Middle Devonian fishes from the Lemhi Range of Idaho

UNIVEHSITY OF

ILLINOIS LIBRARY

AT URBANA-CHAMPA1GN

GEOLOGY

OCT f 3
00*1 9

[976
1977

L161 — O-1096

FIELDIANA • GEOLOGY

Published by
FIELD MUSEUM OF NATURAL HISTORY

Volume 16 December 12, 1968 No. 10

Middle Devonian Fishes

from the
Lemhi Range of Idaho

Robert H. Denison

Curator, Fossil Fishes

In 1960 Mortimer H. Hait, Jr., then a graduate student in the
Department of Geology of Pennsylvania State University, sent to me
for identification some Devonian fishes from east-central Idaho. In
July of that year he guided a Field Museum party, including Dr.
E. S. Richardson, Jr., and myself, to the locality and additional speci-
mens were obtained, and in 1964 I returned briefly for further col-
lecting. The locality is at the head of Spring Mountain Canyon on
the crest of the Lemhi Range, l]/2 miles south of Gilmore in Lemhi
County. By extrapolation on an unsurveyed portion of the Gilmore
quadrangle of the U. S. Geological Survey, it is near the southwest
corner of sec. 10, T 12 N, R 27 E. The fishes occur in sandstones
and dolomites in a channel-like deposit called by Hait the basal unit
of the Jefferson formation. The rocks have been subjected to suffi-
cient pressures so that well-developed jointing makes it hard to ob-
tain entire specimens. Preparation is difficult, particularly in the
sandstones, which are very firmly cemented. All of the specimens
are in the Field Museum collection.

Mr. Hait's discovery and his assistance have been acknowledged
in part by naming one of the new species in his honor. I wish to ac-
knowledge also the help of Dr. Tibor Perenyi, Staff Artist, in making
the drawings, and of John Bayalis and Homer Holdren, Staff Pho-
tographers, in making some of the photographs.

Library of Congress Catalog Card Number: 68-59028
No. 1062 269

The Library of the

MAY 1 5

Umvers4ty of Illinois

GEOLOGY U-BBARjC

270 FIELDIANA: GEOLOGY, VOLUME 16

HETEROSTRACI
Pteraspididae

Psephaspis idahoensis, new species

Type. — FMNH, PF 5680, a nearly complete dorsal disc with at-
tached pineal plate and a fragment of the dorsal spine (fig. 1).

Referred specimens. — Fragmentary dorsal discs, PF 5671, 5674
(fig. 8B), 5675 (fig. 8A), 5681-2, 5683 (fig. 7B) ; fragmentary ventral
disc, PF5661 (fig. 3).

Occurrence. — As given above.

Diagnosis. — A species of Psephaspis probably attaining a smaller
size than P. williamsi 0rvig, and differing in ornament in that the
denticles tend to be higher, with the crown more strongly convex, or
commonly with an upper surface that slopes in one direction while
the opposite sides are very steep.

Description and discussion. — In his original description of Pse-
phasis 0rvig (1961, pp. 526-533) referred it to the Drepanaspida.
Considering the poor preservation and fragmentary nature of his ma-
terial there was justification for this assignment, for the ornament
typically consists of denticles rather than ridges, and there are "blis-
ters" of second generation denticles, such as are common in drepa-
naspids, but hitherto unknown in pteraspids. However, the material
from the Lemhi Range of Idaho to be described here is unmistakably
pteraspid in the arrangement and manner of growth of its plates and
in the presence of a dorsal spine, though it agrees closely with Pse-
phaspis williamsi in its ornament and in the presence of "blisters."

0rvig's type of P. williamsi is a fragment of a disc which he
thought to be probably the lateral part of a ventral disc. Tarlo
(1965, pp. 38-39) considered it to be a dorsal median plate, but I
follow 0rvig in identifying it as a ventral disc, though my recon-
struction of it (fig. 2) differs from his. As I have restored it, the pro-
portions are relatively very broad for a pteraspid ventral disc, though
actually what is preserved gives little information about the length
of the plate. However, a very fragmentary ventral disc of P. idaho-
ensis (fig. 3), though distorted, was clearly relatively short and broad,
with small growth increments posteriorly and postero-laterally, and
the type of P. williamsi has been restored to resemble this. The ven-
tral disc of P. williamsi was much larger, however.

The dorsal disc of Psephaspis idahoensis is best preserved in the
type, PF 5680 (fig. 1), where its shape is typical of pteraspids, but

DENISON: DEVONIAN FISHES FROM IDAHO

271

Fig. 1. Psephaspis idahoensis, n. sp., type, PF 5680, dorsal disc with attached
pineal plate and a fragment of the dorsal spine( X5/8). A, anterior view; B, dorsal

its proportions are extremely broad ; as measured and estimated along
its convex outer surface, the ratio of width/length is probably about
1.7-1.8. Anteriorly, there is a deep notch for the pineal plate which
is apparent at an early growth stage. The antero-lateral and lateral
margins are broadly curved with no identations for branchial open-
ings, which suggests that the latter were at the postero-lateral corners
of the shield as in some Protaspis. The posterior margin is lost in the
type, but is preserved in part in PF 5681-2, both of which show this
edge extending in an antero-lateral direction from the midline, slop-

272

FIELDIANA: GEOLOGY, VOLUME 16

Fig. 2. Psephaspis vnlliamsi 0rvig, a new restoration of the ventral disc,
based on the type specimen, Naturhistoriska Riksmuseet, Stockholm, no. 1357
(X7/16); the preserved part of the shield is indicated by stippling.

ing downward posteriorly, and possessing rather broad posterior
growth increments. The posterior margin is deeply notched medi-
ally for the dorsal spine. The latter is incompletely preserved in the
type and other specimens, but has a long, narrow base that is unusu-
ally deep at its insertion into the dorsal disc, which is greatly thick-
ened in this region. Only a few pores of the lateral line system are
recognizable in the type, but on PF 5683 two rows of elongate slits
(fig. 7B) are seen on the right half of the dorsal disc. These are pre-
sumably the pores of the principal and dorsal lateral lines, both of
which are far from the midline, as is the case also in some species of
Pteraspis.

DENISON: DEVONIAN FISHES FROM IDAHO

273

Fig. 3. Psephaspis idahoensis, n. sp., restoration of the ventral disc, PF 5661
(X 15/16); the preserved part of the shield is indicated by stippling.

Of other plates of the shield, only the ventral disc mentioned
above (fig. 3), and the pineal plate attached to the type dorsal disc
(fig. 1) can be referred surely to this species. The pineal plate is small,
shaped like a broadly open V, and must have been widely separated
from the orbital plates. In the collection are a few isolated plates
and scales, some of which may belong to P. idahoensis. Two incom-
plete rostral plates, PF 5673 and 5684 (fig. 4), are similar in general

P5

X

274

DENISON: DEVONIAN FISHES FROM IDAHO

275

Fig. 5. Psephaspis sp., incomplete branchial plate, PF 5677 ( X3/2).
tral view; B, medial view.

A, ven-

shape, but differ in proportions. The broader one would fit better
the broad shield of P. idahoensis, but it is not impossible that the
narrower proportions of the other have resulted from distortion. A
single, incomplete orbital plate (PF 5676) could belong to a small in-
dividual of this species. Three partial branchial plates, PF 5677-8,
5685, are referable to Psephaspis sp. Of these, PF 5677 (fig. 5) is
an incomplete posterior end, and shows that this plate was very broad
posteriorly, both in its ventral and dorsal laminae, and that the bran-
chial duct extended to the posterior end. Thus, the external branchial
opening was probably at the posterior end of the branchial plate as in

Fig. 6. Psephaspis sp., ridge scale, PF 5679 ( X3J^).

276

FIELDIANA: GEOLOGY, VOLUME 16

Fig. 7. Ornamentation of Psephaspis ( Xll). A, P. williamsi 0rvig, type ven-
tral disc, Naturhistoriska Riksmuseet, Stockholm, no. 1357; B, P. idahoensis, n. sp.,
dorsal disc, PF 5683, showing long slits of the lateral line system.

Protaspis erroli and Europrotaspis crenulata, rather than on the dorsal
face of the shield between the branchial plate and the dorsal disc, as
in certain Protaspis from Utah (Denison, 1953, pp. 320-323). Two
isolated ridge scales, PF 5679 (fig. 6), have anteriorly, just behind the
narrow overlapped area, a band of denticles similar to those of the
of the central area of the dorsal or ventral disc. Behind are many-
transverse rows of longitudinally elongated denticles, with the rows
separated by sinuous grooves, a type of ornamentation that is char-
acteristic of pteraspids.

DENISON: DEVONIAN FISHES FROM IDAHO 277

The typical ornamentation of Psephaspis idahoensis (fig. 7B) is
very similar to that of P. williamsi (fig. 7A) as described by 0rvig
(1961, p. 529). One minor difference is that the denticles of P. ida-
hoensis are usually higher and more strongly convex, and as a result,
their crenulated margins are difficult to see, though they are obvious
on P. williamsi. Commonly, the denticles have a relatively flat sur-
face sloping in one direction, while their other sides are very steep.
The denticles may show a tendency to elongate into ridges near the
margins of a plate, though this is not shown as well as in the type of
P. williamsi. The denticles are of comparable size in the two species.
It is interesting to note that at the center of growth of the dorsal disc
there is an area ornamented with fine, continuous ridges of dentine
(fig. 8A, r), similar to the ornament of Pteraspis or Protaspis. This
ornament grades through a zone of small, elongate denticles into the
typical larger, rounded denticles. In PF 5682, the center of growth
is an elongate oval area as in Pteraspis carmani (Denison, 1960,
p. 572), but in PF 5681 the oval is constricted laterally, suggesting
that there were two initial growth centers; this recalls the three cen-
ters of growth described by Fahlbusch (1957, pp. 23-24) in Rhino-
pteraspis dunensis.

One of the drepanaspid features of Psephaspis williamsi, though
not shown on the type specimen, is the presence of "blisters," or
patches of second generation denticles overgrowing the original den-
ticles (0rvig, 1961, pp. 529-532). The second generation denticles of
this species are very similar to those of the first generation, and are
arranged in concentric rows. Four dorsal discs of P. idahoensis also
show blisters. In two of these, PF 5674 (fig. 8B, b) and 5683, the
blisters may be very irregular in shape and are composed of denticles
that are also irregular in size and shape. On PF 5675 (fig. 8A, b) the
blisters occur near the center of growth and their denticles are more
regular in shape, though considerably larger than the first generation
denticles of the same area. The majority of the blisters overgrow a
worn, elevated growth line, which supports the contention of Gross
(1935, p. 16) that the secondary denticles form in response to super-
ficial injury, though 0rvig (1961, p. 529, footnote) has questioned this.

As far as can be determined from the arrangement of the rows of
denticles and the numerous growth lines, the growth of the shield and
of the individual plates of Psephaspis was similar to that of better
known pteraspids (Fahlbusch, 1957; Denison, 1960). There is no
indication in any shield plates of P. idahoensis, nor in 0rvig's mate-
rial of P. williamsi, of scales or tesserae such as are incorporated into

Fig. 8. Psephaspis idahoensis, n. sp., ornamentation of the dorsal disc, show-
ing "blisters" (X7J^); A, PF 5675, near center of growth; B, PF 5674, with an ir-
regular "blister." b, "blisters"; r, fine ridges near center of growth of dorsal disc;
s, anterior part of dorsal spine.

278

DENISON: DEVONIAN FISHES FROM IDAHO 279

the margins of the plates of some drepanaspids during growth (Heintz,
1957). Tarlo (1965, p. 39, pi. 14, fig. 3) claimed to see such tesserae
in the margin of a plate from southeastern Idaho, which he referred to
P. williamsi. These are clearly only cracks resulting from the crush-
ing of the plate during or after burial, such as commonly occur in
specimens from this locality.

The histology of a dermal plate of P. williamsi, though poorly
preserved, was described by 0rvig (1961, p. 533, fig. 10). He con-
cluded that the plates did "not differ in any essential respect from
those of the other Drepanaspida where their microstructure is con-
cerned," and Tarlo (1965, p. 39) stated that the histology "appears
to be typically psammosteid rather than pteraspid." I cannot agree
with either 0rvig or Tarlo in this matter. The middle layer is crushed
in 0rvig's material (as it is in a sectioned specimen of P. idahoensis,
FMNH, slide 5188), so it is difficult to determine whether it originally
had large cancellae as is usual in pteraspids, or a typical spongiosa
as in drepanaspids; in any case, it surely was relatively thin, in which
it differs strikingly from typical drepanaspid histology. The basal
layer, on the other hand, is relatively very thick in both P. williamsi
and P. idahoensis; this feature is not typical of either drepanaspids or
pteraspids and may be related to the age of the individual sectioned.

? Psephaspis sp.

PF 5672, a ventral disc (fig. 9) preserved in counterpart from the
basal unit of the Jefferson formation at the head of Spring Mountain
Canyon, Idaho, has ornamentation of the Psephaspis type, in fact
agreeing very closely with that of P. idahoensis. The relatively nar-
row proportions of this disc (width/length=0.66) make it impossible
to refer it to P. idahoensis, nor certainly to Psephaspis, and indicates
the presence of a second species. Since a ventral disc exhibits so few
characters, I prefer not to base a new species on this specimen.

Revised diagnosis of Psephaspis

Psephaspis includes moderate to large-sized pteraspids, with the
shield usually relatively broad, and with the ornament consisting typ-
ically of denticles with crenulated margins. Blisters of second gen-
eration denticles may occur. The branchial openings are at the
postero-lateral corners of the shield, at the posterior ends of the
branchial plates. The lateral line system is poorly developed, with
the principal and dorsal lateral lines placed far from the midline at the
anterior end of the dorsal disc.

280

FIELDIANA: GEOLOGY, VOLUME 16

Fig. 9. ? Psephaspis sp., PF 5672, ventral disc (X7/10).

Range and significance of the genus Psephaspis

The association of P. idahoensis with Holonema and an astro-
lepid indicates that its age is Middle Devonian. This is of interest
since the Pteraspididae have hitherto been restricted to the Lower
Devonian. Psephaspis may then be looked upon as a genus that has
persisted after the extinction of the rest of its family, or after some
of them evolved into drepanaspids. It acquired certain of the char-
acters of drepanaspids, such as the broad proportions, the type of
ornament, and the ability to produce a second generation of dentine.
These were presumably parallelisms, since it surely did not acquire
other characteristics of drepanaspids, such as the bands of tesserae
between the larger plates, nor separate postorbital plates.

DENISON: DEVONIAN FISHES FROM IDAHO 281

The type species, P. williamsi, came from the Water Canyon for-
mation of northeastern Utah. The typical fauna of this formation
comes from the lower or Card Member and is Lower Devonian, lower
or middle Siegenian in age (Denison, 1958, p. 500), but P. williamsi
was found in the upper or Grassy Flat member whose age is uncer-
tain. It could be anywhere in the late Lower or Middle Devonian,
because the lower part of the overlying Jefferson formation here con-
tains an Upper Devonian invertebrate fauna. A specimen that I
collected in 1959 in the Water Canyon formation of southeastern
Idaho has been identified as P. williamsi by Tarlo (1965, p. 38, pi. 14,
fig. 3). This came from the west slope of the ridge on the east side
of the North Fork of St. Charles Creek, about 6)^ miles west of
St. Charles, Bear Lake County. Some other fragments from this
locality (FMNH, PF 5689-90) tend to confirm Tarlo's identification.
They came from the lower member of the Water Canyon formation
as described by Coulter (1956, p. 51), though well above its base.
However, here they are associated with what appears to be the typ-
ical Siegenian fauna, tentatively identified as Protaspis, Allocrypt-
aspis, and arthrodires indet., so this occurrence is presumably Lower
Devonian.

Another species of Psephaspis, P. bystrowi, has been named by
Tarlo (1964, pp. 79, 113; 1965, pp. 39-40) and requires some com-
ment. This was based on a single tessera, covered with rounded
tubercles, originally described from the Lower Devonian of Siberia
by Bystrow (1959, pp. 67-68) as Drepanaspis sp. This should not
be referred to Psephaspis for three reasons: 1) Psephaspis surely
lacked such tesserae; 2) the ornament appears to have a different
character from that of Psephaspis; and 3) in histological structure,
the tessera has a thinner basal layer, a much thicker spongiosa, and
perhaps more widely spaced denticles (though Bystrow's figure 7A
shows them closely crowded). This plate had best be called ? Dre-
panaspis sp., until such time as more specimens reveal its relationship.

I wish also to criticize the reference by Tarlo (1965, p. 64, pi. 14,
fig. 5) of a specimen of Psephaspis idahoensis from the Lemhi Range
of Idaho to Schizosteus wellsi. Tarlo (1964, pp. 115-116) based S.
wellsi on an oral plate from the Middle Devonian bone beds of Ohio.
The latter is ornamented with rounded denticles, but as far as one
can judge from Wells' figures (1944, pi. 10, figs. 7-8; pi. 11, fig. 1),
the denticles differ in shape and in spacing from those of Psephaspis,
and the histology of another fragment is quite different in its thin
basal layer, thick spongiosa, and widely spaced denticles. So what-

282 FIELDIANA: GEOLOGY, VOLUME 16

ever may be the merit of Schizostens wellsi as a species, it is certain
that it has nothing to do with Psephaspis.

ARTHRODIRA

Holonematidae

Holonema haiti, new species

Type. — FMNH, PF 5669, an incomplete median dorsal plate
(fig. 10).

Occurrence.- — As given on p. 269.

Diagnosis. — The median dorsal plate is proportionately longer
and narrower than in other species of this genus, with an estimated
ratio of width/length of 0.28. It is also more strongly arched than
in other species, having a median angulation of about 90° near the
anterior end.

Description and discussion. — The type specimen clearly shows the
distinctive characteristics of the median dorsal plate of Holonema.
One of these is the long, narrow proportions which are extreme in
H. haiti; in other species the ratio of width/length ranges from 0.40
in H. radiatum Obruchev to an estimated 0.50 in H. farrowi Stevens.
The ornamentation is formed mainly by ridges instead of the tuber-
cles usual in Arthrodira, and has a characteristic pattern which agrees
well with that described by Wells (1943, p. 5) in H. rugosum (Clay-
pole). This consists of 1) a long central ridge, the "club" of Wells;
2) concentric ridges posterior and lateral to 1; 3) lateral zones of
ridges mainly at right angles to the edge, but showing many irregu-
larities and sinuosities; 4) anterior fanned ridges; 5) also a posterior
zone of fanned ridges, not developed in H. rugosum.

Since the type was preserved in an extremely hard sandstone it
was necessary to prepare it by grinding, resulting in the removal of
the tops of most ridges. As a consequence, the ridges appear wider
and their interspaces narrower than was originally the case. In one
small marginal area where the ridges were exposed by weathering,
they are round crested and have slight swellings that make them
appear like a series of fused elongate tubercles. Here the interspaces
are only slightly narrower than the ridges, and there are occasional
commissures between adjacent ridges.

Another characteristic feature of the Holonema median dorsal
plate is the sensory canal, which extends from near the center of
growth first anterolateral^, and then curves strongly laterally. In

X

".2

1*

pq

283

284 FIELDIANA: GEOLOGY, VOLUME 16

the type of H. haiti it is seen only on the right side; on the left it is
either absent, more superficial, or located asymmetrically more pos-
teriorly in a poorly preserved area. This canal is a dorsal branch of
the main lateral line canal, and is known in Coccosteidae and some
other Placodermi.

On the inner side there is a median keel at the posterior end (fig.
10B), probably similar to that figured by Obruchev (1932, fig. 5) in
H. radiatum. Whether there was also a small anterior keel, as in H.
radiatum, is uncertain; it is not indicated at the point where the an-
terior end of the plate has been broken off.

Probably very little of the posterior end of the type is missing,
and this can be restored with some confidence (fig. 10A). Little more
of the anterior end has been lost, but the most anterior part preserved
of the right margin shows a definite approach toward the midline,
suggesting that only a short length is missing here. On the basis of
these conclusions, the total length of the median dorsal is estimated
to have been 28.0 cm. This is longer than the described median dor-
sals of H. rugosum (17.5 cm.) and H. farrowi (20 cm.), and approxi-
mately the size of H. radiatum (29.5 cm.). The posterior width is
estimated to be 8 cm. The height near the anterior end is about
3.1 cm., but posteriorly the plate is much lower. The thickness of
the bone is 4-5 mm. at the midline (anterior to the keel), and thins
gradually toward the lateral margins. However, it should be noted
that a thin-section (FMNH, slide 5187) shows evidence of compac-
tion of the spongiosa, so presumably the plate was originally some-
what thicker. The histology, which is poorly preserved, resembles in
general that of H. radiatum as figured by Obruchev (1932, pi. 8, fig. 3).

Other species of Holonema are known from the Middle and Upper
Devonian, approximately Givetian and Frasnian. Thus H. rugosum
(Claypole) occurs in the Middle Devonian (Erian) of Michigan and
Wisconsin, as well as the Upper Devonian (Senecan) of New York
and Pennsylvania. H. farrowi Stevens occurs in the Erian of Mich-
igan. H. radiatum Obruchev is found in the lower Upper Devonian
of Russia, and perhaps of Spitsbergen. H. harmae Mark comes from
the Middle Devonian of Estonia. Other species, H. obruchevi Mark,
H. horridum Cope, H. ornatum Traquair, and H. eifeliense (Kayser),
do not belong in this genus. The genus Deirosteus is very closely re-
lated, and was distinguished by Wells (1942, p. 655) because on the
median dorsal the ornamentation was essentially concentric, with
narrow ridges and broad interspaces, and because the sensory canal
was not indicated. As pointed out by Stevens (1964, p. 173), these

DENISON: DEVONIAN FISHES FROM IDAHO

285

Fig. 11. Astrolepididae indet., PF 5668, centro-nuchal plate (X3).

differences may not be of generic importance, especially as H. farrowi
and D. omaliusii are intermediate between typical species of the two
genera. The occurrence of the species referred to Deirosteus is as
follows: D. abbreviatus (Eastman) is from the lower Upper Devonian
(Senecan) of New York; D. omaliusii (Agassiz) is from the Givetian
of Belgium; D. angustatus Obruchev is from the Eifelian of Siberia,
and is thus the oldest species of these two genera. In conclusion, the
age of Holonema haiti is most probably Eifelian to Frasnian.

Arthrodira indet.

The presence of more conventional Euarthrodira is indicated by
a number of fragmentary tuberculated plates on PF 5603. None are
identifiable.

ANTIARCHA

Astrolepididae indet.

Antiarcha are represented in the collection by two centro-nuchal
plates (PF 5667-8) and a median ventral plate (PF 5662). The cen-
tro-nuchals (fig. 11) are most closely similar to those of Astrolepis
Eichwald and Pterichthys Agassiz. The former is known from the

286 FIELDIANA: GEOLOGY, VOLUME 16

Middle Devonian and lower Upper Devonian of Great Britain, Rus-
sia, Spitsbergen, Greenland, China, Australia, and a single record
from the eastern United States. The latter is known from the Middle
Devonian of Scotland and the Baltic states.

? HOLOCEPHALI

Spine indet.

A small spine (fig. 12), 14.9 mm. in length, is associated with the
Psephaspis dorsal disc on PF 5675. Its exserted portion is strongly
curved, ornamented with very fine longitudinal ridges, and has a
single blunt-pointed tubercle near the tip on the concave side. Its
inserted portion lacks ornamentation, broadens rapidly, and con-
tinues the curvature of the exserted portion. As far as I know, no
acanthodian has its fin spines as strongly curved, nor possesses a
tubercle on the concave edge. This spine is suggestive in certain
respects of the spines of some genera referred to the Menaspididae,
and so it is possible that it belonged to an early representative of the
Holocephali.

Fig. 12. ? Holocephalian spine, on PF 5675 (X7.3).

Age and Environment of Deposition

As pointed out above, both the astrolepid and Holonema indicate
a Middle or early Upper Devonian age for the basal unit of the Jeffer-
son formation at the head of Spring Mountain Canyon. Psephaspis,
on the other hand, as a member of the Pteraspididae, suggests a
Lower Devonian age. I conclude, however, that the age of this unit
is Middle Devonian, and that some Psephaspis outlasted the other
genera in its family.

DENISON: DEVONIAN FISHES FROM IDAHO 287

The basal unit of the Jefferson formation at this locality occurs
in what has been described by Hait (personal communication) as a
channel-like deposit. Its fishes, unfortunately, do not give any defi-
nite answer to the environment of deposition. Psephaspis occurs in
the Water Canyon formation of Utah, which is in part, at least,
marine, though some beds may have been deposited in marginal or
even in freshwater environments. The latter is possibly true of the
bed containing the type specimen of P. williamsi, which is a reddish,
poorly sorted sandstone. On the other hand, in southeastern Idaho,
specimens referred to this species occur in beds containing Lingula
and are surely marine in part. Holonema (and the doubtfully dis-
tinct Deirosteus) occurs in both marine and freshwater formations.
The European species occur largely or entirely in freshwater deposits,
while the American species are predominantly marine, though some
occurrences of H. rugosum appear to be continental. Astrolepididae
are predominantly freshwater fishes, though marine genera are known.
One might conclude that the basal unit of the Jefferson formation at
this locality is a marginal deposit with a mixed fauna, perhaps estu-
arine in origin.

Summary

The basal unit of the Jefferson formation at the head of Spring
Mountain Canyon, Lemhi County, Idaho contains the following
fishes: a heterostracan, Psephaspis idahoensis n.sp., which is shown
to belong to the Pteraspididae rather than the Drepanaspididae; the
arthrodire, Holonema haiti n.sp., and undetermined arthrodires; an
antiarch referred to the Astrolepididae indet.; and a spine possibly
belonging to a holocephalian. The age is considered to be Middle
Devonian, and the depositional environment possibly estuarine.

288 FIELDIANA: GEOLOGY, VOLUME 16

REFERENCES

Bystrow, A. P.

1959. The microstructure of skeleton elements in some vertebrates from Lower
Devonian deposits of the USSR. Acta Zool., 40, pp. 59-83, figs. 1-15.

Coulter, H. W.

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

Denison, R. H.

1953. Early Devonian fishes from Utah. Part II. Heterostraci. Fieldiana:

Geol., 11, pp. 291-355, figs. 61-85.
1958. Early Devonian fishes from Utah. Part III. Arthrodira. Fieldiana:

Geol., 11, pp. 461-551, figs. 86-116.

1960. Fishes of the Devonian Holland Quarry shale of Ohio. Fieldiana: Geol.,
11, pp. 555-613, figs. 117-149.

Fahlbusch, K.

1957. Pteraspis dunensis Roemer. Eine Neubearbeitung der Pteraspidenfunde
(Agnathen) von Overath (bez. Koln). Palaeontogr., 108, Abt. A, pp. 1-56,
figs. 1-27, pis. 1-7.

Gross, W.

1935. Histologische Studien am Aussenskelett fossiler Agnathen und Fische.
Palaeontogr., 83, Abt. A, pp. 1-60, figs. 1-30, pis. 1-7.

Heintz, A.

1957. The dorsal shield of Psammolepis paradoxa Agassis. Jour. Paleontol. Soc.
India, 2, pp. 153-162, figs. 1-3, pis. 17-19.

Obruchev, D.

1932. Holonemidae des russischen Devons. Acad. Sci. URSS, Inst. Paleozool.,
Trav., 2, pp. 97-116, figs. 1-27, pis. 5-8.

0RVIG, T.

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

Stevens, M. S.

1964. Thoracic armor of a new arthrodire (Holonema) from the Devonian of
Presque Isle county, Michigan. Mich. Acad. Sci. Arts, Letters, Papers, 49,
pp. 163-175, fig. 1, pis. 1-2.

Tarlo, L. B. H.

1964. Psammosteiformes (Agnatha) — A review with descriptions of new mate-
rial from the Lower Devonian of Poland. I. General part. Palaentologia
Polonica, 13, pp. i-vii, 1-135, figs. 1-32, pis. 1-14.

1965. Psammosteiformes (Agnatha) — A review with descriptions of new mate-
rial from the Lower Devonian of Poland. II. Systematic part. Palaeontolo-
gia Polonica, 15, pp. i-ix, 1-168, figs. 1-48, pis. 1-19.

Wells, J. W.

1942. Arthrodiran fish plates from the Enfield formation (Upper Devonian) of
New York. Jour. Paleontol., 16, pp. 651-656, pis. 95-97.

1943. A median dorsal plate of Holonema from the Upper Devonian of New
York. Bull. Amer. Paleontol., 27, no. 107, pp. 1-8, 1 fig., pi. 26.

1944. Fish remains from the Middle Devonian bone beds of the Cincinnati arch
region. Palaeontogr. Amer., 3, pp. 103-160, figs. 1-9, pis. 1-8.