valve, while the
primary lobe extended forwards and budded off the new secondary lobe 8. The whole
process of growth of the lobate apparatus could have been achieved simply by differen-
tial rates of secretion of shell material over the surface of the valve as it increased in
thickness and over-all size: there is no need to postulate any shell resorption (except
possibly on the posterior edge of the ‘bridge’).

342

PALAEONTOLOGY, VOLUME 11

This analysis of growth justifies the terminology so far used in a purely descriptive
sense. The primary lobe may be so called because it must have persisted throughout onto-
geny in a submedian position, growing progressively forwards. The secondary lobes
must all have been derived from the primary lobe by lateral budding, in order from the
most posterior lobe forwards. The numbering of the secondary lobes thus corresponds
to their order of appearance (text-fig. 6b).

The over-all effect of the growth of the lobate apparatus was to keep the internal
surface of the valve fully covered (except in the peripheral zone) with a series of lobes
and indentations of a constant standard width, whatever the absolute size and area of
the valve.

There is a close analogy between the lobate apparatus of Bactrynium and the brachia-
bearing grooves and ridges in living Thecideacea. On the inner surface of the dorsal
valve there is a lobed groove, two-lobed in Thecidellina, four-lobed in Lacazel/a (Elliott
1965, fig. 742, 3a; 744, 2a). On the outer side the groove is separated by a conspicuous
slope from an outer ridge, though the latter may not be clearly distinguishable from the
peripheral zone of the valve. Postero-laterally the outer ridge extends freely above the
postero-median cavity, projecting towards the mid-line, or it may fuse with that pro-
jecting from the opposite side to form a complete bridge over the cavity. This bridge is
clearly separate from the ‘socket plates’ and cardinal process. In the median indentation
the ridges of left and right sides gradually coalesce and project backwards as a spike
overhanging the postero-median cavity. Indeed in Lacazella all the ‘septa’ in the
indentations may overhang in this way at their inner ends. Internal to the lobed groove
is a second, inner ridge, the so-called ‘brachial ridge’; within each lobe this is clearly
double in origin, but may be coalesced into a single ridge running along the axis of the
lobe. This inner ridge, like the outer, may project freely so as to overhang the postero-
median cavity. Its crest is often distinctly pustular.

Although the anatomy of living thecideaceans remains imperfectly known in many
respects, it is at least clear that the brachial axis of the lophophore is attached throughout
its length to the dorsal mantle, and that it lies in the lobed groove. This can be seen in
both Thecidellina (cf. Elliott 1965, fig. 742, 3a) and in Lacazella (Lacaze-Duthiers 1861,
pi. 3, fig. 1). The frontal side of the axis, bearing the lip, faces inwards, towards the
inner ‘brachial’ ridge; while the outer side, bearing the filament-row, faces outwards,
towards the outer ridge. The distal growing tips of the brachial axis are located close
together at the tip of the median indentation (Lacaze-Duthiers 1861, pi. 2, fig. 7). The
mouth is located medially, on the anterior side of the bridge over the postero-median
cavity. The arrangement of the lobes is such that the lophophore of Thecidellina is a
schizolophe, while that of Lacazella is a simple (four-lobed) ptycholophe. (Williams’s
and Rowell’s (1965, p. H38) novel definition of the trocholophous stage, to include all
lophophores in which the filaments are arranged in a single series, seems to me to be
unwarranted: lophophore growth stages were originally defined by the arrangement
of the brachial axis, and should remain so.)

The brachial axis of Bactrynium may be reconstructed in a comparable position,
attached to the dorsal mantle and lying in the lobed groove, with the filament-row
facing the outer bounding ridge and the lip facing the inner pustular area (text-fig. 7d).
Then proximally, at the mouth, the axis would have had the normal orientation, with
the filament-row on the posterior or ventral side of the mouth and the lip on the anterior

RUDWICK: TRIASSIC BRACHIOPODS THECOSP1RA AND BACTRYNIUM 343

or dorsal side. Thus the lophophore would have been a complex ptycholophe, with up to
twenty or more lobes (text-fig. 4b). It may be assumed that the lophophore would have
grown, like those of all living brachiopods, only by the lengthening of the brachial axis
and the formation of new filaments at the extreme tip of the brachia, in the median
indentation. Any given filament would therefore have shifted continually in its relative
position during the growth of the lophophore. The growth zones of the lophophore and
those of the supporting lobate apparatus were thus entirely distinct from one another.

text-fig. 7. Lophophore and inferred feeding mechanism of Lacazella (a-c) and Bactrynium (d).
a, dorsal valve of Lacazella with lophophoral filaments in contracted state, as illustrated by Lacaze-
Duthiers (1861, pi. 3, fig. 1); b, c, lophophore of Lacazella in inferred feeding orientation, based on
Lacaze-Duthier’s illustrations of opened shell (1861, pi. 1, figs. 2, 7) and on analogy with Megathiris
(Atkins 1960) ; d, reconstruction of brachial axis and filaments of Bactrynium, shown as section through
two lateral lobes; note brachial axis (dark stippled) in groove in surface of dorsal valve (light stippled),
with filaments arching over ‘lateral septa’ to form ‘tunnels’; arrows show water currents and food

currents.

FEEDING MECHANISMS

Some elements of the foregoing reconstructions of the lophophore are necessarily
conjectural. But of the assumptions made, only three are essential for the validity of the
functional interpretation given here : (n) that the basic anatomy of the lophophores of
Thecospira and Bactrynium resembled that of all living brachiopods; (b) that their basic
orientation relative to the body likewise resembled that of all living brachiopods; and
( c ) that in each genus the grooves in the skeletal apparatus bore single rows of the
brachial axis, so that the lophophore was spirolophous in Thecospira and ptycholophous
in Bactrynium. For the reconstruction of the feeding mechanisms, i.e. of the lophophores

344

PALAEONTOLOGY, VOLUME 11

in operation, one further assumption is necessary: ( d ) that the basic functional mechan-
isms of the lophophores were normal, i.e. that they resembled those of all living brachio-
pods (Rudwick 1962, and references therein). These assumptions of anatomical and
physiological uniformity are not of course capable of direct proof. But it is methodo-
logically a sound procedure to begin by testing them indirectly, by determining whether
they give a consistent and intelligible explanation of the morphology of the fossils. Only
if they fail to do so should departures from uniformity be postulated (cf. Rudwick 1964).

For effective ciliary filter-feeding in a brachiopod, it is essential that the tips of all the
filaments (except those projecting at the gape and delimiting the apertures) should
touch the mantle surface or the tips of other filaments. Only by so doing can the lopho-
phore divide the mantle cavity into separate inhalant and exhalant chambers and thereby
prevent wasteful re-circulation and re-filtering of the water (Rudwick 1960). The
possible ways in which a pair of linear brachia can divide the mantle cavity is limited by
inherent topological considerations, and only three essentially different arrangements
are known in living brachiopods (Rudwick 1962). The spirolophe and ptycholophe,
here posulated for Thecospira and Bactrynium respectively, are two of these three known
arrangements.

Current-system o/Thecospira. A topological argument developed for the reconstruction
of any spirolophe (Rudwick 1960) can be used to infer the attitude of the filaments in
Thecospira. Inhalant and exhalant chambers could not be isolated unless the filaments
on the spiral brachia touched the brachial axis of either the next proximal or the next
distal whorl. But on the latter assumption the filaments on the first or most proximal
whorl would have touched the axis of the second whorl, and no filaments would have
been available to bridge the gap between the axis of the first whorl and the dorsal mantle.
Therefore only the former alternative is feasible.

On the distal whorls, the tips of the filaments would thus have touched the brachial
axis of the next proximal whorl; on the medial part of the proximal whorl they would
have touched either the tips of the corresponding filaments on the other brachium, or
else the ventral mantle; anteriorly on this whorl, they would have passed across the
gape, projecting freely as apertural filaments; and laterally they would have touched the
dorsal mantle. Posteriorly, in the most proximal part of the lophophore, they would
have touched the lateral body wall and finally the anterior body wall or ventral mantle
(text-figs. 8, 9).

Only by this precise orientation of all the filaments could the spirolophe of Thecospira
have divided the mantle cavity into separate inhalant and exhalant chambers; by no
other arrangement is it topologically possible for this essential functional condition to
have been met. Of the only two possible varieties of the spirolophous system, the
orientation of the brachidial lamellae thus shows that the interior of the spirals must
have been an exhalant space, so that Thecospira must have operated the ‘ exhalant ’ or
‘ Spirifer' type, which has also been inferred for most spiriferides (Rudwick 1960) and
is known in the living Discinisca (Paine 1962).

The inhalant chamber in Thecospira would thus have comprised the ventral part of
the mantle cavity, external to the spirals; while the exhalant chamber would have occu-
pied the dorsal part, together with the interior of the spirals, and would also have
included a narrow exhalant space between the proximal filaments and the body wall.

RUDWICK: TRIASSIC BRACHIOPODS THECOSPIRA AND BACTRYNIUM 345

text-fig. 8. A. Reconstruction of T/iecospira, lateral view, cut along median plane. Inhalant water
(tailed arrows) is seen entering laterally and circulating around exterior of spiral brachium; exhalant
water (tailless arrows) is seen emerging from interior of spiral and leaving mantle cavity by a median
aperture. To clarify the current system, brachial axes are shown more slender, and filaments more
widely spaced, than in reality. Digestive glands omitted.

b. Similar reconstruction, with shell cut in plane parallel to median plane on right side, but with
right brachium preserved intact (water currents are those that would operate with whole shell intact);
showing crus embedded in lateral body wall, through which muscles and gut are faintly visible. For

further explanation see text.

Given this orientation of the filaments, the current system of Thecospirci follows
necessarily (text-figs. 8, 9). Water must have entered the mantle cavity laterally, being
drawn down into the inhalant chamber by the pumping action of the filaments removing
water into the exhalant chamber. After being filtered between the filaments on the
spirals, the water would have flowed as an exhalant current up the axis of each spiral,
and then forwards to leave the mantle cavity medially. Water filtered by the most
proximal filaments would have flowed along the narrow space between those filaments
and the body wall, emerging to join the main exhalant current on either side of the
dorsal part of the body wall (text-fig. 9, current ‘X’): a similar current can be observed
in living articulates (cf. Rudwick 1962, text-figs. 7c, 9a, c, 12b).

Thus the morphology of the brachidium of Thecospirci is consistent with a ciliary
feeding mechanism essentially similar to that of living brachiopods.

The rate at which a lophophore can pump water through a mantle cavity is roughly
proportional to the total area of the filament-row. On a spirolophe, the rate is therefore
approximately proportional to the area of the roughly conical surfaces which are formed
by the spiral brachia when all the filaments are in their natural orientation. Thus in

a a

C 5586

346

PALAEONTOLOGY, VOLUME 11

a spire-bearing brachiopod the larger the conical ‘framework’ formed by each spiralium,
the greater the filtering capacity of the brachium must have been. For a shell of given
size, the most effective arrangement of the spiralia would be that which most fully
occupies the available space of the mantle cavity, while leaving sufficient space for the
circulation of the water currents outside the spirals.

text-fig. 9. Reconstruction of Thecospira , ventral view of dorsal valve, and dorsal part of ‘body’,
with lophophore (right brachium cut off near base to clarify course of exhalant currents). Other

conventions as in text-fig. 8.

The observed arrangement of the brachidium in Thecospira approximates to this
paradigm. The spiralia are so closely moulded to the available space that when the shell
was closed the filaments must have been almost or quite in contact with the ventral
mantle. When the shell opened, however, the brachidium and lophophore, being
attached rigidly to the dorsal valve, would have been raised clear of the ventral mantle,
thus allowing the circulation of water outside the spiral brachia.

At the degree of opening selected for illustration here (text-figs. 8, 9), the circulation

RUDWICK: TRIASSIC BRACHIOPODS THECOSPIRA AND BACTRYNIUM 347

space outside the spiral brachia approximates to that observed outside the spirolophes
of the living rhynchonellide Notosaria and the living Crania (Rudwick 1962, text-fig. 7;
Atkins and Rudwick 1962, text-fig. 1). If the dorsal valve opened through a wider angle
than that shown in this reconstruction, the same current system would have operated,
except that the lateral inhalant apertures would have coalesced anteriorly, on the ventral
side of the exhalant aperture, to form a single inhalant aperture.

The functional advantage of developing a brachidial support for the spirolophe is not
entirely clear. A possible explanation is that a calcareous support for the brachial axis
takes up less space than the rather bulky hydrostatic and muscular skeletons seen in the
spirolophes of living brachiopods (e.g. Atkins 1963, fig. 8a). Consequently the successive
whorls of the spiral brachia can be set more closely together without substantially
impeding the flow of water; and thus a greater area of filaments can be accommodated
within a mantle cavity of given form and volume. But, on the other hand, this increased
efficiency of the lophophore is obtained at the expense of having a supporting structure
that is less flexible and more vulnerable to damage than is a hydrostatic skeleton. Thus,
a spiral brachidium may on balance be advantageous in some circumstances; but, of
course, it is only an available option in those brachiopods that have developed the power
of resorption in the epithelia of the mantle.

Current-system of Bactrynium. In a schizolophe or ptycholophe the only effective
arrangement of the filaments is for those in the indentations to be flexed abfrontally so
that their tips are contiguous. Thus each indentation forms a ‘tunnel’ of exhalant water,
blind-ended proximally but opening peripherally.

Such an orientation is in fact known in all living species in which schizolophes or
ptycholophes have been observed in operation. In these lophophores the filaments are
flexed abfrontally all along the brachial axis. In a schizolophe the tips of the filaments
touch each other across the median indentation, forming a median ‘tunnel’ which is
blind-ended posteriorly but open anteriorly; this has been observed in Argyrotheca and
Pumilus and in the schizolophous growth-stages of Notosaria and Waltonia (Atkins
1960, fig. 2; Rudwick 1962, fig. 6, 7a, 9a). The schizolophe of Thecidellina has not yet
been described in operation. In a ptycholophe there are similar ‘tunnels’ in the lateral
indentations as well; this has been seen in Megathiris (Atkins I960, fig. 6). Lacaze-
Duthiers (1861) only observed Lacazel/a with the filaments contracted frontally (text-
fig. 7a), but their probable natural orientation can be inferred (text-figs. 7b, c) by
analogy with Megathiris.

In all these schizolophes and ptycholophes the space near the dorsal valve surface is
divided into two parts; the space corresponding to the lobes is filled with inhalant
water, whereas the space corresponding to the indentations is filled with exhalant water.
Water is filtered through the filament-row around the periphery of the lophophore, but
also passes into the blind-ended tunnels in the indentations, from which it emerges peri-
pherally (cf. Atkins 1960, fig. 6; Rudwick 1962, fig. 6).

The wide gape between the valve edges of Megathiris (and probably LacazeUa also)
is divided in effect into two apertures. The exhalant aperture extends all round the gape
as a fairly narrow zone near the dorsal valve edge; the inhalant aperture occupies the
remainder of the gape, nearer the ventral valve but not extending as far posteriorly. The
apertures are separated only by the tips of the filaments around the periphery of the

348 PALAEONTOLOGY, VOLUME 11

lophophore. In principle there might be re-circulation around these apertural filaments;
but in practice the wide opening of the shell exposes the periphery of the lophophore to
sufficient external currents and turbulence to render any re-circulation insignificant.

It is reasonable to reconstruct the filaments of Bactrynium in a similar orientation,
forming ‘tunnels’ across each of the indentations (text-fig. 7d). Then its current system
would also have been similar (text-fig. 10). Water would have been drawn towards the
dorsal valve surface, and would then have been filtered either through the filament-row
around the periphery or else into the tunnels; within each tunnel an exhalant current
would have flowed laterally or anteriorly to emerge near the edge of the valve. Thus in
effect the part of the gape nearest the dorsal valve edge would have been an exhalant

A B

text-fig. 10. Reconstruction of Bactrynium in feeding position, showing inferred orientation of
filaments, and consequent paths of inhalant and exhalant currents, a, anterior view; b, lateral view.

aperture, and the rest of the gape a wide inhalant aperture. Unless Bactrynium lived in
exceptionally still water there would have been little risk of any substantial re-circulation
of the water.

As in Thecospira, therefore, the observed morphology of Bactrynium is consistent
with a ciliary feeding mechanism essentially similar to that of living brachiopods.

On this interpretation Bactrynium developed the ptycholophous alternative to a
greater degree of elaboration than in any living brachiopod. This may be related func-
tionally to its greater size. Observations on living brachiopods of various species and
at various growth stages suggest that the pumping capacity of filaments is roughly

RUDWICK: TRIASSIC BRACHIOPODS THECOSP1RA AND BACTRYNIUM 349

constant, as might be expected of a property dependent on a mechanism (i.e. the ciliary
action) operating on the cellular level (Rudwick 1962, p. 611). But the pressure dif-
ference (between inhalant and exhalant chambers or spaces) at which a lophophore
can operate is limited by the ability of the slender and unfused filaments to withstand
the pressure and remain in position. Therefore if a blind-ended tunnel formed from two
apposed stretches of filament-row were below some critical width, the rate at which
water was being pumped into the tunnel might exceed the rate at which it could escape
down the tunnel, and the water might then force the filaments out of position and flow
back into the inhalant space. On the other hand it is clearly advantageous to the effi-
ciency of the whole ciliary feeding mechanism that each tunnel should be as narrow as
possible without incurring the problem mentioned above: for only so can the maximum
total length of filament-row, and hence the maximum filtering capacity, be accommo-
dated within a shell of given size. There must, therefore, be some optimal width for the
tunnels, which would be the same for any normal ptycholophe. Similarly, it is clearly
advantageous that the inhalant ‘troughs’ between the exhalant tunnels should also be
parallel-sided and of some uniform width, so that as many as possible can be provided
in a given space.

Thus the entire form of a ptycholophe can be interpreted as a means of maintaining
the maximum effective filtering capacity within a shell of limited size. Any ptycholophe
is inherently limited by being attached throughout to the mantle, so that its elaboration
to provide greater filtering capacity can only be in two dimensions, making the best use
of the area of the dorsal valve surface (cf. Rudwick 1962, p. 611). The only means of
lengthening an attached filament-row within such a limited area is to buckle it into
lobes and indentations. But if there are also definite optimal widths for these lobes and
indentations, the lophophore must be elaborated during growth by the sequential
addition of new lobes and indentations to an initially simple schizolophe. Hence it is
to be expected that the degree of elaboration of the ptycholophe should be directly
related to the absolute size of the dorsal valve, both during the ontogeny of a single
species and also among species of different adult sizes. It should thus show ‘size-
required allometry’ (Gould 1966). This seems to be true of Bactrynium as far as the
limited number of specimens allows such a conclusion (text-fig. 5); and a comparison
with some thecideaceans (text-fig. 11) shows that they too have lobes and indentations
of about the same size.

Unfortunately, quantification of this allometric relation is hardly possible, in view
of the difficulties of estimating the size of the true ‘body’ of a brachiopod, as opposed
to the size of its mantle and shell.

AFFINITIES AND EVOLUTION

Affinities o/ Thecospira. Williams (1953n) pointed out that Thecospira had such strik-
ingly strophomenide characters that it could not be excluded from that group solely
because of its possession of a spiral brachidium, the only character which suggested
affinity with the spiriferides. Thecospira was therefore included within the Orthotetacea
(now Davidsoniacea).

Like the davidsoniaceans of the Palaeozoic, Thecospira was cemented by the ventral
valve during the earlier part of its life history; it has a strophic hinge-line, well-developed

350

PALAEONTOLOGY, VOLUME 11

teeth and sockets, and a large rectangular cardinal process. Most Palaeozoic davidsonia-
ceans (except the earliest) are pseudopunctate, but Schuchertella is said to be impunctate
(Williams 1965, p. 408), and an apparently punctate structure has been reported in
Streptorhynchus (Thomas 1958). Against this background of diverse shell structure, the
existence of punctate, impunctate, and obscurely pseudopunctate species of Thecospira
is consistent with an argument for affinity.

text-fig. 11. Drawings of the grooves and ridges on the dorsal valve of various species of Thecidiopsis,
to show the uniformity of dimensions of the lobes. All x5; compare with Bactrynium, text-fig. 5.
(Sources: a, Nekvasilova 1964, pi. XI, fig. 4, 6, 8; pi. XII, fig. 5; b-f, Backhaus 1959, b, pi. 5, fig. 4;
c, pi. 4, fig. 2; d, pi. 7, fig. 1,2; e, pi. 4, fig. 8; f, pi. 5, fig. 7.) Lobes of e and F marked with inferred
order of development ; compare text-fig. 6c. Scale represents 5 mm.

Thecospira is exceptional among davidsoniaceans in having an ‘entire’ (i.e. plane)
pseudodeltidium (Williams 1965, p. H366), but this is a development that certainly
occurred several times among the strophomenides, and apparently related shells have a
convex pseudodeltidium. Likewise it has lost all but a faint trace of the costellate shell
surface characteristic of davidsoniaceans, but this too occurred in other strophomenide
groups (Williams 1965, p. H364). When the costellae are most clearly visible they show
a pattern of increase by intercalation, as in other davidsoniaceans, and not by lateral
branching. The pustular or spinule-bearing shell surface which apparently replaced the

RUDWICK: TRIASSIC BRACHIOPODS THECOSPIRA AND BACTRYNIUM 351

costellae is not found, so far as I am aware, in any other davidsoniacean, and is more
reminiscent of productaceans such as Waagenoconcha. But Thecospira lacks the larger
tubular spines that seem to have been universal among the productaceans and stropha-
losiaceans.

The most important character distinguishing Thecospira from other davidsoniaceans
is, of course, the spiral brachidium. But the unique structure of the lamellae suggests
that the brachidium developed independently from those of the Spiriferida.

Spiral brachidia have recently been reported in Cadomella (Cowen and Rudwick
1966), which is generally regarded as related to the chonetaceans, so that there is now
evidence that spiral brachidia evolved at least twice outside the Spiriferida. Even among
the Spiriferida themselves spiral brachidia may have evolved more than once, the
atrypaceans being in many respects closer to the rhynchonellides than to most other
spiriferides (Copper 1965).

Palaeozoic davidsoniaceans probably possessed spirolophes supported by a hydro-
static or other ‘soft’ skeleton, for spiral impressions are known in Davidsonia itself.
Davidsoniaceans existed in considerable abundance and variety even in Upper Permian
time. The teeth of Thecospira , being unsupported by dental lamellae, suggest the
Schuchertellidae or Orthotetidae as possible ancestors. The cardinal process and ‘socket
plates’ have perhaps the greatest resemblance to those of Streptorhynchus ; Orthotetes is
another possible ancestor.

From some such Permian ancestor, the evolution of Thecospira would have involved
(a) acquisition of brachidium ( b ) loss of convexity in pseudodeltidium and chilidium

(c) obsolescence of costellae and (in some species) replacement by pustules, and possibly

( d ) loss of pseudopunctate structure and (in some species) acquisition of punctae. There
is, however, no reason to suppose that these changes took place simultaneously. As
already mentioned, the diversity of shell structure in Thecospira was foreshadowed in the
Permian davidsoniaceans. Even the spiral brachidium may have evolved earlier than is
yet apparent, and it might be worth examining Permian davidsoniaceans for any trace
of the bases of crura. In any case, the evolution of Thecospira would not have involved
any major changes in the relation of the shell to the substratum, in the hinge and
musculature, or in the ciliary feeding mechanism. Only the supporting structures of the
brachial axis would have changed, from a probably hydrostatic to a purely calcareous
skeleton. As already suggested, there may have been an adaptive advantage in this
change, in allowing a greater filtering capacity within a shell of given size.

The Koninckinidae, which have generally been regarded as abnormal Spiriferida,
should probably now be placed with Cadomella as post-Palaeozoic descendants of the
chonetaceans (Cowen and Rudwick 1966). Like Thecospira , the koninckinids are first
known from Middle Triassic strata, but they are significantly different in the structure
of the brachidium and many other characters. If this interpretation of koninckinids is
correct, it implies that two relict strophomenide groups evolved spiral brachidia inde-
pendently at about the same time, after most of the related ‘normal’ genera had become
extinct.

Affinities of Bactrynium. As already mentioned, the affinities of Bactrynium have re-
mained more controversial than those of Thecospira. Affinity to Lyttoniacea and to
Thecideacea are interpretations that have been almost equally supported for many

352

PALAEONTOLOGY, VOLUME 11

text-fig. 12. Diagram to show inferred phyletic and functional relationships of Bactrynium and T/ieco-
spira to other brachiopods. Outlines of dorsal valves of respresentative species are shown at uniform

PERMIAN TRIASSIC JURASSIC CRETACEOUS TERTIARY

RUDWICK: TRIASSIC BRACHIOPODS THECOSPIRA AND BACTRYNIUM 353

decades. The form of the lobate apparatus of Bactrynium has an immediate resemblance
to that of the lyttoniaceans. The strongly concave form of the dorsal valve gives it an
especially close resemblance to Oldhamina. It is necessary, however, to examine these
similarities critically, in order to assess their value as clues to affinity.

In a lyttoniacean the highly modified dorsal valve (‘internal plate’ of Williams 19536,
1965) has a pair of sub-parallel submedian primary lobes, from which secondary lobes
branch laterally. Between the lobes are lateral indentations and a single median indenta-
tion. The degree of symmetry is variable, but in Oldhamina and Leptodus, for example,
it is as great as in Bactrynium. In such genera as these, there is also an obvious approach
to uniformity in the width of the lobes and of the indentations. The mode of growth of
the lobes can be inferred, by analogous reasoning, to have been similar to that of
Bactrynium, with the secondary lobes extending laterally, while each primary lobe
extended forwards and periodically ‘budded off’ a new secondary lobe. This is con-
firmed by the evidence of the growth-lines on the lobes. The similarities even extend to
the detailed morphology, for the inner side of the dorsal valve of lyttoniaceans has a
contiguous ridge and groove (the ridge being external to the groove) running around the
edge of the lobes and indentations, while the area enclosed by the groove is often pustu-
lar in appearance (PI. 68, fig. 10; see also Stehli 1956, pi. 41, fig. 3; Williams 1965,
fig. 393, 3d).

But the differences between the lobate apparatus of Bactrynium and the dorsal valve
of a lyttoniacean are as striking as the resemblances.

The strophic hinge and articulation of Bactrynium are apparently quite normal, and
are certainly unlike the highly aberrant hinge structures of the lyttoniaceans. The dorsal
valve of Bactrynium is a thick massive plate, whereas that of a lyttoniacean is excep-
tionally thin and delicate. More significantly, the dorsal valve of Bactrynium has a
normal ‘entire’ outline, whereas that of a lyttoniacean is highly indented and corre-
sponds to the course of the lobed ridge. It is true that its indented outline is often
modified by ‘bridges’ of shell material across the indentations, either occasional as
in Oldhamina (cf. Williams 1965, fig. 293, 2b), confined to the median indentation as in
Gubleria (Termier and Termier 1960), or regularly across all the indentations as in
Coscinophora (cf. Williams 1965, fig. 393, la, lc). The inner ends of the posterior

magnification (x2), with course of brachial axes (reconstructed from brachial grooves or brachidial
lamellae) marked in thicker black lines (hatched where reconstruction is tentative). Note fairly uni-
form dimensions of lobes of Thecideacea (1-17) and spiralia of Thecospira (18) in contrast to much
larger lobes of Lyttoniacea (20). Variations in adult body size are indicated approximately by positions
on horizontal scale, smallest species being roughly central. (If space allowed, the lyttoniacean (20)
would be much further to right.) Stratigraphical horizons of species indicated approximately against
Geological Society time-scale (.Quart. J. geol. Soc. Loud. 120S, 260-2). Broad outline of suggested
phylogeny shown by functional ‘zones’ based on type of lophophore in adult stage: specific phyletic
pathways are not given. For further explanation see text. Key to genera: thecideacea: 1. Lacazella ;
2. Glazewskia; 3-5. Thecidellina\ 6. Parathecidea; 1-9. Thecidiopsis', 10. Bifolium ; 11. Glazewskia',
12. Eudesel/a; 13. Elliottina; 14. Davidsonella : 15. Moorellina; 16. Cooperina ; 17. Bactrynium ; david-
soniacea: 18. Thecospira ; 19. A Permian davidsoniacean (based on Diplanus), with (a) juvenile stage
with inferred schizolophe, and (b) adult with inferred spirolophe. lyttoniacea: 20. Leptodus (Frag-
ment only). (Sources: 2, Pajaud 1965; 4, Thomson 1915; 5, Toulmin 1940; 6-9, Backhaus 1959;
10, Nekvasilova 1964; 11, Glazewski and Pajaud 1965; 16, Termier, Termier and Pajaud 1966; 1, 3,

12-15, 17-20 original.)

354

PALAEONTOLOGY, VOLUME 11

indentations may also be filled up progressively, as in Leptodus (cf. Williams 1965, fig.
393, 3d). But in no lyttoniacean were the indentations completely filled in: even in
Coscinophora gaps of uniform length were maintained between the regularly spaced
‘bridges’ across the indentations. As Sarycheva (1964) has pointed out, Stehli’s (1956)
suggested derivation of Bactrynium from Coscinophora is therefore improbable.

The indented edge of the dorsal valve of lyttoniaceans rests on a series of septa rising
from the floor of the ventral valve. The interior of the ventral valve of Bactrynium is not
well known, but it certainly has no such septa rising to make contact with the lobate
apparatus on the dorsal valve (text-fig. 2b; PI. 68, fig. 9).

Another difference of great importance is that of absolute size. The lobes and in-
dentations on a lyttoniacean are nearly three times as large (in linear dimensions) as
those of Bactrynium (text-fig. 12; PI. 68, compare fig. 10 with fig. 1). This difference is
independent of the absolute size of the whole shell, since in both groups the width of the
lobes and the width of the indentations were clearly kept more or less constant during
growth. As Gould (1966, p. 604) suggests, such a difference should draw attention to
a probable adaptive discontinuity between the groups concerned.

The morphology of lyttoniaceans is so abnormal that any reconstruction of their
feeding mechanism on grounds of analogy alone would be hazardous. Only a full
functional analysis of their morphology will give rational grounds for any such recon-
struction. In the meantime, it should not be assumed that their feeding mechanism was
necessarily the same as that of Bactrynium. In particular the very thin and delicate
dorsal valve, recessed within a more robust ventral valve, and the persistently main-
tained slots or perforations between the lobes of the dorsal valve, suggest a significantly
different functional organization; and the much larger dimensions of the lobes and
indentations point to the same conclusion.

On the other hand it is probable that the groove running round the lobes and indenta-
tions bore a brachial axis as in Bactrynium , and that the lophophore therefore was a
ptycholophe. If this is correct, the same topological considerations could have been
responsible for the very similar arrangement and mode of growth of the lobes. The
over-all similarity could, therefore, be due to simple functional parallelism.

There is now no good evidence for the survival of true lyttoniaceans after the end of
the Permian period. The large stratigraphical gap between the last undoubted lyttonia-
ceans in the Dzjulfian (late Permian), and Bactrynium in the Rhaetian (late Triassic), is
therefore no longer a problem.

A much closer comparison can be made between Bactrynium and the thecideaceans.
In both, the dorsal valve is a massive plate with a thickened sub-marginal rim and
‘entire’ outline. The ventral valve is cemented to the substratum, at least in earlier
growth-stages. The rectangular cardinal process of Bactrynium , with flanking sockets
and socket plates, is closely similar to the cardinalia of thecideaceans. The shell surface
of thecideaceans has no radial ornament. The hinge is strophic, with a low dorsal inter-
area and higher ventral interarea.

Many other characters of Bactrynium can be matched among the earlier thecideaceans
of the Lower Jurassic. Thus Davidsonel/a has already been reported as pseudopunctate
(Elliott 1965), and there is a similar obscurely pseudopunctate structure in Moorel/ina
leptaenoides (Deslongchamps).

The cardinal process and sockets are similar to those of Bactrynium, and lateral to

RUDWICK: TRIASSIC BRACHIOPODS THECOSPIRA AND BACTRYNIUM 355

the cardinal process is a pair of elliptical areas similar to the postulated lateral adductor
scars of Bactrynium. The ventral valve is generally strongly convex, and the dorsal
valve may be concave as in Bactrynium (e.g. Davidsonella sinuata (Deslongchamps)).
There is a similar lobed ridge and groove in Eudesella (PL 68, fig. 3). The area within the
groove is generally pustular (PI. 68, figs. 3, 7, 8). In Davidsonella pustules are highly
developed (PI. 68, figs. 4, 5), but do not appear to be separate spicules as implied in the
Treatise description (Elliott 1965, p. H859). Posteriorly the ridge projects across a
postero-median cavity, often as a complete ‘bridge’, which is quite distinct from the
‘socket plates’ (PI. 68, fig. 4). The median indentation of the ridge may project pos-
teriorly as a distinctly doubled spur or spike closely resembling that of Bactrynium
(PI. 68, figs. 4-6).

In thecideaceans there is, as already mentioned, a clear approximation to a uniform
width for the lobes and for the indentations, both in the different growth stages of a
species and between different species; and these uniform dimensions are closely com-
parable to those of Bactrynium (compare text-fig. 1 1 with text-fig. 5; also PI. 68, compare
figs. 3-5, 7, 8 with fig. 1). This suggests a common feeding mechanism for Bactrynium
and thecideaceans, with ‘tunnels’ of exhalent water stabilized at the optimum dimen-
sions inherent in any normal (i.e. non-lyttoniacean) ptycholophous system.

The chief difference between the lobes of Bactrynium and those of the thecideaceans
lies in their arrangement. In genera with several lobes the lobes project anteriorly or
antero-medially, and never laterally (text-fig. 11). It can, however, be inferred that they
grew in the same manner as that postulated for Bactrynium. The difference is that the
primary lobe of thecideaceans is that furthest from the mid-line, not that nearest the
mid-line. The primary lobe extends parallel to the postero-lateral sector of the valve edge ;
the secondary lobes project anteriorly or antero-medially from it, and must have been
‘budded’ from it serially, the postero-median lobes being the oldest (text-figs. 6c-d).
Expressed another way, the secondary lobes were budded from the medial side of the
primary lobe in thecideaceans, but from the lateral side in Bactrynium. Beginning from
a schizolophous stage, these are, in fact, the only two distinct ways in which the number
of lobes can be increased (without involving an unlimited amount of shift in the absolute
positions of all the lobes). As already shown, it appears that Bactrynium actualized one
of these alternative possibilities, with secondary lobes being formed serially on the
lateral side of each primary lobe; whereas the thecideaceans actualized the other
alternative, with secondary lobes being formed serially on the medial side of each
primary lobe (in some thecideaceans the budding pattern may have been irregular — see
text-figs. 11a-c).

These alternatives, though topologically equivalent, are probably not equal in func-
tional efficiency. On the first alternative ( Bactrynium ) all the secondary lobes project
laterally, and hence all the ‘tunnels’ (except the median one) opened laterally. On the
second alternative (Thecideacea) all the secondary lobes project antero-medially,
anteriorly, or at most antero-laterally, so that all the ‘tunnels’ open around the anterior
part of the gape. In terms of the current systems, the greatest outflow of exhalent water
would be, on the first alternative, in the lateral part of the gape, and on the second, in
the anterior part. But the area available to act as exhalant aperture is limited laterally
by the degree of separation of the valve edges, whereas anteriorly the much wider total
gape can be divided between inhalant and exhalant apertures simply by an appropriate

356

PALAEONTOLOGY, VOLUME 11

orientation of the apertural filaments. In other words it would seem simpler to ensure
an unimpeded outflow of exhalant water on the second alternative than on the first.
This suggests a possible adaptive advantage in favour of the thecideaceans, which may
be reflected in their longer range and greater diversity and abundance.

It is interesting that within the lyttoniaceans there are genera (e.g. Paralyttonia,
Rigbyella ) with forwardly projecting lobes like those of the thecideaceans, as well as the
better-known genera (e.g. Oldhamina, Leptodas ) with laterally projecting lobes like those
of Bactrynium. This fact was used by Wanner (1935) as an argument for affinity between
lyttoniaceans and thecideaceans (including Bactrynium). He maintained that any other
interpretation would involve postulating a double convergence. But the present analysis
of the growth of ptycholophes implies on the contrary that the parallel development of
both alternative arrangements of the lobes is by no means improbable.

These similarities together make a strong case for affinity between Bactrynium and
Thecideacea. Indeed, in my opinion there is now no difference warranting supra-
familial recognition. I therefore suggest that Bactrynium, while retaining its family
Bactryniidae in recognition of the distinctive form of its lobate apparatus, should be
assigned to the Thecideacea.

In phylogenetic terms Bactrynium can best be regarded as a derivative of small and
simple schizolophous thecideacean ancestors. Such species are known in the Permian
( Cooperina ) and Rhaetian (e.g. Moorel/ina), and are probably represented in inter-
mediate strata among the poorly known Triassic ''Thecidea' spp. From some such
ancestor, Bactrynium could have evolved allometrically by elaboration of the schizo-
lophe into a ptycholophe and by concurrent increase in absolute size. Later ptycholo-
phous forms, such as Eudesella in the Lower Jurassic, utilized the other alternative
arrangement of lobes, and therefore probably evolved independently from schizolophous
forms. Bactrynium might, however, have left some schizolophous descendants by neo-
teny; the pseudopunctate Davidsonella, for example, might have had such an origin.

Without attempting to reconstruct the detailed course of phylogeny in the Thecideacea,
the functional analysis given here leads to an interpretation in terms of ‘functional zones’
(text-fig. 12). Each ‘zone’ represents the utilization of one possible mode of organiza-
tion of the lophophore, without any necessary effect on the mode of life of the whole
organism; functional zones are thus not synonymous with adaptive zones. Using this
concept, it would seem that from the earliest thecideacean ( Cooperina ) onwards, there
has always been a ‘zone of schizolophes’, occupied by species small enough for a
schizolophe to be an adequate form for the lophophore. At certain times there have also
been species which, by utilizing one or other ptycholophous arrangement, were able to
increase in absolute size and evolve into ‘zones of ptycholophes’. Gould (1966) has
already cited some thecideaceans as an example of such ‘size-required allometry’. Such
species would, of course, have traversed the ‘zone of schizolophes’ during ontogeny, as,
for example, Lacazella is known to do at the present day. I have made a somewhat
arbitrary distinction between ‘zones of simple ptycholophes’, containing species with
no more than three pairs of lobes, and ‘zones of complex ptycholophes’, containing
species with larger numbers of lobes. The ‘envelopes’ outlining the zones on text-fig. 12
are merely to show the limits of known species in each zone: they do not imply that any
zone was in any sense unavailable at other times, nor that all the specimens in a given
zone are closely related to each other (e.g. Eudesella in the Lower Jurassic may have

RUDWICK: TRIASSIC BRACHIOPODS THECOSPIRA AND BACTRYNIUM 357

evolved from schizolophous ancestors independently from the much later, Cretaceous,
ptycholophous species).

Affinities ofThecideacea. Further discussion of the affinities of Bactrynium thus involves
the question of thecideacean affinities. Lacazeda and its fossil relatives formed an im-
portant element in the earlier concept of the ‘Protremata’, and underlying this was the
belief that the Thecideacea were related to the brachiopods now grouped together as
Strophomenida. More recently, however, this early view has been regarded as doubtful,
and an affinity with the Terebratulida or Spiriferida has been favoured (Williams and
Rowell 1965).

In this interpretation undue weight has perhaps been given to the fact that most
Thecideacea are punctate. But it is now recognized that the punctate structure must have
evolved independently several times during the history of the Brachiopoda (Williams
and Rowell 1965, p. H68; Wright 1966). Therefore the punctation ofThecideacea is
not by itself a reliable criterion of affinity to the Terebratulida or punctate Spiriferida.
The existence of diverse shell structures among the earlier Thecideacea increases the
likelihood that punctation has evolved independently in this group.

Two other important characters of Thecideacea, their cementation and their lobed
brachial grooves, find no parallel among Spiriferida or Terebratulida. There is now no
authenticated case of cementation attachment in any articulate brachiopod outside the
Strophomenida; for Dagis (1965) has justly thrown doubt on Thecocyrtella, and
Bittnerula is now known to have a pedicle foramen (Cowen and Rudwick 1967). The
cemented attachment of Thecideacea is, therefore, strong evidence against any but a
strophomenide affinity. Linked with the possession of cementation is the fact that
Thecideacea have a pseudodeltidium and no pedicle foramen; and their cardinalia
cannot be matched closely among terebratulides or spiriferides. The brachial supporting
structures of Thecideacea, with the brachial axes in grooves excavated in the dorsal
valve surface, are entirely different from any of the varied spiral and looped brachidia
of Spiriferida and Terebratulida, all of which are composed of calcareous lamellae
growing independently from the dorsal valve floor, though attached to the cardinalia
and also (in some forms) to septa.

A far stronger case can be made for affinity between the Thecideacea and the Stropho-
menida. Of strophomenide brachiopods, the Davidsoniacea show the greatest resem-
blances to the Thecideacea. Both groups have cementation attachment and therefore no
pedicle foramen, and both lack the tubular spines of chonetoids and productoids.
Both have strophic hinges, with strong articulation flanking a fairly massive cardinal
process which is generally covered by a convex pseudodeltidium.

This conclusion would be given taxonomic recognition by the assignment of the Theci-
deacea to the Strophomenida. Their possession of brachial grooves and related struc-
tures, implying schizolophous or ptycholophous but never spirolophous lophophores,
their lack of costellate ornament, and their generally (but not universally) punctate shell
structure are collectively sufficient grounds for retaining them as a superfamily distinct
from the Davidsoniacea.

Until recently the Thecideacea were unknown before the Rhaetian (Elliott 1965);
but with the discovery of the simple but true thecideacean Cooperina in the Word
Formation (Upper Permian) of Texas (Termier, Termier, and Pajaud 1966), it is now

358

PALAEONTOLOGY, VOLUME 11

clear that the ancestors of the group must be sought among Lower Permian or even
earlier davidsoniaceans. These, like the Devonian Davidsonia itself, probably had
spirolophous lophophores supported only by a hydrostatic skeleton or by muscular
and connective tissue. By analogy with living brachiopods, most Permian davidsonia-
ceans were almost certainly too large for a schizolophe to have been adequate for their
needs; but like all living spirolophous species they would have passed through a schizo-
lophous stage early in ontogeny while they were still small in size. In the light of these
functional considerations a neotenous origin for the Thecideacea seems most probable.
Only the development of a sessile schizolophe, accommodated in a groove in the dorsal
valve, would have been required to convert a small and young davidsoniacean into a true
thecideacean (text-fig. 12).

CONCLUSIONS

If the foregoing arguments are correct, the great Palaeozoic order of Strophomenida,
which suffered a drastic degree of extinction at about the end of the Permian period,
survived into the Mesozoic only as three groups of small, rare, and inconspicuous
brachiopods. One group was the Koninckinidae, modified descendants of the Chone-
tacea (Cowen and Rudwick 1966). The second group consisted of small and simple
davidsoniaceans together with Tliecospiro. The spiral brachidium of the latter genus may
have evolved at about the same time (early or middle Triassic) as the parallel develop-
ment in the Koninckinidae, both being independent of the spiral brachidium of the
Spiriferida. The brachidium of Thecospira may have increased the efficiency of the
pumping action of its lophophore, but did not entail any essential change in its ciliary
feeding mechanism, or in the exhalant spirolophous current system by which that
mechanism operated. Apart from its spiral brachidium Thecospira was functionally
similar to more ‘normal’ davidsoniaceans. The third group, the Thecideacea, was
already in existence even before the end of the Permian, and was probably derived by
neoteny from the davidsoniaceans. In late Triassic time Bactrynium evolved from some
small and simple thecideacean by elaboration of a schizolophe into a ptycholophe and
by corresponding increase in size. This ‘size-required allometry’ involved no essential
change in its ciliary feeding mechanism, but its concavo-convex shell form adapted it
to a free-lying mode of life. Bactrynium itself became extinct at the end of the Triassic,
though some neotenous descendants may be represented among the varied Thecideacea
of the Lower Jurassic. These also included a new and independent development of
ptycholophous forms ( Eudesella ). The Thecideacea survived thereafter, though incon-
spicuously, up to the present day. On this interpretation the Strophomenida are not
yet extinct.

REFERENCES

atkins, d. 1960. The ciliary feeding mechanism of the Megathyridae (Brachiopoda), and the growth
stages of the lophophore. /. mar. biol. Ass. U.K. 39, 459-79.

1963. Notes on the lophophore and gut of the brachiopod Notosaria (formerly Tegulorhynchid)

nigricans (G. B. Sowerby). Proc. zool. Soc. Lond. 140, 15-24.

and rudwick, m. j. s. 1962. The lophophore and ciliary feeding mechanism of the brachiopod

Crania anomala (Muller). J. mar. biol. Ass. U.K. 42, 469-80.
backhaus, e. 1959. Monographic der cretacischen Thecideidae (Brach.), Mitt. geol. St Inst. Hamb.
28, 5-90.

RUDWICK: TRIASSIC BRACHIOPODS THECOSP1RA AND BACTRYNIUM 359

bittner, A. 1890. Brachiopoden der alpinen Trias. Abh. geol. Bundesanst., Wien, 14.

broili, f. 1916. Die permischen Brachiopoden von Timor. In wanner, j. (ed.), Palciontologie von Timor,

7, 12.

copper, p. 1965. Unusual structures in Devonian Atrypidae from England. Palaeontology, 8, 358-73.
cowen, R., and rudwick, m. j. s. 1966. A spiral brachidium in the Jurassic chonetoid brachiopod
Cadomella. Geol. Mag. 103, 403-6.

1967. Bittnerula Hall and Clarke, and the evolution of cementation in the Brachiopoda.

Geol. Mag. 104, 155-9.

dagis, a. s. 1965. Triassic brachiopods of Siberia. Nauka, Moscow, 1-186. (In Russian.)
elliott, g. f. 1965. Suborder Thecideidina. In moore, r. c. (ed.). Treatise on Invertebrate Paleontology,
part (H) Brachiopoda, H857-62.

emmrich, h. 1855. Notiz liber den Alpenkalk der Lienzer Gegend. Jb. geol. Bundesanst., Wien, 6,
444-50.

glazewskj, k., and pajaud, d. 1965. Sur une nouvelle espece de Thecideidae, Glazewskia sp. (Brachio-
pode) du Jurassique de Podolie. Bull. Soc. geol. Fr. (7) 6, 262-8.
gould, s. j. 1966. Allometry and size in ontogeny and phylogeny. Biol. Rev. 41, 587-640.
grant, r. e. 1966. Spine arrangement and life habits of the productoid brachiopod Waagenoconcha.
J. Paleont. 40, 1063-9.

kozlowski, R. 1929. Les brachiopodes gothlandiens de la Podolie polonaise. Palaeont. pol. 1, 1-254.
LACAZE-DUTfflERS, h. de. 1861. Histoire naturelle des Brachiopodes vivants de la Mediterranee.
Annls Sci. nat. (4), 15, 259-330.

makridin, v. p. 1960. Superfamily Thecideacea. In sarycheva, t. g. (ed.). Os navy Paleontologii,
Mshanki, Brakhiopody, 1-305. (In Russian.)

nekvasilova, o. 1964. Thecideidae (Brachiopoda) der Bohmischen Kreide. Sb. geol. Ved. P3, 1 19-62.
noetling, f. 1905. Untersuchungen liber die Familie der Lyttoniidae Waagen emend. Noetling.
Palaeontographica, 51, 129-53.

paine, R. t. 1962. Filter-feeding pattern and local distribution of the brachiopod Discinisca strigata.

Biol. Bull. mar. biol. Lab., Woods Hole, 123, 597-604.
pajaud, d. 1965. Remarques sur les Thecideidae (Brachiopodes) tertiaires. Bull. Soc. geol. Fr. (7) 6,
258-61.

rudwick, m. j. s. 1959. The growth and form of brachiopod shells. Geol. Mag. 96, 1-24.

1960. The feeding mechanism of spire-bearing fossil brachiopods. Geol. Mag. 97, 369-83.

1962. Filter-feeding mechanisms in some brachiopods from New Zealand. J. Linn. Soc. Lond.

( Zool .) 44, 592-615.

■ 1964. The inference of function from structure in fossils. Brit. J. Phil. Sci. 15, 27-40.

1965. Ecology and paleoecology. In moore, r. c. (ed.). Treatise on Invertebrate Paleontology,

part (H) Brachiopoda, H199-H214.

sarycheva, t. g. 1964. Oldhaminoid brachiopods from the Permian of Transcaucasia. Paleont. Zh.
1964, 3, 58-72. (In Russian.)

stehli, f. g. 1956. Notes on oldhaminid brachiopods. J. Paleont. 30, 305-13.

suess, e. 1854. Uber die Brachiopoden der Kossener Schichten. Denkschr. Akad. Wiss., Wien, 7, 29-65.
termier, h., and termier, g. 1960. Fes Oldhaminides du Cambodge. Bull. Soc. geol. Fr. (7) 1, 233-44.

and pajaud, d. 1966. Decouverte d’une Thecidee dans le Permien du Texas. C.R. Acad. Sci.

Paris, 263, D332-5.

Thomas, g. a. 1958. The Permian Orthotetacea of Western Australia. Bull. Bur. Miner. Resour. Geol.
Geophys. Aust. 39.

Thomson, j. a. 1915. On a new genus and species of the Thecidiinae (Brachiopoda). Geol. Mag. (6),
461-4.

toulmin, l. d. 1940. Eocene brachiopods from the Salt Mountain limestone of Alabama. J. Paleont.
14, 227-33.

waagen, w. 1879-87. Productus Fimestone Fossils. Palaeont. indica (13), 1.

wanner, j. 1935. Fyttoniidae and ihre biologische und stammesgeschichtliche Bedeutung. Neues Jb.
Miner. Geol. Paldont. BeilBd. 74B, 201-81.

williams, a. 1953u. The classification of the strophomenoid brachiopods. J. Wash. Acad. Sci. 43, 1-13.
19536. The morphology and classification of the oldhaminid brachiopods. Ibid. 43, 279-87.

360 PALAEONTOLOGY, VOLUME 11

williams, a. 1956. The calcareous shell of the Brachiopoda, and its importance to their classification.
Biol. Rev. 31, 243-87.

1965. Strophomenidina, Oldhaminidina. In moore, r. c. (ed.), Treatise on Invertebrate Paleonto-
logy, part (H) Brachiopoda, H362-H412, H510-21.

and wright, a. d. 1961. The origin of the loop in articulate brachiopods. Palaeontology , 4, 149—

176.

and rowell, a. j. 1965. Anatomy, morphology, evolution and phylogeny. In moore, r. c. (ed.),

Treatise on Invertebrate Paleontology, part (H) Brachiopoda, H6-H57, H57-H138, HI 64-99.
wright, a. d. 1966. The shell punctation of Dicoelosia biloba (Linnaeus). Geol. For. Stockh. Fork. 87,
548-56.

zugmayer, h. 1880. Untersuchungen fiber rhatische Brachiopoden. Beitr. Palaont. Geol. Ost.-Ung.
1, 1-42.

M. J. S. RUDWICK

Department of Geology
Sedgwick Museum

Typescript received from author 28 April 1967 Cambridge

UPPER CRETACEOUS COCCOLITHOPHORIDS
FROM ZULULAND, SOUTH AFRICA

by RICHARD N. PIENAAR

Abstract. A detailed study of some Upper Cretaceous calcareous nannoplankton of Zululand was undertaken,
utilizing the optical as well as the electron microscope. Nine new species belonging to five genera are described
as viewed in the electron microscope. These are Coccolithus cribosphaerella, Coccolithus zuluensium, Cyclolithus
zulua, Discolilhus cristallinus, Discolilltus rhabdosphaericus, Discolithus spiralis, Maslovella africana, Maslovella
blackii, and Maslovella pulchra. Using the sequential occurrence of the coccoliths through the top 800 ft. of
'Zululand Oil Exploration’ Borehole 'A’ the age was determined as being Cretaceous Maestrichtian.

The material on which this study was based was supplied by the Anglo Transvaal
Consolidated Investment Company of South Africa. It belongs to part of this company's
oil prospecting project and is from Borehole ‘A’ drilled near Lake Sibaya, Zululand,
South Africa.

The method used to extract and prepare the coccoliths for observation under both
the optical as well as the electron microscope are similar to those described by Pienaar
(1966).

All samples are housed in the Bernard Price Institute for Palaeontological Research,
University of the Witwatersrand, while the prepared slides and electron micrographs are
housed in the Department of Plant Biology, University of Natal, Durban, South Africa.

TAXONOMIC PROBLEMS

It has been shown by numerous workers in this field that the systematics of this group
of algae is in a state of turmoil. This has arisen mainly out of the fact that there have
been many approaches to this field of study. The use of phase contrast, polarizing, and
ordinary optical microscope as well as the electron microscope have resulted in con-
fusion within the group because these algal remains appear so completely different
when viewed under the different microscopes.

It has been shown that the only satisfactory way to study this group systematically,
is by utilizing the electron microscope which with its high power of resolution reveals
the detailed structure of the coccoliths. This is vitally important as the systematics of
the Coccolithophoridae is based on hard-part morphology; I do not suggest that the
optical microscope be abandoned, and have in fact stressed (Pienaar 1966) that it is
invaluable in the stratigraphic application of the group.

In the present report no attempt has been made to place the coccoliths under any
hierarchical systematic framework. I consider it premature to do so until there is a good
collection of electron micrographs available, and for this reason the coccoliths are
described in alphabetical order.

[Palaeontology, Vol. 11, Part 3, 1968, pp. 361-7, pis. 69-71.]

C 5586 B b

362

PALAEONTOLOGY, VOLUME 11

DESCRIPTIVE TERMINOLOGY

Distal shield. Shield which is the furthest away from the point of attachment of the
coccolith to the coccosphere, and is always convex.

Longitudinal axis. Line joining the two longitudinal poles.

Longitudinal poles. Situated at each end of the longitudinal axis of the theoretical
ellipse that contains the coccolith.

Plates. Elements which comprise the proximal and distal shields; they vary from being
wedge-shaped to rectangular in outline.

Pore. Region where there are no crystals of calcium carbonate, usually within the central
area of the coccolith.

Proximal shield. Part of the coccolith which is in direct contact with the coccosphere;
always concave.

Shield. Main structural element of a coccolith, always composed of plates.

Shield area. Width of the shields when the coccolith is seen in plan view.

Transverse axis. Line joining the two transverse poles.

Transverse poles. Situated at each end of the transverse axis of the theoretical ellipse
that can contain the coccolith.

SYSTEMATIC DESCRIPTIONS
Genus coccolithus Schwarz 1894
Coccolithus cribosphaerella sp. nov.

Holotype. Plate 70, figs. 4, 5.

Locus typicus. Borehole ‘A’, Lake Sibaya, Zululand, South Africa. Assemblage 1382, depth 280 ft.;
Cretaceous.

Diagnosis. Elliptical placoliths composed of two well-developed shields; distal shield
larger, convex, composed of 21-22 plates. The exact detail of the smaller proximal
shield is not known. The central area is sculptured with numerous pores aligned parallel
to the longitudinal axis of the ellipse.

Description. The scalloped outline is due to the overlapping plates of the distal shield.
The proximal shield is presumed to be similarly composed. The central area is sculptured

EXPLANATION TO PLATE 69

Figs. 1, 5. Maslovelta btackii sp. nov. non-replicated. 1, X 28,000. 5, X 10,700.

Fig. 2. Discolithus cristallinus sp. nov., non-replicated; X 21,660.

Fig. 3. Maslovelta pulchra sp. nov., non-replicated; X 27,000.

Fig. 4. Cyclolithus zulua sp. nov. non-replicated; X 13,330.

Figs. 6, 7. Coccolithus zuluensium sp. nov., non-replicated. 6, X 13,330. 7, X 13,750.
Fig. 8. Maslovelta africana sp. nov., non-replicated; X 12,850.

Fig. 9. Discolithus rhabdosphericus sp. nov., non-replicated; X 13,125.

Palaeontology, Vol. 11

PLATE 69

PIENAAR, Cretaceous coccolithophorids

PIENAAR: UPPER CRETACEOUS COCCOLITHOPHORIDS FROM ZULULAND 363

with 3-5 rows of pores aligned parallel to the longitudinal axis of the ellipse; the pores
vary from circular to hexagonal in shape.

Size. Longitudinal axis, 4-5-6-0 p. Transverse axis, 3-1-4-8 p. Width of shield area, 0-75-10 p.

Remarks. This coccolith is well represented in the assemblages studied. The pore size is
variable. Stradner (1961) described Coecolithus opacus which resembles the South
African species and could well be a related form.

Coecolithus zuluensium sp. nov.

Holotype. Plate 69, figs. 6, 7.

Locus typicus. Borehole ‘A’, Lake Sibaya, Zululand, South Africa. Assemblage 1388, depth 330 ft.;
Cretaceous.

Diagnosis. Elliptical coccolith made up of two well-developed shields; distal shield
larger. The central area is spanned by two cross-bars which widen centrally and corre-
spond to the axes of the ellipse.

Description. The distal shield is composed of 27-29 overlapping plates. The longitudinal
polar plates are wedge-shaped and are larger than the remaining rectangular plates. The
proximal shield is similar in construction to the distal shield.

The central area is the characteristic feature of the species; the cross-bars are com-
posed of calcium carbonate crystals arranged in an irregular manner.

Size. Longitudinal axis, 3-25-4-4 p. Transverse axis, 2-3-3-7 p. Width of shield area, 0-8 1-0 p.

Remarks. The coccoliths grouped together in this species resemble Disco/iihus rhabdo-
sphaericus sp. nov. but differ in that they have a proximal and a distal shield and lack
a central boss. The species is assigned to the genus Coecolithus because of the two-
shielded nature. It has been noticed that during the course of the study of the South
African Coccolithophorids that the members belonging to the genus Coecolithus
usually possess large wedge-shaped polar plates.

Genus cyclolithus Kamptner 1948
Cyclolithus zulua sp. nov.

Holotype. Plate 69, fig. 4; Plate 70, fig. 1.

Locus typicus. Borehole ‘A’, Lake Sibaya, Zululand, South Africa. Assemblage 1382, depth 280 ft-;
Cretaceous.

Diagnosis. Elliptical coccolith composed of a single shield abutting against a raised
central rim. The shield is made up of 16-17 plates.

Description. The 16-17 plates are rectangular and overlap slightly. Towards the central
area is a raised rim, against which the plates abutt. The raised portion is levigate and
devoid of any sculpture.

Size. Longitudinal axis, 4-4-6 0 p. Transverse axis, 3-6-4T p. Width of shield area, 0-9-10 p.

364

PALAEONTOLOGY, VOLUME 11

Remarks. On first impression this specimen appears to have two shields but on closer
examination the second shield is seen to be a slightly raised rim with practically no plate
structure. Because of its one-shielded nature and the central area devoid of any sculp-
ture, the specimens are placed in the genus Cyclolithus Kamptner 1948.

Remarks. Under the Botanical Code of Nomenclature, the name Discolithus is retained
for these algal fossils (cf. Loeblich and Tappan 1963).

Genus discolithus Kamptner 1948
Discolithus cristallinus sp. nov.

Holotype. Plate 69, fig. 2; Plate 70, fig. 2.

Locus typicus. Borehole ‘A’, Lake Sibaya, Zululand, South Africa. Assemblage 1382, depth 280 ft.;
Cretaceous.

Diagnosis. Elliptical one-shielded coccolith composed of 24-40 overlapping plates. The
central area is infilled with crystals of calcium carbonate arranged in an irregular order.

Description. The average number of plates is between 28 and 30.

Size. Longitudinal axis, 2-4-2-75 p. Transverse axis, 1 -4-7-0 p. Width of shield area, 0-25-0-45 p.

Remarks. Occasionally specimens of Discolithus cristalinus were found with a distinct
row of crystals following the outline of the central area. Their appearance was almost
like the beginning of a second shield. In addition some forms have larger crystals of
calcium carbonate covering the smaller crystals. This species is easily recognized by its
single shield and the infilled central area.

Discolithus rhabdosphaericus. sp. nov.

Holotype. Plate 69, fig. 9; Plate 71, figs. 1, 2, and 6.

Locus typicus. Borehole ‘A’, Lake Sibaya, Zululand, South Africa. Assemblage 1382, depth 280 ft.;
Cretaceous.

Diagnosis. Elliptical one-shielded coccolith composed of 35 overlapping plates. The
central area is spanned by a solid parallelogram-like structure, the diagonals of which
correspond with the axes of the ellipse.

Description. The sides of the parallelogram are concave and each composed of six plates.
The diagonals are raised and unite in the centre to form a boss which is perforated by
an axial pore. The proximal surface of the coccolith is markedly concave and no central

EXPLANATION TO PLATE 70
Fig. 1. Cyclolithus zulua sp. nov., replicated; X 30,000.

Fig. 2. Discolithus cristallinus sp. nov., non-replicated ; X 22,000.

Fig. 3. Discolithus spiralis sp. nov., replicated; X 12,000.

Figs. 4, 5. Coccolithus cribosphaerella sp. nov., replicated. 4, X 14,000. 5, X 12,000.
Fig. 6. Maslovella af icana sp. nov., replicated; X 17,500.

Palaeontology, Vol. 11

PLATE 70

PIENAAR, Cretaceous coccolithophorids

PIENAAR: UPPER CRETACEOUS COCCOLITHOPHORIDS FROM ZULULAND 365

boss is observed when the coccolith is found in this position. The distal surface is
markedly convex in comparison with the proximal surface.

Size. Longitudinal axis, 3-5-4-2 p. Transverse axis, 2-7-3-2 p. Width of shield area, 04-0-6 /a.

Remarks. This is a very common and distinctive species and was assigned to the genus
Discolithus Kamptner 1948 because of its one-shielded nature and the infilled central
area.

Discolithus spiralis sp. nov.

Holotype. Plate 70, fig. 3; Plate 71, fig. 4.

Locus typicus. Borehole ‘A’, Lake Sibaya, Zululand, South Africa. Assemblage 1387, depth 320 ft.;
Cretaceous.

Diagnosis. Ellipitical one-shielded coccolith with the central area infilled with crystals
of calcium carbonate arranged in a sigmoid pattern.

Description. Shield composed of approximately 56 overlapping plates. The central area
is large and elliptical and completely infilled.

Size. Longitudinal axis, 5-4-5-7 /a. Transverse axis, 3-6-3-7 p. Width of shield area, 0-3-0-4 p.

Remarks. This is a very distinctive coccolith and may always be recognized by the
sigmoid arrangement of the crystals infilling the central area.

Genus maslovella Tappan and Loeblich 1966
Synonym. Covillea Black 1964.

Maslovella africana sp. nov.

Holotype. Plate 69, fig. 8; Plate 70, fig. 6; Plate 71, figs. 3, 5.

Locus typicus. Borehole ‘A’, Lake Sibaya, Zululand, South Africa. Assemblage 1387, depth 320 ft.;
Cretaceous.

Diagnosis. Circular to subcircular asymmetrical coccolith composed of two well-
developed shields. At the locality of the axial pore the central area is infilled with crystals
of calcium carbonate. The distal shield is larger than the proximal shield and is placed
asymmetrically on top of it.

Description. The distal shield is composed of 25-29 overlapping plates, the average
number of plates being 25. The plates are wedge-shaped and those situated at the
longitudinal polar regions are larger and more markedly wedge-shaped than the re-
maining plates. The proximal shield is smaller than the distal shield and composed of
25-29 plates of variable size. In the region of the one longitudinal pole are wedge-shaped
plates which are only a little smaller than the distal shield plates. At the opposite longi-
tudinal pole all the plates are smaller and less than half the size of the distal shield plates.
The central area is infilled with irregularly arranged crystals of calcium carbonate.

Size. Longitudinal axis, 3T-3-2 p. Transverse axis, 2-5-2-7 p. Proximal shield, 1 -9—2-6 p X 2-75-2-9 p.
Distal shield, 3T-2-5 /ax4-1-3-7 p.

Remarks. Maslovella africana is common in most of the assemblages studied and
characterized by the asymmetrically placed shields. It is tentatively placed in the genus

366

PALAEONTOLOGY, VOLUME 11

Maslovella Tappan and Loeblich 1966 which it most closely resembles. Black (1964),
however, did not mention in his description of the type specimen any asymmetry in the
genus, and thus a new genus might have to be erected. This form has also been found
by the author in Type Maestrichtian material sent to him by Dr. E. Martini.

Maslovella blackii sp. nov.

Holotype. Plate 69, figs. 1, 5.

Locus typicas. Borehole ‘A’, Lake Sibaya, Zululand, South Africa. Assemblage 1382, depth 280 ft.;
Cretaceous.

Diagnosis. Elliptical two-shielded coccolith. The distal shield is larger than the proximal
shield and is composed of 17-27 plates. The central area is completely infilled with
crystals of calcium carbonate.

Description. The plates situated at the longitudinal polar region are larger and wedge-
shaped while the remaining plates are rectangular in shape and overlap to varying
degrees. The structure of the proximal shield is not known but it is thought to be
similar to that of the distal shield. The central area is large, elliptical, and completely
infilled.

Size. Longitudinal axis, 1-75-2-8 p. Transverse axis, 1 -5-2-0 p. Width of shield area, 0-5-0-6 p.

Remarks. During the study of the South African sediments a number of coccoliths were
found which were similar to the type description of the genus Maslovella (Black)
Tappan and Loeblich 1966, except for a few minor differences. Some had only fourteen
plates in the distal shield; others had only one well-developed shield and a robust rim
in the place of the second shield. The patterning of the central area also varies by having
slightly regular to irregular crystals infilling the central area or with two rows of crystals
meeting along the longitudinal axis of the ellipse. These forms are thought to be
intermediates or broken forms and are all grouped together into Maslovella blackii.
Maslovella blackii differs from M. africana in that the latter form possesses a distinct
asymmetry of the two shields, whereas the former species is symmetrical.

Maslovella pulclira sp. nov.

Holotype. Plate 69, fig. 3.

Locus typicus. Borehole ‘A’, Lake Sibaya, Zululand, South Africa. Assemblage 1387, depth 320 ft.;
Cretaceous.

Diagnosis. Elliptical coccolith composed of two well-developed shields. The distal
shield and the smaller proximal shield are both composed of seventeen plates. The
central area is elliptical and infilled with prolongations of the proximal shield plates.

EXPLANATION TO PLATE 71

Figs. 1 , 2, 6. Discolithus rhabdosphericus sp. nov., replicated. 1, Distal view, X 17,500. 2, Proximal view,
x 15,000. 6, Proximal view, x 20,000.

Figs. 3, 5. Maslovella africana sp. nov. replicated; 3, Proximal view, x 25,000. 5, Distal view, X 16,000.
Fig. 4. Discolithus spiralis sp. nov., replicated; Proximal view, X 12,000.

Palaeontology, Vol. 11

PLATE 71

PIENAAR, Cretaceous coccolithophorids

PIENAAR: UPPER CRETACEOUS COCCOLITHOPHORIDS FROM ZULULAND 367

Description. The three longitudinal polar plates at each end are larger than the remaining
plates and are distinctly wedge-shaped. The proximal shield plates have prolongations
which dip inwards towards the central area and meet along the longitudinal axis of the
ellipse. The prolongations are rectangular and are on all proximal plates except the six
longitudinal polar plates.

Size. Longitudinal axis, 2 0 /x. Transverse axis, L4 /x. Width of shield area, 0-5 /x.

Remarks. This form is tentatively placed in the genus Maslovella (Black) Tappan and
Loeblich 1966 because of the two distinctive shields and the infilled central area. It
differs from all the previously described species belonging to this genus in the detail and
delicate construction of the central area.

CONCLUSION

The South African material studied is very rich in the remains of these algae and
affords excellent opportunity for a more detailed investigation of the coccolith flora of
both earlier and later horizons. The work reported in this paper is only a portion of the
work done on the Upper Cretaceous of Zululand, South Africa. The remainder of the
new species and variations of existing species will be described in later papers.

At present a detailed account of the microstratigraphy of the Cretaceous of Zululand
is being prepared. The age of the strata from which the coccoliths were described was
determined by the presence of Maestrichtian foraminifera (De Gasparis 1967); this
conclusion was supported by the occurrence of Maestrichtian coccoliths.

Acknowledgements. The author is indebted to Anglo-Transvaal Consolidated Investment Company
who sponsored this study and provided the material. To Dr. G. F. Hart the author is grateful for his
advice and help.

REFERENCES

black, m. 1964. Cretaceous and tertiary coccoliths from Atlantic seamounts. Palaeontology, 7, 307-16.

bramlette, m. m., and MARTINI, E. 1964. The great change in the calcareous nannoplankton fossils
between the Maestrichtian and the Danian. Micropalaeontology, 10, 291-322.

de gasparis, m. l. 1967. The microstratigraphy of the Cretaceous System of Zululand. M.Sc. Thesis,
Univ. Witwatersrand.

loeblich, a. r. jr., and tappan, h. 1963. Type fixation and validation of certain calcareous nanno-
plankton genera. Proc. biol. Soc. Washington, 76, 191-6.

pienaar, r. n. 1966. Microfossils from the Cretaceous System of Zululand as studied with the electron
microscope. S. Afr. J. Sci. 62, 147-57.

stradner, h. 1961. Vorkommen van Nannofossilien im Mesozoikum und alttertiar. Erdoel-Ztschr.
77, 3, 77-88.

tappan, h., and loeblich, a. r. jr. 1966. Maslovella nom. nov. Taxon, 15, 43.

RICHARD N. PIENAAR

Department of Plant Biology
University of Natal
Durban
South Africa

Manuscript received from author 10 July 1967

A NEW EOCENE C ASSIG ERINELLA FROM

FLORIDA

by W. G. GORDEY

Abstract. A new species of Cassigerinella (C. eocaenica ) is proposed for specimens obtained from an Eocene
sample taken off Blake Plateau, Florida. The species is similar, but quite distinct from the Oligocene marker
C. chipolensis (Cushman and Ponton 1932). Therefore the stratigraphical value of the overlap in ranges of
C. chipolensis with Pseudohastigerina micra (Cole 1927) as indicators of basal Oligocene is not affected.

The joint occurrence of the species Pseudohastigerina micra (Cole 1927) and Cassi-
gerinella chipolensis (Cushman and Ponton) has been regarded as a means of recognizing
the basal Oligocene, at least within the tropical and semi-tropical belts. This overlap was
first recognized by Blow and Banner (in Eames et ah 1962, p. 68) as a further criterion for
the recognition of their zone of Globigerina sellii Borsetti 1959 (= G. oiigocaenica Blow
and Banner 1962). Saunders and Cordey (1965) also recognized this overlap in a study
of the Oceanic Formation of Barbados. It occurred in samples immediately overlying
deposits of definite Eocene age containing such forms as Hantkenina spp., Globorotalia
centralis, and G. cerro-azulensis. Bolli (1966), in a general review of world-wide plank-
tonic zonations, added further support to the stratigraphic value of this observed overlap.

Saito and Be (1963) established an Oligocene age (sensu Eames et al., op. cit.) for the
Vicksburg group of the Gulf Coast region. However, they stated (p. 704) that C. chipo-
lensis and P. micra occurred together with Eocene forms (e.g. Hantkenina alabamensis,
H. primitiva, Globorotalia cerro-azulensis) from a core taken off the Florida coast
(Lamont Core A167-21, 29° 49' N., 79° 39' W.). This record therefore cast considerable
doubt on the stratigraphic value of this overlap as far as the recognition of the Eocene-
Oligocene boundary or basal Oligocene was concerned. Blow examined these cassi-
gerinellids and concluded (pers. comm.) that they were different from C. chipolensis.
Through the kindness of Drs. Be and Saito the writer obtained material which contained
the same form of Cassigerinella with Hantkenina spp., but from a location to the north
of the Lamont Core, at Blake Plateau, 30° 04-8' N., 79° 14-5' W., Sample J.6B.

A careful comparison of the cassigerinellids in the Blake Plateau material with
C. chipolensis from both the Globigerina ampliapertura zone of Trinidad and the Oceanic
Formation of Barbados indicates that the Eocene specimens differ from C. chipolensis.
In view of the stratigraphic significance of C. chipolensis , it is desirable that a new species
should be erected for these Eocene occurrences.

SYSTEMATIC DESCRIPTIONS
Genus cassigerinella Pokorny 1955
Cassigerinella eocaenica sp. nov.

Text-fig. 1 a-e

Description. Test very small, calcareous, perforate; 8-9 chambers in the final whorl,
showing a gradual and uniform increase in size, moderately inflated. Initial two (possibly

[Palaeontology, Vol. 11, Part 3, 1968, pp. 368-70.]

CORDEY: A NEW EOCENE C ASS IG ERIN ELL A FROM FLORIDA

369

three) chambers planispirally arranged, later becoming biserial alternating, but coiled
in the same plane. Periphery lobulate, initially sub-acute, becoming sub-rounded,
sutures distinct, depressed, curved. Aperture a latero-marginal, extra-umbilical arch.

TEXT-FIG. 1.

a-e, Cassigerinella eocaenica sp. nov. Sample J-6b, 30° 04-8' N., 79° 14-5' W., 215'
from top. a-c, Holotype, BMNH P46838; d, e, Paratype, BMNH P46839; both
X 300.

f-m, Cassigerinella chipolensis (Cushman and Ponton), f-h. Dissected hypotype
showing penultimate whorl, early chambers planispirally coiled, later chambers
alternating; G. ampliapertura Zone, Cipero Coast, Trinidad; BMNH P46840,

X 213. i-k. After Cushman and Ponton, X 150. /, m, Hypotype, G. ampliapertura
zone, Cipero Coast, Trinidad; BMNH P46841, x275.

Remarks. C. eocaenica is similar to C. chipolensis, but differs in being consistently
smaller, the greatest breadth of the final whorl varying from 0-1 to 0-12 mm., the average
being 0-1 mm. Measurement of 45 specimens of C. chipolensis shows a variation in the
breadth of the final whorl from 0T3 to 0T9 mm., the average being 0-16 mm. Secondly,
the chambers in the final whorl are less inflated than in C. chipolensis. Specimens of
C. chipolensis from Barbados (Oceanic Formation), from the Cipero Formation of
Trinidad and from the Oligocene of Lindi (Blow and Banner in Eames et al. 1962, pi. 15,
figs, m-n), show a more rapid increase in the size of the chambers of the final whorl

370

PALAEONTOLOGY, VOLUME 11

than in C. eocaenica. C. chipolensis also has a more rounded periphery throughout.
Finally, the chamber arrangement in the final whorl of C. eocaenica varies from an
initial planispiral arrangement to alternating, whereas the chamber arrangement in
C. chipolensis is entirely alternating throughout the final whorl. The planispiral arrange-
ment is only developed in the first two or three chambers of the penultimate whorl of
C. chipolensis (text-fig. 1 ,f-h). The only other species which shows any morphological
similarity to C. eocaenica is C. globolocula Ivanova 1958. Pokorny (in Eames et al.)
agreed that his species C. boudecensis was probably conspecific with chipolensis , and
considered that globolocula was ‘certainly conspecific with boudecensis ’ (op. cit., p. 83).
The writer would agree with this conclusion and therefore the above remarks on
C. chipolensis and C. eocaenica apply equally to C. globolocula (and also C. boudecensis).

Deposition of types. Holotype and paratype specimens are deposited in the British Museum (Natural
History). An unfigured paratype is deposited in the United States National Museum, No. 643514.

Material. Fifteen specimens of C. eocaenica', forty-five specimens of C. chipolensis.

Acknowledgements. The author is grateful to Drs. Be and Saito (Lamont Geological Observatory,
Columbia University, New York) for the donation of the samples upon which this study is based;
Dr. R. Lagaaij and J. A. Postuma (Bataafse Internationale Petroleum Maatschappij N.V., The Hague),
for their critical reading of the manuscript; and the Bataafse Internationale Petroleum Maatschappij
N.V., for permission to publish this paper.

REFERENCES

bolli, h. m. 1966. Zonation of Cretaceous to Pliocene marine sediments based on planktonic Fora-
minifera. Bo In inf. Asoc. Venezolana de Geologia, Mineraria y Petroleo, 9(1), 3-32.

eames, F. e. etal. 1962. Fundamentals of Mid-Tertiary Stratigraphical Correlation. Cambridge University
Press.

ivanova, l. G. in bykova, n. k. 1958. Novye Rody i Vidy Foraminifera. Trudy vnigri, no. 115,
Mikrofauna SSR, 9, 4-81.

saito, t. and be, a. w. h. 1963. Planktonic Foraminifera from the American Oligocene. Science, 145,
703-4.

saunders, j. b. and cordey, w. g. 1965. The biostratigraphy of the Oceanic Formation in the Bath
Cliff section, Barbados. Proc. 4th Caribbean Geol. Congress, Port of Spain, 1965. (In press.)

w. G. CORDEY
Bataafse Int. Petr. Mij.

EP/12, Carel van Bylandtlaan 30
The Hague

Typescript received from author 29 June 1967 Netherlands

MORPHOLOGY AND PHYLOGENY OF
ORBULINOIDES BECKMANNII (SAITO 1962)

by w. G. CORDEY

Abstract. An examination of dissected specimens of the Eocene species Orbulinoides beckmannii (Saito 1962),
formerly Porticulasphaera mexicana (Cushman 1925), reveals certain morphological features not hitherto
described. A study of the ontogeny of O. beckmannii and a comparison with the contemporaneous species
Globigerapsis kugleri Belli, Loeblich and Tappan, and Globigerinatheka barri Bolli, Loeblich and Tappan, indi-
cate that O. beckmannii is unlikely to be related to either of these species. It appears to have developed from some
globorotaloid ancestor.

This study is based on specimens obtained from a block of extraneous Eocene in the
Miocene Nariva Formation of Trinidad. This same material was used by Bolli (19576,
p. 1 59) in his monograph on the Eocene faunas of Trinidad. Bolli, Loeblich and Tappan
(1957, p. 34) gave a full and accurate description of their new genus Porticulasphaera,
and Bolli (1957u, p. 116) in placing the Miocene species Globigerinoides glomerosa Blow
1965 in this new Eocene genus commented that detailed comparative studies would
probably \ . . reveal differences between the Eocene and Miocene forms . . (op. cit.,
p. 115). Bolli, Loeblich and Tappan designated Globigerina mexicana Cushman 1925
as the type species of their new genus Porticulasphaera. Saito (1962) considered that the
Cushman type should be referred to the genus Globigerapsis Bolli, Loeblich and Tappan.
Further examination of G. mexicana by Saito, and independently by Blow, has led to
the conclusion that G. mexicana Cushman is conspecific with Globigerapsis semiinvo/uta
(Keijzer 1945). The writer has also examined the holotype of G. mexicana and agrees
with their conclusion. Blow and Saito have therefore assigned the specimens which
Bolli ( 1957a) previously referred to as Porticulasphaera mexicana (Cushman 1925), and
upon which this study is based, to the new genus Orbulinoides as O. beckmannii (Saito
1962).

Olsson (1965) subsequently erected the genus Praeorbulina for the Miocene forms
(i.e. Porticulasphaera of Bolli 1957(7, p. 115) with the type species Globigerinoides
glomerosa glomerosa Blow 1956. The object of this paper is to discuss certain morpho-
logical features observed in O. beckmannii, and which have not previously been
described.

Morphology. Externally O. beckmannii shows numerous apertures at the base of the
final chamber. The early trochospiral part of the test shows between one and four
supplementary apertures; occasionally no such apertures were present. When
present they are usually located at the junction of the spiral and intercameral sutures.
The specimen illustrated by Bolli (1957a, pi. 37, fig. 1 a) showing eight such apertures is
considered atypical (cf. Bolli, Loeblich and Tappan, 1957, pi. 6, figs. 9a, b).

A removal of the final chamber and part of the walls of the last three chambers of the
initial trochospire (text-fig. lc) reveals the numerous supplementary apertures described
by Bolli, Loeblich and Tappan (1957) and Olsson (1965). However, a closer examination

[Palaeontology, Vol. 11, Part 3, 1968, pp. 371-5.]

372

PALAEONTOLOGY, VOLUME 11

text-fig. 1. a-d, Orbulinoides beckmannii (Saito 1962). a. Dissected specimen showing earliest whorl
with an umbilical-extra-umbilical aperture, BMNH P46842, x 76. b, c. Dissected specimens,
BMNH P46843, 46844, x 105. d. Dissected specimen showing trochospiral supplementary aperture
(ts) opening into the vestibule (v), BMNH P46846, x93.
e-g, Globigerapsis kugleri Bolli, Loeblich and Tappan 1957. Dissected specimens showing globigerinid
primary aperture ; e. Specimen with Bulla partly removed, X 1 20 ;/, Bulla and two chambers removed,
x 120; g. Bulla and five chambers removed; BMNH P46847, X 105.
h-j, Globigerinatheka barri Bolli, Loeblich and Tappan 1957. Dissected specimens showing a globi-
gerinid primary aperture; h, Bulla partly removed, X 100; i, Bulla and two chambers removed, x 100;
j. Bulla and five chambers removed; BMNH P46848, X 73.

All specimens from the O. beckmannii Zone, Navet Formation, Eocene, Point-a-Pierre, Trinidad.

CORDEY: O RB ULINO ID ES BECKMANNII (SAIT O 1962) 373

of these apertures shows that in no instance are they directly connected with the
outside of the test. They open into a small cavity (here termed vestibule) between the
thick outer wall and the delicate wall of the initial trochospirally arranged chambers
(text-fig. le). This vestibule is usually situated at the junction of the spiral and inter-
cameral sutures, but probably also extends some way along the intercameral suture.
It is also seen that the external supplementary apertures on the initial trochospire were
never aligned with the internal apertures. Therefore, there is only an indirect connexion,
that is via the vestibule, between the inside and outside of the test, as far as the early
supplementary apertures are concerned.

Bolli’s comparison (op. cit., p. 115) of the supplementary apertures of Orbu/inoides
and Globigerinoides cannot be upheld. There is a basic difference in that, in the latter
genus, these apertures are never covered by subsequent thickening of the test and are in
direct communication with the inside of the test. It would seem that early supplementary
apertures became vestigial structures, in the sense that the degree of direct connexion
between the inside and outside of the test in this area was much reduced. This appears
to be in contrast with species of Globigerinoides (and also Orbulina suturalis Bronnimann
1951) where a direct connexion is maintained for each of the apertures.

The method of test thickening in Orbulinoides also appears to be in contrast with
certain fossil species of Globigerinoides (e.g. G. subquadratus Bronnimann 1954 = G.
ruber (d’Orbigny 1826) of Bolli 1957a). In this genus thickening is progressive rather
than occurring only after the adult stage has been reached.

A consideration of the morphology of the initial trochospire of Orbulinoides (e.g. its
large primary aperture, inflated spinose chambers, numerous large supplementary
apertures, and thin wall) strongly suggests that during the trochospiral stage the animal
inhabited the epipelagic zone ( sensu Hedgepeth 1957, p. 18, fig. 1). The subsequent test
thickening, and the reduction in the number of supplementary apertures, might be
correlated with its migration to deeper levels. The work of Be and Ericson (1963) and
Be (1965) offers some support for this view, since a correlation between the thickness of
the test with depth in the Recent species Globorotalia truncatulinoides and Globigeri-
noides sacculifera (Brady) was observed. The reduction in the number of ‘functional’
supplementary apertures may have been one means of increasing its weight in order to
occupy lower levels. It is equally possible that the thickening of the test and reduction
in supplementary apertures are associated with reproduction. The final inflated chamber
may represent a type of brood pouch.

The presence of such an inflated final chamber in G/obigerapsis and Globigerinatheka
may indicate a similar mode of reproduction in these genera. Le Calvez(1936) showed that
in Orbulina minima (d’Orbigny) there was a gradual reduction in the globigerine part of
the test with depth. Furthermore, in specimens from the deepest levels (about 300 m.)
the globigerine chambers had completely disappeared, and all that remained was the
spheroidal final chamber which was frequently found to be filled with gametes. There-
fore, there is some support for the view that the inflated final chambers of these Eocene
genera might be connected with reproduction.

Phylogeny. Bolli (1957a, p. 160) stated that ‘. . . Globigerapsis, Globigerinatheka and
Porticulasphaera obviously represent a related group’. He considered that Globigerapsis
index (Finlay) gave rise to G. kugleri Bolli, Loeblich and Tappan, from which O.

374

PALAEONTOLOGY, VOLUME 11

beckmannii developed. His further support for this view was the fact that the three genera
showed a 90 per cent, tendency to coil dextrally. Bolli is correct in the case of Globiger-
apsis and Globigerinatheka, and Eckert’s (1963) study of these genera supports this con-
clusion. However, in the writer’s opinion, Orbulinoides is unrelated to either of these
genera. The gross morphological similarity of these three species is more likely to be a
function of convergent evolution than any genetic affinity (at least as far as beckmannii is
concerned). This view is based on a comparison of the ontogeny of G. kugleri , G. barri ,
and O. beckmannii (text-fig. 1). It is clear that both kugleri and barri have arisen from a
globigerinid ancestral form, while O. beckmannii shows an unmistakable umbilical-
extra-umbilical primary aperture initially, and therefore is derived from a globorota-
loid ancestor. The striking differences in the morphology of the initial whorls, parti-
cularly the distinctive shape of the later trochospiral chambers of beckmannii , and
the large primary aperture, adds further support to this view (text-fig. 1, cf. figs.
a , b with f-j). Bolli considered the gross similarity of the adult tests of these three
species the most important phylogenetic factors. The writer, however, considers that
the evidence of the ontogeny of these species outweighs the adult similarity in gross
morphology.

Acknowledgements. The writer is indebted to J. B. Saunders (Texaco Trinidad Inc.) for the material on
which this study is based; Dr. R. Lagaaij and J. A. Postuma (Bataafse Internationale Petroleum
Maatschappij N.V., The Hague) and Dr. J. F. Noorthoorn van der Kruijff (Koninklijke/Shell Explo-
ratie en Produktie Laboratorium, Rijswijk) for much helpful discussion and criticism; Dr. W. H.
Blow for providing a copy of the Blow and Saito MS.; Dr. R. Cifelli (U.S. National Museum) for
permission to examine the types; and Bataafse Internationale Petroleum Maatschappij N.V. for per-
mission to publish this paper.

REFERENCES

bandy, o. l. 1965. Restrictions of the ‘Orbulina’ datum. Micropaleontology, 12, 77-86, pi. 1.

be, a. w. h. 1965. The influence of depth on shell growth in Globigerinoides sacculifera (Brady). Ibid.
11, 81-97.

and ericson, d. b. 1963. Aspects of calcification in planktonic foraminifera (Sarcodina). Ann.

N.Y. Acad. Sci. 109, 65-81.

blow, w. h. 1956. Origin and evolution of the foraminiferal genus Orbidina d’Orbigny. Micropaleonto-
logv, 2, 57-70.

and saito, t. 1967. The morphology and taxonomy of Globigerina mexicana Cushman. (In press.)

bolli, h. m. 1957a. Planktonic Foraminifera from the Oligocene-Miocene Cipero and Lengua Forma-
tions of Trinidad, B.W.I. Bull. U.S. natn. Mus. 215, 97-123, pis. 22-29.

19576. Planktonic Foraminifera from the Eocene Navet and San Fernando Formations, Trinidad,

B.W.I. Ibid. 215, 155-72, pis. 35-39.

loeblich, a. r., and tappan, h. 1957. Planktonic foraminiferal families Hantkeninidae, Orbu-

linidae, Globorotaliidae, and Globotruncanidae. Ibid. 215, 3-50.

eckert, h. r. 1963. Die obereozanen G/obigerinen- Schiefer (Stad- und Schimbergschiefer) zwischen
Pilatus und Schrattenfluh. Eclog. geol. Helv. 56, 1001-72.

hedgepeth, w. 1957. In ‘Treatise on Marine Ecology and Paleoecology’. Mem. geol. Soc. Am. 67.

le calvez, j. 1936. Modifications du test des Foraminiferes pelagiques en rapport avec la reproduction
Orbidina universa d’Orb. et Tretomphalus bulloides d'Orb. Ann/s Protist. 5.

CORDEY: ORBULINOIDES BECKMANNII (SAITO 1962) 375

loeblich, a. r., and tappan, H. 1964. Treatise on Invertebrate Paleontology, ed. r. c. moore, Part C,
Protista 2, 2. Geol. Soc. Am. and Univ. Kansas Press.
olsson, r. k. 1965. Praeorbulina Olsson, a new foraminiferal genus. J. Paleont. 38, 770-1.
saito, t. 1962. Eocene Planktonic Foraminifera from Hahajima (Hillsborough Island). Trans. Proc.
palaeont. Soc. Japan , n.s. no. 45, 209-25.

w. G. CORDEY
Bataafse Int. Petr. Mij.
EP/12, Carel van Bylandtlaan 30
The Hague

Typescript received from author 29 June 1967 Netherlands

THE GASTRIC CONTENTS OF AN
ICHTHYOSAUR FROM THE LOWER LIAS OF
LYME REGIS, DORSET

by JOHN E. POLLARD

Abstract. The partial skeleton of a small ichthyosaur associated with the gastric mass is described from the
Lower Lias of Lyme Regis. The gastric mass was oval in shape and composed of minute dibranchiate cephalopod
hooklets in random orientation. Four distinct types of hooklet are recognized in these gastric contents.

Examination of published records and museum specimens suggests that gastric contents composed of cepha-
lopod remains are more commonly preserved than those of fish remains. A study of ichthyosaur coprolites shows
a predominance of defecated fish remains and an absence of hooklets from these structures. The diet, mode of
feeding, and digestive mechanism of Liassic ichthyosaurs, in comparison with a teuthophagous cetacean, the
sperm whale, are considered.

In April 1963 the skeleton of a small ichthyosaur was found in the shales of the Lower
Lias on the foreshore west of Lyme Regis, Dorset. Unfortunately due to the exposed
location of this specimen, in soft shaly-mudstone at about the half-tide level, only a
short time was available for its extraction. Only the anterior part of the skeleton could
be recovered consisting of parts of the skull, pectoral girdle, vertebral column, and ribs.
Careful preparation of this material showed that the skeleton was crushed and slightly
dismembered, but that the stomach contents were preserved as a dark discrete area under-
neath the vertebral column and ribs. Such occurrences are fairly well known, but the
good state of preservation and lack of dispersal of the stomach contents of this specimen
make them worthy of detailed description, quantitative analysis, and discussion in terms
of the feeding habits and digestive mechanism of the Liassic ichthyosaurs.

DESCRIPTION

Horizon and locality. The specimen was collected from the shales of the lower part of the Psiloceras
planorbis Zone of the Lower Lias (Woodward and Ussher 1911, p. 38), at the south-east corner of
Pinhay Bay, two miles west of Lyme Regis (National Grid Reference: SY 325907). The enclosing
sediment was a poorly fossiliferous silty and shaly mudstone, which was interbedded with thin lime-
stones and shales containing Liostrea liassica, Hemicidaris spines, and rarely P/agiostoma gigantea and
Psiloceras planorbis. No other vertebrate remains or coprolites were observed at this horizon.

Skeletal remains. The partial skeleton of the ichthyosaur extracted was 2 ft. (60 cm.) in
length and consisted of the skull and parts of the vertebral column, pectoral and pelvic
girdles, rib cage, and a paddle. (The specimen is now preserved in the collections of the
Geology Department, University of Manchester, registration number SF.l.) Text-fig. 1
is drawn from a field photograph of the specimen in situ, and shows the relative positions
of the bones and the gastric mass from the dorsal aspect. The prepared skeleton can be
examined both dorsally and ventrally, Plate 72, figs. 1 and 2, and enables the individual
bones to be identified.

[Palaeontology, Vol. 11, Part 3, 1968, pp. 376 88, pis. 72-73.]

J. E. POLLARD: GASTRIC CONTENTS OF AN ICHTHYOSAUR

377

The skull is 10 in. (25 cm.) long, crushed dorso-ventrally and twisted sinistrally. The
anterior part of the snout is missing, but the premaxilla and eleven upper jaw teeth and
thirteen lower jaw teeth on the right side of the mouth are visible on the upper surface
(PI. 72, fig. 1). On the under surface of the skull both dentary bones are present and
twenty-three upper jaw teeth, and sixteen lower jaw teeth, from the left side of the mouth
(PI. 72, fig. 2). The teeth appear to be well formed typical ichthyosaur teeth, up to 13 mm.
in length exposed, with smooth apices and bifurcating grooves on the crown. The form
of the tooth crown is close to that of Ichthyosaurus communis Conybeare as figured by
Owen (1881, pi. 24, fig. 5). The anterior part of the right orbit was present dorsally

0 6

1 r 1 r

0 10 20

12

+

30

40

18

_L_

~r

50

24 in.

60 cm.

text-fig. 1. Ichthyosaur skeleton in situ in the Lower Lias west of Lyme Regis, showing the relation-
ship of the various bones to the gastric mass. Widely spaced fine stippling represents the shale matrix,
while the closely spaced coarser stippling represents the gastric contents.

(text-fig. 1) but no sclerotic plates were seen. The hind part of the skull is badly crushed,
and the only other bones clearly recognizable are displaced fragments of the articular
and basioccipital (PI. 72, fig. 1).

The post-cranial skeleton is represented by a total of twenty-eight vertebrae and
numerous fragments of single and double ribs. On the upper (dorsal) surface of the
specimen, PI. 72, fig. 1, seven thoracic vertebrae occur in a row from 4 to 8 in. (10-12
cm.) behind the skull, and bound the gastric mass dextrally. The dorsal left boundary of
the gastric mass is formed by a series of parallel double ribs (text-fig. 1 and PI. 72, fig. 1),
while a complex of complete and broken double and single ribs are elsewhere compressed
into the gastric mass dorsally.

The anterior boundary of the gastric mass on the ventral side of the specimen is
formed by the bones of the pectoral girdle (PI. 72, fig. 2). Parts of the left coracoid,
humerus, and scapula are clearly recognizable and are impressed into the gastric mass
ventrally. The interclavicle is present, and fragments of fifteen phalangeal bones of the
left anterior paddle were found just beyond the humerus. The only other recognizable
bones collected were a displaced pubis and three phalangeal bones of a posterior paddle,
Plate 72, fig. 1, all occurring postero-dextrally of the gastric mass.

The nature and arrangement of these skeletal remains suggest that there had been
a fair amount of displacement of the bones during burial and that the gastric mass must

c c

C 5586

378

PALAEONTOLOGY, VOLUME 11

have been trapped in an unusually anterior position, crushed between the anterior-
dorsal thoracic rib cage and the pectoral girdle.

Gastric mass. The gastric mass of this specimen, text-fig. 1, was broadly oval in shape,
compressed dorso-ventrally, 13-5 cm. long from anterior to posterior, and 8 -5-9-0 cm.
wide from right to left of the skeleton. Only the anterior part of the mass (8 cm. anterior
to posterior by 7 cm. right to left) was collected and prepared for further study (PI. 72,
fig. 1 ; PI. 73, fig. 1). Estimates of the dorsal area of the stomach mass, measured from

the field photograph vary from 85-5 to 95-5
sq. cm. or approximately 90±5 sq. cm. The
depth or thickness of the dorso-ventral cross
section of the gastric mass was measured
accurately on the prepared specimen (PI. 73,
fig. 1), using a travelling microscope, and
varied from 0-25 to 0-75 cm., with a mean
value of about 0-33 cm.

The cleaned and prepared dorsal and ven-
tral surfaces of the gastric mass, Plate 72,
fig. 2; Plate 73, figs. 1 and 2, show that the
stomach contents preserved consist of a
densely packed mass of dibranchiate cepha-
lopod hooklets and rare large quartz grains.
These hooklets are packed in random orien-
tation (PI. 73, fig. 2) in a matrix of finely
crystalline calcite. The quartz grains are sub-
angular or sub-rounded in shape from 0-25
to 1 -40 mm. in diameter, sparsely distributed
on the dorsal surface, but occurring in considerable concentration in patches of the
ventral surface of the mass (i.e. at point X on PI. 72, fig. 2).

Three, or possibly four, distinctly shaped types of hooklet can be recognized in these
contents, types A, B, C, and D of text-fig. 2. Type A is relatively short straight spinose
form with a strongly bifid base, rather like an odontaspid shark’s tooth in shape. Type B
is longer and more slender than type A, with a less pronounced base and a gentle curve
along its length. Type C is broader bladed than types A and B, has distinct lateral
flattening, and a strong, nearly 90°, hook. The base of type C is much less pronounced
than on types A or B, but this character may be suppressed due to lateral flattening.
Type D of text-fig. 2 is extremely rare in the gastric contents, about 1 mm. or less in
size, and a specimen from a different horizon and locality is figured here for comparative
purposes, the significance of which will be discussed later. Each of these hooklet types

text-fig. 2. Cephalopod hooklets from the
Lias. Types A, B, and C are all drawn from
hooklets in the gastric contents shown on Plate
73, figs. 1 and 2, while type D is drawn from
specimen OUM. J. 14800, in mudstone from the
Upper Lias at Dumbleton, Gloucestershire.

EXPLANATION OF PLATE 72

Fig. 1. Dorsal view of prepared ichthyosaur skeleton preserved with gastric contents. Lower Lias,
Planorbis Zone, Lyme Regis, Dorset. SF.l. Geology Dept. Collections, University of Manchester.
a., articular; v., vertebra; rib; sea., scapula; pd., paddle; pub., pubis.

Fig. 2. Ventral view of ichthyosaur specimen SF.l. Symbol ‘X’ indicates the region of the ventral
surface of the gastric mass with a concentration of quartz grains, d., dentary; sa, supra-articular;
icl., interclavicle; cor., coracoid; hum., humerus.

Palaeontology, Vol. 11

PLATE 72

2

POLLARD, Ichthyosaur gastric contents

J. E. POLLARD: GASTRIC CONTENTS OF AN ICHTHYOSAUR 379

in the gastric contents shows considerable size range, for instance in terms of length,
type A varies from 1-0 to 1-90 mm. (10 measured), type B from 0-70 to 3-00 mm. (16
measured), and type C from 0-9 to 2-90 mm. (22 measured). Types B and C appear to
be commoner than type A in the mass.

All these hooklets seem to be composed of a jet black and very brittle organic,
possibly chitinous, material. They are all hollow although sometimes filled with crystal-
line calcite. Due to their brittle nature and hollow centre, most of the hooklets are
cracked and partially crushed and splinter if any attempt is made to separate them from
the mass.

In an attempt to determine the approximate number of hooklets on the dorsal surface
of the gastric mass, the distribution of the hooklets in an area 2 cm. square was plotted
from the enlarged photograph, Plate 73, fig. 2. The frequency of hooklets on this surface
varied between 450 and 540 per sq. cm., with a mean of about 500 per sq. cm. This
number represents only those hooklets that could be clearly identified and is, therefore,
a minimal estimate. The total number of hooklets on the dorsal surface of the gastric
mass, area 90±5 sq. cm., is about 45,000±7,000 (i.e. 90±5 X 500±50).

It has proved very difficult to estimate the total number of hooklets in the gastric
mass due to their being crushed and randomly orientated. The cross-sectional diameter
of a number of hooklets of various sizes, uncrushed on the dorsal surface, varied from
0-20 to 0-50 mm. The mean depth of the gastric mass is 0-33 cm., so that allowing for
parallel packing and no crushing, the hooklets would be from approximately 16 (3*30/0*2)
to 7 (3-30/0-5) layers deep. Making an allowance for crushing and random packing,
from 6 to 14 or 10±4 layers deep, would seem to be a reasonable estimate. Therefore,
the total number of hooklets in the gastric mass is 45,000^7,000 x 10±4 = 478,000^
250,000 or 478,000±53 per cent. Such a large error is unavoidable in such approximate
calculations, but the figure gives some idea of the correct order of magnitude.

DISCUSSION

In order to understand the signifiance of the gastric contents previously described,
and the precise nature of the dibranchiate remains they contain, a search has been made
in the literature and other specimens have been examined in several British museums.
The author does not intend this as an exhaustive treatment of the subject, but more as
a spur to examination and comment by other workers.

Other Liassic ichthyosaurs with gastric contents. Ichthyosaur remains with associated
gastric contents preserved have been known for more than a hundred years from the
Liassic shales of Lyme Regis and Whitby in England, and Holzmaden in Germany.
Buckland in the Bridgewater Treatise (1836) is among the earliest English records. He
described and figured (pi. 13 and 14) two ichthyosaur specimens from Lyme Regis that
contained a coprolite mass with fish scales, preserved within the abdominal cavity.
These specimens are in the collections of the Oxford University Museum and will be
discussed later in this paper.

Probably the earliest description of the preserved cephalopod hooklets associated
with ichthyosaur bones is that of Coles (1853). He describes a layer of carbonaceous
material made up of ‘ minute black points ’ — hollow and filled with calcite, that was found

380

PALAEONTOLOGY, VOLUME 11

adhering to an ichthyosaur vertebra from the Lias of the Tewkesbury district. His
excellent figures (pi. 5, figs. 2 and 13) show shape, size, and crack patterns identical to
the hooklets described and figured in this paper (PI. 73, fig. 2 and text-fig. 2). This
material was wrongly identified by Coles (1853, p. 81) as ‘setiform or bristly scales’ of
the ichthyosaur integument, and was reported by him to be known associated with
ichthyosaur skeletons from the Lias of Lyme Regis and Ilminster as well as Tewkesbury.
Cole’s error was corrected by Moore (1856), who reported finding stomach contents
composed of cephalopod arm hooklets in sixteen out of twenty-three Liassic ichthyosaur
skeletons he had prepared for his museum. Moore examined the gastric contents further,
and suggested that they consisted of the desiccated ink and arm hooklets of naked
Jurassic cuttle-fish allied to Onyehoteutliis.

Buckman (1879), when describing a new species of fossil dibranchiate Belemnoteuthis
montefiorei from the Lower Lias of Charmouth, mentions the frequent occurrence of
ichthyosaur stomach contents and coprolites full of cephalopod arm hooklets. Similar
general statements recording gastric contents composed largely of cephalopod hooklets
have been made by several workers studying ichthyosaurs from the Holzmaden Lias
(Seeley 1880, Branca 1908, Drevermann 1914, Huene 1922, Hofmann 1958, and Augusta
1964). Wurstemburger (1876) described a Holzmaden specimen of Stenopterygius
quadriscissus, with head 50 cm. long, vertebral column 240 cm. long, where a large
stomach mass of fish and cephalopod remains was found only 20 cm. behind the head.
This unusually anterior thoracic position of the stomach is very similar to that of the
specimen described here. Williston (1914, p. 123) refers to an ichthyosaur skeleton in
the Stuttgart Museum that has preserved in its stomach contents the remains of more
than 200 belemnites. Dr. K. D. Adam (pers. comm.) informs me that no such specimen
exists in the Stuttgart Museum, but Williston’s comment is probably a mistaken reference
to a well preserved specimen of the shark Hybodus from the Upper Lias of Holzmaden
described by Brown (1900) and later Shimanskiy (1949). The gastric contents of this
shark contain over 250 belemnite rostra.

Many British museums possess in their collections ichthyosaur skeletons with well
preserved gastric contents. On other specimens the gastric contents have obviously been
cleaned off in the preparation of the skeleton, and so it would appear that these contents
are of much commoner occurrence than the literature would suggest. The ichthyosaurs
figured by Buckland (1836, pis. 13 and 14) are preserved in the Oxford University
Museum, numbered specimens J. 13587 and J. 13593 respectively. Re-examination of
these specimens by the author confirms that Buckland’s figures and descriptions are
extremely accurate and that the gastric contents consist largely of scales and spines of
the Liassic fish Pholidophorus sp., set in a matrix of a pale buff coprolitic clay. The
larger specimen J.13593 (Buckland 1836, pi. 14) does have a very sparse scattering of
type C hooklets over the whole dorsal surface of the gastric mass. Three other specimens

EXPLANATION OF PLATE 73

Fig. 1 . Dorsal view of the gastric mass and associated bones of ichthyosaur specimen SF. 1 (compare
with PI. 72, fig. 1). Scale of 1 cm.

Fig. 2. Magnified view of part of the dorsal surface of the gastric mass of specimen SF. 1 showing the
various types of cephalopod hooklets present. (This field of view may be orientated on PI. 73, fig. 1,
by the arcuate row of five large quartz grains in the lower half of the picture.) Scale of 1 cm.

Palaeontology, Vol. 11

PLATE 73

POLLARD, Ichthyosaur gastric contents

J. E. POLLARD: GASTRIC CONTENTS OF AN ICHTHYOSAUR 381

of Liassic ichthyosaurs from Lyme Regis in Oxford University Museum, J. 12125,
J. 13592, and J. 10348, all contain patches of gastric material, composed of types A, B,
and C hooklets, in the thoracic or anterior abdominal regions. The gastric mass in each
case is identical to the specimen described here, just hooklets without any matrix of the
coprolitic clay seen in Buckland’s specimens.

Specimens of various species of ichthyosaur from the Lower Lias at Lyme Regis on
display in the public galleries of the British Museum (Natural History) show gastric
contents of densely packed hooklets devoid of matrix (e.g. BMNH 36256, BMNH R1614,
BMNH R1072, BMNH 38523, BMNH 43006, and BMNH R1896). In the Manchester
Museum an excellent specimen from the Upper Lias at Whitby has a large gastric
mass containing A, B, and C type hooklets, just posterior to the pectoral girdle.

The conclusion to be derived from a study of these listed, and other specimens, is
that gastric contents of densely packed dibranchiate cephalopod hooklets are much
commoner in prepared specimens than the fish remains in a matrix of coprolite clay
described by Buckland. The gastric contents of many Jurassic plesiosaurs are also
known to be composed largely of dibranchiate hooklets (Juravlev 1943, Hekker and
Hekker 1955, and Tarlo 1959), but here large gastroliths usually occur as well (Seeley
1877, Brown 1904, and Williston 1904). Gastroliths have not been found preserved in
ichthyosaur stomach contents.

Contents of coprolites from Lower Lias. Well-preserved coprolites, usually assigned to
ichthyosaurs or plesiosaurs, have been known from the Lower Lias at Lyme Regis since
before Buckland's classic paper of 1835. Most of these coprolites are assumed to have
been formed by ichthyosaurs, on account of their similarity in composition to material
described by Buckland (1836) from within the ichthyosaur abdominal cavity. Other
workers (Fraas 1891 and Woodward 1917) have questioned the assignment of these
coprolites to ichthyosaurs. They argue that spirally folded coprolites are rarely found
associated with ichthyosaur skeletons and are more likely to have been formed by the
spiral intestine of Liassic sharks than by the typical reptilian intestines which the
ichthyosaurs probably possessed. The following analysis shows that the majority of
Liassic coprolites do not have well-formed spiral folds but have faunal, lithological
and chemical features identical to the ichthyosaur gastric contents described by Buck-
land. Moreover, the hybodont sharks, suggested by Woodward (1917) as probable
producers of the coprolites, are believed on account of their dentition (Romer 1966,
p. 40) to have been benthonic or necto-benthonic scavenger feeders, and not nectonic
fish feeders as were the producers of the Liassic coprolites.

Buckland (1835) showed that Lower Lias coprolites contain fish remains, bones of
young ichthyosaurs, and possibly the sucker rings of fossil cuttle-fish. He did not observe
or describe any dibranchiate hooklets. The matrix of these coprolites, which I have
called lbuff coprolitic clay’, was shown by Buckland to be phosphatic material derived
from the digestion of fish and reptile bones. Buckland suggested that the strong spiral
involutions frequently seen on coprolites indicate that the small intestine of the ichthyo-
saur was ribbon like and twisted into a spiral. Firtion (1938) analysed the contents of
coprolites from the Lower Lias of Alsace. He found that the undigested contents were
mainly crinoids, gastropods, or pelecypods with less abundant foraminifera, ostracods,
fish remains, and brachiopod shells. These coprolites rarely had spiral folds and were

382

PALAEONTOLOGY, VOLUME 11

obviously formed by a benthonic feeding animal. However, he assigned them to ichthyo-
saurs, although stomach contents of the above composition are unknown in ichthyo-
saurs. Such coprolites could as easily belong to teleosaurs or plesiosaurs (Drevermann
1914). Ager (1963, p. 120) mentions that a study of coprolites suggests that Mesozoic
ichthyosaurs included belemnites in their diet, but he does not refer to any actual
records of this fact.

Fifty well-preserved coprolites from Buckland’s collection in the Oxford University
Museum have been examined by the author. These specimens vary from 1 to 6 in. in
length, mainly 2-3 in. long, and their contents can be examined on the cleaned surface,
or internally where they have been sectioned and polished. Forty-five specimens contain
recognizable fish remains, mainly scales, fin rays, and spines of Pholidophorus sp., less
commonly remains of Dapedium sp. and Lepidotus sp. Two specimens, those figured by
Buckland (1835, pi. 29, figs. 2, 3, 4, and 5), contain reptilian bones and fifteen specimens
have well-formed spiral involutions. None of these coprolites contain visible remains of
dibranchiate hooklets and the possible sucker rings figured by Buckland (1835, pi. 30,
figs. 1 , 2, and 3) are considered to be transverse sections of fin rays and small vertebrae
of fish. Examination of sixteen well-preserved coprolites in Manchester Museum and
Geology Department, University of Manchester, shows that all these contain fish scales
and spines, none contain reptilian bones, and only four specimens have well-formed
spiral folds. One of the Manchester Museum specimens has a small patch of shale
matrix with type B and C hooklets adhering to its surface, but they are not contained
in the adjacent coprolite material. Lydekker (1889, pp. 1 1 4—17) lists sixty-six coprolites
from the Lower Lias in the British Museum collections, but only mentions fifteen con-
taining fish scales, only two with reptilian bones, and only one showing well-formed
vascular impressions.

From this survey it is suggested that ichthyosaurs from the British Lower Lias
primarily defecated fish remains in their coprolites. Undigested cephalopod remains
have not been seen in these coprolites, so it may be inferred that they accumulated in
the stomach, as their predominance in gastric contents would suggest. The commonest
form of fish eaten, Pholidophorus , is presumed to have been a nectonic or necto-
benthonic form, not a deep bodied benthonic fish like Dapedium or Lepidotus. The
possible significance of this latter observation will be discussed later.

Nature of the cephalopod remains. Throughout the preceding part of this paper the
hooklets found in the gastric contents of the ichthyosaur have been broadly described
as belonging to Liassic dibranchiate cephalopods. It is of some importance to establish
the precise nature of the dibranch iates possessing these hooklets before discussing their
relationship to their ichthyosaur predators.

As well as occurring in gastric contents of ichthyosaurs, these hooklets are known
preserved in their life position on the arms of predominently soft bodied dibranchiates
that are rarely found in Liassic and other Jurassic argillaceous sediments. Pearce (1842)
named one of these soft-bodied dibranchiates with arm hooklets from the Oxford Clay
as Belemnoteuthis. The arm hooklets of Belemnoteuthis were figured by Owen (1844,
pi. 6, fig. 2) and Mantell (1852, fig. 4) as all possessing elongated pointed bases, and one
form similar in shape to type D of text-fig. 2.

It has already been mentioned that Coles (1853) seems to have been the first person

J. E. POLLARD: GASTRIC CONTENTS OF AN ICHTHYOSAUR 383

to describe the hooklets from the Lias, although he was in error about their nature. The
suggestion of Moore (1856), that these hooklets belonged to naked dibranchiates allied
to Onychoteuthis, does not seem to be supported by any earlier published records of
specimens of this genus from the Liassic rocks of Britain. Huxley (1864) was the first
person to unquestionably associate these forms with belemnoid arm hooklets. He
figured (1864, pi. 1, fig. 5a) forms identical with the types A, B, and C of text-fig. 2
when he described two specimens of belemnites from the Lias (BMNH 74106 and BMNH
39855), where soft parts were preserved in association with the guard, phragmocone,
and proostracum. One of these specimens, BMNH 74106, must be interpreted with some
caution as it has been restored in preparation. Dr. K. A. Joysey (pers. comm.) has
informed me that one such specimen (J42835) in the Sedgwick Museum, Cambridge,
is a ‘well intended’ forgery by a preparator, for the belemnite guard has been artificially
shaped to improve its appearance before being set in an artificial matrix in association
with a genuine group of hooks. However, Huxley’s interpretation appears to be correct
and has been accepted by such later workers as Crick (1907), Naef (1922), and Jeletsky
(1966).

Buckman (1879) described a specimen of a head of hooked arms from the Lower Lias,
which he named Belenuioteuthis montefiorei. This specimen is in the collection of the
British Museum (BMNH C5026) and the hooklets are identical with the types A, B, and C
described here. However, in affinity this specimen seems to be Belemnites as suggested
by Crick (1902) and not the Belenmoteuthis of Pearce (1842).

The most complete study of the arms of Liassic and other Jurassic dibranchiates is
undoubtedly that of Crick (1907). He examined seventeen specimens of ‘belemnite’
arms in the British Museum collections and described and figured six of these specimens
in detail in his paper. Crick concluded that the belemnites possessed six arms bearing
rows of hooklets with swollen bases, as types A, B, and C, while Belenmoteuthis (= Acan-
thoteuthis) had eight or ten arms bearing hooklets with pointed bases. This latter form
of hooklet is characteristic of fossil dibranchiates known from the Upper Jurassic
Oxford and Kimmeridge Clays in Britain, and the lithographic stone of Solenhofen in
Germany (Pearce 1842, Owen 1844, Mantell 1852, Crick 1897 and 1907).

Several of the standard textbooks on palaeontology (Zittel 1913, Woods 1946, and
Piveteau 1953) figure dibranchiate cephalopod hooklets from Mesozoic sediments but
give little idea of possible affinities of the various forms. Naef (1922) in his authoritative
work on fossil dibranchiates figures and discusses various forms, including types from
the Upper Lias of Holzmaden similar to types B and C of this paper, but is uncertain
of any definite correlation of hooklet form with type of dibranchiate. Both Naef (1922,
p. 219) and Jeletsky (1966, p. 138) disagree with Crick’s (1907) interpretation of six-
arm belemnites and suggest that they had eight or ten arms in common with the belem-
noteuthids. Jeletsky (1966, p. 138) believes that all members of the order Belemnitida
possessed arm hooklets.

From a detailed study of the literature and museum specimens it is here suggested
that some broad association of hooklet form with three major groups of Mesozoic
dibranchiate cephalopod may be possible. Members of the Belemnitidae may have had
hooklet types A, B, and C as described here, characterized by a gentle curved shape and
a swollen bifid base. Such forms are known mainly from the Lias (Huxley 1864, Crick
1907, Naef 1922, and Jeletsky 1966). Dibranchiates of the family Belemnoteuthidae may

384

PALAEONTOLOGY, VOLUME 11

have had gently curved or recurved hooklets with an elongate obliquely pointed base.
This hooklet type is known from Upper Jurassic specimens (Pearce 1842, Owen 1844,
Mantell 1852, Crick 1897, 1907, Hekker and Hekker 1955). Genera of the order
Phragmoteuthida probably had hooklets of a belemnitid type (Jeletsky 1966, p. 31).
Separated hooklets of a variety of shapes are frequently found in microfaunas of Jurassic
and Cretaceous age and are known as ‘ Onychites' sp. (Quenstedt 1885, Naef 1922,
Piveteau 1953, Hekker and Hekker 1955). Such hooklets probably belonged to other
little-known members of the order Belemnitida as fossil teuthoid squids and sepioid
cuttle fish seem to have been devoid of arm hooklets in Mesozoic seas (Jeletsky 1966).
Therefore, the types of hooklets described in the earlier part of this paper would seem
to have belonged to dibranchiate cephalopods of the family Belemnitidae, and not the
family Belemnoteuthidae or the order Phragmoteuthida.

TABLE 1

Aims and arm hooklets of fossil belemnoids from the Lower Lias. Details from Woods (1946, fig. 169)

and Crick (1907, pi. 23, figs. 1, 3, and 5).

Number Total no. Hooklets Number of hooklets per arm of

of of per arm lengths

Specimens

arms

hooklets

mean

max.

> 40 mm.

> 30 mm.

> 20 mm.

SM. J37812

8

211

29

34

32 (5)

0

17 (3)

BM. C3007

6

143

24

28

26(3)

21 (3)

0

BM. 82895

4

123

31

36

31 (4)

0

0

BM. 47020

6

156

25

30

26 (2)

26 (2)

24 (2)

Approx, ‘mean’

6

c. 150

27

32

29

23

20

Examination of published figures and museum specimens of ‘belemnite’ arms with
hooklets (e.g. Woods 1946, fig. 169, and Crick 1907, pi. 23, figs. 1 to 6) has enabled
observations to be made regarding the number of hooklets per arm, the total number
of hooklets per belemnite individual, and the arrangement of hooklet types along
the arms.

The varied state of preservation of specimens with arms from the Lower Lias makes
detailed analysis very difficult. Crick (1907) showed that of the seventeen specimens in
the British Museum collections only six were worthy of description, and of these six
only three are considered sufficiently complete by the present author for detailed analysis
(Table 1). Crick (op. cit.) showed that the belemnite arms varied in length, and that the
hooklets were arranged in two parallel rows on the inner surface of the arms. In many
of the known specimens the arms are either incompletely preserved or superimposed,
so that it is impossible to be certain of the original arrangement of the arm hooklets.
The specimens listed in Table 1, although varying in the number of arms, all possess
arms that appear to be complete, as the hooklets are arranged in parallel rows, and are
largest in the mid-length of each arm, gradually diminishing in size towards each end
(Crick 1907, p. 271). From Table 1 it appears that the number of hooklets per arm
varies with the length of the arm, but about thirty (fifteen pairs) hooklets per arm is
an average number. Specimen SM. J37812 possesses eight distinct arms, and therefore
confirms that there must have been eight or more arms in the belemnites in agreement
with Naef (1922) and Jeletsky (1966). There must have been, therefore, at least 300
hooklets on an individual ten-armed belemnite in Liassic times.

J. E. POLLARD: GASTRIC CONTENTS OP AN ICHTHYOSAUR 385

The detailed arrangement of the hooklets along the arms can only really be seen
on two of the specimens included in Table 1, Wood’s fig. 169 (SM. J37812) and
BMNH 82895 (Crick 1907, pi. 23, fig. 3). On these specimens the proximal hooklets
are seen to be small examples of types A and B. The hooklets at mid-length of the
arms are large examples of types B and C which decrease in size towards the distal
end. This broad pattern of hooklet arrangement can be seen to a lesser extent on other
less well preserved specimens (e.g. Crick 1907, pi. 23).

Significance of the gastric contents
(a) Feeding habits

Diet. The evidence of the gastric contents and coprolites suggests that Liassic ichthyo-
saurs fed mainly on fish and/or dibranchiate cephalopods. Among fish-eaters stomach
contents were rarely preserved while coprolites are commonly found. A reverse situation
possibly exists regarding dibranchiate eating forms, stomach contents being commonly
preserved but coprolites rarely. Both these diets suggest that ichthyosaurs in Liassic
seas were predators on nectonic not benthonic animals. If at least two distinct dietary
habits were established amongst Liassic ichthyosaurs, they could be explained by either
selective predation, or feeding at different levels in the sea as in the sperm whale (Clarke
1962, p. 186). Drevermann (1914, p. 42) has suggested that in Upper Jurassic seas,
virtually toothless ichthyosaurs like Ophthahnosaurus may have fed exclusively on
naked cuttle fish. The suggestion put forward in an earlier section of this paper that
some Liassic ichthyosaurs fed mainly on belemnoids is made with some reservation, as
the known fossil belemnoids from the Lias where arm hooklets and hard parts occur in
association are few and none too well preserved. The similarity of the ichthyosaurs in
mode of life and diet to odontocete Cetacea, especially the sperm whale, has been sug-
gested by several workers (e.g. Buckland 1835, p. 227, Kukenthal 1892, Branca 1908,
and Wiman 1946).

Volume of food eaten. In the previous sections it has been shown that the gastric contents
contain 478,000±250,000 hooklets, and that each belemnite probably had about 300
arm hooklets. The gastric contents described here, therefore, could represent between
760 and 2,430 digested individual belemnites. The undigested organic hard parts of
these belemnites accumulated in the stomach, much in the same way as arm hooklets
and beaks of modern dibranchiates accumulate in the stomach of the sperm whale.
Akimuskin (1955) records 28,000 squid beaks, representing 14,000 squids, and Clarke
(1962) records 4,000 beaks plus 28 undigested squids, representing 2,160 squids, found
in the stomachs of sperm whales caught in the North Pacific and Atlantic respectively.
The length of time of accumulation of these belemnoid hooklets in the ichthyosaur
stomach is impossible to ascertain. It could represent several meals or a lifetime’s
accumulation. Judging from the length of skull and relative size of the various bones it
is probable that the ichthyosaur described was only a young specimen. In this particular
specimen, therefore, the gastric contents might represent a lifetime’s accumulation.

Mode of feeding. Several workers (i.e. Buckland 1836 and Seeley 1880) have suggested
that the ichthyosaurs were voracious feeders, where the prey was swallowed whole
without mastication, as in the sperm whale (Clarke 1956 and Clarke 1962). In the

386

PALAEONTOLOGY, VOLUME 11

teuthophagous whales the presence or absence of teeth makes very little difference to the
efficiency of feeding (Clarke 1956). If a parallel situation existed amongst dibranchiate-
eating Jurassic ichthyosaurs, the presence or absence of teeth could not be explained
purely in terms of diet as suggested by Drevermann (1914). The sperm whale frequently
has facial scars from squid tentacles (Clarke 1956), suggesting that the squids were
swallowed head first. If belemnoids were eaten in this manner by Liassic ichthyosaurs,
biting of the heads would cause separation of the hooked arms from the body with hard
parts. Further significance of this comment is discussed below.

(b) Digestive mechanism

The ichthyosaurs appear to have swallowed their prey whole into a large expandable
stomach where all the digestive breakdown took place (Buckland 1836). Chyme, including
softened fish and reptile bones, passed into the intestine, and then the indigestible mater-
ial was defecated as coprolites. In outline this digestive process is similar to that of the
sperm whale (M. R. Clarke, pers. comm.). Undigested dibranchiate remains, hooklets,
and possibly beaks accumulated in the stomach as they could not be passed on. The
reasons for this accumulation are difficult to understand. Possibly the process was a
defence mechanism on the part of the ichthyosaur to protect the delicate tissues of the
posterior part of the digestive system from damage by the sharp undigested hooklets.
The gastric contents could represent a gravity accumulation of indigestible material on
the ventral side of the stomach in a very fluid chyme, produced by the digestive break-
down of cephalopod tissue. Such an explanation would account for the pockets of
quartz grains found on the ventral side of the gastric mass (PI. 72, fig. 2). A very fluid
chyme might not have been able to transport the undigested matter through the pyrolic
valve, as undoubtedly occurred with the viscous chyme produced from the digested fish
remains, now preserved as coprolites. A third possibility is that the hooklets gripped in
the ventral stomach wall and formed an interlocking network, trapping the quartz
grains. This last suggestion could explain why the gastric contents of this specimen were
not dispersed before burial, despite the slight dismembering of the skeleton.

The accumulation of such remains raises a number of problems for the ichthyosaur
which can only be answered by further analogies with the sperm whale. Did these
remains accumulate to the detriment of the animal, perhaps blocking the digestive tract
and causing death? In the sperm whale the accumulated squid beaks are periodically
vomited from the stomach (M. R. Clarke, pers. comm.). Ambergris frequently contains
squid beaks and so may also aid this regurgatory process. Such mechanisms as these
may have existed in the ichthyosaurs to clear inconvenient accumulations of hooklets
from the stomach.

If Liassic ichthyosaurs frequently ate belemnoid dibranchiates, as has been suggested,
another problem arises concerning the digestion of the crystalline calcite guards.
Modern teuthophagous cetaceans have little difficulty in digestion of the conchiolinic or
weakly calcified ‘pens’ of squids by solution in the stomach, but densely crystalline
guards of the belemnites are much more difficult to destroy, as witnessed by their
abundance in Mesozoic clastic sediments. As gastric contents with belemnoid guards
are rare, or unknown, in ichthyosaurs three possible explanations are suggested:

1. Heads were bitten off and so guards were not swallowed. This seems an unlikely
mechanism as in reptiles and modern teuthophagous cetaceans prey is swallowed whole.

J. E. POLLARD: GASTRIC CONTENTS OF AN ICHTHYOSAUR

387

2. The body with guard, phragmocone, and proostracum separated from the head in
the stomach and was regurgitated. The habits of the sperm whale give some support for
this suggestion as regurgitation of squids frequently occurs on capture of the whales,
and in the digestive process in the stomach the bodies and heads of the squids separate
at an early stage (Clarke 1956 and Clarke 1962).

3. The guards were dissolved, or broken up by gastroliths and then dissolved, by the
chemical environment of the stomach. Difficulties with this possibility are that gastro-
liths are unknown in ichthyosaurs, and a very acid stomach environment would be
necessary to have dissolved dense primary crystalline calcite.

However, of these suggestions processes 2 and 3 would seem to be the more likely.
Other possible explanations may simply be that the ichthyosaurs, as modern cetaceans,
primarily ate naked dibranchiates, or chemical solution of guard posed no problem for
the digestive mechanism of the Liassic ichthyosaur.

Acknowledgements. I thank Dr. F. M. Broadhurst for help in extraction of the specimen and critically
reading the manuscript, Dr. K. D. Adam, Dr. K. A. Joysey, and Dr. M. R. Clarke for offering many
helpful suggestions, and Mr. P. A. Washbourne for help with photography. Mr. J. M. Edmunds and
Mr. P. Powell of the Oxford University Museum gave access to Buckland’s original material; the
Assistant Keeper and Dr. M. K. Howarth of the British Museum (Natural History) and Dr. R. M. C.
Eager of the Manchester Museum allowed me to examine specimens in their collections.

REFERENCES

ager, d. v. 1963. Principles of Palaeoecology. New York. 371 pp.

Akimushkin, i. i. 1955. [The feeding of the Cachalot.] Dokl. Akad. Nauk S.S.S.R. 101, 1139-40.
(In Russian.)

AUGUSTA, j. 1964. Prehistoric Sea Monsters. London. 65 pp., 21 pi.

branca, w. 1908. Nachtrag zur Embryonenfragen bei Ichthyosaurus. Sber. preuss. Akad. IT7vv(1908), 1 ,
392-6.

brown, b. 1904. Stomach stones and food of plesiosaurs. Science, 20, 184-5.

brown, c. 1900. Uber das Genus Hybodus und seine systematische Stellung. Palaeontographica, 46,
pp. 163 ff.

buckland, w. 1835. On the discovery of Coprolites, or fossil faeces, in the Lias at Lyme Regis, and in
other formations. Trans, geol. Soc. Load. (2), 3, 223-38, pis. 28-31.

1836. The Bridgewater Treatises', Geology and Mineralogy. London. 2 vols.

buckman, j. 1879. On the Belemnoteuthis montefiorei. Proc. Dorset nat. Hist, antiq. Fid. Club. 3, 141-3,

pi. 1.

clarke, m. r. 1962. Stomach contents of a Sperm Whale caught off Madeira in 1959. Norsk. Hval-
f angst t id. 5, 173-91.

clarke, r. 1956. Sperm Whales of the Azores. ‘ Discovery ’ Rep. 28, 237-98.
coles, h. 1853. On the skin of Ichthyosaurus. Q. JI. geol. Soc. Lond. 9, 79-82, pi. 5.
crick, g. c. 1897. Acanthoteuthis speciosa Munster. Geol. Mag. (4), 4, 1-4, pi. 1.

1902. Note on the type specimen of Belemnoteuthis montefiorei. Proc. malac. Soc. Lond. 5, 13-16,

pi. 1.

1907. On the arms of the belemnite. Proc. malac. Soc. Lond. 7, 269-79, pi. 23.

drevermann, f. i. 1914. Die Meersaurier im Senkenberg Museum. Senkenberg Naturforsch. Ges. Ber.
45, 35-48, 12 figs.

firtion, f. 1938. Coprolithes du Lias Lnferieur d’Alsace et de Lorraine. Bull. Serv. Carte, geol. Als.
Lorr. 5, 27-43, pis. 4-8.

fraas, e. 1891. Die Ichthyosaurier der sud-deutschen Trias und jura-ablagerunge. Tubingen.
hekker, e. l., and hekker, r. f. 1955. Ostatki Teuthoidea iz verkhnei yuryi nizhnego mela Povolzhya.
Vop. Paleont. 2, 36-44, pis. 1 and 2. (In Russian.)

388 PALAEONTOLOGY, VOLUME 11

hofmann, j. 1958. Einbettung und Zerfall der Ichthyosaurier im Lias von Holzmaden. Meyniana , 6,
10-55. 30 figs.

huene, F. von. 1922. Die ichthyosaurier ties Lias und ihre Zusammenhange. Berlin. 114 pp., 22 pis.
huxley, t. h. 1864. On the structure of the Belemnitidae; with a description of a more complete
Belemnites than any hitherto known, and an account of a new genus of Belemnitidae Xiphoteuthis.
Mem. Geol. Surv. U.K. Monogr. II. 22 pp., 3 pis.

jeletsky, j. a. 1966. Comparative morphology, phylogeny and classification of fossil Coleoidea.

Paleont. Contr. Univ. Kans., Mollusca 7, 1-162, pis. 1-25.
juravlev, k. j. 1943. The remains of Upper Jurassic sea reptiles at the Seveljevra shale mine Acad.

Sci. U.S.S.R. Biol. 5, 293-306, 4 figs. (In Russian, English summary.)
kukenthal, w. 1892. Ichthyosaurier und Wale. Neues. Jb. Miner. Geol. Paleont. Beil Bd. 1, 161-6.
lydekker, R. 1889. Catalogue of fossil Reptilia and Amphibia in the British Museum. Part II. Ichthyo-
pterygia and Sauropterygia. London. 303 pp.

mantell, g. a. 1852. A few notes on the structure of the Belemnite. Ann. Mag. nat. Hist. (2), 10, 14-19.
moore, c. 1856. On the skin and food of ichthyosauri and teleosauri. Rep. Br. Ass. Advmt. Sci. (1856),
69.

naef, a. 1922. Die fossilen Tintenfische. Jena. 322 pp.

owen, r. 1844. A description of certain Belemnites preserved with a great proportion of their soft
parts in the Oxford Clay. Phil. Trans. R. Soc. 134, 65-85, pis. 2-8.

1881. The reptilia of the Liassic formations. Part III. Ichthyosaurus. Palaeontogr. Soc. {Mono-
graph). pp. 83-130, pis. 21-33.

pearce, c. 1842. On mouths of ammonites and fossils contained in laminated beds of the Oxford Clay,
discovered in cutting of G.W.R. near Christian Malford in Wiltshire. Proc. Geol. Soc. Loud. 3,
592-4.

piveteau, j. 1953. Trade de paleontologie. Vol. 2. Paris.
quenstedt, f. a. 1885. Handbuch der Petr efaktenkunde. Tubingen.
romer, a. s. 1966. Vertebrate Paleontology. Chicago. 468 pp.

seeley, h. g. 1877. Mauisaurus gardneri, an elasmosaurian from the base of the Gault at Folkestone.
Q. Jl. Geol. Soc. Land. 33, 541-7, pi. 23.

1880. Report on the mode of reproduction of certain species of ichthyosaurs from the Lias of

England and Wurtemburg. Rep. Br. Advmt. Sci. (1880), 68-76, pi. 1.
shimanskiy, v. n. 1949. O sistematicheskom polozhenii rinkolitov. Trudy Paleont. Inst. 20, 199-208.
tarlo, l. b. 1959. Pliosaurus brachyspondylus from the Kimmeridge Clay. Palaeontology, 1, 283-91,
pis. 51, 52.

williston, s. w. 1904. The stomach stones of the plesiosaurs. Science, 20, 565.

1914. Water reptiles of past and present. Chicago. 251 pp.

wiman, c. 1946. Uber Ichthyosaurier und Wale. Senckenbergiana, 27, 1-11, pis. 1-3.
woods, h. 1946. Palaeontology. Invertebrate. 8th edn. Cambridge. 477 pp.

woodward, a. s. 1917. The so-called coprolites of ichthyosaurs and labrinthodonts. Geol. Mag. (6). 4,
540-2, pi. 34.

woodward, h. b., and ussher, w. a. e. 1911. The geology of the country near Sidmouth and Lyme
Regis. Mem. geol. Surv. U.K.

wurstemberger, A. R. c. von. 1876. Ueber Lias Epsilon. Jh. Ver. vaterl. Naturk. Wurtt. 32, 193-358,
pi. 4, figs. 1-16.

zittel, k. a. von. 1913. Textbook of Palaeontology. 2nd English edn. London.

JOHN E. POLLARD
Department of Geology
The University
Manchester 13

Typescript received 19 December 1966

MANTLE CANAL PATTERNS IN
SCHIZOPHO RIA (BRACHIOPODA) FROM
THE LOWER CARBONIFEROUS OF
NEW SOUTH WALES

by JOHN ROBERTS

Abstract. The brachiopod Schizophoria verulamensis Cvancara, from the Lower Carboniferous of the Gresford
district, New South Wales, has distinctive mantle canal patterns in small, medium, and large specimens. The
different patterns are interpreted as stages in the ontogeny and are thought to result mainly from the growth and
enlargement of the genitalia.

Changes in the morphology of the canals in pedicle valves throughout six horizons in the Lower Carboniferous
sequence show a trend, probably genetically controlled, towards the earlier maturity of the genitalia. The super-
imposition of larger or mature mantle canal systems on smaller or immature systems indicates that the canals
are impressed by periodic resorption, and it is argued that this probably took place during the winter when
there was a stable mantle canal system and little shell deposition. In large specimens there are connexions between
the vascula genitalia, vascula media, and lateral canals in the pedicle valve, and between the vascula genitalia
and vascula myaria in the brachial valve. It is suggested that the connexions may have increased the amount of
oxygenated coelomic fluid circulating over the gonads, possibly causing them to ripen earlier.

Pathological variations in the mantle canals are described, and it is shown that the canals are able to adjust
to injuries or defects in the shell.

The mantle canals in living brachiopods are extensions of the body cavity into the
mantle. Their function is mainly respiratory and in articulates they also act as receptacles
for the gonads. In both living and fossil brachiopods the pattern of canals is frequently
impressed on the inner surface of the shell, particularly that of older specimens. In
fossil brachiopods impressions of the mantle canals are common in orthoids and
strophomenoids, are rarely seen in the spiriferoids and chonetoids, and only recently
have been observed in two species of productoids from the Lower Carboniferous of
the Bonaparte Gulf Basin, northwestern Australia. As a result their use in systematics
is limited.

Schuchert and Cooper (1932) described the morphology of the mantle canals in the
orthoids and used the canal patterns as a basis for the separation of a new subfamily.
Opik (1934) further demonstrated their use in his study of the Clitambonitidina, and
proposed a detailed terminology for the various parts of the mantle canal system.
Williams (1956) was the first to present a comprehensive review of the mantle canal
patterns in articulate brachiopods. He modified Opik’s terminology and showed that the
mantle canal patterns in articulates could be referred to standard types which were
probably derived from Cambrian forms. Williams (1956) and Williams and Rowell
(1965) illustrated many different types of mantle canal patterns.

The present study deals with the morphology, development, and probable genetic
changes in the mantle canals of the Lower Carboniferous orthoid Schizophoria verula-
mensis Cvancara (1958). Extremely well-preserved material has made possible the
recognition of a sequence of young, mature, and old stages in the mantle canal system
of 5. verulamensis. These stages have in turn led to the recognition of a trend towards an
earlier maturity of the reproductive system during the Lower Carboniferous, and a

[Palaeontology, Vol. XI, Part 3, 1968, pp. 389-405, pis. 74-75.]

390

PALAEONTOLOGY, VOLUME 11

method of impression or superimposition of the mantle canals which had not pre-
viously been recognised in the Brachiopoda. Because of the exceptionally fine preserva-
tion it has also been possible to recognize an intricate interconnexion between the
various parts of the mantle canal system, a feature not recognized in Williams’s and
Rowell’s treatment of mantle canal patterns.

Schizophoria verulamensis Cvancara is one of the most common brachiopods in the
Lower Carboniferous (Tournaisian and Visean) of New South Wales, and has been
collected from every known fossil horizon in the Gresford district, about 35 miles
north-west of Newcastle (text-fig. 1). The majority of the specimens described in this
paper are preserved in indurated grey siltstone; the rest occur in medium-grained lithic
sandstone. They are exceptionally well preserved, and have very finely detailed impres-
sions of the mantle canals and setal follicles. Internal moulds only are figured in the
plates because the impressions of the canals, represented by ridges, are particularly
prominent when illuminated by an oblique light source.

Schizophoria verulamensis is especially well represented in collections from the
following localities (with grid references): L. 53 Greenhills (46609790), L. 206 Lewins-

J. ROBERTS: MANTLE CANAL PATTERNS IN SCHIZOPHORIA

391

brook syncline (46279968), L. 270 Trevallyn (45729864), L. 218 Gresford Quarry
(45809912), L. 50 Gresford Quarry (45789913), L. 215 Lewinsbrook (46089882).

With the exception of L. 53 Greenhills, which is on the Paterson One-Mile Sheet, the
grid references are taken from the Dungog One-Mile Sheet. The stratigraphic positions

1 1 TlZ'one' [I_.l Sands'°n° |f Conglomerol, |

Crinoidol

limestone

Oolitic

limestone

text-fig. 2. Diagrammatic stratigraphic sections of the marine rocks at Trevallyn and in the Lewins-
brook Syncline.

of the horizons are shown in text-fig. 2 on diagrammatic stratigraphic sections of the
marine rocks in the north-western part of the Gresford district. All of the collections
are housed in the Geology Department, University of New England, Armidale, N.S.W.

The stratigraphy of the Gresford district has been described by Roberts (1961 and
1965c) and the faunas by Roberts (1963, 1964, and 1965n and b ), Campbell and Roberts
(1964), and Brown, Campbell, and Roberts (1965).

The specimens of Schizophoria verulamensis were collected from a single horizon at
each fossil locality except L. 53, where two horizons were sampled. An attempt was

392

PALAEONTOLOGY, VOLUME 11

made to collect randomly at each locality, and all the specimens collected were examined
in the laboratory. Width-frequency plots for pedicle and brachial valves are given in
text-fig. 3. The samples are characterized by the almost total absence of small specimens,
and the presence of a large number of disarticulated valves, features which suggest that
the populations had been affected by current action and sorting prior to burial.

20 30 40 50 20 30 40 50 60

WIDTH OF VALVES IN MM
• Pedicle valves * Brachial valves

text-fig. 3. Width-frequency polygons for Schizophoria verulamensis Cvancara specimens from six

localities in the Gresford district.

The post-mortem modification is also borne out by the unequal numbers of pedicle
and brachial valves and small number of articulated shells collected from each locality.
These figures are given in Table 1. The collection from L. 270 is the only sample having
a comparable number of pedicle and brachial valves and also contains the greatest
number of articulated shells, suggesting that it is perhaps the closest to a life assemblage.

Further evidence that most of the assemblages are death assemblages is provided by
the occurrence, at L. 218, L. 270, and L. 206, of shell beds containing large numbers of

J. ROBERTS: MANTLE CANAL PATTERNS IN SCHIZOPHORIA 393

mature specimens of Schizophoria to the virtual exclusion of smaller specimens and
other small species. The L. 53 horizon contains several other large species as well as
Schizophoria , but also has very few small species. Smaller species and polyzoan debris
accompany the Schizophoria specimens at L. 215 and L. 50.

TABLE 1

Disarticulated valves

Locality

Pedicle

valve

Brachial

valve

Articulated

shells

L. 53

25

6

—

L. 206

28

18

1

L. 270

31

30

10

L. 218

30

23

1

L. 50

25

20

5

L. 215

54

31

7

MANTLE CANALS

The mantle canals in living brachiopods are situated in the mesenchyme layer of the
mantle and are lined with ciliated epithelium. They are filled with coelomic fluid, which
is rapidly circulated by the ciliated epithelium (Hyman 1959, p. 593), the circulation in
some species being facilitated by a median ridge dividing the in- and out-flowing currents.
The coelomic fluid contains free coelomocytes, including a spherical type which has a
respiratory function and may be equivalent to a red blood corpuscle (Hyman 1959,
p. 557). The pigment in the ‘respiratory cells’ absorbs oxygen and on reduction releases
it to oxygenate the tissues.

The pattern of the mantle canals varies in different brachiopod groups (Williams
1956), and there is usually a different pattern in the pedicle and brachial valves. In most
cases the canals originate as wide branches from the body cavity; some of these divide
repeatedly until they extend to the proximal ends of the setal follicles around the edge
of the mantle; other form sac-like receptacles and act as gonocoeles.

Most living brachiopods have a complicated system of mantle canals extending
throughout the mantle, providing a large surface area through which oxygenation of the
coelomic fluid can take place. An adaptation to increase the surface area of the mantle
canals is illustrated by Williams and Rowell (1965, fig. 24) in which the mantle canals in
the inarticulate Glottidia subtend gill ampullae into the mantle cavity.

In some fossil groups, however, the mantle canals have a relatively small surface area,
and respiration may have been largely carried out by the lophophore. The lophophore
in modern articulates has a plentiful supply of coelomic fluid provided by the coelomic
canal and the small brachial canal (Williams and Rowell 1965, fig. 30). The small
brachial canal gives off a branch to each of the filaments on the lophophore which have
a large surface area and are ideally suited to respiration. Spiriferoids, for example, had
narrow pinnate mantle canals with a small surface area combined with a large spiral
lophophore. The lophophore probably had a large surface area suitable for oxygen
exchange, was bathed in a stream of fresh water brought in by the moving filaments,
and was an ideal organ for respiration as well as food gathering. Mantle canals appear
to be poorly developed in the productoids, which probably also had a spiral lophophore

d d

C 5586

394

PALAEONTOLOGY, VOLUME 11

(Williams 1956, fig. 5 (6); Brunton 1966, figs. 8 and 9), and the lophophore may have
acted as the main respiratory organ. Chonetoids occasionally exhibit mantle canals
(Muir-Wood 1962, pi. 6, fig. 7), but they too probably used a spiral lophophore for
respiration as well as for food gathering.

The mantle canals are also the receptacles for the gonads in living articulate brachio-
pods, and so the gonads are continuously bathed in oxygenated coelomic fluid.

Blood vessels, which are distinct from the mantle canals, extend from the body
cavity into the mantle canals and run along the margins nearest the mantle cavity
(i.e. along the inner margins of the mantle canals). These vessels, which are only seen
in living brachiopods, follow all the branches of the mantle canals and extend to the
margin of the mantle, and also form anastomosing blood sinuses to the gonads
(Hancock 1859, pi. 55, fig. 1; pi. 63, fig. 1; and pi. 56, fig. 4). The blood system in
brachiopods is an open system, the blood returning to the body cavity by way of tissue
spaces. The blood itself is generally free of cells (Hyman 1959, p. 558) and consists of a
colourless lymph-like fluid. Because the blood seeps back to the coelomic cavity before
being recirculated it is probably coelomic fluid from which the coelomocytes have been
filtered off. The blood system is responsible for the supply of nutriment to the organs
within the body cavity, the gonads, and to the mantle, especially to the cells at the grow-
ing edge of the mantle. Waste products are probably transported by the blood as it
filters through the tissues back to the coelome, where they are removed by the nephridia.
The blood must also contain dissolved minerals, mainly calcium carbonate, for deposi-
tion as shell material by the cells at the growing edge of the mantle.

MANTLE CANALS IN SCHIZOPHORIA VERULAMENSIS

The morphology of the mantle canal impressions in both valves of Schizophoria
vendamensis is shown in text-fig. 4. In the following description of the mantle canal
system I propose to use the term ‘canals’ for the impressions of the mantle canals on
the inner surfaces of the valves so as to avoid another descriptive term.

The mantle canals of small and apparently sexually immature specimens of Schizo-
phoria vendamensis consist of small canals of vascula media and vascula myaria as well
as a large number of narrower linear canals originating from the anterior and antero-
lateral margins of the muscle field and extending to the margins of the valve (text-fig.
5 a , b). There is no differentiation of vascula genitalia.

In the pedicle valve the impressions of the vascula media arise from the anterior
margins of the diductor muscle scars and extend with only minor branching to the front
of the valve. An arcuate lateral canal such as that in older individuals (text-fig. 5c) does
not appear to have been developed and the lateral margins of the valve were presumably
supplied with coelomic fluid by the narrow linear canals extending from the body
cavity.

Impressions of the mantle canals are less frequently preserved in brachial valves.
Small brachial valves have canals of vascula media originating from between the inner
margins of the anterior adductor muscle scars or, more rarely, from the inner parts of
the adductor scars, and extending towards the front of the valve. The vascula myaria
are simple, straight canals arising from between the anterior and posterior adductor
muscle scars and extending towards the postero-lateral margins.

J. ROBERTS: MANTLE CANAL PATTERNS IN SCHIZOPHO RIA

395

Adductor muscle scars Posterior adductor muscle scar

PEDICLE VALVE BRACHIAL VALVE

text-fig. 4. Morphology of the mantle canal impressions in Schizophoria
verulamensis Cvancara.

A

text-fig. 5. Mantle canal patterns in Schizophoria verulamensis Cvancara showing the development
from small to large individuals. The pedicle valve is on the left and the brachial valve on the right.

396

PALAEONTOLOGY, VOLUME 11

Larger, apparently sexually mature individuals (text-fig. 5 c, d) have an interconnected
system of vascula genitalia in the pedicle valve (PI. 74, fig. 3), an incipient system of
vascula genitalia in the brachial valve (PI. 74, fig. 6), and more complex vascula media
and vascula myaria canals.

In the pedicle valve of mature individuals the main canals of vascula media are well
separated from one another and are much broader and more deeply impressed than in
the younger form. Each of the main canals branches distally, giving off a number of
canals which run to the front of the valve, and an arcuate lateral canal which extends
parallel with the margin of the valve to the hinge. The lateral canal has a crenulate
pattern and at the peak of each crenulation gives off a major branch, which usually sub-
divides several times before reaching the margin of the valve. Towards the hinge the
lateral canal is irregularly crenulate and terminates in a group of diverse branches. The
vascula genitalia cover an area enclosed by the muscle field, the vascula media, and
the inner margins of the arcuate lateral canals, and are connected by numerous branches
with the vascula media and lateral canals; the significance of these connexions will be
dealt with below.

The vascula genitalia in brachial valves having the same size and presumably the
same age as mature pedicle valves do not appear to be as strongly impressed as those in
the pedicle valve. The genital markings consist of pit-like depressions or incipient inter-
connected canals on the postero-median parts of the valve, and give rise to numerous
branching canals, similar to those of the vascula media and vascula myaria, which
extend to the lateral margins (PI. 74, fig. 6). There are up to ten canals of vascula media,
and ten canals of vascula myaria extending with variable bifurcation towards the anterior
margin. Both the vascula media and vascula myaria leave tracks on the anterior adductor
muscle scars.

In the largest and probably gerontic specimens (text-fig. 5 e, f ) the mantle canal
system is frequently more strongly impressed on the inner surface of the shell and is
characterized by even broader main canals and an expansion of the vascula genitalia.

i

EXPLANATION OF PLATE 74
Schizophoria verulamensis Cvancara

Fig. 1. X 1. F.7145a. Internal mould of pedicle valve showing the connexions between the vascula media
and vascula genitalia; L. 215 Lewinsbrook.

Fig. 2. X 1. F.7145r. Internal mould of pedicle valve; note the truncation of the narrow canals by the
main canals of vascula media; L. 270 Trevallyn.

Fig. 3. X 2. F.4677a. Internal mould of pedicle valve showing a sexually mature mantle canal system
superimposed over an immature system; L. 215 Lewinsbrook.

Fig. 4. X 1. F.8036. Internal mould of pedicle valve showing a third canal of vascula media extending
to the front of the valve; L. 50 Gresford Quarry.

Fig. 5. x 1. F.7147q. Internal mould of brachial valve showing a sexually immature pattern of narrow
linear mantle canals; L. 270 Trevallyn.

Fig. 6. X 1. F.7141a. Internal mould of brachial valve showing the early development of the vascula
genitalia; L. 215 Lewinsbrook.

Fig. 7. xl. F.7153b. Internal mould of mature brachial valve showing the connexion between the
vascula genitalia and the vascula myaria; L. 50 Gresford Quarry.

Fig. 8. X 1. F.7153c. Internal mould of a mature brachial valve; note the connexion between the
vascula genitalia and the vascula myaria; L. 50 Gresford Quarry.

Fig. 9. x2. F.7153a. Internal mould of an old brachial valve; note the narrow canals emerging from
beneath the wide superimposed canals of vascula media; L. 50 Gresford Quarry.

Palaeontology, Vol. 11

PLATE 74

ROBERTS, Schizophoria

J. ROBERTS: MANTLE CANAL PATTERNS IN SCHIZOPHORIA 397

In the pedicle valve the main vascula media canals are close together and branch
distally, the arcuate lateral canals are closer to the margin than in younger specimens,
and the vascula genitalia have the same ramifying canals as those in mature specimens,
covering almost the entire inner surface of the valve (PI. 75, figs. 4-7).

The canals of vascula media in the brachial valve are usually wider than in smaller
specimens but do not appear to have significantly altered their positions. The vascula
myaria are also broader and run straight to the antero-lateral margins, or are bent
inwards at their distal extremities. Many small branches along the outer margins of the
vascula myaria provide a connexion with the vascula genitalia. The genitalia have the
same morphology as those in the pedicle valve. They cover broad flabel late areas on
the lateral and postero-lateral parts of the valve, but have a smaller area than the genitalia
on the pedicle valve because they are confined by the vascula myaria.

A system of narrow linear canals at the margins of mature shells (PI. 74, figs. 1-3, 9;
PI. 75, figs. 1, 2) are continuous with small branching canals in the middle part of the
shell (viz. the small canals between the vascula media in PI. 74, fig. 3). These small
branching canals are interpreted as impressions of a juvenile mantle canal system, and
hence their distal parts are regarded as impressions of mantle canals, probably impres-
sions of the termination of the canals at the proximal ends of the setal follicles. These
impressions were apparently continuously recorded by a process of differential thicken-
ing at the margin. The linear canals are joined by branches of a superimposed mantle
canal system (see below), indicating their function as a canal rather than the trace of
a setal follicle. Had they been merely traces of setal follicles they would intercalate as new
setae were introduced around the margin; when they increase, these canals branch
(PI. 74, fig. 3; PI. 75, fig. 3).

Changes in the mantle canals. In the pedicle valve the main canals of vascula media
extending in front of the muscle field can be divided into three intergrading morpho-
logical types (text-fig. 6):

A. Canals convex outwards and widely separated.

B. Canals straight and widely separated.

C. Canals straight and close together.

It has been shown above that the A and B types of canals typify small- to medium-
sized individuals and C type canals larger individuals, and hence the different shapes
of the vascula media can be explained as stages in the ontogenetic development of the
individual. Accompanying the change from the A to C types of canals is the enlargement
of the complex network of the vascula genitalia, and it is suggested that this enlarge-
ment within the mesenchyme layer of the mantle gradually forced the vascula media
canals together and the lateral canals to move towards the margin as the shell became
older. Specimens having each of these types of canals are present in all of the horizons
in the Gresford district.

This explanation of the development of the canals does not account for their distribu-
tion as outlined in Table 2, unless evolution has taken place.

An analysis of Table 2 is simplified if it is realized firstly that the C type canal is that
seen in the largest specimens, and secondly that the L. 53 sample is dominated by large
specimens (text-fig. 7). Table 2 shows a decrease in A type canals and an associated

398

PALAEONTOLOGY, VOLUME 11

increase in B type canals from the oldest to the youngest horizons. This trend could be
explained in at least two ways: as the expression of the ontogeny, as outlined above,
in which sampling errors gave the impression of a genetic change, or as a true genetic
change towards the earlier maturity of the genitalia.

An examination of the width-frequency polygons (text-fig. 3) shows that the samples
are essentially the same in terms of size distribution, except for L. 53. Also, the specimens
from the highest stratigraphic horizon with B type canals are about the same size as
specimens from the lower horizons characterized by A type canals. In the absence of
genetic change the specimens having the B type canals should have been larger.

TABLE 2

A

Canal types

B

C

L. 53 Greenhills

Occasional

30%

70%

L. 206 Lewinsbrook Syncline

25%

75%

occasional

L. 270 Trevallyn

65%

35%

occasional

^ j Gresford Quarry

65%

35%

occasional

L. 215 Lewinsbrook

70%

30%

occasional

More conclusive evidence in favour of the genetic change is available from younger
horizons (probably middle to upper Visean) at Barrington, N.S.W., in which moderate
to large specimens of S. verulamensis have mainly straight, sub-parallel (or B type)
canals (Cvancara 1958).

The earlier maturity of the genitalia would have increased the number of larvae pro-
duced by each individual and probably significantly increased the S. verulamensis popula-
tion to one of even greater dominance during the Visean.

Superimposition of mantle canals. In mature specimens of Schizophoria verulamensis,
particularly in pedicle valves, the greater part of the mantle canal system is super-
imposed over small radial canals. The small canals originate from the body cavity
frequently branch, and run irregularly towards the front of the shell. In mature valves
the posterior parts of these canals (viz. the narrow branching canals between the main
vascula media in PI. 74, fig. 3) are interpreted as the traces of juvenile mantle canals.

EXPLANATION OF PLATE 75
Schizophoria verulamensis Cvancara

Fig. 1. x2. F.7152a. Internal mould of pedicle valve showing narrow canals extending from the front
of the muscle field and being truncated by the superimposed mature mantle canal system; L. 50
Gresford Quarry.

Fig. 2. x2. F.7146k. Internal mould of pedicle valve; L. 270 Trevallyn.

Fig. 3. x2. F.4677b. Internal mould of pedicle valve; L. 215 Lewinsbrook.

Fig. 4. X 1. F.7144o. Internal mould of an old pedicle valve; L. 53 Greenhills.

Fig. 5. X 1. F.4784. Internal mould of an old pedicle valve with C-type vascula media and large areas
of vascula genitalia; L. 53 Greenhills.

Fig. 6. X 1. F.7145b. Internal mould of an old pedicle valve; L. 270 Trevallyn.

Fig. 7. xl. F.7146c. Internal mould of an old pedicle valve with large vascula genitalia, C-type
vascula media and the lateral canals situated close to the margin; L. 270 Trevallyn.

Palaeontology , Vol. 11

PLATE 75

ROBERTS, Schizophoria

J. ROBERTS: MANTLE CANAL PATTERNS IN SCHIZOPHORIA 399

Because all the canals were connected to the bases of the setal follicles (Williams and
Rowell 1965) these juvenile canals are continuous with the traces, recorded at the
margin, of the termination of the mantle canals at the setal follicles.

text-fig. 6. The three types of vascula media canals in the pedicle valve of Schizophoria verulamensis

Cvancara.

so r

5

UJ 40 -

>

_l

<

>

• L53
o L206
+ L270
x L2I5

O

Q

LU

CL

I

o

z

LU

20 -

o

o

• ®
t*

x + f»

• * liVt0 x
x#+x x x °x x

n 4°
<»XX+* o
X +» 4-

+ +

, *4**

x xx o X

. ° 8

o

o

o t n

—I 1 I

30 40 50

WIDTH PEDICLE VALVE MM

60

text-fig. 7. Scatter diagram of length plotted against width for pedicle valves of Schizophoria verula-
mensis Cvancara from L. 215 Lewinsbrook, L. 270 Trevallyn, L. 206 Lewinsbrook Syncline, and

L. 53 Greenhills.

In mature pedicle valves impressions of the small canals extend from the front of the
muscle field between the large canals of vascula media, branching as they run towards
the front of the valve (PI. 74, fig. 3; PI. 75, figs. 1, 2). In the middle and mid-anterior

400

PALAEONTOLOGY, VOLUME 11

parts of the valve they are truncated by superimposed larger canals: viz. the inner
branch of the right-hand vasculum medium on the specimen in Plate 74, fig. 3 in which
small canals extend on either side of a branch of the superimposed larger canal. Near
the margins, the large canals join with the narrow linear traces left by the mantle canals
at the ends of the setal follicles, and extend to the depressions left by the setal follicles.
The junction between these canals results from both systems supplying the same setal
follicles at the margin; the small linear marks were impressed as the shell grew, and the
larger canals, which had been formed by the modification of the smaller ones, were
impressed at a later time.

The superimposition of the mantle canal systems is less clear in the brachial valve.
One large specimen, however, has a system of narrow linear canals beneath broad canals
of vascula media (PI. 74, fig. 9). The U-shaped branch on the left hand side of the
vascula media truncates at least six of the narrow canals, and other narrow canals
emerge from beneath the sites of bifurcation of the larger canals and are apparently
continuous with the traces of the canals at the bases of the setal follicles.

The two main systems (i.e. the younger and older systems) described above are inter-
preted as impressions of the mantle canals at two different periods during the life of the
shell, probable before and after sexual maturity. Because of the impression of a large or
mature system on top of a smaller and probably immature system it is inferred that the
major part of the canal system was impressed at certain well-spaced intervals instead of
being continuously recorded on the inside of the shell.

Williams (1956, p. 273) thought that mantle canals were impressed on the inner
surface of the shell by differential secretion of shell material by the outer epithelium
covering the canals. In S. verulamensis this mechanism may have been responsible for the
continuous impression of the distal parts of the canals at the margin, but it could not
have been responsible for the impression of the main part of the mantle canal system,
because there are two and sometimes three systems superimposed over one another,
unless rapid differential secretion took place at specific times; this is unlikely because
it means that rather than being gradually thickened the shell would have been periodically
rapidly thickened.

A more convincing explanation is that the outer epithelium over the mantle canals
resorbed the shell material at specific times. Resorption is most likely to have taken
place during the periods of limited shell growth when the mantle canal system had a
stable morphology. During periods of rapid growth there would have been continuous
reorganization of the mantle canals at the mantle edge and the development of addi-
tional setal follicles at the growing margin. To maintain efficient circulation to the
margin the main canals of vascula media and vascula myaria would have enlarged by
reorganizing the connective tissue in the mantle and coalescing with smaller canals.
Many modern shallow-water marine organisms, particularly molluscs, cease depositing
shell during the colder months of the year (Epstein and Lowenstam 1954), and hence it
is reasonable to assume that the winter was the most likely time for the retardation of
shell deposition and the impression of the mantle canals in 5. verulamensis.

Abnormal mantle canals. In the pedicle valve figured in Plate 74, fig. 4, the left-hand
canal of vascula media divides a short distance in front of the muscle field, giving a
third main canal extending to the median anterior part of the valve. All three canals

J. ROBERTS: MANTLE CANAL PATTERNS IN SCHIZOPHORIA 401

branch in a similar manner to those in other valves. Functionally this additional large
branch may have provided a greater supply of coelomic fluid to the growing edge of
the front of the mantle. The two outer canals of vascula media give off arcuate lateral
canals.

Another specimen showing an abnormal development of the vascula media (text-fig.
8) has impressions of the two main canals of vascula media arising from the anterior
margin of the left diductor muscle scar and running, parallel with one another, to the
antero-lateral part of the valve. There is no canal originating from near the right
diductor scar, presumably because of some injury or functional defect to that side of
the body cavity, and the vascula genitalia extend in front of the muscle scar. The

text-fig. 8. Pathological pedicle valve of Schizophoria verala-
mensis Cvancara showing the two main canals of vascula media
extending from the left diductor muscle scar, and the vascula
genitalia extending in front of the right diductor muscle scar.

preservation is too poor to trace the junction of the right-hand canal with the right
lateral canal. This variation indicates a certain adaptability in the organization of the
mesenchyme layer of the mantle, and supports the arguments put forward above that
major reorganizations took place within the mantle during the development of the
species. Other features probably associated with this deformity are a deep furrow on
the antero-median part of the left diductor muscle scar and an exceptionally narrow
median ridge which does not expand anteriorly.

In a large pedicle valve (PI. 75, fig. 6) branches from the arcuate lateral canal in the
posterior part of the valve, instead of extending normal to the margin of the valve, turn
posteriorly and run almost parallel with the margins, diverging only slightly from the
main lateral canal. In other valves (PI. 74 ,fig. 3 jonly the posterior-most branches of the
lateral canal turn and run more or less parallel with the canal.

Connexion between the mantle canals. The description and figures of mantle canals of
fossil brachiopods given by Williams (1956) and Williams and Rowell (1965) suggest
that each of the three pairs of extensions of the coelomic cavity into the mantle — the
vascula media, vascula myaria, and vascula genitalia — is a discrete unit, unconnected
with other canals.

402

PALAEONTOLOGY, VOLUME 11

Larger specimens of Schizophoria verulamensis with well-preserved mantle canals
have vascula media and arcuate lateral canals possessing well defined inner and outer
margins. The vascula genitalia contained within the arc of the lateral canals is connected
by many fine branches with the main vascula media canals and the arcuate lateral
canals, the connexions varying from narrow constricted branches (the branches on the
right-hand vasculum medium of the specimen in PI. 74, fig. 3) to broad canals (the
connexion with the right-hand lateral canal on the same specimen). In another specimen
(PI. 74, fig. 1) the branches into the vascula media canals are more pronounced and have
a herringbone pattern. As well as being connected with the vascula media, the vascula
genitalia are presumably connected directly with the coelomic cavity adjacent to the
muscle-field.

There is no connexion with the vascula media in the brachial valve, but instead the
vascula genitalia are linked with the main canals of vascula myaria (PI. 74, figs. 7, 8).
The branches are comparable with those in the pedicle valve and appear to extend along
the entire outer margin of the vascula myaria canals.

Similar connexions occur in Ordovician orthid brachiopods described by Cooper
(1956) (see ChciulistomeUa magna (Schuchert and Cooper), pi. 70, fig. 21; and Mimella
globosa (Willard), pi. 89, figs. 10, 12, 14). One other possible example of an intercon-
nected system in Schizophoria is given by Pocock (1966, fig. 17), who figured a specimen
of S. striatula (Schlotheim) in which lateral branches arise from the vascula myaria and
extend towards the vascula genitalia.

This type of interconnexion between the genitalia and the other mantle canals is
apparently unknown in modern brachiopods. The interiors of modern brachiopods
figured by Hancock (1859) show that the genitalia are either situated in the mid-part of
a sac-like mantle canal, such as that in Terebratulina caputserpentis Linne (Hancock
1859, pi. 53, figs. 1, 2), or lie within the narrow canals of vascula genitalia or vascula
media as in Macandrevia cranium (Muller) (Hancock 1859, pi. 53, fig. 4). In specimens
with sac-like mantle canals, closest to those in Schizophoria , the vascula media and
lateral canals are simply the lateral extremities of the gonocoel; they have no inner
walls, and hence differ from the discrete canals in S. verulamensis. There are no con-
nexions with the network of genitalia, but the genitalia are bathed in coelomic fluid.
In S. verulamensis , coelomic fluid must have actually passed along the tubular network
of vascula genitalia.

The connexion of the vascula genitalia with other canals in S. verulamensis probably
increased the circulation of oxygenated fluid over the gonads, the fluid coming from the
vascula media and lateral canals (or the vascula myaria in the brachial valve) as well as
from within the vascula genitalia itself. The gonads in brachiopods without intercon-
nected canals are oxygenated only by the coelomic fluid contained within the vascula
genitalia, and sometimes the vascula media, meaning that there is probably less oxy-
genated fluid passing over them. Additional aeration could be responsible for the forma-
tion of larger gonads and possibly for a faster rate of production of sex cells, and it may
also cause them to ripen earlier. In the modern brachiopod Terebratulina caputserpentis,
already mentioned above, the gonads are bathed in the coelomic fluid of a large sac-like
gonocoel, which is probably an even more efficient method of aeration.

The interconnexion of the canals seems to be a favourable attribute, and on a purely
morphological basis it would seem that a system, such as that in S. verulamensis, would

J. ROBERTS: MANTLE CANAL PATTERNS IN SCHIZOPHORIA 403

have been more efficient for respiration and reproduction than, for example, a more
restricted, discrete system such as that in the recent brachiopod Fal/ax (Williams and
Rowell 1965, fig. 23d). Some of the respiratory functions of the mantle could, however,
have been taken over by the lophophore in other brachiopod groups, so that there was
less need to develop an interconnected system and the mantle canals acted as repositories
for the gonads and as an accessory respiratory system.

The present work shows that a closer examination of living brachiopods is desirable
to see if any species have an interconnected mantle canal system like that in S. verula-
mensis and other orthid brachiopods.

Interpretation of the mantle canals. The development of the mantle canals in S. verula-
mensis can be interpreted as follows. During the first period of growth, probably about
one year, the shell was sexually immature and had a relatively undifferentiated system of
mantle canals. In the first winter, when there was little shell growth, narrow linear canals
including vascula media and vascula myaria were impressed on the inner surface of the
shell, probably by resorption (PI. 74, fig. 5). During the following summer the shell grew
rapidly and became sexually mature. The rapid growth and the introduction of many
setal follicles at the mantle edge, combined with the growth of the gonads, resulted in
the reorganization of the mantle canal system. In the pedicle valve the vascula media
became wider and developed complex lateral canals, and the vascula genitalia changed
from narrow linear canals to a network of ramifying channels which branched with one
another and with the main branch and lateral canal of the vascula media (PI. 74, fig. 3).
The changes in the brachial valve were not as marked; the canals of vascula media and
vascula myaria becoming wider, and vascula genitalia becoming pitted near the postero-
median part of the valve and also having wider canals extending to the lateral margins
(PI. 74, fig. 6). These canals were impressed during the second winter, when the shell was
about two years old.

The changes in mature individuals, especially the development of the lateral canals
and the network of vascula genitalia in the pedicle valve, must have resulted in a com-
plete reorganization of the connective tissue in the mantle as well as the introduction of
many blood vessels to the new canals and the genitalia. According to Hyman (1959,
p. 533) the connective tissue of living brachiopods consists of a ‘homogeneous material
containing stellate mesenchymal cells that may unite into loose networks’; the mesen-
chyme was thus probably easily modified to allow the introduction of new or larger
canals.

Further growth in the following summer resulted in an increase in size, the enlarge-
ment of the muscle-field, and a thicker shell, especially at the posterior. In the pedicle
valve the vascula genitalia increased enormously in size, forcing the main vascula media
canals close together (PI. 75, figs. 4-7) and the lateral canals towards the margin, so that
the impressions of the mantle canals covered almost the whole inside of the valve. In
the brachial valve the vascula media and vascula myaria canals became wider and the
vascula genitalia developed a ramifying network similar to that in the pedicle valve
(PI. 74, figs. 7, 8, 9). These enlarged canals were probably impressed during the third
winter.

Little is known about the rate of growth and maturity of modern brachiopods.
Rudwick (1962, p. 334; 1965, p. H209) suggested that modern articulate brachiopods

404

PALAEONTOLOGY, VOLUME 11

can reach sexual maturity after two or three years, and that they live for as long as
seven or eight years or even longer. Amongst the inarticulates Lingula is sexually mature
after a year or a year and a half (Chuang 1959), and Glottidia after only six to nine
months (Paine 1963). It is therefore reasonable to suggest that S. verulamensis became
sexually mature at about two years old, and that the mature mantle canal system was
impressed during its second winter. Mantle canals of older individuals cover almost the
entire surface of the shell and leave no room for the expansion of the vascula genitalia,
so that after about the third year the mantle canals became stable and could not be
modified during further growth of the shell. There is no indication of how long the
largest shell lived.

SUMMARY AND CONCLUSIONS

The morphology of the mantle canals in Schizophoria verulamensis Cvancara changes
during ontogeny, mainly because of the growth of the gonads. The canals start in
immature specimens as narrow linear canals over the inner surface of the valves and,
with the growth of the gonads, are later modified in larger specimens into several major
feeder canals which distribute coelomic fluid from the body cavity to the margins of the
shell, around the areas of genitalia. In the largest shells the genitalia cover the greater
part of the inner surface of the valves and have enlarged to such an extent that the large
distributive canals have been forced together in the middle of the valve, or moved
outwards towards the margins. The pedicle valve is affected earlier by the growth of the
genitalia than the brachial valve.

An analysis of the distortion to the main vascula media caused by the enlargement of
the vascula genitalia in pedicle valves from six horizons throughout the Lower Carboni-
ferous sequence indicates a trend, probably genetically controlled, towards the earlier
maturity of the genitalia.

The main part of the mantle canal system was impressed periodically by the resorption
of shell beneath the mantle canals rather than by differential secretion as suggested by
Williams (1956). Resorption most likely took place during the winter when there was
limited shell growth and a stable mantle canal system, recording several distinct mantle
canal patterns during the development of the individual.

Larger specimens of S. verulamensis have the vascula genitalia connected by numerous
small branches with the vascula media and lateral canals in the pedicle valve and the
vascula myaria in the brachial valve. This type of connexion has not been observed
in recent brachiopods. It is thought to have increased the circulation of oxygenated
coelomic fluid over the gonads, and may have been responsible for their earlier
ripening.

The mantle canals in pathological specimens indicate that not only was modification
possible during ontogeny, but the mantle canal system could also be adapted to the
effects of injuries or defects in the shell during life.

Acknowledgements. This work was begun at the University of Western Australia and finished at the
Bureau of Mineral Resources. I am indebted to Drs. K. S. W. Campbell and P. J. Coleman for their
interested discussion of the problem, and to Drs. J. J. Veevers, J. M. Dickins, and K. S. W. Campbell
for reading the manuscript. This paper is published with the permission of the Director, Bureau of
Mineral Resources, Geology, and Geophysics.

J. ROBERTS: MANTLE CANAL PATTERNS IN SCHIZOPHORIA

405

REFERENCES

brown, d. a., Campbell, k. s. w., and Roberts, j. 1965. A Visean cephalopod fauna from New South
Wales. Palaeontology, 7, 682-92.

brunton, c. h. c. 1966. Silicified productoids from the Visean of County Fermanagh. Bull. Brit. Mas.
(Nat. Hist.) Geol. 12, 175-243.

Campbell, K. s. w., and Roberts, j. 1964. Two species of Delepinea from New South Wales. Palaeon-
tology, 7, 514-24.

chuang, s. h. 1959. The breeding season of the brachiopod Lingula unguis (L). Biol. Bull. 117, 202-7.
cooper, G. a. 1956. Chazyan and related brachiopods. Smithson, misc. Coll. 127, 1-1245.
cvancara, a. m. 1958. Invertebrate fossils from the Lower Carboniferous of New South Wales.
J. Paleont. 32, 846-88.

epstein, s., and lowenstam, h. a. 1953. Temperature-shell-grown relations of Recent and interglacial
Pleistocene shoal-water biota from Bermuda. J. Geol. 61, 424-38.

Hancock, a. 1859. On the organization of the Brachiopoda. Phil. Trans. Roy. Soc. Lond. 148, 791-869.
hyman, l. H. 1959. The Invertebrates: Smaller coelomate groups. 5, 1-783. McGraw-Hill, New York.
muir-wood, H. M. 1962. On the morphology and classification of the brachiopod Suborder Chonetoidea.
Brit. Mus. (Nat. Hist.), London, 1-132.

opik, a. a. 1934. Uber Klitamboniten. Acta et Commentationes, Univ. Tartu, 26, 1-239.

Paine, R. T. 1963. Ecology of the brachiopod Glottidia pyramidata. Ecological Monographs, 33, 187-213.
pocock, y. p. 1966. Devonian Schizophoriid brachiopods from Western Europe. Palaeontology , 9,
381-412.

Roberts, j. 1961. The geology of the Gresford district, N.S.W. /. Proc. roy. Soc. N.S.W. 95, 77-91.

1963. A Lower Carboniferous fauna from Lewinsbrook, New South Wales. Ibid. 97, 1-31.

1964. Lower Carboniferous brachiopods from Greenhills, New South Wales. J. geol. Soc. Aust.

11, 173-94.

1965u. Lower Carboniferous faunas from Wiragulla and Dungog, New South Wales. J. Proc.

roy. Soc. N.S.W. 97, 193-215.

19656. A Lower Carboniferous fauna from Trevallyn, New South Wales. Palaeontology, 8, 54-81 .

1965c. Lower Carboniferous zones and correlations based on faunas from the Gresford-Dungog

district. New South Wales. J. geol. Soc. Aust. 12, 105-22.
rudwick, m. j. s. 1962. Filter-feeding mechanism in some brachiopods from New Zealand. J. Linn.
Soc. Lond. Zoology, 44, No. 300, 592-615.

1965. In Treatise on Invertebrate Paleontology. Part H, Brachiopoda. Ed. R. C. Moore. Univ.

Kansas Press.

schuchert, c., and cooper, g. a. 1932. Brachiopod genera of the Suborders Orthoidea and Pentame-
roidea. Mem. Peabody Mus. Nat. Hist. 4(1) 1-270.
williams, A. 1956. The calcareous shell of the Brachiopoda and its importance to their classification.
Biol. Rev. 31, 243-87.

and rowell, a. j. 1965. In Treatise on Invertebrate Paleontology. Part H, Brachiopoda. Ed. R. C.

Moore. Univ. Kansas Press.

JOHN ROBERTS

Bureau of Mineral Resources, Geology
and Geophysics
Canberra City, A.C.T.

Typescript received from author 1 May 1967 Australia

ON ‘ DENDROCRINUS ’ CAM BRIENSIS HICKS,
THE EARLIEST KNOWN CRINOID

by DENIS E. B. BATES

Abstract. Dendrocrinus cambriensis is re-described from new material and referred to Ramseyocrinus gen. nov.,
and placed in the family Eustenocrinidae.

The crinoid remains described by Hicks from the lower Ordovician rocks of Ramsey
Island and the adjacent mainland of Wales are the earliest yet reported, and, as such,
might be expected to yield significant information on the origin of the class. Several
writers (Moore 1950, p. 32, Ubaghs 1953, p. 744, Regnell 1960, p. 172) have referred to
their early stratigraphical occurrence, and Ramsbottom (1960, p. 5) has re-investigated
Hicks’s specimens, without being able to add significantly to Hicks’s information.
Unfortunately none of Hicks’s specimens shows the posterior side, and hence the anal
tube was unknown. The writer has had the opportunity of examining the extensive
collection of material from Ramsey Island made by Dr. F. J. North for the National
Museum of Wales, kindly made available by Dr. D. A. Bassett, and has based his
description on specimens from it.

Class crinoidea J. S. Miller 1824
Subclass inadunata Wachsmuth and Springer 1881
Order disparata Moore and Laudon 1943
Family eustenocrinidae Ulrich 1924

Monocyclic crinoids with a slender cylindrical crown about equal in diameter to the
stem; calyx composed of three (?) to five basals, four radials, and an equal-sized infer-
radianal; four arms, branching isotomously; infer-radianal bearing a series of un-
branched quadrate plates presumably supporting an anal sac.

Genus Ramseyocrinus gen. nov.

Diagnosis. A genus of the Eustenocrinidae with a four-lobed stem composed of colum-
nals of irregular thickness; basals probably three in number, separated by well-developed
sutures.

Type species. Dendrocrinus cambriensis Hicks 1873.

Discussion. The described species of Eustenocrinus have a round smooth column, with
two alternating sets of thin columnals, one set twice as thick as the other. The five basal
plates are all imperfectly fused one to another. In Ramseyocrinus there are, apparently,
fewer basals, and the sutures between them are as well defined as those between the
radials.

[Palaeontology, Yol. 11, Part 3, 1968, pp. 406-9, pi. 76.]

D. E. B. BATES: ON ‘ D EN D ROCRI N US' CAMBRIENSIS HICKS

407

Ramseyocrinus cambriensis (Hicks)

Plate 76, figs. 1-5

1873 Dendrocrinus cambriensis Hicks, p. 50, pi. 4, figs. 17-20.

1960 Jocrinus ? cambriensis (Hicks) Ramsbottom, pp. 5-6, pi. 3, figs. 9-1 1.

Diagnosis. Basals low, over twice as wide as high; rectangular radials wider than high,
and highly convex; nine or more primibrachs, arms branching at least three times.

Type specimens. The syntypes (Ramsbottom 1960, p. 5) are preserved in the Sedgwick Museum, with
some gutta-percha impressions in the Geological Survey Museum. Ramsbottom’s description is mainly
based on a specimen preserved in the British Museum (BMNH E3). The author has also discovered
in the Sedgwick Museum the counterpart mould of one of Hicks’s syntypes, and the specimen now
bears the numbers SM 16739a-b. The description below is mainly based on a specimen from the
National Museum of Wales, preserved as counterpart moulds (29 308 G296 and G318).

Type locality. All the specimens figured by Hicks come from the Porth Gain Beds of Bay Ogof Hen,
Ramsey Island (Pringle 1930, p. 12), although specimens are recorded from near Llanveran Farm, on
the mainland near Pen Berry. No specimens have been seen or found by the writer from the latter
locality.

Description. Cup slightly wider than high, sides flaring slightly upwards, surface of
plates smooth. Basals apparently not more than four in number, and probably only
three, just less than half the height of the calyx and over twice as wide as high, penta-
gonal, with the upper surfaces concave, and the lower surfaces sharply bent inwards
between the petals of the stem. Radials four in number, quadrangular, wider than high,
with convex lower and concave upper surfaces, highly convex, especially towards their
lateral margins. Anal series commencing with an infer-radianal similar in width and
convexity to the radials, but only half their height, the succeeding three plates of the
anal series similar in size and height. The radials and first primibrachials almost equal
in height to the infer-radianal and the succeeding two plates of the anal series. Lower-
most brachials not interlocking and probably not included in the cup.

Primibrachs nine to twelve, wider than high, almost circular in cross-section, with
a deep narrow ventral groove, secundibrachs four to nine, as high as wide; tertibrachs
seven or more; quadribrachs at least eight; arms branching isotomously at least three
times.

Stem wide, petaloid, four lobed, the lobes proximally almost parallel-sided with
deep indentations, distally becoming less marked so that the stem becomes rectangular
in cross-section; columnals of irregular height.

Discussion. Moore and Laudon (1943, p. 26) described the two lowermost plates in each
ray of Eustenocrinus as infer-radial and super-radial, homologizing them with the com-
pound radials of genera such as Ectenocrinus (op. cit., fig. 3). However, in a later paper
(1950, p. 30) Moore has proposed a simpler nomenclature, regarding the lowermost
plate in each ray (excepting that beneath the anal tube) as radial, and the plates above
and incorporated in the cup as fixed brachials. The latter procedure is followed here.

The chief differences from the described species of Eustenocrinus are in the form of the
stem and the number of basals, which warrant separation of the Welsh species into a new
genus. There are at least three basals, one of which, on the anterior side, may be formed

408

PALAEONTOLOGY, VOLUME 11

from two plates fused together and showing traces of the suture between them (PI. 76,
fig. 1). In addition E. springeri (Ulrich 1924, p. 99) has arms which bifurcate only once,
and its diameter at the top of the basals appears to be less than at either the base or top
of the cup (op. cit., fig. 14a). E. milled (Wetherby 1880, p. 153) has the radial plates
higher than wide, with a thickening of their upper extremity which gives the species
a somewhat swollen appearance at the top of the calyx.

Eustenocrinus was considered by Moore and Laudon (1943, p. 25) as the most
primitive member of the homo-synbathrocrinids, the major stock of the Disparata.
Stratigraphically, however, it post-dates the earliest members of that stock, and hence
is not itself the ancestor. The features considered by Moore to be primitive include the
steep-sided conical cup, the convex base showing the lowest plates, isotomous arm
branching, and the lack of any fused super and infer-radial plates. The occurrence of
Ramseyocrinus at the base of the Ordovician, and its similar structure, further strengthens
Moore’s claim for this stock to be the ancestor of the Disparata.

The most striking features of Ramseyocrinus are the four-lobed stem and the four
arms. It was hoped that Ramseyocrinus would prove to be a possible ancestor for the
whole class Crinoidea, but these features indicate that it post-dates the splitting up of
the ancestral crinoid stock into at least the sub-classes Inadunata and Flexibilia. Most
other contemporary or older pelmatozoans have a well-developed five-fold symmetry,
and thus it is likely that the Eustenocrinidae arose from an ancestor having five arms,
one arm being replaced by an anal tube. Thus the Eustenocrinidae must first have lost
the right posterior arm during the late Cambrian and then regained it during the early
Ordovician, if they were, in fact, ancestral to the bulk of the Disparata. That this hap-
pened is questionable; it seems more probable that the Eustenocrinidae constitute an
early simplified offshoot from the Disparata.

REFERENCES

hicks, h. 1873. On the Tremadoc Rocks in the neighbourhood of St. David’s, South Wales, and their
fossil contents. Quart. J. geol. Soc. Loud. 29, 39-52, pis. 3-5.
moore, R. c. 1950. Evolution of the Crinoidea in relation to major Paleogeographic changes in Earth
History. Int. Geol. Congress, 18th Session, London, 1948, pt. XII, pp. 27-53.

and laudon, l. r. 1943. Evolution and Classification of Paleozoic Crinoids. Spec. Pap. geol. Soc.

Amer. 46, 1-153, pis. 1-14.

pringle, j. 1930. The geology of Ramsey Island (Pembrokeshire). Proc. Geol. Ass. Lond. 41, 1-31,
pis. 1-3.

ramsbottom, w. h. c. 1961. A monograph of British Ordovician Crinoidea. Palaeontogr. Soc. (Monogr).
1-37, pis. 1-8.

EXPLANATION OF PLATE 76

Ramseyocrinus cambriensis (Hicks) from the Porth Gain Beds of Bay Ogof Hen, Ramsey Island. All
photographs are of latex casts, whitened with ammonium chloride. None has been retouched.

Figs. 1-3. Specimen 29 308 G296/G318 (Nat. Museum of Wales). 1. Anterior view, showing stem of
irregular columnals, and two (?) basal plates partially fused together, X 5-8. 2. Posterior view,
showing the anal tube, two of the four (?) basals, and food grooves on two of the arms, X 5-5.
3. Stem viewed from below, showing the four lobed cross-section, x4-5.

Figs. 4-5. Syntype, specimen SM. 16739a-b (Sedgwick Museum, Cambridge). 4. Posterior (?) view,
x 2-0. 5. Anterior (?) view, X 31. The basal and radial plates cannot be positively identified, nor can
the anal series, which may be missing.

Palaeontology, Vol. II

PLATE 76

BATES, Ramseyocrinus cambrensis (Hicks)

1

D. E. B. BATES: ON ‘ D END RO C RIN US' CAMBRIENSIS HICKS 409

regnell, G. 1960. The lower Palaeozoic Echinoderm Faunas of the British Isles and Balto-Scandia.
Palaeontology, 2, 161-79.

ubaghs, g. 1953. Classe des Crinoides in Trade de Paleontologie, 3, 658-773. Ed. J. Piveteau. Paris.
ulrich, e. o. 1924. New classification of the Heterocrinidae in Foerste, A. F., Mem.geol. Surv. Can.
138, 82-104.

wetherby, a. g. 1880. Remarks on the Trenton Limestone of Kentucky with descriptions of new
fossils from that formation and the Kaskaskia (Chester) group, Sub-carboniferous. /. Cincinnati
Soc. Nat. Hist. 3, 144-60, pi. 5.

DENIS E. B. BATES

Department of Geology
University College of Wales

Typescript received 4 February 1967 Aberystwyth

E e

C 5586

REVISION OF TWO UPPER CAMBRIAN

TRILOBITES

by A. W. A. RUSHTON

Abstract. Conocoryphe ? bucephala Belt, 1868, from the Upper Ffestiniog and Lower Dolgelly Beds of Wales,
and possibly England, is re-illustrated; the pygidium is described for the first time; the species is transferred
from the Olenid genus Beltella to which it was referred by Lake (1919) to Parabolinoides Frederickson, a genus
widely known in the approximately contemporaneous Conaspis Zone in the U.S.A.

Sphaerophthalmus major Lake, 1913, though regarded by Henningsmoen (1957) as of questionable validity,
is distinguishable from other species of Sphaerophthalmus ; S. major occurs in England, Wales, and Sweden, and
probably in Norway and eastern Canada, in the Zone of Peltura scarabaeoides.

In his monograph on the Olenidae, Henningsmoen (1957) commented on the two
Upper Cambrian trilobite species Conocoryphe ? bucephala Belt and Sphaerophthalmus
major Lake, but for lack of data was unable to frame a diagnosis for either. In connexion
with work on the Cambrian faunas from the Institute of Geological Sciences Merevale
No. 1 borehole (Rushton, 1966), 1 examined the type-material of both these species
including material from the following museums: British Museum (Natural History)
(BM), Geological Survey Museum (GSM), Oxford University Museum (Ox), and the
Sedgwick Museum, Cambridge (SM). The terminology used in the description below is
that of Henningsmoen (1957).

Family parabolinoididae Lochman
Genus parabolinoides Frederickson 1949

Type species (by original designation). Parabolinoides contractus Frederickson 1949.

Parabolinoides bucephalus (Belt 1868)

Text-fig. 1 ; Plate 77, figs. 1-10, 1 1 ?

1868 Conocoryphe ? bucephala Belt, p. 10, pi. 2, figs. 1-6.

1873 Conocoryphe Willicunsoni Salter, p. 12 [Not Conocoryphe Williamsonii Belt, 1868],

1919 Beltella bucephala (Belt); Lake, p. 106, pi. 12, figs. 1 1-15.

1919 Beltella verisimilis (Salter) [part im]; Lake, p. 107, pi. 13, figs. 4, 5 only.

? 1927 Beltella ? sp. nov. Cobbold, p. 557.

? 1930 Beltella cf. bucephala (Belt); Stubblefield, p. 57.

1957 Olenusl bucephalus (Belt 1868); Henningsmoen, p. Ill (with further synonymy).

Lectotype (selected by Henningsmoen 1957, p. 111). The original of Belt’s plate 2, fig. 1. The only
specimen extant which can be considered as this original is BM I 7578 from the Upper Ffestiniog Beds
near Dolgelly, shown here on Plate 77, fig. 7 ; it is the only relatively large and complete specimen in
Belt’s collection, is the same size as his figure, and also agrees in being rather obscure in the posterior
part of the shield. It is assumed that the figure is restored.

Remarks on the material studied. The specimens of Conocoryphe ? bucephala collected
by Belt and now in the British Museum (Natural History) and the Geological Survey

[Palaeontology, Vol. II, Part 3, 1968, pp. 410-20, pis. 77-78.]

RUSHTON: REVISION OF TWO UPPER CAMBRIAN TRILOBITES 411

Museum show the effects of flattening, crushing, and deformation to varying extents:
the lectotype is obliquely compressed and the preglabellar field is crushed on to the
anterior extension of the doublure of the slightly displaced free cheek; the cranidia in
Plate 77, figs. 3 and 4, are laterally and frontally compressed respectively, and that shown
in fig. 1 is crushed flat but seems to be relatively undistorted. The glabella of the type
specimen of Conocoryphe williamsoni Salter non Belt (PI. 77, figs. 10a and b) is crushed
on to the hypostoma, and the preglabellar field on to part of the free cheek, as in the
lectotype of Parabolinoides bucephalus ; the free cheeks are displaced backwards and
inwards relative to the cranidium and the pygidium is steeply inclined relative to the
thorax, but otherwise the specimen is good, being but little flattened and deformed.

In an attempt to determine the original proportions of P. bucephalus the method
suggested by Henningsmoen (1960, p. 207) was applied to two mature cranidia on one of
Belt’s slabs of Upper Ffestiniog Beds (GSM Za 4843) which satisfy the necessary con-
ditions except that one is somewhat larger than the other. The calculation indicates that
the original ratio of length to breadth of the cephalic axis was 4:3, which is in good
agreement (less than 5 per cent, difference) with the ratios obtained from the apparently
undeformed specimens shown in Plate 77, figs. 1, 6, and 10a; in contrast, direct measure-
ment of the obviously deformed specimens in figs. 3 and 4 gives ratios of 2-0:1 and
0-9 : 1 respectively.

Description. Lake’s description of the species is mainly correct but it requires some
amplification, particularly in the account of the cranidium.

Cephalic axis prominent, tapered, rounded in front, breadth three-quarters of length;
SI weak or distinct, long, oblique backwards, straight or slightly convex; S2 weak,
short, less oblique; a few specimens show traces of S3, as stated by Belt (1868, p. 10);
the appearance of the glabellar furrows varies much according to the state of pre-
servation. Occipital furrow deepest medially, composite; occipital ring apparently
without node or spine. Frontal area one-third to nearly half as long as cephalic axis;
border-furrow slightly bowed forwards and, in some specimens, with a weak median
in-bend; anterior border crescentic or subtriangular, longer at the mid-line than the
preglabellar field. Eye-ridges distinct or weak. Eyes opposite S2 or L2. Inter-ocular
cheeks half or less than half as wide as the glabella at the eye-line; post-ocular cheeks
rounded postero-laterally, slightly narrower than the occipital ring. Pre-ocular facial
sutures straight, divergent forwards to the border-furrow, strongly convergent over
anterior border; post-ocular facial sutures divergent, slightly convex.

Free cheek with broad border and long anterior extension of the doublure (PI. 77,
fig. 8). Ventral sutures unknown.

There are twelve thoracic segments in the type-specimen of Conocoryphe williamsoni
Salter and twelve or thirteen in the lectotype of C. ? bucepha/a, in which the joint between
the thorax and pygidium is not very clear. The pleural tips are pointed and the axis lacks
nodes except in young specimens.

Pygidium poorly known; pygidial axis with about four rings (including terminal);
pleural regions with about three pleural and two interpleural grooves, three pairs of
marginal spines and a striated doublure (PI. 77, fig. 106).

Young specimens in Belt’s collection have a convex glabella which, compared with
adult specimens, is larger in proportion to the rest of the cranidium (PI. 77, fig. 9).

412 PALAEONTOLOGY, VOLUME 11

Measurements in mm.

SM A3095

BM 59284

BM It 2585

Length of cranidium

110

100

6-1

Width of cranidium

17-2

16T (est.)

90

Length of cephalic axis

c. 7-6

7-2

4-2

Width of cephalic axis

5-8

5-3

3-2

The lectotype is about 34 mm. long; SM A3095 is about 30 mm. long (restored).

Remarks. I agree with Lake (1919) in regarding Salter’s Conocoryphe williamsoni as a
synonym of Belt’s C. ? bucephala. Compared with C. ? bucephala the glabella of C.
williamsoni seems to be slightly smaller in proportion to the rest of the cranidium but
this may be because in C. williamsoni it is preserved convex (except where crushed on
to the hypostoma), whereas in the relatively undeformed specimens of C. ? bucephala
it is ‘spread’ by flattening and crushing. Otherwise, so far as the species can be com-
pared, they agree quite well. The small cranidium collected by Stubblefield (1930) from
Bentleyford Brook in Shropshire is slightly flattened compared with young specimens in
Belt’s material and has a less prominent anterior border, but is otherwise very similar.
The specimens collected by Stubblefield from concretions in Comley Brook, Shropshire
(Cobbold 1927, p. 557) and considered by him to be the same species as the Bentleyford
cranidium are very small but may provisionally be retained as Parabolinoides cf.
bucephalus.

I have not seen the specimens from the Ffestiniog Beds illustrated by Lake (1919) as
belonging to the Tremadoc species Beltella verisimilis (Salter), but from the illustrations
they appear to be laterally compressed specimens of Parabolinoides bucephalus.

Generic position. The chief features of Conocoryphe ? bucephala — the tapered glabella
with oblique furrows and rounded front, wide anterior border, small eyes close to the
glabella, the course of the facial suture and the spinose pygidium — are those of Para-
bolinoides Frederickson.

Belt's original reference to Conocoryphe ? followed Salter’s example in using the name
for what would now be called ptychoparioid trilobites. Subsequent writers (Reed 1900;
Lake 1919; Henningsmoen 1957) agreed in assigning P. bucephalus to the Olenidae, but
the wide cephalic border and divergent pre-ocular sutures are not typical of that family.
Nevertheless, P. bucephalus otherwise resembles late Olenus species such as O. cata-
ractes Salter which, however, has more thoracic segments and stronger glabellar furrows ;

EXPLANATION OF PLATE 77

Figs. 1-10. Parabolinoides bucephalus (Belt). 1. Cranidium, Ffestiniog Beds near Penmaenpool, Merio-
neth, X 2\, BM 59284. 2-5. Deformed cranidia, Upper Ffestiniog Beds near Dolgelly (Belt Collection),
x 2\\ GSM Za4844, BM I 7580, BM I 7588, BM I 7590 respectively. 6. Crushed cranidium with part
of thorax; Ffestiniog Beds? X 2£, BM It 2585. 7, 8. Lectotype and free cheek, Upper Ffestiniog Beds
near Dolgelly (Belt Collection), x 2\, BM I 7578, BM I 7583. 9. Convex immature cranidium, prob-
ably this species, X 5, BM 58499. 10. Complete specimen (type specimen of Conocoryphe williamsoni
Salter), Lower Dolgelly Beds, Rhiwfelyn SM A3095; lOu, X24; 106, view of pygidium, X 5. 11. P. cf.
bucephalus (Belt); immature cranidium from Orusia Shales (Lower Dolgelly Beds) in tributary of
Bentleyford Brook, Shropshire, X 5, GSM D2655.

Photographs by Mr. C. Friend.

Palaeontology, Vol. II

PLATE 77

RUSHTON, Upper Cambrian trilobites

RUSHTON: REVISION OF TWO UPPER CAMBRIAN TRILOBITES 413

species of the (typically Tremadoc) olenid genus Beltella Lake differ in having truncate
pleural tips and an entire pygidial margin. Leptoplcistides Raw, which is related to
Beltella (Henningsmoen 1957, p. 265) but has spinose pleurae, differs from P. bucephalus
in having axial spines and a different type of free cheek. Beltella solitaria Westergard
(1922, pi. 14, fig. 1) has a less prominent anterior border and a truncate glabella, and
seems to be a true Olenid.

text-fig. 1. Diagrammatic drawings of cranidia of Parabolinoides species, about X 2-|. ( a ) P. contractus
Frederickson (after Frederickson 1948). (b) P. hebe Frederickson (after Frederickson 1948). (c) P.
palatus Berg (after Berg 1953). (d-f )P. bucephalus (Belt), restored from various specimens.

P. bucephalus is placed in Parabolinoides rather than any other genus of the Para-
bolinoididae such as Maustonia Raasch in Lochman or Taenicephalus Ulrich and Resser
in Walcott because the post-ocular sutures are convex rather than sinuous (see Loch-
man in Harrington et al. 1959, p. 0272). Bernia Frederickson which has convex post-
ocular sutures is considered a synonym of Parabolinoides by Bell and Ellinwood (1962,
p. 398).

Comparison with other species (text-fig. 1). Parabolinoides bucephalus differs from other
species of Parabolinoides in having the inter-ocular cheeks about half as wide as the
glabella, whereas in other species they are only about one-third as wide. The most
similar species are P. contractus Frederickson (1949) which, however, also differs in
having a longer frontal area and an occipital node, and P. hebe Frederickson which has
a relatively short anterior border and a more truncate glabella. P. palatus Berg (1953)
has a relatively short, square glabella and more transverse post-ocular sutures. P. ?
cordillerensis Lochman (1950) has relatively very narrow inter-ocular cheeks and some-
what sinuous post-ocular sutures.

414

PALAEONTOLOGY, VOLUME 11

Occurrence. Parabolinoides bucephalus has been found in North Wales and possibly in
Shropshire at horizons at or near the top of the Olenus Zone and in the Parabolina
spinulosa Zone (Upper Ffestiniog and Lower Dolgelly Beds).

The specimens in Belt’s collection are all labelled ‘Upper Ffestiniog, Dolgelly’ and
in his account of the species Belt (1868) mentions the localities Gwern-y-barcud (near
Penmaenpool), Mynydd Gader, and near Craig-y-Dinas. G. J. Williams collected
P. bucephalus at the same horizon in Nant Cistfaen and on Trinant, near Arenig (Lake
1919, pp. 107, 109). Mr. S. W. Hester of the Geological Survey has collected a specimen
identified by Stubblefield as Beltella cf. bucephala from the basal Lower Dolgelly Beds
at Ogof Ddu, one mile east of Criccieth, Caernarvonshire. The type of Conocoryphe
mUiamsoni Salter is stated by Lake (op. cit., p. 107) to be from the ‘Upper Lingula
Flags’ (= Dolgelly Beds) at Rhiwfelyn in the Upper Mawddach Valley; a variety of
horizons is exposed near Rhiwfelyn (Wells 1925, map, pi. 32) but the lithology of the
specimen agrees with that of the Lower Dolgelly Beds which yield Parabolina spinulosa
(Wahlenberg).

Immature cranidia of Parabolinoides cf. bucephalus have been collected from a sep-
tarian nodule from Comley Brook, south Shropshire (Cobbold 1927, pp. 556-7) where
they were associated with Parabolina spinulosa and Orusia lenticularis (Wahlenberg). In
a section at Bentleyford Brook a single immature cranidium was collected near the base
of a 63-ft. thickness of micaceous shales which, at higher horizons, yielded also Orusia
lenticularis and Parabolinites ? [Parabolinella] aff. williamsonii (Belt).

Apart from P. bucephalus , recorded species of Parabolinoides are confined to the Con-
aspis Zone in the central parts of the United States of America (Lochman and Wilson
1958, see text-figs. 9, 10, 1 1). If, as has been suggested, the top of the Elvinia Zone is not
higher than the top of the Olenus Zone (Rushton 1967), P. bucephalus and the American
species of Parabolinoides occur at about the same horizon.

North Welsh European North American

succession Zones Zones

Lower

Doglelly

Parabolina

spinulosa

Conaspis

Ffestiniog

Upper

Maentwrog

f

Olenus

e

(6 sub-zones
a-f)

d

Elvinia

Family olenidae Burmeister
Subfamily leptoplastinae Angelin
Genus sphaerophthalmus Angelin 1854

Type species (subsequently designated by Linnarsson 1880). Trilobites a/atus Boeck, 1838.

According to Henningsmoen (1957, p. 21 1) there are four species of Sphaerophthalmus,
all occurring in the Upper Cambrian Peltura Zones of the North Atlantic Province. He

RUSHTON: REVISION OF TWO UPPER CAMBRIAN TRILOBITES 415

lists S. alatus (Boeck), Olenus humilis Phillips 1848, S. majusculus Linnarsson 1880, and
S', major Lake 1913. Henningsmoen discussed the nomenclature of the species; his con-
clusions may be summarized as follows :

1. In 1880 Linnarsson stated that Olenus humilis Phillips was a junior synonym of
Sphaerophthalmus alatus (Boeck). Henningsmoen showed that this was a mistake, the
species being distinct.

2. Other workers accepted Linnarsson’s statement; one such was Lake (1913) who
figured and described material from Phillips’s type-area for Olenus humilis under the
name S. alatus (Boeck). Henningsmoen suggested that the name humilis should be
revived for S. alatus of Lake (not Boeck). Lake (op. cit.) also described a new species,
S. major.

text-fig. 2. Reconstruction of cephalon of Sphaerophthalmus
humilis (Phillips) showing attitude of free cheek, X 10,
anterior and lateral views. Cranidium after Westergard.

3. Westergard (1922) likewise figured Phillips’s species under the name S. alatus and
he tentatively assigned a second species to S. major Lake; he also re-illustrated S. maju-
sculus Linnarsson. Henningsmoen pointed out that although the Scandinavian specimens
of ‘S’, major ’ should be referred to S. alatus (Boeck) the same was not necessarily true
of the type-material of S. major.

The morphology of Sphaerophthalmus. The cranidia of species of Sphaerophthalmus are
convex so that the differences between specimens preserved in limestone and those in
shale are quite important, as can be seen in the figures given by Westergard (1922, pi. 13,
figs. 9-35). For example, the transversely arched anterior border, when flattened, be-
comes an emargination in dorsal view (Westergard 1922, pi. 13, figs. 20 and 23), and
the post-ocular cheeks are bent up from their original steeply down-sloping attitude so
that the width of the cranidium is increased. When the cranidium is flattened the facial
suture describes an elongated S-shape, but when preserved convex the course is, in
dorsal view, curved anteriorly and nearly straight behind and, in side view, arched up at
the palpebral lobe.

The free cheek of Sphaerophtha/musbears, a hemispherical eye which projects beyond the
general line of the facial suture; it appears that this can only fit under the arched palpe-
bral lobe if the free cheeks slope steeply, as I have attempted to show in text-fig. 2. As
the cheek spine is bent down in relation to the general plane of the margin of the free
cheek it seems that, in life, these spines curved under the cephalon (text-fig. 2), in which
position they would prevent the trilobite from crawling on the sea floor; Henningsmoen
(1957, p. 78) has suggested that Sphaerophthalmus may have been a strong swimmer.

416

PALAEONTOLOGY, VOLUME 11

Sphaerophthalmus major Lake 1913

Text-fig. 3b; Plate 78, figs. 1-8

1913 Sphaerophthalmus major Lake, p. 77, pi. 8, figs. 7, 9-13; ? non fig. 8.

1949 Sphaerophthalmus major Lake; Edmonds, p. 60.

1952 Sphaerophthalmus major Lake; Hutchinson, p. 90, pi. 4, fig. 17?; non fig. 16 [= Ctenopyge
fletcheri (Matthew)].

1957 Sphaerophthalmus major Lake 1913; Henningsmoen, p. 217.

1957 Sphaerophthalmus humilis (Phillips 1848) \partim\ ; Henningsmoen, pi. 22, figs. 7, 11?,
15 only.

1957 Sphaerophthalmus minor [.y/c] ; Henningsmoen, p. 218 (error for S. major).

Lectotype. As lectotype I select the cranidium GSM 8903 (PI. 78, fig. 2) illustrated by Lake (1913,
pi. 8, fig. 7) which was collected from the White Leaved Oak Shales at White Leaved Oak, Malvern,
and presented to the Geological Survey by Miss M. Lowe.

Other material. Apart from BM I 14849 (the original of Lake’s pi. 8, fig. 10) which has not been seen,
all of Lake’s syntypes have been examined and are considered to be conspecific except possibly I 14851
(the original of Lake’s pi. 8, fig. 8); this consists of two free cheeks, probably of a Ctenopyge species,
and a thorax, possibly of S. major, upside-down and back-to-front in relation to them. Topotype
material in the collections of the Geological Survey Museum, the Sedgwick Museum, the Oxford
University Museum, and the British Museum (Natural History) has also been examined.

Diagnosis. A species of Sphaerophthalmus with palpebral lobes which do not project
laterally and whose centres are opposite or slightly behind the anterior ends of SI ; eye-
ridges weak or absent; inter-ocular cheeks more than half as wide as glabella (about
two-thirds as wide as glabella in shale-preserved specimens) at eye-line ; occipital spine
relatively long; free cheek semicircular with long spine springing from the middle of the
lateral margin; thorax with axial nodes and pleural spines, pleural regions about two-
thirds as wide as the axis; pygidial outline rounded behind, pleural regions about two-
thirds as wide as axis which is composed of about three rings.

Description. The general features of the cranidium are very like those of S. humilis.
Glabella prominent but commonly flattened in shale-preserved specimens, in which cases
the axial furrow may be deepened. SI oblique backwards, concave, joined across
glabella in an even curve; S2 very short or indistinct impressions in the sides of the
glabella. Occipital spine long but in many specimens only a tubercle is preserved.

EXPLANATION OF PLATE 78
Sphaerophthalmus species, X 6.

Pigs. 1 and 5 are of specimens from the Upper Dolgelly Beds near Dolgelly, Merioneth (Belt Collec-
tion); the remainder are from the White Leaved Oak Shales of the Malvern Hills.

Pigs. 1-8. S. major Lake. 1. Small cranidium showing long occipital spine, BM I 7619 partim (see
Fig. 5). 2. Lectotype cranidium, GSM 8903. 3. Cranidium showing mould of occipital spine, SM
A53115. 4. Cranidium, SM A53088. 5. Cephalon and thorax, paratype, BM I 7619 partim. 6, 7.
Free cheeks, SM A52564, 52562. 8. Thorax and pygidium, paratype, SM A509.

Figs. 9, 10. S. cf. a/atus (Boeck), external moulds of cranidium and free cheek, SM A53130.

Figs. 1 1-15. S. humilis (Phillips). 11,12. Cranidia, SM A53093,GSM 8963. 13. Cranidium from Phillips’s
Collection, Ox A287a. 14. Free cheek, SM A52546. 15. Thorax and pygidium figured by Lake,
GSM 8913.

Photographs by Mr. J. M. Pulsford.

Palaeontology, Vol. 11

PLATE 78

RUSHTON, Upper Cambrian trilobites

RUSHTON: REVISION OF TWO UPPER CAMBRIAN TRILOBITES 417

Anterior border short (sagittally), strongly arched in anterior view but probably only
slightly emarginate in dorsal view; in the lectotype it is strongly emarginate and up-
turned because compressed. Eye-ridges very weak on both internal and external moulds,
moderately oblique. Palpebral lobe narrow, arched up in lateral view, emarginate in
dorsal view; centres opposite or just behind the anterior ends of SI. Inter-ocular cheeks
about two-thirds of the glabellar width at the eye-line, but probably little more than half
in specimens preserved with their full convexity; post-ocular cheeks as wide as to four-
fifths the width of the occipital ring, but probably about two-thirds in convex specimens.

c

d

text-fig. 3. Diagrammatic drawings of flattened cranidia, free cheeks and pygidia of Sphaerophthalmus
species, about X 4. (a) S. humilis (Phillips), partly after Westergard. ( b ) S. major Lake (the pygidium
is partly conjectural), (c) S. majusculus Linnarsson; pygidium, associated cranidium (after Wester-
gard) and doubtfully attributed free cheek (after Henningsmoen). (d) S. alatus (Boeck), after Wester-
gard (pygidium unknown).

Free cheek semicircular with a curved spine up to twice as long as the genal field
springing from the middle of the lateral margin. Border of about even width. The
hemispherical eye is placed far back but distinctly forward of the border-furrow.

The thorax and pygidium have been adequately described by Lake, but it should be
noted that the pleural and axial spines are generally not preserved, and that the pygidial
flanks may bear more than one pair of furrows.

Comparison with other species (text-fig. 3). Compared with S. major S. humilis (text-fig.
3a; PI. 78, figs. 11-15) has narrower fixed cheeks; the inter-ocular cheeks in convex
specimens are less than half as wide as the glabella at the eye-line and even when

418

PALAEONTOLOGY, VOLUME 11

flattened they are only slightly more than half. The eyes are further back in S. humilis
and the eye-ridges are consequently more oblique, the occipital and genal spines are
shorter, the border of the free cheek tends to widen forwards, the pleural regions of
thorax and pygidium are narrower and the pygidium is triangular rather than rounded
behind. Measurements made on several specimens from the White Leaved Oak Shales
have shown that the width of the inter-ocular cheeks of S. humilis is 0-53 or less of the
width of the glabella, whereas that of S’, major is 0-63 or more.

S. alatus (text-fig. 3d; cf. PI. 78, figs. 9, 10) resembles S. major in having the palpebral
lobes opposite SI and has fixed cheeks of comparable width, but it differs in having
palpebral lobes that project laterally and are joined to the glabella by distinct eye-ridges;
the anterior ends of SI are further forward in relation to the length of the cephalic axis
in S. alatus than S. major , and the palpebral lobes are correspondingly further forward
on the cranidium. In S. alatus the anterior part of the fixed cheeks narrows forwards;
the occipital spine is shorter than that of S. major and the cheek spine is set further
back; the posterior part of the margin of the free cheek is nearly straight, whereas in
S. major it is curved.

S. majusculus (text-fig. 3c) has wider fixed cheeks than S. major, distinct eye-ridges,
palpebral lobes that project slightly laterally, and a node in place of the occipital spine.
The free cheek tentatively assigned to the species by Henningsmoen (1957, p. 219) has
a shorter spine set further back than that of S. major. The pygidium of S. majusculus
is relatively wider than that of S. major and is triangular rather than rounded behind.

Measurements in mm.

Length of cranidium

GSM 8903
(lectotype)
4-2

SM A53088
c. 3-4

Width of cranidium at eyes

5 0

4-1

Width of cranidium posteriorly

6-6

5-6

Length of cephalic axis

3-5

2-9

Width of cephalic axis

2-6

1-9

Sphaerophtlialmiis major probably reached a total length of about 10 mm.

Occurrence. Sphaerophthalmus major is found within the Zone of Peltura scarabaeoides
in England, Wales, Sweden, possibly in Nova Scotia (Canada), and apparently also in
Norway; such evidence as there is suggests that it occurs in the Subzone of Ctenopyge
linnarssoni. In discussing these occurrences reference is made to Henningsmoen’s (1957,
p. 39) subdivisions of the P. scarabaeoides Zone:

(Peltura par ado xa Subzone
Parabolina lobata Subzone
Ctenopyge linnarssoni Subzone
Ctenopyge bisulcata Subzone

In Scandinavia Ctenopyge bisulcata (Phillips) and C. linnarssoni Westergard are known
only from their own subzones, whereas C.pecten (Salter), P. scarabaeoides (Wahlenberg),
and Sphaerophthalmus humilis (Phillips) occur in both subzones.

England. The lectotype of S. major is from the White Leaved Oak Shales of the Malvern
Hills. The detailed stratigraphy of these Shales is unknown and though the presence of

RUSHTON: REVISION OF TWO UPPER CAMBRIAN TRILOBITES 419

the C. bisulcata Subzone is indicated by the association on single pieces of shale of
C. bisulcata, C. pecten, S. humilis , and P. scarabaeoides, the presence of the C. linnarssoni
Subzone is not proved. S. major has not been seen in association with undoubted frag-
ments of C. bisulcata, but occurs with C. pecten and fragments of a Ctenopyge species
recalling C. bisulcata, C. linnarssoni, or C. falcifera Lake.

In the Geological Survey Merevale No. 1 Borehole, which penetrated the Monks
Park Shales (Upper Stockingford Shales) in Warwickshire (Kellaway 1966), poorly
preserved specimens of S. major were collected at depths of 190 ft. 6 in. and 193 ft. 0 in.,
from horizons slightly higher than those yielding S’, humilis and C. bisulcata (depths of
205 ft. 6 in. and 206 ft. 6 in.).

Wales. One paratype of S. major is from the Upper Dolgelly Beds of the Dolgelly area
in North Wales; the precise locality and zonal position are unknown. Lake records
S. major from Moel Gron, near Dolgelly.

Sweden. A block of Upper Cambrian bituminous limestone in the Sedgwick Museum,
collected by Mr. S. C. A. Holmes from Vilske-Klefua, Vastergotland, Sweden (SM
A56378-87), carries an association of S. major with Ctenopyge fletcheri (Matthew) and
Peltura scarabaeoides westergaardi Henningsmoen, suggesting a horizon near the junc-
tion of the Subzones of C. linnarssoni and Parabolinci lobatci (see Henningsmoen 1957,
p. 39).

Norway. Henningsmoen (1957, pi. 22, figs. 7, 15) illustrated two cranidia from the Sub-
zone of C. linnarssoni at Viul, Ringerike, Norway, under the name of Sphaerophthalmus
humilis ; the fixed cheeks are too wide and the occipital spine rather too large for that
species and they are more probably cranidia of S. major.

Nova Scotia. Hutchinson (1952, p. 90) records S. major from the upper part of the
MacNiel formation ( Peltura Zone), and figures two specimens on his plate 4; one (fig. 16)
is probably C. fletcheri incorrectly identified as S. major but the other (fig. 1 7) may be
S. major (Henningsmoen 1957, pp. 205, 218).

Acknowledgements. I thank Mr. A. G. Brighton (Sedgwick Museum, Cambridge), Dr. W. T. Dean
(British Museum [Natural History]), and Mr. H. P. Powell (Oxford University Museum) for lending
me specimens in their care. Sir James Stubblefield, F.R.S., and Mr. R. V. Melville kindly read the
manuscript and made many helpful suggestions. This paper is published with the permission of the
Director, Institute of Geological Sciences.

REFERENCES

bell, w. c., and ellinwood, h. l. 1962. Upper Franconian and Lower Trempealeauan Cambrian
trilobites and brachiopods, Wilberns Formation, Central Texas. J. Paleont. 36, 385-423, pis. 51-64.
belt, t. 1868. On the ‘Lingula Flags' or ‘Festiniog Group’ of the Dolgelly District. Part III. Geol.
Mag. 5, 5-11, pi. 2.

berg, r. r. 1953. Franconian trilobites from Minnesota and Wisconsin. J. Paleont. 27, 553-68, pis.
59-61.

boeck, c. 1838. Uebersicht der bisher in Norwegen gefundenen Formen der Trilobiten-Familie; in
Keilhau: Gaea Norvegica, I, 138-45. Christiania.

cobbold, e. s. 1927. The Stratigraphy and Geological Structure of the Cambrian Area of Comley
(Shropshire). Quart. J. geol. Soc. Lond. 83, 551-73, pi. 43.

420

PALAEONTOLOGY, VOLUME 11

edmonds, j. e. 1949. Types and figured specimens of Lower Palaeozoic Trilobites in the University
Museum, Oxford. Geol. Mag. 86, 57-66.

frederickson, E. a. 1949. Trilobite fauna of the Upper Cambrian Honey Creek Formation. J. Paleont.
23, 341-63, pis. 68-72.

Harrington, H. j., et al. 1959. Treatise on Invertebrate Paleontology, Part O, Arthropoda 1. Univ.
Kansas Press.

henningsmoen, G. 1957. The Trilobite family Olenidae. Skr. norske Vidensk.-Akad., Mat.-naturv. Kl.

I, 1-303, pi. 1-31.

1960. The Middle Ordovician of the Oslo Region, Norway, 13. Trilobites of the family Asaphidae.

Norsk geol. Tidsskr. 40, 203-57, pis. 1-14.

Hutchinson, r. d. 1952. The stratigraphy and trilobite faunas of the Cambrian sedimentary rocks of
Cape Breton Island, Nova Scotia. Mem. geol. Surv. Canada, 263, 1-124, pis. 1-7.
kellaway, g. a. 1966. In Ann. Rep. Inst. geol. Sci. Part 1, Summ. Prog. geol. Surv. G.B. for 1965, 49.
lake, p. 1906-46. A Monograph of the British Cambrian Trilobites. Palaeont. Soc. [Monogr.]: (4),
1913, 65-88, pis. 7-10; (5), 1919, 89-120, pis. 11-14.
linnarsson, j. g. o. 1880. Om forsteningarne i de svenska lagren med Peltura och Sphaerophthalmus.

Sver. geol. Unders. Avh. Ser. C, 43, 1-31, pis. 1, 2.
lochman, Christina. 1950. Upper Cambrian faunas of the Little Rocky Mountains, Montana.

J. Paleont. 24, 322-49, pis. 46-51.

and wilson, j. l. 1958. Cambrian biostratigraphy in North America. J. Paleont. 32, 312-50.

Phillips, j. 1848. The Malvern Hills compared with the Palaeozoic Districts of Abberley, Woolhope,
May Hill, Tortworth, and Usk. With Palaeontological Appendix. Mem. geol. Surv. G.B. 2 (1),
1-386, pis. 1-30.

reed, f. r. c. 1900. Woodwardian Museum notes: On the British Species of the Genus Conocoryphe.
Geol. Mag. (4), 7, 250-7.

rushton, a. w. a. 1966. In Ann. Rep. Inst. geol. Sci. Part 1, Summ. Prog. geol. Surv. G.B. for 1965, 69.

1967. The Upper Cambrian trilobite Irvingella nuneatonensis (Sharman). Palaeontology, 10,

339-48, pi. 52.

salter, J. w. 1873. A Catalogue of the Collection of Cambrian and Silurian Fossils contained in the Geo-
logical Museum of the University of Cambridge. Cambridge.

Stubblefield, c. J. 1930. A new Upper Cambrian section in South Shropshire. Summ. Prog. geol. Surv.
G.B. for 1929— Part II, 54-62.

wells, a. k. 1925. The Geology of the Robell Fawr District (Merioneth). Quart. J. geol. Soc. Loud. 81,
463-538, pis. 30-32.

westergArd, a. h. 1922. Sveriges Olenidskiffer. Sver. geol. Unders. Avh. Ser. Ca, 18, 1-205, pis. 1-16.

A. W. A. RUSHTON

Institute of Geological Sciences
Exhibition Road
London, S.W. 7

Typescript received 25 February 1967

PROBABLE ANGIOSPERM POLLEN FROM
BRITISH BARREMIAN TO ALB IAN STRATA

by ELIZABETH M. KEMP

Abstract. Two species of the genus Clavatipollenites Couper are described. One is a redescription of the type
species, Clavatipollenites luighesii, from the Upper Barremian part of the Wealden Series. The other, Clavati-
pollenites rotundas, is a new species from the Middle Albian Lower Gault. Three species of tricolpate angio-
spermous grains are described. Only one of these, Tricolpites albiensis, from the Upper Albian, occurs in
sufficient quantity to be formally named as a new species. These forms constitute the earliest record of recogniz-
able angiosperm grains from England.

During an investigation into the spore and pollen assemblages of the marine Lower
Greensand and Gault of southern England, several pollen species were recovered which
are of particular botanical and stratigraphical interest. Two of these are monosulcate
pollens referable to the genus Clavatipollenites Couper; the others are tricolpate
angiospermous forms, the earliest undoubted pollen of this group from England. Of
these last forms, only one was recovered in sufficient quantity to permit formal designa-
tion as a new species. The monosulcate pollens described here are of interest in that they
possess some angiosperm characters and their appearance slightly predates that of
grains of undoubted angiospermous affinity.

Acknowledgements. Research facilities in the Department of Geology, Sedgwick Museum, Cambridge,
were made available by Professor O. M. B. Bulman. Mr. N. F. Hughes took a constant interest in the
course of this work and provided the sample from the Kingsclere Borehole. Financial assistance was
provided by a British Commonwealth Scholarship from 1963 to 1966.

LOCATION AND STRATIGRAPHY OF SAMPLES

The samples yielding the type material of the new species described here are part of
a more extensive collection of Lower Greensand, Gault and Upper Greensand. The
systematic and stratigraphic palynology of these sediments has been investigated in
detail for an unpublished thesis. The sediments are of Aptian and Albian age, with a
rich fauna which permits their subdivision in terms of ammonite zones.

The Lower Greensand samples have been collected chiefly from the Isle of Wight,
with minor sampling in the Weald and in the northern sedimentary basin near Leighton
Buzzard. The Gault and Upper Greensand have been collected over a wider geographic
area, extending from near Leighton Buzzard, through sections in the northern Weald
to the Isle of Wight and the Dorset coast in the south and south-west. In addition, the
sample from 475 ft. in the Kingsclere Borehole, previously studied by Couper (1958),
has been re-examined as it provided the type material of the species Clavatipollenites
hughesii. This species persists into younger strata. The type sample comes from the
Wealden, and is probably of Upper Barremian age.

[Palaeontology, Vol. 11, Part 3, 1968, pp. 421-34, pis. 79 81.]

422

PALAEONTOLOGY, VOLUME 11

More detailed references to the samples used in this study and to the sections from
which they were collected are given below:

Sample FI 54 — Calcareous pale yellowish-grey clay with shell fragments. Gault at Ford Place, Wrotham
(Nat. Grid, reference 51/636591). Bed 20, near base; niobe Subzone of the dentatus Zone; Middle
Albian. The detailed stratigraphy of this clay pit is given by Milbourne (1963), who subdivided the
section into 82 beds on the basis of lithology.

Sample K475 — Medium-grey siltstone with small dark patches, possibly plant fragments. Wealden
Series, 475 ft. in the D'Arcy Exploration Company's Kingsclere No. 1 Borehole; probably Upper
Barremian. Further details of the palynology and stratigraphic position of this sample are given by
Hughes (1958) and Couper (1958).

Sample F312 — Calcareous friable pale yellowish-green silty sandstone. Upper Greensand, west side
of Lulworth Cove, Dorset (Nat. Grid reference 30/825799). 18 ft. beneath the base of the Exogyra Rock,
which underlies a series of stone bands at the top of the Upper Greensand and serves as a convenient
local marker (Wright, in Arkell 1947). Beds above the Exogyra Rock can be referred to the dispar Zone.
The zonal position of the beds below is less certain, but they are probably of Upper Albian age.

Sample F270 — Grey and cream mottled sandstone with ferruginous pebbles and patchy iron staining.
Munday’s Hill pit, Leighton Buzzard (Nat. Grid reference 42/930265). 6 in. beneath base of Carstone
Breccia in the upper part of the Woburn Sands. The precise age of the sample is unknown. The Car-
stone Breccia is partly of Lower Albian age; the sample in question comes probably from the Lower
Albian.

Location of type material. Rock samples, strew slides and individual mounts made from preparations
are lodged in the Sedgwick Museum, Cambridge. Slides are deposited in the Palynology Collection of
that museum. Formally named species are based on examination of 100 specimens in each case, and
all of these are available for restudy. The co-ordinates quoted are those of the Leitz Dialux microscope
Serial No. 526724 in the Sedgwick Museum.

Preparation procedures. Spore assemblages were extracted by solution of silicates in cold
hydrofluoric acid following the removal of carbonates, where necessary, with 10-15 per
cent. HC1. Residues were oxidized in Schulze solution, two short periods of oxidation
usually being used. The total oxidation time did not in any case exceed twenty minutes.
Residues were briefly cleared in less than 1 per cent. NH4OH and the microfossils
separated from the remaining mineral debris by centrifuging in zinc bromide solution.

SYSTEMATIC PALYNOLOGY

Anteturma pollenites Potonie 1931
Turma plicates Naumova 1939 emend. Potonie 1960
Subturma monocolpates Iversen and Troels-Smith 1950
Genus clavatipollenites Couper 1958

1958 Clavatipollenites C ouper, p. 159.

1961 Retimonocolpites Pierce, p. 47.

Type species. Clavatipollenites liughesii Couper 1958, p. 159, pi. 31, figs. 19-22.

Diagnosis. As given by Couper (1958).

Remarks. On a morphological basis Clavatipollenites resembles Liliacidites Couper
(1953, p. 56) which was proposed to incorporate monosulcate grains of reticulate exine
pattern, or, more precisely, ‘for the reception of fossil pollens of liliaceous affinities that

E. M. KEMP: PROBABLE ANGIOSPERM POLLEN

423

cannot be more accurately placed’. The species originally referred to Liliacidites formed
part of an Upper Cretaceous and Tertiary complex which had a large component of
grains of unquestionably angiospermous origin. It would seem that the retention of a
separate genus is valuable for Lower and lower Upper Cretaceous forms the affinity of
which is in question.

Species from several widely separated localities have been referred to Clavatipollenites.
Not all of these, however, conform to the diagnosis of the genus. Pocock (1962, p. 74,

a b

c

text-fig. 1. Grain orientation in Clavatipollenites. L = length, W = width, D = depth. Designa-
tions P, Ex, and E3 correspond with those of Couper (1958) for describing monosulcate and tricolpate

pollens. Orientation: a, polar view; b, equatorial longitudinal view; c, equatorial transverse view.

pi. 12, figs. 190-2) described a large form which he named Clavatipollenites couperi ,
from the Upper Jurassic-Neocomian Vanguard Formation of western Canada. The
specimens he figured do not show the presence of a tectate exine convincingly, and the
species may perhaps be more accurately referred to Monosulcites or Ginkgoeyeadophytus.
Helal (1965, pi. 17, fig. 19; 1966, pi. 34, figs. 54, 55) recorded both C. couperi and C.
hughesii from subsurface Jurassic sediments in the Western Desert, Egypt. Both of the
grains figured by this author appear to be distinctly three-furrowed, suggesting Eucoin-
miidites rather than Clavatipollenites.

The type species, C. hughesii , has been recorded by Brenner (1963) together with a
smaller form which he designated C. minutus, from Potomac Group sediments in Mary-
land. Von der Brelie (1965, p. 151, pi. 13, figs. 6-8) found C. hughesii in sediments of
probable Aptian-Albian age in Germany. Archangelsky and Gamerro (1967) reported
the presence of the species in assemblages of probable Barremian age in southern
Argentina.

Pierce (1961, p. 47, pi. 3, fig. 87) described a monosulcate pollen from Cenomanian
clays and lignites in Minnesota. This species, which he named Retimonocolpites dividuus,

424

PALAEONTOLOGY, VOLUME 11

Pierce designated as type species of the genus Retimonocolpites. This genus is regarded as
synonymous with Clavatipollenites on the basis of Pierce’s definition.

Both the type species, C. hughesii, and a new species, C. rotundus, were recorded from
samples investigated during the present study. The expanded description of C. hughesii
given below is based on re-examination of specimens from Couper’s type sample.

The terminology of the grain dimensions and orientations used in the descriptions is
shown in text-fig. 1 . Comparisons of the size and shape of grains assigned to the two
species are shown in the scatter diagram text-fig. 4, histograms of size frequency dis-
tributions in text-fig. 5.

text-fig. 2. Diagrammatic representation of Clavatipollenites rotundus sp. nov. Large figures, X 2,000.
A, proximal face, LO pattern, position of sulcus dotted, b, distal face, left-hand side showing reticulum,
right-hand side showing exine stratification in section. Dark zone adjacent to the sulcus is stippled,
c, equatorial longitudinal view, d, equatorial transverse view.

Clavatipollenites rotundus sp. nov.

Plate 79, figs. 1-19; Plate 80, figs. 1-8; text-fig. 2

Diagnosis. Monosulcate pollen grains of subcircular to elliptical outline, with a dis-
tinctly tectate exine. The sulcus is conspicuous, round-ended, and bordered by thicken-

All figures at magnification X 1,000 unless otherwise stated.

Figs. 1-19. Clavatipollenites rotundus sp. nov. All specimens from sample F154, niobe Subzone of the
dentatus Zone, Lower Gault at Ford Place, Wrotham. 1-5, FI 54/2; 35-0, 92-7. 1, Median focus
showing tectate exine in section at equator (x 2,000). 2, Fligh focus on sulcus. 3, 4, Deeper foci.
5, Proximal focus on reticulum with sexine partly detached. 6, 7, F154/1 1 ; 27 0, 108-2; grain obliquely
compressed. 6, Deep focus. 7, High focus showing end of sulcus and darkened zone. 8, FI 54/Separate
15, 40 0, 97-6; grain showing pronounced infolding associated with sulcus. 9-11, FI 54/Separate 1 ;
38-3, 99 2. 9, Proximal focus showing reticulum. 10, Median focus on sulcus and associated struc-
ture. 11, Distal focus. 12, 13, FI 54/10 ; 34-2, 90-2. 12, High focus on proximal face, sexine partly
removed. 13, Equatorial focus. 14, 15, FI 54/12; 35-8, 89-8. 14, Median focus of specimen in end-on
view. 15, High focus on reticulum. 16, F154/Separate 8; 39 0, 98-1; median focus showing folding
adjacent to sulcus. 17-19, FI 54/Separate 9; 39-2, 98-5; Holotype. 17, High focus on distal face.
18, Median focus on sulcus with exine stratification distinct at equator. 19, Proximal focus.

EXPLANATION OF PLATE 79

Palaeontology , Vol. 11

PLATE 79

KEMP, Mid-Albian pollen

E. M. KEMP: PROBABLE ANGIOSPERM POLLEN

425

ings or infoldings of the nexine. The reticulate pattern of the sexine is formed from the
fused tips of bacula; the mesh diameter is from 0-7 to 2-0 p. The mean grain width lies in
the range 22-27 p, mean grain length 25-30 p. Sexine shows a tendency to separate
from nexine.

Holotype. Sample F154, separate mount 9; 39-2, 98-5. Plate 79, figs. 17-19. Lower Gault, Ford Place,
Wrotham; Middle Albian. Grain orientated with sulcus uppermost, extending full length of grain,
bordered by darkened zone 2-4-2-5 p wide on either side. Sulcus slightly expanded and rounded at
extremities. Nexine 08 p thick, sexine 1-2 /a. Length x width 26x25 p.

Dimensions. (Measured on 100 specimens from Sample F154; see text-fig. 5.) Grain length 20(27)32 yu.,
grain width 18(25)31 /a, grain depth 17(22)28 /a.

Orientation. (For grain positions see text-fig. 1.) Grains in position a — 68 per cent.,
position b — 7 per cent., position c — 25 per cent.

Description. The original grain shape, as deduced from measurements and from grain
orientations appears to have been subspheroidal to oblate spheroidal. The ratio grain
length to grain width is shown in the scatter diagram text-fig. 4.

The sulcus extends for almost the full length of the grain and is parallel-sided through-
out most of its length, but frequently shows slight expansion at the extremities. The ends
are usually rounded. Most specimens show a darkened zone 1 -5-2-0 p wide on either
side of the sulcus throughout its length. This either represents a thickening of the
nexine adjacent to the sulcus or a doubling of this layer caused by infolding consequent
on rupture of the grain along the sulcus. Optical sections of grains (text-fig. 3) suggest
that the last interpretation is more likely to be correct. Some specimens, for example,
that illustrated in text-fig. 3b , show a small gap in the thickened nexine layer adjacent to
the sulcus, suggestive of breakage during infolding and compression. These sections also
demonstrate that the sexine is not involved in the infolding.

The stratification of the exine is distinct. The nexine is 0-8-T8 p thick. The bacula
arising from this layer expand and coalesce at their tips to form the reticulum of the
ectosexine. The meshes of the reticulum are irregularly polygonal, separated by muri
0-3-0-5 p wide, which are formed from single rows of fused bacula. The meshes are gene-
rally of uniform size over the entire grain surface. The sexine characteristically separates
from the nexine; separation is usually more extreme at the grain ends. Some grains were
observed in which the sexine had been entirely removed.

Remarks. Clavatipollenites rotundas appears similar to Pierce's species Retimonocolpites
dividuus , although the single illustration given by Pierce (1961, pi. 3, fig. 87) makes
comparison difficult. Pierce does, however, mention the tendency for the exine layers to
separate. Brenner (1963) recombined Pierce’s species with Liliacidites, with little
descriptive comment. The specimens from the Patapsco Formation show some evidence
of infolding adjacent to the sulcus but not to the degree seen in the English forms. The
American species appears less spherical and thinner and consequently tends to fold more
readily than its English counterpart.

Distribution. In England C. rotundas is confined to Albian sediments. Its earliest recorded appearance
is in Sample F270, which is probably of Lower Albian age. This occurrence is the only one observed
beneath the base of the Gault. The species was noted in greatest abundance in the vicinity of the niobe

C 5586 F f

426

PALAEONTOLOGY, VOLUME 11

Subzone at Wrotham and at Folkestone, although it persists throughout the Albian. These concentra-
tions, which do not in any case exceed 5 per cent, of the total spore and pollen content, may be due to
a facies factor as they are associated with a rich benthonic fauna and ferruginous bands.

The distribution in time of C. dividuus is similarly restricted in America. Brenner (1963) records the
species as being confined to the Patapsco Formation (Albian).

a

b

c

d

text-fig. 3. Optical sections of specimens of Clavatipollenites
rotundus sp. nov., drawn from photographs to show the structure
associated with the sulcus. All x 1,000. Nexine shown solid black.
a, FI 54/Separate 8; 39-0, 98T; nexine appears as thickening
adjacent to sulcus, b , FI 54/Separate 1 ; 38 3, 99-2; optical section
of grain in equatorial view. Slight obliquity causes sulcus to
appear in section at top of drawing. Sexine is detached, not in-
volved in infolding; the nexine adjacent to the sulcus appears
split. Darkened zone next to the remainder of sulcus shown
stippled, c, F 1 54/10 ; 36-6, 106 0; split grain showing a similar
structure to b. d, F154/2; 32 0, 100 1 ; grain in end-on view.

Clavatipollenites hughesii Couper emend.

Plate 80, figs. 9-19

Remarks. Couper (1958) gave no separate diagnosis for C. hughesii, so his description
has been emended slightly and is here restated as a diagnosis.

Emended diagnosis. Monosulcate pollen grain, sulcus extending full length of grain,
gaping in its central region and tapering slightly towards extremities. Equatorial outline
elliptical to subcircular. Exine consisting of an inner unsculptured nexine 0-5-1 -0 /x thick

E. M. KEMP: PROBABLE ANGIOSPERM POLLEN

427

and a sexine formed of baculate projections approximately 1 -0 p long, which either remain
discrete or fuse at their tips to form a microreticulum. Lumina of reticulum irregularly
polygonal, 1 -0-2-0 p in diameter.

• Clavatipollenites hughesii
+ C. rotundus

c/1

a

o

i_

u

E

30-

c

JZ

20-

+

+ + +

+ + 4+

+ +++4+

+ +-H-++ 44 +

+ 4 +

+ +

+ 4+4+444 +4+4-4 4
44 + +

+ + + +++ + +

• + + +

+ +

• • •• Rtt • + +

• V •• * + +

•••• •• • •

cn

c

<D

cn

1 1 T-

10 20 30

grain width in microns

text-fig. 4. Scatter diagram showing differences of shape and size between Clavatipollenites hughesii
Couper and C. rotundus sp. nov. Based on 100 specimens from each of the type samples, K475 Kings-
clere Borehole, and FI 54, Ford Place, Wrotham.

Holotype. Couper (1958, pi. 31, figs. 21, 22). Slide C128/9, specimen K5087; 35-8, 114-2. Kingsclere
Borehole at 475 ft. Grain orientated with sulcus uppermost. Length x width 28-Ox 17-5 p.

Dimensions. (Measured on 100 specimens from the type sample.) Histograms of size frequency dis-
tributions are shown in text-fig. 5. Grain length 12(18)30 p, width 8(16)19 p , depth 12(13)17 p.

Description. The outline of the grains is variable, ranging from almost circular to strongly
elliptical. Margin of sulcus often ragged. The sulcus is frequently indistinct, or repre-
sented as an indeterminate tear in the exine. The grains are often folded with the long
axes of the folds parallel to the sulcus.

428

PALAEONTOLOGY, VOLUME 11

A

text-fig. 5. Histograms showing frequency distribution of grain lengths and widths in a, Clavatipollenites
hughesii and b, C. rotundus sp. nov. Graphs smoothed using a three-point moving average.

EXPLANATION OF PLATE 80

All figures at magnification x 1,000, unless otherwise stated.

Figs. 1-8. Clavatipollenites rotundus sp. nov. All specimens are from sample FI 54, niobe Subzone of
the dentatus Zone. Lower Gault at Ford Place, Wrotham. 1, 2, F154/2; 32-2, 100-1. 1, Median focus,
end-on view. 2, High focus on reticulum and extremity of sulcus. 3, FI 54/10 ; 36-6, 106-0; specimen
split along sulcus, infolded zone adjacent to sulcus visible. 4, FI 54/7; 27-3, 91-8; sexine has been
partly removed. 5, F154/7; 26-6, 101-1; sexine partly removed. 6, F154/Separate 4; 39-9, 97-3; part
of grain in equatorial focus, sexine loose. 7, 8, FI 54/4; 39-9, 94-2; specimen from which sexine has
been completely lost. 7, High focus showing dark zone adjacent to sulcus. 8, Median focus.

Figs. 9-19. Clavatipollenites hughesii Couper. 9, 10, Slide BP174A (IX); 60 0, 104-4 (N. F. Hughes
Coll. No.). 9, High focus showing bacula in surface view. 10, Deep focus showing sulcus and exine
stratification. Kingsclere Borehole at 475 ft. 11, 12, F154/1 ; 26-7, 107 0. 11, Median focus showing
sulcus and exine stratification. 12, The same at x2,000. 13, F154/5; 28-1, 92-5; median focus,
sulcus and exine stratification shown. Ford Place, Wrotham; Middle Albian. 14, 15, K475/2; 23-2,
101-1. 14, Equatorial focus showing bacula in section. 15, High focus showing surface pattern;
Kingsclere Borehole at 475 ft. 16, 17, F154/2; 54-2, 92-8. 16, High focus showing sulcus. 17, Equa-
torial focus showing exine stratification. Ford Place, Wrotham; Middle Albian. 18, FI 54/5; 23-9,
90-0; median focus on compressed grain. Ford Place, Wrotham. 19, F099/1 ; 55 0, 104 0; high focus
on grain with irregular, gaping sulcus. Compton Bay, Isle of Wight; Middle Albian.

Palaeontology, Vol. 11

PLATE 80

KEMP, Albian and Barremian pollen

E. M. KEMP: PROBABLE ANGIOSPERM POLLEN

429

Remarks and comparison. The dimensions quoted are smaller than those quoted by
Couper and cover the range of the species C. minutus Brenner, which may be conspecific
with C. hughesii. The histograms of size-frequency distribution do not suggest any
ready subdivision of the species on a size basis. The larger size quoted by Couper may
be due to some extent to the maceration technique employed. The specimens measured
in the present study were isolated from a fresh preparation of the type material which
was subjected to only fifteen minutes oxidation time in contrast to the minimum of six
hours quoted by Couper.

C. hughesii and C. rotundas are readily separable on the basis of their size ranges and
on the shape of the grains. Text-fig. 4 illustrates these differences. In C. hughesii most
grains are longer than they are wide. C. rotundas shows greater variation in the ratio
length to width, with slightly less than half the grains being wider than long.

In C. rotundas the sulcus is a conspicuous feature, rounded at the extremities and
modified by probable infolding at its margins. In some of the grains referred to C.
hughesii the sulcus is barely discernible. Where visible it is a ragged opening, tapering
towards its extremities and lacking the marginal rigidity of that in C. rotundas.

A third feature which separates the species is the degree of development of the ecto-
sexinous reticulum. The bacula of the sexine in C. rotundas almost invariably unite at
their tips to form the muri of an unbroken reticulum; in C. hughesii they often remain
discrete or unite in short broken rows only to form an imperfect reticulum.

Distribution. The first occurrence of Clavatipollenites hughesii in the Barremian has been recorded by
Hughes (1958). In about half the Lower Greensand samples examined during the present study it
occurred in concentrations somewhat less than 1 per cent, of the total spore and pollen assemblages.
It persists throughout the Gault to Upper Albian horizons, although with a much lower sample fre-
quency. The highest recorded concentrations occur near the top of the Atherfield Clay (up to 10 per
cent.).

Affinity o/Clavatipollenites. The feature of greatest botanical interest in fossil pollen of
the Clavatipollenites type is the presence of a tectate exine. It was this feature which
led Couper (1958) to suggest an angiospermous origin. He compared C. hughesii to
grains of the extant dicotyledon Ascarina lucida and also drew attention to structural
similarities of pollen of certain members of the Liliaceae. C. hughesii represents, to
present knowledge, the earliest record of this type of exine organization, although it
may be argued that parallels exist among disaccate pollens. Clavatipollenites first
appears in spore and pollen assemblages which lack any other obvious angiosperm
elements, a fact which justifies close attention being paid to the genus.

In spite of the angiospermous character of some of the features of Clavatipollenites
it still remains possible that the grains were produced by a member of some extinct
gymnosperm group. The over-all shape of the grains resembles that of both living and
fossil cycadophyte pollen. The form of the sulcus, particularly where this is rounded at
its extremities, is closer to that of living cycad pollens than to most angiosperm pollen.

The orientation of the grains observed in C. rotundas shows a pronounced tendency
towards proximo-distal compression. This suggests that the proximo-distal axis E2
(designations as used by Couper, 1958, pp. 162-3) is shorter than the other equatorial
axis Ex. This observation indicates a similarity of shape to that of the grains of Andro-
strohus manis and Gingko liuttoni as these were analysed and presented by Couper.

430

PALAEONTOLOGY, VOLUME 11

It is apparent that species referred to Clavatipollenites are characterized by features
suggesting the possibility of either angiosperm or gymnosperm parentage. All discussion
of affinity is, however, purely speculative, and the problem can only be resolved by
discovery of the grains in association with other reproductive structures.

Subturma triptycha Naumova 1939

Remarks. Potonie (1960, p. 92) has treated Naumova’s category as a subturma, although
it was originally intended as a broadly defined form genus. Potonie has remarked that
the category is used as Tricolpites in Erdtman’s (1947) sense; i.e. as a broad supra-
generic category incorporating tricolpate grains of indeterminate affinity.

Genus tricolpites Cookson ex Couper 1953

Type species. Tricolpites reticulatus Cookson 1947, p. 134, pi. 15, fig. 45. (Subsequent designation,
Couper 1953.)

Remarks. The use of a broadly circumscribed form genus for early tricolpate pollens is
considered adequate and non-committal. The diagnosis of Tricolpites given by Couper
(1953, p. 61) is as follows: ‘Grains free, anisopolar, tricolpate. Exine variable in thick-
ness and sculpture. Size variable.’

Other authors dealing with early tricolpate pollens have referred them to Tricolpopol-
lenites Thomson and Pflug (see Groot and Penny 1960, Brenner 1963). The diagnosis of
this form genus implies the presence of some form of germinal structure in the equatorial
region of the colpi, although it is not precise as to the form of this feature; ‘Exitus nur
als meridionale Colpen ausgebildet’ (Thomson and Pflug 1953, p. 95).

Pierce (1961) referred Cenomanian tricolpate grains to Retitricolpites van der Ham-
men, but this genus is not regarded as valid under the International Rules of Nomen-
clature since the type of R. vulgaris, the type species, is a grain of the extant species
Neea macrophylla.

Tricolpites albiensis sp. nov.

Plate 81, figs. 1-22. text-fig. 6

Diagnosis. Small (less than 20 p diameter) tricolpate pollen grains in which the ratio of
polar to equatorial diameter lies between 2 : 1 and 1:1. The bacula of the sexine form an
imperfect microreticulum. The total exine thickness is around 1-0 p.

Holotype. Sample F312/4; 46-8, 90 0. Plate 81, figs. 2-4. From 18 ft. below the base of the Exogyra
Rock, Upper Greensand, Lulworth Cove, Dorset. Probably Upper Albian, zonal position uncertain.
Grain in obliquely polar orientation, with three colpi almost meeting at pole. Grain diameters 16 X
17 p. Exine slightly less than 1 p thick, microreticulate.

Dimensions. Equatorial diameter 7 0(1 1 -4)14-5 p, polar diameter 10-5(1 3-5)1 7-5 p, ratio of polar to
equatorial diameter 1-0(1 -35) 1-65 (based on measurement of 100 specimens from the type sample).

Orientation. Grains in polar view — 73 per cent., equatorial view — 19 per cent., oblique
view — 8 per cent.

Description. Grains are tricolpate, trilobate and radially symmetrical in polar view, oval
in equatorial view. The colpi are narrow and the acolpia small in relation to the rest of

E. M. KEMP: PROBABLE ANGIOSPERM POLLEN

431

the grain; rare grains are almost syncolpate. The margins of the colpi are slightly
ragged. The sexine and nexine are of approximately equal thickness. Diameter of meshes
of microreticulum 0-3-04 p.

Comparison. In its general form and size Tricolpites albiensis
resembles Tricolpopollenites micromunus Groot and Penny.

The original description of the species was brief, so close
comparisons are difficult. T. micromunus , however, was
described as having a coarsely reticulate exine, which justifies
the separation of the English species from it. Brenner (1963,
p. 93, pi. 39, fig. 7; pi. 40, fig. 1) gave a fuller description of
specimens of T. micromunus from the Patapsco Formation,
for which he quoted a mean polar diameter of 16 p, and an
equatorial one of 13 p, which indicates that the American
species is slightly larger.

It is possible that the species Tricolpites minutus Brenner
is conspecific with T. albiensis although it appears to be
slightly smaller and more spherical than the last species.

Affinity. Speculation as to the affinity of dispersed angio-
spermous grains of the generalized form of Tricolpites can-
not be meaningful. Brenner (1963) compared the pollen of
the extant species Tetracentron sinense to the dispersed
species Tricolpopollenites micromunus. Similar small tricol-
pate grains occur in the families Trochodendraceae, Ranun-
culaceae, Salicaceae, etc.

Distribution. No specimens of T. albiensis have been recorded from
the Lower Greensand. Rare specimens have been recovered from the
Gault (Middle and Upper Albian). In the type locality, the Upper
Greensand at Lulworth Cove, the species represented 15 per cent, of
the total spore and pollen assemblage.

Tricolpites sp. 1
Plate 81, figs. 25, 26

Description. Grains tricolpate, oval in equatorial view, polar views not observed. Colpi
distinct, open to a width of 2-3 p in their equatorial regions and acutely pointed at their
extremities. A slight thickening of the exine is visible at the margins of the colpi. The
ectosexine forms a reticulum in which the meshes are irregular in shape, 1-3 p in
diameter. The sexine tends to detach and stand away from the grain in compressed
specimens.

Dimensions. Polar diameter 21-23 p, equatorial diameter 18-21 p (3 specimens only).

Remarks and comparisons. This grain does not resemble any previously described Creta-
ceous form, but its rarity precludes its designation as a new species. Brenner (1963, p. 91,
pi. 38, figs. 6-7) referred tricolpate forms with a coarse reticulum to Retitricolpites

text-fig. 6. Diagrammatic
representation of Tricolpites
albiensis sp. nov., x 2,000. a,
polar view with exine shown
in section on left, nexine
black. Surface view of reti-
culum on right, LO pattern.
b, equatorial view, exine in
section on left, one colpus
shown in centre. Reticulum
in surface view on right.

432

PALAEONTOLOGY, VOLUME 11

georgensis Brenner, but that species is larger than the English form and has a finer, more
closely appressed reticulum. Large reticulate forms from the Patapsco Formation were
assigned to R. geranioides (Couper) by Brenner (pi. 38, fig. 8; pi. 39, fig. 1), a species
originally described from the Miocene of New Zealand. This form, in addition to being
much larger than the British species, has a more complex reticulum.

Distribution. From Sample F312 only, Lulworth Cove, Upper Albian.

Tricolpites sp. 2

Plate 81, figs. 23, 24

Description. Grains tricolpate, oval in equatorial view, trilobate and radially symmetrical
in polar view. Colpi simple, margins smooth, extremities pointed. Exine baculate, bacula
expanded at their extremities and fused to form a microreticulum of mesh diameter
0-5-1 -0 p. Sexine approximately 0-8-1 -0 p thick, i.e. about twice as thick as nexine.

Dimensions. Polar diameter 21-27 p, equatorial diameter 18-24 p (5 specimens).

Comparison. The grains described above are fairly close to those designated Tricolpopol-
lenites virgeus Groot, Penny, and Groot. Brenner (1963, p. 92, pi. 39, figs. 4, 5) reassigned
the species to Retitricolpites although the specimens described by Brenner possess a
finer reticulum than that described by the original authors.

Distribution. On the basis of five specimens distributional records are only provisional. The chief
interest of the grain lies in the fact that it appears to be the earliest occurrence of tricolpate grains
encountered during this study. It has been recorded only from Leighton Buzzard, in the uppermost
sample from the Woburn Sands (Sample F270) and in one sample from the overlying Gault. The first
occurrence is thus in the Lower Albian. Associated with the tricolpate grains are well-preserved speci-
mens of Clavatipollenites rotundus.

EXPLANATION OF PLATE 81

All figures at magnification x 1 ,000, unless otherwise stated.

Figs. 1-22. Tricolpites a/biensis sp. nov. All specimens are from Sample F312, Upper Greensand,
Lulworth Cove, Dorset: Upper Albian. 1, F312/4; 49 8, 90-8 ; equatorial focus. 2-4, F312/4: 46 8,
90 0; Flolotype, grain slightly oblique. 2, High focus showing microreticulum. 3, Focus on pole
showing colpi almost meeting. 4, The same, X 2,000. 5-7, F312/5; 37 0, 99-7. 5, Equatorial focus.
6, Higher focus approaching pole. 7, Equatorial focus showing bacula, x 2,000. 8, 9, F312/4; 43-6,
964; specimen torn along one colpus. 8, Near equatorial focus. 9, High focus. 10, 11, F312/5; 20-0,
102-0. 10. High focus close to pole. 11, Equatorial focus. 12, 13, F312/5; 24-9, 95-0. 12. Equatorial
focus. 13, High focus on polar area. 14, 15, F312/5 ; 33-0,93-1 ; equatorial view. 14, Deep focus on two
colpi. 15, High focus on one colpus, microreticulum distinct. 16, 17, F312/5; 29-2, 93-2; obliquely
orientated specimen. 16, High focus showing two colpi. 17, Median focus showing exine stratifica-
tion. 18, 19, F312/4; 34-9, 89-4; polar view. 18, High focus. 19, Equatorial focus. 20, F312/5; 33-8, 98-4;
equatorial view, median focus. 21, 22, F312/3; 41-0, 90-4; equatorial view. 21, High focus on one
colpus. 22, Deep focus on two colpi, exine stratification visible.

Figs. 23, 24. Tricolpites sp. 2. F270/6; 58-8, 93-8; Munday’s Hill, Leighton Buzzard; Albian. 23, Focus
close to equator. 24, Higher focus close to pole.

Figs. 25, 26. Tricolpites sp. 1. F312/3; 471, 91-5; Lulworth Cove; Upper Albian. 25, High focus show-
ing one of the colpi and the coarse reticulum. 26, Deep focus on reverse face of grain.

Palaeontology, Vol. 11

PLATE 81

KEMP, Albian pollen

E. M. KEMP: PROBABLE ANGIOSPERM POLLEN

483

SUMMARY AND CONCLUSIONS

Tricolpate pollen grains of undisputed angiospermous origin first appear in England
in sediments of Middle Albian age (apart from one known occurrence in what is probably
the Lower Albian). They do not become abundant enough for formal description until
the Upper Albian. Their appearance is pre-dated, in the Aptian or slightly earlier, by
monosulcate grains referable to the genus Clavatipollenites. Grains belonging to this
genus possess a tectate exine which suggests that they have an angiospermous affinity,
although other features of the grains are more gymnospermous in nature.

Although the Lower Greensand (Aptian and Lower Albian) has been intensively
studied palynologically, it has not yielded any recognizable angiosperm pollens. This is
of interest since it was allegedly from this formation that Stopes (1912, 1915) described
petrefactions of angiosperm woods of an advanced type. These fossils have frequently
been cited as indicating a long, pre-Cretaceous history of development for the angio-
sperms as a whole. The absence of recognizable angiosperm pollen from the formation
is difficult to reconcile with the presence of highly evolved woods.

The early angiosperm grains which appear in the overlying formation are of simple
form, tricolpate, with a microreticulate sexine. The grains are all notably small, not
exceeding 20 /x in diameter. Their general lack of specialization and small size is shared
by late Lower Cretaceous angiosperm grains recorded from other parts of the globe,
notably from North America. This apparent lack of specialization in the earliest angio-
sperm grains does not readily fit in with the idea of a long pre-Cretaceous history for the
group. Rather, grain morphology in the observed earliest occurrences of angiosperm
grains is more compatible with the view of an early Cretaceous origin (see Hughes 1961 ).

REFERENCES

archangelsky, s., and gamerro, j. c. 1967. Spores and pollen types of the Lower Cretaceous in
Patagonia. Rev. Palaeobotan. Palynol. 1, 211-17.

arkell, w. j. 1947. The geology of the country around Weymouth, Swanage, Corfe and Lulworth.
Mem. Geol. Sitrv. U.K. 386 pp.

brelie, g. von der. 1964. Eine unterkretazische Mikroflora aus dem nordlichen Sauerland. Fortschr.
Geol. Rheinld. Westf. 12, 117-68.

brenner, g. j. 1963. The spores and pollen of the Potomac Group of Maryland. Bull. Md. Dept. Geol.
Mines , 27, 1-215.

couper, r. a. 1953. Upper Mesozoic and Cainozoic spores and pollen grains from New Zealand.
Palaeont. Bull. Wellington, 32, 1-77.

1958. British Mesozoic microspores and pollen grains. Palaeontographica, B 103, 75-179.

erdtman, g. 1947. Suggestions for the classification of fossil and recent pollen grains and spores.
Svensk. bot. Tidskr. 41, 104-14.

groot, j. j., and penny, j. s. 1960. Plant microfossils and the age of non-marine Cretaceous sediments
of Maryland and Delaware. Micropaleontology, 6, 225-35.
helal, a. h. 1965. Jurassic spores and pollen grains from the Kharga Oasis, Western Desert, Egypt.
Neues Jb. Geol. Palaont. Abb. 123, 115-219.

1966. Jurassic plant microfossils from the subsurface of Kharga Oasis, Western Desert, Egypt.

Palaeontographica , B 117, 83-98.

hughes, n. f. 1958. Palaeontologic evidence for the age of the English Wealden. Geol. Mag. 95, 41-49.

1961. Fossil evidence and angiosperm ancestry. Sci. Prog. Lond. 69, 193, 84-102.

milbourne, r. a. 1963. The Gault at Ford Place, Wrotham, Kent. Proc. Geol. Ass. 74, 55-81.

434 PALAEONTOLOGY, VOLUME 11

pierce, r. l. 1961. Lower Upper Cretaceous plant microfossils from Minnesota. Bull. Minn. geol.
Surv. 42, 1-86.

pocock, s. A. j. 1962. Microfossil analysis and age determination of strata at the Jurassic-Cretaceous
boundary in the western Canada plains. Palaeontographica, Bill, 1-95.
potonie, r. 1960. Synopsis der Gattungen der Sporae dispersae. III. Teil. Sporites, Pollenites. Beih.
geol. Jb. 39, 1-189.

stopes, m. c. 1912. Petrefactions of the earliest European angiosperms. Proc. R. Soc. B 203, 75-100.

1915. Catalogue of the Cretaceous plants in the British Museum (Nat. Hist.) London. Part II.

Thomson, p. w., and pflug, h. 1953. Pollen und Sporen des mittel-europaischen Tertiars. Palaeonto-
graphica, B 95, 1-138.

E. M. KEMP

West Australian Petroleum Pty. Ltd.
Perth

Western Australia

Manuscript received 20 March 1967

SINDULITES, A NEW GENUS OF THE
NUMMULITIDAE (FORAMI NIFERI D A)

by F. E. E AMES

Abstract. Operculina sindensis L. M. Davies 1927 is made the type of a new genus Sindulites (Nummulitidae).

The subgenus Chordoperculinoides Arni 1965 of the genus Nummulites was proposed
to incorporate all real ‘nummulites cordelees’, and has as its type species Operculina
bennudezi Palmer, a Caribbean Palaeocene species. However, Arni (1966) in discussing
his subgenus gave two plates of illustrations of what he regarded to be ‘ Chordopercu-
linoides bennudezi (Palmer)’, but all his illustrated specimens had their origin in the
Palaeocene of Libya. On the other hand, the writer’s experience of Palaeocene foramini-
fera from Libya indicates to him that (a) none of the Libyan specimens illustrated by
Arni (1966, pi. 1-2) are either conspecific or congeneric with O. bennudezi , and ( b ) the
specimens illustrated by Arni (1966, pi. 2, figs. 9-11) are not the megalospheric genera-
tion of the microspheric form he illustrates (ibid., figs. 7-8, 13-14), and probably belong
to the genus Operculina.

DISCUSSION

The species O. bennudezi has been well illustrated by Cole (1953, pi. 1 , figs. 5-7 ; pi. 2,
fig. 4; pi. 3, figs. 2-12) and by Sachs (1957, pi. 14). The spiral lamina is coarsely per-
forate throughout, a feature which has led some authors to suggest that it belongs to
the genus Miscellanea , but the presence of a well-developed perforate marginal cord
precludes placing it in the family Miscellaneidae. The subgenus Chordoperculinoides as
exemplified by its type species O. bennudezi is apparently restricted to the Palaeocene
of the Caribbean region.

The species to which the Libyan megalospheric forms illustrated by Arni (1966)
belong is known to the writer from work he has carried out on material from West
Pakistan, India, and Libya; it is the form originally described by L. M. Davies (1927,
cum syn.) as ‘ Operculina sindensis ’, a species sometimes referred by some subsequent
authors to the genus Ranikotlialia Caudri, 1944. The species O. sindensis is here made the
type of a new genus.

Family nummulitidae de Blainville, 1825
Genus sindulites gen. nov.

Type species. Operculina sindensis L. M. Davies (1927, p. 274, pi. 19, figs. 10-13), Palaeocene.

Diagnosis. Involute, axial section showing very long alar prolongations; wall structure
as in Nummulites (i.e. not coarsely perforate except at the marginal cord); septal
filaments visibly continuous from margin to centre on external surface; marginal cord

[Palaeontology, Vol. 11, Part 3, 1968, pp. 435-8.]

436

PALAEONTOLOGY, VOLUME 11

developed into a strongly swollen and coarsely perforate structure; microspheric form
little larger than and similar in form to the rather flat megalospheric form.

Remarks. It was Davies (1952) who first suggested that there might be a link between
the Palaeocene larger foraminiferal faunas of the East and West Indies when he described
Ranikothalia sahnii and its megalospheric form R. savitriae from West Africa. This
species is very closely related to S. sindensis, is obviously congeneric with it, and might
perhaps by some be considered conspecific with it. Nummulites nuttalli Davies 1927 is
the type species of Ranikothalia , and it differs from the above forms in lacking a strongly
swollen marginal cord and in its megalospheric form ( N . thalicus Davies 1927) being
considerably smaller and more inflated than the microspheric form. Prior to 1927,
N. nuttalli had been referred to N. planulatus (Lamarck), 1804, and the writer considers
that both planulatus and nuttalli are just lenticular Nummulites with rather numerous
septa and high chambers in equatorial section, and that Ranikothalia is a synonym of
Nummulites.

De Cizancourt and Cuvillier (1954) reported Nummulites ( Operculinoides ) bermudezi
(Palmer) from the Palaeocene of Senegal. Only one text-figure, a line drawing of an
axial section, was given. There is no indication of the coarse perforations in the spiral
lamina, so well illustrated by Cole (1953) and Sachs (1957). De Cizancourt and Cuvillier
also recorded several other Caribbean species from the Palaeocene in Senegal, but in
all cases the illustrations consisted merely of line drawings, and in all cases but one only
an axial section was illustrated. The writer would suggest that it is undesirable to identify
firmly in Senegal a nummuloid species from the other side of the world on the basis of
an axial section only, without any good indication of the wall structure and characters
of the equatorial section.

Haynes (1962) recorded and illustrated from the Palaeocene of Libya a form he re-
ferred to as ‘ Operculina canalifera sindensis (L. M. Davies)’. The quality and clarity of
his illustrations are such that one can be quite sure that the wall structure is the same
as that of O. sindensis (Davies) and that it lacks the coarse perforations in the spiral
lamina so characteristic of O. bermudezi and excellently illustrated by Cole (1953) and
Sachs (1957). However, Davies considered O. canalifera and O. sindensis to be different
species, and Haynes evidently overlooked the fact that O. canalifera is completely
evolute, whereas O. sindensis is completely involute. From what the writer has seen of
these two forms in India and West Pakistan, he can confirm Haynes’s identification of
the taxon ‘ sindensis ’ in Libya.

Castelain (1965) recorded the occurrence of ‘probable’ Operculinoides bermudezi
from the Palaeocene of Senegal, and also the occurrence of four discocyclinid forms:
Discocyclina grimsdalei Vaughan and Cole, D. harrisoni Vaughan, D. barkeri Vaughan
and Cole, and D. (vel Pseudophragmina ( Athecocyclina )) soldadensis Vaughan and Cole.
There were no descriptions or illustrations at all, and the writer feels that it would be
inadvisable to accept these identification of species from the other side of the world
without further detailed information.

Arni (1966) recorded Chor doper culinoides bermudezi from the Palaeocene of Libya.
The microspheric forms illustrated by Arni (1966, pi. 1, figs. 1-6; pi. 2, figs. 7-8, 13-14)
have all the characters of the genus Sindulites and of the taxon referred to as sindensis by
Haynes (1962). The megalospheric forms illustrated by Arni (1966, pi. 2, figs. 9-11)

F. E. EAMES: SINDULITES GEN. NOV.

437

seem to be an Operculina rather than the A Form of S. sindensis. The megalosphere of
S. sindensis is fairly large, that of a representative specimen from the Palaeocene of
Libya being bilocular, the outside dimensions being: protoconch 0-42x0-34 mm.,
deuteroconch 0-4x0-21 mm. The Palaeocene in Libya is richly fossiliferous, and quite
a number of species of Operculina are known to occur in association with S. sindensis.

In my opinion the above information lends considerable doubt to the occurrence of
Caribbean species of shallow water larger foraminifera in West Africa, and, if correct,
it would avoid all the difficulty of trying to understand how they could have got from
one side of the Atlantic Ocean to the other. All the forms now suggested to belong to
the genus Sindulites occur in beds of Palaeocene age, the genus being known from
West Pakistan, Somalia (unpublished information), Iran (unpublished information),
Iraq (unpublished information), Libya, Sicily (unpublished information), and Togoland.
It is relatively easy to comprehend how these species of Sindulites could have developed
as a closely knit group in geographical continuity during Palaeocene times when it is
recalled that the Palaeocene sea with its faunas extended from Indonesia through
Burma and West Pakistan to the Middle East and East Africa, thereon through Egypt
and Libya and the Sahara (Lake Tchad) to West Africa.

The species sindensis has at various times been referred to the genera Operculina,
Operculinoides, Nummulites, Ranikothalia, Miscellanea, and the subgenus Chordopercu-
linoides, and now to the new genus Sindulites', the reasons why the species sindensis
should not be referred to any of the first six taxa are:

(a) Except occasionally for the first whorl or two, Operculina is completely evolute, whereas sindensis
is completely involute.

( b ) Operculinoides and Operculinella, both of which are synonyms of Palaeonummulites, are miniature
forms, even the microspheric form of which rarely exceeds 5 mm. in diameter, the nucleoconch of the
megalospheric form rarely exceeding 0-15 mm. ; sindensis is quite large, has a double nucleoconch each
chamber of which is large, attaining a diameter of 0-4 mm. approximately, and the species possesses
a very strongly swollen marginal cord which is not present in Palaeonummulites.

(c) Nummulites does not have a very strongly swollen marginal cord, and its megalospheric forms are
considerably smaller and often much more inflated than the microspheric forms, in contrast to sin-
densis; on these bases Ranikothalia would be a synonym of Nummulites.

{d) Miscellanea lacks the perforate marginal cord of the Nummulitidae, whereas sindensis possesses
this; moreover, Miscellanea has a coarsely perforate spiral lamina, whereas sindensis does not.

(e) In Chordoperculinoides, as exemplified by its type species, the spiral lamina is coarsely perforate
throughout, whereas it is not in sindensis.

For these reasons the writer considers that Sindulites, as exemplified by its type
species O. sindensis, is generically distinct from Operculina, Palaeonummulites (synonyms
Operculinella and Operculinoides), Chordoperculinoides, Miscellanea, and Nummulites
(synonym Ranikothalia).

Acknowledgements. The writer is indebted to the British Petroleum Co., Ltd. for permission to publish
this paper; also to his colleagues Dr. W. J. Clarke, Dr. A. H. Smout, and Dr. F. T. Banner for reading
and commenting on the manuscript.

REFERENCES

arni, p. 1965. Contribution a la systematique des Nummulites s.l. Mem. Bur. Rech. Geol. Min. no. 32,
21-28. (Colloque Int. de Micropal. Dakar, 1963.)

1966. Contribution to the history of growth of the Chordoperculinoides Shell. Eclog. geol. Helvet.

59, 339-46, pis. 1-2. {cum bibl.)

438

PALAEONTOLOGY, VOLUME 11

castelain, j. 1965. Apergu stratigraphique et micropaleontologique du bassin de Senegal — Historique
de la decouverte paleontologique. Mem. Bur. Rech. Geol. Min. no. 32, 135-59, pis. 1-4. (Colloque
Int. de Micropal. Dakar, 1963.)

cizancourt, m. de, and cuvillier, j. 1954. Les Nummulites cordelees du Senegal occidental. C.R.
Seanc. Soc. geol. Fr. 7, 130-3.

cole, w. s. 1953. Criteria for the recognition of certain assumed camerinid genera. Bull. Anier. Paleont.
35, 27-46, pis. 1-3.

davies, l. m. 1927. The Ranikot Beds at Thai (North-West Frontier Provinces of India). Quart. J. geol.
Soc. London, 83, 260-90, pis. 18-22.

1952. Ranikothalia sahnii n. sp. and R. savitriae, n. sp. : A possible link between the Paleocene

faunas of the East and West Indies. Palaeobotanist , 1, 144-58, pi. 1.
haynes, J. 1962. Operculina and associated Foraminifera from the Paleocene of the N.E. Fezzan,
Libya. Contrib. Cushman Found. Foram. Res. 13, 90-97, pis. 17-18.
sachs, K. n. 1957. Restudy of some Cuban larger Foraminifera. Ibid. 8 (3), 106-20, pis. 14-17.

F. E. EAMES

c/o B.P. Research Centre
Exploration and Production Research Division
Chertsey Road
Sunbury-on-Thames

Typescript received from author 28 April 1967 Middlesex

A REVISION OF THE CARBONIFEROUS
LYCOPOD GENUS ESKDALIA KIDSTON

by B. A. THOMAS

Abstract. Eskdalia minuta Kidston has previously been described as a fern stem, but closer examination of its
macroscopic features and a study of its cuticle shows it to be a ligulate lycopod stem. The genus of Russian paper
coal cuticles, Porodendron Zalessky, is shown to be a synonym of Eskdalia.

The type specimen was originally described by Kidston (1883) as Caulopteris minuta.
This specimen, together with three others, later provided Kidston with the basis for his
new genus Eskdalia and all four specimens were named E. minuta. Kidston still believed
them to be probably fern stems, but subsequent examination of the specimens and of
their cuticles has shown the species to consist of ligulate lycopod shoots (Chaloner
1967).

Kidston’s specimens are the type, No. 5594, Geological Survey, Edinburgh, from
Kershope Burn, Liddesdale, Dumfriesshire and Nos. 2739, 2740, and 6731, Kidston
collection, Geological Survey and Museum, London, from Clencartholm, Eskdale,
Dumfriesshire. The species has also been recorded from the river Coquet, half a mile
north-north-east of Holystones, Northumberland (Kidston 1903, p. 820). The known
distribution is therefore limited to the Cementstone group of the Lower Carboniferous.

Cuticle was prepared from the type specimen and No. 2740 by macerating portions of
coalified compression in concentrated nitric acid and potassium chlorate. Large pieces
were obtained from No. 2740 but only small fragments from the type specimen as its
carbon split and sometimes completely broke up during maceration. No. 6731 is pre-
served as an impression so no cuticle could be prepared from it and No. 2739 although
a compression gave no cuticle.

Nos. 5594, 2740, and 6731 are large stems but No. 2739 is much narrower than the
others and its description is best kept separate. The three large specimens are straight
unbranched shoots with oval elongated scars. The stem surface between the scars is
flat but ornamented in all three specimens with large numbers of minute lumps, up to
100 ft wide. These lumps are sparsely scattered in No. 5594 but rather more crowded in
the other two specimens. No stomatal pits are visible. The leaf scars are in a low-angle
spiral of about 10-15° to the horizontal. They are separated from those scars above and
below by about a length of a scar and from those on either side by one to two widths of
a scar. The scar surface is slightly raised in Nos. 2740 (PI. 82, figs. 3-5) and 6731 (PI. 82,
fig. 6) but slightly sunken in No. 5594 (PI. 82, figs. 1, 2). The scars have a sharp margin
and the stem cuticle ceases abruptly at this margin, there being none over the surface
of the scar; this shows that it is truly a scar. The scar outlines are almost perfectly oval,
but may be interrupted at the apex by a slight notch. A minute groove, about 0-4 mm.
long and 0-15 mm. wide is sometimes visible running down from the scar apex. These
grooves are plainly seen in most of the leaf scars of No. 2740, but not in Nos. 6731 and
5594 where the notch alone is seen in a few scars and there only obscurely. Maceration

[Palaeontology, Vol. 11, Part 3, 1968, pp. 439-44, pi. 82.]

440

PALAEONTOLOGY, VOLUME 11

of coaly material from the top of the scar proves that a ligule pit is present. It is situated
under the groove and is the same shape and size; I imagine that the groove is not an
original feature but is caused by a collapse during compression. The cuticle of the ligule
pit is attached to the margin of the stem cuticle surrounding the scar (PI. 82, fig. 7, and
text-fig. 1). The vascular prints are shown most clearly in No. 5594 and their structure

text-fig. 1. Eskdalia minuta Kidston. No. 2740, Kidston collection, a, complete specimen showing the
arrangement of the leaf scars. X 1. b, single leaf scar showing the position of the underlying ligule
pit (1) and the vascular print (v), x4. c, stem cuticle from the apical region of a leaf scar showing the
outline of the leaf scar (s) and the ligule pit cuticle (1). Some, but not all, of the resistant excrescences
have been included to give an indication of their size. Slides PF.4417 and PF.4418, x20. d, stem
cuticle showing epidermal cells and overlying resistant excrescences. Slide PF.4417, x 300.

is described and discussed later. In No. 2740 the vascular prints are very faint, being only
visible as raised ridges, and in No. 6731 they are indistinguishable.

The fourth stem figured by Kidston (No. 2739) is much narrower than the other
specimens, being only 4 mm. broad. The scars are oval and slightly raised but only a few
have visible vascular prints and no ligule pits can be seen. The stem surface is smooth
possessing no lumps and unfortunately no cuticle could be prepared to show epidermal
details. However, the shedding of leaves on such a small shoot is an important feature
and can be taken as evidence of its lycopod affinity.

The cuticles of both Nos. 5594 and 2740 are rather thick. The only cells visible are
elongated and have straight, rather finely marked outlines; the anticlinal walls only
projecting slightly. In part of the cuticle of both specimens there is a border to the cell
outline especially at the cell corners, making the cell interior rounded. The cell surfaces
are flat but many solid lumps of resistant matter are present on the outer surface of the
cuticle. These are most probably the same as the bulges which ornament the stem

c

B. A. THOMAS: REVISION OF ESKDALIA KIDSTON

441

surface. These lumps are oval-rounded, usually slightly elongated, and vary in size to
about 100 [jl. They are irregularly arranged and in many places partly obscure the
underlying epidermal cells. The lumps often overlap and where the cuticle is torn many
project beyond the broken edge, suggesting that they are somewhat separate excrescences.
In No. 6731, where the stem has been lost and the imprint alone remains, many of these
lumps are sill embedded in the rock.

DISCUSSION

The reason Kidston gave for believing Eskdalia to be a Caulopteris- like stem was
that he thought the vascular prints to be horseshoe shaped. The lower part of the holo-
type clearly shows oval areas within each leaf scar which are often interrupted at the
apex to resemble horseshoes and are evidently the features referred to by Kidston.
These ‘horseshoes’ are raised, but it must be made clear that the lower half of the speci-
men is really an impression as most of the specimen except a little carbonized material
has been lost. Chaloner (1967, p. 525) suggested that they represented matrix infilling
of lacunae which more or less surrounded the leaf traces as they passed through the
outer cortex. However, the ‘horsehoes’ and the shallower central areas they surround
are entirely carbonaceous. Thus it seems more probable that the ‘horseshoes’ of
Kidston represent the remains of a strong relatively non-compressible tissue. Graham
(1935, p. 605, figs. 50, 51) has described such a condition in a lycopod leaf where the
xylem and phloem are almost surrounded by an arc of sclerenchyma. This might well
have been the condition in Eskdalia ; Kidston’s ‘horseshoe’ being on this interpretation
sclerenchyma and not an arcuate scar as Kidston supposed. The differential compression
of the tissues has probably been brought about by the infilling of the stem with mineral
matter. In contrast, No. 2740 does not show horseshoe-shaped marks because very little
mineral matter ever got inside this specimen and the surfaces came to lie in close
proximity like a collapsed tube.

The cuticle of E. minuta is very similar to that described from the Russian paper coal
of the Lower Carboniferous of the Moscow basin. Both are flat sheets with regularly
spaced perforations marking the positions of leaf scars. A small tube of cuticle hangs
from the edge of the cuticle at the apex of each perforation. These tubes of cuticle were
originally described as rolled up portions of leaf cuticle (e.g. Zeiller 1882, Zalessky 1915)
or rolled up cuticle which had formed over the leaf abscision area (Bode 1929). However,
Walton (1925) and Wilson (1931) have shown them to be really the internal linings of
ligule pits. The Russian cuticles have been given various names. They were originally
described as a species of Lepidodendron by Eichwald (1860), Auerbach and Traustschold
(1860), and Goeppert (1861), then as Bothrodendron punctatum by Zeiller (1880, 1882),
but Zalessky (1915) instituted a new genus Porodendron for them and is followed in his
nomenclature by modern Russian workers (e.g. Orlov 1963). Nathorst (1894) had sug-
gested Porodendron as a possible alternative name to Bothrodendron but did not finally
use it, nor did he give it a diagnosis. He should not, therefore, be regarded as its author
as quoted by Bode (1929). The species of Porodendron have been defined on leaf scar
shape which varies from round to oval or heart-shaped and on the structure of
the epidermal cells which are normally hexagonal, isodiametric, or slightly elongated.
No species has been described with any resistant excrescences on their outer surface.

og

C 5586

442

PALAEONTOLOGY, VOLUME 11

Distinction between the Russian paper coal cuticles and that from E. minuta is there-
fore limited to leaf scar shape, epidermal cell shape and size and presence or absence
of resistant excrescences. Such differences are of value only at the specific and not
the generic level and thus Porodendron Zalessky is a synonym of Eskdcdia Kidston.

Nathorst (1914) described a small lycopod cone compression as Porostrobus zeilleri
believing it to be the fructification of Porodendron. Bhardwaj (1958) has subsequently
shown the holotype to be heterosporous as might have been expected for a ligulate
shoot such as Porodendron (= Eskdalia). However, as Chaloner (1962, p. 83) has already
pointed out, the cone was not found in organic connexion with the stem and mere
association is not sufficient evidence for believing them to be parts of the same plant.

SYSTEMATIC DESCRIPTION

Genus Eskdalia Kidston

1903 Eskdalia Kidston, p. 750.

1909 Porodendron Zalessky, p. 5.

1914 Porodendron Zalessky; Nathorst, p. 68.

1929 Porodendron Nathorst (non Zalessky); Bode, p. 134.

Type species. Caulopteris minuta Kidston, p. 541, pi. 31, figs. 1, la.

Emended diagnosis. Lycopod stem with round-oval or heart-shaped leaf scars in low
angle spirals. Leaf scars almost level with stem surface. Ligule pit at apex of leaf scar
and projecting beneath it, forming a narrow cutinized tube. One simple vascular print
present slightly above middle of leaf scar. No parichnos present.

Eskdalia minuta Kidston
Plate 82, figs. 1-9; text-fig. 1

1883 Caulopteris minuta Kidston, p. 541, pi. 31, figs. 1, la.

1903 Eskdalia minuta Kidston, pp. 750, 820, pi. 1, figs. 4-8.

1967 Eskdalia minuta Kidston; Chaloner, pp. 525-6, fig. 361.

Holotype. Specimen No. 5594, Geological Survey of Scotland, Edinburgh.

Emended diagnosis. Leaf scars oval, length about one and a half times breadth. Scars
separated from those above and below by about a scar length, from those on either
side by one to two scar widths. Epidermal cells elongated, about 130 p long, 25-35 p
broad. Anticlinal walls 1 p thick, straight, smooth. Periclinal walls flat. Surface bearing

EXPLANATION OF PLATE 82

Figs. 1-9. Eskdalia minuta Kidston. 1, 2. Holotype, No. 5594, Geological Survey, Edinburgh. 1, Near
side of specimen visible in upper half and the impression of its far side in the lower half, X 1. 2, Por-
tion showing the structure of the leaf scars and vascular prints, x 3|. 3-5, No. 2740, Kidston collec-
tion. 3, Complete specimen, x 1 . 4, 5, Parts of bark surface showing leaf scar and stem details, X 4.
4, Illuminated from above to show leaf scars and vascular prints. 5, Illuminated from below to show
ligule pits and surface markings between the leaf scars. 6, No. 6731, Kidston collection, X 1. 7-9,
Cuticle from No. 2740. 7, Cuticle from immediately above a leaf scar showing the outline of the
scar and point of attachment of a broken ligule pit, X 20. 8, 9, Cuticle showing cellular detail and
the resistant excrescences on the outer surface, x 100. Slide PF.4417.

Palaeontology , Vol. 11

PLATE 82

THOMAS, Carboniferous lycopod Eskdalia

B. A. THOMAS: REVISION OF ESKDALIA KIDSTON

443

resistant lumps up to about 100 X 50 /x. No stomata. Ligule pit about 400 p long, 150 /x
broad with rectangular cells 1 5-20 /x long, 20 /x broad in longitudinal files. Anticlinal
walls 1 /x thick, straight, smooth. Periclinal walls flat, smooth.

Stratigraphic occurrence. Cementstone Group, Calciferous Sandstone Measures, Lower
Carboniferous Limestone Series of Northern England and Southern Scotland.

COMPARISON

There are many other arborescent lycopod genera rather similar to Eskdalia. Of
these Cyclostigma Haughton, Pinakodendron Weiss, Scutellocladus Lele and Walton,
and Angarodendron Zalessky are described as eligulate and thus differ in a very impor-
tant respect. Bothodendron Lindley and Hutton differs in having parichnos prints on the
leaf scar. Moreover, most of the species of this genus have the ligule pits situated a little
above the leaf scar and separated from it by a narrow band of stem epidermis. The
described epidermises of Bothrodendron also differ in having many stomata (Thomas
1967).

Acknowledgements. I thank Dr. W. G. Chaloner for suggesting this study and for discussing the
results, Professor T. M. Harris, F.R.S., for his helpful criticism, and the Director of the Geological
Survey for the loan of the specimens. The work was carried out during the tenure of the Lord Adams
Fellowship from the University of Newcastle upon Tyne.

REFERENCES

bhardwaj, d. c. 1958. On Porostrobus zeilleri Nathorst and its spores with remarks on the systematic
position of P. bennholdi Bode and the phylogeny of Densosporites Berry. The Palaeobotanist, 7,
67-75.

bode, h. 1929. Zur kenntnis der Gattung Porodendron Nathorst (non Zalessky). Palaeontographica,
72, 125-39.

chaloner, w. G. 1962. A Sporangiostrobus with Densosporites microspores. Palaeontology, 5, 73-85.

1967. Trade de Paleobotanique, n, 854 pp. Paris.

eichwald, e. 1860. Lethea Rossica , ancienne period. 1,657 pp., Atlas pt. 1 (1859).

goeppert, h. 1861. Uber die kohlen von Malowka in Zentral-Russland. Sber. bayer. Akad. Wiss.
(math.-nat.), 1, 199-209.

graham, r. 1935. An anatomical Study of the Leaves of the Carboniferous Lycopods. Ann. Bot. 49,
557-608.

kidston, r. 1883. Report on Fossil Plants collected by the Geological Survey of Scotland in Eskdale
and Liddesdale. Trans. R. Soc. Edinb. 30, 531-50.

1903. The Fossil Plants of the Carboniferous Rocks of Canonbie, Dumfriesshire and of parts of

Cumberland and Northumberland. Ibid. 40, 741-833.

nathorst, a. g. 1894. Zur palaozoichen Flora der arktischen Zone. K. Svenska. Vetensk. Akad. Hand!.
26, 1-80.

1914. Zur fossilen Flora der Polarlander, Teil I, Lief. 4. Nachtrage zur Paldozoischen Flora

Spitzbergen, 110 pp.

orlov, iv. a. 1963. Osnovy Pa/eontologii, Moscow, 698 pp. [In Russian.]

thomas, b. a. 1967. The cuticle of two species of Bothrodendron [Lycopsida: LepidodendralesJ.
/. Nat. Hist. 1 , 53-60.

walton, l 1925. A note on the structure of the plant cuticles in the Paper-Coal from Toula in Central
Russia. Manchester Mem. 70, 119-23.

wilson, j. a. r. 1931. Some new facts about the structure of the cuticles in the Russian Paper-Coal and
their bearing on the systematic position of some fossil Lycopodiales. Proc. R. Soc. Edinb. 51, 104-1 5.

444

PALAEONTOLOGY, VOLUME 11

zalessky, m. d. 1909. Note sur les debris vegetaux du terrain carbonifere de la chaine de Mugodzary.
Izv. geogr. inst., Petrograd. 28, 11 pp.

1915. Observations sur le Lepidodendron Olivieri Eichwald et le Lepidodendron tenerrimum A. &

T. Mem. Com. Geoi. St. Petersbourg, N.s. 125, 25-46.
zeiller, r. 1880. Note sur des Cuticules fossiles du Terrain carbonifere de la Russie centrale. Bull.
Soc. bot. Fr. 27, 348-53.

1882. Observations sur quelques cuticules fossiles. Aunis. Sci. not. 6 serie, Bot. 13, 217-38.

B. A. THOMAS

Department of Botany
The University
Newcastle upon Tyne 1

Manuscript received 24 April 1967

A SPECIES OF COMPRESSED LYCOPOD
SPOROPHYLL FROM THE UPPER COAL
MEASURES OF SOMERSET

by M. C. BOULTER

Abstract. Compression fossils of a new species of Lycopod sporophyll, LepidostrobophyUum alatum, are
interpreted in terms of their compression history. Three superficially different types of fossil are assigned to the
same species as a result of this interpretation. These, it is suggested, are determined by the original orientation
of the sporophylls in the sediment prior to compression; either upright, inverted, or lateral. The species has an
alated pedicel, and the evolutionary significance of this is discussed.

The appearance of plant compression fossils, as seen on a broken rock surface, is
governed by three factors. First, the orientation of the plant organ in the original
sediment; second, the collapse of the fossil resulting from compaction of the matrix;
and third, the actual face of the fossil which is revealed by cleavage of the sediment.

This paper is an account of some Carboniferous lycopod sporophylls, assigned to
LepidostrobophyUum Hirmer 1928, the varied appearance of which is interpreted in
terms of these three factors.

An interpretation of the original form of the sporophyll described here is shown in
text-figs. 1 and 2. Its general character is that frequently demonstrated for petrified
Lepidostrobus (Balbach 1966, Snigirevskaya 1964, Scott 1920) and for compressed
sporophylls (Abbott 1963, Allen 1961) in which a distal leaf-like lamina extends from
a shorter proximal region, called the pedicel. The end of this pedicel region was
originally attached to the strobilar axis. Such a sporophyll could have fallen into the
Carboniferous mud in three different orientations; in the first, the sporophyll remained
more or less upright; in the second, it settled in an inverted position (the lower surface
uppermost); in the third, it settled on its side so that the lamina of the sporophyll came
to lie more or less on edge in the matrix. These three states are here described as up-
right, inverted, and lateral, respectively (text-fig. 3).

Walton (1936) suggests that the appearance of plant compression fossils is deter-
mined primarily by the original shape of the organ’s undersurface in the matrix. On this
basis, the present sporophylls, together with the matrix of mud, were subjected to
vertical compression from the weight of overlying sediment, which compressed the
vertical structure but left the horizontal outlines more or less unchanged. Accordingly,
it is here suggested that the upright, inverted, and lateral states of orientation in the
sediment resulted in the sporophylls being compressed in different ways, depending on
the shape of the lower surface in the mud (text-fig. 3).

Splitting of the rock will roughly follow the bedding, and will expose the plant itself
by cleaving along the plane of weakness represented by the coaly layer (or layers) of the
fossil. The fossil is accordingly represented by an upper counterpart and a lower counter-

[Palaeontology, Vol. 11, Part 3, 1968, pp. 445-57, pis. 83-84.]

446

PALAEONTOLOGY, VOLUME 11

part. The three orientation states and these two counterparts mean that a sporophyll
may be represented by six different possible states on an exposed rock surface (text-fig.
3). Even more fossil states can exist, due to there being two layers of coal in the collapsed
proximal region of the sporophyll, one representing the collapsed pedicel, and the other
the collapsed sporangium wall (text-fig. 2). The rock may cleave along either the coaly

TEXT-FIG. 1.

TEXT-FIG. 2.

text-fig. 1 . Reconstruction of Lepidostrobophyllum alatum. A large sporophyll which it is suggested
would fall in an inverted position in the substratum.

text-fig. 2. Reconstruction of Lepidostrobophyllum alatum. A small sporophyll which would fall in
an upright position in the substratum. The dehisced sporangium is shown in section at the proximal
end. This line of section can be compared to Plate 84, fig. 6.

layers of the pedicel (PI. 84, figs. 1 and 5) or along the coaly layers of the sporangium
wall (PL 84, fig. 3), giving two quite different fossil forms to the pedicel region. Alterna-
tively, the exposure of the fossil might have started by cleavage along the coaly layer
of the pedicel surface, which then jumped across to cleave along the sporangium wall

BOULTER: COMPRESSED LYCOPOD SPOROPHYLL FROM SOMERSET 447

layer (PI. 84, fig. 3). On this interpretation it can be seen that a single species of sporo-
phyll can be represented by several fossilization states of rather different appearance.

As might be supposed, the frequency of occurrence of the upright, inverted and lateral
states is not random. The frequency of one orientation against another will be governed
by the mechanical properties of the sporophyll (centre of gravity, hydrodynamic pro-
perties), the physical properties of the Carboniferous sediment (mud, sand, etc.), and by
the environment of deposition (speed of deposition, current action). A sporophyll with
a small lamina is most likely to fall upright, since the centre of gravity is at the lower part
of the pedicel region. A sporophyll with a larger lamina will have its centre of gravity at
the upper part of the pedicel region, and so will more readily fall either inverted or
laterally. A sporophyll with a tall, narrow sporangium is more likely to fall in a lateral
position.

Acknowledgement. I am indebted to Dr. W. G. Chaloner, University College London, for suggesting
this topic, and for numerous discussions thereon.

METHODS

Specimens were photographed macroscopically under xylol. An adaptation to Wal-
ton’s transfer technique (Walton 1923) prepared selected specimens for microscopic
examination. A 4 mm. layer of Ceemar embedding resin was poured into a polythene
mould. After about one hour, when the surface became tacky, the fossil was placed face
downwards on the resin surface. A further 1 mm. layer of resin was immediately poured
on to the tacky resin. When the absence of air bubbles was ensured, the block was left
overnight to set, and then immersed in hydrofluoric acid in the usual way. This adapted
technique yields results as good as Walton’s original Canada balsam method, and has
the advantage that there is no problem of temperature control in the balsam ‘cooking’
process. A recent detailed description of plastic transfers of fossil plant compressions
(Cridland and Williams 1966) is essentially similar to the method used here. In order to
investigate parts of the fossils concealed in the matrix (e.g. the heel and sporangium
wall), specimens were cut and smoothed in the relevant position at a surface perpendicu-
lar to the bedding plane. When viewed under xylol, these sections show considerable
detail (PI. 84, fig. 6). Selling (1944) used a similar procedure to examine Calamitean
cone compressions, though he used a microtome to produce thin sections. Block sec-
tions such as were used in this investigation require no preliminary softening treatment,
and produce results similar to those of Selling. Maceration was attempted, but did not
reveal cuticle or spores in any specimen.

THE GENUS LEP I D O S TRO BO P H YLLU M

Brongniart (1828) instituted the genera Lepidostr obits and Lepidophyllum to describe
respectively the supposed cones and isolated leaves of Lepidodendron. Later authors
used Lepidostrobus and Lepidophyllum rather indiscriminately for what were believed
to be isolated cone scales (i.e. sporophylls) of Lepidodendron. Arber (1922) in a revision
of British Lepidostrobus compressions referred many of the species based on isolated
sporophylls to Lepidostrobus, arguing that they are not foliar leaves, as was at one time
thought, but the sporophylls of a cone. Hirmer (1928) subsequently set up the genus

448

PALAEONTOLOGY, VOLUME 11

Lepidostrobophyllum for those species which were based on isolated sporophylls, using
L. major as type species. He retained LepidophyUum for what were believed (from lack
of a basal sporangium) to be vegetative leaves. Snigirevskaya (1958) has recently pointed
out that LepidophyUum was invalid in its application to fossils (having previously been
used for a living plant) and substituted the generic name Lepidophylloides for vegetative
lepidodendroid leaves. Allen (1961) emended Lepidostrobophyllum to describe a species
of sporophyll containing a number of large megaspores.

Abbott (1963) introduced two new genera to describe unisexual lycopod strobili and
their isolated sporophylls, megasporangiate, or microsporangiate. Lepidostrobopsis has
a horizontally alated pedicel, but otherwise is similar to Lepidostrobus. Lepidocarpopsis
has a horizontal and upturned alation to the pedicel, much wider than the lamina, and
bears Cystosporites type megaspores or Lycospora type microspores. Abbott regards
Lepidocarpopsis as being ‘closer to genus Lepidocarpon than Lepidostrobus' . A later
section in this paper explains why the name Lepidostrobophyllum is used for the speci-
mens described here.

SYSTEMATIC DESCRIPTION

Class LYCOPSIDA
Order lepidodendrales
Genus lepidostrobophyllum Hirmer 1928

Lepidostrobophyllum alatum sp. nov.

Plate 83, figs. 1-8; Plate 84, figs. 1-10

Diagnosis. Isolated sporophylls consisting of distal lamina and proximal pedicel.
Lamina: 1 -0-5-0 cm. long; 0-4-1 -0 cm. wide at base, directed obliquely towards apex.
Central mid-rib extending the whole length of the lamina. Pedicel: 0-4-1 -2 cm. long,
0-4-1 -0 cm. wide at edge joining lamina. Prominent keel on whole length of lower sur-
face. Heel on lower surface at junction with lamina. Hairy alation at edges, hairs up to

EXPLANATION OF PLATE 83

Figs. 1-8. Lepidostrobophyllum alatum sp. nov., specimens compressed in the inverted state. All the
specimens figured are in the Geological Survey collection, South Kensington, London. Specimens in
figs. 1, 2, 3, 6, 7, and 8 were photographed under xylol, and figs. 4 and 5 in air. 1, Upper counterpart
with hastate base to lamina, and prominent sporangium wall at pedicel region; RC 4829, X 5.
2, Syntype; upper counterpart specimen in which the coal layer of the pedicel has broken away,
presumably on to the fossil’s lower counterpart. The outer part of the sporangium wall remains on
this specimen as can be seen at the right hand side of the figure; RC 4830, X 5. 3, Smallest specimen
of the inverted type which was collected; lamina mid-rib is seen to protrude down into the shale,
and pedicel region is at a lower level than lamina, therefore a specimen from a lower counterpart ;
RC 4831, X 5. 4, Transfer of pedicel region The hastate lamina and the mid-rib are seen on the right ;
the hairy pedicel alation leads to the heel, which protrudes upwards (transfer of upper counterpart) ;
RC 4832, x 6. 5, Margin of lamina, showing teeth pointing towards apex; RC 4832, X 100. 6, longi-
tudinal section of upper counterpart showing step between lamina (on left) and pedicel, and the heel
embedded in the matrix; RC 4833, X 30. 7, Transverse section across pedicel region of a lower
counterpart, showing one side of the sporangium wall extending just below the surface of the
cleavage plane ; RC 4834, x 12. 8, Transverse section across pedicel region as in fig. 7 ; RC 4835, X 12.

Palaeontology, Vol. 1 1

PLATE 83

BOULTER, Carboniferous lycopod sporophyll

BOULTER: COMPRESSED LYCOPOD SPOROPHYLL FROM SOMERSET 449

3 mm. long. Sporangium wall extending laterally from pedicel, semi-rhomboidal in
transverse section. No spores have been found.

Locality. Tip heap, Kilmersden Colliery, Radstock, Somerset. For the past twelve years, this colliery
has been working the no. 10 seam, which occurs in the Farrington Group of the Upper Coal Measures,
and represents Plant Zone H of Dix’s divisions (Dix 1934). There is no doubt that the sporophylls
described in this paper come from within this group.

Syntypes. RC 4830, PI. 83, fig. 2; RC 4837, PI. 84, fig. 2; Geological Survey, London.

Description of sporophylls oriented in the upright state. Detached sporophylls which are
seen in this state consist of two parts, the distal part or lamina, and the proximal part
which was originally attached to the strobilar axis, called the pedicel. The lamina is of
variable size, 1 -0-4-2 cm. long and 0-4-0-7 cm. wide at the base. The lateral margins are
more or less straight, and have teeth about 0-2 mm. long, pointing towards the apex
(PI. 83, fig. 5). The mid-rib, which is about 0-7 mm. wide, passes from the base to the
apex, and is made up of two parallel ridges (PI. 84, figs. 1, 2).

The pedicel is of variable size, 0-4-1 -0 cm. long, and 0-4-0-7 cm. wide at the edge
joining the lamina, tapering down in nearly all specimens to 0-2 cm. wide at the proximal
end. A hairy alation extends from each of the lateral margins, the hairs being up to
3 mm. long at the distal end (PI. 84, fig. 4). These hairs are best seen in transfers, but in
many of the specimens which were studied during this investigation they were not pre-
served. The keel extends for the whole length of the pedicel, and protrudes either up-
wards or downwards in upper or lower counterparts respectively. It develops into a heel
at the juncture with the lamina, as can be seen in section (PL 84, fig. 7). The sporangium
wall is preserved as two lobes, one on either side of the pedicel, and extending further
than the pedicel alation (PI. 84, fig. 4). Transverse sections across the pedicel reveal the
sporangium wall embedded in the matrix (PI. 84, fig. 6), though, as would be expected,
this is only seen on the upper counterpart specimens.

Description of sporophylls oriented in the inverted state. The lamina is of variable size,
2-0-4-8 cm. long, 0-6-1 -0 cm. wide at the base. The lateral margins are parallel for about
the first two thirds of the length; then they curve to form an acuminate apex. Just as in
the upright state specimens, these have a serrate edge and a mid-rib, though in the
inverted state the lamina is hastate at the base (PI. 83, figs. 1 and 4). The pedicel is of
variable size, 0-7-T2 cm. at the distal end. As in the upright state specimens, the pedicel
has a hairy alation, though here the hairs are slightly thicker and less regular (PI. 83,
fig. 4). Collapse, during compression, has taken place into the (open) sporangium, and
so the sporangium wall is more prominent in this type of preservation than in the up-
right orientation state. Transverse sections of some lower counterpart specimens show
the embedded sporangium wall (PL 83, figs. 7, 8), though this is often absent when
cleavage has occurred along the plane of the wall. There is a difference in level between
the lamina and pedicel of up to 2 mm. On the lower counterpart the lamina is on a
higher level than the pedicel, whilst on the upper counterpart the lamina is lower than
the pedicel.

Description of sporophylls oriented in the lateral state. The lamina is from 1-0 to 3-0 cm.
long. The distal end is often broken by cleavage or concealed in the rock. It is preserved

450

PALAEONTOLOGY, VOLUME 11

so as to lie more or less perpendicular to the bedding plane, and so is preserved on the
cleaved rock face only as a narrow line, or steeply dipping surface of about 3 mm. width.
Such characters as the original width, the serrate margin, and the mid-rib cannot be
seen. The pedicel proper is from 2 to 4 mm. high throughout its whole length (PI. 84,
fig. 8). On the lower edge the keel is about 3 mm. deep in the centre, becoming deeper
at the proximal end for attachment to the strobilar axis. It is also deeper at the distal
end where it forms the heel (PI. 84, fig. 9). The sporangium wall extends from 0-5 to
TO cm. above the whole length of the upper surface of the pedicel (PI. 84, figs. 8-10).

INTERPRETATION OF THE FOSSIL STRUCTURES

It was explained at the beginning of this paper that the appearance of the sporophyll
fossils has been determined by the original shape of the undersurface in the mud. The
sporophyll could have fallen in the mud in three principal orientations, each with a
different undersurface, and these are correlated with the three different fossil states of
the original structure which are found to occur.

Sporophylls falling into the mud in an upright position have the keel, heel, lower
pedicel surface, pedicel alation, the outer extremity of the dehisced sporangium wall, and
the entire lamina, making up the lower surface in the matrix. These structures have been
preserved without distortion in the horizontal plane. The original upper surface of the
pedicel and the inner parts of the sporangium wall will collapse to the level of the lower
surface. But since the sporangium was semi-rhomboidal in transverse section, it became
filled by, and separately enclosed in mud, and so is preserved as a separate extension
from the upper surface of the pedicel (text-fig. 3). Thus, we would expect to see the
sporangium buried in the matrix of the upper counterpart specimens (PI. 84, fig. 6).
The original adaxial lamina mid-rib is compressed to form a ridge protruding upwards
in the matrix.

EXPLANATION OF PLATE 84

Figs. 1-10. Lepidostrobophyllum alatum sp. nov. All specimens figured are in the Geological Survey,
London. All but figs. 2 and 4 were photographed under xylol. 1-7, Specimens fossilized in the up-
right state. 1, Upper counterpart in which the sporangium wall is not visible at the surface; it is
suggested in this article that the sporangium wall is embedded in the shale; RC 4836, X 4. 2, Syntype.
The lamina mid-rib protrudes into the matrix whilst the pedicel protrudes outwards, suggesting the
specimens to be from the upper counterpart; the outer parts of the sporangium wall can be seen
passing into the shale; RC 4837, x4. 3, Upper counterpart. At top left of figure, the sporangium
wall extends beyond the pedicel alation; the opposite part of the sporangium wall has not followed
the cleavage plane and is embedded in the matrix of the counterpart; RC 4838, x4. 4, Transfer
showing hairy alation to pedicel and sporangium wall extending beyond this alation; RC 4839,
x4. 5, Small specimen from lower counterpart with no sporangium wall; RC 4840, x4. 6, Trans-
verse section of upper counterpart across pedicel region, showing the two lobes of the sporangium
wall embedded in the matrix; RC 4841, X20. 7, Longitudinal section of lower counterpart; heel
prominent, and no step between lamina and pedicel; RC 4842, x20. 8-10, Specimens fossilized in
the lateral state. The lamina is broken off in three specimens (top left in 8 and 10, top right in 9).
8, Pedicel alation extends to half way up the sporangium wall; RC 4843, x4. 9, Sporangium wall
almost three times higher than the pedicel alation; heel (bottom right) and keel are prominent
below the pedicel; RC 4844, x4. 10, Specimen showing further variation in the size of the spor-
angium wall; RC 4845, x4.

Palaeontology, Vol. 1 1

PLATE 84

BOULTER, Carboniferous lycopod sporophyll

BOULTER: COMPRESSED LYCOPOD SPOROPHYLL FROM SOMERSET 451

«3 y

C/5 3-

<L>

O

c -5

•— <u

c &
£ ^
o .5

•C X)

w t« 3

« E
2-2 2
£ 2

o

a
3
C a>

« £
(Z X

8 -

3 "SL

x> p

2 a

a 2
— o
a

<U >%

<S1 SZ

g a <u

O O rt
o £: 2
.. O v>

y o.

£ ™ si
_ 60
$0 03 'C

3 e a
a Si 3

C ' 7* *■<— '

O o O

c' &

5j “

< «

. 73
C

- 03

■&£
o v
o 3
& £

“ c

^ £

'o js

co oo

.S c

£ ^
03 D

i_ J-l

^ o

<D U-i

y> cl

CO

13 -T

£ §
73 •— (

<U CO
CL <D

^ Ui

73

S.S

g-S

.5 CL

03 73

2 8.
£ 73
CO C
"Ch CC
& o3

3 .£

<L) £
•— 03

b £

03

5 W (U
^ fc CL

£. £• a

^ I 3

-§ U ^

O .

$ 2? <lT

l-s g

•S n, 3

k"J C/3 • —

«+h ■>. t
° £<2
c g
2 £
o.g-
8 “

O * .

— q=

<L>

a> a>

£ %
<D

X> .

C/3

CL >.
<L> 03
to £

o 13
c ^

73 C3

C >
o3 X

. a
X> o

? o

73 CL

• -h on

£ a>
o3 ctf
■ .£ to
) £ 73

r£ ^

CO y
.£ c

73

3 <■+-<
^ O

O ^
.s <D
73 £

(D £

CL

co

If

3 «
i- 3

O g
v- O
D. ,
3 .-2

sz V

.3 XI

3 3

Oj f—

•£ £
a> &
CL

co CO
<1
si

•3 O

£ p

3 °-

i 8

1 c

i w

; x)

> y

- c/5

I U

452

PALAEONTOLOGY, VOLUME 11

For sporophylls which fell in an inverted position, the dehisced sporangium wall and
upper pedicel surface will form the lower compression surface. The pedicel and its
hairy alation, together with the sporangium wall, will be preserved. But the position of
the sporangium wall in the matrix will be influenced by the pedicel collapsing on top of
it. The wall will thus be compressed close to the pedicel surface (PI. 83, figs. 7, 8). The
original lower surface of the pedicel and the keel will collapse on to the underlying
pedicel surface and will not be preserved. The heel collapses to a more or less vertical
position, but remains separated from the pedicel surface (PI. 83, fig. 6). The originally
adaxial lamina mid-rib collapses into its morphologically upper surface; that is, down-
wards in the matrix. It is suggested that the step between the lamina and pedicel in these
inverted specimens is due to the original morphology at the juncture of lamina and
pedicel. Petrified specimens (Arber 1914, pi. 22, fig. 9; pi. 24, fig. 20; pi. 26, fig. 43)
show that the distal face of the heel is a straight continuation of the lower surface of the
lamina, and that there is a slight depression on the upper surface at the distal end of the
lamina, directly above the heel. This is likely to be preserved after compression, as a
difference in level between lamina and pedicel (text-fig. 3).

This interpretation is again consistent with the specimens studied. Specimens in which
the lamina mid-rib is seen as a groove depressed into the matrix (i.e. in the lower
counterpart) have no heel in section, though the sporangium wall can be seen in trans-
verse section (PI. 83, figs. 7, 8). This is never as well defined as in the upright specimens
because, as is suggested above, the pedicel collapses on top of the sporangium wall
during the process of compression. And when the mid-rib appears as a ridge projecting
upwards from the cleaved fossil surface (i.e. in the upper counterpart) the heel is present
in the matrix, as seen in section, and the sporangium wall is absent. The keel is never
seen in either upper or lower counterpart fossils and the pedicel surface is always more
or less flat.

Sporophylls which fell in the sediment in an upright position tend to have a smaller
lamina length pedicel length ratio than those that fell inverted (text-fig. 4). This is pre-
sumably because the relatively greater lamina length favoured settling in the inverted
position. The size of the sporophyll would probably be a function of its position on
the strobilus; the larger sporophylls towards the base, the smaller ones towards the
apex. This conclusion is evident from studies on petrified material (Arber 1914, pi. 23,
fig. 13).

Specimens of sporophylls which have been compressed laterally show the original
vertical aspect of the keel, heel, pedicel, sporangium wall, and lamina. The pedicel
alation is preserved in most specimens (PI. 84, figs. 8, 10) as a rugose margin above the
pedicel, a very doubtful consequence of compression of the hairs which are seen in the
upright and inverted states. From the evidence available in these compression fossils,
it is not possible to ascertain the original morphology of the pedicel alation; whether it
was horizontal, or upturned at the edges to protect (even partially) the sporangium.
Indeed it must be accepted that the attribution of the laterally compressed specimens to
this species (which is based primarily on upright and inverted specimens) is somewhat
arbitrary. But at least all three states can be reconciled with a single reconstruction such
as that offered in text-fig. 1. Until spores are found enclosed on the sporangium, or
other indirect evidence is obtained, the correlation of the sideways compressed fossils
with the other two states remains hypothetical.

BOULTER: COMPRESSED LYCOPOD SPOROPHYLL FROM SOMERSET 453

COMPARISON WITH PREVIOUSLY DESCRIBED SPOROPHYLLS

The majority of specific, and indeed generic, diagnoses of detached lycopod sporophylls
are concerned principally with the size and shape of the lamina and pedicel, and in some
cases with the type of spore extracted. No spores have been found in the present study,
and so the only comparable characters are size and shape. Both these have proved to be

10

CD

<9. o-8

1 06

O'

CD

o °

\ x

X o o O t
X O O

o 1

. O °

o o

o

o X o

x o
x

X

O

3

CO

0-4

02

10 20 30 40 50

length of lamina (cms.)

text-fig. 4. Length of sporophyll lamina plotted against length of pedicel, for
specimens of Lepidostrobophyllum alatum sp. nov. compressed in upright (o) and
inverted (x) states. The distribution shows that the longer lamina give a preferred
orientation towards the inverted state (clustering of symbol .v at right of diagram).

extremely variable in Lepidostrobophyllum alatum and so the comparison with other
sporophylls is limited by our ignorance of the details of sporangium and pedicel for
most of the earlier described species.

Lepidostrobus triangularis Zeiller. The original description by Zeiller (1886) and sub-
sequent descriptions (Arber 1922, Crookall 1966) show that this species is very similar
in shape and size to the smaller specimens described in this paper as having been fos-
silized in the upright state. The pedicel resembles that in specimens of L. alatum which
have cleaved across the layer of the pedicel and not along the sporangium wall (PI. 84,
fig. 5). But specimens labelled as L. triangularis in the Kidston Collection, Geological
Survey, London (nos. 819a, 6249, 8196) show a clearly defined edge to the pedicel,
which is too narrow to have alation. Also these specimens have a completely flat con-
tour, and so their compression history cannot be compared to that of L. alatum.

454

PALAEONTOLOGY, VOLUME 11

Lepidostrobus goodei Jongmans (Jongmans 1931). In his recent review Crookall (1966)
describes this species as having a wedge-shaped pedicel 5-7 mm. long and 2-3 mm. wide.
This again could represent upright specimens of L. alatum in which the pedicel alation
and sporangium wall are compressed in the matrix, if they are present at all; sporophylls
at the extreme tip of the strobilus may not have had a well-developed alation or spor-
angium. The lamina of L. goodei is of a comparable size to upright specimens of L.
alatum.

Lepidophyllum graeile Lesquereux. Lesquereux’s type figure (Lesquereux 1884) has
similar shape and size to upright specimens of L. alatum which have cleaved along the
sporangium wall. The species has not been described from Britain.

Lepidostrobus brevifolius Lesquereux (Lesquereux 1857). This has a wedge-shaped,
keeled pedicel, 0-8-1 -0 cm. long, and a triangular lamina, 6-8 mm. long (Crookall 1966).
This species can again be compared to small specimens of L. alatum.

Lepidostrobus lancifolius Lesquereux. The type figure of this species (Lesquereux 1870)
shows identical features to the inverted state of L. alatum : acuminate apex to lamina,
hastate base to lamina, prominent sporangium wall to pedicel, and similar dimensions.
No details are available in the literature of the step between lamina and pedicel or
whether the mid-rib formed a ridge or a groove.

Lepidophylloides acuminatus (Lesquereux) Crookall (Crookall, 1966). Although no
specimens of L. alatum as large as those described for this species have been collected,
the shape and proportions of the two species are identical.

Lepidocarpon mazonensis Schopf. Schopf (1938) described a side-compressed sporophyll
containing a seed megaspore which has a rugose texture on the margin of the integu-
ments. The comparable structure in laterally compressed specimens of L. alatum is what
is here regarded as the pedicel alation. Except for the megaspore in Schopf’s specimens,
the two types seem identical.

If the lateral states of L. alatum had a single functional seed megaspore and an integu-
ment, rather than the alated pedicel which is suggested in this paper, then they should
be reassigned to Lepidocarpon sp.

Alternatively the interpretation of the lateral state of L. alatum given above could be
correct. And if the specimens described by Schopf are equivalent (i.e. they have an
alated pedicel rather than an integument), both Schopf’s Lepidocarpon mazonensis and
the specimens described here as Lepidostrobophyllum alatum must be reassigned to a
different genus, describing sporophylls with a single functional megaspore and no
integument, namely Acid amy do carp on Schumacker-Lambry (see below).

Thirdly, L. alatum may not have contained a single functional megaspore, in which
case the present separation of Lepidostrobophyllum alatum from Lepidocarpon sp.
holds good.

Cantheliophorus spp. Bassler. This genus (Bassler 1919), described as a ‘sporangiophoric
lepidophyte’, has a similar appearance to the lateral states of L. alatum. Schopf (in
Janssen 1940) refers Bassler’s genus to Lepidocarpon by reinterpreting the structure.
This was supported later (Schopf 1941, p. 559) by the identification of single megaspores

BOULTER: COMPRESSED LYCOPOD SPOROPHYLL FROM SOMERSET 455

from the illustrations in Bassler’s paper. Photographs of LepidophyJlum waldenburgensis
in Nathorst (1914), refigured by Bassler (1919), have a similar character to the specimens
of L. alatum which were laterally compressed.

All the above-mentioned species are defined principally in terms of size and shape,
factors which are variable in L. alatum. It is therefore suggested that the specimens
described in this paper should not be assigned to any previously described species,
unless characters of the pedicel structure and the compression process are investigated
for these established species. Unfortunately, it is not possible to transfer or develop
type specimens to obtain this information, and so full descriptions of species described
over forty years ago are unlikely ever to be obtained. Inevitably, new species must be
made which may overlap the ill-defined limits of other species.

GENERIC ATTRIBUTION OF L. ALATUM

The assignation of the species here described to Lepidostrobophyllum is uncertain,
due to the absence of spores in all specimens. If the sporangium is ever found to have
contained one or several tetrads of functional megaspores, or numerous microspores,
then the assignation is correct. But if the sporangium contained only one functional
megaspore, with three aborted spores (i.e. a Cystosporites type megaspore), then the
sporophylls belong to a genus in the Lepidocarpaceae (Schopf 1941).

In all specimens of L. alatum the sporangium wall extends beyond the pedicel alation
(PI. 83, fig. 4; PI. 84, fig. 4); the alation could never have completely enclosed the
sporangium. Therefore the genus Lepidocarpon is not applicable to any of the specimens
dealt with here.

Schumacker-Lambry (1966) describes a new genus of petrified lepidocarps as Achla-
mydocarpon. The pedicel alation is not developed into an integument, rather, it is as in
the species described here, or absent. The single functional megaspore is completely
protected by a sporangium wall four layers thick. If a single functional megaspore is
ever found in any L. alatum specimens, the generic assignation might be changed to
Acldamydocarpon , though the identification of a thick sporangium wall in compression
fossils presents a severe difficulty. It is evidence that Schumacker-Lambry’s genus
Acldamydocarpon may be partially synonymous with Abbott’s genus Lepidocarpopsis ;
they both represent ‘unintegumented lepidocarps’, although the evidence for lack of
integument is clearer in Schumacker-Lambry’s petrified material than in Abbott’s
compressions.

In his reinterpretation of Bassler’s Cantheliophorus, Schopf refers it to the genus
Lepidocarpon. He did not differentiate between the structures of an alated pedicel and
a thick sporangium wall, so it is possible that some, if not all of the specimens described
by Bassler are members of Acldamydocarpon. In this case, ‘the surface rugosity of the
sclerotic cells on the integument surface’ (Schopf 1941, p. 560) would be cells of the
thick sporangium wall, or of the pedicel alation.

EVOLUTIONARY SIGNIFICANCE

The structure of the pedicel in L. alatum represents an intermediate stage in the
evolutionary change from the unprotected sporangium of many species of Lepidostrobus

456

PALAEONTOLOGY, VOLUME 11

sporophylls to the integumented Lepdidocarpon. Abbott (1963) has instituted a new
genus for such an intermediate, Lepidocarpopsis, whilst Balbach (1966) suggests that
the morphological changes involved are only of specific importance.

The evolutionary tendencies towards specialization of the pedicel and sporangium
structure in arborescent lycopod sporophylls seems to be more diverse than was pre-
viously thought. The simplest Carboniferous Lepidostrobus had completely unprotected
sporangia containing microspores and megaspores. Lepidostrobus takhtajanii , L. masleni,
and Ach! dmy do carp on belgicum seem to represent one extreme in the evolutionary
development with a single fertile megaspore protected by a thick sporangium wall and
no integument, whilst Lepidocarpon , with a single fertile megaspore protected by an
integument around the sporangium, represents another.

The broad extent of the hairy alation in Lepidostrobophyllum alatum is in some degree
morphologically intermediate between Lepidostrobus and Lepidocarpon. The lack of
information concerning the thickness of the sporangium wall and the type of spores in
these specimens make it impossible for us to appreciate the significance of the species,
at the present.

REFERENCES

abbott, m. l. 1963. Lycopod fructifications from the upper Freeport (no. 7) coal in southeastern Ohio.
Palaeontographica, B 112, 93-118.

allen, k. c. 1961. Lepidostrobophyllum fimbriatum (Kidston 1883) from the Drybrook Sandstone
(Lower Carboniferous). Geol. Mag. 98, 225-9.

arber, a. 1914. An anatomical study of the Palaeozoic cone-genus Lepidostrobus. Trans. Linn. Soc.
ser. 2, B 8, 205-38.

arber, e. a. n. 1922. Critical studies of Coal-measure Plant-impressions. J. Linn. Soc. B 46, 171-217.
balbach, m. k. 1 966. Paleozoic Lycopsid fructifications. II. Lepidostrobus takhtajanii in North America
and Great Britain. Amer. J. Bot. 53, 275-83.

bassler, h. 1919. A Sporangiophoric Lepidophyte from the Carboniferous. Bot. Gaz. 68, 73-108.
brongniart, a. 1828. Prodrome d'un Histoire des Vegetaux Fossiles, p. 87. Paris.
cridland, a. c., and williams, j. l. 1966. Plastic and epoxy transfers of fossil plant compressions.
Butt. Torrey Bot. Club , 93, 311-22.

crookall, r. 1966. Fossil plants of the Carboniferous of Great Britain. Mem. Geol. Surv., Palaeon-
tology, IV, part 4.

dix, e. 1934. The sequence of floras in the Upper Carboniferous, with special reference to South Wales.

Phil. Trans. Roy. Soc. Edin. 57, 789-838.
hirmer, m. 1927. Handbuch der Paldobotanik. R. Oldenbourg, Munich.

janssen, r. e. 1940. Some fossil plant types of Illinois. Illinois State Museum, Sci. Pap. 1. Springfield,
Illinois.

jongmans, w. 1931. ‘ Namenandering bei Lepidostrobus' Jaarverslag over 1930, p. 91. Geol. Bureau,
Heerlen.

lesquereux, L. 1857. Fossil Plant Coalfields Pennsylvania. Boston J. Nat. Hist. 6, 4, 430.

1870. Fossil Plants of Illinois. Geol. Surv. of Illinois , 4, 2, 442.

1884. Description of the coal flora of the Carboniferous formation in Pennsylvania, vol. III.

2nd Geol. Surv. Pennsylvania, Report of Progress.

nathorst, a. g. 1914. Zur fossilien Flora der Polarlander, i. 4. Nachtrage zur Paldozoischen Flora
Spitzbergen, 110 pp. Stockholm.

schopf, j. m. 1938. Two new Lycopod seeds from the Illinois Pennsylvanian. Trans. III. State Acad.
Sci. 30, 139-46.

1941. Notes on the Lepidocarpaceae. Am. Midi. Nat. 25, 548-63.

schumacker-lambry, j. 1966. Etude d'un cone de Lepidocarpaceae du houiller beige : Achlamydocarpon
belgicum g. et sp. nov. Acad. Roy. Belgique Memoires, Coll. 4°, Deuxieme series, 17, 1.

BOULTER: COMPRESSED LYCOPOD SPOROPHYLL FROM SOMERSET 457

scott, d. h. 1920. Studies in Fossil Botany. Part 1. Black, London.

selling, o. h. 1944. Studies on Calamitean Cone Compressions by means of serial sections. Svensk
Botanisk Tidskrift, 38, 295-330.

snigirevskaya, N. s. 1964. An anatomical study of the plant remains from the Donets coalballs. I.

Lepidodendraceae. Paleobotanika Akad. Nauk., U.S.S.R. Trudy, Ser. 8, Pt. 5, 1-36. [In Russian.]
walton, j. 1923. On a new method of investigating fossil plant impressions or incrustations. Ann. Bot.
37, 379-91.

1936. On the factors which influence the external form of fossil plants; with descriptions of the

foliage of some species of the Palaeozoic Equisitalian genus Annularia, Sternberg. Phil. Trans. Roy.
Soc. Lond. B 226, 218-37.

zeiller, r. 1886. ‘Bassin houil. de Valenciennes.’ Etudes Gites Min. France, 508 pp. (1888); Atlas
(1886).

M. C. BOULTER

Department of Biology
West Ham College of Technology
Romford Road, London E. 15

Manuscript received 16 June 1967

FUNCTIONAL STUDIES ON THE CRETACEOUS
OYSTER ARCTOSTREA

by R. M. CARTER

Abstract. The three Cretaceous species-groups Arctoslrea colubrina (Lamarck), A. ungulata (Schlotheim), and
A. diluviana (Linnaeus) are described. Particular attention is paid to the detailed morphology of A. colubrina ,
from which its life history is reconstructed. It is inferred that the unusual characters of the genus (especially the
arcuate shape, the zigzag commissure and the funnel spines) relate to the size and importance of the gills as
primary food-collecting organs. The subspecies A. colubrina ricordeana (d'Orbigny) possesses, in addition to
these special adaptations, a set of long spines secreted sub-parallel to the plane of the commissure on the lower
valve; these are interpreted as a specific response to the hazards of inhabiting a substrate of soft ooze. Behavioural
and structural adaptations used by Recent oysters for combating conditions of high turbidity are discussed,
and it is suggested that similar methods were utilized by many extinct species. Funnel spines occur in other
bivalves, including Pinna and Eiheria ; it is likely that they also are connected with inhalent current streams.
Some taxonomic implications of the functional interpretation of Arctostrea are presented in an appendix.

Many malacologists consider Recent Ostrea to represent the peak of bivalvian evolu-
tion (e.g. Atkins 1938), and few would dispute that its integrated gill/palp complex
represents a feeding organ of outstanding power and efficiency. It is one of the very few
Recent bivalve genera for which a large amount of published material is available,
owing this distinction mainly to its economic importance. It appears that such spec-
tacular success as oysters have achieved has at least partly been made possible by their
adoption of a mode of life involving cementation to a hard substrate. Yet because of
their irregular shell form consequent upon such cementation, oysters have understand-
ably always been the bete noire of most systematists, palaeontologists included. It is
the purpose of this paper to apply some of the available information on Recent Ostrea
to a study of a distinctive group of extinct oysters in an attempt to better understand
their morphology, and hence to clarify their taxonomy. It has become apparent to me
during this study that however well known the biology of common European and
American oysters may be, there is still a large amount of primary research needed on
most of the more exotic tropical oysters before our knowledge of the family can be
claimed to be reasonably complete.

The genus to be studied, Arctostrea, has its morphology exemplified by the species
A. colubrina Lamarck, a common oyster in the Cenomanian rocks of Western Europe
(PI. 85, fig. 5). Particularly noteworthy are the superbly developed zigzag commissure
(PL 86, fig. 1), and the highly characteristic arcuate shape in plan view (PL 85, fig. 5).
Other groups of oysters have zigzag commissures, and virtually all lineages of oysters
from the Jurassic onwards have at some time exhibited a tendency to arcuate shape,
but only a relatively small group of species ranging through the Jurassic and Cretaceous
was able to combine these two trends into a successfully adapted working complex.

To detail the biological and stratigraphical details of presently known populations of
Arctostrea is far outside the scope of this paper. For the purposes of morphological
description it will suffice to recognize three major infrageneric categories, conveniently
termed species groups. They are the Arctostrea colubrina group (PL 85, fig. 5), the

[Palaeontology, Vol. 11, Part 3, 1968, pp. 458-85, pis. 85-90.]

CARTER: FUNCTIONAL STUDIES ON CRETACEOUS OYSTER ARCTOSTREA 459

A. ungulata group (PI. 87, fig. 8), and the A. diluviana group (PI. 85, fig. 1), and will be
discussed in turn.

Terminology. In order to facilitate unambiguous discussion of detailed ontogenetic changes, the terms
proximal and distal are preferred to the more normal dorsal and ventral (text-fig. 4); the shell margins
are termed the anterior and posterior arcs. The term ‘vertical zone’ is useful in description of shells
with steep (though not necessarily rigorously vertical) slopes around the valve edge. A vertical zone
may occur virtually all round the commissure (as in Arctostrea colubrina ), or may be limited to
specific areas on the commissure (as in Arctostrea diluviana). It is the inevitable result of shell
secretion along sectors of mantle edge where no mantle cell generation is taking place (see Carter
1967).

Note on captions and material studied. The following abbreviations are used consistently in all plate
and text figure captions: UMZC , University Museum of Zoology, Cambridge University; SM, Sedg-
wick Museum, Cambridge University; BM, British Museum (Natural Elistory), London. It should be
emphasized that this study is based upon the examination of specimens in these museum collections;
statements as to the type of substrate inhabited by different populations of Arctostrea are, in the main,
inferences made from the matrix that still adheres to many specimens.

THE ARCTOSTREA COLUBRINA GROUP

This species group may be recognized by its regularly arcuate shape (PI. 85, fig. 5),
well-developed vertical zone, the numerous sharp zigzags around the commissure (PI. 88,
fig. 5), and the presence of small tubular spines situated at the crest of each commissural
zigzag (PI. 86, fig. 2); the growth track of these spines results in a characteristic divaricat-
ing ornament pattern on the fiat shell top. The group was widely distributed, and com-
mon, in Lower Cretaceous times, and is known from the Middle East, North Africa,
Austria, Germany, France, Portugal, and England.

A. colubrina ricordeana ( d'Orbigny )

The availability of specimens of ricordeana preserving every detail of their life history
in the form of well defined growth-lines enables an accurate ontogeny to be recon-
structed. Though this reconstruction is based on populations of colubrina ricordeana
inhabiting the Lower Chalk Sea of England, the greater part of it applies with only
minor modifications to other populations of colubrina s.s., and also with slightly greater
modifications to members of the ungulata and diluviana species groups.

The Chalk specimens of ricordeana are large, often very inflated, oysters with strong
zigzag valve edges (PI. 86, fig. 1). Adult specimens may possess a dimension greater than
110 mm. in the plane of the commissure, and have a maximum transverse dimension
(‘inflation’) of up to 70 mm. During growth, mantle expansion is at a maximum distally
on the commissure, in a narrow arc here named the generative arc of the mantle
edge (text-fig. 1); elsewhere on the commissure, outside of the generative arc, the
introduction of new mantle cells is minimal, though not altogether absent. Thus shell
secretion during the life of the animal results in the building of a vertical zone around
the greater part of the commissure; this vertical zone is crossed by steep plicae that
represent the tracks of earlier growth stages of the zigzag commissure (PI. 86, fig. 2).

The generative arc is not only the site of localized generation of new epithelial cells,
it is also the site of introduction of all new plicae (PI. 85, fig. 5). Any zigzag on the current

460

PALAEONTOLOGY, VOLUME 11

commissure can always be traced back up its own plica to the point of its inception as
a gentle undulation in the generative arc of an earlier commissure (e.g. plica a, PL 86,
fig. 1).

h.a.

* *1 —

text-fig. 1 . Comparison of the shape of the gill of ( a ) Pecten and
( b ) Ostrea with the adult shell outline of four specimens (c-f in-
clusive) of Arctostrea colubrina (Lamarck). All X i.

All upper (right) valves viewed from above; stippled area
represents the unornamented, pre-zigzag dissoconch, correspond-
ing to the area of attachment on the left valve, h.a., hinge axis;

/, ligament pit; g, generative arc of the mantle edge, arrow indi-
cating direction of mantle expansion.

Ontogeny. The significant stages in the life history of ricordeana are conveniently sum-
marized in discrete steps:

1 . Spatfall on to available hard attachment sites ; commonly pieces of shell and echino-
derm test. This was accompanied by immediate cementation by the surface of the left
valve; by analogy with Recent oysters, it was followed in the next few days by metal
morphosis from the veliger type organization to the adult anatomical form.

2. Continuing shell secretion eventually resulted in the size of the animal exceeding
the size of the attachment object. At some stage after this the mantle edges changed

CARTER: FUNCTIONAL STUDIES ON CRETACEOUS OYSTER ARCTOSTREA 461

from being roughly planar to being sharply zigzag, passing through stages from gently
flexed to more sharply folded (PI. 86, fig. 4). This change in the morphology of the mantle
edge may have been partly under muscular control, but it was probably also connected
with localized epithelial generation. As shell secretion continued unabated, a clear
record of these changes is shown by the successive growth-lines on the valve surface.
In particular, there is always preserved an unornamented pre-zigzag dissoconch that is

sharply demarcated from the following growth stages and gives a clear indication of
the size and shape of the attachment object (PI. 88, fig. 3).

3. When the shell outgrew the area of cementation laterally, the mantle edges of the
left (lower) valve on either side of the mantle isthmus secreted tubular spines which were
attached in a root-like bundle either to the initial attachment object, or to closely
adjacent objects (text-fig. 2). These root-spines were first secreted about the time of the
inception of the zigzag commissure, and obviously served to attach the animal tightly
to the substrate. They continued to be secreted so long as the animal was still attached
to the substrate proximally. Some specimens do not progress beyond this ontogenetic
stage.

4. At this stage in ontogeny, which occurred at very different shell sizes in different
individuals, the shell either broke free from its proximal attachment or, if it was attached
to a relatively small object, overbalanced, and hence came to lie free on the sea floor.
Presumably stimulated by contact with the chalky ooze, the mantle edges of the left

462

PALAEONTOLOGY, VOLUME 11

valve secreted a set of long tubular spines at the crests of the zigzag of that valve (i.e. in
the troughs of the commissure if the shell is viewed in the life position). These spines
(PI. 85, fig. 4; PI. 87, figs. 6, 7; text-fig. 6) extended at right angles to the shell outline,
were sub-parallel to the plane of the commissure, and sometimes half as long as the
shell itself. They were only secreted along the anterior arc of the commissure. Occasional
specimens (e.g. SM B6461) that did not progress beyond the attached phase of ontogeny
(stage 3 above) were able to utilize their anterior arc spines as additional ‘roots’ or
‘props’ to further stabilize their attachment. In these cases, of course, the spines were
not sub-parallel to the plane of the commissure, but became irregular, and were gener-
ally directed downwards towards the surface of attachment.

5. In large, and presumably old specimens of ricordeana there was a slowing of the
rate of introduction of new epithelial material in the generative arc, whilst over-all
shell secretion continued. As a result extremely high vertical zones were built up. When
this was the case, there was often secretion of further sets of spines around the commis-
sure; on any one plica there may have been up to four successive spines, each vertically
above (with respect to life orientation) its immediate predecessor (PI. 86, fig. 3).

Other populations of colubrina

The long spines sub-parallel to the plane of the commissure, diagnostic of A. colubrina
ricordeana , appear to be restricted to very local populations from particular horizons
in the Cenomanian Chalk Marl of England and Normandy. However, populations of
the colubrina group agreeing with ricordeana in all other aspects are common from rocks
of Cretaceous age.

1. The Haslingfield population. A rich fauna of Arctostrea is known from the locality
of Haslingfield, a few miles south of Cambridge. Specimens from this population (PI.
85, fig. 5) are virtually indistinguishable from the Folkestone ricordeana , apart from
their lack of long spines along the anterior arc. Occasionally broken spine stumps suggest
that some specimens did secrete spines in this position, but the majority of the popula-
tion possess very strong clumps of root spines proximally, and were clearly attached
throughout their ontogeny.

EXPLANATION TO PLATE 85

Fig. 1. Arctostrea diluviana (Linnaeus), Senonian, Essen, Germany, xl. Internal view of a typical
upper (right) valve of this species; note the well-developed gill-gutter anterior to the adductor scar,
with pustulose shell texture along its bottom. BM L61234.

Fig. 2. Arctostrea angulata (Schlotheim), Maastrichtian, Maastricht. X 2. Internal view of an upper
(right) valve to show the flat floor; note that the marginal zigzag is of amplitude equal to the depth
of the valve. BM LL8600.

Fig. 3. Arctostrea imgulata (Schlotheim), Upper Cretaceous, Buzi Valley, Portuguese East Africa.
x2. This specimen shows well the characteristic flat-topped shell, and open crested plicae, that are
so typical of the species group. BM L56928.

Fig. 4. Arctostrea colubrina ricordeana (d'Orbigny), grey Chalk Marl, Folkestone, Kent. Approximately
X 4. View of double valved specimen from below, looking up at the divaricating ornament and
snowshoe spines of the under (left) valve. SM B.6451 ; specimen figured by Woods, text-fig. 122.

Fig. 5. Arctostrea colubrina (Lamarck), Lower Chalk Marl, Haslingfield, Cambridgeshire. X 1. A
specimen showing the characteristic divaricating plicae and branchiform shape of the upper (right)
valve when viewed from above, g , present position of the generative arc (growing edge) of the
mantle margin; g', a previous position of the generative arc. SM B.6559.

Palaeontology, Vol. II

PLATE 85

CARTER, Functional morphology of Arctostrea

CARTER: FUNCTIONAL STUDIES ON CRETACEOUS OYSTER ARCTOSTREA 463

2. The Le Mans population. It was noted earlier that on the shell of an adult Arcto-
strea any given plica starts its history in the generative arc of the mantle edge as a small
undulation in the growth-lines. In many populations of the colubrina group, for example,
that from Cenomanian calcareous sandstones of Le Mans, this may develop next into
a marked, short, tubular ‘funnel spine’ (PI. 88, fig. 5), and only later grow into a sharp
zigzag. The plica, of course, retains this funnel spine at its crest, and several more may
be secreted in homologous positions as growth continues (PI. 88, fig. 4). It is striking
that these funnel spines are extremely well developed on the anterior margin, but are
only rudimentary on the posterior (PI. 86, fig. 2); they also tend to be rather better
developed on the right (upper) valve.

THE ARCTOSTREA UNGULATA GROUP

Though the over-all shell morphology of A. ungulata is similar to A. colubrina, there
are significant differences in detail.

New plicae are again introduced distally in the generative arc, but the commissural
zigzag is of much greater wavelength and amplitude than in colubrina. There is thus an
apparent tendency for new plicae to be introduced in pairs on the anterior and posterior
sides of the distal growing edge (PI. 85, fig. 3). The zigzag also has broad, rounded
extremities in contrast to the sharp, pointed extremities of colubrina.

The initial mantle edge reflection at the start of a new plicae (PI. 85, fig. 3; PI. 86,
fig. 6) is broad and strong, and it retains a similar aspect throughout ontogeny. This,
linked with the angle at which the mantle edge is held, results in the characteristic ‘open-
crested’ plicae that typify the species. A similar open-crested morphology may be seen
in the A. pusilla group (PI. 90, fig. 9). The growth lines on the shell of ungulata beauti-
fully display the early stage of differential mantle expansion, during which an initially
gently undulose commissure is transformed into the broad, strong mantle reflection at
the head of each plica (PI. 86, fig. 6). However, after this initial stage of differential
mantle expansion, further shell secretion results in the building up of a vertical zone in
the usual fashion (PI. 85, fig. 3).

It is likely that the differences in morphology between colubrina and ungulata are at
least partly related to a difference in habitat, ungulata being usually collected from
medium-grained or coarser calcareous sandstones, whilst colubrina is often common in
finer grained sediments. The attachment scar is generally small and not well marked,
and whilst some specimens may have been attached proximally throughout life, the
majority probably lay free on the sea floor.

The A. ungulata group is restricted to Upper Cretaceous rocks and has a wide dis-
tribution, ranging from Crimea, Bulgaria, and East Africa, as far north as Holland,
Belgium, and Scandinavia. It seems to replace the colubrina group at this stratigraphic
level. There is little doubt that ungulata developed from an attached species similar to
A. pusilla (Nilsson) (see appendix).

THE ARCTOSTREA DILU VIA N A GROUP

Members of this group have a larger and more solid shell than members of the
previous two, this being a direct result of their being permanently cemented to the
substrate throughout post-larval ontogeny.

464

PALAEONTOLOGY, VOLUME 11

It is characterized by its somewhat irregular, oval shape (PI. 85, fig. 1) and its large
size. Although there may be a zigzag all around the commissure, it is always best
developed over the anterior margins, and it is only here that there is any vertical zone
built up.

The diluviana group is in an Upper Cretaceous development, and seems to have a more
limited distribution than ungulata or colubrina', specimens that I have examined all
derive from northern Europe (Belgium, Holland, Germany, and Sweden).

text-fig. 3. To illustrate the style of shell secretion in an equilateral
bivalve; the growth increments (ruled area) on each valve are of
wedge shape, tapering towards the hinge axis, h.a., hinge axis; v, ven-
tral margin; 5, rate of secretion of shell material.

CONSIDERATIONS ON SHELL FORM

Fundamental tenets. Analysis of shell form in the Bivalvia becomes hopelessly confused
unless a clear distinction is maintained between the basic phenomena of shell secretion
per se on the one hand and generation of mantle cells (i.e. expansion of the area of
mantle epithelium) on the other (Carter 1967). However, provided this distinction is
maintained, it is possible to understand in detail the growth of even complex Bivalvia
such as Arctostrea.

EXPLANATION OF PLATE 86

Fig. 1. Arctostrea colubrina ricordeana (d’Orbigny), grey Chalk Marl, Folkestone, Kent. x2. Arrow
marks the inception of plica ‘a’ in the generative arc; the shell is of a juvenile animal which had not
secreted snowshoe spines at the time of its death. BM LL14751.

Fig. 2. Arctostrea cf. colubrina (Lamarck), Cenomanian, Trouville, Normandy. X 2. Note the develop-
ment of funnel spines (more conspicuous on the anterior than on the posterior margins) in this view
of the posterior vertical flank. BM 65748.

Fig. 3. Arctostrea colubrina ricordeana (d'Orbigny), Chalk Marl, Norman Cement Works, Cambridge.
X 11. View of part of the vertical zone developed on the anterior arc of the lower (left) valve, showing
up to four broken snowshoe spine bases (one above the other) on individual plicae. SM B.6557.

Fig. 4. Arctostrea colubrina ricordeana (d'Orbigny), same specimen as 1. x4. Note the suppression of
the more dorsal plicae.

Fig. 5. Pinna rugosa J. de C. Sowerby, Recent, unlocated. X 11. View of the ventral valve edge to show
the funnel spines. Note how the earlier spines are sealed off by the secretion of a sheet of shell
across their base. UMZC 2008 (Saul Collection).

Fig. 6. Arctostrea ungulata (Schlotheim), same specimen as Plate 85, fig. 2. X 4. View of the posterior
arc, showing the characteristic open crested plicae and the high amplitude of the zigzag.

Fig. 7. Arctostrea colubrina ricordeana (d’Orbigny), grey Chalk Marl, Folkestone, Kent. X 4. Showing
a plica whose growth lines display a zigzag that is starting to diminish in amplitude (3) after an
initial increase from its inception in the generative arc (1) up to a point of maximal amplitude (2).
Plica is situated on the posterior arc of the upper (right) valve. BM L.80734.

Palaeontology, Vol. 11

PLATE 86

CARTER, Functional morphology of Arctostrea

WM.

CARTER: FUNCTIONAL STUDIES ON CRETACEOUS OYSTER ARCTOSTREA 465

Practical implications. In a bivalve of simple equilateral form (e.g. Glycymeris), secretion
of shell is at a maximum somewhere along the ventral margins, and decreases dorsally
on either side of this around the lateral shell margins (text-fig. 3). By virtue of its most
unusual pattern of mantle cell generation, Arctostrea (text-fig. 4) has its point of maximal
shell secretion (Z) situated on the anterior arc of the shell, secretion diminishing at points
around the commissure both distally and proximally to this; then there is a further

proximal

text-fig. 4. An idealized adult specimen of Arctostrea marking in selected growth-lines ( a through to e )
and plicae (1 through to 5). The growth stages corresponding to each growth-line are drawn separately
as growth series b' to e' \ adductor muscle, solid black; h.a., hinge axis; v, visceral mass; gill, ruled
lines; arrow marks functional centre of the inhalent current stream bathing the gill; feathered arrow
centre of exhalent stream. Note how this point migrates relatively in a distal direction during growth.
Z: point of maximal amplitude of zigzag on the anterior arc of the commissure, diminishing antero-
proximally and antero-distally. Z’ \ point of maximal amplitude on the posterior arc of the commissure,
diminishing postero-proximally and postero-distally. (It should be emphasized that growth-lines a to e
and plicae 1 to 5 are only selected examples from the continuous series of growth-lines and plicae that
occur on the shell surface; see also PL 85, fig. 5.)

maximum of secretion (Z\ but note that secretion here is itself absolutely less than at Z)
on the posterior arc, secretion diminishing both distally and proximally from this point
as well.

A further point of interest may be mentioned here: the hinge axis in Arctostrea is
situated at the proximal end of the shell, and hinging takes place by rotation about this
axis (text-fig. 4). However, the extension of the axis itself has a limiting effect upon the
adult shell form of the animal for, if efficient hinging is to be maintained, it is impossible
for the distal growing edge of the shell to transgress over the extension of the hinge axis.
Individual specimens of Arctostrea show many modifications in late adult life to cope
with this restriction, two of the commonest of which can be seen on text-fig. 1. Fig. 1<?
shows an individual that has solved the problem by increasing the curvature of its

466

PALAEONTOLOGY, VOLUME 11

growing edge so as to grow directly back towards the umbones — extreme specimens
may actually fuse their shell into a complete ‘circle’. Fig. 1 / shows the other alternative,
where the growing edge is directed away from the umbones.

Consequences of cementation. Populations of Arctostrea convincingly demonstrate the
close relation that exists between shell cementation and irregularity of form; the
greatest regularity of adult shape is always found in those members of the colubrina
and ungulata groups that were unattached during the major part of their adult life (PI.
85, fig. 4). However, all extremes of variation can be found throughout the Jurassic and
Cretaceous, from specimens that are cemented throughout life, and hence have a mor-
phology mimicking the ‘ diluviana group style’ (even though they may actually be ungu-
lata or colubrina group), to specimens that are briefly cemented but break free early in
adult life; these may be said to have a ‘ colubrina group style’. It is possible to find speci-
mens approaching these two morphological end members from a single population at
almost any locality possessing Arctostrea in abundance. It is small wonder, in the light
of this, that the systematics of the group is burdened by a plethora of unnecessary
names; nor is there any hope of substantial improvement until the adaptive significance
of these morphological trends is understood.

In oysters that remain attached throughout ontogeny by a considerable portion of
their left valve, the shape of the left valve is mainly dependant upon the form of the
surface to which the oyster is attached ; in turn, the shape of the right valve is influenced
by the fact that its edge everywhere has to meet that of the left. Consequently the form
of the whole animal is strongly influenced by the shape of the substrate to which it is
cemented. This is clearly demonstrated in the specimens of A. pusilla figured on PI. 90,
figs. 3-4, 6-10.

All adult Arctostrea have a sharply zigzag valve edge. Assuming that the shell is
cemented to a reasonably flat surface, it is, of course, only possible for the animals to

EXPLANATION OF PLATE 87

Figs. 1, 2, and 3. Arctostrea alaeformis (S. Woodward), Upper Chalk, Norwich. X 1. Internal, ventral
and external views of the lectotype. NCM F.C.2100; specimen of Woodward 1833, plate 6, fig. 3.

Fig. 4. Arctostrea colubrina ricordeana (d’Orbigny), grey Chalk Marl, Folkestone, Kent. X 2. Internal
view of the upper (right) valve. Note especially the lack of room for a promyal passage dorsal to the
adductor (arrow); and the characteristic pustulose nature of the shell on the floor of the valve.
SM B.6462.

Fig. 5. Arctostrea pusilla (Nilsson), Upper Chalk, Norwich. x2. A specimen with the more usual
open textured plicae along the lateral margins, but with large open folds of the 'larva' type along
the distal margins. View looking into the generative arc. NCM. F.C.2129a; specimen figured by
Woods, plate 58, fig. 2.

Fig. 6. Arctostrea colubrina ricordeana (d'Orbigny), grey Chalk Marl, Folkestone, Kent. x4. Specimen
showing a funnel spine located at the top of a plica (arrow). In the adult growth stage, the same plica
is the site of secretion of a snowshoe spine, subparallel to the plane of the commissure. SM B.6459.

Fig. 7. Arctostrea colubrina ricordeana (d’Orbigny), grey Chalk Marl, Folkestone, Kent, x 2. Typical
development of short snowshoe spines in a specimen of only moderate overall size. SM B.6451;
specimen figured by Woods, text-fig. 122.

Fig. 8. Arctostrea ungulata (Schlotheim), same specimen as Plate 85, fig. 3. X 1. View from antero-
dorsal aspect.

Fig. 9. Chlamys asperrimus (Lamarck), Recent, Tasmania. X 2. Note the presence of a true zigzag
commissure (and not merely an interlocking valve edge) of low amplitude; this is one of the very
few examples of this feature in the Bivalvia outside of the Ostreacea. SM D.20869.

Palaeontology , Vol. 11

PLATE 87

CARTER, Functional morphology of Arctostrea

CARTER: FUNCTIONAL STUDIES ON CRETACEOUS OYSTER ARCTOSTREA 467

produce a regularly zigzagged valve edge if shell secretion has been such that after a
particular stage in ontogeny the growing edge of the shell is not in contact with the
substrate. This may be accomplished in one of two ways: either the rate of mantle
expansion is such that the shell soon comes to grow outside the edge of the object to
which it is attached, which hence must be relatively small; or, if the attachment object
is large, there must be a cessation of mantle cell generation during continued shell
secretion, resulting in the formation of a marginal ‘vertical zone’. The former is the
dominant solution utilized by members of the colubrina and ungulata species groups;
the latter by the diluviana species group. In either case, adult animals possess a regular
and very sharply zigzagged commissure, the form of which is entirely independent of
the form of the animal’s attachment surface.

h.a.

text-fig. 5. Theoretical types of zigzag commissure, (a) zigzag with diminu-
tion of amplitude dorsally, but with straight flanks; efficient hinging of the
shell is impossible; (b) a geometrically graded zigzag, with both diminution of
amplitude dorsally, and curved flanks to enable unimpeded opening and
closing of the shell; (c) zigzag without diminution of amplitude dorsally; note
that such a zigzag cannot reasonably cover any length of the lateral shell
margins, and that it involves a non-linear secretion gradient across the shell.

The problem of grading. Interesting questions are raised by the interaction of the secre-
tion pattern of Arctostrea with its possession of a zigzag commissure.

In a closely reasoned analysis of zigzag commissures in the Brachiopoda, Rudwick
(1964) has introduced the concept of a graded zigzag s/it, which is defined as a slit ‘of
uniform width all round the commissure, from one suppression point [the point where
the zigzag dies away to nothing on the lateral shell margins] to the other’.

It is useful to isolate the morphological components that go to make up a graded
zigzag slit:

1. Diminution. For a zigzag slit that occupies any appreciable length of the whole
commissure there is a gradual diminution of amplitude of the zigzag as the hinge
axis is approached (text-fig. 5a). This is inevitable if a linear secretion pattern is to be

468

PALAEONTOLOGY, VOLUME II

maintained across the shell, though it is, of course, possible to imagine a shell form in
which the zigzag does not diminish in amplitude dorsally and hence does not possess a
linear secretion pattern (text-fig. 5c).

2. Curvature of the flanks. As Rudwick has already clearly pointed out (1964), by
virtue of the hinging motion of any bivalved shell, it is impossible for a zigzag situated
on the lateral margins to have straight flanks. Such a shell (text-fig. 5 a) would be unable
to open because of the occlusion of the shell edges in the crests and troughs of the
zigzag. It is therefore necessary that the flanks of the zigzag be curved, concave dorsally,
in order to overcome this occlusion (text-fig. 5b). An inevitable result of such flank
curvature is that there is a tendency to equalization of width of the slit.

Rudwick is of the opinion that the actual flank curvature is due ‘partly to the . . .
(requirements) ... of a slit of uniform width, and partly to the fact that the valve edges
move apart by rotation’. In other words, he suggests that there was a positive selective
value in favour of a slit of uniform width during the evolution of zigzag slits, and that
their flank curvature cannot be explained on a basis of shell secretion tenets alone.

Whilst this may be the case in the Brachiopoda, within the limits of accuracy of the
measurements it is possible to make, it appears that the curvature of the flanks of the
zigzag in species of Arctostrea is no more than is necessary to ensure the adequate
hinging of the shell. An inevitable result of this is a slit of roughly equal width all round
the commissure, but economy of hypothesis precludes treating this equality of width
as of prime selective significance. As the term ‘grading’ is of considerable usefulness in
morphological description, it is necessary to distinguish ‘adaptive grading’ where the
equalizing of the width of the slit is thought to have adaptive significance over and above
that concerned with shell hinging (i.e. grading sensu Rudwick), from ‘ geometric grading ’,
where all elements of the grading, including both diminution of amplitude and curvature
of the zigzag flanks, are thought to be explicable in terms of basic tenets of shell secre-
tion.

EXPLANATION OF PLATE 88

Fig. 1. Etheria elliptica Lamarck, Recent, Africa. X 1. Well developed funnel spines at spaced intervals
around the edge of the commissure. These spines must be secreted by reflected lobes of mantle tissue,
and are sealed off as soon as they are no longer right at the growing edge of the shell. UMZC 2000
(McAndrew Collection).

Fig. 2. Crassostrea echinata (Quoy and Gainrard), Recent, Australia. X 6. A juvenile shell (umbo
arrowed) showing the extreme development of flaring, funnel spines on the ventral side of the shell.
The long, thin funnel spines of a slightly later growth stage can be seen out of focus in the back-
ground. UMZC 2007.

Fig. 3. Arctostrea colubrina (Lamarck), Lower Chalk, Cherry Flinton, Cambridgeshire. X 1. Note the
presence of two folds (arrowed) on the posterior ear; these probably reflect the presence of persistent
pseudo-siphons in the mantle edge. SM B.358; figured by Woods, text-fig. 135, and Rudwick (1964),
plate 29, fig. 5.

Fig. 4. Arctostrea cf. colubrina (Lamarck), Cenomanian, Le Mans, Normandy. X 7. A specimen
exhibiting the successive secretion of several funnel spines on individual plicae; note the very low
amplitude of the zigzag. BM LL26833.

Fig. 5. Arctostrea cf. colubrina (Lamarck), Cenomanian, Le Mans. X 7. View of the generative arc
(distal end) of the right valve of a young individual. Note the sharp commissural zigzag with funnel
spines developed at the crests. BM LL26832.

Palaeontology, Pol. 11

PLATE 88

CARTER, Functional morphology of Arctostrea

CARTER: FUNCTIONAL STUDIES ON CRETACEOUS OYSTER ARCTOSTREA 469

The uniqueness of individual specimens. The shell of any specimen of A. colubrina has a
unique morphological construction, yet its secretion was none the less governed by certain
basic conditions:

1. The vast majority of new epithelial cells were introduced in a narrow generative
arc of the mantle edge.

2. The absolute secretion of shell material was a function of distance from the hinge
axis, being maximal at the point on the commissure furthest from the hinge axis.

3. After a certain growth stage had been passed, the shell possesssed a zigzag com-
missure; at all further growth stages this was geometrically graded in such a way as to
ensure the free hinging of the shell and the tight approximation of the valve edges when
shut.

The consequence of these conditions is that every single plica on the shell of a typical
Arctostrea has a morphologically unique ontogeny.

Considering (for convenience only) the hinge axis as a fixed direction, it is apparent
that the generative arc of the mantle edge occupies a geometrically unique position for
every successive commissure throughout the life of the animal (text-fig. 4, positions a
through to e). Since the generative arc is the site of inception of all new plicae, each
plica starts its ontogeny in a geometrically unique position, and its detailed structure is
a reflection of its exact positions on either the anterior or posterior margins.

These detailed structural changes are mirrored on the valve surface in the form of
growth-lines (PI. 86, figs. 1, 7). A study of these reveals that the earlier formed plicae (e.g.
plica 1 of text-fig. 4) pass through a series of growth stages which may be idealized for
conceptual convenience. This series generally follows the pattern illustrated by the plica
of PI. 86, fig. 7.

Initially the amplitude of the zigzag is small, corresponding to its inception on the
generative arc of the commissure (position 1 of PI. 86, fig. 7), but this rapidly increases
towards a maximum. Meanwhile, further mantle expansion has been taking place
distally (in the generative arc), causing the relative migration of the zigzag around the
commissure in a proximal direction. The amplitude of the zigzag reaches a maximum
(position 2 of PI. 86, fig. 7), and then starts to come under the influence of the diminu-
tion of amplitude along the antero-proximal part of the commissure; it hence decreases
in amplitude (position 3 of PI. 86, fig. 7). On the very earliest formed plicae (e.g. plica 1
of text-fig. 4; PI. 86, fig. 4) this decrease in amplitude may be carried as far as total
suppression of the zigzag.

THE FUNCTIONAL INTERPRETATION OF ARCTOSTREA

Plicae. Viewed from above, the shape of an adult Arctostrea is crescentic (text-fig. 1 ;
PI. 85, fig. 5); it is immediately obvious that the inner arc of the crescent (i.e. the
posterior margin) is shorter than the outer arc of the crescent (the anterior margin).
It follows that for a zigzag of given wavelength, on any given commissure there are fewer
zigzags (and therefore plicae) on the posterior margin than on the anterior. The more
sharply crescentic is the over-all shell shape, then the fewer plicae there are posteriorly
compared with anteriorly. For example, on a fairly sharply recurved specimen (SM
B.6557) there are 13 posterior plicae, and 19 anterior plicae. There is no need to seek

470 PALAEONTOLOGY, VOLUME 11

a functional interpretation of this fact; it is an inevitable consequence of the mode of
growth.

Spines. There are three different types of spine on A. colubrina ricordeana, and it is
possible to attribute to each a discrete functional role:

( a ) Attachment spines. It is self-evident that the complex of spines secreted at the
proximal end of the shell around the attachment area served to strengthen the attach-
ment of the shell to the substrate. Their restriction to this part of the shell, their irregular

text-fig. 6. Shell outlines of Arctostrea colubrina ricordeana to show the disposition
of spines on the left valve with respect to the over-all shape of the shell; specimens
viewed from above. Note especially their concentration around the umbonal end of
the shell (attachment spines), and their disposition at right angles to the shell
margins (snowshoe spines). Stippling as for text-fig. 1 ; specimen without stippling
is a left valve only. X f . All specimens from the grey Chalk Marl, Folkestone, Kent.

BM L4855.

and intertwined shape, and the fact that they are occasionally preserved attached to
hard objects such as other oysters, all support such an interpretation.

(, b ) Snowshoe spines. Similarly, it was self-evident to Woods over fifty years ago that
‘the long regular outgrowths from the margin of the valves occur [only] in specimens
from the Chalk Marl, and were no doubt adopted for the purpose of fixation in the
soft sediment of the sea floor’, but it is worth noting how closely the observed mor-
phology of the spines approaches the paradigm for a ‘snowshoe’ function. By virtue of
its arcuate shell shape, the distal and proximal ends of the shell in an adult colubrina

CARTER: FUNCTIONAL STUDIES ON CRETACEOUS OYSTER ARCTOSTREA 471

morphologically define one ‘side’ of the shell, whilst the curved anterior margin defines
the other (text-fig. 6). Most of the weight of the shell is concentrated on the anterior
margins, and it is clearly there that the animal was in most danger of sinking into the
substrate. Hence it is there that the spines are developed, one at the trough of every
zigzag. Their length, their sub-parallelism to the plane of the commissure, their dis-
position at right angles to the shell outline, and the successive secretion of more than
one spine on individual plicae in large adults with high vertical zones (PI. 86, fig. 3), all
add up to unequivocal evidence that the spines prevented the animal from being
smothered by sinking into the soft ooze.

(c) Funnel spines. The function of the small funnel-like spines, developed at the crest
of the zigzags on the right (upper) valve, will be discussed in a later section on the func-
tion of the zigzag itself.

The arcuate shape. From the very early stages of ontogeny of Arctostrea , the localized
generation of epithelial material in the generative arc of the mantle edge exercised a
controlling influence on the shell form; it is directly responsible for the characteristic
arcuate shape in plan view. This arcuate shape approaches a logarithmic spiral, since the
shell always tends to grow in such a way as to preserve over-all similarity of shape.

However, the tendency of oyster shells to arcuate shape is by no means confined to
Arctostrea , and Lang (1923) has coined the term lunate to describe it. Functional
explanations are not common; most writers seem content to follow the suggestion of
Douville (1910) that this shape of shell is ideally suited to resist water currents. At its
best this postulate is difficult to understand; and the fact that a lunate trend can be
recognized so frequently in groups of oysters at such different points in time and space
(e.g. Liostrea cmabarensis Bodyl from the Valanginian of Russia (text-fig. 7); Crassostrea
ameghinoi roccmci von Ihering from the Danian of Argentina (text-fig. 8); and the Recent
Crassostrea angulata (Lamarck)) leads one to be dissatisfied with Douville’s explanation
and to suspect that some basic adaptation is involved.

The soft-part anatomy of Ostrea edulis (text-fig. 9) is fairly characteristic of the Ostre-
acea as a whole, and most members of the Pectinacea have a grossly similar anatomy.
This anatomical similarity is undoubtedly a result of their mode of life on the substrate
with the commissure horizontal, possessing the ability to take in water (and hence food)
over a large part of their commissure. Many bivalves that live with their commissure at
right angles to the substrate (e.g. Area) have gills which hang vertically in the mantle
cavity, and which are not fused laterally to the general mantle surface. Generally
speaking, however, the assumption of a horizontal mode of life necessitates the lateral
fusion of the gills (particularly the upper) with the mantle surface; otherwise gravity
would cause them to fold together in a most inefficient way. It is thus perhaps significant
that the advanced pseudo lamellibranch gill condition is characteristic of the Ostreidae,
Pectinidae, and Pteriacea — all groups which live with the commissure sub-parallel to
the substrate.

The presence of an inhalent current along a considerable part of the commissure
requires the development of larger and more complex organs to catch, sort, and process
the available food supply. The Pectinacea and Ostreacea have dealt with this require-
ment by enlarging the gill and expanding it into a characteristic crescentiform organ
fitting snugly between the major adductor muscle and the free valve margins. A

472

PALAEONTOLOGY, VOLUME 11

comparison of the shape of the gill of a typical pectinid and ostreiid with the shell shape
of 'lunate’ oyster species (text-fig. 1) can leave no doubt that this lunate shape is a direct
reflection of the size and importance of the gill in fossil species such as A. colubrina.
This organ was of such paramount importance to these fossil species that it assumed a
dominance over all other soft parts.

text-fig. 7. A lineage of oysters from the Upper Jurassic and Lower Creta-
ceous of Siberia that exhibits a continuous trend towards branchiform shape;
notice the concurrent reduction in size of the attachment scar (stippled).

(a) Liostrea ex. gr. delta (Smith), Kimmeridgian; (b) Liostrea praeanabarensis
Zakharov, Volgian; (c) Liostrea anabarensis Bodyl, Valanginian. All X %

(after Zakharov 1965, fig. la, e, and i).

The gill of Recent bivalves, including oysters, grows by the addition of filaments to
its distal end (Yonge 1960). This corresponds exactly with the position of the ‘generative
arc’ in fossil Arctostrea; as the gill grows by distal addition of new filaments, so must the
shell have been expanded in a similar direction in order to provide protection for the
growing gill. Examination of Recent oysters (e.g. Pycnodonte hyotis (Linnaeus)) in
which the over-all shell shape is not arcuate, reveals the presence on the floor of the
valves, particularly the left, of a marked arcuate gutter in which the gill is situated. There
is a similar ‘gill gutter’ in many fossil species (e.g. Arctostrea diluviana (Linnaeus),
PL 85, fig. 1). In both fossil and Recent species the floor of this gutter is often lined with
shell of a peculiar pustulose character; in specimens of Arctostrea similar pustulose
structure can be observed along the length of the shell (PI. 87, fig. 4).

In view of the close functional relationship between the gill and shell shape, the term
branchiform will be used to describe species that reflect this relationship in their external
shell form.

Sanitation. Sanitation is a general problem that arises amongst animals that have
adopted a horizontal mode of life on the sea floor. In the Bivalvia it is usually effected

CARTER: FUNCTIONAL STUDIES ON CRETACEOUS OYSTER ARCTOSTREA 473

by the accumulation of pseudofaeces around the edge of the mantle cavity, and their
periodic ejection by sharp contractions of the adductor, but such a mechanism may be
insufficient in turbid waters. It appears to be accepted amongst writers on Recent oysters
(e.g. Menzel 1955) that the presence of a promyal chamber in Recent Crassostrea is a
specific adaptation to allow a greater flow of water
through the shell, and hence to aid with sanitation and
to prevent the exhalent chamber from being invaded
by sediment. (The promyal chamber (present in the
right valve only) is an additional passage for the ex-
halent current situated dorsal to the adductor muscle;
it is formed by the local separation of the mantle from
the underlying body-tissues.)

Certain extinct oysters certainly possessed promyal
chambers. For instance, the Danian Crassostrea
ameghinoi rocana has an adductor scar situated distant
from the dorsal margin (text-fig. 8), a sure sign of the
promyal chamber in Recent species. In addition there
is a conspicuous pedal retractor scar in this species,
suggesting that the foot was an active aid to mantle
cavity cleansing.

However, the morphology of Arctostrea is such
that we can be equally certain that it lacked a promyal
chamber. The adductor scar is generally very close to
the dorsal margin, and there is no room at all for an
efficient flow of water dorsal to it (PI. 87, fig. 4). Yet
the most branchiform specimens of Arctostrea in the
Jurassic and Cretaceous are found in fine-grained
muddy sediments such as the Oxford and Kimmeridge
Clay, and the Chalk. It is in just that type of environ-
ment that efficient water circulatory and cleansing
mechanisms are of paramount importance; in the absence of a promyal chamber
it must be assumed that the large size of the gill itself was adaptive in this respect,
since it would ensure a very powerful current of water through the branchiform
shell.

The zigzag commissure. Rudwick (1964) concluded that in the Brachiopoda zigzag com-
missures serve two purposes, both fundamentally protective: firstly, for any given gape
a zigzag prevents the entry of particles above a certain size, whilst conferring a relative
increase in the area of slit actually functioning as intake; secondly, since the mantle
tissue lining a zigzag slit is sensitive, there is also a relatively larger amount of sensitive
warning tissue available to the animal in proximity to its shell edge.

One would expect it to be significant that the only Bivalvia that possess plicate com-
missures homoeomorphic with the brachiopods are those that have adopted the brachio-
pod mode of life and ecologic niche. Whatever the significance of zigzag commissures
might be, one would a priori expect it to be similar in the two groups. Yet it does not
seem to me that Rudwick’s explanation is applicable to Arctostrea.

i i

text-fig. 8. Crassostrea ameghinoi
rocana von Ihering, Danian, Argen-
tine. xl. Reconstruction of the gill
and feeding currents. I, inhalent cur-
rent; E, exhalent through current;
EP , exhalent current from promyal
chamber ; PR, pedal retractor muscle ;
adductor in solid black; gill, ruled
lines. BM LL26838.

C 5586

474

PALAEONTOLOGY, VOLUME 11

The first of the above postulates is unsatisfactory on two counts. Firstly, Arctostrea
lives dominantly in finer-grained sediments, generally of or below sand grade; 1 know
of only one locality where it is collected associated with small pebbles. Secondly, and
more seriously, bivalves do not remain feeding whilst being pelted with a miniature
hailstorm of sediment, no matter how paradigmatically rigorous their zigzag commis-
sures may be. In conditions of high turbidity, the average bivalve firmly closes its shell.
It is fair to conclude that Arctostrea , when inhabiting its usual facies, would have had its
valves gaping more widely than the average size of particles present in the substrate
and therefore that the zigzag commissure would have offered virtually no protection at
all against ingestion of such material.

The second of the postulated functions has the same structural paradigm as the first
(one of the problems of mechanistic analyses is that differing functional interpretations
can have very similar structural paradigms), and there is little that can be added to
Rudwick’s original discussion. Undoubtedly, given that the mantle edges are sensitive
to clouds of (say) silt, and will cause the shell to close on contact with such a cloud,
then a zigzag commissure is relatively more efficient than a similarly long planar com-
missure. But whether this gain in efficiency is sufficient to explain the evolution of a
zigzag slit is open to some doubt, particularly as this interpretation fails to account for
one of the more striking features of the morphology of Arctostrea — that the zigzag is
conspicuously just as well developed over the exha lent posterior borders as it is over the
inhalent anterior borders (PI. 86, figs. 2, 6, 7). (This was not as marked a feature of
the earlier Jurassic members of the group, where the zigzag is noticeably not as well
developed over the exhalent as over the inhalent currents.)

Assuming that the zigzag is a significant feature of the morphology, how then is it to
be interpreted ? It would not seem likely that the zigzag was for the purpose of preventing
anything from leaving the shell. We are left with two alternatives:

1. The zigzag was physically preventive against the entry of harmful agents. We have
already concluded that it is unsatisfactory to believe that the zigzag slit served to keep
out large ‘ball-bearing’ type particles, particularly since it is equally developed all
round the shell. The implication is that, if the zigzag was to prevent the entry into the
shell of some harmful agent or other, then this agent must have been capable of pene-
trating against the powerful exhalent stream of the oyster. This immediately precludes
sediment and suggests a motile, and hence organic, agent. Thus it might be considered
that the zigzag served to prevent the penetration of small irritants such as Crustacea and
fish into the mantle cavity (M. J. S. Rudwick, personal communication). Whilst this is
a function which a zigzag commissure must inevitably fulfil by nature of its morphology,
it is open to some doubt whether or not this is an explanation that can explain the evolu-
tion of the zigzag.

2. The zigzag had a beneficial hydrodynamic effect on feeding currents. A zigzag slit
has an inevitable hydrodynamic effect on the water flowing through it. First, and most
important, due to the geometry of the zigzags there is an effective vertical spatial spread
of the inhalent and exhalent currents; this is so whether the current was an external one
or was merely the feeding current intrinsically generated by the oyster itself. Accompany-
ing this, there would probably be a tendency for jet streams to form at the crests and
troughs of the zigzags; any such tendency would be of considerable value, particularly
on the exhalent margin where it would assist in carrying ejected sediment away from the

CARTER: FUNCTIONAL STUDIES ON CRETACEOUS OYSTER ARCTOSTREA 475

edge of the shell (cf. the exhalent gill streams of the Lamprey). It must also be remem-
bered that it is not only the topology of the shell edges that affects the inhalent current
stream of a feeding oyster. In all Recent oysters the muscular flexibility of the actual
mantle edges is an important factor influencing the direction and strength of feeding
currents. By approximating its mantle edges along the flanks of the zigzag, but leaving
them parted at the crests and troughs, Arclostrea could have greatly enhanced any
natural tendency for jets to form at these points.

s

text-fig. 9. (a) Soft part anatomy of Ostrea edulis Linnaeus (after Yonge 1960). p, palps; g, gill;
am, adductor muscle; in, mantle edge; x, distal point of junction of the gill and mantle lobes; a, anus;
h, heart; s, visceral mass. I, inhalent current stream; E, exhalent current stream. ( b ) Diagrammatic
cross-section through the mantle cavity and gills of Ostrea edulis, with feeding currents indicated by

arrows, c.p., plane of the commissure.

The pseudolamellibranch gill of an oyster is roughly W-shaped in cross-section, each
V corresponding to a demibranch (text-fig. 9b). There are two such gills, which are
attached to the mantle surface marginally and distally, and thus effectively partition
the mantle cavity into two chambers and act as a ‘sieve-strainer’ to the inhalent current
stream. In a typical oyster the gills are fairly close together and form thin high Ws
(text-fig. 9b). Since the lateral cilia on each filament draw the water through the gill at
right angles to the gill lamellae, the general inhalent current stream is impinging on the
gill lamellae at a very oblique angle and filtering is somewhat inefficient. Obviously
more efficient feeding, and in particular a higher rate of water filtration, will be possible
if the gill-W becomes wider, for this would ensure that the current stream impinges on
C 5586 I 1 2

476

PALAEONTOLOGY, VOLUME 11

the individual gill lamellae at a high angle (text-fig. 10). Of course, in order for this
adaptation to be possible, the animal first has to have evolved a deeper body cavity.
However, given such a deeper body cavity, feeding would obviously still be relatively
inefficient if the commissure were to remain planar, for the inhalent current would not
spread laterally far enough to meet directly the majority of the gill surface. But should
the valve edge become plicate, then these problems are solved at a single stroke: the

text-fig. 10. Reconstructed cross-sections through the gill and mantle cavity of Arctostrea. ( a-a ') an
inhalent current entering the shell at the crest of a zigzag is best situated to bathe the upper gill;
(b-b') an inhalent current entering medially on the flank of a zigzag is best situated to bathe the inner
demibranchs of both gills; (c-c') an inhalent current entering at the trough of a zigzag is best situated
to bathe the lower gill, c.p., theoretical plane of the commissure.

gill-W can become relatively wider and still be bathed directly in strong inhalent current
streams. The effect of a zigzag slit such as that of PI. 86, fig. 6, must be to effectively
separate the inhalent current into a set of streams (at the crests of the zigzags) for the
upper gill, and a set of streams (at the troughs of the zigzags) for the lower gill. And since
it is equally important that there be efficient exhalent current streams the zigzag is
developed over the exhalent length of the commissure as well. Phylogenetically one
would expect the inhalent zigzag to be the first to be selected for, which is just what is
observed in primitive populations of Arctostrea.

Further comments on diminution. It was earlier noted (p. 468) that it is theoretically

CARTER: FUNCTIONAL STUDIES ON CRETACEOUS OYSTER ARCTOSTREA 477

possible to construct a shell in which the zigzag does not diminish in amplitude along
the lateral flanks, but only by assuming that the secretion pattern across such a shell be
non-linear (text-fig. 5c). In view of the almost axiomatic presence of linear secretion
patterns in the Bivalvia (this contrasts with their common absence in the Brachiopoda),
it is possible to argue that no specific functional (or adaptive) significance be attributed
to diminution of amplitude of a zigzag on the lateral shell margins, for such diminution
is an inevitable result of a linear secretion rate. By terming the zigzag commissure of
Arctostrea a ‘geometrically graded’ zigzag, I have aligned myself with this view. How-
ever, should it be felt that a more strictly adaptive explanation is necessary in order to
satisfactorily account for the diminution of zigzag amplitude, the interpretation given
above for the over-all shell shape and initial presence of the zigzag (that they are linked
with the size and importance of the gills) may also be applied here.

Several writers (e.g. Yonge 1926) have noted that in Recent Ostrea the inhalent
current is strongest in the middle of the gill, and falls off considerably to either side of
the mid-point (cf. text-fig. 9). The diminution of amplitude of zigzag in Arctostrea is
such that it occurs towards the two ends of the gills, thus automatically ensuring that
the inhalent current be relatively weaker here than in the centre of the gills. Obviously,
though a shell such as text-fig. 5c is theoretically possible, such a zigzag would furnish
undesirably strong inhalent currents right at the ends of the gills. It may also be noted
here that the complete suppression of the more dorsal zigzags on the commissure of
Arctostrea (cf. PI. 86, fig. 4) only takes place after the zigzags in question have migrated
relatively dorsal of the functional end of the gill (Plica 1, text-fig. 4).

Zigzag commissures in other bivalve genera. The presence of a true zigzag commissure
is rare in the Bivalvia; the only other occurrence known to me is in the Australian
pectinid Chlamys asperrimus (Lamarck) where a sharp zigzag of low amplitude and
relatively constant wavelength can be seen (PL 87, fig. 9). It seems probable that zigzags
are related to the demands made by inhalent currents spread round most of the com-
missure, and hence are restricted to bivalves that live lying on the substrate with their
commissures horizontal.

Funnel spines

1. Arctostrea. Granted that the zigzag valve edge has the functional significance
attributed to it above, we have an important clue as to the function of the small ‘funnel
spines’ of Arctostrea. It will be remembered that these spines are invariably developed at
the crests of plicae on the anterior valve margins but are usually poorly developed over
the posterior valve margins (cf. PI. 86, fig. 2; PI. 88, fig. 5). Thus they are concentrated
over the inhalent arc of the commissure and are only developed during that part of the
ontogeny of a plica before the true zigzag of the plica has fully developed. It would,
therefore, seem likely that they serve the same function as the zigzag itself does later;
in other words they enable all parts of the gills to be irrigated with currents of water
impinging as far laterally as possible. It is worthy of further note that those few popula-
tions that habitually secrete several of these funnel spines on a single plica, one beneath
the other (PI. 88, fig. 4), are just those populations that have a zigzag of conspicuously
small amplitude.

It should be stressed at this point that though the above discussion has attributed
a function to the shell spines themselves, it is, of course, flaring funnels of epithelium

478

PALAEONTOLOGY, VOLUME 11

along the mantle edge that have the primary functional significance as far as inhalent
current streams are concerned; the actual funnel spines on the shell are doubtless
secreted partly as support, and partly as protection for such epithelial funnels. It is
possible, particularly in fossil species, that flaring funnels of epithial material could have
reached well away from the shell edge; similar funnel spines are known in fossil pro-
ductoids (Muir-Wood and Cooper 1960, pi. 21).

2. Other Bivalvia. Discussions of inhalent and exhalent currents in the Bivalvia
usually tacitly assume that the ideal inhalent current is one of constant velocity and
magnitude, covering as great a length of commissure as is consistent with other struc-
tural demands. There is considerable circumstantial evidence that this view is erroneous.
The vast majority of infaunal bivalves receive their inhalent current either down a
closely confined siphonal passage or else through a narrow restricted part of the mantle
edge. It is likely that gill morphology and function has evolved in step with this, and is
in many cases better suited to deal with strong localized inhalent currents than with
weak, widely spread ones.

There are other bivalves apart from Arctostrea that have conspicuous tubular spines.
Pinna rugosa Sowerby (PI. 86, fig. 5) and Etheria elliptica Lamarck (PI. 88, fig. 1) are
two of the more striking examples, and it is reasonable to interpret their spines too as
reflecting the presence of epithelial funnels that provide the gills with a series of laterally
spread, spaced inhalent streams. It is noteworthy that in both these species the spines
are only open to the exterior when they are situated right at the very edge of the shell;
as soon as they have been displaced even slightly dorsally by further mantle expansion
and shell secretion they are sealed off with a thin layer of shell. There is a definite need
for observational and experimental evidence to test this theory; but even should it
ultimately prove to be untenable it is certainly the most reasonable explanation for these
tubular spines in our present state of knowledge. None of the paradigms for other
favourite functions attributed to spines (sensory, camouflage, protective, etc.) is closely
approached by the morphology of the spines of rugosa and elliptica. In particular, none
of these functions will fit in with the known ecology of Etheria — cemented to rocks in
the rapids of fast flowing, turbid freshwater rivers. Yet the funnel hypothesis might well
be particularly applicable here; for Etheria has a conspicuously elongated shell in a
dorso/ventral direction. If this should be oriented parallel to the flow of the river, with

EXPLANATION OF PLATE 89

Figs. 1, 2. Lopha folium Linnaeus, phenotype folium , Recent, Farsan Islands, Red Sea. X 1. External
and lateral view of a left valve, attached to a thin twig by the secretion of clasping spines. Note the
dorsal diminution of amplitude of the zigzag commissure. SM D.20866.

Figs. 3, 4. Lopha folium Linnaeus, phenotype cristagalli , Recent, unlocated. X 1. External and lateral
view of a double valved specimen. Note the large amplitude of the zigzag commissure and the secre-
tion of spines as ‘props’ for assisting in stability of fixation. UMZC 2002 (Saul Collection).

Figs. 5, 6. Lopha folium Linnaeus, phenotype bresia. Recent, unlocated. X 1. External and lateral view
of a double valved specimen. Attachment is in this case by cementation of a large area of the left
valve (reflected in the large unornamented area on the right); the zigzag commissure is of a moderate
amplitude. UMZC 2003 (Saul Collection).

Fig. 7. Pycnodonte hyotis (Linnaeus), Recent, unlocated. X 2. An extreme example of development of
funnel spines at the crests and troughs of a strong marginal zigzag. UMZC 2001 (McAndrew
Collection).

Palaeontology, Vol. 11

PLATE 89

CARTER, Functional morphology of Arctostrea

CARTER: FUNCTIONAL STUDIES ON CRETACEOUS OYSTER ARCTOSTREA 479

the dorsal end upstream, then an advantageous eddy effect could well be created around
the spines on the ventral periphery.

Many other species of oyster possess funnel spines of greater or lesser significance.
Pycnodonte hyotis (Linnaeus) is one species that has beautifully moulded funnel spines
placed in the crests and troughs of its marginal zigzag (PI. 89, fig. 7). Even more striking
is the juvenile stage of Crcissostrea echinata (Quoy and Gaimard) (PI. 88, fig. 2). At a very
early stage after spatfall this species has secreted a most complex set of funnel spines.
In the light of its favoured habitat of mangrove swamps, it would seem certain that the
epithelium lining these spines directly aids the circulation of uncontaminated water
and indirectly helps to provide an adequate oxygen supply to the animal. At a slightly
later growth stage, the spines become less obviously funnel-shaped, but project as long
tubular structures reaching well away from the shell edge. The animal is presumably
at this stage even more concerned with gathering an inhalent current stream from as far
away from its mangrove attachment site as possible (the roots to which these animals
are attached are generally covered with an evil-smelling slime). There are no spines at all
in the adult growth stage, perhaps because the left valve of the shell is by then large
enough to ensure that the mantle edge is removed from immediate proximity to the
substrate.

RECENT OYSTERS: TURBID WATER ADAPTATIONS

Students of Recent oysters currently recognize two major groupings into the genera
Ostrea and Crcissostrea. These genera are distinguishable on several counts, but the
major structural difference is the presence in Crcissostrea of an exhalent passage dorsal
to the posterior adductor in the right valve; this passage is termed the promyal chamber ,
and ever since its discovery by Nelson (1938) it has been envisaged as an adaptation
enabling Crcissostrea to inhabit turbid waters. This hypothesis agrees well with the
known preference of Ostrea for offshore clear water, whilst Crassostrea favours a
muddy estuarine habitat. Such an ecological distribution might also be suggested to be
linked to the other major difference between the two genera: Crassostrea is oviparous,
and Ostrea is larviparous.

In a lucid comparison of local American species of Ostrea and Crcissostrea , Menzel
(1955) listed several adaptations, additional to the presence of a promyal chamber,
that he considered to increase the efficiency of life in turbid waters; most of these
adaptations were connected with the development of better cleansing mechanisms.

It is especially noteworthy that Crassostrea has more frequent snapping of the valves
to eject pseudo-faeces, regardless of the turbidity of the water that it is placed in, and
this is linked to a tendency to form and expel much larger quantities of mucus and
pseudo-faeces than does Ostrea. Two further adaptations come into play in really
turbid water: first, the tentacles of the mantle edge are intermeshed, and thus form a
partial guard against excessive amounts of sediment entering the mantle cavity; and
second, there is the formation of what Nelson (1938) has termed ‘pseudo-siphons’ in
the presence of very heavy suspensions. These pseudo-siphons are formed by contrac-
tion of the mantle edges of the right valve to form a groove along which entering particles
are then funnelled. In Crassostrea there is one inhalent pseudo-siphon, and two exhalent,
the second corresponding to the presence of the promyal chamber.

480

PALAEONTOLOGY, VOLUME 11

It should further be noted that of all the adaptations serving for greater efficiency in
turbid waters, only the promyal chamber involves a structural adaptation; all the other
mechanisms, whose efficacy can scarcely be doubted, involve only behavioural adapta-
tion. There is, therefore, no reason to suppose such adaptations to be limited to Cras-
sostrea, and, indeed, Menzel makes it clear that Recent Ostrea also possesses most of
these behavioural adaptations, but to a very much lesser degree than does Crassostrea.

Because it involves a structural adaptation, the presence of a promyal chamber can
be inferred in fossil species, and, as pointed out earlier, there can be no doubt that
Arctostreo did not possess a promyal chamber. However, it is virtually certain that the
environment favoured by Arctostrea colubrina ricordeana was one that could become
locally extremely turbid should something disturb the bottom sediment. It would there-
fore seem inherently highly likely that ricordeana (and probably many other species of
Arctostrea ) possessed behavioural adaptations similar to those outlined above for Recent
oysters. These would include tentaculate mantle edges, the frequent ejection of pseudo-
faeces, and the formation of pseudo-siphons. It is likely that in spite of the large flow
of water through the whole shell, brought about by the large gill and branchiform shape,
there was a concentrated exhalent flow just posterior to the adductor muscle where the
anus debouches its load. The presence of a double plica in the shell edge of occasional
specimens of Arctostrea (e.g. PI. 88, fig. 3) at this point is strong evidence for the
presence of persistent pseudo-siphons here.

One further facet of the habits of Recent oysters is relevant to discussion of Arctostrea.
Oysters are in constant danger of being buried under their own pseudo-faeces, together
with encroaching sediment. Nelson (1938) observed that ‘through. . . movements of the
valves, together with adaptive adjustments of the mantle borders, the oyster on soft
bottom clears an open zone about itself, and wards off encroaching mud’; and Lund
(1957) has shown that even at relatively low turbidities (70-95 per cent, light transmis-
sion), oysters on the bottom of an aquarium which has no through currents will com-
pletely bury themselves in 36 days.

EXPLANATION OF PLATE 90

Fig. 1. ‘ Lopha' semiplana (J. de C. Sowerby), Upper Chalk, Hartford Bridge, Norwich. X IT Left
valve of a double valved specimen that was attached in life to a straight cylindrical object. Note the
few, spaced, strong ribs. SM B.6803.

Fig. 2. ‘ Lopha' semiplana (J. de C. Sowerby), same locality as 1. X IT Internal view of a right valve
displaying well the gill attachment scar ventral to the adductor; note also that the dorsal margin of
the adductor scar is convex dorsally. SM B.6796.

Figs. 3 and 4. Arctostrea pusilla (Nilsson), Upper Chalk, Norwich. X IT Right and left valves of a
well-preserved specimen probably attached in life to a belemnite. SM B.1122; figured by Woods,
plate 47, fig. 11.

Fig. 5. Arctostrea pusilla (Nilsson), Upper Chalk, Fareham, Hants. xlT Internal view of a right valve.
Note the lack of any gill attachment scar, and the fact that the adductor scar has a dorsal border
concave dorsally. SM B. 66558.

Fig. 6. Arctostrea pusilla (Nilsson), same specimen as figs. 3, 4. X24. Lateral view. Note the open
crested plicae, and the dorsal diminution of amplitude of the zigzag.

Figs. 7 and 8. Arctostrea pusilla (Nilsson), Cambridge Greensand. X 2. Specimen attached to an am-
monite shell, and hence reflecting the shape of an ammonite in its own shell form. BM LL26835.

Fig. 9. Arctostrea pusilla (Nilsson), grey Chalk Marl, Folkestone, Kent. X 2\. BM LL26830.

Fig. 10. Arctostrea pusilla (Nilsson), Cambridge Greensand. x2L. Specimen attached to an ammonite
shell. BM LL26837.

Palaeontology, Vol. 11

PLATE 90

CARTER, Functional morphology of Arctostrea

CARTER: FUNCTIONAL STUDIES ON CRETACEOUS OYSTER ARCTOSTREA 481

It is obvious that the narrow, high shell of Arctostrea is ideally shaped to avoid
problems of this sort. Though one would hesitate to assign any primary functional
significance to this aspect of over-all shell shape, it must certainly have been a beneficial
side-effect of the over-all morphology.

THE ITERATIVE EVOLUTION OF ZIGZAG VALVE EDGES

It is easy to understand how ‘pre-adapted’ the ostreiiform mode of life is for the
iterative evolution of zigzag valve structures. As oysters usually settle on flatfish sur-
faces, and as initial shell form is entirely dictated by the
shape of the attachment surface, the gills of a juvenile
oyster of any kind must of necessity be a thin, high W.

But should there be at any subsequent stage a tendency
for shell secretion to continue whilst mantle expansion
is inhibited, i.e. any tendency for the formation of a
vertical zone, then the gill of necessity (because it is
attached to the valve interior laterally) becomes pheno-
typically a broader, less acutely angled W. Now the
attachment surface, of course, is rarely quite flat and
may in fact be possessed of considerable relief; this
relief, again inevitably, is reproduced in the growing edge
of the vertical zone, and we thus have a ready-made
undulose commissure.

These two phenomena, lateral widening of the gill-W
and the inception of an undulose commissure, may both
be predicted to occur phenotypically in a proportion of
any population of oysters; the first is an inevitable result
of building a vertical zone, and the second an additional
result of this plus the irregularity of most attachment
surfaces. On selection, there will be a tendency for muta-
tions in a similar direction to accrue in the gene pool of
a population in which it was advantageous to possess a
zigzag valve edge and a widened gill. Thus an initially
phenotypic undulose commissure might be rapidly trans-
formed into a genetically stable and functionally more
efficient zigzag commissure.

There is obviously a limit to the amount of lateral widening that the gill can accept
without detrimental effects on its basic physiology. It would seem that Arctostrea has
exploited this structural plasticity almost to that limit; internally the genus secretes a
series of calcareous plates (or septa, text-fig. 11) which serve to keep the floor of the
valve (and therefore the lateral edges of the gills) progressively in line with the crests
and troughs of the zigzag (PL 85, fig. 2). In other words, the depth of the body cavity
is adjusted so as to coincide with the amplitude of the peripheral zigzags, thus ensuring
that the most lateral parts of the gills are kept in a zone of continued strong inhalent
currents.

text-fig. 1 1 . Arctostrea colubrina
(Lamarck), Chalk Marl, near
Cambridge, xf. A cross-section
through the closed shell some
distance distal to the adductor;
view looking distally. Note the
deep body cavity (infilled with
sediment, stippled) and the pre-
sence of a series of calcareous
plates, or ‘septa’ (all calcareous
shell material in solid black),
separated by cavities. SM B.6594.

482

PALAEONTOLOGY, VOLUME 11

CONCLUSION

The interpretation of the morphology of Arctostrea as dominantly due to increased
efficiency of feeding and cleansing habits has the major advantage that it enables the
morphology to be understood as an integrated whole. The animal can be viewed as ‘all
adaptation’, and not merely as a bundle of discrete adaptations. The shape and detailed
morphology of Arctostrea are the result of selection for an ever larger and more efficient
set of gills. The subsidiary, but important, adaptation found in A. colubrina ricordeana
was initially a phenotypic response to a specific and difficult environmental regime;
though it is not possible to demonstrate rigorously, it would seem likely that the spines
characteristic of this subspecies became incorporated in the genotype before its early
extinction.

Acknowledgements. I should like to thank Professor H. B. Whittington for making available the facili-
ties of the Sedgwick Museum, Cambridge; and Dr. J. D. Taylor, Dr. N. J. Morris, Mr. A. G. Brighton,
and Dr. A. Bidder who kindly gave access to collections in their care. I am indebted to Drs. R. Cowen,
K. A. Joysey, and M. J. S. Rudwick for their critical and extremely constructive reading of the manu-
script of this paper. The work was carried out during the tenure of a Commonwealth Scholarship,
awarded by the British Council.

APPENDIX: TAXONOMIC NOTES ON ARCTOSTREA

It is customary in England to use the generic name Lopha for oysters throughout the Jurassic and
Cretaceous that possess a plicated valve edge of any type. This usage obscures the fact that there
are, in fact, several morphologically discrete groups of oysters which happen to possess in common
the character of a more or less plicate valve edge. Indeed, if anything is commoner in the Ostreiidae
than the iterative evolution of gryphaeate shape, it must be the repeated evolution of plicate valve
edges.

The type species of Lopha Bolten 1798 (and also of Alectryonia Fischer 1807) is Ostrea crist agalli
(Linnaeus), a Recent species from the Indo-Pacific with very coarsely plicated commissure, and a
generally rounded valve outline. Thomson (1954) has placed this species in synonomy with Ostrea
folium Linnaeus, which species is thought to contain three phenotypic variants formerly considered as
discrete species — cristagalli Linnaeus (PI. 89, figs. 3, A), folium Linnaeus (PI. 89, figs. 1, 2), and bresia
Iredale (PI. 89, figs. 5,6). Even allowing for the notoriety that oysters have achieved as variable animals,
there is still sufficient morphological range here to tempt even the most conservative to establish two
genera for its accommodation. Yet Thomson, after a close study of the animals as well as shells, is
quite specific that it ‘is inescapable that this is one species which responds variably according to the
substrate and to degree of immersion’ (p. 148).

Oysters with plicated valve edges are common in the Chalk, and fall into three quite distinct morpho-
logical groups. Only one of these groups — that of Lopha semiplana { J. de C. Sowerby)(Pl. 90, figs. 1, 2) —
is at all reasonably placed in Lopha s.s. ; the other two groups are best placed in the genus Arctostrea
Pervinquiere 1910. They are Arctostrea colubrina ricordeana d'Orbigny and A. pusilla (Nilsson).

‘ Lopha ’ semiplana (J. de C. Sowerby)

In his detailed monograph on Cretaceous Bivalvia, Woods (1899-1913) carefully discussed the forms
of Lopha known in the Upper Chalk horizons and concluded that there was only one valid common
species, that of L. semiplana. That this species was very variable was both explicitly noted and implied
by the vast synonomy that was quoted. Examination of large numbers of Chalk oysters has convinced
me that there are two quite distinct forms presently included in L. semiplana (Sowerby) sensu Woods,
and this is confirmed by an examination of type material held in the Norwich Museum and British
Museum (Natural History). Whether these two forms represent genetically different species, or whether
they are merely phenotypic variants of a single species, is a moot point. In the light of Thomson's

CARTER: FUNCTIONAL STUDIES ON CRETACEOUS OYSTER ARCTOSTREA 483

work on the Recent L. folium it is virtually impossible to state categorically that any two oysters
collected from the same horizon are not conspecific, no matter how different their morphology. How-
ever, amongst several hundreds of specimens I have not seen a single one that is not obviously one
form or the other, and this points to their being true genetically isolated species; it is, however, con-
ceivable that they represent phenotypes responding to two quite different environmental niches, and
that the lack of intermediates is due to the lack of intermediate environmental niches.

The two forms are illustrated on Plates 87 and 90. The true L. semiplana has a large irregularly
rounded shell (PI. 90, figs. 1, 2) and is usually attached either to a large flattish surface, or to other
specimens of its own kind, or to thin branching objects. It possesses a few (commonly less than 12)
broad undulations around the free margins of the commissure which give rise to rounded spaced ribs
on the shell surface during shell growth. There is conspicuous crenulation of the margins on either side
of the ligamental area, often spreading some distance distally around the commissure; often a well-
marked and characteristic gill attachment scar is visible just antero-ventral to the adductor muscle
(PI. 90, fig. 2). The adductor muscle scar itself is large, and has a dorsal margin that is usually convex
dorsally (PI. 90, fig. 2); in rare specimens it may be straight.

Of the two specimens originally figured by Sowerby (1812-46, pi. 489) only one, the larger and left
hand of the two specimens, is still known. It is here designated lectotype, and is in the collections of
the British Museum (Natural History).

The other form has a small lozenge-shaped shell generally reflecting, as Woods pointed out, the
preferential attachment to a Be/emnite/la , crinoid stem or similar object. When the attachment surface
allows, this species develops the branchiform shape so characteristic of Arctostrea (PI. 90, figs. 3, 4). It
possessses a plicated commissure, but the plicae are sharp, numerous (rarely if ever less than 17;
commonly over 40), and of the type described as open-textured earlier in this paper (cf. PI. 86, fig. 6;
PI. 90, figs. 6, 9). The ligament area is small, and without the very marked central pit of semiplana s.s.,
and though crenulations on either side of the ligament area may be developed, they are generally
somewhat inconspicuous. The adductor scar is of moderate size and has a dorsal border that is markedly
concave dorsally (PI. 90, fig. 5).

The earliest British name available for this form is Ostrea alaeformis S. Woodward 1 833. Of the three
specimens originally figured by Woodward, only one, that of his Plate 90, fig. 3, is still extant. It is here
re-figured (PI. 87, figs. 1-3) and designated lectotype. However, there is an earlier named species of this
type from the Upper Cretaceous of Scandinavia — Ostrea pusilla Nilsson 1 827. Unfortunately the type of
pusilla is missing from the collection of Nilsson’s types at the University Geology Museum, Lund, but
there are many specimens of a small Arctostrea conspecific with alaeformis in Swedish Upper Cre-
taceous (‘mammillatuskrita’) faunas. Following Woods’s monograph, these specimens are usually
labelled as semiplana Sowerby, but occasionally an early label reveals that this is the form called
pusilla Nilsson by Scandinavian workers of the pre-Woods era. On balance, it seems best to adopt the
name pusilla for this species and to place alaeformis in synonomy accordingly.

Though pusilla is common in Britain only in the Middle and Upper Chalk, rare specimens from the
Cambridge Greensand (PI. 90, figs. 7, 8, 10) and the Lower Chalk (PI. 90, fig. 9) demonstrate that it
was present in British seas as early as the Cenomanian. These early occurrences are interesting for
they serve to show that the species was sympatric with its relative Arctostrea colubrina. These early
specimens seem to show a strong preference for attachment to ammonite shells (PI. 90, fig. 10), though
this may be due to the small number of specimens available for study.

The introduction of Arctostrea. This name was first introduced as a subgenus of Alectryonia by Per-
vinquiere 1910c/. The subgenus — based on Ostrea carinata Lamarck from the Cenomanian of Le Mans
— received the immediate approval of Douville.

Alectryonia, as noted earlier, is an absolute synonym of Lopha, the two genera being based on the
same type species; but, anyway, modern European practice is to use Lopha and Arctostrea as discrete
genera, which course is followed in this paper.

Ostrea carinata Lamarck 1806 (PI. 86, fig. 2; PL 88, figs. 4, 5) possesses well the characters that
Pervinquiere and Douville used to distinguish the group they called Arctostrea. Particularly diagnostic
is the branchiform shape, the sharply zigzagged commissure, and the presence of low, rounded,
divaricating ribs on the flattened shell top (PL 86, fig. 2). Unfortunately, the name carinata is pre-
occupied by Ostrea carinata O. C. Schroeter 1802. However, Pervinquiere (191 0A) has already provided

484

PALAEONTOLOGY, VOLUME 11

reasons for treating carinata Lamarck 1806 and colubrina Lamarck 1819 as merely varieties of a single
species, and his excellent plates enable one to agree with this conclusion. Thus, though the genus
Arctostrea must continue to be based on the holotype of O. carinata Lamarck, it is permissible to use
the name colubrina for this group of Cenomanian forms. It may well transpire that a new subspecific
name is necessary for the Le Mans carinata population, but this must await a thorough revision of the
whole genus.

Cox (1962) has adopted the name colubrina Lamarck for the English Cenomanian populations of
Arctostrea. However, there are definite differences between the fossils from Le Mans and the typical
English specimens. Conspicuously the English specimens (PI. 86, fig. 1) have a more open commissural
zigzag of greater wavelength, besides possessing the long ‘snowshoe’ spines described elsewhere in this
paper; for their part, the Le Mans specimens have a zigzag of small wavelength, and the presence of
conspicuous funnel spines (PI. 88, fig. 4) results in the strong development of divaricating ribs on the
flat top of the valve (PI. 86, fig. 2). Accepting, on the authority of Coquand (1869), that the name
Ostrea ricordeana d’Orbigny 1850 refers to a Cenomanian Arctostrea possessing long ‘snowshoe’
spines, it is convenient to distinguish the English forms as Arctostrea colubrina ricordeana (d’Orbigny).

It remains to give brief mention to other English members of Arctostrea. The common Jurassic
species ‘ Lopha' gregarea (J. de C. Sowerby) is certainly an Arctostrea, and ranges from the Bathonian
up to the Kimmeridgian. The genus is not found again in any numbers until the Aptian, where it is
a common constituent of the faunas of the Lower Greensand on the Isle of Wight, at Farringdon, and
Upware. Woods (1899-1913) considered that these forms were conspecific with the chalk populations,
but there are significant differences that warrant the resurrection of the name macroptera J. de C.
Sowerby. The chalk populations themselves are concentrated at two localities — those of Haslingfield
and Folkestone. Specimens of Arctostrea are known from other scattered quarries in the Lower Chalk
(e.g. Norman Cement Works, Cambridge; Cherry Hinton quarry, Cambridge), but are very rare.

REFERENCES

atkins, d. 1938. On the ciliary mechanisms and interrelationships of lamellibranchs. Part 7. Latero-
frontal cilia of the gill filaments and their phylogenetic value. Quart. J. Micr. Sci. 80, 346-436.

carter, r. m. 1967. On Lison’s model of bivalve shell form, and its biological interpretation. Proc.
malacol. Soc. Lond. 37, 265-78.

cox, L. R. 1962. British Mesozoic Fossils. London (British Museum, Natural History).

coquand, h. 1 869. Monographic du genre Ostrea. Paris.

douville, h. 1910. Observations sur les Ostreides, origine et classification. Bull. Soc. geol. France (4),
10, 634-46.

lang, w. d. 1923. Evolution: a resultant. Proc. Geol. Assoc. 34, 7-20.

lund, e. j. 1957. Self-silting by the Oyster and its significance for Sedimentation Geology. Publ. Inst,
mar. Sci. (Texas), 4, 320-7.

menzel, r. w. 1955. Some phases of the biology of Ostrea equestris Say and a comparison with
Crassostrea virginica (Gmelin). Ibid. 4, 69-153.

Muir- wood, h., and cooper, g. a. 1960. Morphology, classification and life habits of the Productoidea
(Brachiopoda). Geol. Soc. America, memoir, 81.

nelson, t. c. 1938. The feeding mechanism of the oyster. 1. On the pallium and branchial chambers
of Ostrea virginica, O. edulis and O. angulata. J. Morph. 63, 1-61.

pervinquiere, l. 19 10n. Seance de la Soc. geol. de France, 20 June 1910, 645-6.

19106. Palaeontologia Universalis, fasc. 2, ser. 3.

rudwick, m. j. s. 1964. The function of zigzag deflexions in the commissures of fossil brachiopods.
Palaeontology, 7, 135-71.

sowerby, j. 1812-46. Mineral Conchology. London, vols. 1-7.

Thomson, j. m. 1954. The Genera of Oysters and the Australian Species. Austr. J. Mar. Freshw. Res. 5,
132-68.

woods, h. 1899-1913. A monograph of the Cretaceous Lamellibranchia of England. 2 vols. Palaeont.
Soc.

woodward, s. 1833. An Outline of the Geology of Norfolk. Norwich.

CARTER: FUNCTIONAL STUDIES ON CRETACEOUS OYSTER ARCTOSTREA 485

yonge, c. m. 1926. Structure and physiology of the organs of feeding and digestion in Ostrea edulis.

Jour. war. Biol. Ass. U.K. 14, 295-386.

1960. Oysters. Collins, 209 pp.

zakharov, v. a. 1965. Late Jurassic and early Cretaceous Ostreidae in Arctic seas of Siberia. Int. Geol.
Rev. 7, 1075-83.

R. M. CARTER
Sedgwick Museum
Cambridge
Present address:
Department of Geology
University of Otaga

Typescript received from author 19 October 1967 Dunedin, New Zealand

SHELL STRUCTURE OF THE
BIELINGSELLACEAN BRACHIOPODS

by ALWYN WILLIAMS

Abstract. Sections of Cambrian articulate brachiopods from the U.S.S.R. show that the calcareous shell of
the billingsellacean Nisusiidae probably consisted of normally developed primary and secondary layers with
orthodoxly stacked fibres. Disposition of the recrystallized fibres further suggests that whereas Nisusia was
impunctate, the related Kotujella was permeated by simple canals which must have been indistinguishable from
the caeca accommodated within endopunctae of younger, unrelated articulate brachiopods.

This note is intended as an addendum to the Special Paper on ‘Evolution of the shell
structure of articulate brachiopods’ (Williams 1968) in that it supplements the informa-
tion contained in that paper on the ultrastructure of the primitive articulate shell. When
comparative studies of shell fabric of living and extinct articulate brachiopods were
started, it was hoped to conduct a preliminary survey of the subphylum by examining
representative genera drawn from all currently recognized superfamilies. This aim was
largely achieved although, as stated in the Special Paper (p. 37), the taxa sampled did
not include any billingsellaceans or, for that matter, any Lower Cambrian articulates.
Such omission was particularly tantalizing because it involved Nisusia and related forms
which, as the most primitive articulates yet discovered, are morphologically and strati-
graphically nearest to the ancestral prototype. Suitable material was not lacking for
want of help from many colleagues. Dr. J. Bergstrom of the Palaeontological Institute,
Lund University, Dr. H. Brunton of the British Museum (Natural History), Dr. G. A.
Cooper of the U.S. National Museum, and Dr. H. Mutvei of the Riksmuseum, Stock-
holm had all loaned or given specimens of Billingsella (?) lindstromi Linnarsson and
Oligomys exporrecta Linnarsson from the Middle Cambrian of Sweden, and Eorthis
remnicha Winchell and Billingsella perfecta Cooper from the Upper Cambrian of the
United States. But when sectioned, the shell fabric of all these specimens was found to
have been destroyed by gross recrystallization or by sihcification or dolomitization.

It is therefore a matter of great satisfaction to have recently received from Dr. O. N.
Andreeva, to whom I am indeed greatly indebted, over 80 specimens of articulate
brachiopods from various Cambrian limestone formations of the Siberian Platform, and
to find that four of the five genera represented in the collection are well enough preserved
to indicate the original style of the shell fabric. Variation in preservation is not obviously
related to differences in rock matrix. Specimens of Upper Cambrian Bajanorthis tuko-
landiea Andreeva and Lower Cambrian Nisusia ferganensis Andreeva contained in grey,
coarse, partially dolomitized calcarenites with conspicuous bioclastic constituents up to
1 mm. in size, usually display a discernible fabric, whereas those of Bobinella kulumbensis
Andreeva which occur with the former species are less well preserved. Moreover,
although the shell fabric of the Lower Cambrian Matutella grata Andreeva and
Nisusia cf. kotujensis Andreeva, which have been collected from a pale-green marly
limestone with more finely divided bioclasts, has been destroyed by recrystallization,

[Palaeontology, Vol. 11, Part 3, 1968, pp. 486-90, pis. 91, 92.]

WILLIAMS: SHELL STRUCTURE OF BILLINGSELLACEAN BRACHIOPODS 487

contemporaneous Kotujella calva Andreeva, found in the same matrix, is not greatly
affected. It thus seems likely that processes of recrystallization or replacement of these
shells were as much controlled by local geological conditions as by their texture and
the nature of the rock containing them.

Nearly all the shell of B. tukolandica and B. kulumbensis is made up of recrystallized
fibres (PI. 91, figs. 4-6; PI. 92, fig. 1) which originally must have been identical in
morphology and stacking with those composing the secondary layer of the majority
of articulate brachiopods, including the Orthacea and Porambonitacea to which super-
families the two species respectively belong. Also, where the shell exterior is covered by

text-fig. 1. Tracing of montage showing the relationship between recrystallized secondary fibres in
a transverse section of a pedicle valve of Nisusia ferganensis Andreeva, with the inferred outlines of the
original fibres; the areas occupied by figs. 1 and 2 of Plate 91 are indicated.

rock, a consistently thin layer of finely crystalline calcite, which is quite distinct from the
enclosing granular cement, is seen to form an external skin to the fibres. If this skin
represents the primary layer, as seems likely, secretion of Bajanorthis and BobineUa
shells proceeded in the same way as in living rhynchonellides.

Judging from the ultrastructure of the shell, the mantle of N. ferganensis is also likely
to have functioned in the same way as that of living rhynchonellides. A consistently thin
coat of equidimensional grains, sharply distinguishable from the granular cement of
the rock, forms an external skin to better-preserved valves and has been taken to repre-
sent a recrystallized primary layer. The fibres of the secondary layer, although recrystal-
lized and tending to form aggregates in places, still retain much of the characteristic
shape of orthodoxly stacked units (text-fig. 1; PI. 91, figs. 1-3).

The most unexpected fabric was that found in K. calva. Again recrystallization could
not disguise the fact that the shell almost certainly consisted of normally secreted
primary and secondary layers. Yet the fibres are disposed in arches about 10 /x wide and
30 i x apart which are convex outwards (PI. 92, figs. 2-6). Such deflections could only
have been brought about by secretion of fibres around evaginations of the mantle
(Williams 1968, p. 30). Their frequency and attitude are strongly reminiscent of the
effects of caecal outgrowths from the mantle of endopunctate shells; since they indicate
unbranched canals which appear to penetrate both internal and external shell surfaces,
they probably accommodated caeca or related structures (text-fig. 2).

488

PALAEONTOLOGY, VOLUME 11

In assessing the significance of these Cambrian fabrics to the phylogeny of the articulate
shell, some consideration must be given to the affinities of the stocks concerned. The relation-
ships of both Bajanorthis and Bobinella are incontestible and simply confirm that orthodox
primary and secondary shell was as characteristic of the late Cambrian orthaceans and
porambonitaceans as of Ordovician representatives of the superfamilies (Williams 1968,
pp. 34, 37). It is equally certain that the early Cambrian Nisusia with its rudimentary teeth
and cardinalia, supra-apical foramen and strong pseudodeltidium is one of the most
primitive articulate brachiopods known. These features are also characteristic of three
contemporaneous genera, Eoconcha Cooper, Kotujella Andreeva, and Matutella Cooper.
On the reasonable assumption that all four stocks are closely related, they have been
classified as members of the Nisusiidae which, in turn, has been regarded as the root
stock of the Billingsellacea and, more remotely, of all articulate brachiopods (Williams
in Williams et al. 1965, p. H306). Andreeva (1962, p. 90), however, has drawn attention
to the external resemblance between Matutella and Kotujella and porambonitaceans,
and concluded that these two genera should be classified as pentamerides. No one can
dispute the superficial likeness between the genera just referred to and certain early
Ordovician stocks like Alimbel/a Andreeva. But apart from Alimbella having too many
orthide characteristics to be generally accepted as a pentameride, there are much more
fundamental links between Matutella , Kotujella , and Nisusia, like their possession of
supra-apical foramina, poorly developed teeth, and other primitive internal features.
In particular, one cannot attach as much importance as did Andreeva (1962, p. 94) to
the ‘open’ delthyrium of Kotujella because it was permanently separated from the
supra-apical pedicle passage and foramen by a ‘concave plate’. It may well be that
Matutella and Kotujella stocks were ultimately ancestral to the porambonitaceans, but
that does not diminish the intimacy of their affinities with Nisusia and the Lower
Cambrian Virginian genus Eoconcha (Cooper 1951, p. 5).

In view of the ultrastructure of the shell of Kotujella and Nisusia it now seems

EXPLANATION OF PLATE 91

Electron micrographs of single stage negative replicas — cellulose acetate/carbon : shadowed with gold-
palladium at 1 in 1. Linear scale, at the bottom left-hand corner of each figure, equivalent to 2 ju..

Figs. 1-3. Transverse and oblique sections of recrystallized fibres of the secondary layer of Nisusia
ferganensis Andreeva; Lower Cambrian Lensky Beds, Madygen region, Ferganskaya Dolina. Figs. 1
and 2 are part of montage illustrated in text-fig. 1.

Figs. 4, 5. Oblique sections of recrystallized fibres of the secondary layer of Bajanorthis tukolandica
Andreeva; Upper Cambrian limestones, R. Kulyumbe, Vostochnaya, Siberia.

Fig. 6. Oblique sections of recrystallized fibres of the secondary layer of Bobinella kulumbensis
Andreeva; Upper Cambrian limestones, R. Kulyumbe, Vostochnaya, Siberia.

EXPLANATION OF PLATE 92

Electron micrographs of single stage negative replicas — cellulose acetate/carbon: shadowed with gold-
palladium at 1 in 1. Linear scale at the bottom left-hand corner of each figure, equivalent to 2 /x.

Fig. 1 . Oblique section of recrystallized fibres of the secondary layer of Bobinella kulumbensis Andreeva ;
Upper Cambrian limestones, R. Kulyumbe, Vostochnaya, Siberia.

Figs. 2-6. Oblique to transverse sections of recrystallized fibres of the secondary layer of Kotujella
calva Andreeva; Lower Cambrian Lensky Beds, R. Dakhoi, Vostochnaya, Siberia. Figs. 2 and 4
are part of montage showing vertical range of punctum; figs. 5 and 6 are part of montage illus-
trated in text-fig. 2.

Palaeontology, Vol. 11

PLATE 91

WILLIAMS, Shell structure of billingsellacean brachiopods

Palaeontology, Vol. 11 PLATE 92

WILLIAMS, Shell structure of billingsellacean brachiopods

WILLIAMS: SHELL STRUCTURE OF BILLINGSELLACEAN BRACHIOPODS 489

certain that the nisusiid exoskeleton was secreted as a triple-layered shell composed
of an outermost periostracum, a finely crystalline primary layer, and a secondary
layer made up of orthodoxly stacked fibres. This being so, the programme of secretion

text-fig. 2. Tracing of montage showing the disposition of recrystallized secondary fibres in a trans-
verse section of a pedicle valve of Kotujella calva Andreeva in relation to two adjacent canals inferred
to have accommodated caeca; the areas occupied by figs. 5 and 6 of Plate 92 are indicated as

1 and 2 respectively.

followed by the outer epithelial cells of the mantle of archetypal articulates was the
same, even in detail, as that determining the growth of the shell in living rhyncho-
nellides and terebratulides. It is also intriguing to discover that whereas Nisusia was
impunctate, Kotujella bears all the evidence of having been endopunctate, thereby
providing yet another example of the independent development of this condition
during brachiopod evolution.

490

PALAEONTOLOGY, VOLUME 11

REFERENCES

andreeva, o. n. 1962. Some Cambrian brachiopods of Siberia and Northern Asia. Paleont. Zhurn. 2,
87-96.

cooper, g. A. 1951. New brachiopods from the Lower Cambrian of Virginia. J. Wash. Acad. Sci. 41,
4-8.

williams, A. et at. 1965. Treatise on Invertebrate Paleontology (ed. moore, r. c.), Part H, Bracltiopoda,
1, H1-H521; 2, H523-H927. Lawrence (Univ. of Kansas).

1968. Evolution of the shell structure of articulate brachiopods. Spec. Pap. Palaeont. No. 2, 1-55,

pis. 1-24.

ALWYN WILLIAMS
Department of Geology
The Queen’s University

Typescript received 6 November 1967 Belfast, N. Ireland

THE PALAEONTOLOGICAL ASSOCIATION

COUNCIL 1968-9

President

Professor Alwyn Williams, The Queen’s University, Belfast
Vice-President

Dr. W. S. McKerrow, University Museum, Oxford
Treasurer

Dr. C. Downie, Department of Geology, The University, Mappin Street, Sheffield, 1

Membership Treasurer

Dr. A. J. Lloyd, Department of Geology, University College, Gower Street, London, W.C. 1

Secretary

Dr. J. M. Hancock, Department of Geology, King’s College, London, W.C. 2

Assistant Secretary

Dr. W. D. I. Rolfe, Hunterian Museum, The University, Glasgow, W. 2

Editors

Mr. N. F. Hughes, Sedgwick Museum, Cambridge
Dr. Gwyn Thomas, Department of Geology, Imperial College, London, S.W. 7
Dr. I. Strachan, Department of Geology, The University, Birmingham, 15
Professor M. R. House, The University, Kingston upon Hull, Yorkshire
Dr. R. Goldring, Department of Geology, The University, Reading

Other members of Council

Dr. F. M. Broadhurst, The University, Manchester
Mr. M. A. Calver, Geological Institute of Sciences, Leeds
Dr. C. B. Cox, King’s College, London
Mr. D. Curry, Eastbury Grange, Northwood, Middlesex
Dr. Grace Dunlop, Bedford College, London
Dr. G. F. Elliott, 60 Fitzjohn Avenue, Barnet, Herts.

Dr. A. Hallam, University Museum, Oxford
Dr. Julia Hubbard, King’s College, London
Dr. J. D. Hudson, The University, Leicester
Dr. R. P. S. Jefferies, British Museum (Natural History), London
Dr. J. D. Lawson, The University, Glasgow
Dr. A. H. Smout, British Petroleum Company, Sunbury-on-Thames
Professor H. B. Whittington, Sedgwick Museum, Cambridge

Overseas Representatives

Australia : Professor Dorothy Hill, Department of Geology, University of Queensland, Brisbane
Canada : Dr. D. J. McLaren, Institute of Sedimentary and Petroleum Geology, 3303-33rd Street
NW., Calgary, Alberta

India : Professor M. R. Sahni, 98 The Mall, Lucknow (U.P.), India

New Zealand : Dr. C. A. Fleming, New Zealand Geological Survey, P.O. Box 368, Lower Hutt

West Indies and Central America : Mr. John B. Saunders, Geological Laboratory, Texaco Trinidad,
Inc., Pointe-^-Pierre, Trinidad, West Indies

Western U.S.A. : Professor J. Wyatt Durham, Department of Paleontology, University of California,
Berkeley 4, Calif.

Eastern U.SA.: Professor J. W. Wells, Department of Geology, Cornell University, Ithaca, New
York

PALAEONTOLOGY

VOLUME 11 • PART 3

CONTENTS

The feeding mechanisms and affinities of the Triassic brachiopods Thecospira
Zugmayer and Bactrynium Emmrich. By M. j. s. rudwick 329

Upper Cretaceous coccolithophorids from Zululand, South Africa. By
R. N. PIENAAR 361

A new Eocene Cassigerinella from Florida. By w. g. cordey 368

The gastric contents of an ichthyosaur from the Lower Lias of Lyme Regis,
Dorset. By J. E. pollard 376

Mantle canal patterns in Schizophoria (Brachiopoda) from the Lower
Carboniferous of New South Wales. By j. Roberts 389

On ‘ Dendrocrinus' cambriensis Hicks, the earliest known crinoid. By D. E. bates 406
Revision of two Upper Cambrian trilobites. By a. w. a. rushton 410

Probable Angiosperm pollen from the British Barremian to Albian strata. By
ELIZABETH M. KEMP 421

Sindulites, a new genus of the Nummulitidae (Foraminiferida). By F. E. eames 435
A revision of the Carboniferous lycopod genus Eskdalia Kidston. By
B. A. THOMAS 439

A species of compressed lycopod sporophyll from the Upper Coal Measures of
Somerset. By M. c. boulter 445

Functional studies on the Cretaceous oyster Arctostrea. By r. m. carter 458

Shell structure of the Billingsellacean brachiopods. By alwyn williams 486

PRINTED IN GREAT BRITAIN AT THE UNIVERSITY PRESS, OXFORD
BY VIVIAN RIDLER, PRINTER TO THE UNIVERSITY

VOLUME 11 • PART 4

Palaeontolo

NOVEMBER 1968

PUBLISHED BY THE
PALAEONTOLOGICAL ASSOCIATION
LONDON

Price £3

THE PALAEONTOLOGICAL ASSOCIATION

The Association was founded in 1957 to further the study of palaeontology. It holds
meetings and demonstrations, and publishes the quarterly journal Palaeontology.
Membership is open to individuals, institutions, libraries, etc., on payment of the
appropriate annual subscription:

Institute membership . . . . . £7. 0.y. (U.S. $20.00)

Ordinary membership . . . . . £5. Os. (U.S. $13.00)

Student membership . . . . . £3. 0s. (U.S. $8.00)

There is no admission fee. Student members are persons receiving full-time instruction
at educational institutions recognized by the Council; on first applying for member-
ship, they should obtain an application form from the Secretary or the Treasurer. All
subscriptions are due each January, and should be sent to the Membership Treasurer,
Dr. A. J. Lloyd, Department of Geology, University College, Gower Street,
London, W.C.l, England.

Palaeontology is devoted to the publication of papers (preferably illustrated) on all
aspects of palaeontology and stratigraphical palaeontology. Four parts at least are
published each year and are sent free to all members of the Association. Members
who join for 1968 will receive Volume 11, Parts 1 to 5.

All back numbers are still in print and may be ordered from B. H. Blackwell,
Broad Street, Oxford, England, at £3 per part (post free). A complete set, Volumes
1-11, consists of 38 parts and costs £129.

Special Papers in Palaeontology is a series of substantial separate works published
by the Association. The subscription rate is £6 (U.S. $16.00) for Institute Members
and £3 (U.S. $8.00) for Ordinary and Student Members. Subscriptions and orders
by members of the Association should be placed through the Membership Treasurer.

The following Special Papers are available. Members may obtain them at reduced
rates through the Membership Treasurer. Non-members may obtain them from
B. H. Blackwell, Broad Street, Oxford, England, at the prices indicated.

Special Paper Number One (for 1967): Miospores in the Coal Seams of the
Carboniferous of Great Britain, by A. H. V. Smith and M. A. Butterworth.
324 pp., 72 text-figs., 27 collotype plates. Price £8 (U.S. $22.00), post free.
Special Paper Number Two (for 1968): Evolution of the Shell Structure of
Articulate Brachiopods, by Alwyn Williams. 55 pp., 27 text-figs., 24 collotype
plates. Price £5 (U.S. $13.00).

Special Paper Number Three (for 1968): Upper Maestrichtian Radiolaria of
California, by Helen P. Foreman. 83 pp. with 8 collotype plates. Price £3 (U.S.
$8.00).

Typescripts on all aspects of palaeontology and stratigraphical palaeontology are
invited. They should conform in style to those already published in this journal, and
should be sent to Mr. N. F. Hughes, Department of Geology, Sedgwick Museum,
Downing Street, Cambridge, England, who will supply detailed instructions for
authors on request (these are published in Palaeontology, 10, pp. 707-12).

© The Palaeontological Association 1968

THREE NEW TETHYAN D AS YCLAD ACEAE
(CALCAREOUS ALGAE)

by GRAHAM F. ELLIOTT

Abstract. Epimastopora malaysiana sp. nov., from the Malayan Permian, is described from a unique whole
solid specimen: normally remains of any Epimastopora sp. are extremely fragmentary. Also described are
Harlanjohnsonella annulata gen. et sp. nov., and Suppiluliiimaella polyreme gen. et sp. nov., from the Yugoslav
and Turkish Cretaceous respectively, and both considered related to Dissocladella.

1. EPIMASTOPORA

Remains of the genus Epimastopora are common at many levels in the Upper Carboni-
ferous and Permian of southern Europe and North Africa, the Middle East, Malaya
and Japan, and the south-western United States. They are almost always fragmentary,
consisting of small pieces of thin calcareous wall penetrated by numerous close-set
canals or pores, and usually seen in thin-sections of limestones. This fragmentation has
led to much speculation as to the original form of the thallus (Pia 1923, 1937; Wood
1943; Kochansky and Herak 1960, etc.), and has not been without influence on the
synonymy (ref. summary in E. Fliigel 1966).

Thin-section remains of whole thalli are extremely rare, though figured by EL Fliigel
(1963) and Elliott (in press): both of these are of the closely related Pseudoepimastopora
(Endo 1961).

It is therefore of interest to describe a solid, nearly complete, three-dimensional
specimen of Epimastopora. This rarity was found in Malayan Permian material collected
by Dr. D. J. Gobbett (University of Malaya) and sent by him to me, and I am very much
indebted for his permission to describe it.

Family dasyclad aceae Kiitzing 1843 orth. mut Hauck 1884
Tribus cyclocrineae Pia 1920
Genus epimastopora Pia 1923

Epimastopora malaysiana sp. nov.

Plate 93, figs. 3, 4

Diagnosis. Elongate cylindrical dasycladacean test with rounded distal termination:
calcareous walls very thin and fragile, perforated by very numerous close-set fine
cylindrical pores or canals which are themselves subparallel, each with constricted
terminal openings to inner and outer wall-surfaces: horizontal (peripheral) rows of
pores grouped into fascicules of about seven rows each by sinuous horizontal lines of
interruption; these separate but do not space apart the fascicules.

[Palaeontology, Vol. 11, Part 4, 1968, pp. 491-7, pis. 93-5.]

C 5934 K k

492

PALAEONTOLOGY, VOLUME 11

Holotype. The specimen figured in Plate 93, figs. 3, 4 from the Middle Permian, H.S. Lee Mine no. 8,
Kampar, Perak, Malaya (Jones, Gobbett, and Kobayashi 1966, p. 324). Brit. Mus. (Nat. Hist.) Dept.
Palaeont. V53442.

Other material. One fragmentary specimen, same locality and horizon.

Description. The thallus is hollow-cylindrical in form, with gently curved longitudinal
axis, circular cross-section and a rounded sub-paraboloid distal termination; the
observed length (incomplete) is 35 mm. and the diameter 7-0-7 -5 mm. The thin, very
fragile wall is from 0-26-0-31 mm. thick. It is perforated by very numerous close-set
fine radial branches or canals, set at right-angles to the longitudinal axis. These canals
are straight-sided, the diameter sometimes widening very slightly outwards, but they
are rounded at both ends, constricting to open on inner and outer wall-surfaces as
circular pores of lesser diameter than the main portion of the intra-mural canal. A thin-
section in the very fine-grained powdery dolomite replacing the wall shows the form
of the canals, somewhat indistinctly, and suggests that their diameter is probably
0-078 mm. with separating interstices of 0-039 mm.

The wall has broken away distally to reveal the inner core, also of fine dolomite: this
shows an outer surface-pattern due to the impression of the inner wall-surface, with
detail in reverse. This pattern is of tiny raised circular areas (casts of the inner canal
pores), very uniform in size and close-set spacing, about 220 to a complete horizontal
peripheral ring. Longitudinally (vertically) the horizontal rows are separated into
fascicules of six or seven by sinuous horizontal peripheral lines. These have a charac-
teristic irregular appearance due to the sinuosities being independent as between
consecutive lines. The distance apart of any two successive lines varies from 0-68-0-78
mm., and this distance, accommodating six or seven rows of pores, is compatible with
the inner canal and interstice dimensions inferred from the thin-section.

The outer wall-surface shows a comparable pattern, but the canal terminations show
as minute pore-openings in gentle depressions separated by interpore convexities. On
rubbed or worn portions the pores are much larger, since the abraded surface intercepts
the inner, larger canal-diameters. For this reason, although the pores of the outer
surface correspond each to a pore-cast of the inner core-surface, the outer surface has
a much less uniform appearance. The sinuous horizontal lines occur, but are much
fainter and less obvious. Where a broken edge of the calcareous wall terminates across
a sinuous line on the core, the sinuous line of the outer surface, when visible, may be
seen to correspond with that on the core.

Remarks. This unique specimen shows that in the living Epimastopora calcification was
confined to a very thin peripheral zone. The thick core may have been a large non-
calcified juicy stem-cell, with presumed endospore reproduction, in which case the

EXPLANATION OF PLATE 93

Figs. 1, 2. Harlanjohnsonella annulate gen. et sp. nov. Upper Cretaceous; Fetrebovo, Zlatibor, Serbia,
Yugoslavia. 1. Near-transverse thin-section; 2. Tangential thin-section. Both x30; syntypes.
B.M. (N.H.), V53441 .

Figs. 3, 4. Epimastopora ma/avsiana sp. nov. Middle Permian; H.S. Lee Mine no. 8, Kampar, Perak,
Malaya. 3. Holotype, complete specimen, x3. 4. Distal portion, enlarged to show detail, x 10.
B.M. (N.H.), V53442.

Palaeontology, Vol. 11

PLATE 93

ELLIOTT, Tethyan Dasycladaceae

G. F. ELLIOTT: TETHYAN D ASYCLADACEAE

493

existing wall-canals described above were minute crowded branches, and their sinuous-
fasciculate arrangement may have been a transitional stage from aspondyl to
euspondyl branch-arrangement. Alternatively, a thinner non-calcified stem-cell of un-
known diameter, could have given rise to euspondyl whorls of non-calcified primaries,
in which case the preserved wall-canals are to be interpreted as crowded secondaries or
tertiaries in sinuous zones corresponding each to an inner whorl. The fascicules as now
seen are to be interpreted not as six or seven adpressed horizontal rows, but as irregular
horizontal zones six or seven branches wide. This interpretation is compatible with
either of the two reconstructions suggested above.

Although traces of matrix were feebly effervescent with weak acid, both core and wall
were completely replaced by fine-grained, friable dolomite. This was difficult to section,
and the details of the wall were not distinct, although as stated, the inferred dimensions
are compatible with external evidence. Since all other Epimastopora spp. are based on
clearly sectioned wall-fragments with measurable canals and little else (see tables in
Johnson 1963), the present specimen shows so much more that it is described as a new
species.

This new evidence on the external form of Epimastopora invites comparison with
Koninckopora, of somewhat similar morphology and also of Upper Palaeozoic age.
This latter genus was described in detail by Wood (1943) from solid specimens and many
thin-sections.

Both genera show a thin marginal calcification only, giving a thin sheath-like calcified
thallus. In Koninckopora the infilled cavities representing the branch-terminations are
proportionally large, very thin-walled, and adpressed to give a polygonal structure in
cross-section. This appearance had led to its earlier interpretation as a coral and as
a bryozoan, and only later was it recognized as algal (Wood 1943). In Epimastopora the
corresponding structures, ‘pores’, are separate filled cavities within the calcified wall,
and although often close-set, are proportionally better spaced and not adpressed. In
Koninckopora the cells show evidence of sporangial structures in their outer parts,
interpreted by Wood from the differential calcification of the infilling This phenomenon
has never been seen by me in very many thin-sections of Epimastopora, nor in the related
Pseudoepimastopora, although in this latter genus the pores have been suggested as
sporangial from the swollen morphology.

Wood (op. cit., p. 209) suggested the close relationship and possible identity of
Koninckopora and Epimastopora. In view of the very limited palaeontological evidence
from the surviving calcified structures, I regard the two as valid separate genera, both
referable to Pia’s ‘tribe’ (or subfamily) Cyclocrineae.

2. HARLANJOHNSONELLA

In material examined by me for Dr. J.-P. Rampnoux (Universite d’Orleans-Tours),
certain thin-sections showed a profusion of dasyclad debris and fragments: most of this
proves to be referable to a new genus. This sample is from transgressive basal Upper
Cretaceous, possibly Cenomanian, in the Zlatibor area of SW. Serbia; associated
microfossils are fragments of a small Clypeina sp. ( Dasycladaceae), gastropod and other
molluscan shell-debris, and rare crustacean fragments. I am much obliged to Dr.
Rampnoux for his kind permission to study and describe the alga.

494

PALAEONTOLOGY, VOLUME 11
Tribus thyrsoporelleae Pia 1927 emend. Pia 1936
Genus harlanjohnsonella gen. nov.

Diagnosis. Weakly calcified thin-walled tubular and annular dasyclad, with successive
verticils showing numerous swollen primaries, the presumed secondaries not being
calcified. Upper Cretaceous; type-species H. annul at a sp. nov.

Harlanjohnsonella annulata sp. nov.

Plate 93, figs. 1, 2; Plate 94

Description. Thin-walled, weakly and irregularly calcified dasyclad, length (incomplete)
up to 6 mm. seen, external diameter up to and commonly 2-0 mm., exceptionally
2-25 mm., full wall-thickness 0-26 mm., giving a d/D ratio of about 74%, but commonly
the calcification is incomplete and hence the stem-cell cavity appears even wider. The
dasyclad is annular in structure; these annuli or rings, about 0T 9-0-25 mm. thick,
readily came apart, and by reason both of their proportions and their weak calcification
they fragmented easily, so that the species is largely represented by debris. Each com-
plete annulus shows about 48-50 primary branch-cavities, somewhat crowded so that
neighbouring branches are sometimes directed at slight angles greater or lesser than
normal from the stem-cell axis. Because of this the alternation in position of branches
in successive verticils may appear irregular in tangential section. The branches are pyri-
form or flask-shaped, communicating with the stem-cell by a narrow pore-neck of
0-013 mm. diameter, and widening rapidly within the wall-thickness to a globular or
ovoid cavity of 0-13-0-14 mm. diameter. These cavities are usually open to the exterior
by reason of the ragged incomplete calcification of the outer annulus-surface, though
exceptional preservation seems to suggest that the swollen cavities divided externally
into thinner secondaries.

Syntypes. The specimens figured in Plate 93, figs. 1, 2, Plate 94, fig. 1 from the transgressive basal
Upper Cretaceous, possibly Cenomanian, of Fetrebovo, Dlaglica, SE. of the Zlatibor massif, SW.
Serbia, Yugoslavia. Brit. Mus. (Nat. Hist.) Dept. Palaeont. V53441, 53443.

Other material. Five thin-sections, rock from same horizon and locality.

Remarks. Harlanjohnsonella , thin-walled, the annuli crowded with simple swollen
branches, and fragmentary, brings to mind especially the Carboniferous Coelosporella ,
the Permian Epimastopora and Pseudoepimastopora, and certain forms referred to the
Cretaceous Neomeris cretacea Steinmann. From the former three genera it differs in its
markedly annular character, the Palaeozoic fragments being random pieces of the thin
tubular or flask-shaped thalli. From such forms as the annular Neomeris cretacea
Delmas and Deloffre non Steinmann, from the Cretaceous (Delmas and Deloffre 1962),

EXPLANATION OF PLATE 94

Figs. 1, 2. Harlanjohnsonella annulata gen. et sp. nov. Upper Cretaceous; Fetrebovo, Zlatibor, Serbia,
Yugoslavia. 1. Syntype, Longitudinal thin-section, X 30. B.M. (N.H.), V53443. 2. Thin section of
matrix with numerous broken and fragmentary Harlanjohnsonella, x 10. B.M. (N.H.), V53444.

Palaeontology, Vol. 11

PLATE 94

ELLIOTT, Tethyan Dasycladaceae

G. F. ELLIOTT: TETHYAN DASYCLAD ACEAE

495

it differs in the apparently more simple branch-structure, Neomeris spp. showing globu-
lar sporangia flanked by thinner sterile branches. A closer comparison can be made with
Dissocladel/a, especially the Palaeocene D. savitriae (Pia 1936). This species shows
annuli of similar size and proportions, each with a similar number of branches, but
more strongly calcified. Within this calcification each branch is seen to communicate
with the stem-cell cavity by a narrow pore and to swell distally to a spherical body:
from this a bunch of four to six thin short secondaries communicates with the outer
surface. DissocladeUa is now known to be represented by species from older Mesozoic
to Palaeocene (Ott 1965); it seems likely that the form now described was, in fact, a
dasyclad of very similar plant morphology to D. savitriae but more weakly calcified, so
that no certain evidence of the original secondary branches remains. Geological age
and the observed nature of the calcification alike fit with this interpretation. Since
there is no direct fossil evidence of the secondary branch-structure the species cannot
correctly be referred to DissocladeUa , all of whose species show calcified remains of
the secondary branching. I have pleasure therefore in referring it to Harlanjohnsonella
gen. nov., in tribute to the distinguished American palaeophycologist Professor J. Harlan
Johnson.

3. SUPPILULIUMAELLA

This new and striking Lower Cretaceous genus occurred in Turkish material reported
on by me for Professor Fuat Baykal of Ankara University, and I am much indebted to
him for permission to describe it.

Genus suppiluliumaella gen. nov.

Diagnosis. Strongly calcified thick-walled tubular dasyclad, showing verticils of long
thin primaries with large conspicuous terminal swellings, each dividing into very short
swollen secondaries. Lower Cretaceous: type-species S', polyreme sp. nov.

Suppiluliumaella polyreme sp. nov.

Plate 95, figs. 1-4

Description. Strongly calcified tubular dasyclad; length 6 mm. (incomplete), external
diameter 1-82 mm., internal diameter 0-62 mm., with consequent d/D ratio of 34%.
Numerous close-set verticils of branches, about five or six verticils per millimetre of
tube-length. The calcification extends inwards adjacent to each primary, so that the
stem-cell cavity in vertical section has a regularly notched outline, showing as annular
bars in vertical section tangential to the stem-cell cavity wall. The primaries joined the
stem-cell on the undersides of these annular bars.

Each verticil is estimated to contain about twenty-four primary branches. These
branches are initially almost straight, about 0-065-0-078 mm. thick, circular in cross-
section and set at an angle of 65-70° from the horizontal, with consequent long oblique
courses in the wall thickness. At about the outer third of wall-thickness they turn to an
angle of 45° from the horizontal, and form conspicuous swellings, of up to 0-18 mm.
diameter. In vertical section the swellings are rounded triangular or shield-shaped, with

496

PALAEONTOLOGY, VOLUME 11

apices inwards: in transverse section they are rounded squares or pentagons. Finally,
each divides into four or five short closely adjacent rounded-cylindrical secondaries of
0-065-0-078 mm. diameter, which extend to the outer surface.

Holotype. The specimen figured in Plate 95, fig. 1 , from the Lower Cretaceous limestones exposed about
16 km. SSE. of Amasya, a Turkish town about 260 km. ENE. of Ankara. Brit. Mus. (Nat. Hist.),
Dept. Palaeont. V53445.

Paratypes. The specimens figured in Plate 95, figs. 2-4, from the same thin-section as the holotype.

Remarks. This peculiar and distinctive dasyclad shows the basic branch-pattern or plan
of a thyrsoporellid, and since it divides after the primary swelling into secondaries only,
is closest to Dissocladella. However, the proportions of primary, swelling and secondary,
and the steep angle at which they occur is quite different to that in any known species of
Dissocladella, which have very short horizontally set branches. For this reason the
Turkish species is made the type of a new genus. The generic name commemorates
King Suppiluliuma of the ancient Hittites, whose territory lay in what is now Turkey,
and the specific name refers to the numerous oar-like branches as seen in thin-section.

REFERENCES

delmas, m. and deloffre, r. 1962. Un niveau a algues calcaires au passage Albien-Cenomanien en
Aquitaine. Revue Micropaleont. 5, 214-23, 4 pi.

elliott, o. f. (in press). Fossil calcareous algae, family Dasycladaceae, of the Middle East, Permian to
Palaeocene. Bull. Brit. Mus. (Nat. Hist.) Geol.

endo, r. 1961. Stratigraphical and palaeontological studies of the later Paleozoic calcareous algae in
Japan, XV. A restudy of the genus Epimastopora. Scient. Rep. Saitama Univ., (B), Endo Comm.
Vol. 267-70, pi. 44.

flugel, e. 1966. Algen aus deni Perm der Karnischen Alpen. Carinthia II, 25, 76 pp., 1 1 pi.
flugel, h. 1963. Algen und Probleniatica aus deni Perm Sfid-Anatoliens und Irans. S.B. Akad. Wiss.
Wien, (1), 172, 86-95, 2 pi.

Johnson, j. h. 1963. Pennsylvanian and Permian algae. Colo. Sell. Min. Quart. 58 (3), 211 pp., 81 pi.
jones, c. r., gobbett, d. j., and kobayashi, t. 1966. Summary of fossil record in Malaya and Singapore,
1900-1965, 309-59 in T. Kobayashi, Geology and Palaeontology of Southeast Asia , 2, Tokyo.
kochansky, v. and herak, m. 1960. On the Carboniferous and Permian Dasycladaceae of Jugoslavia.
Geoloski Vjesn. 13, 65-94, pi. 1-9.

ott, e. 1965. Dissocladella cretica, eine neue Kalkalge (Dasycladaceae) aus dem Mesozoikum der
griechischen Inselwelt und ihre phylogenetischen Beziehungen. Neues Jb. Geol. Paldont. Mh. Jhrg.
1965, 683-93.

pia, j. 1923. Einige Ergebnisse neuerer Untersuchungen fiber die Geschichte der Siphoneae verticillatae.
Z. indukt. Abstamm.- u. VererbLehre, 30, 63-98, 1 pi.

1936. Description of the Algae, in L. Rama Rao and J. Pia, Fossil Algae from the uppermost

Cretaceous Beds ( The Niniyur Group) of the Trichinopoly District, S. India. Mem. geol. Surv. India
Palaeont. indica (n.s.) 21, (4), 49 pp., 6 pi.

explanation of plate 95

Figs. 1-4. Suppiluliumaella polyreme gen. et sp. nov. Lower Cretaceous; 16 km. SSE. of Amasya,
Turkey. Thin-sections, x 30. 1. Holotype: longitudinal section. 2. Paratype: outer tangential section
to show swellings and secondary branchlets. 3. Paratype: oblique-transverse cut to show form of
branches. 4. Paratype: inner tangential section to show annular structures around stem-cell cavity.
B.M. (N.H.), V53445.

Palaeontology, Vol. 11

PLATE 95

ELLIOTT, Tethyan Dasycladaceae

G. F. ELLIOTT: TETHYAN D AS YCLAD ACE AE 497

pia, J. 1937. Die wichtigsten Kalkalgen des Jungpalaozoikums und ihre geologische Bedeutung.

C. /•. 2me Congr. Avang. Et. strat. Carb. 2, 765-856, pi. 85-97.
wood, a. 1943. The algal nature of the genus Koninckopora Lee; its occurrence in Canada and western
Europe. Q. Jl geol. Soc. Loud. 98, 205-22, pi. 8-9.

G. F. ELLIOTT

60, Fitzjohn Avenue
Barnet

Typescript received from author 14 December 1967 Herts.

COLOUR MARKINGS IN PHACOPS AND
GREENOPS FROM THE DEVONIAN OF NEW YORK

by GEORGE C. ESKER, III

Abstract. The colour markings in Phacops rana and Greenops boolhi are described and discussed. It is suggested
that P. rana had the ability to change its pigmentation. Only four examples of colour markings in trilobites
have been mentioned in the literature.

Three specimens of more or less complete dorsal exoskeletons of trilobites with colour
markings have been discovered in the Middle Devonian Hamilton shale at a locality
near Alden, New York. Two of these specimens are identified as Phacops rana Green
1832, and the third as Greenops boothi (Green 1837). The specimens exhibit colour
markings in the form of black spots wherever the dorsal exoskeleton is intact and
unweathered. The black spots are presumed to be original rather than diagenetic because
of their linear arrangement on the pleura and not on the axis and glabella of the speci-
men of G. boothi. The irregularly circular black spots vary in size, are randomly arranged
in P. rana , and have a linear arrangement in part in G. boothi. The black spots are
probably melanophores.

The two specimens of P. rana have the black spots scattered over the entire exo-
skeleton. This general pattern appears to be the same as that described by Teichert
(1944) as occurring in a pygidium of Ditomopyge meridionalis from the Permian of
Australia. One of the two specimens of P. rana has spots noticeably smaller in size than
in the other specimen. Many of the spots are very irregular and show black branches
extending out from them. The specimen with larger spots has very short black branches
extending outwards. It is assumed that these two specimens belonged to the same
species possessing the ability to change its pigmentation, rather than that they show
different degrees of fixed pigmentation. This suggests that the specimen with the smaller
spots had the pigment contracted into the central area and the main branches. This
would make the whole surface of this trilobite lighter to match its background. The
specimen with the larger spots has the pigment expanded to occupy nearly all the
branches of the melanophores; this individual would therefore be darker to match a
darker background. Such an ability to change colour to match its background would

EXPLANATION OF PLATE 96

Figs. 1, 3-5. Phacops rana Green. 1, Dorsal view of complete specimen, X 3, Louisiana State University
Geology Museum, Type Collection No. 8269. 3, Lateral view of enrolled specimen, x4, Louisiana
State University Geology Museum, Type Collection No. 8271. 4, Portion of left thoracid pleura,
Xl65, same specimen as 1, contracted pigment. 5, Portion of left pygidial pleura, X 165, same
specimen as 3, expanded pigment.

Fig. 2. Greenops boothi (Green). Dorsal view of complete specimen, x4-5. Louisiana State University
Geology Museum, Type Collection No. 8270.

[Palaeontology, Vol. 11, Part 4, 1968, pp. 498-9, pi. 96]

Palaeontology, Vol. 11

PLATE 96

ESKER, Colour markings in trilobites

G. C. ESKER: COLOUR MARKINGS IN PHACOPS AND GREENOPS 499

clearly be an advantage to any trilobite. Trilobites having this ability were most probably
active, vagrant benthonic forms and not burrowers.

The specimen of G. boothi has the colour markings preserved on most of the axis, on
a portion of the glabella, and on part of the pleura. On the glabella and axis, the spots
are small and entirely random in arrangement. On the pleura, the spots have a secon-
darily symmetrical arrangement due to the two, generally linear, transverse rows per
segment, though the number and the position of the spots varies.

It is probably safe to assume that D. meridional is had the same ability as P. rana to
change its colour to match its background. Perhaps the eyes would be sensitive to colour
changes in the background and served to inform the trilobite of such changes. It is
unlikely that trilobites which were either blind or which had reduced vision had this
type of camouflage ability. Whether or not G. boothi and Phil/ipsial tenuituberculata
as described by Williams (1930) had this same ability, remains open to speculation. Both
of these species show at least a partial arrangement of the black spots in rows.

The other two recorded examples of colour markings in trilobites appear to be of an
entirely different type. Raymond (1922) described a pygidium of Anomocare vittata from
the Cambrian of Alabama with slightly irregular .light and dark bands becoming
narrower across the axis. Wells (1942) reported several examples of Isotelus maximus
from the Ordovician of Ohio, showing deeper shades of the same colour on different
areas of the dorsal exoskeleton. These two types of colour markings are most probably
due to carotenoid pigments, different from the melanin pigment forming the black
spots occurring in the other cases of colour markings. The banding may have served to
conceal the trilobite if it could be matched against an appropriate background.

In conclusion, it seems probable that there were many different types of colour
markings in such a diversified group as the trilobites, as there are in modern Crustacea.
Of special interest is the apparent ability of individuals of Phacops rana to conceal
themselves by expanding or contracting the pigment in their chromatophores to match
a light or a dark background.

Acknowledgements. The author wishes to thank Dr. W. T. Dean of the British Museum (Natural
History) for his criticisms of this paper; also Mr. D. Nugent and Mr. L. Nichols for photographic
work.

REFERENCES

green, J. 1961. A Biology of Crustacea. London.

Raymond, p. e. 1922. A trilobite containing color markings (Cambrian, Cherokee County, Alabama).
Am. Jour. Sci., 5 ser., 4, 461-4.

teichert, c. 1944. Permian trilobites from Western Australia. J. Paleont. 18, 445-65.
wells, j. w. 1942. Supposed color markings in Ordovician trilobites from Ohio. Am. J. Sci., 5 ser.,
240 (10), 710-13.

williams, j. s. 1930. A color pattern on a new Mississippian trilobite. Ibid. 5 ser. 20, 61-4.

GEORGE C. ESKER, III
Geology Department
Louisiana State University
Baton Rouge, La. 70803, U.S.A.

Typescript received from author 5 February 1968

STUDIES ON TRIASSIC FOSSIL PLANTS
FROM ARGENTINA. IV.

THE LEAF GENUS DICROIDIUM AND ITS
POSSIBLE RELATION TO RHEXOXYLON STEMS

by SERGIO ARCHANGELSKY

Abstract. The cuticle of four species of the genus Dicroidium is described. The material has been found in
close association with Rhexoxylon stems in the locality of Ischigualasto, San Juan Province, Argentina. Two
new combinations are proposed: Dicroidium zuberi (Szajnocha) and Dicroidium elongation (Carruthers). It is
proposed that Dicroidium leaves are related to Rhexoxylon stems (a) because they are associated in four Triassic
localities of South Africa and Argentina, and ( b ) because there is a structural relation of both organs with the
Palaeozoic pteridosperms.

This paper reports the very striking association of Rhexoxylon petrifactions with
Dicroidium compressions and impressions in two localities in Argentina, and (though
with less clear information) in South Africa.

The Ischigualasto Formation (of NW. Argentina) has yielded abundant petrifactions
of Rhexoxylon stems (Archangelsky and Brett 1961, Brett 1968). The stems can be about
10 m. high and about 1 m. in diameter near the base. Most of the petrifactions observed
in the field belong to this genus. In the same beds, pockets of loose sandstone with
many leaves were also found. Some of these pockets include stems covered by compact
masses of cuticles, and the stems are sometimes further surrounded by a thick layer of
gypsum in which similar cuticles are embedded. The leaves are fragments consisting
of cuticles with only a little internal substance. Four lots of material collected at different
times, have been studied. The Archangelsky and Brett collection (1961) and the Romer
collection (1964-5) belong to the La Plata Museum Palaeobotanical Collection (LP-PB),
while the Archangelsky (1958) and Flerbst (1960) material, belongs to the Lillo Institute
Collection of Tucuman University (LIL-PB). In the same formation, and in the same
area, a very rich vertebrate fauna has been found. Some of the skeletons are also covered
by compact masses of cutinized leaves (Romer 1962).

Four species of leaves were found in Ischigualasto, associated with Rhexoxylon ,
and I refer them, on the ground of size and shape of pinnae, and cuticular structure, to
the following species of Dicroidium :

Dicroidium zuberi (Szajnocha) comb. nov.

Dicroidium elongation (Carruthers) comb. nov.

Dicroidium coriaceum (Johnson) Townrow

Dicroidium sp.

Only the first three species may be considered as abundant; the fourth is rare and
difficult to determine precisely.

The second Argentine locality which has yielded Rhexoxylon petrifactions and

[Palaeontology, Vol. 11, Part 4, 1968, pp. 500-12, pis. 97-8].

ARCHANGELSKY : TRIASSIC FOSSIL PLANTS FROM ARGENTINA. IV 501

Dicroidium impressions is Barreal, San Juan Province. This association is here reported
for the first time; the petrifactions are now being studied by Dr. Bonetti (Buenos Aires
Natural History Museum) who has already published the study of the leaf impressions
(1963).

From the literature (Du Toit 1954, Walton 1956) it seems that the Rhexoxylon -
Dicroidium association is also present in the Triassic of South Africa.

On independent grounds, both Rhexoxylon and Dicroidium (together with their known
male and female fructifications, Pteruchus and Umkomasia respectively) may be derived
from Palaeozoic pteridosperms.

DESCRIPTION OF THE MATERIAL

Genus dicroidium Gothan 1912 emend. Townrow 1957

1943 Dicroidiopsis Frenguelli
1943 Diplasiophyllum Frenguelli
1943 Zuberia Frenguelli
1 943 Xylopteris Frenguelli
1957 Hoegia Townrow

Discussion. Townrow (1957) emended the original diagnosis of Dicroidium (Gothan
1912), making a clear definition of its morphological and anatomical characters.
I accept here this definition with the exception of two characters: {a) the cuticle may be
thicker than 3 p and ( b ) the stomata may be imperfectly dicyclic as well as dicyclic. In
the emended diagnosis of Dicroidium , there is no statement of the shape of the pinnules.
As Xylopteris is here included in Dicroidium , pinnules of the genus may be rhomboidal,
elongated or linear, with one or several veins in them.

The thickness of a cuticle is of no generic value; it may even vary in one species.
Also, the stomata may be variable (from mono- to di-cyclic) in one species of a genus;
when the cycles of subsidiary cells are constant, the character may be regarded as
specific rather than generic. Accepting these two emendations, the genus Hoegia
(Townrow 1957) is a synonym of Dicroidium. (The forking of the main rachis of the
frond was not observed by Townrow in his Hoegia specimens, probably because he
dealt with small fragments of pinnae. Large fronds of Dicroidium are commonly found
broken and it is not easy to find the main rachis with its fork.)

Frenguelli (1943, 1944) defined the genus Zuberia on external morphology of pinnae;
there are no obvious reasons, however, to differentiate this genus from Dicroidium , as
the shape of pinnules and venation are identical. Townrow (1957) believes that Zuberia
feistmanteli sensu Frenguelli is in fact Dicroidium. Bonetti (1966) goes further and places
all four species of Frenguelli’s Zuberia in Dicroidium feistmanteli , but I unite them as
a separate species, Dicroidium zuberi.

The genera Dicroidiopsis (Frenguelli 1943) and Diplasiophyllum (Frenguelli 1943)
are also placed in Dicroidium by Townrow and by Bonetti; I agree with these authors.

Xylopteris (Frenguelli 1943) is here considered as belonging also to Dicroidium.
Although not expressed in synonymy both Townrow (1962n) and Bonetti (1966) have
already recorded this opinion. The cuticular structure of X. elongate i for instance, in
particular the stomatal apparatus, is similar to the cuticular structure of Dicroidium

502

PALAEONTOLOGY, VOLUME 11

coriaceum (formerly known as Johnstonia coriacea). The forking of the main rachis in
Xylopteris agrees with all known species of Dicroidium. The shape of the pinnules, which
are linear, is here considered as a character of specific rank.

Dicroidium zuberi (Szajnocha) comb. nov.

Plate 98, figs. 1,2; text -figs. 1 a, Id, e

1888 Cardiopteris zuberi Szajnocha, p. 233, pi. 2, fig. 1.

1943 Zuberia zuberi (Szajnocha) Frenguelli, p. 300.

1944 Zuberia zuberi (Szajnocha) Frenguelli; Frenguelli, p. 9, pi. iv-xi, pi. xii, figs. 1, 2.

1944 Zuberia feistmanteli (Johnston) Frenguelli; Frenguelli, p. 3, pi. i-iii.

1944 Zuberia barrealensis Frenguelli, p. 20, pi. ix, figs. 1, 3 ; pi. xiii.

1944 Zuberia sahnii Frenguelli, p. 23, pi. xii, fig. 3.

1957 Hoegia papillata Townrow, p. 49, pi. ii c, iii b, text-figs. 8 b-d, g, k; 10 c-j; 1 1 a-c.

71957 Hoegia autevsiana Townrow, p. 50, pi. iii c, text-figs. 8 e; 9 l; 10 k; 1 1 d.

Description. This is the most frequent species in several beds of the Ischigualasto
Formation. It is also present in association with all Rhexoxylon stems. The better
fragments show the forking of the main rachis (Frenguelli 1944, pi. iv, pi. vii, 1), but
since the whole leaf is large, small specimens appear bipinnate or simply pinnate. The
size, and especially the shape of the pinnules are the most important characters of the
species; usually the pinnules are rhomboidal or somewhat square, with slightly con-
tracted bases and the distal margin may be entire or irregularly lobed. The pinnules of
one row may overlap or be separated. Elongated pinnules do occur, but they are not
common, their shape being slightly falcate. Venation is also characteristic; one or more
groups of veins arise at the base of the pinnule from the rachis, each vein forking usually
once or twice and curving slightly towards the lateral margins. Commonly the veins
tend to concentrate all in one basiscopic point. Our specimens are in complete agree-
ment in size, shape, and venation with the material described by Frenguelli as Zuberia
zuberi and also with Townrow’s Hoegia papillata. By transparency a distinct border
is seen along the margins of the pinnules, devoid of venation and inner coaly substance.

One cuticle is slightly thicker than the other ; both are very thick (about 8 p). Epidermal
cells on both cuticles are typically isodiametric and polygonal, with straight walls,
about 50-60 /x. Cuticular projections from anticlinal walls present; they are most
irregular in their thickness. Short, round papillae are commonly found on the cell
surface (more frequent on the thicker cuticle). The stomata are indistinctly orientated,
present on both cuticles and evenly distributed, rarely with subsidiary cells in contact.
There are about 70-110 stomata per square millimetre. The stomatal apparatus is
typically monocyclic, circular to oval. The guard cells are sunken in a pit formed by
4-6 subsidiary cells which coalesce around the mouth of the pit, forming usually a
strong rim of cutine. Subsidiary cells normally not papillate (but sometimes they are).

The papillae are sometimes almost absent on a whole cuticle; or they may vary
greatly in concentration over the same cuticle. It seems that this is a variable character.
The papillae are usually small and solid.

Material studied. San Juan Province, Ischigualasto Basin. Bed 1: LP 7449, 7450, 7453; LIL 2768,
2769. Bed 2: LP 7451, 7454; LIL 2770. Bed 3: LP 7439, 7440-3; LIL 2767, 2772. Bed 4: LP 7430-4,

* I § $ f $ *
t, tM i* **

text-fig. 1. Compressions of Dicroidium leaves, xf. a, D. zuberi (Frenguelli) Archangelsky; slide
LP425, fragmentary pinnae, b, D. elongation (Carruthers) comb. nov. (the four central linear pinnae),
and D. coriaceum (Johnston) Townrow; slide LP 427 shows extreme types and transitional forms
between the two species, c, Dicroidium sp.; slide LP 428, fragmentary pinnae.

504

PALAEONTOLOGY, VOLUME 11

7436-8, 7441, 7448; LIL 2765, 2776. Bed 7: LP 7452. Slides LP 417, 424, 425; LIL 504-8. Age:
Middle to Upper Triassic, Ischigualasto Formation, Agua de la Pena Group.

Discussion. Frenguelli (1944) described four species of his previously established genus
Zuberia. Two of the species, Z. sahnii and Z. barrealensis, based on very fragmentary
material. Z. zuberi and Z. feistmanteli, as far as the Argentinian material is concerned,
are very similar and there are no important characters differentiating them. D. feistman-
teli (Johnston) Gothan, is distinguished by the entire margins of its pinnules, less
crowded veins, much thinner cuticle and fewer stomata (25 against 70-100 per mm. in
D. zuberi).

The Ischigualasto specimens, though fragmentary, are identified with Frenguelli's
material (which I have examined in the La Plata Museum Collection) because they
agree in size, shape, and venation of pinnules. Moreover, Frenguelli’s four species inte-
grate in these respects and for this reason I have merged them all in D. zuberi.

The original specimen (type) of Zuberia zuberi is the one that Szajnocha described as
Cardiopteris zuberi (from Argentina) and Frenguelli combined in his genus Zuberia.
This specimen should therefore be the holotype, but 1 do not know where it is located,
and Szajnocha gave no collection reference or numbers in his paper (1888). I select a
neotype, LP 9520, a large leaf with a main bifurcate rachis (figured by Frenguelli
1944, pi. iv).

Our material also coincides in morphology and cuticular structure with that described
as Hoegia papillata by Townrow. As D. zuberi has a forked main rachis of leaves, there
is no reason to keep a different species for pinnate fragments, which are too small to
show the main rachis of the whole leaf.

Some specimens (particularly those found in Bed 2) present a very thick cell border
of the pinnules, resembling in this respect Hoegia antevsiana Townrow. I suggest that
Townrow’s species might be regarded as a mere variation of D. zuberi, because they
coincide in stomatal structure and the size of epidermal cells.

Dicroidium elongation (Carruthers) comb. nov.

Plate 97, figs. 1, 3; Plate 98, fig. 4; text-fig. lb, 2 a

1872 Sphenopteris elongata Carruthers, p. 355, pi. 27, fig. 1.

1943 Xylopteris elongata (Carruthers) Frenguelli, p. 318, text-figs, a, b.

1962 Xylopteris elongata (Carruthers) Frenguelli; Townrow 1962, p. 123.

Description. The Ischigualasto material is comparable in shape and size with the
lectoparatype figured and described by Townrow (19626), having very elongated
segments which sometimes fork dichotomously. The segments are almost parallel-sided
and taper very gradually to a short but rounded apex. The best specimens are fairly

EXPLANATION OF PLATE 97

Figs. 1, 3. Dicroidium elongation (Carruthers) comb. nov. 1, Specimen showing bifurcation of the
main rachis; LP 7456, X 1. 3, Two stomata; LP 416, x 300.

Figs. 2, 5. D. coriaceum (Johnston) Townrow. 2, Fragment of leaf showing bifurcate rachis; LP 7457,
x 1. 5, Two stomata; LP 437, x 300.

Fig. 4. D. sp., a few stomata showing papillae overhanging the mouth of the pit; LP 432, x 300.

Palaeontology, Vol. II

PLATE 97

ARCHANGELSKY, Triassic Dicroidium leaves

text-fig. 2. Drawings of Dicroidium leaves, a, D. elongation (Carruthers) comb. nov. ; epidermis show-
ing three stomata, slide LP 14. b, D. coriaceum (Johnston) Townrow; epidermis showing stomata and
cuticular ridges, slide LP 437. c, D. sp.; epidermis showing stomata and papillae overhanging the
mouth of the pit, slide LP 432. d, D. zuberi (Szajnocha) comb, nov.; outline of pinnules showing vena-
tion, slide LP 425 ; X 1 -5. e, As last, epidermis showing stomata and papillae on cells, slide LP 41 8.

506

PALAEONTOLOGY, VOLUME 11

complete, and they are preserved as impressions; compressions are usually very frag-
mentary, represented by only the single linear segments. Though a midrib is not at first
visible in these segments, a suggestion of one is visible when the fossil (consisting of
little more than cuticles) is cleared as a transparency. Also by transparency a clear
marginal zone may be seen, with no coaly substance in it. Some fragments of pinnae
seem to have wider segments, showing what may be an incipient marginal lobing. Such
fragments show, by transparency, a clear median vein.

The cuticles may be of equal thickness, or one (which Townrow regards as the lower)
is rather thinner. The epidermal cells on both cuticles are alike, usually rectangular or
elongated, with their longitudinal axis parallel to the long axis of the segments. They
form ill-defined rows. Isodiametric epidermal cells also do occur. Long cutin ridges
are present on both cuticles, and they are also longitudinally orientated, although
oblique and transversely orientated stomata have been observed. They are widely
separated and rarely have subsidiary cells in contact. I have counted about 10-15
stomata per square millimetre. Epidermal cells are 100-150 p longx 20-40 /x wide
(and 50-70 p when isodiametric). Their outline is straight, showing small sinuosities
due to different thickness of the anticlinal walls. Papillae are rarely seen. The stomatal
apparatus is oval, typically monocyclic. The guard cells are feebly cutinized and are
sunken in a pit formed by 4-6 unspecialized subsidiary cells. The mouth of the pit
usually has a distinct cutin border, sometimes forming a continuous rim. The poles of
the guard cells are sometimes exposed on the surface. Occasionally, a circular stomatal
apparatus was seen, with strongly sunken guard cells. In the apical parts of the segments,
the epidermal cells tend to be more isodiametric.

Materia! studied. San Juan Province, Ischigualasto Basin. Bed 2: LP 7451; Bed 4: LP 7429, 7438;
LIL 2766. Bed 6: LP 7418-28; LIL 2763, 2764, 2771. Bed 7: LP 7452, 7456, 7463, 7464, 7466. Slides
LP 412-16, 426, 427; LIL 501-3. Age: Middle to Upper Triassic, Ischigualasto Formation, Agua
de la Pena Group.

Discussion. Our material agrees in all respects with the type material of D. elongatum as
described by Townrow (19626), in the shape and size of the segments and in the cuticular
structure.

In Ischigualasto, D. coriaceum and D. elongatum occur together in some beds, while
D. elongatum is found alone in some others (Beds 4 and 7).

In the Barreal region (San Juan Province) three plant beds have been recognized
(Bonetti 1963). D. elongatum does occur together with D. coariaceum in Bed I, but in
Bed II, D. elongatum occurs alone.

In some other Triassic localities in Argentina, Xylopteris species have been described,
but there is no mention of the Johnstonia type of leaves (Neuquen, Santa Cruz, etc.).

All these additional references seem to indicate that D. coriaceum and D. elongatum
are probably two different species, although they have some morphological similarities.

Dicroidium coriaceum (Johnston) Townrow
Plate 97, figs. 2, 5; Plate 98, fig. 3; text-figs. 1 b, 2b

Description. The leaf is very variable. The less-lobed segments (pinnae) are somewhat
similar to the D. elongatum linear segments, although they are wider. The lobing of the

ARCHANGELSKY: TRIASSIC FOSSIL PLANTS FROM ARGENTINA. IV 507

segments (differentiation of the pinnules) may be just incipient, or it may be quite
developed; in such a case, the resulting pinnules may be again similar to unlobed pinnae.
All three degrees of variation have been observed in the abundant material. The size
and the shape of the Australian material described by Townrow (1957) may be matched
in our specimens. The pinnules of the well-developed pinnae are linear and have rounded
apices. The segments are bordered by a rim of cutin, clearly seen by transparency. One
vein penetrates each segment, and it may fork once or twice (veins were seen by trans-
parency).

Both cuticles are of a similar thickness. The epidermal cells are of two kinds : elongated
(more than 100 /x long x 40 /x wide) and isodiametric (50-70 /x in diameter). Low
cuticular papillae, one per cell, are sometimes seen. The stomata are usually longi-
tudinally orientated, sometimes obliquely or transversely. The walls of epidermal cells
are straight, sometimes with irregular cutinization of anticlinal walls. Strong longitudinal
ridges of cutin are also present. The stomata are irregularly disposed, separated and
never with subsidiary cells in contact. There are about 30 stomata per square millimetre.
The stomatal apparatus is typically imperfectly dicyclic, with feebly cutinized guard
cells, little or not at all sunken. There are 4-6 subsidiary cells, the polar usually differen-
tiated from the lateral. Subsidiary cells often having papillae. When the guard cells are
little sunken, the subsidiary cells (usually the lateral) form a rim of cutin at the opening
of the epistomatal chamber.

Material studied. San Juan Province, Ischigualasto Basin. Bed 6: LP 7424, 7425, 7428; L1L 2764, 2771.
Material Romer, LP 7457, 7458, 7460-2, 7465. Slides, LP 426, 427, 436, 437 (Bed 2); LIL 515. Age:
Middle to Upper Triassic, Ischigualasto Formation, Agua de la Pena Group.

Discussion. Our material seems to have more variation in cuticular structure than that
described by Townrow. However, Australian specimens are very fragmentary, while
our materia] presents a wide variety of leaves and different segment types. I believe that
epidermal cells tend, in this species as a rule, to be longitudinally orientated in long
pinnae or pinnules. Townrow ’s material shows only short, not fully developed pinnules,
where the epidermal cells are usually more isodiametric. The papillae may be fairly
common or may be lacking in large sectors of leaves; it is also a very variable
character.

Extreme types of D. elongation and D. coriaceum are quite distinct in shape. Inter-
mediate forms can, however, be confused, and their determination becomes more or
less arbitrary. The cuticles of both species are alike. The elongation of epidermal cells
depends on the shape of segments and pinnules. Similar segments of both species have
similar epidermal cells. Papillae are more abundant in D. coriaceum, but they are also
present, though rarely, in D. elongation. Stomata are similarly orientated in both
species, and their concentration is alike. The structure of the stomatal apparatus is
similar, though in D. elongation guard cells seem to be more sunken. Differentiation
of polar and lateral subsidiary cells is more clear in D. coriaceum.

It is difficult and may even be arbitrary to distinguish these species, for there are
intermediate types. The cuticular structure is of no particular help in this respect. It
seems, however, wise to have the two forms separated into two species. Certainly there
are no firm grounds for the use of separate genera.

l 1

C 5934

508 PALAEONTOLOGY, VOLUME 11

Dicroidium sp.

Plate 97, fig. 4; text -figs, lc, 2c

Discussion. Although there is a lack of certain characters, particularly the gross mor-
phology of the fronds, this species is described here in order to complete the presentation
of all the taxa which are found in association with the petrified Rhexoxylon stems. The
only exception is Neocalamites, which is found with these compressions but which
cannot be related to either Dicroidium or Rhexoxylon.

Description. This species is the least frequent in Ischigualasto. The size and the shape
of pinnules and pinnae coincides with Dicroidium superbum (Shirley), as described by
Townrow (1957). Pinnules are usually longer than wider, and their margins are sinuous
or lobed. One complete pinna shows a narrow acute apex. Usually one decurrent vein
was seen by transparency, running up to near the apex of the pinnule. Forking of veins
was not clearly seen, but it may be possible in larger pinnules. By transparency the
pinnules show a clear border of cutin without coaly substance.

Both cuticles are of an even thickness. Epidermal cells are usually isodiametric, except
on the rachis and on the main vein, where they tend to be rectangular or elongated.
Anticlinal walls are straight. On the surface, single round papillae are sometimes seen ;
one, rarely two, per cell. The stomata are irregularly orientated and distributed, some-
times with subsidiary cells in contact. The stomatal apparatus is usually oval, sometimes
subcircular, imperfectly dicyclic. The guard cells are cutinized and sunken in a pit
formed by 4-7 subsidiary cells, which can be wedge-shaped or of irregular shape.
Subsidiary cells each bearing characteristically one round or slightly elongate papilla.
Papillae usually overhanging mouth of pit, almost closing it. Guard cells slightly thick-
ened near the aperture. Polar and lateral subsidiary cells not differentiated.

A few large pinnules are included in this species. In gross morphology they resemble
D. zuberi , but they differ in their stomatal structure, and in the presence of strongly
developed papillae and unicellular hairs. The cuticle of these pinnules is thick, up to
10 p (one of the two is slightly thinner); the epidermal cells are isodiametric, often
square or rectangular, forming definite longitudinal rows for a variable distance, about
40-50 p in diameter, when elongated up to 60 p long x 20-30 p wide. Epidermal cells
as a rule bear one strong, usually hollow, papilla; towards the margins papillae becoming
high, in some cases resembling unicellular hairs up to 50 p long. Cell walls with a
uniform thick border. Stomata present on both cuticles, up to 60 per square millimetre.
Stomata slightly sunken, surrounded by 4-5 subsidiary cells. Mouth of pit square,
rectangular sometimes oval. Normally, subsidiary cells with strongly developed papillae
overhanging the mouth of the pit, sometimes almost closing it. Epidermal cells on rachis,
elongated and bearing papillae. Longitudinal cutin ridges also present.

EXPLANATION OF PLATE 98

Figs. 1, 2. Dicroidium zuberi (Szajnocha) comb, nov., LP 418. 1, Epidermis showing distribution of
stomata, X 120. 2, Epidermal cells with papillae and two stomata, X 300.

Fig. 3. D. coriaceum (Johnston) Townrow, LP 437; epidermis with thickened bands of cutin, X 120.
Fig. 4. D. elongation (Carruthers) comb, nov., LP 416; epidermis, x 120.

Palaeontology, Vol. 11

PLATE 98

ARCHANGELSKY, Triassic leaf cuticles

ARCHANGELSKY: TRIASSIC FOSSIL PLANTS FROM ARGENTINA. IV 509

There are variations: some pinnules although having stomata with protective papillae,
have no papillae on ordinary epidermal cells. Sometimes, most stomata are devoid of
protective papillae, but after careful search, a few are found to be present.

It is difficult to establish whether these specimens are merely ecological varieties of
D. zuberi or if they do belong to another species.

Material studied. San Juan Province, Ischigualasto Basin. Slides LP 428-35; LIL 509-15 (Beds 1, 3,
4, 6). Age : Middle to Upper Triassic, Ischigualasto Formation, Agua de la Pena Group.

Discussion. The size and the shape of the small pinnae resemble D. superbum (Shirley)
Townrow. However, the cuticle shows some differences, mainly in the stomatal appara-
tus. No papillae overhanging the stomatal pit are known in D. superbum. Such papillae
are normal in Lepidopteris but for other reasons our leaves must be placed in
Dicroidium. The rachis of our specimens is not covered by lumps; the shape of the pin-
nules is not alike (neither is the venation) and the stomatal apparatus is not circular
as a rule but often tends to be oval. On the other hand, similar stomata have been
reported and figured by Townrow for D. odontopteroides (1957, figs. 5d, 6m).

As our material is scanty, I prefer to make no specific assignation until better speci-
mens become available from this and other Triassic localities.

ASSOCIATION OF DICROIDIUM LEAVES AND RHEXOXYLON STEMS
San Juan Province , Ischigualasto Basin.

The material described here was collected from about 10 lenses of plant-bearing
sandstones, over a distance of about 4 km. There are also compressions of Neocalamites
stems. No other compressions have so far been found in these particular beds, although
in other strata of the formation there are different plant elements. There are several
petrifactions in close association with the compressions; most of them belong to
Rhexoxylon. One small petrifaction belongs to a Cycad-like stem, Michelil/oa wait on ii
Arch, and Brett (1963), which is unlikely to have borne Dicroidium, because the leaf
base scar is too small. Finally, one petrified Conifer was recently described as Proto-
juniperoxylon ischigualastensis Bonetti (1966), which is also unlikely to have borne
Dicroidium leaves. The overwhelming abundance of Rhexoxylon as petrifaction, and
Dicroidium as compression, provides strong support of probable relationship.

San Juan Province , Bar real

In the locality Quebrada de la Tinta, La Cortaderita Formation, Bed III. 36 (Bonetti
1963), also of Triassic age, some Rhexoxylon petrified stems have been found. They have
not yet been described, but Dr. Bonetti who collected them, kindly showed the material
to me. It undoubtedly belongs to Rhexoxylon, and possibly to the Ischigualasto
species R. piatnitzkyi. In the same bed, Bonetti reports the following plant impressions:
Dicroidium lancifolium and D. feistmanteli (which in her sense is equivalent to Zuberia
zuberi of Frenguelli) and Pseudoctenis ctenophyl/oides as the most abundant; less
abundantly are found Cladophlebis mendozaensis, Dicroidium narrabeenense, and
Pteruchus dubius. Therefore, there are four of the six species belonging to the Dicroidium

510 PALAEONTOLOGY, VOLUME 1 1

complex, two of them being dominant. Cladophlebis, being a fern, is not related to
Rhexoxylon. Only Pseudoctenis could perhaps be related to the petrified stems, apart
from Dicroidium species.

South Africa

In South Africa the evidence of association is less clear, the facts not being available
to me. Species of Rhexoxylon certainly do occur in the same formations with Dicroidium
leaves. The stratigraphic references as given by Walton (1956) and Du Toit (1954)
indicate that Rhexoxylon is present in several localities; all findings of the stems are
referred to the Triassic Molteno and Red Beds in the Cape Province, and to the Soma-
bula Beds of Rhodesia. Both units are part of the Stormberg Series. In the Red Beds,
only silicified woods of Rhexoxylon and Dadoxylon are known. In the Molteno Beds,
a large number of leaf impressions are known, including several species of Dicroidium.
But in the taphoflora of the Somabula Beds in Rhodesia, apart from Rhexoxylon , three
species of Dicroidium , one of Taeniopteris and one of Schizoneura have been reported.
So, again only two genera can be related to Rhexoxylon, and the more important of
them, having several representatives, is Dicroidium.

Resuming all the facts, Rhexoxylon is always associated with Dicroidium in two locali-
ties of Argentina and in two of South Africa. In Ischigualasto, they are the only taxa
which might be related. In two other localities (Barreal, Argentina and Rhodesia, South
Africa) Rhexoxylon and Dicroidium are almost exclusive elements in association, apart
from Pseudoctenis in one and Taeniopteris in the other. Therefore, all four plant associa-
tions from South Africa and Argentina are consistent with the idea that Rhexoxylon
bore Dicroidium leaves.

Further evidence is needed for making a much better case of this association. We are
not able at present to relate particular species of Dicroidium leaves with species of
Rhexoxylon. Indeed, different species of Rhexoxylon may represent different stages of
development of one stem. An organic connexion is therefore needed.

STRUCTURAL RELATION OF DICROIDIUM AND RHEXOXYLON

On general grounds Rhexoxylon may be related in its structure to some Palaeozoic
pteridosperms (Medullosae; Archangelsky and Brett 1961). On the other hand, all
other organs of Dicroidium (male and female fructifications) are also related to the
Palaeozoic pteridosperms. The leaves are also related to the pteridosperms as they
have a fern-like appearance, with dichotomous main rachis and a thick cuticle with
specialized stomata. It may be added that there is agreement between the leaf base of
Dicroidium and the scars on Rhexoxylon, but only in their similar size and shape, for
we have no supporting evidence of the vascular bundles.

The geological and geographical distribution shows Rhexoxylon and Dicroidium
confined to the Gondwana Palaeofloristic Region, during the Triassic Period. Therefore,
the data so far available are not against considering these two taxa as belonging to the
same plant. The peculiar structure of Rhexoxylon is not in favour of its abundant
representation in the sediments.

ARCHANGELSKY: TRIASSIC FOSSIL PLANTS FROM ARGENTINA. IV 511

CONCLUSION

This paper makes a strong suggestion that the Dicroidium leaf was borne on the
Rhexoxylon stem, but I do not claim that the present evidence constitutes overwhelming
proof. It is merely circumstantial evidence, each item no more than suggestive, but gain-
ing greatly in weight by repetition. 1 publish this idea mainly in the hope that it will
cause others to notice further evidence bearing on this question. On the assumption
that the idea is right, 1 examine some of the taxonomic consequences.

The Corystosperms constitute an important stock of information. Several Mesozoic
groups, such as the Peltaspermaceae and the Caytoniaceae may be linked with them in
varying degree. The Corystosperms are obviously an evolutionary step forward from
the Palaeozoic pteridosperms, and attention should be paid to them when new relation-
ships are to be established with other Mesozoic pteridosperms. Dispersed information
available at present, seems to indicate that the Mesozoic pteridosperms played an
important role not only in a particular palaeogeographical area, but all over the world.
The stems of Hermanophyton from the Jurassic of U.S.A., Pteroma and Pachypteris
(Harris 1964) from the Jurassic of England, and even the Peltaspermaceae, clearly
indicate the world-wide occurrence of plants whose allies are .the Palaeozoic pterido-
sperms.

Finally, I believe that further research work with the Mesozoic pteridosperms is of
utmost importance for Palaeobotany, not only to establish the real taxonomic status
of the different groups, but also to look for further evidence in relation to the origin of
the angiosperms, which are claimed as descendants of the pteridosperms.

Acknowledgements. I express my gratitude to Professor Thomas M. Harris (Reading University) who
encouraged this work, and supplied an important source of information. I am indebted to Dr. Alfred
Romer for his kindness in offering for study the plant material collected during the 1964-1965 La
Plata-Harvard Expedition to Ischigualasto.

REFERENCES

archangelsky, s. and brett, d. w. 1960. Nota preliminar sobre el hallazgo de Rhexoxylon en la
cuenca de Ischigualasto, limite de las provincias San Juan y La Rioja. Acta Geol. Lilloana 3, 187-90.

1961. Studies on Triassic fossil plants from Argentina. I. Rhexoxylon from the Ischigualasto

Formation. Phil. Trans. Roy. Soc. Load. B 244, 1-19, 2 pi.

1963. Idem II. Michelilloa waltonii nov. gen. et sp. from the Ischigualasto Formation. Ann.

Bot. N.s. 27, 147-54, 2 pi.

Arnold, c. A. 1962. A Rhexoxylon- like stem from the Morrison Formation of Utah. Amer. Journ. Bot.
49, 833-6.

bonetti, m. i. r. 1963. Contribucion al conocimiento de la Flora Fosil de Barreal, Departamento
Calingasta (Prov. de San Juan). Ph.D. Thesis, Buenos Aires University.

1 966 a. Consideraciones sobre algunos representantes de la Familia Corystospermaceae. Ameghi-

niana, 4.

19666. Protojuniperoxylon ischigualastensis sp. nov. del Triasico de Ischigualasto (San Juan). Ibid.

211-18.

brett, d. w. 1968. Studies on Triassic fossil plants from Argentina. III. The trunk of Rhexoxylon.
Palaeontology, 11, 236-45, pi. 44.

carruthers, w. 1872, in daintree, r. Notes on the Geology of the Colony of Queensland. Quart. J.
geol. Soc. Loud. 28, 271-360.

512

PALAEONTOLOGY, VOLUME 11

frenguelli, j. 1943. Resen a crltica de los generos atribuidos a la Serie de Thinnfeldia. Revta Mus. La
Plata n.s. 2, Pal. 12, 225-342.

- 1944. Las especies del genero Zuberia en la Argentina. An. Mus. La Plata, Pal. sec B 1, 1-30,
13 pi.

gothan, w. 1912. Uber die gattungen Thinnfeldia Ettingshausen. Abh. naturhist. Ges. Niirnberg 9 (3),
67-80, 4 pi.

Harris, T. M. 1964. The Yorkshire Jurassic Flora. II. Claytoniales, Cycadales and Pteridosperms.
British Museum (Natural History), London.

romer, a. s. 1962. The fossiliferous Triassic deposits of Ischigualasto, Argentina. Breviora, Mus. Comp.
Zoology, Cambridge, Mass. 156, 1-7.

szajnocha, l. 1888. Uber pflanzenreste aus Cacheuta in der Argentinischen Republik. Sber. Akad.

\ Viss. Wien, Math. Naturwiss. Klasse, 97, 1, 219-45, 3 pi.
thomas, h. hamshaw. 1933. On some Pteridospermous plants from the Mesozoic Rocks of South
Africa. Phil. Trans. Roy. Soc. Loud. B 222, 193-265, 2 pi.
toit, a. l. du. 1954. The Geology of South Africa. Oliver and Boyd. Edinburgh-London.
townrow, j. a. 1957. On Dicroidium, probably a Pteridospermous leaf, and other leaves now removed
from this genus. Trans. Geol. Soc. S. Africa, 60, 21-56, 2 pi.

- 1962. On Pteruchus, a microsporophyll of the Corystospermaceae. Bull. Brit. Mus. (Nat. Hist.)
Geology, 6, 289-320, 3 pi.

- 1962. Note on the type material of Xylopteris elongata (Carruthers) Frenguelli. Proc. Roy. Soc.
Q'land, 72, 123-7.

walton, j. 1956. Rhexoxylon and Dadoxylon from the Lower Shire Region of Nyasaland and Portu-
guese East Africa. Colon. Geol. Miner. Resour. 6, 159-68, 1 pi.

SERGIO ARCHANGELSKY
Research Fellow
National Research Council,
Museo de Ciencias Naturales,
La Plata,

Typescript received from author 14 November 1967 Argentina

ASTROCYSTITES DISTANS SP. NOV., AN
EDRIOB LASTOID FROM THE ORDOVICIAN OF
EASTERN AUSTRALIA

by B. D. WEBBY

Abstract. A new Ordovician edrioblastoid, Astrocystites distans sp. nov., and other miscellaneous pelmatozoan
plates are described from the lower part of the Cliefden Caves Limestone WNW. of Mandurama, New South
Wales. The edrioblastoid is similar to the only other species of Astrocystites known, A. ottawaensis Whiteaves
from the Trenton Limestone of Ottawa, which was hitherto the sole representative of the group. This represents
the first record of an edrioblastoid outside North America. Sections of the Australian material have revealed
details of internal structures not previously known in the group.

During recent geological work in the Lower Palaeozoic successions of the central- west
of New South Wales by G. H. Packham and the writer, an abundance of pelmatozoan
material was found in a band of fawn-grey thinly bedded limestones. Unfortunately
virtually all of it is fragmentary, consisting mainly of stem ossicles, incomplete thecae
and disarticulated plates. The thecae and plates form the basis of this paper. The
band of limestones also contains polyzoans, brachiopods, corals, gastropods, trilo-
bites, and nautiloids, and lies within the lower thinly bedded part of the Cliefden
Caves Limestone. The localities from which the collection was made are situated on
the property of Boonderoo, about \ mile ESE. of the entrance to Cliefden Caves
(Bathurst sheet, 1:250,000; S155-8, Edit. 1-AAS, Series R 502; Grid ref. 186848),
and near Little Boonderoo (Grid ref. 186844).

The stratigraphy of the area was described by Stevens (1952), but the faunas have
received little attention. Some of the corals in the Cliefden Caves Limestone were
described by Hill (1957) and suggest a Trenton age. However, the presence of graptolites,
probably of the Nemagraptus gracilis zone, in the overlying shales and limestones of the
Malongulli Formation (Sherrard 1954) implies an older age. The faunas are being
studied by Packham and the writer, and it is hoped to offer a more exhaustive statement
on the age and correlation of units in the succession.

SYSTEMATIC DESCRIPTION

The registration numbers of specimens in the University of Sydney, Geology Department, palaeon-
tological collections have the prefix USGD.

Subphylum pelmatozoa Leuckart 1 848
Class edrioblastoidea Fay 1962
Family steganoblastidae Bather 1900
Genus astrocystites Whiteaves 1897 ( = steganoblastus Whiteaves 1898)

Type species. Astrocystites ottawaensis Whiteaves.

[Palaeontology, Vol. 11, Part 4, 1968, pp. 513-25, pis. 99, 100.]

514

PALAEONTOLOGY, VOLUME 11

Nomenclature and previous work. Whiteaves (1897) first described Astrocystites ottawa-
ensis from the Trenton Limestone of Ottawa. Jn the same year Bather raised objections
to Whiteaves’s generic name on the grounds of possible confusion with Asterocystis
Haeckel, and suggested that Whiteaves substitute the name Steganoblastus. The new generic
name was introduced by Whiteaves (1898), and the family name Steganoblastidae of the
Class Edrioasteroidea added by Bather (1900). Whiteaves’s original generic name, Astro-
cystites, was restored by Bassler (1935) on grounds of priority, and he introduced a new
family name, Astrocystitidae, to replace Steganoblastidae. However, since Stegano-
blastidae, which is based on a junior objective synonym, has priority and has been
a more widely used name, it would seem desirable that it be retained. This view accords
with Art. 40 of the International Code of Zoological Nomenclature.

Astrocystites was interpreted as an edrioasteroid by Bather (19146), and as a blastoid
by Hudson (1925). Bassler (1935, 1936) also classified the genus in the Edrioasteroidea,
but observed that, since only a few specimens are known and pending further discoveries
it might well be assigned to the Protoblastoidea. Fay (1962) contended that it did not
belong either to the edrioasteroids or the blastoids, and raised a new class, the Edrio-
blastoidea, to accommodate it. He noted its resemblances to blastoids in having five
petaloid ambulacra with pores between ambulacral plates, five radials, five deltoids,
five orals, and a primitive stem, and its differences in exhibiting five basals, numerous
infradeltoids, a hydropore, and deep infolds in plates of the theca. His reasons for
removing Astrocystites from the edrioasteroids are not clearly stated. Judging from
Bather’s and Fay’s descriptions of A. ottawaensis it is evident that there are close similari-
ties with edrioasteroids, especially in the nature and arrangement of floor and cover
plates, and pores in the ambulacral areas, and in the presence of a cluster of small plates
around the anal opening and hydropore. The genus may be distinguished from edrio-
asteroids by having a pyriform theca, a stem, five basals, five radials, and five deltoids.
Regnell (1966, p. 150) preferred a non-commital term ‘third aperture’ for the structure
usually referred to as the hydropore in edrioasteroids, because it may alternatively be
interpreted as a gonopore or a hydrogonopore. In the following account of the new
edrioblastoid, the term hydropore has been retained, although it is recognized that the
aperture may also have acted as an outlet for sexual products.

The original descriptions of Astrocystites ottawaensis by Whiteaves (1897) and Bather
(19146) were based on three complete specimens. However, Wilson (1946) and Fay (1962)
indicated that only one specimen is now known to exist, the others having been mis-
placed or lost. Wilson listed the specimen from the Cobourg beds, in Booth Street,
Ottawa, the topmost unit of the Ottawa Formation, and of Trenton age. Fay founded
the Class Edrioblastoidea on this single remaining specimen. Details of the internal
structures remain unknown, yet there seems to be no good reason why this should be so,
for Spencer (1938, p. 291) observed that the Trenton rocks of Ottawa contain beds
‘composed almost entirely of separate ossicles of species of Steganoblastus' ; Bassler
and Moodey (1943, p. 197) listed A. ottawaensis from three different localities and
horizons, namely, from the Upper Trenton Hormotoma trentonensis beds of Ottawa,
from the Hull Fimestone of Kirkfield, Ontario, and from the Curdsville Fimestone of
Mercer County, Kentucky. From these statements, it would, therefore, be surprising
if no other material is available either in existing collections or in outcrops. It is to be
hoped that further material will be found, and sections cut in order to elucidate the

B. D. WEBBY: ASTRO CYSTITES DISTANS SP. NOV.

515

nature of the internal structure of A. ottawaensis, enabling closer comparisons to be
made with the fragmentary Australian material.

The writer has reservations about the need to separate Astrocystites from the Edrio-
asteroidea and to erect a new class, the Edrioblastoidea (Fay 1962). Even greater doubt
is felt about the assignment of edrioasteroids and edrioblastoids to different subphyla
(Fell 1965; Fell and Moore 1966). Although Fay’s classification has been followed, the
obvious morphological similarities of Ec/rioaster and Astrocystites , the appearance of
Astrocystites in the Middle Ordovician during the period of maximum diversification
of the edrioasteroids, and the presence of several forms with comparable sac-like form
and straight ambulacra (e.g., Cystaster and Cyathocystis) still regarded as edrioasteroids
(Regnell 1966), support the view that edrioblastoids should be grouped in an order or
a suborder of the Class Edrioasteroidea.

Astrocystites distorts sp. nov.

Plate 99, figs. 1-10; Plate 100, figs. 1-8; text-figs. 1-4

Diagnosis. A species of Astrocystites with slit-like ambulacral pores, averaging 25-30
per cm., and small, circular hydropore.

Material. Two partially complete specimens (holotype, USGD 2302, and paratype A, USGD 2303),
five incomplete specimens showing sections across the theca (paratypes B-D, USGD 2309; other
paratypes, 2304-6, 2312, 3303), two serially sectioned specimens (paratypes, USGD 2310-11) and
plate fragments (paratypes, USGD 2308, 2313-14). All the type specimens come from the lower
part of the Cliefden Caves Limestone f mile ESE. of the entrance to Cliefden Caves, excepting USGD
3303, which is from the same horizon near Little Boonderoo.

Description. Theca medium-sized, incompletely preserved. Holotype (USGD 2302)
measures more than 18 mm. in height and 20 mm. in width; paratype A (USGD 2303)
is more than 23 mm. high and 25 mm. wide. Base of theca not observed. Basals small,
convex plates with strong radial crenulations and small apical openings (probably
caused by wear on this more prominent part of the plate); basals 3-6 mm. wide and
2-5-5 mm. in height. Radials large, subpentagonal, estimated to be 8-14 mm. wide
and of similar height; lower and lateral parts exhibit gentle to deep crenulations,
seemingly at right angles to sutures, and upper medial part of plate underlies distal end
of each ambulacrum. Small, tightly fused infradeltoids appear to be represented in
posterior interradius around possible anal opening of holotype. These have also been
seen in other interradii in cross-section. Deltoids large, subtriangular, thickened plates,
approximately 5-8 mm. in both height and width, and with deep to undulating crenula-
tions on external surface. Hydropore circular, 0-5 mm. in diameter, situated near apex
of posterior interradius.

Straight petaloid ambulacra; in holotype ambulacrum 15 mm. long and 6 mm. wide,
and in paratype A, more than 22 mm. long and 6-5 mm. wide; ambulacra gently curved
around theca in distal and medial regions, but more strongly arched and thickened in
proximal region leading to mouth; ambulacral groove bounded by row of floor plates
on either side, and united by very close suture; individual floor plates cannot be differen-
tiated, owing to extreme fusion of plates and nature of preservation; regularly spaced,
slit-like pores extend along each row of floor plates, about 25-30 per cm., and become

516

PALAEONTOLOGY, VOLUME 11

smaller and more closely spaced proximally. It is also virtually impossible to recognize
junctions between floor plates and deltoids because of the complete fusion and crystallo-
graphic continuity of calcite in the plates; they can seldom be distinguished even in
thin sections under a polarizing microscope. In the distal part of the ambulacrum pores
pass through floor plates into a canal lying between the radial and the floor plates; it
presumably opens into the thecal cavity at the top of the radial. Floor plates in proximal
part of ambulacrum and margins of deltoids are thickened, producing an inwardly
directed flange-like extension through which pores penetrate and communicate with
the thecal cavity. Flanges seem to form part of the frame surrounding the peristome.

text-fig. 1. Astrocystites distans sp. nov. a, longitudinal section of paratype B, USGD 2309, X 5,
showing large, vertical interradial element and the nature of the pores inside the theca. The circum-oral
ring of the water vascular system may have crossed the element in the lower part, below the level of
adjacent pores, b, oral view of holotype, USGD 2302, X 5, showing right postero-lateral ambulacrum
and posterior interradius with hydropore. At the top of the ambulacrum there is a plate which may be
interpreted as a radial element in the frame, or a foreign plate fragment lying inside the peristome.

Larger vertical interradial elements are found in the angles between floor plates and
adjacent radii, and make up the interradial component of the frame (text-fig. la); the
upper part of the interradial element is smooth and may have been connected in the
frame, but the lower part, just below the level of adjacent pores, is irregular, and may
have been free. Floor plates on either side of the interradial element are apparently
narrower and taper inwards, as associated pores are smaller and slope obliquely in
towards larger vertical canals on either side of the interradial element. Structure inter-
preted as stone canal extends from the interior in an adoral direction between inter-
radial element of the posterior interradius and left posterior ambulacrum. From the
interior a narrow canal expands into a double chamber, which is linked by a narrow
obliquely inclined canal to a funnel-shaped chamber (PI. 100, fig. 3). It opens to the
exterior by the centrally placed hydropore. A radial element is suggested from the oral
region of the holotype (text-fig. 1 b) but it has not been confirmed in other specimens.
Cover plates and oral plates have not been observed. Tiny pits scattered on the external
surface of parts of the theca possibly represent attachment points for small spines.

A series of cross-sections, unfortunately slightly oblique, was prepared from paratype
USGD 2310; the specimen is fragmentary, lacks basals, and a prominent joint inter-

B. D. WEBBY: ASTROCYSTITES DISTANS SP. NOV.

517

sects one side of the theca, resulting in recrystallization and disruption of plates. Radials,
deltoids, infradeltoids, and floor plates are represented (text-fig. 2). Pores, although
mainly cut obliquely, appear to extend through floor plates along the length of the
ambulacra, and canals between radials and floor plates are clearly exhibited in the distal
parts of ambulacra (text-fig. 2 f-h).

text-fig. 2. Astrocystites distcms sp. nov. Slightly oblique cross sections of paratype, USGD2310, x 2.
Camera lucida drawings from cut sections and cellulose peels at intervals of 0-5 (a), 2-5 (b), 6 0 (c),
8-3 (d), 9-9 (e), 12-2 (/), 14-7 (g), 17 0 (//), and 20 0 mm. (/) below the top of the theca. Sections are
not oriented according to convention as the position of the anus cannot be determined, but the radials
( R) have been numbered clockwise from above. Deltoids (D) and infradeltoids (ID) are also indicated.
Floor plates above the radials can be easily differentiated by the suture and the central canal (dark
shaded), but in the upper part of the theca, floor plates fuse with deltoids and cannot be separated.

Similar canals are shown in a section of another fragmentary specimen, paratype
USGD 2311 (text-fig. 3c). However, a somewhat different arrangement of canals can

518

PALAEONTOLOGY, VOLUME 11

be seen in the other sections of this specimen (text-figs. 3, 4). In one of these, the floor
plates resting on radial R3 have two small tear-drop shaped canals instead of a larger
central canal (text-figs. 3a, 4a). Two sutures which extend from the canals to the ambulac-
ral groove seem to parallel cleavages in calcite of the floor plates; they separate the
floor plates into inner and outer parts. In the other, the floor plates lying on the radial

C

text-fig. 3. Astrocystites distans sp. nov. Slightly oblique cross sections of paratype, USGD 2311, x 2.
Camera lucida sketches of thin sections at intervals of 2 (a), 7 ( b ), and 12 mm. (c) below upper surface
of specimen shown in Plate 100, fig. 5. Radials (R) are numbered, and deltoids (D) and infradeltoids

(ID) are also indicated.

R4 have three small canals, two lateral and one medial which lies on the line of suture of
floor plates just inside the ambulacral groove (text-fig. 3b, 4b). This medial canal appears
to be connected with the laterals by sutures or narrow passages, which thus separate the
inner and outer parts of the floor plates, as in the previous example; the inner super-
ficially resembles the lancet plate of blastoids. This is directly underlain by what seems
to be another thin plate, but on closer study under polarised light this proves to be
a part of R4.

EXPLANATION OF PLATE 99

Figs. 1-10. Astrocystites distans sp. nov. 1, 2, Flolotype, USGD 2302, x 3, side and oral views showing
straight, petaloid right postero-lateral ambulacrum with slit-like pores and ambulacral groove, and
posterior interradius with crenulations and hydropore near the apex. 3, Holotype, USGD 2302, X 4,
detail of posterior interradius showing gentle to deep crenulations with vague, small tightly fused
infradeltoids surrounding the anal opening and a small, circular hydropore near the apex. 4, Paratype
A, USGD 2303, X 4, showing straight ambulacrum and radials with gentle crenulations. 5, Paratype
A, USGD 2303, X 8, detail of a row of slit-like pores in the medial part of the ambulacrum. 6, Paratypes
B-D, USGD 2309, X 3, two longitudinal sections (paratype B, top left; paratype C, centre) and basal
plate (paratype D, bottom right) ; B, section showing large vertical interradial element ; C, longitudinal
section intersected nearer to the outer margin of the theca than in paratype B, and therefore only
exhibiting traces of the interradial element; D, fragmentary basal with small apical opening. 7, Para-
type, USGD 2304, X 4, oblique view of a single row of slit-like pores penetrating tightly fused floor
plates with pores on either side and interradial element at the apex. 9, Paratype, USGD 2313, x4,
small, convex basal with radial crenulations and small apical opening. 10, Paratype, USGD 2314,
x 4, convex basal with strong radial, crenulations and small apical opening. Photographs by G. Z.
Fold vary and B. D. Webby.

Palaeontology , Vol. 11

PLATE 99

WEBBY, Ordovician edrioblastoid

B. D. WEBBY: ASTROCYSTITES DISTANS SP. NOV.

519

Remarks. Astrocystites ottawaensis may be distinguished from A. distans by having
pores with a circular cross-section, a wider spacing of pores (15 per cm.), and a slit-like
hydropore. Few other features can be compared closely because only three specimens
of the more perfectly preserved A. ottawaensis have previously been studied and details
of internal structures are unknown. There are also differences of interpretation of
certain structures which must be clarified before close comparisons can be made. For
instance, Bather (1914/?, pp. 196, 199) stated that the radials slope upwards towards the

text-fig. 4. Enlarged sections of parts of paratype, USGD 2311, x7 .a, radial R3 consists of a large
calcite crystal with prominent cleavages. It is overlain by floor plates also exhibiting cleavages. Each
floor plate has a tear-drop shaped canal, the left also having a connecting pore. An area of recrystallized
calcite (lightly stippled) lies between the canals. In addition to the vertical suture between floor plates,
there are inclined sutures extending towards the ambulacral groove which seem to parallel cleavages
in the calcite of the floor plates, b, radial R4 of text-fig. 3 b consists of a large plate with two faint sutures
(one irregular) of unknown significance, and prominent cleavages. Pores in the overlying floor plates
seem to be connected to three small canals, two lateral and one medial. The medial canal lies on the
suture between adjacent floor plates just below the ambulacral groove, and seems to be joined by sutures
or passages to the lateral canals. They divide the floor plates into inner and outer parts. The lightly

stippled area consists of recystallized calcite.

ambulacral grooves from the top of the interradial sutures, and are notched for the
excavation of the ambulacral grooves ‘so that the pores between the floor-plates pass
through into the thecal cavity’. Hudson (1925, 1927) and Fay (1962), on the other hand,
considered that the radials slope downwards from the top of the interradial sutures
rounding the distal ends of the ambulacral grooves.

Another edrioblastoid has recently been collected from the Trelawney Beds of northern
New South Wales, which Philip (1966), on the basis of conodont work, regarded as
Upper Caradocian (i.e., Trentonian-Maysvillian of North America). The edrioblastoid
specimen (USGD 3304) consists of the upper part of an incomplete theca cut in longi-
tudinal section, 1 cm. wide, and exhibiting fused deltoids and floor plates with slit-like

520

PALAEONTOLOGY, VOLUME 11

pores. It probably represents a species of Astrocystites comparable with A. distans, but
until more material is found closer comparisons cannot be made.

Discussion of edrioblastoid relationships. Bather (19146, p. 202), while observing the
similarities in the nature and arrangement of the floor and cover plates in Astrocystites
and Edrioaster, noted that the ambulacra in Astrocystites are not sinuous, the cover
plates in the oral region are arranged in a different manner, and the floor plates are more
rigidly fused. The interradial element of E. bigsbyi, formed by the fusion of four floor
plates (Bather 1914m p. 165, text-fig. 3), is very different from that in A. distans , which
is a large rod-like structure flanked by shorter, thin plates inclining towards the element
because of the curvature of the theca. The pores in Edrioaster pass directly into the
thecal cavity through the entire length of the ambulacra, whereas, at least in A. distans ,
they only pass directly into the thecal cavity in the proximal part of the ambulacra.
Distally they lead into a canal which apparently opens into the thecal cavity at the top
of the radial.

In A. distans the hydropore is centrally situated and near the apex of the posterior
interradius. Leading from it the stone canal consists of an outer, centrally placed
funnel-shaped chamber linked by a narrow oblique canal to an inner double chamber.
This lies between the interradial element of the posterior interradius and the left
posterior ambulacrum, and is connected to the interior of the theca by a short, narrow
canal (text-fig. 5). This arrangement differs from that described by Kesling and Mintz
(1966) in the edrioasteroids Isoroplius cincinnatiensis (Roemer) and Carneyella pi lea

EXPLANATION OF PLATE 100

Figs. 1-8. Astrocystites distans sp. nov. 1, 2, Paratype, USGD 2310, X 3, slightly oblique cross-sections
at intervals of 2-5 and 17 mm. below the top of the theca; photographed from cellulose peels; 1,
Section through two ambulacra showing thickened deltoids and solidly fused floor plates penetrated
by pores, which gradually diminish in size proximally (see text-fig. 2b). 2, Section showing two
ambulacra near their distal ends resting on radials; pores extend through floor plates into a canal
overlying the radial ; outer surface of radials show strong crenulations and interior surface is smooth
(see text-fig. 2h). 3, Paratype, USGD 3303, X 3, internal view of adoral part of posterior interradius
showing nature of the stone canal and its relationship to the interradial element. 4, Paratype, USGD
2305, X 3, longitudinal section showing thickened deltoids and fused floor plates penetrated by slit-
like pores. 5, Paratype, USGD 2311, X 1-5, side view of specimen before sectioning showing two
ambulacra separated by an inter-ambulacral area; ambulacra rest on radials R1 and R5 respectively
(see text-fig. 3). 6-8, USGD 2311, X 5, thin sections at intervals of 2, 7, and 12 mm. below upper
surface shown in fig. 5. 6, slightly oblique section across thecal plates showing surface crenulations
and calcite cleavages in the plates; two ambulacra lie on radials R2 and R3 (for interpretations see
text-figs. 3n, 4a). 7, section showing radials, R4 and part of R5, and overlying floor plates (see text-
figs. 36, 46). 8, oblique section near distal ends of ambulacra showing them lying in grooves on
radials R1 and R5 (see text-fig. 3c). Note prominent central canal in each ambulacrum below the
floor plates, and well-developed cleavages in calcite of the radial plates.

Figs. 9-16. Miscellaneous pelmatozoan plates. 9, USG D 2307, x 3, large thecal plate. 10, USGD 3300,
X 3, circular structure with stellate grooving, and another obliquely cut structure. 11, USGD 2315,
x 3, large round columnal with stellate axial canal. 12, USGD 3301, x3, round columnal with
stellate axial canal and crenellae on part of the articulating surface. 13, USGD 3302, X 3, small
round columnal with round axial canal. 14, USGD 2307, X 3, round columnal with round axial canal
and concentrically grooved articulating surface. 15, USGD 2316, X 3, part of stem showing alternate
nodals and internodals. 16, USGD 2315, X 3, small, round articulated columnals. Photographs by
G. Z. Foldvary and B. D. Webby.

Palaeontology, Vol. 11

PLATE 100

WEBBY, Ordovician edrioblastoid

B. D. WEBBY: ASTROCYSTITES DISTANS SP. NOV.

521

(Hall), where the stone canal extends from the hydropore inwards between the posterior
interradius and the right posterior ambulacrum. A similar arrangement is exhibited by
Lepidodiscus fistidosus (Anderson 1939). On the other hand, Bather (1914<7, p. 168)
observed that in Edrioaster bigsbyi the hydropore becomes deeper adorally and in the
direction of the left posterior ambulacrum, suggesting a similar pattern to A. dislans.

Bather (1915, p. 266) concluded that in
edrioasteroids (including Astrocystites ) the
ambulacra exhibited an open system (similar
to that in asteroids), viz., ampullae in the
thecal cavity connected by way of the pores
to tube feet and the perradial canal lying in
the ambulacral groove. The tube feet were
presumed to have acted both in respiration
and feeding of the animal. Cover plates were
evidently opened when Edrioaster was feeding.

Anderson (1 939, p. 75), in contrast, considered
that in Lepidodiscus fistidosus and forms de-
scribed by Bather, including Astrocystites ,
a closed system (as in echinoids) was repre-
sented, viz., a perradial canal below the
ambulacral groove connected to external tube
feet and internal ampullae. The evidence from
the study of A. distorts supports Anderson’s
suggestion for a closed ambulacral system,
but the nature of the closed system differs.

Whereas in L. fistidosus the theca is flexible
and the ampullae and perradial canal lie in-
ternal to the skeleton, in A.distans the theca is
rigidly fused and both ampullae and perradial
canal are enclosed between radial and floor
plates in the distal parts of the ambulacra,
along the lines of the ophiuroid pattern
(Nichols 1966, p. 109, fig. 15c), and internal
to the skeleton in proximal parts of the
ambulacra. In a section across the radial and
floor plates in the distal part of one ambula-
crum (see text-fig. 46), side branches from the
medial (perradial) canal lead through the floor
plates to pores on the surface. Presumably the
side branches contained enclosed ampullae
which were surrounded by muscles capable
of protraction, and were connected to tube
feet occupying the pores.

In text-fig. 5 the inferred water vascular system in A. distans is represented. Judging
from the nature of the pores, the tube feet in the proximal part of the ambulacrum were
probably thinner and longer (text-fig. 5, t.f. 1) than those situated medially (t.f. 2), and

text-fig. 5. Schematic perspective diagram of
part of one ambulacrum and the inferred water
vascular system in Astrocystites distans. Del-
toids are represented by horizontal lines, the
radial by vertical lines and the floor plates by
crossed lines. Cover plates are not shown.
Hydropore (/?) and stone canal (s.c.) are situated
in the posterior interradius. The stone canal
consists of funnel-shaped and double chambers
linked by a narrow canal, and connected to a
ring canal with a perradial canal (p.c.) leading
into each ambulacrum. Three types of tube feet
are suggested: long, slender forms (t.f. 1 ) which
penetrate the thickened plates in the proximal
region; shorter, thicker forms (t.f. 2) in the
medial region; and shorter, thinner forms (t.f.
3) which protude through the rather thin floor
plates lying on the radial in the distal region.

522

PALAEONTOLOGY, VOLUME 11

penetrated floor plates and possibly margins of deltoids as well. Distally, the tube feet
appear to become thinner and shorter (t.f. 3). From the morphology of the inner,
double chamber of the stone canal it may be tentatively suggested that it contained valves
which regulated the flow of water through the system.

Specimen B of Bather’s description of A. ottawaensis exhibited the most complete
set of cover plates, the proximal and medial parts being covered but not the distal.
Bather (19146, p. 199) added that there was ‘no reason to doubt that they were con-
tinuous ... to the extreme end of the floor plates.’ It seems probable that the cover plates
opened along the margins of each ambulacrum, allowing the tube feet to protrude as in
edrioasteroids. The cover plates in the oral region, however, are more irregular in size
and shape, with smaller plates intercalated. There are, in addition, triangular oro-
tegminal plates lying with their broader ends on interambulacral areas, and narrower
ends reaching almost to the oral pole. It may be assumed from this arrangement that
the circlet of oro-tegminal and cover plates in the oral region was fixed, and the smaller
pores underlying this region, as seen in A. distans, contained slender, elongated tube
feet. The tube feet of this region may have been mainly sensory and respiratory in
function, whereas those in medial and distal parts of the ambulacrum served also in
feeding.

Edrioasteroids range from the Lower Cambrian to the Lower Carboniferous, evolving
most rapidly in the Ordovician. From the close similarities in plating of the ambulacra
and in the nature of the posterior interradius, there seems to be little doubt that during
the Middle Ordovician the small, stalked group of edrioblastoids evolved from the basic
edrioasteroid stock. Both Astrocystites and Edrioaster exhibit a hydropore and a cluster
of plates around the anus in the posterior interradius. However, Astrocystites may be
distinguished from Edrioaster by the pyriform shape of the theca, the rudimentary stem
and, to a certain extent, by the stabilization of the number of thecal plates — five basals,
five radials, and five deltoids.

Astrocystites superficially resembles blastoids like Pentremites but exhibits funda-
mental differences. No hydrospires, brachioles or spiracles can be recognized. Even if
an internal soft-bodied chamber analogous to a hydrospire existed, it must have been
restricted to the region below the proximal part of each ambulacrum, for the pores in
the distal part communicate with a medial (perradial) canal lying between the radial
and floor plates. If the pores had solely a respiratory function, some type of soft struc-
tures, like brachioles attached to floor plates and capable of extension on opening of
cover plates, must have been present to catch and direct food particles towards the
mouth. However, there is no direct evidence to support any of these speculations, and
it must be concluded that edrioblastoids are morphologically distinct from blastoids.

Fay (1962, p. 205; 1967, pp. 242, 286) suggested that some form related to Astro-
cystites could have been ancestral to the blastoids which, according to him, appear in
the Middle Silurian and range into the Permian. For an AstrocystitesAike form to give
rise to spiraculate blastoids the following changes would be required: (1) the develop-
ment of a stem with definite columnals, (2) resorption of two sutures to produce three
basals; (3) fusion of the infradeltoids with the deltoids; (4) reduction by fusion in the
number of anal plates in the posterior interradius; (5) atrophy of the tube feet and in-
folding of the deltoids and radials to form the hydrospire; (6) replacement of a single
hydropore at the upper extremity of the posterior interradius by a spiracle at the top

B. D. WEBBY: ASTROCYSTITES DISTANS SP. NOV.

523

of each deltoid (some blastoids develop paired spiracles); (7) migration of radial
canal inward or outward, and consequent secretion of calcium carbonate to form the
lancet plate; and (8) development of brachioles. This, however, would entail more
radical changes than by deriving blastoids, in the more generally accepted fashion, from
cystoids through the transitional Middle Ordovician form Blastoidocrinus (Nichols 1966,
pp. 146-8). Moore (1954) suggested derivation from a rhombiferan Cystoblastus- like
ancestor by shifting of radials downward, laterals upward and partial fusion of basals,
Brachioles were already developed, and hydrospires have been shown to be homolo-
gous with pore rhombs. However, the derivation of the lancets by modification of the
radials seems too conjectural for the ready acceptance of this view.

Miscellaneous pelmatozoan plates. In addition to the thecal plates and fragments which
can be positively recognized as parts of A. distans , there are numerous stem ossicles
and a few plates from the Cliefden Caves locality which cannot be so confidently
determined. Two main types of stem columnal can be distinguished; those with a
stellate axial canal and those with a circular axial canal. The largest columnal (USGD
2315) is round, 5 mm. in diameter and has a stellate axial canal; part of the articulating
surface shows faint crenellae, but it is otherwise smooth (PI. 100, fig. 11). Another round
columnal (USGD 3301), 3-8 mm. in diameter, also shows a stellate axial canal and a
patchy development of crenellae on the articulating surface (PI. 100, fig. 12). The
columnal USGD 3302 is 2-0 mm. in diameter, has a small round axial canal and
a smooth flat articulating surface (PL 100, fig. 13). Another (USGD 2307) is 2-3 mm. in
diameter, has a larger round axial canal, and a concentric groove on the articulating
surface (PI. 100, fig. 14). In specimen USGD 2316 a pattern of round alternating nodal
and internodal columnals is shown; the nodals have a diameter of 2-5 mm. and the
internodals 2-0 mm.; the columnals are spaced about six in 4 mm. (PI. 100, fig. 15).
A thinner articulated specimen (USGD 2315) has round columnals, 1-2 mm. in diameter,
and of uniform size, averaging five in 4 mm. (PI. 100, fig. 16).

It is not possible at the present time to assign any of these columnals to A. distans.
None of them seem to be comparable with the elements of the stem in A. ottawaensis.
Bather (19146) described the stem as circular and composed of pentameres, and Fay
(1962) similarly observed that it was made up of curved polygonal plates grouped into
three columnal rings each with five or more plates. Certainly the variety of stem colum-
nals from the Cliefden Caves locality strongly suggests the presence of at least one other
pelmatozoan in the fauna. This is confirmed by two other finds. The first is a large plate
(USGD 2307) 14-5 mm. wide and 8 mm. high, with a radiating pattern of fine crenula-
tions, three on the left side, six medial, and three on the right side; it is up to 3 mm.
thick on the upper edge (PI. 100, fig. 9). It resembles a cystoid thecal plate, and it is
possible that a pore rhomb has been eroded from the thickened upper edge. Secondly,
in USGD 3300, there is a circular structure with five grooves radiating from the centre
(PI. 100, fig. 10). The grooves subdivide away from the centre, and there seems to be
a development of partitions across the grooves towards the periphery, forming facets
possibly for the attachment of brachioles. To one side of the circular structure there
appears to be a less well-defined, obliquely cut structure. The circular structure resembles
the oral region of certain rhombiferan cystoids, for example, Glaphyrocystis Jaeckel and
Echinoenerinites Meyer (Hecker 1964, pi. 3, fig. 9b; text-fig. 19), or the eocrinoid

C 5934

m m

524

PALAEONTOLOGY, VOLUME 11

Cryptocrinus (Bather 1900, p. 70, text-fig. 37). It is possible that the isolated large plate
and the circular structure belong to the same species. The only rhombiferan cystoid
previously described from the Ordovician of Australia is Cheirocrinus merrileesi Brown
1964, which has very different ornamentation on thecal plates.

Eocrinoids have not yet been reported from Australia.

Acknowledgement. This work was supported by a grant from the Australian Research Grants Com-
mittee.

REFERENCES

anderson, f. w. 1939. Lepidodiscus fistulosus sp. nov. from Lower Carboniferous rocks, Northumber-
land. Bull. geol. Surv. Gt. Br. 1, 67-81.

bassler, r. s. 1935. The classification of the Edrioasteroidea. Smithsonian misc. collns, 93 (8), 1-11.
1936. New species of American Edrioasteroidea. Ibid. 95 (6), 1-33.

and moodey, m. w. 1943. Bibliographic and faunal index of Paleozoic pelmatozoan echinoderms.

Spec. Pap. geol. Soc. Am. 45, 1-734.

bather, f. a. 1900. Echinoderma, In lankester, e. r. A Treatise on Zoology. Part 111. 344 pp.
London.

19 14z7. Studies in Edrioasteroidea, Part IV. The Edrioasters of the Trenton Limestone. Geol.

Mag. [6] 1, 115-25, 162-71.

19146. Studies in Edrioasteroidea, Part V. Steganoblastus. Ibid. 1, 193-203.

1915. Idem. Part VII. Morphology and bionomics of the Edrioasteridae. Ibid. 2, 211-15,

259-66.

brown, i. A. 1964. An Ordovician cystoid (Pelmatozoa, Echinodermata) from Australia. J. Proc. R.
Soc. West. Aust. 47, 3-7.

fay, r. o. 1962. Edrioblastoidea, A new class of Echinodermata. J. Paleont. 36, 201-5.

1967. Evolution of the Blastoidea. In teichert, c. and yochelson, e. l., Eds. Essays in Paleon-
tology & Stratigraphy. R. C. Moore Commemorative Volume. 242-86, Univ. Kansas Press.
fell, h. b. 1965. Early evolution of the Echinozoa. Breviora, 219, 1-17.

■ and moore, r. c. 1966. General features and relationships of Echinozoans. In moore, r. c., ed..

Treatise on Invertebrate Paleontology, Part U, Echinodermata 3, vol. 1. U108-18. Geol. Soc. Amer.
and Univ. Kansas Press.

hecker, r. f. 1964. Class Cystoidea. In orlov, y. a., Osnovy Paleontologii. Echinodermata, Hemi-
chordata, Pogonophora and Chaetognatha. 30-45. Moscow.
hill, d. 1957. Ordovician corals from New South Wales. J. Proc. R. Soc. N.S.W. 91, 97-107.
Hudson, g. h. 1925. The need of improved technique in illustration. J. Geol. 33, 642-57.

1927. The surface characteristics of Astrocystites ( Steganoblastus ) ottawaensis. Rep. St. Geol.

Vermont, 15, 97-100.

kesling, r. v. and mintz, l. w. 1966. Internal structures in two edrioasteroid species, Isorophus
cincinnatiensis (Roemer) and Carney ella pilea (Hall). Contr. Mus. Paleont. Univ. Mich. 15 (14),
315-48.

moore, r. c. 1954. Status of invertebrate paleontology, 1953, IV. Echinodermata; Pelmatozoa. Bull.

Mus. comp. Zool. Harv. 112 (3), 125-49.
nichols, d. 1966. Echinoderms. 200 pp. London.

philip, G. m. 1966. The occurrence and palaeogeographic significance of Ordovician strata in northern
New South Wales. Aust. J. Sci. 29, 112-13.

regnell, g. 1966. Edrioasteroids. In moore, r. c., ed., Treatise on Invertebrate Paleontology, Part U,
Echinodermata 3, vol. 1. U1 36-73. Geol. Soc. Am. and Univ. Kansas Press.
sherrard, k. m. 1954. The assemblages of graptolites in New South Wales. J. Proc. R. Soc. N.S.W.
87, 73-101.

spencer, w. k. 1938. Some aspects of evolution in Echinodermata. In de beer, g. r., Evolution. Essays
on Aspects of Evolutionary Biology, 287-303. Oxford.
stevens, n. c. 1952. Ordovician stratigraphy at Cliefden Caves, near Mandurama, N.S.W. Proc. Linn.
Soc. N.S.W. 77 (3-4), 114-20.

B. D. WEBBY: ASTROCYSTITES DISTANS SP. NOV. 525

whiteaves, j. f. 1 897. Description of a new genus and species of Cystidean from the Trenton Limestone
at Ottawa. Can. Rec. Sci. 7, 287-92.

1898. Postcript. Ibid. 395.

wilson, a. e. 1946. Echinodermata of the Ottawa Formation of the Ottawa-St. Lawrence Lowland.
Bull. geol. Surv. Can. 4, 1-61.

B. D. WEBBY

Department of Geology and Geophysics
University of Sydney
Sydney, N.S.W.

Typescript received 26 September 1967 Australia

C HU B BIN A, A NEW CRETACEOUS ALVEOLINED
GENUS FROM JAMAICA AND MEXICO

by E. ROBINSON

Abstract. Chubbina jamaicensis gen. et. sp. nov. and Chubbina macgillavryi sp. nov. from the Cretaceous rocks
of Jamaica and southern Mexico are described. Borelis cardenasensis Barker and Grimsdale is placed in the
new genus. Chubbina appears to be a useful index for the Upper Cretaceous (Upper Campanian to Maastrich-
tian) of the Greater Antilles and Central America.

In 1937 Barker and Grimsdale described a new alveolinid species, Borelis cardenasensis ,
from the Cardenas beds of north-eastern Mexico. In this description they doubtfully
referred the species to Borelis, mainly on the grounds that it bore a superficial resem-
blance to Borelis jamaicensis Vaughan and Borelis matleyi Vaughan (1929). Cole (1956)
showed that both of Vaughan's species belonged to the genus Fabularia.

Under Cosine/la, Reichel (1937, p. 136 n.) mentioned a manuscript species of Schlum-
berger and compared it with B. cardenasensis but no description of Cosinella was given
either then or subsequently. Reichel has indicated (pers. comm., 1966) that he is still in
possession of Schlumberger’s manuscript material for Cosinella, which must remain
a nomen nudum. Nevertheless the name Cosinella was used, without further description,
by Dunnington et al. (1960, p. 61) for specimens subsequently described by Smout as
Pseudedomia globularis (1963), and was also considered by Seiglie and Ayala (1963) for
specimens resembling B. cardenasensis found in the Cuban Maastrichtian. Smout (1963,
p. 225) in emending the genus Pseudedomia , briefly discussed B. cardenasensis, suggest-
ing, on the basis of information received from Grimsdale, that the species had a uni-
serial termination and was, therefore, possibly related to Raadshoovenia van den Bold
(1946). On the other hand Colalongo (1963) included B. cardenasensis in her new genus
Sellialveolina, apparently on the basis of the original description.

During studies of the Cretaceous inliers of Jamaica, undertaken by the Geology
Department of the University of the West Indies, numbers of larger foraminifera have
been collected, amongst which are numerous examples of an alveolinid related to
‘ Borelis ’ cardenasensis. The internal structures of the Jamaican form are also similar to
those of Pseudedomia and Sellialveolina, although the mode of coiling is basically
different.

On the basis of an examination of the Jamaican material, which contains a large
number of microspheric individuals, the new genus Chubbina is proposed, with Chubbina
jamaicensis sp. nov. as the type. ‘ Borelis ’ cardenasensis is considered to be congeneric
with the Jamaican form. Recently Professor H. J. MacGillavry kindly sent me thin
sections of a complanate alveolinid which he collected in southern Mexico. These
specimens also belong to Chubbina and are further described below.

Acknowledgements. I wish to thank Dr. A. H. Smout for reading part of the manuscript of this paper
and Mr. R. DeSouza who assisted with the photography. The new genus is named for Dr. L. J. Chubb,
in appreciation of his work on the Cretaceous of the Caribbean area.

[Palaeontology, Vol. 11, Part 4, 1968, pp. 526-34, pi. 101-3.]

E. ROBINSON: CHUBBINA, A NEW ALVEOL1NID GENUS

527

Location. The species described have come from the following localities.

Locality 1. ER 108, on the railway between Catadupa and Cambridge, Jamaica, W.I., at the tenth
telegraph pole, plus 20 ft., after milepost 96.

Locality 2. ER 213A, three-quarters of a mile west of Frankfield, Parish of Clarendon, Jamaica, W.L,
at Guinea Corn, in the bed of the Rio Minho, in the type section of the Guinea Corn Formation
(Coates 1965), about 40 ft. stratigraphically below the top of the limestone.

Locality 3. Type locality for Chubbina cardenasensis (Barker and Grimsdale 1937).

Locality 4. On the road between Tuxtla Gutierrez and Ocozocuautla, state of Chiapas, Mexico, at
stop 2, K.1061, (fig. 1, Excursion C-15-b, 20th International Geological Congress, Mexico, 1956).

Figured specimens are deposited at the British Museum (Natural History) (BMNH).

SYSTEMATIC DESCRIPTIONS

Family alveolinidae Ehrenberg, 1839
chubbina gen. nov.

Type species, Chubbina jamaicensis sp. nov.

Diagnosis. Test with walls composed of microcrystalline calcite, imperforate, probably
originally porcellaneous; in the megalospheric generation subglobular or lenticular,
becoming discoidal in the mature microspheric form; chamber coiling markedly
streptospiral in the early stages, becoming almost planispiral in the last whorls of the
microspheric generation; involute, often becoming pseudevolute in the later whorls
and strongly recurved but not cyclical; the number of chambers per whorl increasing
with the addition of each whorl and always being greater than two in the ephebic stage.

Dimorphism pronounced; the single, spherical proloculus of the megalospheric
generation followed by a single tubular chamber, without internal structures; micro-
spheric generation with a milioline nepiont; succeeding chambers subdivided into
spirally directed, tubular chamberlets, embedded in a solid microcrystalline endoskeleton
or ‘couche basale’; chamberlets multiple, initially arranged in one or more tiers or
layers, later becoming irregular in layering, with varying differentiation into a subepi-
dermal layer of primary chamberlets and multiple supplementary chamberlets in the
‘couche basale’; pre-septal canals present, irregularly equipped with buttresses;
communication between chamberlets in successive chambers effected by means of
rounded multiple apertures, more or less in alignment with the chamberlets.

Chubbina jamaicensis sp. nov.

Plate 101, figs. 1-6; Plate 102, figs. 1-5

Microspheric Form. Test free, subglobular to subdiscoidal, consisting of 6 to 8 strepto-
spirally coiled whorls after the nepionic stage; early stages involute, forming a sub-
spherical, immature test; at about the fifth whorl becoming flaring, with the development
of a peripheral flange with an acute sub-angular margin ; in the final whorl pseudevolute
and flaring, with a broad, complanate flange partly surrounding an excentric, umbonate
region containing the earlier part of the coil; peripheral margin of the final whorl usually
acquiring a stoutly rounded rim which may become equal in thickness to the umbonate
region of the test; the distal, recurved face of the final whorl flattened, with sub-angular

528 PALAEONTOLOGY, VOLUME 11

margins, and provided with numerous round apertures, scattered more or less uniformly
over the whole face.

Coil subdivided into chambers by septa which are the turned-in continuations of the
spiral wall; chambers about 6 in number in the fourth whorl, rising to 15 in the final,
complanate whorl; septa becoming increasingly recurved in the later whorls; alar
prolongations of the chambers of the final whorl slightly vorticiform.

The microspheric nepiont was not clearly seen but appears to be milioline. The
succeeding chambers are filled with a solid mass of microcrystalline calcite forming
the ‘couche basale’, pierced by spirally directed, tubular chamberlets. The chamberlets
are continuous between the distal and proximal septal walls of each chamber and the
chamberlets of any one chamber are connected with the chamberlets of the preceding
and succeeding chambers by more or less directly aligned rounded apertures in the septa.
In each chamber the material of the ‘couche basale’ does not extend as far as the distal
septum, but leaves a space, a pre-septal canal, allowing free communication between the
chamberlets. However, the pre-septal canal is traversed irregularly by projections from
the ‘couch basale’, which reach the distal septum forming buttresses.

The 3 or 4 chamberlets in the chambers of the first whorl are arranged in a single, more
or less regular tier. In the second or third whorl the increasing number of chamberlets
gives rise to a second tier below the first. At about the fourth whorl a third tier of
chamberlets is gradually inserted. In the last-formed whorl the number of tiers or rows
rises from 4 or 5 to 80 or 90. Subdivision of the chamberlets occurs in the region of each
pre-septal canal, by the addition of apertures through the distal septum, from which
additional chamberlets project into the next-formed chamber. The addition of chamber-
lets is a more or less continuous process, so that the second tier of chamberlets is
acquired gradually and only becomes clearly defined over the space of 2 or 3 chambers.
The later addition of tiers occurs in a similar manner and these are largely acquired
through the multiplication of chamberlets, along the inner margin of the whorl. Although
all the chamberlets are spirally directed, a degree of undulation or meandering occurs
so that, beyond about the fourth whorl, they are no longer arranged in regular tiers but
occur scattered uniformly in the ‘couche basale’. The distance between adjacent
chamberlets is approximately the same as the diameter of the chamberlets themselves
in the final whorl. In the early whorls it is less. The chamberlets nearest to the spiral wall
tend to be aligned more regularly along the wall.

EXPLANATION OF PLATE 101

Fig. 1. Chubbina jamaicensis gen. et sp. nov., cotype, microspheric form, sectioned normal to the axis
of coiling of the final whorl, x 20. BMNH P. 48049.

Fig. 2. Chubbina jamaicensis gen. et sp. nov., cotype, microspheric form, sectioned parallel to the axis
of coiling of the final whorl, x 30. BMNH P. 48048.

Fig. 3. Chubbina jamaicensis gen. et sp. nov., cotype, megalospheric form, sectioned to show the
progressive change in direction of coiling in successive whorls, 'canal flexostyle’ and buttresses,
x 45. BMNH P. 48052.

Fig. 4. Section through portion of terminal flange of C. jamaicensis, showing some differentiation of
a subepidermal layer of chamberlets as seen in Pseudedomia globularis Smout. x 20. BMNH P. 48050.

Fig. 5. Natural internal mould of a portion of several chambers from a microspheric individual of
C. jamaicensis, showing tubular chamberlets and preseptal canals. X 16. BMNH P. 48055.

Fig. 6. Chubbina jamaicensis gen. et sp. nov., cotype, external view, X 18. BMNH P. 48056.

All specimens from Locality 1 , Jamaica.

Palaeontology , Vol. 11

PLATE 101

ROBINSON, Chubbina gen. nov.

E. ROBINSON: CHUBBINA, A NEW ALVEOLINID GENUS 529

Megalospheric form. Test much smaller than in the microspheric form, sub-globular to
inflated lenticular, consisting of 4 to 6 streptospirally coiled, involute whorls, the last-
formed whorl frequently showing a tendency to acquire a flaring end stage; megalo-
spheric specimens with a fully complanate, pseudevolute final stage not observed.

Internally there is a single, spherical proloculus, followed by a single, tubular chamber
without internal structure, a ‘canal flexostyle’. Further internal development closely
parallels that of the microspheric form, but megalospheric specimens with more than 10
rows of chamberlets in the last whorl, when seen in spiral section, have not been observed.

Critical measurements of C . jamaicensis are included in Table 1. Text-fig. 1 shows the
size range of specimens of Chubbina from Jamaica and elsewhere.

Type locality. Locality 1. Additional material from locality 2.

Chubbina cardenasensis (Barker and Grimsdale 1937)

1937 Borelis cardenasensis Barker and Grimsdale, p. 173, pi. 9, figs. 1-5.

Examination of Barker and Grimsdale’s original figures and of additional syntypes
reveals that pre-septal canals or passages are present, and occupy the space in each
chamber in front of the incomplete transverse septa mentioned by Barker and Grimsdale.
No true buttresses are visible on any of the original figures, but one can see that
buttresses are probably present from an examination of figures 3 and 4 of the
original description. These show a certain irregularity in the shape of the spiral and
transverse septa of Barker and Grimsdale, where parts of the sections are cut through
the pre-septal canals. Dr. C. G. Adams of the British Museum (Natural History) has
confirmed that no specimens likely to represent a microspheric form are present
amongst Grimsdale’s syntypes.

Chubbina cardenasensis differs from C. jamaicensis primarily in the rate of chamber
enlargement. This is more rapid in C. jamaicensis resulting in the insertion of the third
tier of chamberlets in the third or fourth whorl. In C. cardenasensis the third tier of
chamberlets is not inserted until the fifth whorl or later.

Chubbina macgillavryi sp. nov.

Plate 102, figs 8; Plate 103, figs. 3, 4; see also Plate 102, figs. 6, 7

Thin-sections of a limestone collected in southern Mexico by Professor H. C. Mac-
Gillavry and sent to the author, contained numerous random sections of Chubbina.
They differ from C. jamaicensis and C. cardenasensis. The coiling is streptospiral, but
the degree of eccentricity of the coiling is more variable than in the other two species.
Some specimens, typified by that illustrated on Plate 103, fig. 3, are similar in their
coiling mode to C. jamaicensis. Other specimens may approach a planispiral mode
(PI. 102, fig. 6). C. macgillavryi resembles C. cardenasensis in the relatively low rate of
chamber enlargement. The third tier of chamberlets is normally inserted at about the
fifth whorl, in some specimens even the second tier is not inserted until the fourth or
fifth whorl. The preseptal canals of C. macgillavryi are narrower than in either C.
jamaicensis or C. cardenasensis and the proloculus of the magalospheric generation is
also larger. Table 1 compares some of the critical measurements. A single, uncentred

530

PALAEONTOLOGY, VOLUME 11

section of a microspheric individual (PI. 103, fig. 4) is larger than the largest example of
C. jamaicensis yet found. All the available thin sections show that the chamberlets of the
first few chambers are relatively rectangular in cross-section with thin walls.

Type locality. Locality 4.

table 1.

Comparison of some critical measurements of species of Chubbina.

(A) C. jamaicensis, (B) C. cardenasensis, (C) C. macgillavyri.

A

B

C

P

P

P

Diameter of chamberlets

40-55

60

25-50

Spacing between chamberlets

30

10-30

10-20

Distance across pre-septal canal

90-140

120

25-100

Diameter of proloculus

100-180*

90-120

1 50-300f

* Only one specimen of C. jamaicensis has a prolocuius reaching 180 p in diameter. The next largest
measured was 140 p.

t Although all the magalospheric specimens of C. macgillavryi had large proloculi, there was some
degree of recrystallisation at the centres of many individuals, making accurate measurement difficult.

Discussion. The specimens illustrated by Seiglie and Ayala-Castanares (1963, pi. 1,
figs. 2 and 3 appear to belong to Chubbina cardenasensis. The section illustrated by
these authors as Rhapidionina sp. (ibid., p. 27 ; pi. 1, fig. 1) is rather indistinct, but could
be a portion of the terminal flange of Chubbina.

Most of the skeletal features mentioned in the preceding descriptions have already
been noted in discussions of the complanate genus Pseudedotnia by Eames and Smout
(1955), Smout (1963) and Reiss et al. (1964). The major difference between Pseudedotnia
and Chubbina lies in the streptospiral mode of coiling in the latter genus. This coiling
eccentricity continues on a diminishing scale throughout ontogeny and is readily
distinguishable from the slight irregularities which may develop in a large specimen of
Pseudedotnia, due for instance to environmental conditions. The eccentricity is apparent
in most thin sections of Chubbina but it is only when a centred thin section happens to be
cut normal to the axis about which the coiling plane rotates that the degree of rotation
can be measured (PI. 101, fig. 3).

EXPLANATION OF PLATE 102

Figs. 1, 2, and 4. Chubbina jamaicensis gen. et sp. nov., cotypes, all megalospheric individuals from
Locality 1, showing limits of variation in rate of chamber enlargement and a progressive deteriora-
tion in the state of preservation of successive chamber walls, possibly due to diagenetic effects.
Fig. 1, BMNH P. 48051, fig. 2, BMNH P. 48053, fig. 4, BMNH P. 48047.

Fig. 3. Chubbina jamaicensis , megalospheric form, from Locality 2. x45. BMNH P. 48054.

Fig. 5. Off-centred section of C. jamaicensis from Locality 1, showing preseptal canals, buttresses and
some differentiation of a subepidermal layer of chamberlets in the last whorl, x 45. BMNH P. 48046.

Fig. 6. Chubbina sp. cf. C. macgillavryi sp. nov., Locality 4, showing canal flexostyle, relatively low
rate of chamber enlargement and relatively thin walls between chamberlets. x 30. BMNH P. 48042.

Fig. 7. Chubbina sp. cf. C. macgillavryi sp. nov., Locality 4. ? microspheric individual. X 30. BMNH
P. 48040.

Fig. 8. Chubbina macgillavryi sp. nov., cotype, Locality 4. Off-centred section of megalospheric speci-
men. x 30. BMNH P. 48041.

Palaeontology, Vol. 11

PLATE 102

ROBINSON, Chubbina gen. nov

E. ROBINSON: CHUBBINA, A NEW ALVEOLIN ID GENUS 531

The partitioning of the chambers, with the resultant chamberlets, preseptal canals, and
buttresses, is very similar to that of Pseudedomia globularis Smout. Unlike those seen
in P. drorimensis Reiss et al., the chamberlets resemble tubes excavated in a solid ‘couch
basale’ rather than chamberlets enclosed by regular, spirally directed partitions. The
irregularity in chamberlet formation in the later chambers suggests that chamber division
is not ordered by the secretion of subepidermal partitions, but occurs through the
somewhat random development of spirally directed tubular filaments which surround
themselves with a solid skeletal framework. The loss of regular tiering in the later
stages is due to the confining influence of the spiral wall, which becomes progressively
both higher and relatively narrower, forcing the tubular filaments to accommodate
themselves within the available space. It is for this reason that the distinction between
primary and secondary chamberlets, noted by Smout, for instance in the case of P.
globularis (1963), is not here considered important. The distinction is sometimes seen
in specimens of Chubbina (PI. 101, fig. 4) but the development of primary chamberlets
is irregular. The outermost chamberlets in any one chamber are confined in their
position by the enclosing spiral wall, which is secreted first, by a process separate from
that producing the ‘couche basale’.

Chubbina resembles Sellialveolina vialli Colalongo in its gross internal structure and
it is possible that S. vialli possesses a yet undiscovered complanate end stage. However
the diagnosis for Sellialveolina should be emended to exclude Chubbina cardenasensis.
Colalongo specifically mentions (1963, p. 7) as one of her distinctions between S. vialli ,
the genus type, and ‘ Borelis ’ cardenasensis , that in ‘ B\ cardenasensis: ‘the coiling of the
first whorls is streptospiral but in S. vialli it is always planispiral’ (trans.). This distinc-
tion is considered here to be of generic importance, particularly as the stratigraphical
and geographical distributions of the two forms are markedly different. On the other
hand there is some justification for erecting a new subfamily to include complanate
species of alveolinid.

Fourcade (1966) described the new genus Murciella from the Upper Cretaceous of
Spain. The internal structures show that this form belongs correctly to the Pseudedomia j
Chubbina group, and in my view, should also be included in the Alveolinidae, but the
uniserial termination of its coil allows separation at the generic level. At the same time
it should be noted that rare, megalospheric individuals in populations of Chubbina
jamaicensis also terminate their growth with one or two uncoiled chambers. The ap-
parently earlier appearance of Murciella in Europe (Bignot 1967) would preclude any
direct connection with Chubbina.

Reiss et al. (1964) suggested that Pseudedomia could be related to the peneroplids
through the development of excessive thickening of the subepidermal partitions to
produce a ‘couche basale’. It seems that a ‘couche basale’, with associated multiple
rows of chamberlets, is developed characteristically in only two groups, the Alveolinidae
(including complanate forms) and the so-called Fabulariinae, and this may suggest a
close relationship. However, both groups evolve their multiple rows of chamberlets in
different geological eras. Moreover, the Tertiary genus Fabularia retains a biloculine
type of coiling, so that direct descent from a multiloculine Cretaceous alveolinid
ancestor, with multiple rows of chamberlets, is unlikely. In Jamaica, Fabularia is
seen to evolve from a primitive form with a single row of chamberlets {Fabularia
matleyi (Vaughan)) through intermediate forms, retaining the single row of

532

PALAEONTOLOGY, VOLUME 11

chamberlets, but with basal thickening ( Fabularia vaughani Cole), to advanced forms
with multiple rows of chamberlets in a ‘couche basale’ ( Fabularia verseyi Cole).

Another American species which appears to belong to the complanate alveolinid
group is Borelis gunteri Cole. The type figures (Cole 1941, pis. 2, 18) are rather indistinct,
but B. gunteri appears to be a planispiral form, like Pseudedotnia. It may also be related
to Raadshoovenia van den Bold, but both types need re-examination.

E

E

LO *

to

V)

(D

c

u

o® • °# o o

o

o

• n ft ° ° °

+ +oS 8 °°°

XV

o o

o

megatospheric

oo o
o

o o

mi crospheric

o

o

greatest diameter 8mm

text-fig. 1. Measurements of thicknesses and greatest diameters of Chubbina.

Dots are C. jamaicensis. Locality 1 .

Circles are C. jamaicensis, Locality 2.

X signs are syntypes of C. cardenasensis (Barker and Grimsdale), Locality 3.

+ signs are C. macgillavyri. Locality 4.

Distribution. Chubbina has been recorded from Mexico (Barker and Grimsdale, op. cit.
and this paper), Cuba (Seiglie and Ayala, op. cit.) and Jamaica. There are a number of
other records of Cretaceous alveolinids in the Caribbean and Central American region,
which, on further investigation, would probably prove to be species of Chubbina (e.g.
Bronnimann in Dixon 1956, p. 81, as Cosinella , from the upper Cretaceous of British

EXPLANATION OF PLATE 103

Fig. 1. Ayalaina rutteni (Palmer), from the lower part of the Guinea Com Formation, near Locality 1.
Jamaica. X 12. BMNH P. 48058.

Fig. 2. Undetermined alveolinid from Locality 4, having only a single layer of chamberlets in each
chamber, x 30. BMNH P. 48039.

Fig. 3. Chubbina macgillavryi sp. nov., megalospheric cotype from Locality 4, showing large pro-
loculus, canal flexostyle and low rate of chamber enlargement, x 30. BMNH P. 48043.

Fig. 4. Chubbina macgillavryi sp. nov. microspheric form, from Locality 4, section cut tangentially,
normal to the axis of coiling of the final whorl, x 20. BMNH P. 48045.

Fig. 5. Part of a thin section from Locality 4, showing abundant specimens of Chubbina. rudistid
fragments and an off-centred transverse section of Ayalaina rutteni (Palmer). X 8. BMNHP. 48044.

Palaeontology, Vol. 1 1

PLATE 103

ROBINSON, Chubbina and Ayalaina

E. ROBINSON: CHUBBINA, A NEW ALVEOLINID GENUS 533

Honduras). Brown and Bronnimann (1957, p. 35). reported alveolinids from the type
locality of Kathina jamaicemis (Cushman and Jarvis). This locality is only 20 metres
from the type locality of Chubbuna jamaicensis (which also contains K. jamaicensis).

In Jamaica C. jamaicensis is an index fossil for the group of limestones, including the
Guinea Corn Formation of Coates (1965), which contains the Tilanosarcolites rudistid
fauna of Chubb (1955, pp. 178, 183) considered by him to be of Maastrichtian age.
It is normally associated with K. jamaicensis, miliolids, and ostracodes. Vaughanina
cubensis minor Seiglie and Ayala and Sulcoperculina dickersoni (Palmer) are also found
at this horizon. Recently Ayalaina rutteni (Palmer) has been found for the first time in
the same limestone group ( PI. 103, fig. 1, specimen donated by Mr. A. D’Aguilar of the
University of the West Indies). Previously this species (as IMeandropsina rutteni) had
been recorded in Jamaica only from the limestones containing the Barrettia rudistid
fauna, considered by Chubb to be of Campanian age (in Zans et a/. 1962, pp. 1 1—15).

In the material from locality 4, C. macgillavryi occurs with K. jamaicensis, Sulcoper-
culina sp. and Ayalaina sp. In Cuba Seiglie and Ayala (op. cit.) consider the range of their
‘ Borelis ’ cf‘B\ cardenasensis (C. cardenasensis of this paper) as being upper Campanian
to lower Maestrichtian. The types of C. cardenasensis come from a section which Barker
and Grimsdale considered to be of Campanian to Maastrichtian age (op. cit.).

Chubbina , then, appears to be of regional stratigraphical value in indicating a late
Campanian to Maastrichtian age for rocks in which it is found. In Jamaica the genus
is also restricted to rocks whose general appearance and faunas indicate a shallow shelf
or lagoonal environment (Coates 1965).

Although Chubbina may be the most common late Cretaceous alveolinid in the
Americas, other genera do occur. One of these is a planispirally coiled form, apparently
with only one row of chamberlets to each chamber. A single section of this form,
collected from Locality 4, is illustrated on Plate 103, fig. 2.

REFERENCES

barker, r. w. and grimsdale, t. f. 1937. Studies of Mexican fossil Foraminifera, III. A new Alveo-
linellid from the Upper Cretaceous of Mexico. Ann. Mag. Nat. Hist. (10) 19, 173-8, pi. 9, figs. 1-5.
bignot, g. 1967. Presence de Murciella cuvillieri Fourcade dans le Liburnien des environs de Trieste.
C.r. Somme Seane. Soc Geol. Fr. 2, 50, 51.

brown, N. K. and bronnimann, p. 1957. Some Upper Cretaceous rotalids from the Caribbean region.
Micropaleont. 3, 29-38, pi. 1.

chubb, l. j. 1955. The Cretaceous succession in Jamaica. Geol. Mag. 92, 177-95.
coates, A. G. 1965. A new section in the Maestrichtian Guinea Corn formation near Crawle River,
Clarendon. Quart. J. Geol. Soc. Jamaica, 7, 28-33.
colalongo, m. L. 1963. Sellialveolina vialli n. gen. n. sp. di alveolinide cenomahiano delfAppennino
meridionale. Giorn. Geol. Bologna (2) 30, 1-10, pi. 1.
cole, w. s. 1941. Stratigraphic and paleontologic studies of wells in Florida. Geol. Surv. Bull. Florida,
19, 1-34, pi. 2, 18.

1956. Jamaican larger foraminifera. Bull. Amer. Paleont. 36, 205-33, pi. 24-31.

congreso geologico internacional, 1956. Geologia a lo largo de la carretera entre Tuxtla Gutierrez,
Chiapas, y Mexico D.F., y visita a monumentos precoloniales de Oaxaca, Oaxaca. In Libreto-Guia
de la Excursion C-15B, XX Congreso Geologico Internacional, Mexico. 53 pp.
dixon, c. G. 1956. Geology of southern British Honduras. Govt. Printer, Belize, 91 pp.
eames, f. e. and smout, a. h. 1955. Complanate alveolinids and associated Foraminifera from the
Upper Cretaceous of the middle East. Ann. Mag. Nat. Hist. (12) 8, 505-12, pi. 9-1 1.

534

PALAEONTOLOGY, VOLUME 11

fourcade, e. 1966. Murciella cuvillieri n. gen. n. sp. nouveau foraminifere du Senonien superieur du
sud-est de EEspagne. Revue Micropaleont. 9, 147-55.
reichel, m. 1936, 1937. Etude sur les Alveolines. Mem. Soc. Paleont. Suisse, 57, 59, 1-147, pi. 1-1 1.
reiss, z. hamaoui, m. and ecker, a. 1964. Pseudedomia from Israel. Micropcdeont. 10, 431-37, pi. 1-2.
seiglie, g. a. and ayala-castanares, a. 1 963. Systematica y biostratigrafia de los foraminiferos grandes
del Cretacico superior (Campaniano y Maastrichtiano) de Cuba. Paleont. Mex. 13, 1-56, pi. 1-43.
smout, a. h. 1963. The genus Pseudedomia and its phyletic relationships, with remarks on Orbitolites
and other complex foraminifera. In Evolutionary Trends in Foraminifera. New York, Elsevier.
224-81, pi. 1-4.

vaughan, t. w. 1929. Additional new species of Tertiary larger foraminifera from Jamaica. J. Paleont.
3, 373-82, pi. 39-41.

ZANS, V. A., CHUBB, L. J., VERSEY, H. R., WILLIAMS, J. B., ROBINSON, E., and COOKE, D. L. 1962. Synopsis
of the geology of Jamaica. Geol. Surv. Jamaica Bull. 4, 1-72.

E. ROBINSON

Department of Geology
University of the West Indies
Mona

Typescript received from author 1 March 1968 Jamaica

FAMENNI AN AMMONOIDS FROM NEW SOUTH

WALES

by T. B. H. JENKINS

Abstract. Seven species and two subspecies belonging to four genera of ammonoids (Clymeniina and Goniati-
tina) are described from a Famennian horizon, Platyclymenia Zone, a few hundred feet below the erosive base
of the Carboniferous in the vicinity of Keepit Dam, near Tamworth, New South Wales. Three new species of
Clymeniina are described: Genuclymenia keepitensis, Platyclymenia (Platyclymenia) teicherti, and P. (P.) alterna.

Only in recent years have ammonoid cephalopods been reported in the Upper Devonian
strata of eastern Australia. As Teichert remarked in 1948, the earlier records of am-
monoids from those strata are probably erroneous. The first acceptable record of such
fossils is that of Pickett (1960) who described a new species of clymenid from the Borah
Limestone in northern New South Wales. Lately the index goniatite Cheiloeeras has
been recorded by the present writer and reference has been made to the fauna now
described, both occurrences being in the same northern New South Wales province.

The stratigraphic horizons of these ammonoids have been previously indicated in
a column for the area near Keepit Dam (Jenkins 1966, text-fig. 2), the horizon of the
fossils now described being about 3,000 ft. higher, in an expanded, largely clastic
succession, than that of the Cheiloeeras occurrence. The persistent Borah Limestone
seems to be absent in the belt of country running meridionally through Keepit Dam
which has yielded the later ammonoid finds; its absence there is presumably due to the
disconformity between the Devonian and Carboniferous systems. The inferred position
of the Borah Limestone is thus an unmeasured distance, possibly as much as a thousand
feet, above that of the fauna now described. The general geology of the surrounding
country has been described by Voisey and Williams (1964).

All the fossils here described come from a 25 ft. lens of clayey coarse sandstone,
containing patches of shale breccia, in the basal portion of the Mandowa Mudstone
outcropping in Spring Gully at a point 3-13 miles N. 24° W. of Keepit Dam and TO
miles W. 27° S. of Gartmore Homestead. It is 1-5 miles N. of the Cheiloeeras spot
shown in text-fig. 1 of Jenkins (1966), the scale of which locality map should be corrected
to 1 in. = 0-7 mile.

Recent discoveries of Late Devonian ammonoids in northern New South Wales
provide a few firm correlations with other areas but the sequence of faunas is as yet
only very imperfectly known in eastern Australia, in marked contrast to the position
in western Australia, where ammonoids are now known from each of the major divisions,
Manticoceras-Stufe to Wocklumeria-Stufe. (Glenister and Klapper 1966).

Described specimens are in the fossil collections of The University of Sydney Geo-
logical Department and are referred to by USGD catalogue numbers.

SYSTEMATIC DESCRIPTIONS

Suborder clymeniina Hyatt 1884
Superfamily clymeniaceae Edwards 1849

[Palaeontology, Vol. 11, Part 4, 1968, pp. 535-48, pis. 104-5.]

536

PALAEONTOLOGY, VOLUME 11

Family cyrtoclymeniidae Hyatt 1884
Genus genuclymenia Wedekind 1908

Type species. Clymenia frechi Wedekind 1908, p. 617, by subsequent designation of Schindewolf 1957.

In its sutural character this genus is intermediate between the two other genera,
Cyrtoclymenia and Cymaclymenia, included in the Cyrtoclymeniidae by Schindewolf,
(in Moore, 1957). Genuclymenia is previously recorded from Germany and the Urals
(U.S.S.R.).

Genuclymenia keepitensis sp. nov.

Plate 104, figs. 1-4; text-fig. 1 g, h

Derivation of name. From Keepit Dam, near Gunnedah, N.S.W.

Holotype. USGD 6827.

Preservation. The holotype is an internal mould in clayey sandstone; three other specimens are simi-
larly preserved.

Dimensions (in mm.)

D

WW

WH

UW

Holotype USGD 6827

16-8

c.4-5

6-8

c. 5-5

USGD 6830

16-3

4+

6-9

5-7

Diagnosis. Genuclymenia with lateral lobe separated from umbilical seam by short
straight radial segment of suture, flatly rounded whorl sides converging from maximum
width near mid-flanks to a narrowly rounded venter.

Description. Shell form is subdiscoidal and rather narrowly umbilicate with whorls
deeply overlapping; in cross-section the whorls are compressed with flatly rounded
sides having maximum width at about mid-height and converging gradually to a narrowly
rounded venter. Low ribs on part of one mould and faint growth-lines on its counterpart
seem to mark a shallow lateral sinus passing into a broad ventro-lateral salient.

The external suture line runs radially from the umbilical seam and passes abruptly
into a semicircular lateral lobe which is separated from the flat ventral saddle by a
shallow lobe. The internal suture has not been observed.

EXPLANATION OF PLATE 104

Figs. 1-4. Genuclymenia keepitensis sp. nov. 1, internal mould, USGD 6829, x4; 2, internal mould,
holotype, USGD 6827, x3; 3, a partial internal mould, uncoated, showing suture and low ribs,
USGD 6830, x 2; 4, plasticine mould from counterpart of USGD 6830, showing low growth ridges,
X 2.

Figs. 5, 6. P. (Platy clymenia) annulata annulata (Munster). 5, internal mould, USGD 6834, X 1 ; 6,
fragment of an internal mould, USGD 6836, X 2.

Figs. 7, 8. P. (Platy clymenia) annulata densicosta Freeh. 7, internal mould, USGD 6837, x2; 8,
external mould, USGD 6880, x 2.

Figs. 9-13. P. ( Platy clymenia ) teicherti sp. nov. 9, crushed specimen preserving fine growth-lines,
syntype, USGD 6839, X 1; 10, internal mould, (subsequently sectioned for text-fig. 2 e, f),
syntype, USGD 6891, x2; 11, internal mould (see also text-fig. 2 a-d), syntype, USGD 6890,
x2; 12, internal mould showing septal face of body chamber, USGD 6851, xfl; 13, partly crushed
internal mould showing portions of three septal faces, USGD 6850, x 1J.

Palaeontology, Vol. 11

PLATE 104

JENKINS, Famennian ammonoids from New South Wales

JENKINS: FAMENNIAN AMMONOIDS FROM NEW SOUTH WALES 537

Remarks. Previously figured sutures of species assigned to Genuclymenia (see text-fig. 1 )
show an umbilical saddle passing immediately into a rounded lateral lobe. In G.
keepitensis the lateral lobe is separated from the umbilical seam by a nearly straight
radial portion of the suture which could be considered a flat saddle. Wide separation of
lateral lobe and umbilical seam characterizes the related Cymaclymenia which has a
broad saddle in this position. In the character of its dorso-lateral suture, the new species
is thus intermediate between Cymaclymenia and other species of Genuclymenia. Of
figured specimens it is closest to Cymaclymenia pseudogoniatites (Sandberger 1853,
pi. 7, fig. 4 (others excluded)) which shows a slightly asymmetrical and incipiently
acuminated lateral lobe (text-fig. 1 i herein).

In shell form G. keepitensis is distinguished from Wedekind’s species G. frechi and
G. discoidalis by the absence of any flattening along its venter and from G. angelini by
its more compressed whorls. G. karpinskii Perna has wider whorls (WW/WH = 0-81
at D = 25 mm.) and straighter lateral growth-lines meeting the umbilical seam more
orthogonally than in keepitensis. G. bond Schindewolf is closest in shell shape but has
maximum whorl width near the umbilical shoulder and lacks even the low ribs seen in
some of the Keepit specimens.

Occurrence. According to Schindewolf (in Moore 1957) Genuclymenia is confined to the
Platyclymenia zone of the Famennian, i.e. to Stufen III and IV of Wedekind (1913).
G. frechi was recorded at Enkeberge by Wedekind (1908, table after p. 634) from beds
that were later termed Stufen IV and V. But these tabulated records do not appear in
the text, were not repeated by Wedekind (1914) and were not confirmed by Lange’s
(1929, p. 12 n., p. 88) later work in the same area. Lange’s observations agreed with
Wedekind’s (1914) revision in restricting Genuclymenia to 1 1 1/3, wherein it is abundant.
Schindewolf’s record of G. frechi from Ilia at Gattendorf is the only authentic occur-
rence of the genus known to the writer outside III/). Schindewolf’s (1923, p. 435)
record of G. dunkeri (Munster) from Va at Gattendorf refers to the type species of
Protoxyclymenia Schindewolf 1922.

Family clymeniidae Edwards 1849

Genus platyclymenia Hyatt 1884

Type species. By original designation Goniatites annulatus Munster 1832.

This genus receives the species of widely umbilicate clymenids having growth-lines
which are concave on the lateral areas and simple sutures consisting of a broad ventral
saddle passing evenly through a broadly rounded lateral lobe to a sharper, often angular
umbilical saddle and a moderately deep, undivided dorsal lobe.

The genus is currently subdivided by Schindewolf (1 934, 1957 in Moore) and Bogoslovsky
(1962) into four subgenera: P. (Pleuroclymenia),P.(Trigonoclymenia),P .(Spinoclymenia),
and P. ( Platyclymenia ), of which only the last is represented in my material. Species
have been defined mainly on whorl section and ornamentation. It is well established
that both characters change during the ontogeny of many included species. Specimens
which combine the features of several nominal species are common ( vide Lange 1929,
pp. 93-6).

G. frechi

G. karpinskii

G. keepitensis

C. pseudogoniatites

text-fig 1 . Comparison of G. keepitensis sp. nov. with other species.

a, b , G. frechi Wedekind, the suture at maturity, from Schindewolf 1957 (text-fig. 40, 3b) and whorl
section from Wedekind 1914, pi. 1, fig. 76.

c, d, G. angelini Wedekind, the suture from Perna 1914 (text-fig. 81) and whorl section from Wedekind
1914 (pi. 1, fig. 6b).

e, f G. karpinskii (Perna), suture and whorl section from Perna 1914 (text-fig. 86 a, and pi. 3, fig. 18tf,
respectively).

g, h, G. keepitensis sp. nov., suture X 6, whorl section x4, based on holotype, USGD 6827.

j, C. pseudogoniatites (Sandberger), suture and whorl section from Sandberger (1853, pi. 7, figs.
4b, 4a, respectively).

JENKINS: FAMENNIAN AMMONOIDS FROM NEW SOUTH WALES

539

Platyclymenia is widespread in the northern hemisphere, being found in North
America (U.S.A., Canada?), Europe, North Africa, and Asia (Kazakstan, China). It is
also recorded from Western Australia (Teichert 1941) and a doubtful specimen from
New South Wales has been figured by Pickett (1960).

Platyclymenia ( Platyclymenia ) annulata (Munster)

Neotype. By designation of Wedekind 1914: Clymenia annulata Munster; Giimbel 1863, pi. 15, fig. 12.

Diagnosis. A species of Platyclymenia with strong, sharp, unpaired, and clearly separa-
ted ribs on the flat or arched whorl sides; ribs absent or greatly reduced in later growth
stages.

Remarks. Interpretation of this, the type species, was stabilized by Wedekind’s selection
of a neotype and clarified by Schindewolf’s (1923) redefinition. Lange has discussed
the morphological variation on the basis of material from around Enkeberg, close to
the locality of the neotype and has described intergradations with P. ( P .) richteri
Wedekind which seemingly would be better regarded as a subspecies of annulata.
Freeh’s (1902) annulata var. densicosta and Wedekind’s (1914) richteri var. densicosta
as well as Petter’s (1960) annulata var. semperornata belong within the redefined species,
but Perna’s (1914) var. correct a and var. rustica are excluded.

Occurrence. Oberdevonstufe IV of Wedekind, more usually termed nowadays the
annulata Zone, which is the uppermost division of the Platyclymenia- Stufe.

1832
1834
1863
1910
[non] 1914
1914
1923
1929
1956
1960

Platyclymenia ( Platyclymenia ) annulata annulata (Munster)

Plate 104, figs. 5, 6

Goniatites annulatus Munster, p. 3, pi. 6, fig. 6 [fide auctt.].

Goniatites annulatus Munster; Munster, p. 95, pi. 6, fig. 6.

Clymenia annulata Munster; Giimbel, p. 130, pi. 15, fig. 12 [non figs. 11, 13].

Clymenia annulata Munster; Rzehak, p. 169, pi. 2, fig. 1 [non figs. 2-5].

Clymenia annulata Miinst. ; Perna, p. 75.

Platyclymenia annulata (Munster) Giimbel; Wedekind, p. 35.

Platyclymenia annulata Giimb. red.; Schindewolf, p. 447, pi. 17, figs. 7, 8.
Platyclymenia annulata Giimb. red. Schind.; Lange, p. 107.

Platyclymenia ( Platyclymenia ) annulata (Miinster 1832); Muller, p. 70.

Platvclvmenia annulata { Munster) 1832 red. Giimbel ; Petter, p. 23, text-fig. 3e, pi. 4,
figs. 4, 7, 9, 10.

Diagnosis. A subspecies of P. ( P .) annulata having 26 or fewer ribs per whorl in the
coarsely ribbed portion of the shell and a subrectangular or trapezoidal whorl section.

Material. Two specimens, one a fragment of an internal mould, the other two whorls of an internal
mould and its external mould counterpart, preserved in a clayey sandstone.

Dimensions (in mm.)

D

63

N n

C 5934

USGD 6834 (internal mould)

WW

WH
c 1 5

UW

38

540

PALAEONTOLOGY, VOLUME 1 1

Description. The more complete specimen is large (maximum diameter over 88 mm.),
widely evolute with a rounded rectangular whorl section in its strongly ribbed stage.
The ribs are in the form of sharp ridges, radial to slightly prorsiradiate and weakly
concave towards the aperture; they increase in height and width, but maintain their
sharpness, towards the ventro-lateral shoulder and there flatten out abruptly, thus
seemingly accentuating the sub-rectangular whorl shape. There are 21 ribs in the whorl
ending at WH = 9-5 mm. Adjacent ribs are separated by completely smooth sections
which are twice to four times the width of a rib. In the next whorl, ending with
WH = c .15 mm., the ribs are much less sharp, about twice as numerous and are highest
towards the umbilical shoulder. The flank of the last preserved whorl, on the external
mould, is smooth except for low radially elongate swellings which diminish in definition
and elevation towards the aperture.

Part of one suture is preserved within the strongly ribbed stage; it shows the shallow,
evenly rounded, lateral lobe which is characteristic of the genus.

Remarks. There is close correspondence with P. annulata red. Schindewolf and especially
with a large specimen carefully described by Lange (1929, p. Ill) under the name
Platyclymenia richteri , but which he regarded as possibly belonging to P. annulata,
presumably because the whorl section was not clearly observable.

Platyclymenia ( Platyclymenia ) annulata densicosta Freeh
Plate 104, figs. 7, 8

1863 Clymenia annulata Miinst. Gtimbel, p. 130, pi. 15, fig. 13, 13a [non figs. 11, 12].

1902 Clymenia annulata var. densicosta Freeh, p. 31, pi. 1, fig. 7.

1914 Platyclymenia annulata var. densicosta Freeh emend.; Wedekind, p. 36, pi. 3, fig. 2.
1914 Platyclymenia richteri var. densicosta Wedekind, p. 35.

1923 Platyclymenia annulata var. densicosta Freeh; Schindewolf, p. 449.

1929 Platyclymenia annulata var. densicosta Freeh; Lange, p. 108.

1960 Platyclymenia annulata var. densicosta Freeh emend. Wedekind; Petter, p. 24, pi. 4,
fig. 5; pi. 5, figs. 13, 23.

1960 Platyclymenia richteri var. semperornata Petter, p. 25, pi. 4, figs. 6, 1 1.

Diagnosis. A subspecies of P. (P.) annulata having 27 or more ribs per whorl in the
coarsely ribbed portion of the shell.

Material. Two specimens preserved in clayey sandstone, one a fragment of an external mould and the
other an internal mould of two inner whorls.

Dimensions (in mm.)

USGD 6837 (internal mould)

D WW WH UW

21-7 c.l 6-9 12-6

Description. The more complete specimen is a widely umbilicate small mould of approxi-
mately circular whorl section. Fourteen ribs in the last-preserved half whorl are sharp
and are highest and broadest towards the venter, but abruptly end in a slightly pror-
siradiate attitude to leave the venter unnbbed and rounded. The inner of the two
preserved whorls has somewhat more numerous but less sharp ribs.

The second specimen (USGD 6880) is a fragment of an external mould showing ten
ribs of the annulata pattern in about one-third of a whorl.

JENKINS: FAMENNIAN AMMONOIDS FROM NEW SOUTH WALES

541

Remarks. The close similarity of P. ( P .) annulata and P. (P.) richteri in features other
than whorl shape, and the existence of specimens of intermediate whorl section has been
pointed out by Lange and Petter. Each species has a named variety distinguished from
its type by more numerous ribs : annulata var. densicosta Freeh and rieliteri var. densicosta
Wedekind. In view of their similarity, intergradation, corresponding varieties and ranges
I consider that annulata and richteri belong to one species and, further, that the more
closely ribbed nominal varieties constitute a single taxon of subspecific rank which
includes also Petter’s var. semper ornata.

In publishing the name var. densicosta Freeh cited two specimens, one fragmentary
(his pi. 1, fig. 7) and the other GumbePs plate 15, fig. 15 (sic), which, as there is no
fig. 15 or 14 on Gumbel's plate 15, must be read plate 15, fig. 13, and then corresponds
with the text. The latter is a specimen of five whorls ending at WH = 8-5 mm, with
28 ribs in the last preserved whorl. I hereby designate it as lectotype of P. (P.) annulata
densicosta Freeh.

Platyclymenia (Platyclymenia) teicherti sp. nov.

Plate 104, figs. 9-13; Plate 105, figs. 1-4; text-fig. 2a-j

Derivation of name. For Professor Curt Teichert of the University of Kansas Paleontological Institute.
Syntypes. USGD. 6839, 6890, 6891.

Material and preservation. About forty more or less complete shells in clayey sandstone, crushed and
decalcified to varying degrees.

Dimensions {in nun.)

D

WW

WH

UW

USGD 6839

c. 42

(crushed)

13

21

USGD 6890

25

c. 5-5

6-9

130

USGD 6891

23

c. 5

5-6

13

Diagnosis. A species of Platyclymenia which is unribbed and very widely evolute, with
an umbilical diameter equal to or greater than one-half the total diameter, i.e.
UWx2 > D.

Description. The smooth widely umbilicate shells have whorls which vary in cross-section
between sub-circular, oval and sub-rectangular; impressed area is always shallow
(text-fig. 2).

Rarely preserved growth-lines are almost rectiradiate faint undulations showing
a weak lateral sinus (PI. 104, fig. 9); they have not been observed on the venter.

Suture lines consist of an almost flat ventral saddle, broad, evenly rounded lateral
lobes, moderately sharp umbilical saddles, and a widely open dorsal lobe. The
umbilical saddle is usually crested forward of the ventral saddle; in several specimens
this forward projection of the umbilical saddles is accentuated in late growth stages
(text-fig. 2b). One sectioned specimen shows septa near the dorsum to be strongly
deflected forward to meet the next succeeding septum, and apparently failing to form
complete transverse partitions; this curious configuration may possibly have resulted
from distortion. In some specimens (e.g. USGD 6846) the curvature of the lateral
lobe is localized on the mid-flank, forming a very broad V.

542

PALAEONTOLOGY, VOLUME 11

Septal necks are short.

Late growth stages usually show internal constrictions numbering up to 5 per whorl.

text-fig. 2. Comparison of P. (P.) teicherti sp. nov. with P. (P.) pattisoni (M‘Coy) and Clymenia
laevigata (Munster).

a-j, P. (P.) teicherti sp. nov. x4. a-d, based on USGD 6890; a, whorl section at D = 24-3 mm.;
b , the last suture, at D = 24 mm.; last but one suture, at D = 23-5 mm.; last but four suture at
D = c. 20-5 mm. ; e,f based on USGD 6891 ; e, whorl section,/, external suture, both at D = c. 19 mm. ;
g, h, based on USGD 6843 ; g, whorl section, h, suture, both at D = c. 15-5 mm. ; i,j, based on USGD
6862; /, whorl section,/, suture, both at D = 1 6-9 mm.;

k-m, sutures of C. laevigata, k, copied from Sandberger, 1853, pi. 7, fig. 1 e-f; I, copied from Gumbel,
1863, pi. 16, fig. 5c, ‘Munster’s Original’; m, copied from Bogoslovsky et al. in Orlov 1962, fig. 187u;

n, suture of C. placida Perna, copied from Perna 1914, fig. 79.

o, whorl height plotted against diameter for P. (Pi) teicherti and P. (P.) pattisoni.

p, umbilical width plotted against diameter for the same species. Specimens of P. (P.) teicherti represen-
ted in these diagrams are indicated by USGD numbers at the left margin. Represented specimens ofP.
(P.) pattisoni are indexed by letters a-g; a, holotype of pattisoni , dimensions from Selwood 1960, p. 165;
b, c, P. placida from Perna 1914, p. 77; c, P. subnautilina from Petter 1960, p. 27; c1 from Petter 1960,
pi. 5, fig. 20; d, e, P. quenstedti, from Wedekind 1914, p. 45; f, P. subnautilina schleizi from Muller,
1956, p. 74; G, C. subnautilina from Sandberger 1855, pi. 1, fig. If.

JENKINS: FAMENNIAN AM MONOIDS FROM NEW SOUTH WALES 543

They are rectiradiate to rursiradiate, straight or with a slight forward concavity, and
are formed by wave-like thickenings which may be symmetrical or have crests located
forward of their mid-line. Constrictions are not detectable externally.

Remarks. Smooth species of Platyclymenia are not always readily distinguished from
species of Clymenia. Wedekind’s criterion of relative straightness of growth-lines in
Clymenia is unsatisfactory since they are rarely preserved in either group and may be
nearly straight in late growth stages of Platyclymenia (Schindewolf 1923, p. 462;
Muller 1956, p. 74). More practical are the criteria proposed by Schmidt (1924, p. 124)
based on sutural and septal form: the wider dorsal lobe and shorter septal necks of P.
subnautilina in comparison with C. laevigata. Suture diagrams showing dorsal lobes
which are about as deep as they are wide have been figured also for several other species
of P. ( Platyclymenia ), (e.g. Perna 1914, figs. 71, 72, 77, 79 and Schindewolf 1934, hg. 12);
in this characteristic they contrast sharply with the narrow parallel-sided dorsal lobe of
Clymenia laevigata (e.g. Sandberger 1853, pi. 6, hg. 6; pi. 7, hg. 1 ; Gtimbel 1863, pi. 16,
hg. 5c; Freeh 1902, text-hg. 4b). The relatively wide dorsal lobes and the short septal
necks of the new species therefore assign it to Platyclymenia.

Nominal species of Platyclymenia having oval or round cross-section and lacking
ornamentation have been justihably synonymized by Selwood (1960) under the name
P. ( P .) pattisoni M’Coy 1851 who also rehgured the holotype (pi. 26, hg. 10, not hg. 11
as stated in text). Perna’s species Clymenia placida also belongs, at least in part, to
P. ( P .) pattisoni (cf. Schindewolf 1922, p. 124). But the new species described above
differs from pattisoni in being more widely umbilicate and in its slower rate of increase
of whorl height (text-hg. 2 o and p). In these characters of general form, and in the
presence of constrictions, P. (P.) teicherti resembles C. laevigata, but in the present
material it does not reach the size of the latter species. It could be regarded as a possible
ancestor of Clymenia.

Occurrence. Specimens of P. (P.) teicherti outnumber the total of all the other species
in the fauna here described.

Platyclymenia ( Platyclymenia ) alterna sp. nov.

Plate 105, figs. 8-1 1

Derivation of name. Latin alternus, alternating, referring to the ribs.

Holotype. USGD 6838, the only specimen found of this species.

Preservation. Internal mould of segment of one whorl and the more complete external mould of two
whorls, preserved in clayey sandstone.

Dimensions (in mm.)

D WW WH IJW

Holotype, USGD 6838 23-5 4-5 c. 7 11

Diagnosis. A species of Platyclymenia with intercalated ribs and a subrectangular whorl
section.

Description. The holotype is widely umbilicate with a high subrectangular whorl section.
It is slightly distorted and WW as measured may be somewhat reduced thereby. Whorls

544

PALAEONTOLOGY, VOLUME 11

overlap only slightly, the impressed area being about 1 mm. deep at WH = 7 mm.
Whorl sides are parallel and the venter is weakly arched.

Ribs are concave, projected and in two alternating sets, both ending at the ventro-
lateral shoulder to leave the venter smooth. The ribs of the more prominent set are
usually continuous across the flanks; they increase in height and width towards the
venter and, to a lesser extent, towards the umbilical seam, so that the ribs are relatively
weak in the middle portion of the whorl. Intercalated ribs are usually confined to the
outer half of the whorl, but may occasionally appear weakly near the umbilical seam.
Crests of adjacent ribs are about 1-5 mm. apart at WH = 7, the accentuated rib termina-
tions being there about 0-5 mm. wide.

Remarks. Two other specimens from the type locality having high subrectangular whorl
sections and umbonally accentuated ribs may belong to this species (USGD 6835 and
6881); both represent a growth stage which is later than is preserved in the holotype and
so cannot be confidently identified until more complete material becomes available.

In whorl shape the species resembles P. ( P .) sandbergeri and P. (P.) walcotti, but the
intercalated ribs of a/terna are distinctive.

Family rectoclymeniidae Schindewolf 1923
Genus rectoclymenia Wedekind 1908

Type species. Rectoclymenia roemeri Wedekind 1908 by subsequent designation of Schindewolf 1957.

Rectoclymenia ? sp.

Plate 105, figs. 15-18

Description and remarks. Together with the other described species occur numerous
fragments of a large coiled cephalopod. Its siphuncle has not been observed; the simple
partial suture preserved on one fragment, the straightness of the low radial undulations
on the whorl flanks and the general form suggest Rectoclymenia.

A quite exceptionally large size is indicated. One short portion of a whorl which is
incomplete ventrally measures 74 mm. in the radial direction, has a whorl thickness

EXPLANATION OF PLATE 105

Figs. 1-4. P. {Platyclymenia) teicherti sp. nov. 1, 2, internal mould, USGD 6862, x2; 3, 4, internal
mould, USGD 6843, X 2. 2, 4 are dorsal views of the detached incomplete body chambers showing
septal face and internal suture.

Figs. 5-7. Sporadoceras cf. rotundum Wedekind. 5, internal mould, USGD 6831, x3; 6, 7, internal
mould, USGD 6832, x3.

Figs. 8-1 1. P. ( Platyclymenia ) a/terna sp. nov. 8, 9, external and internal moulds of holotype USGD
6832, x2; 10, 11, the same, X 1.

Figs. 12-14. Sporadoceras inflexion Wedekind. 12, part of impressed area of body whorl showing
spiral striae, x3. 13, 14, lateral and ventral views, the latter showing the deformed venter, due to
crushing; USGD 6866, X 1.

Figs. 15-18 . Rectoclymenia ? sp. 15, small internal mould showing several partial sutures, USGD 6853,
x 2; 16, part of an external mould of an outer whorl photographed to show low rectiradiate undula-
tions, USGD 6899, xf; 17, portion of an outer whorl, USGD 6865, xf; 18, USGD 6864, X 1.

Palaeontology, Vol. 11

PLATE 105

JENKINS, Famennian ammonoids from New South Wales

JENKINS: FAMENNIAN AMMONOIDS FROM NEW SOUTH WALES 545

(internal mould) of 24 mm. and an impressed area 12-5 mm. deep; its umbilical margin
forms an arc which approximately fits a radius of 70 mm. A diameter of about 300 mm.
is thus indicated.

Whorl section in late growth stages is compressed high oval, the very gently arched
sides converging ventrally from the inner flanks to an evenly rounded unkeeled venter.
Sculpture seemingly changes with growth from rectiradiate ribs, evenly rounded in
profile and strongest towards the umbilicus, to low dorso-lateral bullae which pass
radially into low rectiradiate swellings at a later growth stage. Faint radial growth lines
parallel the ribs and also indicate a probable ventral sinus.

Three small unribbed specimens may represent an early growth stage. One shows
simple sutures with shallow rounded lateral lobes (PI. 105, fig. 15), is moderately evolute,
and has the following dimensions (in mm.).

R WH WW UW
USGD 6853 12 6-7 c. 2 10

(WW may be reduced by compressive distortion).

Closest of the described species of Rectoclymenia in general form to these small
specimens is R. acuta (Schmidt), but none is recorded as attaining the exceptionally large
size indicated for the material described above.

According to Schindewolf (in Moore 1957) the genus is confined to the Platyclymenia
Zone. It is previously recorded from Europe, Asia, and North America.

Suborder goniatitina Hyatt 1884
Superfamily cheilocerataceae Freeh 1897
Family cheiloceratidae Freeh 1897
Genus sporadoceras Hyatt 1884

Type species. Goniatites bidens G. and F. Sandberger by original designation.

Sporadoceras inflexion Wedekind
Plate 105, figs. 12-14; text-fig. 3 a, b

1908 Sporadoceras inflexion Wedekind, p. 595, pi. 39, fig. 43; pi. 42, figs. 3, 2>a.

[non] 1914 Sporadoceras inflexion Wedekind; Perna, p. 36, 98, pi. 1, fig. 14; text-fig. 21 [= A.
muensteri var brachyloba Frecfi]

1918 Sporadoceras inflexion Wedekind; Wedekind, p. 149, text-fig. 47 d.

1959 Sporadoceras inflexion Wedekind; Petter, p. 274.

Material. One specimen (USGD 6866), an internal mould, somewhat crushed.

Diagnosis. A species of Sporadoceras in which the second lateral lobe is pointed and is
two or three times as wide as the first lateral lobe which is shorter and skewed towards
the venter.

Dimensions {in mm.)

D WW WH UW
USGD 6866 (internal mould) 45-8 11+ 16 nil

546

PALAEONTOLOGY, VOLUME 11

Description. The mould shows a closed umbilicus, the whorls being completely over-
lapping; the body chamber extends through three-quarters of a whorl. Whorl section is
interpreted as having been approximately oval with an evenly rounded venter; crushing
has resulted in a narrowed venter but the original rounded venter is preserved at some
points.

A detached fragment of the body chamber preserves a sector of its impressed area

which clearly shows fine and regular spiral striae
(about 30 per cm.).

The suture is characteristic of the genus (text-fig.
3 a, b) with rounded saddles and more or less pointed
lobes. The inner lateral saddle is pointed and asym-
metrical with a straighter outer limb; the outer
lateral saddle is shorter, deep and narrow, parallel
sided, ending bluntly or sub-acutely with a slight
outward direction.

Remarks. The specimen described resembles hetero-
lobatum Lange but its first lateral lobe lacks the
consistent sharpness shown in Lange’s species, and its
first lateral saddle (E/A1L) is relatively higher. The
ventral deflection of the first lateral lobe in my speci-
men, is less than in Wedekind’s figure of the type of
S. inflexion , but this is judged to have been possibly
diminished by the partial crushing which has con-
verted an originally rounded venter to an acute form.
S. muensteri var. brachyloba Freeh and S. pseudo-
carinatum Petter are somewhat similar to my specimen in shell-form and suture but have
a relatively wider first lateral lobe.

Sporadoceras inflexion, is previously recorded from 111/3 in Germany and from an
unrecorded locality and horizon in North Africa (Petter 1959).

Sporadoceras cf. rotundum Wedekind
Plate 105, figs. 5-7; text-fig. 3c

1908 Sporadoceras rotundum Wedekind, p. 594, pi. 39, fig. 21 ; pi. 42, fig. 1.

1918 Sporadoceras rotundum Wedekind; Wedekind, p. 148, text-fig. 47c.

Two similar small distorted specimens of Sporadoceras are comparable with S.
rotundum in form and suture-line. The widely rounded venter and the dorso-ventral
direction of distortion by compaction indicate an originally sub-sphaerical form. A
partial external suture (text-fig. 3c) on the smaller specimen shows a deep and pointed
second lateral lobe and a shallow rounded first lateral; the umbilical sector is indistinct
but probably as shown by the dashed line, whereas the ventral sector is not observable.
Both specimens show three constrictions about 90° apart; they are rectiradiate on the
flanks and show a constant shallow rounded ventral sinus. Sporadoceras biferum is
similar in its suture-line and has a named variety ( sulcifera Lange) distinguished by

text-fig. 3. Sutures of Sporadoceras.
a, b, S. inflexum Wedekind, based on
USGD 6866 at D = c. 47 ; x 2. c, 5.
cf. rotundum Wedekind, based on
USGD 6832 at D = c. 9; x4-25.

JENKINS: FAMENNIAN AMMONOIDS FROM NEW SOUTH WALES

547

constrictions, but its shell-form is more compressed than that indicated for the specimens
here recorded.

S. rotundum is recorded from Zones Ilia and III/3 in Germany and the lower Prolobites
Zone (= Ilia) in the Urals. 5. biferum is recorded from Zones II/3-I1I/3 in Germany and
Zones II-IV in the Sahara.

Acknowledgements. For discussions and information the writer thanks Professors B. F. Glenister and
W. M. Furnish of Iowa State University, Iowa City, U.S.A., Professor M. R. Flouse of the University
of Hull and Dr. E. B. Selwood of the University of Exeter, England. Financial assistance from the
University of Sydney Research Grant is also acknowledged.

REFERENCES

Bogoslovsky, b. i. 1962. Rare type of ornamentation in clymenids. Paleont. Zh. 1962 (1), 166-8.
[In Russian.]

frech, F. 1902. Uber devonische Ammoneen. Beitr. Paldont. Geol. Ost.-Ung. 14, 27-1 1 1, pi. 2-5.
glenister, b. f. and klapper, g. 1966. Upper Devonian conodonts from the Canning Basin, Western
Australia. J. Paleont. 40, 777-842, pi. 85-96.

gumbel, c. w. 1863. Ueber Clymenien in den Uebergangsbilden des Fichtelgebirges. Palaeontographica,
11, 85-165, pi. 15-21.

hyatt, a. 1883-4. Genera of fossil cephalopods. Proc. Boston Soc. nat. Hist. 22, 253-338.
jenkins, t. b. h. 1966. The Upper Devonian index ammonoid Cheiloceras from New South Wales.
Palaeontology, 9, 458-63, pi. 72.

lange, W. 1929. Zur Kenntnis des Oberdevons am Enkeberge und bei Balve (Sauerland). Abh. preufi.

geol. Landesanst., n.f. 119, 1-132, pi. 1-3.
m‘coy, f. 1851. On some new Devonian Fossils. Ann. Mag. Nat. Hist. (2) 8, 481-9.
moore, R. c. (ed.) 1957. Treatise on Invertebrate Paleontology, Part L — Mollusca 4, Cephalopoda,
Ammonoidea, xxii + 490. Geol. Soc. Amer. and Univ. Kansas Press.
muller, k. j. 1956. Cephalopodenfauna und Stratigraphie des Oberdevons von Schleiz und Zeulen-
roda in Thiiringen. Beih. Geol. Jb. 20, 1-93, pi. 1 and 2.
munster, G. Graf zu. 1832. Uber die Planiditen und Goniatiten ini Ubergangskalk des Fichtelgebirges.
1-38, pi. 1-6. Bayreuth.

1834. Memoire sur les Clymenes et les Goniatites du calcaire de transition du Fichtelgebirge.

Ann. Sci. naturelles, 2(2), Zool. 65-99, pi. 1-6.

(A French translation of the preceding item.)

orlov, y. a. (ed.) 1962. Principles of Palaeontology, Mollusca-Cephalopoda. 1. Moscow, 1-425, pi. 68.
[In Russian.]

perna, e. 1914. Die Ammoneen des oberenNeodevonsvom Ostabhang des Siidurals. Com. Geol.. Mem.

S.N. 99, 1-114, pi. 1-4. [In Russian, German summary.]
petter, g. 1959. Goniatites devoniennes du Sahara. Mem. Carte geol. Alger. Paleont. 2, 1-313, pi. 1-26.
1960. Clymenies du Sahara. Ibid. 6, 1-58, pi. 1-8.

pickett, J. W. 1960. A clymenid from the Wockhaneria zone of New South Wales. Palaeontology, 3,
237-41, pi. 41.

rzehak, a. 1910. Der Brunner Clymenienkalk. Zeit. mdhr. Landesmus. 10, 149-216, pi. 1-3.
sandberger, g. and f. 1850-6. Systematische Beschreibung and Abbildung der Versteinerungen des
rheinischen Schichtensystems in Nassau. 1-564, pi. 1-39 (Atlas). Wiesbaden.
sandberger, g. 1853. Einige Beobachtungen uber Clymenien; mit besonderer Riicksicht auf die west-
phalischen Arten. Verh. natarhist. Ver. preufi. Rhein. Westphal. 10, 171-216, pi. 6-8.

1855. Clymenia subnautilina n. sp., die erste und bis jetzt einzige Art aus Nassau. Jb. nassau. Ver.

Naturk. 10, 127-36, pi. 1.

schindewolf, o. h. 1922. Einige Randbemerkungen zu E. Perna's Abhandlung ‘Die Ammoneen des
oberen Neodevon vom Ostabhang des Siidurals’. Senckenbergiana, 4, 185-96.

548

PALAEONTOLOGY, VOLUME 11

schindewolf, o. h. 1923fl. Beitriige zur Kenntnis des Palaozoicums in Oberfranken. Ostthiiringen und
dem Sachsischen Vogtlande, 1, Stratigraphie und Ammoneenfauna des Oberdevons von Hof a. S.
Neues Jb. Min. Geol. Paldont. 49, 250-357, 393-509, pi. 14-18.

19236. Entwurf einer natiirlichen Systematik der Clymenoidea. Centralbl. Min., Geo!., Paldont.

1923, 23-50, 59-64.

1934. Uber eine oberdevonische Ammoneen-Fauna aus den Rocky Mountains. Neues Jb. Min.

Geol. Paldont. B. B. (B) 72, 331-50.

schmidt, h. 1924. Zwei Cephalopodenfaunen an der Devon-Carbongrenze im Sauerland. Jb. preufi.
geol. Landesanst. 44, 98-171, pi. 6-8.

selwood, e. b. 1960. Ainmonoids and Trilobites from the Upper Devonian and Lowest Carboniferous
of the Launceston Area of Cornwall. Palaeontology, 3, 153-85, pi. 26-29.
teichert, c. 1941. Upper Devonian Goniatite Successions of Western Australia. Am. Journ. Sci. 239,
148-53.

1948. Middle Devonian goniatites from the Buchan district, Victoria. J. Paleont. 22, 60-7, pi. 16.
voisey, a. h. and williams, k. l. 1964. The geology of the Carroll-Keepit-Rangari area of New South
Wales. J. Proc. Roy. Soc. N.S. W. 97, 65-72.

wedekind, r. 1908. Die Cephalopodenfauna des hoheren Oberdevon am Enkeberge. Neues Jb. Min.
Geol. Paldont. 26, 565-634, pi. 39-45.

1913. Die Goniatitenkalke des unteren Oberdevon von Martenberg bei Adorf. Ges. Naturf.

Freunde Berl. 1913, 23-77, pi. 4-7.

1914. Monographic der Clymenien des rheinischen Gebirges. Abh. Ges. Wiss. Gottingen, math.-

physik. Kl. n.f. 10, 1-73, pi. 1-7.

1918. Die Genera der Palaeoammonoidea (Goniatiten), (mit Ausschlu/? der Mimoceratidae,

Glyphioceratidae und Prolecanitidae). Palaeontographica, 62, 85-184, pi. 14-22.

T. B. H. JENKINS

Department of Geology and Geophysics
The University of Sydney
Sydney, N.S.W.

Typescript received 24 October 1967 Australia

A Q UILAPOLLENITES IN THE BRITISH ISLES

by A. R. H. MARTIN

Abstract. Species of Aquilapollenites Rouse emend. Funkhouser are present in the interbasaltic lignites of
Shiaba and Bremanoir, Mull, Scotland. The bearing of this on the age of the lignites is discussed; it is believed
that this may be pre -Eocene.

The genus Aquilapollenites Rouse (1957) emend. Funkhouser (1961) has been regarded
as an important indicator pollen fossil of the uppermost Cretaceous, Palaeocene, and
Lower Eocene of parts of the Northern Hemisphere, at least Western and Plains states
of North America and North-west and Central Siberia. Funkhouser (1961) and Stanley
(1961) refer to the lack of records of the genus from western Europe. However, at about
the same time, a posthumously published paper by J. B. Simpson (1961) described the
genus Taurocephalus as new and ascribed to it the single species T. proteus , which was
regarded as an extinct member of the Proteaceae. This very characteristic pollen is
clearly Aquilapollenites , but as the description differs somewhat from the published
description of Aquilapollenites , a brief explanation is offered.

Simpson apparently saw in polar view, the fine furrows which in Aquilapollenites
have been described as demicolpoids, but interpreted these as a proximal-face triradiate
mark such as he believed to occur in some Proteaceae. The two polar lobes of the grain
(PI. 106, fig. 6), superimposed one over the other, appear to have been taken for a
single distal pore resembling the three equatorial protrusions which were also described
as pores. These were correctly described as differing from all extant Proteaceae in that
ectexine covered the supposed pores. Simpson’s own photographs (1961, pi. xii, figs.
I, 2, and 5, for example) show that two polar protrusions are present. Taurocephalus
is described as sometimes having four equatorial arms; this is the only way of reconcil-
ing the appearance of Simpson’s figs. 1 and 5 with his description since these are equa-
torial views. However, his plate xii, fig. 1 1 is a specimen with four true equatorial arms
seen in polar view. The specimen has been re-located and re-photographed (PI. 106,
figs. 1-3). This feature had not been observed before in Aquilapollenites. Four-armed
grains are probably of low incidence.

In examining the material from Mull, in which Aquilapollenites occurs, a second
species and a few grains of a third were observed. Indeed Taurocephalus proteus is likely
to be a compound of two of these species (see taxonomic section). To find these in
sediments believed to be of Miocene or Pliocene age (the interbasaltic lignites of Shiaba
and Bremanoir) suggests that the dating of the lignites should be reconsidered. An age
between Maastrichtian and Lower Eocene seems most likely. The reliance on high
percentages of coryloid, quercoid, and Hamamelidaceae-type pollens seems misplaced
when it is realized that these pollens have a very extended time range. Manum (1962), for
example, reports many such grains in the Tertiary of Vestspitsbergen, which on zoo-
logical grounds is thought to be late Palaeocene to Eocene. British phytogeographers,
reluctant to accept the Mull interbasaltics as contemporary with the London Clay

[Palaeontology, Vol. II, Part 4, 1968, pp. 549-53, pis. 106.

550

PALAEONTOLOGY, VOLUME 11

flora, would probably be sympathetic towards the earlier rather than the later end of the
range in time indicated by AquilapoIIenites i.e. uppermost Cretaceous rather than Eocene.

Several other pollen grain types present in the interbasaltic beds are consistent with
the earlier dating here proposed. These include Classopollis sp. (PI. 106, figs. 15-16),
which though uncommon, does not show any obvious signs of redeposition and, in
lignite, is less likely to be re-deposited. The known range is Triassic to Eocene but it is
uncommon after the mid-Cretaceous. Haloragis bremanoirensis Simpson is conspecific
with Nudopollis thiergartii Pflug, an indicator of the Palaeocene in Germany, though
ranging from the Maastrichtian to Lower Eocene.

A search for AquilapoIIenites on slides, prepared by Simpson, of the pre-basaltic
lignite of Ardslignish, Ardnamurchan, failed to reveal any. All the slides quoted in the
paper of 1961 were re-examined. Most were in fair condition, with pollen grains still
stained by safranin, though some had developed small crystals in the glycerine jelly
mountant.

SYSTEMATIC DESCRIPTION
angiospermae Incertae sedis

Genus aquilapollenites Rouse emend. Funkhouser (April 1961)

Taxonomic synonyms. Parviprojectus Mtchedlishvili (July 1961); Taurocephalus Simpson (December
1961).

AquilapoIIenites spimdosus Funkhouser ( Taurocephalus proteus Simpson pro parte)

Plate 106, fig. 7

Sice range (10 grains). Polar axis 35-43 p (av. 38 p)\ tips of equatorial projections to polar axis 19-25 p
(av. 23 p); (ratio 0-5-0-67).

Location. Bremanoir and Shiaba interbasaltic lignites, Mull, Scotland.

Known time range. Palaeocene-Eocene (Funkhouser 1961); Maastrichtian (Srivastava
1966).

AquilapoIIenites attenuatus Funkhouser ( Taurocephalus proteus Simpson pro parte)

Plate 106, figs. 1-6

Size range (5 grains). Polar axis 28-42 p (av. 36 p)\ tips of equatorial projections to polar axis 21-24 p
(av. 23 p)\ (ratio 0-5-0-8). Four-armed grains are of rare occurrence.

Location. Bremanoir and Shiaba interbasaltic lignites. Mull, Scotland.

Known time range. Maastrichtian (Funkhouser 1961, Srivastava 1966).

Distinction. The characters held to distinguish A. spimdosus from A. attenuatus are the
shorter equatorial protrusions, the absence of puncta and the scattering of spinules over
the whole surface instead of being grouped. In the Scottish material, while it is possible
to find specimens corresponding to the two forms, there are a good many which combine
features of both (text-fig. 1). The term ‘punctate’ could be applied to 70% of the
Mull specimens, including both long and short-armed ones and to some with scattered

MARTIN: AQUILAPOLLENITES IN THE BRITISH ISLES 551

as well as those with grouped spinules. It is too early to suggest that a single polymorphic
form is represented, but clearly there is a need for further study of the variability shown
by these forms.

Ratio: Length of projection tip to polar axis
Length of polar axis

text-fig. E Number of Aquilapollenites spinulosus and A. attematus grains falling into different polar
length/equatorial radius classes. Crossed squares = punctate grains. R indicates grain with spinules

restricted to parts of surface.

Aquilapollenites pachypolus nom. nov.

Plate 106, figs. 8-14

Nomenclatural synonym. Parviprojectus striatus Mtchedlishvili 1961, pi. 73, fig. 1 a-c.
The genus Parviprojectus is a later homonym of Aquilapollenites as emended by Funk-
houser and within that genus the name striatus is preoccupied by A. striatus Funkhouser.
It is therefore necessary to give a new name to this very distinctive form. A translation
of Mtchedlishvili’s diagnosis and description of Parviprojectus striatus follows:

Diagnosis. Pollen grains rather large sized, tricolpate. Body broadly ellipsoidal in form,
equatorial projections short, colpi resembling cracks. Exine thick, pilate, striate, but having
thickenings on the polar areas.

552

PALAEONTOLOGY, VOLUME 11

Description. Length of grain 46-0-46-6 p; diameter of body 27-4-31-0 p, length of equatorial
projections 9-7 p, their width 7-7-9'0 p.

Pollen grains tricolpate. Outline of body in equatorial view broadly ellipsoidal. Exine
thickness about 2-Op. Nexine layer only slightly thinner than sexine layer, tapering from
the proximal end of the projection. Sexine consisting of pila which have grown together to
form comb-like lirae. In sectional view are seen the thinner parts of the columns and the
round heads. The lirae are situated at a slight angle to each other and sometimes branch,
forming the characteristic striate structure of the grain. At the poles the sexine layer becomes
considerably wider, forming thickened polar areas which have a corrugated structure. On
the equatorial projections the sexine layer becomes considerably thinner. Colour of grains
brown.

Size range of Scottish specimens {5 grains). Polar axis 35-43 p (av. 40 p); equatorial axis 28-33 p
( c . 38 p tip to tip of protrusions); thickness of the polar area 4-5-6 0 p.

Location. Shiaba interbasaltic lignites, Mull, Scotland.

Known time range. Maastrichtian-(?) Danian (Samoilovich and Mtchedlishvili 1961).

Distinction. The grains are distinguishable from A. nntrus Stanley, and A. bertillonites
Funkhouser by the thickened polar areas. They are partially distinguishable by their
generally larger size and shorter projections. The colpoids are very narrow, rather in-
distinct, and in one specimen (PI. 106, fig. 10), pores appear to be present. The size of
Scottish specimens is smaller than in the two Siberian specimens.

EXPLANATION OF PLATE 106

Magnification of all figures, X 1,000.

Figs. 1-3. Aquilapollenites attenuatus Funkhouser ( Taurocephalus proteus Simpson pro parte). Specimen
with four equatorial arms, also showing demicolpi; slide Bremanoir 5 a (= Simpson 1961, pi. xii,
fig. 11). 1, High focus, polar area. 2, High focus, exine pattern, showing baculi (puncta) and spines.
3, Low focus, exine pattern (LO-pattern).

Figs. 4-6. Aquilapollenites attenuatus Funkhouser. 4, Equatorial view, showing well-defined double
pattern of spines and baculi (LO-pattern) ; slide Shiaba 29 a. 5, Equatorial view of small grain showing
typical attenuatus shape, but lacking distinct LO-pattern; slide Bremanoir 5a. 6, Polar view; slide
Bremanoir 5a.

Fig. 7. Aquilapollenites spinulosus Funkhouser ( Taurocephalus proteus Simpson pro parte). Equatorial
view, showing scattered spines and lack of distinct LO-pattern; slide Bremanoir 5a.

Figs. 8-14. Aquilapollenites pachvpolus nom. nov. ( Parviprojectus striatus Mtched.). 8, Equatorial view,
showing striate pattern, short projecting arms, and thickened polar area; slide Shiaba 1 a. 9, Surface
and median view to show thickening of both ectexine and endexine in polar area; slide Shiaba 22 a.
10, Small grain, with apparent pore-apertures on equatorial arms, showing some convergence
towards A. bertillonites Funkhouser; slide Shiaba 22 a. 11, Oblique view, sculpture of polar area
insulous, colpoid visible in facing projection; slide Shiaba 1 a. 12-13, Successive views through
specimen with upward projecting arm, showing the continuation of the striate pattern over the
narrow indistinct colpoid, thus demonstrating that the colpoid is restricted to the endexine. 14,
The same, focused on pattern of lower surface of grain; slide Shiaba 36/.

Figs. 15-16. Classopollis sp. High and low focus of striate exine surrounding pore, which faces to
right; slide Shiaba 1 a.

All photographs from original slides prepared by J. B. Simpson, and housed in Scottish Geological

Survey Office, Edinburgh.

Palaeontology , Vol. 11

PLATE 106

MARTIN, Aquilapollenites from Scotland

MARTIN: AQUILAPOLLENITES IN THE BRITISH ISLES

553

Acknowledgements. I gratefully acknowledge the receipt of an honorary associateship at University
College, London University, the hospitality of Dr. W. G. Chaloner in offering the facilities of his
section in the Botany Department, and the kindness of the Scottish Geological Survey Office, Edinburgh
in making freely available the slides and materials left by Dr. J. B. Simpson.

REFERENCES

funkhouser, j. w. 1961. Pollen of the genus AquilapoIIenites. Micropaleontology, 7, 2, 193-8.
manum, s. 1962. Studies in the Tertiary flora of Spitsbergen, with notes on Tertiary floras of Ellesmere
Island, Greenland and Iceland. Norsk Polarinstitutt Skrifter, Oslo, 125.
samoilovich, s. and mtchedlishvili, n. d. 1961. Pollen and spores of western Siberia, Jurassic to
Palaeocene. Trudy V.N.I.G.R.Inst., Leningrad, 177, 657 pp. [In Russian.]
simpson, j. b. 1961. The Tertiary pollen-flora of Mull and Ardnamurchan. Trans. Roy. Soc. Edinburgh ,
64, 16, 421-68.

srivastava, s. k. 1966. Upper Cretaceous microflora (Maestrichtian) from Scollard, Alberta, Canada.
Pollen et Spores , 8, 2, 497-552.

Stanley, e. a. 1961. The fossil pollen genus AquilapoIIenites. Pollen et Spores 3, 2, 329-52.

A. R. H. MARTIN

School of Biological Sciences
University of Sydney

Typescript received 28 October 1967 Australia

DISCOHELIX (ARCHAEOGASTROPODA,
EUOMPH ALACEA) AS AN INDEX FOSSIL IN THE

TETHYAN JURASSIC

by J. WENDT

Abstract. Morphological features, growth, mode of life, and stratigraphical distribution are examined in more
than 300 extremely well-preserved specimens of Discohelix from the Sicilian Lower and Middle Jurassic. In
some species parabolic ribs similar to those in ammonites are formed owing to resorbtion during brief dis-
continuities in growth. Deviated peristomes recurring at an angle of more or less 72 0 and resulting in a penta-
gonal outline of the shell, are developed in two species for which the new subgenus Pentagonodiscits (type species
D. (P.) angusta n. sp.) is established. Ten species of Discohelix sensu stricto are described, five of which are new:
D. dictyota, D. conica, D. centricosta , D. costata, and D. levis. The short stratigraphical ranges of the represen-
tatives of the genus are proved here for the first time and may be useful for the division of cephalopod-free
series in the Tethyan Jurassic.

Few genera of the widely distributed Palaeozoic Euomphalacea crossed the Permo-
Triassic boundary and the superfamily became extinct in the Upper Cretaceous. In
contrast to its Palaeozoic representatives, amongst which Knight (1934, p. 145) was
even able to recognize some with value as index forms, those in the Mesozoic have
neither been investigated in quantity nor in relation to their stratigraphical occurrence.
Both these omissions can now be corrected using an individually and specifically rich
gastropod fauna from the Jurassic of Sicily, of which members of the genus Discohelix
are the chief element. The accompanying ammonites have usually allowed all the
specimens to be zonally dated, so far as strong condensation permits such accuracy.
A total of more than 300 Discohelix, extremely well preserved and with complete shells,
have been collected from several Jurassic sections in western Sicily (Rocca Busambra).
The only drawback is the recrystallization of the shell material which has left its struc-
ture difficult to recognize. Several specimens from the Liassic of the Northern Cal-
careous Alps, the originals of Stoliczka (1861) from the north alpine Hierlatz facies,
and those of Quenstedt (1884) and Brosamlen (1909) from the south German epi-
continental Jurassic have been available for comparison.

Acknowledgements. I am indebted to Professor Dr. A. Seilacher (Tubingen) for his interest in the work
and for critical reading of the manuscript, to Professor Dr. O. H. Schindewolf (Tubingen) and Dr.
E. L. Yochelson (Washington) for stimulating suggestions and discussions. The loan of specimens
from the Geologische Bundesanstalt Vienna and the Bayer. Staatssammlung Pal. Hist. Geol. Munich
was kindly arranged by Professor Dr. R. Sieber and Dr. D. Herm. The fieldwork was made possible
by the financial support of the Deutsche Forschungsgemeinschaft. Finally I would like to thank Dr.
D. A. B. Pearson for his translation of the German text, and W. Wetzel (Tubingen) who prepared the
photographs. The specimens labelled Ga 1347/1-56 are deposited in the Tubingen Geological Institute
collection.

MORPHOLOGY

Protoconch

Because of the more or less planispiral coiling of the shell, the larval whorls are almost
always preserved, so that the sculpture of the earliest ontogenetic stages can be precisely

[Palaeontology, Vol. 11, Part 4, 1968, pp. 554-75, pis. 107-10.]

WENDT: DISCOHELIX (ARCH AEOG ASTROPOD A, EUOMPH ALACEA) 555

followed. After the nucleus, which is always smooth, two or three fine spiral lines appear
on the upper and lower sides in the majority of the investigated individuals. After one
or two whorls they disappear even if a renewed spiral ornament occurs later on the teleo-
conch (PI. 107, fig. 16). In only three species is the protoconch completely smooth
(PI. 107, fig. 15). A boundary between the proto- and teleoconch can neither be traced
through a particularly accentuated interruption in growth (varix), nor through a sudden
onset of the adult sculpture. The number of larval whorls can therefore not be deter-
mined. When the shell is biconcave the protoconch is commonly slightly raised (text-fig.
3o), and when convexo-concave a little depressed (text-fig. 3c).

Teleoconch

Simple ribs. After one or two whorls, ribs appear more or less contemporaneously with
the onset of clear growth-lines. They consist at first of weak swellings in the keel region,
but with progressive growth they become enlarged as periodic ridges reaching to the
suture on the upper and lower side, and running parallel with the growth-lines. On the
interior of the shell this sculpture is either totally unrefiected or only barely detectable,
so that steinkerns of strongly ribbed forms may be perfectly smooth. As the growth-lines
are not noticeably denser on the ribs themselves than between them, they are more
likely to have been produced by periodically increased shell deposition than by a
deceleration of growth. The number and strength of the ribs can vary considerably and
is therefore of only restricted specific value.

Parabolic ribs. Long known in certain ammonites (especially perisphinctids) the signifi-
cance of so-called parabolic ribs has often been discussed. In both ammonites and gas-
tropods they are morphogenetically distinguishable from simple ribs. In form they are
similar to simple ribs, but the course of the growth-lines is different. At regular intervals
a growth-line (parabolic line) swings backwards in the keel region forming a parabolic
notch and truncating the earlier growth-lines (text-fig. 1a). This is even more clearly
seen on shells with a spiral sculpture since the sculpture itself is also cut off in the area
of the parabolic notch. To some extent the spiral elements curve in sympathy with the
notch and thus hint at the following discontinuity in growth (text-fig. 1b). The para-
bolic notch may be developed as a shell ridge (D. a/binatiensis), or as a node ( D . cf.
gumbeli ), or may take the form of a longitudinally open spine ( D . ( P .) angusta). In the
last case the growth-lines are practically continuous.

The truncation of earlier growth-lines and the development of special sculpture
features show that the normal growth of the shell was interrupted at these points. The
finger-like protrusions of the mantle concomitant with the parabolic nodes or notches
possessed the ability, on the one hand for resorbtion, and on the other, for shell secretion
As is often the case with the siphon it may have projected beyond the calcareous envelope
which it depositated. In normal shell formation the parabolic notch was first closed
before new material was added to the shell margin. Only then did a marked cessation
of growth occur, since: firstly, in D. (P.) angusta the end of a strong inner varix (former
peristome) does not lie immediately beneath the parabolic notch but shortly mouth-
wards of it; secondly, the existing peristome was only formed after the infilling of the
parabolic notch (however, in this final growth stage the last 2-3 ribs are often simple);
and thirdly, a clear concentration of growth-lines owing to decelerated growth is only

C 5934 O O

556

PALAEONTOLOGY, VOLUME 11

visible immediately after the parabolic line. In this area shell damage which affected the
earlier peristome is often found (text-fig. 1b).

text-fig. 1. Parabolic nodes and course of the growth-lines from the suture to the middle of the outer
side (raised parts of the shell in black), a. Discohelix (£>.) cf. albinatiensis Dumortier, Ga 1347/14, X 13.
b. Discohelix ( Pentagonodiscus ) angusta n. sp., with healed injury in a former peristome, Ga 1347/50,

X 12, 5.

From this it can be concluded that the parabolic notch may not itself be equated with
a former aperture, in other words, a halt in growth (of course, strictly speaking each
growth-line, and therefore also the parabolic line, represents a temporary shell margin).
It is rather a part of a not wholly continuous process of shell secretion between two
protracted interruptions in growth, although locally the contrary, namely resorbtion,
occurs. Observation of recent gastropods has shown that shell secretion between two
varices occupies only a few days and that a long pause follows thereafter, in which the
peristome is only strengthened by internal varices or ribs (Fretter and Graham 1962,
p. 64). For this reason gastropods in which the shell margin lies within the rapid
growth phase are rarely found. The same growth process explains why a parabolic line
has never been observed as the final peristome in ammonites. As the extension of the
mantle withdrew the parabolic notch which it had produced was infilled, a process that
bearing in mind the speed of growth in gastropods could probably have been completed
within a few hours. Only later was growth interrupted, which in perisphinctids is com-
monly indicated by deep constrictions equivalent to the inner varices of gastropods. It
remains unknown whether the temporary finger-like extensions of the mantle had a

WENDT: DISCOHELIX (ARCH AEOG ASTROPODA, EUOMPHALACEA) 557

physiological function in addition to the formation of parabolic notches and hollow
spines. Nothing is known of this in gastropods either.

The interpretation of the parabolae of perisphinctids given by Schairer (1967, p. 26)
is incorrect. He proposed as an explanation that ‘the shell at certain points was more
slowly deposited than at others’ (author’s translation). If such were the case the growth-
lines would be of different density around the curve of the parabolic notch, but this is
not so. Moreover, from Schairer’s fig. 11, one can see that on the parabola the sculpture
(in this case ribs as the author had only steinkern material) is indeed truncated. To my
knowledge only Schindewolf (1934, p. 342, text-fig. 9) has demonstrated that the
parabolae of ammonoids (here Upper Devonian clymenias) originated through local
resorbtion.

Spiral ornament. Beginning with an intermediate growth-stage, commonly on the last
2-3 whorls, fine spiral lines are formed as result of a folding of the mantle edge. They
appear first in the keel region and gradually encroach onto the sides, and together with
the growth-lines and ribs produce a retiform sculpture typical of many gastropods.
Their number and strength may vary considerably between specimens of equal dimen-
sions, and spiral lines are therefore of only restricted value in the differentiation of
species of Discohelix. For example in the case of D. albinatiensis they can be either
equally spaced over the visible part of the whorl or only developed in the keel region,
and thus show a transition to specifically inseparable shells without any spiral ornament
(D. cf. albinatiensis).

Peristome (text-fig. 2). Owing to the pure calcareous facies from which the whole of the
material was collected, in no specimen could the peristome be completely cleared and
the structures on the inside of the shell made visible. They could only be investigated
in section, and the various types are best illustrated when cut in a longitudinal direction.

In the simplest case the shell margin can only be recognized through a marked
concentration of the growth-lines, without any alteration of the peristome profile (e.g.
D. dictyota , text-figs. 2a, 3e). Probably not to be specifically separated from this type
is a peristome which is gradually thickened by the addition of new shell substance on
the inner side, and which therefore appears somewhat narrowed in cross-section (e.g.
D. albinatiensis, D. cornea, D. cf. crenulata — text-figs. 2b, 3b, d, g, c). It may be preceded
on the last whorl by one or two similar deposits formed during a pause in growth, but
their position is random (text-figs. 2c, d).

The broadened, trumpet-like peristome is strengthened by an inner varix on the
outer lip (e.g. D. levis, text-fig. 2e, 3l), and the aperture may also be narrowed by an
inner varix on the inner lip (e.g. D. cotswoldiae, text-fig. 2f), The highest degree of
specialization is achieved when these last two aperture types are repeated at regular
intervals ( D . ( P .) reussii and D. (P.) angusta — text-figs. 2g, h).

Cross-section (text-fig. 3). A cross-section passing exactly through the protoconch illus-
trates best the considerable variety of form within Discohelix. The mean spiral angle
may lie between the wide boundaries of 125° to 220°, and the mean umbilical angle
between 65° and 145°. The comparison of these two angles in one specimen, and the
weak trochospiral coiling of the protoconch, show that perfect bilateral symmetry is not

558

PALAEONTOLOGY, VOLUME 11

achieved in Discohelix. Transitions exist between a slightly concave, a plane, and a convex
spire, so that this character cannot always be given specific significance. D. albinatiensis
shows the greatest variability in this way and may occur with a plane (sometimes even
slightly concave) to a clearly convex spire (text-figs. 3b, a). In one and the same individual

text-fig. 2. Types of peristomes (longitudinal sections), x3, 1. A. Discohelix (D.) dictyota n. sp., Ga
1347/4. b. Discohelix (£>.) albinatiensis Dumortier, Ga 1347/12. c. Discohelix (D.) conica n. sp., Ga
1347/28. d. Discohelix (D.) conica n. sp., Ga 1347/29. e. Discohelix (£>.) lev is, n. sp., Ga 1347/41. f.
Discohelix (D.) cotswoldiae (Lycett), Ga 1347/48. g. Discohelix ( Pentagonodiscus ) reussii (Hornes, Ga
1347/56. h. Discohelix ( Pentagonodiscus ) angusta n. sp., Ga 1347/51.

the degree of involution may change so that the cross-section loses its axial (mirror-
image) symmetry.

Septa. Septa closing off the uninhabited parts of the shell were shown by Knight (1934,
p. 140) to be typical for the Euomphalidae. However, although many longitudinal
sections were prepared for the species investigated here, septa were observed in only
a few cases. They were found in the first 2-3 whorls but showed no regularity in either
number (3-15) or distance from one another (text-fig. 4b). In one specimen of D. (P.)
angusta three septa were found lying against the inner varices towards the end of the
penultimate whorl (text-fig. 4a). This very late deposition of septa is surprising and
permits no generally applicable conclusion about the length of the visceral mass.

text-fig. 3. Cross-sections, x 3, 1. A. Discohelix ( D .) albinatiensis Dumortier, Ga 1347/9. b. Disco-
helix (D.) albinatiensis Dumortier, Ga 1347/7. c. Discohelix ( D .) conica n. sp., Ga 1347/24. d. Discohelix
(D.) cf. albinatiensis Dumortier, Ga 1347/15. e. Discohelix (D.) dictyota n. sp., Ga 1347/2. f. Discohelix
(D.) cf. crenulata (Moore), Ga 1347/19. G. Discohelix ( D .) cf. crenulata (Moore), Ga 1347/23. h.
Discohelix ( D .) cf. calculiformis Dunker, Ga 1347/42. i. Discohelix (D.) centricosta n. sp., Ga 1347/31.
k. Discohelix (£>.) costata n. sp., Ga 1347/35. l. Discohelix ( D.) levis, n. sp. Ga 1347/37. m. Discohelix
(D.) cotswolcliae (Lycett), Ga 1347/45. n. Discohelix ( D.) cf. giimbeli Ammon, Ga 1347/49. o. Discohelix
( Pentagonodiscus ) angusta n. sp., Ga 1347/52. p. Discohelix ( Pentagonodiscus ) reussii (Hornes), Ga

1347/55.

560

PALAEONTOLOGY, VOLUME 11

text-fig. 4. Discohelix ( Pentagonodiscus ) angusta n. sp. a. Ga 1347/54, longitudinal section with three
septa towards the end of the penultimate whorl (inner whorls not preserved), X3-2. b. Ga 1347/51,

inner whorls with septa, x 25.

GROWTH

In smooth shells no clearly marked interruptions in growth can be read from the
density of the growth-lines. Therefore, the formation of the shell must have been
more or less continuous. The same applies to shells with simple ribs, but those with
parabolic ribs grew discontinuously, as previously discussed (p. 555). Similarly the
regular inner varices following one another at intervals of 69° to 75° (mean 71-5°) in
D. ( P .) reussii, indicate periodic phases of shell deposition between long pauses (text-
fig. 5).

text-fig. 5. Discohelix ( Pentagonodiscus ) reussii (Hornes), Ga 1347/56, longitudinal section, x6-3.

WENDT: DISCOHELIX ( ARCHAEOG ASTROPODA, EUOMPHALACEA) 561

These two growth-rhythms, expressed in the one case by parabolic ribs and in the
other by inner varices, occur together and interfere with one another in D. ( P .) angusta.
As in D. ( P .) reussii the inner varices exhibit pentagonal symmetry (means of 704°,
70-7°, and 76° in three measured specimens), and between them three (on inner whorls
sometimes two) parabolic ribs are intercalated. These ribs tend to arrange themselves
in groups of three towards the last whorl. In this way the first formed parabolic rib of
such a group coincides with an inner varix (text-figs. 6a, 7). (The outer varices, seen
externally, cannot be exactly transferred into the longitudinal section; their position
may therefore be slightly different.)

text-fig. 6. Discohelix ( Pentagonodiscus ) angusta n. sp. A. Ga 1347/53, longitudinal section, x 6 3
(arrows indicate outer varices), b. Ga 1347/51, extrapolation of the normal logarithmic spiral (hatched
line) from the regularly coiled inner whorls in comparison to the actual spiral (unbroken line), x 6.

Thus if one considers the parabolic features as brief discontinuities and the inner
varices as drawn out breaks in shell secretion, then D. ( P .) angusta shows complicated
double-phased growth, the control of which is unknown.

The subgenus Pentagonodiscus assumes a pentagonal outline in the last 2-2 f whorls.
In text-fig. 6b the normal logarithmic spiral of the regularly coiled inner whorls has
been extrapolated omitting the various irregularities of the shell (inner varices, ribs).
This clearly shows that the aberrant shell form commences with the first inner varix
and becomes more accentuated as the varices increase in strength towards the aperture.
The most prominent deviations from a logarithmic spiral occur at the points of periodic

562

PALAEONTOLOGY, VOLUME 11

constrictions of the former peristomes. In the intervening growth-phases the spiral is
again followed and thus shows itself to maintain control of the pentagonal coiling.

Among the gastropods, there appear to be only two cases of angular instead of regular
spiral coiling: Triangularia paradoxa Freeh and Tremanotus polygonus Barrande. The
first has a triangular, and the latter a pentagonal outline. Nothing is known of the
internal structure of either of these two species. Periodically constricted apertures were
probably also formed and might be revealed by inner varices.

text-fig. 7. Discohelix ( Pentagonodiscus) angusta n. sp. Ga 1347/53. Diagrammatic view of the
intervals of the inner and outer varices on the last whorl (n = no. of outer varices, counted from the

centre).

Deviations from a logarithmic spiral are widely distributed among ammonoids, but
commonly affect only the last body chamber. Triangular coiling in early ontogenetic
stages, producing an extremely aberrant shell form, is found solely in a few clymenids
{Soliclymenia, Wocklumeria, Parawock/umeria ) and goniatites ( Paralegoceras ) (see
Schindewolf 1937, p. 110 et seq.). It is conceivable that periodic narrowing of the
aperture was also responsible. Deep constrictions, signifying a former aperture at the
points of strongest deviation from the normal spiral, occur in two of these genera
( Wocklumeria , Parawocklumeria ) and support this supposition. A narrowed body
chamber in early ontogenetic stages, remaining unresorbed, and periodicity in growth
lead automatically to an angular instead of a spiral shell form. If one imagines further
shell growth after the last known peristome of e.g. Oecoptychius refractus (Reinecke),
an angular outline must necessarily be the result. Not to be confused with such genuine

WENDT: DISCOHELIX (ARCH AEOGASTROPOD A, EUOMPHALACEA) 563

angular forms are those in which the umbilicus has a polygonal outline as a result of
deep constrictions (e.g. Morphoceras, some perisphinctids). In longitudinal section they
have a perfectly regular logarithmic spiral.

No attempt has been made here to fully explain the phenomenon of triangular am-
monoids. The control of such aberrant growth remains as speculative as in polygonally
coiled gastropods.

MODE OF LIFE

The planispiral coiling of many Discohelix species is reminiscent of ammonoids and
suggests adaption to a nektonic mode of life. Thus Koken (1897, p. 42) proposed that
the Upper Triassic KokeneUa (Pleurotomariacea), which is first coiled in a trochospiral
and later in a planispiral, was adapted to a free swimming in deep sea. However, no
recent gastropod with an almost symmetrical shell is known to be capable of active
swimming. A single exception are the pelagic Heteropoda (e.g. Atlanta) with a weakly
calcified, very thin or wholly vestigial shell. They cannot be compared with the thick-
shelled, often strongly sculptured Discohelix , for which only a benthonic mode of life
can be considered. If the planispiral coiling were an adaption to a nektonic mode of life,
then this feature would normally tend to be realized as a result of selection. It turns out,
however, to be extremely variable, not only on the specific but also on the generic level.

The disc shape, in fact, would appear to be a disadvantage since it is much more
susceptible to water movement than a helicocone. This can scarcely have been com-
pensated for by the narrow whorls and the small aperture, indicating a correspondingly
small and weak foot. For this reason the most favourable environment for Discohelix
may have been in quiet sublittoral water or in the shelter of reefs rather than in a dis-
turbed littoral. In a vertical position the pentagonal form can be seen to have an
advantage over the spiral form, since the angularity provides a kind of resting surface
(cf. Murex ) allowing the animal a greater stability on the sea-floor. The efficiency of
this is increased through the support of lateral spines in D. ( P .) angusta.

DEVELOPMENT AND VALUE AS INDEX FOSSILS

There are still great gaps in our knowledge of the development of the Euomphalidae
in the Triassic. For this reason the exact origin of the genus Discohelix has not yet been
elucidated. Several Triassic forms described as Euomphalus belong to Woehrmatmia,
but the majority are insufficiently known for an exact generic determination.

Typical biconcave shells with more or less bilaterally symmetrical whorls and upper
and lower keel appear for the first time in the lowermost Liassic. A greater variety of
forms are known from the Sinemurian, especially from the Northern Calcareous Alps
(Hierlatz-facies), amongst which are some with a plano-concave cross-section. The
transition from biconcave to convexo-concave shells probably occurred several times
independently, as has been observed in D. albinatiensis (p. 558). A special off-shoot of
pentagonally coiled forms (subgenus Pentagonodiscus) also commenced in the Sine-
murian but with only two stratigraphically isolated species, it has not yet been fully
verified. The observed variation of individual features reduces their phylogenetic value
and makes it necessary to separate species according to character combinations. Despite

564

PALAEONTOLOGY, VOLUME 11

the approximately forty known (many insufficiently) Discohelix species their lines of
evolution can only be traced at a few points. Their stratigraphical distribution has not
been proved because of their sparsity, occurrence as fissure-faunas, and the failure of
accompanying index fossils. Therefore the well dated material presented here assumes
a special stratigraphical value. From text-fig. 8 it can be seen that the species have only
a very short range, extending through only one or a few zones. Similar conclusions can
be drawn from the stratigraphical information given by Hudleston (1892, pp. 315-21)
for the English Middle Jurassic.

D. dictyota

D. a/binatiens/s

D. cf a/binatiens/s

!

1

D. conica

D, centr/costa

D. costa ta

1

-s?

1

D.cfca/cu/iformis

D.(P.)angusta

-cs

1

27

50

52

28

> 100

7

13

9

6

n

17

1

Malm

Oxford/an

u

L

1

1

■

■

■

Dogger

Calloi/ian

U

M

L

Bathonian

U

M

L

Bajocian

U

M

16

T

1

L

28

1

■

■

1

1

■

1

1

1

s

1

■

I

B

■

■

L

r

■

■

■

i

Aa/enian

U

L

Lias

Toarcian

U

~120

■

1

T

M

148

T

T

1

1

r

1

L

text-fig. 8. The stratigraphical distribution of Discohelix in the condensed Jurassic of Rocca Busambra
(Western Sicily); numbers indicate number of specimens.

The strong condensation in the sections made it necessary to use substages instead
of zones for the time-scale in text-fig. 8. In three cases the boundaries of the substages
could not be drawn since up to five zones are inseparably concentrated in one strati-
graphical unit (Wendt 1965, p. 300). For this reason the exact range of some of the
species cannot be given. The representatives of Discohelix which begin the fossil record
in the Middle Toarcian (Bifrons Zone) may have originated earlier. The absence of
species from the base of the Bathonian to the Upper Callovian is owing to a stratigraphic
gap characteristic of the west Sicilian Jurassic. A diminution of the Discohelix fauna is
already evident earlier, and the last representative on the Dogger-Malm boundary is
almost coincident with the extinction of the genus in the Lower Oxfordian.

WENDT: D1SCOHELIX (ARCH AEOGASTROPODA, EUOMPH ALACEA)

565

It should not be overlooked that the stratigraphical range of the species investigated
here is based on a series of sections from a single locality. To date it has been verified in
two cases by specimens from the Northern Calcareous Alps. Furthermore well-dated
material of Discohelix is wanted to complete and, if necessary, correct the ranges of the
various species.

Only two imperfectly preserved examples were available to Dunker when he estab-
lished the genus Discohelix in 1847, so that even he himself was uncertain of its indivi-
duality five years later (see footnote in Reuss 1852, p. 116). Rich material allowed
Stoliczka (1861, p. 180 et seq.) to distinguish Discohelix from Euomphalus and to define
it precisely for the first time. Without altering this concept of the genus Cossmann
(1915, p. 133 et seq.) included in it both Triassic and Cretaceous species, although they
did not completely pass within his diagnosis. In the same publication he proposed the
genus Colpomphalus (1915, p. 136) for forms with a convex spire transitional to Disco-
helix. This gradation is in fact so continuous, even in one and the same species (see
p. 558), that Colpomphalus can no longer be maintained as an independent unit. Among
the Triassic predecessors of Discohelix, Amphitomaria Koken differs through its prosocyrt
growth-lines on the outer side, and Anisostoma Koken through its peculiar, deflected
peristome. (Specimens placed in the latter genus, from which the aperture has been
broken, could, in fact, be early members of Discohelix.)

Type species. D. calculiformis Dunker 1847.

Diagnosis. Widely to extremely widely umbilicate representatives of the Euomphalidae,
with numerous, not overlapping whorls of trapezoidal cross-section, with an upper and
lower keel. Mostly dextral, rarely sinistral. Spire concave, plane or convex. Protoconch
homeostrophic, not sharply separated from the teleoconch. Growth-lines on the upper
and lower sides prosocline, on the outer side opistocyrt. Ornament collabral and/or
spiral. Peristome more or less tangential to foregoing whorl, its outline parallel to the
axis of coiling, simple or with thickened outer and/or inner lip. Operculum unknown.

Distribution. Triassic? Jurassic (Hettangian to Oxfordian).

Synonym. Colpomphalus Cossmann 1915.

SYSTEMATIC DESCRIPTIONS

Genus discohelix Dunker 1847

Discohelix ( Discohelix ) dictyota n. sp.
Plate 107, figs. 1-8; text-figs. 2a, 3e

Derivation of name, dictyotus = retiform.
Holotvpe. Ga 1347/1. Plate 107, figs. 1-4.

Ga 1347/1
Ga 1347/2
Ga 1347/3

No.

dp d h/d ha/wa nw nt sa ua
200 0-29 1 -49 9 34

0-28 1-33 8 (30)

0-29 1-78 8 33

18-3

16-4

214° 145°

566

PALAEONTOLOGY, VOLUME 11

Diagnosis. Dextral, spire concave; protoconch spirally ornamented, teleoconch with
retiform sculpture; peristome unthickened.

a

text-fig. 9. Shell measurements, a = axis of coiling, dp = maximum diameter measured at peri-
stome (in mm.), /; = height of shell, ha — height of aperture, p = cross-section through peristome
(hatched), sa = mean spiral angle, ua = mean umbilical angle, wa = width of aperture.

Additional abbreviations in the measurement-tables: d = diameter (in mm.), nt = number of tubercles
(nodes), numbers in ( ) referring to half whorls, nw = number of whorls.

Description. Fine ribs appear after the protoconch contemporaneously with the onset
of clear growth-lines. They develop on the keel as nodes, and with good preservation
may be recognized on the last whorls as a parabolic feature. On this last whorl the spiral
lines, 7-10 on the upper and lower sides and 16-20 on the outer side, are strongest; they
fail on the inner whorls of the teleoconch. The final peristome shows no special structure.
Before it, on two specimens, a short pause in growth is indicated by a gentle angulation
in the shell outline.

Discussion. D. reticulata Stoliczka shows a similar sculpture, the spiral lines being
restricted to the outer side, however, the whorls are more compressed, and broader
than high in cross-section. D. orbis (Reuss) can be distinguished by its much finer
ornament and clearly differentiated keel, and D. dunkeri Moore by its fewer nodes and
plano-concave cross-section.

Distribution. Middle Toarcian (Bifrons Zone).

Discohe/ix ( Discohelix ) albinatiensis Dumortier 1874

Plate 107, figs. 9-15; Plate 108, figs. 8-10; text-figs. 2b, 3a-b

1874 Discohelix Albinatiensis Dumortier, p. 284, pi. 59. figs. 3-5.

? 1892 Straparol/us pulchrior Hudleston, p. 318, pi. 25, fig. 9.

EXPLANATION OF PLATE 107
Figs. 1-4. Discohelix (£>.) dictyota n. sp., holotype. Ga 1347/1, X2.

Figs. 5-8. Discohelix (D.) dictyota n. sp., Ga 1347/3, X 2.

Figs. 9-12. Discohelix (D.) albinatiensis Dumortier, Ga 1347/6, X3.

Figs. 13-14. Discohelix (D.) albinatiensis Dumortier, Ga 1347/8, x3.

Fig. 15. Discohelix (D.) albinatiensis Dumortier, Ga 1347/11. Inner whorls with protoconch, X 20.
Fig. 16. Discohelix (D.) cf. albinatiensis Dumortier, Ga 1347/14. Inner whorls with protoconch, X 20.

Palaeontology, Vol. 11

PLATE 107

WENDT, Discohelix from the Tethyan Jurassic

WENDT: DISCOHELIX (ARCHAEOG ASTROPOD A, EUOMPEIALACEA) 567
Lectotype (here selected). Dumortier 1874, pi. 59, figs. 4-5.

No.

dp

d

Ipd

ha/wa

nw

nt

sa

ua

Ga 1347/5

15-3

0-29

1-55

8

(16)

184°

120'

Ga 1347/6

14 2

0-32

1 43

8

27

Ga 1347/7

14-2

0-30

1-48

8

(13)

188°

117'

Ga 1347/8

13-2

0-42

1 -22

8

25

Ga 1347/9

14-4

0-48

1-47

9

(18)

132°

81

Ga 1347/10

12-4

0-33

1-63

8

12

Diagnosis. Dextral, spire weakly concave, plane, or weakly convex; protoconch smooth,
retiform sculpture on teleoconch; peristome thickened.

Description. The majority of the available specimens are plano-concave, several speci-
mens, however, have a slightly concave or a more or less strongly convex spire. The
larval whorls are free of sculpture, with an angulation near the suture, slightly concave
upper side, and a smooth keel. The ornament of the teleoconch, beginning after 1 \-2
whorls, is similar to, but somewhat coarser than in D. dietyota. 20-27 parabolic nodes
may be counted on the keel. 10-20% fewer nodes are present on the lower keel, thus
emphasizing the asymmetry of the shell. Seven examples varying in diameter between
7 and 13 mm. stand apart from the remainder of the material in that the nodes are more
widely spaced (12-15 instead of 20-23 on typical individuals, see PI. 108, figs. 8-10).
A clear spiral ornament begins only after a diameter of 10 mm. has been reached. It may
completely fail on the upper and lower sides and be restricted to a small area adjacent
to the keel on the outer side. The peristome is internally thickened and therefore
narrowed.

Discussion. The larval whorls and peristome were not preserved in the material used by
the authors cited in the synonymy, so that the present specimens can only be named with
some reservation. Those with the most raised spire cannot be distinguished in cross-
section and sculpture of the teleoconch from D. exsertus (Hudleston 1892, p. 320,
pi. 26, figs. 3-4).

Distribution. Middle Toarcian (Bifrons Zone). The type was thought by Dumortier to have come
from the Opalinum Zone, but was found in a loose block of uncertain stratigraphic origin. According
to Hudleston (1892, p. 319) D. pulchrior occurs rarely in the Murchisonae Zone.

Discohelix ( Discohelix ) cf. albinatiensis Dumortier 1874
Plate 107, fig. 16; Plate 108, figs. 1-7; text-figs. 1a, 3d
? 1935 Discohelix albinatiensis Dumortier; Roman, p. 32, pi. 6, figs. 1, 1 a.

No.

dp

d

hid

lia/wa

nw

nt

sa

ua

Ga 1347/13

22-9

0-37

1-40

9

29

Ga 1347/14

210

041

1 96

9

30

Ga 1347/15

19-4

0-39

1-52

9

(16)

171°

99'

Ga 1347/16

16-3

0-47

1-73

8

(14)

165°

90

Description. In contrast to typical members of the species spiral sculpture is totally
lacking on the last whorls, but 2-3 spiral lines may be traced on the upper side of the

568 PALAEONTOLOGY, VOLUME 11

protoconch. From these two characters a decisive separation from D. albinatiensis is
not possible.

Distribution. Middle Toarcian (Bifrons Zone).

Discohelix ( Discohelix ) cf. crenulata (Moore 1867)

Plate 108, figs. 11-16; text-figs. 3f, g
cf. 1867 Solarium crenulatum Moore, p. 90, pi. 4, figs. 19-20.

No.

dp

d

hid

ha/wa

nw

nt

sa

ua

Ga 1347/19

14 4

0-46

1-41

9

(30)

148°

CC

O

Ga 1347/20

1 3 3

0-38

1-42

8

50

Ga 1347/21

12-8

0-52

1 -50

54

Ga 1347/23

7-7

0-57

1-36

7

147°

76°

Description. The shell form of this dextrally coiled species varies from almost plano-
concave ( sa = c. 170°) to clearly convexo-concave. The mean umbilical angle varies
correspondingly between relatively wide boundaries. The protoconch has distinct spiral
lines, and after \b-2 whorls fine ribs appear which rise to small nodes on the keel. Later,
spiral lines reappear so that the sculpture is similar to that of D. dictyota and D. dun-
driensis (Tawney), as Fludleston (1892, pi. 26, fig. 2) figured in detail. However, the
external spiral lines are seldom regularly distributed over the whole outer side of the
last whorl, and a smooth band is left free in the centre.

In the available material two varieties differing in their whorl number can be distin-
guished: one having 7 whorls at a maximum of approximately 8 mm. in diameter, and
the other 8-9 whorls at 13-15 mm. in diameter. The aperture of the small examples is
thickened (text-fig. 3g) and that one of the large ones is simple and unthickened (text-
fig. 3f).

Discussion. Moore’s figure and description do not give a clear impression of this species
and it has not since been mentioned. Therefore no comparison of the shell cross-sections
can be made, but aside from this the present specimens have a much wider umbilicus.
A similar umbilicus is possessed by D. crussoliensis Roman (1935, p. 102, pi. 3, figs.
5, 5a), but this species is considerably more conical. D. helenae (Dumortier 1874, p. 141,
pi. 36, figs. 1-4) has a comparable shell shape but the nodes on the keels are coarser and
more widely spaced (18 in contrast to 50-60 in D. cf. crenulata).

Distribution. Middle Toarcian (Bifrons Zone). The type was said to be from the upper Middle Liassic.

EXPLANATION OF PLATE 108

Figs. 1-2. Discohelix (D.) cf. albinatiensis D urn order, Ga 1347/13, x2

Figs. 3-4. Discohelix (£>.) cf. albinatiensis Dumortier, Ga 1347/18, x2,

Figs. 5-7. Discohelix (D.) cf. albinatiensis Dumortier, Ga 1347/17, X2,

Figs. 8-10. Discohelix ( D .) albinatiensis Dumortier, Ga 1347/10, X 3.

Figs. 11-13. Discohelix (D.) cf. crenulata (Moore), Ga 1347/21, X 3.
Figs. 14-16. Discohelix (D.) cf. crenulata (Moore), Ga 1347/22, x3.
Figs. 17-21. Discohelix ( D .) conica n. sp., holotype, Ga 1347/25, x3.
Fig. 21. Discohelix (D.) conica n. sp., Ga 1347/27, x3.

Palaeontology, Vol. 11

PLATE 108

WENDT, Discolielix from the Tethyan Jurassic

WENDT: DISCOHELIX (ARCHAEOGASTROPODA, EUOMPHALACEA) 569
Discohelix ( Discohelix ) conica n. sp.

Plate 108, figs. 17-21; text -figs. 2c-d, 3c
Holotype. Ga 1347/25. Plate 108, figs. 17-20.

No.

dp

hid

hajwa

nw

so

UCl

Ga 1347/24

91

0-64

1 -52

1

124°

63°

Ga 1347/25

9-3

0-65

1-42

7

Ga 1347/26

10-7

0-53

1 31

7 h

132°

74°

Diagnosis. Small, dextral, spire convex; protoconch spirally ornamented, on teleoconch
only growth-lines, smooth upper and finely tuberculated lower keel ; peristome thickened.

Description. Three spiral lines can be traced on the depressed protoconch and disappear
after 1-1| whorls. From approximately the fourth whorl fine nodes appear close to the
suture, stretched out in the direction of the growth-lines but not reaching the smooth
upper keel. The lower keel is formed by a succession of fine tubercles which are accom-
panied on the outer side by 2-3 undulating spiral lines. The externally thickened peri-
stome can sometimes be seen to have been preceded on the last whorl by 1-2 growth
interruptions (text-figs. 2c-d).

Discussion. D. conica contrasts with the above species in having a higher spire. Further-
more the failure of spiral ornament and the smooth upper keel are good specific charac-
ters.

Distribution. Middle to lower Upper Toarcian (Bifrons to c. Thouarsense Zone).

Discohelix ( Discohelix ) centricostan n. sp.

Plate 109, figs. 1-7; text-fig. 3i
Holotype. Ga 1347/30. Plate 109, figs. 1-3.

No.

dp

d

hid

ha/wa

nw sa ua

Ga 1347/30

230

0 31

1 51

8

Ga 1347/31

25-6

0-35

1-29

9 217° 136°

Ga 1347/32

18-4

0-32

1-75

8

Diagnosis. Dextral, spire concave; protoconch spirally ornamented, nepionic whorls
ribbed, neanic whorls smooth; peristome trumpet-like.

Description. The shell consists of 7-9 whorls and the spire is almost as depressed as the
umbilicus resulting in a nearly symmetrical cross-section. The 2-4 whorls following the
spirally ornamented protoconch have weak ribs and finely tuberculate keels. These
elements become less accentuated with growth. The last whorls are completely smooth
or show only traces of a slight undulation. The aperture, which is only preserved in two
cases, is broadened and thickened to an almost circular cross-section in the larger
specimen (text-fig. 3i), and in the smaller (Ga 1347/32) is similar in cross-section to the
preceding whorls.

570 PALAEONTOLOGY, VOLUME 11

Discussion. Without knowledge of the protoconch this species cannot be separated from
D. calculiformis Dunker.

Distribution. Middle to lower Upper Toarcian (Bifrons to c. Thouarsense Zone).

Discohelix ( Discolielix ) costata n. sp.
Plate 109, figs. 12-18; text-fig. 3k
Holotype. Ga 1347/34. PI. 3, figs. 15-18.

No.

dp

d

hid

ha/wa

nw

nt

Ga 1347/34

24-8

0-42

1-83

9

29

Ga 1347/35

250

0-38

1-46

8

(13)

Ga 1347/36

19-8

0-32

1-42

8

27

Diagnosis. Dextral, spire concave; protoconch spirally ornamented, teleoconch ribbed;
peristome thickened, somewhat broadened.

Description. 4-5 spiral lines are present on the upper and lower side of the protoconch,
but are quite absent from the following whorls. They are ornamented by regular ribs
which may be traceable in the nepionic stage or only on the last whorl. The outer side,
between the strongly corrugated keels, is completely smooth except for fine growth-lines.

Discussion. The biconcave shell form, the failure of spiral lines, and the strong sculpture
make D. costata a very distinctive species. A phylogenetic connection with D. centri-
costa seems probable, D. costata having originated from this species through a persis-
tence of the ribbing to the last whorl.

Distribution. Upper Toarcian (c. Variabilis to Levesquei Zone).

Discohelix ( Discohelix ) lev is n. sp.
Plate 109, figs. 8-11 ; text-figs. 2e, 3l
Derivation of name, levis = smooth.

Holotype. Ga 1347/38. Plate 109, figs. 8-9.

No.

dp

d

h\d

ha/wa

nw sa ua

Ga 1347/37

22-7

0-38

1-29

8 211° 124°

Ga 1347/38

19-7

0-36

1-51

8

Ga 1347/39

15-6

0-33

1-30

Ga 1347/40

19-8

0-37

1-59

8

EXPLANATION OF PLATE 109

Figs. 1-3. Discohelix (D.) centricosta n. sp., holotype, Ga 1347/30, X2.
Figs. 4-6. Discohelix (£>.) centricosta n. sp., Ga 1347/32, X 2.

Fig. 7. Discohelix (D.) centricosta, n. sp., Ga 1347/33, X 3.

Figs. 8-9. Discohelix (D.) levis n. sp., holotype, Ga 1347/38, X 2.

Figs. 10-11. Discohelix (D.) levis n. sp., Ga 1347/40, X 2.

Figs. 12-14. Discohelix (D.) costata n. sp., Ga 1347/36, x2.

Figs. 15-18. Discohelix (D.) costata n. sp., holotype, Ga 1347/34, x2.

Palaeontology, Vol. 11

PLATE 109

WENDT, Discohelix from the Tethyan Jurassic

WENDT: D1SCOHELIX ( ARCHAEOG ASTROPOD A, EUOMPH ALACEA) 571

Diagnosis. Dextral, spire concave; protoconch and teleoconch smooth; peristome
broadened and thickened.

Description. The smooth shell is only ornamented with growth-lines, which become
stronger towards the keel causing it to be corrugated. The slight asymmetry of the
cross-section is seen in the greater depression of the umbilicus relative to the spire, and
in the irregular curvature of the outer side. In three specimens the trumpet-like aperture
is preserved and is reinforced by an inner varix on the outer lip.

Discussion. The shape is very similar to D. excavata (Reuss) without being as high as
this species {lt\d = 043-0-48 in Stoliczka’s originals). Apart from this the keel in D. levis
remains smooth as far as the peristome, whilst in D. excavata more or less strong nodes
appear on the last half-whorl. The sculpture of the protoconch of this last species is
unknown.

Distribution. Aalenian to Lower Bajocian (condensed Opalinum to Sauzei Zone).

Discohelix ( Discohelix ) cf. calculiformis Dunker 1 847
Plate 110, figs. 1-4; text-fig. 3h

cf. 1884 Discohelix calculiformis Dunker; Quenstedt, p. 325, pi. 197, figs. 32, 33, 35.

No.

dp

d

hid

ha/wa

nw set net

Ga 1347/42

16-1

0-36

1 -80

«n

m

O

m

<N

00

Ga 1347/43

20-4

0-31

1-66

8

Ga 1347/44

21-1

0-36

1 -48

8

18-6

0-34

1 64

Description. The umbilicus is somewhat more depressed than the spire and only this
reveals the dextral coiling of the species. The cross-section of the whorls is regularly
trapezoidal. The sculpture of the protoconch, consisting of 5-6 spiral lines on the first
two whorls, is clearly different from that of the teleoconch. The following whorls are
ornamented as in D. levis but the keel is more sharply expressed and the outer side less
strongly curved. The aperture, preserved in only one specimen, is somewhat broadened
and thickened on the outer side.

Discussion. D. calculiformis Dunker, founded on two badly preserved specimens and
an unsatisfactory figure, was regarded by its author five years later as probably identical
with D. orbis (Reuss 1852, p. 1 16). On the one hand Quenstedt’s originals of D. calculi-
formis and on the other hand Stoliczka’s of D. orbis (1861, p. 182, pi. 3, figs. 8-10), as
well as other topotypic material of the last species are available to me for comparison.
The first seem to have a smooth protoconch and no trace of spiral ornament on the
later whorls; the last show a very fine spiral sculpture which can also be seen in Stoliczka’s
figures. The present specimens differ from both these species in the spiral ornament on
the protoconch and the smooth teleoconch. They have here been attached with open
nomenclature to D. calculiformis sensu Quenstedt, from which they cannot be confidently

pp

C 5934

572

PALAEONTOLOGY, VOLUME 11

separated without knowledge of the protoconch. The erection of a new species does not
seem to be justified so long as D. calculiformis remains so poorly known.

Distribution. Aalenian to Middle Bajocian (condensed Opalinum to c. Subfurcatum Zone).

Discohelix ( Discohelix ) cotswoldiae (Lycett 1850)

Plate 110, figs. 5-12; text-figs. 2f, 3m

1850 Solarium Cotswalcliae Lycett, p. 419, pi. 11, figs. 2, 2a.

1892 Discohelix Cotswoldiae Lycett; Hudleston, p. 316, pi. 25, fig. 7.

No.

dp

d

hid

hajwa

nw

Ga 1347/45

120

0 31

1 -84

7

Ga 1347/46

16-2

0-33

1-23

71

13 6

0-35

2-40

Ga 1347/47

14-7

0-32

1-57

7i

13-7

0-29

1-60

Diagnosis. Dextral, spire concave; protoconch spirally ornamented, retiform sculpture
on teleoconch; peristome broadened with inner varices on inner and outer lips.

Description. The umbilicus of the almost perfectly biconcave shell is commonly some-
what more depressed than the upper side. The slightly raised nucleus indicates that the
species is dextral. The fine spiral lines of the protoconch reappear only on the last two
whorls, but may sometimes be seen earlier on the keel. The carinae are more or less
clearly marked and finely corrugated in the nepionic stage. On the last whorl a fine
tuberculation may be developed, as can be seen in Hudlestone’s figure. However, the
keel can remain almost perfectly smooth. Externally the peristome is trumpet-like, and
internally thickened by strong varices on the inner and outer lips.

Discussion. Specimens with smooth keels resemble the inner whorls of D. orbis (Reuss),
from which they probably cannot be separated without knowledge of the protoconch
and peristome: D. orbis having apparently a smooth protoconch, and an un- or only
slightly broadened peristome which is not narrowed internally. The fine retiform sculp-
ture differentiates the present species from D. calculiformis.

Distribution. Aalenian to Lower Bajocian (condensed Opalinum to Sauzei Zone).

EXPLANATION OF PLATE 110

Figs. 1-3. Discohelix ( D .) cf. calculiformis Dunker, Ga 1347/44, x2.

Fig. 4. Discohelix (D.) cf. calculiformis Dunker, Ga 1347/43, x2.

Figs. 5-8. Discohelix (D.) cotswoldiae (Lycett), Ga 1347/46, x 2.

Figs. 9-11. Discohelix (D.) cotswoldiae (Lycett), Ga 1347/47, x2.

Fig. 12. Discohelix ( D .) cotswoldiae (Lycett), Ga 1347/48, x2.

Figs. 13-16. Discohelix (£>.) cf. giimbeli Ammon, Ga 1347/49, x3.

Fig. 17. Discohelix ( Pentagonodiscus ) angusta n. sp., Ga 1347/53, X 3.

Figs. 18-21. Discohelix (Pentagonodiscus) angusta n. sp., holotype, Ga 1347/50, x2.

Figs. 22-24. Discohelix ( Pentagonodiscus ) reussii (Hornes), lectotype, Geol. Bundesanst. Vienna 2980,
x 2.

Palaeontology, Vol. 1!

PLATE 110

WENDT, Discohelix from the Tethyan Jurassic

WENDT: DISCOHELIX (ARCHAEOG ASTROPOD A , EUOMPHALACEA) 573
Discohelix (Discohelix) cf. gumbeli Ammon 1893

Plate 1 10, figs. 13-16; text-fig. 3n

cf. 1893 Discohelix Gumbeli Ammon, p. 215, text-fig. 39.

No. dp h/d ha/wa nw nt sa ua

Ga 1347/49 140 0-38 1-40 6\ 17 222° 142°

Description. That the perfectly biconcave shell is dextral, can only be seen in the some-
what asymmetrical outline of the peristome. The spiral lines of the protoconch disappear
after \ \ whorls with the onset of strong ribs which swell to parabolic nodes on the keel.
Spiral ornament reappears only on the last two whorls and remains restricted to the
keels leaving a smooth band between them on the external side. The peristome, in con-
trast to the growth-lines, lies almost in a shell radius and has thickened margins.

Discussion. A comparison with the holotype must be restricted to its figure, since the
specimen itself, formerly in the Bayer. Staatssammlung Pal. hist. Geol. in Munich,
has been lost. From this figure it appears that D. gumbeli has higher whorls (h/d is said
to be 0-5) and a more asymmetrical cross-section. D. sapplio (Orbigny) is similarly
ornamented, but is plano-concave.

Distribution. Upper Callovian/Lower Oxfordian (together with Distichoceras bicostatum (Stahl)).
The holotype is from the Lias-Dogger boundary.

Discohelix ( Pentagonodiscus ) subgen. nov.

Type species. D. (P.) angusta n. sp.

Diagnosis. Subgenus of Discohelix in which the last 2-3 whorls are pentagonal in
outline owing to periodical deviation of the peristome at intervals of approximately
72°. Inner whorls coiled in a normal logarithmic spiral. Spire concave, dextral. Growth-
lines and ornament as in Discohelix ( Discohelix ).

Discohelix ( Pentagonodiscus ) angusta n. sp.
Plate 110, figs. 17-21; text-figs. 1b, 2h, 3o, 4a-b, 6a-b, 7
Derivation of name, angustus = narrow (referring to the peristome).
Holotype. Ga 1347/50. Plate 110, figs. 18-21; text-fig. 1b.

No.

dp

d

h/d

ha / wa

nw nt sa ua

Ga

1347/50

14-6

0-38

1-38

7 12

Ga

1347/51

12 9

0-33

1-28

6

Ga

1347/52

12-6

0-37

1 25

6 210° 129°

Diagnosis. Protoconch smooth, retiform sculpture, and parabolic ribs on teleoconch;
peristome with inner varices on inner and outer lip.

574

PALAEONTOLOGY, VOLUME 11

Description. The sculpture is similar to that of D. cf. gumbeli , but much more irregular
through the arrangement of the ribs into groups of three. Between the nodes the whorl
cross-section is almost round. In contrast D. (P.) reussii has broad trapezoidal whorls
and is also distinguishable by the failure of ribs. (See also sections ‘Parabolic ribs’ and
‘Growth’).

Distribution. Aalenian to Middle Bajocian (condensed Opalinum to c. Subfurcatum Zone).

Discohelix ( Pentagonodiscus ) reussii (Hornes 1853)

Plate 110, figs. 22-4; text-figs. 2g, 3p, 5
1853 Euomphalus Reussii Hornes; Hauer, p. 760.

1861 Discohelix Reussi Hornes; Stoliczka, p. 184, pi. 3, figs. 13, 14 a-c, non fig. 14 d.
1897 Discohelix Reussi Hornes; Krafft, p. 21 1.

1911 Discohelix Reussi Hornes; Gemmellaro, p. 215, pi. 9, fig. 14.

1935 Discohelix Reussi Hornes, Wahner, p. 22.

Material. Eight specimens, of which five are Stoliczka’s originals.

Lectotype. (here selected) Stoliczka 1861, pi. 3, figs. 14 a-b (here refigured PI. 110, figs. 22-4).

No. d h/d ha/wa nw sa ua

2980 14-5 0-43 1 -97 6

Ga 1347/55 15-8 0-42 L58 6 221° 123°

Diagnosis. Fine retiform sculpture on teleoconch; peristome narrowed with inner varix
on outer lip.

Description. Even without a longitudinal section Stoliczka supposed the occurrence of
inner varices. The aperture, recurring periodically at mean intervals of71-5°(Ga 1347/56)
is depressed on the outer side and somewhat broadened on the upper and lower sides.
This results in regular swellings which are especially strong in the specimen figured by
Stoliczka on pi. 3, fig. 13. In the remaining examples the keel has a regularly undulat-
ing profile. The peristome is only slightly thickened on its inner lip, but narrowed by a
strong inner varix on the outer lip as in D. ( P .) angusta.

Distribution. Sinemurian-Pliensbachian. According to their label Stoliczka's specimens came from the
Oxynotum Zone (Hierlatz) and the Obtusum Zone (Kratzalpe). The horizon of Wahner’s material is
not known; a personal specimen from the same area (Sonnwend Mountains) was found with Arietites
sensu lato. The examples of Krafft and Gemmellaro are said to be younger (Middle Lias), as is that
from the Tubingen Pal. Inst, collection which is the only known extra-alpine occurrence of the species.

CONCLUSION

The stratigraphical division of thick sequences of the Jurassic in the Tethys region is
often made extremely difficult by the sparsity or failure of zonal ammonites, and the
knowledge of other fossil groups is still insufficient to replace them as stratigraphical
indicators. The present investigation of the little-known gastropod genus Discohelix may
help to fill this gap, since, although this genus is only sporadic in the epicontinental
Jurassic, it is occasionally very frequent in the Alpine-Mediterranean area (e.g.

WENDT: DISCOHELIX (ARCHAEOG ASTROPODA, EUOMPH ALACEA) 575

Hochfelln Limestone, Hierlatz Limestone, Liassic Reef Limestone of Sicily). The value
of its species as index fossils, shown here, may make possible the division of such cephalo-
pod free or poor series in the Tethyan Lower and Middle Jurassic.

REFERENCES

ammon, l. 1893. Die Gastropodenfauna des Hochfelln-Kalkes und iiber Gastropoden-Reste aus
Ablagerungen von Adnet, vom Monte Nota und den Raibler Schichten. Geogn. Jh. 5, 161-219.
brosamlen, r. 1909. Beitrag zur Kenntnis der Gastropoden des schwabischen Jura. Palaeontographica,
56, 177-321, pi. 17-22.

cossmann, M. 1915. Essais de Paleoconchologie comparee, 10, 292 pp., 12 pi.

dumortier, E. 1874. Etudes paleontologiques sur les depots jurassiques du bassiit du Rhone. 4me partie.
Lias superieur. 339 pp., 62 pi.

dunker, w. 1847. Ueber einige neue Versteinerungen aus verschiedenen Gebirgsformationen. Palae-
ontographica, 1, 128-33, pi. 18.

fretter, v. and graham, a. 1962. British Prosobranch Molluscs, xvi+755 pp., London.
gemmellaro, M. 1911. Sui fossili degli strati a Terebratula aspasia della contrada Rocche Rosse presso
Galati (prov. di Messina). Cefalopodi (fine) — Gasteropodi. G. Sci. nat. econ. Palermo, 28, 203-46,
pi. 8-10.

hauer, f. 1853. Ueber die Gliederung der Trias-, Lias- und Juragebilde in den nordostlichen Alpen.
Jb. K. K. geol. Reichsanst. 4, 715-84.

hudleston, w. h. 1887-96. A monograph of the Inferior Oolite Gastropoda, being part I of the British
Jurassic Gastropoda. Palaeontogr. Soc. (1886-96), 1-514, pi. 1-44.
knight, j. b. 1934. The gastropods of the St. Louis, Missouri, Pennsylvanian outlier: VII. The Euom-
phalidae and Platyceratidae. J. Paleont. 8, 139-66, pi. 20-6.
koken, e. 1897. Die Gastropoden der Trias um Hallstatt. Abh. K. K. geol. Reichsanst. 17, H. 4, 1-112,
pi. 1-23.

krafft, A. 1897. Ueber den Lias des Hagengebirges. Jb. K. K. geol. Reichsanst., 47, 199-224, pi. 4.
lycett, j. 1850. Tabular view of fossil shells from the middle division of the Inferior Oolite in
Gloucestershire. Ann. Mag. nat. Hist., ser. 2, 6, 401-25, pi. 11.
moore, ch. 1867. On the Middle and Upper Lias of the South West of England. Proc. Somerset,
archaeol. nat. Hist. Soc. 13, 1-128, pi. 1-7.

quenstedt, f. a. 1884. Petrefaktenkunde Deutschlands. 1. Abt. 7. Bd: Gasteropoden. 867 pp., pi. 1 85—
218, Leipzig.

reuss, a. 1852. Ueber zwei neue Euomphalusarten des alpinen Lias. Palaeontographica, 3, 113-16,
pi. 16.

riche, a. and roman, f. 1921. La Montagne de Crussol. Etude stratigraphique et paleontologique.
Trav. Lab. Geol. Univ. Lyon, 1, 1-196, pi. 1-8.

roman, f. 1935. La faune du minerais de fer des environs de Privas. Ibid. 27, 1-52,
pi. 1-8.

schairer, G. 1967. Biometrische Untersuchungen an Perisphinctes, Ataxioceras, Lithacoceras der
Zone der Sutneria platynota (reinecke) (unterstes Unterkimmeridgium) der Frankischen Alb.
Inaug.-Diss. naturw. Pak. Univ. Munchen, 131 pp., 18 pi.
schindewolf, o. h. 1934. Uber eine oberdevonische Ammoneen-Fauna aus den Rocky Mountains.
Neues Jb. Miner. Geol. Paldont. Bei/Bd. 72, Abt. B, 331-50.

1937. Zur Stratigraphie und Palaontologie der Wocklumer Schichten (Oberdevon). Abh. preufi.

geol. Landesanst., n.f. 178, 3-132, pi. 1-4.

stoliczka, F. 1861. Uber die Gastropoden und Acephalen der Hierlatz-Schichten. Sber. Akad. Wiss.
Wien. math, naturw. Cl. 43, 157-204, pi. 1-7.

wahner, f. 1935. Das Sonnwendgebirge im Unterinntal. Ein Typus alpinen Gebirgsbaues. Zweiter Teil
bearb. u. voll. v. e. spengler). xvi+200 pp., 29 pi., Leipzig and Wien.
wendt, J. 1965. Synsedimentare Bruchtektonik im Jura Westsiziliens Neues Jb. geol. Paldont. Mh.
(1965), 286-31 1. j. wendt

Institut fur Geologie und Palaontologie

Typescript received 18 January 1968 Tubingen, Germany

PEDICELLARIAE OF TWO SILURIAN ECHINOIDS
FROM WESTERN ENGLAND

by D. BRYAN BLAKE

Abstract. Pedicellariae have been discovered on the Silurian (Ludlovian) echinoids Echinocystites pomum and
Palaeodiscits ferox from Leintwardine, England. All pedicellariae of both species are tridentate in form and
simple in morphology. Most are interambulacral in position. Pedicellariae of E. pomum are abundant, all are
probably three-valved, and of variable size. Those of P. ferox are relatively rare, apparently two-valved, and small
in size. These pedicellariae, and an Australian occurrence, are the oldest known echinoid pedicellariae. The
unusually large size of one fossil pedicellaria, combined with a general morphology of all pedicellariae inter-
mediate between that of spines and modern pedicellariae suggests that pedicellariae were derived from spines.

While studying materials in the Echinoderm Collections of the British Museum
(Natural History), Dr. J. Wyatt Durham observed pedicellariae on specimens of the
Silurian echinoid Echinocystites pomum and made latex casts of four specimens (three
with pedicellariae) of this species and of two specimens of Palaeodiscus ferox. In subse-
quent studies, the writer observed the presence of pedicellariae on both specimens of
P. ferox. All fossils are from the Ludlow Flags of Leintwardine, Herefordshire. These
occurrences, with that of Philip (1963), who recorded pedicellariae associated with
a lepidocentroid from rocks regarded as Ludlovian in age from New South Wales,
Australia, represent the oldest known echinoid pedicellariae and thus merit detailed
consideration.

The Leintwardine material has been studied by many workers including Gregory
(1897), Thomson (1861), and Hawkins (1927) but the presence of pedicellariae has not
been previously noted in spite of the fact that one of the specimens is the lectotype of
Echinocystites pomum. Mortensen (1928-51) considered pedicellariae important in
echinoid systematics in his monograph of the group but noted and regretted the rarity
of fossil pedicellariae (op. cit., vol. 1, p. 33). He did not mention pedicellariae in his
discussion of either of these two genera (op. cit., vol. 2, pp. 54-6, 60-3). Reports of
the few described Palaeozoic pedicellariae are reviewed by Geis (1936).

Acknowledgements. Dr. Durham made available latex casts of specimens of E. pomum and P. ferox', the
original external moulds of these specimens are in the collections of the British Museum (Natural
History) in London. Dr. Porter M. Kier loaned a specimen of E. pomum belonging to the Geological
Survey of Great Britain which he had prepared for study and identified. Thanks are due to Dr. R. P. S.
Jefferies and the authorities of the British Museum (Natural History) for permitting Dr. Durham to
prepare latex casts of their specimens and to Dr. Jefferies for information on the histories of some of
the specimens.

Echinocystites pomum

All pedicellariae appear to be of the tridentate type, and all probably originally had
three valves. (‘Tridentate’ is one of the four basic types of pedicellariae; most, but not
all, have three valves (Mortensen, op. cit., vol. 2, p. 33).) They are numerous; sixty were
counted on one incompletely preserved test. The blades of a pedicellaria are in contact

[Palaeontology, Vol. 11, Part 4, 1968, pp. 576-9, pis. 1-2.)

BLAKE: PEDICELLARIAE OF TWO SILURIAN ECHINOIDS

577

throughout their length; they are slender and elongate with a circular cross-section, and
are attenuated distally. The blades bear longitudinal striae but apparently no marginal
dentition; their general morphology is essentially identical to that of the spines. The
bases are triangular and grade evenly into the blades. Unlike blades of many tridentate
pedicellariae, there is little or no constriction near the base. The interior of the base is
massive, apparently without a sharp keel-like apophysis. Many heads are located near
to tubercles suggesting that stems were lacking, very short, or had contracted upon
death. Small pedicellariae are most common, but a considerable range in size exists;
there are no distinct size classes. Bases range in diameter from under 0-25 mm. to
about T75 mm. In general, head length in essentially complete examples ranges from
about 0-75 mm. in specimens with small bases to about 3-75 mm. on a specimen of base
diameter of 0-75 mm. However, one unusually large pedicellaria (text-fig. 1a) has a
base diameter of 1-75 mm. and a length slightly over 11 mm. Blade length is not neces-
sarily directly proportional to base diameter because blade length on bases of a given
size is variable.

Only two valves are visible on a number of pedicellariae. However, the great variability
of preservation of the three-valved forms and the essential similarity of morphology
between those with three valves and those with only two implies that the apparent
two-valved pedicellariae are probably incompletely preserved typical three-valved
pedicellariae.

With two possible exceptions, all pedicellariae are interambulacral in position. They
are widely distributed on the interambulacral columns; on one specimen, pedicellariae
occur on plates which border the madreporite, while on another, pedicellariae occur on
plates near the peristome. They may be densely grouped, occurring on neighbouring
plates, or in some cases, more than one on a single plate. There appears to have been no
areal restriction of pedicellariae of certain sizes, for large and small pedicellariae occur
on neighbouring plates.

Material examined. Casts of B.M. (N.H.) BMNH 40158; E34352 (lectotype); 40156; GSM 102622.

Palaeodiseus ferox

Pedicellariae are rare on specimens of this species; only two well-preserved pedicel-
lariae have been observed on each of the two specimens. In addition, a number of
doubtful pedicellariae are present.

All pedicellariae are tridentate in form but with only two valves; there is no evidence
for a third valve. The heads are elongate, triangular, and rounded proximally. The blades,
like those of Echinocystites, are morphologically very similar to the spines; they are
slender, elongate, circular, or sub-circular in cross-section, with longitudinal striae,
and in contact throughout their length. There appears to have been no marginal
dentition. Bases appear to have been triangular and only slightly enlarged. The heads
are very small with a basal dimension of approximately 0-2 mm. and a length of about
0-8 mm. One of the well-preserved pedicellaria is ambulacral in position located near the
groove between the rows of ambulacral plates; the others are interambulacral in position.

Although relatively few pedicellariae can be identified, the similarity of morphology
between the blades and the spines suggests more pedicellariae may be present but are
incompletely preserved or exposed and therefore unrecognized.

Material examined. Casts of Palaeodiseus ferex, BMNH E34360 and E34362.

578

PALAEONTOLOGY, VOLUME 11

DISCUSSION

The presence of previously unrecorded pedicellariae, in one case abundant, on much-
studied echinoid species suggests fossil pedicellariae are probably more common than

text-figs. 1a-g. Echinocystltes pomum pedicellariae. 1, GSM 102622, x7-5. 2-5, 7, B.M. 40156, X 20.
6, from B.M. E34352, x 20. h. Palaeodiscus ferox pedicellaria. B.M. E34360.

previously has been thought. Further, because two families of Silurian echinoids are
involved, evolution of these structures must have been relatively early and fairly wide-
spread.

In many features, the Silurian pedicellariae are spine-like. The valves are circular in
cross-section, the surface ornamentation is similar to that of the spines, marginal
dentition is absent, the base of the blade is not strongly constricted, blades of a pedicel-
laria are in contact throughout their length, the bases are massive without a sharp
apophysis. Further, one unusually long spine-like pedicellaria is present. These
features render the Silurian pedicellariae morphologically intermediate between spines
and modern types of pedicellariae and suggest the possibility that pedicellariae were
derived from spines.

BLAKE: PEDICELLARIAE OF TWO SILURIAN ECHINOIDS 579

REFERENCES

geis, h. l. 1936. Recent and fossil Pedicellariae. J. Paleont. 10, 427-48, pi. 58-61.

Gregory, j. w. 1897. On Echniocystis and Palaeodiscus — two Silurian genera of Echinoidea. Quart. J.
geol. Soc. Loud. 53, 123-36, 7 text-fig., pi. 7.

hawkins, H. L. and hampton, s. m. 1927. The occurrence, structure, and affinities of Echinocystis and
Palaeodiscus. Ibid. 83, 574-603.

mortensen, t. 1928-51. A monograph of the Echinoidea. 5 vols. Copenhagen.

philip, G. m. 1963. Silurian echinoid pedicellariae from New South Wales. Nature, Lond. 200, 1334.

Thomson, w. 1861. On a new Palaeozoic group of Echinodermata. Edinb. New. Phil. J. N.s. 13, 106-18.

D. BRYAN BLAKE
Department of Geology
University of Illinois
Urbana, Illinois 61801
U.S.A.

Typescript received 10 November 1967

MACROCYSTELLA CALLAWAY, THE EARLIEST
GLYPTOCYSTITID CYSTOLD

by C. R. C. PAUL

Abstract. Macrocystella mariae Callaway 1877, type species of Macrocystella, has a stem which is divisible
into proximal and distal portions; a theca composed of 4 basal, 5 infra-lateral, 5 lateral, 6 radial, and some oral
plates ; a large periproct surrounded by 5 thecal plates ; biserial unbranched brachioles grouped into 5 ambulacra
and arising from the margins of the flattened oral surface. In all these respects it agrees with Mimocystites
bohemicus Barrande 1887, type species of Mimocystites which becomes a subjective junior synonym of Macro-
cystella. Macrocystella azaisi (Thoral) has 7 orals and thus Macrocystella differs from the rhombiferan Cheiro-
crinus Eichwald only in the absence of pectinirhombs. The Macrocystellidae are therefore transferred to the
rhombiferan superfamily Glyptocystitida.

Macrocystella evolved into Cheirocrinus by the acquisition of pectinirhombs. In Macrocystella respiration
probably took place through all the thecal plates which are very thin. In Cheirocrinus respiration was restricted
to the pectinirhombs thus allowing much thicker and stronger thecal plates to develop.

Macrocystella led a freely vagrant existence and may have had internal buoyancy devices. The stem did not
provide permanent fixture and may have been used as a organ of locomotion in conjunction with the brachioles.

The cystoidea, as currently defined (Kesling 1963) is probably an artificial group. The
main character which is used to unite the cystoids as a class is the possession of pore-
structures (rhombs and dipores) developed in the thecal plates. However similar pore-
structures are found in at least some representatives of other Palaeozoic echinoderm
classes (blastoids, crinoids, paracrinoids, eocrinoids, for example) and one genus of
rhombiferan cystoids entirely lacks pore-structures. This paper deals with another
genus, Macrocystella Callaway 1877, which lacks true pore-structures but which is
thought to be the oldest known representative of the Glyptocystitida, one of the three
major rhombiferan superfamilies. Macrocystella has a complex taxonomic history (see
below) and has been variously regarded as an eocrinoid, a rhombiferan cystoid or as a
link between these classes. Close comparison indicates that Macrocystella is identical
to the rhombiferan genus Cheirocrinus Eichwald in all details except the possession of
pectinirhombs. Hence Macrocystella is regarded as a rhombiferan. It is believed that the
absence of pectinirhombs in Macrocystella is a primitive character. Many Ordovician
pelmatozoans independently developed thecal or calycinal pore structures, apparently
in response to respiratory needs. To group all such echinoderms together obscures
their true relationships, it is essential to consider other characters in addition to the
possession of pore-structures, especially when the latter are so variable.

Acknowledgements. The author is indebted to Dr. R. P. S. Jefferies and Mr. H. G. Owen, British
Museum, Natural History (BMNH), Professor B. Kummel and Mr. J. Sprinkle, Museum of Compara-
tive Zoology, Harvard (MCZ), Mr A. G. Brighton, Sedgwick Museum, Cambridge (SM), and Dr.
R. A. Robison, United States National Museum, Washington (USNM) for the loan of, or access to,
specimens. Dr. I. Strachan, Birmingham University (BU) supplied latex impressions of Callaway’s
original specimens of M. mariae and Professor K. E. Caster, University of Cincinnati (UC) impressions
of Macrocystella and Cheirocrinus from France and Bohemia. Part of this work was undertaken during
the tenure of a National Environmental Research Council research studentship at the Sedgwick
[Palaeontology, Vol. 11, Part 4, 1968, pp. 580-600, pis. 111-13.]

C. R. C. PAUL: MACROCYSTELLA CALLAWAY 581

Museum, Cambridge and part during the tenure of a Post-doctoral Fellowship at the University of
Michigan (NSF grant GB-3366). Both are gratefully acknowledged.

PREVIOUS RESEARCH

MacrocysteUa (type species M. mariae Callaway 1877) was first described from the
Lower Ordovician (Tremadoc) Shineton Shales of Shropshire. Barrande (1887, p. 163)
described a closely similar genus, Mimocystites, for a single species, M. bohemicus
Barrande. Jaekel (1899, p. 171) suggested that these two genera were synonymous but
used the name Mimocystites. Jaekel regarded Mimocystites as the progenitor of the
cystoids and most closely related to the rhombiferan Cheirocrinus Eichwald. Bather
(1899, p. 920) proposed the family Macrocystellidae which he assigned to the Rhombi-
fera. He did not elucidate the composition of the Macrocystellidae but later (1900, p. 56)
included MacrocysteUa, Mimocystites and Lichenoides Barrande 1887. Although Bather
thought MacrocysteUa and Mimocystites hardly differed he used both names. Bather’s
reconstruction of MacrocysteUa mariae (1900, p. 95, fig. 18) was inaccurate in depicting
branched arms and considerably influenced subsequent opinions on the affinities of
MacrocysteUa.

Jaekel (1918, p. 27) retained both MacrocysteUa and Mimocystites (possibly because
of Bather’s reconstruction of the former) and placed them, with his new genus Polypty-
chella, in the Macrocystellidae. A separate family was proposed for Lichenoides. The
Macrocystellidae and Lichenoidae were assigned respectively to the orders Plicata and
Reducta of the Eocrinoidea. Thoral (1935, p. 113) considered MacrocysteUa and Mimo-
cystites to be distinct but based his opinion of the former on the original description.
Bassler and Moodey (1943) reverted to Bather’s classification, included Lichenoides in
the Macrocystellidae and that family in the Rhombifera. Cuenot (1948) also assigned
MacrocysteUa to the Rhombifera and later (1953) regarded Mimocystites as a junior
synonym. Moore (1954, p. 127, fig. 2d) published an inaccurate plate diagram of
MacrocysteUa which depicts 5 basals, 5 radials, 5 orals, and a minute periproct. Moore
regarded MacrocysteUa as an eocrinoid. Sdzuy (1955) accepted MacrocysteUa and
Mimocystites as separate genera only if published reconstructions were accurate. He
based his opinion of MacrocysteUa on Bather’s (1900) and Moore’s (1954) work, the
accuracy of which he doubted. More recently Prokop (1966) and Ubaghs (1968) have
suggested that MacrocysteUa and Mimocystites are synonymous. Ubaghs regards
MacrocysteUa as most closely related to Cheirocrinus.

Quite apart from the synonymy of MacrocysteUa and Mimocystites another problem
arises as to the status of Cystidea Barrande 1868. Barrande (1867, p. 179) published
two nomina nuda, Cystidea sedgwicki and C. bohemicus. Later he introduced C. bavarica
(Barrande 1868, p. 106) this time accompanied by a description and figures. Barrande
made it quite clear in both publications that he intended Cystidea as a collective group
name not a formal generic name and he so used it again (1887), erecting several more
species. Cystidea bavarica Barrande is a valid binomen and could be construed to be
type species of the genus Cystidea Barrande by monotypy. Pompeckj (1896, p. 90) and
Sdzuy (1955, p. 170) have attributed Cystidea bavarica to MacrocysteUa. There is no
doubt they are correct in this action and therefore Cystidea Barrande 1868, if accepted
as an available generic name, should take precedence over both MacrocysteUa and
Mimocystites. Such action is not in the interests of nomenclatorial stability. No author

582

PALAEONTOLOGY, VOLUME 11

has accepted Cystidea as a valid generic name whereas MacrocysteUa has been widely
used and is figured and described in standard text-books in English, French, and
German. Application has therefore been made to the International Commission on
Zoological Nomenclature for the suppression of Cystidea Barrande 1868 under the
Plenary Powers (Paul 1967fi). In anticipation of a favourable decision, MacrocysteUa
is used throughout this work.

The systematics and composition of the Macrocystellidae and the suggested synonymy
between MacrocysteUa and Mimocystites can only be settled after a revised account of
the morphology of MacrocysteUa mariae has been given. Latex impressions of the
original specimens of M. mariae have been re-examined and additional material studied.
This has been compared with Barrande’s (1887) and Jaekel’s (1899) descriptions and
figures of Mimocystites bohemicus, with latex impressions of some of Barrande’s original
material and additional material of M. bohemicus in the Schary Collection, Museum of
Comparative Zoology, Harvard. Latex impressions of M. azaisi (Thoral) have added
further information.

MacrocysteUa and Mimocystites are identical and quite distinct from Lichenoides to
judge from Ubagh’s (1953) account of L. priscus Barrande. Polyptychella Jaekel was
founded on isolated plates and its systematic position cannot be settled without further
information. The Macrocystellidae thus contains the single genus MacrocysteUa. As
previously stated the Macrocystellidae is assigned to the Rhombifera (Glyptocystitida).

SYSTEMATIC PALAEONTOLOGY
Superfamily glyptocystitida Bather 1899

Diagnosis. A superfamily of Rhombifera with well-developed stem divided into proximal
and distal portions; theca composed of 4 basals, 5 infra-laterals, 5 laterals, 4-6 radials,
and 7 orals; with pectinirhombs (when pore structures are developed).

All Glyptocystitida are characterized by a theca composed of 25-7 thecal plates
arranged in five circlets termed basal, infra-lateral, lateral, radial, and oral. All but two
genera — MacrocysteUa and Amecystis Ulrich and Kirk — have pectinirhombs. These
characters distinguish glyptocystitids from members of the other two major rhombi-
feran superfamilies, the Hemicosmitida and Caryocystitida. The former have thecal
plates arranged in three or four circlets and slightly different rhombs. The latter have
a large variable number of plates, some of which may be added during growth, and a
completely different type of rhomb (Paul 1968).

Family macrocystellidae Bather 1899 emend. Jaekel 1918

Diagnosis. A family of Glyptocystitida without pectinirhombs; with cylindrical theca
having 6 radials; large periproct surrounded by 5 thecal plates and covered with a
flexible plated integument; brachioles confined to oral surface, grouped into 5
ambulacra.

Genus macrocystella Callaway 1877

1868 Cystidea Barrande, p. 106.

1877 MacrocysteUa Callaway, p. 669, pi. 24, fig. 13.

C. R. C. PAUL: MAC ROCYSTELLA CALLAWAY

583

1880 Mcicrocystella Callaway; Zittel, p. 420.

1887 Mimocystites Barrande, p. 163, pi. 28 (1), figs. 1-20.

1891 Macrocystella Callaway; Carpenter, p. 13.

1891 Mimocystis [,s7c] Barrande; Carpenter, p. 13.

1896 Mimocystis [s/c] Barrande; Haeckel, p. 149.

1899 Macrocystella Callaway; Jaekel, p. 171.

1899 Mimocystites Barrande; Jaekel, p. 172, fig. 33.

1900 Macrocystella Callaway; Bather, p. 56, fig. 18.

1900 Mimocystis [s/c] Barrande; Bather, p. 56.

1913 Macrocystella Callaway; Springer, p. 157.

1918 Macrocystella Callaway; Jaekel, p. 27.

1918 Mimocystites Barrande; Jaekel, p. 27.

1935 Mimocystites Barrande; Thoral, p. 1 10, 1 13.

1943 Macrocystella Callaway; Bassler and Moodey, p. 6.

1943 Mimocystites Barrande; Bassler and Moodey, p. 6.

1948 Macrocystella Callaway; Regnell, p. 11.

1948 Macrocystella Callaway ; Cuenot, p. 18, fig. 17.

1953 Macrocystella Callaway; Cuenot, p. 619.

1953 Mimocystites Barrande; Choubert, Termier, and Termier, p. 137.

1954 Macrocystella Callaway; Moore, p. 127, fig. 2a.

1954 Mimocystites Barrande; Termier and Termier, p. 92, figs. a-e.

1955 Macrocystella Callaway; Sdzuy, p. 269.

1966 Macrocystella Callaway; Prokop, p. 820.

non Cvstidea Barrande 1867 (nomen nudum) nee Barrande 1887 nee Haeckel 1896 (inde-
terminate echinodernt fragments).

Diagnosis. As for family.

Regional distribution and stratigraphic range. Macrocystella is recorded from the Tremadoc of England
and Wales (M. mariae Callaway), Bavaria (M. bavarica (Barrande) 1868), Bohemia (M. bohemicus
Barrande 1887), and France (M. azaisi (Thoral) 1935). Macrocystella is also recorded from Greenland,
the South American Cordillera and Korea (for detailed references see Regnell, 1948, pp. 11-12).
Choubert, Termier, and Termier record Macrocystella from the Llandeilo of Morocco. Available
specimens confirm the genus from the Llandeilo of Pu-piao, Northern Shan States, Burma, (SM),
from the Tremadoc of Mexico (USNM) and possibly from the Caradoc of Corwen, Wales, and
Girvan, Scotland (BMNH). M. pachecoi Melendez (1944) from the Ashgill of Aragon, Spain is
probably a Heliocrinites. Macrocystella thus ranges from the Tremadoc to the Llandeilo and possibly
Caradoc (Lower-Middle Ordovician).

Macrocystella mariae Callaway 1877

Plate 111, figs. 1, 3-6; Plate 1 12, figs. 1-3, 5-10; Plate 113, fig. 2

1877 Macrocystella mariae Callaway, p. 670, pi. 24, fig. 13.

1896 Macrocystella mariae Callaway; Haeckel, p. 149, pi. 4, fig. 30.

1900 Macrocystella mariae Callaway; Bather, p. 56, fig. 18.

1905 Macrocystella mariae Callaway; Fearnsides, p. 617.

1911 Macrocystella mariae Callaway; Kirk p. 16, pi. 2, fig. 17.

1913 Macrocystella mariae Callaway; Springer, p. 157, fig. 249.

1927 Macrocystella mariae Callaway; Stubblefield and Bulrnan, pp. Ill, 118.

1943 Macrocystella mariae Callaway; Bassler and Moodey, pp. 27, 175.

1952 Macrocystella mariae Callaway; Termier and Termier, p. 363, fig. 9.

1953 Macrocystella mariae Callaway; Cuenot, p. 618, fig. 15.

1955 Macrocystella mariae Callaway; Sdzuy, p. 270, pi. 1, fig. 14.

1964 Macrocystella mariae Callaway; Castell, p. 58, pi. 3, fig. 6.

584

PALAEONTOLOGY, VOLUME 11

Diagnosis. A species of Macrocystella of small size; with 10-15 brachioles, triangular in
section; circular outer proximal columnals with thin, blade-like external flanges.

Type. BU 409 (PI. 113, fig. 2) is selected as lectotype. It is possibly the original of Callaway 1877, pi. 24,
fig. 13 and is from the Shineton Shales of Shineton, Shropshire. Parts of other specimens on this slab
are accepted as paralectotypes.

Horizon and locality. Stubblefield and Bulman (1927) record M. marine from the Clonograptus tenellus
and Slnanardia pusi/la zones (Middle and Upper Tremadoc respectively) of the Wrekin. M. mariae is
also recorded from the Slmmardia pusilla zone of Arenig (Fearnsides, 1905, p. 617) and Macrocystella
sp. from the same zone near Portmadoc, Caernarvonshire (Fearnsides 1910, pp. 161-2).

Material. Crushed remains of four more or less complete thecae, one complete stem and many isolated
thecal plates and fragments.

Description, a. Stem

The stem has a proximal and a distal portion. One complete proximal portion (PI. 112,
fig. 8) has 20 outer proximals each with a blade-like unornamented external flange.
This crushed portion tapers from 5 mm. adorally to 2 mm. in approximately 15 mm. At
the junction with the distal stem small distal columnals appear between the flanged
columnals. Throughout the preserved portion of the distal stem, flanged and unflanged
columnals alternate but this alternation becomes less obvious distally.

The distal portion of the stem (PI. Ill, fig. 1) tapers gradually from 1 mm. proximally
to 0-5 mm. at the tip. It is about 35 mm. long. The topmost distal columnal is a thin
annulus; the terminal distals are cylindrical and about three times as high as wide. There
is no alternation of flanged and unflanged distals in this stem. Preserved proximal stems
are straight or quite strongly curved. Curved distal stems are also preserved but there
is no evidence to support Bather’s (1900) interpretation of the distal stem with a distinct
distal coil.

Counterparts of isolated columnals indicate the mode of construction and articula-
tion of the stem. Both proximal and distal portions are composed of two types of
columnals. Outer proximals (text-figs. 1 a-b) are annular. Each has a smooth, sharp-
edged outer flange and a narrow inner flange (text-fig. 2, PI. 1 12, fig. 7). Outer proximals
alternate with inner proximals and the latter abut against the inner flanges of the former
(text-fig. 2). The inner wall of each outer proximal has two sockets set opposite each
other on both the upper and lower surfaces. The inner flanges are thickened adjacent to
these sockets to form fulcra (text-fig. 2). If the sockets on one surface are orientated
N.-S., those on the opposite surface of the same columnal lie NW.-SE. (text-figs.
1 a-b). Both upper and lower surfaces of the inner proximals are flattened to form facets.

EXPLANATION OF PLATE 111

Stereophotos of Macrocystella mariae Callaway and M. azaisi (Thoral)

Figs. 1, 3-6. M. mariae Callaway. 1. Complete stem showing proximal and distal portions. BMNH
E291 13. 3. Left lateral view of crushed theca. BMNH E29110a. 4. Right lateral view of same theca.
BMNH E29109a. 5. Anterior lateral view of another crushed theca. BMNH E29109b. 6. Posterior
view of crushed theca to show outline of periproct and small periproctal plates. BMNH E29113.
Fig. 2. M. azaisi (Thoral). Proximal and part of distal stem to show ornament of flanges on outer
proximals. BMNH E23697.

All figures of latex impressions whitened with ammonium chloride sublimate. All X 2.

Palaeontology, Vol. 11

PLATE 111

PAUL, Macrocystella

C. R. C. PAUL: MACROCYSTELLA CALLAWAY

585

at the same two opposite points (text-figs, lc, 2). The outer margin of an inner proximal
protrudes at these points and the protrusions key into the sockets in the inner wall of
the outer proximals above and below (text-fig. 2). The fulcra on the outer proximals and
the facets on the inner proximals articulate and the axis of articulation changes by
approximately 45° with each outer proximal. This results in a right-handed spiral
arrangement of articulation facets in M. cizaisi and presumably in other MacrocysteUa.

outer proximal columnal (OP) in the same orientation to show different orientations of fulcra (Fu).
c. Inner proximal columnal (IP). Ef, external flange; Fa, facet; If, internal flange.

text-fig. 2. Diagrammatic reconstruction of part of proximal stem of MacrocysteUa mariae Callaway
to show arrangement of outer (OP) and inner (IP) columnals and spiral arrangement of facets (Fa)
and fulcra (Fu). Ef, external flange; If, internal flange; IW, inner wall of outer proximal columnal.
This has been drawn as a left-handed spiral although M. azaisi is known to show a right-handed spiral.

Each inner proximal is keyed into the sockets of the outer proximals above and
below. Thus although highly flexible, the stem was quite resistant to rotation about its
axis. Both inner and outer proximals are annular and the proximal stem has a wide
lumen. Flexing of the proximal stem was probably achieved by muscles housed in this
lumen. The mechanical keying of the columnals prevented rotation which would have
sheared such muscles.

Larger (sometimes flanged) and smaller (unflanged) distals alternate in the distal stem
but this becomes less apparent distally. The most proximal distals, which appear to be

586 PALAEONTOLOGY, VOLUME 11

newly formed, are annular but most distals are cylindrical. All distals have narrow lumina
and the articulating surfaces are smooth (PI. 112, fig. 3).

The nature of the outer proximals may prove to be a useful specific character. In
M. mariae the outer proximals are approximately circular with thin, blade-like outer
flanges (PI. 1 1 1, fig. 1, PI. 112, fig. 7). In M. bavarica the outer flanges are also thin and
sharp-edged but are produced into ten angles or incipient spines (Sdzuy, 1955, pi. 1,
figs. 8-10, text-fig. 1//). In M. azaisi the outline is circular but the flanges are thicker and
have fine irregular granules or spines encircling them (PI. 11 1, fig. 2, PI. 1 13, figs. 1, 8).
The outer flanges of M. sp. nov. from Mexico have four rounded lobes arranged in
two pairs.

b. Theca

All known thecae are crushed and it is impossible to describe all thecal plates from
one specimen. BMNH E29 109-10 are counterparts which show two crushed thecae in
different orientations. There is another theca which shows details of the periproct.

text-fig. 3. Diagrammatic reconstruction of plate arrangement of Macrocystella
mariae Callaway. Based on BMNH E29109a-b, E29110a-b, and E29113. B1-B4
basals, IL1-IL5 infra-laterals, L1-L5 laterals, R1-R6 radials.

A composite plate arrangement and a reconstruction based on these specimens are
depicted in text-figs. 3 and 15. The theca was composed of five circlets of plates, four
of which can be seen in M. mariae. The subvective system was confined to the oral
surface from the margins of which the brachioles arose in five groups.

Some details of text-fig. 3 are restored but all plates shown existed. The four basals
(BB) unite aborally to form an invagination around the stem. One basal (presumably
B4) was hexagonal (PI. 112, fig. 2). The infra-laterals (ILL) form a closed circlet; IL4
and IL5 contribute to the periproct border. IL1, IL2, and IL3 are roughly hexagonal.
The five laterals (LL) apparently form a closed circlet; LI, L4, and L5 contribute to
the periproct border. L5 is distinctly smaller than the other laterals and has a radial (R5)

C. R. C. PAUL: MAC ROCYSTELLA CALLAWAY

587

directly adoral to it. One pair of counterparts (PI. Ill, figs. 3, 4) seem to have three
hexagonal laterals which means either there were six laterals in this specimen or only
LI and L5 contributed to the periproct border. Unfortunately the relevant portion of
this theca is crushed and this appearance may be misleading. The six radials (RR)
form a closed circlet; one is directly adoral to L5. There are seven orals (00) in M.
azaisi.

text-figs. 4-5. Camera lucida drawings of the periproct (Pe) of Macrocystella mariae Callaway. 4.
BMNH E291 10b cf. PI. 2, fig. 9. 5. BMNH E29113 cf. Plate 1, fig. 6. IL4-IL5 infra-laterals, LI, L4, L5

laterals.

The mouth, gonopore, and hydropore have not been detected in M. mariae but were
all on the oral surface in M. azaisi (text-figs. 1 1 a, b ). Critical details of the periproct show
in BMNH E291 10b, and more clearly, in BMNH E29113 (text-figs. 4, 5). There are five
plates around the periproct which is large. The periproct was covered by a thin, flexible,
plated integument in life. Some small periproctal plates are preserved (PI. Ill, fig. 6)
but the position of the anal pyramid in unknown.

The outlines of the individual thecal plates vary with their positions in the theca.
Periproct border plates can be recognized easily, as can basals and radials. The remain-
ing five laterals and infra-laterals are difficult to distinguish from each other. All plates
have raised umbones from which ridges radiate to the middles of the sides, connecting
centres of adjacent plates (text-fig. 6). Auxiliary ridges are developed parallel to the
primary ridges to form ‘rhombs’. These ‘rhombs’ are not true rhombs as they are
composed of folds in the thecal plates. There is no development of thin-walled thecal
canals. The external ridges of Macrocystella are formed by the folds in the plates and
are not solid strengthening struts such as occur in Cheirocrinus. The thecal plates of
M. mariae are approximately OT mm. thick.

Q9

C 5934

588

PALAEONTOLOGY, VOLUME 11

text-fig. 6. Isolated thecal plate of Macrocystella
mariae Callaway to show folds. BMNH E29119.

c. Subvective system

The exothecal portion of the subvective system consists entirely of brachioles. These
are long, slender, biserial, unbranched structures. The most complete brachioles (PI. 112,
fig. 10) have approximately 120 brachiolar plates and are slightly longer than the thecal
height (up to 16 mm.). The brachioles have a triangular cross-section with sides twice
the width of the adoral surface (text-fig. 7). The food groove ran down the centre of the
adoral surface and was covered by lappets in life. The lappets are between one and a half
times and twice as numerous as the brachiolars and apparently were flexible. Some have
been preserved covering the food groove; others in an ‘open’ position (PI. 112, fig. 10).
The lappets alternate and each one imbricates over its more distal neighbour (text-fig. 8).

EXPLANATION OF PLATE 112

Figs. 1-3, 5-10. M. mariae Callaway. 1. Internal and external views of two isolated thecal plates. The
internal view (above) shows the folds in the plates. The other plate was one of five bordering the
periproct (probably LI). BMNHE291 12, X 5. 2. External view of B4 and an isolated distal columnal.
BMNHE29112, x5. 3. Three isolated thecal plates. All originally bordered the periproct. BMNH
E29112, x 5. 5. Internal mould of isolated thecal plate BMNH E29 11 9, X3.6. Inner proximal colum-
nal within outer proximal columnal. BMNH E291 12, x 2. 7. Outer proximal columnal showing inner
flange and fulcra. BMNH E7574, X 1 -5. 8. Crushed stem and theca. Note the larger distals are flanged.
BMNH E7574, X 1-5. 9. Stereophotos of posterior lateral view of crushed theca showing periproct.
Counterpart to Plate 1, fig. 5. BMNH E29110b, x2. 10. Detail of brachioles to show lateral, ab-
and adoral views. BMNH E29109, X 3.

Fig. 4. M. aiaisi (Thoral). Lateral view of theca and stem. BMNH E23697, X 2.

All except Fig. 5 latex impressions, all whitened with ammonium chloride sublimate.

Palaeontology , Vol. 11

PLATE 112

PAUL, Macrocyste/la

C. R. C. PAUL: M AC RO CYSTELLA CALLAWAY

589

They coalesce towards the margins of the brachiole to form a continuous narrow band
(text-fig. 9). The free portions were able to curl in on themselves. This curling in is not
preservational and the lappets may have been only partially calcified in life.

7

text-figs. 7-9. Brachioles of Macrocystella marine Callaway. 7. Diagrammatic section through
brachiole without lappets. 8. Camera lucida drawing of portion of brachiole with lappets (La) closed
over food groove. Note that the lappets alternate and imbricate. 9. Camera lucida drawing of aboral
view of brachiole with lappets in ‘open’ position. Br brachiolar plate. Text-figs. 8-9 based on BMNH

E29113.

No more than three brachioles arise in any one ambulacrum and there were presum-
ably 10-15 brachioles in all. There is no evidence to support Bather’s (1900, fig. 18)
reconstruction of branched brachioles. BMNH E29110a shows three brachioles in one
radius (PI. Ill, fig. 3): all three are separate entities from their origin at the margin of
the theca.

Macrocystella mariae Callaway is characterized by the following: 1 . A stem which is
divisible into two portions: a short, rapidly tapering, highly flexible, proximal portion
composed of two types of annular columnals with a wide lumen; and a long distal
portion composed of cylindrical columnals with a narrow lumen.

2. A theca with plate formula 4BB, 5ILL, 5LL, 6RR, 700.

3. A large periproct surrounded by five thecal plates and covered with a flexible
plated integument.

4. Biserial, unbranched brachioles which arise from the margins of the flat oral
surface and are grouped into five ambulacra.

COMPARISON WITH MIMOCYSTITES BARRANDE

Mimocystites boliemicus Barrande has a subvective system which is confined to the
flat oral surface and consists of about twenty brachioles grouped into five ambulacra.

590

PALAEONTOLOGY, VOLUME 11

Three brachioles were present in one radius of an available specimen (UC latex impres-
sion of original of Barrande 1887, pi. 28 (i), fig. 14). The theca is composed of 4BB,
5ILL, 5LL, 6RR, and some 00. Jaekel’s analysis of the arrangement of thecal plates
(1899, p. 201, fig. 36) is reproduced here in a slightly modified form (text-fig. 10). This
interpretation differs slightly from that of Macrocystella mariae. Only two laterals

text-fig. 10. Jaekel’s (1899) interpretation of the plate arrangement of
Mimocystites bohemicus Barrande with modern notation of plates. B1-B4
basals, IL1-IL5 infra-laterals, L1-L5 laterals, R1-R6 radials. Cf. text-fig. 3.

(LI and L5) contribute to the periproct border and LI has a straight upper border with
R6 directly adoral to it. This is an unexpected arrangement. Two specimens in the
Schary collection (MCZ) apparently show the arrangement figured for Macrocystella
mariae. However one specimen of M. mariae (Counterparts BMNH E29109a and 291 10a)
has apparently three hexagonal laterals as shown in Jaekel’s figure of Mimocystites
bohemicus. It seems possible that the plate arrangement varied slightly in different
specimens. Mimocystites azaisi Thoral has a plate arrangement identical to that in
Macrocystella mariae. Both the present and Jaekel’s interpretations agree in most
respects, particularly in the five plates around the periproct.

EXPLANATION OF PLATE 113

Stereophotos of M. mariae Callaway, M. azaisi (Thoral) and M. azaisi multicristata (Thoral).

Fig. 2. M. mariae Callaway. Lateral view of lectotype. BU 409.

Figs. 1, 3, 5, 8. M. azaisi (Thoral). 1. Anterior lateral view of theca to show well-developed folds in
B2 and ornament of stem flanges. CU. 3. Lateral view of another theca. CU. 5. Oral view of same.
8. Lateral view of another theca to show well-developed folds in B2 and ornament of stem flanges.
CU.

Figs. 4, 6, 7. M. azaisi multicristata (Thoral). 4. Oblique oro-lateral view to show ornament of orals
and lateral food grooves alternating in ambulacrum IV (left). CU. 6. Interior view of oral surface
of same theca. 7. Lateral view of same theca to show more strongly developed folds in radial plates.
All figures of latex impressions whitened with ammonium chloride sublimate. All X 2.

Palaeontology, Vol. 11

PLATE 113

PAUL, Macrocystella

C. R. C. PAUL: MAC ROCYSTELLA CALLAWAY 591

The thecal plates of Mimocystites bohemicus and M. azaisi are identical to those of
MacrocysteUa mariae except in the number of folds, which is variable in each species.

The proximal stem of Mimocystites bohemicus is identical to that of MacrocysteUa
mariae except that the outer proximals have thicker flanges which are less blade-like.
Mimocystites azaisi has still thicker flanges, the peripheries of which are granulose or
spinose (PI. 1 1 1 , fig. 2, PI. 113, figs. 1 , 8).

in

A

text-fig. 11. Camera lucida drawing of oral surface of MacrocysteUa azaisi (Thoral). a, entire surface
to show oral plates and arrangement of food grooves. Cf. Plate 113, fig. 5. CU. b, detail of gonopore and
hydropore area of same. FG, food groove; G, gonopore; H, hydropore; LFG, lateral food groove;
M, position of mouth which was probably much larger than shown; 01-07, orals. I-V, ambulacra;
l1, I2, I3, facets of ambulacrum I; IV1, IV2, facets of ambulacrum IV.

The type species and one other species of Mimocystites therefore exhibit all the features
which characterize MacrocysteUa mariae. All three are congeneric.

Specimens from the Montagne Noire, France (UC latex impressions) have yielded
additional information on the morphology of MacrocysteUa. One example of M. azaisi
(Thoral) has an almost complete oral surface (text-fig. 11a, PI. 113, fig. 5). Another

592

PALAEONTOLOGY, VOLUME 11

example of M. azaisi multicristata (Thoral) shows both the internal and external
surfaces of the oral area (text-fig. 12, PL 113, figs. 4, 6). There were seven orals, arranged
as shown in text-fig. 1 \a. A slit-like hydropore and an oval gonopore are developed
across the common suture of Ol and 07 (text-fig. Mb). There are five main ambulacral
grooves each of which has lateral branches leading to brachiole facets. Apparently the
branches are consistently to the left of ambulacra I and IV in the example of M. azaisi
(PI. 113, fig. 5) but regularly alternate in ambulacrum IV in the example of M. azaisi
multicristata (PI. 113, fig. 4). Details of the other ambulacra are not well preserved in
either specimen.

In internal view the mouth is large
(4-6 mm.xl-7 mm.) and is covered
by ambulacral cover plates. The orals
and ambulacral flooring plates are
visible (PI. 113, fig. 6). The latter are
between, not on, the orals and form
part of the thecal wall. Only primary
ambulacral flooring plates can be
detected. Close to the mouth is a deep
pit which apparently connected to
the hydropore. This pit is separated
from the mouth and the supposed
gonopore by two internal ridges, one
on each side. The gonopore is ap-
parently represented by a small
circular pit some distance from the
mouth (text-fig. 12). Unfortunately
the critical area of the external oral
surface is not preserved and it is not
possible to match up external and
internal openings. Ambulacra IV
and V are more deeply impressed than I, II, and III. Another pit is developed obliquely
under the aboral of the two internal ridges near the hydropore (PI. 113, fig. 6). The
significance of this is unknown.

COMPARISON WITH OTHER PELMATOZOA

Several authors have grouped Macrocystella with Lichenoides while others have placed
them in separate families. In addition Macrocystella has been variously assigned within
the Pelmatozoa.

The most recent and most complete account of the morphology of Lichenoides is that
of Ubaghs (1953) who showed that it completely lacks a stem. The thecal plates are
arranged in four circlets but the total number is variable, 5-12 basals, 5 infra-laterals,
5-7 laterals, and 5-7 radials. (The homologies with the Glyptocystitida implied by the
use of the same terms for the plate circlets are unjustified.) All the infra-laterals, laterals,
and radials bear epispires. These are a type of pore-structure with a single sutural pore
which leads to a narrow channel in the external surface of both adjacent plates (see

III

text-fig. 12. Camera lucida drawing of internal oral
surface of Macrocystella azaisei multicristata (Thoral).
cf. Plate 1 1 3, fig. 6. CU. G, supposed internal opening
of gonopore; H, supposed internal opening of hydro-
pore; M, mouth; 01-07, orals; I-V ambulacra.

C. R. C. PAUL: MACROCYSTELLA CALLAWAY

593

Ubaghs 1953, figs. 3, 11). There is no lateral periproct in Lichenoides. The brachioles
arise from both lateral and radial plates and apparently they are not grouped into five
radii.

Thus while Lichenoides resembles MacrocysteUa in having definite plate circlets and
biserial, unbranched brachioles it differs in the absence of a stem and lateral periproct,
in the presence of epispires and brachioles on lateral and radial plates, and in the total
number and position of the thecal plates. These differences are considered to be impor-
tant taxonomically. The Lichenoidae and Macrocystellidae are maintained as separate
families as suggested by Jaekel (1918) and Ubaghs (1953).

text-fig. 13. Plate arrangement in Cheirocrinus radiatus Jaekel. Based on Jaekel, 1899,
p. 213, fig. 36. B1-B4, basals; IL1-IL5, infra-laterals; L1-L5, laterals; R1-R6, radials.

Among cystoids MacrocysteUa most closely resembles Cheirocrinus Eichwald. This
latter genus is characterized by a stem with proximal and distal portions. The construc-
tion and articulation of the stem are identical to that of MacrocysteUa and the helical
arrangement of the articulations of the proximal stem was described by Billings (1858)
in Cheirocrinus anatiformis (Hall) (= Glyptocystites logani Billings).

The theca of Cheirocrinus is composed of 27 plates arranged in 5 circlets: 4BB, 5ILL,
5LL, 6RR, and 700. The periproct is large, surrounded by 5 thecal plates (1L4, IL5, LI,
L4, and L5) and was covered by a flexible plated integument in life. L5 is directly adoral
to the periproct and has R5 directly adoral to it (text-fig. 13). The subvective system is
restricted to the flat oral surface and there are 20-5 brachioles grouped into 5 ambulacra
whose flooring plates lie between the orals, not on them. Cheirocrinus differs from Macro-
cysteUa in the possession of pectinirhombs.

MacrocysteUa and Cheirocrinus have in common many distinctive features of which
perhaps the most important is the detailed structure of the stem. This type of stem is
characteristic of and confined to the superfamily Glyptocystitida. It is most unlikely
that such a complex organ developed independently in two groups which share other
common characteristics. MacrocysteUa probably gave rise to Cheirocrinus and through
it to the other Glyptocystitida as originally suggested by Jaekel (1899).

594

PALAEONTOLOGY, VOLUME 11

Thoral (1935) thought it possible to recognize in the Montague Noire a lower horizon
with Macrocystella azoisi (upper Tremadoc) and a higher horizon with Cheirocrinus
languedocianus (basal Arenig). In Britain M. marine occurs below the oldest Cheiro-
crinus. Macrocystella seems to be a characteristic fossil of the Tremadoc whereas the
oldest known Cheirocrinus are all Arenig. Stratigraphic evidence agrees with the idea
that Macrocystella evolved into Cheirocrinus.

In the past the main objections to the inclusion of Macrocystella in the Rhombifera
were its branched arms and lack of rhombs. The former was an error but the latter is
more important. Regnell (1945) has stressed that the main character which unites the
cystoids as a class is the presence of pore-structures. Detailed study of cystoid pore-
structures (Paul, 1968) suggests the Rhombifera should be regarded as a distinct class.
The rhomb-less Macrocystella is included in the Rhombifera on the same grounds that
led Kesling (1963) to include Amecystis in the Rhombifera. Amecystis is effectively a
Pleurocystites without pectinirhombs, just as Macrocystella is a Cheirocrinus without
pectinirhombs. The many similarities outweigh this single distinction.

THE EVOLUTION OF PECTINIRHOMBS

The rhombs of Cheirocrinus are fully developed pectinirhombs (which Bather regarded
as a highly specialized type of rhomb) even in the earliest species known. There is no
evidence of a gradual evolution of pectinirhombs. If Macrocystella evolved into Cheiro-
crinus, pectinirhombs either appeared ‘ suddenly ’ or pre-existing structures broke through
the thecal plates to appear as pectinirhombs. The internal surfaces of isolated plates in
both Macrocystella and Cheirocrinus show no features which could be incipient pectini-
rhombs. Pectinirhombs were functional throughout their growth; they are present as
external features from the earliest stages. They did not develop internally and become
external features later in growth. Rather sudden appearance therefore seems more
likely.

All rhombs have generally been accepted as respiratory organs. In the simplest case
respiration would have taken place through the thecal wall. The amount of oxygen
required would have been proportional to the volume, and the amount of respiratory
exchange to the surface area, of the theca. The oxygen requirements would thus increase
with growth faster than the amount of exchange. This difficulty can be overcome, without
materially altering the over-all thecal shape, by the production of evaginations or
invaginations of the thecal wall. Macrocystella has the former in the folds of the thecal
plates. The dichopores of pectinirhombs are invaginations and produce a slightly better
volume to surface area ratio.

Exchange is facilitated by a large surface area and a thin exchange surface. Either
invaginations or evaginations are almost equally effective in increasing the surface area
but the latter are exposed and liable to mechanical damage. The ridges in the thecal plates
of Macrocystella probably facilitated exchange by increasing the surface area. The
thecal plates were extremely thin (OT mm.) and although strengthened by the ridges they
were still very fragile. The dichopores of pectinirhombs are within the theca and there-
fore protected. Dichopore walls are much thinner than thecal plates (usually 0-01 mm.).
In Macrocystella the entire thecal wall probably took part in exchange. In Cheirocrinus
a differentiation of function is seen. Without decreasing the amount of exchange it

C. R. C. PAUL: MACROCYSTELLA CALLAWAY

595

became possible to have much thicker and stronger thecal plates. Exchange was re-
stricted to specialized areas, namely the pectinirhombs. The thecal plates of Cheiro-
crinus, especially in early species like C. languedocianus, are still thin compared with
later glyptocystitids but they are thicker (usually more than 0-5 mm.) than those of
Macrocystella.

text-fig. 14. Pectinirhombs of a specimen of Cheirocrinus languedocianus
Thoral. Note the incomplete pectinirhombs. Based on camera lucida
drawings of the individual plates. CU. Symbols as in text-fig. 3.

In Macrocystella all thecal plates are ridged and all probably contributed to respira-
tion. However the distribution of ridges in M. azaisi is uneven. More and better-
developed ridges occur on radial plates (PI. 113, fig. 7) and associated with B2 (PI. 113,
figs. 1, 8). This is quite independent of the variation between M. azaisi and M. azaisi
multicristata. Sdzuy (1955, figs. 1 d-g) figured a similar concentration of ridges on B2
in M. bavarica. In Cheirocrinus languedocianus , probably the earliest known species of
Cheirocrinus , dichopores are developed across more sutures than in later species
(text-fig. 14). Many sutures bear one or two demi-rhombs and some have only a few
randomly spaced dichopores. These latter form incomplete pectinirhombs which have
not been recorded in later glyptocystitids. Clearly the arrangement is more random than

596

PALAEONTOLOGY, VOLUME 11

in later species which have a reduced number of pectinirhombs. All species of Cheiro-
crinus have pectinirhombs on radials and on B2 however. The similarity between the
distribution of ridges in Macrocystella and of pectinirhombs in Cheirocrinus indicates
a concentration of respiratory activity in the same areas of their thecae, probably
reflecting a similar internal organization, and confirming the idea that Macrocystella
and Cheirocrinus are related.

MODE OF LIFE

The stem of Macrocystella lacks a root structure or anchoring device and there is
little evidence for Bather’s reconstruction of a distal coil to the stem. Whatever benefits
were conferred on Macrocystella by the possession of a well-developed stem, they were
not associated with permanent attachment. M. mariae and other species of Macro-
cystella are found in fine-grained sediments in which there is little evidence for suitable
attachment surfaces. Nevertheless the general morphology of the theca and subvective
system suggests Macrocystella habitually held its theca upright. Only free-floating or
free-swimming modes of life can satisfy both lines of reasoning.

The distal stem in Macrocystella and in other early glyptocystitids tapers gradually
to a small diameter and ends abruptly. The terminal portion of the stem is not modified
as in Brockocystis Foerste or Lepocrinites Conrad and there is no way to be certain that
the last distal columnal preserved was the terminal columnal in life. It is possible that
Macrocystella broke free prior to death and was buried some distance from its point of
attachment. However M. mariae and other apparently free glyptocystitids are frequently
preserved in clay-grade sediments with little associated fauna which could have acted
as substrata for attachment. No Macrocystella , Cheirocrinus nor Pleurocystites has ever
been found with a root structure or anchoring device, yet many specimens are excellently
preserved with brachioles, periproctal membranes, and other delicate structures intact.
Some specimens from the Trenton Limestone of Ottawa and the Starfish Bed of Girvan,
Scotland (for example) suggest that death was due to burial alive and therefore that
the specimens are complete. The terminal diameter of the distal stem (0-5 mm.) compared
with the total length (50 mm.), in M. mariae indicates that if attached the stem was
extremely weak at this point. Macrocystella may have been attached early in its develop-
ment but the evidence strongly suggests that it was free for the major portion of its life.

Pentameral symmetry is well developed in Macrocystella'. there are circlets of five
plates and five ambulacra. The theca is cylindrical not flattened on one side. Bather
argued that pentameral or radial symmetry was developed in fixed animals and he
associated departures from pentamery in echinoderms, particularly flattening of the
theca, with departures from an upright fixed position. Pleurocystites , another free
glyptocystitid, has a markedly flattened theca which apparently lay on the sea floor. If
Macrocystella and Cheirocrinus also lay on the sea floor one would expect a similar
flattening of their thecae. The persistance of Macrocystella and Cheirocrinus after the
appearance of Pleurocystites suggests the former were adapted to a different mode of
life. The subvective system in Macrocystella and Cheirocrinus is confined to the oral
surface and forms a cone of collection. Paul (1967#) suggested that such an arrangement
is better adapted to collect fa ling food particles in a relatively still sea and requires an
upright theca. Both Macrocys el/a and Cheirocrinus have respiratory surfaces developed

C. R. C. PAUL: M AC ROCYSTELLA CALLAWAY

all round the theca and some of these would have been fouled
if the theca rested on the sea floor. Plewocystites however has
all its pectinirhombs on one side of the theca, presumably
the upper side in life. All these lines of evidence indicate that
Macrocystella and Cheirocrinus habitually held their thecae
upright. However, when upright, Macrocystella is very top
heavy (text-fig. 15) and the theca would have rested on the
sea floor if the stem were free.

Without a root structure or anchoring device the stem could
exert no leverage against the substrate and would have been
useless to maintain the theca upright. The distal stem may
have been partially buried in the substrate but its whiplike
form would provide little resistance in soft mud. There is no
evidence for partial burial of the stem during life in specimens
which apparently died by entombment. Other pelmatozoans
which were fixed in soft sediments have well-developed
branching root structures (e.g. Eucalyptocrinites and Caryo-
crinites). Even recent brachiopods which live in soft sedi-
ments have branching root-like pedicles. If Macrocystella and
Cheirocrinus were free and habitually held their thecae up-
right they could only achieve this if they floated or swam
actively. Both modes of life are known in other pelmatozoans.

Recent free-swimming crinoids have negative buoyancy (i.e.
they are denser than sea water) and this has generally been
assumed to obtain in all echinoderms. If the coelomic fluids
and organic tissues are of neutral buoyancy, the only element
which contributes to negative buoyancy is the skeleton, the
effective density of which is 0-8 gm. per cc. when fully im-
mersed in sea water. An echinoderm with negative buoyancy
will sink when at rest. Recent free crinoids maintain their
position by swimming. They are characterized by a well-
developed subvective system with multiple branched arms
and, in most cases, by the complete absence of a stem and
marked reduction of the calyx: certainly none has a calyx
comparable in size to a cystoid theca. Swimming is wholly
achieved by the arms, buoyancy devices are absent (there is
nowhere to house them) and dead weight in the form of stem
and thecal plates is minimal.

In Macrocystella there is a relatively undeveloped subvective
system, a large theca, and a well-developed stem. Dead weight
was considerable and the brachioles may have been rela-
tively inefficient organs of locomotion. The flattening of the
brachioles in M. marine is in the opposite direction to that
one would expect if they were used actively in swimming.

text-fig. 15. A reconstruction of Macrocystella mariae Callaway. Br, brachiole;

DS, distal stem; Pe, periproct; PS, proximal stem.

598

PALAEONTOLOGY, VOLUME 11

However if the lappets were folded in during an upstroke and extended during a down-
stroke the brachioles could have been fairly efficient swimming organs. The relatively
large theca has very thin plates and could well have housed buoyancy devices. Quite small
gas bubbles would have significantly altered the total buoyancy: in M. mariae a bubble
3 mm. in diameter would completely compensate for the weight of a theca 10 mm.
in diameter and 15 mm. high. Brockocystis Foerste, another apparently free glypto-
cystitid, has hollow, bulbous thecal plates each of which could have housed a gas
bubble. The stem bulb in Brockocystis is also hollow. It is not impossible that Macro-
cystella had buoyancy chambers but there is no direct evidence for this.

Function of the stem. Most authors, with the notable exception of E. Kirk (1911),
have taken stems in pelmatozoans to imply permanent fixture. It is therefore surprising
to find among early stemmed pelmatozoans an almost total lack of definite root struc-
tures or anchoring devices. What benefit did the possession of a stem confer if it was not
attachment ?

If an ancestral stem-less form were permanently attached, the development of a stem
could have raised the theca off the substrate. This would reduce the chances of accidental
burial, of fouling by mud-laden currents, and it could have placed the echinoderm above
other benthonic organisms competing for food particles as suggested by Kesling and
Mintz (1961). What may have been attempts to raise the theca, or at least the oral
surface, are seen in edrioasteroids and some diploporites but none of these produces
a true stem, only aboral stem-like projections of the theca. Alternatively if an ancestral
stem-less form were free, a stem would weight the theca aborally. This would maintain the
animal upright if the theca were buoyant or if the subvective system were used in swim-
ming. However, a simple weight placed aborally would satisfy this requirement.

The form of the stem in Macrocystella could have allowed considerable variation in
buoyancy when the stem was in contact with the substrate. If the theca was buoyant
but the weight of the stem gave the whole animal negative buoyancy, equilibrium
could have been reached if part of the stem rested on the sea floor. The flexible proximal
stem would still allow the theca to be held upright; buoyancy would hold the theca
upright without any effort on the part of the animal. Macrocystella probably rested
with its theca near but not on the substrate and moved about by means of its brachioles
and stem. This interpretation seems preferable to one without a buoyant theca even
though there is no direct evidence for buoyancy devices. If Macrocystella used its
brachioles to maintain an upright posture it could not have rested in an upright position.
As recent free crinoids feed when at rest it is difficult to imagine how Macrocystella fed
under these circumstances.

Macrocystella is imagined to have had a slightly buoyant theca but to have been
weighed down by the stem. Under these circumstances it could have rested with the
stem on the sea floor, the theca upright and the brachioles extended to feed. In times of
short food supply Macrocystella swam away actively by means of its brachioles and
stem, moving just above the sea floor. Macrocystella may have drifted under the
influence of gentle currents but current action was not strong in the environment of
deposition of M. mariae. Occasionally M. mariae was overwhelmed by a sudden influx
of mud possibly brought about by turbidity currents. Macrocystella seems to have lived
in much deeper water than most later glyptocystitids (see Paul 1967a).

C. R. C. PAUL: MACROCYSTELLA CALLAWAY

599

REFERENCES

barrande, J. 1867. Systeme silurien du centre de la Boheme. le partie: recherches paleontologiques, 3.
Prague and Paris.

1868. Silurische Fauna aus der Umgebung von Hof in Bayern. Neues Jb. Miner. Geo!. Pa/dont.

(for 1868), 641-96, 2 pi. (also a separately published edition in French with different pagination).

1887. Systeme silurien du centre de la Boheme. \e partie: recherches paleontologiques, 7. Leipzig

and Prague.

bassler, r. s. and moodey, m. w. 1943. Bibliographic and faunal index of Paleozoic pelmatozoan
echinoderms. Spec. Pap. geol. Soc. Am. 45, vi+733 pp.
bather, f. a. 1899. A phylogenetic classification of the Pelmatozoa. Rep. Br. Ass. Advmt Sci. 68,
916-23.

1900. The echinoderms. in lankester, e. r. A treatise on zoology, 3. London.

billings, e. 1858. On the Cystideae of the Lower Silurian rocks of Canada. In Figures and descriptions
of Canadian organic remains, Decade 3, 9-74, pi. 1-8. Montreal.
callaway, c. 1877. On a new area of Upper Cambrian rocks in South Shropshire, with a description
of a new fauna. Q. J! geol. Soc. Lond. 33, 652-72, pi. 24.
carpenter, p. h. 1891. On certain points in the morphology of the Cystidea. J. Linn. Soc. (Zool.) 24,
1-52, pi. 1.

castell, c. p. (compiler) 1964. British Palaeozoic Fossils. B.M. (N.H.) London.
choubert, h., termier, g., and termier, h. 1953. Presence du genre Mimocystites Barrande dans
l'Ordovicien du Maroc. Notes Mem. Serv. Mines Carte geol. Maroc, 117, 137-43.
cuenot, l. 1948. Classe des Cystides. In grasse, p., Traite de Zoologie, 11, 15-25. Paris.

1953. Echinodermes. Classes des Cystides. In piveteau, j., Traite de Paleontologie, 3, 607-28.

Paris.

fearnsides, w. g. 1905. On the geology of Arenig Fawr and Moel Llyfnant. Q. Jl geol. Soc. Lond.
61, 608-37, pi. 41.

1910. The Tremadoc slates and associated rocks of South-east Caernarvonshire. Ibid. 66,

142-87, pi. 15-17.

haeckel, E. 1896. Amphorideen und Cystideen. In Festschrift zum siebenzigsten Geburt stage von Carl
Gegenbaur am 21st. August 1896, 1, 1-80, pi. 1-5. Leipzig.
jaekel, o. 1899. Stammesgeschichte der Pelmatozoen, 1, Thecoidea und Cystoidea. Berlin.

1918. Phylogenie und System der Pelmatozoen. Paldont. Z. 3, 1-128.

kesling, r. v. 1963. Key for the classification of cystoids. Contr. Mus. Paleont. Univ. Mich. 18, 101-16.

and mtntz, l. w. 1961. Notes on Lepadocystis moorei (Meek), an Upper Ordovician callocystitid

cystoid. Contr. Mus. Paleont. Univ. Mich. 17, 123-48, 7 pi.
kirk, e. 1911. The structure and relationships of certain eleutherozoic Pelmatozoa. Proc. U.S. natn.
Mus. 41, 1-137, pi. 1-11.

melendez, b. m. 1944. Nuevos datos para la estratigrafia del Paleozoico Aragones. Boln R. Soc. esp.
Hist. nat. 42, 129-50, pi. 24-9.

moore, r. c. 1954. Echinodermata : Pelmatozoa. In kummel, b., Status of invertebrate paleontology.
Bull. Mus. comp. Zool. Harv. 112, 125-49.

paul, c. r. c. 1967n. The British Silurian cystoids. Bull. Br. Mus. nat. Hist. (Geol.) 13, 297-356, 10 pi.

1961b. Cystidea Barrande 1868 (Cystoidea: Glyptocystitida) : proposed suppression under the

Plenary Powers. Bull. zool. Nom. 24, 304-7.

1968. The morphology and function of dichoporite pore-structures in the cystoids. Palae-
ontology, 11, 697-730, pi. 134-140.

pompeckj, J. f. 1896. Ein neuentdecktes Vorkommen von Tremadoc-Fossilien bei Hof. Ber. nordober-
frank. Ver. Naturk. 1, 89-101.

prokop, r. j. 1966. Class Eocrinoidea. In spinar, z. et a/., Systematicka paleontologie bezobratlych,
pp. 817-20, figs.

regnell, g. 1945. Non-crinoid Pelmatozoa from the Paleozoic of Sweden. A taxonomic study. Meddn
Lunds geol-miner. Instn, 108, 255 pp., 15 pi.

1948. An outline of the succession and migration of non-crinoid pelmatozoan faunas in the

Lower Paleozoic of Scandinavia. Ark. Kemi Miner. Geol. 26, (13), 1-55.

600

PALAEONTOLOGY, VOLUME 11

sdzuy, k. 1955. Cystoideen aus den Leimitz-Scheifern (Tremadoc). Senckenberg. leth. 35, 269-76, 1 pi.

springer, f. 1913. Cystoidea. In Eastman, c. r. (ed.), Textbook of Palaeontology, 1, 145-61. London.

Stubblefield, c. j. and bulman, o. M. b. 1927. The Shineton Shales of the Wrekin district: with notes
on their development in other parts of Shropshire and Herefordshire. Q. Jl geol. Soc. Load. 83,
96-145, pi. 3-5.

termier, g. and termier, H. 1952. Histoire geologique de la biosphere. 721 pp. 8 pi. Paris.

1954. A propos de la structure des Eocrino'ides, C. r. Seanc. Soc. geol. Fr. (1954) 4, 92-4.

thoral, M. 1935. Contribution a V etude paleontologique de FOrdovicien inferieur de la Montague Noire
et revision sommaire de la faune cambrienne. 362 pp. 35 pi. Montpellier.

ubaghs, G. 1953. Notes sur Lichenoides priscus Barrande, eocrinolde du Cambrien Moyen de la
Tchecoslovaquie. Bull. Inst. r. Sci. nat. Belg. 29 (34), 1-24.

1968. Eocrinoids. In moore, r. c. (editor) Treatise on invertebrate paleontology, S. Echinodermata

1, S455-S495.

zittel, k. A. von, 1880. Handbuch der Palaeonto/ogie. 1. Miinchen and Leipzig.

C. R. C. PAUL
Geology Department
Indiana University Northwest
Gary, Indiana

Typescript received from author 10 January 1968

A REVISION OF SOME UPPER DEVONIAN
FORAMINIFERA FROM WESTERN AUSTRALIA

by james e. conkin and Barbara m. conkin

Abstract. Arenaceous foraminifera described by Crespin (1961) from the Upper Devonian Virgin Hills and
Gogo Formations of the Fitzroy Basin of Western Australia are revised. Four genera and eight species are
recognized; these include Oxinoxis Gutschick 1962 and Sorosphaeroidea Stewart and Fampe 1947, which were
previously recorded only from the Middle Palaeozoic of the United States. The similarity of the fauna to Upper
Devonian foraminiferal faunas in the United States and Europe is described.

Several new occurrences of foraminifera are reported from the Silurian and Devonian of New South Wales,
Victoria, and Tasmania.

In 1964-5, we had the opportunity to collect from the Palaeozoic rocks of Victoria,
New South Wales, and Tasmania, south-eastern Australia, in order to assess the
stratigraphic occurrence and geographic distribution of arenaceous foraminifera. A
preliminary report on the Permian foraminifera of Tasmania has been made (Conkin
and Conkin 1965a). We were able, through the kindness of Dr. Irene Crespin and Dr.
D. J. Belford of the Bureau of Mineral Resources, Geology, and Geophysics at Canberra
to study the types of Dr. Crespin’s (1961) Upper Devonian foraminifera from the Virgin
Hills and Gogo Formations of the Fitzroy Basin, Western Australia. It was discovered
that some of the forms would require revision in the light of work done on arenaceous
foraminifera after the completion of Dr. Crespin’s work. With these revisions made, the
relationships between this important fauna of Crespin (1961) and similar faunas in the
United States and Europe can be more clearly seen.

Past work. Crespin (1961, p. 397) summarized the record of Devonian foraminifera in
Australia as follows:

In 1918 Chapman described and figured three genera and five species of what he considered to be
foraminifera from thin sections of a limestone of Devonian age from the Nemingha area, Tamworth,
northern New South Wales. However, Wood (1957) stated that these forms were ‘oolite grains more
or less affected by dolomitization and by mechanical distortion’. Parr (in Teichert and Talent 1958)
reported poorly preserved arenaceous foraminifera from Devonian rocks in the Buchan area, east
Gippsland, Victoria.

Chapman (1933) described two species of foraminifera from the Silurian of Victoria,

‘ Trochammina bursaria ’ and ‘ Hemigordius lilydalensis ’ ; the former is perhaps conspecific
with Thuramminoides sphaeroidalis Plummer 1945, which may not be a foraminiferan,
but possibly a radiolarian or spore-like body (Conkin, Conkin, and Canis 1968); the
latter is an indeterminate arenaceous foraminiferan and certainly not the calcareous
Hemigordius or even a siliceous replacement of Hemigordius.

Crespin’s (1961) fauna, the first important foraminiferal fauna from the Devonian of
Australia, contains four genera and eight species as revised in the present paper.

[Palaeontology, Vol. 11, Part 4, 1968, pp. 601-9, pis. 114-17.]

602

PALAEONTOLOGY, VOLUME 11

Present work. Silurian and Devonian foraminifera were observed by us in the following
states and stratigraphic horizons :

Tasmania, upper part of the Middle Devonian Point Hibbs Limestone at Point Hibbs: Sorospha-
eroidea sp.

Victoria, east Gippsland, Devonian Pyramid Member of the Buchan Cave Limestone at Bindi : Toly-
pammina spp., Sorosphaeroidea sp., and Hyperammina ? sp.

New South Wales, Yass-Taemas area, Devonian Spirifer yassensis beds: Hyperammina sp. and Colon-
ammina ? sp.

Victoria-New South Wales border area, Silurian Cowombat Formation at Cowombat Flat: Hyper-
ammina sp., Tolypammina sp., and Psammosphaeral sp.

The presence of these foraminifera in the Palaeozoic of Australia shows the need for
further work on this group along the lines of Crespin (1958, 1961). We are in the
process of describing these foraminifera from the Silurian and Devonian of Australia
as well as making a detailed study of the magnificent arenaceous foraminiferal faunas
from the Permian of Tasmania.

DISCUSSION OF GENERA

Saccammina Sars 1869. This genus consists of a simple agglutinated chamber with an
aperture which may or may not have a short neck; as thus understood, Saccammina has
been reported from Silurian to Recent.

Sorosphaeroidea Stewart and Lampe 1947 has a distinctive aperture which was not
originally noticed by Stewart and Lampe or subsequently by Summerson (1958).

Oxinoxis Gutschick 1962, emended Conkin and Conkin 1964. The range of Oxinoxis,
previously reported only from the United States, is upper Middle Devonian (middle
and upper Hamiltonian) to Lower Mississippian (Kinderhookian and lower Osagean).
Middle Devonian forms are almost exclusively single chambered and small. Upper
Devonian forms are single to multiple chambered, and small to large. Kinderhookian
and lower Osagean forms are generally much like the Upper Devonian, although there
is in addition a distinct single chambered species of Oxinoxis in the Kinderhookian
(Conkin, Conkin, and Canis 1968).

Tolypammina Rhumbler 1895. This attached genus ranges from Ordovician to Recent.
Pre-Upper Devonian forms appear to wind or twist in a haphazard, disorganized manner,
while some of the post-Middle Devonian forms wind, at least partially, in a more
organized pattern, it is probable that the even more ‘organized’, but similar genus,
Ammovertella Cushman 1928, which apparently arose in the Kinderhookian, was
derived from Tolypammina.

Hyperammina Brady 1878 is known from Ordovician to Recent. It is not present in
Crespin’s Upper Devonian fauna (1961), but occurs in the Silurian Cowombat Formation
in the New South Wales-Victoria border area. The restricted generic definition of
Hyperammina of Loeblich and Tappan (1964, p. C190) is not used. Hyperammina is
employed in the sense of Brady 1878, as emended by Conkin (1954, 1961), and discussed
by Conkin and Conkin (19656, p. 211). It is basically a test with a free proloculus and

CONKIN AND CONKIN: REVISION OF UPPER DEVONIAN FORAMINIFERA 603

a free second chamber of more or less tubular shape, tapering and/or slightly constricted
in some species. Hyperammina has great potential as a tool in stratigraphic palaeontology
within the Palaeozoic; a beginning along this line has been made within the Lower
Mississippian of the United States (Conkin 1961).

Psammosphaera Schulze 1875 consists of a simple agglutinated spherical or globular
chamber with no obvious aperture. It has been reported from Ordovician to Recent.

AGE OF THE VIRGIN HILLS FORMATION BASED ON FORAMINIFERA

Crespin (1961) reported only two specimens from that portion of the Upper Devonian
of Western Australia which may be referred to the Gogo Formation. Most of her
specimens came from the overlying Upper Devonian Virgin Hills Formation.

Considered strictly in terms of the known stratigraphic ranges of foraminiferal genera
included, the Virgin Hills Formation could be of late Middle Devonian, Late Devonian,
or Early Mississippian (Kinderhookian) age. Sorosphaeroidea, previously thought by us
to be restricted to beds older than Mississippian, has now been discovered in rare
numbers in the Kinderhookian McCraney Limestone of north-eastern Missouri.
Oxinoxis has not been found below the upper Middle Devonian, but it ranges into the
lower Osagean.

On a specific level, however, Tolypammina helina, with a test consisting of an initially
coiled portion and a final extended portion, is an ‘organized’ kind of Tolypammina
which is post-Middle Devonian in its stratigraphic occurrence. The foraminiferal fauna
of the Louisiana Limestone (Upper Devonian) from Missouri and Illinois is the oldest
fauna from which this type of Tolypammina has been described (Conkin and Conkin
1964) in North America; a similar species is found in the Upper Devonian of Germany
(Blumenstengel 1961). Thus, T. helina places the Virgin Hills Formation in the Upper
Devonian or Lower Mississippian, rather than the Middle Devonian.

SYSTEMATIC PALAEONTOLOGY

The following observations are based on a study of Dr. Crespin’s type specimens of Upper Devonian
foraminifera at the Bureau of Mineral Resources, Geology and Geophysics, Canberra, A.C.T.,
Australia, and on a discussion of the fauna with Dr. Crespin. Text-fig. 1 is based on drawings made at
that time and on photomicrographs taken subsequently and forwarded to us by Dr. D. J. Belford.
Type numbers refer to Commonwealth Palaeontological Collection (CPC), and holotype and para-
type designations are the original ones of Crespin (1961).

Family saccamminidae
Genus saccammina M.Sars 1869
Saccammina glenisteri Crespin 1961

Plate 1 16, figs. 1-7.

1961 Saccammina glenisteri Crespin, pp. 401, 402, pi. 65, figs. 3-7.

The holotype (CPC 401 ; Pi. 1 16, figs. 1-3) has slit-like? aperture as figured by Crespin, and a few
holes of secondary nature; the test, one grain thick, is composed of coarse grains of quartz in a fine-
grained matrix; there is no evidence of attachment.

R r

C 5934

604

PALAEONTOLOGY, VOLUME 11

Paratype A (CPC 402; PI. 116, figs. 6, 7) has a coarse-grained, subglobular test with an elongate,
lunate aperture; there is no evidence of attachment.

Paratype B (CPC 403) has a rounded aperture; there is no evidence of attachment.

Paratype C (CPC 404; PI. 1 16, figs. 4, 5) has a knob-like structure on the inside of the broken test
which is like the structure found on the interior of the test and just below the aperture in Soros-
phaeroidea, but which in this case may be secondarily formed.

Paratype D (CPC 405) consists of two globular tests attached to a Tolypammina fragment and
composed of coarse grains of quartz; the apertures are indistinct, and the tests are smaller than the
other types of this species; this specimen came from the Gogo Formation which underlies the Virgin
Hills Formation from which the other specimens came. The identity of Paratype D is not certain since
the tests are attached on a small surface and are thus not completely free.

Genus sorosphaeroidea Stewart and Lampe 1947
Sorosphaeroidea adhaerens (Crespin) 1961
Plate 116, figs. 8-11; text -fig. lc

1961 Sorosphaera adhaerens Crespin, pp. 403, 404, pi. 66, figs. 1-5.

The holotype (CPC 406; Pi. 1 16, figs. 8, 9) consists of chambers which are attached to a Tolypammina
fragment and the nature of the surface of attachment on the test (whether covered with a floor or open),
is not known. Its reference to Sorosphaeroidea is uncertain, but probably correct.

Paratype A (CPC 407) consists of several chambers which are indistinct on one side, and attachment
areas are somewhat obscure.

Paratype B (CPC 408; PI. 1 16, figs. 10, 11 ; text-fig. lc) possesses broad, flattened attachment scars
(not originally figured), which are somewhat finer grained than the upper side of test; there is a definite
aperture on the upper, unattached side of each chamber. A fragment of a Tolypammina is attached to
the specimen.

Paratype C (CPC 409) is attached to a Tolypammina fragment, but the attached surface is not visible;
an aperture is present on the upper side.

Paratype D (CPC 410) has an aperture on the upper unattached side; the attached side is not visible
as it is fastened to a fragment of Tolypammina.

Remarks. Paratype B is definitely referable to Sorosphaeroidea ; the other types have the
attachment surface hidden, or it is obscure (Paratype A), but they probably belong to
this genus also. We place this species in Sorosphaeroidea because of the large areas of
attachment, and because the apertures generally are like those which we have found in
the original types of Sorosphaeroidea of Stewart and Lampe; the aperture appears
inverted, as if pushed through from the outside.

EXPLANATION OF PLATE 114

Figs. 1-4,9. Tolypammina devoniana (Crespin) 1961. 1, Paratype A; basal, attached side showing rather
large attachment scars, smoothly polished along the edges. 2, Paratype D, top view. 3, Side view of
holotype showing attachment scar on the proloculus and small areas of attachment along the length
of the second chamber. 4, Top view of holotype showing slight notch on proloculus. 9, Paratype E,
top view showing break in test.

Figs. 5, 6. Tolypammina ? sp. 5, Plesiotype (holotype of ‘ Rhabdammina virgata ’ Crespin 1961), showing
fragmentary nature and coarse-grained texture of the test. 6, Reverse side of same specimen;
proloculus and aperture are not preserved.

Figs. 7, 8. Tolypammina sp. 7, Plesiotype of ‘ Marsipella sp.’, Crespin 1961 ; top view of hemitubular
fragment showing rough and coarse-grained test. 8, Basal view of same specimen showing attached
lower surface.

All figures approximately X 105.

Palaeontology, Vol. 11

PLATE 114

9

C0NK1N and CONK1N, Upper Devonian arenaceous foraminifera

CONKIN AND CONKIN: REVISION OF UPPER DEVONIAN FORAMINIFERA 605

The Middle Devonian forms of Sorosphaeroidea described by Stewart and Lampe
(1947) and Summerson (1958) are more regularly shaped and arranged than the present
forms.

text-fig. 1. Schematic drawings of attached sides of Tolypammina devoniana(A, e), Oxinoxis ampullacea
(b, d, f), and Sorosphaeroidea adhaerens (c). All figures approximately X 50. Hatched areas are flattened
surfaces of attachment. Black areas indicate openings into test. White areas are unattached portions

of test.

Family reophacidae

Genus oxinoxis Gutschick 1962, emend. Conkin and Conkin 1964
Oxinoxis ampullacea (Crespin) 1961

Plate 117, figs. 1-11; text-fig. 1b, d, f

1961 Lagenammina ampullacea Crespin, pp. 404, 405, pi. 66, figs. 6-8.

1961 Colonammina imparilis Crespin, pp. 405, 406, pi. 65, figs. 10-13.

1961 Proteonina sp., Crespin, p. 404, pi. 65, fig. 8.

The holotype of ‘’Lagenammina' ampullacea (CPC 412; PI. 117, figs. 1-3) has a faint flattened
attachment scar at the base, but otherwise the chamber is nearly spherical with a small apertural neck.

Paratype A of ‘ Lagenammina ' ampullacea (CPC 413; PI. 117, fig. 11) is slightly flattened on the
basal side, with a somewhat finer-grained texture there than on the upper side, but with some larger
quartz grains studded on it; there is a small apertural neck.

Paratype B of ‘ Lagenammina ' ampullacea (CPC 414) has a rather prominent apertural neck with
the basal side flattened; there is also a small, slightly roughened attachment area at the base of the
chamber.

The holotype of ‘ Colonammina imparilis' (CPC 415; PI. 117, figs. 4-6; text-fig. Id) consists of a
nearly spherical chamber with a prominent wedge-shaped attachment scar forming a re-entrant into
the test below the neck; the prominent apertural neck is hemitubular with the basal surface flattened
and attached.

Paratype A of ‘ Colonammina imparilis' (CPC 416) has a flattened attachment scar on the basal side
which is finer grained than the unattached, upper, convex side; the prominent tubular apertural neck
has no attachment scar.

Paratype B of ‘ Colonammina imparilis' (CPC 417; PI. 117, figs. 7, 8; text-fig. 1b) has a large basal
attachment scar which is fine grained and has a polished lustre; the apertural neck is only partly
attached on its basal side; the convex upper unattached side of the test is coarse grained.

Paratype C of ‘ Colonammina imparilis' (CPC 418) is an obscure specimen; the apertural neck
figured by Crespin (1961, pi. 65, fig. 13) appears to be part of another specimen and thus foreign to
paratype C; the base of the test is flattened and has a prominent fine-grained attachment scar with a
polished lustre; this specimen is not typical of the species, but does apparently belong to O. ampullacea.

606

PALAEONTOLOGY, VOLUME 11

‘ Proteonina sp.’ (CPC 411; PI. 117, figs. 9, 10; text-fig. If) is a compressed test with holes and
depressions in the wall left after calcareous grains were leached out; the apertural neck is partly
hemitubular and has a small attachment scar on the basal side at the apertural end, and after which the
neck is tubular adjacent to the chamber, which is itself unattached; this specimen is from the Gogo
Formation.

Remarks. The presence of an attachment scar on the chamber of the test and/or on the
hemitubular to tubular apertural neck serves to place these specimens in Oxinoxis.

O. ampuUacea differs from the type species O. ligula (Gutschick, Weiner, and Young
1961), which occurs in the Upper Devonian and Lower Mississippian in the United
States, in having a more coarse-grained test and in being nearly twice as large as the
single-chambered, more solidly-walled forms of O. ligula.

Family ammodiscidae
Genus tolypammina Rhumbler 1895
Tolypammina devoniana (Crespin) 1961

Plate 114, figs. 1-4, 9; text-figs. 1a, e

1961 Hyperammina devoniana Crespin, pp. 406, 407, pi. 64, figs. 1-6.

The holotype (CPC 419; PI. 114, figs. 3, 4) has a proloculus which is asymmetrical, shaped rather
like the head of a golf club, owing to a notch-like attachment scar on its basal surface where it joins
the second chamber; this basal side of the proloculus is partially attached and there are attachment
scars at irregular intervals and distances along the elongate second chamber.

EXPLANATION OF PLATE 115

Figs. 1-4. Tolypammina helina Crespin 1961. 1, Holotype, basal flattened and attached side of hemi-
tubular test. 2, Top view of holotype showing proloculus, 2\ planispiral whorls, and extended por-
tion. 3, Paratype A, basal view showing the completely attached hemitubular test, and an early
unsuccessful attempt to uncoil (after 1J planispiral whorls) with return to planispiral coiling for
1 1 whorls more before finally uncoiling. 4, Top view of paratype A, showing proloculus, rather obscure
whorls, and extended portion.

Figs. 5-8. Tolypammina nexuosa Crespin 1961. 5, Paratype B of Tolypammina ‘ helina', top view.
6, Basal view of same specimen showing complete attachment of hemitubular test; no proloculus
preserved. 7, Holotype of Tolypammina nexuosa, top view. 8, Basal view of holotype, showing
attached side of test.

All figures approximately x 105.

EXPLANATION OF PLATE 116

Figs. 1-7. Saccammina glenisteri Crespin 1961. 1, Holotype, upper apertural view, showing slit-like
aperture in centre of test. 2, Basal view of holotype. 3, Side view of holotype. 4, Paratype C, broken
test showing knob-like structure underneath the aperture on the interior of the test. 5, Exterior view
of paratype C showing aperture in the centre of the test. 6, Paratype A, upper apertural view of test
showing slit-like aperture in the centre of the test. 7, Basal view of paratype A.

Figs. 8-11. Sorosphaeroidea adhaerens (Crespin) 1961. 8, Holotype, view of aggregate of chambers
shown by Crespin (1961, pi. 66, fig. 1) showing attachment to foreign body. 9, Reverse side of
specimen. 10, Paratype B, upper unattached surface of test showing aperture piercing middle
chamber and an attached fragment of Tolypammina. 1 1 , Basal view of paratype B showing flattened
chambers on attached side of test.

All figures approximately X 105.

Palaeontology, Vol. 11

PLATE 115

CONKIN and (TONKIN, Upper Devonian arenaceous foraminifera

Palaeontology , Vol. 11

PLATE 116

CONKIN and CONKIN, Upper Devonian arenaceous foraminifera

CONKIN AND CONKIN: REVISION OF UPPER DEVONIAN FORAMINIFERA 607

Paratype A (CPC 420; PI. 1 14, fig. 1) is like the holotype; a prominent attachment scar beginning
about midway along the proloculus and smoothly polished along the edges, extends approximately
half the length of the test.

Paratype B (CPC 421) consists of a nearly symmetrical test which is, however, slightly flattened with
an attachment scar on the basal side; the test is somewhat longer than originally shown.

Paratype C (CPC 422) is an asymmetrical test with a slightly flattened attachment scar on the basal side.

Paratype D (CPC 423; PI. 1 14, fig. 2; text-fig. 1e) is an asymmetrical and crooked test; the proloculus
has a large notched attachment scar which extends a short distance along the second chamber; the
edges of the scar are polished and the scar itself has a fine-grained texture.

Paratype E (CPC 424; PI. 1 14, fig. 9; text-fig. 1a) is a quite crooked test with the second chamber
hemitubular over most of its length owing to attachment scars on one side; short intervals of the second
chamber are tubular with no attachment along these portions; the attachment scar extends only
slightly on the proloculus.

Remarks. T. devoniana is a good example of the genus even though the attachment scars
are only slightly developed in some specimens. Tolypammina is attached at intervals
along or throughout, the length of the second chamber, while the proloculus may be
free, attached, or partially attached as in the present species. The hemitubular nature of
the second chamber is characteristically developed to some degree on the attached
portions, but this is not so in Hyperammina, which is entirely free.

T. devoniana differs from a similar species, T. bulbosa (Gutschick and Treckman 1959),
which occurs in the Upper Devonian and Lower Mississippian of the United States, in
having a more irregularly shaped and partially attached proloculus, while the pro-
loculus of T. bulbosa is free and quite spherical.

Tolypammina helina Crespin 1961

Plate 115, figs. 1-4

1961 Tolypammina helina Crespin (in part), p. 407, pi. 67, figs. 1-4.

The holotype (CPC 425; pi. 115, figs. 1, 2) has an attached proloculus and a hemitubular second
chamber which is flattened all along the basal attached side and is only partly floored; whorls are all
roughly in the same plane; the upper, unattached side is somewhat obscure and was shown slightly
incorrectly in the original drawing; the present figure (PI. 1 15, fig. 2) shows the nature of the whorls
on this dorsal side.

Paratype A (CPC 484; PL 115, figs. 3, 4) is a completely attached test which is flattened on the basal
side, with the second chamber hemitubular throughout and only partly floored; the whorls are roughly
in the same plane.

Remarks. T. helina is remarkably similar to T. gersterensis Conkin and Conkin 1964,
which occurs in the Upper Devonian and Lower Mississippian of the United States.
A similar species, in part at least, is T. irregularis Blumenstengel 1961, from the Upper
Devonian and Lower Carboniferous of Germany; the exact relationships between
these species are not certain.

Tolypammina nexuosa Crespin 1961

Plate 115, figs. 5-8

1961 Tolypammina nexuosa Crespin, pp. 407, 408, pi. 67, figs. 6-8.

1961 Tolypammina helina Crespin (in part), p. 407, pi. 67, fig. 5.

608

PALAEONTOLOGY, VOLUME 11

The holotype (CPC 486; PI. 115, figs. 7, 8) displays an irregular pattern of twisting and winding.

Paratype A (CPC 487) possesses a flat, attached, basal side; no proloculus is visible; the rounded
hump on the specimen is not the proloculus, but merely a bend in the hemitubular second chamber.

Paratype B (CPC 488) has a hemitubular test which is flat and attached on the basal side; no pro-
loculus is present.

Paratype B of Tolypammina ‘ helina ’ (CPC 485; PI. 1 15, figs. 5, 6) probably belongs to T. nexuosa
for it does not possess a planispiral portion; the basal side is flat and the second chamber is hemi-
tubular.

Remarks. This species is of doubtful status, for a proloculus should be seen in order to
erect a species and to make specific identification in the genus Tolypammina ; the early
relationship of the second chamber to the proloculus is usually distinctive in a given
species.

Tolypammina sp.

Plate 1 14, figs. 7, 8

1961 Marsipella sp., Crespin, p. 401, pi. 65, fig. 9.

This specimen (CPC 400; PI. 1 14, figs. 7, 8) consists of a broken test with no proloculus preserved;
the test is hemitubular and attached, with the attached side open, without a floor; upper unattached
side is of coarse and clear quartz grains.

Remarks. This specimen clearly belongs to Tolypammina, even though it is a fragment,
for it is hemitubular and attached.

Tolypammina ? sp.

Plate 1 14, figs. 5, 6

1961 Rhabdammina virgata Crespin, pp. 400, 401, pi. 64, figs. 7-10; pi. 65, figs. 1, 2.

The holotype (CPC 394; PI. 1 14, figs. 5, 6) consists of a test which is broken at both ends so that the
nature of the proloculus and aperture is not known; this fragment is tubular, with no evidence of
attachment noted, but the test is only partly preserved; the test is composed of medium to coarse
quartz grains. Although the evidence for positive generic assignment is not present in this specimen,
we prefer to regard this specimen as belonging to Tolypammina rather than Rhabdammina.

Paratype A (CPC 395) is a limonitic cast of an orange colour; the test is broken at both ends and is
curved; no proloculus is present.

EXPLANATION OF PLATE 117

Figs. 1-11. Oxinoxis ampullacea (Crespin), 1961. 1, Holotype of ‘ Lagenammina' ampidiacea, side view.
2, Upper surface of same specimen. 3, Basal attached side of same specimen. 4, Holotype of ' Colon-
ammina imparilis ’ Crespin; oblique view of basal attachment and upper surface showing hemi-
tubular apertural neck and partially obscured wedge-shaped attachment scar. 5, Basal attached
surface of same specimen showing attachment scar. 6, Upper surface of coarse-grained test, same
specimen, showing a few coarser quartz sand grains incorporated in the test. 7, Paratype B of
‘ Colonammina imparilis upper convex surface of test showing holes in the wall, not apertural,
due to breakage. 8, Basal attached surface of same specimen showing large attachment scar.
9, Plesiotype of ‘ Proteonina sp.’, Crespin 1961, basal attached and pitted surface of test. 10, Upper
pitted surface of test, same specimen. 11, Paratype A of ‘ Lagenammina ’ ampidiacea Crespin 1961,
upper surface.

All figures approximately x 105.

Palaeontology, Vol. 11

PLATE 117

CONKIN and CONKIN, Upper Devonian arenaceous foraminifera

CONKIN AND CONKIN: REVISION OF UPPER DEVONIAN FORAMINIFERA 609

Paratype B (CPC 396) is a bent and curving test, broken at both ends, with no proloculus or aperture;
tube is composed of coarse-grained quartz.

Paratype C (CPC 397) is broken at both ends and is curved; no proloculus is preserved.

Paratype D (CPC 398) consists of an irregularly coiled test, broken at both ends; no proloculus is
preserved.

Paratype E (CPC 399) is broken at both ends with no proloculus preserved; the test is red-orange
in colour and is composed of quartz sand grains.

Remarks. These specimens appear to be fragments of free portions of Tolypammina, but
specific identification is impossible.

Acknowledgements. We are grateful to: Dr. Irene Crespin for her kind permission to revise these
foraminifera; Dr. D. J. Belford for photographing the specimens; Dr. John Talent, who led us to the
fossiliferous outcrops of the Silurian Cowombat Formation in Gippsland and southern New South Wales ;
Mr. Edmund Gill, who facilitated examination of type specimens in the Victoria Museum, Melbourne,
and Dr. K. S. W. Campbell for guiding us in a study of the Yass-Taemas area, New South Wales.

This work was made possible by a Fulbright Senior Research Fellowship to the University of
Tasmania and sabbatical leave from the University of Fouisville in 1964-5.

REFERENCES

blumenstengel, v. h. 1961. Foraminiferen aus dem Thiiringer Oberdevon. Geologic, 10 (3), 316-35,
pi. 1-3.

chapman, F. 1933. Some Paleozoic fossils from Victoria. Proc. R. Soc. Viet. 45 (2), 245-8, pi. 11.
conkin, j. e. 1954. Hyperammina kentuckyensis, n. sp. from the Mississippian of Kentucky, and discussion
of Hyperammina and Hyperamminoides. Contr. Cushman Fdn. foramin. Res. 5 (4), 165-9, pi. 31.

1961. Mississippian smaller foraminifera of Kentucky, southern Indiana, northern Tennessee,

and south-central Ohio. Bull. Am. Paleont. 43 (196), 131-368, pi. 18-27.

and conkin, b. m. 1964. Devonian foraminifera: Part 1, The Fouisiana Fimestone of Missouri

and Illinois. Ibid. 47 (213), 53-105, pi. 12-15.

1965 a. Permian foraminifera in Tasmania. Aust. J. Sci. 28, 312.

19656. Ordovician (Richmondian) foraminifera from Oklahoma, Missouri, Illinois, and

Kentucky. Okla. Geol. Notes, 25 (8), 207-21, pi. 1, 2.

and canis, w. 1968. Mississippian foraminifera of the United States. Part III, The lime-
stones of the Chouteau Group in Missouri and Illinois. Micropaleontology, 14, 133-78, pi. 1-4.
crespin, i. 1958. Permian foraminifera of Australia. Bull. Bur. Miner. Resour. Geol. Geophys. Aust.
48, 1-207, 33 pi.

1961. Upper Devonian foraminifera from Western Australia. Palaeontology, 3, 397-409, pi. 64-7.

loeblich, a. R. andTAPPAN, H. 1964. Sarcodina, chiefly ‘Thecamoebians’ and Foraminiferida. In moore,
r. c., ed. Treatise on Invertebrate Paleontology, Part C, Protista l.Geol. Soc. Am. and Unix. Kans. Press.
stewart, G. A. and lampe, l. 1947. Foraminifera from the Middle Devonian bone beds of Ohio. J.
Paleont. 21, 529-36, pi. 78, 79.

summerson, c. h. 1958. Arenaceous foraminifera from the Middle Devonian limestones of Ohio.
Ibid. 32, 544-88, pi. 81, 82.

JAMES E. CONKIN
Department of Geology
University of Fouisville
Fouisville, Kentucky 40208
U.S.A.

BARBARA M. CONKIN

Jefferson Community College
University of Kentucky
Fouisville, Kentucky 40202
U.S.A.

Typescript received from authors 27 February 1968

A NEW MEDUSOID (?) FROM THE SILURIAN

OF ENGLAND

by ISLES STRACHAN

Abstract. Duodecimedusina palmeri sp. nov. is described from the Upper Llandovery of the Malverns and West
Midlands. This extends the stratigraphic range of the genus previously described from the Carboniferous and
Lower Devonian. It is also the first record from Europe.

One of the specimens described here was picked up some years ago on a student field
excursion to the Lickey Hills, just south of Birmingham. Like so many interesting
specimens it was a loose block so that there is no direct evidence of which way up the
specimen was, or more important, from which geological horizon it came, both Cam-
brian and Lower Silurian being exposed in the quarry. In June 1967, a similar specimen
(for which the stratigraphic horizon is more precise) was collected by Mr. K. F. Palmer
on a student trip to the Malverns. The confirmation of geological age and morphological
features make recording of the specimens more important since the range of the genus
is considerably extended in time and space.

Genus duodecimedusina King

This genus was proposed by King (1955) for three species, D. typica, D. wycherleyi,
and D. ulrichi. The first two of these, known from a single specimen each, come from
Upper Carboniferous of Kansas, and the third species, also from a single specimen, is
from the Lower Devonian of Bolivia. Avnimelech (1966) has recently described a
fourth species, D. aegyptica from the Lower Carboniferous of Egypt, again from a
single specimen. The present specimens therefore extend the range of the genus down
to the Lower Silurian as well as being the first record from Europe.

Harrington and Moore (1956) classified Duodecimedusina amongst the ‘Medusae
Incertae Sedis’ but Avnimelech has proposed transferring the genus to the Proto-
medusae on the grounds of similarity to Brooksella which forms the whole class. The
latter genus, however, frequently shows supplementary lobes on the sub-umbrellar
surface which are not found in the Egyptian and British specimens. The American
specimens are all attached to a ‘pedestal’ so that the sub-umbrellar surface cannot be
seen. The general category of ‘Medusae Incertae Sedis’ therefore seems to me to be a
better resting place for Duodecimedusina for the present.

Duodecimedusina palmeri sp. nov.

Material. Holotype, BU 302, from base of Upper Llandovery, Gullet Quarry, Malverns, in a slightly
micaceous siltstone, way-up not known. Paratype, BU 303, probably from Rubery Sandstone (Upper
Llandovery), Rubery Quarry, Leach Green Lane, Birmingham, in a medium to coarse sandstone.

Description. Body roughly circular, 16-19 mm. in diameter. There is a central raised

[Palaeontology, Vol. 11, Part 4, 1968, pp. 610-11.]

STRACHAN: NEW MEDUSOID (?) FROM THE SILURIAN OF ENGLAND 611

area, rather poorly defined, occupying about half of the diameter, surrounded by twelve
lobes. The lobes are distinct at the periphery, measuring about 3 mm. across, but merge
centrally into the more raised area. The undersurface is poorly preserved in the holo-
type and not seen on the paratype.

text-fig. 1. a, b, c. Hclotype, BU 302, exumbrellar, side, and subumbrellar views, d. Paratype, BU 303,

exumbrellar view. All X 1|.

Discussion. D. palmeri is similar to D. typica King in the poorly delimited area of the
exumbrellar surface but is little more than half its size. D. wycherleyi King is even
smaller (10-1 1 mm. in diameter) but has a central depression, which is also found in the
Lower Devonian, D. ulrichi King. D. aegyptica Avnimelech is similar in size to D.
palmeri but is much more convex in lateral view and has more clearly marked off lobes.

Since the above was written, Dr. Goldring has sent me another specimen from the
Gullet Quarry. This specimen (Reading Univ. Geol. Dept. no. 14601) was collected as
a loose specimen by Mr. D. C. Smith while mapping the area. It is somewhat larger
(about 28 mm. in diameter) and the lobes show weak, but fairly constant, concentric
bands, a feature not reported on any other of the species.

REFERENCES

avnimelech, M. A. 1966. A new medusoid fossil from the Lower Carboniferous of Egypt. J. Paleont.
40, 742-5.

Harrington, H. J. and moore, r. c. 1956. Protomedusae, etc. In moore, r. c. Treatise on invertebrate
Paleontology, Part F, Coelenterata. Kansas.

king, r. h. 1955. In harrington, h. j. and moore, R. c. Fossil jellyfishes from Kansan Pennsylvanian
rocks and elsewhere. Bull. geol. Surv. Kansas. 114, 153-64, pi. 1, 2.

ISLES STRACHAN
Geology Department
University of Birmingham
P.O. Box 363
Birmingham, 15

Typescript received from author 8 March 1968

THE ATRYPIDINE BRACHIOPOD DAYIA
NAVICULA (J. BE C. SOWERBY)

by E. V. TUCKER

Abstract. The atrypidine brachiopod Dayia navicula (J. de C. Sowerby) is described and a neotype is designa-
ted. The musculature is reinterpreted and shown to differ from previous interpretations. Comparisons are made
between Dayia navicula s. str. and other forms referred to the species. Between the Ludlow and the Lower
Devonian the shells increase in size and incurvature of the umbo becomes less marked.

Prior to 1881 the brachiopod now designated Dayia navicula rested uneasily in
palaeontological works first under the style of Terebratula navicula and later Rhyn-
chonellal navicula. JVTCoy in 1846 anticipated in part the probable relationship between
this form and the atrypids when he referred it to Atrypa navicula. In 1881 Davidson,
studying material skilfully prepared by the Revd. N. Glass, recognized the distinctiveness
of the brachidium and erected the new genus Dayia. Since then material referred to the
species D. navicula has received attention especially in the hands of Kozlowski (1929,
p. 179) and Alexander (1947). The latter undertook a morphological study of the
characteristic British form but her conclusions contain certain misconstructions.

In stratigraphy, Dayia navicula has been widely used as an index of Ludlovian rocks,
particularly for the Leintwardine Beds (formerly the Dayia or Mocktree Shales of Elies
and Slater (1906)) which underlie the Upper Ludlow or Whitcliffe Beds. Straw (1937),
however, noted its presence in rocks ranging from the nilssoni zone into the base of the
Upper Ludlow at Builth. This extended range was re-emphasized by Earp (1938) and
Shirley (1939 and 1952) for other parts of Wales and the Welsh Borderland. Davidson
(1869, p. 191) in the monograph of Silurian brachiopods records Wenlock occurrences
given in Murchison’s ‘Siluria’ but only the general locality of Builth is mentioned.

A similar stratigraphical range has been suggested in continental Europe. Barrande
(1879) records the species from the Upper Ordovician of Bohemia but this claim needs
re-examination. In Scandinavia, Lower Ludlow forms, although smaller than the
typical Dayia navicula, closely resemble certain British Elton Bed (Lower Ludlow)
forms. Koslowski’s description (1929, p. 179) and illustrations of Podolian specimens
from the Marnes de Dzwinogrod, now believed by Jaeger (1965) and others to be
post-Ludlovian and possibly Gedinnian in age, indicate close agreement with the
characteristic British form. The existence of other sub-species, or at least variants of
Dayia navicula, is suggested, however, by the work of Boucek (1940) and Shirley (1962).
Shirley emphasizes the differences between varieties from the Schistes de Lievin and the
Kobbinghauser Schichten of the Artois basin and Westphalia respectively, which have
previously been referred to as Dayia navicula, and the characteristic British forms; a
view endorsed by the present writer. This misidentification and the usage of Dayia
navicula as a Ludlovian index fossil is partly responsible for the discordant views which
have been expressed on the correlation of French-Rhenish and British successions.

[Palaeontology, Vol. 11, Part 4, 1968, pp. 612-26, pis. 118-21.]

TUCKER: ATRYPIDINE BRACHIOPOD DAYIA NAVICULA (J. DE C. SOWERBY) 613

In her account, Alexander referred (1947, p. 305) to a specimen in the Geological
Survey collection as holotype. This specimen does not match Sowerby’s illustration in
Murchison’s Silurian System (1839, pi. 5, fig. 17) however, differing in size, detail of
ornamentation, nature of ventral umbo, and in its geographical location. Search else-
where has failed to reveal the type specimens and they are presumed to have been lost.
To assist future comparison a neotype for Dayia navicula is designated here, the mor-
phology is re-described, the musculature reinterpreted and some aspects of variation in
the genus are outlined.

The material on which the study is based has been collected by the writer, chiefly from the Leint-
wardine Beds of Downton Gorge (Herefordshire) and various Silurian inliers in the Welsh Borderland,
particularly those of Woolhope and Usk, and from the Elton Beds and Whitcliffe Beds. These personal
collections were supplemented by collections housed in the Geological Survey Museum (GSM) and
Mrs. Alexander’s collections of silicified specimens in the Sedgwick Museum, Cambridge (SM).
(Manuscript notes of Alexander suggest that this collection was made near Greenway Cross (Nat.
Grid. SO/461 827), although repeated visits to the same locality by the present writer have revealed
few specimens of Dayia navicula. The fauna indicates a Lower Leintwardine age for the beds.) The source
of foreign material is indicated in the appropriate section.

SYSTEMATIC DESCRIPTION

Superfamily dayiacea Waagen 1883
Family dayiidae Waagen 1883
Subfamily dayiinae Waagen 1883
Genus dayia Davidson 1881

Type species. Terebratula navicula (J. de C. Sowerby) 1839.

Diagnosis. Small, strongly inequivalve, pedicle valve larger; hinge line gently curved;
shell generally elongated medially but sometimes spherical. Brachial valve with sulcus
expanding to one-third of shell width at anterior margin; weakly inflated laterally.
Pedicle valve highly arched transversely, smoothly curved medially. Umbo normally
incurved. Lateral commissure dorsally curved, anterior commissure sulcate. Surface
ornamentation of very fine costae and weak growth-lines, indistinct.

Spires directed towards sides of pedicle valve but contained within the first coil. In
the pedicle valve a platform of secondary callus supported the diductor muscles. Muscle
fields deeply impressed, of chevron form. Adductor scars ill-defined. Prominent median
septum in the brachial valve extends to three quarters valve length and separates well-
developed adductor muscle fields. A septalium divides a subdued cardinal plate at the
centre of the hinge-line.

Range. Silurian (Ludlovian, possibly Wenlock) to Devonian (Gedinnian). Dayia has been recorded
from many Western European countries along a belt from the Dingle peninsula in south-west Ireland
to Podolia in the U.S.S.R., and from Scandinavia. It is not known from extra-European areas.

Discussion. The pre-Ludlovian records of Dayia require re-investigation. Dayia cymbula
(Davidson) has been recorded from the Caradoc Beds of Hendre Wen (North Wales)
and from the Drummuck Group (Ashgillian) near Girvan. Reed (1917) described a
variant ‘ girvanensis’ associated with it near Girvan. Externally there are many

614

PALAEONTOLOGY, VOLUME 11

similarities between these forms and Dayia navicula. The musculature of the pedicle valve
differs in detail however and the teeth are more prominent. Dental plates, observed by
Davidson, are lacking in Dayia. The nature of the brachidium of Dayia cymbula is
not known.

Dayia navicula (J. de C. Sowerby)

Plates 118-21

1839 Terebratula navicula Sowerby, p. 611, pi. v, fig. 17.

1846 Atrypa navicula (Sowerby); M‘Coy, p. 40.

1848 Hypothyris navicula (Sowerby); Phillips, p. 281.

1852 Hemithyris navicula (Sowerby); M‘Coy, p. 204.

1859 Rhynchonella ? navicula (Sowerby); Lindstrom, p. 366.

1869 Rhynchonella ? navicula (Sowerby); Davidson, p. 191.

1881 Dayia navicula (Sowerby); Davidson, p. 291.

Neotype. A specimen collected by the writer and deposited in the Geological Survey Museum (GSM
103291) is selected as neotype (PI. 1 18, fig. 2). Sowerby’s type locality is vaguely defined. The neotype
is from the Lower Leintwardine Beds of Downton Gorge, Herefordshire (Nat. Grid SO/431 733)
approximately 55 ft. above the base.

Occurrence. Dayia navicula is present throughout the Leintwardine Beds of the Welsh
Borderland, being most common in western areas. It extends upwards into the Lower
Whitcliffe Beds, especially in adjacent parts of Wales. Closely allied forms also occur in
the Elton Beds of these areas and at similar levels in Scandinavia whilst other variants
range upwards into Lower Devonian rocks of central Europe.

Dimensions of neotype. Length 10-4 mm., width 8 0 mm., thickness 7-2 mm.

Diagnosis. As for the genus. The umbo characteristically is gibbous and rests upon the
brachial valve, effectively sealing the foramen. The diductor muscle platform is pro-
minent in the pedicle valve and the chevron scars occupy much of the shell width.

Description. Considerable shape variation can be recognized. Ontogenetic studies
demonstrate a progressive relative increase in elongation with age. (PI. 118, figs. 1-5).
This is generally accompanied by a deepening of the sulcus in the brachial valve and
accentuation of umbonal incurvature in the pedicle valve. This incurvature ultimately
caused pedicle atrophy during maturity (Tucker 1965). In juvenile growth stages the
sulcus remains shallow and smoothly curved transversely. It becomes steeper sided in
mature forms, where there is an equal division between forms with a deep, flat-bottomed
sulcus and those more sharply angular. Variation in mature shells is expressed also by
differences in degree of elongation and degree of incurvature of the ventral umbo. Most
shells have a pronounced elongation (the width to length ratio exceeding 1:1-2) but

EXPLANATION OF PLATE 118

Figs. 1-5. Dayia navicula (J. de C. Sowerby). All un-numbered specimens in author's collection. Lower
Leintwardine Beds, Downton Gorge, Herefordshire. National Grid: SO/431 733. Dorsal, anterior,
and lateral views of five growth stages. Fig. 2 is designated neotype (GSM 103291) (all x4).

Palaeontology, Vol. 11

PLATE 118

TUCKER, Dayia

TUCKER: ATRYPIDINE BRACHIOPOD DAYIA NAVICULA (J. DE C. SOWERBY) 615

some retain more squarish outlines. All mature shells have a strongly incurved umbo
but this is accompanied by excessive elevation of the posterior part of the ventral valve
in some specimens, producing an acute transverse arch which accentuates elongation
of the shell.

Dentition. The teeth diverge at an angle of less than 160°. Their length is almost one-
quarter the shell width. They are most prominent posteromedially, hence articulation
is best defined towards the centre-line where the inward and partly downward facing
edge of each tooth locks into a laterally opening notch within the sockets (see text-fig.
lg). Thickening of the inner socket-ridge beneath, and to a lesser extent above, the
socket is accompanied by a swelling into the visceral cavity, intensifying the socket and
assisting articulation. The stages of development of the inner socket ridge are demon-
strated by the pattern of growth lines shown in text-fig. 3 d.

The crenulations on the dorsal surface of the teeth referred to by Alexander (1947,
p. 307), are rarely apparent although similar features, lying on the inner face, have been
observed on one silicified mould in her collection. There is no clear indication of these
features in the micro-structure of other, calcareous, shells.

Lophophore supports. These comprise crura and laterally directed spires united by a
comparatively simple jugum (see text-fig. 1).

The crura arise from the inner margin of, and fuse with, geniculiform inner shell
surfaces which flank a septalium in the cardinal plate. Geniculation commences on a
level with the hinge axis and immediately adjacent to the shell mid-line. Its development
is reflected in the micro-structure of the shell where a close relationship is seen to exist
with the development of the septalium, beneath which concentration of growth-lines
indicates a retarded rate of shell secretion (see PI. 121, figs. 1,2; text-fig 3).

The crura develop from the septalium walls as narrow rods and follow the line of the
walls as they diverge at an angle of about 55°. Complete fusion and continuity of shell
structure is rapidly achieved between each crus and the parent shoulder, which provides
additional support.

The crura extend ventrally a distance equal to about one quarter the depth of the
cavity, and curve gently with the convex edge facing anteriorly; they descend anterior to
their point of origin. At their distal end they produce the primary lamellae which first
extend laterally to a position near the dental sockets before descending anteriorly.
Peels taken from the crura reveal lamellae of fibrous calcite parallel to the surfaces.
There is complete continuity with the primary coil.

In longitudinal sections the micro-structure of the supporting crural base reveals a
sharp projection within the secondary layer representing an early stage of crural
development (see text-fig. 2). The base is supported on a flat spoon-shaped depression
within the brachial shell.

The primary lamellae descend almost vertically at a position close to the outer socket
ridge before producing the first coil. The first four or five coils are strongly flexed along
their dorsal arc (text-fig. 1 and PI. 119, fig. 10). Up to nine coils are contained in each
spire, fiexuring being less pronounced on each successive coil. The coils are more widely
spaced dorsally than ventrally, hence the axis of rotation of successive coils migrates
ventrally. Alexander (1947, p. 308) aptly describes each spire as being ‘. . . shaped like
a sombrero with one half of the brim turned up’.

text-fig. 1. Davia navicula (J. de C. Sowerby), Lower Leintwardine Beds, Downton Gorge, Hereford-
shire. Series of transverse sections drawn from cellulose acetate peels. X 5.

TUCKER: ATRYPIDINE BRACHIOPOD DAYIA NAVICULA (J. DE C. SOWERBY) 617

The jugum, arising from a position low on the first coil, is V-form, the apex pointing
postero-ventrally. It is sharply recurved at its union with the primary coils (text-figs.
\p-r). A jugal stem points posteriorly. There appears to be perfect continuity in the
micro-structure of jugum and primary lamellae.

text-fig. 2. Features of shell micro-structure in Dayia navicula (J. de C. Sowerby). ( a ) longitudinal
section through muscle platform of ventral valve to show the position of an earlier valve floor and
muscle attachment region. ( b ) longitudinal section adjacent to crus showing crural base formed of
callus and ‘embryonic' crus. These are in turn supported by a spoon-shaped depression in the brachial
valve, (c) longitudinal section through crus. All x 16.

Fimbria extend from the anterior part of the primary coils and are directed anteriorly.
They originate in pustule-like growths within the coils which produce surface corruga-
tions when fimbria are not fully developed. The pustules generally point medially.
Discrete pustules occur rarely on the lower part of the jugum.

Silicified specimens (PI. 119, fig. 9) clearly show the fimbria; they attain a maximum
length of over 0-175 mm.

Median septum. At its anterior end the median septum has the form of a low rounded
ridge and is little more than an internal bulge complementary to the external sulcus in
the brachial valve. In a posterior direction the septum is more prominent and ridge-like
with broad, low, supporting buttresses on its flanks. Towards the hinge line secondary
callus builds up over the median septum producing a more triangular form in

618

PALAEONTOLOGY, VOLUME 11

cross-section before the median septum fades into the lower part of the cardinal plate,
the transverse profile becoming increasingly subdued.

Ontogenetic changes of the median septum are preserved in the micro-structure of the
shell (text-fig. 3a, b). Within the secondary layer calcite was secreted about the mid-line,
successive growth layers being inclined from the shell outer surface laterally inwards
and gently convex outwards. The following stages can be recognized in the development
of the median septum from this layer:

(i) Simple fold of shell (see above);

(ii) Addition of secondary callus producing buttresses which are to serve as supports
to the median septum;

(iii) Successive addition of calcite layers to produce the characteristic ridge form,
there being little lateral addition of calcite at this stage;

(iv) Concentration of growth lines over crest of median septum indicating suppression
of development ventrally but the addition of calcite on the flanks producing a
triangular form;

(v) Extensions of process operating in (iv) to markedly increase the width; continua-
tion of this process finally flattening out the floor of the brachial valve at the
mid-line.

The full sequence is discernible within the brachial shell just below the hinge line. Peels
show darker areas of more concentrated lamellae at intervals. They increase in number
posteriorly and must represent growth pauses. One shell revealed at least five and
possibly six such growth pauses.

EXPLANATION OF PLATE 119

Figs. 1-7, 9-10. Dayia navicitla (J. de C. Sowerby). 1, Upper Llanbadoc Beds, Darran, Usk, Monmouth-
shire. National Grid ST/327 979 (GSM 103292). Internal mould of ventral valve showing chevron
diductor field, with the accessory diductor field in its apex, and the adductor muscle field. Between
the adductor muscle impressions faint ridges (? vascula media) exist (x7). 2, Lower Leintwardine
Beds, Greenway Cross, near Norton, Shropshire. National Grid SO/461 827. (SM A1 1207). Internal
mould of ventral valve showing diductor muscle field in an immature specimen (x5). 3-4, Lower
Leintwardine Beds, Greenway Cross, near Norton, Shropshire. National Grid SO/461 827. (SM
A1 1202 and All 206). Internal moulds of brachial valve showing median septum, adductor muscle
fields and position of dental sockets (both x 5). 5, Lower Leintwardine Beds, Greenway Cross, near
Norton, Shropshire. National Grid SO/461 827. (SM A1 1655-64 one of twelve specimens bearing
number 2000). Lateral view of internal mould of ventral valve showing part of chevron and vascular
markings ( ? gonadal sacs) ( x 5). 6, Lower Forest Beds. Llancayo, Usk, Monmouthshire. National
Grid SO/372 036. Internal mould of ventral valve showing chevron diductor impression (x6).
7 and 1 1, nilssoni-scanicus zone, S. side of Mynydd-y-Gaer, Denbigh Moors, Denbighshire. National
Grid SH/974 715. Internal moulds of ventral valves showing chevron diductor impression. Fig. 7 is
an immature specimen also showing the adductor impressions (x4 and x6 respectively). 9-10,
Lower Leintwardine Beds, Greenway Cross, near Norton, Shropshire. National Grid SO/461 827.
(SM A1 1655-664 (specimen numbered 2042) and All 201 ). Silicified specimens showing form of
spiralium. 9, lateral view exhibiting pustules and fimbria on spire, photographed through the silica
mould (this specimen has been cut along the left side). 10, looking from the anterior (x7 and x6
respectively).

Fig. 8. Dayia sp. Kobbinghauser Beds (Lower Devonian), Huinghausen, W. Germany. Internal
mould of posterior part of ventral valve showing relatively small diductor muscle field ( x 4).

Palaeontology , Vol. 1 1

PLATE 119

TUCKER, Day la

TUCKER: ATRYPIDINE BRACHIOPOD DAYIA NAVICULA (J. DE C. SOWERBY) 619

Further changes can be recognized in the shell micro-structure at the mid-line but
these are related to the changes in the cardinal plate.

Cardinal plate. Serial sections and latex casts provide no indication of the kidney-shaped
cardinal process ascribed to Dayia navicula by Alexander (1947, p. 307). Instead, it is
seen to have the form of a flattish or gently arched area lying within the brachial hinge
plate, divided by a pronounced cleft or septalium. Suppression of the median septum
by lateral addition of callus, as described above, reaches its ultimate conclusion in the

text-fig. 3. Micro-structure of shell of Dayia navicula (J. de C. Sowerby) at mid-line of dorsal valve
(m.s., median septum; i.s., inner shell surface; o.s., outer shell surface; s., septalium; c., crura;
d.s., dental socket), (a) transverse section through median septum x 45 (b) transverse section through
posterior part of median septum immediately below hinge line x 50. (c) transverse section through
cardinal plate below crura x 50, and (d) including crura X 40.

production of the cardinal plate. From then onwards crowding of growth-lines and
growth pause at the centre-line reverses the form of the inner shell surface to produce
the septalium (text-fig. 3 and PI. 121, fig. 2). This trough is flanked by the crura and is
in part a result of the development of the crura and expansion of the inner socket ridge.

Musculature. There is clear evidence for the former position of certain muscle fields but
the interpretation of these fields and of the particular function of the muscles is less easy
to deduce.

The pedicle valve contains a prominent chevron muscle-scar impressed on a platform
produced by the addition of secondary callus to the inner shell surface. The muscle field
occupies much of the shell width in mature specimens but only one-third of the width
in juveniles (PI. 119, figs. 1, 2). The apex of the chevron is directed posteriorly and
extends to a point level with or even above the hinge-line; however the major part of
the muscle field lies below this line. Two muscles occupied this field, one each side of

s s

C 5934

620

PALAEONTOLOGY, VOLUME 11

the shell mid-line, and they diverged at angles between 90° and 100°. One or more
narrow troughs, or sometimes ridges, cross the field parallel to its long axis. Weaker,
but more numerous, corrugations cross this line obliquely subparalleling the shell

b.

text-fig. 4. Muscle fields and postulated form of muscles in Dciyia navicula (J. de C. Sowerby).

mid-line. These are muscle tracks. The first set probably result from the anterior exten-
sion and migration of the muscle attachment areas during growth. This migration is
demonstrated by the micro-structure of the supporting platform where the secondary
addition of calcite is seen to obliterate earlier muscle fields which now lie ghost-like

TUCKER: ATRYPIDINE BRACHIOPOD DAYIA NAVICULA (J. DE C. SOWERBY) 621

within the layer of callus (see text-fig. 2). The oblique structures are believed to be due
to the lateral extension of the muscle areas.

Situated in the apex of the chevron is a small elliptical area often divided by a trough
(text-fig. 4 and PI. 1 1 9, fig. 1). The limits of the area are strongly defined and the area must
have provided an attachment region for separately functioning tissue — probably muscular.

Together, the above muscle fields are here interpreted as the diductor muscle attach-
ment areas, the main diductors being attached to the chevron field and accessory
diductors to the elliptical field. This interpretation differs from that placed upon the
muscle-fields by Alexander (1947, p. 308, text-fig. 3).

Anterior of the diductor muscle platform in the ventral valve two slightly divergent
troughs are weakly impressed into the shell: they are not always discernible. They extend
from a position close to the lower apex of the diductor chevron halfway towards the ante-
rior margin of the shell. The troughs exist in most growth stages but are most prominent
in mature shells (PI. 119, figs. 1, 7). These features were not recorded by Alexander;
the present writer interprets them as the ventral muscle-fields of the adductor muscles.

In the umbonal region of the ventral valve other corrugations exist on the inner shell
surface. Certain of these seem to have been interpreted as diductor muscle fields by
Alexander but it is difficult to understand how the muscles could function mechanically
if this interpretation is correct. These corrugations are possibly responses to folding of
the mantle following pedicle atrophy. Alternatively they may represent pedicle attach-
ment areas during early stages of growth, although their size tends to preclude such a
simple interpretation.

No muscle scars are preserved on the cardinal plate. However, the adductor muscles
left prominent impressions on the floor of the brachial valve. Although exhibiting some
variation in form, these are normally elongate twin fields, the posterior field embracing
half its companion (PI. 1 19, figs. 3, 4). The muscle fields lie partly on the flanks of, and
are separated by, the median septum which acts in part as a myophragm.

The writer’s interpretation of the muscle fields is shown in text-fig. 4, together with
a reconstruction of the form of the muscles as they crossed the visceral cavity. During
youth the disposition of the (small) diductor muscle fields in the ventral valve and the
cardinal plate in the dorsal valve permitted shell opening probably by uniform contrac-
tion of the muscles. In mature forms however the posterior arching of the ventral valve
transposed part of the main diductor field to a position posterior of the hinge axis. It was
inevitable therefore, that the main functional part of the muscle was that part lying
below or anterior of the hinge axis and the main diductor muscles became extended
anteriorly, with accompanying anterior extension of the muscle platform.

Vascular markings. Faint vascular markings are discernible between the diductor
muscle field and hinge extremity in the ventral valve of mature individuals and also
anterior of the muscle fields in both valves (see PI. 119, figs. 1, 5). Saccate or lemniscate
gonadal sacs can possibly be inferred from fig. 5 whilst fig. 1 shows divergent ridges
between the ventral adductor scars which may indicate the former site of the vascula
media. Other sinuses have not yet been recognized in Dayia navicula.

General micro-structure of the shell. Certain aspects of the micro-structure of the shell
have already been discussed and many points of detail can be interpreted from text-figs.
2 and 3 and Plate 121.

622

PALAEONTOLOGY, VOLUME 11

The inner secondary layer of the shell, composed of fibrous calcite, is more prominent
than the primary layer. The lamellae lie parallel or subparallel to adjacent shell surfaces,
notably the inner surface. Considerable secondary callus has been added at the inner
surface particularly around the median septum and on the ventral muscle platform. In
the latter instance, as well as assisting change of muscle position it partly accounts for the
highly arched and elevated form of the posterior region of the ventral valve. Within
the umbo of the ventral valve growth layers are concentrated together as a result of
growth pauses.

The very thin primary layer is frequently lost. Lamellae are closely spaced and seem
to be sub-parallel to the outer shell surface, though sometimes lying obliquely to it.
Alexander (1947, p. 309) suggested that the lamellae are ‘. . . directed forward and
inward’.

In the umbonal region of the ventral valve the continuity of shell structure is inter-
rupted by a number of ‘suture lines’, seen in transverse section as distinct seams
(sometimes followed by cracks (text-fig. 1 a-b)). They may be represented by inflections
in the shell microstructure in longitudinal section (see text-fig. 2). The most prominent
suture crosses the shell transversely and descends antero-ventrally through the secondary
layer meeting the valve floor at the diductor attachment area. Two weaker sutures
diverge ventrally from near the mid-line of the transverse suture; they appear to fuse
anteriorly and rapidly fade away. These structures are difficult to interpret but bearing
in mind the modifications affecting the muscle platform during ontogeny it is not
unreasonable to suggest that they either had a similar supporting function or at least
defined the limits of the muscles during some early stage of the individual’s develop-
ment. In this respect it is worth noting that a somewhat similar structure, a transverse
plate (shoe-lifter process), is exhibited by the Ordovician brachiopod Aulidospira (see
Williams 1962, p. 253) another member of the Dayiacea.

ALLIED FORMS

Dayia is not uncommon in rocks of Eltonian (Lower Ludlow) age in many parts of
Wales and the Welsh Borderland. The major difference between these forms and Dayia

EXPLANATION OF PLATE 120
Variants of Dayia navicula (J. de C. Sowerby).

Figs. 1-3. Dayia navicula cf. bohemica Boucek, upper part of Pridoli Beds (e/L), Bohemia. Dorsal,
anterior, and lateral views of three growth stages (all X 4).

Fig. 4. Dayia navicula (J. de C. Sowerby); Hemse Group (Lower Ludlow) Lau, Gotland, Sweden.

Dorsal, anterior, and lateral views of a mature specimen. (x4).

Fig. 5. Dayia navicula (J. de C. Sowerby); Skala Beds, Ruchotina, Podolia. Dorsal, anterior, and lateral
views. Ventral umbo imperfect (x4).

EXPLANATION OF PLATE 121

Microstructure of shell in Dayia navicula (J. de C. Sowerby). Photographs of cellulose acetate peels of
calcareous shell, fig. 1 and 2 slightly retouched. All specimens from Lower Leintwardine Beds, Down-
ton Gorge, Herefordshire. National Grid SO/431 733.

Fig. 1. Cardinal plate, showing shallow septalium (t.s. x 100).

Fig. 2. Growth line pattern associated with socket, crus (proximal end) and septalium (t.s. X 125).
Fig. 3. Pustules and fimbria on the lower part of the first coil of the spiralium (l.s. X 100).

Palaeontology, Vol. 11

PLATE 120

TUCKER, Dayia

Palaeontology, Vol. 11

PLATE 121

TUCKER, Dayia

TUCKER: ATRYPIDINE BRACHIOPOD DAYIA NAVICULA (J. DE C. SOWERBY) 623

navicuJa is their smaller size. Mature shells have an average length of between 7 and 8
mm. The degree of incurvature of the ventral umbo is similar although the posterior
part of this valve is normally proportionately less elevated above the hinge line. Most
specimens occur as internal moulds hence the external morphology is difficult to assess
fully but there appear to be no fundamental differences from the typical form.

The internal structures are also the same although it has not been possible to interpret
the spiralium fully. The intensity, but not always the length, of the chevron muscle scars
and platform is similar, and the presence of additional impressions within the chevron
supports the interpretation of the musculature. Specimens collected from the Mono-
graptus scanicus zone of the Denbigh Moors (North Wales) show chevron muscle
scars occupying at least two-thirds of the shell width (PI. 119, fig. 11), a relationship
which occurs in the Leintwardine Bed forms. Other specimens collected from the upper
part of the Elton Beds (locally, Lower Forest Beds) near Usk possess proportionately
smaller muscle fields which occupy approximately one-third of the shell width (PI. 1 19,
fig. 6). There is thus minor variation between the typical Dayia navicu/a and these
earlier forms and also variation within the latter. It would however be premature to
append varietal names until more comprehensive collections from rocks of Eltonian age
are available for examination.

I am indebted to Dr. Anders Martinsson and Dr. H. Jaeger for providing me with
specimens of Dayia from the Hemse Group (considered by Hede 1919 to belong to the
zone of Monograptus nitssoni ) of Scandinavia. No attempt at a full description is made
here but the following points are worth noting. The average dimensions are similar to
those of British Eltonian age but the posterior elevation of the ventral valve is greater
than in the Usk form: the umbo is again strongly incurved (PI. 120, fig. 4). The Scandi-
navian forms possibly possess closer relationships with the Denbigh Moor forms (which
unfortunately have not yielded good externals although the moulds show a prominent
umbonal feature). The chevron muscle field shows similar proportions to the Denbigh
variety also. One characteristic feature of the Scandinavian forms is the almost horizontal
disposition of the teeth; this is extremely rare in British forms. Almost all mature shells
possess a deep sulcus and a strongly flexed lateral commissure.

Few Podolian specimens of Dayia are available for comparison. Kozlowski’s descrip-
tion and illustration of this form in 1929, supplemented by material loaned to the writer
by Dr. O. Nikiforova, show the close similarity between it and Dayia navicula. Minor
differences in external morphology include a squarer outline to the brachial valve
resulting from the straighter hinge and the fact that maximum width is attained closer
to the hinge (PI. 120, fig. 5). The shell is relatively less wide than British examples but
with a deeper ventral valve, strongly incurved but less elevated at the umbo. The species
occurs in the Marnes de Dzwinogrod of the Skala Formation.

Similarities exist between certain Podolian and large Bohemian specimens (see
below) but the Podolian specimens are closer to the British forms.

In 1940, Boucek discussed variation within Dayia navicula in Bohemia. He recognized
two subspecies which he named Dayia navicula minor and Dayia navicula bohemica.
They are recorded from zones e^x and e/32 of the Budnany Beds which, according to
Jaeger (1962 and 1965) range from the Ludlovian into younger beds. It is clear from
Boucek’s illustrations that the shape of both varieties is different from the typical British
form. Horny (1955, p. 438) suggested that a transition exists between the two subspecies.

624

PALAEONTOLOGY, VOLUME 11

A collection of Bohemian Dayia from the Pridoli Beds (belonging to e/32) have been
studied by the writer. They include many growth stages. Based on external morphology

L DEVONIAN (PRIDOLY BEDS)
BOHEMIA

text-fig. 5. Scatter diagrams to show the relationship between incurvature of the ventral umbo and
overall size (height + width -(-thickness) in four assemblages of Dayia. Assemblages b and c occur
22 and 55 ft. respectively above the base of the Lower Leintwardine Beds. N, no. of individuals in
sample; r, degree of correlation between related regression lines.

alone, both Boucek’s subspecies appear to be present. With the exception of one
assemblage which may represent the true minor , mature individuals generally attain
greater dimensions than the British form. The ventral umbo is less strongly incurved
(see PI. 120, figs. 1-3) and the foramen probably remained unimpaired until later in

TUCKER: ATRYPIDINE BRACHIOPOD DAYIA NAVICULA (J. DE C. SOWERBY) 625

life; hence pedicle atrophy was delayed. The diductor muscle platform is weakly
developed. There is a suggestion of vestigial deltidial plates in some small juvenile
shells, although the observed features could result from slight curvature of the ventral
interarea at the delthyrium edge. Deltidial plates are not apparent in mature shells.

The hinge-line is straighter and proportionately longer than in British forms so
producing distinctly squarish cardinal extremities. The shells are less elongate and the
lateral margins are more sharply angular. Surface ornamentation is similar but the
costae are more prominent.

The Kobbinghauser Beds of Germany and their age equivalent the Schistes de
Lievin of Artois yield ‘ Dayia navicula'. This form is being studied by Dr. J. Shirley who
in 1962 expressed the view \ . that the “Z>. navicula ” of these areas and Lievin differs
specifically and constantly from the Middle Ludlow (i.e. Leintwardine Beds ) form and
is now to be referred to a new species D. tenuisepta sp. nov. (nomen nudum)’. A
specimen from the Kobbinghauser Schichten is figured on Plate 119, fig. 8. It is larger
than British forms but possesses relatively small and less prominent muscle attachment
areas.

STRATIGRAPHICAL CONSIDERATIONS

The true nature of the sequence of faunas in the various Siluro/Devonian successions
of Great Britain and Continental Europe is now more fully understood than a few
decades ago and closer agreement is being reached on correlation. Following Jaeger
(1962 and 1965) it is apparent that Dayia- bearing strata in the Rhineland, Artois,
Bohemia, and Podolia are younger than the British Ludlovian. Jaeger shows that
Monograptus extends upwards at least to the top of the Middle Siegenian in Germany
and Bohemia. Teller (1964, table 5) for Poland, also shows graptolites in beds younger
than the British Ludlow Bone Bed. Correlation between graptolitic and shelly facies in
these areas and adjacent Podolia demonstrates the high stratigraphic level attained by
Dayia and emphasizes the inapplicability of the genus as an index of Silurian rocks. The
characteristic British forms and the, in part contemporary, Scandinavian forms are
ancestral to these other varieties.

A study of assemblages separated in time brings out a fundamental change in external
morphology during phylogeny. Text-fig. 5 shows the relationship between incurvature
of the ventral umbo (see Tucker 1965, fig. 2) and overall size for four assemblages
ranging in age from Lower Ludlow to Lower Devonian. Regression lines have been
computed for each assemblage. The lower degree of correlation of samples (#) and (c) is
due to the low number of individuals and lack of juvenile forms in both samples. How-
ever, it can be seen that during phylogeny individuals tended to attain greater dimen-
sions whilst incurvature was progressively delayed and less intense. These changes are
most pronounced in the assemblage of Dayia navicula bohemica from the Lower
Devonian of Bohemia. The reduced rate of hooking would clearly postpone pedicle
atrophy, a condition facilitating survival of the individual under a wider range of
environment than the quiet water conditions suggested for Dayia navicula sensu stricto
by Tucker (1965).

Acknowledgements. The author is most grateful to Mr. J. D. D. Smith and Mr. D. E. White, and Mr.
A. G. Brighton for making available specimens in their care at the Geological Survey Museum and

626

PALAEONTOLOGY, VOLUME 11

Sedgwick Museum respectively. Mr. White drew the author’s attention to the uncertain status of the
assumed holotype. He is also grateful to those others mentioned in the text who supplied him with
specimens from abroad. Dr. F. A. Middlemiss kindly read the manuscript.

REFERENCES

Alexander, f. e. s. 1947. On Dayia navicula (J. de C. Sowerby) and Whitfieldella canalis (J. de C
Sowerby) from the English Silurian. Geol. Mag. 84, 304-16.
barrande, j. 1879. Systeme silurien du centre de la Boheme. v. Brachiopodes. Prague.
boucek, b. 1940. O variability ramenonozcu Dayia navicula (Sow.) a Cyrtia exporrecta (Wahl) a o
pouziti metod variacni statistiky v paleontologii. Rozpr. Ceske Akad. Ved Umeni, II, 50, no. 22, 1-27.
davidson, T. 1864-84. A monograph of the British fossil Brachiopoda, vols, 3 and 5. Palaeontogr. Soc.

\ Monogr .].

- 1881. On the Genera Merista, Suess, 1851, and Dayia Dav. 1881. Geol. Mag. 18, 290-3.

earp, J. R. 1938. The higher Silurian rocks of the Kerry District, Montgomeryshire. Q. Jl. geol. Soc.
Load. 94, 125-60.

elles, g. l. and slater, i. l. 1906. The highest Silurian rocks of the Ludlow District. Ibid. 62, 195-220.
hede, j. e. 1919. Om en forekomst of colonusskiffer vid Skarhult i Skane. Geol. For. Stockh. Forh. 41,
113-60.

horny, r. 1955. Studie o vrstvach budnanskych v zapadni casti barrandienskeho siluru. Sb. Ustred.
Ust. geol. 21, 315-447.

jaeger, h. 1962. Das Silur (Gotlandium) in Thiiringen und am Ostrand des Rheinischen Schiefer-
gebirges (Kellerwald, Marburg, Giessen). Symposiums-Band 2. Internat. Arbeit stagung iiber die Silur I
Devon-Grenze und die Stratigraphie von Silur und Devon, Bonn- Bruxelles 1960. pp. 108-35. Stuttgart.

- 1965. Referate. Symposiums-Band der 2. Internationalen Arbeitstagung iiber die Silur/Devon-
Grenze und die Stratigraphie von Silur und Devon, Bonn-Bruxelles 1960. Geologie, 14, 348-64.

KOZtowsKi, r. 1929. Les brachiopodes gothlandiens de la Podolie polonaise. Palaeont. pol. 1.
UNDSTROM, G. 1859. Bidrag till Kannedomen om Gotlands Brachiopoden. K. svenska Vetenskakad.
Hand/. 17, 337-82.

m‘coy, f. 1846. A synopsis of the Silurian fossils of Ireland. Dublin.

1852. Description of the British Palaeozoic fossils in the Geological Museum of the University of

Cambridge. London.

Phillips, j. 1848. The Malvern Hills compared with the Palaeozoic Districts of Abberley, etc. Mem.
Geol. Surv. G.B. 2, pt. 1.

reed, f. r. c. 1917. The Ordovician and Silurian brachiopods of the Girvan district. Trans. R. Soc.
Edinb. 51, 795-8.

shirley, j. 1939. Note on the occurrence of Dayia navicula (J. de C. Sowerby) in the Lower Ludlow
Rocks of Shropshire. Geol. Mag. 76, 360-1.

1952. The Ludlow rocks north of Craven Arms. Proc. Geol. Ass. Loud. 63, 201-6.

— — • 1962. Review of the correlation of the supposed Silurian strata of Artois, Westphalia, the Taunus
and Polish Podolia. Symposiums-Band 2. Internat. Arbeitstagung iiber die Silur/ Devon-Grenze und
die Stratigraphie von Silur und Devon, Bonn-Bruxelles 1960. pp. 234-42. Stuttgart.
sowerby, j. de c. 1839. In Murchison, r. The Silurian System. London.

straw, s. h. 1937. The higher Ludlovian rocks of the Builth District. Q. JL geol. Soc. Loud. 93,
406-53.

teller, l. 1964. Graptolite fauna and stratigraphy of the Ludlovian deposits of the Chelm Borehole,
Eastern Poland. Studia geol. pol. 13, 7-88.

tucker, e. v. 1965. The ecology of the brachiopod Dayia navicula (J. de C. Sowerby). Ann. Mag. nat.
Hist. (13) 7, 339-45.

williams, a. 1962. The Barr and Lower Ardmillan Series (Caradoc) of the Girvan District, South-west
Ayrshire, with descriptions of the Brachiopoda. Mem. geol. Soc. Lond. 3.

e. v. TUCKER
Geology Department
Queen Mary College
London, E. 1

Typescript received from author 2 November 1967

PLANICARDINIA , A NEW SEPTATE
DALMANELLID BRACHIOPOD FROM THE
LOWER DEVONIAN OF NEW SOUTH WALES

by N. M. SAVAGE

Abstract. Planicardinia is described as a new genus of septate dalmanellid occurring in deposits of probable
early Siegenian age in New South Wales. The new genus, known only by the type species Planicardinia canoli sp.
nov., is placed in the Mystrophoridae together with the genera Mystrophora , Kayserella , and Hypsomyonia.
It is the earliest known representative of the family and its presence at this horizon indicates that the origins of
the group are at least as old as the early lower Devonian.

Septate dalmanellid brachiopods have received considerable attention since 1953
when Havlicek described the new Middle Devonian genus Prokopia. In 1955 Cooper
re-examined the Middle Devonian forms Kayserella and Mystrophora and proposed the
new Middle and Upper Devonian genera Phragmophora, Hypsomyonia , and Mone-
lasmina. More recently Johnson (1966) described the new genus Vallomyonia from the
Middle Devonian of Nevada and Johnson and Talent (1967a) described the new
genus Muriferella from the late Lower Devonian of south-east Australia, Nevada, and
Arctic Canada.

The new genus described herein is from the early Siegenian Mandagery Park Forma-
tion near Manildra, New South Wales, where it occurs infrequently in a large silicified
brachiopod fauna (Savage 1967, 1968). It is the earliest known representative of the
Mystrophoridae and is therefore of particular interest. Little is known of the phylogeny
or origin of this family and the new genus does not directly throw new light on the
problem of a possible precursor. However it does show that a highly specialized mystro-
phorid existed in the middle Lower Devonian and that the ancestral forms of the family
can be expected in even older strata.

Attempts to classify septate dalmanellid genera have resulted in several different
schemes which disagree markedly. Nevertheless most authors believe the group to be
polyphyletic with septate dorsal valves evolving independently in more than a single
lineage. Cooper (1955) referred Mystrophora and Kayserella to the Mystrophoridae,
Phragmophora and Hypsomyonia to the Onniellidae, and Moneiasmina to the Schizo-
phoriidae. Wright (in Moore 1965) proposed that Kayserella , Prokopia , Moneiasmina ,
and Phragmophora be grouped into a new family, the Kayserellidae, with Hypsomyonia
the sole member of another new family, the Hypsomyoniidae, leaving Mystrophora the
only remaining member of the Mystrophoridae. Johnson (1966) and Johnson and Talent
(1967a) referred the new genera Vallomyonia and Muriferella to the Schizophoriidae.
These latter authors discuss possible phylogenetic relationships in some detail (1967a,
pp. 44-6) and it is clear that they envisage three distinct groups. The genera Muriferella ,
Vallomyonia , Moneiasmina , and Hypsomyonia are all placed in the subfamily Draboviinae
together with Salopina which is suggested as a possible ancestor, Prokopia and Phrag-
mophora are separated as a second lineage, and a third group consists of Mystrophora ,
Kayserella , and the new genus described herein (see Johnson and Talent 1967a, p. 46).

[Palaeontology, Vol. II, Part 4, 1968, pp. 627-32, pi. 122.]

628

PALAEONTOLOGY, VOLUME 11

It seems very probable that Planicardinia is closely related to Mystrophora and to
KaysereUci. The three genera have a number of features in common, of which most
notable is the possession of a cruralium in the dorsal valve. Cooper (1955, p. 47) does
not appear to have been convinced that the small cruralium in Kayserella was the sole
attachment surface for the dorsal adductor muscles. Furthermore, Wright (1965, in
Moore, p. h338) refers to the cruralium as a septalium and suggests that the adductor
muscles were attached to the valve floor. However, Biernat (1959, p. 38) has demon-
strated that some Polish specimens of Kayserella possess a cruralium extending half the
length of the dorsal valve (Biernat 1959, pi. iii, fig. 7) and it now seems probable that
this structure was a functional muscle platform.

Hypsomyonia appears to be more closely related to Mystrophora than to Muriferella
and Monelasmina. It has a large elevated cruralium which clearly functioned as a
muscle attachment surface and a typical mystrophorid pentagonal outline with an
emarginate anterior margin. Hypsomyonia has therefore been assigned herein to the
Mystrophoridae.

It is probable that additional material will eventually be found which will indicate the
true relationships of the various septate dalmanellids. Meanwhile any classification is
likely to be only tentative. It appears to the author that the septate genera with a
cruralium form a distinct phylogenetic group and the classification below differs from
that of earlier authors in this respect.

SYSTEMATIC DESCRIPTION

Phylum BRACHIOPODA
Suborder dalmanelloidea
Superfamily dalmanellacea Schuchert 1913
Family mystrophoridae Schuchert and Cooper 1931

Revised diagnosis. Small, ventribiconvex, dorsally sulcate Dalmanellacea with a long
apsacline or anacline ventral interarea and a shorter anacline or hypercline dorsal
interarea. Both the delthyrium and notothyrium are open. The ornamentation is costel-
late to subfascicostellate. The dorsal interior has a high median septum supporting a
large cruralium, a bilobed cardinal process, and ventrally or ventro-laterally elongate
brachiophores. The ventral interior has a short, wide muscle field with the adductor
scars not enclosed by the diductor scars.

Genera included : Mystrophora Kayser, 1871

Planicardinia gen. nov.

Kayserella Hall and Clarke, 1892
Hypsomyonia Cooper, 1955

Genus planicardinia nov.

Type species. Planicardinia carroli sp. nov.

Diagnosis. A plano-convex mystrophorid with an anacline ventral interarea and a
hypercline dorsal interarea. Heavy triangular teeth are supported by strong dental

N. M. SAVAGE: PLANICARDINIA , A NEW SEPTATE DALMANELLID 629

lamellae. The dorsal interior has a small bilobed cardinal process, ventrally elongate
brachiophores and a vertical spoon-shaped cruralium supported on a variable median
septum.

Discussion. Known only by the type species, Planicardinia most closely resembles the
Western European Middle Devonian genus Mystrophora, particularly in the features of
the dorsal interior. Comparison with the type species of Mystrophora (Cooper 1955,
pi. 11, figs. 45-9) shows a marked similarity in the form of the ventrally elongate
brachiophores with a crescentic cross-section and the high spoon-shaped cruralium. The
ventral interior of Planicardinia also closely resembles that of Mystrophora with a
similar short thickened muscle field. However, the shell shape of Planicardinia is distinc-
tive with the anacline ventral interarea and hypercline dorsal interarea together forming
a prominent part of the dorsal surface of the plano-convex shell. Another difference is
in the orientation of the cruralium which in Planicardinia is inclined almost normal to
the dorsal valve floor, a feature clearly related to the extreme curvature of the ventral
umbo which necessitates adductor muscles positioned almost parallel to the com-
missural plane (text-fig. 1). The ventral ‘median septum’ in Mystrophora , described and
figured by Cooper (1955, pi. 11, fig. 50), has been observed in only a single specimen
where it showed as a narrow slit in the internal filling of a calcined specimen. This slit
probably represents the position of the high dorsal septum extending to the floor of
the ventral valve.

Hypsomyonia is easily distinguished from Planicardinia for it possesses a much less
curved ventral umbo and an interarea which is apsacline and not anacline. Internally
the brachiophores are more laterally directed so that they almost parallel the hinge line
and the cruralium is more gently inclined.

Kayserella is also very distinct from Planicardinia externally. Like Hypsomyonia it
has posteriorly directed interareas and a less rounded outline. Internally the dorsal
median septum is longer and the cruralium is much smaller and less upright. In addition,
the brachiophores are much shorter.

Planicardinia carroli sp. nov.

Plate 122

Material. The species occurs very infrequently and only 16 specimens have been recovered from
several thousand shells dissolved from a silicified limestone horizon. Of these 2 are complete shells
with the valves conjoined, 8 are ventral valves, and 6 are dorsal valves. The specimen numbers used are
those of the Palaeontology Collection, Department of Geology and Geophysics, University of Sydney.
Specimen SU 19540 is designated the holotype. Other illustrated specimens are paratypes.

Description. Exterior. The shell is subcircular to subpentagonal in outline with the
greatest width at about midlength. The cardinal margins are very obtuse and the
anterior margin is emarginate to slightly rounded. In lateral profile the shell is plano-
convex (PI. 122, fig. 25).

The ventral valve is strongly convex with the umbo swollen and prominent. The
beak is erect or incurved and appears to be commonly resorbed by the pedicle (PI. 122,
fig. 15). A long concave interarea is anacline and has an apical angle of about 130°.

630 PALAEONTOLOGY, VOLUME 11

The delthyrium, which has a width of about one-quarter that of the hinge-line, includes
an angle of about 35°.

The dorsal valve is transversely oval to pentagonal in outline and planar or weakly
convex. In the mature dorsal valves there is a distinctive convexity at the valve margin
(PI. 122, fig. 10). The umbo is poorly differentiated and the small distinct beak is postero-
dorsally directed (PI. 122, fig. 25). A long, slightly concave, hypercline interarea makes
a very obtuse angle with the valve surface. (PI. 122, fig. 25). It has an apical angle of
about 150°. The notothyrium is open and includes an angle of about 50°.

A low broad sulcus in the dorsal valve extends from the beak to the gently sulcate
anterior commissure. No distinct fold is present in the ventral valve. The surface is
costellate with about thirty-four costellae 2 mm. from the dorsal beak. The costellae
increase chiefly by intercalation.

Interior of ventral valve. In the ventral interior a broad, deep delthyrial cavity is bounded
by short vertical dental lamellae which diverge anteriorly at about 40° (PI. 122, fig. 26).
The anterior edges of the lamellae recede slightly and meet the strongly curved valve
floor at about one-sixth the distance from the beak. Deep lateral umbonal cavities are
present. The teeth are strong and pointed with a triangular profile (PI. 122, fig. 15).
A short sub-triangular muscle field is slightly elevated on the thickened floor of the
delthyrial cavity but individual scars have not been observed (PI. 122, figs. 14, 26).

Interior of dorsal valve. The dorsal interior has a small bilobed cardinal process set
deeply in the apex of the notothyrium. The brachiophores are ventrally elongate and
blade-like, diverging from the valve floor at about 50° and continuous with the margins
of the notothyrium for part of their length (pi. 122, figs. 7, 10). They are crescentic in
cross-section and bluntly pointed. The sockets are deeply excavated into the posterior
of the brachiophore bases just ventral to the interarea margin (PI. 122, figs. 10, 21).

EXPLANATION OF PLATE 122

Planicardinia carrolli sp. nov.

Figs. 1-5. Ventral, dorsal, anterior, posterior, and lateral views of complete specimen SU 22665. Note
the dorsally directed interareas and the strongly convex ventral valve (fig. 5) (x 6-5).

Figs. 6-11. Ventral, dorsal, postero-dorsal, antero-ventral, lateral, and postero-ventral views of the
isolated mature dorsal valve SU 19541 showing the very high brachiophores (fig. 10), large spoon-
shaped cruralium, and the small bilobed cardinal process (fig. 9) (x 6-5).

Figs. 12-14. Antero-dorsal, ventral, and dorsal views of broken ventral valve SU 22666. Note the
deep, lateral, umbonal cavities and the elevated muscle-field ( X 6-5).

Figs. 15-17. Dorsal, ventral, and lateral views of ventral valve SU 19542 showing the resorbed beak,
strong triangular teeth (fig. 15), and external ornament (x6-5).

Figs. 18-21. Ventral, dorsal, antero-ventral, and lateral views of mature dorsal valve SU 22667.
Fig. 18 shows the spoon-shaped form of the cruralium and the degree to which the median septum
has been almost totally resorbed (x 6-5).

Figs. 22-5. Posterior, dorsal, anterior, and lateral views of complete specimen SU 19540 (holotype)
showing the large dorsally directed interareas (fig. 23), the strongly convex ventral valve (fig. 25),
and the gently sulcate anterior commissure (x 6-5).

Fig. 26. Antero-dorsal view of broken ventral valve SU 19543. Note the strong dental lamellae and the
thickened muscle area ( x 6-5).

Figs. 27-32. Ventral, dorsal, postero-ventral, antero-ventral, lateral, and posterior views of the young
dorsal valve SU 19544 showing the relatively larger median septum extending into the cruralium
(fig. 27), and sockets excavated into the bases of the brachiophores (fig. 29) (x 8).

Palaeontology , Vol. 11

PLATE 122

SAVAGE, Planicardinia

N. M. SAVAGE: PLANICARDINIA, A NEW SEPTATE DALMANELL1D 631

A high, almost vertical cruralium arises close to the brachiophore bases and extends
anteriorly to about half the valve length.

The spoon-shaped cruralium is supported
by a high median septum which extends
a little into the cruralium cavity and in
young forms has a maximum elevation
near the anterior valve margin (PI. 122,
figs. 1 8, 27, 3 1 ). The septum is much shorter
in mature specimens (PI. 122, figs. 8, 18).

Adductor muscles were presumably at-
tached to the inner surface of the crura-
lium but distinct scars have not been
observed in the few specimens available.

The interior of both valves usually reflects the external costellation.

Measurements (in mm.)

SU 19540

Complete shell

Length

4-2

Width

4-3

Thickness

21

SU 19541

Dorsal valve

3-6

61

—

SU 19542

Ventral valve

4-6

51 (est.)

—

SU 19544

Dorsal valve

1-5

2-6

—

text-fig. 1. Planicardinia carroli gen. et sp. nov.
Sketch to illustrate the position of the cruralium,
brachiophores, and dental lamellae within a com-
plete shell and the probable position of the ad-
ductor and diductor muscles.

Ontogeny. The very few specimens of this species include a young dorsal valve and two
mature dorsal valves. In the mature specimens the median septum appears to be
partly resorbed both anterior and posterior of the almost vertical cruralium, but in the
young specimen it extends anteriorly almost to the valve margin and posteriorly well
into the cruralium cavity (PI. 122, figs. 27, 29). Another feature of the mature specimens
is the extreme ventral elongation of the brachiophores and cruralium which greatly
exceeds that in the small form (PI. 122, figs. 10, 21). The large spoon-shaped cruralium
seems to be deeply resorbed in one of the mature specimens by the two principal trunks
of the lemniscate vascu/a media (PI. 122, figs. 8, 11).

Occurrence. The species occurs in the silicified limestone horizon at the base of the
Mandagery Park Formation 3 miles south of the town of Manildra, New South Wales,
at Locality 1 of Savage (1968).

Acknowledgements. The author would like to express his gratitude for the assistance of Professor C. E.
Marshall and Dr. G. H. Packham of the Department of Geology and Geophysics, University of Sydney
where this work was commenced during the tenure of a Teaching Fellowship; also to Professor
F. H. T. Rhodes for the use of the facilities in the Department of Geology, University College of
Swansea. It is also a pleasure to acknowledge the helpful comments of Dr. J. A. Talent of the Geological
Survey of Victoria, and Drs. A. J. Boucot and J. G. Johnson of the California Institute of Technology.

REFERENCES

biernat, g. 1959. Middle Devonian Orthoidea of the Holy Cross Mountains and their ontogeny.
Palaeont. polon. 10, 1-78, pi. 1-12.

cooper, g. a. 1955. New Genera of Middle Paleozoic brachiopods. J. Paleont. 29, 45-63, pi. 1 1-14.

632

PALAEONTOLOGY, VOLUME 11

havucek, v. 1953. O nekolika novych ramenonozcich ceskeho a moravskeho stredniho devonu. Vest,
ustred. Ust. geol. 28, 4-9, pi. 1-2.

Johnson, j. g. 1966. Middle Devonian brachiopods from the Roberts Mountains, central Nevada.
Palaeontology, 9, 152-81, pi. 23-7.

and talent, j. a. 1967«. Muriferella, a new genus of Lower Devonian septate dalmanellid. Proc.

R. Soc. Viet. 80, 43-50, pi. 9, 10.

19676. Cortezorthinae, a new subfamily of Siluro-Devonian dalmanellid brachiopods.

Palaeontology, 10, 142-70, pi. 19-22.

kayser, e. 1871. Die Brachiopoden des Mittel- und Ober-Devon der Eifel. Z. dt. geol. Ges. 23, 491-647
pi. 9-14.

savage, n. m. 1967. Studies in the Silurian and Devonian of the Manildra District, New South Wales.
Unpublished Ph.D. thesis, Univ. Sydney.

1968. The Geology of the Manildra District, New South Wales. J. Proc. R. Soc. N.S. W. (in press).

schuchert, c and cooper, g. a. 1932. Brachiopod genera of the suborders Orthoidea and Penta-
meroidea. Mem. Peabody Mus. Yale, 4, pt. 1, 1-270, pi. 1-29.
wright, a. d. 1965. Superfamily Enteletacea. In moore, r. c. (ed.). Treatise on invertebrate paleon-
tology, Part H, Brachiopoda, EI328-H346, Lawrence, Kansas.

n. m. savage
Department of Geology
University of Natal
Durban, Natal

Typescript received from author 18 March 1968 S. Africa

PROBABLE DISPERSED SPORES OF
CRETACEOUS EQUISETITES

by D. J. BATTEN

Abstract. Dispersed spores that are considered to be those of Cretaceous Equiselites have been recovered from
Wadhurst Clay ( ?VaIanginian) sediments of southern England. They are described as Pilasporites allenii sp. nov.

In the course of a study of the palynology of British Wealden delta sediments, the plant
microfossil content of Equisetites soil-beds (in which rootlets, rhizomes, and stems of
Equisetites, in various combinations, are preserved in situ) and partings with aerial
debris of Equisetites (Allen 1941, 1947, 1959, 1967) was examined in the hope that the
spores of these plants, which have not previously been recognized in Wealden sediments,
might be recovered.

Samples used in this study came from Wadhurst Clay horizons in the Cuckfield No. 1
Borehole (Stubblefield 1967) and exposures at the Railway Brickyard, Sharpthorne,
and in the lower clay pit at Freshfield Lane Brickworks, Danehill (text-figs. 1, 2). The
soil-bed at Freshfield Lane Brickworks was recorded by Gallois (Stubblefield 1963, p. 37)
as being ‘the High Brooms E. lyelli soil-bed, some 25 ft. below the top of the Wadhurst
Clay’. Fourteen soil-bed and fragment parting samples yielded a characteristic palyno-
logical assemblage dominated (> 29-5%, based on a count of 200 specimens) by alete
spores described herein from one of these assemblages as Pilasporites allenii sp. nov.,
and referred to elsewhere as cfA. P. allenii if not from the type assemblage, following
the procedure recommended by Hughes and Moody-Stuart (19676). Seventeen adjacent
samples yielded assemblages in which these spores are frequent (>4-5-<30%).

Although recently some authors have been using the names of living genera for
Mesozoic plants, I consider that the convention of using a different name for Mesozoic
plants is a useful one and should be retained. Thus, the name Equisetites has been used
here rather than Equisetum.

The apparent absence from the palynological record of spores of Equisetites , the com-
mon Wealden macrofossil, has in the past been variously attributed to lack of preservation,
recognition, or production. The restriction of assemblages in which the inaperturate spores
mentioned above are dominant, to Equisetites soil-beds and fragment partings, is con-
sidered to be evidence in favour of an association of these spores with Equisetites. Of the
350 preparations I have made from Wealden samples, and of the many preparations in the
Sedgwick Museum palynology collection, only those prepared from samples associated
with soil-beds and fragment partings have yielded these spores in abundance. They have
rarely been recorded from other palynological facies. The relationship was tested
after the association had been recognized from the Cuckfield and Sharpthorne assem-
blages. I decided to prepare a further series of samples collected from Danehill where

[Palaeontology, Vol. 11, Part 4, 1968, pp. 633-42, pi. 123.]

A

B

Cuckfield Nol Borehoie
Sidnye Farm (TQ 296 273)
Sussex

rootlet beds
ironstone around roots

shell, ostracod and planty
laminae(mainly Equisetites
remains)

grey(N5)

rhizomes, roots, rootlets

rootlets

ostracods

ostracods

SCALE IN FEET
Silt,medium and fine 3 —

(laminated)

SiltyClay

Clay 2

Plant Fragments

Sample number prefix : CUC
x

Not. Grid Ref.

Shells
SOIL BED

E_q uisetites (lyelli ) , s tern s
rhizomes,root(let)s

o-J

text-fig. 1. a and b, Wadhurst Clay sediment cores from Cuckfield No. 1 Borehole. Details of litho-
logies (diagrammatic, obtained in part from the Geological Survey log) and horizons from which
samples have been prepared for microscopical examination. The colour classification (see also text-fig.
2) is from the Rock-Color Chart (1963) published by the Geological Society of America. Colours have
generally been noted when there is a departure from the predominantly grey (N3-N7) colours of the
Wadhurst sediments. For the sedimentary rock nomenclature, a modified Wentworth grade scale has
been followed. In beds where rhizomes and stems have not been seen but rootlets are recorded, it is
likely that the latter are those of Equisetites. The medium grey clays (sometimes shales) are generally
interlaminated with paler, generally very thin ( < 1 mm.) siltstones. The presence of shells and ostracods

has only been noted when they are abundant.

BATTEN: PROBABLE DISPERSED SPORES OF CRETACEOUS EQUISETITES 635

soil-beds and fragment partings had been observed in the field. The first samples proces-
sed were the fragment parting samples (DB 367, 366, 365, and 358) and I predicted that

Railway Brickyard

SHARPTHORNEfTQ 375330)
Sussex

Freshfield Lane Brickworks
DANEHILL(TQ382 266)

rootlets, especially concentrated in
medium grey clay partings(in situ?) 369-

medium grey clay(N5)
abundant plant remains (Equiseti tes)
scattered ostracods

moderate brown (5 YR 4/4)
Equisetites stems, rhizomes,rootlets
tbin plant fragments partings
ostracods

== Equisetites fragments

greyish yellow(5Y 8/4) and
lightgrey (N 7)

?stems, rhizomes, root(let)s of
Equisetites sp.

SCALE IN FEET
3— |

2-

1-

light brownish grey
(5YR 6/l)rootlets

o-J

rootlets,? rhizomes

ostracods,plant debris

medium dark grey (t
rootlets rhizomesof E

slightly calcareous

medium lightgrey (N6)
Equisetites stems(?), rhizomes
rootlets

rootlets

slight ly calca reous

stems (?), rhizomes,
rootlets

dark grey(N3)slightly calcareous

Equisetites lyelli stems,
rhizomes, rootlets
scattered ostracods
medium grey (N5)

Sample number prefix : DB

Sample number prefix:DB

text-fig. 2. a, Part of the (lower) Wadhurst Clay succession exposed at the Railway Brickyard,
Sharpthorne, Sussex, b, (upper) The Wadhurst Clay succession exposed at Freshfield Lane Brickworks,
Danehill. Details of lithologies (diagrammatic) and of horizons from which samples have been pre-
pared for microscopical investigation. For ‘Key’ and other comments, see text-fig. 1.

they would yield assemblages in which cfA. P. allenii is dominant (> 29-5%). The
prediction proved to be correct. Several of the soil-bed samples were subsequently
found to contain an abundance of these spores (see Table 1 and text-fig. 2b).

Tt

C 5934

A Cuckfield BH

SAMPLE No (prefix CUC)

789

790

791

791/11

792

792/1

792/2

792/3

792/6

793

794

794/6

796

797

cf A. P. allenii

0

8

3

31

58

25

22

60

5

X

3

X

11

X

Other inaperturates

7

14

10

10

X

10

12

8

11

11

7

12

13

10

bisaccates

54

32

34

13

9

27

20

5

28

26

29

21

30

43

Classopoll is

4

8

X

2

X

2

2

X

3

2

4

X

5

2

total ‘fern1 spores

28

45

46

31

23

31

35

20

45

55

51

60

25

24

Cicatricosisporites

X

15

14

4

5

7

9

6

12

18

17

16

X

2

B 'Cuckfield BH

SAMPLE No (prefix CUC)

863

864

865/3

866

867

868

869

cf A. P. allenii

2

28

60

67

5

X

X

Other inaperturates

7

8

7

10

8

7

8

bisaccates

49

15

12

8

29

28

41

Classopol 1 is

10

2

3

3

X

3

11

total 'fern' spores

23

38

14

9

54

54

25

Cicatricosisporites

2

4

2

X

8

11

5

C Sharpthorne

SAMPLE No (prefix DB)

332

331

330

347

329

349

328

327

325

324

321

320

319

318

317

316

cfA. P. al 1 enii

56

54

9

6

5

30

X

X

2

2

Other inaperturates

2

4

4

9

6

6

X

7

10

12

16

10

14

X

5

12

bisaccates

25

30

16

16

27

21

43

20

41

40

25

49

24

35

40

26

Classopollis

40

35

5

X

32

3

25

18

16

10

18

20

30

10

25

40

total 'fern' spores

26

22

11

16

19

40

18

22

12

25

31

18

16

16

25

14

Cicatricosisporites

10

6

X

4

X

15

2

X

X

X

2

X

X

X

X

X

D Danehill

SAMPLE No (prefix DB)

369

368

367

366

365

364

363

362

361

360

359

358

357

355

356

cf A . P. al lenii

11

13

35

38

64

6

6

2

6

36

46

60

6

5

1

Other inaperturates

11

20

17

10

13

14

11

12

8

6

n

12

8

7

7

bisaccates

38

34

20

24

6

41

28

24

36

24

24

8

24

30

36

Classopoll is

3

4

3

7

X

4

X

4

X

2

1

2

total 'fern' spores

33

25

19

18

8

26

49

49

39

28

14

16

54

54

48

Cicatricosisporites

5

4

3

X

X

8

10

11

8

8

2

X

10

12

10

table 1 . Distribution of cfA. Pilasporites allenii and major spore and pollen groups in some of the
samples studied (see text-figs. 1 and 2) recorded as percentages, based on counts of 200 specimens.
A and b, Cuckfield No. 1 Borehole, c, Railway Brickyard, Sharpthorne. D, Freshfield Lane Brickworks,
Danehill. X indicates presence, but 1% or less. The absolute bisaccate and fern spore counts are fairly
constant (see Hughes and Moody-Stuart 1967a) and are only apparently reduced in numbers when
cfA. P. allenii or Classopollis are abundant due to their non-dispersal. The sample numbers are as
they appear in text-figs. 1 and 2, from the highest point to the lowest point in the successions and not

in numerical order.

BATTEN: PROBABLE DISPERSED SPORES OF CRETACEOUS EQUISETITES 637

DISCUSSION OF THE OBSERVATIONS

Although spores cfA. P. allenii are usually frequent in the soil-beds, some horizons lack
an abundance of these spores (Table 1). This may be for several reasons. The spore species
in most of the soil-bed assemblages are poorly preserved. The poor preservation is
partially related to sediment type which probably in turn partly reflects the environment
of deposition. Allen (1947) stated that the Wealden Equisetites plants probably grew in
shallow waters, colonizing a specific locality for a short period of time, perhaps for as
little as three or four years. If, like living Equisetum, they grew predominantly in aerated
waters, many of their spores may thus have been deposited in an environment, at times
unfavourable for preservation, in which they were largely destroyed. When and where
conditions were more favourable for preservation, indicated by the generally better
state of preservation of most of the spore species, cfA. P. allenii is found preserved in
abundance. Winnowing and subsequent dispersal of these spores in the waters of the
Wealden basin could account for a lack of abundance in some horizons, and so could
a low rate of spore production, or seasonal production and/or a rapid sediment accumula-
tion rate in the soil-bed localities. Since the Equisetites communities were probably
relatively local and only existed for short periods of time, even if the plants dispersed
large quantities of spores frequently, the total number produced would be small when
compared with the number produced by the plants living on the near-by land. Thus if
Equisetites spores were dispersed over a large part of the Wealden basin, the chances
of obtaining an assemblage in which they constitute more than 5% of the total' number
of spores counted is slight. If the plants did not produce large numbers of spores, as is
possible, the likelihood is even less. They have not been observed to be more susceptible
to corrosion than other dispersed spores, and in fact, when abundant, they are generally
better preserved than the majority of bisaccates and some of the ‘fern’ spores in
the assemblage. The poor preservation of the latter forms may however be due to
reworking.

Some movement in the water (Allen 1941, p. 371), perhaps the result of storm wave
action, at times caused the destruction of the aerial stems of the Equisetites plants and
subsequent deposition of the fragment beds and partings. The fragments are preserved
as flattened carbonaceous remains as opposed to stems preserved ‘in situ’ in which the
pith cavities have generally been filled with sediment forming pith casts. The cones
(strobili) of living species of Equisetum rapidly fragment after spore dispersal. If
Equisetites cones similarly decomposed, it is unlikely that a mature cone would have
been preserved in Wealden sediments. Only when a cone was prevented from reaching
maturity, for example if the supporting stem was broken or the entire plant became
suddenly inundated with sediments, would preservation be possible. Although a strobi-
lus has yet to be recognized, cfA. P. allenii spore masses are frequently found, especially
in assemblages extracted from fragment partings, and are probably microscopic remains
of strobili (sporangia) broken up prior to or at the time of deposition.

Elaters which characterize the spores of living Equisetum have not been seen surround-
ing cfA. P. allenii or separated from them in any of the assemblages. They may have been
destroyed during the preparation procedure. However, unless their chemical make-up
is such that they are readily destroyed by the chemicals used during preparation, or were
removed from the spores at the time of deposition and rarely fossilized, they should

638

PALAEONTOLOGY, VOLUME 11

be present in the palynological preparations. It is possible that the Wealden Equisetites
species concerned produced spores which lacked elaters.

Apart from some Eocene Equisetum spores (Chandler 1964), and Elciterites triferens
Wilson 1943, a Pennsylvanian tri-radiate spore, no other fossil spore species recorded
to date possess elaters. They were not found surrounding the spores described from
Equisetites compressions by Halle (1908), Hartung (1933), or Gould (1968).

Specimens recorded were all those present along selected traverses of strew slides.
Stage co-ordinates refer to Leitz Laborlux (LI) microscope, number 557187, Depart-
ment of Geology, Cambridge University. The slides have been deposited in the Sedgwick
Museum palynology collection.

RECORD

Anteturma sporites H. Potonie 1893
Turma aletes Ibrahim 1933

Subturma azonaletes (Luber 1935) Potonie and Kremp 1954
Genus pilasporites Balme and Hennelly 1956

Type species. Pilasporites calculus Balme and Hennelly 1956, p. 64, pi. 3, figs. 60-4.

Remarks. Several genera have been erected for inaperturate miospores but their diag-
noses are not distinct. I consider that it is not practical or helpful to attempt to distin-
guish this new spore from all published inaperturates of all ages; it is however, desirable
and necessary to group it with and distinguish it from other Mesozoic inaperturates.
Pilasporites , the late Palaeozoic (Permian) genus has therefore been selected instead
of a Tertiary-based name. The alternative solution of erecting a new genus has been
rejected because the primary purpose of this paper is to present an accurate account of
the new species, rather than to classify it.

Pilasporites allenii sp. nov.

Plate 123, figs. 1-4, 6, 10-14

Type sample. CUC 792, Cuckfield No. 1 Borehole, Sidnye Farm, Surrey, England (TQ 296 273) from
depth 792 ft. Wadhurst Clay, ?Valanginian. Medium dark grey (N4) clay, size of coarse fraction 25 /a;

explanation of plate 123

Figs. 1-4, 6, 10-14. Pilasporites allenii sp. nov.; spores from rock sample CUC 792, preparation
T 210, X 1,000. 1, Loose ?ektexine, dark area; T 210/10, LI 33-3 124-8. 2, Smooth form; T 210/1,
LI 32-0 118 0. 3, Scabrate exine, development of major fold; T 210/5, LI 49-7 120-1. 4, Scabrate
exine; T 210/1, LI 34-1 116-1. 6, More or less smooth exine, scabrate perine; T 210/1, LI 42-0 1 16-2.
10, Smooth exine, smooth crinkled perine;T 210/1, LI 34-7 117-1. 1 1 , Smooth exine, closely adhering
perine with granules and small verrucate elements attached to the perine; T 210/10, LI 29-0 123-4.
12, Dark area, scabrate, secondary folds developed around dark area; T 210/14, LI 55-0 126-8. 13,
Split form; T 210/2, LI 30-8 123-3. 14, Split form; T 210/1 43-5 116-0.

Figs. 5, 7, 8, 9, 15, cf A. Pilasporites allenii sp. nov.; spores from rock sample CUC 792/3, preparation
T 244, x 1,000. 5, Showing dark area, only faintly visible; T 244/4, LI 27-7 118-1. 7, Major fold,
loosened ?ektexine; T 244/4, LI 51-6 125-2. 8, Scabrate exine, major fold; T 244/4, LI 58-0 116-0. 9,
Smooth exine, dark area, closely adhering smooth, crinkled perine; T 244/4, LI 28-1 116-0. 15,
Scabrate exine, loose ?ektexine; T 244/4, LI 31-9 121-0.

Palaeontology, Vol. 11

PLATE 123

BATTEN, Lower Cretaceous dispersed spores

BATTEN: PROBABLE DISPERSED SPORES OF CRETACEOUS EQU ISETITES 639

thin fine silt laminations; Equisetites ‘in situ’ and fragments of the same. Preparation: oxidation 10
minutes Schulze solution, 3 minutes 5% ammonium hydroxide, mineral separation using zinc bromide;
short centrifuging; strew slides, Clearcol and Euparal. Palynological facies: Pilasporites allenii well
preserved, the bisaccates and some of the ‘fern’ spores less well preserved; 58% P. allenii, 23% total
‘fern’ spores of which 7% are Ischyosporites and 5% Cicatricosisporites, 9% bisaccates (based on
a count of 200 specimens).

Diagnosis. Miospores, alete, mean diameter 36-7 p, standard deviation 5-36 p (200
specimens), amb circular. Exine 1-25-1 -75 p thick, smooth, or scabrate. Frequently only
one (major) fold present, usually developed close to and sub-parallel to the margin of
the spore. Spores sometimes split (7% of the specimens), may be surrounded by a
crinkled and folded perine c. 0-25 p thick. Perine smooth or scabrate. An indistinct
darker coloured area is seen on 4-5% of the specimens, averaging 14/x in diameter.

Holotype. Slide preparation T 210/1, LI 42-0 1 16-2; Plate 123, fig. 6.

Description. Maximum diameter, observed limits 28-48 p\ coefficient of variation 14-6%.
94-5% of specimens fall between 31 and 44 p (text-fig. 3). Shape may be distorted by
folding (PI. 123, fig. 7), 2% of spores unfolded. Exine thickness not always readily
measured owing to difficulty in obtaining an optical section. The outer layer of the
exine (?ektexine) which carries the sculpture, may be loosened (brought about by
oxidation) and simulate a closely adhering perine. No elaters seen. The darker area may
result from a thickening of the exine. Usually these spores show folds developed only
around and not across the area (PI. 123, fig. 12).

Preservation and compression. Preservation good. Spores occasionally unfolded but folds
usually developed as mentioned above, probably resulting from the dehydration of the
spore leading to the collapse of the spore wall in the subsequent compression. Spores
very compressed.

Probable affinity. Equisetites lyelli (and ? other species of Cretaceous Equisetites).

Similarity of recent Equisetum spores. Observed acetolysed spores of 11 of some 25
living species of Equisetum (E. hiemale , E.fluviatile , E. pratense, E. moorei, E. sylvaticum ,
E. arvense, E. te/mateia, E. variegatum, E. scirpoides, E. palustre , E. ramosissimum)
lack a tri-radiate mark, and frequently lack elaters which separate from the spores
during acetolysis. They are apparently smooth or scabrate spherical grains mainly
30-50 p in diameter, surrounded by a sometimes scabrate folded perine. Specimens of
some species show a tendency to split. Histograms of the diameters of spores of an
individual species are unimodal, with a small standard deviation. Hauke (1963) states
that in any one strobilus the size range of the spores is small, but strobili of different
specimens of the same species may vary considerably perhaps representing differences
in the age of the cones from which they come. Abortive spores also increase the size
range obtained from an individual strobilus. Acetolysed specimens are frequently
folded. Often only one major fold present (prep. TR002, Equisetum telmateia).

Distinction. Local: Couper (1958) erected the genus Perinopo/lenites for monoporate
perinaceous grains, the presence of a pore being a diagnostic character although not

640

PALAEONTOLOGY, VOLUME 11

necessarily seen on all grains. The presence of the pore and a consistently firmly attached
perine distinguishes this genus and its contained species from Pilasporites allenii.
Colamospora mesozoica Couper 1958, a spore that is rarely definitely recognized in
Mesozoic sediments, is distinct in possessing, by definition, a small tri-radiate mark
similar to the dispersed spores of Palaeozoic Colamospora. SpheripoJlenites scabratus
Couper 1958, differs from P. allenii in having a poorly defined pore, an internally
scabrate exine, and in being smaller. ExesipoIIenites tumulus Balme 1957 possesses a
circular depression surrounded by an area of exinal thickening.

2 1 MEAN I 1 2SD

text-fig. 3. Frequency distribution histogram of maximum diameters in microns of Pilasporites
allenii sp. nov. Preparation T 210, 200 specimens, mean 36-7 p, standard deviation 5-36 p, coefficient

of variation 14 6%.

Literature: Satisfactory comparisons are frequently difficult or impossible due to the
inadequate descriptions and illustrations of the types. Laevigatasporites maximus
Delcourt and Sprumont 1955, and the Tertiary species L. intrapunctatus Kedves 1961,
differ from P. allenii in being much larger. L. reissingeri , a Tertiary species which Kedves
(1961) refers to the ‘Equisetaceae cf. Equisetum' , is larger and has a thicker ‘exospore’.
I. dubius Pflug and Thomson 1953 has a thinner exine than P. allenii ; /. limbatus Balme
1957, is larger, has a thinner exine, and possesses a dark peripheral band; Arctu-
cariacites australis Cookson 1947 is larger and has a different sculpture; Pilasporites
calculus (Balme and Hennelly 1956, p. 64) has a differentially thickened exine; Pila-
sporites phtrigens Balme and Hennelly 1956 is smaller; Aulisporites (Leschik) Klaus
I960, possesses an occasionally observable, faintly developed tri-radiate mark.

BATTEN: PROBABLE DISPERSED SPORES OF CRETACEOUS EQUISETITES 641

Comparison. Local: Similar spores from other samples or localities are attributed as
suggested by Hughes and Moody-Stuart (19676, p. 353).

cfA. P. allenii sp. nov.

1. Cuckfield No 1 Borehole, (TQ 296 273), depth 792 ft. 3 in., Wadhurst Clay,
general size of coarse fraction of sediment 50 p, Equisetites sp. soil-bed, Equisetites
fragments; sample CUC 792/3, prep. T 244/4, preservation good, Pilasporites dominant.
Maximum diameter 28 (35-5) 45 p (200 specimens).

2. Cuckfield No 1 Borehole, depth 866 ft., Wadhurst Clay, general size of coarse
fraction 8 p, Equisetites sp. fragment parting; sample CUC 866, prep. T 269/2, preserva-
tion good, Pilasporites dominant. Maximum diameter 28 (36-8) 47 p (50 specimens).

3. Sharpthorne, Sussex (TQ 375 330), (lower) Wadhurst Clay, general size of coarse
fraction 50 p, Equisetites sp. fragment parting; sample DB 330, prep. T 189/1, preserva-
tion moderately good, some pyrite and other corrosion, Pilasporites dominant. Maximum
diameter 27 (35-8) 46 p (75 specimens).

Literature: Reissinger (1950), Couper (1953), and Rogalska (1954) described and/or
illustrated spores from Jurassic sediments which resemble P. allenii , and which they
referred to the Equisetaceae ( lEquisetum ). P. allenii is also similar to Pilasporites
marcidus Balme 1957, an Australian Cretaceous species, and to Equisetosporites hallei ,
the species that Nilsson (1958) erected for dispersed spores which he thought were
similar if not identical to those produced by Equisetites ( Equisetostacliys ) suecicus
(Nath.) Halle 1908. Some of the specimens of supposed Perinopollenites elatoides grains
which Danze-Corsin and Laveine illustrate (in Briche et al. 1963, pi. 8, figs. 4-6) lack
a pore and resemble P. allenii. Gould (1968) has described some remarkably similar
spores from a cone of Equisetum laterale Phillips 1829 recovered from the Marburg
Sandstone (Jurassic) of Australia.

Acknowledgements. This paper was prepared in the Department of Geology, Cambridge University
with the help of a N.E.R.C. studentship. Material from the Cuckfield No. 1 Borehole of the Geological
Survey and access to the log was made available to Mr. N. F. Hughes by Dr. R. G. Thurrell. I thank
Mr. J. G. Duckett of the Cambridge University Botany School for information. I am indebted to Mr.
N. F. Hughes for discussion and critical reading of the manuscript.

REFERENCES

allen, p. 1941. A Wealden soil bed with Equisetites lyelli (Mantell). Proc. Geol. Ass., London, 52,
362-74.

• 1947. Notes on Wealden fossil soil-beds. Ibid. 57, 303-14.

1959. The Wealden environment; Anglo-Paris basin, Phil. Trans. R. Soc. London, 242B, 283-346.

1962. The Hastings beds deltas: recent progress and Easter field meeting report. Proc. Geol. Ass.,

London, 73, 219-43.

1967a. Origin of the Hastings Facies in North-western Europe. Ibid. 78, 27-105.

19676. Strand-line pebbles in the mid-Hastings beds and the geology of the London Uplands.

Old Red Sandstone, New Red Sandstone and other pebbles. Conclusion. Ibid. 241-76, 3 pi.
Anderson, f. w., bazley, r. A. b., and shephard-thorn, e. r. 1967. The sedimentary and faunal
sequence of the Wadhurst Clay (Wealden) in boreholes at Wadhurst Park, Sussex. Bull. geol. Surv.
Gt. Br. London, 27, 171-235.

balme, b. e. 1957. Spores and pollen grains from the Mesozoic of Western Australia. Rep. Fuel Res.
Sect. C.S.I.R.O. Aust., Chatswood, N.S.W., 1-48, 11 pi.

642

PALAEONTOLOGY, VOLUME 11

balme, b. e. and hennelly, j. p. f. 1956. Monolete, monocolpate and alete sporomorphs from
Australian Permian sediments. Aust. J. Bot. Melbourne, 4, 54-67, 3 pi.

Baxter, r. w. and leisman, g. a. 1967. A Pennsylvanian Calamitean cone with Elat elites triferens
spores. Am. J. Bot. Lancaster, Pa. 54, 748-54, 2 pi.
briche, p., danze-corsin, p., and laveine, j.-p. 1963. Flore infraliasique du Boulonnais. Mem. Soc.
Geol. N. Lille, 13, 7-143, 11 pi.

chandler, m. e. j. 1964. The Lower Tertiary floras of Southern England. 4. A summary and survey of
findings in the light of recent botanical observations. Brit. Mus. (Nat. hist.) London, xii+ 151 pp. 4 pi.
couper, r. a. 1953. Upper Mesozoic and Cainozoic spores and pollen grains from New Zealand.
Palaeont. Bull. Wellington, 22, 3-77, 9 pi.

1958. British Mesozoic microspores and pollen grains. Palaeontographica, Stuttgart, B103,

75-129, 17 pi.

dettmann, M. e. 1963. Upper Mesozoic microspores from south eastern Australia. Proc. R. Soc. Viet.
Melbourne, 77, 1-148, 27 pi.

could, r. e. 1968. Morphology of Equisetum laterale Phillips 1829, and E. bryanii sp. nov. from
the Mesozoic of south-eastern Queensland. Aust. J. Bot. Melbourne, 16, 153-76, 3 pi.
halle, t. g. 1908. Zur Kenntnis der mesozoischen Equisetales Schwedens. K. svenska Vetensk. Akad.
Hand!. Uppsala, 43, (1), 3-56, 9 pi.

Harris, t. M. 1961. The Yorkshire Jurassic Flora. I. Thallophyta-Pteridophyta. Brit. Mus. (Nat. hist.)
London, 212 pp.

hartung, w. 1933. Die Sporenverhaltnisse der Calamariaceen. Arb. Inst. Palaobot. u. Petr. Brenn-
steine, Preuss. geol. Landesanstalt, 3, 96-149.

hauke, r. 1963. A taxonomic monograph of the genus Equisetum subgenus Hippochaete. Nova Hed-
wigia, Weinheim, 8, 1-123, 22 pi.

hughes, n. f. and moody-stuart, j. c. 1967a. Palynological facies and correlation in the English
Wealden. Rev. Paiaeobotan. Palynol. Amsterdam, 1, 259-68.

- 19676. Proposed method of recording pre-Quarternary palynological data. Ibid. 3, 347-58.
kedves, m. 1961. Etudes palynologiques dans le bassin de Dorog. — II. Pollen Spores, Paris, 101-53,
10 pi.

nilsson, t. 1958. Uber das Vorkommen eines mesozoischen Sapropelgesteins in Schonen. Acta Univ.
land. Lund. N.F. Avd. 2, (54-10), 3-111, 8 pi.

potonie, R. 1966. Synopsis der Gattungen der Sporae dispersae, Teil IV. Beih. geol. Jb. Hannover, 72,
1-244, 15 pi.

reissinger, a. 1950. Die ‘Pollenanalyse’ ausgedehnt auf alle Sedimentgesteine der geologischen Ver-
gangenheit II. Palaeontographica, Stuttgart, B90, 99-126, 9 pi.
rogalska, m. 1954. Spore and pollen analysis of the Liassic coal of Blanowice in Upper Silesia. Biul.

Inst. Geol. Warszawa, 89, 5-46, 12 pi. (in Polish, English summary).

Stubblefield, c. J. 1963. Mem. geol. Surv. Summ. Prog. 1962, London, 91 pp.

1967. Rep Inst. geol. Sci. 1966, London, 197 pp.

wilson, l. r. 1943. Elater-bearing spores from the Pennsylvanian strata of Iowa. Am. Midi. Nat.
Notre Dame, 30, 518-23.

1963. Elaterites triferens from a Kansas coal ball. Micropaleontology, New York, 9, 101-2.

DAVID J. BATTEN
Department of Geology
Sedgwick Museum

Typescript received 9 November 1967 Cambridge

THE PALAEONTOLOGICAL ASSOCIATION

COUNCIL 1968-9

President

Professor Alwyn Williams, The Queen’s University, Belfast
Vice-President

Dr. W. S. McKerrow, University Museum, Oxford
Treasurer

Dr. C. Downie, Department of Geology, The University, Mappin Street, Sheffield, 1

Membership Treasurer

Dr. A. J. Lloyd, Department of Geology, University College, Gower Street, London, W.C. 1

Secretary

Dr. J. M. Hancock, Department of Geology, King’s College, London, W.C. 2

Assistant Secretary

Dr. W. D. I. Rolfe, Hunterian Museum, The University, Glasgow, W. 2

Editors

Mr. N. F. Hughes, Sedgwick Museum, Cambridge
Dr. Gwyn Thomas, Department of Geology, Imperial College, London, S.W.7
Dr. I. Strachan, Department of Geology, The University, Birmingham, 15
Professor M. R. House, The University, Kingston upon Hull, Yorkshire
Dr. R. Goldring, Department of Geology, The University, Reading

Other members of Council

Dr. F. M. Broadhurst, The University, Manchester
Mr. M. A. Calver, Geological Institute of Sciences, Leeds
Dr. C. B. Cox, King’s College, London
Mr. D. Curry, Eastbury Grange, Northwood, Middlesex
Dr. Grace Dunlop, Bedford College, London
Dr. G. F. Elliott, 60 Fitzjohn Avenue, Barnet, Herts.

Dr. A. Hallam, University Museum, Oxford
Dr. Julia Hubbard, King’s College, London
Dr. J. D. Hudson, The University, Leicester
Dr. R. P. S. Jefferies, British Museum (Natural History), London
Dr. J. D. Lawson, The University, Glasgow
Dr. A. H. Smout, British Petroleum Company, Sunbury-on-Thames
Professor II. B. Whittington, Sedgwick Museum, Cambridge

Overseas Representatives

Australia : Professor Dorothy Hill, Department of Geology, University of Queensland, Brisbane
Canada: Dr. D. J. McLaren, Institute of Sedimentary and Petroleum Geology, 3303-33rd Street
NW., Calgary, Alberta

India: Professor M. R. Sahni, 98 The Mall, Lucknow (U.P.), India

New Zealand: Dr. C. A. Fleming, New Zealand Geological Survey, P.O. Box 368, Lower Hutt

West Indies and Central America: Mr. John B. Saunders, Geological Laboratory, Texaco Trinidad,
Inc., Pointe-^-Pierre, Trinidad, West Indies

Western U.S.A. : Professor J. Wyatt Durham, Department of Paleontology, University of California,
Berkeley 4, Calif.

Eastern U.S.A.: Professor J. W. Wells, Department of Geology, Cornell University, Ithaca, New
York

PALAEONTOLOGY

VOLUME 11 • PART 4

«r

CONTENTS

Three new Tethyan Dasycladaceae (Calcareous Algae). By g. f. elliott 491

Colour markings in Phacops and Greenops from the Devonian of New York.

By g. c. esker hi 498

Studies on Triassic fossil plants from Argentina. IV. The leaf genus Dicroidium
and its possible relation to Rhexoxylon stems. By S. archangelsky 500

Astrocystites distans, sp. nov., an edrioblastoid from the Ordovician of Eastern
Australia. By B. D. webby 513

Chubbina, a new Cretaceous alveolinid genus from Jamaica and Mexico. By
E. ROBINSON 526

Famennian ammonoids from New South Wales. By t. ,b. h. jenkins 535

Aquilapollenites in the British Isles. By a. r. h. martin 549

Discohelix (Archaeogastropoda, Euomphalacea) as an index fossil in the
Tethyan Jurassic. By J. wendt 554

Pedicellariae of two Silurian echinoids from western England. By D. b. blake 576

Macrocystella Callaway, the earliest glyptocystitid cystoid. By c. R. c. PAUL 580

A revision of some Upper Devonian foraminifera from Western Australia.

By j. e. conkin and Barbara m. conkin " 601

A new medusoid (?) from the Silurian of England. By isles strachan 610

The atrypidine brachiopod Dayia navicula (J. de C. Sowerby). By E. v. tucker 612

Planicardinia, a new septate dalmanellid brachiopod from the Lower Devonian
of New South Wales. By n. m. savage • 627

Probable dispersed spores of Cretaceous Equisetites. By D. J. batten 633

PRINTED IN GREAT BRITAIN AT THE UNIVERSITY PRESS. OXFORD
BY VIVIAN R1DLER, PRINTER TO THE UNIVERSITY

VOLUME 11 • PART 5

Palaeontology

DECEMBER 1968

PUBLISHED BY THE
PALAEONTOLOGICAL ASSOCIATION
LONDON

Price £3

THE PALAEONTOLOGICAL ASSOCIATION

The Association was founded in 1957 to further the study of palaeontology. It holds
meetings and demonstrations, and publishes the quarterly journal Palaeontology.
Membership is open to individuals, institutions, libraries, etc., on payment of the
appropriate annual subscription :

Institute membership . . . . . £7. Os. (U.S. $20.00)

Ordinary membership £5. Os. (U.S. $13.00)

Student membership . . . . . £3. 0s. (U.S. $8.00)

There is no admission fee. Student members are persons receiving full-time instruction
at educational institutions recognized by the Council ; on first applying for member-
ship, they should obtain an application form from the Secretary or the Treasurer. All
subscriptions are due each January, and should be sent to the Membership Treasurer,
Dr. A. J. Lloyd, Department of Geology, University College, Gower Street,
London, W.C. 1, England.

Palaeontology is devoted to the publication of papers (preferably illustrated) on all
aspects of palaeontology and stratigraphical palaeontology. Four parts at least are
published each year and are sent free to all members of the Association. Members
who join for 1969 will receive Volume 12, Parts 1 to 4.

All back numbers are still in print and may be ordered from B. H. Blackwell,
Broad Street, Oxford, England, at £3 per part (post free). A complete set, Volumes
1-11, consists of 43 parts and costs £129.

Special Papers in Palaeontology is a series of substantial separate works published
by the Association. The subscription rate is £6 (U.S. $16.00) for Institute Members
and £3 (U.S. $8.00) for Ordinary and Student Members. Subscriptions and orders
by members of the Association should be placed through the Membership Treasurer.

The following Special Papers are available. Members may obtain them at reduced
rates through the Membership Treasurer. Non-members may obtain them from
B. H. Blackwell, Broad Street, Oxford, England, at the prices indicated.

Special Paper Number One (for 1967): Miospores in the Coal Seams of the
Carboniferous of Great Britain, by A. H. V. Smith and M. A. Butterworth.
324 pp., 72 text-figs., 27 collotype plates. Price £8 (U.S. $22.00), post free.

Special Paper Number Two (for 1968): Evolution of the Shell Structure of
Articulate Brachiopods, by Alwyn Williams. 55 pp., 27 text-figs., 24 collotype
plates. Price £5 (U.S. $13.00).

Special Paper Number Three (for 1968): Upper Maestrichtian Radiolaria of
California, by Helen P. Foreman. 82 pp., 8 collotype plates. Price £3 (U.S.
$8.00).

Special Paper Number Four (for 1969): Lower Turonian Ammonites from Israel,
by R. Freund and M. Raab. 83 pp., 1 5 text-figs., 10 plates. Price £4 (U.S. $11.00).
Available January 1969.

Typescripts on all aspects of palaeontology and stratigraphical palaeontology are
invited. They should conform in style to those already published in this journal, and
should be sent to Mr. N. F. Hughes, Department of Geology, Sedgwick Museum,
Downing Street, Cambridge, England, who will supply detailed instructions for
authors on request (these are published in Palaeontology, 10, p. 707-12).

© The Palaeontological Association, 1968

THE STRUCTURE OF VERTEBRARIA
INDICA ROYLE

by DIVYA DARSHAN PANT and R. SHANKER SINGH

Abstract. Pulls of carbonised substance of Royle’s type of V. indica and 250 other axes referable to the same
species have been studied. The other specimens were collected from the Raniganj and Giridih coalfields (Per-
mian; Upper Damuda Group, Gondwana System). They all show the same kinds of secondary xylem, large
parenchyma, and phloem-like cells. The different kinds of pitting and their frequency in the tracheids, the fre-
quency of uniseriate rays of different heights, their length, and the range of variation in xylem character, have
been determined. It has been found that characters of xylem described for V. raniganjensis are within the range
of V. indica and this species may be regarded as a synonym of V. indica. None of our several axes of V. indica
shows any trace of pith.

Royle (1833) founded two species of Vertebraria from India ( V. indica and V. radiata),
that were later recognized as lateral and sectional views of the same fossil. By a strange
coincidence McCoy (1847) and Dana (1849) made the same error in giving two names
( V. australis and Clasteria australis) to laterally exposed and sectional views respectively
of the same fossil. Arber (1905) and others have pointed out that all axes of this kind
belong to a single species, Vertebraria indica, but Surange and Maheshwari (1962) have
recognized two new species of the genus among compressions and Schopf (1965) has
established a fourth species for a petrified axis, described earlier by Kraiisel. However,
a review of the literature on Vertebraria gave us the impression that the characters of
the various species of the genus were neither sufficiently known nor clearly demarcated.

MATERIAL AND METHODS

A large number of specimens of Vertebraria were collected from localities in the coal-
fields of Raniganj and Giridih. About 300 celloidin pulls of coaly material were pre-
pared from 250 compressions. The pulls were mounted in Canada balsam and examined
in ordinary transmitted light under a compound microscope and also under phase
contrast and dark field illumination. Averages have been generally calculated from 300
random counts but where this was not possible, the smaller number of counts has been
mentioned against the averages.

All figured specimens and slides of this paper form part of Divya Darshan Pant
Collection of plant fossils, located in the Botany Department of the Allahabad
University.

DESCRIPTION OF EXTERNAL FEATURES

Compressions of Vertebraria axes are either found lying almost vertical to the bedding
plane and seen in sectional views or they lie horizontally in the layers of the Lower
Gondwana rocks and are exposed in a lateral view (PI. 124, PI. 125, figs. 1-2; text-fig.
1a-c). Sectional views of Vertebraria axes show rays of carbon, all joined and radiating
from a centre and separated from each other by rather wide bays of rock matrix. The

[Palaeontology, Vol. 11, Part 5, 1968, pp. 643-53, pis. 124-7.1

C 6055 U u

644

PALAEONTOLOGY, VOLUME 11

rays are simple and almost equal, their ends may be slightly flattened, occasionally the
surface of a bay may be covered by a film of carbon which joins two adjacent rays
(PI. 125, fig. 1). The largest vertically preserved Vertebraria in our collection is 2-7 cm.
in diameter. Lateral views of horizontally compressed Vertebraria axes are more com-
mon. The thickest specimen in our collection is 9 cm. wide. The axes show 2 to 4 series
of rectangular areas with 1 to 3 intervening longitudinal furrows or ridges. As a rule
the rectangular areas are of almost equal size but sometimes they may appear very
unequal (text-fig. 1b). Branching axes are not uncommon. As described by Oldham (1897)
the branching of these axes takes place in two ways. Either the branching axis divides
into two or more almost equal branches or the axes produce thinner branches from their
sides (text-fig. 1a-c). The branches may show similar rectangular areas but some of the
thinner axes have a central longitudinal groove or ridge and only occasional obscure
horizontal marks (PI. 124, fig. 5). Some of these axes show attached roots (PI. 124,
fig. 3). An attached root is also seen in a vertically compressed axis (PI. 125, fig. 1).

As suggested by Pant (1956) the preservation of horizontally compressed Vertebraria
is not fundamentally different from that of vertically compressed axes, except that their
xylem rays radiating dorsally and ventrally become shorter or narrower in the radial
direction by undergoing greater radial compression, while the radial dimension of their
horizontally placed rays remains practically uncompressed. Naturally such axes are
usually exposed along the plane of easier splitting, i.e., along the wider horizontally
placed rays and these appear as two series of typical rectangular areas on either side
of a median ridge or furrow representing the central longitudinal core of Vertebraria
and the transverse sides of the rectangles represent the upper and lower limits of the
wide parenchymatous bays. Occasionally the rock is fractured along one of the dorsally
or ventrally placed vertically compressed rays (Pant 1956, text-fig. 2a, where towards the
top left a few narrower rectangular areas are those of a vertically compressed ray).
Additional longitudinal rows of rectangular areas may be seen at a point where a
Vertebraria branches (text-fig. 1b; PI. 124, fig. 2). Many of our axes lie among leaves
of Glossopteris, Rhabdotaenia, Pteronilssonia, and other specimens. In one of these
a Glossopteris leaf at first sight appeared to be connected with a Vertebraria but a closer
examination reveals that the apical end of the leaf was actually facing the axis.

INTERNAL STRUCTURE
Horizontally compressed axes

Celloidin pulls of carbon from horizontally compressed axes and their branches with
or without typical rectangular areas usually show the same kind of secondary xylem as

EXPLANATION OF PLATE 124

Figs. 1-5. Vertebraria indica Royle. 1, Forked axis, No. 1804, xf. 2, Trifurcate axis showing wide
rectangular areas in two series below and triseriate areas above; rectangular areas are not seen for
some distance near the lower end; No. 1756, xf. 3, Axis with elongated rectangular areas and a
few thin branches (roots) with scalariform xylem (cf. Lithorhiza tenuirama Pant), No. 1805, xf.
4, Fragment of a thick axis showing tetraseriate rectangular areas, No. 1808, X 5, Axis with a
median furrow but its rectangular areas are ill defined due to the presence of only a few rather obscure
transverse marks. No. 3052, xf.

Palaeontology, Vol. 11

PLATE 124

PANT and SINGH, Permian Vertebraria

PANT AND SINGH: STRUCTURE OF VERTEBRARIA INDICA ROYLE

645

s'

text-fig. 1, a-c. Vertebraria indica. A, Axis forked into two almost equally thick branches. No. 1804,
X2. b, Trifurcate axis showing short and wide rectangular areas above; rectangular areas not seen
near the lower end (also PI. 124, fig. 2), No. 1756, X c, Axis showing thinner branch, No. 1789b, X 2.

646

PALAEONTOLOGY, VOLUME 11

is seen in a radial section. Naturally, in compressed wood, such pulls also show the
radial or tangential walls of the tracheids often overlapping each other. At a few places,
the xylem is even seen in a tangential view or in a partially radial and partially tangential
view (text-fig. 2h; PI. 127, fig. 6).

Radial views of xylem. (a) Royle's type. A pull from Royle’s type specimen No. V. 4189
in the British Museum (Natural History) shows the typical xylem of Vertebraria as if in
a radial longitudinal section (PI. 125, fig. 3). The tracheids have uniseriate to tri-seriate
oval to circular pits placed far apart or contiguous. Biseriate and triseriate pits are
usually opposite but alternately arranged pits have also been observed (text-fig. 3d, e).
Rays are uniseriate and ray fields show pits usually without any border (text-fig. 3f).
A few large parenchyma cells like those reported by Pant (1956) are mostly broken
down.

( b ) Specimens from Raniganj and Giridih coalfields. All our pulls of the carbon from
horizontally preserved specimens of Vertebraria axes show well-preserved secondary
xylem covering the entire surface of the rectangular areas (where carbon is preserved),
from one side of the axis through the median ridge or furrow right up to the opposite
side. On all our slides most of the xylem is seen as in a radial longitudinal section. A pith
should have been visible in such radial views but no pull from the axes in our collection
shows any trace of parenchyma cells in the centre, although ray parenchyma, phloem-
like thin- walled cells and the typical large parenchyma cells of Vertebraria are usually
well preserved. Instead, numerous pulls show undisturbed tracheids filling the central
parts of the axes.

Primary xylem. Primary xylem tracheids consisting of scalariform, annular, or spiral elements were
described from pulls of Vertebraria by Walton and Wilson (1932) and Pant (1956). Pant described
them as occurring towards the periphery of the axes, where they could even have belonged to associated
or attached roots.

Secondary xylem. The wood is mostly secondary and pycnoxylic. The tracheids are almost all of a
uniform type without growth rings. The length of most tracheids is indeterminable in the pulls and this
may suggest that they were usually longer than the size of our pulls. However, 60 complete tracheids
have been observed; their length is 77-679 /a, (average 250 p , 8 105). As both ends of most tracheids are
not seen, we presume that these sizes may represent the shorter elements in the xylem of Vertebraria.
The width of the tracheids ranges from 10-56 p (average width of tracheids is 31 p, 8 20 5). The trachei-
dal walls are up to 4 p thick. The ends of the tracheids taper gradually to a narrow point, the end walls
being very oblique (PI. 126, fig. 1). The tracheids over the rectangular areas on the surface of the
axes are straight and vertical but in the region of the horizontal ridges or furrows they are slightly
curved.

Arrangement of pits. The pits on radial walls are uniseriate, or multiseriate up to 5 series of pits
(text-fig. 2 b, c, and g). A single tracheid with 6 seriate pits was also seen (text-fig. 2 e, PI. 126, fig. 2).

EXPLANATION OF PLATE 125

Figs. 1-6, Vertebraria indica Royle. 1, Vertically compressed axis towards the top right-hand corner;
part of transverse diaphragm (black) and an attached root also seen; No. 1375, X 11. 2, Vertically
compressed axis showing marks of peripheral tissues, No. 1810, x2. 3, Pull from Royle’s holotype
showing biseriate and triseriate pits and a two-cell high ray, V. 4189, x300. 4, Xylem of root
attached to axis in Plate 124, fig. 3, showing scalariform tracheids, Slide No. 1805, X 350. 5, Portions
of tracheidal walls of axis showing crossed pit pores, Slide No. 1809 Ba, X 500. 6, Tracheids show-
ing uniseriate pits or mixed uniseriate and biseriate opposite pits, Slide No. 1808c, X 800.

Palaeontology, Vol. 11

PLATE 125

PANT and SINGH, Permian Vertebraria

PANT AND SINGH: STRUCTURE OF VERTEBRARIA INDICA ROYLE

647

text-fig. 2, a-h. Vertebraria indica. A, Portion of tracheid showing scalariform pitting (also PI. 127,
fig. 4); Slide No. 586, X 1,000. b, Part of tracheid with uniseriate non-contiguous pits; Slide No. 1784
Da, X750. c, Portion of tracheid showing uniseriate and biseriate pits; Slide No. 1784 Da, X750.
d. Irregularly-pitted tracheid showing oval pit-pores, some with crossed pit-pores; Slide No. 1807a,
X 1,000. e, Portion of wide tracheid showing 6-seriate pits (also PI. 126, fig. 2); Slide No. 1807a, X 750.
F, Portion of tracheid showing oval groups of pits with a central pit; Slide No. 1765 C2, X750. g,
Tracheid showing tetraseriate and pentaseriate pits; Slide No. 1767c, X750. h, Uniseriate 3-cell-high
ray in tangential view; Slide No. 1784 Da, x450.

648

PALAEONTOLOGY, VOLUME 11

The pits may be contiguous or far apart (text-fig. 2 b; PI. 125, fig. 5) and the portions of the same
radial wall may be pitted in some regions and unpitted for varying distances in other regions. Multi-
seriate pits may be opposite, araucarioid (i.e. with crowded alternate pits having hexagonal outlines)
or irregularly arranged (PI. 126, fig. 3). As a rule unpitted portions of the wall are seen in tracheids
with irregular pits. Some of these may show circular or oval groups of 3 to 9 pits (text-fig. 2 d, f). The
same tracheid may show different kinds of pitting in different regions i.e., uniseriate or multiseriate,
opposite or alternate, contiguous or non-contiguous (text-fig. 2 c; PI. 125, fig. 6; PI. 126, fig. 3).
A random count of 300 tracheids showed about 11% with uniseriate pits, 47% opposite multiseriate
pits, 12% alternate crowded multiseriate pits of araucarioid type, 28% irregular pits, and 2% showed
portions of walls without pits. Out of the 47% tracheids with opposite pits 19% are biseriate, 18%
triseriate, 8% tetraseriate, and 2% pentaseriate.

Shape and size of pits. The typical shape of bordered pits may be circular or oval (text-fig. 2 b). Oval
pits in multiseriate tracheids are placed horizontally but where uniseriate they may be horizontal or
oblique. The pit pores are generally oval but occasionally rounded. The two opposite pores of a pit
pair are usually crossed (text-fig. 2d; PI. 125, fig. 5). In the tracheids of the thinner axes pit pores are
wide with a very thin border.

In some regions of the pulls, the tracheids show large horizontally extended oval pits where a border
is not discernible. Such pits may be uniseriate or multiseriate. Where uniseriate they almost look like
scalariform thickenings (text-fig. 2a; Plate 125, fig. 4; Plate 127, fig. 4) and where multiseriate they often
appear transitional between typical scalariform and pitted elements. Pant (1956) described such
tracheids as in contact with large parenchyma cells.

The circular pits are 4-8-5 p (average 5 p, 8 0-5). The length of the oval pits is 8-5-26-5 p (average
length is 13 p, 8 2-5) and the breadth is 4-10 p (average breadth is 6 p, 8 1). The length of the oval
pits whose borders are not clear is 10-31 p (average length is 21-5 p, 8 1-5) and the breadth is 4-10 p
(average breadth is 7 p, 8 1).

Secondary xyleni rays. Only uniseriate secondary xylem rays have been observed. These are 1-13 cells
high (PI. 126, fig. 5; PI. 127, fig. 1), but more commonly they are only 1-4 cells high:

Percentage of uniseriate rays of differing heights
Height of rays in number of cells 1 2 3 45 6 7 13

Percentage 33 29 17 10 6 2 2 1

Frequency of rays per 10 mm.2 is 4-18 averaging 9 (60 readings, S 4). The rays are 1-48 cells long
(average 1 1 cells, 8 4).

Crossfield pits. Crossfields of V. indica where clearly visible show 1-3 elongated oval pits (text-fig. 3a;
PI. 127, fig. 2). The oval pits are horizontally placed and seemingly simple. On the basis of our observa-
tions, we believe that the numerous bordered crossfield pits described by Walton and Wilson (1932) and
by Surange and Maheshwari (1962) are pits of tracheidal wall, other than the common wall between
a ray cell and a tracheid. In a compression, this wall may overlap a crossfield and then the pits can be
mistaken for crossfield pits. The length of the crossfield pits ranges from 11-32 p (average 18 p, 8 2)
and breadth from 4 to 14 p (average 5 p, 8 1).

Phloem like cells. Many pulls show thin elongated cells by the side of xylem which look like phloem
(text-fig. 3 c; PI. 127, fig. 3). The length of these cells is from 17 to 238 p (average 61 p, 8 17) and
breadth 4-35 p (average 10 p, 8 3). The thickness of the walls of these cells is up to 3 p.

EXPLANATION OF PLATE 126

Figs. 1-5. Vertebraria indica Royle. 1, Xylem showing tapering ends of two tracheids and longitudinal
walls of others with multiseriate pits, Slide No. 1807a, X 800. 2, Portion of wide tracheid (on the
right side) showing six-seriate pits, Slide No. 1 807a, X 800. 3, Xylem showing triseriate pits, opposite
in some regions but alternate and araucaroid in other parts, Slide No. 1754c, X 350. 4, Pull from
thinner axis with obscure rectangular areas showing typical Vertebraria type of xylem, Slide No. 1784
Db, X 140. 5, Xylem showing a 7-cell-high ray. Slide No. 1807a, X 200.

Palaeontology, Vol. II

PLATE 126

PANT and SINGH, Permian Vertebraria

PANT AND SINGH: STRUCTURE OF VERTEBRARIA INDICA ROYLE

649

text-fig. 3, a-f. Vertebraria indica. a, Portion of xylem showing simple, horizontally elongated, oval
pits in cross-fields; Slide No. 1809 Ba, x 600. b. Pull showing xylem with simple pits overlapped by
broad parenchyma cells, which tend to be in radial series; Slide No. 586, x 150. c, Pull showing thin-
walled phloem like cells; Slide No. 1767a, X 150.D, Portion oftracheid from holotype showing triseriate
opposite pits; V. 4189, X750. e, Pull from holotype showing tracheid with uniseriate, biseriate, and
irregularly arranged pits; V. 4189, X 750. F, Xylem showing horizontally elongated, simple oval pits

in crossfield; Holotype, V. 4189, X750.

650

PALAEONTOLOGY, VOLUME 11

Large parenchyma cells. In addition, sheets of large rectangular to polygonal parenchyma cells were
seen overlying xylem almost in all preparations. Very often the cells tend to arrange in radial rows
(text-fig. 3 b; PI. 127, fig. 5). The length of these cells is 21-228 p (average 79 p, 8 49) and breadth
21-84 p (average 34 p, 8 17).

Tangential views of xylem. In portions of the pulls where the xylem appears in a partially or wholly
tangential view, the xylem rays are seen and they are invariably uniseriate (text-fig. 2h). Where the
rays are obliquely compressed, presenting a combination of tangential and radial views, the ray field
pits appear to be simple and horizontally oval. The tangential walls of the tracheids are devoid of pits
and their surface presents a uniformly granular texture (PI. 127, fig. 6). The xylem of axes with ob-
scure areas is essentially similar, showing identical pitting, uniseriate rays, crossfield pits, phloem-like
cells and large parenchyma cells.

Vertically preserved axes

Xylem and other cells from vertically preserved axes are also indistinguishable from
those of horizontally compressed Vertebraria except for the presence of short tracheids
with tapering or truncated ends over the transverse side of the rectangular areas.

COMPARISON AND DISCUSSION

Surange and Maheshwari (1962) described a new species V. raniganjensis, with xylem
differing specifically from that of V. indica as described by Walton and Wilson (1932),
Pant (1956), and Sen (1958). Surange and Maheshwari (1962) assumed that Royle’s
type specimen of V. indica, No. V. 4189 in British Museum (Natural History) was
without carbon. However, since one of us (D. D. P.) had examined Royle’s type of
V. indica and seen carbon in it, we obtained through the courtesy of Dr. Pettitt a small
pull from Royle’s type. Therefrom, we have been able to confirm that it shows exactly
the same kind of xylem as has been described by Walton and Wilson (1932), Pant
(1956), Sen (1958), and by us here from other specimens assignable to V. indica. This
xylem is also similar to that of Surange and Maheshwari’s V. raniganjensis.

Surange and Maheshwari described the following peculiarities of the xylem of V.
raniganjensis :

(a) The pits are ‘often’ six seriate.

( b ) The pits are ‘more than often’ arranged irregularly and in groups of 2 to 9 which
are sometimes of stellate shape.

(e) ‘Xylem rays if not common are also not scarce as in V. indica’ . They are usually
1-5 cells high, rarely up to 8 cells high.

(d) Crossfield pits are 7-9 in number, bordered and with an oval pore and some
fields show no pits.

(e) Absence of large parenchyma cells.

They did not, however, mention the frequency of tracheids with six seriate or irregu-
larly arranged pits or of the xylem rays. The frequent occurrence of irregularly pitted

EXPLANATION OF PLATE 127

Figs. 1-6. Vertebraria indica Royle. 1, Xylem showing 13-cell-high ray, Slide No. 1777b, x150. 2,
Xylem showing horizontally elongated, simple oval pits in crossfields, Slide No. 1809 Ba, x800.
3, Pull showing thin-walled phloem-like cells, Slide No. 1767a, X 300. 4, Portion of tracheid showing
scalariform pits, Slide No. 586, X 350. 5, Pull showing broad parenchyma cells, tending to be in
radial series, Slide No. 1781c, x40. 6, Portion of xylem showing unpitted tangential walls, Slide
No. 1789c, x 230.

Palaeontology, Vol. 11

PLATE 127

PANT and SINGH, Permian Vertebraria

PANT AND SINGH: STRUCTURE OF VERTEBRARIA INDICA ROYLE 651

elements of xylem and the occurrence of six seriate tracheids (even though rare) in
axes whose xylem is otherwise exactly like that of V. indica suggests that such tracheids
may not be used to distinguish another species. The frequency of rays in V. indica is
rather variable. Calculated on the basis of figures and photographs given by Walton
and Wilson (1932), Pant (1956), and Surange and Maheshwari (1962), the frequency of
rays in the axes described by these authors is within the range determined for V. indica.
As mentioned above we believe that there is an error in the description of crossfield pits
by Walton and Wilson and by Surange and Maheshwari. The only remaining difference
is the absence of large parenchyma cells in V. raniganjensis and their presence in V.
indica ; however, the negative evidence of the absence of such cells in a pull may be mis-
leading for even in two pulls of the same axis, such cells are often seen to be preserved
in some regions and not in other portions. We have also re-examined Pant’s original
pulls of V. indica from Mhukuru coalfield, Tanganyika, and they show the same kind
of the xylem as the type. Accordingly we are convinced that axes described as V. rani-
ganjensis and V. indica belong to the same ‘species’.

Surange and Maheshwari (1962) described another new species, V. myelonis , which
differed from V. indica in having a pith in the centre of the axes but no parenchymatous
tissue was recorded in the ‘pith’ region. None of our pulls from the Raniganj and Giridih
coalfields shows any trace of a parenchymatous pith although the xylem and parenchyma
cells are usually well preserved. Surange and Maheshwari based their species on a single
specimen which does show a narrow space (pith) in the centre between the two lateral
series of rectangular areas, but they did not describe any parenchyma from this region
and the space possibly represents the width of the solid central part of the xylem because
we have actually observed solid xylem in this region in numerous specimens and also
in the centre of vertically preserved axes. In the absence of any positive evidence of a
pith it would be simpler, if this species too is regarded as a synonym of V. indica.

MORPHOLOGY OF VERTEBRARIA

Structural features

Oldham believed that Vertebraria had a solid central core joined to a peripheral
bark by a series of radiating longitudinal septa which could be branched (Oldham 1 897,
pi. 4, fig. 2), and according to his description eight of these septa could usually be seen
in a sectional view in vertically preserved axes. He thought that the longitudinal septa
were joined at intervals by transverse partitions so that the axes of Vertebraria were
visualized as having a ring of air chambers around a central solid core. In his view the
transverse partitions are responsible for the septate appearance in horizontal preserva-
tion. Zeiller (1896) thought that Vertebraria axes had an externally ridged form with
transverse septa between the longitudinal ridges as in the rhizomes of the living fern
Strutbiopteris.

Both Oldham and Zeiller agreed in regarding the rectangular areas as representing
empty spaces (internal or external). Walton and Wilson (1932), however, suggested that
the axes of Vertebraria had a rayed stele practically without a pith; each carbonaceous
ray in a vertically preserved Vertebraria was formed by compression of compact
secondary xylem having uniseriate parenchymatous rays between its tracheids. These
rays of xylem were joined with adjacent xylem rays on either side; rock-filled spaces

652

PALAEONTOLOGY, VOLUME 11

(bays) between the carbonaceous rays were not empty but originally occupied by a
soft tissue. Pant (1956) agreed with Walton and Wilson on the basis of his study of a
Stigmaria with rays similarly preserved, and he described large parenchyma cells which
occupied the bays.

An interpretation of the features of Vertebraria which is similar to that given earlier
by Oldham has recently been given by Plumstead (1962), but she agrees with Walton
and Wilson (1932) in believing that the chambers were filled with a soft tissue. However,
her reconstruction visualizes a solid ‘central column communicating by means of large
oval cavities with the rest of the axis’, and also a woody bark. Recently, Schopf (1965)
has generally confirmed the reconstructions of the structural features of Vertebraria
suggested by Walton and Wilson (1932).

Our investigation of the extensive new material of Vertebraria is fully consistent with
the views of Walton and Wilson (1932). Pulls from rays of carbon in vertically preserved
axes show the same kind of xylem as the surface of the rectangular areas in horizontally
compressed axes. Celloidin pulls of carbon from rectangular areas of horizontally
preserved axes show tracheids which tend to be short, wide, and curved near the trans-
verse portions but elsewhere the tracheids run uniformly in the longitudinal direction.
We have also confirmed that the transverse connections joining adjacent carbonaceous
rays are also formed by xylem. As already described by Pant (1956) xylem and phloem
over rectangular areas is often overlapped by large parenchyma cells, which are clearly
the remnants of a soft tissue which originally filled the bays. The absence of rectangular
areas in thinner branches of Vertebraria leads us to the conclusion that the parenchy-
matous bays may have been formed after secondary growth as in the axes of some
modern lianas like Aristolochia. This may also explain the regular seriated arrangement
of the large parenchyma cells.

Stem or root

Vertebraria has been generally recognized as the stem or rhizome of Glossopteris after
Zeiller (1896) and Oldham (1897) described axes showing apparent attachment of
Glossopteris leaves, and an axis described by Dolianiti (1954) seemingly confirms this.
On the contrary Glossopteris bearing axes of Glossopteris described by Etheridge (1894),
Seward (1910), and Walton and Wilson (1932), do not show any Vertebraria features.
An explanation for this inconsistency has been offered by Arber (1905) and Pant (1956,
1962) who suggest that the features of Vertebraria are only those of its vascular cylinder
and accordingly axes where the peripheral tissues are intact would appear different.
Walton and Wilson (1932), and Thomas (1952) have, however, doubted the attachment
of Glossopteris leaves on Vertebraria axes.

On the basis of his study of some roots called Lithorhiza tenuirama , Pant (1958) sug-
gested by implication that Vertebraria could be a root, and pointed out that the roots of
L. tenuirama could sometimes be found attached to axes of Vertebraria and sometimes
to axes which combined the features of roots as well as those of a Vertebraria. Such
axes are somewhat thicker than undoubted roots but thinner than axes showing typical
Vertebraria features. Roots arise from them endogenously. Unlike roots they show large
parenchyma cells characteristic of Vertebraria but they entirely lack its rectangular
areas. Lately, Schopf (1965) has strongly supported the root nature of Vertebraria on
the basis of a petrified axis. In this connection he emphasizes the lack of leaf traces or

PANT AND SINGH: STRUCTURE OF VERTEBRAR1A INDICA ROYLE

653

nodes by Vertebraria axes and also the endogenous insertion of its roots (but adventi-
tious roots in stems are also endogenous). However, Vertebraria axes are sometimes
dichotomised or trichotomised like stems or adventitiously branched roots, although
Pant (1958) has also described some similarly forked roots. Pant (1967) has recently
described a number of axes lacking rectangular areas of Vertebraria type but showing
attached leaves of Glossopteris and this may also serve as a negative evidence favouring
the root character. An investigation of the structural features of Glossopteris bearing
stems which appear to differ from Vertebraria axes may help in solving the problem.

Acknowledgements. We are grateful to Dr. J. M. Pettitt of the British Museum (Natural History),
London, for sending us a pull from Royle’s type specimen and to Drs. D. D. Nautiyal, G. K. Srivastava,
B. K. Varma, and Mr. K. B. Singh for help in the collection of Vertebraria axes and Dr. (Miss) P. F.
Kidwai for help in taking some enlargements. We thank the State Council of Scientific and Industrial
Research, Uttar Pradesh, for financial assistance.

REFERENCES

arber, E. A. N. 1905. Catalogue of the fossil plants of the Glossopteris flora in the Department of Geology,
British Museum ( Natural History). London.

DANA, J. D. 1849. Wilkes' United States exploring expedition, during the years 1838-1842 under the
Command of Charles Wilkes, U.S.N. 10, Geology, New York.
dolianiti, e. 1954. A Flora do gondwana inferior em Santa Catarina, IV O genero Vertebraria. Div.
Geol. Min. Notas Prel. E. Estudos, 81, 1-5.

etheridge, r. jun. 1894. On the mode of attachment of the leaves or fronds to the caudex in Glossop-
teris. Proc. Linn. Soc. N.S.W., Ser 2, 9, 228-58.

mccoy, F. 1847. On the fossil botany and zoology of the rocks associated with the coal of Australia.

Ann. Mag. Nat. History, 20 (132), art 15, 145-57; 20 (134), art 28, 298-312.
oldham, r. d. 1897. On a plant of Glossopteris with part of the rhizome attached, and on the structure
of Vertebraria. Rec. Geol. Surv. India, 30 (1), 45-50.
pant, d. d. 1956. On the two compressed Palaeozoic axes: Stigmaria ficoides in Gymnostrobus con-
dition and Vertebraria indica. Ann. Bot. N.s. 20 (79), 419-29.

1958. Structure of some roots and spores from the Lower Gondwana (Permo-Carboniferous) of

East Africa. Vijnana Parishad Anushandhan Patrika, 1 (4), 231-44.

1962. Some recent contributions towards our knowledge of the Glossopteris flora. Proceedings of

the Summer School of Botany, Darjeeling, 302-10.

1967 (In press). On stem and attachment of Glossopteris leaves. Phytomorphology, 17.

plumstead, E. p. 1962. Fossil flora of Antarctica. Trans-Antarctic Expedition {London), Rept. 9,
Geology, pt 2, 154 p.

royle, J. F. 1833. Illustrations of the botany and other branches of natural history of the Himalayan
mountains, and of the flora of Cashmere. London .

schopf, j. m. 1965. Anatomy of the axis in Vertebraria. Geol. and Paleont. of the Antarctic. Antarctic
Research Series, 6, 217-28.

sen, j. 1958. Further studies on the structure of Vertebraria. Bot. Not. 3, (2), 436-48.
seward, a. c. 1910. Fossil plants, vol. 2. Cambridge.

surange, k. r. and maheshwari, h. k. 1962. Studies in the Glossopteris flora of India-1 1 . Some observa-
tions on Vertebraria from the Lower Gondwanas of India. Palaeobotanist 9, (1, 2), 61-7.
thomas, h. H. 1952. A Glossopteris with whorled leaves. Ibid. 1, 435-8.

walton, j. and wilson, j. a. r. 1932. On the structure of Vertebraria. Proc. Roy. Soc. Edinburgh,
52, pt. 2, 200-7.

zeiller, r. 1896. Etude sur quelques plantes fossiles, en particulier Vertebraria et Glossopteris, des
environs de Johannesburg (Transvaal). Bull. Soc. Geol. France, Paris, ser. 3, 24, 349-78.

DIVYA DARSHAN PANT
R. SHANKER SINGH

Department of Botany,

The University, Allahabad, India

Typescript received 11 January 1968

THE GRAPTOLITE ASSEMBLAGES AND
ZONES OF THE BIRKHILL SHALES
(LOWER SILURIAN) AT DOBB’S LINN

by P. TOGHILL

Abstract. The graptolite assemblages and lithologies of the Birkhill Shales (Lower Silurian) at Dobb’s Linn
are described. The condensed 141 ft. (43 m.) euxinic sequence is divided into eight graptolite zones in descending
order : Rastrites maximus, Monograptus sedgwickii, M. convolutus, M. gregarius, M. cyphus, Cystograptus vesi-
culosus, Akidograptus acuminatus, and Glyptograptus persculptus. These zones are compared with the original
work of Lapworth (1878), and with equivalent zones elsewhere in Great Britain. The base of the Middle Llan-
dovery at the type locality for the graptolitic Llandovery (Valentian) is taken at the base of the gregarius Zone
( triangulatus Zone in Wales), and not at the base of the magnus Zone, as is the current practice in Wales.

The classic locality of Dobb’s Linn, and the surrounding Moffat area, have been left
relatively untouched by research workers since the admirable account of Lapworth
(1878).

Because of the lack of new information from the Moffat area, recent reviews of
Southern Uplands geology (Walton 1963, 1965) have had to discuss the Llandovery
geology of the area using information obtained in, and unaltered since, the late 1800s,
and thus their discussions are rather speculative. The present work does not, however,
suggest any major errors by Lapworth. It is only the application of knowledge obtained
since 1878 which has allowed alterations and additions to his classic results. The founda-
tion for a Llandovery graptolite zonal scheme was laid down by Lapworth (1878, 1880),
and his original research was quickly followed up in the Lake District by Marr and
Nicholson (1888). Table 2 is a correlation of the large number of zonal schemes which
have been used in Great Britain for the Llandovery (Valentian) up to the base of tur-
riculatus Zone. The most recent scheme given by Curtis (1961) is in fact a compilation
of most zones recognized in Wales, and based on the original research of O. T. Jones
(1909), and Jones and Pugh (1916, 1935). The present work reappraises the graptolite
fauna of the condensed 141 ft. (43 m.) euxinic Birkhill Shale sequence at the type locality,
redefines the Llandovery graptolite zones exposed at Dobb’s Linn, and compares them
with equivalent zones elsewhere. The condensed sequence at Dobb’s Linn is almost
completely fossiliferous and thus is the most suitable place to define Llandovery grap-
tolite zones, not only because it is the type locality for the graptolitic Llandovery
(Valentian) but also because the equivalent sequence in Wales is a thick series (1,000 ft.)
of coarse sediments in which the graptolite horizons are separated by considerable
thicknesses of unfossiliferous strata (Jones 1909, p. 504).

POSITION OF SECTIONS

Dobb’s Linn lies 10 miles north-east of Moffat, and at the head of the Moffat Water.
A complicated isoclinal fold structure, with Caledonoid trend, brings the graptolitic con-
densed sequence of Moffat Shales (Caradoc-Llandovery) to the surface from below

[Palaeontology, Vol. 11, Part 5, 1968, pp. 654-68.]

TOGHILL: GRAPTOLITE ASSEMBLAGES AND ZONES OF BIRKHILL SHALES 655

the overlying Gala Greywackes (Upper Llandovery). This work is only concerned with
the highest (Birkhill) division of the Moffat Shales, from the base of the Silurian
( persculptus Zone) up into the maximus Zone of the Upper Llandovery.

The south-east limb of the fold is highly contorted and no detailed measurements are
possible. The north-west (inverted) limb however shows an excellent section of the Birk-
hill Shales, particularly in the Linn Branch, and on the North and Main Cliffs (text-
fig. 1). A large waterfall has developed in the Linn Branch at the junction of Gala
Greywackes and Birkhill Shales, and below this all but the lowest beds of the Birkhill
Shales are exposed, the beds being inverted and dipping downstream at between 40° and
70°. The basal beds are cut out by a strike fault, and there is also a considerable amount
of strike faulting in the convolutus, gregarius , and upper cyphus Zones. The Corrie and
North Cliff show a continuous section from the lower sedgwickii Zone down to the base
of the Silurian with little or no strike faulting. The Main Cliff (text-fig. 1) provides an
excellent section across the Ordovician-Silurian boundary, and is by far the best place
for measuring up the basal beds of the Birkhill Shales. Using these sections a detailed
lithological sequence (text-fig. 2) has been worked out for the Birkhill Shales. Descending
from the Gala Greywackes, the maximus and sedgwickii Zones have been measured up
in the Upper Linn Branch (section A, text-fig. 1). Conspicuous lithological horizons in
the lower sedgwickii Zone then allowed the section to be transferred to the Corrie and
North Cliff (section B, text-fig. 1) where the remainder of the sequence was measured
down to the base of the persculptus Zone. The lower part of the North Cliff' section
( persculptus to cyphus Zones) was correlated with, and supplemented by, collections
from the Lower Linn Branch section (C) in the cyphus, vesiculosus and upper acuminatus
Zones, adjacent to the West Fault, and the Main Cliff section (D) in the persculptus ,
acuminatus and vesiculosus Zones.

THE LITHOLOGIES OF THE BIRKHILL SHALES

Three main lithological types are present: massive black pyritic graptolitic mudstone,
massive barren grey mudstone, and beds of soft, pale grey to orange, clay-like deposits.
Lapworth (1878) referred to this latter conspicuous lithology under a variety of names
including, ‘white clay bands’, ‘soft shaly clay’, and ‘white lines’, but the term claystone
has recently been applied to these deposits by Walton (1965, p. 185). These beds, which
often contain lines of calcareous nodules, weather out very easily and may be of volcanic
origin (Walton 1963, pp. 85, 89; 1965, p. 185). This lithology makes up a large propor-
tion of the Birkhill Shales and also occurs in the underlying Hartfell Shales, and persists
into the Gala Greywackes. If volcanic in origin it indicates a period of vulcanicity from
the Caradoc into the Upper Llandovery. The lower 65 ft. of the succession is made up
of alternations of black mudstone and pale soft claystone, with occasional nodule bands,
but without any grey mudstone at all. These lower Birkhill Shales follow directly on top
of grey mudstones of the Hartfell Shales. The black mudstones are generally between
3 in. and 1 ft. thick and the intervening claystones are usually between 1 in. and 3 in.
thick, but occasionally up to 1 ft. Within the upper acuminatus, vesiculosus, and lower
cyphus Zones is a conspicuous 10-ft. thickness of massive black mudstone with very few
claystones. These massive beds, Lapworth’s vesiculosus Flags, are a conspicuous unit
of the lower Birkhill Shales well exposed in the Lower Linn Branch. The upper 76 ft. of

656

PALAEONTOLOGY, VOLUME 11

text-fig. 1. Locality map of Dobb’s Linn, showing position of sections.

TOGHILL: GRAPTOLITE ASSEMBLAGES AND ZONES OF BIRKHILL SHALES 657

o o o

2E55-I

Block mudstone

Grey mudstone
with cloystone

Claystone/with
calcareous nodules

0 12 3 4

Vertical scale in metres

text-fig. 2. Vertical section through the Birkhill Shales, showing the position of fossiliferous horizons,
etc. Although an attempt has been made to indicate the frequency of claystones, the section is diagram-
matic, particularly below 55 ft. The positions of all nodular claystones, and of individual fossiliferous

beds above 55 ft. are precise.

the Birkhill Shales is made up of alternations of grey and black mudstones, varying in
thickness from mere bedding plane bands to beds of 1 ft., and many soft claystones with
occasional nodule bands. The inter-relationships of grey and black mudstones are shown
in text-fig. 2. The highest graptolitic horizons in the Birkhill Shales ( maximus Zone)
occur as thin bands | in. to 1 in. thick in a dominantly grey mudstone sequence. About
12 ft. above the highest black band the first greywacke appears in the Upper Linn Branch,

658

PALAEONTOLOGY, VOLUME 11

below the waterfall. It is 6 in. thick and is taken as the base of the Gala Greywackes, as
indicated by Lapworth (1878, p. 322). This bed is followed by 14 ft. of massive grey
mudstones, thin greywackes, and claystones, before the massive greywackes at the base
of the waterfall are reached.

From the above description, and from text-fig. 2, it can be seen that within the
Birkhill Shales there is a gradual change from the exclusively euxinic facies of the lower
Birkhill Shales, through the grey and black mudstones of the convolutus and sedgwickii
Zones, into the dominantly grey mudstones of the rnaximus Zone, and finally into grey-
wacke facies. Throughout the Moffat area the base of the Gala Greywackes has been
found to be highly diachronous, and this will be further explained when describing the
fauna of the R. maximus Zone.

THE CHARACTERISTIC GRAPTOLITE FAUNA OF THE
BIRKHILL SHALES

The various graptolite zones have been divided up into fossiliferous horizons (1-46)
which are indicated on text-fig. 2. The detailed fauna of any horizon can be ascertained
from Table 1 .

Zone of Glyptograptus persculptus

This zone is well exposed on the Main CliiT (section D, text-fig. 1), but is less accessible
at the top of the North Cliff (base of section B). On the Main Cliff the black mudstones
of the Birkhill Shales follow barren grey mudstones of the anceps Zone of the Hartfell
Shales. The highest fossiliferous band with D. anceps is approximately 10 ft. below the
base of the Birkhill Shales. The persculptus Zone (horizon 46, text-fig. 2) comprises
3-|- ft. (T06 m.) of black mudstone with thin claystones. The mudstones are often
weathered to a drab grey colour and the fauna is then extensively destroyed. At the base
of the zone is a 4-in. bed crowded exclusively with Climacograptus scalar is normalis.
The remainder of the zone is characterized by this species with the zone fossil, and
Climacograptus medius, C. scalaris miserabilis and Diplograptus modestus sd.

Zone of Akidograptus acuminatus

16| ft. (5-0 m.) of black mudstones with many thin claystones are assigned to this
zone (horizons 43-5). It is best exposed on the Main Cliff (section D), but the highest
beds are well exposed in the Lower Linn Branch (section C). The whole zone is exposed
on the North Cliff (section B) but is rather inaccessible. The lower part of the zone
weathers in a similar fashion to the persculptus Zone, but the highest beds are hard and
massively bedded, and form the lowest part of Lapworth’s vesiculosus Flags.

The base of the zone is marked by the appearance of Akidograptus and the zone is
characterized by A. acuminatus sd. and A. ascensus, neither of which occur outside the
zone. The latter occurs commonly before the main burst of A. acuminatus, which com-
mences in the middle of the zone. Glyptograptus persculptus is common in the lower part
and Climacograptus trifilis in the middle of the zone. Cystograptus vesiculosus appears
in the middle and is common at the top of the zone, and Diplograptus modestus occurs
throughout. Climacograptus scalaris normalis and C. medius are common throughout,
and C. rectangularis appears in the upper part.

TOGHILL: GRAPTOLITE ASSEMBLAGES AND ZONES OF BIRKHILL SHALES 659
Zone of Cystograptus vesiculosus

This zone is represented by 4 ft. 3 in. (1*3 m.) of massive black mudstone with a few
claystones (horizons 41-2), and is best exposed in the Lower Linn Branch (section C).
The base is marked by the most conspicuous horizon in the whole of the Silurian, the
appearance of the Monograptidae, together with Dimorphograptus and Rhaphidograptus.
The zone is characterized by the earliest monograptids, with the zone fossil, and dimor-
phograptids with long uniserial portions: D. elongatus and R. extenuatus. One particular
bedding plane at the base of the zone is crowded with Monograptus cypluts praematurus
Toghill (1968), which with M. atavus Jones is the earliest monograptid. The latter is
common throughout the zone and occurs in the upper part with M. acinaces and M.
incommodus. Climacograptus innotatus appears at the base of the zone and occurs
throughout as does Diplograptus modestus sd. Glyptograptus tamariscus sd. appears and
climacograptids are abundant including C. medius, C. rectangular is, and C. scalaris
normalis.

Zone of Monograptus cyphus

This zone is represented by 24 ft. (7-3 m.) of massively bedded black mudstones with
yellow and orange claystones, which are rare towards the base. The zone (horizons
34-40) has been measured up in the Lower Linn Branch (section C) and on the North
Cliff (section B). The higher parts of the zone in the Lower Linn Branch are extensively
strike-faulted. The overlying gregarius Zone has at its base a most conspicuous 1 ft.
claystone with large calcareous nodules, and this, the lowest nodular claystone, is well
exposed on the North Cliff.

At the base of the zone dimorphograptids with short uniserial portions appear,
including: D. confertus, D. confertus swanstoni, D. erectus, D. decussatus sd., and D.
longissimus. Of these the first three are the commonest, and the first two persist into
the upper parts of the zone. D. physophora is rare but probably occurs throughout as
does Rhaphidograptus toernquisti. Monograptus cyphus is rare at the base of the zone but
soon becomes abundant and ranges throughout the zone as do M. sandersoni, M. incom-
modus, M. atavus, and M. acinaces. M. gregarius and M. revolutus appear in the middle
of the zone and are common at the top. Cystograptus vesiculosus and Climacograptus
innotatus are limited to the basal beds where the latter is common. Climacograptus
medius and C. rectangularis are abundant in the lower part of the zone, but do not reach
the top, and neither does C. scalaris normalis. Glyptograptus tamariscus sd. occurs
throughout the zone.

Zone of Monograptus gregarius

26 ft. (7-9 m.) of strata are assigned to this zone (horizons 29-33) as measured on the
North Cliff (section B). Black mudstones with thick nodular claystones occupy the
lower 18 ft. but these grade upwards into banded grey and black mudstones with
many claystones.

The base of the zone is marked by the appearance of Monograptus triangulatus sd.
which occurs throughout the zone and is particularly common in the lower part.
M. gregarius and M. revolutus occur throughout the zone, but M. sandersoni, M. incom-
modus, and M. atavus are limited to the lower part. M. communis is characteristic of the
upper part of the zone, and Rastrites longispinus and R. peregrinus appear in the middle.

x x

C G055

660

PALAEONTOLOGY, VOLUME 11

A conspicuous horizon (30) 8 ft. from the top of the zone yields abundant specimens of
Monograptus fimbriatus with many petalograptids, including Petalograptus minor , P.
palmeus latus, P. palmeus ovato-elongatus, as well as a few specimens of Diplograptus
magnus. This is the only horizon with D. magnus and M . fimbriatus, and the main burst
of the latter occurs above that of M. triangulalus s.l., as suggested by Sudbury (1958).
This horizon is presumably equivalent to the magnus Band in Wales, although here it is
only 3 in. thick. The unfossiliferous beds (text-fig. 2) which overlie horizon 30 but
underlie the highest fossiliferous horizon (29) of the gregarius Zone might be considered
part of the magnus Zone. This 3-ft. thickness of beds compares with 27 ft. assigned to
the magnus Zone in Wales, although there the actual magnus Band at the base is only
6 in. thick (Jones and Pugh, 1935, p. 275), the remainder of the zone being relatively
unfossiliferous.

In horizon 29 Monograptus leptotheca is common, M. denticulatus appears, and this
is the only horizon with M. argenteus. It can be equated with the argenteus Zone of the
Lake District (Marr and Nicholson 1888), and the leptotheca Zone of Wales, which are
here considered to be equivalent. M. argenteus in abundance seems to be restricted to the
Lake District, its place in Wales and Scotland being taken by M. leptotheca.

Zone of Monograptus convolutus

The foot of the North Cliff (section B) provides an excellent section of this zone
(horizons 22-8) which is 17J ft. (5-35 m.) thick and made up of two groups of highly
fossiliferous black mudstones with claystones separated by barren grey mudstones and
claystones (text-fig. 2). The upper group of fossiliferous beds (horizons 22-5) is, in fact,
Lapworth’s original clingani Band, and the lower group his cometa Zone. The detailed
fauna of any horizon can be ascertained from Table 1.

Faunally the base of the zone is marked by the appearance of Monograptus lobiferus
in abundance, with M. convolutus. The former is limited to the basal beds of the zone
but the latter persists, with M. clingani , throughout the zone and is found as high as
the lower maximus Zone. M. triangulatus, M. communis, and M. leptotheca are all present
in the basal beds, but M. gregarius has disappeared. Cephalograptids, and petalograptids
with protracted proximal ends are characteristic including in the lower part, Petalo-
graptus folium and Cephalograptus tubulariformis, and C. cometa in the upper part of the
zone. Rastritids are common and include R. hybridus, R. peregrinus, R. longispinus,
and R. approximatus geinitzi. In the higher group of black mudstones the following
monograptids are common: M. clingani, M. limatulus, M. crenularis, M. convolutus,
and M. decipiens, together with Cephalograptus cometa, Glyptograptus tamariscus
incertus, and Orthograptus bel/ulus. Of these species M. crenularis is very typical of the
upper convolutus Zone.

Zone of Monograptus sedgwickii

27J ft. (8-4 m.) of strata are assigned to this zone (horizons 8-25) which is best exposed
in the Upper Linn Branch (section A), and the Corrie under the North Cliff (section B).
Lithologically the zone can be divided into three units. A lower 7 ft. of barren grey
mudstones with claystones, best seen in the Corrie; a middle 6 ft. of fossiliferous black
mudstones with claystones (horizons 14—21) exposed in the Corrie and Upper Linn
Branch; and an upper 14J ft. of grey mudstones with claystones and a few fossiliferous

TOGHILL: GRAPTOLITE ASSEMBLAGES AND ZONES OF BIRKHILL SHALES 661

Q

O

A.ac.

u

M. cyphus

M.gregarius

M. convolutus

M. sedgwick ii

R.max imus

Fossiliferous horizons

5

-

"t

o

■s

ft

ft

5

s

ft

5

o

ft

s

in

ft

a

5

£

ft

5

o

ft

2

£

2

2

2

2

2

-

-

-

-

-

-

-

o

r

r

•

o c

r

r

r

o

9

o

•

r

©

o

r

o

o

r

o

•

o

O

,

©

o

o

o

o

O

,

O

r

vor. normal is Lapw.

o

o

o

o

o

o

miserabilis E.and W.

o

o

o

Cystograptus vesiculosus (Nich.)

o

•

9

o

o

o

9

Orthograptus bellulus ( Tornq.)

cyperoides (Tbrnq.)

r

o

o

o

O

o

O

r

insectiformis ( Nich.)

r

mutabilis E.and W.

o

o

o

truncatus var abbreviatus E.and W.

Glyptograptus persculptus (Salter)

sinuatus ( Nich J
tamariscus (Nich.)

var. incertus E.andW.
serratus E.and W.

var. barbatus E.andW.

Diplograptus modestus L a p w.

var. pa r v ul 0 s ( H . La pw.)

Petalograptus palmeus (Barr.)

var. tenuis (Barr.)
latus (Barr.)
ovato- elongatus (Kurck)

minor Elies
fol ium ( H i s.)

Cepha lograptus cometa ( Gei n .)

tubulariformis (Nich.)

Akidograptus acuminatus s.l.(Nich.)

ascensus Davies

Rhaphidograptus toernquisti (E.andW.)

extenuatus (E.andW.)

Dimorphograptus confertus (Nich.)

var swanstoni Lapw.
erectus E.and W.
decussatus E.and W.

var partiliter E.andW.
elongatus Lapw.
longissimus ( Kurck)
physophora (Nich.)

Rastrites maximus Carr.

linnaei Barr.

distans Lapw.

fugax Barr.

hybridus Lapw.

peregrinus Barr.

longispinus Perner

approximatus var. geinifzi Tornq.

Monograptus cyphus Lapw.

var. praematurus Toghill
gregarius Lapw.
acinaces TiJrnq.
leptotheca Lapw.
regularis TcJrnq.
jaculum Lapw.
nudus Lapw.
concinnus Lapw.
revolutus Kurck
difformis Tdrnq.
argenteus ( Nich.)
limatulus Tttrnq.
atavus Jones
sandersoni Lapw.
incommodus Tornq.
argutus Lapw.
crenularis Lapw.
gemmatus ( Barr.)
attenuatus (Hopk.)
sedgwickii ( Port 1 . )
ha 1 1 i (Barr.)
lobi ferus ( M'C oy )
clingani (Carr.)
millipeda (M'Coy)
convolutus ( H is.)
decipiens Tornq.
triangulatus s.l. (Hark.)
denticulatus Tbrnq.
spiralis (Gein.)
involutus Lapw.
circularis E.and W.
communis Lapw.
fimbriatus (Nich.)

. i nter m ed ius L apw.

Diploqraptus maqnus H Lapw.

®

.

•

?

0

9

?

9

?

o

o

o

o

r

o

o

o

o

o

9

o

o

O

r

r

r

r

o

9

•

o

r

o

o

o

o

o

o

o

o

o

r

r

r

f

r

r

r

r

r

o

O

r

o

o

o

•

•

r

o

r

9

r

o

o

o

•

o

o

o

o

•

o

•

o

I

r

r

1

o

o

©

r

•

•

•

v

•

o

o

o

o

o

o

r

r

r

r

r

>

v

,

o

O

9

9

r

9

9

r

O

o

o

r

o

o

o

o

o

o

r

r

o

r

r

•

o

9

©

©

9

•

9

9 r

?

o

o

o

o

?

©

o

o

?

r

r

o

9

©

o

o

«

9

9

r

•

9

o

o

r

r

r

o

9

9

o

o

9

9

9

o

r

O

?

r

o

O

r

?

O

7

©

o

o

o

o

o

o

i

,

o

o

r

©

9

r

•

©

©

©

o

©

o

r

r

r

©

•

9

o

«

o

o

®

•

o

o

o

o

o

o

•

9

i

o

O

o

o

r

o

o

9

o

9

©

9

9

r

r

r

9

o

e

o

r

r

r

o

o

:

e

o

o

o

r

•

r

©

o

o

o

o

9

o

9

9

O

9

2.

r

r

o

o

9

9

9

O

o

r

r

o

•

•

•

o

o

r

o

o

©

o

O

r

r

r

o

o

o

o

o

?

-

9

©

©

o

o

r

O

•

o

•

o

r

O

r

r

r

r

9

9

o

• common o occurs

r rare ? questionably present

§

3

3

-

o

?;

s

3.

con

CN

-

o

a

IS

o

£

s

;;

ft

O

o-

co

K

o

«n

*

r>

CN

-

o

■>

-

o

-

-

table 1. The vertical distribution of the Graptoloidea in the Birkhill Shales at Dobb’s Linn.

662

PALAEONTOLOGY, VOLUME 11

black seams (horizons 8-13) exposed in the Upper Linn Branch. The middle unit is
characterized by an abundance of Monograptus sedgwickii, M. regularis, and M. jacu-
lum., with triangulate monograptids in the lower part, including M. convolutus, M.
involutus, and M. decipiens. Glyptograptids are common and Petalograptus palmeus
temds occurs with Climacograptus sealaris. The upper 144 ft. contain a similar but poorer
fauna in which M. spiralis and M. attenuatus appear.

Zone of Rastrites maximus

The highest 214 ft. (6-55 m.) of the Birkhill Shales is assigned to this zone. Grey
mudstones and claystones predominate, and grade up into the Gala Greywackes, the
base of which is taken at the first greywacke seen in the Upper Linn Branch (section A),
as indicated by Lapworth (1878, p. 322). Black mudstones are restricted to the lowest
9 ft. Seven separate fossiliferous horizons occur, and of these the two lowest yield a
large fauna (Table 1). In these basal beds Monograptus sedgwickii, M. regularis, M.
nudus, M. convolutus, and M. circularis are common, but the latter two species only occur
commonly on one particular bedding plane. These common species occur with M.
spiralis, M. attenuatus, M. gemmatus, M. decipiens, M. intermedius, Climacograptus
extremus, and Petalograptus palmeus tenuis. At this lower level Rastrites maximus and
M. halli are rare.

The highest fossiliferous horizon of the Birkhill Shales at Dobb’s Linn is a 4-in.
composite band of four black mudstones up to 1 in. thick separated by thin claystones.
It yields Rastrites maximus, R.fugax, R. linnaeil, R. distansl, R. hybridus, Monograptus
halli, M. sedgwickii, M. convolutus, M. spiralis, and M. intermedius.

Owing to the diachronism of the overlying Gala Greywackes (see below), higher
horizons not represented by fossiliferous strata at Dobb’s Linn, but occurring south east
of the Caledonoid strike line through Dobb’s Linn, are characterized by the following
common species: Rastrites maximus, Monograptus halli, and M. spiralis, with M. run-
cinatus, M. nudus, and M. turriculatus minor.

THE LLANDOVERY GR APTOLITE ZONES AT DOBB’S LINN

The graptolite zonal scheme used here has been arrived at by considering the original
scheme of Lapworth (1878, 1880), as modified by Elies and Wood (1913), and the large
number of schemes erected in Wales by Jones (1909), Jones and Pugh (1916), Davies
(1926), and others, as summarized by Curtis (1961). Table 2 shows an attempted cor-
relation between the zones used here, and other schemes used elsewhere.

The Glyptograptus persculptus Zone was defined at Dobb’s Linn for the first time by
Toghill (1968), although the species was first recorded from there by Davies (1929).
The validity of the zone is emphasized by the absence of Akidograptus, which does not
appear until the overlying acuminatus Zone.

The Akidograptus acuminatus Zone as defined here is approximately the same as
defined by Lapworth (1878, pp. 250, 252, 318-20), although the basal beds have been
separated off as the persculptus Zone.

The base of the Cystograptus [Orthograptus] vesiculosus Zone has been taken at
the incoming of the monograptids, as suggested by Elies (1925, p. 344). The re-
stricted thickness (4 ft. 3 in.) used here is only the lower part of the zone as originally

TOGHILL: GRAPTOLITE ASSEMBLAGES AND ZONES OF BIRKHILL SHALES 663

defined by Lapworth (1878, pp. 250, 252, 319-20), but the zone has been strictly defined
on the ranges of the earliest monograptids, as given by Elies (1925), and dimorpho-
graptids. The fact that the zone is so thin compared with the 16 ft. assigned to it by
Lapworth is due to the more precise fixing of monograptid ranges, particularly those of
M. cyphus, M. sander soni, M. incommodus, and M. atavus. Lapworth (1878, p. 319) did
not give a very detailed monograptid fauna from his vesiculosus Zone and lower gre-
garius Zone, recording only M. ‘ tenuis' and M. attenuatus from the vesiculosus Zone, and
of course M. incommodus and M. atavus were not described until later. Elies (1925) did
not give any thicknesses for the vesiculosus Zone, and thus this is the first time that the
zone has been defined at Dobb’s Linn (the type locality) using a comprehensive mono-
graptid fauna. The monograptid ranges agree with Elles’s definition of the zone (1925),
but she recorded Dimorphograptus confertus and D. confertus swanstoni as being charac-
teristic of the vesiculosus Zone. These are here taken as indicating the base of the over-
lying cyphus Zone together with the appearance of M. cyphus and M. sandersoni.

Thus, the vesiculosus Zone as used here is only the lower part of the zone as defined
by Lapworth (1878), Elies and Wood (1913), and Elies (1925). It is probably equivalent
to the atavus Zone in Wales (Jones 1909, Curtis 1961). The cyphus Zone includes the
remainder of Lapworth’s vesiculosus Zone and part of his gregarius Zone. It is equivalent
to the cyphus and acinaces Zones in Wales (Jones, 1909). From Table 1 it can be seen
that the ranges of M. atavus and M. acinaces are almost identical, the former having the
larger range, and the latter occurs, in the later part of its range, with M. cyphus. Although
the vesiculosus Zone as defined here is equivalent to the atavus Zone in Wales, the former
title is preferred on historical grounds, Dobb’s Linn being the type locality.

The M. gregarius Zone as defined here is equivalent to the triangulatus, magnus, and
leptotheca Zones in Wales (Curtis, 1961). The base is taken at a conspicuous faunal
horizon, the appearance of triangulate monograptids, and is probably the base of Lap-
worth’s Upper Division of his gregarius Zone (1878, pp. 319, 321), where he records
M. triangulatus ‘in profusion’. The main horizon (29) with M. leptotheca is only 1 ft.
thick, and that (30) with Diplograptus magnus only 3 in. thick. The former horizon con-
tains M. argenteus and is presumably equivalent to the argenteus Zone in the Lake
District. M. triangulatus s.l. ranges right through the gregarius Zone as defined here as
does the zone fossil, and the title gregarius Zone is preferred on historical grounds,
although it must be noted that it is equivalent to the triangulatus s.l. Zone of W. D. V.
Jones (1945). Jones and Pugh (1935, pp. 275, 276) stated that the magnus Zone was made
up of 27 ft. of mainly unfossiliferous strata with a 6-in. magnus Band at the base.
Similarly the leptotheca Zone was given as 20 ft. of relatively unfossiliferous beds with
a 2|-ft. leptotheca Band at the base. The practice of erecting zones based on thin
graptolite bands separated by barren strata is not considered correct, particularly as one
has no idea of the fauna of the intervening beds. Although the horizons with M.
leptotheca and Diplograptus magnus at Dobb’s Linn provide direct correlation with the
sequence in Central Wales, they should only be regarded as bands and not zones.

Dr. R. B. Rickards records (in lift .) that the lithologies and faunas at the level of the
gregarius Zone in the Howgill Fells (eastern Lake District) are very similar to Dobb’s
Linn. He recognizes the three zones triangulatus, magnus, and argenteus ( leptotheca ), and
reports that there is a marked lithological change from black to grey mudstones at the
level of the magnus Zone, a state of affairs comparable with the sequence at Dobb’s Linn

664

PALAEONTOLOGY, VOLUME 11

table 2. A correlation of Llandovery (Valentian) graptolite zonal schemes, up to the base of the M. turriculatus Zone.

TOGHILL: GRAPTOLITE ASSEMBLAGES AND ZONES OF BIRKHILL SHALES 665

where the first grey mudstones appear immediately above the 3-in. band with Diplo-
graptus magnus.

The M. convolutus Zone as used here is equivalent to the same zone as defined by
Elies and Wood (1913), and Elies (1925), the cometa Band being included in the zone.
In some cases, in Wales, the cometa Band has been treated as a separate zone (Jones,
1909, 1929; Curtis, 1961). The zone is equivalent to the cometa Zone of Lapworth (1878,
pp. 322-3) together with the clingani Band of his sedgwickii Zone.

The base of the M. sedgwickii Zone as defined here is equivalent to the same horizon
in Wales, being taken at the appearance of M. sedgwickii. The clingani Band of Lapworth’s
sedgwickii Zone has been relegated to the convolutus Zone.

The base of the Rastrites maximus Zone has been taken at a similar level to Lapworth
(1878, pp. 322, 325-7), but the two species which according to Lapworth characterize
the zone at Dobb's Linn, R. maximus and M. hal/i, have not been found to be at all
common in the zone, but occur abundantly in localities south-east of Dobb’s Linn (see
below). The zone is considered to be equivalent to the halli Zone in Wales (Jones and
Pugh 1916, W. D. V. Jones 1945). Some authors have suggested in the correlation charts
of review papers (Jones 1921, 1929; Curtis 1961) that a maximus Zone has been defined
in Wales above the halli Zone. In fact this has never been done, although within the
halli Zone around Machynlleth ( Jones and Pugh 1916) there is a separate Rastrites Band
above a halli Band. The only use of a R. maximus Zone in Wales was by Elies (1909)
in the Conway area where it lay between the sedgwickii and crispus Zones.

Localities south-east of the Caledonoid strike-line through Dobb’s Linn (Craigmichan
Scars, Thirlstane Scar, and Beldcraig Burn) expose fossiliferous horizons which are
higher in the maximus Zone than any fossiliferous strata at Dobb’s Linn. Lapworth com-
pared these sections directly with the maximus Zone at Dobb’s Linn (1878, pp. 326-7)
as if they were exactly the same horizon. He was unaware of the diachronous nature
of the Gala Greywackes south-east of Dobb’s Linn which meant that their base had
‘stepped up’ in relation to Dobb’s Linn. He was, however, quite aware that the Gala
Greywackes ‘stepped down’ north-west of Dobb’s Linn. These localities south-east of
Dobb’s Linn yield Rastrites maximus and Monograptus halli in abundance with M.
runcinatus, M. spiralis and rarely M. turriculatus minor, and further substantiate the
suggestion that the halli and maximus Zones are equivalent. The latter name should take
priority on historical grounds. Higher horizons in the Gala Greywackes south-east
of Dobb’s Linn yield a fauna which is typical of the Monograptus turriculatus Zone,
including, M. exiguus, M. barrandei, M. nodifer, and the zone fossil.

THE TERMS LOWER, MIDDLE, AND UPPER LLANDOVERY
APPLIED TO THE BIRKHILL SHALES

The terms Llandovery (Murchison 1859) and Valentian (Lapworth 1876) are now
considered to be synonymous, and the former term has priority. However, the Valentian
is usually considered to be the graptolitic Llandovery, and as the divisions of the Llan-
dovery are based on a shelly fauna and unconformities in the type area (Jones 1925),
it is unlikely that the divisions can be readily applied to the graptolitic sequence.

When the Valentian was further defined by Jones (1921) the base was taken at the
base of the G. persculptus Zone (1921, pp. 173-4), this zone having been originally

666

PALAEONTOLOGY, VOLUME 11

defined in the graptolitic sequence of Central Wales (Jones 1909, p. 482). This level is
now accepted as the base of the Llandovery (Curtis 1961, p. 129). The base of the Middle
Llandovery is usually equated to the base of the magnus Zone in the graptolitic sequence
of Central Wales (Curtis 1961, p. 130), but this is entirely based on lithological considera-
tions (Jones, 1925, p. 366; 1947, p. 3; 1954, p. 252; Jones and Pugh 1935, pp. 274-5).
The only variation from this opinion was by Jones in 1929 when he included the magnus
Zone in the Lower Llandovery (1929, p. 97). The highest graptolites recorded from the
Lower Llandovery at Llandovery are Monograptus incommodus and Climacograptus
hughesi from the top of A4 (Jones 1925, p. 360; 1929, p. 94), and these were considered to
indicate thetop of the acinaces Zone (Jones 1925, p. 360). The Middle Llandovery at Llan-
dovery has yielded Monograptus decipiens, M. cf. lobiferus, and M. cf. regularis. These
were collected from near to the top of the division and considered by Jones (1925, p. 366)
to indicate the convolutus Zone. The abundance of M. sedgwickii at 1 50 ft. above the base
of the Upper Llandovery at Llandovery (Jones 1925, p. 370) suggests that the present
accepted level for the base of the Upper Llandovery in the graptolitic sequence, i.e.
the base of the sedgwickii Zone, is quite acceptable. However, there is no such fixed
level for the base of the Middle Llandovery in the graptolitic sequence which could, on
the evidence stated above, lie as low as the base of the cyphus Zone, or even as high as
the base of the convolutus Zone (Jones 1925, p. 366). This is an unsatisfactory situation,
and the problem cannot really be solved until more graptolite evidence is obtained from
the Llandovery area, or shelly evidence from the graptolitic sequence. Therefore as the
type area cannot offer any faunal evidence which will fix the Lower-Middle Llandovery
boundary with any certainty in the graptolitic sequence, there can be no harm in con-
sidering the graptolite fauna of the Llandovery (Valentian) as a whole, in order to suggest
an equivalent horizon in the graptolitic sequence as a working base for the Middle
Llandovery.

Within the graptolite fauna of the Birkhill Shales are four distinct divisions cor-
responding to those erected by Elies (1922) and further elaborated by Bulman (1958).
The lowest part of the Birkhill Shales below the incoming of monograptids ( persculptus -
acuminatus Zones) is equivalent to the Ortho-Climacograptid sub-fauna of Bulman
(1958, pp. 170-1), the highest division of Elles’s Diplograptid Fauna. The next division
of the Birkhill Shales, above the incoming of monograptids, but below the appearance of
triangulate monograptids, vesiculosus-cyphus Zones, ( atavus-cyphus Zones in Wales)
represents the lowest division, A, of Bulman’s Monograptid Fauna. The next division,
gregarius- convolutus Zones, (triangulatus- convolutus Zones in Wales), represents the
second division, B, of Bulman’s Monograptid Fauna, and the final division, sedgwickii-
maximus Zones represents part of the third division, C, of the Monograptid Fauna
(Bulman 1958, pp. 170-1).

The base of the Llandovery is well fixed in the graptolitic sequence at the base of the
persculptus Zone, which was defined for the first time in Scotland by Toghill (1968).
It is worth noting that Elies (1922, p. 195) suggested that the lowest two zones of the
Llandovery ( persculptus-acuminatus Zones) being devoid of monograptids could well
be relegated to the Ordovician on faunistic grounds, but this has never been accepted.
The base of the Middle Llandovery if taken at the base of the magnus Zone is some way
above the appearance of triangulate monograptids, and some way above the base of the
gregarius Zone as defined here. The appearance of triangulate monograptids at the base

TOGHILL: GRAPTOLITE ASSEMBLAGES AND ZONES OF BIRKHILL SHALES 667

of the gregarius Zone ( triangulatus Zone in Wales) corresponds to the base of the second
division, B, of the Monograptid Fauna. It would be more convenient if the base of the
Middle Llandovery in the graptolitic sequence were to be taken at the base of the
triangulatus Zone in Wales, and thus equate with the base of the gregarius Zone at
Dobb’s Linn. If acceptable the base of the Middle Llandovery would then correspond
to the base of the second division of Bulman’s Monograptid Fauna (Bulman 1958,
pp. 170-1). The base of the Upper Llandovery taken at the base of the sedgwickii Zone
is a convenient horizon as it is the base of the third division of the Monograptid Fauna.
Thus the stage boundaries of the Llandovery in the graptolitic sequence would now
correspond to significant faunal horizons, though within the Lower Llandovery is the
most conspicuous horizon of all, the appearance of the monograptids.

In conclusion, the Lower Llandovery at Dobb’s Linn is represented by the zones of
Glyptograptus persculptus, Akidograptus acuminatus, Cystograptus vesiculosus, and Mono-
graptus cyphus, and is 48 ft. (14-6 m.) thick. The Middle Llandovery includes the zones
of M. gregarius and M. convolutus and is 44 ft. (13-4 m.) thick. Only part of the Upper
Llandovery is present, represented by the zones of M. sedgwickii and Rastrites maximus.
These two zones are 49 ft. (15 m.) thick and pass up into the great thickness of the Gala
Greywackes which (together with the Hawick Rocks?) represent the remainder of the
Upper Llandovery.

Acknowledgements. I am extremely grateful to Dr. I. Strachan for his constant advice throughout the
research, a grant for which was obtained from the Department of Scientific and Industrial Research
(now N.E.R.C.). I would also like to thank Professor O. M. B. Bulman for the loan of supplementary
specimens from the Sedgwick Museum.

REFERENCES

bulman, o. m. b. 1958. The sequence of graptolite faunas. Palaeontology , 1, 159-73.
curtis, m. l. K. 1961. In Lexique Stratigraphique International, 1, Europe, (3aV), Silurian.
davies, K. a. 1926. The geology of the country between Drygarn and Abergwesyn (Breconshire).
Q. Jl geol. Soc. Lond. 82, 436-64.

1929. Notes on the graptolite faunas of the Upper Ordovician and Lower Silurian. Geol. Mag.

66, 1-27.

elles, G. l. 1909. The relations of the Ordovician and Silurian rocks of Conway (North Wales). Q. Jl
geol. Soc. Lond. 65, 169-94.

— — 1922. The graptolite faunas of the British Isles. A study in evolution. Proc. Geol. Ass. Lond. 33,
168-200.

1925. The characteristic assemblages of the graptolite zones of the British Isles. Geol. Mag. 62,

337-47.

and wood, e. m. r. 1901-18. A monograph of British graptolites. Pa/aeontogr. Soc. [Monogr.].

jones, o. t. 1909. The Hartfell-Valentian succession in the district around Plynlimon and Pont Erwyd
(North Cardiganshire). Q. Jl geol. Soc. Lond. 65, 463-537.

1921. The Valentian series. Ibid. 77, 144-74.

1925. The geology of the Llandovery district: Part I — The southern area. Ibid. 81, 344-88.

1929. Silurian. In Handbook of the geology of Great Britain, j. w. evans and c. J. Stubblefield

(eds.), London, 88-127.

1933. The lower Palaeozoic rocks of Britain. Int. geol. Congr. 16th session, Washington, 1,

463-84.

1947. The Llandoverian graptolite succession in Britain. Ball. Mus. r. Hist. nat. Belg. 23, (22), 1-4.

1954. The use of graptolites in geological mapping. Lpool. Manchr. geol. J. 1, 246-60.

and pugh, w. j. 1916. The geology of the district around Machynlleth and the Llyfnant Valley.

Q. Jl geol. Soc. Lond. 71, 343-85.

668

PALAEONTOLOGY, VOLUME 11

jones, o. T. and pugh, w. j. 1935. The geology of the districts around Machynlleth and Aberystwyth.
Proc. Geol. Ass. Lond. 46, 247-300.

jones, w. d. v. 1945. The Valentian succession around Llanidloes, Montgomeryshire. Q. Jl geol. Soc.
Lond. 100, 309-32.

and rickards, r. b. 1967. Diplograptus penna Hopkinson 1869, and its bearing on vesicular

structures. Palaont. Z. 41, 173-85.

lapworth, C. 1876. The Silurian system in the south of Scotland. In Catalogue of the Western Scottish
fossils, j. Armstrong et al. (ed.), Glasgow, 1-28.

1878. The Moffat Series. Q. Jl geol. Soc. Lond. 34, 240-346.

1880. On the Geological Distribution of the Rhabdophora. Ann. Mag. nat. Hist. (5), 5, 45-62,

358-69; 6, 16-29, 185-207.

lapworth, h. 1900. The Silurian sequence of Rhayader. Q. Jl geol. Soc. Lond. 56, 67-137.

marr, j. e. and nicholson, h. a. 1888. The Stockdale Shales. Ibid. 44, 654-732.

murchison, r. i. 1859. Siluria. 2nd ed. London.

sudbury, m. 1958. Triangulate Monograptids from the Monograptns gregarius Zone (Lower Llan-
dovery) of the Rheidol Gorge (Cardiganshire). Phil. Trans. R. Soc. B. 241, No. 685, 485-555.

toghill, p. 1968. The stratigraphical relationships of the earliest Monograptidae, and the Dimor-
phograptidae. Geol. Mag. 105, 46-51.

walton, e. k. 1963. Sedimentation and structure in the Southern Uplands. In The British Caledonides,
m. r. w. Johnson and f. h. stewart (eds.), Edinburgh, 71-97.

1965. Lower Palaeozoic rocks— stratigraphy, palaeogeography and structure. In The geology of

Scotland, g. y. craig (ed.), Edinburgh, 161-227.

P. TOGHILL

Department of Palaeontology,
British Museum (Nat. Hist),
Cromwell Road,

Typescript received 5 February 1968 London S.W. 7

NUMMULITES (FOR AMINIFE R A) FROM
THE UPPER EOCENE KOPILI FORMATION
OF ASSAM, INDIA

by B I M A L K. SAMANTA

Abstract. Four species of Nummulites are described and illustrated front the Kopili Formation, Garo Hills,
Assam, India. This is the first account of the genus Nummulites from the Upper Eocene Pellatispira- bearing
horizon in the Indian region.

Outcrops of marine Upper Eocene rocks with larger foraminifera are known to occur
in three areas in the India-Pakistan region (Samanta 1968, fig. 1): Surat-Broach in
Western India (Rao 1941), the Sulaiman Range in West Pakistan (Eames 1952) and
Assam in Eastern India (Nagappa 1951, Samanta 1965). Nummulites has been reported
to occur in association with the typical Upper Eocene genus Pellatispira in all three
areas but so far there is no published account of the genus from this horizon.

In Assam the Kopili Formation contains a rich Upper Eocene larger foraminiferal
assemblage including such stratigraphically important genera as Asterocyclina, Dis-
coeyclina, Nummulites , and Pellatispira. An investigation of the larger foraminifera of
the Kopili Formation in the Garo Hills has been carried out by the writer and an account
of the genus Nummulites is given in the present paper.

KOPILI FORMATION

Evans (1932, pp. 173—5) first called this unit the Kopili alternations ‘Stage’ and some-
times Kopili ‘Stage’. Later workers have changed the name to Kopili Formation, since
by original designation it is basically a rock unit. In the type section (Kopili River section
of the Kopili-Khorungma region) the succession is reported to be about 450 m. thick
and consists of alternations of sandstone, mudstone, shales, carbonaceous rocks, and
shell-bearing sandstone. It conformably overlies the Sylhet Limestone and is apparently
conformably overlain by the Barail group of rocks. The formation outcrops along the
southern fringe of the Shillong Plateau, from the Garo Hills in the west to the Mikir
Hills in the east.

In the Garo Hills the Kopili Formation is best exposed in the Simsang River section
between Siju Artheka (90° 41' E., 25° 20' N.) and Matmagitik (90° 40' E., 25° 18' N.).
It conformably overlies the Siju Limestone and is apparently conformably overlain by
Barail-equivalent rocks (Samanta 1968, p. 128, table 1). The lower part of the formation
is richly fossiliferous and contains abundant larger foraminifera, including such strati-
graphically important genera as Asterocyclina , Discocyclina, Nummulites , and Pellati-
spira. Of these, Discocyclina is the most abundant. Because of their much larger size in
comparison to other larger foraminifera, discocyclines constitute the most conspicuous
element of the foraminiferal fauna. Nummulites is represented by small to medium sized
striate and reticulate forms and occurs in almost all foraminiferal samples. In contrast

[Palaeontology, Vol. 11, Part 5, 1968, pp. 669-82, pis. 128-9]

670

PALAEONTOLOGY, VOLUME 11

to Discocyclina and Nummulites, Asterocyclina and Pellatispira occur in fewer samples
and are much less abundant in numbers of individuals.

The following larger foraminifera are identified from the Kopili Formation, Garo
Hills (see also Samanta 1968, p. 129, table 2):

Asterocyclina matanzensis Cole
Discocyclina archiaci (Schlumberger)
D. assamica Samanta
D. august ae Weijden
D. dispansa (Sowerby)

I). eamesi Samanta
D. javana (Verbeek)

D. omphalus (Fritsch)

D. pygmaea Henrici
D. sella (d’Archiac)

D. sowerbyi Nuttall
D. sp. cf. D. trabayensis Neumann
Nummulites chavannesi de la Harpe
N. sp. aft. N. chavannesi de la Harpe
N. fabianii (Prever)

N. pengaronensis Verbeek
Pellatispira inflata Umbgrove
P. sp. cf. P. irregularis Umbgrove
P. madaraszi (Hantken)

P. sp. cf. P. orbitoidea (Provale)

Of these, D. augustae, D. sella, N. chavannesi, N. fabianii, and P. madaraszi are recorded
from the Priabonian of North Italy, while A. matanzensis, D. javana, D. omphalus,
D. pygmaea , D. sella, N. pengaronensis, and the four species of Pellatispira are abundantly
represented in the T b of the Indonesian region. The larger foraminiferal assemblage,
therefore, indicates a definite Upper Eocene age for the lower part of the Kopili
Formation.

Material. The material was collected from five localities in the Garo Hills, previously described
(Samanta 1965, p. 416, text-fig. 3). All the four species of Nummulites are represented by sufficient
material. Presence of free specimens permits a detailed study of these forms. Table 1 shows the
distribution of the species in the Garo Hills.

table 1. Distribution of Nummulites in the Kopili Formation, Garo Hills, Assam

Localities

Species

Sa

Rn

Rgt

N

K

Nummulites chavannesi

X

de la Harpe

N. sp. atf. N. chavannesi

X

X

X

de la Harpe

N. fabianii (Prever)

X

X

X

N. pengaronensis Verbeek

X

X

X

X

X

Acknowledgements. The author is indebted to Dr. J. R. Haynes for critically reading the manuscript;
Dr. F. E. Eames for helpful discussions; Professors H. Hagn and E. Montanaro Gallitelli for com-
parative material; Drs. F. Bieda and V. Roveda for literature; Professor Alan Wood for providing
facilities in his Department; and Mr. H. Williams for help in preparing the plates.

SYSTEMATIC PALAEONTOLOGY

Family nummulitidae de Blainville 1825
Subfamily nummulitinae de Blainville 1825
Genus nummulites Lamarck 1801

Nummulites chavannesi de la Harpe

Plate 128, figs. 11, 12; Plate 129, figs. 9-14; text-fig. 1

1877 Nummulites chavannesi de la Harpe, p. 232 ( nom . nud.).

1883fl Nummulites bouillei var. riitimeyeri de le Harpe, pi. 6. figs. 5-11.

1883fl Nummulites chavannesi de la Harpe, pi. 6. figs. 22-41.

18836 Nummulites riitimeyeri de la Harpe, pp. 162, 163, pi. 30, figs. 9-11.

SAMANTA: NUMMULITES, KOPILI FORMATION, ASSAM 671

1883 b Nummulites chavannesi de la Harpe; de la Harpe, pp. 163, 164, pi. 30, figs. 12-18.

19116 Nummulites chavannesi de la Harpe; Boussac, pp. 37, 38.

1934 Nummulites cf. chavannesi de la Harpe; Flandrin, pp. 254, 255, pi. 14, figs. 15, 16.

1934 Nummulites rutimeyeri de la Harpe; Flandrin, p. 254, pi. 14, fig. 17.

1938 Nummulites rutimeyeri de la Harpe; Flandrin, pp. 34, 35, pi. 3, figs. 9, 10.

1951 Nummulites rutimeyeri de la Harpe; Daci, pp. 209, 210, pi. 2, figs. 7, 8.

1951 Nummulites chavannesi de la Harpe; Daci, pp. 210, 211, pi. 2, fig. 9.

1957 Nummulites chavannesi de la Harpe; Bieda, pp. 46, 47, pi. 4, figs. 8, 9.

1960 Nummulites chavannesi de la Harpe; Hagn, p. 70, pi. 1, fig. 2; pi. 2, figs. 4, 5.

1961 Nummulites chavannesi de la Harpe; Roveda, pp. 177-81, pi. 14, figs. 1-8.

1963n Nummulites chavannesi de la Harpe; Bieda, pp. 71, 72, 186, pi. 6, figs. 5-7 ; pi. 7, figs. 1-3.

text-fig. 1. Nummulites chavannesi de la Harpe. a, Part of the equatorial section of a microspheric
specimen. x20 approx, b, Equatorial section of a megalospheric specimen, X 25 approx, c, Axial
section of a megalospheric specimen, X 25 approx. All from locality K.

Material. Megalospheric form — 25 specimens examined externally, 5 specimens studied in equatorial
section, and 5 in axial section. Microspheric form — 6 specimens examined externally, 3 specimens
studied in equatorial section, and 2 in axial section.

Description. Megalospheric form. Test small, lenticular, with slightly elevated polar
region surrounded by sloping peripheral part; margin acute. Surface ornamented with
well-developed polar pustules from which thin, straight to gently curved septal filaments
radiate. Diameter of test varies from T9 to 3-4 mm., thickness from 0-9 to T4 mm.,
ratio of diameter to thickness from 2-2 to 2-7, and diameter of polar pustules from 0-4
to 0-6 mm.

About 4\ to 6 regularly coiled whorls open rapidly. Spiral lamina thin and in outer

672

PALAEONTOLOGY, VOLUME 11

whorls height of spiral cavity about 4 to 6 times thickness of spiral lamina. Septa nearly
perpendicular to spiral lamina, straight with sharp curvature near distal end. About
8-11 septa occur in 1st whorl; 15-20 in 2nd; 19-28 in 3rd; 24-9 in 4th; and 28-32
in 5th.

Small, subcircular first chamber followed by subequal, reniform second chamber.
Separating wall either straight or curved outwards. Diameters of first chamber vary
from 0-055 x 0-050 mm. to 0-095x0-075 mm. and those of second chamber from
0-055 x 0-045 mm. to 0-100x0-060 mm. Distance across both chambers varies from
0-105 to 0-180 mm. Equatorial chambers quadrate in shape and about twice as high as
long.

In axial section first chamber circular and about 0-05 mm. in height. Chamber cavity
triangular in shape. Alar prolongations wide open. Marginal cord distinct. Wedge-
shaped polar plugs always very conspicuous and about 0-6 mm. in diameter near surface.

Microspheric form. Test small with well-developed, slightly elevated polar pustules;
margin acute. Septal filaments thin, radiate, nearly straight. Diameter of test varies
from 3-8 to 5-0 mm., thickness from 1-8 to 2-0 mm., ratio of diameter to thickness from
2-1 to 2-4 mm., and diameter of polar pustules from 0-8 to TO mm.

There are about 9 whorls in diameter of 4-2 mm. Whorls regularly coiled and open
rather rapidly. In outer whorls height of spiral cavity about 3 times thickness of spiral
lamina. Septa nearly perpendicular and straight with sharp curvature at top. Chambers
quadrate, about 2 to 3 times higher than long.

In axial sections alar prolongations wide open. Marginal cord distinct. Well-developed
polar plugs wedge-shaped, about TO mm. in diameter near surface.

Remarks. The presence of well-defined polar pustules, high equatorial chambers between
almost straight septa, and wide alar prolongations distinguish this species from the
associated nummulites in the Kopili Formation. The Assam specimens have been com-
pared with European material, provided by Professor H. Hagn. They are closely similar
to the Priabonian material described and illustrated by Roveda (1961).

Distribution. N. chavannesi has been reported from Italy, Spain, France, Switzerland, Poland, Hungary,
Albania, Turkey, Algeria, Egypt, and Somaliland. Its known range is from Upper Eocene to Oligocene.

In the Garo Hills N. chavannesi occurs only in the Upper Eocene Kopili Formation (Table 1). There
is no report of its occurrence in the other Upper Eocene localities in India and adjacent countries. The
present record of N. chavannesi extends its geographic distribution considerably.

EXPLANATION OF PLATE 128

Figs. 1-10. Nummulites pengaronensisNe rbeek. 1, External view of microspheric specimen, x6. 2, 3,
External views of megalospheric specimens; 2, inflated variety, x6; 3, compressed lenticular
variety, x9. 4, 5, Equatorial sections of megalospheric specimens, x 15; 4, inflated variety; 5, len-
ticular variety. 6-8, Axial sections of megalospheric specimens showing variation in transverse views
of tests ; 6, X 1 5 ; 7, 8, X 2 1 . 9, Axial section of microspheric specimen, X 9. 10, Equatorial section of
microspheric specimen, x9. 1, 3, 8, 10, from locality Sa; 2, 7, from locality K; 4, 5, from locality
Rn; 6, 9, from locality N (see Samanta 1965, p. 416).

Figs. 11, 12. Nummulites chavannesi de la Harpe. 11, Axial section of megalospheric specimen, x21.
12, Equatorial section of megalospheric specimen, x21. Both from locality K.

Figs. 13-15. Nummulites sp. aff. N. chavannesi de la Harpe. 13, External view of megalo-spheric
specimen, X 15. 14, Equatorial section of megalospheric specimen, x21. 15, Axial section of megalo-
spheric specimen, x 21. All from locality K.

Palaeontology, Vol. 11

PLATE 128

SAMANTA, Upper Eocene Nummulites

SAMANTA: NUMMULITES, KOPILI FORMATION, ASSAM 673

Nummulites sp. aft'. N. chavannesi de la Harpe
Plate 128, figs. 13-15; text-fig. 2

Material. Only megalospheric specimens were observed. 1 5 specimens examined externally, 5 specimens
studied in equatorial section, and 2 in axial section.

Description. Megalospheric form. Test very small, lenticular, with acute margin. Septal
filaments thin and radial. Diameter of test varies from T6 to 2-0 mm., thickness from
0-56 to 0-80 mm., and ratio of diameter to thickness from 2-0 to 3-4.

Spire is regular with about 44 to 54 whorls increasing regularly in height. Spiral lamina
thin, about } height of spiral cavity in thickness. Septa are perpendicular to wall and

text-fig. 2. Nummulites sp. aft'. N. chavannesi de la Harpe
(megalospheric form.) A, Equatorial section, b. Axial
section. Both from locality K, x25 approx.

nearly straight with curvature at top. About 8-10 septa occur in 1st whorl; 15-18 in
2nd; 19-25 in 3rd; 24-6 in 4th; and 24-30 in 5th.

Subcircular first chamber followed by subequal, crescentic second chamber. Separat-
ing wall gently curved outwards. Diameters of first chamber vary from 0-05 x 0 03 mm.
to 0-08. x 0-07 mm.; those of second chamber from 0-055 x 0-035 mm. to 0-10x0-055 mm.
Distance across both chambers varies from 0-09 to 0-15 mm. Equatorial chambers
quadrate and higher than long.

In axial section first chamber circular, about 0-06 mm. in height. Alar prolongations
wide open. Marginal cord distinct.

Remarks. The present specimens are closely similar to N. chavannesi in internal charac-
ters, but can easily be distinguished from the latter by the shape of the test and the
absence of polar pustules. They are provisionally identified as N. sp. aff. N. chavannesi.

Distribution, The species occur in localities Rn, N, and K in the Garo Hills (Samanta 1965, p. 416).
The presence of similar forms has also been noticed by the writer in the Upper Eocene of Surat-
Broach, Western India.

Nummulites fabianii (Prever)

Plate 129, figs. 1-8; text-fig. 3, 4

1905 Bruguieria fabianii Prever in title, Fabiani, pp. 1805, 1811, 1824.

1905 Bruguieria sub-fabianii Prever in lift.; Fabiani, pp. 1811, 1824.

674

PALAEONTOLOGY, VOLUME 11

1906 Nummulites fabianii (Prever in Fabiani); Boussac, pp. 88-90, pi. 1, figs. 1-5, 7-9; pi. 3,
fig. 6.

191 la Nummulites fabianii Prever in Fabiani; Boussac, pp. 40, 72, pi. 10, figs. 1, 2, 28; pi. 17,
figs. 8, 11, 13.

19116 Nummulites fabianii Prever in Fabiani; Boussac, pp. 79-84, pi. 1, figs. 6, 13; pi. 4, figs.
9, 10.

1928 Nummulites fabianii Prever; de Cizancourt, p. 294, pi. 2, fig. 10.

1930 Nummulites fabianii Prever; de Cizancourt, pp. 209, 210, pi. 22, figs. 4, 7; pi. 23, fig. 5.

1934 Nummulites fabianii Prever; Flandrin, p. 259, pi. 1, fig. 20.

1951 Nummulites fabianii Prever; Daci, pp. 221, 222, pi. 3, figs. 1, 2.

1951 Nummulites subfabianii Prever; Daci, pp. 222-4, pi. 3, figs. 4-7.

1957 Nummulites fabianii Prever; Bieda, p. 30, pi. 5, fig. 5.

1959 Nummulites retiatus Roveda, pp. 201-7, pi. 1, figs. 1-11.

1960 Nummulites fabianii (Prever); Hagn, p. 149, pi. 2, figs. 2, 3, 7.

1961 Nummulites fabianii (Prever)', Roveda, pp. 161-9, pi. 15, figs. 15, 16; pi. 17, figs. 8, 9;

pi. 18, figs. 4, 5; pi. 19, figs. 1, 6-8, 14-16.

1963a Nummulites fabianii Prever; Bieda, pp. 101-4, 195, 196, pi. 15, fig. 9; pi. 16, figs. 1-4.
19636 Nummulites fabianii Prever; Bieda, pp. 201-4, 214-15, pi. 13, figs. 3, 4.

1965 Nummulites fabianii Prever; Bozorgnia and Kalantari, pp. 17, 18; pi. 20, figs. 1-7.

Material. Megalospheric form — 20 specimens examined externally, 5 studied in equatorial section, and
7 in axial section. Microspheric form — 10 specimens examined externally, 4 studied in equatorial
section, and 1 in axial section.

Description. Megalospheric form. Test small, lenticular, with subacute margin. Surface
ornamented with spirally arranged rectangular meshes produced by intersections of
radial filaments with raised spiral line. Spirally arranged granules joined together by
‘transverse lamina’ produce raised spiral line. In some specimens granules cluster at
poles to form polar pustules. Diameter of test varies from 1-8 to 3-0 mm., thickness from
1-25 to 1-85 mm., and ratio of diameter to thickness from 1-4 to 1-9.

About 5 to regularly coiled whorls occur, increasing slowly in height. Spiral
lamina thick, and in some inner whorls may be as thick as height of spiral cavity. Near
periphery height of spiral cavity about 2 to 3 times thickness of spiral lamina. Septa
slightly inclined to spiral wall, straight to gently curved in their course. About 6-7 septa
occur in 1st whorl; 9-13 in 2nd; 12-16 in 3rd; 15-20 in 4th; and 16-22 in 5th.

Subcircular first chamber followed by smaller, semicircular to reniform second cham-
ber. Separating wall either straight or curved outwards. Diameters of first chamber vary
from 0-130x0-095 mm. to 0-20x0-20 mm.; those of second chamber from 0*10x0-05
mm. to 0-175 X 0-095 mm. Distance across both chambers varies from 0-16 to 0-28 mm.
Chambers quadrate in shape. Near centre, chambers almost as long as high, but in
ontogeny chambers become considerably longer, so that in outer whorls chambers
become twice as long as high.

EXPLANATION OF PLATE 129

Figs. 1-8. Nummulites fabianii (Prever). 1, External view of microspheric specimen, X 6. 2, 3, Equatorial
sections of microspheric specimens, x9. 4, External view of megalospheric specimen, Xl5. 5, 6,
Axial sections of megalospheric specimens, x21. 7, Equatorial section of megalospheric specimen,
x 15. 8, Axial section of microspheric specimen, x7-5. All from locality Rn (see Samanta 1965,
p. 416).

Figs. 9-14. Nummulites chavannesi de la Harpe. 9, External view of microspheric specimen, x6.
10, 11, External views of megalospheric specimens, x 12. 12, Axial section of megalospheric speci-
men, x21. 13, Equatorial section of megalospheric specimen, x21. 14, Equatorial section of
microspheric specimen, X 15. All from locality K.

Palaeontology, Vol. 11

PLATE 129

SAMANTA, Upper Eocene Nummulites

675

SAMANTA: NUMMULITES , KOPILI FORMATION, ASSAM

text-fig. 3. Nummulites fabianii (Prever) (megalospheric form). A, Equatorial
section, b. Axial section. Both from locality Rn, x25 approx.

text-fig. 4. Nummulites fabianii (Prever). Part
of the equatorial section of a microspheric
specimen from locality Rn, x 18 approx.

In axial section first chamber circular, about 0-10 to 0-13 mm. in height. Spiral lamina
rather thick. There may be reduction in thickness of spiral lamina at periphery. Alar
prolongations narrow to moderately open. Marginal cord distinct. Pillars well-developed,
start from marginal cord of each whorl, and of uniform thickness throughout length.
At poles, pillars cluster together to form polar plugs. Diameter of pillars varies from
0-050 to 0-075 mm. and polar plugs from 0-25 to 0-50 mm.

Microspheric form. Test medium-sized, lenticular, with subacute margin. Surface of

vy

C 6055

676

PALAEONTOLOGY, VOLUME 11

test ornamented with thin, reticulate septal filaments. In young individuals rectangular
meshes are discernible but in adult specimens branching filaments produce complex
network. Diameter of test varies from 5-1 to 8-4 mm., thickness from 2-7 to 4-2 mm., and
ratio of diameter to thickness from 1 -9 to 2-2.

In equatorial section about 9 to 13 whorls occur, coiled regularly and increasing
slowly in height during ontogeny. Spiral lamina rather thick. In adult whorls height of
spiral cavity usually greater than thickness of spiral lamina. Septa inclined to whorl
wall, and straight to gently curved in their course. Equatorial chambers longer than
high, and in outer whorls 3 to 4 times as long as high.

In axial section, alar prolongations narrow to moderately open. Marginal cord dis-
tinct. Pillars moderately developed. Each pillar starts from marginal cord and extends
up to surface. Diameter of pillars varies from 0-05 to 0-15 mm. In polar region pillars
cluster together to form polar plug-like structures about 0-8 mm. in diameter near
surface.

Remarks. Both in external and internal features the present form is distinctive. The
reticulate ornamentation, the long equatorial chambers, and the pillared axial section
enables the species to be distinguished from the associated nummulites in the Kopili
Formation. The Assam specimens were compared with those of N. fabianii from North
Italy provided by Professor Montanaro-Gallitelli.

Because of their distinctive morphological features and wide geographic distribution
in the rocks of Upper Eocene to Oligocene age, the reticulate Nummulites have received
particular attention and several species have been described. But at present there is
considerable difference of opinion about the validity of a number of these forms (see
Eames et al. 1959; Bieda 196 3b); consequently, application of reticulate Nummulites
species in the finer biostratigraphic zonation of Upper Eocene-Oligocene rocks is
lacking.

Distribution . N. fabianii is one of the most widely distributed representatives of the genus, reported
from the Upper Eocene of Italy, Spain, France, Switzerland, Poland, Hungary, Albania, Rhodes
Island, Turkey, Morocco, Algeria, Tunisia, Libya, Egypt, and Iran.

In the Garo Hills N. fabianii occurs in the Kopili Formation at localities Sa, Rn, and K (Table 1).
It occurs also in two other Upper Eocene localities in the Indian region; in the Sulaiman Range its
presence has been noted by Bayliss (1961), while the writer has observed it in Surat-Broach in associa-
tion with Pellatispira spp., etc.

There is no authentic record of reticulate Nummulites from the Upper Eocene of the Malayan
Archipelago (Cole 1963, Adams 1965). The only report of an occurrence in association with a typical
Upper Eocene assemblage from this region was that by Cole (Cloud and Cole 1953, p. 323) who later
(1963, pp. E4, E14) postulated that the Upper Eocene species in the assemblage are reworked specimens
and that reticulate Nummulites do not occur in the Eocene of the Malayan Archipelago. Thus, Assam
is the easternmost locality with N. fabianii.

Nummulites pengaronensis Verbeek
Plate 128, figs. 1-10; text-figs. 5, 6

1871 Nummulites pengaronensis Verbeek, pp. 3-6, pi. 1, figs. 1 a-k.

1892 Nummulites nanggoelani Verbeek, pp. 116, 118.

1896 Nummulites nanggoelani Verbeek; Verbeek and Fennema, p. 1152, pi. 8, figs. 111-13.
1896 Nummulites pengaronensis Verbeek; Verbeek and Fennema, pp. 1153, 1 154.

1912 Nummulites pengaronensis Verbeek; Douville, pp. 284, 285, pi. 24, fig. 6.

SAMANTA: NUMMULITES, KOPILI FORMATION, ASSAM

677

1921 Nummulites cf. pengaronensis Verbeek; Yabe, pp. 104, 105, pi. 18, fig. 8.

1929 Nummulites pengaronensis Verbeek; Vlerk, pp. 20, 21, figs. 12, 35a, b.

1932 Camerina pengaronensis (Verbeek); Doornink, pp. 283, 284, pi. 4, figs. 1-3; pi. 6, fig. 12.
1934 Camerina pengaronensis (Verbeek); Henrici, pp. 29, 30, pi. 1, fig. 10.

1934 Camerina cf. pengaronensis (Verbeek); Caudri, p. 52.

1953 Camerina saipanensis Cole, pp. 20, 21, pi. 2, figs. 7-19.

1957 Camerina pengaronensis (Verbeek); Cole, pp. 753, 754, pi. 231, figs. 1-17.

1959a Nummulites pengaronensis Verbeek; Nagappa, pp. 163, 166, pi. 10, figs. 3-5.

1965 Nummulites cf. saipanensis (Cole); Adams, p. 313, pi. 23, fig. c.

text-fig. 5. Nummulites pengaronensis Verbeek (megalospheric form), a-c, Axial sections, x25
approx.; a, from locality Sa; b, from locality K; c, from locality N. d, Split specimen from locality
Sa, x 20 approx, e, Equatorial section from locality Rn, x 25 approx.

678

PALAEONTOLOGY, VOLUME 11

Material. Megalospheric form — 75 specimens examined externally, 5 studied in equatorial section, and
12 in axial section. Microspheric form — 12 specimens examined externally, 3 studied in equatorial
section, and 4 in axial section. In addition, 23 split specimens of megalospheric form and 3 of micro-
spheric form were studied in equatorial view.

Description. Megalospheric form. Test small, compressed, lenticular to globose, with
acute margin. Surface marked by thin, radial septal filaments, straight to gently curved
at ends. Diameter of test varies from 2-0 to 5-2 mm., thickness from 0-8 to 3-3 mm., and
ratio of diameter to thickness from 1-3 to 3-6.

Spire more or less regular with about 5 to 7 whorls increasing regularly in height,
with exception of last whorl, which may be narrower than preceding one. Height of
spiral cavity in outer whorls about 3 times thickness of spiral lamina. Septa nearly
perpendicular to inclined at their base, straight for about half their course, then curve
sharply backwards. Thickness of septa decreases considerably from proximal to distal

end. About 6-7 septa occur in 1st whorl; 1 1-14
in 2nd; 17-19 in 3rd; 23-6 in 4th; 24-7 in
5th; 26-9 in 6th; and 29-32 in 7th.

First chamber circular to elliptical in equa-
torial section, followed by usually smaller, cres-
centic to reniform second chamber. Separating
wall curved outwards. Diameters of first
chamber vary from 0-125 xO-120 mm. to
0-275x0-225 mm., those of second chamber
from 0-105x0-050 mm. to 0-175x0-075 mm.
Distance across both chambers varies from
0-200 to 0-325 mm. Equatorial chambers
subquadrate to falciform in shape, usually
higher than long, although reverse also quite
common.

In axial section first chamber circular, about
0-100 to 0-175 mm. in height. Alar prolongations
extremely narrow. Appreciable reduction in
thickness of spiral lamina at periphery. Mar-
ginal cord distinct. Traces of weakly developed pillar-like structures in polar region.

Microspheric form. Test small- to medium-sized, lenticular, with acute margin. Septal
filaments radial, straight. Diameter of test varies from 4-6 to 8-9 mm., thickness from
1-9 to 3-3 mm., and ratio of diameter to thickness from 2-0 to 3-3.

In equatorial section about 9 to 13 whorls occur, regularly increasing in height.
Height of spiral cavity greater than thickness of spiral lamina. Septa nearly perpendi-
cular at base, straight for about half their course, then curve sharply backwards. Thick-
ness of septa decreases considerably from proximal to distal end. Chambers sub-
quadrate to falciform, usually higher than long.

In axial section chamber cavity triangular in shape. Alar prolongations extremely
narrow. Considerable reduction in thickness of spiral lamina at periphery. Marginal
cord weakly developed. In polar region traces of pillar-like structures.

text-fig. 6. Nummulites pengaronensis Ver-
beek. Part of the equatorial section of a
microspheric specimen from locality Sa,
x 18 approx.

Remarks. Although the Assam specimens show considerable variation in external form

SAMANTA: NUMMULITES, KOPILI FORMATION, ASSAM 679

of the test, they are identical in internal structures and are included here under one
species. They are characterized externally by the presence of radial septal filaments and
the absence of distinct polar pustules, and internally by the characters of the septa in
equatorial section and very narrow alar prolongations in axial section.

Nummulites saipanensis (Cole), originally described from Saipan, Mariana Islands
(Cole and Bridge 1953), was later considered by its author (Cole 1957) to be synonymous
with N. pengaronensis Verbeek. This is accepted here. Cole (op. cit.) also included
Nummulites semiglobula (Doornink), described from Java (Doornink 1932), in the
synonymy of N. pengaronensis. However, although N. semiglobula , as well as N. gerthi,
bear some resemblance to N. pengaronensis, a more detailed study based on the topo-
type material of these two species is needed before considering them as junior synonyms
of N. pengaronensis.

Among the European species, Nummulites stel/atus Roveda described from the
Priabonian of North Italy (Roveda 1961), is very closely similar to N. pengaronensis in
internal morphology.

Sen Gupta (1965), while working on some Middle Eocene Nummulites from Western
India, regarded N. pengaronensis as a junior synonym of Nummulites beaumonti
d’Archiac and Haime. Under the remarks on N. beaumonti , Sen Gupta (op. cit., p. 93)
wrote: ‘Another synonym of N. beaumonti is N. pengaronensis Verbeek, a wide-spread
Indo-Pacific form. It shows a tight coiling of spiral wall, which is almost uniformly thick,
and small embryonic chambers, as does typical N. beaumonti. These features are clearly
seen in the figures of N. pengaronensis published by Cole (1957m pi. 231, figs. 1-17) and
have been confirmed by an examination of Cole’s material from Eniwetok.’ Sen Gupta,
therefore, neither examined the type or topotype materials nor consulted the type
description and illustrations of N. pengaronensis to support his remarks. Further, he
did not compare his Indian specimens of N. beaumonti with ones from the Indian region
identified as N. pengaronensis by other workers. (There are good illustrations of micro-
spheric and megalospheric specimens of N. pengaronensis from Eastern India (Nagappa
1959a, b). In both these publications, there are good illustrations of N. beaumonti too,
and in the latter Nagappa has pointed out (p. 158, pi. 21, figs. 1, 2) the conspicuous
difference in the character of the septa in these two species.) A thorough comparison
between the two species, including such taxonomically important features as the charac-
ters of the equatorial chambers and the septa as seen in equatorial section, has not been
made and Sen Gupta’s remarks do not appear to be justified.

A comparison of the description and illustrations of N. beaumonti provided by Davies
(1940) with those of N. pengaronensis given by Verbeek (1871), Doornink (1932), and
others, shows clearly that they are two distinct species. In equatorial section they can
always be separated by the characters of the septa and the equatorial chambers, and in
axial section by the width of the alar prolongations and the degree of development of
polar plugs. The writer believes that these two species are not only distinct but that they
belong to two different groups of species. If N. pengaronensis is considered to be a
synonym of N. beaumonti, or in other words, if the morphological differences between
them are not considered to be of specific importance, then the usefulness of species of
Nummulites in the stratigraphic analysis and correlation of the Lower Tertiary will be
greatly reduced. With N. pengaronensis as a junior synonym, the stratigraphic range
of N. beaumonti would be from Middle Eocene to Oligocene (not Middle to Upper

680

PALAEONTOLOGY, VOLUME 11

Eocene as mentioned by Sen Gupta (1965, p. 92)) and it would then be difficult to use it
as a ‘key’ species in stratigraphy.

Distribution. N. pengaronensis is a widely distributed Indo- Pacific form and has been reported from the
Central Pacific Islands, the East Indies, Burma, Eastern India, and Western Pakistan. Its known
stratigraphic range is from the upper part of the Middle Eocene to Oligocene.

In the Garo Hills N. pengaronensis ranges from the Upper Member of the Siju Limestone (Middle
Eocene) to the overlying Kopili Formation (Upper Eocene). It is the most abundant representative of
the genus in the Kopili Formation and occurs in all the five localities (see Table 1).

Although it is known to occur in the Sulaiman Range (Eames 1952), there is no report of the species
from the Upper Eocene of Surat-Broach, Western India.

GENERAL REMARKS

The four species of Nummulites recorded from the Kopili Formation belong to three
different groups. N.fabianii (Prever) belongs to the reticulate group of forms charac-
teristic of Upper Eocene to Oligocene. This is the only pillared form in the present
assemblage. Of the three remaining striate forms, N. pengaronensis with its strongly
curved septa and very narrow alar prolongations is distinctly different from N. cliavan-
nesi and N. sp. aff. N. chavannesi , characterized by rapid opening of the whorls, straight
septa, and wider alar prolongations. In all these four forms the marginal cord is only
moderately developed and the size of the tests does not exceed 9 mm. The striate forms
are much more abundant than the reticulate one, and occur in almost all foraminiferal
samples.

The assemblage of Nummulites in the Kopili Formation is markedly different from
that in the underlying Upper Member of the Siju Limestone (Samanta 1968). In the
latter horizon the assemblage is characterized by the presence of large, highly evolved
species showing three ‘parts’ in the spire as recognized by Schaub (1963). These forms
are totally absent in the Kopili Formation. Also, the number of species of Nummulites
is fewer in the Kopili Formation than in the underlying Siju Limestone. Throughout
the range of the genus in the Indian region the most striking change in the assemblages
occurs at this horizon. The total absence of the typical representatives of the genus
characterizing the older horizons, together with the appearance of a new group of forms
in the Kopili Formation, makes the assemblage more akin to that of the Oligocene than
to that of underlying Middle Eocene horizon. Indeed, in the absence of reticulate species
it is difficult to distinguish the Nummulites assemblage of the Upper Eocene from that of
the Oligocene.

In the presence of N. fabianii and abundant small to medium striate forms, the present
assemblage is closely comparable to that recorded in the Priabonian of Europe. It is
distinguished from the latter essentially by the absence of the Nummulites striatus-
garnieri group of forms, which are common in the European Upper Eocene. The Num-
mulites assemblage in the Kopili Formation is, however, quite distinct from that known
from the Upper Eocene of the Far East. N. pengaronensis is the only species common to
the two regions. The Nummulites yawensis-djokdjokartae group of forms described from
the Upper Eocene of the Malayan Archipelago are absent in the Kopili Formation of
Assam. The absence of the well-known and widely distributed Upper Eocene reticulate
Nummulites in the Far East constitutes the most striking difference between the Upper
Eocene Nummulites assemblages of the two regions.

SAMANTA: NUMMULITES , KOPILI FORMATION, ASSAM

681

REFERENCES

adams, c. G. 1965. The Foraminifera and stratigraphy of the Melinau Limestone, Sarawak, and its
importance in Tertiary correlation. Q. Jl geol. Soc. Loud. 121, 283-338, pi. 21-30.
bayliss, d. d. 1961. An investigation of certain larger fossil foraminifera from Pakistan. Unpubl. Ph.D.
thesis, Univ. Wales, 1-253, pi. 1-26.

bieda, f. 1957. Die Fauna Grosser Foraminiferen im Obereozaen der Slowakei. Geol. Sb., Bratislava,
8, 28-71, pi. 2-6.

1963a. Larger foraminifers of the Tatra Eocene. Inst. Geol. Peace, 37, 1-215, pi. 1-26.

19636. Septieme niveau de grands Foraminiferes dans le Flysch des Karpates Polonaises. Ann.

Soc. geol. Pologne, 33, 189-218, pi. 11-14.

boussac, J. 1906. Developpement et morphologie de quelques foraminiferes de Priabona. Bull. Soc.
geol. Fr. (4) 6, 88-97, pi. 1-3.

— 1911a. Etudes stratigraphiques et paleontologiques sur le Nummulitique de Biarritz. Ann.
Hebert, 5, 1-95, pi. 1-24.

19116. Etudes stratigraphiques sur le Nummulitique alpin. Mem. Carte geol. Fr. 1 122, pi. 1-5.

bozorgnia, f. and kalantari, A. 1965. Nummulites of parts of Central and East Iran — A thin-section
Study. 1-28, pi. 1-24. Tehran, Iran.

caudri, c. m. b. 1934. Tertiary deposits of Soemba. Diss. Leiden Natl. Mas. geol., Amsterdam, 1-224,
pi. 1-5.

cizancourt, m. de. 1928. Sur quelques nummulites du flysch Karpatique et sur leur signification pour
la stratigraphie des Karpates. Kosmos, Warsz. 53, 287-312, pi. 1, 2.

- 1930. Sur la stratigraphie et la faune Nummulitique du Flysch de l’Albanie. Bull. Soc. geol. Fr.
(4) 30, 195-212, pi. 22, 23.

cloud, p. e. jun. and cole, w. s. 1953. Eocene Foraminifera from Guam, and their implications.
Science, N.Y. 117, 323-4.

cole, w. s. 1957. Larger Foraminifera from Eniwetok drill holes. Prof. Pap. U.S. geol. Surv. 260- V,
743-84, pi. 231-49.

1963. Tertiary larger Foraminifera from Guam. Ibid. 403-E, 1-28, pi. 1-11.

and bridge, j. 1953. Geology and larger Foraminifera of Saipan Island. Ibid. 253, 1-45, pi. 1-15.

daci, a. 1951. Etude paleontologique du Nummulitique entre Kiigukgekmece et Catalca. Rev. Fac.
Sc. Univ. Istanbul, (B) 16, 89-246, pi. 1-5.

davies, l. m. 1940. The Upper Khirthar beds of north-west India. Q. Jl geol. Soc. Lond. 96, 199-230,
pi. 9-12.

doornink, h. w. 1932. Tertiary Nummulitidae from Java. Verb, geol.-mijnb. Genoot. Ned. 9, 267-315,
pi. 1-10.

douville, h. 1912. Quelques foraminiferes de Java. Samml. geol. Reichsmus. Leiden, (1) 8, 279-94,
pi. 22-24.

eames, f. e. 1952. A contribution to the study of the Eocene in Western Pakistan and Western India:
D. Discussion of the faunas of certain standard sections, and their bearing on the classification
and correlation of the Eocene in Western Pakistan and Western India. Q. Jl geol. Soc. Lond. 107,
173-96.

clarke, w. j. and banner, f. t. 1959. Nummulites retiatus, a synonym of Nummulites fabianii.

Revue Micropaleont . 2, 113.

evans, p. 1932. Explanatory notes to accompany a table showing the Tertiary succession in Assam.

Trans. Min. geol. metall. Inst. India, 27, 155-260.
fabiani, r. 1905. Studio geo-paleontologico dei Colli Berici. Atti 1st. veneto Sci. 64, 1797-1839.
flandrin, J. 1934. La faune de Tizi Renif pres Dra el Mizan (Algerie). Bull. Soc. geol. Fr. (5) 4,
251-71, pi. 14-16.

1938. Contribution a l’etude paleontologique du Nummulitique algerien. Mater. Carte geol.

Alger. (1) Paleont. 8, 1-155, pi. 1-15.

hagn, h. 1960. Die stratigraphischen, palaeogeographischen und tektonischen Beziehungen zwischen
Molasse und Helvetikum im ostlichen Oberbayern. Geologica Bav. 44, 1-208, pi. 1-12.
harpe, p. de la. 1877. Note sur les Nummulites des Alpes occidentales. Verb, schweiz. naturf. Ges.
60, 227-32.

682

PALAEONTOLOGY, VOLUME 11

harpe p. de la. 1883a. Etude des Nummulites de la Suisse et revision des especes eocenes des genres
Nummulites et Assilina. Abh. schweiz. palaont. Ges. 10, 141-60, pi. 3-7.

18836. Monographic der in Aegypten und der libyschen Wiiste vorkommenden Nummuliten.

Palaeontographica, 30, 157-216, pi. 30-35.

henrici, H. 1934. Foraminiferen aus dem Eozan und Altmiozan von Timor. Ibid. Suppl. 4, pi. 4,
1-56, pi. 1-4.

nagappa, y. 1951. The stratigraphical value of Miscellanea and Pellatispira in India, Pakistan and
Burma. Proc. Indian Acad. Sci. (B) 33, 41-8.

1959a. Foraminiferal biostratigraphy of the Cretaccous-Eocene succession in the India-Pakistan-

Burma region. Micropaleontology, 5, 145-92, pi. 1-11.

19596. Note on Operculinoides Hanzawa 1935. Palaeontology, 2, 156-60, pi. 21-23.

rao, s. r. n. 1941. The Tertiary sequence near Surat and Broach (Western India), with description of
foraminifera of the genus Pellatispira from the Upper Eocene of this region. Half-yrly J. Mysore
Univ. N.s. (B) 2, 5-17, pi. 1, 2.

roveda, v. 1959. Nummulites retiatus nouvelle espece de nummulite reticulee des Abruzzes (Italie).
Revue Micropaleont. 1, 201-7, pi. 1.

1961. Contributo alio studio di alcuni macroforaminiferi di Priabona. Riv. ital. Paleont. Stratigr.

67, 153-224, pi. 14-19.

samanta, b. k. 1965. Discocyclina from the Upper Eocene of Assam, India. Micropaleontology, 11,
415-30, pi. 1-4.

1968. The Eocene Beds of Garo Hills, Assam, India. Geol. Mag. 105, 124-35.

schaub, h. 1963. Uber einige Entwicklungsreihen von Nummulites und Assilina und ihre stratigra-
phische bedeutung. Evolutionary trends in Foraminifera, 282-97. Amsterdam.
sen gupta, b. k. 1965. Morphology of some key species of Nummulites from the Indian Eocene.
J. Paleont. 39, 86-96, pi. 15-17.

verbeek, r. d. m. 1871. Die Nummuliten des Borneo-Kalksteines. Neues Jb. Miner. Geol. Palaont.
Jahr. 1871, 1-14, pi. 1-3.

1892. Voorloopig bericht over Nummuliten, Orbitoiden en Alveolinen van Java, en over den

ouderdom der gesteenten waarin zij optreden. Natuurw Tijdschr. Ned.-Indie, 51, 101-39, pi. 1.

and fennema, r. 1896. Description geologique de Java et Madoura. 2 vols. 1-1183, pi. 1-11.

Amsterdam.

vlerk, i. m. van der. 1929. Groote foraminiferen van N. O. Borneo. Meded. Dienst Mijnb. Ned.-Indie
9, 1-44.

yabe, H. 1921. Note on some Eocene Foraminifera. II. Notes on two Foraminiferal Limestone from
E. D. Borneo. Sci. Rep. Tdhoku Univ. II Series (Geol.), 5 (4), 100-6, pi. 17, 18.

BIMAL K. SAMANTA
Department of Geology
University College of Wales

Typescript received 3 January 1968 Aberystwyth, Wales

A NEW PLANT FROM THE LOWER OLD RED
SANDSTONE OF SOUTH WALES

by DIANNE EDWARDS

Abstract. A new plant is described from the Senni Beds of the Lower Devonian of South Wales. The naked
axes were pseudomonopodially and dichotomously branched and contained protosteles in which the protoxylem
was central. Terminal fructifications consisted of sporangia alternating with sterile bracts; the plant was homo-
sporous. A comparison is made with other Devonian genera, which show similar organization in the fertile
regions, and it is concluded that the plant should be placed in a new genus, Krithodeophyton, assigned to the
Barinophytaceae (Incertae sedis).

The fossils described in this paper were among those collected by Croft from the Brecon
Beacons Quarry, also called the Storey Arms Quarry (Nat. Grid Ref. SO 972208) and
are now in the Department of Palaeontology, British Museum (Natural History). The
quarry is in the Senni Beds, which form the lower part of the Breconian Stage of the
Lower Old Red Sandstone in South Wales (Croft 1953) and are probably equivalent
to the Siegenian of Europe. Among the plants previously described from this locality
are Gosslingia breconensis, Zosterophyllum Hanover anum, and Drepanophycus spinae-
formis (Heard 1927 and 1939, Croft and Lang 1942, Edwards and Banks 1965, Edwards
1967). The majority of the plants were preserved as compressions in a fine-grained,
blue-grey sandstone, but parts of the axes were sometimes petrified.

Small pieces of cuticle were recovered after bulk maceration of the rock in commercial
strength (40%) hydrofluoric acid. These were then treated with Schulze’s solution (con-
centrated nitric acid and potassium chlorate) for 1-8 hours and, when the carbon had
oxidized and the outlines of cells were visible, the fragments were washed, immersed in
Diaphane solvent and mounted in Diaphane (Distributors: Will Scientific Inc., New
York 52, N.Y., U.S.A.). Pieces of carbon were also picked oft' both axes and sporangia
with steel needles, macerated in Schulze’s solution and mounted in the same way. Film
pulls were made from those fossils, which were exposed on the surface of the rock.
A solution of cellulose nitrate in amyl acetate was poured over the fossil and left to dry
overnight. The resulting rough, transparent film was peeled off and mounted in Harleco
Synthetic Resin (H.S.R.). (Distributors: Arthur H. Thomas Company, Philadelphia,
Pa., U.S.A.).

The anatomy of the pyritised axes was investigated using a modification of the method
described by Beck in 1955. A small piece of rock containing a petrified fossil was first
embedded in the synthetic resin, Ceemar, because the rock matrix tended to crumble
when sawn (Leclercq and Noel 1953). The transparent block of plastic was trimmed to
size and cut into sections about a millimetre thick. These were ground smooth on a glass
plate with grade 600 carborundum powder, washed and then treated in chromic oxide
powder until a good polish was obtained. After washing they were dried, soaked in
xylol and mounted in H.S.R. The sections of the axes were then examined using a Leitz
Ultropak microscope. All the preparations have been deposited in the Department of
Palaeontology, British Museum (V26578, V26579, V521 37-54).

[Palaeontology, Vol. 11, Part 5, 1968, pp. 683-90, pis. 130-32.]

684

PALAEONTOLOGY, VOLUME 11

SYSTEMATIC DESCRIPTION

Family barinophytaceae Krausel and Weyland 1961
Genus krithodeophyton gen. nov.

Type species. Krithodeophyton croftii sp. nov.

Diagnosis. Plant consisting of naked, pseudomonopodially branching axes with some
dichotomous branching in distal parts. Simple protostele composed of mainly scalari-
form, but some reticulate tracheids; protoxylem central. Sporangia in terminal spikes.
Dichotomous branching within the base of the fertile region or just below it. Oval
sessile sporangia borne in two rows, one on either side of the axis. Narrow sterile
appendages, attached at right angles to the axis, alternate with the sporangia. Bracts
straight throughout their length or with distal parts curving downwards. Sporangium
wall composed of isodiametric cells. Plant homosporous. Spores assignable to the dis-
persed spore genus Apiculiretusispora ( sensu Streel 1964).

Krithodeophyton croftii sp. nov.

Plate 130, figs. 1-12; Plate 132, figs. 1-10

Diagnosis. Plant, at least 10-0 cm. high, consisting of naked, pseudomonopodially
branching axes, 1 -5-4-3 mm. wide, with some dichotomous branching in the distal
parts; wide angles (< 80°) at branching points. Small bud-like lateral branches about
4-0 mm. long. Simple,(circular protosteleTjaverage diameter 0-5 mm., composed of mainly
scalariform, with some reticulate, tracheids; smallest elements at the centre of the xylem.
Outer cortex composed of elongate, thick-walled cells, 36-60 /a in diameter. Epidermis
composed of short, fusiform cells. Sporangia aggregated into terminal spikes, 2-5-3-0
mm. wide and up to 1-3 cm. long. Dichotomous branching within the base of the fertile
region or just below it. Oval sessile sporangia, 1-25-1-5 mm. long and 0-8-1 -0 mm. wide
borne in two rows, one on either side of the axis. Narrow sterile appendages up to 2-5
mm. long, given oft' at right angles to the axis, alternate with the sporangia. Bracts
straight throughout their length or with distal parts curving downwards. Sporangium
wall composed of isodiametric cells, of average diameter 27 p. Homosporous. Spores
approximately circular in outline, average diameter is 60 /a; simple trilete f-f of the spore
radius long; variable wall elements up to 3 p high (the majority being under 1 /a);
ornament reduced or absent in contact area. Assignable to the dispersed spore genus
Apiculiretusispora.

EXPLANATION OF PLATE 130

Figs. 1-12. Krithodeophyton croftii sp. nov. 1,4, Rock bearing fertile spikes, xl-1 (V26578 and V26579
respectively). 2, 5, Sterile axes, x 0-9 (V52154 and V52152 respectively). 3, Film pull showing epider-
mal cells, X 80 (V52151, BMP 25). 6, Film pull from axis showing small lateral branch. (Inset =
specimen before treatment), x6 (V52154, BMP 12). 7, Film pull from axis showing cortical cells,
x 50 (V52144, BMP 1). 8, Central strand after treatment with Schulze’s solution, x 200 (V52137,
BMM 22). 9, Central strand on film pull, X 100 (V52145, BMP 2). 10, T.S. axis showing xylem strand
before a division, x 108 (V52150, RSI/8). 11, T.S. pyritised xylem strand with centre not preserved,
x 108 (V52138, RSI/4). 12, T.S. part of outer cortex, X 108 (V52150, RSI/5).

Palaeontology, Vol. II

PLATE 130

EDWARDS, Lower Old Red Sandstone plants

DIANNE EDWARDS: A NEW PLANT FROM SOUTH WALES 685

Locality. Brecon Beacons Quarry, abandoned roadside quarry on the A470 between Brecon and
Merthyr, approximately 1\ miles south of Brecon.

Horizon. Senni Beds, Breconian, Lower Old Red Sandstone of South Wales (= Siegenian).

Holotype. Specimen V26579, Department of Palaeontology, British Museum (Natural History),
London.

Description of vegetative parts. The over-all height, branching frequency, and basal
parts of the plant are unknown. The most complete fertile specimen found was 10 cm.
long, with axes 2 mm. in diameter. There were two branching points (3-4 cm. apart) in
the vegetative region, and a further dichotomy occurred within the base of the spike of
sporangia. The diameter of the sterile axes ranged between T5 and 4-0 mm. and axes
of similar size were always found together. Hence it is probable that larger ones formed
the basal, and small axes the more terminal, parts of the plant. Branching was pseudo-
monopodial in these wider axes, and the narrower ‘lateral’ branch formed a wide angle
with the ‘main’ axis (PI. 130, fig. 2). These axes tended to be slightly flexuous and, in
addition to a central line, 0-5 mm. in diameter, faint longitudinal striations were some-
times visible. Small, lateral bud-like structures up to 4-0 mm. long, which tapered in
width from base to apex were also present (PI. 130, fig. 6). Each was supplied with a
central strand of xylem. The axes associated with the fertile regions were much narrower
(< 20 mm. wide) and branched dichotomously.

Film pull and maceration preparations revealed some cellular detail of the outer
layers of the axes. The outer cortex was the tissue most frequently recovered and con-
sisted of very long cells with thick walls, 36-44 p apart. Occasionally the tapering, over-
lapping ends of these cells were seen (PI. 130, fig. 7). Scattered among the thick parallel
walls were thinner, irregular lines probably representing the remains of underlying
tissue. The outer cortical cells were represented in the transverse sections of petrified
axes by 1 or 2 layers of angular cells (35-60 p in diameter) surrounding a structureless
mass of pyrites, containing a circular protostele. The outer walls of the cortical cells
were often broken down giving the surface of the axis a hairy appearance (PI. 1 30, fig. 1 2).
In one instance only, a piece of cuticle was found where the cells had thinner walls,
were fusiform in shape and relatively short (PI. 130, fig. 3). It is probable that this was
the epidermis.

The central strand of xylem was composed mainly of scalariform tracheids in which
pits on adjacent walls were opposite (PI. 130, figs. 8 and 9). A few elements showed
reticulate pitting. Sections through petrified xylem revealed similar anatomy. The
central part was usually not preserved (PI. 130, fig. 11), but a decrease in tracheid
diameter from the outside of the xylem inwards was always apparent. In a few cases,
very small elements could be seen at the centre. The protoxylem is therefore considered
to be central. The largest tracheids measured 35 p in diameter and the distance between
the horizontal bars of the scalariform tracheids was sometimes as great as 10 p.

In one axis only could stages in the division of the xylem be seen. A few millimetres
below the branching point, the xylem became oval in cross section (PI. 130, fig. 10) and
divided to give two circular strands of approximately the same diameter. The axis itself
then divided into two.

I have found well-preserved pyritised axes with similar anatomy to those described
above at the same locality. I include a detailed account of them here to emphasize this

686

PALAEONTOLOGY, VOLUME 11

close anatomical and morphological similarity, but in the absence of any reproductive
parts in the new specimens, I feel it would be unwise to conclude that the two sets of
remains belong to the same plant.

The axes were at least 10-0 cm. long and 2-0-3-0 mm. wide. They branched dichoto-
mously and were slightly flexuous. One axis bore a small protuberance 2-0 mm. long
and T5 mm. wide (PI. 131, fig. 1) comparable to the small lateral branch in Krithodeo-
phyton. The surfaces of the axes were striated. Film pulls showed thick longitudinal
walls similar to those described above and sometimes smaller fusiform cells were seen
(PI. 131, fig. 4).

Transverse and longitudinal sections through petrified axes showed a circular pro-
tostele composed predominantly of scalariform tracheids (PI. 131, figs. 2 and 5). Some
reticulate pitting was seen. The widest tracheids were found to the outside of the xylem.
The cells at the centre were either not preserved or very small and crushed (PI. 131, fig. 5).
Immediately outside the xylem was a narrow band of squashed thin-walled cells. One or
two layers of thick-walled cells, circular to angular in cross-section, with an average
diameter of 49 p, formed the outermost layer of the axis. The outer walls of these cells
were eroded away so that the radial walls project into the matrix, giving the axis a hairy
outline (PI. 131, fig. 6) which is also seen in Krithodeophyton. In longitudinal sections,
the cortical cells had tapering ends (PI. 131, fig. 3). They were, on average, 350 p long.

Stages in the division of the xylem below a branching point are illustrated in Plate
131, figs. 5-10.

Description of fructifications. The sporangia were aggregated into terminal spikes. A few
millimetres below the fertile region an axis dichotomised and this was followed by a
further dichotomy immediately below or within the base of the fructification (PI. 130,
fig. 4). Isolated fructifications were common in the matrix (PI. 130, fig. 1). The spikes
were at least 1-3 cm. long and, on average, were 2-75 mm. wide. The majority were
incomplete at the apex and 6-0-8-0 mm. long. At the base they were parallel-sided, but
there was a gradual decrease in width in the distal parts and the apices were rounded
(PI. 132, fig. 4). No organization was apparent in the distal parts. The sporangia were
oval in outline, 1 -25-1 -5 mm. long and 0-8-1 -0 mm. wide (PI. 1 32, fig. 2). They were sessile
on a central axis (0-3 mm. diameter) which was often obscured by the sporangia them-
selves. The sporangia appeared to be arranged in two rows one on either side of the axis,
each row containing at least eleven sporangia. It is unlikely that this arrangement was
produced by compression of an originally spiral or whorled organisation as the spor-
angia are quite distinct. At the base of the spike the long axes of the sporangia were
orientated at right-angles to the central axis, in the distal part they were directed toward
the apex. Alternating with the sporangia and extending beyond them out into the matrix
were thin, possibly spine-like appendages up to 2-5 mm. long (PI. 132, figs. 3 and 5).

EXPLANATION OF PLATE 131

Figs. 1-10. Krithodeophyton croftii sp. nov. 1, Branching sterile axes from Brecon Beacons Quarry
(small lateral projection indicated by arrow), x 1 (DE 32/1). 2, Section through petrified axis,
L.S. xylem strand composed of scalariform tracheids, x 108 (32-4-33). 3, As last, L.S. outer cortical
cells, X 108 (32-4-35). 4, Film pull of surface of axis showing epidermal cells, X 80 (DE32. P77).
5-10, Series of sections of petrified axis showing division of xylem at a branching point, x51
(Series 32-4).

Palaeontology, Vol. 11

PLATE 131

EDWARDS, Lower Old Red Sandstone plants

DIANNE EDWARDS: A NEW PLANT FROM SOUTH WALES

687

These arose at right angles to the fructification axis and sometimes curved downwards
at their tips. The exact relationship between sporangium and lateral appendage or bract
could not be determined with certainty as the structures were very heavily carbonised.
In some cases, the base of a sterile bract extended down the central axis partially covering
the sporangium below, giving the impression that the sporangium and the bract above
it formed a single unit (PI. 132, fig. 1). Plate 132, fig. 2 shows a fructification, where
sporangia only are visible, while in Plate 132, fig. 6 only bracts are present. Where
sporangia and bracts were slightly superimposed the sporangia appear beaked. At the
moment it seems very likely that the appendages were not attached to the sporangia,
but that both were borne separately on a central axis.

Description of spores. Maceration and film pull preparations showed spores adhering
to the sporangium wall. The cellular structure of the walls was not preserved, but some
very fragmentary remains indicated that part of the wall was composed of isodiametric
cells, around 27 p, in diameter. Imprints left by the spores could frequently be seen in the
sporangium wall (PI. 132, fig. 7). A small number of spores were recovered by dissolving
the film pull bearing a sporangium (PI. 132, figs. 8-10). They were all approximately the
same size, with an average diameter of 58 /x (range: 55-68 /x). The spores were almost
circular in outline. The position of the trilete was usually difficult to establish because
of heavy folding, but in one case a simple Y-mark was observed. The length of the trilete
was almost equal to the spore radius. The wall ornament was made up of rather variable
elements ranging from coni to spinae, which were 1 /x or less high. Narrower parallel-
sided elements with truncated tips occasionally reached a height of 3 /x. There was a
well-developed contact area, where the ornament was much reduced or absent. These
spores are assignable to the dispersed spore genus Apiculaliretusispora ( sensu Streel
1964) and are comparable to the species A. brandtii (Streel 1964).

In conclusion it must be emphasized that no anatomy has been detected in the axes
attached to the fertile regions. The sterile and fertile parts are thought to belong to the
same plant because they are always found in close association, no other plant being
present, and where fructifications terminate relatively long axes, the latter have the same
dimensions, surface features, and branching angles as do the sterile axes.

Discussion. Plants with a similar arrangement of sporangia in terminal spikes and, in
some cases, with sterile and fertile appendages intermixed, are known from other
Devonian deposits. These include Barinophyton, Protobarinophyton, Pectinophyton, and
the possible reproductive parts of Enigmophyton. Hoeg (1942) described some fertile
axes, which he considered were probably attached to Enigmophyton superbum. These
plants came from the upper Middle Devonian of Spitzbergen. As in the Welsh plant,
the sporangia, which alternated with sporophylls, were arranged in two rows, with
dichotomous branching occurring both below and within the base of the fructification.
Hoeg, however, thought that the two rows were produced by compression of an origin-
ally verticillate or spiral arrangement of appendages. He considered the sporangia to be
borne proximally on the upper surfaces of the sterile bracts. Unlike Krithodeophyton,
this plant was heterosporous, and Vigran (1964), described the spores in detail. She
identified the microspores as Phyllothecotriletes microgranulcitus (average diameter 72 /x)
and called the megaspores Enigmospora simplex (200-70 p diameter). Although the
microspores are comparable in size to those spores isolated from Krithodeophyton , the

688

PALAEONTOLOGY, VOLUME 11

ornament is quite different. Nothing resembling the fan-shaped leaves of Enigmophyton
superbum has been found at the Brecon Beacons Quarry.

Earlier, Hoeg (1935) had described a Middle Devonian plant from Norway, in which
naked axes, with mainly pseudomonopodial branching had fructifications terminating
the ‘lateral’ branches. The sporangia, or perhaps the organs bearing the sporangia, were
arranged in two rows, both borne on the lower surface of the fructification axis. Hoeg
called this plant Pectinophyton norvegicum and Ananiev (1957), described another species
P. bipectinatum, where the axis of the fructification bore two rows of long, lateral
appendages, twisted towards the axis. On the inside of the resulting coils were attached
the sporangia. The axis had a central strand of annular tracheids. Whereas Petrosyan,
in Lepekhina, Petrosyan, and Radchenko (1962) considered the above two species to
be identical, Hoeg (1967) maintained that they were two distinct species on the basis of
differences in sporangial shape and arrangement. Not enough is known of Hoeg’s plant
to make a satisfactory comparison with Krithodeophyton , but P. bipectinatum with its
apparently complex organisation in the sporangial region, is quite unlike the Welsh
plant.

The genus Barinophyton was erected by White (1905), when he re-investigated some
plants from the Upper Devonian of Perry, Maine. He distinguished three species,
Barinophyton richardsonii ( Lepidostrobus richardsonii Dawson 1861, Lycopodites richard-
sonii Dawson 1863), B. obscurum ( Pecopteris (?) obscurum Dun 1898), and B. perryanum.
The most completely described species in the literature are B. citrulliforme from the
Upper Devonian of New York (Arnold 1939) and B. richardsonii, re-described by Pettitt
(1965). Arnold showed that branching in the sterile region was pseudomonopodial but
that the fructifications, much larger than in Krithodeophyton, terminated lateral branches.
In both Barinophyton species the fructification axis bore on its dorsal surface two rows
of appendages between which were found the sporangia. They were unlike Krithodeo-
phyton in that the appendages did not extend beyond the sporangia and were discoidal
in shape. In addition, in Pettitt’s re-construction two sporangia occurred between suc-
cessive bracts in each row. Where the sporangia were borne in two rows, one on either
side of the fertile axis, i.e. the normal Krithodeophyton arrangement, Arnold thought the
arrangement was produced during fossilization. Both species were heterosporous.
A Middle Devonian species, B. sibiricum (Lepekhina et al. 1962) differed from the Welsh
plant in having long, lax, pendant spikes of sporangia.

In 1954, Ananiev described B. obrutschevii from the Lower Devonian of the U.S.S.R.
This differed from previously described species of Barinophyton in that branching was
dichotomous not pseudomonopodial. On the basis of this difference alone, he transferred
the plant to a new genus Protobarinophyton obrutschevii (Ananiev 1957). Although in
1954, Ananiev had briefly mentioned sporophylls associated with the sporangia, in later
papers by him and other authors no mention is made of them. Certainly Protobarinophyton
obrutschevii f. mucronata (originally Distichophytum mucronatum Ananiev) transferred

EXPLANATION OF PLATE 132

Figs. 1-6. Fructifications of Krithodeophyton croftii sp. nov.; s, sporangium; b , bract; d, branching
within the base of the fertile region. 1, x 5 (V26579). 2 and 6, x 2-3 (V26578). 3-5, x 5 (V26578). 7,
Sporangium wall with spores attached, x 500 (V52146, BMM 12). 8-10, Spores isolated from sporan-
gia. 8, x 500 (V52146, BMS 1). 9 and 10, x 1000 (V52146, BMS 1).

Palaeontology, Vol. 11

PLATE 132

EDWARDS, Lower Old Red Sandstone plants

DIANNE EDWARDS: A NEW PLANT FROM SOUTH WALES

689

to the genus by Lepekhina et a/. (1962) had no bracts in the fertile region. Ananiev
(1963) stated that the sporangia of P. obrutschevii contained spores of one size only
(70 p diameter).

Krithodeophyton and Protobarinophyton resemble each other in many ways. Indeed,
considering such characters as their similar age, the aggregation of oval, sessile sporangia
into terminal spikes, homospory and dichotomous branching in the fertile regions, they
might be included in the same genus. However, until more is known of the detailed
organization of the spikes (e.g. whether or not bracts are present), about the spores and
the anatomy of the axes of Protobarinophyton , I feel justified in erecting a new genus for
the Welsh plant. Differences include the absence from Krithodeophyton of U- and H-
branching, the absence of dehiscence lines on the sporangia and the presence of scalari-
form not annular tracheids in the protostele. In addition, the spikes in Krithodeophyton
were much smaller, more compact and had distinctly protruding bracts. The specimen
illustrated in Plate 132, fig. 2, in which no bracts are visible, most closely resembles the
Russian plant. It is possible that this specimen represents a late stage in the maturation
of a spike, where the sporangia were large and completely obliterated the bracts. This
explanation could be applied to the apparently bractless condition seen in Protobarino-
phyton. Variation within the species has been indicated by Lepekhina et al., who dis-
tinguished a P. obrutschevii f. minuta from P. obrutschevii f. typica , but little is known
of the variation in fructification morphology within a single population.

In conclusion it is stressed that not enough is known of the details of anatomy and
morphology of any of the plants discussed above. It may be that ultimately they will all
be placed in the same genus. Indeed Arnold believed that this was true for Pectinophyton
and Barinophyton. On the information available at the moment, I propose to place this
relatively completely described Lower Devonian plant in the new genus Krithodeophyton
(Derivation: Krithodes — like barley).

Classification. Ananiev (1963) included Protobarinophyton, with Zosterophyllwv and
Bucheria in the Zosterophyllaceae. My own observations on the anatomy of Z. llano-
veranum show that Krithodeophyton is not related to this species, because the latter had
an elliptical exarch protostele, and therefore should not be placed in the same family.
In his contribution to the second volume of the Traite de Paleobotanique (ed. Boureau
1967), Hoeg included the Barinophytales as an order, Incertae Sedis, containing
two families, Barinophytaceae and Barrandeinaceae. The former included, in addition
to Barinostrobus, all the genera mentioned above with the exception of Enigmophyton.
It is suggested that Krithodeophyton should also be placed in this family.

Acknowledgements. I thank Dr. K. R. Sporne for his help in preparing this paper, and Dr. J. B.
Richardson and Dr. F. A. Hibbert for their assistance in identifying the spores. I am indebted to the
Keeper of Palaeontology of the British Museum (Natural History) for the loan of specimens. This
work formed part of a Ph.D. thesis submitted to the University of Cambridge and was carried out
during the tenure of a N.A.T.O. Research Studentship.

REFERENCES

ananiev, a. r. 1954. On the Lower Devonian flora of the south-east part of Western Siberia. Voprosy
Geologii Azii, 1, 287-324, pi. 1-5. [Russian Translation Service, National Lending Library.]

— 1957. New fossil plants from the Lower Devonian deposits of the village of Torgashino, south-
eastern zone of Western Siberia. Bot. Zh. Moscow, 42, 691-702, pi. 1-3. [Ibid.]

690 PALAEONTOLOGY, VOLUME 11

ananiev, a. r. 1963. Psilopsida in V. A. Vakhrameev (ed.) Osnovy Paleontologii, 14, 315-43, pi. 1-8
[In Russian.]

Arnold, c. a. 1939. Observations on fossil plants from the Devonian of Eastern North America, IV.
Plant remains from the Catskill Delta deposits of Northern Pennsylvania and Southern New York.
Contr. Mus. Geol. Univ. Mich. Ann Arbor, 5, 271-313, pi. 1-10.
beck, c. b. 1955. A technique for obtaining polished surfaces of sections of pyritized plant fossils.
Bull. Ton. hot. Cl. 82, 286-91.

croft, w. n. 1953. Breconian: a stage name in the Old Red Sandstone. Geol. Mag. 90, 429-32.

and lang, w. h. 1942. The Lower Devonian flora of the Senni Beds of Monmouthshire and

Breconshire. Phil. Trans. R. Soc. Loud. B221, 131-63, pi. 9-11.
dawson, j. w. 1861. On the pre-Carboniferous flora of New Brunswick, Maine and Eastern Canada.
Can. Naturalist, 6, 161-80.

1863. Further observations on the Devonian plants of Maine, Gaspe and New York. Q. Jl geol.

Soc. Loud. 19, 458-69, pi. 17-19.

dun, w. s. 1898. On the occurrence of Devonian plant-bearing beds on the Genoa River, County of
Auckland. Rec. geol. Surv. N.S. Wales, 5, 117-21, pi. 10, 11.
edwards, d. 1967. An investigation of certain Devonian plants. Ph.D. thesis. University of Cam-
bridge.

and banks, H. p. 1965. Branching in Goss/ingia breconensis. Am. J. Bot. 52, 536.

heard, a. 1927. On Old Red Sandstone plants showing structure from Brecon (South Wales). O. Jl
geol. Soc. Load. 85, 195-209, pi. 13-15.

1939. Further notes on Lower Devonian plants from South Wales. Ibid. 95, 223-9, pi. 12, 13.

hoeg, o. a. 1935. Further contributions to the Middle Devonian flora of Norway. Norsk geol. Tidsskr.
B15, 1-18, pi. 1-8.

1942. The Downtonian and Devonian flora of Spitzbergen. Skr. Svalb. og Ishavet 83, 1-228,

pi. 1-62.

1967. Psilophyta. In Traite de Paleobotanique, vol. 2. Boureau ed. Paris. [In French.]

krausel, r. and weyland, h. 1961. Uber Psilophyton robustius Dawson. Palaeontographica 108B,
11-21, pi. 4-5.

leclercq, s. and noel r. 1953. Plastic: a suitable embedding substance for petrographic study of coal
and fossil plants. Phytomorphology, 3, 222-3.

lepekhina, v. G., Petrosyan, n. m., and Radchenko, g. p. 1962. Main Devonian plants of the Altay-
Sayan Mountain region. Trudy vses. naucho-issled. geol. Inst. 70, 61-189, pi. 1-24. [Russian Transla-
tion Service, National Lending Library.]

pettitt, j. m. 1965. Two heterosporous plants from the Upper Devonian of North America. Bull.
Brit. Mus. ( nat . hist.) Geol. 10, 83-92, pi. 1, 2.

streel, m. 1964. Une association de spores du Givetien inferieur de la Vesdre, a Goe (Belgique). Annls.
Soc. geol. Belg. 87 B1-B30, pi. 1, 2.

vigran, j. o. 1964. Spores from Devonian deposits, Mimerdalen, Spitzbergen. Skr. norsk. Polarinst.
132, 1-33, pi. 1-6.

white, d. 1905. In G. O. Smith and D. White. The geology of the Perry Basin in South Eastern Maine.
Prof. Pap. U.S. geol. Surv., Washington, 35, 9-92, pi. 2-6.

D. EDWARDS

Botany School
Downing Streel

Typescript received 1 March 1968 Cambridge

DENCKM ANNITES (TRILOBITA) FROM THE
SILURIAN OF NEW SOUTH WALES

by L. SHERWIN

Abstract. The trilobite Denckmannites Wedekind is redefined and a new species, Denckmannites rutherfordi, is
described from Siluro-Devonian shales near Orange in the Central West District of New South Wales.

The genus Denckmannites has hitherto been recorded only from the Siluro-Devonian
of Central Europe and the Devonian of Morocco (Richter, in Moore 1959, p. 0 467).
The species described in this paper, Denckmannites rutherfordi , was found in the Wallace
Shale (Stevens and Packham 1952) at a locality about 18 miles west of Orange in the
Central West district of New South Wales. The two bands bearing the trilobites were
separated by a thickness of 100 ft. of apparently unfossiliferous olive-green shales.
Some uncinate monograptids accompanying the trilobites in both bands indicated either
a late Silurian, or less probably, an early Devonian age. Other fossils in the lower band
include Dictyonema sp., Encrinurus mitchelli Foerste, an indeterminate odontopleurid
trilobite, orthoconic nautiloids, and small orthid and linguloid brachiopods. Apart from
Monograptus sp., the only associates of Denckmannites in the upper band are Dictyonema
sp. and Encrinurus sp.

The morphological terms are as in the Treatise, Part O.

SYSTEMATIC PALAEONTOLOGY

Order phacopida Salter 1864
Genus denckmannites Wedekind 1914

1931 Phacopidella ( Denckmannites ) Wedekind; Richter and Richter, p. 143 (cum syn).

1959 Denckmannites Wedekind; Richter, p. 0 467.

Type species. Phacops volborthi Barrande 1852; SD Vogdes 1925.

Emended diagnosis. Cephalon with broad border and shallow border furrow along
posterior edge, both continued antero-laterally as far as axial furrows; border furrow
joins axial furrow in front of eye; in front of glabella border is reduced or absent. Mar-
gin of cephalon sharp, upper surface meeting doublure at acute angle; 3p and 2p
glabellar furrows vestigial, equal in depth, unconnected with axial furrows; lp furrows
initially as deep as axial furrows, becoming much shallower axially. Intercalating ring
tripartite, distinct from glabella. Eyes small, of cryptothalmus pattern. Entire posterior
section of facial suture with forwardly directed convex curvature; genal angles rounded.
Vincular furrow crenulate, developed only on postero-lateral doublure. Pygidium large,
rounded, with narrow border and smooth surface.

Remarks. In 1914 Wedekind erected the subgenus Glockeria ( Denckmannites ) without
selecting a type, referring the definition instead to the ‘group of Phacops volborthi
Barrande’ which also included P. miser Barrande and P.fugitivus Barrande. P. volborthi

[Palaeontology, Vol. 11, Part 5, 1968, pp. 691-6, pi. 133.]

C 6055 Z Z

692

PALAEONTOLOGY, VOLUME 11

was subsequently designated as the type by Vogdes in 1925; Richter and Richter (1931)
confirmed this action by applying Art. 30, la of the then current Rules for Zoological
Nomenclature. In defining G. ( Denckmannites ) Wedekind stated that the cephalon is
totally surrounded by a border. However, this feature is found only in P. volborthi and
possibly P. fugitivus, but not in P. miser, although Wedekind included them within his
subgeneric conception. In addition, he stated that the Ip glabellar furrows did not
coalesce, although they are shown connected in figures of P. volborthi and P. miser by
Barrande (1852). Vogdes (1925) designated P. volborthi as the type for Phacopidella
( Denckmannites ), but retained Wedekind’s diagnosis unaltered. Richter and Richter
(1931), in revising the subgenus, omitted any reference to the cephalic border and lp
glabellar furrows. They distinguished Phacopidella ( Denckmannites ) from P. ( Phacopi-
della) by its ’uniform glabellar surface anterior to the lp furrows as in Phacops (i.e., an
undivided glabellar surface with only mere vestiges of 2p and 3p furrows), eyes of a
cryptothalmus pattern, and a tripartite intercalating ring’ (p. 143, in German). Richter
(in Moore 1959, p. O 467) raised Denckmannites to full generic status, at the same time
modifying the 1931 diagnosis by referring to the intercalary furrow as tripartite instead
of the intercalating ring. The genal angles were described as truncate, while the entire
posterior section of the facial suture was specified as concave; i.e., with the concave side
directed posteriorly. Alberti (1966) designated P. (£>.) micromma (A. Roemer) as the
type species for Phacopidella ( Struveaspis ) thus restricting the scope of the genus Denck-
mannites ( sensu Richter 1959). This new subgenus is characterized by fusion of the
reduced central lobe of a tripartite intercalating ring to the base of the glabella by medial
reduction of the lp furrows. For the sake of conformity in the two following lists,
Struveaspis is regarded as being of the same generic status as Denckmannites.

Species retained in genus Denckmannites :

Denckmannites volborthi (Barrande) 1852
Denckmannites miser (Barrande) 1852
Denckmannites rutherfordi sp. nov.

Species transferred to Struveaspis

Struveaspis micromma (A. Roemer) 1852
Struveaspis fugitivus (Barrande) 1872
Struveaspis thuringica (Kegel) 1932
Struveaspis prantli (Ruzicka) 1946
Struveaspis n.sp. A. Alberti 1966

Denckmannites rutherfordi sp. nov.

Plate 133

Source of name. After the Rutherford family, on whose property the fossils were found.

Material. Six more or less complete exuviae preserved as internal moulds with fragments of the exo-
skeleton adhering; USGD (University of Sydney Geology Department) 8914, 8916, 8934, and MMF
(Geological and Mining Museum, Sydney) 14467. The external counterparts of USGD 8929 and 8934
have also been preserved. The only deformation is that caused by burial pressure, with crushing of the
glabella. In addition, abundant imperfect specimens were also collected from the same locality.

Holotype. USGD 8927 a, b.

Paratypes. USGD 8929, 8934, MMF 14467.

Locality. Portion 216, Parish Boree Nyrang, County Ashburnham.

Sydney University Geology Department Locality catalogue MN/II/27.

SHERWIN: DENCKMANNITES FROM NEW SOUTH WALES

693

Diagnosis. Denckmannites with very small eyes (about 20 lenses in each), deep trans-
glabellar lp furrows becoming shallow axially, and exceedingly narrow anterior cephalic
border; pygidium with 6-7 discrete axial rings.

Description. The carapace possesses a long oval outline and, except for a strongly
convex axis, has only moderate relief. The cephalon has a micro-granulose surface and
a semicircular outline, with rounded genal angles which do not project noticeably
beyond its posterior edge. The interior of the cephalon is pitted. The wide border is
defined by a shallow border furrow, and is continuous along the posterior and lateral

A B

text-fig. 1. A. Reconstruction of cephalon showing inferred position of facial
sutures, b. Reconstruction of ventral surface of cephalon, showing approximate

shape of doublure.

edges of the cephalon, maintaining a more or less uniform width. In front of the glabella
the border narrows considerably; none of the specimens was sufficiently well preserved
to determine if it disappears altogether. The border furrow runs parallel to the margin
as far as the anterior corner of the cheeks, where it swings sharply into the axial furrow.
Anterior to the lp furrows the axial furrows are straight, and are generally deeper than
the border furrow. Near the cephalic margin they fade, and outline the glabellar curve
to the edge of the cephalon before disappearing entirely.

The glabella is large, pentagonal, and slopes forward gently to meet the doublure in
an acute angle. The 2p and 3p furrows are weak, and do not meet the axial furrows. In
different specimens the 2p furrows are represented by straight or curved furrows, the
curve always being directed anteriorly. The posterior branches of the 3p furrows are
generally more strongly curved than the 2p furrows; the 3p anterior branch is straighter
and inclined at about 30 degrees to the axial furrows. The lp furrows cut deeply into the
glabella for about one-third of their length, where they divide into two branches. The
anterior branches swing forward decreasing in depth, being joined axially by a shallow
depression. The posterior branch cuts the intercalating ring and joins the occipital
furrow. The intercalating ring is thus trisected into a broad medium lobe and two ovoid
lateral lobes. This division is reflected by the vaguely tripartite occipital ring. The cheeks
have a slightly lower relief than the glabella; approximately triangular in shape, they
support very small eyes at their anterior corners. Preservation does not indicate whether
an anterior section of the facial suture exists or not; the posterior section has a forwardly
directed curvature throughout, and intersects the cephalic margin just in front of the
genal angles.

694

PALAEONTOLOGY, VOLUME 11

The eyes are oval in shape, supported on a reniform palpebral lobe. Preservation does
not allow an accurate count, but each eye probably has between 1 9 and 2 1 lenses arranged
in six vertical rows as follows (front-rear): 2, (?3), 3, 3, (?4), 4, 4, 3. In the holotype
(USGD 8927) a portion of the exoskeleton partially covering one of the eyes shows no
expression of the underlying lenses, unlike other species even in the same genus where
the lenses are quite prominent surface features. Unless the exoskeleton in this region was
transparent, it seems unlikely that the eyes could have functioned as light sensitive
organs.

The hypostome is triangular in shape with a slightly convex anterior margin. The
anterior wings are particularly well developed, with appreciable thickening at their tips.
On the ventral surface there is a small longitudinal ridge which runs from the lateral
notch, fading at the anterior margin: on the dorsal surface the ridge is expressed as a
shallow furrow. The border is continuous posterior to the lateral notches, and is broadest
along the posterior edge. Any other details of the posterior border were obscured in all
specimens by either cephalon or thorax. The median body is subcircular, apparently
undivided, and strongly convex. No maculae were observed. The entire surface of the
hypostome is finely granulose.

The cephalic doublure is crescentic in shape, broadest beneath the glabella, with a
paraboloid posterior margin which commences near the genal angles and passes just
behind the eyes. The crenulate vincular furrow is widest near the genal angles, tapers
forward, and disappears near the axial furrow. The doublure on small specimens seems
comparatively rigid, as in such cases the posterior portion of the glabella, and to a lesser
extent that of the cheeks, is crushed along an arc corresponding to the portion of the
cephalon unsupported by the doublure. In large specimens, crushing of the cephalon is
quite independent of the doublure. As regards ornament, the doublure seems comparable
to the dorsal cephalic surface.

The thorax is virtually parallel-sided, showing an appreciable taper in only the last
three segments. The axial rings are strongly convex, and have a tripartite division into
a central lobe, flanked by smaller, globose, lateral lobes. The central lobe carries a par-
ticularly well-formed articulating half-ring, separated from it by a deep but broad
transverse furrow. At the junction of central and lateral lobes, the extension of this

EXPLANATION OF PLATE 133
Denckmannites rutherfordi sp. nov.

Figs. 1 , 2, 5, 8, 9. Holotype USGD 8927. 1 , composite internal mould of cephalon, showing eye covered
by exoskeleton which gives no indication of underlying lenses, X 5. 2, Composite internal mould of
cephalon, X 1-5. 5, hypostome, x 3. 8, Composite external mould of cephalon, and internal mould
of thorax and pygidium. Short hooks are visible on the fragments of exoskeleton adhering to the
pleural extremities. 9, Silastic cast of external mould and hypostome, X 1-5. The anterior granulation
is caused by pitting on the interior surface of the exoskeleton.

Fig. 3, Paratype MMF 14467. Internal mould exhibiting phacopid or Salterian mode of moulting,
x 2.

Fig. 4, Paratype USGD 8934. Latex cast of external mould, showing border and interpleural furrows on
pygidium, xl-5.

Figs. 6, 7, Paratype USGD 8929a, b. 6, Internal mould, showing crushing of portion of cephalon unsup-
ported by doublure, X 3. 7, Interior of dorsal carapace showing apodemes, x 3. An attempt to make
a latex cast resulted in damage to the exoskeleton.

Palaeontology, Vol. 11

PLATE 133

SHERWIN, Denckmannites

SHERWIN: DENCKMANNITES FROM NEW SOUTH WALES

695

furrow projects downward to form a robust apodeme. These apodemes are flange shaped
and taper downwards, with a slight expansion at their tips. The pleurae are more or less
uniform in width throughout their length, with deep pleural furrows extending from
axial furrow to articulating facet. These facets are large, extending about one half the
length of the leading edge and a third the length of the posterior edge of the segments.
The first five segments have rounded extremities; the remaining six show progressive
development of a short spine or hook, possibly a feature associated with enrollment.

The pygidium has a very rounded outline, marked by a flat border. On internal moulds
this border is exaggerated by the imprint of the doublure which extends about one-third
the distance across the pleural region. Articulating facets occupy about half the length
of the leading ribs. The axis contains six discrete axial rings, a seventh vaguely discernible
in some specimens, followed by a fused terminal piece. The first ring is partially obscured
by the last thoracic segment. Both the first and second rings are transversely tripartite,
though the division is much less pronounced than in the thoracic axial rings. Each
pleural field bears five rather flattened ribs, separated by straight and narrow pleural
furrows. Interpleural furrows are seldom observed except on external mounds, and are
parallel to the pleural furrows. The surface is free of ornament.

Dimensions (in mm.). Holotype USGD 8927:

Width

Length

Cephalon

22

12

Thorax

24

20

Pygidium

17

9

Remarks. The bohemian species, Denckmannites miser (Barrande), most closely resembles
D. rutherfordi, but has much larger eyes, and the 3p glabellar furrows, as shown in
Barrande’s illustrations, appears to be whole, not bipartite. D. volborthi (Barrande) has
eyes of about the same size, with a similar number of lenses in each, as D. rutherfordi ,
but its cephalon is wholly surrounded by a border, while the pygidium has almost
twice as many axial segments as D. rutherfordi.

Acknowledgements. I wish to thank Dr. J. W. Pickett of the Geological and Mining Museum, Sydney
and Dr. B. D. Webby of Sydney University for a critical review of this paper; Dr. K. S. W. Campbell
of the Australian National University, Canberra, for much helpful discussion and criticism; Dr. G. H.
Packham of Sydney University for forwarding useful information concerning the stratigraphic position
of this species. Permission to publish this paper was given by the Under-Secretary, New South Wales
Department of Mines.

REFERENCES

alberti, g. k. b. 1966. Uber einige neue Trilobiten aus dem Silurium und Devon, besonders von
Marokko. Senck. leth. 47, 1 1 1-21, pi. 4.

barrande, J. 1852. Systeme Silurien du Centre de la Boheme: Iere Partie, Crustaces, Trilobites. 1,
935 pp., 51 pi. Prague-Paris.

richter, r. 1959. In R. C. Moore (ed.), Treatise on invertebrate paleontology. Part O, Arthropoda 1,
Lawrence, Kansas.

and richter, E. 1931. Unterlagen zum Fossilium Catalogus, Trilobitae. V. Senckenbergiana, 13,

140-6.

stevens, N. c. and packham, g. h. 1952. Graptolite zones and associated stratigraphy at Four Mile
Creek, S.W. of Orange, N.S.W. /. Proc. R. Soc. N.S.W. 86, 95-8.

696

PALAEONTOLOGY, VOLUME 11

vogdes, a. w. 1925. Palaeozoic Crustacea. Pt. 11. A list of the genera and subgenera of the Trilobita.
San Diego Soc. Nat. Hist. Trans. 3, 87-115.

wedekind, r. 1914. Palaontologische Beitrage zur Geologie des Kellerwaldes. Abb. preuss. geol.
Landesanst. 69, 1-84, pi. 1-5, figs. 1-26.

L. SHERWIN

Geological and Mining Museum
36-64 George Street North
Sydney, New South Wales, 2000

Typescript received from author 18 April 1968 Australia

MORPHOLOGY AND FUNCTION OF
DICHOPORITE PORE-STRUCTURES
IN CYSTOIDS

by C. R. C. PAUL

Abstract. The cystoids, a diverse and artificial group of extinct Palaeozoic echinoderms, are characterized by
the possession of pore-structures developed in the thecal wall and composed of thecal canals which open in thecal
pores. Thecal canals may occur singly to form dipores, or in sets perpendicular to plate sutures to form rhombs.
Five basic types of pore-structure are recognised: pectinirhombs, cryptorhombs, humatirhombs, humatipores, and
diplopores. The first two are composed of dichopores, thecal canals which connect external pores and through
which sea-water flowed in life. The dichopores of pectinirhombs open in slits; those of cryptorhombs open in
pores, one of which is sieve-like. Pectinirhombs and cryptorhombs are characteristic of, and confined to, the
super-families Glyptocystitida and Hemicosmitida respectively; humatirhombs characterize the superfamily
Caryocystitida. Five types of pectinirhombs and one type of cryptorhomb are recognized.

Pectinirhombs and cryptorhombs agree closely with the paradigm of an exchange system and were respiratory
structures. Evolution of pectinirhombs proceeded from less to more efficient types independently in all families
of Glyptocystitida. The three rhombiferan superfamilies probably acquired their rhombs independently.
Rhombifera and Diploporita are regarded as separate classes. The former contains two Orders: Dichoporita
and Fistuliporita.

The cystoids, an extinct group of Palaeozoic echinoderms, are characterized, inter alia,
by the possession of pore-structures in the calcareous wall of the theca. Although
various authors from Muller (1854) to Regnell (1945) and Kesling (1963) have em-
phasized the importance of these pore-structures in cystoid morphology and taxonomy,
they have not been the subject of complete morphological or functional analyses. Stain-
brook (1941) and Sinclair (1948) considered the function of pectinirhombs in one genus
each and Delpey (1942) the function of pectinirhombs and cryptorhombs. The last-
named author suggested that pectinirhombs and cryptorhombs were balancing organs
but most authors have accepted a respiratory function for all pore-structures. Function-
ally all cystoid pore-structures fall into two groups: dichoporite and non-dichoporite.
This paper deals with the dichoporite pore-structures (pectinirhombs and crypto-
rhombs) after a brief review of the morphology of the five basic types of pore-structure
recognized. Function in non-dichoporite pore-structures will be the subject of a future
paper.

The cystoids of this account correspond to the class Cystoidea of Kesling (1963) but
only to the subclass Hydrophoridea of Regnell (1945). The hydrospires of blastoids and
epispires of certain eocrinoids and crinoids are not considered nor are the peculiar
pore-structures of paracrinoids like Comarocystites Billings. Knowledge of the mor-
phology of cystoid pore-structures has grown piecemeal, virtually every researcher add-
ing some contribution. One result of this is a rather imprecise terminology often related
to ill-defined concepts. In reviewing the morphology of pore-structures an attempt has
been made to introduce a systematic terminology and to define the concepts upon which
this terminology is based. Previous usage has been followed as far as possible. Through-
out this account the classification is that of Kesling (1963) unless otherwise stated.

[Palaeontology, Vol. 11, Part 5, 1968, pp. 697-730, pis. 134^40.]

698

PALAEONTOLOGY, VOLUME 11

Acknowledgements. The author is grateful to the following for the loan of, or access to, specimens in
their care: Dr. R. P. S. Jefferies and Mr. H. G. Owen, British Museum, Natural History (BMNH);
Dr. I. Strachan and Dr. G. R. Coope, Birmingham University, Department of Geology (BU); Dr. T. E.
Bolton and Dr. G. W. Sinclair, Canadian Geological Survey (CGS); Dr. R. V. Melville and Dr. A.
Rushton, Geological Survey and Museum, London (GSM); Dr. W. D. I. Rolfe, Hunterian Museum,
Glasgow (HM); Dr. J. C. Harper, Liverpool University, Department of Geology (LU); Dr. H. Mutvei,
Naturhistoriska Riksmuseet, Stockholm (RM); Dr. C. D. Waterston, Royal Scottish Museum, Edin-
burgh (RSM); Mr. A. G. Brighton, Sedgwick Museum, Cambridge (SM); Dr. K. E. Caster, University
of Cincinnati, Department of Geology (UC); Mr. H. L. Strimple, University of Iowa, Department of
Geology (UI); Dr. D. B. Macurda Jr., University of Michigan, Museum of Paleontology (UMMP);
Dr. P. M. Kier and Mr. T. Phelan, United States National Museum, Washington (USNM); and Dr.
Katherine G. Nelson, University of Wisconsin, Milwaukee, Department of Geology (UWM).

Parts of this work were completed during the tenure of a Harkness Scholarship and a National
Environmental Research Council research scholarship at the Sedgwick Museum, Cambridge; and a
post-doctoral fellowship at the Museum of Paleontology, University of Michigan (NSF GB-3366).
All three are gratefully acknowledged.

Mr. P. Minton, Imperial College, Department of Civil Engineering and Dr. R. P. S. Jefferies con-
tributed much useful discussion on the functioning of rhombs. Dr. Jefferies kindly reviewed the manu-
script and suggested several improvements.

MORPHOLOGY OF CYSTOID PORE-STRUCTURES
A. GENERAL

The term pore-structure embraces all the structures dealt with. All pore-structures are
composed of thecal canals and thecal pores. These have not always been distinguished
from each other in the past but Kesling (1963) noted that two separate concepts are
involved in the terms pore and canal.

Thecal pores are perforations in the surfaces of the thecal wall. They cannot have a
separate existence without the thecal canals to which they give rise, i.e. thecal pores bear
the same relationship to thecal canals that the two ends of a stick do to the stick. Pores
in the external surface of the thecal wall are external pores (text-fig. 1) and correspond-
ingly, internal pores (text-fig. 2) are developed in the internal surface of the thecal wall.
External pores may be slit-like, and are then referred to as slits (text-fig. 12), or they may
be either simple or sieve-like (text-figs. 3, 20). When simple a single thecal canal ter-
minates in a single pore, when sieve-like the canal divides near the external surface of the
thecal wall and opens in a cluster of small pores. Apparently internal pores are always
simple and circular in cross-section.

Thecal canals are tubular structures which connect , or terminate in, thecal pores.
Thecal canals are analogous to the entire stick plus its ends. The portions of thecal
canals which are approximately perpendicular to the surface of the thecal wall are termed
perpendicular canals (text-fig. 2) and those tangential to the surface, tangential canals
(text-fig. 2). All thecal canals are composed of one or more tangential canals which
connect a pair of perpendicular canals and they all commence and terminate in two
thecal pores, both of which lie in the same surface of the thecal wall (with the possible
exception of haplopores). Tangential canals may be incompletely calcified or uncalcified,
i.e. made of soft tissue only as in diplopores, in which case only the pair of perpendicular
canals is preserved in fossils. Thecal canals may be compound (text-fig. 4) when several
tangential canals connect a pair of perpendicular canals, or simple (text-figs. 1, 2) when
one tangential canal connects the perpendicular canals. Here thecal canals are considered

PAUL: DICHOPORITE PORE-STRUCTURES IN CYSTOIDS

699

as functional units rather than morphological units. Portions of canals composed entirely
of soft tissue, as is the case with external papulae, are considered integral parts of the
canal even though their former existence can only be inferred in fossils.

Previously two types of pore-structures have been recognized: ‘pore rhombs’ (or
more simply and accurately, rhombs ) and ‘diplopores’. The latter term has been used
in both a specialized and a generalized sense. The new term dipore is proposed to cover
all types of pore-structures within the Diploporita and diplopore is restricted to one type
of dipore following current usage.

1 2

EP S TC

text-figs. 1-4. Thecal pores and thecal canals.

1. Diagrammatic longitudinal section through a pectinirhomb dichopore to show external pores.

2. Diagrammatic longitudinal section through a simple fistulipore to show internal pores.

3. Diagrammatic longitudinal section through a cryptorhomb dichopore to show simple and sieve-
like external pores.

4. Diagrammatic plan view of a compound fistulipore with surface layer removed to show two
perpendicular canals connected by a pair of tangential canals.

CP, sieve pore; EP, external pore; IP, internal pore; SP, simple pore; PC, perpendicular canal; TC,
tangential canal; S, plate suture. In text-figs. 1-3 and all other figures unless otherwise stated, the
external surface of the thecal wall is towards the top of the figure.

Rhombs (fig. 5) may be defined as pore-structures composed of a set of thecal canals each
of which arises in one thecal plate, crosses a plate suture and terminates in an adjacent
plate , the whole set having a rhombic outline in plan view. Rhombs have their thecal canals
developed more or less perpendicular to the plate suture and were randomly orientated
with respect to other features in the theca since plate sutures occur in almost all possible
orientations. Rhombs are therefore orientated and described with respect to their sutures
(text-fig. 5).

The outline of a rhomb may be described as indicated in text-figs. 6-11. Rhombs may
be composed of two types of thecal canal. Jaekel (1899) proposed the term dichopore for
the canals (and their terminal pores) of the Rhombifera, Jaekel’s Dichoporita. Sub-
sequently dichopore has been restricted to the canals of pectinirhombs. A fundamental
difference exists between canals which terminate in internal pores and those with external
pores. No complete canal can have both. Here the term dichopore is restricted to rhomb
canals with external pores (text-figs. 1, 3) and the new term fistulipore (fistula = canal)
is proposed for rhomb canals with internal pores (text-figs. 2, 4). There are two types
of rhombs composed of dichopores and a third composed of hstulipores. Pectinirhombs
(text-figs. 15-19) are composed of dichopores which open in slits, whereas cryptorhombs

700 PALAEONTOLOGY, VOLUME 11

(text-fig. 20) are composed of dichopores which open in pores. Humatirhombs are rhombs
composed of fistulipores.

Dipores may be defined as pore-structures composed of a single thecal canal randomly
orientated with respect to the thecal plates. Dipores are composed of one type of thecai
canal, which resembles a fistulipore in connecting internal pores but differs in rarely
crossing plate sutures. There were two types of dipore. A diplopore is a dipore composed

text-fig. 5. Diagram of the principal features of a rhomb. Every rhomb is developed in two thecal
plates (1 and 2). A plate suture divides the rhomb into two half-rhombs and the width is measured along
it. Perpendicular to the suture lies the rhomb axis which divides the rhomb into two demi-rhombs and
along which the length is measured. Thecal canals adjacent to the axis are adaxial, those away from the
axis lateral and the outermost one or two are marginal. Positions and directions ( adsutural , adaxial,
etc.) are defined with respect to the plate suture and the rhomb axis.

of a simple thecal canal in which the tangential portion was uncalcified, i.e. the tangential
portion formed a papula in life and only the two perpendicular portions of the canal are
found fossil. Humatipores are dipores composed of compound thecal canals in which
all the tangential canals were fully calcified.

These five major types of pore-structure are recognized: pectinirhombs, crypto-
rhombs, humatirhombs, diplopores, and humatipores. The first two differ fundamentally
from the other three in that they are composed of thecal canals (dichopores) which connect
external pores. Hudson (1911, 1915) recognized this distinction and described dicho-
porite pore-structures as endothecal and non-dichoporite as exothecal. These general
terms may be applied to all types of pore-structures in primitive echinoderms.

A sixth type of pore-structure, haplopores, has been attributed to cystoids. Haplopores,

PAUL: DICHOPORITE PORE-STRUCTURES IN CYSTOIDS

701

as here understood, consist of a single, blind (in life), perpendicular canal which
opens in a single internal pore. No true haplopores have been observed in any cystoids
examined by the writer but this does not imply that none ever existed.

The first information on the structure of the thecal wall, which is intimately related to
the morphology of pore-structures, was given by Barrande (1887). Barrande maintained
that there were three layers to the thecal wall and that the outer and inner layers sealed

text-figs. 6-11. Outlines of rhombs.

Rhombs in which the length exceeds the width are compressed (6); the converse, depressed (7).
Rhombs with unequal half-rhombs are unequal (8); the converse, equal (6, 7). Rhombs with unequal
demi-rhombs are asymmetrical (9); the converse, symmetrical (6, 7, 8). These features occur in combina-
tion. A symmetrical, equal rhomb which is neither depressed nor compressed is ideal (10). Rhombs
with thecal canals oblique to the suture are oblique (11).

A, rhomb axis (an imaginary line); S, plate suture. Text-figs. 6-11 represent disjunct pectinirhombs
in plan view but the terms may be applied to all types of rhombs. Rhombs which are ‘depressed’ into
the theca may be termed sunken to avoid confusion with those with depressed outlines.

the thecal canals which were confined to the middle layer. Several subsequent authors
have described an epithecal layer sealing canals externally but the present writer has
found no evidence to support Barrande’s statement. Barrande described the surfaces
of his layers (as preserved in moulds in most cases) and not cross-sections. All sub-
sequent authors also described surfaces not sections. When the external surface was
without pores Barrande claimed an epithecal layer was present; when pores were visible
on the external surface he claimed the epitheca was eroded before preservation. How-
ever, external surfaces with and without pores may result from the type of canal
developed. A rhombiferan with dichopores will have pores in the external surface but
one with fistulipores will not unless the outer walls of the tangential canals are eroded.

702

PALAEONTOLOGY, VOLUME 11

A similar distinction may be made between diplopores, with tangential canals which
were of soft tissue and never preserved, and humatipores with fully calcified tangential
canals. The description of sealed dichopores in cryptorhombs may possibly result from
post-mortem calcification. However, it is more probably due to a misinterpretation of
sieve pores as a progressive sealing of the canals. Finally sealed canals do not necessarily
indicate the presence of a separate layer sealing them. All thin sections prepared by the
author lack any indication of separate layers sealing the canals but the thecal wall is
definitely not uniform in structure.

text-fig. 12. Diagram to illustrate the principal features and measurements of pectinirhomb dicho-
pores and slits. Dichopore and slit lengths and widths are measured in the same direction as rhomb
lengths and widths. In this figure and text-figs. 15-20 the external surface of the thecal wall faces the
top of the figures and part of the rhomb is removed to show internal features.

B. THE MORPHOLOGY OF PECTINIRHOMBS

Pectinirhombs are rhombs composed of dichopores which open externally in slits (text-fig.
12). The term pectinirhomb was first used by E. Forbes (1848, as 'pectinated rhomb’)
for the rhombs of the superfamily Glyptocystitida. The various features of pectinirhomb
morphology will now be described.

The pores (text-figs. 1, 12) of pectinirhombs are slits. The principal measurements of
slits and dichopores are shown in text-fig. 12. There may be a single conjunct slit along
the entire length of a dichopore (text-figs. 15, 17). Alternatively the dichopore may open
in a pair of disjunct slits, one at each end (text-figs. 16, 18). One disjunct slit is generally
longer than the other and all the longer slits occur in one half-rhomb, a fact first noted
by Stainbrook (1941, p. 93) for the pectinirhombs of Strobilocystites White. The slit
lengths are also approximately proportional to the length of the dichopores to which
they give rise and thus marginal slit lengths are less than adaxial slit lengths (text-fig. 32).

PAUL: DICHOPORITE PORE-STRUCTURES IN CYSTOIDS

703

The canals (text-figs. 1, 12) of pectinirhombs are dichopores. Pectinirhomb
dichopores are discrete (text-fig. 1 3) when each dichopore is a separate entity and the inter-
dichopore space is covered by a thin strip of unmodified thecal plate. The inter-dicho-
pore spaces may vary in width in a pectinirhomb with discrete dichopores, and the
dichopores open externally in slits each provided with an individual slit rim. Confluent
dichopores (text-fig. 14) are dichopores in which the walls of adjacent dichopores are con-
fluent and the inter-dichopore spaces are covered with dichopore walls. In such cases the
inter-dichopore widths are constant and equal to the dichopore widths. Confluent
dichopores open in slits grouped into vestibules (text-figs. 17, 18) and collectively

13 14

text-figs. 13-14. Discrete and confluent dichopores.

13. Discrete. Each dichopore is separated from adjacent dichopores by a strip of unmodified thecal
plate.

14. Confluent. The walls of adjacent dichopores are continuous and the pectinirhomb is made of
isoclinal folds. Both figures represent sutural views of conjunct pectinirhombs.

surrounded by a vestibule rim. In disjunct pectinirhombs with confluent dichopores
one half-rhomb has a complete closed rim which surrounds the slits on both the ab-
and ad-sutural sides. The other half-rhomb has an open rim on the ab-sutural side only
(text-fig. 18). A number of types of pectinirhomb may therefore be recognized on the
character of the slits and dichopores. Five types are in fact recognized:

1. Conjunct pectinirhombs with discrete dichopores (text-fig. 15)

2. Disjunct ,, „ ,, „ (text-fig. 16)

3. Conjunct ,, ,, confluent ,, (text-fig. 17)

4. Disjunct ,, ,, ,, ,, (text-fig. 18)

5. Multi-disjunct pectinirhombs (text-fig. 19)

The structure of multi-disjunct pectinirhombs is not yet fully known.

Conjunct pectinirhombs are much less common than disjunct pectinirhombs and
discrete dichopores are only found in Ordovician genera. Conjunct pectinirhombs with
discrete dichopores are only found in the genus Cheirocrinus s./.: seven species have this
type of pectinirhomb. Conjunct pectinirhombs with confluent dichopores occur in the
genera Homocystites Barrande, Pleurocystites Billings, and Regulaecystis Dehm. Dis-
junct pectinirhombs with discrete dichopores occur in the Cheirocrinidae, Rhombi-
feridae, Cystoblastidae, and Echinoencrinitinae. The Scoliocystinae, Callocystitidae,
and Praepleurocystis Paul have disjunct pectinirhombs with confluent dichopores. Multi-
disjunct pectinirhombs occur in two Upper Ordovician species of Cheirocrinus s.l .,
C.jamesi (McCoy), and C. interruptus (Jaekel). The transition from discrete to confluent
dichopores seems to have been gradual and to have taken place independently in several

704

PALAEONTOLOGY, VOLUME 11

text-figs. 15-18. Diagram to illustrate the morphology of pectinirhombs.

1 5. Conjunct with discrete dichopores.

16. Disjunct with discrete dichopores.

17. Conjunct with confluent dichopores.

18. Disjunct with confluent dichopores.

text-fig. 19. Diagram to illustrate the interpretation of the morphology of a multidisjunct

pectinirhomb.

PAUL: DICHOPORITE PORE-STRUCTURES IN CYSTOIDS

705

families. In particular the pectinirhombs of Glyptocystites spp. are intermediate in
character. Isolated demi-rhombs only occur in species which have pectinirhombs with
discrete dichopores.

All five types of pectinirhomb are confined to and characteristic of the superfamily
Glyptocystitida. Members of this superfamily have thecae composed of a definite number
of plates arranged in five circlets. There are 4 basals (BB), 5 infra-laterals (ILL), 5 laterals
(LL), 4, 5, or 6 radials (RR), and 7 orals (OO). The plates are numbered clockwise around
the theca from the left of the periproct. The positions of pectinirhombs on the theca may

text-fig. 20. Diagram of the morphology of a cryptorhomb.

be denoted by citing the two plates in which the half-rhombs are developed. Thus pec-
tinirhomb L4:R3 has one half-rhomb in L4 and the other in R3. Pectinirhombs are only
developed across certain sutures and some positions are occupied more frequently than
others. No orals ever bear pectinirhombs and there is no marked preference for pec-
tinirhombs to develop across sutures between different circlets (inter-circlet sutures) as
opposed to sutures between plates of the same circlet (intra-circlet sutures).

C. THE MORPHOLOGY OF CRYPTORHOMBS

Cryptorhombs are rhombs composed of dichopores which open externally in pores
(text-fig. 20). The term cryptorhomb is proposed because the entire length of the tan-
gential part of all the dichopores is hidden from external view so that in many species
the rhombs are inconspicuous.

The pores of cryptorhombs (text-figs. 3, 20) may be simple or sieve-like (compound
pores of text-fig. 20). In almost all cases examined each dichopore opened in a sieve-
pore at one end and a simple pore at the other. In all complete cryptorhombs one half-
rhomb has sieve-pores only and the other has simple pores only. A few dichopores may

706

PALAEONTOLOGY, VOLUME 11

be reversed with respect to the others in some incomplete cryptorhombs. The dichotomous
branching of the perpendicular canal which opens in a sieve-pore may proceed evenly in
all branches to produce clusters of 2, 4, 8, 1 6, or 32 individual pores, or unevenly, produc-
ing intermediate numbers of individual pores. The individual pores may be circular,
oval, or dumb-bell shaped : this depends on the stage of branching. Individual pores are
usually surrounded by pore rims and may be set in shallow depressions or raised in
rounded tubercles. Within a sieve-pore, individual pores may be arranged randomly or in
rows. Simple pores may be developed on tubercles, or in or beside ridges on the plates.
Nearly all simple pores have strongly developed rims.

Cryptorhomb dichopores (text-figs. 3, 20) resemble discrete dichopores of pectini-
rhombs in the sense that the inter-dichopore spaces are covered by unmodified thecal
plate. Dichopores may be evenly developed throughout a cryptorhomb, which then has
a regular rhombic outline and is complete. Incomplete cryptorhombs have few, irregularly
spaced, dichopores and irregular outlines. Cryptorhomb dichopores may be weakly
calcified within the theca.

Only one type of cryptorhomb (text-fig. 20), which corresponds in structure to a
disjunct pectinirhomb with discrete dichopores, has been recognized. However, crypto-
rhombs may be complete or incomplete and the unusual rhombs of Polycosmites Jaekel
may represent a second type of cryptorhomb, corresponding to multi-disjunct pectini-
rhombs in structure. Unfortunately no specimens are available.

Cryptorhombs are confined to and characteristic of the superfamily Hemicosmitida,
members of which have thecae composed of a definite number of plates arranged in
3 or 4 circlets. These are designated by the same symbols as in the Glyptocystitida but the
arrangement of plates is less constant. There may be 3 or 4 basals, 6 or 10 infra-laterals,
8 or 9 laterals, and, in the Hemicosmitidae, 9 radials. In the Caryocrinitidae the oral
area is covered by a special ‘tegmeiT which apparently replaces the radial circlet. Crypto-
rhombs are not developed in tegminals but may be present in all other plates. In many
species complete cryptorhombs are predominantly developed across inter-circlet sutures
while incomplete cryptorhombs are confined to intra-circlet sutures of the lateral and
infra-lateral circlets. There are thus rings of complete cryptorhombs which correspond
to the inter-circlet sutures. Furthermore the arrangement of the pores of complete
cryptorhombs is constant (text-fig. 21). In the BB:ILL ring sieve-pores occur in the
basals; in the ILL: LL ring they are in the laterals; in the LL: RR ring (Hemicosmitidae
only) they are in the radials. Incomplete cryptorhombs complicate this arrangement but
functionally they are of minor importance.

THE FUNCTION OF DICHOPORITE PORE-STRUCTURES
A. THE BASIS OF FUNCTIONAL INTERPRETATIONS

The principles on which this functional analysis of cystoid pore-structures has been
based were outlined by Rudwick (19646) but a brief resume will now be given. It is
necessary first to assume that any organ under investigation was functional and to
suggest a function for it. This assumption may be tested in fossils only if it is possible
to predict a structure which would serve the function efficiently and which is directly
comparable with the preserved hard parts of the fossil. This prediction has been termed
a paradigm by Rudwick (1960) and is assumed to have maximum efficiency within the

PAUL: DICHOPORITE PORE-STRUCTURES IN CYSTOIDS

707

limits imposed by the materials of which the fossil structure is made. A close morpho-
logical similarity between the fossil structure and the paradigm would be expected if the
former functioned efficiently. Where such similarity is found it is probable that the postu-
lated function was served in life but this cannot be said to have been proved absolutely.

Three points arise in this connection. 1. The preserved hard parts of the fossil did not
constitute the entire structure in life, i.e. at the time of functioning. It is necessary to
reconstruct the unpreserved soft parts and in doing so it is assumed that the soft

text-fig. 21. Diagrammatic representation of the distribution of simple and sieve-pores in the theca
of Hemicosmites extraneus Eichwald. Only complete cryptorhombs are shown. F, arm facet; G, gono-
pore; N, node in lateral plates; Pe, periproct. Based on RM. Ec5381.

(unpreserved) parts of the fossil bore the same relationship to the preserved hard parts
that the soft parts of related living organisms do to their hard parts. For example,
Sinclair’s (1948, p. 306) suggestion that the dichopores of the pectinirhombs of Glypto-
cystites were lined with ciliated epithelium may be accepted since dichopores are
invaginations of the external surface of the thecal wall and nearly all modern echino-
derms have an external ciliated epithelium.

2. The second assumption that the fossil structure functioned efficiently is only a
working assumption and may be modified to fit any case. To test any postulated func-
tion a close morphological similarity between the fossil structure and the paradigm is
sought. If it is found, the fossil structure could have served the function efficiently. If
not, it cannot be concluded that the structure did not serve the postulated function but
only that it was inefficient if it did.

The early functional morphologists evolved as an empirical law the idea that any
given organ is only developed to the minimum efficiency conducive with survival

C 6055 3 A

708

PALAEONTOLOGY, VOLUME 11

(i minimum survival efficiency ) and not the maximum potential efficiency. A low relative
efficiency in an organ may result from a low minimum survival efficiency, perhaps due
to a lack of competition or low selective pressures. At present this is an hypothetical
consideration since evaluation of minimum survival efficiencies in fossils is impossible
and there is little information from living animals.

Any structure may form part of two (or more) functional systems, or may have two
(or more) concomitant functions, each with different structural requirements. Such
structures may become a compromise between the ideal cases. If it is possible to suggest
the two (or more) functions, it may become possible to test whether departure from one
paradigm is due to approximation to another.

A low relative efficiency in one structure may be compensated for by a high efficiency
elsewhere in the system. Equally minimum survival efficiency may be achieved by a few
efficient organs or a large number of less efficient organs. Thus the second assumption
that the fossil structure functioned efficiently is only a working assumption and con-
siderable information on the functioning of a fossil structure may be gained even when
there is not a close comparison with the paradigm.

3. There is always a danger of using teleological language in functional analyses since
‘function’ and ‘purpose’ are sometimes used synonymously. Yet no fossil ever evolved
a structure to serve a purpose. To investigate the ‘effect’ of a structure need neither
imply function nor purpose. Thus if a chance rock fall produces a narrow crack between
two boulders through which the tide washes, the crack will prevent particles of a dia-
meter greater than its width from passing between the boulders. The crack may be
analysed in the same way as zigzag slits in brachiopods (Rudwick 1964#) or pectini-
rhomb slits (p. 715) and will be found to have the same effect. Since it was a chance
rock fall which produced the crack there can be no purpose behind it nor does it serve
any function. Rudwick (1964fl) concluded that if zigzag slits in brachiopods were pro-
tective they could have served this function efficiently. That is: the effect of zigzag slits
in brachiopods was to reduce the maximum size of particles that could enter (or leave)
the mantle cavity for any given gape. If fossil brachiopods behaved in a similar manner to
living brachiopods, those with zigzag slits were better ‘protected’ from large particles
than those without (for any given gape). This may have been beneficial to these brachio-
pods. The ‘function’ of a structure may be considered as a ‘beneficial effect’ or better
still as the most beneficial of its several effects. Although the purpose of functional
analyses is to determine the ‘function’ of structures, it is the mode of functioning, i.e. the
sum of the ‘effects’ of the detailed morphology of the structures, which is investigated.
In the next sections attempts are made to interpret the effects of the detailed morphology
of pectinirhombs and cryptorhombs, having assumed that they were functional in life.

B. THE GENERAL FEATURES OF EXCHANGE SYSTEMS

Most authors have accepted a respiratory function for all cystoid pore-structures. It
is true that Delpey (1942) has suggested that pectinirhombs and cryptorhombs were
balancing organs and Hyman (1955) among others, has suggested that some canals may
have been nutritive. However, Sinclair (1948, pp. 307-8) has shown how unlikely it is
that pectinirhombs or cryptorhombs were balancing organs. If thecal canals were
nutritive, they would be evenly distributed or at least present in all thecal plates. This
is not always the case with pectinirhombs, cryptorhombs, and diplopores but apparently

PAUL: DICHOPORITE PORE-STRUCTURES IN CYSTOIDS 709

it is the case with humatipores and humatirhombs. The most plausible function for
pore-structures is respiration. If pore-structures were respiratory this was achieved by
diffusion of oxygen and carbon dioxide and the pore-structures formed part of an
exchange system. To test this hypothesis the paradigm of an exchange system must be
examined.

The most essential feature of an exchange system is that only the ‘exchange substance’
or ‘substances’ should be exchanged; direct mixing must be prevented. There must be
an exchange surface which allows the passage of the exchange substances while keeping
the donor and acceptor fluids separate. In respiration the donor current for oxygen is
also the acceptor current for carbon dioxide. Throughout donor current is used with
reference to oxygen exchange. The rate of exchange will depend on the following factors:

1. The area of the exchange surface. The larger the area of the exchange surface, the
greater the amount of exchange. The exchange surface should be as large as its strength
will allow. Echinoderm skeletal calcite is a meshwork of calcite rods and soft tissue.
Oxygen and carbon dioxide diffused not through the skeletal calcite but rather through
the soft tissues. In echinoderms no more than half the area of an exchange surface was
available for exchange because of this meshwork structure.

2. The resistance to exchange of the exchange surface. The resistance of an exchange
surface will be proportional to its thickness and will depend on its nature. Although the
walls of the thecal canals are heterogeneous only the soft tissue functioned as exchange
surface. The thinner the walls are, the lower their resistance to exchange, and their
strength, will be. The paramount function is to maintain separation between two fluids;
rupture of the exchange surface could be fatal. The exchange surface may be as thin as
is compatible with this stipulation.

3. The concentration gradient across the exchange surface. No exchange will take place
unless there is a gradient across the exchange surface, i.e. unless the concentrations of
the exchange substances in the donor and acceptor fluids are different. In a static system
the exchange substances would flow from higher concentrations to lower until the con-
centrations were equal when flow would cease. It is therefore necessary to have a current
system to maintain the gradient. The most efficient system is a counter-current system
(text-fig. 22). In a closed (internal) current system (text-fig. 24) an animal has control
over the composition of the fluids circulating. In an open current system (text-fig. 24)
however, devices to prevent recirculation of exhausted fluids and choking by extraneous
particles are desirable.

Thus the paradigm of an exchange system will have a large area of exchange surface
which is as thin as is compatible with its strength and a counter-current system (with
provisions to prevent recirculation and choking where necessary). This paradigm will
fit any exchange system such as a radiator (heat exchange), a kidney (urea excretion) or
a gill (oxygen and carbon dioxide diffusion). It has not been developed purely for the
present analysis.

C. DETAILED FUNCTIONAL INTERPRETATIONS

From a purely functional point of view there are two types of cystoid pore-structures:
dichoporite and non-dichoporite. In dichoporite pore-structures the current which flowed
through the thecal canals was the donor current (for oxygen) and formed part of an

710

PALAEONTOLOGY, VOLUME 11

open current system. The exchange surfaces were within the thecal cavity (endothecal)
and protected from mechanical damage. In this section the detailed morphology of
dichoporite pore-structures (pectinirhombs and cryptorhombs) will be compared with
the paradigm of an exchange system and estimates of the relative efficiencies of the
various types of rhomb made.

22

D

100

0

100-)*90

T

80 4 70

60

1

40 ->30

20 -> 10

l

10 -• 0

t

100-

• 40

t

,30

1

■20

t

10

24

text-figs. 22-4. Current and exchange systems.

22. A counter-current system. In the ideal case all the exchange substance is transferred from the
donor to the acceptor current.

23. A system with both currents flowing in the same direction. The maximum potential exchange
is 50%.

24. Diagram to illustrate closed and open current systems. In the donor current there is a potential
danger of the entrance becoming choked with particles and of re-circulation from the exit.

A, acceptor current; D, donor current; ES, exchange surface. Figures represent concentrations of

exchange substance; heavy arrows represent exchange; light arrows represent current directions.

1. The area of the exchange surface. An increase in the area of the exchange surface
means an increase in the amount of exchange. In the simplest hypothetical case exchange

would take place through the thecal wall. Oxygen require-
ments would be proportional to the volume, and rate of
exchange to the surface area, of the theca. Throughout
growth oxygen requirements would increase faster than the
amount of exchange, thus limiting maximum size. Evagi-
nations or invaginations of the thecal wall can counteract
this effect without involving a change of gross thecal shape.
Both are found in the Glyptocystitida. Maerocystella
Callaway, the earliest known form, has externally ridged
thecal plates (text-fig. 25). The ridges are formed by folds
in the plates and are evaginations of the thecal wall. They
increased the area, and also the strength, of the plates
which are uniformly very thin (OT mm. in M. mariae
Callaway). Exchange probably took place through the
entire area of the thecal wall in Maerocystella.

Pectinirhombs first appear in Cheirocrinus Eichwald, a direct descendent of Macro -
cy Stella (see Paul 1968). Dichopores are invaginations of the thecal wall which allow a
differentiation of function between the thicker thecal plates (rarely if ever less than 0-5
mm.) and the thin dichopore walls (usually 0-01 mm.). Exchange was probably restricted

text-fig. 25. An isolated thecal
plate of Maerocystella mariae
Callaway to show external
ridges which are effectively
evaginations of the thecal wall.

PAUL: DICHOPOR1TE PORE-STRUCTURES IN CYSTOIDS

711

to the dichopore walls. These allow a large area for exchange within a small area of
thecal surface. In Cheirocrinus granulatus (Jaekel) RM. Ec5384 (PI. 135, figs. 3-5) one
half-rhomb has a ratio of exchange area to thecal surface area of 7-84 : 1 .

This ratio indicates the relative exchange efficiencies of different types of rhomb. In
dichoporite rhombs the closer the spacing of the dichopores and the greater their length
and depth, the higher the ratio will be. Available measurements of spacing are sum-
marized in Table 1. Dichopore width (W) would probably be given by W = 2 (h+1),
where h = height of lining epithelial cells and 1 = length of cilia attached to the epi-
thelial cells. At the limit of closeness of spacing the dichopores would be evenly spaced
and the dichopore and inter-dichopore widths would be equal. This is the case in pectini-
rhombs with confluent dichopores and in complete cryptorhombs but not always in
pectinirhombs with discrete dichopores nor in incomplete cryptorhombs. These latter
were less efficient and tend to have higher values for spacing (Table 1).

Large pectinirhombs tend to have depressed outlines, a large number of dichopores,
and to occur in large thecae. In any exchange system there is a distance within which all
the exchange substance will be removed from the donor current as it flows along the
exchange surface. This limits the length of a dichopore and the longest dichopores of
both compressed and depressed pectinirhombs are about the same length (7-9 mm.).
The exchange area of a pectinirhomb or cryptorhomb may be more efficiently increased
by the addition of new dichopores than by the lengthening of existing dichopores. This
may partly account for the tendency of large pectinirhombs to have depressed outlines
(i.e. width greater than length). Maximum dichopore lengths in cryptorhombs are greater
than those of pectinirhombs and oxygen exchange may have been less rapid in the
former.

To summarize, the structure of dichoporite rhombs compares closely with the para-
digm of an exchange system as regards the area of the exchange surface. Pectinirhombs
with confluent dichopores were relatively more efficient than those with discrete dicho-
pores in terms of exchange area.

2. The resistance to exchange of the exchange surface. For maximum efficiency an
exchange surface should be as thin as its strength will allow. Dichopore walls vary in
thickness from less than 0-01 mm. up to 0-03 mm. in rare cases, whereas the thecal wall
is 0T mm. thick in Macrocystel/a mariae Callaway, rarely, if ever, less than 0-5 mm. in
Cheirocrinus and many Glyptocystitida have thecal walls 2-3 mm. thick. Dichopore
walls are very much thinner than other skeletal elements and pectinirhombs were poten-
tially weak areas of the theca. The morphology of dichopores, as thin tubes or isoclinal
folds, allows great strength and rigidity for a thin surface. It also allows a large surface
area without increasing the volume of the theca and thus satisfies both this and the pre-
vious requirements.

Pectinirhombs with conjunct slits were particularly susceptible to fracture or distor-
tion by forces acting on the surface of the theca. It was essential to maintain rigidity for
normal functioning and this could be achieved by solid ridges in the positions shown
in text-fig. 26. Such ridges impart both strength and rigidity. Ridges are found in the
positions suggested in some cystoids with conjunct pectinirhombs. In species of Cheiro-
crinus with discrete dichopores the ridges are often incorporated in the plate ‘ornament’
of radiating ridges. These ridges strengthened the plates in general and may only

712

PALAEONTOLOGY, VOLUME 11

table 1 . Spacing of Dichopores in Pectinirhombs. The values for spacing are determined by measuring
10 dichopores and 10 inter-dichopore spaces and expressing the average as 1 slit per O.xxx mm. In
pectinirhombs with fewer than 10 dichopores the average was based on the maximum number of
dichopores measurable. The averages for each type of pectinirhomb are based on the total number of

measurements (No.)

Spacing

No. of

Species

Average

Range

measurements

A. Conjunct pectinirhombs with discrete dichopores
Cheirocrinus sp. 0-316

0-217-0-415

2

C. giganteus Leuchtenberg

0-273

1

C. granulatus (Jaekel)

0-253

1

C. cf. atavus (Jaekel)

0 106

1

C. languedocianus Thoral

0-370

0 350-0390

4

B. Conjunct pectinirhombs with confluent dichopores
Homocystites alter Barrande 0-215

0205-0-225

2

H. constrictus (Bather)

0-2025

0-201-0-204

2

Plearocystites elegans Billings.

0196

1

P. filitextus Billings

0-241

1

P. cf. rugeri Salter

0-20 approx.

1

P. squamosus Billings

0-171

1

C. Disjunct pectinirhombs with discrete dichopores
Cheirocrinus sp. 0174

1

C. anatiformis (Hall)

0-257

0-25-0-27

4

C. forbesi Billings

0-273

0-25-0-29

6

Echinoencrinites angulosus Pander

0-27

0-21-0-31

3

E. senckenbergii von Meyer

0-33

0-32-0-34

3

E. reticulatus Jaekel

0-427

0-41-0-44

3

Erinocystis sp.

0-38

1

D. Glyptocystites type

Glyptocystites ehlersi Kesling

0-197

0-19-0-20

3

G. multiporus Billings

0-193

0-18-0-20

10

G. regnelli Sinclair

0-206

0-20-0-22

5

G. batheri Sinclair

0-225

0-205-0-245

7

E. Disjunct pectinirhombs with confluent
Sphaerocystites multifasciatus Hall

dichopores

0-223

0-212-0-225

4

Jaekelocystis hartleyi Schuchert

0-168

0-150-0 180

4

Lepadocystis moorei (Meek)

0 169

0-165-0-175

6

Apiocystites pentrematoides Forbes

0-208

0-200-0-214

3

Strobilocystites calvini White

0-276

0-265-0-285

6

Lovenicystis angelini (Haeckel)

0-218

0-201-0-236

2

Lipsanocystis traversensis

0-152

0 151-0 154

14

Ehlers and Leighley

Tetracystis oblongus (Forbes)

0-156

0 153-0 159

3

Lepocrinites gebhardii Conrad

0-250

2

Pseudocrinites gordoni Schuchert

0-167

0-160-0 174

6

P. bifasciatus Pearce

0-177

0 168-0 187

3

P. pyriformis Paul

0-257

0-233-0-272

3

Staurocystis quadrifasciatus (Pearce)

0-193

0-190-0-201

6

Glansicystis baccata (Forbes)

0205

0 190-0-220

6

Over -all
averages

Range

No.

Type A

0-306

0-106-0-415

9

B

0-205

0-171-0-241

8

C

0-300

0-174-0-440

21

D

0-205

0-193-0-245

25

E

0-192

0-151-0-285

68

PAUL: DICHOPORITE PORE-STRUCTURES IN CYSTOIDS

713

incidentally have strengthened the pectinirhombs. In Homocystites constrictus (Bather)
(PI. 134, fig. 5) and Pleurocystites spp. (PI. 134, figs. 7, 8) rims occur which are indepen-
dent of the plate ridges and whose strengthening effect was confined to the area immedi-
ately adjacent to the pectinirhombs.

In fully developed disjunct pectinirhombs the intra-rhomb areas between the slits
would resist forces acting on the thecal surface. In disjunct pectinirhombs with discrete
dichopores the intra-rhomb areas tend to be large (PI. 135, figs. 1, 9) and this partly
restores the strength of the rhomb-bearing plates. In cryptorhombs (PI. 138,fig. 4; PI. 139,
fig. 1), with only rows of pores, this restoration is virtually complete. In disjunct pectini-
rhombs with confluent dichopores the intra-rhomb areas tend to be smaller but the

text-fig. 26. A conjunct pectinirhomb with ridges
which deflect forces acting within the plate and
impart rigidity to the pectinirhomb.

adsutural portions of the vestibule rims develop within them (PI. 136, fig. 3; PI. 137,
fig. 1). During growth of a disjunct pectinirhomb some marginal slits were conjunct
and subject to the same weakness as conjunct pectinirhombs. In disjunct pectinirhombs
with confluent dichopores the absutural portions of vestibule rims arise early in develop-
ment in exactly the position postulated to strengthen conjunct pectinirhombs (PI. 137,
fig. 1). No such rims are present in disjunct pectinirhombs with discrete dichopores.

Uncalcified dichopores would allow a greater exchange area than calcified dicho-
pores since only the soft tissues function as exchange surface. In pectinirhombs
distortion or disarticulation would probably occur if the dichopores were uncalcified
but in cryptorhombs strength is provided by the covering of thecal plate. It is still
essential to have rigid dichopore walls to prevent the dichopores themselves from dis-
torting (under pressure differences or due to gravity) since this would interrupt the func-
tioning of the cryptorhomb. A maximum movement of only 0-05 mm. by two adjacent
dichopore walls would completely close a dichopore or inter-dichopore space. Calcifica-
tion of dichopores imparts rigidity at the expense of the available area of exchange.

Most pectinirhombs and cryptorhombs have strengthening structures closely associ-
ated with them indicating they were areas of weakness. The conclusion that this weak-
ness was mainly due to the dichopore walls seems inescapable. Dichopore walls were not

714

PALAEONTOLOGY, VOLUME 11

only thin absolutely but relatively so thin that the limits of their strength were reached
and strengthening devices became necessary to counteract this weakness. In this respect
dichoporite pore-structures agree closely with the paradigm of an exchange system.
Further conclusions are that conjunct pectinirhombs were intrinsically weaker than
disjunct pectinirhombs and all pectinirhombs were weaker than cryptorhombs. Streng-
thening devices in pectinirhombs with confluent dichopores apparently allowed a
weaker structure with a better exchange area to thecal surface area ratio. Disjunct pec-
tinirhombs with confluent dichopores were strengthened throughout a longer period of
their growth than those with discrete dichopores.

3. Protective devices and devices to prevent recirculation. The currents which flowed
within dichopores formed part of an open current system. Therefore devices to prevent
choking by extraneous particles and recirculation of deoxygenated sea-water would have
been beneficial to exchange. Such devices are also very useful in determining current
directions since they indicate which pores were entrances and which exits. Clearly it is
more efficient to modify the entrance (rather than the exit) with a protective device since
this will prevent entry of harmful particles. Equally, exhausted sea-water in the vicinity
of the entrance will be sucked in by the in-current. Recirculation can only be prevented
by modifying the exit to direct the out-current away from the entrance and into the
ambient sea-water for remixing. Protective devices, which will be considered first,
indicate entrances and devices to prevent recirculation indicate exits.

A protective device is a structure which reduces the maximum size of particles that
may pass through an aperture without interfering with the functioning of the aperture.
There are only three classes of structure which fulfil these requirements : meshes, grilles,
and narrow slits. Paradigms for all three have been defined by Rudwick (1961, meshes
and grilles; 1964a, slits). The maximum particle which can pass through the device is
termed the critical particle and is of critical diameter. In an ideal protective device the
critical diameter is the same at all points within the aperture.

Two types of harmful particle may be recognized. Passive particles are any particles
in suspension (animal or plant debris, sediment, etc.) which may choke the dichopores
but not attack the cystoid. Active particles are organisms which may attack the cystoid

EXPLANATION OF PLATE 134
Stereophotos of conjunct pectinirhombs

A. with discrete dichopores.

Fig. 1. Cheirocrinus giganteus Leuchtenberg. BMNH E16130, basal rhombs.

Figs. 2, 4 Cheirocrinus sp. SM A3 138 (fig. 4), SM A3 139 (fig. 2). External moulds of isolated plates.
Fig. 3. Cheirocrinus granulatus (Jaekel). RM Ec5384.

Fig. 9. Cheirocrinus languedocianus Thoral. Ubaghs coll, latex impression showing incomplete pec-
tinirhombs and demirhombs in radial and lateral plates.

B. with confluent dichopores.

Fig. 5. Homocystites constrictus (Bather). HM 3541a, latex impression showing rhombs B2:IL2,
LI :L2, and R1 :R2. Note the rim on L2 (middle rhomb).

Fig. 6. Pleurocystites rugeri Salter. LU 3175, latex impression of L3 :L4.

Fig. 7. Pleurocystites elegans Billings. SM A53059. L3:L4.

Fig. 8. Pleurocystites filitextus Billings, BMNH E16047. L4:L3.

Figs. 1, 8, 9 x 2, figs. 2-7 x 3. All whitened with ammonium chloride sublimate.

Palaeontology , Vol. 11

PLATE 134

PAUL, Cystoid pore structures

PAUL: DICHOPORITE PORE-STRUCTURES IN CYSTOIDS 715

by boring into it, attaching to it or browsing off cilia, etc. Dichopores may be regarded
as long thin tubes whose critical diameter at any given point is the minimum width.
Complete protection against passive particles may be achieved if the critical diameter
of the entrance is less than that of any other point along the tube. Any spherical
particle capable of entering the tube will pass right through without becoming lodged.
Complete protection against active particles, which may be able to swim against cur-
rents, can only be achieved if the critical diameters of both the entrance and the exit
are less than the diameter of the smallest active particle. As nothing is known about

table 2. Ratios of slit lengths and areas in disjunct pectinirhombs

Ratios

Species

Rhomb

Lengths
longer slit
shorter slit

Areas
longer slit
shorter slit

A. Discrete dichopores

Echinoencrinites angulosus BMNH E29102

B2

IL1

1-27 (7)

0-82 (5)

B2

IL2

I 21 (11)

0-84(11)

E. reticulatus RM Ec5515

B2

IL1

1-25 (6)

0-94 (6)

B2

IL2

1-27 (8)

1-31 (8)

L4

R3

Ml (9)

0-94 (9)

E. senckenbergii BMNH E29095

B2

IL1

1-54 (8)

0-97 (8)

B2

IL2

1 43 (6)

1-06(6)

L4

R3

116 (10)

0-87 (10)

Cheirocrinus anatiformis

R2

R3

M4(15)

1-08 (15)

B. Confluent dichopores

Glyptocystites multiporus

LI

L2

1 -46 (9)

1-46 (9)

Staurocystis quadrifasciatus

LI

R5

1-37 08)

0-97 (18)

Lovenicystis angelini RM Ec5053

LI

R5

1-77 (13)

1 -38 (13)

RM Ec5066

LI

R5

1-30 (17)

0-98 (17)

Glansicystis baccata Bu Ho55

LI

R5

115 (9)

1 09 (9)

L4

R3

1-43 (9)

1 19 (9)

GSM 7380

LI

R5

2-63 (8)

1 91 (8)

Figures in brackets represent the number of measurements upon which each ratio was based.

active particles it is impossible to investigate protection against them. Only passive
particles are considered below.

The dichopores of pectinirhombs open as slits and if protected the entrances should
form protective slits. In conjunct pectinirhombs each slit runs the entire length of the
dichopore and acts as both entrance and exit. In Cheirocrinus latavus (Jaekel) BMNH
E23515 and Pleurocystites filitextus Billings BMNH E7600b (PI. 140, fig. 2) slit widths
are slightly less than dichopore widths within the theca. In these cases the slits were
protective.

In disjunct pectinirhombs the slits of the two half-rhombs were of different lengths.
In many cases the longer slits are also narrower. To satisfy the requirements of a protec-
tive device the critical diameter of the incurrent slit must be reduced without reducing
current flow. If protection is achieved by reducing the width, constant current
velocity may be maintained by a corresponding increase in slit length since the velocity
is inversely proportional to the cross-sectional area of the slit. In the ideal case the ratio
of entrance to exit slit length will be greater than unity but the ratio of the areas will
equal unity. Available measurements are shown in Table 2. In cases where the longer

716

PALAEONTOLOGY, VOLUME 11

slits were also narrower they could have been protective. In cases where the slit widths
are equal they could still have been protective if both were narrower than the dichopore
width within the theca.

The pectinirhombs of Lipsanocystis magnus Stumm (PI. 140, fig. 4) show microscopic
bars across the slits of both half-rhombs. These bars reduce the critical diameter of both
apertures. Preservation is not quite adequate to determine which slits are better pro-
tected or if any difference exists. The formation of these microscopic grilles may be
unique to L. magnus; perhaps a response to its muddy environment or to active particles
since both apertures are modified. Alternatively it may be that the preservation of most
glyptocystitids is not adequate for these delicate structures to be detectable. In the latter
case the arguments for protective devices in pectinirhombs may need some modification.

table 3. Critical diameters in pectinirhombs and cryptorhombs
Entrance Exit

Rhomb type
Pectinirhombs

Average

Range

No.

Average

Range

conjunct discrete

0 115

0 02-0- 19

6

conjunct confluent

0066

005-008

8

disjunct discrete

0 113

005-0 18

10

0-162

0-12-0-20

confluent disjunct

0075

0-07-0-095

5

0-104

0-09-0012

Cryptorhombs

0058

0-04-0-07

13

0-20

010-0-35

No. = Number of rhombs of each type measured. In pectinirhombs with variable slit widths the average
width of all slits was calculated.

In cryptorhombs the dichopores open as pores, at least one of which is sieve-like,
being composed of a cluster of fine pores. If sieve-pores were protective they formed
protective meshes. Measurements given in Table 3 show that the critical diameters of
sieve-pores are much less than those of simple pores. The areas (Table 4) of simple and
sieve-pores are often approximately equal however. Again protection could be provided
while interfering to the least extent possible with current flow.

Particles prevented from entering dichopores by protective devices would choke the
slits and pores themselves unless removed. Clearance could be achieved by surface ciliary
currents, the particles being carried in the same direction as the currents within the

EXPLANATION OF PLATE 135
Stereophotos of disjunct pectinirhombs

A. with discrete dichopores.

Fig. 1. Cheirocrinus sp. nov. SM A15966. Latex impression of isolated plate.

Figs. 2-5. Cheirocrinus granulatus (Jaekel). RM Ec5384. Fig. 2, internal surface of basal plate; figs.

3-5, views of isolated half-rhomb to show relationship between dichopores and slits.

Fig. 6. Cheirocrinus sp. BMNH El 5983.

Fig. 7. Echinoencrinites senckenbergii von Meyer. BMNF1 E29102, rhombs B2:IL1 and B2:IL2.

Fig. 8. Echinoencrinites reticulatus Jaekel. BMNH E29095. B2:IL1 and B2:1L2.

Figs. 9, 12. Cheirocrinus anatiformis (Hall). GSC. B2:IL2 (fig. 9) and radials (fig. 12).

Fig. 10. Echinoencrinites reticulatus Jaekel. RM Ec5515. B2 : 1 L 1 .

B. with confluent dichopores.

Fig. 11. Glyptocystites multiporus Billings. GSC 1387g (Lectotype). B2:IL1, B2:IL2, and R1 : R2. Note
the adsutural rim on R2 (large upper rhomb).

Figs. 1, 6-9, 11, 12 x2, figs. 2-5, X 3, fig. 10, x5. All whitened with ammonium chloride sublimate.

Palaeontology, Vol. 11

PLATE 135

PAUL, Cystoid pore structures

PAUL: DICHOPORITE PORE-STRUCTURES IN CYSTOIDS

717

table 4. Areas of simple and sieve-pores in cryptorhombs

Area of sieve-

Area of simple

Specimen

pore ( entrance )

pore {exit)

Hemicosmites sp. RM Ec5364

0 0707 sq. mm.

0 0471 sq. mm.

H. extraneus RM Ec5283

00226

00201

H. sp. nov. RM Ec2280

0 01 54 (radials)

00314—00572

0 0502 (laterals)

H. cf. verrusosus BMNH El 5994

0 0491

0 0433 00201-

H. malum RM Ec5517

0 0201 (basals)

0 0402 (laterals)

00254

H. cf. pyriformis BMNH E7592

00314

00227

Caryocrinites ornatus SM A50958*

00308

00154

C. roemeri BMNH E29105*

00628

00628

C. septentrionalis RM Ec25493

00226

00314

Thomacystis tuberculata BMNH E16300

00176

00176

Unless otherwise stated the area of sieve-pores was taken from lateral plates: all areas of simple pores are
from infra-lateral plates.

* These specimens have sieve-pores at both ends of dichopores in complete cryptorhombs. However, there
are fewer larger pores per cluster where one would normally expect simple pores.

text-fig. 27. Supposed paths of extraneous particles in conjunct pectinirhombs. Solid arrows indicate
external currents; dashed arrows internal currents.

dichopores. In conjunct pectinirhombs particles strained from the in-current would be
passed along the slit until they reached the out-current and would then be carried away
from the theca (text-fig. 27). Similar currents could have been present in disjunct pec-
tinirhombs and cryptorhombs but the slit, vestibule, and pore rims would render them
less effective (text-figs. 28, 29).

We may conclude that the narrower slits of disjunct pectinirhombs and the sieve-
pores of cryptorhombs reduced the critical diameters of these apertures, which were

718 PALAEONTOLOGY, VOLUME 11

text-figs. 28, 29. Possible paths of extraneous particles in disjunct pectinirhombs. Note that the slit
and vestibule rims make it more difficult for the excurrent to remove the particles.

PAUL: DICHOPORITE PORE-STRUCTURES IN CYSTOIDS 719

therefore entrances. Absolute critical diameters shown in Table 3 indicate that crypto-
rhombs were better protected than pectinirhombs. However the latter, particularly con-
junct pectinirhombs, were better able to prevent the protective devices themselves from
becoming choked. This may account for the occurrence of conjunct pectinirhombs in all
but one rhomb-bearing member of the Pleurocystitidae. Members of this family were
probably vagrant with the theca resting on the substrate (Paul 1967c) and may have
encountered many more particles than more typically pelmatozoan cystoids.

For maximum efficiency devices to prevent recirculation of exhausted sea-water must
do so while interfering with current flow to the least possible extent. The most funda-
mental requirement is to have two apertures and to keep them apart. How far apart will
depend on the volume, velocity, and direction of the out-current. Even if the entrances
were capable of detecting oxygen-depleted water they could not prevent it from entering
without arresting or reversing the currents. Recirculation can only be prevented effec-
tively by modifying the exit to direct the out-current away from the entrance.

Recirculation may occur within a single rhomb or from one rhomb into another. In
a theca with many rhombs recirculation of the second type may be prevented by group-
ing exits together. In general where two or more apertures are found within a single
plate they are either exits or entrances (text-figs. 30, 31). Recirculation within a single
rhomb may be prevented by mechanical barriers or by devices to project the out-current
through the boundary layer into the ambient sea-water for remixing. The boundary layer
is the layer of fluid immediately adjacent to the surface of an immersed solid, within
which viscosity is the most important single force controlling flow. Cilia beat within the
boundary layer and without them water in contact with the external surface of an
echinoderm would remain unchanged in all but the most turbulent seas. Water emerging
from the dichopores would be recirculated by surface cilia unless forced through this
layer. Mechanical barriers between the entrance and exit were therefore probably less
efficient than were devices to project the out-current through the boundary layer.

The best method to determine the effects of such structures is to construct models
and test them under simulated conditions as close as possible to those under which they
operated in life (cf. Rudwick 1961, Jefferies and Minton 1965). Time has not been avail-
able to do this and the conclusions outlined below are tentative.

The conjunct pectinirhombs of Homocystites constrictus (Bather) (PI. 134, fig. 5) have
distinct rims on one half-rhomb only. Water passing along the dichopores towards
these rims would be deflected away from the theca. Where two half-rhombs are developed
in a single plate they are either rimmed or unrimmed. Current directions shown in text-
fig. 30 are derived from the presence of these rims.

Sinclair (1948) suggested that the adsutural ridges on one half-rhomb of each pectini-
rhomb in Glyptocystites spp. were mechanical barriers to recirculation. This suggestion
seems likely although as yet without practical confirmation. Again where two or more
half-rhombs occur in one plate they are either with or without ridges. In text-fig. 31 the
arrows point to the plates with ridges.

Disjunct pectinirhombs with confluent dichopores have a closed vestibule rim on one
half-rhomb and an open rim on the other. The absutural portion of the closed rim would
have directed the out-current away from the theca. The adsutural portion would have
acted as a mechanical barrier to recirculation as Sinclair suggested for the pectinirhombs
of Glyptocystites. In addition the closed vestibule rims united all the currents of the

720

PALAEONTOLOGY, VOLUME 11

constituent dichopores and probably reinforced them. If out-currents they would have
passed through the boundary layer more readily. The pectinirhombs of Jaekelocystis
hartleyi Schuchert (PI. 137, figs. 7, 9, 10) have closed vestibule rims which form a pore
0-5 mm. in diameter. This pore would considerably increase the velocity of currents
passing through it. In disjunct pectinirhombs with a constant slit width in both half-
rhombs, the shorter slits would have had faster currents. Fast in-currents would be more
likely to suck in extraneous particles but fast out-currents would be more likely to pass
through the boundary layer and disperse. Shorter slits and closed vestibule rims therefore
probably indicate exits.

text-figs. 30, 31. Current patterns in Cheirocrinus and G/yptocystites.

30. Diagram to show the arrangement of thecal plates and pectinirhombs in Cheirocrinus anatiformis
(Hall).

31. Diagram to show the arrangement of thecal plates and pectinirhombs in Glyptocystites ehlersi
Kesling.

Arrows indicate supposed current directions in the pectinirhombs. Note where more than one
half-rhomb is developed within a single plate they are usually of the same type (i.e. either entrance or
exit). Plate outlines based on Kesling 1962 (text-fig. 30) and 1961 (text-fig. 31).

Many cryptorhombs have simple pores developed beside or within ridges on the plate
surface. These ridges could have directed currents away from the theca. In Caryocrinites
septentrionalis Regnell (PI. 139, fig. 5) the ridges beside simple pores are more strongly
developed than those associated with sieve-pores. In Hemicosmites spp. (PI. 138, figs.
1, 5) simple pores are developed into chimney-like tubercles and may be interpreted as
chimneys. The further away from the thecal surface the out-current emerges the less
likely it is to be recirculated by surface ciliary currents. These were therefore probably
exits.

EXPLANATION OF PLATE 136

Stereophotos of disjunct pectinirhombs with confluent dichopores

Figs. 1, 7. Pseudocrinites gordoni Schuchert. BMNH E23122. LI :R5 (fig. 1) and B2:IL2(fig. 7). Note
the vestibule rims typical of this type of pectinirhomb.

Fig. 2. Cal/ocystites jewetti Hall. GSC 14686. L4:R3.

Fig. 3. Pseudocrinites pyriformis Paul. BU HolO. LI :R5.

Figs. 4, 5. Lovenicystis angelini (Jaekel). RM Ec5066. L4:R3 (fig. 4) and LI :R5 (fig. 5).

Fig. 6. Strobilocystites calvini White. UI. L4:R3.

Fig. 8. Sphaerocystites multifasciata Hall. BMNH 22960. B2:IL2.

Fig. 9. Pseudocrinites perdewi Schuchert. USNM 35072. L4:R3.

Figs. 1, 3-5, 7, 8 X 3, figs. 2, 6, 9 X 2. All whitened with ammonium chloride sublimate.

Palaeontology, Vol. 11

PLATE 136

PAUL, Cystoid pore structures

PAUL: DICHOPORITE PORE-STRUCTURES IN CYSTOIDS 721

To conclude, devices which prevented recirculation are apparently present in most
cryptorhombs and disjunct pectinirhombs but are generally absent from conjunct
pectinirhombs. Suggested entrances were described under protective devices and exits
under devices to prevent recirculation. In all cases these interpretations should be, and
in fact are, complementary. Thus in disjunct pectinirhombs longer slits were interpreted
as entrances and shorter slits as exits. In cryptorhombs sieve-pores were interpreted as
entrances and simple pores as exits. Closed vestibule rims in pectinirhombs were inter-
preted as being associated with exits and they always occur on the half-rhombs with
shorter slits. All these interpretations confirm each other and also confirm the presence
and direction of donor currents within dichoporite pore-structures.

4. The maintenance of a concentration gradient. A counter-current system is the most
efficient method of maintaining a concentration gradient. Ciliary currents within the
dichopores would be expected on a priori grounds as suggested by Sinclair (1948). The
viscous effect of the boundary layer would be even more marked in a restricted space
such as a dichopore. No matter how turbulent the seas, water within the dichopores
would not be disturbed without cilia.

text-fig. 32. Graph to show relationship between slit length and dichopore length in pectinirhomb
LI :R5 of Stauroeystis quadrifasciatus (Pearce), BU Ho 1. Lower scale reversed about the rhomb axis
for clarity. Note that the slit lengths are proportional to the dichopore lengths and the slit lengths in

LI are constantly less than those of R5.

If currents were present, sea-water flowed within the dichopores forming the donor
current which was part of an open current system. The best evidence for the presence
and direction of the donor current has been outlined in the preceding section. Additional
evidence is found in the relationship between slit and dichopore lengths in pectini-
rhombs and between pore counts and dichopore length in cryptorhombs. Adaxial
dichopores are longer and deeper than marginal dichopores. If the current velocity were
constant in all dichopores more water would pass through the larger dichopores per
unit time. The capacity of the dichopore would be proportional to its cross-sectional
area and, since the width is constant, capacity would be proportional to dichopore

722

PALAEONTOLOGY, VOLUME 11

depth. For constant velocities of in-current and out-current across a rhomb, slit length
must be proportional to dichopore capacity. Dichopore depth is not readily measurable
but in some instances e.g. Staurocystis quadrifasciatus (Pearce) (PI. 137, fig. 1, text-fig. 32)
slit length is almost exactly proportional to dichopore length. In the cryptorhombs of
Caryocrinites ornatus Say ( PI . 1 38, fig. 3) the number of pores per sieve-pore cluster increases
adaxially as does the total area of the entrance. The morphology of these rhombs would
allow constant current velocities across the rhombs and within all dichopores, if currents
were present. Ciliary currents are likely to have been of similar velocities in all dichopores.

It is more difficult to confirm the presence and direction of internal acceptor currents
as they formed a closed current system, but there is some indirect evidence. In rhombs
with closely spaced dichopores the inter-dichopore spaces were entirely within the
boundary layer. Ciliary currents between dichopores would be expected on the same
a priori grounds used to postulate their presence within dichopores. The width of dicho-
pores was probably controlled by the combined thicknesses of ciliated epithelium, and
lengths of cilia, lining them. If the dichopore width were much greater, dead water would
accumulate within the dichopore. Alternatively cilia on opposite sides of the dichopore
would interfere with each other if the dichopores were thinner. Where inter-dichopore
widths equal dichopore widths both were probably controlled by the presence of cilia.
This suggests that internal ciliated currents were present. The most efficient internal
current is a counter-current.

Internal currents within a rhomb could only act efficiently if the external currents in
all the dichopores were counter to them. With very rare exceptions in some incomplete
cryptorhombs, current directions in all dichopores of one rhomb were apparently the
same. Currents in all dichopores could have been counter to an internal current.

EXPLANATION OF PLATE 137

Stereophotos of disjunct pectinirhombs with confluent dichopores

Fig. 1. Staurocystis quadrifasciatus (Pearce). BU Hoi. LI :R5.

Figs. 2, 4. G/ausicystis baccata (Forbes), typical form. GSM 7380. B2:IL2 (fig. 2) and LI :R5 (fig. 4).

Fig. 3. Lipsanocystis traversensis Ehlers and Leighley. UMMP 56222. L4:R3.

Fig. 5. Schizocystis armata (Forbes). GSM 7381. B2:IL2.

Figs. 6, 8. Glansicystis baccata (Forbes), common form. BU Ho55. L4:R3 (fig. 6) and LI :R5 (fig. 8).

Figs. 7, 9, 10. Jaekelocystis hartleyi Schuchert. BMNH E23121. L4:R3 (fig. 7), L1:R5 (fig. 9), and
B2:IL2 (fig. 10). Note that the vestibules in IL2, LI, and L4 are small pores.

Figs. 1, 2, 4-10 X 3, fig. 3x2. All whitened with ammonium chloride sublimate.

EXPLANATION OF PLATE 138
Stereophotos of cryptorhombs

Fig. 1. Hemicosmites sp. RM Ec5364. Infra-lateral with simple pores.

Fig. 2. Hemicosmites extraneus Eichwald. RM Ec5283. Laterals with simple pores and radials with
sieve-pores.

Figs. 3, 6. Caryocrinites ornatus Say. BMNH E29091. Internal (fig. 6) and external (fig. 3) views of
isolated lateral plate. Note that the number of individual pores in a cluster increases towards centre
of plate (fig. 3).

Fig. 5. Hemicosmites cf. verrucosus Eichwald. BMNH El 5994. Infra-laterals and laterals. Note chimney-
like simple pores.

Figs. 4, 7. Caryocrinites ornatus Say. BMNH E29084. Internal (fig. 7) and external (fig. 4) views of
isolated infra-lateral plate. Note that only simple pores occur in the infra-lateral.

Figs. 1,2 x 3, figs. 3-7 x 2. All whitened with ammonium chloride sublimate.

Palaeontology, Vol. 11

PLATE 137

PAUL, Cystoid pore structures

Palaeontology, Vol. 11

PLATE 138

PAUL, Cystoid pore structures

PAUL: DICHOPORITE PORE-STRUCTURES IN CYSTOIDS

723

In any cystoid relying on diffusion alone to distribute oxygenated fluids internally it
would be advantageous to have rhombs over the entire external surface as all internal
organs would be equidistant from an oxygen source. An approximation to this distribu-
tion of rhombs is found in Cheirocrinus and Glyptocystites (pectinirhombs) and in most
cryptorhomb bearing genera. As the latter grew, relatively large areas without crypto-
rhombs were developed adjacent to the intra-circlet sutures of the infra-lateral and lateral

text-figs. 33-6. Current directions in cystoids with three pectinirhombs.

33. Staurocystis quadrifasciatus (Pearce). 35. Pleurocystites sp.

34. Glansicystis baccatus (Forbes). 36. Erinocystis sp.

Note the current directions in all four have the same basic pattern. Plate outlines after Paul 1967a
(text-figs. 33-4), Breimer 1963 (text-fig. 35), and Jaekel 1899 (text-fig. 36).

circlets. It is within just these areas that incomplete cryptorhombs develop relatively
late in ontogeny. The distribution and development of cryptorhombs, and of pectini-
rhombs in Cheirocrinus and Glyptocystites , suggests that internal circulations were
inefficient or lacking. Cystoids with only a few rhombs probably required some sort
of internal circulation. Independent evidence for this is found in the Callocystitinae
(Paul 19676). In all cystoids with few rhombs the same external current pattern is found
(text-figs. 33-6). This could reflect a fixed internal circulation but only if there were a
constant relationship between donor and acceptor current directions. The most efficient
relationship is a counter-current system.

3 B

C 6055

724

PALAEONTOLOGY, VOLUME 11

As regards maintenance of a concentration gradient, there is considerable direct
evidence for the presence of external currents within dichopores and some indirect
evidence for internal counter-currents between dichopores. Again dichoporite pore-
structures agree closely with the paradigm of an exchange system in this respect.

In general most details of dichoporite pore-structures agree remarkably well with the
paradigm of an exchange system and would have been beneficial to exchange. Oxygen,
and corresponding carbon dioxide, diffusion is the most likely form of this exchange and
hence it seems that pectinirhombs and cryptorhombs were most probably respiratory
organs.

D. EVOLUTIONARY AND TAXONOMIC IMPLICATIONS

The distribution of different types of pectinirhomb within the Glyptocystitida is
summarized in text-fig. 37. Throughout the detailed functional analysis it was suggested
that conjunct pectinirhombs were inherently less efficient than disjunct pectinirhombs
(except in clearing the slits of extraneous particles) and that discrete dichopores were less
efficient than confluent dichopores.

With the possible exception of Scoliocystis Jaekel, conjunct pectinirhombs are con-
fined to the Cheirocrinidae and Pleurocystitidae. Their presence in the early (Arenig,
L. Ordovician) cheirocrinids is probably primary. The appearance of conjunct pectini-
rhombs with confluent dichopores in later cheirocrinids like Homocystites constrictus

EXPLANATION OF PLATE 139
Stereophotos of cryptorhombs

Fig. 1. Hemicosmites malum Pander. RM Ec5517. Note the ‘hidden rhombs’.

Fig. 2. Hemicosmites sp. RM Ec5702. Internal surface of isolated infra-lateral.

Figs. 3, 6. Thomacystis tuberculata Paul. BMNH E16300. Complete theca (fig. 3) and external mould of
isolated basal plate (fig. 6). Latter shows cast of two dichopores which bifurcate near external surface
of plate to form sieve-pore (cf. PI. 140, fig. 5).

Fig. 4. Caryocrinites roemeri Jaekel. BMNH E29105. Infra-lateral and lateral plates; both show
sieve-pores.

Fig. 5. Caryocrinites septentrionalis Regnell. RM Ec25493. Note rims beside simple pores are more
strongly developed than those by sieve-pores.

Fig. 7. Caryocrinites ornatus Say. SM A50958.

Fig. 8. Hemicosmites cf. pyriformis von Buch. BMNH E7592.

Figs. 1, 4, 8 x2, figs. 2, 3, 5-7 x 3. All whitened with ammonium chloride sublimate.

EXPLANATION OF PLATE 140

Fig. 1. Echinoencrinites angulosus Pander. RM Ec5490, section 9. Vertical section through one
dichopore in B2:IL2. Note structure of thecal plate. x35.

Fig. 2. Pleurocystites filitextus Billings. BMNH E7600b. Vertical section through B2:IL2 X 35.

Fig. 3. Cheirocrinus jamesii (M‘Coy). RSM 1870.12.879. External mould of isolated plate showing
multi-disjunct pectinirhomb. X 6.

Fig. 4. Lipsanocystis magnus Stumm. UI 32001. Detail of slits in L4:R3 to show microscopic bars
across slits. x25.

Figs. 5, 6. Hemicosmites sp. nov. Sections through thecal plates to show sieve-pore (fig. 5) and simple
pore (fig. 6). X 30 approx.

Fig. 7. Strobilocystites calvini White. UI. Sutural view of isolated plate to show dichopores. Photo-
graphed under xylol, x 25.

Fig. 8. Echinoencrinites angulosus Pander. RM Ec5490, section 3. Tangential section through
B2:IL1 x 3.

Palaeontology, Vol. 11

PLATE 139

PAUL, Cystoid pore structures

2SS

Palaeontology, Vol. 11

PLATE 140

PAUL, Cystoid pore structures

PAUL: DICHOPOR1TE PORE-STRUCTURES IN CYSTOIDS

725

pectinirhombs and of confluent and discrete dichopores.

726

PALAEONTOLOGY, VOLUME 11

(Bather) (Ashgill, U. Ordovician), may be due to the mode of life. These pectinirhombs
are more highly developed than those of earlier species. The retention of conjunct
pectinirhombs in the Pleurocystitidae is probably due to the mode of life as suggested
by Paul (1967c).

Confluent dichopores are unknown before the Llandeilo (M. Ordovician) while
discrete dichopores are unknown after the Ashgill or possibly even the Caradoc (Late
Middle Ordovician). Between these time limits there was a change from discrete to
confluent dichopores independently in all evolutionary lines. The change may have been
gradual in some cases: the pectinirhombs of G/yptocystites Billings are somewhat inter-
mediate in character. All Silurian and Devonian glyptocystitids (with the exception of
two species of Pleurocystitidae) have disjunct pectinirhombs with confluent dichopores,
the most efficient type. Apparently throughout the evolution of the Glyptocystitida
there was a progression from less to more efficient types of pectinirhomb. There is no
evidence for a gradual change from conjunct to disjunct pectinirhombs but this is not
surprising since it is difficult to see, on geometrical grounds, how a transition could exist.

At present no estimate of the functional efficiency of multi-disjunct pectinirhombs can
be made. It is also much more difficult to compare the relative exchange efficiencies of
pectinirhombs and cryptorhombs. The latter show much less variation and as yet no
evolutionary trends have been recognized.

Within the Glyptocystitida (but not within the Rhombifera as a whole) there is evi-
dence of a progressive decrease in the number of pectinirhombs in one theca. Bather
(1913) explained this reduction in the Pleurocystitidae as a result of the anus partly
taking over respiration. However, this explanation cannot be applied to the Callocysti-
tidae, for example, which lack the extensive periproctal membrane supposedly used in
respiration per anum. The reduction in the number of pectinirhombs could be due to the
development of an internal circulation to distribute more efficiently oxygen gained from
the pectinirhombs. Some independent evidence for such a circulation system has been
described by Paul (19676). The Glyptocystitida alone show this trend.

Rudwick (1961, 1964a) has emphasized that there may be a limited number of different
structures which can perform a given function efficiently. In these cases it is inherently
likely that the same or similar structures may have evolved independently several times.
This was undoubtedly the case with confluent dichopores. It is also undoubtedly the
case with rhombs as a whole. Fistulipores and dichopores represent basically different
structures and there is no feasible way to evolve one from the other. The fistuliporite
and dichoporite rhombs of the Rhombifera were independently acquired and the pos-
session of rhombs is a polyphyletic character within the class. The Glyptocystitida and
Hemicosmitida (dichoporite) can be separated from the Caryocystitida (fistuliporite) on
several characters, e.g. plating arrangement, mode of growth, presence of true stem,
besides the type of thecal canal present. The dichoporite superfamilies form natural
units but it is still possible that they acquired their rhombs independently. It is easy to
envisage that cryptorhombs evolved from disjunct pectinirhombs but there is no evidence
at all for such a transition and both types of rhomb appear simultaneously in the
Arenig. Furthermore, a rhomb-less ancestor for the Glyptocystitida is known ( Macro -
cystella ) which has an identical plate arrangement to later Glyptocystitida but which is
quite distinct from all Hemicosmitida.

The cystoidea as currently understood (Kesling 1963, 1968) form a polyphyletic group

PAUL: DICHOPORITE PORE-STRUCTURES IN CYSTOIDS

727

despite the recognition of the Eocrinoidea, Paracrinoidea, etc. as distinct classes. It is
suggested that the Diploporita and Rhombifera should also be regarded as separate
classes. Within the Rhombifera two orders, the Dichoporita (Jaekel 1899 emend.) and
Fistuliporita nov., may be recognized. The former corresponds to the Pectinirhombi-
fera (Delpey 1942, p. 207) but as the distinctive character is the possession of dichopores
not pectinirhombs, an emended use of Jaekel’s Dichoporita is preferred.

Since its first inception as a class to include all non-crinoid, non-blastoid Pelmatozoa,
the Cystoidea has been subject to the removal of constituent groups as their distinctive-
ness was realized. This process is here carried to the limit and as neither Diploporita nor
Rhombifera can be considered the more typical cystoids, the formal name Cystoidea
must fall into disuse. The concept of a ‘cystoid’ is still useful in a general sense similar
to that now reserved for ‘Dinosaur’. Cystoid may be used to embrace the Eocrinoidea,
Paracrinoidea, Rhombifera, Diploporita, Parablastoidea, Edrioblastoidea, and possibly
the Blastoidea s.s. These classes share many common characteristics which separate
them from ‘carpoids’ and ‘edrioasteroids’, two other groups once included within the
Cystoidea s.I.

APPENDIX

List of species with dichoporite pore-structures examined.

Class Rhombifera Muller

Order Dichoporita Jaekel emend. Rhomb type

Superfamily Glyptocystitida Bather Pectinirhombs

Family Macrocystellidae Bather emend. Jaekel slits dichopores

Macrocystel/a mariae Callaway — ■ —

M. bohemicus (Barrande) —

M. azaisei (Thoral) — ■

Family Cheirocrinidae Jaekel

Cheirocrinus sp. (Ngwetung, Burma)
C. atavus (Jaekel)

C. giganteus (Leuchtenberg)

C. languedocicmus Thoral

conjunct

discrete

55

55

C. gramilatus (Jaekel)

conjunct and

,,

disjunct

C. anatiformis (Hall)

C. sp. nov. (Tramore, Eire)
C. sp. (Katlino, U.S.S.R.)

disjunct

55

”

C. interruptus (Jaekel)
C. jamesi (M‘Coy)

multi-disjunct

unknown

Homocystites alter Barrande
H. const rictus (Bather)

conjunct

confluent

Family Echinoencrinitidae Bather
Subfamily Echinoencrinitinae Paul

Echinoencrinites angulosus Pander disjunct discrete

E. senckenbergii von Meyer „ „

E. reticulatas Jaekel ,, ,,

Erinocystis sp. „ „

Subfamily Scoliocystinae Jaekel emend. Paul

Schizocystis armata (Forbes) „ confluent

Glansicystis baccata (Forbes) „ „

Osculocystis monobrachiolata Paul „ ,,

728

PALAEONTOLOGY VOLUME 11

Family Rhombiferidae Kesling
Rhombifera bohemica

Family Pleurocystitidae
Amecystis laevis (Raymond)

A. sp. nov. (Neebish Channel, Michigan)

Praepleurocystis ivatkinsi (Strimple)

Pleurocystites elegans Billings

P. filitextus Billings

P. squamosus Billings

P. rugeri Salter

P. quadratus (Bather)

P. gibbus (Bather)

Regulaecystis pleurocystoides Dehm

Family Glyptocystitidae Bather
Glyptocystites ehlersi Kesling
G. multiporus Billings
G. regnelli Sinclair

G. batheri Sinclair

Family Callocystitidae Bernard
Cal/ocystites jewetti Flail
C. canadensis Billings
C. snbglobosus (Hall)

Sphaerocystites nmltifascialns Hall

S. bloomfieldensis Schuchert
Hallicystis imago (Hall)

H. elongata Jaekel
H. attenuata Paul
Lepadocystis moorei (Meek)

Brockocystis tecumseth (Billings)
Jaekelocystis hartleyi Schuchert
Lovenicystis angelini (Haeckel)

Apiocystites pentrematoides Forbes
Lepocrinites gebhardii Conrad
Prunocystites fletcheri Forbes
Lipsanocystis traversensis Ehlers & Leighley
L. magnns Stumm

Tetracystis fene stratus Troost

T. chrysalis Schuchert
T. oblongns (Forbes)

T. elegans (Hall)

Strobilocystites calvini White
Staurocystis quadrifasciatus (Pearce)
Pseudocrinites bifasciatns Pearce

P. pyriformis Paul
P. gordoni Schuchert
P. perdeivi Schuchert
P. stellatus Schuchert
P. clarki Schuchert

Superfamily Hemicosmitida Jaekel
Family Hemicosmitidae
Hemicosmites pyriformis von Buch
H. extraneus Eichwald

slits

disjunct

dichopores

discrete

disjunct confluent

conjunct ,,

disjunct ?confluent

99

disjunct

confluent

cryptorhombs throughout

PAUL: DICHOPORITE PORE-STRUCTURES IN CYSTOIDS

729

H. sp. nov. (Kullsberg, Sweden) cryptorhombs throughout

H. malum Pander
H. cf. verrucosus Eichwald
H. sp. (Karrast, Estonia)

H. sp. (Uxnorm, Estonia)

Family Caryocrinitidae Bernard
Caryocrinites ornatus Say
C. roemeri Jaekel
C. septentrionalis Regnell
C. elongatus (Rowley)

Family Thomacystidae Paul
Thomacystis tuberculata Paul

REFERENCES

barrande, j. 1887. Systeme Silurien du centre de !a Boheme. I'' par tie: Recherches paleontologiques.

7. Classe ges Echinodermes. Ordre des Cystides. xvii+233 pp., 39 pi., Leipzig and Prague.
bather, f. a. 1913. Caradocian Cystidea from Girvan. Trans. R. Soc. Edin. 49, 359-530, 6 pi.
breimer, a. 1963. On a specimen of Pleurocystites (Cystoidea) from an erratic boulder near Westerhaar
(Netherlands). Proc. K. ned. Akad. Wet. (B) 66, 296-302, 2 pi.
delpey, g. 1942. Organes speciaux d’Echinodermes primitifs: les pectinirhombes. Bull. Soc. geol. Fr.
[5] 11, 207-17, 5 text-figs.

forbes, e. 1848. On the Cystideae of the Silurian rocks of the British Isles. Mem. geol. Surv. U.K. 2,
483-538, pi. 11-23.

Hudson, G. H. 1911. Studies of some early Siluric Pelmatozoa. Bull. N.Y. St. Mus. 149, 195-258, 7 pi.

1915. Some fundamental types of hydrospires with notes on Porocrinus smithi Grant. Ibid. 177,

163-73, 2 pi.

huxley, T. h. (translator) 1854. On the structure of the echinoderms by Prof. J. Muller. Ann. Mag.

nat. Hist. [2] 13, 1-24, 112-23, 241-56. (See Muller 1854.)
hyman, l.h. 1955. The invertebrates, 4. Echinodermata. The coelomate Bilateria. vi+763 pp. New York,
Toronto, and London.

jaekel, o. 1899. Stammesgeschichte der Pelmatozoen. 1. Thecoidea and Cystoidea. x+442pp., 18 pi.,
88 text-figs. Berlin.

jefferies, R. p. s. and minton, p. 1965. The mode of life of two Jurassic species of ‘ Posidonia'. Palaeon-
tology, 8, 156-85, pi. 19.

kesling, r. v. 1961. A new Glyptocystites from the Middle Ordovician strata of Michigan. Contr.
Mus. Paleont. Univ. Mich. 17, 59-76, 3 pi.

1 962. Morphology and taxonomy of the cystoid Cheirocrinus anatiformis (Hall). Ibid. 18, 1-21 , 4 pi.

1963. Key for the classification of cystoids. Ibid. 101-16.

1968. Cystoids, in R. C. Moore (ed.), Treatise on invertebrate paleontology. Part S. Echinodermata

1, S85-S267.

muller, j. 1854. liber den Bau der Echinodermen. Abh. preuss. Akad. Wiss. (for 1853), 123-219,
9 pi. (See Huxley 1854.)

paul, c. r. c. 1967o. The British Silurian cystoids. Bull. Br. Mus. nat. Hist. (Geol.) 13, 297-356, 10 pi.

19676. Hallicystis attenuata, a new callocystitid cystoid from the Racine Dolomite. Contr. Mus.

Paleont. Univ. Mich. 21, 231-53, 4 pi.

■ 1967c. The functional morphology and mode of life of the cystoid Pleurocystites, E. Billings, 1854.

Symp. zool. Soc. Loud. 20, 105-21, 22 figs.

1968. Macrocystella Callaway, the earliest glyptocystitid cystoid. Palaeontology, 11, 580-600,

pi. 111-13.

regnell, g. 1945. Non-crinoid Pelmatozoa from the Paleozoic of Sweden. A taxonomic study.

Meddn. Lunds geol. -miner. Instn 108, 255 pp., 15 pi.

RUDWiCK, m. J. s. 1960. The feeding mechanisms of spire-bearing fossil brachiopods. Geol. Mag. 97,
369-83, 8 text-figs.

730

PALAEONTOLOGY, VOLUME 11

rudwick, m. j. s. 1961. The feeding mechanism of the Permian brachiopod Prorichthofenia.
Palaeontology, 3, 450-71, pi. 72-4.

1964n. The function of zigzag deflections in the commissures of fossil brachiopods. Palaeontology,

7, 135-71, pi. 21-9.

19646. The inference of function from structure in fossils. Br. J. Phil. Sci. 15, no. 57, 27-40.

Sinclair, G. w. 1948. Three notes on Ordovician cystoids. J. Paleont. 22, 301-14, pi. 42-4.
stainbrook, m. a. 1941. Last of great phylum of the cystids. Pan-Am. Geol. 76, 83-98, pi. 7.

C. R. C. PAUL

Geology Department
Indiana University Northwest
Gary, Indiana 46408
U.S.A.

Typescript received 12 February 1968

AUSTRALIRHYNCHIA, A NEW LOWER
DEVONIAN RHYNCHONELLOID BRACHIOPOD
FROM NEW SOUTH WALES

by N. M. SAVAGE

Abstract. Australirhynchia cudalensis gen. et sp. nov., family Rhynchotrematidae, is described and figured from
the early Siegenian Mandagery Park Formation, New South Wales.

Study of a Lower Devonian brachiopod fauna from the Mandagery Park Formation,
near Manildra, New South Wales, has led to the recognition of a new rhynchonelloid
genus, Australirhynchia. The Mandagery Park Formation consists of interbedded lime-
stones and tuffaceous sandstones of probable early Siegenian age (Savage 1967, 1968).
A silicified horizon within the basal limestone of the formation, 3 miles south of Manil-
dra, at locality 1 of Savage (1968), contains a rich brachiopod fauna from which the
material described herein has been selected.

SYSTEMATIC PALAEONTOLOGY

Superfamily rhynchonellacea Gray 1 848
Family rhynchotrematidae Schuchert 1913
Subfamily orthorhynchulinae Cooper 1956
Genus australirhynchia nov.

Type species. Australirhynchia cudalensis sp. nov.

Diagnosis. A plicate rhynchonelloid with strongly angular costae which multiply by
bifurcation on the dorsal fold and intercalation in the ventral sulcus. The dorsal interior
has a small boss-like cardinal process between ponderous inner socket ridges. No median
septum is present. The ventral interior is without distinct dental lamellae and the long
heavy teeth, which are fused to the valve walls, bear deep longitudinal grooves.

Comparison. The ponderous inner socket ridges and small cardinal process, together
with the massive, longitudinally grooved teeth and absence of distinct dental lamellae,
suggest a relationship to the genera grouped as the Orthorhynchulinae by Schmidt
(1965).

Of genera possessing a cardinal process but without a septalium the nearest is Macha-
eraria Cooper 1955, for this genus also lacks a dorsal median septum and has similarly
long curved crura, crescentic in cross-section; moreover, there is a suggestion of bifurca-
tion on the dorsal umbones of the specimens figured by Cooper, though this feature
is much less prominent and more variable than in Australirhynchia. There are, however,
several major differences; the outline of Machaeraria is more triangular with the greatest
width anterior, and not posterior, of mid-length, the shell contours are more rounded,
and the costae are more numerous. Internally the hinge plate of Machaeraria is deeply

[Palaeontology, Vol. 11, Part 5, 1968, pp. 731-5, pi. 141.]

732 PALAEONTOLOGY, VOLUME 11

cleft, the inner socket ridges are much smaller, and the teeth are less massive with no
deep longitudinal grooves.

Another closely related genus is Sicorhyncha Havlicek 1961, from the Lower Devonian
of Bohemia. This resembles Australirhynchia in lacking a septalium and a dorsal median
septum. However, it differs greatly in general shell outline, being bluntly rounded posteri-
orly and very broad anteriorly with concave lateral margins. Furthermore, it does not
possess the strongly inflated inner socket ridges and massive teeth of Australirhynchia.

Latonotoechia Havlicek 1961, also from the Lower Devonian of Bohemia, is another
genus without a septalium and a dorsal median septum. In addition, it has massive
teeth and inflated inner socket ridges (Havlicek 1961, p. 25, text-fig. 1). It is easily dis-
tinguished from the Manildra genus, however, for it has rounded contours quite unlike
those of Australirhynchia and lacks the bifurcating and intercalating costae.

Of the genera referred by Schmidt (1965) to the Rhynchotrematinae, the Lower
Silurian genus Stegerhynchus Foerste 1909, and the upper Silurian genus Stegorhynchella
Rzhonsnitskaya 1959, both show a general resemblance to Australirhynchia. The
Manildra form is distinct from these genera because of its unusual ornamentation and
the absence of a median septum or callosity supporting the hinge plates. Moreover,
both Stegerhynchus and Stegorhynchella have distinct umbonal cavities in their
ventral valves, a feature completely absent from Australirhynchia where dental
lamellae cannot be distinguished in any of the specimens examined.

Australirhynchia cudalensis sp. nov.

Plate 141

Material. Of a total of 167 silicified specimens 146 are complete shells with the valves conjoined. The
remaining few specimens consist of 9 dorsal valves, 8 ventral valves, and 4 posterior fragments showing

EXPLANATION OF PLATE 141
Australirhynchia cudalensis gen. et sp. nov.

(All figures x 6)

Figs. 1-4. Dorsal, ventral, posterior, and lateral views of SU 19579, a large specimen with a distinctly
pentagonal outline, an emarginate anterior margin, and an anacline ventral interarea.

Figs. 5-9. Ventral, dorsal, lateral, anterior, and posterior views of SU 19578 (holotype).

Fig. 10. Ventro-anterior view of broken dorsal valve SU 19584 showing the small, rounded, posterior
adductor scars and the larger, ovate, anterior adductor scars.

Figs. 11, 12. Dorso-anterior and dorsal views of specimen SU 19581, a ventral valve with the socket
plates and cardinalia of the dorsal valve still attached to show the articulation. The small, rounded,
adductor scar is prominent within the long, faintly impressed, diductor field.

Fig. 13. Dorsal view of broken ventral valve SU 19585 showing the deep delthyrial cavity bounded by
long, thick, deeply grooved teeth from which the distal parts are broken.

Figs. 14-18. Dorsal, lateral, ventral, anterior, and posterior views of young specimen SU 19577. Here
the maximum thickness is more posteriorly placed than in older specimens and the ventral interarea
is distinctly apsacline.

Fig. 19. Ventral view of dorsal valve SU 19580 showing the small cardinal process between large inflated
inner socket ridges, and the elongate adductor muscle field.

Fig. 20. Posterior view of specimen SU 19580 showing further the inflated inner socket ridges.

Figs. 21-5. Posterior, lateral, anterior, dorsal, and ventral views of specimen SU 19576, a young stage
in which the bifurcation of the two central costae of the dorsal valve and the intercalation of two
new costae in the ventral valve have just occurred. The fold and sulcus are poorly developed at this
stage.

Palaeontology, Vol. 11

PLATE 141

18

19

25

24

23

SAVAGE, Australirhynchia

N. M. SAVAGE: AUSTRALIRHYNCHIA GEN. NOV. 733

the valve articulation. The specimen numbers used are those of the Palaeontology Collection, Depart-
ment of Geology and Geophysics, University of Sydney.

Holotype. Complete shell SU 19578.

Description. Exterior. In outline the shell is transversely ovate to subpentagonal with the
greatest width at the hinge line. The cardinal margins are sharply rounded, the lateral
margins are gently rounded, and the anterior margin is straight or emarginate. In lateral
profile the shell is subequally biconvex with the greatest thickness between mid-length
and the anterior margin (PI. 141, figs. 4, 7).

The ventral valve is moderately convex with a strongly curved umbo and a prominent,
suberect beak (PI. 141, fig. 4). An apsacline to anacline interarea has a width about half
the shell width and an apical angle of about 1 10°. The delthyrium includes an angle of

text-fig. 1. Australirhynchia cudalensis gen. et sp. nov. A. Interior of the dorsal valve showing the
small cardinal process between the large inflated inner socket ridges. Note also the elongate adductor
muscle field, b. Interior of the ventral valve showing the long, deeply grooved teeth and the small,
rounded adductor field set within the larger diductor muscle field.

about 80° and is partly closed by small triangular deltidial plates. These roof over the
dorsal umbo and almost meet to leave an oval foramen extending to the posterior
extremity of the valve (PI. 141, figs. 1, 6). The dorsal valve is strongly convex with a broad
umbo and no interarea.

A broad ventral sulcus extends most of the valve length to the rectangularly uniplicate
anterior commissure where it is bordered by very steep lateral walls (PI. 141, figs. 5, 8).
The dorsal fold is less pronounced posteriorly but very distinct anteriorly (PI. 141, fig. 8).
About 10-12 strong angular costae are present on the dorsal valve, with 4 on the fold,
and about 9-11 costae are present on the ventral valve, with 3 in the sulcus. The costae
on the fold invariably bifurcate to increase from 2 to 4 (PI. 141, figs. 1,6, 14, 24), whilst
the single initial costa in the sulcus increases to 3 by an intercalation on either side
(PI. 141, figs. 2, 5, 16, 25). In both valves the increase occurs at about the 1-5 mm. growth
stage. Numerous faint concentric growth lines cross the costae.

Interior of ventral valve. The ventral interior has a deep, narrow, delthyrial cavity
bounded by long thick teeth, each bearing a prominent longitudinal groove with which
the corresponding dorsal inner socket ridge articulates (PI. 141, fig. 1 1). The teeth diverge
anteriorly at about 45°. A long and anteriorly expanding diductor muscle field extends
almost half the valve length (text-fig. 1b). The more strongly impressed adductor field

734

PALAEONTOLOGY, VOLUME 11

is very small and rounded, with a length about one-quarter that of the diductor field in
which it is almost centrally placed (PI. 141, fig. 12).

Interior of dorsal valve. The dorsal interior has a small boss-like cardinal process placed
between large inflated inner socket ridges (PI. 141, fig. 19). The sockets, which are small,
triangular, and widely divergent, are bounded postero-laterally by the low overhanging
valve margin. Thin, ventrally curved crura, which arise from the inside of the socket
ridges, are slightly crescentic in cross-section with the convex side directed outwards
(PI. 141, fig. 19). No median septum is present. The adductor muscle scars are clearly

-o 5-
£

i 5-

0< , , , , , , ■

0 5 10

Length

04 . 1 . . . . . . . .

0 5 10

Length

A

B

text-fig. 2. Australirhynchia cudalensis gen. et sp. nov. Scatter diagrams of the dimensions of 60 speci-
mens plotted in mm. a. Plot of width to length, b. Plot of thickness to length.

impressed and extend about half the distance to the anterior margin (text-fig. 1a).
Rounded, widely spaced, posterior adductors are separated from larger, longitudinally
ovate anterior adductors by narrow, transverse ridges (PI. 141, fig. 10).

Measurements (in mm.)

Length

Width

Thickness

SU 19577

Complete shell

3-5

4-1

2-6

SU 19578

Complete shell

4-5

5-4

40

SU 19579

Complete shell

5-3

7-3

4-6

SU 19580

Dorsal valve

6-5

8-6

—

SU 19581

Ventral valve

60

7-6

—

Variation. Shells display little variation in external form apart from differences in thick-
ness and width related principally to ontogeny (text-fig. 2). The multiplication of the
costae by bifurcation on the fold and intercalation in the sulcus is invariably present and
shows remarkably little variation. The costation on the flanks also shows a low vari-
ability so that in most shells the dorsal valve possesses 10 costae and the ventral valve
11 costae. Internally even the younger specimens show the massive teeth and socket
ridges.

Ontogeny. A good range of growth stages has been collected and the later part of the
ontogeny of the species can be followed in detail. In the youngest stage available the

N. M. SAVAGE: AUSTRALIRHYNCHIA GEN. NOV.

735

maximum width occurs just anterior to mid-length, no bifurcation or intercalation of
costae has occurred, and the ventral interarea is apsacline. At a slightly larger stage
(PI. 141, figs. 24, 25) the bifurcation of the two central costae of the dorsal valve and the
intercalation of costae on either side of the central costa of the ventral valve is already
apparent. As the shell approaches maturity it assumes a more pentagonal outline with
the maximum width posterior to mid-length (PI. 141, fig. 6). In a mature specimen the
ventral interarea is distinctly concave and more anacline than apsacline. The maximum
thickness of a mature specimen is well towards the anterior of the shell, largely because
of the strong development of the dorsal fold and ventral sulcus (PI. 141, figs. 7, 8).
Internally, large inflated inner socket ridges and deeply grooved teeth are present in all
but the youngest stages. In mature specimens the adductor muscle scars in both valves
are deeply impressed.

Phylogeny. Although Australirhynchia is distinct from Stegerhynchus and Machaeraria
in several important respects, these two genera appear to be the most closely related
forms. It seems possible that both Australirhynchia and Machaeraria evolved indepen-
dently from Stegerhynchus , although there is little evidence for this apart from the general
morphological similarities mentioned above and the absence of possible alternative
precursors during the Silurian. It is unlikely that Australirhynchia was derived from
Machaeraria as it retains such features as inflated inner socket ridges and a relatively
angular shell shape; features characteristic of Stegerhynchus but not present in Macha-
eraria.

Acknowledgements. The author would like to acknowledge the assistance of Professor C. E. Marshall
and Dr. G. H. Packham of the Department of Geology and Geophysics, University of Sydney, where
this work was commenced as part of a larger faunal study during the tenure of a Teaching Fellowship.
It is also a pleasure to thank Professor F. H. T. Rhodes for his encouragement and for the use of facilities
in the Department of Geology, University College of Swansea; also Dr. A. J. Boucot and Dr. J. G.
Johnson of the California Institute of Technology for their helpful comments.

REFERENCES

cooper, G. a. 1955. New genera of Middle Paleozoic brachiopods. /. Paleont. 29, 45-63, pi. 11-14.
havlIcek, v. 1961. Rhynchonelloidea des bohmischen alteren Paleozoikums (Brachiopoda). Rozpr.
ustred. Ust. geol. 27, 1-211, 27 pi.

savage, n. m. 1967. Studies in the Silurian and Devonian of the Manildra District, New South Wales.
Unpublished Ph.D. thesis, Univ. Sydney.

— — 1968. The Geology of the Manildra District, New South Wales. J. Proc. R. Soc. N.S. W. (in press).
schmidt, h. 1965. Family Rhynchotrematidae. In R. C. Moore (ed.), Treatise on invertebrate paleon-
tology, Part H. Brachiopoda, H572-6, Fawrence, Kansas.

N. M. SAVAGE

Department of Geology
University of Natal
Durban S. Africa

Typescript received from author 1 April 1968

THE LLANDOVERY TRANSGRESSION OF
THE WELSH BORDERLAND

by A. M. ZIEGLER, L. R. M. COCKS, and W. S. McKERROW

Abstract. The study of the evolution of several brachiopod genera has enabled the early Silurian shelf sequences
of the Welsh Borderland to be correlated with the type area of Llandovery, and, to some extent, with the grap-
tolite zonal sequence. The transgression across the Borderland started at the beginning of Llandovery time,
when areas in Montgomeryshire became inundated after a short break in deposition at the end of the Ashgill.
The sea reached Shropshire, and possibly as far east as the Malverns and May Hill, by Middle Llandovery times.
By late Upper Llandovery times, much of the English Midlands were covered.

Fossil communities indicate the relative depths in which the Llandovery sediments were deposited. With the
advance of the sea, most sequences show a progressive increase of depth with time, although minor reversals
are known. The many gaps in the local sequences are probably due to submarine erosion or non-deposition,
rather than to uplift and sub-aerial erosion, because they are characteristically followed by progressively
deeper-water communities. The community distribution indicates that a continuous gradient was maintained
from the coast to the shelf margin, once the topographic relief of the original surface had been filled in.

During the Lower Silurian there was a spread of the shelf sea from Central Wales
over the Welsh Borderland (text-fig. 1). Brachiopods are the dominant element in the
shelf faunas of this region. Now that the evolution of several brachiopod genera is
known in detail, it is possible to correlate the shelly deposits with the standard section
at Llandovery.

Animal communities have been defined for this period, and they have been shown to
reflect the depth of the sea. These communities are diachronous, and migrated as the
transgression advanced.

The twin concepts, of evolving lineages as a basis for correlation and of animal
communities as a basis for interpreting the environment, provide the foundation of this
work.

BASIS OF CORRELATION

The type section of the Llandovery is to be seen south-east of the town of this name
in Wales (Jones 1925). Brachiopods are the dominant faunal element in the area, but
correlation of other shelly areas with the type section has had to await the study of the
commoner brachiopod lineages. This approach contrasts with the tendency for workers
in the shelf areas to have correlated Silurian beds by the assemblages of species present.

Evolutionary sequences are now known for the stricklandiids (Williams 1951, St.
Joseph 1935), the pentamerids (St. Joseph 1938), the atrypide Eocoelia (Ziegler 19666)
and the strophomenide Leptostroplria (Cocks 1967#).

Some brachiopod lineages and their relationship with the type Llandovery are shown
in text-fig. 2. The only brachiopods on this list which do not occur at Llandovery itself
are E. sulcata, C. lirata typica, and Pentameroicles; their assignment to C6 is based on the
fact that they are known to occur between C5 and basal Wenlock beds in adjacent areas
of the Welsh Borderland.

Most of the Lower Silurian of the British Isles carries a fauna of graptolites. Lapworth
(1878) and Elies and Wood (1901-18) established a sequence of graptolite zones which

[Palaeontology, Vol. 11, Part 5, 1968, pp. 736-82.]

ZIEGLER, COCKS, AND McKERROW: LLANDOVERY TRANSGRESSION 737

have provided an accepted basis for correlation. A few graptolites of zonal value have
been found at Llandovery (text-fig. 2) but are only recorded from three horizons. Jones
(1925, p. 360) found graptolites referable to his cicinaces Zone in A4 (this is now equated
with the lower part of the cyphus Zone — Toghill 1968): Jones also (1949, p. 52) cor-

text-fig. 1. Location map of the Llandovery rocks of Wales and the Welsh Borderland, showing

traverse lines and position of local sections.

relates B3 with the C. cometa Band, which is in the upper part of the convolutus Zone;
and he further records graptolites of the sedgwickii Zone (1925, p. 370) from beds
assigned to Cx. This last fauna has been confirmed by our recent collecting from the
other known locality of this restricted facies near an old ford across the River Sefin
(Grid Ref. SN/7419 2811).

738

PALAEONTOLOGY, VOLUME 11

In addition to these finds from Llandovery itself, other links between the brachiopod
and graptolite sequences have been determined in the upper half of the Llandovery of
Shropshire (Cocks and Rickards, in press); but, as yet, firm ties are not proved for the
whole Llandovery succession.

Brachiopod Graptolite

Lineages Zones

Wenlock

1

Eocoelia t

Cyrtograptus

murchisoni

c6

sulcata* Costistncklandia

1 irata . typica *

* l ' Prntnn

Monoclimacis

crenulata

Monoclimacis

griestoniensis

Monograptus

crispus

Monograptus

turriculatus

Monograptus

sedgwickii

Monograptus

convolutus

Pristiograptus

gregarius

Pristiograptus

cyphus

Cystograptus

vesiculosus

Akidograptus

acuminatus

Gl yptograptus
persculptus

| ^ rGnTQr

Eocoelia Costistncklandia

curtisi lirata, alpha

1 4.

>er

n

-p*

"1 ^ 1
Eocoelia Stricklcndia

u.

~> ^3

intermedia lens ultima

1 1

C2

Pentar

Eocoelia Stricklandia

nerus

C,

■s.

hemisphaerica lens

progressa

*

* O B3

*

° *o R

T3 T3 D,
c —

Strickla

lens

nd ia

ntermedia

A4

Strickla

ndia

0) A3
> .

lens typica

l

Lov

>

ro

Stricklandia
lens prima

A,

Ashgill

Dicellograptus

anceps

text-fig. 2. The correlation of brachiopod lineages and graptolite zones with the standard sequence
at Llandovery. The asterisks against some brachiopods indicate that they have not been found at
Llandovery; the tie-arrows indicate graptolite finds at Llandovery.

The limits of the Llandovery Series at Llandovery have not yet been properly defined,
nor have they been firmly correlated with the sequence of graptolites. However, in this
paper, we take the base of the series as equivalent to the base of the Glyptograptus
persculptus Zone, and the top of the series at the base of the Wenlock; or as equivalent

ZIEGLER, COCKS, AND McKERROW: LLANDOVERY TRANSGRESSION 739

to the base of the C. murchisoni Zone, in the sense of Elies and Wood (1901-18). This
last zone is sometimes now divided into a lower Cyrtograptus centrifugus Zone and an
upper, restricted, C. murchisoni Zone.

DISTRIBUTION OF ANIMAL COMMUNITIES

By means of large collections from single beds (many with over a thousand specimens),
it is possible to quantify the proportions of species present at each locality.

Five main benthic animal communities have been defined by this means from the
Llandovery (Ziegler 1965, Cocks 19676, Ziegler, Cocks, and Bambach 1968). Each
community is characterized by a particular faunal assemblage, consisting chiefly of
brachiopods, and each has been named after a prominent brachiopod genus in the
assemblage. They are the Lingula Community, the Eocoelia Community, the Penta-
merus Community, the Strieklandia Community, and the Clorinda Community. At
any one time, they occurred in this sequence, from the inferred coastline to the outer
margin of the shelf. In beds of C5 and C6 age, Pentamerus and Strieklandia have evolved
into Pentameroides and Costistricklandia respectively, and the community names change
correspondingly.

So far the communities have been described only from the Upper Llandovery, but,
apart from Eocoelia, the characteristic genera also occur in the Middle Llandovery, and
the Strieklandia and Clorinda Communities are present in Lower Llandovery beds.
A Cryptothyrella Community occurs in the Lower and Middle Llandovery with associ-
ates of the Eocoelia Community, and it appears to be an early equivalent of the latter
community; it is abundant in basal beds at several areas, and occurs also in North
America (Ziegler and Boucot, in press).

In addition, during parts of the Middle and Upper Llandovery, the position of the
Eocoelia Community in Shropshire is taken by the parallel ? Rostricel/ula Community
(Cocks and Rickards, in press). Also, in places near the edge of the shelf, the Clorinda
Community dwindles to a sparser 'Marginal Clorinda' Community, with many typical
members absent.

Apart from the burrowing Lingula, all the Llandovery brachiopods were epifaunal.
Thus the majority of Silurian level bottom communities contrast with equivalent modern
communities which are mainly infaunal. These Silurian communities were therefore less
dependent on the type of substrate than their modern counterparts.

Rocky bottom communities are known at the base of some local sequences, although
in all cases where these have been preserved they are accompanied by some soft-bottom
elements.

The distribution of the main communities at three horizons in the Llandovery is shown
in three maps at the end of the paper (text-figs. 12-14). While the actual depths repre-
sented by the communities are uncertain, the lines bounding them are presumed to
parallel the bathymetric contours. Thus we have a picture of the changing configuration
of the shelf sea during the transgression, and of the migration of the communities with
time.

A distinction is drawn between ‘basin’ and ‘graptolitic’ facies. The former is restricted
to areas (for example text-fig. 3, cols. 1 and 10) where rapid sediment supply, great

C 6055 3 c

740

PALAEONTOLOGY, VOLUME 11

water depths, and considerable subsidence has resulted in large thicknesses of sediment.
The term ‘graptolitic’ facies is used for rocks in which the dominant fossils are grapto-
lites, and in which shelly forms, apart from a few orthocones and epiplanktonic bivalves,
are generally absent. This facies represents one or other of two environments, firstly
deposition under water which was too deep for benthic colonization at that time, or,
secondly, water of any depth which had a foul bottom, lower temperature, or other
physical conditions inimical to a shelly fauna. When a graptolitic facies occurs to sea-
ward of a Clorinda Community, we feel justified in selecting the first of these alternatives,
and interpreting the occurrence as being too deep for a shelly fauna ; this situation applies
to all of the Llandovery in the Welsh Borderland. However the second alternative
certainly applied at other times and places during the Lower Palaeozoic, for example
this is how we would interpret the Wenlock Shale to the south of the Long Mynd,
Shropshire.

Acknowledgements. We are indebted to the late Professor W. F. Whittard for his advice, fossil collec-
tions, and loan of field maps. Dr. M. L. K. Curtis kindly made his thesis and collections available, and
conducted us to localities in the Tortworth Inlier. Drs. W. B. N. Berry, R. B. Rickards, and P. Toghill
identified the graptolites we have found, and discussed older records.

The theses by Ziegler (1963) and Cocks (1965), both at Oxford, were supported in part by the
Burdett-Coutts Fund and the Department of Scientific and Industrial Research respectively. We are
grateful to Mr. J. M. Edmonds and Mr. H. P. Powell for housing our thesis collections in the Oxford
University Museum (Appendix 1), and for providing funds for their labelling. Collections made
subsequently are deposited in the British Museum (Natural History) (Appendix 2) and the U.S. National
Museum (Appendix 3). These last were prepared by Ziegler in the laboratories of Dr. A. J. Boucot at
the California Institute of Technology from 1964-6. More recently, Ziegler has been supported by
Grant 910-G from the Petroleum Research Fund, administered by the American Chemical Society;
and by Grant GB-6592 from the National Science Foundation of America.

DESCRIPTIONS OF THE OUTCROPS

Three traverses across part of Wales and the Borderland (text-fig. 1) are represented
by the composite sections in text-fig. 3.

In the following account, the areas studied in detail are treated first; i.e. the sections
from Minsterley to Tortworth (text-fig. 3, cols. 5-6, 11-19, and 25-9). These constitute
the classical Welsh Borderland area. Secondly, we discuss the small inkers to the east
of the Borderland (text-fig. 3, cols. 7-9 and 20). We have examined collections from these
areas, and include them for the sake of completeness. Finally we contrast these shelf
sequences with areas to the west of the Borderland, which in Llandovery times formed
the shelf margin and the basin. These areas are mostly in Wales (text-fig. 3, cols. 1-4,
10, and 21-3); we have collected in all of them, with the exception of the purely grapto-
litic sequences.

text-fig. 3. Three traverses across the Welsh Borderland (see text-fig. 1) showing local sections with
communities, correlation, lithology, and stratigraphical relationships. True thicknesses are shown. The
distances between columns are not the straight-line distances, but the N. 60° W. component of these

distances.

742

PALAEONTOLOGY, VOLUME 11

MINSTERLEY AND BOG
Text-fig. 3, cols. 5 and 11

Exposures of Llandovery sediments to the north and west of the Shelve Inlier, Shrop-
shire (text-fig. 4), occur in a sinuous strip running north-east from north of Chirbury
through Hope Valley and Venusbank to near Minsterley. Various outliers, including

text-fig. 4. Locality map of Shropshire (after many authors, with modifications).

those at Bog, rest unconformably on the Ordovician of the Shelve Inlier. The area has
been re-mapped, but the only major re-interpretation is based on new exposures indicating
that the Venusbank ‘outlier’ is connected northwards with the outcrop in Hope Valley.

Whittard (1932) used the same stratigraphical names, i.e. Pentamerus Beds and Purple
Shale, throughout Shropshire, but the Llandovery rocks of the Minsterley and Bog area
differ in lithology and age from the beds in the Wenlock Edge outcrop (see below).
Accordingly we propose new names: the Bog Quartzite, Venusbank Lormation, and
Minsterley Lormation.

(a) Venusbank Formation and Bog Quartzite. The type section of the Venusbank Lorma-
tion is in Hope Brook, running eastwards from near Hope Quarry. The base is exposed
in both brook (SJ/3554 0207) and quarry (SJ/3550 0207), where a sandstone with
occasional shale fragments rests on Hope Shales of Ordovician (Llanvirn) age. The top
of the Lormation in Hope Brook is taken above the highest of the dominant sandstones
(SJ/3572 0215).

ZIEGLER, COCKS, AND McKERROW: LLANDOVERY TRANSGRESSION 743

The base of the formation may also be seen in an outlier (SJ/3458 0146) two-thirds
of a mile south-west of Hope Quarry (Tyler 1925). Here a basal conglomerate is present,
in which pebbles (up to 5 cm.) of Hope Shales, Stapeley Volcanics (Llanvirn) and
Stiperstones Quartzite (Arenig) occur. This outlier occurs at 900 ft. O.D. (Ordnance
Datum — feet above sea level) but another small outlier is present 1 50 yards to the east,
on the side of the valley and between 650 and 700 ft. O.D. ; Hope Quarry lies at 600 ft.
O.D., and, to the south-east, the Llandovery at Venusbank (SJ/3535 0125) is at 800 ft.
O.D. Thus the conclusions reached by Whittard (1932, pp. 893-5), of an irregular base
to the Llandovery of the area, are confirmed; steep dips are absent and there is no
detectable subsequent faulting. It appears that the present-day Hope Valley has been
re-excavated along the line of a pre-Middle Llandovery depression, and that the Llan-
dovery sea spread over a surface relief of 200-300 ft. in this area. The thickness of the
Venusbank Formation varies between 0 and 200 ft. (0-65 m.).

The formation is well exposed at Hope Quarry (SJ/3550 0207), where a thickness of
30 ft. (10 m.) is visible. Sandstone bands up to 3 ft. (1 m.) thick occur there, sometimes
including shells or mudflakes at or near their base, with each unit fining up into unfossili-
ferous silts. Some of the sandstones have parallel laminae throughout, but cross-bedding
is absent, apart from very occasional ripple cross-lamination on the tops of one or two
sandstones. Unit bases are sharp, with no visible bottom structures apart from worm
traces. Low-angle channelling occurs with occasional silty partings at the base of the
channelling sandstone. These sediments may represent proximal turbidites filling a
previously scoured channel of Pentamerus- and Stricklcmclia- Community depth.

The Llandovery also rests unconformably on the Ordovician in some poorly exposed
outliers near Bog (SO/355 979) and at Round Hill (SO/348 993). Many large loose
blocks of coarse quartzite are present; often containing pebbles and shale chips. These
beds are termed the Bog Quartzite.

Fossil communities. The Hope Valley sections of the Venusbank Formation have mainly
yielded fossils of the Stricklandia Community (Locs. 68-70, 72, 73) ; the exceptions are the
presence of Pentamerus Communities at some places near the base of the formation
(Loc. 74 and at SJ/3555 0107). This suggests that the irregular landscape was covered
by a moderately deep shelf sea ( Pentamerus Community) that soon became slightly
deeper ( Stricklandia Community). However, there may have been some fluctuation in
depth, as Whittard (1932, p. 876) reports Pentamerus above Stricklandia- bearing beds
in Hope Quarry.

In Ox Wood Dingle (Loc. 67), 4 miles west of Hope Valley, the Stricklandia
Community is also present. However, early shallow-water conditions may be indicated
at The Stubbs (2 miles north-west of Hope Quarry — SJ/324 033) where Whittard
records (1932, p. 874) Lingula in growth position near the base of the Venusbank
Formation.

The fauna of the Bog Quartzite (Locs. 65, 66) is a mixture of a near-shore Crypto-
thyrella Community with some rocky-bottom elements (Ziegler, Cocks, and Bambach
1968).

Correlation. All the collections from the Venusbank Formation and Bog Quartzite have
yielded Stricklandia ; the subspecies present at Bog is S. lens intermedia indicating a

744

PALAEONTOLOGY, VOLUME 11

Middle Llandovery age; the subspecies from the Venusbank Formation is at the bound-
ary between S. lens intermedia and progressa and is B3 to Q in age. A collection from
near the base of Hope Quarry (Loc. 70) includes the graptolite Climacograptus aff.
rectangularis (identified by R. B. Rickards), which indicates a Middle Llandovery age
for the lower part of the formation. Whittard (1932, facing p. 896) records Monograptus
runcinatus pertinax from the ‘Pentamerus Beds of the Wilmington-Minsterley’ area;
this suggests that at least some of the Venusbank Formation is of early Upper Llandovery
age.

( b ) Minsterley Formation. The type section goes diagonally across the strike of the
formation and is in Hope Brook. The base is taken above the top sandstone of the Venus-
bank Formation (SJ/3572 0215), and the top above the last purplish bed in the exposure
to the south of Wagbeach (SJ/3642 0270). The thickness of the Minsterley Formation
is about 400 ft.

The lowest beds seen are calcareous siltstones and mudstones, with occasional
coarser beds; these represent a distinct sedimentary environment unknown elsewhere in
the Welsh Borderland. They pass up into maroon, green, or blue mudstones, the top
150 ft. (49 m.) being of the same dominantly purple colour as the Hughley (Purple)
Shales on the main Wenlock Edge outcrop. The mudstones contain some interbedded
siltstone and bands of fragmented shells. Locally, the Minsterley Formation overlaps
the Venusbank Formation, e.g. at Estell (Whittard 1932, p. 877) and east of Minsterley;
in these areas the formation tends to be coarser. The Minsterley Formation is followed
by several hundred feet of unfossiliferous turbidites, well exposed in Flope Valley,
which may be either late Llandovery or early Wenlock in age.

Fossil communities. Collections in Hope Brook (Loc. 71) and Minsterley-Habberley
Lane (Loc. 75) are both in the Clorinda Community; this indicates that the Minsterley
Formation was deposited in deeper water than the underlying Venusbank Formation.

Correlation. The brachiopods collected from the Minsterley Formation provide no
precise evidence of age, but Whittard’s (1932, p. 877) record of graptolites from Locality
75 includes Monograptus halli, M. becki, and M. cf. proteus. This assemblage indicates
a turriculatus Zone age (Cocks and Rickards, in press). A lack of fossils in the upper part
of the formation, and in the overlying turbidites, makes it uncertain whether it ter-
minates, even approximately, at the end of Llandovery time.

NORBURY AND CHURCH STRETTON
Text-fig. 3, cols. 12 and 13

A long winding outcrop of Llandovery rocks borders the south side of the Cambrian
and Ordovician rocks of the Shelve Inlier, and continues all around the southern part
of the Pre-Cambrian rocks of the Long Mynd (text-fig. 4). The formation names which
we use for this area are the same as those used by Greig et al. (1968); the Pentamerus
Beds and the Hughley (or Purple) Shales. The area, sometimes termed the southern
Long Mynd-Shelve Outcrop, was mapped by Whittard (1932, pi. 59, 60) and more
recently by the Geological Survey (Church Stretton Sheet, New Series, no. 166). Six

ZIEGLER, COCKS, AND McKERROW: LLANDOVERY TRANSGRESSION 745

boreholes were drilled to assist the Survey in their mapping, and four of these passed
through beds of Llandovery age (Cocks and Rickards, in press). These are called the
Eaton Farm (text-fig. 3, col. 12) and Robury Ring Boreholes, both of which are to the
west of the southern Long Mynd spur, and the Hamperley (text-fig. 3, col. 13) and
Springbank Farm Boreholes, both of which are to the east of the Long Mynd spur and
in the Church Stretton Valley. A fifth borehole, at Botvyle (also shown on text-fig. 4),
did not penetrate to the local base of the Wenlock.

(a) Pentamerus Beds. Between Plowden (SO/382 874) and Little Stretton (SO/445 921),
up to 60 ft. (20 m.) of conglomerates are intermittently present below mudstones and
other sediments with Pentamerus. These beds, when fossiliferous, contain a Lingula
Community, and, in parts, the formation name Kenley Grit (used in the adjacent Wen-
lock Edge outcrop — see below) might be employed. However, as the conglomerates fill
quite small topographic pockets, and locally vary both in lithology and possible age,
we follow Whittard (1932) and the Geological Survey (Greig et al. 1968) in treating these
beds as a local coarse base to the Pentamerus Beds. The Hamperley and Springbank
Farm Boreholes both revealed complete sections through the Pentamerus Beds, which
are 480 (157 m.) and 170 ft. (56 m.) thick in the respective cores (although these thick-
nesses may be affected by minor faulting).

Whittard (1932) interpreted the basal deposits to the south of the Long Mynd as a
beach with sea stacks. However, from faunal and sedimentological considerations, it
seems more probable that most of the sedimentation occurred sub-tidally, and certainly
a fairly deep-water Pentamerus Community was established at Hillend Farm (Loc. 62),
only 20 ft. (7 m.) above the unconformity with the Pre-Cambrian.

The Pentamerus Beds are absent in the large embayment to the west of the southern
Long Mynd spur (text-fig. 3, col. 12) and also south of The Roveries (SO/323 915);
at both places Hughley Shale rests directly upon the unconformity. However, between
these two places, at Norbury (Locs. 63 and 64), a local lens of calcareous sandstone,
often crowded with Pentamerus, is present between the unconformity and the Hughley
Shale, and this lens is termed ‘Pentamerus Beds’. Further west, near Snead, Whittard
(1932, p. 871) recorded 280 ft. of Pentamerus Beds, but his description shows that the
lowest Silurian rocks contain a limited shelly fauna, perhaps representing an environ-
ment deeper than the Clorinda Community.

Fossil Communities. In the Hamperley Borehole, the Lingula Community is present
between 65 and 115 ft. (21-38 m.) above the base (Cocks and Rickards, in press).
This is followed by a IRostricellula Community, which is a parallel community to that
of Eocoelia, and occupies a similar position, between Lingula and Pentamerus Communi-
ties. In the borehole, a Pentamerus Community succeeds the IRostricellula Community,
and is itself succeeded by a Stricklandia Community. After the Stricklandia Community,
there was a reversion to the Pentamerus Community immediately below the Hughley
Shales. In the Springbank Farm Borehole, the Pentamerus Beds carry a Pentamerus
Community throughout, although at approximately two-thirds up the formation, the
proportion of Stricklandia increases enough to be termed a Pentamerus! Stricklandia
mixture for that part of the core. This mixture was probably contemporary with the
Stricklandia Community in the Hamperley Borehole.

746

PALAEONTOLOGY, VOLUME 11

In the surface exposures, the Lingula Community is developed in various parts of the
basal coarse facies. At Marshbrook (Loc. 61) we have collected a mixed Lingula /? Ros-
tricellula Community. IRostricellula is a close homeomorph of Eocoelia and was listed
as the latter by Whittard (1932, p. 864). Pentamerus Communities occur at Hillend Farm
(Loc. 62) and at Norbury (Locs. 63 and 64). Further to the west, deeper water was pre-
sent near Snead; above a basal Pentamerus Community, Whittard (1932, p. 872) records
a fauna which we interpret as a ‘Marginal’ Clorinda Community.

Correlation. Graptolites from the Hamperley Borehole indicate the convolutus Zone
for the middle part of the formation, with the sedgwickii Zone above. The sedgwickii
Zone is also proved for the middle of the Springbank Farm core. The brachiopod
collections from Hillend Farm and Norbury both contain Eocoelia, but the former
locality carries E. hemisphaerica (indicating C4_2 age), while at Norbury E. intermedia
is present (Ziegler 1966 b), and indicates a C3_4 age for beds near the local base of
the formation.

Whittard (1932, p. 872) records Climacograptus sp., Monograptus halli and M.
dextrorsus from Snead; these indicate the turriculatus Zone. This evidence, together with
Whittard’s statement (1932, p. 873) that ‘there is a gradation along strike from arena-
ceous to argillaceous sediments’, points to a fairly rapid increase in depth westwards
from the Long Mynd during C3_4 times. It also emphasizes that both base and top of
the Pentamerus Beds are diachronous.

( b ) Hughley (or Purple) Shales. These beds are not well exposed south of the Longmynd-
Shelve Inlier. The Hamperley Borehole shows a thickness of 1 50 ft. (49 m.) and Whittard
(1932, p. 866) estimated a thickness of 200-30 ft. of Hughley Shales to the south of the
Long Mynd. To the west of the Long Mynd spur there is a greater thickness; 560 ft.
(184 m.) is seen in the Eaton Farm borehole, although small faults may have affected
this figure. Still further west, near Snead, Whittard (1932, p. 871) estimated that a mini-
mum of 250 ft. were present. The formation is distinguished from the underlying Penta-
merus Beds and overlying Wenlock Shale mainly by its colour, which is a characteristic
maroon or purple, although green and brown bands also occur. The Pentamerus Beds
are usually blue-hearted mudstones, and the Wenlock Shale grey and slightly siltier. The
contacts between the formations are usually sharp.

Fossil Communities. Whittard (1932, p. 869) reports a Clorinda Community fauna
between Plowden and Minton, and this community is also present in the Eaton Farm,
Robury Ring, Hamperley, and Springbank Farm Boreholes. Above the Clorinda Com-
munity in the two boreholes to the west of the southern Long Mynd spur, a ‘Marginal’
Clorinda Community occurs, which probably indicates (Cocks and Rickards, in press)
a continued deepening of the sea.

Correlation. Graptolites from the boreholes show a sequence from within the turri-
culatus Zone, through the crispus Zone and into the griestoniensis Zone. The top 75 ft.
(25 m.) of the Hughley Shale in the Eaton Farm Borehole are above the griestoniensis
Zone, but carry no diagnostic graptolites. Thus the presence or absence of the highest
Llandovery crenulata Zone is not yet confirmed.

ZIEGLER, COCKS, AND McKERROW: LLANDOVERY TRANSGRESSION 747

WENLOCK EDGE
Text-fig. 3, cols. 6 and 14

Llandovery rocks crop out in a strip twenty miles long, which parallels, and directly
underlies, the rocks of Wenlock Edge (text-fig. 4). Whittard (1928) described this ‘Main’
outcrop in detail; the three formation names which he used, i.e. Arenaceous Beds,
Pentamerus Beds, and Purple Shales, have been subsequently named by the Geological
Survey Kenley Grit, Pentamerus Beds, and Hughley Shales. The northern part of this
area was covered by the Shrewsbury Memoir (Pocock et al. 1938), and the southern part
by the Church Stretton Memoir (Grieg et al. 1968). A recent revision of the northern-
most part of the area was prompted by a very large temporary exposure at Devil’s Dingle
(Cocks and Walton 1968).

(a) Kenley Grit. A maximum thickness of 200 ft. (65 m.) is reached near Kenley itself
(SJ/558 004). Four-and-a-half miles to the south-west, the Kenley Grit dies out near
Gretton (SO/519 946). To the north-east, it thins more gradually, and is 35 ft. (11 m.)
thick in Harper’s Dingle (SJ/634 072), near the point where the Llandovery is covered by
the Carboniferous, 6 miles from Kenley.

The formation consists of conglomerates, sandstones, arkoses, and some finer beds,
but to the north-east there is no conglomerate except at the base; these basal pebbles
are mainly derived from Uriconian rocks, but occasional Cambrian and Ordovician
pebbles are present (Whittard 1928, p. 740). The formation rests with strong uncon-
formity upon Ordovician and older rocks, and is distinguished from the overlying
Pentamerus Beds by the abundance of coarse elastics and the absence of calcareous beds.

Fossil communities. The fossils recorded from the Grit (Whittard 1928, p. 741) all belong
to the Lingula Community. Lingula itself occurs near the top of the formation at Mor-
rellswood (near Loc. 46) and Sheinton Brook (near Loc. 50); it is also present in an old
quarry near Kenley (Pocock et al. 1938, p. 106). An assemblage with abundant rhyn-
chonellides, referable to the Lingula Community, but without Lingula itself, occurs at
Cressage Park, two miles east of Kenley (Whittard 1928, p. 741).

Correlation. No fossils found in the Grit are of zonal significance; however, at Mor-
rellswood (Loc. 46), beds of C2_3 age occur immediately above the Grit, which must
therefore be earlier.

( b ) Pentamerus Beds. The Kenley Grit passes up gradationally into the Pentamerus
Beds, but to the south-west, where the Grit is absent, the Pentamerus Beds rest uncon-
formably upon Uriconian, Tremadoc, and Ordovician rocks. The Pentamerus Beds
reach a maximum thickness of between 400 and 500 ft. (about 150 m.) in the north-
eastern half of the area, and thin southwards to nothing in the vicinity of Wistanstow
(SO/431 857). A Pentamerus-beanng conglomerate 1 in. thick has been reported (Whit-
tard 1928, p. 749) in the bed of the Onny River (Loc. 60), but it is absent higher up on
the river bank, where Hughley Shales rest directly upon the Ordovician. The lithology
of the Pentamerus Beds consists of blue mudstones, with sporadic shelly limestones, and
less fossiliferous siltstones and concretionary limestones. At the extreme northern end
of the outcrop the ’Dingle Conglomerate’ is seen near the base of the Pentamerus Beds
in Harper’s Dingle (SJ/6328 0704); the bed is confined to this one locality, and consists

748

PALAEONTOLOGY, VOLUME 11

of pebbles of Uriconian rock and Pentamerus shells set in a matrix of quartz sand and
calcite (Whittard 1928, p. 745, Pocock et al. 1938, p. 107). Apart from this distinctive
conglomerate, the Pentamerus Beds may be differentiated from the Kenley Grit by their
palaeontology, finer grain size, and more calcareous sediments, but the contact between
the two formations is gradational. The top of the Pentamerus Beds is usually sharp,
with the colour change to the Hughley Shales, but the contact is sometimes confused by
the presence of small turbidite bands (e.g. at Sheinton Brook — Loc. 51).

Fossil communities. A collection from near Gilberries (Loc. 57), in the lower part of the
Pentamerus Beds, contains abundant Hyattidina. This genus does not fit precisely into
any of the main Llandovery communities as at present defined (Ziegler et al. 1968), but
abundant Hyattidina are associated with the Eocoelia Community in New York State
and also at Presteigne (Locs. 90 and 93); we conclude that the Gilberries collection
represents an intermediate environment between the Lingula Community of the Kenley
Grit, and the deeper-water communities of the remainder of the Pentamerus Beds.

Collections from north-west of Merrishaw (Locs. 52 and 53) and near Ticklerton
(Loc. 59) are typical of the Pentamerus Community. At about the centre of the outcrop,
the deeper-water Stricklandia Community is present (Loc. 54) and in Sheinton Brook
(Loc. 50), although this last collection only contains 1% of Stricklandia. At Morrells-
wood (Loc. 46) an assemblage dominated by IRostricelluIa is present, this assemblage
also occurs in the Church Stretton area (see above), where it is interpreted as a parallel
community to that of Eocoelia.

Correlation. Eocoelia hemisphaerica indicates a C4_2 age for the Pentamerus Communi-
ties at Ticklerton (Loc. 59) and Merrishaw (Loc. 52). The Stricklandia Community at
Sheinton (Loc. 50) and the collection from Merrishaw (Loc. 53) contain S. lens ultima
of C3_4 age. Our evidence is thus consistent with a progressive increase of depth from
C4 to C4 times during the deposition of the Pentamerus Beds of the Wenlock Edge
Outcrop. Graptolites from various parts of the formation (Whittard 1928, Pocock et al.
1938) are all attributable to the turriculatus Zone, which is equivalent to C2-3 times.

(c) Hughley (or Purple) Shales. This formation has a thickness varying from 0 to 500 ft.
(0-175 m.). The lithology is dominantly purplish mudstone, but occasional coarser
turbidite bands also occur locally. The basal contact with the Pentamerus Beds is sharp
and there is overlap on to the Ordovician to the south-east. North of Much Wenlock
there is an apparently conformable upper contact with the Buildwas Beds; Cocks and
Rickards (in press) conclude that these basal Wenlock beds may include equivalents
of all the Lower Wenlock graptolite zones. However, as one proceeds south-west down
the outcrop, the Middle Wenlock overlaps both the Lower Wenlock and the Hughley
Shales and comes to rest unconformably on the Ordovician four miles south of the
Long Mynd.

Fossil Communities. Most of the collections from the Hughley Shale of the Main Outcrop
are typical of the Clorinda Community (Locs. 47-9, 51, 55, 56, 58, and 60), which
reflects a deeper-water environment than that of the Pentamerus Beds. However, near
the top of the formation to the north of the River Severn, Cocks and Walton (1968)
describe a Clorinda! Stricklandia mixture, indicating some slight local shallowing in
latest Llandovery times.

ZIEGLER, COCKS, AND McKERROW: LLANDOVERY TRANSGRESSION 749

Correlation. The oldest brachiopod fauna collected from the Hughley Shale is from the
very base of the formation in Sheinton Brook (Loc. 51), where Eocoelia intermedia
indicates a C3_4 age. E. curtisi, of C5 age occurs in collections from Boathouse Coppice
(Loc. 49) and Wall-under-Heywood (Loc. 58). Costistrieklandia lirata and Eocoelia
sulcata indicates a C6 age for collections from Boathouse Coppice (Loc. 47), Domas
(Loc. 55), and Hughley (Loc. 56), all very near the top of the formation. Thus the brachio-
pod evidence indicates that the formation spanned the period C3 to C6. Graptolites

from the Onny River (Loc. 60) indicate the turriculatus Zone (probably equivalent to
C2_3); at several localities in the north-east part of the outcrop the griestoniensis Zone
has now been proved. This last occurs in conjunction with C5 indicators such as Eocoelia
curtisi, and the probable presence of equivalents of the highest Llandovery crenulata
Zone may be inferred by the presence of the C6 brachiopods.

PRESTEIGNE
Text-fig. 3, cols. 24 and 25

To the south of Presteigne, Radnorshire, the Upper Llandovery crops out along the
B4362 road between Nash Scar and the railway bridge near Corton (text-fig. 5); we
propose the name Folly Sandstone for these beds. They are folded in an anticline with

750

PALAEONTOLOGY, VOLUME 11

its axis extending west-south-west through The Folly (SO/316 633). These Llandovery
beds are overlain by a thin (15-20 ft. — about 6 m.) development of the Nash Scar Lime-
stone to the north; to the south their contact is obscure, and the Folly Sandstone may
either dip below or be faulted against the 200 ft. (65 m.) of limestone seen in Nash
Scar Quarry. The outcrop is covered by drift east of The Folly, and is cut off by an
extension of the Church Stretton Fault to the west of Nash Wood. The only published
map of the area is the Geological Survey, Old Series, no. 56, NE. and SE., of more than
a century ago; but it shows the Llandovery correctly, except for a possible fault north
of Nash Scar quarry. The most recent description of these rocks was published in abstract
form by Kirk (1951).

The Folly Sandstone consists of dark sandstones in beds ranging from a few inches to
a few feet thick. There are many thin conglomerates with rounded pebbles up to 1| in.
(4 cm.) in diameter. The thickness of the Folly Sandstone is in excess of 100 ft., but its
base is not exposed. At Old Radnor, 4 miles to the south-west, this unit is missing alto-
gether and Precambrian conglomerates and mudstones occur unconformably beneath
the Dolyhir Limestone, the Old Radnor equivalent of the Nash Scar Limestone (Kirk,
1951, p. 56; Garwood and Goodyear, 1918, pi. vii). On the north side of the Presteigne
inlier (SO/3152 6347), the contact with the Nash Scar limestone appears to be a dis-
conformity; over an exposed distance of about 20 ft. the limestone rests on various
sandstone beds indicating a relief of about 4 ft.

Fossil communities and correlation. The stricklandiids from this area have recently been
described by Ziegler (1966a). Collections from above Nash Scar quarry (Loc. 93) and
from south of The Folly (Loc. 90) yielded Eocoelia hemisphaerica', the latter locality
also yielded Stricklandia lens aff. progressa. These collections are both representative of
the Eocoelia Community and are bothCx-C2 in age. Pentamerus Community collections
have been made in the Folly area (Locs. 91 and 92) from higher beds than Locality 90,
but the presence of E. hemisphaerica shows they are also of Q-Q age. The Pentamerus
Community is again present in Corton House Quarry (Loc. 89) and in Folly road a few
feet under the base of the Nash Scar Limestone (Loc. 88); both these collections contain
S. lens aff. progressa which shows that the top of the Folly Sandstone is still Q-C2 in age.

The evidence from the Folly Sandstone shows that the Nash Scar Limestone could be
as old as C3 but the presence of Rhipidium (in the Garwood Collection, Geological
Survey Museum) proves that part, at least, is Wenlock. An upper age limit is provided
by the occurrence of Cyrtograptus symmetricus of Wenlock age in the overlying shales
(Kirk 1951, p. 56).

ANKERDINE HILL AND OLD STORRIDGE COMMON
Text-fig. 3, cols. 15 and 16

Llandovery beds are exposed in a faulted block at Ankerdine Hill, 6| miles north of
West Malvern, and again in the core of a north-plunging anticline at Old Storridge
Common, 3 miles north of West Malvern (text-fig. 6). The base of the Llandovery suc-
cession is not exposed in either area. Groom (1910, p. 704) applied the terms, Cowleigh
Park Beds, Wych Beds, and Woolhope Shales to the Llandovery rocks of the contiguous
Malvern Hills area. However, the term Woolhope Shales is confusing and in any case
this relatively thin unit cannot be distinguished with certainty from the Wych Beds.

ZIEGLER, COCKS, AND McKERROW: LLANDOVERY TRANSGRESSION 751

text-fig. 6. Locality map of Ankerdine Hill and Old Storridge Common, Worcestershire.
This map is continuous with text-fig. 7.

752

PALAEONTOLOGY, VOLUME 11

We therefore reject the term Woolhope Shales for these Llandovery beds and regard
the Wych Beds as extending up to the first appearance of the dominantly calcareous
beds of the Woolhope Limestone. The latter is a gradational contact and could, of course,
be diachronous.

(a) Cowleigh Park Beds. Fine green sandstones and siltstones make up the lowest beds
exposed on Ankerdine Hill, while at Old Storridge Common the sediments are coarser
and thicker bedded (up to 2 ft.) sandstones. Lack of exposures and the presence of faults
make thickness estimates very uncertain; minimum exposed thicknesses are about 200 ft.
(65 m.) at Ankerdine Hill and 400 ft. (130 m.) at Old Storridge Common.

Fossil communities. In the Old Storridge Common area, fossils diagnostic of the Lingula
Community have been recorded by Lamont and Gilbert (1945, p. 642) near the axis of
the anticline, and therefore probably low in the Cowleigh Park Beds. The Eocoelia
Community occurs within a few feet of the top of the formation in Leigh Brook (Loc.
36) and in a north-flowing tributary (Loc. 35). It also occurs in lower beds south of Leigh
Brook (Locs. 34 and 87). At Ankerdine Hill, the Eocoelia Community is represented in
all three collections obtained (Locs. 42-4), but the relative stratigraphic positions of
these are not known.

Correlation. All the Eocoelia obtained from the Cowleigh Park Beds at Ankerdine Hill
and Old Storridge Common are E. hemisphaerica of Q-Ca age; the type of this species
comes from Ankerdine Hill (Ziegler 19666). The lower beds with the Lingula Community
may either be of a very early Upper Llandovery or a late Middle Landovery age.

( b ) Wych Beds. The Wych Beds consist of shales with some siltstones (less than 10%
of the formation). The base is exposed in Leigh Brook (Loc. 37) near the disused Gun-
wich Mill (SO/7430 5152) west of Old Storridge Common. The contact is sharp, but,
although there is palaeontological evidence of a non-sequence, there is no sign of erosion
of the underlying Cowleigh Park Beds. The formation appears to be about 400 ft. (130
m.) thick in the section west of Gunwich Mill.

Fossil communities. The basal few feet of Wych Beds at Gunwich Mill (Loc. 37) contain
a Pentameroides Community. The non-sequence at this locality is thus followed by
deeper-water conditions than in the Eocoelia Community of the Cowleigh Park Beds.
The evidence suggests a progressive increase in water depth throughout Upper Llan-
dovery times with a long pause in sedimentation at this locality before the deposition of
the finer grained Wych Beds. A Pentameroides Community has also been obtained from
loose blocks on Ankerdine Hill (Loc. 45) and may represent a similar low horizon in
the Wych Beds.

The Pentameroides Community is followed by the Costistricklandia Community some
10 or 20 ft. (about 5 m.) above the base of the Wych Beds; collections low in the forma-
tion were obtained from a track in Coneygore Coppice (Locs. 38, 39) and in a north-
flowing tributary of Leigh Brook (Loc. 40) to the south and south-west of Old Storridge
Common. A continued increase in the depth of water is indicated by the occurrence of
the Clorinda Community near the top of the Wych Beds to the north-east of Mousehole
Bridge (Loc. 41).

ZIEGLER, COCKS, AND McKERROW: LLANDOVERY TRANSGRESSION 753

Correlation. The basal Wych Beds at Gunwich Mill contain Eocoelia curtisi, showing
that these beds are of C5 age and that the non-sequence covers the whole of C3 and C4
time. The collection from Ankerdine Hill also contains E. curtisi.

The highest Wych Beds near Mousehole Bridge yield E. sulcata of C6 age. No fossils
of C6 age have been collected on Ankerdine Hill, and it is probable that the higher parts
of the Wych Beds have been cut out by faulting.

MALVERN HILLS
Text-fig. 3, cols. 17, 18, and 19

Llandovery rocks crop out continuously from Old Storridge Common south through
West Malvern to Herefordshire Beacon (text-fig. 7). They appear again south of Here-
fordshire Beacon in the area to the east of Eastnor (text-fig. 8). As in the Old Storridge
Common area, we recognize two formations, the Wych Beds (named from The Wyche
1J miles south of West Malvern) and the Cowleigh Park Beds (Cowleigh Park, 1 mile
north of West Malvern contains many small exposures of this lower formation). This
area has been mapped by Groom (1899, 1900) and by Phipps and Reeve (1967). We
differ from these authors in our interpretation of the Llandovery in two main respects:
the upper limit of the Wych Beds is taken immediately below the Woolhope Limestone
(see above) ; and we have no doubt that an unconformity is present below the Silurian
along the west side of the Malverns (Reading and Poole 1961, 1962, Ziegler 1964). The
Cowleigh Park Beds rest on Cambrian shales east of Eastnor and probably again east of
Cowleigh Park; to the south-east of Cowleigh Park they are inferred to rest on Malver-
nian. Between West Malvern and Ragged Stone Hill (at the south end of the Malvern
Hills — SO/759365) the Cowleigh Park Beds are overlapped by the Wych Beds which
rest directly on the Malvernian.

(a) Cowleigh Park Beds. Exposures of this formation are scattered, but our mapping
suggests that there are at least four distinct members :

1. A basal member of up to 15 ft. (5 m.) of red micaceous mudstone interbedded with
several thin (4 in. (10 cm.) or less) green sandstones composed of round quartz grains
about 1 mm. in diameter. These were exposed during a Geologist’s Association field
trip in 1963, leader Mr. N. E. Butcher, with the aid of a mechanical digger, near the
base of the Silurian to the south-west of Bronsil Castle.

2. About 100 ft. of light brown siltstones and sandstones containing fossils of the
Lingula Community and occasional pebbles of Malvernian rock. This lithology crops
out in both Cowleigh Park and in the Eastnor Obelisk — Howler’s Heath area (Loc. 32,
23, and 24).

3. A coarse pink sandstone about 100 ft. (33 m.) thick which forms a ridge extending
south-south-east from Rough Hill (SO/758 482) into Cowleigh Park. A similar topo-
graphic feature extends from Eastnor Obelisk to the eastern margin of Howler’s Heath,
and, though there are no outcrops there, loose blocks of a similar pink sandstone occur,
containing chips of Cambrian shale.

4. Up to 200 ft. (65 m.) of unsorted purple conglomerates with angular pebbles up to
3 cm. in diameter, interbedded with coarse green sandstones up to a foot thick. This
unusual lithology is known only from Cowleigh Park (SO/7616 4723) and may be
restricted to this area which is immediately adjacent to the Malvernian ridge source area.

PALAEONTOLOGY, VOLUME 11

storridge!

COWLEIGH

i- '\park\| j

NORTH
• H I L L .

\wESTj: :

MALVERN

GRE AT\
MALVERN

■WORCESTER-
V .BEACON. ■;

/ WYCHET,
CUTTING

MALVERN^

VWELLSs

STRATA OF UPPER SILURIAN
AND DEVONIAN AGE

L ' 1

V/A

J WYCH BEDS
EZ3 COWLEIGH PARK BEDS
j'lllil CAMBRIAN (UNDIFF.)

. j PRE-CAMBRIAN ROCK
^23^__F0SSIL LOCALITY

T FOLD AXIS

FAULT

... — UNCONFORMITY

text-fig. 7. Locality map of the northern Malvern Hills, West Malvern, and
Cowleigh Park areas. This map is continuous with text-figs. 6 and 8.

ZIEGLER, COCKS, AND McKERROW: LLANDOVERY TRANSGRESSION 755

To the south-west of the Malverns, Holl (1865, fig. 4) and Groom (1899, p. 167) have
described the unconformable relationship at the base of the Cowleigh Park Beds where
the Llandovery rocks rest on the Tremadocian Bronsil Shales between the Obelisk and
Howler’s Heath. Although Cambrian shales have never been mapped to the north-west
of the Malverns, it is probable that a similar unconformity exists in the Cowleigh Park
area; Groom (1900, p. 157) states that Cambrian shale was exposed in a well 1 mile
north of West Malvern, near the foot of North Hill. Our interpretation of the geology
is given in text-fig. 7.

The pebbles in the conglomerates show that a landmass composed of Malvernian
and Cambrian rocks was being eroded — this source area may have extended eastwards
from the present-day Malvern Hills. The thickness of the Cowleigh Park Beds is perhaps
250 ft. (80 m.) at the Eastnor Obelisk. At Cowleigh Park there is probably between 350
and 400 ft. (100-30 m.), but lack of exposures and the possible presence of faults make
this estimate uncertain.

Fossil communities. The Lingula Community is the only fossil community represented
in the Cowleigh Park Beds of the Malverns; the fossils are confined to the second
(siltstone and sandstone) member. They have been collected from Cowleigh Park (Loc.
32), south of the Obelisk (Loc. 23), and north of Howler’s Heath (Loc. 24). This com-
munity indicates very shallow marine conditions, and it is possible that some of the
Cowleigh Park Beds are beach deposits. The coarseness of the sediments and the absence
of the Eocoelia Community indicate environments nearer to the shore-line than those in
correlated beds to the north-north-west near Old Storridge Common and Ankerdine
Hill, and to the south-south-west at May Hill.

Correlation. No fossils of value in correlation occur in the Cowleigh Park Beds of the
Malverns. Comparison with Old Storridge Common to the north and with May Hill
to the south suggests an early Upper Llandovery or late Middle Llandovery age.

( b ) Wych Beds. The Wych Beds are seen resting on the fourth (conglomerate) member
of the Cowleigh Park Beds at the Cowleigh Park football field (SO/7616 4723); the
basal 6 ft. (2 m.) are coarse sandstones (they appear to be re-worked Cowleigh Park
Beds) with occasional broken Fentameroides, which probably indicate a deeper-water
environment than in the Cowleigh Park Beds; these sandstones are followed by shales
and green siltstones typical of the main mass of the Wych Beds. The only other two
exposures of the base of the Wych Beds are on the western margin of the Malvern Hills
at West Malvern (SO/7646 4554) commonly known as the Sycamore tree quarry ( Robert-
son 1926, p. 168), and the Gullet Quarry (SO/7612 3811) (Reading and Poole 1961);
at both these places a basal conglomerate is present, up to 2 or 3 ft. (1 m.) thick. The
conglomerate consists of rounded boulders of various Malvernian rocks, ranging from
4 mm. to 80 cm. in diameter, set in a matrix of green siltstone and mudstone — typical
of the Wych Beds — containing a rich fauna of brachiopods and corals (see below).

The Wych Beds are poorly exposed and their thickness is determined with difficulty;
estimates of about 400 ft. (130 m.) have been obtained from sections near Eastnor
Obelisk and in Cowleigh Park. This is about the same thickness as at Old Storridge
Common. It includes all the beds up to the base of the Woolhope Limestone, that is,
where calcareous nodular beds become abundant.

3 D

C 6055

756

PALAEONTOLOGY, VOLUME 11

WYNDS,
==p’OIN T>;

HEREFORDSHIRI

/.“.'BEACON}3.'

S WIN YARD
JhillI

iULLE

i ■ -a

MIDSUMMER
. • ’ ‘ H I L L - . •

EASTNOR

'HOCtYRUSH*

RA G G E D.
'STONE .
=HILL '

HOWLERS

HEATH

;HASE;

'•END

HILL*

^ NEW RED SANDSTONE

V / / A STRATA OF UPPER SIL
V//\ AND DEVONIAN AGE

WYCH BEDS

a COWLEIGH PARK BEDS

BLACK SHALES \

1 "-I HOLLYBUSH SANDSTONE/
2 PRE-CAMBRIAN ROCK
( 25) FOSSIL LOCALITY
FOLD AXIS
NORMAL FAULT
• THRUST FAULT
UNCONFORMITY

SCALE IN MILES

text-fig. 8. Locality map of the southern Malvern Hills.
This map is continuous with text-fig. 7.

ZIEGLER, COCKS, AND McKERROW: LLANDOVERY TRANSGRESSION 757

Fossil communities. The fossils in the matrix of the conglomerates at West Malvern
(Loc. 30) and the Gullet Quarry (Loc. 26) contain many species that are absent
elsewhere in the Upper Llandovery of the Welsh Borderland, apart from Bog, Shrop-
shire, and are interpreted (Ziegler, Cocks, and Bambach 1968), as a mixture of rocky
bottom and soft bottom animals. Rocky bottom communities are usually in the areas
of erosion, and do not frequently get preserved, but in this case the beds immediately
above the conglomerate at the Gullet Quarry contain a Pentameroides Community
(Loc. 27), and the fauna on the rounded boulders probably lived well below low-tide
level and uncovered by sediment for a long period.

At Cowleigh Park (SO/7616 4723) the basal 6 ft. (2 m.) of the Wych Beds are coarse
sandstones containing broken Pentameroides, and at the same exposure (Loc. 33) the
Costistricklandia Community was found in typical Wych Beds lithology a few feet above
the highest sandstone. Higher horizons in the Wych Beds also yield the Costistricklandia
Community at Gullet Quarry (Locs. 28, 29) and on the south side of Howler’s Heath
(Loc. 25). The Clorinda Community has been found at West Malvern (Loc. 31 ) indicating,
as at Old Storridge Common, still deeper water in the upper part of the Wych Beds.

Correlation. At Gullet Quarry Eocoelia curtisi is present in the basal Wych Beds (Locs.
26, 27). This species is of C5 age, as is Costistricklandia Ur at a alpha which occurs in
higher beds at Gullet Quarry (Locs. 28, 29). The Cowleigh Park (Loc. 33), and West
Malvern (Loc. 31) collections contain C. lirata typica, showing that, in the Malverns,
the greater part of the Wych Beds is of CG age.

WOOLHOPE
Text-fig. 3, col. 26

Squirrel and Tucker (1960) proposed the names Lower Haugh Wood Beds and Upper
Haugh Wood Beds for the Llandovery rocks exposed in the Woolhope dome, east-
south-east of Hereford. The former are at least 280 ft. (90 m.) thick, their base unexposed,
and are lithologically similar to the Wych Beds of the Malverns. The Upper Haugh
Wood Beds are 30 ft. (10 m.) thick; the basal 20 ft. (7 m.) are purple and green muddy
siltstones, but more calcareous beds are present in the top 10 ft. (3 m.).

Fossil communities. In the Lower Haugh Wood Beds the Costistricklandia Community
is represented by collections from Kidley Coppice (Loc. 85) and Rudge Wood (Loc. 86).
Judging from the faunal lists of Squirrel and Tucker this community is also present in
the Upper Haugh Wood Beds.

Correlation. C. lirata typica occurs in both the Lower and Upper Haugh Wood Beds,
indicating a C6 age. The Petalocrinus Limestone is present in the highest 10 ft. (3 m.) of
the Upper Haugh Wood Beds; it was found by Pocock (1930, pp. 60-1) at equivalent
horizons in the May Hill and Malvern areas, but these exposures are now obscure. It is
very close to the lithological boundary at the base of the Woolhope Limestone, and we
conclude that this limestone is of late C6 or early Wenlock age.

MAY HILL
Text-fig. 3, col. 27

Llandovery Beds are exposed in the core of an anticlinal structure at May Hill,
8 miles west of Gloucester (text-fig. 9). Two formation names, Huntley Hill Beds and

758

PALAEONTOLOGY, VOLUME 11

text-fig. 9. Locality map of the May Hill Inlier (after Lawson 1955,
with minor changes).

Yartleton Beds, were proposed by Gardiner (1920) who also proposed the term Huntley
Quarry Beds for ‘Fine and somewhat coarse grits, containing lapilli’ for beds in the
vicinity of the quarry. At the present time, volcanic flows with interbedded red-beds are
seen in Huntley Quarry, and these apparently underlie coarse sandstones with angular
fragments of the volcanic rocks which grade up into fossiliferous Llandovery sediments.

ZIEGLER, COCKS, AND McKERROW: LLANDOVERY TRANSGRESSION 759

The contact is obscure and faulted but it is probable that the Llandovery lies unconform-
ably on these volcanics, which might possibly be correlated with similar volcanics in the
Malvern Hills assigned to the Pre-Cambrian Warren House Series (Groom 1910). We
regard the coarse grits in the vicinity, which have sometimes been included in the
Huntley Quarry Beds, as the basal part of the Huntley Hill Beds, and propose the term
Huntley Quarry Volcanics for the volcanic rocks and their associated red beds.

We also differ from the current literature concerning the relationship of the Yartleton
Beds and the Woolhope Limestone. It has long been known (Phillips 1848, pp. 184-5)
that a series of sandstones occur within the Woolhope Limestone sequence at Little
London (SO/700 184) in the southern part of the inlier. Our mapping shows that the
sandstone at Old Oaks Farm Quarry (Loc. 21) and the stream section near Hill Farm
(Loc. 22) in the north-west and central parts of the inlier are also within the Woolhope
sequence; in these two areas Lawson (1955) mistakenly referred these Woolhope sand-
stones to the Llandovery. Except for these minor points, we agree with Lawson’s map
of the Llandovery beds.

(a) Huntley Hill Beds. Lawson’s ( 1 955) minimum thickness of 600 ft. (200 m.) is probably
an underestimate. The Little London stream section (SO/7085 1832 to SO/6997 1834)
where the beds, with the exception of one outcrop, dip continuously to the west, sug-
gests a thickness of about 850 ft. (280 m.). The oldest beds exposed in this section are on
strike with the basal beds at Huntley Quarry, and are therefore near the base of the
formation.

The Huntley Hill Beds are mostly grey coarse sandstones and conglomerates; they
weather to a buff colour and are often reddened in areas close to the Triassic cover.
The sandstone beds are normally one to 2 ft. thick; they show lateral changes in thick-
ness, and they have little internal structure except towards the top of the formation
where thinner bedded sandstones are present with fine parallel laminations. Shales are
rare, but a few tens of feet of shale are exposed in the Little London stream (SO/7057
1837) at about 300 ft. above the base of the formation.

The pebbles in the conglomerates include the distinctive quartz-orthoclase plutonic
rocks of the Malvern Hills. Volcanic pebbles are also present and could have been derived
from other Malvern rocks or, in the case of the angular andesite prophyry pebbles,
directly from the Huntley Quarry Volcanics. There is no doubt that adjacent pre-
Cambrian rocks were being eroded at this time.

Fossil communities. The lowest 350 ft. (115 m.) of the Huntley Hill Beds are largely
barren of fossils, but Lingula pseudoparallela has been reported from the basal sediments
of Huntley Quarry (Symonds 1872, pp. 147-8). Eocoelia hemisphaerica and other repre-
sentatives of the Eocoelia Community have been collected in the Little London stream
section at about 350 ft. (Loc. 6) and 410 ft. (Loc. 7) above the supposed base of the
formation ; closely similar assemblages also occur near Dursley Cross (Loc. 5) and on the
south-east side of May Hill (Loc. 4). At higher stratigraphical levels there is a reversion
to the Lingula Community; about 440 ft. (140 m.) above the base in the Little London
stream (Loc. 8), and in old quarries on Brights Hill (Locs. 9, 10). The highest Huntley
Hill Beds are poorly exposed; a single fossil locality near old quarries at the top of
Nottswood Hill (Loc. 11) has yielded the Eocoelia Community, with E. intermedia at
about 610 ft. above the base of the formation.

760

PALAEONTOLOGY, VOLUME 11

Correlation. The lowest, largely unfossiliferous 350 ft. (115 m.) may be pre-Upper
Llandovery in age. The beds between 350 and 410 ft. (115-35 m.) from the base of the
formation in the Little London stream section are known to be of C4 or C2 age from the
presence of E. hemisphaerica, as are the beds that have been correlated with these at
Dursley Cross and May Hill. The Nottswood Hill quarries at around 610 ft. (200 m.)
from the base contain E. intermedia of C3 or C4 age.

( b ) Yartleton Beds. The lower contact of the Yartleton Beds is not exposed, but there
appears to be a gradational transition from the coarse sandstones of the Huntley Hill
Beds to the fine laminated sandstones (typically 3-4 in. thick), siltstones and shales of
the Yartleton Beds. The formation is about 500 ft. (160 m.) thick.

Fossil communities. Fossil localities are abundant in the Yartleton Beds, but there are no
well-exposed sections showing the stratigraphic relationships between exposures more
than a few feet apart, as most of the outcrops occur along streams in the vicinity of faults.
The fossil collections however can be arranged stratigraphically by comparing the Eoco-
e/ia, stricklandiids and pentamerinids with contemporaneous sections at Llandovery
and in other parts of the Welsh Borderland. Table 1 summarizes the data for eleven
localities at May Hill. The method used for calculating percentages is described in
Ziegler, Cocks, and Bambach (1968, p. 4).

Except for the two collections (Locs. 18, 19) within 100 ft. (33 m.) of the Woolhope
Limestone, all the Yartleton fossils belong either to the Pentameroides or Costistrick-
landia Communities or to mixtures of the two. Some changes in community composition
apparently took place in a short period of time, for example, Locality 16, with 34%
Pentameroides and 44% Costistrieklandia contains elements of both communities but
occurs a few inches above Locality 15 which consists of Costistrieklandia with 8%
Clorinda. Locality 16 may have a fauna from the boundary between the Costistrieklandia
and Pentameroides Communities, or it may be due to mixing of faunas after death.
Locality 18 is interpreted as a death assemblage; it contains fairly abundant Eocoelia
together with some deeper- water species. The highest locality in the Yartleton Beds,
19, is assigned to the Clorinda Community and represents deeper water than the earlier
faunas.

Correlation. The majority of the Yartleton Beds are of C5 and C6 age (Table 1). It is
possible that the basal parts are C4; Locality 12 has no Eocoelia nor pentamerinids, and
the stricklandiids are fragmentary and cannot definitely be assigned to Stricklandia or
Costistrieklandia, so it could represent an earlier horizon than the remaining collections.

At May Hill there is a transition between the Llandovery and Wenlock Beds. Phillips
(1848, pp. 74-5, 184-5) arbitrarily drew the contact at the first appearance of limestone.
Unfortunately the Woolhope Limestone and the overlying Wenlock Shales in the
southern part of the Welsh Borderland have not yet yielded any graptolites that allow
correlation with the graptolite zones and Phillips’ arbitrary criterion must continue
for the time being.

TORTWORTH
Text-fig. 3, cols. 28 and 29

Curtis (1955) has assigned the Llandovery rocks of this area (text-fig. 10) to two for-
mations: the Damery Beds and the Tortworth Beds. These are separated by the ‘Upper

ZIEGLER, COCKS AND McKERROW: LLANDOVERY TRANSGRESSION 761

Trap’, an andesite flow with a fossiliferous ashy tuff bed on its upper surface (Reed and
Reynolds 1908, p. 515, Reynolds 1924). The basaltic ‘Lower Trap’, considered intru-
sive by Reynolds (1924) but extrusive by Curtis, occurs at the base of the Damery
Beds, resting unconformably on Tremadocian shales. Both basalts thin towards the
north-west, and appear to be absent north of Stone.

( a ) Damery Beds. This unit is comprised of about 500 ft. (160 m.) of alternating fine-
grained sandstone and shale beds (Curtis, 1955), with the former being dominant in the
lower part of the unit.

Table 1. Critical community and dating elements of the collections from the Yartleton Beds
(E, Eocoelia ; P, Pentameroides', S, Costistricklandia ; and C, Clorinda)

Community elements

Com-

Dating

Stratigraphical

Section

Loc.

E

P

5

C

munity

elements

Date

notes

Stream 450 yd.

19

7%

0%

3%

10%

C

E. sulcata

C6

Within 100 ft. of

W. of Hay Farm

C. lirata typica

Woolhope Lst.

(NE. side of

18

18

0

1

1

E. sulcata

G

Along strike from

May Hill)

C. lirata typica

Loc. 19.

84

3

0

8

1

s

E. sulcata

c6

10 ft. above Loc.

C. lirata typica

83.

83

12

18

26

0

P/S

E. sulcata
Pentameroides sp.
C. lirata typica

c6

(See above.)

17

0

0

42

0

s

C. lirata alpha

G

Stream S. of Pit-

20

4

0

4

0

s

E. sulcata ?

G

man’s Farm (W.

C. lirata typica

side of May Hill)

16

0

34

44

0

P/S

Pentameroides sp.

G

A few inches

C. lirata alpha

above Loc. 15.

15

0

0

29

8

s

C. lirata alpha

C5

(See above.)

Stream NW. of

14

3

37

7

0

p

E. curtisi ?

G

Hill Farm (SE.
side of May Hill)

Pentameroides sp.

Stream S. of Holly-

13

3

34

19

0

p

E. curtisi ?

G

bush Farm,
Dursley Cross

12

0

0

34

0

s

Pentameroides sp.

?C5

Fossil Communities. The Eocoelia Community is present in the basal 130 ft. (43 m.) of
the Damery Beds. Collections have been made within a few feet of the top of the Lower
Trap in Damery Quarry (Loc. 82), and in higher beds at Charfield Green (Loc. 1) and
near Damery Bridge (Loc. 77). The Pentameroides Community has been recognized in
two collections, 10 ft. (3 m.) stratigraphically apart in the middle of the formation at
Charfield Green (Locs. 78, 79). Still deeper water is indicated by the presence of the
Costistricklandia Community in the top two-thirds of the Damery Beds; collections
have been made on the road south of Damery Bridge at about 165 and 176 ft. (54-
7 m.) above the base of the formation (Locs. 80, 81), and in the railway cutting at
Charfield Green (Loc. 2) within 25 ft. (8 m.) of the top of the formation.

Correlation. Eocoelia curtisi occurs in all the collections from the lowest third of the
Damery Beds, and Costistricklandia lirata alpha is present in the collections from the
upper two-thirds of the formation. These show that all the Damery Beds are of C5 age.
The stream south of Charfield Green (Loc. 2) is the type locality of E. curtisi (Ziegler
19666, p. 537).

762

PALAEONTOLOGY, VOLUME 11

text-fig. 10. Locality map of the Tortworth Inlier (after Curtis 1955).

(b) Tortworth Beds. The lithology of these beds is similar to the shales and fine-grained
sandstones of the Damery Beds. Exposures are rare except near the base; Curtis (1955)
estimates the thickness at 200 ft. (65 m.), but the upper contact appears to be gradational
with an impersistent limestone forming the base of the Wenlock in the western parts
of the Tortworth inlier.

Fossil communities. The only collections from the Tortworth Beds are from near the
base. In Daniel’s Wood (Loc. 3), the Eocoelia Community is present within a few feet

ZIEGLER, COCKS, AND McKERROW: LLANDOVERY TRANSGRESSION 763

of the Upper Trap which is 150 ft. (50 m.) thick here; but north-east of Stone (Loc. 76)
the Costistricklandia Community occurs in the basal foot of the Toriworth Beds where
the basalt is only about 15 ft. (5 m.) thick. The different thicknesses of the lava appear
to have controlled the local depth of water; Costistricklandia Community depths are
present before and after the flow where it is thin, but where the flow is 150 ft. (50 m.)
thick the deeper-water Costistricklandia Community at the top of the Damery Beds is
replaced by the shallower-water Eocoelia Community above the andesite (Ziegler 1965).

Correlation. Eocoelia sulcata is present in both collections from the base of the Tort-
worth Beds; Costistricklandia lirata alpha is also present north-east of Stone (Loc. 76).
The presence of C. lirata alpha and E. sulcata indicates that the horizon must be near
the C5-C6 boundary.

WALSALL AND GREAT BARR
Text-fig. 3, col. 6

Butler (1937) has described Llandovery rocks in the Walsall Borehole, and Jukes
(1853) and Eastwood et al. (1925, p. 11), mention a very small area of the Llandovery
rocks at Great Barr near the Eastern Boundary Fault of the South Staffordshire
Coalfield. The Walsall Borehole penetrated Wenlock Limestone, Wenlock Shale, Ban-
Limestone, and below this, some 4577 ft. (150 m.) of shales, mudstones, siltstones, and
sandstones before hitting Cambrian quartzite with a dip differing by 8° to 15° from the
Silurian. Some of the beds beneath the Barr Limestone are Wenlock in age as Cyrto-
graptus murchisoni has been recorded 36 ft. (12 m.) below this unit (depth 8317 ft.), but
747 ft. (24 m.) below the Barr Limestone (depth 870 ft.) are purple and green shales,
similar to the Hughley Shales of Shropshire, and we prefer to draw the Wenlock-Llan-
dovery boundary at this point. Butler records 24 bentonitic clays above depth 870 ft.,
and 27 bentonites below this depth; one bed is 7 in. (18 cm.) thick, the remainder range
from 2 in. down to ^ in. (5 cm.-L5 mm.).

Fossil communities. The beds from 28 to 303 ft. (9-5-100 m.) above the base of the Silurian
(depths 1225 to 950 ft.) contain Costistricklandia ; in this sequence, Pentameroides
(previously recorded as Pentamerus oblongus ) is known only over a short thickness
(233-48 ft. (76-81 m.) above base, or depths 1020 to 1005 ft.), so it appears probable that
the Llandovery beds accumulated mainly in the moderate depths characteristic of the
Costistricklandia Community. This interpretation agrees well with Butler’s (1937,
p. 251) picture of the depositional environment. The Costistricklandia Community also
occurs at Great Barr where a Llandovery quartzite is exposed south-south-east of the
house formerly known as the Australian Arms (SP/038957).

Correlation. Costistricklandia lirata alpha occurs between 28 and 93 ft. (9-5-31 m.) above
the base of the sequence (depths 1225 to 1160 ft.) (Butler, 1937, p. 247), showing that the
oldest Llandovery in this area is of C5 age. C. lirata typica has been recorded between
123 and 283 ft. (41-93 m.) above the base (depths 1130 to 970 ft.), which proves a C6
age for these beds. Graptolites occur at several horizons; those that Miss Elies examined
from below 120 ft. (39 m.) from the base (depth 1133 ft.) are ‘suggestive of the Upper
Valentian’, and those below 27 ft. (9 m.) from the base (depth 1226 ft.) suggest a horizon
at least as low as the M. griestoniensis Zone (Butler 1937, p. 246). At the Great Barr
exposure, we have collected C. lirata typica.

764

PALAEONTOLOGY, VOLUME 11

RUBERY
Text-fig. 3, col. 8

Llandovery rocks rest with slight angular unconformity on the Lickey Quartzite
(Cambrian) to the north and south-east of Rubery (Wills et al. 1925, 1938, Eastwood
et al. 1925, p. 12); the latter authors infer that it occurs in hollows in the quartzite.
Wills et al. (1925, p. 68 and 1938, p. 177) describes the Rubery Shales Series as consisting
of a minimum of 40-50 ft. (13-16 m.) of shales and thin sandstones with some calcareous
beds; these rest on 100 ft. (33 m.) of Rubery Sandstone Series which are coarse and fine
sandstones, some shale and a basal conglomerate.

Fossil communities. Costistricklandia is present in both the Rubery Sandstone series and
the Rubery Shale Series; the faunal list of Wills et al. (1925, pp. 72-3) suggests that the
Costistricklandia Community is present throughout the whole Llandovery sequence in
this area, although Pentameroides does occur in some beds.

Correlation. Costistricklandia lirata alpha is abundant in the Rubery Sandstone Series
(St. Joseph 1935, p. 419) and Pentameroides also occurs (previously recorded as
Pentamerus oblongus). The higher Rubery Shale Series contains C. lirata alpha , Mono-
graptus marri and M. nudus. It is concluded that the section described by Wills et al.
(1925, p. 68) is all of C5 age.

The fossils listed (Eastwood et al. 1925, pp. 14-15) as coming from Wenlock beds
near Rubery Hill Asylum (SO/992778) include undoubted Llandovery fossils, but we
have not examined this locality.

COVENTRY
Text-fig. 3, col. 9

Shotton (1927) records Upper Llandovery sandstone pebbles from the Corley Con-
glomerate of Upper Carboniferous age, which is exposed from Coventry to Corley and
Hollyberry End (SP/262842) 6 miles to the north-west. Shotton has shown that,
approaching Coventry from the north-west, the proportion of Llandovery sandstone
pebbles increases from 46% at Hollyberry End, to 70% at Corley (SP/304850) and 85%
at King Street, Coventry (SP/333797). This traverse along the present outcrop of
the Corley Conglomerate runs obliquely to a south-south-eastward extension of the
Nuneaton Ridge (composed of Cambrian and Precambrian rocks), and Shotton suggests
(1927, p. 616) that this ridge was capped by a strip of Upper Llandovery rocks (as at
Walsall and Rubery) which provided the pebbles. It is probable that the source area
extended north-north-west from around Binley, 3 miles east of Coventry.

Fossil communities. The faunal list (Shotton 1927, pp. 609-10) contains 55 species,
which, with one or two possible exceptions, are typical of, or known in, the Eocoelia
Community. Stricklandiids, which are dominant to the west at Rubery and Walsall,
are absent, as are Pentameroides and other elements of the offshore communities.

Correlation. Examination of Eocoelia in pebbles (some kindly lent from the Birmingham
University collections by Professor Shotton, others in the Sedgwick Museum, Cam-
bridge) show the present of E. intermedia , E. curtisi, and E. sulcata , which indicates

ZIEGLER, COCKS, AND McKERROW: LLANDOVERY TRANSGRESSION 765

ages from C3_4 through C5 to C6. This is the only record of Silurian beds as old as C3_4
to the east of the Malvern line.

LOWER LEMINGTON
Text-fig. 3, col. 20

Silurian rocks are recorded below Coal Measures in Strahan’s (1913) description
of a borehole 350 yd. W. 22° N. of Lower Lemington Church, 1 mile north-east of
Moreton-in-Marsh. Grey shales with sandstone and limestone occurred from a depth
of 1546 ft. (510 m.) to the completion of the hole at 1700 ft. (558 m.); allowing for a
dip of 22°, the strata proved are about 140 ft. (46 m.) thick.

Fossil communities. Examination of the fossils in the Geological Survey collections
(JP3143-61) from between 1590 and 1682 ft. show an Eocoelia Community present
throughout this thickness.

Correlation. The Eocoelia present are all E. sulcata , indicating a C6 age for all the
fossiliferous Silurian beds that were penetrated.

EASTERN MENDIPS

Curtis (1955) summarizes the earlier work of Reynolds (1907, 1912, 1929) on the
stratigraphy of the Silurian inlier which extends 3^ miles westwards from Downhead,
Somerset :

3. Wenlock mudstones with some micaceous sandstones (120 ft. — 40 m.).

2. Andesites and tuffs (at least 400 ft. — 130 m.).

1. Tuffs (110 ft. (36 m.) seen). These oldest beds have, in the past been assigned to
the Upper Llandovery.

Fossil communities. A collection from a field 1 mile west of Downhead (Loc. 94) yielded
a dominance of Salopina and Sphaerirhynchia, which suggest a near-shore environment
for beds which our interpretation of the literature suggests are beneath the andesites.

Correlation. It is probable that there are no Llandovery beds exposed in the inlier.
Collections from tuffs (ST6752 4584) suggest an environment similar to the Eocoelia
Community of the Llandovery, but no Eocoelia or other distinctive Llandovery forms
have been found by us or recorded in the literature. We have examined the specimens
(GSM DEW 3833-935) upon which Green (1962) based his report of beds attributable
to the Llandovery, and conclude that they are also of Wenlock or Ludlow age. We thus
tentatively consider these beds to be of Wenlock age. This conclusion is supported by
the presence of Acaste downingiae (Reynolds 1907, p. 227, Curtis 1955, p. 7) which we
believe to be absent in the Llandovery.

RUMNEY

The Silurian inlier at Rumney (Rhymney), 2\ miles north-east of the centre of Cardiff',
has been described by Sollas (1879). The Silurian beds in this inlier have always been
assigned to the Wenlock and Ludlow, although Sollas recorded Pentamerus and
Stricklandia.

The problem has been investigated by Dr. M. G. Bassett, who kindly writes: ‘I am
certain that all the beds are Wenlock-Ludlow. The best exposure at present is at

766

PALAEONTOLOGY, VOLUME 11

Penylan Quarry, where virtually the oldest beds brought to the surface in the inlier are
exposed. These beds are certainly below the Rumney Grit, and hence older than the
horizons from which Sollas quoted his diagnostic Llandovery brachiopods. The beds
at Penylan contain abundant Meristina obtusa, which I consider to be a good indicator
of late Wenlock age. This is supported by the occurrence of other fossils which could
only be Wenlock or Ludlow in age.’ In addition there are many Wenlock brachiopods
from the inlier at the National Museum of Wales which bear old labels mis-identifying
them as ‘P. oblong us' and ‘ Stricklandia' .

BRE1DDEN HILLS
Text-fig. 3, col. 4

Watts (1885, pp. 536-7) gave the first description of the Llandovery beds of the Breid-
dens, and Wade (1911, p. 419) subdivided these beds into a lower unit, the Cefn Beds,
and an upper unit, the Buttington Shales. The use of local names is preferable to the
Shropshire names employed by Whittard (1932, pp. 880-2) in his description of this
area. These beds are exposed sporadically along a 5-mile strip of country north-east
of Buttington Station; the largest exposures are in the Brick Works (SJ/265 100) a
quarter of a mile east of the station, and adjacent road sections.

(a) Cefn Beds. Whittard (1932) states that these consist of about 300 ft. (100 m.) of
sandy mudstones with calcareous sandstones and occasional conglomeratic beds, but
in the south-west the calcareous beds are more common. Ripples exposed at the top of
the formation, over a large face in Buttington Brick Works, give a direction of movement
from east to west.

Fossil communities. Whittard (1932, p. 881) mentions two fossil localities: north of
Middletown in the centre of the outcrop (SJ/3008 1276) within 50 ft. (16 m.) of the base,
and at the south-west end, south-east of Buttington Station (SJ/2633 0997) within
100 ft. (33 m.) of the base of the formation. Our collections show that the Stricklandia
Community is present in both places.

Correlation. The presence of Stricklandia lens cf. intermedia suggests a Middle Llan-
dovery age for the lower part of this formation.

( b ) Buttington Shales. Whittard (1932, p. 881) described maroon, green, and blue
mudstones with some calcareous beds up to 6 in. (15 cm.) thick, becoming uniformly
maroon towards the top; he estimated their thickness to be about 350 ft. (115 m.). The
upper contact is defined by an abrupt termination of the maroon coloured beds; above
this Cocks and Rickards (in press) record the presence of a complete succession of
Lower Wenlock graptolite zones at Buttington Brick Works.

Fossil communities. The Buttington Shales have few fossils, but we have a collection,
from near the top of the sequence in the Brick Works, in the Clorinda Community.
Whittard (1932, p. 882) also found fossils in loose blocks at the north-east end of the
outcrop; his faunal list is also typical of the Clorinda Community.

Correlation. It is not possible to state what horizon within the Llandovery is represented
by these two collections from the Buttington Shales, but they are overlain with apparent
conformity by a basal Wenlock centrifugus Zone fauna.

ZIEGLER, COCKS, AND McKERROW: LLANDOVERY TRANSGRESSION 767

WELSHPOOL
Text-fig. 3, col. 3

Wade (1911) described two formations of Llandovery age in the Welshpool area, the
Powis Castle Beds and the higher Cloddiau Beds. The poorly exposed outcrop extends
from Belan (2 miles south-west of Welshpool) to the Llansantffraid ym Mechain area,
north-east of Meifod (see below).

(a) Powis Castle Beds. This unit consists of 100 ft. (33 m.) of massive bedded con-
glomerates; sandstones, shales and limestones are also present (Wade 1911, p. 431).
It rests on Ashgill beds {Phillip sinella parabola age) in the Gwern y Brain (SJ/219 128)
(Cave 1965, p. 282) and at the Laundry (SJ/197 103) (ibid., p. 295). Towards Welshpool
the uppermost Ordovician beds are overstepped and the Powis Castle Beds rest on
Lower Caradoc rocks (Harnagian Stage). In the Gwern y Brain, the Powis Castle Beds
contain angular pebbles of these Ashgill and Caradoc units.

Fossil communities. A collection from the Gwern y Brain (SJ/2186 1283) within 10 or 15
ft. (3-5 m.) of the base of the Powis Castle Beds has yielded nothing but derived fossils
from the underlying units, but the angularity of the pebbles and boulders is very sug-
gestive of deposition in non-marine conditions. We have found the Cryptothyrella Com-
munity, a probable early Llandovery equivalent of the Eocoelia Community, in a loose
block in the southern part of the area, 600 yd. west of Belan Locks (SJ/2100 0518).
On Cherry Tree Bank (SJ/2226 0831) on the north edge of Welshpool, we have collected
Clorinda from the higher beds of the unit, and this would suggest that some of the Powis
Castle Beds accumulated in deeper water; this conclusion is strengthened by the presence
of Stricklandia lens in the conglomerates in the north of Powis Castle Park. In fact, the
full depth range of communities may be present in the formation.

Correlation. The fact that the overlying Cloddiau Beds contain a Lower Llandovery
graptolite fauna (vesiculosus Zone) suggests that the Powis Castle Beds are of Lower
Llandovery age, and suggests a short pause in sedimentation between the Ashgill beds
in the Gwern y Brain and the earliest Silurian. The presence of Stricklandia lens indicates
a Llandovery as opposed to a late Ordovician age for the unit.

( b ) Cloddiau Beds. Around Welshpool, upper Silurian beds rest directly on the Powis
Castle Beds, but from Cloddiau northwards, some 250 ft. (80 m.) of mudstones and
siltstones are seen above the conglomerates.

Fossil communities. A collection from the Cloddiau Beds at Y Frochas (SJ/1955 0824)
yielded the Clorinda Community.

Correlation. Wade (1911, p. 434) reported Climacograptus innotatus, C. cf. rectangularis
and C. medius from the lower part of the formation near Cloddiau. These indicate the
vesiculosus Zone, but as they were found only 20 ft. (7 m.) above the base of the forma-
tion, it is possible that the Middle and Upper Llandovery is also present — the Cloddiau
Beds extent upwards into Wenlock strata with no change in dip.

768

PALAEONTOLOGY, VOLUME 11

MEIFOD
Text-fig. 3, col. 2

The Llandovery outcrop extends northwards from Welshpool for 7 miles to the
Llansantffraid ym Mechain area (Whittington, 1938), and thence changes direction
south-westwards to Meifod (King, 1928). King recognized four formations which he
labelled Vx, V2, V3, and VS and Whittington employed this terminology at Llansant-
ffraid ym Mechain.

(a) Craig-wen Sandstone (F2). This basal unit is a 30-ft. (10 m.) sandstone with some
large calcareous concretions near the base. The sandstone fills hollows cut in Ashgill
mudstone south-west and west of Meifod at Graig-wen and Glan-yr-afon-isaf (King
1928, pp. 682-3). These hollows are a foot or more in depth and contain angular
fragments of Ashgill mudstone.

Fossil communities. King (1928, pp. 686-7, 699) records two communities from the
Graig-wen Sandstone. In the calcareous beds fragmentary corals and brachiopods are
present, while the sandstone contains 70% orthids, 20% Cryptothyrella (the Meristina
crassa of earlier reports), and about 10% rhynchonelloids. Our collection from the basal
1| ft. (50 cm.) of the unit at Graig-wen Quarry (SJ/0991 0919) has yielded the Crypto-
thyrella Community, the probable early Llandovery equivalent of the Eocoelia Com-
munity.

Correlation. The fossils of this unit are not diagnostic as to age, but the underlying Ash-
gill unit has yielded graptolites normally associated with the Dicellograptus anceps zone
(King, 1928, p. 698) and the overlying formation has yielded late Lower Llandovery
graptolites, so the Graig-wen Sandstone is probably of early Lower Llandovery age.

( b ) Blue silty mudstones (TV)- Blue-black micaceous shales and hard siltstones (V2a)
come at the base of this formation west of Meifod; higher beds (V2u and V2c) in this
630-ft. (205 m.) formation are blue mudstones with silty siliceous bands and calcareous
nodules (King 1928, pp. 687-8). Five miles north-east of Meifod, Whittington (1938,
pp. 438—42) describes blue mudstones and limestones near Tre-wylan House and Gelli
Farm (which he correlates with King’s V2b beds) resting on Ashgill. Similar beds occur
at Godor and it would seem likely that the outcrop is continuous with that of Meifod.
Beds like King’s V2c are recorded at Godor and Sarnau (Whittington 1938, pp. 440-1).

Fossil communities. King (1928, p. 699) lists Stricklandia and Clorinda [ Barrandella ] as
common in V2b; and Clorinda as common in V2o. This suggests deeper water than in the
Graig-wen Sandstone.

Correlation. In the lower part of this unit, Climacograptus normalis, ‘and what Dr. G. L.
Elies thinks may be a specimen of Monograptus acinaees ’ have been found (King 1928,
p. 688). These species would suggest the zone of M. cyphus (late Lower Llandovery), but
the unit may extend into the Upper Llandovery as the lowest graptolite fauna in the
overlying formation is in the turriculatus Zone.

(c) Pale grey and purple-red shales ( V3). King’s (1928, p. 690) V3 beds start with an abrupt
change of colour from grey-blue mudstones to pale dove-grey fine mudstones; they
consist of 330 ft. (110 m.) of pale grey and red shales.

ZIEGLER, COCKS, AND McKERROW: LLANDOVERY TRANSGRESSION 769

Fossil communities. Apart from some fragmentary trilobites and brachiopods near the
base, graptolites are the only fossils present.

Correlation. The zones present range from turriculatus to crenulata (King 1928, p. 690)
and represent the upper two-thirds of the Upper Llandovery. The red shales occur in the
crispus and griestoniensis Zones; they are the oldest dated red beds known in the Welsh
Llandovery succession.

(d) Passage Beds {VS). King (1928, p. 691) records 60 ft. (20 m.) of blue-grey mudstones
above V3 beds.

Fossil communities. Benthic forms are absent from this unit.

Correlation. King found Monoclimacis crenulata in these beds to the north-east of
Meifod, so they are assigned to the topmost Upper Llandovery. They are followed by
lowest Wenlock beds with a similar lithology.

TRANNON AND LLANIDLOES
Text-fig. 3, cols. 1 and 10

These areas are in the middle of the graptolitic-greywacke facies of central Wales
and have been included here because they contrast with the dominantly shelly facies of
the Welsh Borderland areas already described. Wood (1906) and Jones (1945) have
described the stratigraphy of the areas. At Trannon (called Tarannon by Wood) at
least 3650 ft. (1200 m.) of Llandovery sediments (we use Llandovery to include all the
beds up to the base of Wood’s murchisoni Zone) are present, and 9670 ft. (3175 m.) at
Llanidloes. In both areas the whole succession consists of graptolitic shales with grey-
wackes at some horizons. The base of the Silurian is not quite reached near Trannon,
but at Llanidloes a sharp but comfortable base is present.

Greywackes are especially abundant in the griestoniensis Zone. Bassett (1963, pp. 57-
9) interprets the Talerddig Grits of Trannon and the contemporary Moelfre Group
of Llanidloes as turbidites and comments on the problems in estimating the depth of
water in which they were deposited. We would consider that these turbidites could have
been laid down over a very large depth range, provided that the depth was greater than
that of the Clorinda Community, which is completely absent around Trannon.

The highest Llandovery consists of purple and green mudstones which contain a fauna
characteristic of the crenulata Zone. There is a gradational passage upwards into the
Wenlock beds.

BUILTH WELLS
Text-fig. 3, col. 23

Upper Llandovery rocks occur in a very narrow discontinuous belt running north-east
from Park Wells (SO/027 519 — 1 mile west-north-west of Builth Wells) for a distance of
4 miles (Jones 1947). The beds consist of about 50 ft. (16 m.) of coarse sandstone and
mudstones, resting unconformably on Ordovician beds, and overlain unconformably by
beds of Lower Wenlock age. The Wenlock overstep is considered to be responsible for
the discontinuous outcrops (Jones 1947, p. 35).

Fossil communities. A collection from Trecoed (SO/0527 5516) has yielded an abundance

770

PALAEONTOLOGY, VOLUME 11

of Pentamerus and Stricklandia , suggesting a mixture of these two communities. These
came from a coarse sandstone about 14 ft. (5 m.) above the base of the Llandovery.

Correlation. S. lens ultima is present, indicating a C3_4 age for this horizon.

GARTH

Text-fig. 3, cols. 21 and 22

The Garth area is situated some 1 5 miles north-east of Llandovery and is part of the
same belt, which, during Llandovery times, formed the junction of the stable shelf to
the south and east and the subsiding basin to the north and west. The two columns
representing this area were selected to illustrate some of the many lateral changes in
thickness and lithology in the Llandovery rocks of the Garth area. Andrew (1925)
divided the Lower Llandovery into three formations, Aa, Ab, and Ac; designated the
Middle Llandovery B; and divided the Upper Llandovery into Ca, Cb, Cc, and Cd.

(а) Lower Llandovery. The basal Llandovery sandstones and conglomerates rest on Bala
mudstones with a sharp boundary; these are followed by mudstones with occasional
sandy beds, all totalling 1880 ft. (600 m.) in the centre of the area, but thinning to 1200
ft. (400 m.) in the south-west (Andrew 1925, pp. 392, 402).

Fossil communities. No fossils are recorded in Aa. In the southern part of the area, the
Stricklandia Community is present in Ab, and both Stricklandia and Clorinda occur in
Ac, but the shelly faunas become sparse or absent in the north (Andrew 1925, pp. 396-7),
that is, towards the basin.

Correlation. Andrew (1925, p. 403) considered ‘on faunal and lithological grounds’:
that his units ‘Aa, Ab, and Ac appear accurately to represent A4, A,, and A3 of the
southern part of the Llandovery area’. It is of importance to note that the top 20 ft. of
the ‘Bala’ beds of Garth contain ‘ Glyptograptus cf. persculptus and Mesograptus cf.
modestus var. parvulus, which . . . are usually associated with the basal zone of the
Valentian’ (Andrew 1925, p. 392). This fauna indicates the persculptus Zone, the base
of which is now conventionally taken (e.g. Toghill 1968) as the base of the Llandovery in
the graptolite sequence. If Andrew was correct in equating his unit Aa with the Al
sandstone at Llandovery itself, then the base of the formation is probably diachronous.
The latter conclusion is supported by recent unpublished sedimentological work by
Dr. M. A. Woollands.

(б) Middle Llandovery. Only a small thickness of mudstones is present in a restricted
area in the south-east of the Garth area (Andrew 1925, p. 397). The few specimens
recorded are suggestive of the Clorinda Community. Andrew suggested a correlation
with B3 of Llandovery on the basis of lithological resemblance (Andrew 1925, p. 403),
but its age is still not definitively known.

(c) Upper Llandovery. The lower three Upper Llandovery units of Andrew (Ca, Cb, Cc)
comprise a sequence of mudstones, argillaceous sandstones, and rather prominent
conglomerates which is very variable in thickness (0-300 ft., 0-100 m.). The upper unit
(Cd) consists of 150-1000 ft. (50-330 m.) of olive-green and purple mudstones, similar
in colour and lithology to the high Llandovery of many other areas; it overlaps earlier

ZIEGLER, COCKS, AND McKERROW: LLANDOVERY TRANSGRESSION 771

beds, coming to rest on Ab at the extreme south-western and north-eastern ends of the
area, according to the mapping of Andrew. Ca passes up without a break into beds
containing the basal Wenlock zone of Cyrtograptus murchisoni (Andrew 1925, p. 400).

Fossil communities. Andrew's faunal lists (1925, pp. 398-9) show that Ca contains a
Pentamerus Community, which indicates a shallowing after the Middle Llandovery.
Cb has yielded a fauna which includes Pentamerus, Stricklandia, and Clorinda, suggesting
a return to deeper water. It is not possible to determine the community present in Cc,
but increase in depth is again shown by the graptolites present in Cd.

Correlation. The Upper Llandovery of Garth can be tentatively correlated with the
beds of Llandovery on gross lithological and faunal grounds. Andrew and Jones (1925,
p. 412) suggest it may be equivalent to C4-6. The presence of beds with Monograptus
priodon and Monoclimacis crenulata conformably below the basal Wenlock shows that
the highest Llandovery is present in this area.

CHRONOLOGICAL SUMMARY

Text-fig. 1 1 illustrates the chronological sequence of Llandovery deposits in the Welsh
Borderland. It brings out the fact that, although the spread of the sea was continuous
from Lower Llandovery times, there were several gaps in sedimentation. In general,
however, there is evidence of a progressive deepening at any one locality.

Lower Llandovery (T4-4) times

The shelf was narrow during the Lower Llandovery, and deposits are confined to an
area west of Welshpool and Garth (text-fig. 12). In the basin area of Central Wales
(text-fig. 11, cols. 1, 10, 21, and 22), basal Llandovery graptolitic shales follow above
similar Ashgill sediments without a break. At Trannon and Llanidloes this facies con-
tinues throughout the Lower Llandovery, but at Garth Stricklandia and Clorinda Com-
munities are present, showing that shallowing took place there in A2_3 times. This was
probably followed by a depositional break, as no A4 has been proved at Garth.

On the edge of the basin at Meifod and Welshpool (text-fig. 11, cols. 2 and 3), there
was uplift of Ashgill sediments prior to the deposition of beds containing A4_2 shallow-
water Cryptothyrella Communities. In Shropshire, to the east (text-fig. 11, cols. 5, 6,
1 1-14), the Caradoc and earlier rocks are seen to be considerably folded before the
Middle Llandovery. If these movements are correlated with the basal A4 non-sequence
at Meifod and Welshpool, then it may be that the acme of this folding occurred at the
very beginning of the Silurian. After the initial uplift, the shallow-water A4_2 beds at
Meifod and Welshpool are followed by deeper-water A3 Stricklandia and Clorinda Com-
munities. The general subsidence characteristic of the Welsh Borderland throughout
the Lower Silurian thus started as early as A2_3 times.

Middle Llandovery ( fi,-3) times

In the basin (text-fig. 1 1 , cols. 1 and 10), the graptolitic facies continues without break;
this facies spread over the basin margin at Meifod (text-fig. 11, col. 2) in B2time, when
the Clorinda Community was replaced by graptolites. At about the same time, the sea

C 6055 3 E

772

PALAEONTOLOGY, VOLUME 11

Early

Wenlock

C6

v. C5

£.c4

§C3
N °2

S' C,

^ <b d

C ^ B2

■§ ^ 6,

A«
* A,
I A.2

. r* a,

Late 1
Ashgill

2345 6 78 9

~T5 T ~T~ V 1 1 ■ — * — *—> — > > ' V V — 7 — 7

\ ^

\ \ \ \ \

G

\ \ \

G

G

^ V "V

\\

(Buildwas

Bed ////////,

/ / /;? ; ? ;?;

X \ \ \ \ \

G

\ \ \

G

/ // /

-/ / / / / /

*/ / / / / ' ' ' y / sypp ;
///////' ' --

7

• ssss\\sxs f

\ \ \ \ \ N

G

\ \ \

G

/ / / /

//////

C

ilSl f

\ \ \ \ \ '

G

\ \ \

G

\ (

/ / / /

/ / / / /

\ \ \ \ \N

\ \ \
\ \ \

n

\ \ j

' ,

/ / / /
/ / / /

P

’/////A'
/ / / /XA\

o

S \ \ \ \ '

u

\ \ \

o

\ ^ (>

p

/ / / /

L/

n

■////.'.'////'A' Q

= Graptolitic Facies

= Clorinda Community

= Strick/andia or

Costistricklandia 11

= Pentamerus or __

Pentameroides "

= Eocoe/ia or

Cryptothyrella "

= Lingula "

\ \ \ \ \

G

\ \ \

\ \| ^

(/

L-

/

/ / / /

/

AUxy//.

r

\ \ \
\ \ \

\ \y
\ /

J/l

■

NUN

L

pi

.,A _ S

\ \ \ \ \

G

\ \ \

// /

i

s

S

\ \ \ \ \ '

\ /

C

/ / ■

C,

:

p

\ \ \ \ \

[

/ / >

/ 'J

\ \ \ \ \ '
\ \ \ \ \ '

Xp

c,

S

7 F

AT

N VO
i / A

£

E

\ \ \ \ \ '

1

V SE

1 —
o

ON 1 L

\ \ \ \ \ '

?

text-fig. 1 1. Three sections across the Welsh Borderland (see text-fig. 1 and 3) showing communities

plotted against time.

ZIEGLER, COCKS, AND McKERROW: LLANDOVERY TRANSGRESSION 773

deepened at Welshpool (text-fig. 11, col. 3) and spread over West Shropshire (text-fig.
11, cols. 4, 5, 11, and 13) and may have just reached the Malverns (text-fig. 11, cols.
15-17) by the end of Middle Llandovery times. It is worth noting here that there are
no Middle Llandovery sediments in nearby sections at Eaton Farm and Wistanstow

text-fig. 12. Palaeogeography of the area during Lower Llandovery (A2) time. E, Cryptothyrella
Community; S, Stricklandia Community; C, Clorinda Community; G, graptolitic facies.

Note. The coastline is shown as a smooth curve, but we envisage a coastal topography perhaps
similar to Essex at the present time (this applies also to text-fig. 13 and 14). The size of some of the
smaller outcrops is enlarged for greater clarity, and the Upper Severn Estuary is omitted.

(text-fig. 11, cols. 12 and 14). The earliest deposits at these localities contain C2_3
Clorinda Communities; this shows that the sea was deep before sedimentation com-
menced, and that the advance of the sea was not everywhere accompanied by a supply
of sediments. Thus the sea probably extended over the Eaton Farm and Wistanstow
areas in Middle Llandovery time.

774

PALAEONTOLOGY, VOLUME 11
Early Upper Llandovery (Cx-3) times

At Trannon, Meifod, and Llanidloes (text-fig. 11, cols. 1, 2, and 10) the graptolite
facies continues, both in the basin and on the edge of the shelf (text-fig. 13). In West
Shropshire (text-fig. 11, cols. 4 and 5) the sea is seen to have deepened with the replace-

text-fig. 13. Palaeogeography of the area during early Upper Llandovery (Q) time. L, Lingula
Community; E, Eocoelia Community; P, Pentamerus Community; S, Stricklandia Community;
C, Clorinda Community; G, graptolitic facies.

ment of the Stricklandia Community by the Clorinda Community. South of the Long-
mynd (text-fig. 11, col. 13) the communities show a minor reversal, but there was an
over-all deepening from C1 to C3 times. At neighbouring sites (text-fig. 11, cols. 12
and 14) sedimentation commenced for the first time with the deposition of deep-water
Clorinda Communities resting directly on Pre-Cambrian and Ordovician rocks.

In the Wenlock Edge outcrop, and in the Malverns and at Presteigne (text-fig. 11,
cols. 6, 15-17, and 25), there are undoubted early Upper Llandovery sediments; these
contain shallow-water communities showing the progressive deepening so characteristic

ZIEGLER, COCKS, AND McKERROW: LLANDOVERY TRANSGRESSION 775

of this transgression. There is a sequence from a basal Lingula Community (not exposed
at Presteigne) deepening to Eocoelia or Pentamerus Communities. At May Hill (text-fig.
1 1, col. 27) there is a minor reversal in the progressive deepening as Lingula and Eocoelia
Communities alternate.

text-fig. 14. Palaeogeography of the area during late Upper Llandovery (C5) time. E, Eocoelia
Community; P , Pentaineroides Community; S, Costistricklandia Community; C, Clorinda Community;

G, graptolitic facies.

At Garth and Builth (text-fig. 11, cols. 21-3) no beds of C4-3 age are present, this is
probably to be accounted for by non-deposition in a sea of Pentamerus Community
depth, as this community is present above the non-sequence in beds of C4 age.

Late Upper Llandovery (C4-6) times

During this time, basin facies (the Pale Shales) continue at Garth (text-fig. 11, col. 21),
but, in Wenlock times, graptolitic deposits are present at Builth as well (col. 23); an
eastwards spread of this deep environment thus presumably occurred in C6 times (text-
fig. 14). A similar C6 spread probably took place over the edge of the basin from

776

PALAEONTOLOGY, VOLUME 11

Montgomeryshire into West Shropshire (text-fig. 11, cols. 2-4); this deepening of
the sea may also be related to the turbidites at Minsterley (which are of late Upper
Llandovery [post C4] or early Wenlock [pr e-riccartonensis Zone] age).

On Wenlock Edge (text-fig. 11, col. 6), the progressive deepening, mentioned above,
continues through most of C4_6, but in late C6 time there is an indication of slight
shallowing in the beds immediately below the Buildwas Beds.

From sandstone pebbles in the Upper Carboniferous north-east of Coventry (text-fig.
11, col. 9), Eocoelia has been found, ranging in age from C4 to C6. However, at Walsall
and Rubery (text-fig. 11, cols. 7 and 8) the earliest Silurian deposits contain C5 beds with
Costistricklandia Communities and rest directly on Cambrian rocks ; here, as in Shrop-
shire during C4-3 (see above), we have strong indications that the sea was present at
Walsall and Rubery well before sedimentation commenced. The evidence from Lower
Lemington (text-fig. 11, col. 20) suggests a shallow sea there in C6 times, but the local
base of the Silurian is unseen.

In the Malvern areas (text-fig. 11, cols. 15-19), no beds of C3_4 age are known; this
non-sequence is followed by C5 beds with Pentameroides and Costistricklandia Com-
munities which overlap the C4-2 deposits (which contain Lingula and Eocoelia Com-
munities), and rest in places on the Pre-Cambrian. Text-fig. 11 shows that there is a
progressive deepening of the sea across this non-sequence; thus there is no suggestion
of any marine regression during C3_4 time in the Malvern area.

A different type of stratigraphical break is present beneath the Nash Scar Limestone
at Presteigne and Old Radnor (text-fig. 11, cols. 24 and 25); a shallow-water Wenlock
algal limestone rests unconformably on C2 beds (with a Pentamerus Community) at
Presteigne, and on Pre-Cambrian at Old Radnor. There must have been shallowing
during this interval, and local uplift may well have occurred here during the late Upper
Llandovery.

The progressive deepening, seen or inferred for C4_6 beds of most areas, was also
present to the south of the Malverns in the May Hill and Tortworth areas (text-fig. 11,
cols. 27-9). At May Hill the communities range from Eocoelia through Pentameroides
and Costistricklandia to Clorinda, and at Tortworth the Stone succession (col. 28) also
shows a progressive sequence. This last sequence contains a 15-ft. (5-m.) thick basalt
within the beds containing the C6 Costistricklandia Community. But at the other end of
the Tortworth Inlier (at Charfield Green, text-fig. 11, col. 29) the basalt has thickened
to 150 ft. (50 m.), and is followed directly by beds containing an Eocoelia Community.
We conclude that the sudden shallowing of the shelf sea by 150 ft. is directly responsible
for the replacement of a Costistricklandia Community by an Eocoelia Community;
this implies that the depth range of the intervening Pentameroides Community is signi-
ficantly less than 150 ft.

CONCLUSIONS

The elucidation of the evolution of several brachiopod genera has enabled wide cor-
relation of the 13 subdivisions of the Llandovery established in the type area by O. T.
Jones. Recognition of similar assemblages of fossils at different times has led to the
establishment of five main animal communities. In the Silurian the marine communities
of the shelf were largely epifaunal and relatively independant of substrate; depth was
thus the main controlling factor in community distribution.

ZIEGLER, COCKS, AND McKERROW: LLANDOVERY TRANSGRESSION 111

In the shelf regions on the Welsh Borderland, the Llandovery always rests uncon-
formably on older rocks; on the margin of the basin there is only a small break at the
very base of the Lower Llandovery ; and in the basin of Central Wales there is conformity
between the Ordovician and Silurian. From Ax to C6 there was a progressive spread of
the sea to the east and south. In the Lower Llandovery marine deposits are present at
Welshpool and Garth; by Middle Llandovery times West Shropshire had been sub-
merged ; and by early Upper Llandovery (Q) times the sea had transgressed as far as the
Malverns and May Hill; it did not reach Tortworth until the late Upper Llandovery
(C5). The theory (Jones 1938) that there was a sudden submergence of the whole area
at the beginning of the Upper Llandovery cannot now be accepted.

With few exceptions, each local sequence shows a progressive increase of depth with
time. This progressive depth increase can even span periods of non-deposition, for
example the C5 beds above the non-sequence in the Malverns were laid down in deeper
water than in C2 beds below the break. We have concluded that, where a deeper-water
community is present at the base of a local sequence, the sea was present for some time
before deposition commenced. It follows that the spread of the sea with time was even
more uniform than the sedimentary sequences record. It also follows that there were
periods when no sediments were being deposited in certain parts of the sea floor; this
may have been due either to lack of sediment supply or to no deposition or erosion in
current-swept areas.

The Welsh Borderland is on the south-eastern margin of a Lower Palaeozoic geo-
syncline which included the Lake District, the Southern Uplands of Scotland and much
of central Ireland. Both this margin and the western margin in County Galway show
an extension of the sea during Llandovery times, but at the same time north-east New-
foundland (and perhaps also parts of the Midland Valley of Scotland, for example,
Lesmahagow), show a transition from deep-water basin sediments to shelf or even non-
marine conditions. When the Welsh Borderland is viewed as part of this larger region,
it seems improbable that the advance of the sea during Llandovery times was due to
eustatic deepening.

Although there was probably little structural control of sedimentation from place to
place within the Welsh Borderland at this time, there were three major tectonic regions
affecting the area:

(i) The basin region to the west which subsided rapidly, accommodating large thick-
nesses of sediment (nearly 10 000 ft. (3300 m.) of Llandovery beds at Llanidloes).

(ii) The shelf region, which subsided slowly throughout the whole of Llandovery time,
and upon which thinner sediments accumulated.

(iii) A region of uplands to the south and east, which provided the sediment for both
of the other regions.

REFERENCES

Andrew, g. 1925. The Llandovery rocks of Garth, Breconshire. Q. Jl geoL Soc. Loud. 81, 389-406,
pi. 22.

— — and jones, o. t. 1925. The relations between the Llandovery rocks of Llandovery and those of
Garth. Ibid. 407-16.

bassett, d. a. 1963. The Welsh Palaeozoic geosyncline: a review of recent work on stratigraphy and
sedimentation. In The British Caledonides, ed. M. R. W. Johnson and F. H. Stewart, 35-69. Oliver
& Boyd, Edinburgh and London.

778

PALAEONTOLOGY, VOLUME 11

butler, a. j. 1937. On Silurian and Cambrian rocks encountered in a Deep Boring at Walsall, South
Staffordshire. Geol. Mag. 74, 241-57.

cave, r. 1965. The Nod Glas sediments of Caradoc age in North Wales. Geol. Jnl, 4, 279-98, pi. 12.
cocks, l. r. m. 1967m Llandovery stropheodontids from the Welsh Borderland. Palaeontology, 10,
245-65, pi. 37-9.

19676. Depth patterns in Silurian Marine Communities. Marine Geol. 5, 379-82.

— and rickards, r. b. (in press). Five boreholes from Shropshire and the relationships of shelly and
graptolitic facies in the Lower Silurian. Q. Jl geol. Soc. Land.

and walton, g. 1968. A large temporary exposure in the Lower Silurian of Shropshire. Geol.

Mag. 105, 390-7.

curtis, m. l. k. 1955. Lower Palaeozoic. In Bristol and its adjoining counties, ed. C. M. Maclnnes and
W. F. Whittard, 3-7. Brit. Assn Adv. Sci. Bristol.

Eastwood, t., whitehead, t. h., and robertson, t. 1925. The geology of the country around Birming-
ham. Sheet 168. Mem. Geol. Surv. ( U.K. ). 1-152, pi. 1-7.
elles, G. l. and wood, e. m. r. 1901-18. A monograph of British graptolites. Palaeontogr. Soc.
[Monogr.], 1-539, pi. 1-52.

gardiner, c. i. 1920. The Silurian rocks of May Hill. Proc. Cotteswold Nat. Fid Club, 20, 185-218, pi. 1.
garwood, e. J. and Goodyear, E. 1918. On the geology of the Old Radnor District, with special refer-
ence to an algal development in the Woolhope Limestone. Q. Jl geol. Soc. Lond. 74, 1-30, pi. 1-7.
green, G. w. 1962. In Summ. Progr. geol. Surv. (U.K.) (for 1961), p. 29.

greig, d. c., wright, J. e., and mitchell, G. h. 1968. Geology of the country around Church Stretton,
Craven Arms, Wenlock Edge and Brown Clee. Sheet 166. Mem. geol. Surv. (U.K ). 1-379, pi. 1-13.
groom, t. t. 1899. The geological structure of the Southern Malverns, and of the adjacent district
to the west. Q. Jl geol. Soc. Lond. 55, 129-69, pi. 13-15.

1900. On the geological structure of portions of the Malvern and Abberley Hills. Ibid. 56, 138-97,

pi. 8.

— — 1910. The geology of the Malvern and Abberley Hills, and the Ledbury District. Geol. Assoc.,
Jubilee Vol. 698-738, pi. 23.

holl, H. b. 1865. On the geological structure of the Malvern Hills and adjacent districts. Q. Jl geol.
Soc. Lond. 21, 72-102.

jones, o. t. 1925. Geology of the Llandovery District. Part 1 : the southern area. Ibid. 81, 344-88, pi. 21.
1938. On the evolution of a geosyncline. Ibid. 94, lx-cx.

1947. The geology of the Silurian rocks west and south of the Carneddau Range, Radnorshire.

Ibid. 103, 1-36, pi. 1.

1949. Geology of the Llandovery District. Part 2: the northern area. Ibid. 105, 43-64, pi. 3.

jones, w. d. v. 1945. The Valentian succession around Llanidloes, Montgomeryshire. Ibid. 100 (for
1944) 309-32, pi. 14.

jukes, J. b. 1853. On the occurrence of Caradoc sandstone at Great Barr, South Staffordshire. Ibid.
9, 179-81.

king, w. b. r. 1928. The geology of the district around Meifod, Montgomeryshire. Ibid. 84, 671-702,
pi. 52.

kirk, N. h. 1951. The Upper Llandovery and Lower Wenlock rocks of the area between Dolyhir and
Presteigne, Radnorshire. Proc. geol. Soc. Lond. 1471, 56-8.
lamont, a. and gilbert, d. f. l. 1945. Upper Llandovery Brachiopoda from Coneygore Coppice and
Old Storridge Common, near Alfrick, Worcs. Ann. Mag. nat. Hist. [11] 12, 641-82, pi. 3-7.
lapworth, c. 1878. The Moffat series. Q. Jl geol. Soc. Lond. 34, 240-346, pi. 1 1-13.
lawson, j. d. 1955. The geology of the May Hill Inlier. Ibid. Ill, 85-116, pi. 6.

Phillips, j. 1848. The Malvern Hills, compared with the Palaeozoic Districts of Abberley, Woolhope,
May Hill, Tortworth, and Usk. Mem. geol. Serv. (U.K.) 2 (1), 1-330, pi. 1-3.
phipps, c. b. and reeve, f. a. e. 1967. Stratigraphy and geological history of the Malvern, Abberley
and Ledbury Hills. Geol. Jnl, 5, 339-68.

pocock, r. w. 1930. The Petalocrinus Limestone horizon at Woolhope (Herefordshire). O. Jl geol.
Soc. Lond. 86, 50-63, pi. 6, 7.

whitehead, t. h., wedd, c. b., and robertson, T. 1938. Shrewsbury District, including the Han-

wood Coalfield. Sheet 152. Mem. geol. Surv. (U.K.). 1-297, pi. 1-8.

ZIEGLER, COCKS, AND McKERROW: LLANDOVERY TRANSGRESSION 779

reading, h. G. and poole, a. b. 1961. A Llandovery shoreline from the Southern Malverns. Geol. Mag.
93, 295-300, pi. 15, 16.

1962. Malvern structures. Ibid. 99, 377-9.

reed, f. r. c. and Reynolds, s. h. 1908. On the fossiliferous Silurian rocks of the southern half of the
Tortworth Inlier. Q. Jl geol. Soc. Load. 64, 512-45.

Reynolds, s. h. 1907. A Silurian inlier in the Eastern Mendips. Ibid. 63, 217-40, pi. 18.

1912. Further work on the Silurian rocks of the Eastern Mendips. Proc. Bristol Nat. Soc. 3, 76-82.

1924. The igneous rocks of the Tortworth Inlier. Q. Jl geol. Soc. Load. 80, 106-12, pi. 7, 8.

1929. The geology of the Bristol District. Proc. Geol. Assoc. 40, 77-103.

robertson, t. 1926. The section of the new railway tunnel through the Malvern Hills at Colwall.

Summ. Progr. geol. Surv. ( U.K .) (for 1925), 162-75.
st. Joseph, j.k. s. 1935. A critical examination of Stricklandia(= Stricklandinia ) lirata (J. deC. Sowerby)
1839 forma typica. Geol. Mag. 72, 401-24, pi. 16, 17.

1938. The Pentameracea of the Oslo region. Norsk geol. Tidsskr. 17, 225-336, pi. 1-8.

shotton, F. w. 1927. The conglomerates of the Enville Series of the Warwickshire Coalfield. Q. Jl
geol. Soc. Load. 83, 604-21, pi. 47, 48.

sollas, w. j. 1879. On the Silurian district of Rhymney and Pen-y-lan, Cardiff. Ibid. 35, 475-507, pi. 24.
squirrell, h. c. and tucker, e. v. 1960. The geology of the Woolhope Inlier (Herefordshire). Ibid.
116, 139-85, pi. 15.

strahan, A. 1913. Batsford (or Lower Lemington) Boring, near Moreton-in-Marsh. Summ. Progr.
geol. Surv. (U.K.) (for 1912), 90-1.

symonds, w. s. 1872. Records of the rocks. John Murray, London, 433 pp.

toghill, p. 1968. The graptolite associations and zones of the Birkhill Shales (Llandovery) at Dobb’s
Linn. Palaeontology , 11, 654-68.

tyler, w. h. 1925. Notes on Sheet 48 N.W. (Shropshire). Proc. Geol. Assoc. 36, 377.
wade, a. 1911. The Llandovery and associated rocks of north-eastern Montgomeryshire. Q. Jl geol.
Soc. Loud. 67, 415-59, pi. 33-6.

watts, w. w. 1885. On the Igneous and Associated Rocks of the Breidden Hills in East Montgomery-
shire and West Shropshire. Ibid. 41, 532-46.

whittard, w. f. 1928. The stratigraphy of the Valentian rocks of Shropshire: the Main Outcrop. Ibid.
83 (for 1927), 737-59, pi. 56, 57.

1932. The stratigraphy of the Valentian rocks of Shropshire. The Longinynd-Shelve and Breidden

Outcrops. Ibid. 88, 859-902, pi. 58-62.

Whittington, h. B. 1938. The geology of the district around Llansantffraid ym Mechain, Montgomery-
shire. Ibid. 94, 423-57, pi. 38, 39.

williams, a. 1951. Llandovery brachiopods from Wales with special reference to the Llandovery
District. Ibid. 107, 85-136, pi. 3-8.

wills, l. j. and laurie, w. h. 1938. Deep sewer trench along the Bristol road from Ash ill Road near
the Longbridge Hotel to the City Boundary at Rubery, 1937. Proc. Birmingham nat. Hist. phil. Soc.
16, 175-80, pi. 1.

wilkens, l. G. and hubbard, G. H. 1925. The Upper Llandovery Series of Rubery. Ibid. 15, 67-83.

wood, e. m. r. 1906. The Tarannon Series of Tarannon. Q. Jl geol. Soc. Lond. 62, 644-701, pi. 47, 48.
ziegler, a. m. 1964. The Malvern Line. Geol. Mag. 101, 467-9.

1965. Silurian marine communities and their environmental significance. Nature, Lond. 207,

270-2.

1966m Unusual stricklandiid brachiopods from the Upper Llandovery beds near Presteigne,

Radnorshire. Palaeontology, 9, 346-50, pi. 58.

1 9666. The Silurian brachiopod, Eocoelia hemisphaerica (J. de C. Sowerby) and related species.

Ibid. 523-43, pi. 83, 84.

cocks, l. r. m. and bambach, r. k. 1968. The composition and structure of Lower Silurian marine

communities. Lethaia, Oslo, 1, 1-27.

and boucot, a. j. 1968. North American Silurian animal communities. In W. B. N. Berry, and

A. J. Boucot. Silurian of North America. Mem. geol. Soc. Am. (in press).

780

PALAEONTOLOGY, VOLUME 11

APPENDIX 1

Localities of collections in Oxford University Museum

Area Locality

Tortworth 1

2

3

May Hill 4

5

6

7

8
9

10

11

12

13

14

15

16

17

18

19

20
21
22

Malvern Hills 23

24

25

26

27

28

29

30

31

32

33

Old Storridge Common 34

35

36

37

38

39

40

41

Ankerdine Hill 42

43

44

45

Field designation

Grid reference

T-M-A

(ST/7268 9212)

T-R-A

(ST/7233 9234)

T-D-A

(ST/6962 9390)

M-N-A

(SO/7014 2104)

M-D-A

(SO/7055 1987)

M L-A

(SO/7052 1838)

M-L-B

(SO/7049 1838)

M-L-C

(SO/7048 1839)

M-Q-A

(SO/7066 1994)

M-Q-B

(SO/7067 1993)

M-H-A

(SO/7047 1811)

M-B-A

(SO/7041 1974)

M-B-B

(SO/7039 1973)

M-G-C

(SO/7051 2097)

M-F-D

(SO/6874 2154)

M-F-A

(SO/6874 2154)

M S-A

(SO/6918 2261)

M-S-B

(SO/6932 2270)

M-S-C

(SO/6936 2271)

M-F-C

(SO/6872 2155)

M-O-A

(SO/6869 2244)

M-G-B

(SO/7055 2103)

H-O-A

(SO/7533 3764)

H-H-A

(SO/7487 3595)

H-T-A

(SO/7494 3492)

H-G-A

(SO/7612 3811)

H-G-D

(SO/7612 3811)

H-G-E

(SO/7612 3811)

H-G-F

(SO/7608 3811)

H-P-A

(SO/7646 4594)

H-C-A

(SO/7672 4442)

H-R-A

(SO/7613 4784)

H-S-A

(SO/7616 4723)

H-L-A

(SO/7467 5115)

H-F-C

(SO/7444 5124)

H-M-A

(SO/7430 5152)

H-M-B

(SO/7430 5152)

H-L-C

(SO/7464 5108)

H-L-B

(SO/7464 5111)

H-F-B

(SO/7438 5114)

H-M-C

(SO/7405 5167)

A-A-A

(SO/7352 5625)

A-H-A

(SO/7376 5696)

A-Q-A

(SO/7363 5630)

A-K-A

(SO/7347 5690)

Collections from loose blocks.

ZIEGLER, COCKS, AND McKERROW: LLANDOVERY TRANSGRESSION 781

Area

Locality

Field designation

Grid reference

Wenlock Edge

46

Morrellswood

(SJ/6284 0637)

47

Boathouse A

(SJ/6210 0379)

48

Boathouse B

(SJ/6206 0390)

49

Boathouse C

(SJ/6205 0398)

51

Sheinton B

(SJ/6116 0310)

52

Merrishaw A1

(SJ/5805 0090)

53

Merrishaw A2

(SJ/5805 0090)

54

Merrishaw B

(SJ/5852 0065)

55

Domas

(SJ/5936 0062)

56

Hughley

(SO/5605 9747)

58

W all-u-Hey wood

(SO/5120 9276)

59

Ticklerton

(SO/4821 9088)

60

Onny River

(SO/4260 8532)

Church Stretton

61

Marshbrook

(SO/4341 8982)

62

Hillend Farm

(SO/3956 8769)

Norbury

63

Norbury

(SO/3587 9284)

Bog

65

Bog A

(SO/3510 9815)*

66

Bog B

(SO/3510 9815)*

Minsterley

67

Ox Wood Dingle

(SJ/2909 0123)

68

Venusbank

(SJ/3534 0125)

69

Hope Outlier

(SJ/3482 0140)

70

Hope Quarry

(SJ/3551 0208)

71

Hope Brook

(SJ/3578 0212)

72

Josey’s Wood A

(SJ/3656 0213)

73

Josey’s Wood B

(SJ/3653 0221)

APPENDIX 2

Localities of collections in British Museum ( Natural History )

Area

Locality

Field designation

Grid reference

Wenlock Edge

50

Sheinton A

(SJ/6099 0324)

57

Gilberries

(SO/5115 9361)

Norbury

64

Linley Wall

(SO/3587 9284)*

Minsterley

74

Josey’s Wood C

(SJ/3641 0216)

75

Minsterley Lane

(SJ/3803 0487)

APPENDIX 3

Localities of collections in United States National Museum

Area

Locality

U.S.N.M. no.

Grid reference

Tortworth

76

10215

(ST/6879 9574)

77

10217

(ST/7056 9428)

78

10218

(ST/7267 9206)

79

10219

(ST/7267 9206)

80

10220

(ST/7055 9426)

81

10221

(ST/7055 9426)

82

10223

(ST/7045 9440)

* Collections from loose blocks.

782

PALAEONTOLOGY, VOLUME 11

Area

Locality

U.S.N.M. no.

Grid reference

May Hill

83

10225

(SO/6927 2263)

84

10226

(SO/6927 2263)

Woolhope

85

10230

(SO/5804 3730)

86

10232

(SO/5948 3564)

Old Storridge Common

87

10237

(SO/7496 5156)

Presteigne

88

10242

(SO/3177 6343)

89

10243

(SO/3197 6334)

90

10244

(SO/3160 6320)

91

10245

(SO/3153 6316)

92

10246

(SO/3155 6317)

93

10247

(SO/3018 6235)

Eastern Mendips

94

12936

(ST/6755 4585)*

* Collections from loose blocks.

A. M. ZIEGLER

Department of the Geophysical Sciences
University of Chicago
Illinois 60637, U.S.A.

L. R. M. COCKS

Department of Palaeontology
British Museum (Natural History)
London S.W. 7

W. S. MCKERROW

Department of Geology and Mineralogy
Parks Road
Oxford

Typescript received 1 March 1968

THE UNUSUAL BRACHIAL SKELETON OF
ATTENUATELLA CONVEXA S P. NOV.
(BRACHIOPODA)

by JOHN ARMSTRONG

Abstract. Excellently preserved specimens of Attenuatella convexa sp. nov. have enabled reconstruction of
a unique brachial skeleton for Attenuatella. Apart from this characteristic the genus is closest to members of the
Ambocoeliinae (Spiriferida) and its inclusion in this subfamily is still the most plausible supra-generic grouping.
Globally, eleven species of Attenuatella are now documented and the known range of the genus is from the lower
Artinskian (Aktastinian) to the Kazanian or Wordian Stage. The species Attenuatella convexa sp. nov., Attenua-
tella sp. A, and Attenuatella sp. cf. incurvata Waterhouse are here described from Queensland.

Attenuatella Stehli (1954) is a small, very distinctive spiriferid which occurs at
only a few localities throughout the world. The genus has recently been documented
from Australia by Waterhouse (1967) and by Armstrong and Brown (1968) who
describe species from New South Wales and Queensland respectively. Several additional
occurrences are known from the Permian faunas of Queensland and the species from
these are described here. Serial acetate peels of well-preserved specimens of one of these
species, Attenuatella convexa sp. nov., have supplemented the hitherto only fragmentary
information that is available about the brachial skeleton of the genus.

Specimens of Attenuatella from a number of the localities in Queensland were kindly lent by Dr. J. M.
Dickins of the Bureau of Mineral Resources, by Dr. J. F. Dear of the Geological Survey of Queensland,
and by Dr. B. N. Runnegar of the University of Queensland. Dr. Dickins made available specimens
which he collected from three localities in the Bowen Basin, and Dr. Dear and Dr. Runnegar each
supplied a specimen which they had collected from the Barfield Formation in the Bowen Basin, and
from just below the Upper Limestone near Gympie respectively. Bureau of Mineral Resources localities
referred to, occur on the Duaringa and Taroom 1 : 250,000 military maps and are designated by a
number prefixed by Du and T respectively. Geological Survey of Queensland, and University of Queens-
land Department of Geology and Mineralogy locality numbers are prefixed by GSQL and UQL
respectively. All specimens mentioned have individual numbers which follow the initials of the institu-
tion in which the specimens are housed; CPC, Commonwealth palaeontological type collection.
Bureau of Mineral Resources, Canberra; GSQF, Geological Survey of Queensland; UQF, Department
of Geology and Mineralogy, University of Queensland.

Order spiriferida Waagen 1883
Subfamily ambocoeliinae George 1931

Remarks. The brachidia of Attenuatella are unusual and are very distinctive. Serial
acetate peels prepared from ten specimens of Attenuatella convexa sp. nov. do not show
any traces of spiralia but they reveal the structures shown in text-fig. 1. Waterhouse
(1964) has also figured serial sections of specimens of a species of Attenuatella. He found
that each brachidium in specimens of A. incurvata Waterhouse consisted of an anteriorly
directed lamella which at the front of the shell gives rise to a semicircle-like structure
parallel to the hinge of the shell (Waterhouse 1964, fig. 48). The brachidia of the

[Palaeontology, Vol. 11, Part 5, 1968, pp. 783-92, pi. 142.]

784

PALAEONTOLOGY, VOLUME 11

sectioned specimens of A. convexa, although initially similar to the preserved brachidia
of A. incurvata, are more complete and they have enabled reconstruction of brachidia
presumably characteristic of the genus (text-fig. 2) though those of the type species are
unknown. Each brachidium of A. convexa consists of an anteriorly directed lamella

text-fig. 1. Serial acetate peels of four (A, B, C, D) specimens of Attenuatella convexa sp. nov. The
interval between successive sections is 025 mm. except where indicated, and in the latter cases the
dimensions given on the diagram are in millimetres.

which leads into an S-shaped structure (text-fig. 1) from the ventral end of which there
is a ventrally placed posteriorly directed lamella. The consistency of the arrangement
of these features in the sectioned specimens of A. convexa suggests that they represent
the complete brachial skeleton of the species. Spiral lamellae are not present in A. con-
vexa but rather the brachial skeleton is basically loop-like with curious anterior modi-
fications.

The nature of the brachidia of Attenuatella does not necessitate its reclassification.
Indeed except for the fact that the spiralia characteristic of several members of the
Ambocoeliinae (George 1931, Veevers 1959, Vandercammen 1956) are absent from

J. ARMSTRONG: BRACHIAL SKELETON OF ATTENUATELLA 785

Attemiatella, the characteristics of Attenuatella completely support its classification in
this subfamily. Attenuatella has an impunctate shell, a shell form which is comparable
with other ambocoelins, a spinose micro-ornament, areas on both valves, and a non-
striated tuberculate cardinal process. The placement of Attenuatella in the Ambocoe-
liinae is therefore maintained. A spiriferid having abbreviated brachidia resembling
those of Attenuatella is the Middle Ordovician genus Protozyga. Protozyga possesses a
jugum but otherwise each of its brachidia has the same basic form of a single volution
parallel to the median plane of the shell as the brachidia of Attenuatella.

text-fig. 2. Diagrammatic reconstruction of an antero-ventral view of the
brachidia of Attenuatella convexa sp. nov., based on serial acetate peels of
ten specimens, four of which are shown in text-fig. 1 .

The reason for the brachidial deviation in Attenuatella from the normal type of
spiriferid spiral is problematic and it is conjectural whether the lophophore of Attenua-
tella had the form of a spiral. The shape of the brachidia of A. convexa does not pre-
clude the presence of a spiral-like lophophore, and if such an organ did exist it appears
that in general only the first flexured whorl of the lophophore was provided with a cal-
careous support. If the lophophore was not spirolophous but rather was confined to the
reduced brachidia it is possible that the S-shaped brachidial modifications were developed
to compensate the loss of length of the lophophore. In this instance the feeding organ
was probably zygolophous. The brachial skeleton of Attenuatella seems to be of a more
primitive type than the spiralia of other ambocoelins. Perhaps it is relevant that species
of Attenuatella are the youngest known representatives of the Ambocoeliinae and that
reversion to this loop-like arrangement in Attenuatella occurred not long before the
extinction of the subfamily in the Upper Permian.

786

PALAEONTOLOGY, VOLUME 11

Genus attenuatella Stehli 1954

Attenuatella Stehli; Chernyak 1963
Attenuatella Stehli; Waterhouse 1964

Type species (original designation). Attenuatella texana Stehli 1954, pi. 25, figs. 31-3 from the lower
part of the Lower Permian Bone Spring Formation, Texas, text-fig. 3 (1).

Diagnosis. Small essentially plano-convex spiriferids; ventral valve elongate, smooth,
or sulcate, with elongate relatively narrow umbo and more or less incurved beak; ventral
area apsacline; dorsal valve convex, flat, or gently concave with low area; ornament of
small spines often arranged in concentric lines, and fine discontinuous radial grooves;
each groove runs anteriorly for a short distance from a spine; field of muscular attach-
ment in ventral valve an elevated platform; no plates support the teeth; crural plates
sessile and cardinal process tuberculate; shell impunctate.

Other species and specimens

A. acutirostrus (Krotova) 1885, pi. 4, fig. 24 of Artinskian age from the Ural Mountains in Russia,
text-fig. 3 (2).

A. stringocephaloides (Chernysheva and Likharev) in Likharev and Einor 1939, pi. 13, fig. 5a, b from
Abrosimov Bay, Novaya Zemlya, text-fig. 3 (3). According to Likharev and Einor (1939, p. 189) a
number of the species occurring with A. stringocephaloides at Abrosimov Bay (Likharev and Einor
1939, p. 185) also occur in the Spirifer Limestone and Brachiopod Chert in Spitzbergen. Likharev and
Einor further intimate (p. 188) that all of the faunas of Novaya Zemlya are to be equated with the
faunas of the Spirifer Limestone and the Brachiopod Chert for which Stepanov (1957) has suggested
either an upper Lower Permian or a lower Upper Permian age or perhaps both (i.e. belonging to his
Svalbardian Stage).

A. attenuata (Cloud) 1944, pi. 17, figs. 22-5 from the Wordian Waagenoceras Zone in Mexico,
text-fig. 3 (4).

A. stringocephaloides (Chernysheva and Likharev); Chernyak 1963, pi. 42, figs. 3 and 4 from the
lower Baykura sub-horizon of eastern Taimyr, text-fig. 3 (5). Ustritskiy and Chernyak (1963) correlate
the lower Baykura sub-horizon with the lower Vorkuta Series in the Pechora Basin and they consider
that both of these units correspond to the Ufa Suite on the Russian Platform. Stepanov (1957) intimated
that the latter Suite should be included in the Svalbardian Stage which he proposed (1957) for faunas
in Spitzbergen intermediate in aspect between the faunas of the Artinskian Stage and the Kazanian
Stage. For this time interval Ustritskiy (1960) also proposed a new stage, the Paykhoyian Stage based on
the fauna of the Vorkuta Series, and in 1963 Ustritskiy and Chernyak had little hesitation in referring
the fauna of the lower Baykura sub-horizon to the upper part of Ustritskiy’s stage. Ustritskiy and
Chernyak also suggest, in view of the absence of a complete Permian section in Spitzbergen, that it
might be preferable to recognize the Paykhoyian Stage rather than the Svalbardian Stage.

A. taimyrica Chernyak 1963, pi. 42, figs. 5-9 from the Byrranga horizon of eastern Taimyr, text-fig.
3 (6). Ustritskiy and Chernyak (1963) correlate the Byrranga horizon with a composite Artinskian-
Kungurian Stage in the Ural Mountains.

A. incurvata Waterhouse 1964, pi. 20, figs. 1-12, pi. 21, figs. 1-9 from the Arthurton Group in New
Zealand, text-fig. 3 (7). Waterhouse (1964, 1967) considers the fauna in this unit to be of Kazanian age.

A. sp. Landis and Waterhouse, 1966, pi. 1, figs. 1-5 from the Wesney Siltstone in the Eglinton
Volcanics, New Zealand, text-fig. 3 (8). Landis and Waterhouse seem to be uncertain about the age
of these specimens. They firstly state (p. 139) that the specimens 'are probably identical with Aktastinian
(lower Artinskian) Attenuatella in the Takitimu Group', but later (p. 145) intimate a Kazanian age
from their remark regarding the age of the Wesney Siltstone, that ‘a Kazanian age is possible on an
objective consideration of the fossils alone’.

A. sp. Landis and Waterhouse 1966, p. 144 also mention some specimens of Attenuatella from New
Caledonia, and after a preliminary examination of the material they consider that the specimens could
be conspecific with A. incurvata, text-fig. 3 (9).

J. ARMSTRONG: BRACHIAL SKELETON OF ATTENUATELLA

787

A. multispinosa Waterhouse 1967, pi. 24, figs. 1-7 from the Gilgurry Mudstone in the Boorook
Group in northern New South Wales, text-fig. 3 (10). Waterhouse (1967, p. 172) considers that this
species is likely to be of late Kungurian or early Kazanian age.

Australis Armstrong and Brown (1968) describe a new species from below the Gigoomgan
Limestone in the Maryborough Basin, Queensland, text-fig. 3 (11). They suggest a lower Artinskian
(Aktastinian) age for the species.

A. convexa sp. nov. (see p. 788), text-fig. 3 (12).

A. sp. A (see p. 790), text-fig. 3 (13).

A. cf. incurvata (see p. 791), text-fig. 3 (14).

text-fig. 3. Global distribution of Atteiuiatella. Each solid triangle signifies at least one occurrence of
the species nominated by the adjacent number. References to the numbers are in the text.

Comparison. Attenuatella is distinguished from Crurithyris, its closest relative by its
narrower more elongate ventral valve and umbo, and the elevated platform of muscular
attachment in the ventral valve. It is also possible to separate the genera on the nature
of their brachidia. The brachial skeleton of A. texana is unknown. However, from the
work of Waterhouse (1964) and from information gained from A. convexa it is known
that each brachidium of these two species of Attenuatella consists essentially of an
anteriorly directed lamella which eventually curves ventrally and then is deflected pos-
teriorly to form a single volution parallel to the median plane of the shell. In contrast
the brachial skeleton of Crurithyris urii (Fleming), the type species of Crurithyris, con-
sists of laterally directed spires of from three to six whorls (George 1931, p. 40). Water-
house (1964, p. 108) has further noted that on I. incurvata only one series of spines seems
to be present whereas according to George (1931, pp. 51 and 55) there are two series
of spines on C. urii. However, these differences may not be significant for on some speci-
mens of A. convexa the spines are relatively uniform in size, and on others there are both
large and small spines (PI. 142, figs. 5, 12).

Another genus superficially similar to Attenuatella is Moumina Fredericks 1919

C 6055 3 F

788

PALAEONTOLOGY, VOLUME 11

(1924), but unfortunately this genus has been little used or described since its institution.
Moreover, the only description of M. incertia (Chernysheva), the type species of Mou-
mina (OD; Fredericks 1919) is Chernysheva’s original description of the general external
appearance of the species.

Distribution. In addition to its occurrence in Australia, Attenuatella is found in New Caledonia,
Russia, North America, and New Zealand. The global distribution is shown in text-fig. 3.

Range. Lower Artinskian (Aktastinian) to Kazanian or Wordian.

Attenuatella convex a sp. nov.

Plate 142, figs. 1-12, 19

Holotype. UQF53036 from UQL3127 in the lower part of the Tiverton Formation in the north-eastern
part of the Bowen Basin, Queensland. (UQL3127, about half a mile east of Homevale Homestead,
20 miles north of Nebo on the Nebo-Collinsville road.)

Diagnosis. Ventral valve moderately inflated with a distinct sulcus, and a broad mas-
sive umbo that is not incurved over the area; commissure ligate; dorsal valve gently

EXPLANATION OF PLATE 142
All figures are x 4 except where indicated

Figs. 1-12. Attenuatella convexa sp. nov. All specimens are from the Tiverton Formation at UQL3127,
about half a mile east of Homevale Homestead, 20 miles north of Nebo on the Nebo-Collinsville
road, Queensland. 1, Internal mould of ventral valve, UQF53029. 2, 3, 4, Internal moulds of dorsal
valves (note impressions of areas of adductor muscle attachment in figs. 3 and 4), UQF53030,
UQF53031, and UQF53032 respectively. 5, 8, Latex casts from external moulds of ventral valves
(note two sizes of micro-ornamental spines on specimen in fig. 5), UQF53033 and UQF53034
respectively. 6, Lateral view of latex cast of external mould of both valves, UQF53035. 7, 9, Latex
casts of external moulds of dorsal valve and umbo of ventral valve, UQF53036 (holotype) and
UQF53037 respectively. 10, Same specimen as in fig. 9, Natural size. 11, Micro-ornament on a
ventral valve x 17 (note the relative uniformity in the size of the spines), UQF53038. 12, Internal
mould of the umbonal part of a dorsal valve showing the impression of a tuberculate cardinal
process X 12, UQF53039.

Fig. 19. Attenuatella sp. cf. convexa. Postero-ventral view of internal mould of ventral valve. UQF47256,
from the sandstone unit below the Upper Limestone near Gympie. The specimen is from the first
ridge north of Rammut Road, three-quarters of a mile east of Chatsworth, 5 miles north of Gympie,
south Queensland.

Figs. 13-18. Attenuatella sp. cf. incurvata. Waterhouse. 13, 14, 15, Internal moulds of ventral valves.
CPC901 8, CPC9019, and CPC9020 respectively from T1 11,2-7 miles north-north-west of the Cracow-
Theodore road crossing of Delusion Creek, 10 miles north-west of Cracow, Queensland. 16, Latex
cast from an external mould of a ventral valve. CPC9021 from Till. 17, Latex cast from an external
mould of a dorsal valve and the umbo of the ventral valve. CPC9022 from Till. 18, Lateral aspect
of a ventral valve. CPC9023 from Till.

Figs. 20-3. Attenuatella sp. A. 20, Latex cast from external mould of dorsal valve. CPC9027 from
Till. 21, Internal mould of dorsal valve. CPC9026 from Dul51, 1 mile south-west of Bundaleer
Homestead, 37-5 miles north of Bluff, Queensland. 22, 23, Internal moulds of ventral valves. CPC9024
and CPC9025 respectively from Dul51.

Figs. 24-6. Attenuatella sp. 24, Internal mould of ventral valve. CPC9028 from Du764, 2 miles west-
north-west of Bundaleer Homestead 37-5 miles north of Bluff, Queensland. 25, Latex cast of external
mould of specimen in fig. 24. 26, Internal mould of ventral valve (note the absence of ridges along
the margins of the platform of muscular attachment). GSQF3459 from GSQL D16, in the north-
eastern corner of portion 359, Parish Walloon, Queensland.

Palaeontology, Vol. 11

10

PLATE 142

ARMSTRONG, Attenuatella

J. ARMSTRONG: BRACHIAL SKELETON OF ATTENUATELLA 789

to moderately convex having a wide median flattening bordered by low ridges for its
entire length.

Description. The shell is biconvex with an elongate, moderately inflated ventral valve
and a semicircular, gently to moderately convex dorsal valve. Cardinal extremities are
obtusely rounded and the shell is usually widest at the mid-length of the dorsal valve.
The ventral umbo is high, and relatively broad, but is not incurved over the area. The
ventral area is apsacline and is about three times as high as the anacline dorsal area,
both areas being separated from the remainder of their valves by distinct beak ridges.
There is usually a narrow sulcus in the ventral valve corresponding to a median flatten-
ing or depression on the dorsal valve. Flanking the depression on the dorsal valve there
are two low ridges. Otherwise the dorsal valve is gently convex, being slightly more so
umbonally than towards the commissure. The commissure is ligate. The valves are
externally ornamented with small variably sized spines which lie along vaguely concen-
tric lines and are sometimes developed at the edges of growth lamellae. The lines of
spines are between OT and 0-3 mm. apart and along them 9 to 12 spines occur in each
millimetre. On some specimens there are large spines interspersed with smaller ones
(PI. 142, fig. 5) whereas on other specimens the spines are relatively uniformly sized
(PI. 142, fig. 12). A small groove extends anteriorly for a short distance from each spine.

The internal features of the ventral valve of A. convexa are similar to those of At-
tenuateUa sp. nov. of Armstrong and Brown (1968). There is little umbonal cavity
thickening, and in the apical part of the delthyrial cavity there is a small delthyrial plate
somewhat depressed below the level of the area. The musculature is similar to that of At-
tenuatella sp. nov. of Armstrong and Brown except that in A. convexa the ridges border-
ing the muscle platform are less prominent, especially posteriorly; the diductor scars
are wider; and the platform of muscle attachment is generally less elevated posteriorly.

In the dorsal valve the sockets are prominent and they lie along the margins of the
notothyrium. They are defined internally by prominent socket ridges, and medio-
anteriorly confluent with each ridge and extending anteriorly for a short distance along
the floor of the valve are the crural bases. From these arise anteriorly directed crural
lamellae which give rise to the brachidia diagrammatically represented in text-fig. 2.
The arrangement of the scars of adductor muscle attachment is shown in text-fig. 4.
The cardinal process occupies the apical part of the delthyrium and is tuberculate, the
tubercles being roughly arranged in rows radiating from the beak (PI. 142, fig. 13).

Comparison. A. convexa is readily distinguished by its massive and only gently curved
ventral umbo, its broad shell, and its convex dorsal valve. A. taimyrica from Russia is
the species morphologically closest to A. convexa. According to Chernyak’s description
and figures, A. taimyrica shares with A. convexa a massive umbo, a gently concave ventral
area and a narrow but distinct sulcus. A. taimyrica is, however, distinguished by its flat
dorsal valve.

Distribution. A specimen from the sandstone unit underlying the Upper Limestone near Gympie is
a representative of Attenuatella convexa. The specimen is an internal mould of a ventral valve which had
a broad massive umbo and a relatively low muscle platform. This is the only known occurrence in
addition to the type locality.

Age. The fauna occurring with Attenuatella convexa at the type locality is listed by Armstrong et al
(1967, p. 89) in their record of the fauna which occurs with Uraloceras lobulatum Armstrong, Dear, and

790

PALAEONTOLOGY, VOLUME 11

Runnegar. Armstong et al. suggest that the most likely age of this fauna is lower Artinskian (Aktas-
tinian) and the similarity between A. convexa and A. taimyrica provides a measure of support for
this determination.

Attenuatella sp. A
Plate 142, figs. 20-3

Material. Several deformed internal moulds of ventral and dorsal valves (CPC9024-6) and a fragment
of an external mould fromDu!51 (1 mile south-west of Bundaleer Homestead 37-5 miles north of Bluff,
central Queensland).

A B

text-fig. 4. Arrangement of the scars of muscular attachment in two dorsal valves of Attenuatella
convexa sp. nov. A, UQF53031, and B, UQF53032 are figured respectively in Plate 142 figs. 3 and 4.
cp, cardinal process; cb, crural base; md, median depressian; pA, posterior abductor scar; aA,

anterior abductor scar.

Description. Contour of ventral valve similar to that of Attenuatella incurvata; exterior
of ventral valve unknown; dorsal valve gently convex and slightly elongate; it is less
convex and narrower than dorsal valves of A. convexa but is more convex than dorsal
valves of A. incurvata ; on the fragment of external mould the spines number about
10 to 12 per mm. along roughly concentric lines; the interior of the ventral valve is similar
to interiors of ventral valves of A. incurvata in that the ridges along the edges of the
muscle platform are not as prominent as they are in Attenuatella sp. nov. of Armstrong
and Brown, or in A. nmltispinosa Waterhouse.

Comparison. Attenuatella sp. A is distinguished from A. incurvata and Attenuatella sp.
of Landis and Waterhouse by its elongate gently convex dorsal valve. Attenuatella
convexa has a broader shell and a lower, more massive ventral umbo. The most closely
comparable species are A. stringocephaloides and A. attenuata both of which have
virtually non-sulcate ventral valves and flat or gently convex dorsal valves. The lack of
internal details for A. attenuata and A. stringocephaloides of Chernysheva and Likharev
and the lack of ventral exteriors of Attenuatella sp. A precludes complete comparison
of these species. The interiors of the valves of Chernyak’s specimens of A. stringocepha-
loides are known however and again in one of the ventral valves figured by Chernyak

J. ARMSTRONG: BRACHIAL SKELETON OF ATTENUATELLA 791

there are only indistinct ridges along the edges of the muscle platform (Ustritskiy and
Chernyak 1963, pi. 42, fig. 3a). Chernyak’s specimens have flat dorsal valves and
although they are thereby distinguishable from Attenuatella sp. A Chernyak’s specimens
seem to be morphologically closest to Attenuatella sp. A.

Distribution. A dorsal valve (CPC9027) from T1 1 1 (2-7 miles north-north-west of the Cracow-Theodore
road crossing of Delusion Creek, 10 miles north-west of Cracow, Queensland) probably belongs to
this species. It is elongate and gently convex, and diverging from the beak there are two low ridges
separated by a broad median flattening. Spines on this valve number 12 per mm. along vaguely con-
centric lines.

Age. The specimens of Attenuatella which in one or another respect are morphologically close to
Attenuatella sp. A are recorded from lower Upper Permian faunas suggesting possibly a similar age for
Attenuatella sp. A.

Attenuatella sp. cf. incurvata Waterhouse 1964
Plate 142, figs. 13-18

Material. Several specimens (CPC9018-23) from the Barfield Formation at Till (above).

Description. The specimens have elongate almost non-sulcate ventral valves with
moderately inrolled beaks. The only dorsal exterior is very gently convex rostrally and
slightly concave commissurally, and two ridges similar to those on the dorsal valves of
A. convexa occur on its posterior part. Internal moulds of dorsal valves are more flat.
There are about 12 spines per mm. on the valves. The interiors of the ventral valves are
similar to those of specimens of A. incurvata. Only indistinct ridges occur along the
sides of the areas of muscle attachment in the ventral valve, and there is a pronounced
groove separating the adductor scars.

Discussion. The specimens from T1 1 1 are compared with A. incurvata on the basis of their
non-sulcate ventral valves, on the shape and contour of their ventral and dorsal valves,
and on the absence of prominent ridges along the edges of the platform of muscle
attachment in the ventral valve.

Age. Lower Upper Permian?

Attenuatella sp.

Plate 142, figs. 24-6

Material. Two specimens (GSQF3459 and CPC9028). The first specimen is from the Flat Top Forma-
tion at GSQL D16 (in the north-eastern corner of portion 359, Parish Walloon, 3-3 miles north-east
of Theodore, Queensland) and the other specimen is from Du764 (2 miles west-north-west of Bundaleer
Homestead, 37-5 miles north of Bluff, central Queensland).

Description. Both specimens are internal moulds of ventral valves and are characterized
by their posterior elongation and their incurved beaks. The external mould of the speci-
men from Du764 is non-sulcate. In the interiors of the valves there is an elevated plat-
form of muscular attachment, but as in specimens of Attenuatella sp. A and A. sp. cf.
incurvata only indistinct ridges occur along the sides of the platform. In both specimens
the areas of adductor muscle attachment are separated by a prominent groove.

The specimens are probably conspecific with either Attenuatella sp. A or 4. sp. cf.
incurvata but the absence of dorsal valves in association with the specimens from D16
and Du764 does not permit their correct assignment.

792

PALAEONTOLOGY, VOLUME 11

Acknowledgements. I would like to thank Professor D. Hill, F.R.S. for reading the manuscript and
for making a number of helpful suggestions. The work was carried out at the Department of Geology,
University of Queensland, while the author was in receipt of a Commonwealth Post-graduate Award.
The photographs are the work of Mr. J. Coker and Mr. E. Hollywood of the Photographic Depart-
ment of the University of Queensland.

REFERENCES

Armstrong, j. d. and brown, c. d. (1968). A new species of Attenuatella from the Permian of Queens-
land. Mem. Qd Mus. 15 (2), 56-64.

dear, j. f. and runnegar, b. 1967. Permian ammonoids from eastern Australia. J. geol. Soc.

Aust. 14, 87-97.

chernyak, g. e. 1963. In V. I. Ustritskiy and G. E. Chernyak. Biostratigrafiya i brakhiopody verkhnego
paleozoya Taymyra. Trudy nauchno-issled. Inst. Geol. Arkt. 134, 1-139. (Biostratigraphy and
brachiopods of the Upper Permian of Taimyr.)

cloud, p. e. 1944. Permian Brachiopoda. In R. E. King, C. O. Dunbar, P. E. Cloud, and A. K. Millar.
Geology and palaeontology of the Permian area north-west of Las Delicias, south western Coahuila,
Mexico. Spec. Pap. geol. Soc. Am. 52, 1-72.

dear, J. f. 1966. An ammonoid from the Permian of Queensland. Mem. Qd Mas. 14 (5), 199-203.
dickins, j. m. 1964. Correlation and subdivision of the Permian of Western and eastern Australia.
Rep. Int. geol. Congr. 22, Abstracts, 115-16.

Fredericks, G. 1919. Etudes paleontologiques. 2. Sur les spiriferids du Carbonifere superior de l’Oural.
Izv. geol. Korn. 38 (3), 295-324.

george, t. n. 1931. Ambocoelia Hall and certain similar British Spiriferidae. Q. Jl geol. Soc. Lond. 87,
30-61.

krotova, p. 1885. Artinskiy Yarus. Geologopaleontologicheskaya monografiya Artinskogo pescha-
nika. Trudy Obshch. Estest. imp. khar'kov. Univ. 13 (5), 314 pp. (Artinskian Stage. Geological-
palaeontological monograph of the Artinsk sandstones.)
landis, c. a. and Waterhouse, j. b. 1966. Discovery of Permian fossils in the Eglinton Volcanics,
Southland. Trans, roy. Soc. N.Z. Geol. 4 (6), 139-46.
likharev, b. k. and einor, o. l. 1939. Materialy k poznaniyu verkhnepaleozoyskikh faun Novaya
Zemli, Brachiopoda. Trudy arkt. nauchno-issled. Inst. 127, 1-245. (Contributions to the knowledge
of the upper Palaeozoic faunas of Novaya Zemlya, Brachiopoda.)
stehli, f. g. 1954. Lower Leonardian Brachiopoda of the Sierra Diablo. Bull. Am. Mus. nat. Hist.
105 (3), 261-358.

stepanov, d. L. 1957. O novom yaruse permskoy sistemy Arktiki. Vest, leningr. gos. Univ. 24, ser. geol.-
geog. 4, 20-4. (A new stage of the Permian system in the Arctic.)
ustritskiy, v. i. 1960. O granitse nizhney i verkhney Permi v Pechorskom basseyne i v Arktike. Trudy
nauclmo-issled. Inst. Geol. Arkt. 1 14. (On the boundary between the Lower and Upper Permian in the
Peckora Basin and in the Arctic.)

and chernyak, g. e. 1963. Biostratigrafiya i Brakiopody verkhnego paleozoya Taymra. Trudy

nauchno-issled. Inst. Geol. Arkt. 127, 1-245. (Biostratigraphy and brachiopods of the Upper Permian
of Taimyr).

vandercammen, a. 1956. Revision des Ambocoeliinae du Devonien de la Belgique. Bull. Inst. r. Sci.
nat. Belg. 32 (43), 1-51.

veevers, j. 5. 1959. Devonian brachiopods from the Fitzroy Basin, Western Australia. Bull. Bur.
Miner. Re sour. Geol. Geophys. Aust. 45, 220 pp.

Waterhouse, j. b. 1964. Permian Brachiopoda of New Zealand. Bull. geol. Surv. N.Z. Palaeont. 35,
1-287.

1967. A new species of Attenuatella (Brachiopoda) from Permian beds near Drake, New South

Wales. Rec. Aust. Mus. 21 (7), 167-73. john Armstrong

Department of Geology
University of Queensland
St. Lucia

Brisbane, Queensland 4067

Typescript received from author 2 May 1968 Australia

TAXONOMIC PROBLEMS IN THE
STUDY OF COCCOLITHS

by MAURICE BLACK

The Eleventh Annual Address, delivered 6 March 1968

abstract. From a review of Jurassic, Cretaceous, Tertiary, and Recent coccolithophorid and discoaster material,
it is suggested that the abrupt appearance of some complex forms, hitherto difficult to classify, may be explained
by mineralization of already existing unmineralized stocks; it also seems likely that some living un-mineralized
forms may have fossil mineralized ancestors. Some recent work on uncalcified Haptophyceae reveals interesting
parallels with the coccolithophorids. One new family, Stephanolithiaceae, is erected, and four specific names
are recorded in new combination.

Living in the present oceans there are about 200 species of coccolith-bearing algae.
About half of these fall quite naturally into two big families; the rest make up a pecu-
liarly heterogeneous collection, with numerous monospecific genera which are very
difficult to group into families or higher taxa. Attempts to devise a scheme of classifica-
tion for these organisms have not been entirely successful. The simpler schemes tend
to force very diverse genera into an inappropriately small number of families, as may be
seen in many reports on quantitative plankton-surveys. Some of the more elaborate
schemes try to solve the problem by building up a theoretical phylogenetic tree which is
used as a basis of classification. For this purpose, the evidence has been drawn exclusively
from the comparative morphology of living species, and up to the present palaeonto-
logical observations have been almost completely ignored. Twenty years ago, this may
have been excusable, but it is not so today, for there is now enough information at hand
for us to start establishing phylogenetic series on the evidence of the fossil record.

When we attempt to do this, one of the first results is the discovery that the two largest
families, the Syracosphaeraceae and Zygosphaeraceae, each with more than 50 living
species, have virtually no geological record at all. With the smaller taxa, the situation is
quite different; many of the systematically isolated genera are found to have long
ancestries, and some are survivors from once prosperous Mesozoic or Tertiary families.
Clearly, these survivors are more closely related to their own fossil ancestors than they
are to living genera of different parentage, and it is no longer reasonable to lump all the
living coccolith-bearing algae into a single family, as has been done in some of the most
recently published classifications (Kamptner 1958, pp. 68-71, Deflandre 1966, pp. 5-6).

Considerations such as these suggest that the time has now arrived when we must
look carefully at the impact of palaeontological research upon the taxonomy of living
forms. During the last two decades there has been a strong tendency to deny the possi-
bility of recognizing fossil representatives of many living taxa above the rank of species.
This is, of course, a strictly logical attitude, because the micropalaeontologist has to work
largely with individual coccoliths, and some modern genera are defined in terms that
take into account the way in which the coccoliths are assembled to form a complete
skeleton. Nevertheless, I think it is an unnecessarily defeatist attitude, because it assumes
that each fossil coccolith is to be considered as a completely isolated entity, with no

[Palaeontology, Vol. 11, Part 5, 1968, pp. 793-813, pis. 143-54.]

794

PALAEONTOLOGY, VOLUME 11

ancestors and no descendents. This, of course, is not true, and I hope to show that many
living taxa can be traced back with reasonable confidence into the Tertiary, and a few
can be followed still further into the Mesozoic. In this way, family relationships sort
themselves out as phylogenetic lineages, and the theoretical difficulties of working with
isolated coccoliths fall into the background.

PROBLEMS IN THE CLASSIFICATION OF LIVING
COCCOLITHOPHORES

From the early days, it has been found satisfactory to identify and classify the living
species on the basis of coccolith-structure. No two species are known which have exactly
similar coccoliths, and, as a working hypothesis, it is not unreasonable to assume that
this also holds good for extinct species. Coccolith-structure has likewise been relied
upon extensively for the delimitation of genera, but other characters such as the shape
of the cell and the presence or absence of a naked area at the flagellar pole have also
been used to some extent in defining genera and higher taxa. Living taxa defined purely
in terms of coccolith-morphology can be recognized with certainty when they are found
as fossils, but characters involving the arrangement of coccoliths on the cell-surface are
not as a rule observable in fossil material. With this difficulty in mind, some micro-
palaeontologists have argued that all fossil coccoliths should be assigned to provisional
taxa until a complete coccosphere has been discovered. This practice has the serious
disadvantage of requiring a dual nomenclature, with one set of names for living material
and another for fossils, thus creating an artificial break between modern forms and their
fossil ancestors.

In order to resolve this problem, we will first of all look more closely at some of the
taxonomic difficulties presented by living coccolithophores, and then try to discover
just how serious these difficulties really are when we try to work out some sort of phylo-
genetic story.

Dimorphism of coccoliths

Lohmann (1902) found that in some species the coccoliths covering a single individual
were not all alike; for example, those round the flagellar pole might bear spines, whereas
the ordinary body-coccolith did not. Usually the polar coccoliths are easily recognizable
as straightforward modifications of the normal type, and Lohmann did not in general
segregate species with this kind of dimorphism into separate genera. The only genus
which he distinguished on these grounds was Scyphosphaera , which has an equatorial
girdle of very large float-coccoliths. Taxonomic difficulties did not arise until 40 or 50
years later, when Kamptner (1941) and Deflandre (1952) took a much more serious view
of coccolith-dimorphism, and removed all species with dimorphic coccoliths into
separate genera. The consequent practical difficulty in naming fossil coccoliths was
perhaps unnecessarily exaggerated, but the theoretical implications led Deflandre to
abandon the use of a single classification for both living and fossil forms, and to create
an independent taxonomic system for the fossils. The general adoption of this policy
during the decade that followed did much to retard progress in the study of phylogenetic
and taxonomic relationships among the fossil coccoliths.

Coccolith-dimorphism is likely to have been prevalent to much the same extent in

M. BLACK: TAXONOMIC PROBLEMS IN THE STUDY OF COCCOLITHS 795

fossil as in living taxa, and although final proof is usually beyond our reach, it is not
difficult to find fossil examples of morphological variants that have all the earmarks of
dimorphic pairs. These are not uncommon in Cretarhabdus (PI. 150, figs. 4, 5), Deflan-
dr ius (PI. 151, figs. 4, 5), Kamptnerius (PI. 152, figs. 5, 6), and Rhabdolithina\ if fossil
species with dimorphic coccoliths are not recognized as such, the worst that can happen
is that two species with identical ranges appear in the fossil record instead of one.

Pleomorphic life-histories

Observations on the life-histories of species kept in culture have brought to light
several examples in which a motile coccolith-bearing phase alternates with a sessile
uncalcified phase which had previously been accepted as a member of a different genus
(Parke 1961). This is particularly true of coastal species, whose non-motile phases tend
to live attached to rocks. Distinct from these are the pelagic species, whose resting-stages
take the form of calcified cysts, which in the absence of any solid support, remain
suspended in sea-water. One of these species, Crystallolithus hyalinus, which bears organic
scales, sometimes unmineralized (PI. 154, figs. 1, 2), and sometimes covered with minute
rhombohedra of calcite, has proved to be the motile phase of Coccolithus pelagicus , an
encysted form with coccoliths of a totally different kind (PI. 143, figs. 1, 2).

These unexpectedly complicated life-histories are interesting from a taxonomic point
of view, because the individual phases are usually so different from each other that they
were put into separate genera when first discovered. To judge from our knowledge of
living species, we can hardly expect to find tangible evidence of pleomorphism in the
fossil record. In the coastal species, all the known sessile phases are uncalcified, and
hence unlikely to be preserved as fossils. In the pelagic species, on the other hand, it is
only the heavily calcified cysts, or their component coccoliths, that stand much chance
of being preserved; delicate motile forms like Crystallolithus soon disintegrate, and are
never found in bottom deposits. Indeed, material from living cultures needs careful
handling in the laboratory to prevent the organic scales from losing their cover of calcite
crystals. The cysts, on the other hand, are often preserved entire, and good examples are
quite common in rocks as old as the Kimeridge Clay (PI. 143, fig. 6).

Variation in response to external conditions

Changes in external conditions can, in some circumstances, cause the organism to
modify the appearance of its coccoliths, or even to cease calcification.

In most species that have been experimented upon, the principal effect of temperature-
change is to alter the rate of cell-multiplication, without any significant effect upon
coccolith-morphology. In Coccolithus huxleyi, however, the coccoliths grown in warm
water are reported to be distinguishable from those grown at low temperatures (PI. 145,
figs. 1, 2); the two variants have been recorded in laboratory cultures (Watabe and
Wilbur 1966) and in natural populations (McIntyre and Be 1967).

C. huxleyi is kept in culture in several laboratories, and a number of separate strains
are kept under constant observation. Some of these have been found quite suddenly
to stop growing coccoliths, and to persist in a naked condition (Paasche 1964, p. 11).
One cause of this change appears to be an enrichment in the supply of nutrients, par-
ticularly phosphate. An interesting point is that whereas some strains will resume growth

796

PALAEONTOLOGY, VOLUME 11

of coccoliths if the culture medium is adjusted to a composition more like natural sea-
water, others will not, and no known treatment will induce them to do so.

There are several very suggestive possibilities here in connection with the evolution and
taxonomy of these algae; fixation of a single phase out of a pleomorphic life-history,
independent development of ecological variants, and sudden changes from a naked to
a calcified condition, could all lead to the introduction of apparently cryptogenetic
forms into the geological record. Furthermore, some of these changes could take place
very suddenly, owing to the extraordinarily fast rate of multiplication that is prevalent
in these organisms (Parke and Manton 1962).

PHYLOGENETIC HISTORY OF SOME LIVING TAXA

Attempts to trace living coccolithophorids back into geological history lead to curi-
ously divergent results. Some species such as Coccolithus pe/agicus, C. huxleyi, and
Braarudosphaera bigelowi lead back to flourishing and diverse complexes, which in the
present oceans have been reduced to just one or two surviving representatives. The
ancestral stocks show a good deal of evolutionary divergence, but a set of unifying
family characters persists throughout. Pontosphaera provides an extreme example of
divergence, for we not only get a hint of an evolutionary connection with the extinct
Zygolithaceae, but we also find here the first suggestion of a far-distant linkage with
another living family, the Helicosphaeraceae. Other stocks seem to have been peculiarly
stable throughout their geological history. Calciosolenia can be traced back to mid-
Cretaceous times with very little change, and the coccoliths of Braarudosphaera bigelowi
are specifically indistinguishable from fossils preserved in early Cretaceous rocks. In
contrast with these are two flourishing modern families, the Syracosphaeraceae which
are abundantly represented in contemporary Globigerina Ooze, but have not been
found in deposits older than Quaternary, and the Zygosphaeraceae which are not
represented in bottom deposits at all.

Coccolithus pelagicus (Wallich) Schiller
Plate 143, figs. 1, 2

This is the most familiar of all coccoliths; it was discovered in samples of ooze raised
from the floor of the North Atlantic Ocean during the first telegraph surveys, and was
examined by Huxley and Wallich, and later by Murray and Blackman (1898). The living

EXPLANATION OF PLATE 143

Figs. 1,2. Coccolithus pelagicus (Wallich) Schiller, recent Globigerina Ooze, Discovery II Sta. 4269,
Biscay Abyssal Plain. 1, No. 20175, oblique distal view, x 4800. 2, No. 5000, details of central area,
distal view, X 5000.

Fig. 3. Coccolithus sp. cf. C. cavus Hay and Mohler; No. 22279, Upper Oligocene, core DWBG 10,
Pacific Ocean; distal view, x 5300.

Fig. 4. Coccolithus sp. cf. C. marismontium Black; No. 22284, Upper Oligocene, core DWBG 10,
Pacific Ocean; proximal view, X6700.

Figs. 5, 6. Ellipsagelosphaera sp. 5, No. 15230, Campanian Chalk, Belgorod, Russia; distal view,
x8000. 6, No. 22874, Cambridge Greensand, Cherry Hinton Fields, near Cambridge; complete
coccosphere (calcified cyst), x 3700.

Palaeontology, Vol. 11

PLATE 143

BLACK, Living, Tertiary and Cretaceous coccoliths

M. BLACK: TAXONOMIC PROBLEMS IN THE STUDY OF COCCOLITHS 797

alga is confined to the North Atlantic, and is the only species that is known to have a
preference for cold water; all attempts to find it in Antarctic waters of suitable tempera-
ture have failed. This was apparently not the case during the Pleistocene, for Allan Be
and his colleagues at the Lamont Observatory have found it in core-samples from the
southern oceans at levels dated as belonging to the middle of the Wisconsin glacial
episode (McIntyre and Be 1967). In the Tertiary, there are very similar forms, but studies
under the electron microscope show that they are specifically distinct (PI. 143, fig. 3).
The same generic form can be traced back to the Eocene, but once we pass into the
Mesozoic, the species that superficially resemble C. pelagicus are obviously different in
the finer details of their construction. When the coccoliths are examined under a petro-
logical microscope, with the fossil lying flat on the slide, Tertiary and living species give
an interference figure in which the distal shield plays no part, since the optic axes of the
component calcite crystals are at right-angles to the plane of the shield. In the Mesozoic
genus Ellipsagelosphaera this is not the case, and the crystals of the distal shield show
strong birefringence, so that the optical figure takes a complicated form, due to the
superposition of two black crosses, one produced by each shield. In addition, the Meso-
zoic species always have a corona of quadrate or keystone-shaped crystals lying on top
of the distal shield, and marking off the central area (PI. 143, fig. 5). Under an ordinary
microscope the Tertiary and Mesozoic species look very much alike, but under a polariz-
ing microscope or an electron microscope they are so obviously distinct that they are
now placed in separate subfamilies.

In the Jurassic and Cretaceous there are also several other genera such as Sollasites
(PI. 144, figs. 1, 2), which resemble Coccolithus in a general way, but differ considerably
in the finer details of their structure. They are clearly not on the main line of descent
towards the C. pelagicus stock, but some of them may prove to be ancestral to certain
other Tertiary and living forms.

Cyclococcolithus leptoporus (Murray and Blackman) Kamptner
Plate 144, figs. 3, 4; Plate 147, fig. 1

On theoretical grounds, Kamptner has argued that the circular outline is more primi-
tive than the elliptical, and has insisted upon the taxonomic separation of these two
shapes (compare PI. 143, figs. 1-6, PI. 144, figs. 3-7). He remarks:

‘Above all else, one clear-cut taxonomic principle is decisive for subdivision . . . into
tribes and subtribes: a sharper separation of the circular from the elliptical forms. A
primitive character must be attributed to the circular outline, and a derivative character
to the elliptical. ... It is also a priori conceivable that the change-over from the circular
to elliptical types has been achieved polyphyletically, and so to speak on a broad front’
(Kamptner 1958, p. 64).

In accordance with this principle, he removed Coccolithus leptoporus to a new genus,
Cyclococcolithus, and similarly split up other genera so that species with circular
coccoliths could be put into different subtribes from those with elliptical coccoliths.

Cyclococcolithus leptoporus provides a convenient starting point from which to examine
this taxonomic principle in relation to the geological record. This species is abundant
in the living plankton, and is widely distributed. Very similar coccoliths are common in
the Pliocene, and particularly so in core-samples from the ocean floor (PI. 144, fig. 4).

798

PALAEONTOLOGY, VOLUME 11

In such deposits, the two discs of the placolith are apt to become separated, and many
records of Tiarolithus and Calcidiscus from the Upper Tertiary are probably based upon
dismembered specimens of C. leptoporus or closely related species. Other species are
found in the Lower Tertiary, and the earliest record of the genus is from the Danian.
These early forms occur associated with characteristically elliptical species of Cocco-
lithus, and the two stocks were quite distinct from each other at the time of their first
appearance in the early Tertiary. Thus the Tertiary record does not give much support
to the idea that the circular shape is primitive; indeed, it tends to emphasize the complete
independence of the elliptical forms rather than to provide any suggestion of their
descent from circular ancestors.

This independent relationship is emphasized by a curiously similar state of affairs in
the Upper Jurassic, for a different set of circular and elliptical types exist side by side
in the Oxford and Kimeridge Clays. These include the genera Cyclagelosphaera and
Ellipsagelosphaera (Noel 1965), which at first sight appear to be so much like C. lepto-
porus and C. pelagicus that they were originally recorded under these names. This
resemblance, however, is illusory, for there are significant differences in micro-structure.
It is remarkable that the two Jurassic genera each differ from their Tertiary analogues
in exactly the same way: they both have strongly birefringent distal shields, surmounted
by a well-developed corona (compare PI. 143, fig. 3 and PI. 144, fig. 3 with PI. 143, fig. 5
and PI. 144, fig. 5). Consequently, if we focus our attention on the minutiae of crystal-
arrangement, we find that the Jurassic circular forms resemble their elliptical contem-
poraries more closely than they resemble modern circular forms. Are we then to unite
the Jurassic genera, both circular and elliptical, into one family or subfamily, as Noel
(1965) has done, or would it be more reasonable to keep the two shapes separate,
regarding them as two parallel stocks that became independent at an early stage, and
have remained so ever since?

An attempt to answer this question requires a closer look at the geological history of
these two stocks. The elliptical forms are abundant throughout the Mesozoic and
Cainozoic, and it may be that within this multitude of species there exists a continuous
evolutionary thread leading from the Jurassic to the present day. The only serious break
in this record is at the Cretaceous-Paleocene boundary, and there is not enough
information available at present to reach a decision. The history of the circular placoliths
is rather different. After the great abundance of Cyclagelosphaera (PI. 144, fig. 5) in the
Jurassic, there is a paucity of circular forms in the Cretaceous. The few examples that

EXPLANATION OF PLATE 144

Figs. 1, 2. Sollasites horticus (Stradner et at.) comb. nov. ( Coccolithus horticus Stradner et at.),
Cambridge Greensand near Cambridge, X9600. 1, No. 4706, Barrington, distal view. 2, No. 21739,
Cherry Hinton Fields, proximal view.

Figs. 3, 4. Cyclococcolithus spp.. Lower Pliocene, core LSDH 78P, Pacific Ocean. 3, cf. C. leptoporus
(Murray and Blackman) Kamptner, No. 22224, distal view, x4000. 4, No. 22210, group with some
specimens in the ‘Calcidiscus’ condition, x 2500.

Fig. 5. Cyclagelosphaera margereli Noel; No. 17410, Oxford Clay (L. Oxfordian), Cambridge Experi-
mental Borehole; distal view, x 10 000.

Fig. 6. Markalius sp. cf. M. inversus (Deflandre) Bramlette and Martini; No. 17199, Belemnite Marl
(L. Turonian), Cherry Hinton, near Cambridge; proximal view, x 6000.

Fig. 7. Ericsonia sp.; No. 22295, Upper Oligocene, core DWBG 10, Pacific Ocean; proximal view,
X 6000.

Palaeontology , Vo l 11

PLATE 144

BLACK, Cretaceous, Tertiary and Jurassic coccoliths

M. BLACK: TAXONOMIC PROBLEMS IN THE STUDY OF COCCOLITHS 799

are known, such as Markalius (PI. 144, fig. 6), can hardly be regarded as linking Cycla-
gelosphaera with Cyclococcolithus, because they have a more complicated structure than
either of these. They seem rather to be leading towards Ericsonia (PI. 144, fig. 7) and
other still more specialized forms that flourished in the Lower Tertiary. There is thus
a gap extending throughout the Cretaceous period, during which they were apparently
no representatives of the hypothetical Cyclococcolithus lineage. It seems, therefore,
unlikely that the Tertiary species originated as an offshoot from Cyclagelosphaera, and
it is more probable that they developed from an unknown ancestor early in the Tertiary.

The evidence from the geological record thus seems to alienate the Mesozoic circular
forms from their modern analogues, and the evidence of morphology throws them into
association with their elliptical contemporaries, in spite of the difference of shape.

Coccolithus huxleyi (Lohmann) Kamptner
(= Emiliania huxleyi (Lohm.) Hay and Mohler 1967)

Plate 145, figs. 1, 2

The taxonomic position of this species is surrounded by interesting problems. It
stands apart from C. pelagicus and all other living members of the genus in the peculiar
construction of its shields, and its conspicuous central grid, which recalls the similar
structures in many extinct forms. Mary Parke in Parke and Dixon (1964, p. 520) has
pointed out that the motile phase lacks an obvious haptonema, and in this respect it
resembles the Isochrysidales rather than the Coccolithophorales. Paasche (1964, p. 11)
found in his cultures that coccolith-secreting individuals were without visible flagella,
although these could be seen on some of the naked cells.

In the living plankton, it is undoubtedly the most vigorous and successful of the
coccolith-bearing algae. It is distributed over the whole area of the oceans from the
Antarctic convergence to the Arctic, and can invade brackish waters inaccessible to
other pelagic species. Because of its great vigour and tolerance, it has provided material
for more laboratory experiments than any other species.

C. huxleyi is apparently a very recent addition to the oceanic plankton. McIntyre and
Be (1967) have reported its first appearance in deep-sea cores as taking place within
Brunhes Normal Zone (that is, less than 700 000 years ago, and more probably nearer
to 100 000 years). They also report the presence of a coocolith intermediate between
C. huxleyi and Gephyrocapsa oceanica in cores spanning the interval between 300 000
and 100 000 years. G. oceanica (PI. 145, fig. 3) can be traced back into the Pliocene, and
survives in the living plankton. Forms with the oblique bridge characteristic of Gephyro-
capsa are unknown before the Pliocene, but the other characters of the genus, such as
the large central opening with a bilateral grid, and the non-imbricate ray-elements, are
strongly developed in a host of Lower Tertiary species of Tremalithus (PI. 145, figs. 4-6).
There is a great diversity of species with elaborate grids in the Middle and Upper
Eocene, and the same type of coccolith can be traced back still further into the Mesozoic.
The earliest representatives, with rather simpler grids, are found in the Lower Gault
j (Middle Albian) at Folkestone (PI. 145, figs. 7, 8).

The taxonomic isolation of C. huxleyi from the rest of the genus Coccolithus, which is
suspected from the peculiarities of the living alga, is thus confirmed by its long and
independent geological history, and there can be little doubt that this species should be

800 PALAEONTOLOGY, VOLUME 11

removed to a separate genus, and possibly to a family independent of the Coccolitho-
phoraceae.1

Pontosphaera discopova Schiller
Plate 146, figs. 1 , 2

The genus Pontosphaera was created by Lohmann (1902) for species in which the
coccoliths appear to be simple discs, with or without a shallow rim. The outline is
elliptical, and the coccoliths are most commonly shaped like a shallow pie-dish. Specific
diagnoses were originally based upon work under the biological microscope, and later
studies in polarized light or under the electron microscope have led to the removal of
several species to other genera. No electron micrographs of the type species, P. syracu-
sana , have yet been published, but Mrs. Gaarder has generously allowed me to compare
some of her excellent micrographs of this species with my Tertiary material. The cocco-
liths are large, with a floor perforated by about 750 very delicate pores, and the wall is
constructed of about 200 very slender and steeply inclined calcite laths. P. discopova has
similarly constructed but smaller coccoliths, with larger and less numerous pores. The
Tertiary coccoliths discussed below have the same general type of structure, and can with
reasonable confidence be referred to the same genus as P. syracusana and P. discopova.

Coccoliths of various species of Pontosphaera are not uncommon in modern Globi-
gerina Ooze, particularly beneath the warmer parts of the oceans. In the Pliocene a form
closely resembling P. discopova is occasionally found (PI. 146, figs. 4, 5), but other species
are rare or absent, and none of the modern perforated species has yet been identified
with certainty in pre-Pliocene sediments. Nevertheless, coccoliths with all the charac-
teristic features of Pontosphaera except the circular pores in the floor, are widely dis-
tributed throughout the Tertiary. There appear to be several lineages among these
Tertiary species, each with a slightly different pattern in the structure of the floor. One
of these is of special interest, since it appears to connect the living P. discopova with a
complex of Eocene forms whose ancestry can probably be traced back to the Cretaceous.
The proximal surface of the floor in this series has a bilateral arrangement of plates
similar to that originally described by Kamptner in P. scutellum, and now known in a
number of diverse Eocene forms (PI. 146, fig. 6). The distal side has an entirely different

EXPLANATION OF PLATE 145

Figs. 1,2. Coccolithus huxleyi (Lohmann) Kamptner; recent oceanic deposits. 1, warm-water form,
No. 3421, Challenger Sta. 338, S. Atlantic Ocean; distal view, x 16 000. 2, cold-water form, No. 11612,
Discovery II Sta. 3809, Galicia Bank; distal view, X 20 000.

Fig. 3. Gephyrocapsa oceanica Kamptner; No. 18072, modern Globigerina Ooze, Discovery II Sta.
4288, Biscay Mts., proximal view, X 8000.

Figs. 4-8. Tremalithus spp. 4, T. danicus (Black) comb. nov. (= Dictyococcites danicus Black);
No. 11819, Middle Oligocene, Grundfor, Denmark; proximal view, X 6000. 5, T. placomorphus
Kamptner; No. 22838, Lower Oligocene, Rodstenseje nr. Odder, Denmark; proximal view, X 4000.
6, T. dictyodus (Defiandre and Fert) comb. nov. (= Discolithus dictyodus D. and F.); No. 22831,
Lower Oligocene, Rodstenseje nr. Odder, Denmark (possibly reworked from M. Eocene); proximal
view, X 6100. 7, T. burwellensis Black; No. 14622, Cambridge Greensand (Cenomanian), Barrington
near Cambridge; distal view with proximal shield showing through, X 10 000. 8, T. sp., No. 13371,
Lower Gault (M. Albian), Folkestone, Kent; proximal view, X 8000.

1 Since this lecture was delivered, a paper by W. W. Hay et al has been received, in which C. huxleyi has
been formally transferred to a new genus, Emiliania.

Palaeontology , Vol. 11

PLATE 145

BLACK, Living, Tertiary and Cretaceous coccoliths

M. BLACK: TAXONOMIC PROBLEMS IN THE STUDY OF COCCOLITHS 801

appearance, with a pattern of concentrically arranged fibrous elements (PI. 146, figs. 3, 7).
This combination of patterns is also seen in several species of Helicosphaera (PL 147,
figs. 1-3), and will be considered further in connection with the ancestry of that genus.

This constructional pattern can be picked up again in the late Cretaceous. Two species
in the Santonian and Campanian have the same kind of wall-structure, and the same
arrangement of plates on the proximal surface. One of these, Pontolithina moorevillensis,
has two small perforations at the foci of the ellipse (PI. 149, fig. 4). This species may well
prove to be the ancestor of several Tertiary linages. Obliteration of the two pores and
an increase in the number of wall-elements would give a form very much like certain
Eocene species, for example P. versa Bramlette and Sullivan (PI. 146, fig. 6), P. vadosa
Hay et al., and possibly other forms like Discolithus oamaruensis Dell. Enlargement of
the pores to make two circular windows would lead to such forms as D. duoeavus Bram.
and Sull. (L. Eocene) and D. panarium Defl. (M. Eocene).

There are also several Eocene species which differ in having numerous small circular
pores, but which probably have a comparable floor-structure, since they give similar
interference patterns in polarized light; their ultra-structure has not yet been examined
under the electron microscope. D. punctosus Bram. and Sull. (L. Eocene) and D. dis-
tinctly Bram. and Sull. (M. Eocene) have this structure; they foreshadow the modern
perforate species of Pontosphaera, but in the absence of any recognizable intermediates
in the Upper Eocene and Miocene, we cannot say whether there is any direct phylo-
genetic connection. Possibly D. vigintiforatus, the type species of Discolithus from the
Miocene of the Vienna basin, may yet prove to be on this line of descent.

There can be little doubt that Pontosphaera has a long geological history, and that
many Tertiary species that have been referred to the rather unsatisfactory form-genus
Discolithus actually fit quite naturally into one or other of the lineages of this complex.
A derivation from some branch of the Mesozoic Zygolithaceae seems quite possible,
and is indeed suggested by the presence in the Upper Cretaceous of forms that are inter-
mediate between the two families (PI. 149, fig. 4).

Helicosphaera carteri (Wallich) Kamptner
(= Helicopontosphaera kamptneri Hay and Mohler 1967)

Plate 147, figs. 1, 2

In the present-day oceans, the genus Helicosphaera is represented by a single species
whose coccoliths are peculiar in having a spiral brim which terminates in a charac-
teristically flaring wing. Within this brim is an elliptical shield, shaped rather like the
crown of a hat; it has bilaterally arranged plates on the proximal side, and concentric
fibres on the distal, so that the structure is much the same as in many species of Ponto-
sphaera.

This species has been customarily regarded as closely related to, and possibly derived
from taxa such as Coccolithus with typical placoliths. Study of H. carteri under the
electron microscope does not give much support to this idea; the coccoliths are clearly
not mis-shapen placoliths, as is often assumed, and it is difficult to see how the spiral
flange could have been derived from the two shields of a placolith. Compared with other
living coccolithophorids, H. carteri stands very much by itself, and its geological history
abundantly confirms its independence from the Coccolithophoraceae.

802

PALAEONTOLOGY, VOLUME 11

At the present day H. carteri is confined to the parts of the oceans lying roughly
between latitudes 50° N. and S. ; in bottom sediments its range is withdrawn a little
towards the equator (McIntyre and Be 1967, p. 585). In the Pliocene, forms that would
be difficult to separate from this species are widely distributed, both in deep-sea cores
and in samples collected on land. Similar forms, still very much like H. carteri, are
abundant in the Upper Miocene, but here they are associated with several extinct
species which are obviously quite different. In the underlying parts of the Miocene and
in the Oligocene and Eocene there is a varied complex of extinct species, and the genus
was evidently undergoing a burst of evolutionary diversification, possibly at its climax
during the Oligocene.

In the early Eocene there are several species whose coccoliths lack the broad, flaring
wing that is so characteristic of later species; they have a shape that does not depart
much from an ovoid or even a regular ellipse. The earliest of these, H. seminulum , evolves
from a Lower Eocene population in which this simple type (subspecies seminulum ) is
predominant (PI. 147, fig. 4), to a Middle Eocene population in which the predominant
subspecies lophotus is much more like a typical Helicosphaera (Bramlette and Sullivan
1961). The rim in subsp. seminulum shows little more than a suggestion of spiral struc-
ture, and in fact resembles the wall of an early Tertiary Pontosphaera crossed by an
oblique wrinkle. The resemblance between these primitive species of Helicosphaera and
some of the contemporary species of Pontosphaera is so close that they might reasonably
be placed in the same genus, were it not that forms like H. seminulum are so plainly
ancestral to other more typical species of Helicosphaera.

We have thus traced the ancestry of H. carteri back to an early stage when the genus
was barely distinguishable from Pontosphaera, and we are now in a better position to
assess its taxonomic status. It clearly has no close relationship with the Coccolithopho-
raceae, and must be placed in some other family. It has sprung from the same ancestors
as the living Pontosphaera, and for this reason might be included in the Pontosphaera-
ceae. On the other hand, it can be argued that the divergence between the two stocks
since the early Tertiary has been so profound that the living representatives can hardly
be included in a single family, and I have proposed elsewhere that a new family, the
Helicosphaeraceae, should be established for H. carteri and the numerous Tertiary
species of which it is the sole survivor (Black 1968).

Braarudosphaera bigelowi (Gran and Braarud) Deflandre
Plate 147, fig. 5

In 1935 a new species, thought to be a Pontosphaera, was recorded in plankton hauls
from the Bay of Fundy (Gran and Braarud 1935); this was re-examined by Deflandre,

EXPLANATION OF PLATE 146
Figs. 1-7. Pontosphaera spp. 1-3, P. sp. cf. P. discopora Schiller, recent Globigerina Ooze, Challenger
Sta. 338, S. Atlantic Ocean. 1, No. 11393, proximal view, X 7200. 2, No. 11318, distal view, X 8000.
3, No. 11319, details of distal surface, X 24 000. 4, 5, P. sp. cf. P. discopora Schiller, Pliocene, Cisano
nr. Albenga, Italy. 4, No. 16852, proximal view, X4800. 5, No. 16854, oblique proximal view,
X 4000. 6, 7, P. versa (Bramlette and Sullivan) comb. nov. (= Discolithus versus B. and S.), Eocene,
Tuilerie de Donzacq, Landes, France. 6, No. 15855, proximal view, x6000. 7, No. 15834, distal
view, X 6000.

Palaeontology , Vol. 11

PLATE 146

BLACK, Living and Tertiary coccoliths

I

M. BLACK: TAXONOMIC PROBLEMS IN THE STUDY OF COCCOLITHS 803

who showed that its coccoliths were entirely different from any previously known form.
They consist of five peculiarly shaped calcite plates arranged to form a regular pentagon,
and twelve of these pentaliths fit together to make a dodecahedral skeleton enclosing
an encysted cell. The species was re-named Braarudosphaera bigelowi , and a new family
was based upon it (Deflandre 1947). At that time little was known about its geological
history, which we now know to be quite remarkable. Coccoliths with exactly the same
structure have been found at various levels in the Tertiary, and are abundant in the
Upper Bracklesham Beds of the Sussex coast. They have also been found, though less
abundantly, throughout the Cretaceous system, and except for an increase in size in the
older specimens, are specifically indistinguishable from recent material. No other species
of coccolithophorid has a stratigraphical range at all comparable with this, and at first
sight it might be thought that we have here an extreme example of specific stability.
This is no doubt true of the main lineage of B. bigelowi itself, but in the Tertiary we are
confronted not only with a multiplicity of other species of Braarudosphaera itself, but
in addition an exuberant development of two closely related genera, Micrantholithus and
Pemma (PI. 147, fig. 6), all with coccoliths constructed on the same pentalith system.
Detailed phylogenies have still to be worked out, but some of the simpler forms are so
close to B. bigelowi that there can be little doubt about their origin.

B. bigelowi is an unquestionable example of a single living species that is the sole
survivor of an important and diversified Tertiary family whose ancestry can be traced
back well into the Mesozoic. This point is stressed because there are several other species
in the living plankton with very much the same kind of history, resulting in a taxonomic
isolation which is equally real, but by no means always so obvious.

calciosoleniaceae Kamptner

Plate 148, figs. 1, 2

Most coccolithophorids have a spherical, egg-shaped, or pear-shaped body. The
Calciosoleniaceae differ in being cylindrical or fusiform, and their coccoliths instead of
being circular or elliptical, take the form of a narrow parallelogram (PI. 148, fig. 1).
There are probably four or five living species, and although the family characters are
unmistakable, their systematics at generic and specific levels are not easy. Fossil repre-
sentatives are never common, but have been found at intervals in the geological column
down to the Cretaceous, the earliest British occurrence being at the base of the Ceno-
manian (PI. 148, fig. 2); Stradner (in Stradner el al 1968) has recently announced the dis-
covery of similar specimens in the Albianof Flolland.The interesting feature of this record
is that the earliest specimens differ so little from living material: the family characters
with all their eccentricities are fully developed at the first appearance, there is no clue
at all about relationships to other coccolith taxa, and no suggestion of any evolutionary
change during the long interval from the middle Cretaceous to the present day.

LIVING FAMILIES WITH NO KNOWN GEOLOGICAL RECORD

We have now considered a number of isolated species in the living plankton, many of
which have turned out to be survivors of once flourishing families that are otherwise

C 6055 3 G

804

PALAEONTOLOGY, VOLUME 11

extinct. In contrast with these are the two largest living families, the Syracosphaeraceae
and Zygosphaeraceae, which have no fossil record behind them. Both families have more
than 50 species, and together they account for more than half the living coccolitho-
phorids. The Zygosphaeraceae are all holococcoliths: that is, they are constructed of
minute rhombohedra or prisms of calcite, each enclosed in an organic membrane. Such
structures are very delicate, and would probably break down into their component
crystals soon after falling to the sea floor; this may well be the reason why they are not
found in the Globigerina Ooze or other oceanic deposits. This, however, can hardly be
true of the Syracosphaeraceae, which are much more robustly constructed. Many
species of Syracosphaera are, indeed, found abundantly in the modern Globigerina
Ooze, and their absence from the Pliocene and older deposits therefore calls for some
different explanation. (W. W. Hay and his colleagues have recently referred two
Tertiary species to this genus (S. bisecta Hay et al. 1966, p. 393, and S. clathrata, 1967,
p. 449). S', bisecta has been re-examined by Bramlette and Wilcoxon (1967, p. 102),
who regard it as a species of Coecolithus, and until more is known about the wall-
structure of S. clathrata, its reference to Syracosphaera cannot be regarded as fully
established.) The obvious conclusion that the Syracosphaeraceae are, in fact, post-
Tertiary additions to the oceanic plankton is probably correct, and will be considered
later in relation to the abrupt appearance of other cryptogenic families in earlier times.

EXPLANATION OF PLATE 147

Figs. 1-4. Helicosphaera spp. 1, 2, H. carteri (Wallich) Kamptner, recent Globigerina Ooze. 1, No.
18030, Discovery II Sta. 4288, Biscay Mts., proximal view, X 7200. Also Cyclococcolithus leptoporus
(M. and B.) Kamptner, distal view. 2, No. 13137 from Challenger Sta. 338, S. Atlantic Ocean;
distal view, x7200. 3, H. sp.. No. 22839, Lower Oligocene, Rodstenseje nr. Odder, Denmark;
proximal view, X 3200. 4, H. seminulum Bramlette and Sullivan; No. 14293, Middle Eocene, core
DWBG 23B, Pacific Ocean; proximal view, X6700.

Figs. 5, 7. Braarudosphaera spp. 5, B. bigelowi (Gran and Braarud) Deflandre; No. 15730, Yazoo
Formation (Upper Eocene), Clarke County, Mississippi; complete coccolith of five plates, X4000.
7, B. africana Stradner; No. 16057, Sutterby Marl (Aptian), borehole Alford, Lines., 101-7 ft.;
complete coccolith, X 4000.

Fig. 6. Pemma papillatum Martini; No. 15871, Yazoo Formation (Upper Eocene), Clarke County,
Mississippi; single plate, x 5000.

EXPLANATION OF PLATE 148

Fig. 1. Calciosolenia sp., No. 13131, recent Globigerina Ooze, Challenger Sta. 338, S. Atlantic Ocean;
proximal view, X 10 000.

Fig. 2. Scapholithus sp., No. 20678, Chalk Marl, (Cenomanian), Folkestone, Kent; proximal view,
x 13 000.

Figs. 3, 4. Syracosphaera spp. 3, S. hystrica Kamptner; No. 2573, recent Globigerina Ooze, Challenger
Sta. 338, S. Atlantic Ocean; distal view, x 16 000. 4, S. pulchra Lohmann; No. 18062, recent
Globigerina Ooze, Discovery II Sta. 4288, Biscay Mts. ; proximal view, X 6000.

Fig. 5. Loxolithus armilla (Black) Noel, holotype, No. 2807, Burwell Rock (Cenomanian), Great Shel-
ford nr. Cambridge; distal view, X 8000.

Figs. 6, 9. Rhabdolithina spp. 6, R. sp.. No. 18616, Sutterby Marl (Aptian), borehole, Alford, Lines.,
101-7 ft.; oblique proximal view, x6000. 9, R. sp.. No. 13501, Lower Gault (Albian), Folkestone,
Kent; oblique proximal view, x 8000.

Fig. 7. Staurolithites sp., No. 21747, Cambridge Greensand (Cenomanian), Cherry Hinton Fields nr.
Cambridge; distal view, x 8000.

Fig. 8. Zygolithus diplogrammus Deflandre; No. 22337, Mooreville Chalk (Santonian), nr. Eutaw,
Alabama; distal view, X5300.

Palaeontology, Vol. 11

PLATE 147

BLACK, Living, Tertiary and Cretaceous coccoliths

Palaeontology, Vol. 11

PLATE 148

BLACK, Living and Cretaceous coccoliths

M. BLACK: TAXONOMIC PROBLEMS IN THE STUDY OF COCCOLITHS 805

EXTINCT FAMILIES

In examining the fossil representatives of living families, we have not found it
necessary to employ a dual classification with the extinct species segregated into para-
taxa. We can approach the extinct families in much the same way, and follow the
well-tried palaeobotanical practice of treating them as being made up of closely related
organ-genera. Much depends upon the choice of characters that are relied upon for the
definition of genera or families, and there must always be an element of personal opinion
in making this selection. Quite clearly, we cannot follow any hard and fast rules, for some
kinds of criteria that are of great value in certain families are worthless in others. For
example, in the Zygolithaceae and Podorhabdaceae, the number of crystals in the
marginal shields or rings is variable, even within a single species; in the Deflandriaceae,
however, the number is always 16 (PI. 151, fig. 2), and in the Braarudosphaeraceae it is
5 throughout the whole family (PI. 147, fig. 5). In a large number of families the structure
of the outer margin of the coccolith follows a characteristic pattern, and can be used
as a primary basis of classification. Thus, the Zygolithaceae all have an architecture
based upon the loxolith-ring (see below and PI. 148, fig. 5); the Coccolithophoraceae
have two marginal shields of radial elements (PI. 143, fig. 4; PI. 144, fig. 2; PI. 145, fig. 3),
and the Podorhabdaceae have yet another type of rim (PI. 150, figs. 1, 6). The presence
or absence of a central spine is also of unequal significance in ditfieient families. Such
structures are characteristically present in the Rhabdosphaeraceae, but never in the
Coccolithophoraceae. In the Podorhabdaceae, on the other hand, the coccoliths of
some genera, such as Polypodorhabdus (PI. 150, fig. 1) always have spines, but in others,
such as Cretarhabdus, some coccoliths in a single species may have spines and others not
(PI. 150, figs. 4, 5). This state of affairs is also known in some living genera, such as
Syracosphaera.

ZYGOLITHACEAE Noel
Plate 148, figs. 5-9; Plate 149, figs. 1-5

A large number of Mesozoic coccoliths have an outer wall that consists of narrow,
steeply inclined laths of calcite, giving a structure rather like the wall of a barrel that
has been given a sharp twist, so that the staves run obliquely instead of straight up and
down (PI. 148, figs. 6, 9). The resulting loxolith-ring resembles the wall of Pontosphaera
in its general plan of construction, but differs in normally having no floor, and in having
a much smaller number of staves, which are also coarser and more robust. The numerous
genera which make up the family are defined by the various structures that bridge or
partly close the opening within the ring.

The name Loxolithus was proposed by Denise Noel (1965) for species consisting of
an unmodified ring (PI. 148, fig. 5); many fossils preserved in this condition are in fact
damaged specimens of other genera, having lost the bridge or central cross that was
originally present. In Zygolitbus there is a bridge along the shorter axis of the ellipse
on the proximal side (PI. 148, fig. 8). Pontilithus has a grid on the proximal side and
a distal bridge (PI. 149, figs. 2, 3), and Stawolithites has a central cross (PI. 148, fig. 7).
In several genera the proximal side of the loxolith-ring is closed by a floor, sometimes
solid, but frequently pierced by pores (PI. 149, fig. 1). In Rhagodiscus the floor consists

806

PALAEONTOLOGY, VOLUME 11

of a mosaic of little granules, and the addition of a prominent spine gives a rhabdolith-
like appearance to Rhabdolithina (PI. 148, figs. 6, 9).

The earliest Zygolithaceae are found in the Lias, but there is no great variety of struc-
ture until the Lower Cretaceous, when specialized forms begin to appear, leading up to
the spectacular diversity of genera and species in the Albian and Cenomanian. High in
the Upper Cretaceous there are species with a new type of floor-structure, in which the
plates are arranged very much as in early species of Pontosphaera and Helicosphaera
(compare PI. 149, fig. 4 with PI. 146, fig. 6 and PI. 147, fig. 3). These have been mentioned
already in connection with the ancestry of these two genera.

PODORHABDACEAE Noel
Plate 150, figs. 1-6

This is an exclusively Mesozoic family which made its first appearance in the Oxfordian,
when five well-differentiated genera appeared simultaneously without any known
ancestors. Most of these continue throughout the upper part of the Jurassic and the
Lower Cretaceous. An additional genus, Cretarhabdus, appears in the Berriasian and
continues with undiminished vigour to the top of the Maestrichtian, when this genus,
and with it the whole family quite suddenly became extinct (Bramlette and Martini
1964). The characters of the family may be seen in Plate 150, figs. 1-6; there is an ellip-
tical shield with a narrow rim consisting of two layers of keystone-shaped crystals sur-
rounding a very large central area, from which arises a tall spine. The several genera
which make up the family are distinguished from each other by the structures occupying
the central area (compare PI. 150, figs. 1, 2, 3, 5); in most of them the central spine is
supported by four buttresses constructed of fibrous crystals. This is an architectural
feature which is not confined to the Podorhabdaceae (PI. 149, fig. 6; PI. 151, fig. 3;
PI. 152, fig. 1), and its taxonomic significance will be discussed on a later page.

EXPLANATION OF PLATE 149

Fig. 1 . Zygolithus sp. with perforated floor; No. 22407, Cambridge Greensand (Cenomanian), Hauxton
Mill near Cambridge; proximal view, x 8000.

Figs. 2, 3. Pontilitlws flabellosus (Stradner), comb. nov. ( = Coccolitlws flabellosus Stradner), Cam-
bridge Greensand (Cenomanian). 2, No. 22881, Cherry Flinton Fields, proximal view, X 6000. 3,
No. 11510, Barrington, distal view, X 8000.

Figs. 4, 5. Pontolithinci spp. 4, P. moorevillensis Black, No. 22362, Mooreville Chalk (Santonian),
nr. Eutaw, Alabama; proximal view, X4800. 5, P. sp., No. 22899, Prairie Bluff Formation (Mae-
strichtian), Wilcox County, Alabama; distal view, X 4800.

Fig. 6. Eiffellitlms turriseiffeli (Deflandre) Reinhardt, No. 22114, Prairie Bluff Formation (Maestrich-
tian), Wilcox County, Alabama; distal view with characteristically oblique spine, X4000.

EXPLANATION OF PLATE 150

Fig. 1. Podorhabdus cylindratus Noel, No. 22798, Upper Oxford Clay (L. Oxfordian), Cambridge
Experimental Borehole, 220 ft. 3 in. -221 ft. 3 in.; distal view, X4800.

Fig. 2. Polvpodorhabdus madingteyeusis Black, No. 17413, Oxford Clay, locality as 1; distal view,
x 9600. ’

Fig. 3. Ethmorhabdus gallicus Noel, No. 22799, Oxford Clay, locality as 1; distal view, X7200.

Figs. 4-6. Cretarhabdus spp., Cambridge Greensand (Cenomanian) nr. Cambridge. 4, No. 21766 with
spine, Hauxton Mill; distal view, X4800. 5, No. 21763 without spine. Cherry Hinton Fields; distal
view, x 9600. 6, No. 22867, Cherry Hinton Fields, side view, X4800.

Palaeontology, Vol. 11

PLATE 149

BLACK, Cretaceous coccoliths

Palaeontology, Vol. 11

PLATE 150

BLACK, Jurassic and Cretaceous coccoliths

M. BLACK: TAXONOMIC PROBLEMS IN THE STUDY OF COCCOLITHS 807

DEFLANDRIACEAE Black 1967
Plate 151, figs. 1-5

More than a century ago Sorby described a remarkable coccolith from the Chalk;
it has a ring of sixteen granules from which arise four flying buttresses supporting a
peculiarly constructed stalk. Bramlette and Martini (1964) have shown that Sorby’s
coccolith belongs to one of several closely similar species that existed in the Upper
Cretaceous, and which they united into the genus Deflandrius. The earliest species is
found in the Gault (PI. 151, fig. 1); it has all the family characters fully developed, and
has no known ancestors. Three distinct, but very similar species are very abundant in
the Upper Cretaceous right up to the top of the Maestrichtian, when the whole family
became extinct.

The main differences between this family and the Podorhabdaceae are to be found in
details of ultra-structure. Otherwise the two families share the same basic architectural
plan : a spine supported by four buttresses arising from a granular ring. Quite recently,
this structural plan has been found to be developed in a remarkably similar way among
the uncalcified Haptophyceae, notably in several species of Chrysochromulina. This
discovery raises the possibility that the ancestors of both the Deflcindriaceae and Podor-
habdaceae may have belonged to such uncalcified stocks in which the family characters
were already developed, and only needed the ability to calcify in order to make their
appearance on the geological record.

goniolithaceae Deflandre
Plate 151, fig. 6

This is perhaps the most remarkable of the extinct families. It was founded by
Deflandre (1957) upon a single species, G.fluckigeri, which he discovered in the Eocene
of north-west Germany and the Oligocene of Basses Alpes. The coccolith is pentagonal
in outline, with a granular centre and radially constructed border. More recently,
Martini (1964) has discovered a complete cyst in the Clayton Formation (Danian) of
Alabama ; in this unique specimen, twelve of the pentagonal coccoliths are fitted together
to form a regular dodecahedron, giving a configuration exactly like the cysts of Braaru-
dosphaera , but with an entirely different coccolith-structure. Martini has shown that
G.fluckigeri ranges from the Lower Maestrichtian to the Rupelian, in the later occur-
rences apparently re-worked. The only observable difference between earliest and latest
autochthonous specimens is a slight increase in size. Although Goniolithus is always a
rare fossil, it has a wide geographical distribution : Denmark, France, Germany, English
Channel (submarine core-samples), Tunisia, Alabama, and Texas. At the time of writing,
the family still includes only the single species.

stephanolithiaceae fam. nov.

Plate 152, figs. 1-3

The genus Stephanolitliion was created by Deflandre (1939) for a peculiar species of
coccolith from the Oxford Clay of Calvados, which he named S. bigoti. It differs from
other coccoliths in having a set of radial spines extending beyond the marginal ring;

808

PALAEONTOLOGY, VOLUME 11

moreover, each spine consists of a single scalenohedron, a crystal habit that is most
unusual among the coccolithophorids. Two more species have subsequently been dis-
covered : S', speciosum, also from the Oxford Clay, and S. laffittei from the Portlandian
of Algeria. S. bigot i is known from the Callovian, but is only common in the Oxfordian;
in England it appears to be confined to the Upper Oxford Clay. C. laffittei is now known
to occur throughout the Cretaceous System, becoming extinct at the top of the Maestrich-
tian.

Stephanolithion stands so far apart from other coccoliths that it cannot be fitted into
any existing family, and I propose a new family name, Stephanolithiaceae, with the
characters of the single included genus.

discoasteraceae Tan Sin Hok

Plate 153, figs. 1-9

The discoasters are usually considered to be taxonomically distinct from the cocco-
liths, but it would be unrealistic to ignore the liklihood that they are the products of
similar planktonic algae, possibly in an encysted condition. They are mentioned here
because they illustrate particularly clearly the taxonomic importance of crystal-orienta-
tion, which these organisms were able to control with great precision. In most species
of Discoaster no crystal outlines are visible; nevertheless, the orientation of the crystals
can be determined optically, and they are invariably found to be arranged so that the
c-axis lies parallel with the axis of symmetry of the discoaster. In a smear-slide prepared
so that the fossils lie flat against the cover-slip, they consequently remain dark in all
positions between crossed nicols, and show the maximum refractive index for calcite.
This property is constant for all discoasters, and at once distinguishes them from typical
coccoliths.

The general appearance of the discoasters can be seen in Plate 153, figs. 1-8. They

EXPLANATION OF PLATE 151

Figs. 1-5. Deflandrius spp. 1, D. cantabrigensis Black, No. 22863, Cambridge Greensand (Cenomanian),
Cherry Hinton Fields nr. Cambridge; side view, X 6000. 2, 3, D. spinosus Bramlette and Martini.
2, No. 22187, Prairie Bluff Formation (Maestrichtian), Wilcox County, Alabama; oblique distal
view, x 4000. 3, No. 20144, Upper Chalk (Campanian), borehole at Holton nr. Halesworth, Suffolk;
distal view, X 8000. 4, 5, D. cretaceus (Arkhangelski) Bramlette and Martini. 4, No. 18481, Upper
Chalk (Santonian), nr. Salisbury, Wilts.; oblique proximal view of short-stalked form, x 6000.
5, No. 22321, Mooreville Chalk (Santonian). nr. Eutaw, Alabama; oblique proximal view, X 6000.

Fig. 6. Goniolithiis fluckigeri Deflandre, No. 22888, Prairie Bluff Formation (Maestrichtian), Wilcox
County, Alabama; damaged coccolith, X 10 000.

EXPLANATION OF PLATE 152

Figs. 1-3. Stephanolithion spp. 1, S. bigoti Deflandre, No. 22783, Upper Oxford Clay (L. Oxfordian),
Cambridge Experimental Borehole; distal view, X4000. 2, 3, S. laffittei Noel, Lower Gault (M.
Albian), Folkestone, Kent. 2, No. 13503, proximal view, X 10 000. 3, No. 13537, oblique distal
view, x 8000.

Fig. 4. Neococcolithes sp.. No. 22923, Prairie Bluff Formation (Maestrichtian), Wilcox County,
Alabama; proximal view, X 10 000.

Figs. 5-7. Kamptnerius magnificus Deflandre, Prairie Bluff Formation (Maestrichtian), Wilcox County,
Alabama, X 4000. 5, No. 22902, form with narrow wing, distal view. 6, No. 22928, form with
expanded wing, distal view. 7, No. 22925, form with narrow wing, proximal view.

Palaeontology, Vol. 11

PLATE 151

BLACK, Cretaceous coccoliths

Palaeontology, Vol. 11

PLATE 152

BLACK, Jurassic and Cretaceous coccoliths

M. BLACK: TAXONOMIC PROBLEMS IN THE STUDY OF COCCOLITHS 809

consist of calcite crystals grouped round a central axis, giving a snow-flake pattern. In
Marthasterites (PI. 153, fig. 9) the whole structure appears to consist of a single calcite
crystal, fn Discoaster each ray is crystallographically distinct, as may be seen from the
etch-patterns which are all too common on corroded specimens, or from the crystal
faces that are sometimes seen in exceptionally well-preserved specimens (PI. 153, fig. 1).

Marthasterites is found in the Upper Cretaceous and in the Lower Eocene; Dis-
coaster is confined to the Tertiary, appearing in the Paleocene and becoming extinct
at the top of the Pliocene or just within the Pleistocene, according to the horizon adopted
for the base of the Quaternary. Reported discoveries of living Discoasteraceae are at
present unconfirmed.

Constancy of crystal-arrangement is equally prevalent among the coccoliths; its
optical effects, however, are much more complicated when the crystal-axes are not all
parallel, but they are none the less of great taxonomic significance, and are relied upon
extensively in systematic work. In the Braarudosphaeraceae the crystals are arranged
round a central axis, much as in the Discoasteraceae, but the orientation is quite dif-
ferent, with the rhombohedron faces lying approximately in the plane of the coccolith
(PI. 147, fig. 5). In Stephanolithion the radial spines are scalenohedra with the optic axes
at right angles to the central axis of the coccolith (PI. 152, fig. 1). In the majority of
coccoliths, the arrangement is much more complicated, with the optic axes arranged in
a pattern that almost invariably involves some kind of screw symmetry.

RELATIONS WITH UNCALCIFIED STOCKS

The geological histories of the various coccolithophorid stocks discussed above
clearly do not all follow the same pattern. On the one hand there are long-ranging stocks
such as the Coccolithophoraceae, Pontosphaeraceae, and Zygolithaceae, with a great
number of species spread out through a considerable span of geological time. These pass
through phases when generic and specific distinctions are blurred, and evolutionary
divergence was evidently taking place. The Coccolithophoraceae, for example, passed
through such phases in Albian-Cenomanian times, again in the Eocene, and apparently
once more in the Pleistocene. The taxonomic problems involved in this state of affairs
are those familiar in any actively evolving group of organisms, and call for no special
comment here.

In contrast with such normally evolving stocks, there are others which behave quite
differently; these arrive on the scene abruptly, with all their characters ready-made.
Some of them, for example Goniolithus, Deflandrius , and Stephanolithion , persist perhaps
for the length of a geological period with little or no change, and then just as abruptly
become extinct; others, such as the Podorhabdaceae, appear equally abruptly, but
undergo noticeable evolutionary changes before they become extinct. In each of these
cryptogenic stocks, there is a well-defined architectural plan which we have taken as
characteristic of the family, and which is fully perfected in the earliest known species.
It looks very much as though the architectural plan had already been elaborated before
the fossil record started, and I think we have to consider seriously the possibility that the
beginning of the fossil record for such a family simply marks the change-over to calcifi-
cation in a stock whose scales were previously unmineralized.

Of course, we can hardly hope to find any tangible evidence of a palaeontological

810

PALAEONTOLOGY, VOLUME 11

nature to support this speculation, but fortunately there are other lines of investigation
open to us. The most promising of these is an inquiry into the existing unmineralized
stocks themselves, with a view to finding out what linkages may exist between these and
the coccolith-bearing algae.

Recent work on the uncalcified Haptophyceae has indeed brought to light some very
interesting parallels with the coccolithophorids. The various species of Chrysochromulina
have been found to bear delicate organic scales, which in their architectural construction
are very similar to some coccoliths. It is a little surprising to find that the most elaborately
constructed scales, although quite unlike the majority of living forms, can be readily
matched among extinct taxa.

In spite of their extreme tenuity, these scales apparently consist of two layers which
are differently sculptured, and if calcified would produce entirely different crystal
patterns. The simplest scales are elliptical discs, with sharply etched radiating lines on the
proximal surface, and a less clearly defined series of concentric rings o,r a vague spider’s
web pattern on the distal side (PI. 154, figs. 1, 2). Scales of this kind are also well-known
in Crystallolitlms hyalinus , where they provide the support upon which calcite crystals
are secreted (Parke and Adams 1960). In addition to these simple scales, most species
of Chrysochromulina also carry larger and more elaborate scales, in which the distal
layer is modified in various ways, often producing a central spine. The large scales of
C. chiton (Parke et al. 1958, figs. 16, 17, 28, 29) are shaped like pie-dishes, with sloping
walls on the distal surface, much like those of Pontosphaera (PI. 146, fig. 5). C. ericina
(Parke et al. 1956, figs. 17-19) has scales shaped like rhabdoliths with very long spines,
reminiscent of some Tertiary species of Rhabdosphaera. In several other species the
spines are differently constructed, and do not resemble those of Rhabdosphaera , but are

EXPLANATION OF PLATE 153

Figs. 1-8. Discoaster spp. 1, D. sp. showing crystal faces; each of the six rays is a separate rhombo-
hedron, with triad axis at right-angles to page; No. 22291, Upper Oligocene, core DWBG 10,
Pacific Ocean ; X6100. 2, D. elegans Bramlette and Sullivan, No. 11890, Middle Oligocene (reworked
from Eocene?), Grundfor, Denmark; X 3000. 3, D. barbadiensis Tan Sin Hok, No. 11949, Middle
Oligocene (reworked from Eocene?), Grundfor, Denmark; X4500. 4, D. lodoensis Bramlette and
Riedel, No. 15771, Eocene, Tuilerie de Donzacq, Landes, France; X 1900. 5, D. strictus Stradner,
No. 11899, Middle Oligocene (reworked from Eocene?), Grundfor, Denmark; x 3000. 6, D.
deflcmdrei Bramlette and Riedel, No. 15830, Eocene, Tuilerie de Donzacq, Landes, France; x 4000.
7, D. hamatus Martini and Bramlette, No. 16969, damaged specimen. Pliocene, core DWBP 119,
Pacific Ocean; : 4000. 8, D. brouweri Tan Sin Hok, No. 22197, Pliocene, core LSDH 78P, Pacific
Ocean ; x 3000.

Fig. 9. Marthasterites tribrachiatus (Bramlette and Riedel) Deflandre, No. 10485, Lower Eocene,
Rosnaes, Denmark; X2500.

EXPLANATION OF PLATE 154

Figs. 1-3. Crystcillolithus hyalinus Gaarder and Markali (= motile phase of Coccolithus pelagicus ),
strain No. LB 913/12 (Plymouth 182), Culture Collection of Algae and Protozoa, Botany School,
Cambridge; x 16 000. 1, No. 22909, two scales in proximal view. 2, No. 22911, two scales in
distal view. 3, No. 22915, scale and calcite rhombohedra detached during preparation.

Fig. 4. Scyphosphaera sp., No. 18028, ordinary coccolith, proximal view; recent Globigerina Ooze,
Discovery II Sta. 4269, Biscay Mts.; x 6000.

Figs. 5-7. Scyphosphaera sp., ordinary coccoliths; modern Globigerina Ooze, Challenger Sta. 338,
S. Atlantic Ocean. 5, No. 11316, distal view, x 6000. 6, No. 11317, detail, distal surface, x 20 000.
7, No. 11308, detail, proximal surface, x 80 000.

Palaeontology, Vol. 11

PLATE 153

BLACK, Tertiary discoasters

Palaeontology, Vol. 11

PLATE 154

I

3

BLACK, Living uncalcified Haptophyceae and coccoliths

M. BLACK: TAXONOMIC PROBLEMS IN THE STUDY OF COCCOLITHS 811

remarkably like those of certain Mesozoic genera. In these species the spine is supported
by four buttresses which arise from the distal surface of the scale. C. alifera (Parke et a/.
1956, fig. 76) has buttresses which do not reach the marginal rim, much as in the
Cretaceous Eijfellithus (PI. 149, fig. 6). In C. ephippium (Parke et a/. 1956, figs. 38, 39)
the buttresses are arranged exactly as in the Jurassic Polypodorhabdus (PI. 150, fig. 2);
they extend from the marginal rim to the centre, dividing the scale into four quadrants,
each of which is ornamented by radial striae on the proximal surface. In the large spined
scales of C. pringsheimii (Parke et al. 1962, figs. 21, 22, 29) the supports of the tall spine
are no longer in contact with the surface of the scale except at the outer margin, but
behave like flying buttresses; the whole arrangement is an astonishing duplication of the
Upper Cretaceous Deflandrius (PI. 151, fig. 3). C. pringsheimii is unusual in having so
many as four different kinds of scale, each of which could serve as a template for the
orderly disposition of calcite-crystals to give a familiar coccolith pattern.

Enough has been said to show that there is a very significant parallel between the
calcified and uncalcified Haptophyceae. Several of the living species of Chrysochromu-
lina could indeed be regarded as uncalcified survivors of familiar Mesozoic families.
It would be asking too much of coincidence to believe that the few living species of
Chrysoehromulina have contrived to reconstruct in their uncalcified scales just those
features that were required to start off the Podorhabdaceae in the Oxfordian, the
Deflandriaceae in the Albian, and the Pontosphaeraceae and Helicosphaeraceae some-
time near the beginning of the Tertiary. 1 think it is at least as reasonable to suppose that
this remarkable genus is indeed a survival from a protean stock that existed in early
Mesozoic times, and by selective calcification of its organic scales has given birth to
these ancient families.

REFERENCES

black, m. 1968. The systematics of coccoliths in relation to the geological record. In The micro-
palaeontology of oceans , edited by B. M. Funnell and W. R. Riedel, Cambridge University Press.

bramlette, m. n. and martini, e. 1964. The great change in calcareous nannoplankton fossils between
the Maestrichtian and the Danian. Micropaleontology , 10, 291-322.

and sullivan, f. r. 1961. Coccolithophorids and related nannoplankton of the early Tertiary in

California. Ibid. 7, 129-88.

and wilcoxon, j. a. 1967. Middle Tertiary calcareous nannoplankton of the Cipero section,

Trinidad, W.I. Tulane Stud. Geol. 5, 93-131.

deflandre, g. 1939. Les stephanolithes, representants d'un type nouveau de coccolithes du Jurassique
superieur. C. r. hebd. Seanc. Acad. Sci. Paris , 208, 1331-3.

1947. Braarudosphaera nov. gen., type d’une famille nouvelle de coccolithophorides actuels a

elements composites. Ibid. 225, 439-441.

1952. Coccolithophoridae. In Traite de Zoologie, ed. P. P. Grasse. Paris: Masson.

1959. Sur les nannofossiles calcaires et leur systematique. Rev. Micropaleont. 2, 127-52.

1966. Commentaire sur la systematique et la nomenclature des nannofossiles calcaires. Cahiers

Micropaleont., ser. 1, no. 3, C.R.N.S., Paris. 9 pp.

and fert, c. 1954. Observations sur les coccolithophorides actuels et fossiles en microscopie

ordinaire et electronique. Ann/s. Paleont. 40, 1 15-76.

gaarder, k. r. 1967. Observations on the genus Ophiaster Gran (Coccolithineae). Sarsia, 29, 183-92.

gran, h. h. and braarud, t. 1935. A quantitative study of the phytoplankton in the Bay of Fundy
and the Gulf of Maine. J. biol. Bd. Can. 1, 297-467.

halldal, p. and markali, j. 1955. Electron microscope studies on coccolithophorids from the

812 PALAEONTOLOGY, VOLUME 11

Norwegian Sea, the Gulf Stream and the Mediterranean. Avh. norske VidenskAkad. Oslo, 1955 (1),
30 pp.

hay, w. w., mohler, h. p., roth, p. h., schmidt, r. r., and Boudreaux, j. e. 1967. Calcareous nanno-
plankton zonationof the Cenozoic of the Gulf Coast and Caribbean-Antillean area and transoceanic
correlation. Trans. Gulf-Cst Ass. geol. Socs. 17, 428-80.

and wade, mary e. 1966. Calcareous nannofossils from Nal’chik (Northwest Caucasus).

Eel. Geol. Helv. 59, 379-400.

kamptner, e. 1958. Betrachtungen zur Systematik der Kalkflagellaten, nebst Versuch einer neuen
Gruppierung der Chrysomonadales. Arch. Protistenk. 103, 54-116.
lohmann, h. 1902. Die Coccolithophoridae, eine Monographic der Coccolithen bildenden Flagellaten.
Ibid. 1, 89-165.

mcintyre, a. and be, a. w. h. 1967. Modern Coccolithophoridae of the Atlantic Ocean — I. Placoliths
and Cyrtoliths. Deep Sea Res. 14, 561-97.

martini, e. 1964. Ein vollstandiges Gehause von Goniolitluis fluckigeri Deflandre. Neues Jb. Geol.
Paldont. Abh. 119, 19-21.

Murray, g. and blackman, v. h. 1898. On the nature of the coccospheres and rhabdospheres. Phil.
Trans. R. Soc., B, 190, 427-41.

noel, denise. 1965. Sur les coccolithes Jurassiques Europeens et d'Afrique du Nord. C.N.R.S., Paris.
209 pp.

paasche, e. 1964. A tracer study of the inorganic carbon uptake during coccolith formation and
photosynthesis in the coccolithophorid Coccolithus huxleyi. Physio/ogia PL, suppl. 3, 82 pp.
parke, mary. 1961. Some remarks concerning the class Chrysophyceae. Br. phycol. Ball. 2, 47-55.

and adams, Irene. 1960. The motile ( Crystallolithus hyalinus Gaarder and Markali) and non-

motile phases in the life history of Coccolithus pelagicus (Wallich) Schiller. J. mar. biol. Ass. U.K.
39, 263-74.

— and dixon, p. s. 1964. A revised check-list of British marine algae. Ibid. 44, 499-542.

manton, Irene, and clarke, b. 1956-62. Studies on marine flagellates. Ibid. 35, 387-414; 37,

209-28; 42, 391-404.

schiller, j. 1930. Coccolithineae. In Rabenhorst’s Kryptogamen-Flora, 10 (2), 89-266.
stradner, h„ adamiker, d., and maresch, o. 1968. Electron microscope studies on Albian calcareous
nannoplankton. Verb. K. ned. Akad. Wet. 24 (4). 107 pp.
watabe, n. and wilbur, k. m. 1966. Effects of temperature on growth, calcification and coccolith form
in Coccolithus huxleyi (Coccolithineae). Limnol. Oceanogr. 11, 567-75.
wilbur, k. m. and watabe, n. 1963. Experimental studies on calcification in molluscs and the alga
Coccolithus huxleyi. Ann. N.Y. Acad. Sci. 109 (1), 82-112.

APPENDIX

Localities of deep-sea samples mentioned in the explanation of plates

H.M.S. Challenger

Station 338: dredge, 21° 15' S., 14° 02' W., about 900 miles S. of Ascension Is., South Atlantic Ocean.

R.R.S. Discovery II

Station 3809: dredge, northern slope of Galicia Bank, 42° 56' N., 11° 47' W., about 120 miles W. of
Cape Finisterre.

Station 4269: trigger-weight core, Biscay Mts., W. of Bay of Biscay.

Station 4288 : dredge, Biscay Abyssal Plain, Bay of Biscay.

Scripps Institution of Oceanography

DWBG 10: core sample, specimens from 11 to 27 cm. 6° 54' N., 131° 00' W., central Pacific Ocean.
DWBG 23B: core sample, specimens from 10 to 12 cm. 16° 42' S., 145° 48' W., between Tuamotu
Archipelago and Tahiti.

M. BLACK: TAXONOMIC PROBLEMS IN THE STUDY OF COCCOLITHS 813

DWBP 119: core sample, specimens from 398 to 400 cm. 27° 54' S., 106° 53' W., about 200 miles
ESE. of Easter Is.

LSDH 78PG core sample, specimens from 250 to 270 cm. 4° 30' S., 168° 03' E., between Gilbert Is.
and Solomon Is.

MAURICE BLACK

Department of Geology
Sedgwick Museum

Typescript received 21 May 1968 Cambridge

THE PALAEONTOLOGICAL ASSOCIATION

Extracts from the Annual Report of the Council for 1967-8

Membership. On 31 December 1967 there were 1,353 members (725 Ordinary, 131 Student, and 497
Institutional), a net increase of 15 members during the year.

Finance. During 1967 expenditure rose by £2,724 due largely to the cost of reprinting Volume 3, Part 2
and producing Special Papers. This increase was more than offset by an increase in sales from £1,381
to £4,348. There was also an increased income from subscriptions, interest, and donations to specific
papers, but general donations at £525 reached their lowest level.

Special Paper No. 1 was published at a cost of £3,227 ; donations of £1,500 from the N.C.B. and £250
from Shell Oil Company together with sales of £912 left £565 as a charge on the Publications Reserve
at the end of the year. Standing orders for Special Papers now number 135.

Volume 3, Parts 1, 3, and 4 will have to be reprinted as soon as possible and £1,500 has been set aside
for this purpose from the Publications Reserve. This now stands at £5,720, enough to cover 62 per cent
of a single volume compared with 63 per cent at the end of 1966.

Subscriptions for Ordinary and Student Members were raised to meet the increasing cost of printing
Palaeontology which will rise to over £9,000 in 1968.

Publications. The four parts of Volume 10 of Palaeontology were published during 1967; they con-
tained 37 papers plus ‘Notes for Authors’.

Special Papers in Palaeontology No. 1 for 1967 was published; it contained 324 pages and 27 plates.
Meetings. Four meetings took place during 1967-8. The Association is grateful to the Council of the
Geological Society of Fondon, Professor J. Sutton (Imperial College, Fondon), and Professor Scott
Simpson (University of Exeter) for generously granting facilities for meetings, and to Mr. G. W. Green
and Mr. M. Mitchell for leading the field meeting.

a. The Tenth Annua! Genera! Meeting was held in the Rooms of the Geological Society of London,
Burlington House, London, W. 1, on Wednesday, 1 March 1967, at 5.00 p.m. The Annual Report
of the Council for 1966-7 was adopted and the Council for 1967-8 was elected. Professor Alwyn
Williams of the Queen’s University, Belfast, delivered the Tenth Annual Address on ‘Evolution of
the shell structure of articulate brachiopods’.

b. A Field Demonstration Meeting was held at Burrington Combe in the Mendips on Saturday,
6 May 1967. Mr. G. W. Green and Mr. M. Mitchell demonstrated the Carboniferous Limestone
succession.

c. The Tenth Anniversary Meeting and Dinner was held at Imperial College, London, on Wednesday,
8 November 1967. The discussion meeting was on ‘The explosive development of palaeontology’
and the introductory speakers were Professor P. C. Sylvester-Bradley, Dr. D. N. Wood, and
Dr. W. S. McKerrow.

d. A Symposium on ‘Devonian floras and faunas’ was held in the Department of Geology, Univer-
sity of Exeter, on Thursday to Saturday, 14-16 December 1967. About 45 persons attended to
hear 16 papers and see 13 demonstrations. Professor Scott Simpson was local secretary.

Special General Meeting. Immediately before the Discussion Meeting at Imperial College, London,
on Wednesday, 8 November 1 967, a Special General Meeting was held at 3.00 p.m. to consider a change
in Rule 2 of the Constitution. The following motion was proposed from the Chair and passed (nem.
con.):

That, effective from 1 January 1968, Rule 2 of the Constitution be amended to read as follows:
'There shall be Ordinary members. Institutional members, and Student members, and their annual
subscriptions shall be Five Pounds, Seven Pounds, and Three Pounds respectively. Each subscriber
shall be considered a member of the Association, but Institutional members shall not be eligible
to take part in the government of the Association.’

THE PALAEONTOLOGICAL ASSOCIATION

815

Council. The following were elected members of the Council of the Association for 1967-8 at the
Annual General Meeting on 1 March 1967: President : Professor T. S. Westoll, F.R.S.; Vice-Presidents :
Dr. W. S. McKerrow, Professor F. H. T. Rhodes, Professor C. H. Holland; Treasurer : Dr. C. Downie;
Secretary: Dr. J. M. Hancock; Editors: Mr. N. F. Hughes, Dr. Gwyn Thomas, Dr. I. Strachan,
Professor M. R. House, Dr. R. Goldring; Other members: Mr. M. A. Calver, Dr. C. B. Cox, Mr. D.
Curry, Dr. Grace Dunlop, Mr. G. F. Elliott, Dr. T. D. Ford, Dr. A. Hallam, Dr. R. P. S. Jefferies,
Dr. G. A. L. Johnson, Dr. W. D. I. Rolfe, Dr. A. H. Smout, Dr. L. B. H. Tarlo, Professor H. B.
Whittington.

BALANCE SHEET AND ACCOUNTS FOR THE YEAR ENDING

31 DECEMBER 1967

Balance Sheet

Liabilities

Publications Reserve Account

Balance as per Account ......

Special Papers Fund Account .......

Provision for Reprinting ........

Amounts received in advance of Subscriptions for 1968
Provision for cost of publication of Palaeontology , Vol. 10, as per
Income and Expenditure Account ......

Less Expenditure incurred to 31 December 1967

Sundry Creditors .........

£ s. d. £ s. d.

5,720 1 8

1,500 0 0
451 10 8

7,960 8 10
4,860 8 10

3,100 0 0

25 5 1

Note. No amount has been included in these accounts for subscriptions unpaid at

31 December 1967

£10,796 17 5

Assets

Office Equipment at cost .....
Less Depreciation to date ....

Investments at cost :

Equities Investment Fund for Charities — 958 units
5 per cent Defence Bonds ....
Liverpool 6| per cent Bonds
Wagon Investment Loan 6f per cent
5i per cent National Development Bonds

Sundry Debtors and Payments in Advance:

Authors for offprints .....
Advance Payments:

For Palaeontology, Vol. 11.

For Palaeontology, offprints ....

Cash at Bank:

Deposit Account .....

Current Account .....

£

s.

d.

27

6

6

15

6

6

999

18

3

2,000

0

0

1,500

0

0

2,000

0

0

2,000

0

0

83

6

8

84

16

6

106

17

6

1,462

15

7

£ s. d.
12 0 0

8,499 18 3
547 211

168 3 2

1,569 13 1

£10,796 17 5

Report of the Auditors to the Members of the Palaeontological Association. We have examined the above Balance
Sheet and annexed Income and Expenditure Account which in our opinion give respectively a true and fair view
of the state of the Association’s affairs as at 31 December 1967 and of its income and expenditure for the year
ended on that date.

JOSHUA WORTLEY & CO.
Chartered Accountants

816

PALAEONTOLOGY, VOLUME 11

Income and Expenditure Account for the Year ended
31 December 1967

Expenditure

£

s.

d.

£

s.

d.

To Provision for the cost of Publication of Palaeontology, Vol.

10

Part 1 ......

1,925

10

3

Part 2 ......

2,034

18

7

Part 3

2,000

0

0

Part 4 ......

2,000

0

0

7,960

8

10

Add extra costs Vol. 9, Part 4

146

3

4

8,106

12

2

Less over-provision for Vol. 9, Part 3

118

16

9

7,987

15

5

Reprinting Vol. 3, Part 2

554

8

6

Administrative Expenses:

Insurance .....

2

0

3

Postages and Stationery

191

13

6

Audit Fee .....

14

14

0

Honorarium to Assistant Treasurer .

120

0

0

Duplicating and dispatch of circulars .

181

10

2

Meeting expenses ....

32

9

3

542

7

2

Depreciation of Equipment .

6

0

0

Special Papers:

Printing costs .....

3,319

7

9

Less Sales .....

. 920

9

11

Donations:

Shell

. 250

0

0

Foreman .....

. 475

8

9

1,645

18

8

1,673

9

1

Excess of Income over Expenditure transferred to Publications

Reserve Account ........ 1,578 13 9

£12,342 13 11

THE PALAEONTOLOGICAL ASSOCIATION

817

Income

By Subscriptions for 1967 ....

Subscriptions for 1966 .....

Sales of Publications .....
Interest received :

5 per cent Defence Bonds ....
Bank Deposit Account ....
Equities Investment Fund for Charities
Liverpool Corporation ....
Wagon Investments .....

Special Donations .....
General Donations:

British Petroleum Co. ....

Texaco .......

Burmah Oil Co., Ltd. ....

Miscellaneous Receipts ....
Special Papers Fund .....
Excess of Expenditure over Income transferred
Reserve Account .....

£

s .

d.

£

s.

d.

5,923

11

8

202

0

3

6,125

11

11

3,428

15

2

260

0

0

30

18

2

76

4

10

101

9

0

124

18

4

593

10

4

217

14

7

250

0

0

175

0

0

100

0

0

525

0

0

66 11 11
1,385 10 0

to Publications

£12,342 13 11

Publications Reserve Account

To Balance as per Balance Sheet ....

5,720

1

8

Excess of Expenditure over Income

.

-

-

Provision for Reprinting .....

.

-

-

£5,720

1

8

By Balance at 31 December 1966 ....

5,141

7

11

Add Excess of Income over Expenditure

1,578 13 9

Less Provision for Reprinting ....

1,000 0 0 578

13

9

£5,720 1 8

INDEX

Pages 1-162 are contained in Part 1; pages 163-327 in part 2; pages 329-490 in Part 3; pages 491-642 in
Part 4; pages 643-824 in Part 5. Figures in Bold Type indicate plate numbers.

A

Acltlysopsis, 218; hemitora ; 219, 42; liokata, 219, 42.

Africa: non-marine ostracods from Ghana, 295.

Albertella, 212; eiloitys, 213, 39; judithi, 212, 39;
lata, 214, 39.

Albertelloides, 214; mischi, 215, 38; pandispinata, 216,
39; sp., 217, 39.

Alectryonia, 482.

Algae: Cretaceous coccoliths, 361; taxonomy of coc-
coliths, 793; Tethyan Dasycladaceae, 491.

Allen, N. W. See Kilenyi, T. I.

Alokistocare alta, 226, 41.

Alveolites, 49; multiperforatus, 50, 13; obtortus, 51,
17 ; suborbicularis, 49, 12, 14, 15 ; s. forma gemmans,
50, 16.

Amecephalus laticaudum, 227, 40.

Ammobaculites sp., 156, 30.

Ammonoidea: Famennian ammonoids, 535; new
Kimmeridgian ammonite, 16; Visean/Namurian
goniatites, 264.

Andrews, H. N. See Phillips, T. L.

Angiosperm: probable pollen in British Cretaceous,
421.

Antagmus arenosus, 220, 36.

Aquilapollenites, 549; attenuatus, 550, 106; pachypolus,
551, 106; spinulosus, 550, 106.

Archangelsky, S. Studies on Triassic fossil plants from
Argentina. IV. The leaf genus Dicroidium and its
possible relation to Rhexoxylon stems, 500.

Archengraulis, 255 ; productus, 256.

Arctostrea, 458; alaefonnis, 87; colubrina, 85, 86, 88;
c. ricordeana, 459, 85-87; diluviana, 463, 85; pusilla,
87, 90; ungulata, 463, 85-87.

Argentina: leaf genus Dicroidium, 500; trunk of
Rhexoxylon, 236.

Armstrong, J. The unusual brachial skeleton of Atten-
uatella convexa sp. nov. (Brachiopoda), 783.

Arthropoda. See Ostracoda, Trilobita.

Assam : Eocene Nummulites from, 669.

Astartila cytherea, 20.

Asterocyclina asterisca, 287.

Astrocystites, 513; distans, 515, 99, 100.

Athabaskia sp., 203, 43.

Attenuatella, 786; convexa, 788, 142; cf. incurvata,
791, 142; sp. A, 790, 142.

Aulopora, 60; ? tubaeformis, 61, 17.

Auroraspora macro, 124, 27.

Australia: Carboniferous brachiopods, 389; Devonian
ammonoids, 535; Devonian brachiopods, 627, 731 ;
Devonian foraminifera, 601 ; Ordovician edrio-
blastoid, 513; Permian bivalve ligaments, 94;
Permian brachiopods, 783; Silurian trilobite, 691.

Australirhynchia, 731 ; cudalensis, 732, 141.

Austrospirifer variabilis, 64.

Ayalaina rntteni, 103.

B

Bactrynium, 329; bicarinatum, 66-68.

Baculatisporites fusticulus, 117, 25.

Bajanorthis tukolandica, 91.

Banner, F. T. See Eames, F. E.

Barringtonella bulmani, 149.

Bates, D. E. On ‘ Dendrocrinus ’ cambriensis Flicks, the
earliest known crinoid, 406.

Bathyuriscus pe talus, 203, 43.

Batten, D. J. Probable dispersed spores of Cretaceous
Equisetites, 633.

Batten, R. L. See Rollins, H. B.

Biscalitheca musata, 105, 21-24.

Bivalvia: epizoic oysters, 19; functional study of
oyster, 458; ligaments in Permian, 94; micro-
structure of shell, 163.

Black, M. Taxonomic problems in the study of cocco-
liths, 793.

Blake, D. B. Pedicellariae of two Silurian echinoids
from western England, 576.

Blow, W. FI. See Eames, F. E.

Bobinella kulumbensis, 91, 92.

Bonnia copia, 194, 36.

Boucot, A. J. See Walmsley, V. G.

Boulter, M. C. A species of compressed lycopod
sporophyll from the Upper Coal Measures of
Somerset, 445.

Braarudosphaera, 802; bigelowi, 802, 147; africana,
147.

Brachial skeleton: unusual in Attenuatella, 783.

Brachiopoda: Carboniferous schizophoriids, 64, 389;
Dayia navicula, 612; delthyrial cover in Mucrospi-
rifer, 317; feeding mechanisms, 329; mantle canal
patterns, 389; new resserellid genus, 306; new
rhynchonelloid genus, 731; septate dalmanellid,
627 ; shell structure, 486 ; unusual brachial skeleton,
783.

Brett, D. W. Studies on Triassic fossil plants from
Argentina. III. The trunk of Rhexoxylon, 236.

Bulman, O. M. B., and Rickards, R. B. Some new
diplograptids from the Llandovery of Britain and
Scandinavia, 1.

Bythicheilus? sp. indet., 229, 42.

C

Caborcella, 221; clinolimbata, 221, 41; granosa, 221,
39; sp., 222, 43.

Calciosolenia sp., 803, 148.

820

INDEX

Callocystites jewetti, 136.

Cambrian: brachiopod shell structure, 486; revision
of two trilobites, 410; trilobites from Nevada, 183.

Cantheliophorus spp., 454.

Carboniferous: fern, 104; goniatites, 264; lycopod
Eskdalia, 439; lycopod sporophyll, 445; mantle
canal patterns, 389; schizophoriid brachiopods,
64; Tournaisian spore flora, 116.

Carter, R. M. Functional studies on the Cretaceous
oyster, Arctostrea, 458.

Caryocrinites; ornatus, 138, 139; roemeri, 139; sep-
tentrionalis, 139.

Cassigerinella, 368; chipolensis, 369; eocaenica, 368.

Central America: Tertiary larger foraminifera from,
283.

Chaetetes, 46; lonsdalei, 47, 14; multitabulatus, 47, 9;
cf. rotundus, 48, 9.

Chanda vemista, 230, 40.

Cheirocrinus, 593; anatiformis, 135; giganteus, 134;
granulatus, 134, 135; jamesii, 140; languedodanus,
134; spp., 134, 135.

Chlamys asperrimus, 87.

Chubbina, 527; cardenasensis, 529; jamaicensis, 527,
101, 102; macgillavryi, 529, 102, 103.

Clarke, W. J. See Eames, F. E.

Classopollis sp., 106.

Clavatipollenites, 422; hughesii, 426, 80; rotundus,
424, 79, 80.

Cleidogonia antiqua, 256.

Climacograptus scalaris, 10.

Coccoliths: Cretaceous from Zululand, 361; taxo-
nomic problems in, 793.

Coccolithus, 362; cf. cavus, 143; cribosphaereUa, 362,
70; huxleyi, 799, 145; cf. marismontium, 143;
pelagicus, 796, 143; zuluensium, 363, 69.

Cocks, L. R. M. See Ziegler, A. M.

Coelenterata : Devonian tabulate corals, 44; Silurian
nredusoid (?), 610.

Coenites, 51 ; escharoides, 52, 10, 13; gradatus, 53, 17;
sp., 54, 14.

Colour markings: in Devonian trilobites, 498.

Conkin, B. M. See Conkin, J. E.

Conkin, J. E., and Conkin, B. M. A revision of some
Upper Devonian foraminifera from Western
Australia, 601.

Convolutispora, 122; cf. mellita, 122, 26; cf. tuberosa,
123, 26.

Cope, J. C. W. Epizoic oysters on Kimmeridgian
ammonites, 19.

— Propectinatites, a new Lower Kimmeridgian am-
monite genus, 16.

Corals : Devonian tabulates, 44.

Cordey, W. G. A new Eocene Cassigerinella from
Florida, 368.

— Morphology and phylogeny of Orbulinoides beck-
mannii (Saito 1962), 371.

Corynexochides prolatus, 204, 42.

Cowen, R. A new type of delthyrial cover in the
Devonian brachiopod Mucrospirifer, 317.

Crassifimbria walcotti ?, 223, 37.

Crassostrea, 479; ameghinoi rocana, 473; echinata, 88.

Cravenoceras leion, 270, 47.

Cretaceous: alveolinid, 526; angiosperm pollen, 421;
coccolithophorids, 361; Equisetites spores, 633;
functional study of oyster, 458; non-marine ostra-
cods, 259; Wealden marine-brackish bands, 141.

Cretarhabdus spp., 150.

Crinoids: earliest known, 406; function of stem, 275.

Crystallolithus hyalinus, 154.

Cyclagelosphaera margereli, 144.

Cyclococcolithus leptoporus, 797, 144.

Cyclocypris? sp. A, 261, 45.

Cyclolithus zulua, 363, 70.

Cymaclymenia pseudogoniatites, 538.

Cypridea, 158, 260; marina, 29; pumila, 29; sp. A,
260, 45; sp. B, 260, 45; sp. C, 261, 45; tuberculata,
29.

‘ Cypris ’ henfieldensis, 144, 29.

Cystoids: dichoporite pore-structure, 697; Macro-
cystella, 580.

D

Darwinula, 144; leguminella, 144, 29; oblonga, 145, 29.

Davidsonella sinuata, 68.

Dayia, 613; navicula, 614, 118-21; /;. cf. bohemica,
120; sp., 119.

Deflandrius, 807; cantabrigensis, 151; cretaceus, 151;
spinosus, 151.

Denckmannites, 691 ; rutherfordi, 692, 133.

‘ Dendrocrinus' cambriensis, 406.

Devon: tabulate corals from, 44.

Devonian, colour markings in trilobites, 498; del-
thyrial cover, 317; Famennian ammonoids, 535;
foraminifera from Australia, 601; new Lower
O.R.S. plant, 683; new rhynchonelloid, 731 ; septate
dalmanellid, 627 ; sinus-bearing monoplacophoran,
132; tabulate corals, 44.

Dicroidium, 501; coriaceum, 506, 97, 98; elongation,
503, 97, 98; sp., 508, 97; zuberi, 502, 98.

Discoaster, 808; barbadiensis, 153; brouweri, 153;
deflaiulrei, 153; elegans, 153; hamatus, 153; lodoen-
sis, 153; strictus, 153; sp., 153.

Discohelix, 554; ( D .) albinatiensis, 566, 107, 108; cf.
calculiformis, 571, 110; centricosta, 569, 109; conica,
569, 108; costata, 570, 109; cotswoldiae, 572, 110;
cf. crenulata, 568, 108; dictyota, 565, 107; cf. gum-
beli, 573, 110; levis, 570, 109; ( Pentagonodiscus )
angusta, 573, 110; reussii, 574, 110.

Discolithus, 364; cristallinus, 364, 69, 70; rliabdo-
sphaericus, 364, 69, 71; spiralis, 365, 70, 71.

Drozdzewski, G. See Seilacher, A.

Duodecimedusina palmer i, 610.

E

Eames, F. E. Sindulites, a new genus of the Nummuli-
tidae (Foraminifera), 435.

— Clarke, W. J., Banner, F. T., Smout, A. H., and
Blow, W. FI. Some larger foraminifera from the
Tertiary of Central America, 283.

Echinocystites pomum, 576.

Echinodermata. See Crinoids, Cystoids, Echinoids,
Edrioblastoids.

INDEX

821

Echinoencrinites ; angulosus, 140; reticulatus, 135;
senckenbergii, 135.

Echinoids: Silurian pedicellariae of, 576.

Edrioblastoids: Ordovician from Australia, 513.

Edwards, D. A new plant from the Lower Old Red
Sandstone of South Wales, 683.

Eiffellithus turriseiffeli, 149.

Elliott, G. F. Three new Tethyan Dasycladaceae
(Calcareous Algae), 491.

Elliottina deslongchampsi, 68.

Ellipsagelosphaera sp., 143.

Emiliana huxleyi, 799, 145.

Encrinus liliiformis, 48.

England: Bathonian fish otoliths, 246; Cretaceous
angiosperm pollen, 421; ichthyosaur gastric con-
tents, 376; lycopod sporophyll from Somerset, 445;
new Llandovery diplograptids, 1 ; Silurian echinoid
pedicellariae, 576; Silurian medusoid (?), 610;
tabulate corals from Devon, 44.

Epimastopora malaysiana, 491, 93.

Equisetites, 633.

Ericsonia sp ., 144.

Eskdalia, 439; minuta, 442, 82.

Esker, G. C., III. Cloour markings in Phacops and
Greenops from the Devonian of New York, 498.

Etheria elliptica, 88.

Ethmorhabdus gallicus, 150.

Eudesella mayalis, 68.

Eulepidina, 285; favosa, 296, 57; undosa, 285, 49; u.
laramblaensis , 296, 57.

Eumorphoceras, 268; ( Edmooroceras ) dichalocium,
269, 46; pseudocoronula, 270, 46; tornquisti, 268, 46;
( Eumorphoceras ) medusa, 272, 47 ; m. sinuosum, 47.

Europe: Carboniferous schizophoriid brachiopods
from, 64.

F

Fabanella bononiensis, 146, 29.

Feeding mechanisms: in Triassic brachiopods, 329.

Figge, K. A goniatite fauna from the Visean/Namur-
ian boundary, 264.

Fish: Bathonian otoliths, 246; Silurian ostracoderm,

21.

Florida: Eocene Cassigerinella, 368.

Foraminifera: Cretaceous alveolinid, 526; Eocene
Cassigerinella , 368; Indian nummulitid, 669; larger
Tertiary, 283; new nummulitid genus, 435; Orbulin-
oides, 371 ; Upper Devonian, 601.

Fritz, W. H. Lower and early Middle Cambrian trilo-
bites from the Pioche Shale, east-central Nevada,
U.S.A., 183.

G

Gastropoda: Discohelix, Tethyan index fossil, 554;
sinus-bearing monoplacophoran, 132.

Genuclymenia, 536; angelini, 538; f rechi, 538; karpin-
skii, 538; keepitensis, 536, 104.

Gephyrocapsa oceanica, 145.

Germany: pseudoplanktonic crinoid, 275; Visean /
Namurian goniatites, 264.

Ghana: non-marine Cretaceous ostracods, 259.

Glansicystis baccata, 137.

Glauconome rugosa, 98.

Globigeraspis kugleri, 372.

Globigerina ampliapertura, 369.

Globigerinatheka barri, 'ill.

Glossopleura, 205; sp. 1, 205, 43; sp. 2, 205, 43.

Glyptocystites multiporus, 135.

Glyptograptus ( Pseudoglyptograptus ) vas, 13.

Goniatite: Visean/Namurian fauna, 264.

Goniatites schaelkensis, 267 , 47.

Goniolithus, 807 ; fluckigeri, 151.

Grandispora echinata, 126, 27.

Graptolites: assemblages from Birkhill Shales, 654;
new Llandovery diplograptids, 1.

Greenops boothi, 96.

Gyrosteus subdeltoideus, 249.

H

Halkyardia bikiniensis, 292, 51.

Harlanjohnsonella annulata, 494, 93, 94.

Harper, C. W. See Walmsley, V. G.

Haude, R. See Seilacher, A.

Helicolepidina paucispira, 287, 292, 51.

Helicosphaera carteri, 801, 147; seminulum, 147; sp.
147.

Helicostegina soldadensis, 287.

Hemicosmites', extraneus, 138; malum, 139; cf. pyri-
formis, 139; cf. verrucosus, 138; spp., 138-40.

Heterostegina, 290; ( Heterostegina ), 290; ( Vlerkina ),
290; ( Vlerkinella ) kugleri, 291, 51.

Holwill, F. J. W. Tabulate corals from the Ilfracombe
Beds (Middle-Upper Devonian) of North Devon,
44.

Homocystites constrictus, 134.

Hudson, J. D. The microstructure and mineralogy of
the shell of a Jurassic mytilid (Bivalvia), 163.

Hutsonia capelensis, 153, 30.

Hymenozonotriletes, 125; hast ulus, 126, 26.

I

Ichthyosaur, 376, 72, 73.

Illinois: Pennsylvanian fern from, 104.

India: Eocene Nummulites from Assam, 669; structure
of Vertebraria, 643.

J

Jaekelocystis hartleyi, 137.

Jamaica: new Cretaceous alveolinid genus from, 526.

Jamoytius kerwoodi, 21, 3-6.

Jenkins, T. B. H. Famennian ammonoids from New
South Wales, 535.

Jurassic: dinosaur, 40; epizoic oysters, 19; fish oto-
liths, 246; function of crinoid stem, 275; gastric
contents of ichthyosaur, 376; index gastropod, 554;
Kimmeridgian ammonite, 16; microstructure of
mytilid shell, 163.

K

Kamptnerius magnificus, 152.

Kemp, E. M. Probable angiosperm pollen from the
British Barremian to Albian strata, 421.

822

INDEX

Kilenyi, T. I., and Allen, N. W. Marine-brackish
bands and their microfauna from the lower part of
the Weald Clay of Sussex and Surrey, 141.

Kistocare campbellensis, 230, 43.

Knoxisporites pristinus , 123, 27.

Kootenia, 195; aculacauda, 195, 41; brevispina , 196,
40; convoluta, 197, 41; crassa, 198,42; crassinucha,
198, 41; sp., 199, 43.

Kotujella calva, 92.

Krithodeophyton croftii, 684, 130-2.

Krommelbein, K. The first non-marine Lower Cret-
aceous ostracods from Ghana, West Africa, 259.

L

Lepidocarpon mazonensis, 454.

Lepidocyclina, 284; armata, 285, 50; gigas, 299; man-
telli, 295, 56, 57 ; montgomeriensis, 287 ; rdouvillei,
291, 55; yurnagunensis, 284, 49; y. morganopsis,
293, 295, 55, 56; ( Lepidocyclina ), 297 ; canellei, 297,
59; cf. crassicosta, 297, 58; sachsi, 299, 58 ; (Nephro-
lepidina), 294; bikiniensis, 300, 58; b. pumilipapilla,
301, 59; wilsoni, 294, 56.

Lepidophylloides acwninatus, 454.

Lepidophyllum gracile, 454.

Lepidostrobophyllum, 447 ; alatum, 448, 83, 84.

Lepidostrobus, 453; brevifolius, 454; goodei, 454;
lancifolius, 454; triangularis, 453.

Leptolepis, 254; densus, 254; roddenensis, 255;
tenuirostris, 254.

Ligaments: preserved in Permian bivalves, 94.

Liostrea, 472; anabarensis, 472; ex. gr. delta, 472;
praeanabarensis, 472.

Lipsanocystis magnus, 140.

Lopha, 482; folium, 89; semiplana, 90.

Lovenicystis angelini, 136.

Loxolitluis armilla, 148.

Lycopod: Carboniferous genus Eskdalia, 439; com-
pressed sporophyll, 445.

Lycospora torulosa, 123, 26.

Lyttonia sp., 68.

M

Macrocystella, 580; azaisi, 111-13; a. midticristata,
113; mariae, 583, 111-13.

Markalius cf. inversus, 144.

Marthasterites tribrachiatus, 153.

Martin, A. R. H. Aquilapollenites in the British Isles,
549.

Maslovella, 365; africana, 365, 69-71; blackii, 366,
69; pulchra, 366, 69.

McKerrow, W. S. See Ziegler, A. M.

Megadesmus', cuneatus, 20; nobilissimus, 20.

Megalosaurus bucklandi, 7, 8.

Metacypris, 262; sp. A, 262, 45; sp. B, 262, 45.

Mexico: new Cretaceous alveolinid genus, 526.

Mimocystites, 589.

Miogypsina cushmani, 295.

Mollusca. See Ammonoidea, Bivalvia, Gastropoda.

Monoplacophora : sinus-bearing, 132.

Moorellina leptaenoides, 68.

Mucrospirifer, 317; mucronatus, 63, 64.

Myonia; morrisi, 20; valida, 19, 20.

N

Neococcolithus sp., 152.

Nevada: Cambrian trilobites from Pioche Shale, 183.

Newman, B. H. The Jurassic dinosaur Scelidosaurus
Owen, 40.

New South Wales: brachiopod mantle canal patterns,
389; Famennian ammonoids, 535; new rhynchon-
elloid, 731; septate dalmanellid, 627; Silurian tri-
lobite, 691.

New York: colour markings in Devonian trilobites,
498.

Nisusia ferganensis, 91.

Nummulites, 669; chavannesi, 670, 128, 129; fabianii,
673, 129; pengaronensis, 676, 128.

O

Ogygopsis sp., 201, 40.

Old Red Sandstone: new plant from South Wales, 683.

Olenellus gilberti, 193, 36.

Olenoides, 199; steptoensis, 199, 39; sp., 200, 43.

Onchocephalus papulus, 224, 36.

Orbulinoides beckmannii, 371.

Ordovician: earliest known crinoid, 406; edrioblas-
toid from Australia, 513; Macrocystella, 580.

Oryctocephalites typicalis, 202, 41.

Oryctocephalus, 201; maladensis, 202, 41; cf. primus,

201, 40.

Ostracoda: non-marine from Ghana, 259; Wealden
marine-brackish, 141.

Ostracoderm : new evidence on Jamoytius, 21 .

Ostrea ; bononiae, 2; multiformis, 2.

Oxinoxis ampullacea, 605, 117.

Oysters: epizoic on ammonites, 19; functional study
of, 458.

P

Pachyaspis, 231; gallagari, 231, 40; longa, 231, 40.

Paedumias sp., 193, 36.

Pagetia, 189; arenosa, 189, 43; clytia, 190, 43; mala-
densis, 190, 43; mucrogena, 191, 43; resseri, 192, 38.

Palaeodiscus ferox, 577.

Palaeonummulites, 287; antiguensis, 301, 51; kugleri,
287, 50; palmarealensis, 288, 50; stainforthi, 288, 51.

Palynology: Aquilapollenites in Britain, 549; Cret-
aceous angiosperm pollen, 421; dispersed spores of
Equisetites, 633; Tournaisian spore flora, 116.

Pant, D. D., and Singh, R. S. The structure of Verte-
braria indica Royle, 643.

Parabolinoides, 410; bucephalus, 410, 77; contractus,
413; hebe, 413; palatus, 413.

Paul, C. R. C. Macrocystella Callaway, the earliest
glyptocystitid cystoid, 580.

— Morphology and function of dichoporite pore-
structures in cystoids, 697.

Pectinatites ( Virgatosphinctoides ), 19; elegans, 2;
reisiformis densicostatus, 2.

Pemma papillatum, 147.

Permian: brachial skeleton of Attenuatella, 783; pre-

INDEX

823

served ligaments in bivalves, 94; structure of
Vertebraria, 643.

Peytonoceras ? involution, 271, 46.

Phacops rana, 96.

Phillips, T. L., and Andrews, H. N. Biscalitheca
(Coenopteridales) from the Upper Pennsylvanian
of Illinois, 104.

Pholidophorus, 251 ; paradoxicus, 251 ; prae-elops, 252.

Pienaar, R. N. Upper Cretaceous coccolithophorids
from Zululand, South Africa, 361.

Pilasporites allenii, 638, 123.

Pioche Shale: trilobites from, 183.

Planicardinia, 628; carroli, 629, 122.

Plants: Carboniferous lycopod, 439, 445; new Lower
Devonian, 683; Pennsylvanian fern, 104; structure
of Vertebraria, 643; Triassic from Argentina, 236,
500. See also Algae, Palynology.

Platyclymenia, 537; (P.) annulata annulata, 538, 104;
a. densicosta, 540, 104; alterna, 543, 105; pattinsoni,
542; teicherti, 541, 104, 105.

Pleurocystites-, elegans, 1 34 ; filitextus, 134, 140; rugeri,
134.

Pleurodictyum sp., 60, 17.

Pliolepidina, 289; tobleri panamensis, 289, 292, 52-55;
? sp., 289, 54.

Pocock, Y. P. Carboniferous schizophoriid brachio-
pods from Western Europe, 64.

Podorhabdus cylindratus, 150.

Poliella, 206; denticulata, 206, 38; germana, 207, 37;
leipalox, 208, 38.

Pollard, J. E. The gastric contents of an ichthyosaur
from the Lower Lias of Lyme Regis, Dorset, 376.

Poly podorhabdus madingleyensis, 150.

Pontolithina moorevillensis, 149.

Pontosphaera, 800; discopora, 800, 146; versa, 146.

Praemytilus str at hair densis, 163, 31-35.

Propectinatites, 16; websteri, 17, 1.

Pseudo climacograptus, 2; ( Clinoclimacograptus ) retro-
versus, 8; ( Metaclimacograptus ) hughesi, 3; undu-
latus, 6.

Pseudocrinites; gordoni, 136; perdewi, 136; pyriformis,

136.

Pseudo phragmina ( Proporocyclina ) flint ensis, 290.

Ptarmiganoides, 209; araneicauda, 209, 37; lepida?,
209, 37; sp., 210, 38.

Pustulatisporites gibberosus, 118, 25.

Pycnodonta hyotis, 89.

Pyramus laevis, 19, 20.

Q

Queensland: Permian brachiopods from, 783.

R

Raistrickia, 118; clavata, 118, 25; corynoges, 118, 25.

Ramseyocrinus, 406; cambriensis, 407, 76.

Rectoclymenia ? sp., 544, 105.

1 Reptiles: ichthyosaur gastric contents, 376; Jurassic
j] dinosaur, 40.
i Rhabdolithina spp., 148,

] Rhexoxylon, 236, 509; africanum, 239; piatnitzkyi,
240, 44; tetrapteridoides, 239.

Rhinocypris jurassica jurassica, 148, 29.

Rickards, R. B. See Bulman, O. M. B.

Ritchie, A. New evidence on Jamoytius kerwoodi
White, an important ostracoderm from the Silurian
of Lanarkshire, Scotland, 21.

Roberts, J. Mantle canal patterns in Schizophoria
(Brachiopoda) from the Lower Carboniferous of
New South Wales, 389.

Robinson, E. Chubbina, a new Cretaceous alveolinid
genus from Jamaica and Mexico, 526.

Rollins, H. B., and Batten, R. L. A sinus-bearing
monoplacophoran and its role in the classification
of primitive molluscs, 132.

Rudwick, M. J. S. The feeding mechanisms and
affinities of the Triassic brachiopods Thecospira
Zugmayer and Bactrynium Emmrich, 329.

Runnegar, B. Preserved ligaments in Australian
Permian bivalves, 94.

Rushton, A. W. A. Revision of two Upper Cambrian
trilobites, 410.

S

Saccammina glenisteri, 603, 116.

Samanta, B. K. Nummulites (Foraminifera) from the
Upper Eocene Kohili Formation of Assam, India,
669.

Savage, N. M. Australirhynchia, a new Lower Devon-
ian rhynchonelloid brachiopod from New South
Wales, 731.

— Planicardinia, a new septate dalmanellid brachio-
pod from the Lower Devonian of New South Wales,
627.

— See also Walmsley, V. G.

Scandinavia: new Llandovery diplograptids from, 1.

Scapholithus sp., 148.

Scelidosaurus harrisoni, 40, 7, 8.

Schizocystis armata, 137.

Schizophoria, 64, 389; annectans, 77, 18; connivens,
64, 18; gibbera, 69, 18; linguata, 72, 18; resupinata,
80, 18; verulamensis, 74, 75; woodi, 86, 18.

Schopfites claviger, 121, 25.

Schuleridea ( Eoschuleridea ) wealdensis, 151, 30.

Scotland: Lower Silurian graptolite assemblages,
654; Silurian ostracoderm, 21; Tournaisian spore
flora, 1 16.

Scyphosphaera sp., 154.

Seilacher, A., Drozdzewski, G., and Haude, R. Form
and function of the stem in a pseudoplanktonic
crinoid ( Seirocrinus ), 275.

Seirocrinus subangularis, 275, 48.

Shell structure: Billingsellacean brachiopods, 486;
Jurassic mytilid, 163.

Sherwin, L. Denckmannites (Trilobita) from the
Silurian of New South Wales, 691.

Silurian: Birkhill graptolite assemblages, 654; Dayia
navicula, 612; Denckmannites from Australia, 691;
echinoid pedicellariae, 576; Llandovery transgres-
sion, 736; medusoid (?), 610; new Llandovery
diplograptids, 1 ; ostracoderm, 21 ; resserellid
brachiopod, 306.

Sindulites sindensis, 435.

824

INDEX

Singh, R. S. See Pant, D. D.

Sinuitopsis acutilira, 134, 28.

Smout, A. H. See Eames, F. E.

Sollasites horticus, 144.

Sorosphaeroidea adhaerens, 604, 116.

South Africa: Cretaceous coccoliths from Zululand,
361.

South Wales: new plant from Lower Old Red Sand-
stone, 683.

Spencia quadrata, 232, 40.

Sphaerocystites multifasciata, 136.

Sphaeronchus, 250; circularis, 251; dorsetensis, 250.

Sphaerophthalmus, 414; alatus, 417, 78; humilis , 417,
78; major, 416, 78; majusculus, 417.

Sporadoceras, 545; inflexion, 545, 105; cf. rotundum,
546, 105.

Staurocystis quadrifasciatus, 137.

Staurolithites sp., 148.

Stephanolithion, 807; bigoti, 152; laffittei, 152.

Sternbergella ( Parasternbergella ), 147; wolburgi, 148,

30.

Stinton, F. C., and Torrens, H. S. Fish otoliths from
the Bathonian of southern England, 246.

Strachan, I. A new medusoid (?) from the Silurian of
England, 610.

Strobilocystites calvini, 136, 140.

Sullivan, H. J. A Tournaisian spore flora from the
Cementstone Group of Ayrshire, Scotland, 116.

Suppiluliumaella polyreme, 495, 95.

Syracosphaera; hystrica, 148; pulchra, 148.

Syspacephalus ?, 225; cf. uncus, 225, 36; sp., 226, 37.

T

Taxonomy: problems in study of coccoliths, 793.

Tertiary: AquilapoUenites, 549; Eocene Cassigerinella,
368; Indian Nummulites, 669; larger foraminifera,
283; new nummulitid, 435; Orbulinoides, 371.

Tethys: calcareous algae of, 491; index gastropod in
Jurassic of, 554.

Thamnopora , 54; boloniensis, 55, 12, 14; cervicornis,
56, 11, 12; cronigera, 57, 12; irregularis, 58, 11;
polyforata, 57, 11; polymorpha, 59, 10; tumefacta,
60, 11.

Thecospira, 329; haidingeri, 65; tyrolensis, 65; sp.,

66, 68.

Theriosynoecum, 158.

Thomacystis tuberculata, 139.

Thomas, B. A. A revision of the Carboniferous lyco-
pod genus Eskdalia Kidston, 439.

Toghill, P. The graptolite assemblages and zones of
the Birkhill Shales (Lower Silurian) at Dobb’s
Linn, 654.

Tolypammina, 606; devoniana, 606, 114; helina, 607,
115; nexuosa, 607, 115; sp., 608, 114.

Torrens, H. S. See Stinton, F. C.

Tremalithus; burwellensis, 145 ; danicus, 145; dictyodus,
145; placomorphus, 145.

Triassic: brachiopod feeding mechanisms, 329; leaf
genus Dicroidium, 500; trunk of Rhexoxylon, 236.

Tricolpites, 430; albiensis, 430, 81 ; sp. 1 , 431, 81 ; sp. 2,
432, 81.

Trilobita: Cambrian from Nevada, 183; colour mark-
ings in Devonian, 498; Silurian Denckmannites,
691 ; two Upper Cambrian, 410.

Tucker, E. V. The atrypidine brachiopod Dayia navi-
cula (J. de C. Sowerby), 612.

U

?Utriculith, 257.

V

Vacunella ; curvata, 20; ?sp. nov., 20.

Vallatisporites vallatus, 123, 26.

Verrucosisporites, 121; scoticus, 121, 25; variotuber-
culatus, 121, 26.

Vertebraria indica, 643, 124-7.

Vertebrata: fish otoliths, 246; ichthyosaur gastric
contents, 376; Jurassic dinosaur, 40; Silurian
ostracoderm, 21.

Visbyella, 306; cumnockensis, 313, 61; nana, 311, 62;
pygmaea, 310, 61; visbyensis, 307, 60.

W

Walmsley, V. G., Boucot, A. J., Harper, C. W., and
Savage, N. M. Visbyella — a new genus of resserellid
brachiopod, 306.

Wealden: marine-brackish bands from Sussex and
Surrey, 141.

Webby, B. D. Astrocystites distans sp. nov., an edrio-
blastoid from the Ordovician of eastern Australia,
513.

Welsh Borderland: Llandovery transgression of, 736.

Wendt, J. Discohelix (Archaeogastropoda, Euompha-
lacea) as an index fossil in the Tethyan Jurassic, 554.

Williams, A. Shell structure of the Billingsellacean
brachiopods, 486.

Z

Zacanthoides, 211; demissus, 211, 42; sp., 212, 40.

Zacanthopsis levis, 217, 36.

Ziegler, A. M., Cocks, L. R. M., and McKerrow,
W. S. The Llandovery transgression of the Welsh
Borderland, 736.

Zones: Silurian graptolite, 654.

Zululand: Cretaceous coccoliths from, 361.

Zygolithus] diplogrammus, 148; sp., 149.

THE PALAEONTOLOGICAL ASSOCIATION

COUNCIL 1968-9

President

Professor Alwyn Williams, The Queen’s University, Belfast
Vice-President

Dr. W. S. McKerrow, University Museum, Oxford
Treasurer

Dr. C. Downie, Department of Geology, The University, Mappin Street, Sheffield 1

Membership Treasurer

Dr. A. J. Lloyd, Department of Geology, University College, Gower Street, London, W.C.l

Secretary

Dr. J. M. Hancock, Department of Geology, King’s College, London
Assistant Secretary

Dr. W. D. I. Rolfe, Hunterian Museum, The University, Glasgow, W. 2

Editors

Mr. N. F. Hughes, Sedgwick Museum, Cambridge
Dr. Gwyn Thomas, Department of Geology, Imperial College, London, S.W.7 .

Dr. I. Strachan, Department of Geology, The University, Birmingham, 15
Professor M. R. House, The University, Kingston upon Hull, Yorkshire
Dr. R. Goldring, Department of Geology, The University, Reading

Other members of Council

Dr. F. M. Broadhurst, The University, Manchester
Mr. M. A. Calver, Institute of Geological Sciences, Leeds
Dr. C. B. Cox, King’s College, London
Mr. D. Curry, Eastbury Grange, Northwood, Middlesex
Dr. Grace Dunlop, Bedford College, London
Mr. G. F. Elliott, 60 Fitzjohn Avenue, Barnet, Herts.

Dr. A. Hallam, University Museum, Oxford
Dr. Julia Hubbard, King’s College, London
Dr. J. D. Hudson, The University, Leicester
Dr. R. P. S. Jefferies, British Museum (Natural History), London
Dr. J. D. Lawson, The University, Glasgow
Dr. A. H. Smout, British Petroleum Company, Sunbury-on-Thames
Professor H. B. Whittington, Sedgwick Museum, Cambridge
Overseas Representatives

Australia : Professor Dorothy Hill, Department of Geology, University of Queensland, Brisbane
Canada : Dr. D. J. McLaren, Institute of Sedimentary and Petroleum Geology, 3303-33rd Street
NW., Calgary, Alberta

India : Professor M. R. Sahni, 98 The Mall, Lucknow (U.P.), India

New Zealand: Dr. C. A. Fleming, New Zealand Geological Survey, P.O. Box 368, Lower Hutt
West Indies and Central America: Mr. John B. Saunders, Geological Laboratory, Texaco Trinidad,
Inc., Pointe-^-Pierre, Trinidad, West Indies

Western U.S.A. : Professor J. Wyatt Durham, Department of Paleontology, University of California,
Berkeley 4, Calif.

Eastern U.S.A. : Professor J. W. Wells, Department of Geology, Cornell University, Ithaca, New York

PALAEONTOLOGY

VOLUME 1 1 ' PART 5

CONTENTS

The structure of Vertebraria indica Royle. By D. d. pant and R. s. singh

The graptolite assemblages and zones of the Birkhill Shales (Lower Silurian)
at Dobb’s Linn. By p. toghill

Nummulites (F or ammif era.) from the Upper Eocene Kopili Formation of Assam
India. By b. k. samanta
i /A new plant from the Lower Old Red Sandstone of South Wales. By

DIANNE EDWARDS

Denckmannites (Trilobita) from the Silurian of New South Wales. By L. sherwin
Morphology and function of dichoporite pore-structures in cystoids. By
C. R. C. PAUL

Australirhynchia, a new Lower Devonian rhynchonelloid brachiopod from New
South Wales. By N. M. savage

The Llandovery transgression of the Welsh Borderland. By A. M. ziegler,
l. r. m. cocks, and w. s. mckerrow

The unusual brachial skeleton of Attenuatella convexa sp. nov. (Brachiopoda).
By 3. ARMSTRONG

Taxonomic problems in the study of coccoliths. By M. black

643

654

669

683

691

697

673

736

783

793

PRINTED IN GREAT BRITAIN AT THE UNIVERSITY PRESS, OXFORD
BY VIVIAN RIDLER, PRINTER TO THE UNIVERSITY

CO Z CO -•'* Z CO Z to

ion NouniiiSNi_NviNosHims S3iavyan libraries Smithsonian institution Noun

to ~T \ to — CO — CO

sn libraries Smithsonian institution NouniiiSNi nvinoshxiins saiavaan libri

^ ^ — t/) \ ^ (/)

'ion NouniiiSNi nvinoshiiws saiavaan libraries smithsonian institution Noun

to Z CO Z . tO -r ... CO

X A'

O ^

Z vo***'^/ *.

an libraries smithsonian institution NouniiiSNi nvinoshjliws °S3 i avya nzu bri

CO Z 07 to X

tO As ^ M

c c
<

,y 5 2 W I | - "V S o

ion'' NOliniUSNlAviNOSHiinS S3iavaanJLIBRARIESZSMITHS0NIANJ|NSTITUTI0N:rN0Un
z r- v z r- z —

— ^ — GO _ C/> ?

an libraries smithsonian institution NouniiiSNi nvinoshiiins S3iavaan libr;

5 ^ z to z to z

Z A —I IT .V _» T ZcAstoCX. — i

/■

o

to

X

H-

s g » § . 4?^' i '®wv/ a ,

» A I V»y :f > s s XaA >

ION NOimiliSNI_ NVIN0SH1IINS S3IMVaan libraries SMITHSONIAN INSTITUTION NOlin:
to z \ 00 „ = to — to

^ CO X w to UJ JS LU y^I!

c

H v

_ O ~ O XJ^SZ “■

-* 2 J Z _j —

an LIBRARIES SMITHSONIAN INSTITUTION NOlinillSNI NVIN0SH1IWS S3iavaaH LIBRA
— Z r* z r~ z

CO

31
>

3J

m xgNDCjz m m w CO rn

(O £ co 5 — CO \ 5 CO

ION NOlinillSNI NVINOSHHWS S3IMVaan LIBRARIES SMITHSONIAN INSTITUTION NOI10
to Z to z to z ... co

< .. s < a\.. s

X

to tf.

> 2 S v > ^inT^y ^ > 'W

a n_ U 8 rar i es smithsonian_institution NoiiniusNi^NviNOSHinNs^s 3 1 y vy a \\_ li b r/
^ ^ - .... *.Ǥ I .v*fe ? /^USks

■ >>

•A'

NVIN0SH1IWS°°S3 I a vy a nZU BRAR I ES^SMITHSONIAN^INSTITUTION NO!iniliSNI_NVINQSHlllNS

to ~ co — to 2:

- ^ uj co

^ CK —

f <

_J

O NJV\o5 _ ’ O Xftos*£

-I 2 _j 2 -J ^

SMITHSONIAN INSTITUTION NOIlfllllSNI NVIN0SH1IWS S3iaVM9n LIBRARIES SMITHSONIAIs

2 r~ 2 C z

O /T\asvTT\ “ . O — xtwo.TN O

co

■*» ^=* •~) 1*^ ^ 3
m rn

co ± co \ ± <0

nvinoshiiws S3iavaan libraries Smithsonian institution NoiinmsNi nvinoshihns

CO 2 »,. CO 2 ... CO —

^ > vp* 5 ^ > Nsjpms!^ 2 ®\ > '‘W^ 2

SMITHSONIAN INSTITUTION MOIlfUIJlSm NVINOSHIIWS^S 3 I BVB 9 11 LIBRARIES SMITHSONIAN

CO — CO 2 CO

DL
<

«rmr H VcW.07 to Z! " W* m =j m '<»

MviNOSHims^sa lavaan^UBRAR i es^ Smithsonian^ institution 2 NouniiiSNi~JNviMOSHiii/MS

v 2 r~ 2 r* 2 «“ v

m x^vosvvvx rn

co 2 co _ co

SMITHSONIAN INSTITUTION NOlinillSNI NVIN0SHJL1WS S3iaVMail LIBRARIES SMITHSONIAIS

2 » CO 2 CO 2

s 3 Wm>, s £$rm\ 3 «k z

- ■4"m o P

rw w p

-x- «A-V w ss/t/' -tff:, X VC

-5 ' W > ' 3E X^voscOJ^ >

co • ■*“' 2 co 2 to

nvinoshiiws S3 lava an libraries Smithsonian institution NoiinuxsNi_NviNOSHim$

CO — CO ~ CO

c

-> 2 _J 2 -1 ^

SMITHSONIAN^ INSTITUTION NOlinillSNI NVIN0SH1MS S3iaVHail LIBRARIES SMITHSONIAf

yv- ^ /'vfn^v I o AfL> | p ./

rn ^ in g

2i
>

W 30

rn VLjtDv^ w Jr; W u; rn

CO _ — co X E to

NVIN0SH1IINS S3iavaan LIBRARIES SMITHSONIAN. INSTITUTION NOlinillSNI NVINOSHIIIM!

CO 2 CO 2 V- CO 2

SSS&X E , . .*s s

■ny I t I O t 1 -

bMITHSO1^’ INSTITUTION NOIlfllllSNI NVINOSHIIINS^SB I aVH 9 l*lf LI B RAR I ES^SMITHSONIAf

<c?N B,soJv 07 2 CO „ CO

k /W'iP'WX Cc* fE .^SSvV ... UJ Vk X^V«OV>.v tu