Postilla

pha BD h ret Deas ted St $eF 2 it)
ots init FORO Teh si

poly ht Wr Ayia bert an ety ts edd

PTs teh hie)

Rosette
=

tank

rad tof
ed bait.

piper er tis DAC aE

a Me tety
Aviv ters

BeBe Mette USED URS Ar bee
soe ed Teka

Si aterd -latael)
fey Sel

i
NS iab aust) 4 otal akcie
S962 PIPE Rede N rena,

sd eae FRA BlRed 0~

hehehe Oe ee te hires tate SPR
gadis $2 Sele eRe Dobe be
o> trp eet tte gether

YM IREN ORD RT DIV ON ANE Be ETP NRG NS
AVA SADE ASE Mes y LOSS

Se eee oe ee ee
Nateere Siz Poetet 9 tae

nf ain fo Kg PALI OL PTE Si ohh

eee

det gta: te
me Seta

ae ae

as : - - —_—— 2 =
= as - - - - —_~ a
- e x5 eo
< = - .
= ze -
. ig : :
- 2 F ;

ros

HARVARD UNIVERSITY

Els EO
ES

LIBRARY

OF THE

Museum of Comparative Zoology

ST |
VSiHi2
oF Aaa o/%,¢ 2? re
cu > a
aha Dik ae ;
OHM Glyph), a M977 | |
. ae es ae —_
Are ai) ape oe a) a ,
* a
m
; 4
e
an ,
%
a.»

MUS. COMP. ZOOL!
LIBRARY,

FEB 20 1969

HARVARD
UNIVERSITY)

POSTILLA

PEABODY MUSEUM
YALE UNIVERSITY

NUMBER 127. 31 DEC. 1968

VERTEBRAL STRUCTURE IN RHI-
PIDISTIA (OSTEICHTHYES, CROS-
SOPTERYGII) WITH DESCRIPTION
OF A NEW PERMIAN GENUS

KEITH STEWART THOMSON
PETER PAUL VAUGHN

Li

POSTILLA

Published by the Peabody Museum of Natural History, Yale University

Postilla includes results of original research on systematic, evolution-
ary, morphological, and ecological biology, including paleontology.
Syntheses and other theoretical papers based on research are also
welcomed. Postilla is intended primarily for papers by the staff of
the Peabody Museum or on research using material in this Museum.

Editors: Jeanne E. Remington and Nancy A. Ahlstrom

Postilla is published at frequent but irregular intervals. Manuscripts,
orders for publications, and all correspondence concerning publications
should be directed to:

Publications Office
Peabody Museum of Natural History
New Haven, Conn., 06520, U.S.A.

Lists of the publications of the Museum are available from the above
office. These include Postilla, Bulletin, Discovery, special publications,
and available back numbers of the discontinued journal, Bulletin of the
Bingham Oceanographic Collection. All except Discovery are available
in exchange for relevant publications of other scientific institutions
anywhere in the world.

VERTEBRAL STRUCTURE IN RHIPIDISTIA (OS-
TEICHTHYES, CROSSOPTERYGII) WITH DE-
SCRIPTION OF A NEW PERMIAN GENUS

KEITH STEWART THOMSON

Department of Biology and Peabody Museum of Natural
History, Yale University

PETER PAUL VAUGHN

Department of Zoology, University of California, Los Angeles

ABSTRACT

Ectosteorhachis nitidus from the Lower Permian of North America
has holospondylous vertebrae, but a new North American genus
of osteolepid rhipidistian of the same age has compound vertebrae
each comprising a large principal and a small anterior median
dorsal accessory centrum. Attempted embryological analysis of
vertebral structure in Rhipidistia reveals no evidence of sclero-
tomic resegmentation and no direct homology with tetrapod
vertebrae.

POSTILLA 127: 19 p. 31 DECEMBER 1968.

MUS. COMP. ZOOL
LIBRARY,

FEB 20 1969

HARVARD

INTRODUCTION UNIVERSITY,
The Paleozoic fishes of the suborder Rhipidistia (Order Crosso-
pterygii) are known from a variety of forms of Early Devonian to
Early Permian age. They are of special interest because it was
from a Devonian rhipidistian stock that the first Amphibia evolved
(see, for example, Thomson, 1968) and it is well known that in
many features of their structure the Rhipidistia are somewhat
intermediate between other osteichthyan fishes and the tetrapods.
It is therefore important that rhipidistian structure be known in
as great detail as possible in order to interpret better the process
of the origin of the tetrapods. A matter of considerable importance
in this respect is the structure of the vertebral column. The ver-
tebral column must have been subject to a pronounced change in
function when the fishes left the water and began moving on land.
The tetrapods rapidly evolved a variety of different patterns of
vertebral structure in the new environment (for a recent study,
see Parrington, 1968) and the type of vertebral composition is,
in fact, a useful taxonomic character among higher categories of
fossil Amphibia (Romer, 1964). Thus it is of interest to study the
range of vertebral structure in Rhipidistia to see if any indications
of the patterns of tetrapod structure and development were already
present in these fishes.

The present paper describes and discusses the vertebral struc-
ture of two Lower Permian rhipidistian fishes. The description is
based largely on new material, particularly of a new fish collected
by Vaughn in Utah.

2 POSTILLA

MATERIALS

The material studied included the following specimens (for abbre-
viations, see end of paper): Ectosteorhachis nitidus Cope: MCZ
8630, almost complete fish with several vertebrae exposed; MCZ
8930, complete skull and a small postcranial fragment with
4 vertebrae; YPM 5000, almost complete fish lacking posterior
trunk and tail, two vertebrae in situ exposed by preparation; all
material of Ectosteorhachis from the Lower Permian, Wichita
Group, of North-central Texas (see Romer, 1958).

Specimens on which the new genus and species are based are:
YPM 5701, isolated and flattened postparietal shield; YPM 5702,
7 postcranial fragments with scales, 3 with vertebrae in natural

RHIPIDISTIAN VERTEBRAE 3

articulation but slightly dislocated; YPM 5703, left gular plate;
UCLA VP 1688, disarticulated fragments including one with
mandible in natural association. All material of the new genus
is from the Lower Permian, Cutler Group, Halgaito Shale of
Southeastern Utah. Some of this material was formerly referred
to as “Ectosteorhachis” by Vaughn (1962).

ECTOSTEORHACHIS

The original description (Cope, 1880) of the species Ectosteor-
hachis nitidus Cope includes the following description of the
vertebrae: “In Ectosteorhachis they are represented by annular
ossifications resembling somewhat those of the stegocephalous
genus Cricotus, but with a larger foramen chordae dorsalis”. Cope
considered this structure to differ from that of the “completely
ossified”, “biconcave” vertebrae of Megalichthys. Hussakoft (1911)
synonymized Ectosteorhachis and Megalichthys, stating that the
vertebrae “in both ... are narrow rings, but those in Cope’s
specimen ... are not well enough preserved to make it absolutely
certain that they were complete, and not open, above”. In the more
recent literature (e.g. Thomson, 1967) the acceptance of a simple
ring-shaped vertebral structure in Ectosteorhachis has been con-
tinued, while this genus is distinguished from Megalichthys on
other evidence (Thomson, 1964).

Newly available material of Ectosteorhachis nitidus from the
Wichita Group of north-central Texas includes vertebrae in natural
articulation (YPM 5000, MCZ 8930) and also vertebrae from
the immediately postcranial region to the level of the first dorsal
fin.

Each vertebral unit in all material of Ectosteorhachis that we
studied consists of a complete, ring-shaped principal centrum to
which the neural arch is attached (Fig. |). The principal centrum
is slightly wedge-shaped in lateral view, tapering dorsally from a
wide base. The notochordal canal is wide and the wall of the
principal centrum is correspondingly slim. The posterior portion of
the lateral wall of the principal centrum is significantly depressed
and the “step” between the raised and depressed surface is devel-
oped into a slight ridge that is interpreted as having served for the
attachment of the myoseptum (see below). The recessed posterior
region does not extend dorsally to the midline but just short of

4 POSTILLA

Fig. 1. Ectosteorhachis nitidus Cope. Vertebral centrum in left lateral and
posterior view. MCZ 8930. « 5

this point there is a small facet for the atttachment of a rib.
This facet separates the main lateral recessed region on either side
from a dorsal recessed area that perhaps formed the site of
attachment of the neural arch. Immediately in front of this dorsal
recessed area there is a short ridge which forms the posterior rim
of a transverse groove running directly ventrolaterally from the
neural canal; this groove probably carried the ventral spinal nerve
(Fig. 1). In front of this groove, on either side of the midline,
there is a short anterodorsally directed process that apparently
articulated with the rear surface of the neural arch associated with
the principal centrum in front. The lateral surface of the principal
centrum, anterior to the ridge for the myoseptum, is marked by a
series of small foramina (Fig. |) probably for small blood vessels.

The neural arches are not preserved in material at hand, but
their probable association with the principal centra is reconstructed
in Figure 2. Ventrally, in the posterior part of the trunk, there is a
pair of small haemal processes (hpr, Fig. 2) on the posterior
region of each principal centrum; these no doubt became devel-
oped into full haemal arches in the tail region.

In Figure 2 a reconstruction of the soft structures associated
with the vertebrae in Ectosteorhachis is given. It will be noted that
it is necessary to restore a considerable amount of cartilage be-
tween each principal centrum.

RHIPIDISTIAN VERTEBRAE 5

A NEW PERMIAN RHIPIDISTIAN

The material from southeastern Utah represents a fish different
than Ectosteorhachis although in previous studies it had been
tentatively assigned to that genus (Vaughn, 1962). It is only the
second known genus of rhipidistian of unequivocal Permian age.
A formal diagnosis of this new fish is given below, followed by a
complete description of the vertebral structure.

my

Fig. 2. Ectosteorhachis nitidus Cope. Reconstruction of three vertebrae
and associated soft parts in left lateral view. The stippling represents
cartilage.

6 POSTILLA

FAMILY OSTEOLEPIDAE
Lohsania gen. n.
TYPE SPECIES. Lohsania utahensis sp. n.

DERIVATION OF NAME. Lofsania (feminine) from the Navajo
words for fish (loh) and old (sani).

N
a
~~

Fig. 3. Lohsania utahensis gen. et. sp. nov. Left gular plate. « 1.2

DIAGNOSIS. Osteolepid rhipidistian of medium size, estimated maxi-
mum total length 60 cm. Postparietal shield essentially as in
Ectosteorhachis; gular bone more narrow and elongate than in
Ectosteorhachis, maximum width contained 3.3 times in greatest
length (as opposed to 2.3 times in Ectosteorhachis). Each ver-
tebral unit composed of a principal centrum that is incomplete
dorsally and a crescentic anterior accessory centrum lying in the
dorsal midline. Neural arch attached primarily to the accessory
centrum. Posterior recessed area of lateral wall of principal
centrum lacking (in available material).

DESCRIPTION, Lohsania is readily assigned to the osteolepid
Rhipidistia because of the typical structure of the scales and

RHIPIDISTIAN VERTEBRAE 7

Fig. 4. A and B. Lohsania utahensis gen. et sp. NOV. Mandibles and asso-
ciated elements in left and right lateral view. UCLA VP 1688. x 0.8

C. Ectosteorhachis nitidus Cope. Postparietal shield in dorsal view. MCZ

8930. « 0.8

D. Lohsania utahensis gen. et sp. NOV.
YPMs5701. SC1-5

Postparietal shield in dorsal view.

8 POSTILLA

Fig. 5. Lohsania utahensis gen. et sp. nov. Five vertebrae in right lateral
view, slightly displaced. Holotype YPM 5702. « 2

Fig. 6. Lohsania utahensis gen. et sp. nov. Two incomplete vertebrae
in right lateral view, slightly displaced. Holotype YPM 5702. x 2.8

RHIPIDISTIAN VERTEBRAE 9

Fig. 7. Lohsania utahensis gen. et sp. nov. Vertebra in left lateral and
anterior view. YPM 5702. x 3.2

dermal bones. We have been unable to distinguish the scales from
those of Ectosteorhachis in either gross or micro-structure. The
postparietal shield of the dermal skull roof, illustrated in Figure 4
along with the same region in Ectosterorhachis, was not found in
direct association with vertebrae of the characteristic Lohsanta-
type, but is confidently assigned to this taxon. The main points of
difference are in the somewhat slightly broader anterior margin
and the shape of the posterior margin.

A single vertebral unit in Lohsania (Figs. 5 and 6) consists of
three separate elements — a principal and an accessory centrum
and a neural arch. The principal centrum in the available material
is relatively undifferentiated. However, we consider a_ poorly
defined ridge (Figure 7) running slightly diagonally across the
posterior part of the lateral face of the principal centrum to
mark the line of attachment of the myoseptum (Figure 8). The
recessed area posterior to this ridge, seen in Ectosteorhachis and

ac

nch

pc

10 POSTILLA

x Sts po : So a
\

iy,

Fig. 8. Lohsania utahensis gen. et sp. noy. Reconstruction of three ver-
tebrae and associated soft parts in left lateral view. The stippling indicates
cartilage.

other rhipidistians, is lacking. This absence may be associated
with the presence of an accessory centrum in Lohsania (see below).
Also lacking are grooves for the intersegmental arteries or spinal
nerves; this may be due to the imperfect nature of the preserva-
tion. In lateral view the principal centrum (Figure 7) tapers
markedly toward the dorsum. A unique feature of the vertebrae
of Lohsania is the presence of a median accessory central element

RHIPIDISTIAN VERTEBRAE 11

in the dorsal midline. This element is associated with the anterior
face of each principal centrum, as is demonstrated by the con-
stant close association of each accessory element with the prin-
cipal centrum behind even in material (such as YPM 5702,
Fig. 4) where considerable displacement of the vertebrae has
taken place. This constant relationship must be a natural phenom-
enon. The accessory centrum is crescentic in shape, and the
lateral wings curving down between the principal centra taper
sharply on either side. The accessory element bears on its dorsal
surface a pair of parallel, anteroposteriorly directed ridges (Figure
7) which mark the attachment of the neural arch. In two vertebrae
from our material (part of YPM 5702) the accessory centrum.
seems to be fused to the principal centrum. Possibly this is related
to the relative position along the column.

The neural arch is illustrated in Figures 5, 6 and 7. The base
of each arch seems to be associated primarily with the accessory
element but there was also a slight connection with the tips of
the principal centrum. There is a very small, but clearly defined,
canal for the dorsal ligament (Fig. 6). This canal was presumably
oriented horizontally and this allows us to restore precisely the
posterior slope of the neural arches.

Figure 8 is a tentative restoration of the soft structure asso-
ciated with the vertebral column in Lohsania.

Lohsania utahensis sp. n.

Ectosteorhachis aft. E. nitidus Vaughn, 1962, p. 533.

HoLotype. YPM 5702, fragments of trunk in partial articulation
(Figs. 5 and 6).

PARATYPES. YPM 5701, postparietal shield (Fig. 4); YPM 5703,
left gular bone (Figure 3); UCLA VP 1688, partially disarticulated
fragments including mandibles (Fig. 4).

OCCURRENCE. All specimens from the Halgaito Shale, Cutler Group,
Lower Permian of San Juan County, Utah; probably of Wolf-
campian age (see Vaughn, 1962). YPM 5702 —NW %4, NE 4
sec. 34, T. 40 S., R. 19 E. YPM 5701 —NW 4%, NW % sec. 29,
17 40'S:2R, 19 E.. YPM 5703 — NW 4 sec: 3,.-7..41 S.,-R217E.
UCLA VP 1688 — NW %, NE % sec. 34, T. 40 S., R. 19 E.

DIAGNOSIS. As for the genus, above.

12 POSTILLA

COMPARISON AND DISCUSSION

At the present time, the vertebral structure of Ectosteorhachis and
Lohsania may only be compared in detail with that of one other
rhipidistian fish, Eusthenopteron foordi Whiteaves (family Rhi-
zodontidae; Upper Devonian) as described by Jarvik (1952).
As shown in Figure 9 each vertebral unit in Eusthenopteron con-
sists of a neural arch and a principal centrum which are basically
very similar to those of Lohsania, and a set of “accessory ele-
ments” that are not at all similar to the single accessory element
in Lohsania.

The principal centrum of Eusthenopteron is extremely similar
to that of Lohsania, but it is important to note that the latter lacks
the extensive posterior recessed lateral surface. All three forms
possess a myoseptal ridge. This ridge, in Eusthenopteron and
Ectosteorhachis bears a facet for the articulation of a rib, although
the facet in Ectosteorhachis is much smaller. The posterior recessed
area on the principal centrum in Ectosteorhachis is developed in
the same way in Eusthenopteron (Fig. 9). The groove for the
spinal nerves in Ectosteorhachis is not seen in any other form
but a pair of notches in the accessory elements in Eusthenopteron
possibly mark the passage of these elements.

The accessory elements in Eusthenopteron are a pair of small
subcircular elements interposed between the neural arches (Fig.
9). They are also seen, somewhat imperfectly, in the Middle
Devonian genus Glyptolepis (family Holoptychidae) from Scot-
land (specimen OS3.11/2, Museum of Zoology, Cambridge Uni-
versity). The most striking feature of these elements is that they
are not associated with the notochord itself. In slightly disasso-
ciated specimens in which the various elements of the vertebra
become separated one from another (for example, the specimen
of Glyptolepis noted above), the accessory elements are shown
to be mechanically associated with the neural arches, while the
accessory elements in Lohsania are shown to be associated with
the principal centra behind them. The accessory elements in
Eusthenopteron are therefore not completely homologous with the
accessory elements in Lohsania. The accessory elements in
Eusthenopteron have been termed “‘interdorsals” (Jarvik 1952),
“pleurocentra” (Romer, 1964), or “intercalaries” (Schaeffer,
1967). That they are not true interdorsals and particularly that

RHIPIDISTIAN VERTEBRAE 13

Fig. 9. Eusthenopteron foordi Whiteaves. Reconstruction of two vertebrae
and associated soft parts in left lateral view. The stippling represents
cartilage.

they are not homologous with the pleurocentra of tetrapods (see
below) is demonstrated by the fact that they are not fully asso-
ciated with the notochordal sheath and thus are not true central
elements. Schaeffer's interpretation that they are intercalaries,
homologous with the intercalaries found between the neural arches
in certain actinopterygian fishes (e.g. Amia), seems the most
accurate and is accepted here. The accessory elements of Lohsania,
on the other hand, are fully associated with the notochordal sheath.

It is always difficult to attempt to decide the homology of a
particular bony element solely from fossil material. This is
particularly true when it comes to the homology of the vertebrae.
Despite this difficulty, problems that require consideration are
the questions of vertebral homology and of a possible resegmenta-
tion of the vertebral column in Rhipidistia. It is now well accepted
(Williams, 1959; Panchen, 1967; and Schaeffer, 1967) that the
vertebrae of tetrapods undergo an ontogenetic resegmentation of
the original, segmentally arranged sclerotomic material that be-

14 POSTILLA

comes involved in the organization of the adult vertebra within
the perichordal tube. This occurs through the separation of the
cranial and caudal halves of sclerotome segments and their sub-
sequent recombination such that the adult vertebra is formed
from the caudal half of the sclerotomic material of one segment
and the cranial half of the sclerotomic material from the segment
immediately behind. Thus a new intervertebral separation devel-
ops in an originally intrasegmental position. Such sclerotomic
resegmentation does not occur in living fishes (at least not in
exactly the same form; cf. Lepisosteus in Schaeffer, 1967). The
homology of the vertebral elements of the Rhipidistia is of primary
interest in this respect, because of the almost intermediate posi-
tion that they occupy between fishes and tetrapods.

The Gadovian system of vertebral homology involving the
interpretation of vertebral components as being induced by a
series of embryonic elements, has been subjected to considerable
re-examination in recent years (Williams, 1959; Schaeffer, 1967).
While this system is evidently not applicable to most tetrapods,
in fishes such as Amia (Schaeffer, 1967) it is possible to distin-
guish in the very early developmental stages a series of segmentally
arranged anlagen which induce the development of the final os-
seous centrum. We cannot, of course, observe any part of such
a process in fossil forms in cases in which the adult centrum is
holospondylous. However, where the adult vertebra is composed
of more than one element the strong likelihood exists that each
element is induced by a separate anlage and we may attempt such
an analysis in order to try to shed more light on the problem of
vertebral homology.

In considering the fossil Rhipidistia, we have as guides to the
homology of the vertebral elements the position of the interseg-
mental artery, the position of the myoseptum, and the position
of the haemal process on the principal centrum. In all forms that
we know about, the myoseptum is located in the normal embryolog-
ically “primitive” position in the posterior half of the principal
centrum, with the intervertebral artery behind it. The myoseptum
passes directly up onto the neural arch in all forms (with the
possible exception of Lohsania). In the case of the holospondylous
Ectosteorhachis (Fig. 10A), we may see that the portions of the
vertebrae that might normally be considered to be derived from
anlagen in the embryonic caudal sclerotome-half, namely, the

RHIPIDISTIAN VERTEBRAE 15

Fig. 10. Hypothetical analysis of the embryonic derivation of the vertebral
units in (A) Ectosteorhachis, (B) Lohsania and (C) Eusthenopteron.
Portions indicated with open circles are thought to be induced by the
basidorsal, closed circles induced by the interdorsal, horizontal lines by
the basiventral and diagonal lines by the interventral.

neural arch (induced by the so-called basidorsal) and the haemal
arch (induced by the basiventral), are in the same posterior
position in the adult. In continuing this rather academic analysis
in Gadovian terms, we may identify the anterodorsal portion of
the principal centrum (including the groove for the spinal nerve
and the anterodorsal articular process) as having been induced
by an interdorsal and the remaining anteroventral portion of the
centrum as having been induced by an interventral (Figure 10A).
It will be seen that there is no indication here of vertebral reseg-
mentation.

The situation in Lohsania (Figure 10B) is exactly comparable,
except that in this case the accessory centrum (in a position sug-
gesting induction by the interdorsal) is formed as a separate ele-
ment and, presumably for mechanical reasons involved with the
function of the vertebral column, the principal centrum is narrow
dorsally. The neural arch nonetheless retains its posterior position.

In our opinion, the construction in Eusthenopteron may also be
considered to follow the same pattern, with the exception that the
interdorsal anlage, instead of inducing a central element (as in
Lohsania) or a portion of the principal centrum (as in Ecto-

16 POSTILLA

steorhachis), has induced the formation of an intercalary element
(Figure 10C). Even so, it will be noted that the element induced
by the interdorsal also has an asscciation with the spinal nerves.

The condition in Megalichthys and other holospondylous forms
such as Rhizodus and Rhizodopsis is presumably the same as in
Ectosteorhachis. Dr. S. M. Andrews (as quoted in Schaeffer, 1957)
has discovered the existence of a diplospondylous condition in
Osteolepis. While the above interpretations are completely tenta-
tive and will be liable to modification upon full publication of
Dr. Andrew’s results, it may be noted that a full diplospondylous
condition could be derived in the scheme given above by simple
separation of elements induced in the cranial and caudal halves
of a nonresegmented unit, that is, by separation of a unit induced
by the combined interdorsal and interventral instead of one
induced by the interdorsal alone.

The structure of neither Lohsania nor Eusthenopteron seems
to be directly comparable to that of any tetrapods except the
Ichthyostegalia, although it is possible that the embryonic rudi-
ments that induce rhipidistian structures induce different structures
which, through resegmentation, make up the vertebrae of tetrapods.
The extremely close similarity of structure between Eusthenop-
teron and Ichthyostega, with intercalary elements rather than true
accessory centra, must reflect a very close similarity in the func-
tion of the vertebral column in these forms. However, in view of
our conclusion that the intercalary elements of Eusthenopteron
bear no close relationship to the pleurocentra of tetrapods such
as the Rhachitomi, we must note that a similar view must apply
to the intercalaries (the so-called pleurocentra) of /chthyostega.
In fact, from the evidence of the vertebral column we are inclined
to separate the Ichthyostegalia from all other tetrapods and further-
more we consider it unlikely that the ichthyostegals gave rise to
any known later Temnnospondyli.

The above scheme of vertebral homology is entirely compatible
with our knowledge of the structure and development of the
vertebrae of “primitive” actinopterygian fishes, such as Amua
(Goodrich, 1930; Schaeffer, 1967). And insofar as the standard
Gadovian terminology is applicable to such forms as Amia, we
feel justified in using it as a tool for the interpretation of rhipi-
distian vertebrae, especially since there is no sign of resegmenta-
tion or the sort of modification of embryonic structure seen in

RHIPIDISTIAN VERTEBRAE 17

Lepisosteus in these fishes. That the structure of the Rhipidistia
is more closely comparable with that of other osteichthyan fishes
rather than with that of tetrapods is perhaps not surprising, and
there is, in fact, a remarkable resemblance between the vertebrae
of Lohsania and those of the amioid Osteorhachis (Goodrich,
1930, p. 39). Our inability to identify particular tetrapod pat-
terns in Rhipidistia serves only to emphasize the conclusion (see,
for example, Thomson, 1967; in press) that the characteristic
amphibian patterns must have evolved on the “tetrapod side”
of the fish-amphibian transition, The holospondylous and apsido-
spondylous conditions in Rhipidistia may well have evolved in
accordance with the same mechanical situations to which the
Amphibia responded (aquatic and semi-terrestrial locomotion,
respectively: Thomson, 1967; Parrington, 1968) but resegmenta-
tion of the vertebral components seems to have been a particular
tetrapod characteristic and it led to the development of the new
vertebral patterns characteristic of the tetrapod radiations. It is
particularly interesting that while non-resegmented holospondylous
Rhipidistia are known, presumably every instance of holospondyly
in tetrapods (if resegmentation has occurred) is secondary and
not inherited directly from a rhipidistian ancestor.

CONCLUSIONS

We have described the vertebral structure in two late Paleozoic
rhipidistian fishes. From our study of these forms and of
Eusthenopteron, we conclude that there is no evidence of reseg-
mentation of the vertebrae in the Rhipidistia. Furthermore, we
conclude that the type of vertebral structure seen in Eusthenop-
teron and Ichthyostega bears no direct relationship to that seen
in the mainline of tetrapod evolution, that is, the accessory ele-
ments in the adults of rhipidistians and ichthyostegals are not
true pleurocentra.

ACKNOWLEDGEMENTS

We are grateful to Professors A. S. Romer and B. Patterson for
the loan of material from the collection of the Museum of Com-
parative Zoology, Harvard University, and to Dr. F. R. Parrington
for the earlier loan of material from the Museum of Zoology,

18 POSTILLA

Cambridge University. The figures were prepared by Diane C.
McClure. Our studies have been supported by the National Science
Foundation: Grants GB-4818 and GB-7573X to K.S.T.; and
GB-5104 and GB-7784 to P.P.V.

ABBREVIATIONS

MCZ — Museum of Comparative Zoology, Harvard University

YPM — Peabody Museum, Yale University

UCLA VP — University of California, Los Angeles, Department
of Zoology collections

ac — accessory centrum

c lig —canal for dorsal ligament

da — dorsal ligament

ic — intercalary

hpr — haemal process

1 — dorsal ligament

my — myoseptum

na — neural arch

nc — nerve cord

nch — notochord

pe — principal centrum

rf — facet for rib articulation

LITERATURE CITED

Cope, E. D. 1880. Second contribution to the history of the Vertebrata
of the Permian formation of Texas. Trans. Am. Phil. Soc. 19: 38-58.

Goodrich, E. S. 1930 (Reprinted 1958). Studies on the structure and devel-
opment of vertebrates. Dover Publications Ltd., New York.

Hussakof, L. 1911. The Permian fishes of North America. Carnegie Inst.
Publ. 146: 153-178.

Jarvik, E. 1952. On the fish-like tail in the ichthyostegid stegocephalians.
Meddr. Grgnland 114: 5-87.

Panchen, A. L. 1967. The homologies of the labyrinthodont centrum.
Evolution 21: 24-33.

Parrington, F. R. 1968. The vertebrae of early tetrapods, p. 269-279. Jn
J-P. Lehman (ed). Problemes actuels de paléontologie. Centre Nat.
Rech. Scient., Paris.

Romer, A. S. 1958. The Texas Permian red-beds, p. 157-179. In T. S..
Westoll (ed.) Studies on fossil vertebrates. Univ. of London Press,
London.

1964. The skeleton of the Lower Carboniferous labyrinthodont
Pholidogaster pisciformis. Bull. Mus. Comp. Zool. Harvard 131:
129-159.

RHIPIDISTIAN VERTEBRAE Uy

Schaeffer, B. 1967. Osteichthyan vertebrae, p. 185-196. /n C. Patterson and
P. H. Greenwood, |[ed.| Fossil vertebrates. Linnean Society of
London.

Thomson, K. S. 1964. Revised generic diagnoses of the fossil fishes
Megalichthys and Ectosteorhachis (family Osteolepidae). Bull. Mus.
Comp. Zool. (Harvard) 131: 283-311.

1967. Notes on the relationships of the rhipidistian fishes and
the ancestry of the tetrapods. Jour. Paleont. 41: 660-674.

1968. A new Devonian fish (Crossopterygii: Rhipidistia) con-
sidered in relation to the origin of the Amphibia. Postilla no. 124:
1-13.

[In press]. The biology of lobe-finned fishes. Biol. Rev.

Vaughn, P. P. 1962. Vertebrates from the Halgaito tongue of the Cutler
Formation, Permian of San Juan County, Utah. Jour. Paleont. 36:
529-539.

Williams, E. E. 1959. Gadow’s arcualia and the development of tetrapod
vertebrae. Quart. Rev. Biol. 34: 1-32.

REVIEW

STYLE

FORM

nitlLe

ABSTRACT

NOMENCLATURE

ILLUSTRATIONS

FOOTNOTES

TABLES

REFERENCES

JTHOR’S COPIES

PROOF

COPYRIGHT

INFORMATION FOR AUTHORS

The Publications Committee of the Peabody Museum of Natural
History reviews and approves manuscripts for publication. Papers
will be published in approximately the order in which they are
accepted; delays may result if manuscript or illustrations are not in
proper form. To facilitate review, the original and one carbon or
xerox copy of the typescript and figures should be submitted. The
author should keep a copy.

Authors of biological papers should follow the Style Manual for
Biological Journals, Second Edition (Amer. Inst. Biol. Sci.). Authors
of paleontological manuscripts may choose to follow the Sugges-
tions to Authors of the Reports of the U.S. Geological Survey,
Fifth Edition (U.S. Govt. Printing Office).

Maximum size is 80 printed pages including illustrations (= about
100 manuscript pages including illustrations). Manuscripts must be
typewritten, with wide margins, on one side of good quality
8'42 x 11” paper. Double space everything. Do not underline any-
thing except genera and species. The editors reserve the right to
adjust style and form for conformity.

Should be precise and short. Title should include pertinent key
words which will facilitate computerized listings. Names of new
taxa are not to be given in the title.

The paper must begin with an abstract. Authors must submit com-
pleted BioAbstract forms; these can be obtained from the Postilla
editors in advance of submission of the manuscripts.

Follow the International Codes of Zoological and Botanical Nomen-
clature.

Must be planned for reduction to 4 x 642” (to allow for running
head and two-line caption). If illustration must go sideways on
page, reduction should be to 334 x 634”. All illustrations should be
called “Figures” and numbered in arabic, with letters for parts
within one page. It is the author’s responsibility to see that illustra-
tions are properly lettered and mounted. Captions should be typed
double-spaced on a separate page.

Should not be used, with rare exceptions. If unavoidable, type
double-spaced on a separate page.

Should be numbered in arabic. Each must be typed on a separate
page. Horizontal rules should be drawn lightly in pencil; vertical
rules must not be used. Tables are expensive to set and correct;
cost may be lowered and errors prevented if author submits tables
typed with electric typewriter for photographic reproduction.

The style manuals mentioned above must be followed for form and
for abbreviations of periodicals. Double space.

Each author receives 50 free copies of his Postilla. Additional copies
may be ordered at cost by author when he returns galley proof.
All copies have covers.

Author receives galley proof and manuscript for checking printer’s
errors, but extensive revision cannot be made on the galley proof.
Corrected galley proof and manuscript must be returned to editors
within seven days.

Any issue of Postilla will be copyrighted by Peabody Museum of
Natural History only if its author specifically requests it.

it

WLAN

3 2044 066 3

BOUND DEC 1972

ped eB:

“Ye

: Eee

7
Ps toe ee et
aA :

ae ae ER eee

Apitas<0.558 eat

rfid te me | te
IRAP te nim ing