Ci

LIBRARY OF THE

UNIVERSITY OF ILLINOIS

AT URBANA-CHAMPAIGN

550.5

Fl
V. 17

k-lifci!

UNIVERSITY OF

ILLINOIS LIBRARY

AT URBANA-CHAMPAIGN

GEaOGY

(zr &. e i <

*n

)EVONIAN LUNGFISHES

OMING, UTAH, AND IDAHO

ROBERT H. DENISON

****

^DIANA: GEOLOGY
•LUME 17, NUMBER 4

Published by
3EUM OF NATURAL HISTORY
APRIL 24, 1968

i)»:ttisity of illiitois
JUL 111968

•V>

EARLY DEVONIAN LUNGFISHES

FROM WYOMING, UTAH, AND IDAHO

ROBERT H. DENISON

Curator of Fossil Fishes

FEILDIANA: GEOLOGY

VOLUME 17, NUMBER 4

Published by

FIELD MUSEUM OF NATURAL HISTORY

APRIL 24, 1968

Library of Congress Catalog Card Number: 68-26373

PRINTED IN THE UNITED STATES OF AMERICA
BY FIELD MUSEUM PRESS

I. A Lungfish from the Beartooth Butte Formation
of Wyoming

Fragments of Early Devonian lungfishes have been collected on
a number of Field Museum expeditions, some as early as 1949. With
the exception of a few scales, plates, and a single tooth plate from the
Water Canyon formation of Utah and Idaho, all of this material has
come from two localities in Wyoming. They are: 1) Beartooth Butte
in Park County; and 2) Cottonwood Canyon, east of Lovell, NE %,
sec. 4, T 56 N, R 93W, Bighorn County. The first is the type local-
ity of the Beartooth Butte formation, and has previously yielded
plants, a few eurypterids, and many fishes, including a poorly pre-
served and indeterminable lungfish specimen which was doubtfully
listed by Bryant (1932, p. 254) as a skull of Dipterus (Princeton Uni-
versity 13503). A variable series of sediments at the Cottonwood
Canyon locality has also been referred to the Beartooth Butte for-
mation by Sandberg (1961, p. 1305). It contains a considerable
flora and fauna, similar to those of Beartooth Butte, but as yet un-
described except for Cardipeltis (Denison, 1966). The rocks at these
two localities are believed to be approximately the same age, and
have been correlated (Denison, 1958, p. 500; 1964, p. 451) with the
Early or perhaps Middle Siegenian of the western European section.
If this is correct, the Beartooth Butte formation lungfishes are the
earliest ones known, predating the German Dipnorhynchus lehmanni
from the Late Siegenian Hunsruckschiefer (Lehmann and Westoll,
1952), and D. sussmilchi from the Emsian or Late Siegenian of Aus-
tralia (Philip and Pedder, 1964). A new lungfish from northern
Canada, Melanognathus, recently described by Jarvik (1967, pp. 167-
170), is doubtfully dated as late Lower Devonian; however, since it
is associated with paleoniscoid scales, it is very probably younger.

It is interesting and somewhat surprising that both Dipnorhyn-
chus lehmanni and D. sussmilchi (Hills, 1958, pp. 87-88) occur in
undoubted marine rocks. This is not true of the Beartooth Butte
lungfish, since at Beartooth Butte and Cottonwood Canyon defi-
nitely marine invertebrates are absent, and deposition is in channels

353

354 FIELDIANA: GEOLOGY, VOLUME 17

cut into the underlying Ordovician limestones. On the other hand,
it is not possible to conclude definitely that these are deposits of
freshwater streams. One reason for this is that in Utah many of the
same vertebrates occur in the Water Canyon formation, which is a
widespread and at least in part a marine deposit. Dorf (1934,
pp. 735-736) concluded that the channel-fill at Beartooth Butte
was formed in fresh or brackish water under estuarine conditions in
a drowned river valley. Several years ago I concluded (Denison,
1956, p. 414) that it might have been formed in the arm of a marine
bay. Sandberg (1961, p. 1308) suggested that the Beartooth Butte
formation was deposited in many different continental and shallow-
water, marginal marine environments, the latter related to a sea that
invaded Wyoming from the west. When the study of the fauna and
flora of Cottonwood Canyon is completed, it may be possible to come
to more definite paleoecological conclusions. All that can be said
now is that the varied lithology (Denison, 1966, p. 90) indicates depo-
sition in a great variety of conditions, and it is possible that some
were on the margins of the sea and that others were in freshwater
streams and pools.

The fragmentary material from the Water Canyon formation will
be described at the end of this paper (p. 407). All of the identifiable
specimens of lungfishes from the Beartooth Butte formation are re-
ferred to a single species, which is distinguished from previously
described Dipnoi as the new genus Uranolophus. All material is in
Field Museum collections, except for one specimen in the collection
of Princeton University.

Uranolophus1 new genus

Type-species. — Uranolophus wyomingensis, new species.

Diagnosis. — A pineal foramen is absent. There is a single bone D
in the pineal region of the skull roof. The prepineal region of the
skull roof is relatively short (prepineal length /total length of skull
roof is 0.35) . The supratemporal region of the skull roof is relatively
long (bones X - Yj - Y2 occupy about one-half of the total length of
the skull roof). Pterygoid and prearticular tooth plates are absent;
the margin of each pterygoid and prearticular is crowned with a
"tooth ridge," which is a continuous, irregular crest of dentine with
numerous side ridges.

1 From ovpavos, palate; and Xo^os, ridge.

355

356 FIELDIANA: GEOLOGY, VOLUME 17

Uranolophus wyomingensis new species

Type. — PF 3874, a largely articulated fish lacking the posterior
and ventral portions of the body (figs. 1, 2, 4, 19, 21), from the Early
Devonian, Beartooth Butte formation, Cottonwood Canyon, east of
Lovell, Bighorn County, Wyoming.

Referred specimens. — From the Cottonwood Canyon locality: in-
complete skull and jaws, PF 3795; skulls and partial skulls, PF 3792
(figs. 10, 15C-D), 3794, 3805 (figs. 6, 8, 15A-B), 3816 (fig. 5), 3817;
snouts, PF 3843, 3847-8, 3860, 3868 (fig. 7) ; palates, PF 3793 (fig.
9A), 3796 (fig. 9B), 3841; pterygoids, PF 3801, 3803, 3807, 3844,
3867, 5524; operculars, PF 3833-5 (fig. 11A), 3839 (fig. 11B), 3852,
3870; probable suboperculars, PF 3836 (fig. 11C), 3837, 3840, 5518,
5522; gulars PF 3838 (fig. 11F), 3866 (fig. HE), 5519; ceratohyals,
PF5514 (fig. 16), 5520; lower jaws, PF 3797 (fig. 13A), PF 3798
(figs. 13B, 14A), 3804, 3808, 3845, 3849 (fig. 22B), 3850, 3855, 3858,
3861, 3863, 3869, 5551 ; prearticulars, PF 3802, 3857, 3864 (fig. 15E) ;
associated lower jaws, opercular, subopercular, ceratohyal, and scales,
PF 3862 (figs. 11D, 20F) ; cleithra and clavicles, PF 3851 (fig. 17A),
5364; cleithra, PF 3846 (fig. 17B), PF 3856 (fig. 17C), 3859, 4333;
?interclavicle, PF 3865 (fig. 17D) ; flank scales, PF 5542, 5543, 5547
(fig. 20C), 5548 (fig. 20B), 5550; posterior flank scales, PF 5541
(figs. 24-25) ; median dorsal scales, PF 5544-5 (figs. 20D-E), 5549;
various associated scales, PF 5546 (fig. 20A) .

From Beartooth Butte, Park County, Wyoming: skull, PF 1427
(fig. 3A); lower jaws, PF3318; associated scales, PF 3319; flank
scales, PF 140, 275, 278-280; median dorsal scale, PF 5536; also in
the Princeton University collection an impression of an operculum,
PU 13658, originally figured by Ruedemann (1934, p. 3) as a coxal
joint of Pterygotus.

Diagnosis. — As for genus.

Discussion. — Uranolophus is similar in many respects to the typ-
ical Dipnorhynchus (D. sussmilchi) , but differs strikingly in the de-
velopment of its "teeth." This in itself would appear to be adequate
to distinguish it generically, but is reinforced by the consistent ab-
sence of a pineal foramen, and by proportional differences of evolu-
tionary significance. The generic affinities of Dipnorhynchus lehmanni
are uncertain; in its cranial roof it resembles D. sussmilchi, but its
teeth, though poorly known, appear to be quite differently developed.

Dermal bones of the cranial roof.- — The nomenclature of the dermal
bones of the dipnoan skull roof has been discussed recently at some

DENISON: EARLY DEVONIAN LUNGFISHES

357

Fig. 2. Skull and lower jaws of type specimen of Uranolophus wyomingensis,
n. sp., PF3874 (X 9/10).

length by White (1965, pp. 7-31), who uses a modification of the
alphabetical notation introduced by Forster-Cooper in 1937 for Dip-
terus valenciennesi. In this paper White's notation for Dipterus will be
followed with minor exceptions, of which the most important is that
the bone he labels I2 is called simply I. White reasoned that there is
evidence that there was once an Ix behind I2, and indeed it is occa-
sionally preserved in Dipterus and Rhinodipterus. But this bone lay
in the transverse extrascapular series and would better be called H
than I (though this is not the H of Forster-Cooper nor of Westoll's
1949 review).

358

359

360 FIELDIANA: GEOLOGY, VOLUME 17

Those who have compared the skull of Dipnorhynchus with that
of Dipterus have come to varied conclusions regarding the homologies
of the important cranial roofing bones. My own interpretation of
Dipnorhynchus and Uranolophus (fig. 3) differs in details or in im-
portant respects from earlier ones. It is based fundamentally on the
great similarity in pattern between the eight bones forming a tri-
angular area in the relatively stable postero-median part of the skull
roof in these three genera. The bone pattern differs only in that
B in Dipterus has pushed back between I -I to reach the posterior
margin of the skull roof. The lateral-line system differs only in that
the posterior end of the supraorbital canal of Dipnorhynchus and
Uranolophus has become a pit line in Dipterus and has extended
postero-medially onto B. According to this interpretation, the extra-
scapular series, Z-A-Z (or perhaps, Z-H-A-H-Z), was loosely at-
tached to the posterior edge of the skull roof in Dipterus, where it is
sometimes found, and in Dipnorhynchus and Uranolophus, where it
has not yet been discovered.

In the interpretation favored by White (1965, fig. 33; 1966, fig. 4),
the bones at the posterior edge of the skull roof of Dipnorhynchus are
a distinct series (j8-a-a-/3), lost in Dipterus and in other lungfishes.
This hypothetical series of bones introduces an unnecessary compli-
cation into an otherwise simple picture. A second interpretation sug-
gested by White (1965, fig. 32) identifies as I -I the bones that are
here labelled J-J. This makes it necessary to homologize the paired
bones at the posterior margin of the skull with A-A, which is unlikely
because A is a loosely attached median element in Dipterus. Westoll
(1949, p. 141) identified the bones in Dipnorhynchus here called C-C
as "B2-B2." This homology was based on the spatial relations of
"B2-B2" to the lateral-line bones on either side, but it has little valid-
ity because Westoll's identifications of these lateral-line bones are,
in my opinion, incorrect. Campbell (1965) used Westoll's homolo-
gies of Dipnorhynchus cranial bones, but implied reservations as to
their acceptance. The homolgies of Hills (1941, pp. 50-54) are diffi-
cult to understand because he makes his comparisons with osteolepids
rather than directly with Dipterus, but as far as the posterior triangle
under discussion is concerned, his homologies are apparently the
same as my own. Romer's (1936, fig. 1A) homologies in this area
of the skull are the same as my own.

If one accepts the homologies of I-I, B, J-J, C-C, and D, as here
proposed, the lateral bones of the posterior half of the skull roof are
easily identified. The more medial series of lateral-line bones, J-K-

DENISON: EARLY DEVONIAN LUNGFISHES 361

Lx-Lq, have exactly the same relations to B and C in Dipterus, Dip-
norhynchus, and Uranolophus. It is only in the variable more anterior
members of this series that precise homologies cannot be attempted.
The genera differ in that the posterior continuation of the supra-
orbital canal onto K and J of Dipnorhynchus and Uranolophus has
become in part the anterior pit line of Dipterus. The lateral or supra-
temporal series of lateral-line bones shows differences, yet can be
homologized simply. The bone at the postero-lateral corner of the
skull roof is identified as Y2 in all three genera. This implies that
the extra-scapular bones (Z-A-Z), if evolved as distinct elements,
lay behind the skull roof. Yx consistently overlies the fused otic
process of the palato-quadrate. The bone in which the main lateral
line canal turns down to become the infraorbital canal is identified
as X. White (1965, figs. 20, 21, 23) has identified this as Yi in some
specimens of Dipterus, but this homology is open to question. The
position of this sensory canal is known definitely in Dipterus valen-
ciennesi and Uranolophus wyomingensis, and its probable course is
known in Dipnorhynchus sussmilchi. The Y2-Yi-X series, as so iden-
tified, is relatively long in Uranolophus, occupying nearly half of the
total length of the cranial roof. This is presumably the primitive
condition in lungfishes. In Dipnorhynchus sussmilchi and D. leh-
manni, it occupies about 44 per cent of the cranial roof length, while
in Dipterus this series is much shorter, a result, no doubt, of the re-
duction of the subjacent cheek region. In spite of this shortening,
the relations of bones Y2-Yi-X to I-J-K remain much the same,
though in Dipterus X no longer contacts Ilt and Y2 no longer con-
tacts J, as is usual in Dipnorhynchus and Uranolophus. One result
of the shortening of the supratemporal part of the skull roof is that
the supraorbital canal in K has approached the infraorbital canal in
X, and a commisure has formed between them in Dipterus.

None of the previous students of Dipnorhynchus has come to sim-
ilar conclusions regarding the homologies of the lateral series of lateral-
line bones. Hills (1941, fig. 6) correctly identified X, but added an
additional extrascapular bone at the postero-lateral corner of the
skull roof, which modified his homologies of Y2 and Yx. Romer
(1936, fig. 1A) followed Hills in adding an extra bone, and this led
him to place the equivalent of X too far posteriorly. Westoll (in
Lehmann and Westoll, 1952, fig. 5) and White (1965, figs. 32, 33)
have, in my opinion, placed X too far forward, and, as a result, have
misinterpreted the bones behind it in this series.

362

FIELD IANA: GEOLOGY, VOLUME 17

soc

Fig. 4. Skull of type specimen of Uranolophus wyomingensis, n. sp., PF 3874
(X 3/2). ar, articulatory area for lower jaw; co, circumorbital or cheek plates;
ifc, pores of infraorbital sensory canal; 11, pores of main lateral line; occ, pores of
occipital commissure; pi, pit line; pq, palatoquadrate; pr, process on dorsal face
of pterygoid; pt, pterygoid; soc, pores of supraorbital sensory canal; tr, tooth ridge.

As is usual in early lungfishes, there is considerable variation in
the pattern of the cranial bones of Dipnorhynchus and Uranolophus.
The following have been noted in the central and posterior parts of
the skull roof:

C-C are partly fused in the type of D. sussmilchi and in one speci-
men of Uranolophus wyomingensis (PF 3795) . This is known to oc-
cur in Dipterus also (Westoll, 1949, fig. 3E; White, 1965, fig. 7).

In D. lehmanni and in one specimen of U. wyomingensis (PF 1427,
fig. 3A), K and Lx are separate on one side, but represented by only
one bone on the other. In D. lehmanni Y1 and Y2 are separate on
the right, a single bone on the left. As shown by Parrington (1949)

DENISON: EARLY DEVONIAN LUNGFISHES

363

Fig. 5. Incomplete cranial roof of Uranolophus wyomingensis, n. sp., PF 3816
(X 1), showing Westoll lines and areas of cosmine resorption (stippled). Where
dermal bones have been lost, the impressions of their inner surfaces show radiating
lines.

and White (1965, p. 12) these variations may be interpreted as fusions,
or as the loss of one bone and invasion by the other. They are com-
mon in the lateral-line series of Dipterus (Westoll, 1949, p. 132).

In a number of specimens there are supernumerary bones, usually
absent in other lungfishes; they may be interpreted as relics of an an-
cestral stage when the skull roof was covered by a mosaic of small
bones. In the L region of the lateral-line series of the type of Dip-
norhynchus sussmilchi there are three bones on the left, and the usual
two occupying the equivalent area on the right. In the type speci-
men of Uranolophus wyomingensis (PF 3874, fig. 4) there are four

364 FIELDIANA: GEOLOGY, VOLUME 17

instead of three bones in the supratemporal series, two of which oc-
cupy about the same space as the usual Yi. On the left side of a
specimen of U. wyomingensis (PF 3816, fig. 5) J is represented by
two bones lying side by side. A comparable situation has been re-
corded in a specimen of Soederberghia (Lehman, 1959, fig. 12) where
there is a small bone medial to one J. Also, in Dipterus and Scaume-
nacia (Westoll, 1949, figs. 2B, 6B) J may be represented by two
bones, one behind the other. Yx is represented by two bones lying
side by side on the right side of the type of D. sussmilchi. Westoll
(1949, p. 145, fig. 6D) found a similar situation in Scaumenacia,
and considered the more lateral bone to be a "separate prespiracu-
lar bone."

There are variations in the mutual relations of bones. Thus Yx
and I are not in contact in D. sussmilchi and on one side of three
specimens of U. wyomingensis (PF 1427, fig. 3A; 3816, fig. 5; 3874,
fig. 4) . X and J do not contact in the type of D. sussmilchi and on
one side of two specimens of U. wyomingensis (PF 1427, fig. 3A; 3874,
fig. 4).

The differences mentioned above are clearly individual variations,
but two other differences in the posterior triangle of the cranial roof
are considered to be of generic significance. One is the absence in
Uranolophus wyomingensis of a pineal foramen, which is present in
Dipnorhynchus. The other is the presence in U. wyomingensis of a
single and sometimes large D bone in the pineal region, while D. leh-
manni (fig. 3B) has two bones in this area, and D. sussmilchi (fig. 3C)
shows evidence of three or four.

The anterior part of the skull roof has little or no ossification in
some Late Devonian and in known post-Devonian lungfishes. In
Dipterus valenciennesi an ossified, cosmine-covered snout is usually
absent in small specimens and is invariably present in large ones;
according to White (1965, p. 7), this may have been because it formed
only at a certain size or age. When it is present in Dipterus, it is
always a single unit with no evidence of the mosaic of small, separate
ossifications that are believed to have characterized the ancestral
dipnoan (Westoll, 1949, fig. 10A). This mosaic is actually known
in only three genera — one is the Upper Devonian Soederberghia (Leh-
man, 1959, figs. 2-4) where the anterior part of the skull roof is
covered with a large but variable number of bones lacking any cos-
mine coating. The others are Dipnorhynchus and Uranolophus where
separate ossifications are well displayed only in an occasional indi-
vidual. One of the best is the specimen of D. sussmilchi described

DENISON: EARLY DEVONIAN LUNGFISHES

365

by Campbell (1965, fig. 1) where there are about seven lateral-line
bones on each side anterior to L2, five or more marginal bones on
each side, and 29 anamestic bones between the lateral-line bones (in-
cluding the bones in the pineal area). Only the thick, anterior mar-

Fig. 6. Anterior part of skull of Uranolophus wyomingensis, n. sp., showing
mosaic of small plates near snout, PF 3805 ( X 1).

gin of the snout does not show subdivision into individual bones.
One specimen of Uranolophus wyomingensis (PF 3805, fig. 6) has the
entire anterior part of the skull roof (with the exception of a much-
enlarged D) subdivided into what appear to be small distinct ossifi-
cations; here the subdivision extends even onto the anterior margin
of the snout. An isolated snout (PF 3868, fig. 7A) shows a similar
tendency toward apparent subdivision along the rostral margin. In

366

FIELDIANA: GEOLOGY, VOLUME 17

■SB*:

Fig. 7. Snout and upper lip of Uranolophus wyomingensis, n. sp., PF 3868
( X 1). A, dorsal view; B, ventral view.

other specimens of Dipnorhynchus the mosaic of small plates is poorly
developed. In the type of D. sussmilchi only the interorbital area
shows individual ossifications. In the type of D. lehmanni the cen-
tral area anterior to the pineal opening shows about 15 small ossifi-
cations. In other specimens of Uranolophus wyomingensis two, one,
or no distinct ossifications are apparent. The tendency toward sub-
division of the snout is presumably a relic of the ancestral condition,
but in Dipnorhynchus and Uranolophus it is so variably expressed
that it is of little use as a systematic character.

Perhaps the most striking differences in the cranial roofs of Dipno-
rhynchus, Uranolophus, and Dipterus are those of proportions. This
is well shown by Figure 3 where the skulls are all drawn with the
same post-pineal lengths (measured from the pineal opening or the
center of bone D to the posterior edge of bones I in Dipnorhynchus
and Uranolophus or bone B in Dipterus). The most important dif-
ferences are:

1. The prepineal part of the skull is very short in Uranolophus,
slightly longer in Dipnorhynchus, and much longer in Dipterus. The

DENISON: EARLY DEVONIAN LUNGFISHES 367

prepineal length is about 35 per cent of the total skull length (exclud-
ing any extrascapular bones) in Uranolophus wyomingensis, 42 per
cent in Dipnorhynchus lehmanni, and D. sussmilchi, while it is ap-
proximately 50 per cent in Dipterus valenciennesi. There is no sug-
gestion in these genera that the growth of the snout is allometric,
and that it became relatively longer in larger individuals, as has been
shown to be the case in the long-snouted Soederberghia by Lehman
(1959, pp. 35-37).

2. The postero-lateral or supratemporal part of the skull roof,
including bones X-Y!-Y2, is long in Uranolophus and in Dipnorhyn-
chus, and considerably shortened in Dipterus. The length of this
series of bones is approximately half of the total skull length in
Uranolophus wyomingensis, 44 per cent in Dipnorhynchus lehmanni
and D. sussmilchi, and only one-third of the skull length in Dipterus.
The shortening of these bones is related to the shortening of the
cheek region.

3. The orbits lie anterior to midlength of the skull roof in Ura-
nolophus and Dipnorhynchus, while they are behind midlength in
Dipterus. This is related both to the shortening of the cheeks and
the lengthening of the prepineal part of the skull.

Westoll (1949, p. 163) claimed that a shortening of the otic region
of the neurocranium was a general evolutionary trend in Dipnoi, and
specifically that the otic region was shorter in Dipterus than in Dip-
norhynchus. However, though the otic region is known to be very
short in Dipterus and Chirodipterus (Save-Soderbergh, 1952, pp. 15-
19), it is as yet unknown in Uranolophus or Dipnorhynchus, and its
extent cannot be determined easily from the extent of the cranial
roofing bones. In Dipterus, Uranolophus, and Dipnorhynchus the otic
process of the palatoquadrate underlies Ylt and indicates the position
of the anterior part of the otic region to which it is fused. The
position of Yx suggests that the otic region was very short in Urano-
lophus and Dipnorhynchus, also.

Ventral side of the skull. — The dermal shield of the ethmoidal re-
gion turns down at the snout, and then curves postero-ventrally to
form the so-called "upper lip," which is developed in Uranolophus
(figs. 7B, 8, ul) in much the same way as it is in Dipnorhynchus,
Dipterus, Rhinodipterus, Chirodipterus, and Ganorhynchus. Medially,
there is a relatively broad process which forms the anterior border
of the mouth; this is divided in the middle in some specimens of
U. wyomingensis. On either side of this process there is a deep notch
which forms the anterior border of the anterior naris (fig. 8, an) .

368

FIELDIANA: GEOLOGY, VOLUME 17

an U|

Fig. 8. Palate and upper lip of Uranolophus wyomingensis, n. sp., PF 3805
(X 1). an, anterior naris; mr, median ridge; na, position of nasal capsule; PS,
parasphenoid; PT, pterygoid; sa, smooth, untuberculated areas of pterygoids and
parasphenoid; tr, tooth ridge; ul, upper lip; VO, vomer.

The lateral border of each anterior naris is formed by a lateral proc-
ess, which is a continuation of the "upper lip" (="subnasal ridge"
of Jarvik, 1964, p. 44, figs. 12-14). This lies at the lateral corner of
the downturned snout in Uranolophus, but is usually some distance
antero-medial to this corner in Dipterus. The oral edge of the upper
lip in Uranolophus wyomingensis (fig. 7B), including both median and
lateral processes, usually has the dentine developed in sharp ridges,
irregular projections, and tubercles, similar to those on the lateral
edges of the pterygoids. Since this is the only type of "teeth" de-
veloped in this species, it is of interest that they should occur in the

DENISON: EARLY DEVONIAN LUNGFISHES 369

region corresponding to that occupied by the premaxillaries in other
Osteichthyes. In Dipnorhynchus sussmilchi and most Dipterus, this
border is probably smooth and toothless. However, "teeth" have
been reported on the median process of one specimen of Dipterus
valenciennesi (Watson and Day, 1916, p. 33), in D. oervigi (Gross,
1964, pp. 11-14), and in Ganorhynchus woodwardi (Traquair, 1873,
p. 553). Possible vestiges of toothed premaxillaries or maxillaries
have been described in Scaumenacia (Stensio, 1947, p. 140, fig. 32B)
and Phaneropleuron (Watson and Day, 1916, p. 36, fig. 7), in both
of which the snout is otherwise largely unossified. A slender, tooth-
less maxilla has been identified in Dipterus by Watson and Gill (1923,
p. 206, fig. 32).

On the ventral surface of the neurocranium, immediately behind
the upper lip and in front of the palate, there is a median ridge, on
either side of which is a depression forming part of the nasal capsule.
The structure of this region is best known among fossil lungfishes in
Chirodipterus (Save-Soderbergh, 1952, fig. 3) and Ganorhynchus
(Gross, 1965, fig. 1) ; it has been figured in Dipterus (Jarvik, 1954,
fig. 36B), and can be seen in two specimens of Uranolophus wyo-
mingensis (PF 3792, fig. 10B; 3805, fig. 8), though the latter are so
flattened that most details are obscured. The most obvious differ-
ence in Uranolophus is that this region is relatively short. This fact,
together with the corresponding shortness of the prepineal part of
the skull roof, are doubtless related to a comparatively poor develop-
ment of the nasal organs in Uranolophus. The median ridge (fig. 8,
mr) is broad in Uranolophus, as in Dipterus, but is shown by Gross
to be very narrow in Ganorhynchus splendens. In the latter, the roof
of each nasal pit has a recessus anterior leading into the anterior
naris, and a recessus posterior, which was thought by Gross (1965,
p. 117) to correspond in position to the posterior naris. This would
place the posterior naris approximately in the same position as in
Neoceratodus (Gunther, 1872, pi. 34, fig. 3) ; however, Traquair (1878,
pi. 3, fig. 1), Goodrich (1909, fig. 210A), Jarvik (1954, fig. 36B), and
Thomson (1965, p. 222) have given it a more antero-lateral position,
just postero-lateral to the process of the upper lip that forms the
lateral margin of the anterior naris.1 A well-preserved specimen of
Dipterus valenciennesi in Field Museum collections (PF 1293) shows
the anterior and posterior recessi in the same positions as in Gano-
rhynchus, with the posterior one far removed from the position of the

1 Jarvik, 1964, p. 44, has changed his opinion and places the posterior naris as
in recent Dipnoi.

370 FIELDIANA: GEOLOGY, VOLUME 17

posterior naris indicated by Traquair, Goodrich, and Thomson. If
the recessus posterior did actually house the posterior part of the
nasal organ, it is likely that Gross is correct in assuming a position
for the posterior naris well back near the palate and far removed from
the margin of the mouth. Because of flattening, the recessi are not
visible in specimens of Uranolophus, but in all Uranolophus and Dip-
norhynchus the processes that bound the anterior nares laterally lie
at the lateral corners of the upper lip, and there are no notches lateral
to them.

The palate of Uranolophus wyomingensis (figs. 8, 9, 10B) is dis-
tinctive in three respects. First, it is relatively long, reaching from
a point close to the median process of the upper lip at least as far
back as the quadrate, which has a position considerably farther pos-
terior than in Dipterus. Secondly, there are no palatal tooth plates;
the whole ventral surface of the palate is covered with denticles, and
on the lateral margins of the pterygoids the dentine has developed
into an irregular "tooth ridge" or, occasionally, into isolated teeth.
The palate is also distinctive in that the parasphenoid lacks a pos-
terior stem.

The elements of the entire palates of U. wyomingensis are usually
so closely fused that it is impossible to determine the extent of indi-
vidual bones. However, in PF 3805 (fig. 8) and 3796 (fig. 9B), the
former showing the ventral and the latter the dorsal surface of the
palatal bones, the sutures can be determined satisfactorily. These
specimens show that the pterygoids (fig. 8, pt), or more properly the
entopterygoids, are very long and relatively narrow, and taper slightly
and irregularly anteriorly. They are broadest posteriorly where they
expand laterally underneath the quadrates. The posterior part of
the lateral margin of each pterygoid, which underlies the quadrate,
is smooth and concave, but much of the rest of the margin carries a
"tooth ridge" (fig. 8, tr), which will be described further below. An-
teriorly the pterygoids meet or overlie the vomers, and in the midline
behind the vomers the two pterygoids meet in a relatively short
suture. Most of the medial edge of each pterygoid is a suture with
the long parasphenoid. The posterior edges of the pterygoids curve
dorsally and share a smooth, untuberculated area (fig. 8, so) with
the parasphenoid.

The paired vomers are the smallest of the palatal bones. In
PF 3805 (fig. 8, vo) they meet in a short median suture, then curve
backward around the anterior ends of the pterygoids, becoming nar-
rower posteriorly. Their ventral surfaces are tuberculated, but their

Fig. 9. Palates of Uranolophus wyomingensis, n. sp. (X 2/3). A, PF 3793,
ventral surface; P>, PF 3796, dorsal surface.

371

372 FIELDIANA: GEOLOGY, VOLUME 17

lateral margins lack "tooth ridges." The nature of their suture with
the pterygoids is not clear, but it looks in PF 3805 as if the ptery-
goids were underlain by the vomers. This is the relationship shown
by Gross (1964, p. 18, fig. 2B) in Dipterus oervigi between a pterygoid
and a bone identified as a palatine or vomer. The vomers of Dipno-
rhynchus sussmilchi (Campbell, 1965, pi. 91) are similar in general,
forming a blunt, arrow-shaped point at the anterior end of the palate.
In D. lehmanni there is what is described by Westoll (Lehmann and
Westoll, 1952, p. 410) as "a flanged surface, probably for the over-
lapping vomers"; this surface resembles in shape the vomers of U.
wyomingensis. In Dipterus and most other lungfishes, the vomers
are much reduced, perhaps because of the predominant development
of the pterygoid tooth plates. White (1965, pp. 39^41) found no
confirmation of the presence of relatively large vomers such as were
shown in a restoration of Dipterus valenciennesi by Watson and Day
(1916, p. 33, fig. 6). It is interesting that in Fleurentia, where tooth
plates are lacking, the vomers were probably relatively large, as in
Dipnorhynchus and Uranolophus.

The parasphenoid of Uranolophus wyomingensis (fig. 8, PS) is a
long, slender bone, occupying about one-third of the width of the
palate posteriorly, and tapering with some sutural undulations to a
point near the anterior end of the palate. In only two other lung-
fishes is it known to extend as far forward: one of these is Dipno-
rhynchus lehmanni, where Westoll (Lehmann and Westoll, 1952, fig.
4B) shows it as a very narrow plate, though its outline is not known
with any certainty; and the other is Uronemus, where it is shown by
Watson and Gill (1923, fig. 29B) to be very broad. The ventral sur-
face of the parasphenoid of U. wyomingensis is denticulate, except at
the posterior end where there is a wide, smooth area (fig. 8, sa),
shared with the pterygoids, presumably for the overlap of more pos-
terior plates. Since the parasphenoid of Uranolophus completely
lacks the long, posterior stem that is characteristic of other Dipnoi,
it is possible that this area was overlapped by separate parotic plates.
Westoll (Lehmann and Westoll, 1952, p. 415) has already suggested
that separate parotic plates may have been present in Dipnorhyn-
chus, just as they are in crossopterygians ( Jarvik, 1954, figs. 22, 25) .
Parotic plates possibly occur in place in one specimen of Uranolophus
wyomingensis, PF 3792 (fig. 10B), but, unfortunately, this specimen
has lost the posterior end of the parasphenoid and the anterior ends
of the possible parotic plates, though they have been preserved in
part as a plastic cast of their natural mold. What is actually left is

DENISON: EARLY DEVONIAN LUNGFISHES

373

Fig. 10. Nearly complete skull of Uranolophus wyomingensis, PF 3792
(X 3/4). A, dorsal view; B, ventral view.

a broken bone, or more probably a pair of bones, lying at the midline
behind the parasphenoid, and projecting posteriorly well behind the
posterior border of the cranial roof.

The dorsal or cranial surface of the palate is well shown in PF 3796
(fig. 9B). The parasphenoid is largely spongy bone on this surface
and was probably closely applied to the ventral surface of the neuro-
cranium. Centrally, it has an area covered with compact bone, and
in the posterior part of this area is a depression which was probably
occupied by the hypophysis. U. wyomingensis has no hypophysial
canal penetrating to the ventral surface of the parasphenoid, such as
may occur in Dipnorhynchus, Dipterus, and Rhinodipterus (White,
1965, p. 39). The vomers cannot be distinguished as separate ele-
ments in PF 3796, perhaps only because the sutures are obscure, or
perhaps because they are overlain by the pterygoids. The ptery-
goids of PF 3796 and isolated specimens (PF 3807, 3867) are well
defined and largely covered on their dorsal surface by compact bone.
Overlying the "tooth ridge" is a sharp ridge (fig. 4, pr) which rises
above the anterior and posterior ends of the "tooth ridge" into prom-

374 FIELDIANA: GEOLOGY, VOLUME 17

inent processes whose medio-dorsal surfaces lack the compact bone.
These processes are the points of attachment of the pterygoid, but
the nature of this attachment is difficult to ascertain. Similar proc-
esses may occur in Dipnorhynchus sussmilchi (White, 1965, p. 36),
but Save-Soderbergh (1952, figs. 2, 5) shows nothing comparable in
Chirodipterus, where the pterygoids apparently are closely applied to
the ventral surfaces of the palato-quadrate and neurocranium. In
Dipterus White (1965, p. 35, fig. 45) has described two "buttresses"
on the lateral wall of the endocranium, an anterior one formed by the
lateral part of the planum antorbitale, and a posterior one in the an-
terior part of the orbito-temporal region. These buttresses overlie
the tooth plates, just as the dorsal processes of the pterygoids overlie
the ends of the tooth ridges of Uranolophus wyomingensis. They are
assumed to be endocranial ossifications, but it is not possible to deter-
mine their boundary with the dermal bones. In Neoceratodus, where
the endocranium is completely cartilaginous, each pterygoid tooth
plate is braced by processes of dermal bones. One is a process de-
scending from the largest paired cranial roofing bone; this splits into
two processes which extend to the pterygoid, one near the anterior
part of the tooth plate, and the other near its posterior part. The
pterygoid has an ascending process which rises just anterior to the
middle of the tooth plate and inter-fingers with the descending proc-
ess. These processes appear to be parts of the dermal bones to which
they are attached, and they lie against the lateral wall of the neuro-
cranium just anterior to the orbital region. The relationships of the
processes that brace the tooth plates of Neoceratodus are very similar
to the "buttresses" of Dipterus, except that they are formed by der-
mal bones in the former, and probably by endocranial bones in the
latter. It is possible that as ossification of the cranium was reduced
during lungfish evolution, the dermal pterygoids and cranial roofing
bones enlarged their processes to take over the bracing function that
was earlier performed largely by endocranial ossifications.

Cheek, opercular and gular plates. — In addition to several isolated
opercular, gular, and probable cheek plates of Uranolophus wyoming-
ensis, the partially articulated type specimen (PF 3874, fig. 2) shows
some of them, but, unfortunately, scattered from their natural posi-
tions so that it is not possible to determine their precise arrangement.
This is particularly true of the circumorbital and cheek plates, but
while this region cannot be reconstructed, it is possible to demon-
strate that the cheek was relatively much longer in Uranolophus and
Dipnorhynchus than in other lungfishes, as already indicated by Hills
(1933, p. 641) and Westoll (1949, p. 142, fig. 5A). This is shown by

Fig. 11. Operculars and gulars of Uranolophus wyomingensis ( X 2/3). A, in-
ner face of opercular, PF 3835; B, outer face of opercular, PF 3839; C, second
subopercular, PF 3836; D, first subopercular, PF 3862; E, antero-lateral gular,
PF 3866; F, principal gular, PF 3838.

375

376 FIELDIANA: GEOLOGY, VOLUME 17

the position of the fused quadrate, which extends ventrally and a
little anteriorly from the level of the posterior part of Yx. The oper-
cular must have attached immediately behind this suspensorium, and
the relatively wide space between it and the orbit was occupied in
life probably by a number of relatively large cheek and circumorbital
plates. Soederberghia and Rhynchodipterus retain a moderately long
cheek with relatively large plates, but in other adequately known
Dipnoi the cheek is much shortened, and the plates small and reduced
in number. This has resulted from the forward shift of the jaw artic-
ulation and from the more posterior position of the orbits.

The opercular (figs. 11 A-B) is the biggest plate on the skull,
attaining a length of 100 mm. and a height of 73 mm. in the largest
individual in the collection (PF3862). Its shape is variable, and
changes with age, as is indicated by growth lines, which are some-
times prominent on the inner surface (fig. 11 A) . The center of growth
is near the antero-dorsal corner, which in life must have occupied a
position at the edge of the skull roof near the boundary of Yx and Y2.
The dorsal edge lay against Y2, and also against scales or bone Z,
behind Y2. The anterior edge is usually quite straight, and presum-
ably attached at or close behind the quadrate. The posterior and
ventral edges are rounded, and in PF 3870 there is a deep notch in
the ventral edge which, as is indicated by growth lines, appeared
quite late in development. The attachment of the operculum is quite
similar to that of Dipterus, where often there is a laterally projecting
angle on Yi that lay against the antero-dorsal corner of the opercular.

Two suboperculars may be present in Dipterus (Graham-Smith
and Westoll, 1937, fig. 2a) and Scaumenacia (Stensio, 1947, fig. 32B),
and it is thought that this is true in Uranolophus wyomingensis, also.
If these bones are correctly identified and oriented (fig. 12), the sub-
operculars are relatively much larger in Uranolophus, even larger
than restored by Westoll (1949, fig. 5A) in Dipnorhynchus. Sub-
opercular 1 (fig. 11D) is a bone with rather uniform width; dorsally
it has a relatively wide area overlapped by the operculum, and ante-
riorly a small area overlapped by a cheek plate or postmandibular
bone. Subopercular 2 (fig. 11C) is narrow anteriorly and wide poste-
riorly; dorsally it has a wide area overlapped by subopercular 1, ven-
trally a wide area overlapped by the principal gular, and anteriorly
a very small overlap area for postmandibular bone (fig. 12, oa). A
single subopercular is retained in some other fossil lungfishes and in
Neoceratodus, but it is never as largely developed as the subopercu-
lars of Uranolophus.

DENISON: EARLY DEVONIAN LUNGFISHES

377

Fig. 12. Attempted restoration of right operculars and principal gular of
Uranolophus wyomingensis, n.sp. OP, opercular, based on PF 3835; PG, principal,
gular, restored from PF 3838; SOPu first subopercular, based on PF 3862; SOPi,
second subopercular, based on PF 3836 (SOP, about X 1/2; others slightly larger);
oa, overlap areas for cheek plates or postmandibular bones; pi, pit line.

Gular plates are preserved in the partially disarticulated head of
the type of Uranolophus wyomingensis (PF 3874), but they are buried
under the lower jaws, ceratohyal, and other elements, so their num-
ber and shape are not determinable. A few large, isolated plates can
be tentatively identified as gulars on the evidence of the conditions in
Dipterus, as figured by Watson and Gill (1923, fig. 33) and Graham -
Smith and Westoll (1937, fig. 2a). The smaller ones (PF 3866,
fig. HE; 5519) are believed to be paired antero-lateral gulars; they
have an anterior area that may have been overlapped by an antero-
median gular, and a median area that may have been overlapped by
the antero-lateral gular of the opposite side. Their antero-lateral
edges presumably lay against the lower jaws unless there were sub-
mandibular elements in between, such as Jarvik (1963, p. 13) has iden-
tified in crossopterygians. Their posterior edges probably overlapped

378 FIELDIANA: GEOLOGY, VOLUME 17

the principal gular and a median gular. One plate (PF 3838, fig. 11F)
has been tentatively identified as postero-lateral or principal gular.
It is one of the largest known plates of U. wyomingensis, with a pre-
served length of 82 mm. It differs from opercular plates in that its
center of growth, as shown by growth lines, is near the center of its
presumed anterior end, and also in its relatively longer form. Near
one side of the exposed inner surface there are some short ridges
which possibly indicate the pit lines that occur on the principal gulars
of other Dipnoi. As placed in the restoration (fig. 12), this gular
overlaps the ventral edge of subopercular 2; however, neither the
lateral nor medial edge is completely preserved

Lower jaws. — The lower jaws of Uranolophus wyomingensis are
similar in general features to those of Dipterus except that they lack
ridged tooth plates and are relatively longer. The length of the lower
jaws (measured from the midline anteriorly to the articular) is about
78 per cent of the median length of the cranial roof (excluding extra-
scapulars) in the articulated type specimen of U. wyomingensis
(PF 3874) . In Dipterus valenciennesi the same ratio, estimated from
PF 1293, is about 67 per cent. The relatively greater length of the
jaws of Uranolophus is related to the more posterior position of their
articulation.

The names that have been used for the dermal bones of the lower
jaw have been based on assumed homologies with the jaw bones of
other fishes, particularly rhipidistians. In view of the situation of
the cranial roof, where strict homologies with rhipidistian bones are
clearly impossible in most cases, it may be questioned whether it
serves any useful purpose to attempt homologies of the lower jaw
elements. All of the lower jaws from Cottonwood Canyon are cos-
mine covered ventrally and laterally, and so usually no sutures are
visible. However, on the ventral surface of the symphysial region
one large specimen (PF 3797, fig. 13A) shows grooves in the cosmine
which resemble incomplete sutures; these divide this area partially
into what appears to be a mosaic of small plates. This recalls the
snout of PF 3805, which is similarly divided, and suggests that the
ancestral dipnoan may have had a lower jaw covered with a mosaic
of small dermal bones. If so, they may have enlarged or fused during
evolution in a manner different from that in rhipidistians, making
strict homologies impossible. Since non-committal names are not
available, the names used by Watson and Gill (1923) will be em-
ployed, but with reservations.

DENISON: EARLY DEVONIAN LUNGFISHES

379

Fig. 13. Lower jaws of Uranolophus wyomingensis, n. sp. ( X 3/4). A, ventral
view of PF 3797; B, dorsal view of PF 3798.

In a recent paper Jarvik (1967) has introduced new notations for
the dermal bones of the dipnoan lower jaw. These are meritorious in
as far as the notations imply no homologies with the bones of other
fishes, but as Jarvik used them within the lungfishes, they indicate
his assumption of bone fusions, for which evidence is largely lacking.
The notations are based in large part on his new Devonian genus,
Melanognathus, where four distinct sensory canal bones are preserved
on each side, and the presence of a fifth one is assumed. Since this
arrangement is unique among lungfishes, it is hardly proper to assume
that it is primitive, as did Jarvik, or to use it as a basis for compari-
son of other genera.

380 FIELDIANA: GEOLOGY, VOLUME 17

Around the anterior end of the jaw, the dentaries form an erect,
nearly rectangular margin, which is entirely surmounted by a "tooth
ridge" (fig. 13B). Instead of being shaped like a smoothly-rounded
semi-circle, as in Dipterus and Holodipterus, they have a nearly
straight, transverse, anterior edge, which is separated by sharp curves
from the nearly straight, antero-posteriorly directed, lateral rims.
The dentaries are well separated from the prearticulars, except pos-
teriorly where they meet them in well-marked sutures (fig. 14A).
In PF 3797 (but not in PF 3798) the dentaries have small, triangu-
lar, denticulated surfaces between these sutures and the tooth ridges.
Anteriorly, the dentaries are cosmine covered and pass without vis-
ible sutures into the ventral bones of the symphysial region. Later-
ally, each dentary overlies on the side of the jaw an elongate fossa
(fig. 14A, jo) which was probably occupied in life by a fold of the
lower lip (Jarvik, 1964, p. 44).

On the ventral and lateral surfaces of the jaw of specimens from
Cottonwood Canyon (fig. 13A) individual bones usually cannot be
distinguished because of the cosmine coating. One incomplete lower
jaw from Beartooth Butte (PF 3318) has lost much of its bone, but
does show probable sutures dividing the lower and lateral surfaces
into three pairs of bones very similar to the splenials, postsplenials,
and angulars figured in Dipterus by Watson and Gill (1923, fig. 34).
The specimen of IDipnorhynchus from Australia figured by Hills
(1935, fig. 1) shows on one side the radiation of three bones that are
also closely comparable. The splenial and postsplenial (fig. 14B, SP,
PSP) lie below the symphysis and postero-lateral to it. The angular
(fig. 14B, AN) is the most posterior of these bones and extends nearly
to the posterior tip of the jaw, where it lies on the lateral side of the
articular; in front of the articulatory area it forms the lateral margin
of the adductor fossa.

The prearticulars (fig. 14A, PR) cover most of the medio-dorsal
surfaces of the lower jaws. Most anteriorly they do not meet the
dentaries, as in Dipterus and Holodipterus, but form a blunt point
separated from the dentaries by a short distance. Laterally, they
suture with the posterior ends of the dentaries, but in front of each
suture is a sub-triangular space that is floored by what is probably an
ossification of Meckel's cartilage (fig. 14A, MK). The prearticulars
show no suture where they meet in the midline. Their posterior mar-
gin near the midline in the symphysial region is separated from the
underlying splenials by a gap that was probably occupied by Meckel's
cartilage and has been reduced by crushing in Cottonwood Canyon

Fig. 14. Lower jaws of Uranolophus wyomingensis, n. sp. (X 1). A, dorsal
view of PF 3798; B, ventral view, restored largely from PF 3318 and 3798. af, ad-
ductor fossa; AN, angular; AR, articular; co, commissure between mandibular
sensory canals; cs, undetermined foramen; DE, dentary;/o, fossa for fold of lower
lip; MK, ossification of Meckel's cartilage; ml, mandibular sensory canal; ol, oral
sensory canal; pa, processus angularis; PR, prearticular; PSP, postsplenial ;
SP, splenial; tr, tooth ridge.

381

382 FIELDIANA: GEOLOGY, VOLUME 17

material. Here were attached the retractor muscles of the lower
jaws, while the anterior gular lay below them in a superficial position.

Each prearticular is surmounted by a "tooth ridge" (fig. 14A, tr)
which continues that of the dentary, and extends posteriorly about
to the middle of the adductor fossa. In this area of the lower jaw
of Holodipterus, Gorizdro-Kulczycka (1950, p. 88) described three
pairs of conical teeth on what she identified as coronoids. Since no
sutures are shown in her drawings nor described in the text, it is
doubtful whether there were separate coronoid ossifications in this or
any dipnoan. Medial to the tooth ridge, the prearticular is denticu-
late, except on its medio-ventral edge. This medial lamina extends
nearly to the medio-ventral edges of the angular and postsplenial,
but is separated from them by a space filled by a presumed ossifica-
tion of Meckel's cartilage (fig. 14A, MK). Posteriorly, the medial
lamina extends nearly to the posterior end of the jaw, where it sur-
rounds the articular on its medial and ventral sides. Laterally to the
"tooth ridge" the prearticular has a smooth, non-denticulate lamina
tljat becomes very narrow posteriorly where it bounds the medial
side of the adductor fossa (fig. 14A, af) . The lateral lamina reaches
as far back as the articulation, just anterior to which it rises up to
form much of the processus angularis (fig. 14A, pa) . The latter has
globular ossifications on its crown which may represent ossifications
of an ascending or coronoid process of Meckel's cartilage. Just be-
hind the processus angularis the prearticular is pierced by a large
foramen (fig. 14A, cs) comparable to the foramen "cs." described by
Gross (1956, pp. 28, 32) in other lungfish and crossopterygians.

The articular ossification (fig. 14A, AR) appears in the glenoid
fossa for the quadrate, which is a deep groove, not quite transverse,
but oriented a bit anterior of medially. Its ventral surface is nearly
horizontal, while its anterior surface is strongly concave and vertical.
The whole surface of the articular is covered with irregular, globular
ossifications.

Teeth. — The pterygoids and prearticulars of Uranolophus wyo-
mingensis completely lack the ridged tooth plates that characterize
the majority of lungfishes. Instead, they have evolved what has
been described above as "tooth ridges." On the palate each "tooth
ridge" (fig. 8, tr) appears on the lateral edge of the pterygoid just in
front of the sharp curvature where that bone expands laterally under
the quadrate; it extends forward along the lateral edge of the ptery-
goid to the suture with the vomer, but not onto the vomer. On the
lower jaw the "tooth ridge" (fig. 14A, tr) rises on the dorsal edge of

DENISON: EARLY DEVONIAN LUNGFISHES 383

each prearticular opposite the middle of the adductor fossa; it ex-
tends anteriorly nearly continuously along the prearticular, onto the
dorsal edge of the dentary, and around the front of the jaw into the
prearticular "tooth ridge" of the opposite side. The continuity of
this ridge is interrupted slightly at the prearticular-dentary suture,
where the prearticular ridge continues mediad to the dentary ridge
along the margin of the prearticular.

The ridge itself forms a continuous crest, sometimes irregular, and
typically modified by the development on its inner and outer sides
of numerous vertical ridges at right angles to the main ridge (fig. 15).
Both the main ridge and the side ridges are composed of dentine.
The side ridges are often more strongly developed on the lateral side,
while on the medial side and at the posterior end of the "tooth ridges"
they grade into elongate and then conical denticles. In some speci-
mens the side ridges are weakly developed and take the form of elon-
gate denticles. Other specimens (PF 3807, 3867) show lateral to the
"tooth ridge" a row of isolated, conical teeth, resembling those of
Griphognathus minutidens (Gross, 1956, fig. 27) . These may be con-
sidered as side ridges that have developed independently from the
main "tooth ridge," since they are comparable in shape to some of
the adjacent side ridges. One specimen (PF 3816) completely lacks
"tooth ridges" on the pterygoids; where they usually occur are
rounded edges composed apparently of bone only partially covered
by dentine.

The crests of the main and side ridges are usually blunted, pre-
sumably by wear. Such worn ridges usually have a flat crest sep-
arated by sharp edges from the sides. In some specimens the amount
of wear is slight, the ridges remain prominent, and low side ridges,
or the bases of larger side ridges, may remain sharp and unworn.
In one specimen (PF 3792, fig. 15D) the main ridge has worn low
and the side ridges are much reduced.

The anterior, transverse part of the dentary, and of the upper lip
which over-rides it (fig. 7B), are developed differently from the rest
of the "tooth ridges." On the posterior faces of the dentary and
upper lip there may be side ridges, but usually they are not as prom-
inent as more posteriorly. Anterior to these posterior faces the den-
tine shows great variability in its development. It may be relatively
continuous, but hollowed out into irregular, shallow depressions; or
it may be pierced by deep grooves and pores which divide it into a
pattern of irregular areas, simple or branched ridges, projections, or
denticles.

384 FIELDIANA: GEOLOGY, VOLUME 17

The bones of the palate (pterygoids, vomers, and parasphenoid)
and the medial laminae of the prearticulars are covered with small,
conical denticles, composed of dentine. These denticles are com-
monly larger anteriorly, and also near to the "tooth ridges," where
they may be developed as short, sometimes branching ridges. The
palate and prearticular may show an occasional patch of larger den-
ticles, and one palate (PF 3793, fig. 9A) is notable for an oval, ele-
vated area near the center of the parasphenoid covered with sinuous,
branching ridges and elongate denticles composed of dentine.

There are a number of genera of lungfishes in which ridged dental
plates are lacking, but none has evolved anything exactly comparable
to the "tooth ridges" of Uranolophus wyomingensis. The "teeth" of
Dipnorhynchus lehmanni (Lehmann, 1956, pi. 1, fig. 2) and Ganorhyn-
chus splendens (Gross, 1965, fig. IB) are knob-like prominences ar-
ranged along the lateral edges of the pterygoids in two rows; some
of the prominences are confluent at their bases in G. splendens and
perhaps in D. lehmanni, though the only specimen of the latter is too
poorly preserved to be sure. Uronemus splendens has one or two rows
of "large, compressed, conical teeth" on the margins of the ptery-
goids and on the prearticulars (Watson and Gill, 1923, p. 204, fig. 30;
Graham-Smith and Westoll, 1937, p. 261) ; the teeth appear in the
figure to be confluent at their bases and thus to form a sort of "tooth
ridge." Griphognathus minutidens (Gross, 1956, p. 36, fig. 27) has
one or more rows of small, conical teeth on the prearticular, but usu-
ally they remain distinct and do not unite at their bases. Holodip-
terus sanctacrucensis (Gorizdro-Kulczycka, 1950, pp. 89-90, fig. 2:1)
has three pairs of large, isolated, conical teeth, presumably on each
prearticular. Fleurantia denticulata (Graham-Smith and Westoll,
1937, p. 249, fig. 9) has conical teeth arranged in radial rows on the
palate, but not forming a distinct tooth plate. Perhaps the most
aberrant development is in Dipnorhynchus sussmilchi, which has
three large, bulbous masses on the palate (Campbell, 1965, p. 635)
and similar structures on the lower jaws (White, 1966, p. 7, pi. 1,
fig. 1) ; as White has shown, they are composed of dentine underlain
by spongy bone. At least two genera, Soederberghia and Concho-
poma, lack teeth entirely, but have the palate and prearticular cov-
ered with denticles.

Fig. 15. Tooth ridges of Uranolophus wyomingensis, n. sp. (X 8). A-B, left
pterygoid of PF 3805, A, anterior part and B, middle of ridge; C-D, right ptery-
goid of PF 3792, C, middle and D anterior part of ridge; E, probable prearticular
ridge, PF 3864.

385

386 FIELDIANA: GEOLOGY, VOLUME 17

Watson and Gill (1923, p. 214) believed that Uronemus and Con-
chopoma were derived from ancestors with typical tooth plates. They
argued that the whole structure of the dipnoan skull was specially
modified for the use of crushing tooth plates, and that the skull and
tooth plates must have evolved together. According to them, the
presence of isolated teeth in Uronemus was a neotenic character, and
this is possible because in the modern Neoceratodus, as shown by
Semon (1901, pp. 116-122), the teeth arise as isolated denticles and
only later fuse to a bony base to form a tooth plate. Lehman (1959,
p. 34) also concluded that the presence of denticles instead of tooth
plates on the palate of Soederberghia was a neotenic character. To
me it seems more probable the dipnoans without tooth plates evolved
from an ancestor which had not yet evolved these specialized struc-
tures. As the marginal jaw bones were reduced and lost in early
dipnoan history, the palate and prearticulars evolved a variety of
biting structures, of which the typical dental plate with radiating
ridges was only one. Some lungfishes, such as Soederberghia and
Conchopoma, simply retained the primitive denticulation. Others
such as Griphognathus, Ganorhynchus, Uronemus, Holodipterus, and
Dipnorhynchus lehmanni, enlarged denticles on the margins of the
palate and on the prearticulars to form one or more rows of teeth.
In Uranolophus such individual tooth elements were fused to form
a ridge. Dipnorhynchus sussmilchi is quite isolated in developing
whole areas of the palate and lower jaws as bulbous, crushing sur-
faces. This view of dipnoan dental evolution has an important bear-
ing on classification, since those genera without tooth plates would
not be derived from, nor closely related to, those genera that did
possess such structures.

Endocranium and visceral skeleton. — All of the skulls of Uranolo-
phus wyomingensis in the collection have been greatly flattened and,
as a result, their endocrania have been crushed to such an extent that
they show little or none of their structure. One moderately large
skull (PF 3816) was partially prepared on the ventral surface in
order to examine the posterior part of the endocranium. This proved
to be so extensively crushed that it was of little use; it did show, how-
ever, that there were both perichondral ossifications as well as spongy
endochondral bone. In this respect, U. wyomingensis agrees with
Dipterus and Chirodipterus, and differs from later lungfishes in which
the ossification of the endocranium was reduced or lost.

The partially articulated type specimen of Uranolophus wyoming-
ensis (PF 3874, fig. 2) preserves a number of elements of the hyoid

DENISON: EARLY DEVONIAN LUNGFISHES 387

and branchial arches (and possibly cranial ribs), but they are dis-
placed enough so that identification of individual elements, with one
exception, is probably not possible. The largest is surely a cerato-
hyal, an element also known isolated in three other specimens (fig. 16).

Fig. 16. Ceratohyal of Uranolophus wyomingensis, n. sp., PF 5514 (X 1).

It is a long bone, laterally compressed as preserved, tapering at both
ends, and curved so that its ventral edge is convex and its dorsal edge
is nearly straight. At its proximal end is an unossified area, facing
postero-dorsally, presumably cartilaginous in life. There appears
also to be a smaller unossified area at the distal end for attachment
to a hypohyal. In other fossil lungfishes whose ceratohyals have been
described (Fleurantia, Soederberghia, Jarvikia, Uronemus, Concho-
poma) and in the modern Neoceratodus, the ceratohyals have a dif-
ferent and characteristic shape; they are rod-shaped bones distally
and expand to deep, laterally compressed blades proximally. In
Lepidosiren the ceratohyal is similar except that the proximal end,
while laterally compressed, is not much expanded. What I identify
as ceratohyals in Dipterus (PF 1294, UC 2217) are well-ossified,
robust rods, tapering slightly distally, and laterally compressed prox-
imally. In all lungfishes the ceratohyal probably had a cartilaginous
core and the ossification is superficial.

Almost certainly a number of the rod-like bones behind the skull
of the partially articulated U. wyomingensis (PF 3874, fig. 2) are
branchial arch elements. Although these cannot be individually
identified, they are of interest because this is the only lungfish, with
the possible exception of Conchopoma (Weitzel, 1926, p. 168), in
which branchial arches are known to be ossified. This is additional

388 FIELDIANA: GEOLOGY, VOLUME 17

support for the widely accepted view that primitive Dipnoi had a
well-ossified skeleton.

Lateral-line system. — None of the material of Uranolophus wyo-
mingensis from Cottonwood Canyon has been prepared to show the
canals of the lateral-line system, so their course can only be inferred
from the distribution of the pores by which the tubuli open on the
surface. As indicated by White (1965, p. 9), this can lead to errone-
ous interpretations. However, a skull (PF 1427, fig. 3A) and lower
jaws (PF 3318, fig. 14B) from Beartooth Butte, show the actual
course of some of the canals, which are preserved as sedimentary
fillings, or as impressions of fillings on the surface of dermal bones.
On the cranial roof, the arrangement of the canals agrees closely with
that of Dipnorhynchus sussmilchi, as figured by Hills (1941, figs. 1, 5).
The supraorbital canals start in the posterior part of J, extend antero-
lateral^ to Lx and L2, where they curve antero-medially onto M;
more anteriorly they are not indicated on PF 1427. A specimen
from Cottonwood Canyon (PF 3805) has on the anterior part of the
skull roof a pore distribution which can be interpreted to indicate
supraorbital canals with similar courses to those of Dipterus valen-
ciennesi (fig. 3D), as shown by White (1965, fig. 18). The main
lateral-line canal crosses Y2, Y1} and X, where it turns down into the
infraorbital canal. At the point where it turns down, PF 1427
(fig. 3A) has on one side a branch extending antero-medially toward
Lx+K; though this cannot be traced to the supraorbital canal, it
corresponds to part of the commissure connecting supra- and infra-
orbital canals in Dipterus. The middle pit line is seen only as short
grooves on I and Y2 of PF 3874 (fig. 4, pi) . No posterior pit line
or occipital commissure has been identified, probably because it is
developed as a canal within bone I ; this is indicated by the presence
of pores on this bone in PF 3816 and 3874 (fig. 4, occ), and of a short
section of a filled canal on one bone I of the Beartooth Butte skull
(PF 1427, fig. 3A) . The same may apply to Dipnorhynchus sussmilchi
(Campbell, 1965, p. 635). It is possible that the middle pit line is
developed as a canal in PF 3816; no pit line is visible, but some pores
do occur in the anterior part of I (fig. 5).

The sensory canals of the lower jaws are well shown on the Bear-
tooth Butte specimen (PF 3318) as grooves on the impressions of the
inner surfaces of the dermal bones. Each mandibular canal (fig. 14B,
ml) runs parallel to the medio-ventral edges of the angular and post-
splenial ; in the latter bone it turns forward parallel to the postsple-
nial-splenial suture, and from the point of turning also sends a medi-

DENISON: EARLY DEVONIAN LUNGFISHES 389

ally-directed commissure (fig. 14B, co) into the splenial. Presum-
ably the commissure joins the mandibular canal of the opposite side,
but its medial part is not preserved on this specimen. The oral canal
(fig. 14B, ol) runs parallel to the dorso-lateral borders of the angular
and postsplenial. The anterior parts of the lower jaws are not pre-
served, but it is probable from the courses of the canals that the oral
and mandibular canals meet in the anterior part of the postsplenial.
A similar canal pattern is indicated by pores on some of the lower jaws
from Cottonwood Canyon and in Dipterus (Jarvik, 1964, fig. 14B).

Shoulder girdle. — The shoulder girdle of Uranolophus wyomingen-
sis is known from four more or less complete cleithra, two attached
but incomplete cleithra and clavicles, and one probable interclavicle.
The partially articulated type specimen (PF 3874) is of little use
since it shows only part of one cleithrum, overlain by an operculum.

The cleithrum (fig. 17A-C) consists of a rather narrow, elongate,
dorsally-directed blade, and a shorter, antero-ventrally-directed ven-
tral part. The dorsal blade lacks on its external surface (fig. 17C)
any ridges such as occur in Neoceratodus and Sagenodus. Its anterior
edge is concave, and its posterior edge is convex in its dorsal part and
concave in its ventral part, terminating in a prominent angulation
where it curves into the lower part. The outer face is cosmine-
covered only along the posterior edge, and usually only on its dorsal
part. More anteriorly the outer surface is covered with tubercles and
ridges, probably indicating the area overlapped by the operculum.
The inner surface of the cleithrum (fig. 18, CLM) shows a center of
radiation nearly opposite the angulation between the dorsal and ven-
tral parts. From this center a shallow depression extends to the
antero-dorsal corner. The dorsal end of the cleithrum probably over-
lapped a plate, or the most ventral and posterior of a series of plates,
that connected it to the skull. These plates have not been identified,
nor is any overlap area clear. However, PF 3851 (fig. 17A) is unique
in having a deep, V-shaped notch in the dorsal end of the cleithrum
for the reception of such a plate.

The ventral part of the cleithrum has an external lamina that is
continuous with the dorsal blade, and differs from it externally only
in being directed more anteriorly. This lamina is ornamented on its
outer face with tubercles and ridges, and is truncated anteriorly, ex-
cept ventrally where it is prolonged into an anterior process; this
process is strong in one specimen (fig. 17B), but is only weakly de-
veloped in another (fig. 17C). On the inner surface (fig. 18, il) is a
clearly distinct, inner lamina, extending anteriorly from the center of

Fig. 17. Shoulder girdle elements of Uranolophus wyomingensis, n. sp. ( X 2/3).
A, left cleithrum and incomplete clavicle, inner side, PF 3851; B, right cleithrum,
inner side, PF 3846; C, right cleithrum, outer side, PF 3856; D, probable inter-
clavicle, PF 3865.

390

DENISON: EARLY DEVONIAN LUNGFISHES 391

Fig. 18. Inner side of right shoulder girdle of Uranolophus wyomingensis,
n. sp., restored from PF 3846 and 3851. CLA, clavicle; CLM, cleithrum; il, inner
lamina of ventral part of cleithrum; SC-CO, approximate position and extent of
scapulo-coracoid.

growth of the cleithrum along its postero-ventral edge to the tip of the
anterior process. The antero-dorsal edge of the inner lamina is widely
separated from the inner face of the outer lamina. One specimen
(PF 3851, fig. 17A) has part of the scapulo-coracoid preserved, at-
tached to the inner face of the cleithrum near the angulation between
the dorsal and ventral parts, and covering the center of radiation
(fig. 18, SC-CO). It appears to consist of a thin layer of superficial
bone covering spongy bone. Another specimen (PF 3846, fig. 17B)
shows rugosities in the area of attachment of the scapulo-coracoid.
Other lungfishes, as far as is known, have cartilaginous scapulo-
coracoids.

The only knowledge of the clavicles of Uranolophus wyomingensis
comes from two specimens, PF 3851 (fig. 17A) and 5364, which pre-
serve, respectively, the inner surface and an impression of the inner
surface of the postero-dorsal part of this bone. These show a curved
blade, concave medially, with a distinct dorsal process. The clavicle
lies against the inner face of the adjacent part of the cleithrum, and

392 FIELDIANA: GEOLOGY, VOLUME 17

its ventral edge lies against the ridge formed by the antero-dorsal
edge of the inner lamina of the cleithrum. Nothing is known of the
more anterior part of the clavicle.

A single plate, PF 3865 (fig. 17D), is provisionally identified as
an interclavicle. In life it was approximately symmetrical, but has
been bent near the midline, flattened, and distorted somewhat. It
is broadest posteriorly and narrows considerably anteriorly. Most
of its outer surface is cosmine-covered, but along all but the posterior
part of each lateral edge there is a broad overlap area, presumably
for the clavicle; anteriorly there is a narrow area probably overlapped
by a gular plate. Neoceratodus has an interclavicular cartilage rather
than a bone, and apparently this was true of most fossil lungfishes.
A probable interclavicle has been reported in Dipterus (Watson and
Gill, 1923, p. 207; Forster-Cooper, 1937, p. 229), but it has not
been described.

It would be of interest to compare the shoulder girdle of Uranolo-
phus with that of Dipterus, but the latter is not known in sufficient
detail. Later lungfishes whose shoulder girdles are well known, such
as Sagenodus, Ctenodus, and Neoceratodus, show many differences.
However, a close comparison can be made with Devonian rhipidistian
crossopterygians, such as Eusthenopteron (Jarvik, 1944) . The cleith-
rum of the latter can be compared point by point, and the only im-
portant difference is that it has a small ventral ridge on the inner
surface instead of a strong inner lamina. The clavicle is attached in
a similar fashion to the cleithrum, and has a postero-dorsal process
like that of Uranolophus. A small, oval interclavicle is present, but
Glyptolepis may have had a broad one more closely comparable to
that of Uranolophus (Gross, 1936, fig. 8g). It is worth noting that
Gross (1956, pp. 12-13, fig. 4) described from the Baltic Upper De-
vonian some shoulder girdles that are comparable to those of Urano-
lophus; he attributed them to the osteolepid, Latvius, but considered
the possibility that they might belong to the dipnoan, Rhinodipterus.

Body and fins. — One of the most important specimens obtained
at the Field Museum excavations in Cottonwood Canyon in the
northern Bighorn Mountains is the partially articulated type speci-
men of Uranolophus wyomingensis, collected in 1962 by Dr. E. S.
Richardson, Jr. Unfortunately, since the close of the 1960 field sea-
son this specimen had been lying under less than one inch of matrix,
and had been subjected to severe weathering. As a result, the pos-
terior and much of the ventral part of the body was lost, and the only
fins preserved are the first dorsal and part of the second dorsal. What

DENISON: EARLY DEVONIAN LUNGFISHES

393

is left of the body (fig. 1) suggests that it was at least as robust as
that of Dipterus valenciennesi, as restored by Forster -Cooper (1937,
fig. 1).

The position of the dorsal fins appears to be much the same as in
Dipterus, that is, close together and far posterior on the body. The

Fig. 19. First dorsal fin and base of second dorsal fin of type specimen of
Uranolophus wyomingensis, n. sp., PF 3874 ( X 1).

first dorsal is relatively larger than in Dipterus valenciennesi; its
length measured along its axis is approximately two-thirds of the
median length of the cranial roof, while in Dipterus the correspond-
ing ratio is only one-third. The second dorsal of U. wyomingensis is
surely larger than the first, but only part of its base is preserved.
The relatively large size of the first dorsal fin is probably a primitive
feature; all other lungfishes in which the dorsal fins are known, ex-
cept Rhynchodipterus, have the first dorsal reduced, modified, or fused
with the second dorsal.

The first dorsal fin (fig. 19) has at its base a small, sharply delim-
ited lobe, presumably muscular in life; it is covered with cycloid,
cosmine-coated scales which diminish in size distally, but proximally
become larger and blend into the scale rows of the body. The rest
of the fin, except its anterior edge, is covered with cosmine-covered,
dermal fin rays, agreeing in all respects with the lepidotrichia of other
Osteichthyes. Proximally, there is a series of elongate fin rays, ex-
posed for a considerable length, but overlapped at their bases by the
rounded scales of the basal lobe. The rest of the rays are jointed and
occasionally branched, and consist of small, rectangular, cosmine-
covered elements that diminish in size distally. Along the anterior

394 FIELDIANA: GEOLOGY, VOLUME 17

edge of the fin, two or three large, median dorsal scales bound the
base of the fin. Further distally, the anterior edge is bounded by
variably and irregularly shaped scales which are usually slightly
larger than the adjacent elements of the lepidotrichia.

The second dorsal fin (fig. 19) is surely larger, has a relatively
larger scale-covered lobe at its base, and a longer row of large, me-
dian dorsal scales along its anterior edge.

Uranolophus wyomingensis is, in general, comparable to Dipterus
in its dorsal fins, but is presumably more primitive in the relatively
large size of the first dorsal, in the small size of its scaled basal lobe,
and in having the lepidotrichia mostly exposed and covered with cos-
mine. It is interesting to note that in these respects it closely ap-
proaches the conditions in early crossopterygians, such as Porolepis
and Gyroptychius (Jarvik, 1959, figs. 11, 20D, pi. 2).

Scales. — Uranolophus wyomingensis is unique among lungfishes in
having scales that, in general, have a rhombic exposed part and rela-
tively narrow overlapped margins. In these respects it resembles the
primitive crossopterygians, Osteolepididae and Porolepididae, rather
than other lungfishes, which have cycloid scales with very broad
overlapped areas. The scales are also thick, consisting of the same
three layers as the dermal bones. Almost all of the scales in the col-
lection that show the outer surface have the exposed part covered
with cosmine. This is true of the partially articulated type specimen
(PF 3874) and also of PF 3862, which preserves many associated
scales and bones of a large individual. Almost all of the prepared
isolated scales have well-developed cosmine, but PF 5546 (fig. 20A) ,
which shows several scales believed to belong to one individual, has
the exposed part only partially covered with cosmine. Many of the
isolated scales which are exposed on the inner face may lack the cos-
mine layer; where this layer is absent, the exposed part has a denticu-
late surface which undoubtedly adheres better to the matrix than
does the smooth inner surface. A cosmine-coated scale usually has
a narrow anterior band of sloping-topped, dentine-covered denticles;
this band lies between the cosmine and the bony overlapped area,
but on the evidence of the articulated type specimen, was also over-
lapped by the scales anterior to it. Many of the cosmine-covered
scales show Westoll lines and other growth phenomena; these will be
discussed below.

The flank scales are closely comparable to those of early crosso-
pterygians such as Osteolepis and Gyroptychius (Jarvik, 1948, fig. 28) .

Fig. 20. Scales of Uranolophus wyomingensis, n. sp. ( X 2). A, anterior flank
scale, outer side, PF 5546; B, flank scale, outer side, PF 5548; C, anterior flank scale,
inner side; PF 5547; D-F, median dorsal scales; D, PF 5544; E, PF 5545; F,
PF 3862.

395

396 FIELDIANA: GEOLOGY, VOLUME 17

The exposed part is distinctly rhombic, and in the larger anterior
scales (fig. 20A) the antero-dorsal to postero-ventral dimension is
appreciably longer than the opposite dimension. In these scales the
anterior overlapped area is narrow, but the dorsal overlapped area
has a moderately to very large articular process, comparable to, but
much larger than that of Gyroptychius. There is no groove between
the exposed and overlapped areas of the outer surface, such as Jar-
vik (loc. cit.) figures in Gyroptychius. More posterior flank scales
(fig. 20B) have the two dimensions subequal, and may have the
overlapped margins broader, the articular process reduced, and the
posterior angle rounded. The inner surface of the flank scales
(fig. 20C) has a depressed overlap area along its posterior border
and a convex depression on the ventral border for overlap of the ar-
ticular process of the adjacent scale. The ridge that is characteristic
of the inner face of the scales of osteolepids and porolepids is weak or
absent in Uranolophus wyomingensis. The flank scales may reach a
quite impressive size in this species. In the rather small, type speci-
men (PF 3874), whose total length may have been 45 cm., the largest
flank scale, measured along its exposed face in an antero-dorsal to
postero-ventral direction, is nearly 14 mm. long. In PF 3862, which
is estimated to have been 23^ times as long, flank scales reach at least
35 mm. in this dimension. As mentioned above, small flank scales
extend onto the lobe at the base of the first dorsal fin. These are
definitely cycloidal, and diminish in size distally.

Median dorsal scales are, of course, symmetrical, but show a
variety of shapes. One very large one associated with PF 3862
(fig. 20F) is considerably broader than long; presumably it occupied
a position close behind the head. Other relatively large ones occur
a short distance in front of the first dorsal fin of PF 3874 (fig. 1) ;
these, as better seen in isolated specimens (PF 5544-5, fig. 20D-E),
are longer than broad, and taper gradually posteriorly to a rounded
point. The median dorsal scales extend for a short distance up the
anterior edges of the two dorsal fins (fig. 19).

Vertebral column. — A few flank scales were removed from the type
specimen just behind the shoulder girdle in an effort to find the ver-
tebral column and ribs, and a string of ossified vertebrae was revealed
(fig. 21A). Each vertebra consists dorsally of a backwardly-inclined
neural spine fused to a neural arch, and ventrally of a smaller ele-
ment. No ribs were found. Each neural spine (fig. 21B, nsp) is
divided along its length by a sulcus into an anterior median element,
and what is probably a pair of posterior elements that overlap the

DENISON: EARLY DEVONIAN LUNGFISHES

397

Fig. 21. Seven vertebrae of the type specimen of Uranolophus wyomingensis,
n. sp., PF 3874 (X 2). Photograph (A) is interpreted in drawing (B). ic, inter-
centrum; na, neural arch; nsp, neural spine.

median element of the next posterior neural spine. The neural arch
(fig. 21B, na) has a prominent process projecting posteriorly over
the neural arch of the vertebra next behind. The ventral element
(fig. 21B, ic) is probably an intercentrum, but no pleurocentrum has
been recognized. The structure is basically comparable to that of
living Neoceratodus, except that the neural arch and central elements
of the latter are cartilaginous, and a small pleurocentrum is present
anteriorly. It is, in general, similar to the vertebrae of Conchopoma
(Weitzel, 1926, pp. 168-169), though in the latter the neural spines

398

FIELDIANA: GEOLOGY, VOLUME 17

are not fused to the arches. There is nothing resembling the ossified,
amphicoelous centra described by Jarvik (1952, pp. 40-47, figs. 16-
19) in two Upper Devonian lungfishes from Greenland, and it is
probable that the latter are specializations not typical of Dipnoi.

Histology. — There have been a number of studies of the histology
of the dermal bones and teeth of Devonian lungfishes, especially of
Dipterus; the most important are those of Pander (1858), Bystrow
(1942), and Gross (1956, 1965). In general, Uranolophus has a sim-
ilar histological structure to Dipterus and for this reason only certain
points will be discussed.

The superficial layer of Uranolophus, as in Dipterus, may consist
of a continuous sheet of dentine, pierced by pore canals and under-
lain by mesh canals, forming the tissue known as cosmine (fig. 23C).
In some individuals, or in some places, it may consist only of scat-
tered tubercles composed of dentine (fig. 23B), and in these cases the
pore-canal system must have lain in soft tissue between the denticles.
In Dipterus and Rhinodipterus such tubercles are composed of bone
rather than dentine (Gross, 1965, p. 122). In scales, but usually not
in skull bones, such tubercles or denticles may be overgrown and
buried by bone, apparently with little resorption; new denticles or

Fig. 22. Pores of the pore-canal system on dermal plates of Uranolophus
wyomingensis, n. sp. ( X 8). A, very large pores on undetermined plate, PF 5515;
B, usual small pores, lower jaw fragment, PF 3849.

DENISON: EARLY DEVONIAN LUNGFISHES 399

perhaps cosmine may then form on the surface, resulting in some
scales in as many as four generations of buried denticles (fig. 23B-C).
This type of overgrowth also occurs in some crossopterygians and
actinopterygians, but has not been reported previously in lungfishes.

The pores of the pore-canal system are usually quite uniform in
size on a jaw or a skull, or over all of the preserved body of the type.
This is true of moderately small individuals such as the type, as well
as of the largest individual in the collection (PF 3862). There is no
variation clearly related to the size of the individual or to the position
of the pores on the body, such as Gross (1956, p. 72) found in certain
Dipteridae. In most specimens (fig. 22B) the pores range in diam-
eter between .04 and .08 mm., approximately the same as Gross (loc.
cit.) found most commonly in the Dipteridae he studied. There are
two exceptional situations in Uranolophus, however. One occurs in
a very small lower jaw (PF3808), which has exceptionally large,
.10 to .20 mm. in diameter, closely-spaced pores. A few isolated
plates (fig. 22A) and scales have similar large, closely-spaced pores,
and their smooth margins show that the pores have not been en-
larged by resorption. The other unusual situation occurs in a pair
of large lower jaws (PF 3797, fig. 13A) and the anterior part of a
skull (PF 3805, fig. 6), which though disassociated, could well have
belonged to the same individual. Here the pores of the pore-canal
system are larger, .15 to .30 mm. in diameter, but are completely
lacking on the anterior part of the skull and lower jaws. Two
isolated snouts (PF 3848, 3868) also lack the typical pore-canal sys-
tem. Sections of similar fragments (fig. 23A) show that not only the
pore canals but also the mesh canals are absent. The superficial cos-
mine of these individuals is cut by grooves resembling sutures. On
the snout of PF 3805 these grooves are similar to and continuous
with sutures between the dermal bones of the middle part of the skull
roof, but in a fragment (PF 3806) that has been sectioned such a
groove proved to be superficial and no evidence of a suture could be
found in the middle or basal layers. It is not impossible that in some
cases (though probably not in PF 3805) these grooves represent a
specialized remnant of the pore-canal system rather than sutures.

The pores of the lateral-line system are much larger than those
of the pore-canal system, and in some cases are extremely large an-
teriorly. However, I do not find evidence that they had been enlarged
by resorption, as indicated by White (1962, pp. 3-4) in Rhinodipt-
erus. Most of the large anterior pores can be attributed to the lateral-
line system, though it is not impossible that some housed specialized

400 FIELDIANA: GEOLOGY, VOLUME 17

sensory organs such as occur in modern dipnoans (Gross, 1956, pp.
94-96).

On the pterygoid (fig. 23F) and prearticular tooth ridges, and on
the "upper lip" and dentaries (fig. 23E), the dentine becomes very
thick and assumes the characteristic structure of dipnoan teeth ; this
consists of a type of trabecular dentine with numerous pulp canals
perpendicular to the surface and many branched dentine tubules ex-
tending out from the canals. I find no evidence of a prismatic struc-
ture in this dentine, such as that described by Gross (1965, p. 128)
in Dipterus, Rhinodipterus, and Ganorhynchus. The small denticles
that cover the surface of the palate and prearticular are constructed
of a simple orthodentine. The surface of the dentine is covered by a
thin, transparent layer that has often been identified as enamel;
however, it is penetrated by numerous dentine tubules and is more
probably durodentine or enameloid, as has been shown in Ganorhyn-
chus by Gross (1965, p. 124). On the tooth ridges this layer is worn
off by the abrasion of chewing. On other superficial parts of the
skull and jaws the enameloid and outermost dentine may be removed
locally in some specimens, just as in Cardipeltis (Denison, 1966, pp.
111-112). This is demonstrably not the result of abrasion, and may
be due either to resorption or infection.

The middle layer is spongy bone that is generally crushed in
specimens from Cottonwood Canyon. It is usually rather thin, but
in the snout and anterior parts of the lower jaws (fig. 23A) it becomes
very thick. In scales it may include several generations of buried
denticles (fig. 23B-C).

The basal layer is typically a dense, cross-laminated bone, some-
times called isopedine. During growth additional laminae are added
basally, resulting often in a thick basal layer; however, on the snout
and anterior parts of the lower jaws it apparently remains thin.
Usually there are very few vascular canals, but one specimen (slide
5082) has in one region numerous large canals, some clearly enlarged
by resorption. Scales have the usual thick layer of dense, evenly
cross-laminated isopedine, which shows a very regularly alternating
structure between crossed nicols (fig. 23D). At the base in cross sec-
tions there may be a bulge, which is presumably the inner ridge.
The lamination is convex inward in this bulge, but between crossed
nicols there appears a very distinct structure, cutting across the lami-
nation, and sometimes resembling a secondary lamination. This
cross structure indicates a distinct crystalline orientation, and pre-
sumably is the result of an original coarse fibrous structure cutting
across the lamination.

Fig. 23. Thin-sections of Uranolophus wyomingensis, n. sp. (A-E, X 25;
F, X 75). A, longitudinal section through splenial of large lower jaw lacking pores
and canals of the pore-canal system, slide 4609; B, transverse section of a scale
with tuberculate surface and two generations of overgrown denticles, slide 5090;
C, transverse section of a scale with cosmine surface and three or more generations
of overgrown denticles, slide 4258; D, transverse section of a scale and part of an-
other of the type specimen, PF 3874, between crossed nicols to show the lamination
of the basal layer, slide 5085; E, longitudinal section through the tip of the dentary
of PF 3850, slide 5088; F, transverse section through the pterygoid tooth ridge,
slide 4254.

401

402 FIELDIANA: GEOLOGY, VOLUME 17

Growth of scales, skull bones and tooth ridges. — Cosmine, since it
consists typically of a continuous layer of dentine, has imposed severe
restrictions on the growth of those fish that possessed it. Early

Fig. 24. Scale of Uranolophus wyomingensis, PF 5541 ( X 3), showing position
of section illustrated in Figure 25.

Osteostraci, such as Tremataspis, whose carapace was completely
covered by cosmine, simply did not grow after the formation of the
cosmine, and presumably passed through an unarmored larval stage.
The other two groups of fishes with cosmine, Dipnoi and Crossoptery-
gii, did manage to grow by various expedients, though most of them
lost cosmine quite early in their evolutionary history. It is clear that
Uranolophus wyomingensis overcame the cosmine problem and grew.
This is shown by the considerable size range of specimens in the
Field Museum collection (indicating an estimated total length of 35
to 110 cm.), by growth lines, and by histology. Thin sections have
revealed that the manner of growth was different in scales and in
dermal bones of the skull roof and lower jaws, and these will be dis-
cussed separately.

Transverse sections of cosmine-coated scales (fig. 23C), often
reveal several generations of dentine-capped denticles buried beneath
the cosmine. Each generation of denticles was at the surface of the
scale during one stage of growth before cosmine was formed. De-
tailed study of a scale in thin-section shows that the history of growth
may be very complex, and that growth may continue well after the
first formation of cosmine. The sectioned scale (fig. 24), which was
about 16 mm. in diameter, had a central-posterior area of cosmine,

Q

0

— ' efl 05

0)0 0)

05 M

O) 5

— 05

c * c
•S £ +»

o ° 53

3 «~ t-"
-C O

05

£ 5
9 «

c SP c
C 0) o •

'G £ *> x

0! OS 0)

£2 £~
£ g-s *

Sb x •*; «
.2 £ fc g.

C « J 41
0) 05 +■> rt

w"*~, -
.2§3

"M

i

at .

• W "O Or

il-s

£ £ 3

403

404 FIELDIANA: GEOLOGY, VOLUME 17

separated by a Westoll line from a surrounding ring of cosmine (in-
complete posteriorly where the scale was broken) , and two small an-
terior patches of cosmine also separated by a Westoll line from the
cosmine ring. In front was an overlapped area consisting only of
bone most anteriorly, but capped with numerous denticles adjacent
to the cosmine. As interpreted (fig. 25), the section shows two early
generations of denticles (stages 1-2), overlain by the first layer of
cosmine (stage 3), which forms the central-posterior area. A set of
buried denticles is probably related to this generation of cosmine, and
is thought to have covered part of the anterior overlapped area at
this stage of growth. The next three stages (4-6) of growth are
largely restricted to the anterior part of the scale, and are covered
only by denticles. However, the following stage (7) involved the
formation of cosmine in a ring around the exposed area of the scale,
and probably also of a denticulate band in the anterior overlapped
area. The final growth stage shown in the section (8) is indicated
only by a small patch of bone surmounted by a single denticle near
-the anterior end of the scale. The third generation of cosmine was
not included in this section. Though this scale shows a complicated
history, it is probably simpler than larger scales with more numer-
ous Westoll lines.

The manner of growth of Uranolophus scales differs from that of
other lungfishes. Post-Devonian and most Upper Devonian lung-
fishes had completely lost the cosmine and so were not faced with the
growth problems that it imposed. In Dipterus cosmine may be pres-
ent or absent, but in spite of a number of studies it is not yet certain
just how Dipterus scales grew. Westoll (1936, p. 169) and Forster-
Cooper (1937, p. 227) showed that cosmine was periodically resorbed
to permit growth. Bystrow (1942, p. 286) and Jarvik (1950, p. 39) dif-
fered as to whether resorption and regrowth was centrifugal or cen-
tripetal, then Gross (1956, p. 78) demonstrated that areas of cosmine
did not grow, but were formed all at once, with finished, enamel-
oid-coated edges. In whatever manner Dipterus scales grew, major
amounts of cosmine resorption were involved, and cosmine did not
overgrow earlier generations of cosmine or denticles. In Uranolo-
phus, on the contrary, resorption played no important part in scale
growth and cosmine regularly overlies older denticles, though prob-
ably not earlier layers of cosmine. In this respect it is closely similar
to the primitive crossopterygian, Porolepis, in which one or more
generations of dentine-crowned denticles are commonly buried be-
neath later cosmine (Gross, 1966, p. 41, fig. 5A).

DENISON: EARLY DEVONIAN LUNGFISHES 405

Such skull bones as operculars and gulars, which did not fuse to
each other or to the rest of the dermal cranium, could have grown
in a similar manner to scales. Apparently they did not, however, as
is shown by thin sections of small fragments of three operculars.
The smallest sectioned opercular (PF 3839) has a maximum diam-
eter of 61 mm.; its outer surface is largely tuberculate, and the sur-
face tubercles rest on bone which has been eroded on the surface
by resorption; the superficial bone layer covers one buried denticle
whose crown is partly resorbed. The second opercular (PF 3835)
has a maximum diameter of 71 mm. ; its surface is covered with large-
pored cosmine, and the section shows no buried denticles. The larg-
est opercular (PF 3862) has a maximum diameter of 100 mm.; it has
fine-pored cosmine, and the section shows no buried denticles. In
these three operculars, the basal layer shows an increase in thickness
in the larger plates (0.19 -0.39 -0.68 mm., respectively), but the
middle and superficial layers do not, and only one denticle of an
earlier generation was preserved. This suggests that there has been
extensive resorption and replacement of denticles and cosmine dur-
ing growth.

The growth of the plates of the cranial roof and of the ventral
and lateral sides of the lower jaws of Uranolophus is distinctly differ-
ent from that of its scales. Individual plates tend to fuse at their
sutures, and the sutures are commonly overgrown by a sheet of cos-
mine. The relatively small type specimen (PF 3874) has many, but
not all, of its cranial roof sutures covered by cosmine; where covered,
they are indicated by a linearity of pore arrangement along the suture.
The largest specimen, PF 3862, has similar small-pored cosmine cov-
ering the lower jaws, and no sutures are evident. Specimens of inter-
mediate size show a variety of conditions which cannot be related in
any way to size. Some have fine-pored cosmine covering many
sutures. One (PF 3792) is entirely covered with denticles of den-
tine, except for a narrow band of cosmine at the snout. A large skull
(PF 3805) has open sutures and large-pored cosmine, except on the
snout where there are no pores (and presumably no canals) of the
pore-canal system. A pair of lower jaws (PF 3797) is similar. Finally,
another large skull (PF 3816, fig. 5) has fine-pored cosmine, Westoll
lines, and areas of cosmine resorption along sutures and Westoll lines.
Of the thin sections that have been made of skull or jaw plates, only
one shows buried denticles, probably all of one generation; in most
cases buried denticles and cosmine are definitely absent. These sec-
tions, the moderate-sized denticulate specimen (PF 3792), and the
large specimen with partly resorbed cosmine (PF 3816) all indicate

406 FIELDIANA: GEOLOGY, VOLUME 17

that cosmine and denticle resorption was a regular process on the
skull and lower jaws, and was probably extensive, as in Dipterus.
Presumably at times when cosmine was resorbed there was an open-
ing of sutures and a period of growth. In some cases this was followed
or accompanied by the formation of dentine-capped denticles on the
surfaces, while in other cases cosmine was formed again. In certain
individuals there was a modification of the pore-canal system when
the cosmine was absent; in some individuals this resulted in the de-
velopment of large pores, and in others in the loss of the pore-canal
system on the snout and anterior ends of the lower jaws.

The growth of the tooth ridges is a special problem. Since they
are composed of dentine, they cannot grow on their surfaces, and
since they are fused to the pterygoids and prearticulars they cannot
grow at their bases. Yet they are not only longer in larger individ-
uals, but they are also higher and wider. This may have been ac-
complished by a periodical resorption and reformation of the tooth
ridges. One large skull (PF 3816) is at the intermediate stage in the
replacement; it completely lacks tooth ridges and the margins of
the pterygoids are bone, probably only partially covered by dentine.
The replacement of tooth ridges is also supported by the fact that
large skulls of older individuals may have tooth ridges that are rela-
tively sharp and unworn.

Relationships. — The phylogenetic significance of Uranolophus will
be discussed in another paper. At this time it will suffice to state
that it is in most respects a very primitive lungfish, and that it may
be grouped with Dipnorhynchus sussmilchi and D. lehmanni in the
family Dipnorhynchidae. The latter, here characterized by similari-
ties in cranial roof bone pattern and proportions, and by the absence
of tooth plates, was originally proposed by Berg (1940, p. 385) for
Dipnorhynchus alone. Ganorhynchus and Holidipterus may also be
related, as was first suggested by Gorizdro-Kulczycka (1950, p. 95),
but these genera are inadequately known. Griphognathus has been
included by Vorobyeva and Obruchev (1964, p. 307), but is so aber-
rant in its slender jaws, long retro-articular process, and minute, iso-
lated teeth that this is unlikely.

II. Lungfishes from the Water Canyon Formation
of Utah and Idaho

The Card member or lower part of the Water Canyon formation
of Utah (Williams and Taylor, 1964, p. 39) has been correlated ap-
proximately with the Beartooth Butte formation on the evidence of
its fish fauna. In it fragments referable to Uranolophus sp. have been
collected at three localities in Cache County:

Blacksmith Fork, locality A (Denison, 1952, p. 266) : PF 352, a
small scale with cosmine coating and a marginal area with sloping
denticles.

Blacksmith Fork, locality D (loc. cit.) : PF 496, a scale with
Westoll lines and denticulate margin.

Cottonwood Canyon, locality G (op. cit., p. 267) : PF 927, a large,
median dorsal scale with Westoll lines.

An undetermined lungfish plate and scale (PF 497-8) occur also
at locality G.

The Grassy Fork or upper member of the Water Canyon forma-
tion is presumably younger than the Beartooth Butte formation, and
contains a distinct fish fauna that is as yet undescribed except for
Psephaspis williamsi 0rvig (1961, p. 526) and Protaspis erroli Deni-
son (1967). From the ridge south -southeast of Naomi Peak, local-
ity H (Denison, 1952, p. 267), in Cache County has come a fragment
of a snout of a small, undetermined lungfish (PF 340) which is much
smaller than any referred so far to Uranolophus.

A few lungfish fragments have also been obtained from the lower
member of the Water Canyon formation of southeastern Idaho
(Coulter, 1956, p. 30). Scales (PF 5554-8, 5560), a probable antero-
lateral gular (PF 5553), and an incomplete angular bone (PF 5559),
all referable to Uranolophus sp., have been obtained from the west
slope of the ridge on the east side of the North Fork of St. Charles
Creek, about 63^ miles west of St. Charles, Bear Lake County. Un-
determined lungfishes from this locality are skull plates (PF 5562-3)
and scales (PF 5561, 5564) . At a slightly higher level at this locality,
but still within the lower member of the formation, has come a very

407

408 FIELDIANA: GEOLOGY, VOLUME 17

small and incomplete tooth plate (fig. 26) which may be referred to
Dipterus sp. (PF 5566). This is of particular interest because it is
the oldest known dipnoan dental plate, and shows that more con-
ventional lungfishes existed approximately contemporaneously with

Fig. 26. Incomplete tooth plate of Dipterus sp. from the Water Canyon for-
mation of southeastern Idaho, PF 5566 (X 8).

Uranolophus. It preserves parts of seven rows of denticles, and on un-
worn parts the individual denticles of a row are completely distinct.
Each denticle is capped with shiny dentine or enameloid and sepa-
rated from adjacent denticles in its own and neighboring rows by a
spongy tissue, surely bone. In this respect it differs from many other
Dipterus, in which adjacent denticles are contiguous, and resembles
the presumed primitive condition shown in embryonic Neoceratodus
(Semon, 1901, pi. 19, figs. 10-11). On worn rows or parts of rows the
individual denticles are reduced to form a continuous ridge, though
the position of the original denticles is indicated by swellings on the
ridge. The apex of the tooth plate has been worn smooth and shows
no evidence of the original rows of denticles.

Uranolophus sp. also occurs at another locality about 1 mile north
of Beaver Creek along the Beaver Creek to Green Canyon Road,
about 6j/2 miles west-southwest of St. Charles in Franklin County.
From here have come flank scales with sloping denticles and Westoll
lines (PF 5568) and a skull plate with lateral-line pores and Wes-
toll lines (PF 5569).

All of these lungfish fragments have come from rocks that are
probably of marine origin, though there are no diagnostically marine

DENISON: EARLY DEVONIAN LUNGFISHES 409

invertebrates associated with them. They occur in limestones which
are widespread, and locally contain Lingula, pelecypods, gastropods,
and ostracods. Marginal marine environments have been suggested
(Denison, 1956, p. 414; Williams and Taylor, 1964, p. 42), though it
is not impossible that some freshwater sediments are interbedded
with the marine ones.

REFERENCES

Berg, L. S.

1940. Classification of fishes, both recent and fossil. Trav. Inst. Zool., Acad.
Sci. U.R.S.S., 5, pp. 87-517.

Bryant, William L.

1932. Lower Devonian fishes of Bear Tooth Butte, Wyoming. Proc. Amer.
Phil. Soc, 71, pp. 225-254, figs. 1-6, pis. 1-10.

Bystrow, A. P.

1942. Deckknochen und Zahne der Osteolepis und Dipterus. Acta Zool., 23,
pp. 263-289, figs. 1-18.

Campbell, K. S. W.

1965. An almost complete skull roof and palate of the dipnoan Dipnorhynchus
sussmilchi (Etheridge). Palaeontology, 8, pp. 634-637, fig. 1, pi. 91.

Coulter, Henry W.

1956. Geology of the southeast portion of the Preston quadrangle, Idaho.
Idaho Bur. Mines Geol., Pamphlet 107, pp. 1-48, map.

Denison, Robert H.

1952. Early Devonian fishes from Utah. Part I. Osteostraci. Fieldiana:
Geol, 11, pp. 263-287, figs. 50-60.

1956. A review of the habitat of the earliest vertebrates. Fieldiana: Geol.,
11, pp. 359-457.

1958. Early Devonian fishes from Utah. Part III. Arthrodira. Fieldiana:
Geol., 11, pp. 461-551, figs. 86-116.

1964. The Cyathaspididae, a family of Silurian and Devonian jawless verte-
brates. Fieldiana: Geol., 13, pp. 309-473, figs. 90-161.

1966. Cardipeltis, an Early Devonian agnathan of the Order Heterostraci.
Fieldiana: Geol., 16, pp. 89-116, figs. 1-11.

1967. A new Protaspis from the Devonian of Utah, with notes on the classifi-
cation of Pteraspididae. Jour. Linn. Soc. London (Zool.), 47, no. 311, pp.
31-37.

Dorf, Erling

1934. Stratigraphy and paleontology of a new Devonian formation at Bear-
tooth Butte, Wyoming. Jour. Geol., 42, pp. 720-737, figs. 1-7.

FORSTER-COOPER, C.

1937. The Middle Devonian fish fauna of Achanarras. Trans. Roy. Soc. Edin-
burgh, 59, pp. 223-239, figs. 1-10, pis. 1-8.

Goodrich, E. S.

1909. A treatise on zoology. Part IX. Vertebrata Craniata (First Fascicle:
Cyclostomes and Fishes). London, i-xvi, 1-518 pp., figs. 1-515.

GORIZDRO-KULCZYCKA, Z.

1950. Les dipneustes devoniens du massif de S-te Croix. Acta Geol. Polonica,
1, pp. 53-105, figs. 1-2, pis. 1-4.

410

DENISON: EARLY DEVONIAN LUNGFISHES 411

Graham-Smith, W. and Westoll, T. S.

1937. On a new long-headed dipnoan fish from the Upper Devonian of Scau-
menac Bay, P. Q., Canada. Trans. Roy. Soc. Edinburgh, 59, pp. 241-266,
figs. 1-12, pis. 1-2.

Gross, Walter

1936. Beitrage zur Osteologie baltischer und rheinischer Devon-Crossoptery-

gier. Palaont. Z., 18, pp. 129-155, figs. 1-10, pis. 7-8.
1956. Uber Crossopterygier und Dipnoer aus dem baltischen Oberdevon im

Zusammenhang einer vergleichenden Untersuchung des Porenkanalsystems

palaozoischer Agnathen und Fische. K. Svenska Vetenskapsakad. Handl.,

(4), 5, no. 6, pp. 1-140, figs. 1-124, pis. 1-16.

1964. tlber die Randzahne des Mundes, die Ethmoidalregion des Schadels und
die Unterkiefersymphyse von Dipterus oervigi n. sp. Palaont. Z., 38, pp. 7-25,
figs. 1-3, pis. 1-3.

1965. Uber den Vorderschadel von Ganorhynchus splendens Gross (Dipnoi,
Mitteldevon). Palaont. Z., 39, pp. 113-133, figs. 1-7, pi. 19.

1966. Kleine Schuppenkunde. Neues Jahrb., Geol. Pal., Abh., 125, pp. 29-48,
figs. 1-7.

Gunther, Albert

1872. Description of Ceraiodus, a genus of ganoid fishes, recently discovered in
rivers of Queensland, Australia. Phil. Trans., Roy. Soc, London, 161, pp.
511-571, 3 figs., pis. 30-42.

Hills, E. S.

1933. On a primitive dipnoan from the Middle Devonian rocks of New South
Wales. Ann. Mag. Nat. Hist. (10), 11, pp. 634-643, figs. 1-5, pis. 11-12.

1935. Records and descriptions of some Australian Devonian fishes. Proc.

Roy. Soc. Victoria, n. s., 48, pp. 161-171, figs. 1-5, pi. 12.
1941. The cranial roof of Dipnorhynchus sussmilchi (Eth. Fil.). Rec. Australian

Mus., 21, pp. 45-55, figs. 1-6, pi. 9.

1958. A brief review of Australian fossil vertebrates. In Westoll, T. S., ed.
Studies on Fossil Vertebrates. Univ. of London, pp. 86-107.

Jarvik, Erik

1944. On the exoskeletal shoulder-girdle of teleostomian fishes, with special

reference to Eusthenopteron foordi Whiteaves. K. Svenska Vetenskapsakad.

Handl., (3), 21, no. 7, pp. 1-32, figs. 1-9.
1948. On the morphology and taxonomy of the Middle Devonian osteolepid

fishes of Scotland. K. Svenska Vetenskapskad. Handl., (3), 25, no. 1, pp. 1-

301, figs. 1-85, pis. 1-37.

1950. Middle Devonian vertebrates from Canning Land and Wegeners HalvO"
(East Greenland). Part II. Crossopterygii. Medd. Gr0nland, 96, Nr. 4,
pp. 1-132, figs. 1-37, pis. 1-24.

1952. On the fish-like tail in the ichthyostegid stegocephalians with descrip-
tions of a new stegocephalian and a new crossopterygian from the Upper
Devonian of east Greenland. Medd. Gr0nland, 114, Nr. 12, pp. 1-90, figs.
1-36, pis. 1-12.

1954. On the visceral skeleton in Eusthenopteron with a discussion of the para-
sphenoid and palatoquadrate in fishes. K. Svenska Vetenskapsakad. Handl.,
(4), 5, no. 1, pp. 1-104, figs. 1-47.

1959. Dermal fin-rays and Holmgren's principle of delamination. K. Svenska
Vetenskapsakad. Handl., (4), 6, no. 1, pp. 1-51, figs. 1-22, pis. 1-4.

1963. The composition of the intermandibular division of the head in fish and
tetrapods and the diphyletic origin of the tetrapod tongue. K. Svenska
Vetenskapsakad. Handl., (4), 9, no. 1, pp. 1-74, figs. 1-27.

412 FIELDIANA: GEOLOGY, VOLUME 17

1964. Specializations in early vertebrates. Ann. Soc. Roy. Zool. Belgique, 94,

pp. 11-95, figs. 1-28.
1967. On the structure of the lower jaw in dipnoans: with a description of an

early Devonian dipnoan from Canada, Melanognathus canadensis gen. et sp.

nov. Jour. Linn. Soc. London (Zool.), 47, no. 311, pp. 155-183, figs. 1-9,

pis. 1-6.

Lehman, J.-P.

1959. Les dipneustes du Devonien superieur du Groenland. Medd. Gr0nland,
160, nr. 4, pp. 1-58, figs. 1-30, pis. 1-21.

Lehmann, W. M.

1956. Dipnorhynchus lehmanni Westoll, ein primitiver Lungenfisch aus dem
rheinischen Unterdevon. Palaont. Z., 30, pp. 21-25, figs. 1-2, pi. 1.

Lehmann, W. and Westoll, T. S.

1952. A primitive dipnoan fish from the Lower Devonian of Germany. Proc.
Roy. Soc, London, (B), 140, pp. 403-421, figs. 1-5, pi. 24.

0rvig, Tor

1961. Notes on some early representatives of the Drepanaspida (Pteraspido-
morphi, Heterostraci). Ark. Zool., (2), 12, nr. 33, pp. 515-535, figs. 1-10.

Pander, C. H.

1858. Die Ctenodipterinen des devonischen Systems. St. Petersburg, i-viii +
65 pp., 9 pis.

Parrington, F. R.

1949. A theory of the relations of lateral lines to dermal bones. Proc. Zool.
Soc. London, 119, pp. 65-78, figs. 1-2.

Philip, G. M. and Pedder, A. E. H.

1964. A re-assessment of the age of the Middle Devonian of south-eastern
Australia. Nature, 202, pp. 1323-1324.

Romer, A. S.

1936. The dipnoan cranial roof. Amer. Jour. Sci., (5), 32, pp. 241-256, figs. 1-4.

Ruedemann, Rudolph

1934. Eurypterids from the Lower Devonian of Beartooth Butte, Wyoming.
Proc. Amer. Phil. Soc, 73, pp. 163-167, figs. 1-2, pis. 1-3.

Sandberg, Charles A.

1961. Widespread Beartooth Butte formation of early Devonian age in Mon-
tana and Wyoming and its paleogeographic significance. Bull. Amer. Assoc.
Petrol. Geol., 45, pp. 1301-1309, figs. 1-3.

Save-Soderbergh, G.

1952. On the skull of Chirodipterus wildungensis Gross, an Upper Devonian
dipnoan from Wildungen. K. Svenska Vetenskapsakad. Hand!., (4), 3, no. 4,
pp. 1-28, figs. 1-15, pis. 1-7.

Semon, Richard

1901. Die Zahnentwickelung des Ceratodus forsteri. Zool. Forschunsr. in Aus-
tralien, 1, pp. 113-135, figs. A-M, pis. 18-20.

Stensio, Erik A.

1947. The sensory lines and dermal bones of the cheek in fishes and amphibians.
K. Svenska Vetenskpsakad. Handl., (3), 24, no. 3, pp. 1-195, figs. 1-38.

Thomson, Keith S.

1965. The nasal apparatus in Dipnoi, with special reference to Protopterus.
Proc. Zool. Soc. London, 145, pp. 207-238, figs. 1-12, pi. 1.

DENISON: EARLY DEVONIAN LUNGFISHES 413

Traquair, R. H.

1873. On a new genus of fossil fish of the order Dipnoi. Geol. Mag., 10, pp.

552-555, 1 fig., pi. 14.
1878. On the genera Dipterus, Sedgw. & Murch., Palaedaphus, Van Beneden

and De Koninck, Holodus Pander, and Cheirodus, M'Coy. Ann. Mag. Nat.

Hist., (5), 2, pp. 1-17, pi. 3.

Vorobyeva, E. I. and Obruchev, D. M.

1964. Subclass Sarcopterygii. In: Obruchev, D. M.,ed. Fundamentals of Pale-
ontology. Manual for Paleontologists and Geologists of USSR. Agnatha,
Pisces. Moscow. Pp. 268-322, figs. 1-67, pis. 1-5. (In Russian.)

Watson, D. M. S. and Day, Henry

1916. Notes on some Palaeozoic fishes. Mem. Proc. Manchester Lit. Philos.
Soc, 60, pp. 1-52, figs. 1-9, pis. 1-3.

Watson, D. M. S. and Gill, E. L.

1923. The structure of certain Palaeozoic Dipnoi. Jour. Linn. Soc, Zool., 35,
pp. 163-216, figs. 1-34.

Weitzel, Karl

1926. Conchopoma gadiforme Kner, ein Lungenfisch aus dem Rothliegenden.
Abhandl. Senck. Naturforsch. Ges., 40, pp. 159-178, pis. 18-23.

Westoll, T. S.

1936. On the structures of the dermal ethmoid shield of Osteolepis. Geol. Mag.,
73, pp. 157-170, figs. 1-4.

1949. On the evolution of the Dipnoi. In: Jepsen, G. L., Mayr, E., and Simp-
son, G. G., eds. Genetics, Paleontology, and Evolution. Princeton Univ.
Press. Pp. 121-184, figs. 1-11.

White, E. I.

1962. A dipnoan from the Assise de Mazy of Hingeon. Bull. Inst. Roy. Sci.
Nat. Belgique, 38, no. 50, pp. 1-8, figs. 1-3, pis. 1-2.

1965. The head of Dipterus valenciennesi Sedgwick & Murchison. Bull. Brit.
Mus. (Nat. Hist.), Geol., 11, pp. 1-45, figs. 1-51, pis. 1-3.

1966. Presidential address: A little on lung-fishes. Proc. Linn. Soc. London,
177, pp. 1-10, figs. 1-5, pi 1.

Williams, J. S. and Taylor, M. E.

1964. The Lower Devonian Water Canyon formation of northern Utah. Univ.
Wyoming, Contr. Geol, 3, pp. 38-53.

Publication 1044