en

LIBRARY OF THE

UNIVERSITY OF ILLINOIS

AT URBANA-CHA|V1PAIGN

550.5

n

GEOLOGY

yNIVERSITYGF

ILLINOIS LIBRARY

AT URBANA-CHAMPAIGN
OBOiOGY

;.5

FIELDIANA • GEOLOGY

Published by
CHICAGO NATURAL HISTORY MUSEUM

Volume 14 January 22, 1963 No. 7

New Silurian Heterostraci
from Southeastern Yukon

.' Robert H. Denison

\ Curator op Fossil Fishes

INTRODUCTION

By H. R. HovDEBO.i A. C. Lenz,i E. W, Bamber^

During the 1959 and 1960 field seasons, California Standard Com-
pany field parties, under the direction of H. R. Hovdebo, collected
ostracoderm remains in the Beaver River area of Yukon Territory.
The locality at which these collections were made is approximately
60° 27' N., 125° 47.5' W., some 100 feet north of the Beaver River.

The east-dipping homocline (figs. 58, 59) of the area in which the
ostracoderms were found is bounded to the east and to the west by
faults on which substantial vertical displacement can be demon-
strated, but which appear to diminish in magnitude in a southerly
direction. Within this homocline the upper contact of map Unit 5,
which is the top of a carbonate sequence, and the outcrop pattern of
map Units 3 and 7, which are resistant sandstones, have been fairly
closely defined; the other contacts are more loosely defined because
of poor exposure. Ages of the map Units have been determined only
for Unit 4 and the basal part of Unit 2; these are Silurian and Ordo-
vician, respectively. Map Units 5 through 7 appear to encompass,
on the basis of regional information, rocks of Middle and Late De-
vonian age.

The ostracoderms which are the subject of this paper were col-
lected from a dolomite lens within the middle part of Unit 4 (fig. 59,
Pt. C). Detailed relationships at the outcrop locality are shown in
figure 59. The possible fault zone indicated on this figure (Pt. D)
is not considered to have any significant magnitude in view of con-
tinuity of stratigraphically higher formations on the hillside above
the exposure.

1 The California Standard Company, Edmonton, Alberta.

2 Geological Survey of Canada, Ottawa, Canada.

Library of Congress Catalog Card Number. 63-12867

No. 963 105 GEOLOGY LIBRARY

106

FIELDIANA: GEOLOGY, VOLUME 14

Fig. 58. Surface geology of Beaver River area, southeastern Yukon, by H. R.
Hovdebo (scale: % inch=l mile). 1, Pre-Ordovician igneous complex; 2, Ordo-
vician (and younger?) carbonates; 3, Silurian(?) sandstone; 4, Silurian limestones
and graptolitic shale; 5, Lone Mountain formation (Middle Devonian?); 6, Fort
Creek shale (Upper Devonian); 7, Sandstone (Upper Devonian?).

The following graptolites have been collected from the shale unit
which is exposed some 250 feet below the ostracoderm occurrence
(fig. 59, Pts. A to B) : Monograptus vomerinus var. gracilis, M. prio-
don, M. acus, M. "spiralis," M. ex grupo duhius, Retiolites geinitzi-
anus, and Cyrtograptus cf. murchisoni var. bohemicus.

v. 5

htzULVt 7

Co

q>

c;

QJ

^

<»

-5?

.5

c>

■^

■C:

<o

5

-a?
5

;5

^

A3

'!^

^ .ti c -«

•^ E «

O

_. CO «« ^

. w ho

u rt o -2

Sc^
C8 3 V g

W o ^ £3
Q

pq'

0)

^ 2 § "

•SH;S .

' ^" .ti ^ o;
T3 a ^ ■

CQ

^j ti ID a) -55
O O S C8 >)

o +^ --i jS C;

^^ O ca tn (B

.-^ " m >

<^ ^ r.% *^

CO p^ '^ ^

»- "73 +2 "^ 2
^ S ^-2 <«

U C8 T3 .5 ^

^ =^« g^
*!IJ C t/j ni ^

oi i; -tj -s ■♦^

>> i O cfl

w-o; 3-3:3

O cS T3 "2

o P w JS So

^ «l*s g

O ti „ •« a;

fci CO CO p, to
03 S g CO

" 2 § s

oi S £ ~ b

"^ fl i ^ .2

rt G ^ ^
^ cu C CO

107

108 FIELDIANA: GEOLOGY, VOLUME 14

This fauna appears to span the equivalents of zones 24 to 26 of
the British graptolite succession, i.e., upper Llandovery and lowest
Wenlock, or the upper Clinton of the New York section.

Occurring in the same beds as the ostracoderms (fig. 59, Pt. C)
are: Monograptus sp. indet., fConchidium sp., and Atrypa sp.

The presence of the brachiopod fConchidium suggests a Late
Silurian age for these beds.

Monograptus dubius is found in the shale beds above the ostra-
coderm occurrence (fig. 59, Pt. E). The European range of this spe-
cies is Upper Wenlock through Middle Ludlow, i.e., late Middle to
Late Silurian.

All ostracoderm specimens have been deposited in the Princeton
University Geological Museum.

We wish to express our appreciation to The California Standard
Company for permission to publish this introduction to Dr. Deni-
son's paper.

VERTEBRATA: HETEROSTRACI

Family CYATHASPIDIDAE

Vernonaspis Flower and Wayland-Smith, 1952

The original description by Flower and Wayland-Smith (1952,
pp. 373-374) does not adequately define this genus and in many
respects is misleading. The shape of the dorsal shield of their speci-
mens has been so modified by crushing that many of the features
they describe have little significance. The anterior end is probably
rounded rather than obscurely pointed. The posterior margin has a
small, rounded, median lobe, rather than an obtuse point. The dif-
ferentiation of the rostral epitegum is common in Cyathaspididae,
rather than unique. There is almost certainly only a single pair of
lateral epitega, rather than two pairs. The absence of a clear differ-
entiation of lateral and central epitega is common in Cyathaspididae,
not unique. The pustules that modify the dentine ridges are un-
doubtedly the result of some pecuHarity of preservation, not a natural
characteristic. But Vernonaspis is probably a valid genus, though
it is similar to Archegonaspis. It is redefined as follows, incorporating
characteristics shown only by the new species from the Yukon :

Vernonaspis is a member of the Cyathaspididae in which the
dorsal shields are moderate to narrow in proportions (width/length=
.54-.64), the orbits are far forward (orbital length/total length=.10-

DENISON: SILURIAN HETEROSTRACI

109

.13), and the pineal macula has an anterior position (pineal length/
total length =.20-.22). The postbranchial lobes are long. There is

Fig. 60. Vernonaspis bamberi, type, dorsal shield, PU 17081 (X 2).

a median lobe on the rostral edge of the dorsal shield, and the pre-
orbital processes are strongly developed. The dentine ridges are of
moderate width or fine (5-9 per mm.). The rostral epitegum usually
has transverse ridges posteriorly, but its ornament is variable ante-
riorly. The pineal triangle is not distinct from the central epitegum.
The ridges of the central epitegum are essentially longitudinal.

no

FIELDIANA: GEOLOGY, VOLUME 14

Fig. 61. Vemonaspis bamberi, anterior part of dorsal shield of type, PU 17081
(X 4). lep, lateral epitegum; or, orbital notch; pm, pineal region; rep, rostral
epitegum.

Vernonaspis bamberi/ new species

Type. — PU 17081, a dorsal shield, nearly complete, though some-
what crushed and distorted (figs. 60, 61).

Referred specimens. — PU 17082, a dorsal shield, nearly complete
though split in half near the midHne; PU 17083^, incomplete dorsal
shields; PU 17375, the anterior part of a dorsal shield; PU 17086, a
ventral shield, incomplete anteriorly and folded sharply near the
midline; PU 17085, 17087, fragments of ventral shields.

Diagnosis. — The length of the dorsal shield is 43 to 46 mm. The
dentine ridges are 5 to 6 per millimeter. The posterior edges of the
shields are smoothly convex, without a median lobe.

Vernonaspis major, new species

Type.~P\J 17104, a complete dorsal shield (figs. 62, 63).

Referred specimens. — PU 17105, a nearly complete dorsal shield;
PU 17106, the anterior two-thirds of a dorsal shield; PU 17107, an
incomplete and badly crushed dorsal shield; PU 17091, 17376, frag-
ments of dorsal shields.

^ After E. W. Bamber, collector of these specimens.

DENISON: SILURIAN HETEROSTRACI

111

Fig. 62. Vernonaspis major, type, dorsal shield, PU 17104 (X 2).

Diagnosis. — The length of the dorsal shield is 57 mm. in the type.
The dentine ridges are 5 to 6 per millimeter. The dorsal shield has a
rounded median lobe on its posterior edge, and sharply marked post-
branchial processes.

Discussion. — These new species are distinguished by their larger
size and coarser dentine ridges from the previously described Ver-
nonaspis allenae and V. leonardi. The species from the Yukon have
a distinct median rostral process and strongly developed preorbital
and postbranchial processes. None of these features is clearly shown

o-'T^r"*

X

o

i§

>H

c

o

crt

nS

u

>~

ft

o

S

■•->

r/1

O

03

fl)

a.

O

ec

o

en

)-.

t^ s

^

bfl

-s

a

o

<u

fe nS

112

DENISON: SILURIAN HETEROSTRACI 113

in the original material of Vernonaspis, although there is reason to
believe that the preorbital and postbranchial processes, at least, are
similarly developed.

The rostral epitegum shows considerable variability in its ridge
pattern (figs. 61, 63, rep). Almost always there are transverse ridges
on its posterior part. In front of this, V. allenae and V. leonardi
show a fanlike arrangement of ridges directed anteriorly and antero-
laterally. The ridges in the Yukon species show a tendency to be
denticulate, coarse, and irregular anteriorly. There may be one or
more ridges parallel to the anterior margin. Commonly, circular
ridges surround the anterior pores of the supraorbital sensory canal
(fig. 63, soc).

In Vernonaspis major (fig. 63) there is no differentiation of a
pineal triangle, and this area, at least superficially, is incorporated
into the central epitegum. In PU 17104 and 17105, ridges of the
central epitegum curve medially, then postero-medially into the
pineal region. In PU 17106, central ridges continue longitudinally
through the pineal area. This region is not shown as well in the other
species, but apparently in none is there a distinct pineal triangle.
The type of V. allenae has an arrangement similar to PU 17104.
In the type of V. leonardi the anterior ridges of the central epitegum
curve antero-laterally to run parallel to the fanned lateral ridges of
the pineal region.

The type dorsal shield of Vernonaspis major is relatively broader
than in other Vernonaspis, the ratio of width/length being .63. The
breadth in this specimen has undoubtedly been increased by flatten-
ing. The ratio of width/length is estimated to be .55 in PU 17106.

Vernonaspis is most closely similar to Archegonaspis. The latter
differs in its broader shield, more posterior orbits and pineal macula,
absence of median process on the rostrum, weak development of the
postbranchial processes, and coarser dentine ridges (4.5-5 per mm.).
The rostral epitegum usually has a more completely transverse ridge
pattern, and the pineal triangle is distinct, with a fanned ridge pattern.

Ptomaspis,^ new genus

Type species. — P. canadensis.

Diagnosis. — The dorsal shield is of moderate proportions; its or-
bital notches are shallow, its rostrum is smoothly rounded, and its

1 From irrujua, carcass, and aiTTrts, shield.

114

FIELDIANA: GEOLOGY, VOLUME 14

posterior margin is gently convex. The dentine ridges include some
that are slightly higher between others that are lower and sometimes

Fig. 64. Ptomaspis canadensis, type, dorsal shield, PU 17090 (X 2).

slightly narrower; they are subdivided into short lengths or even into
small denticles. Posteriorly the ridges are grouped into scale-like
areas. The dorsal shield is divided by the pattern of ornamentation
into epitega: the rostral epitegum has short transverse ridges ante-
riorly, and grades through a denticulate area into antero-posterior
ridges in the pineal triangle; the central epitegum has a slightly ellip-
tical pattern.

Fig. 65. Ptomaspis canadensis, type, dorsal shield, PU 17090 (X 3). cd, pores
of transverse sensory commissure; Idl, pores of lateral dorsal sensory canal; mdl,
pores of median dorsal sensory canal; or, orbital notch; pm, pineal macula; rep,
rostral epitegum; soc, pores of supraorbital sensory canal.

115

116 FIELDIANA: GEOLOGY, VOLUME 14

Ptomaspis canadensis, new species

Type. — PU 17090, a nearly complete dorsal shield (figs. 64, 65).

Diagnosis. — The length of the dorsal shield is 61 mm.; the ratio
of width to length is .58. Dentine ridges average about 5 per milli-
meter.

Discussion. — Ptomaspis belongs to the group of Cyathaspididae
in which the dorsal shield is divided superficially into distinct epitega,
though its lateral epitega are not very sharply differentiated. The
postbranchial part of the shield is relatively long (the ratio of post-
branchial length to total length is .46), and the postbranchial lobes
are very slightly developed. The dentine ridge pattern, transverse
on the rostrum and slightly elliptical on the central epitegum, is
characteristic of this group of the Cyathaspididae. The subdivision
of the ridges into short lengths or even into denticles, and their
grouping into scale-like areas on the posterior part of the shield, dis-
tinguish Ptomaspis from other Cyathaspididae except Tolypelepis.
In Tolypelepis, however, the scale-like areas are much more distinct
and extend over the whole central epitegum and onto the pineal tri-
angle. Ptomaspis approaches the ridge arrangement of Archegon-
aspis and retains only a vestige of what was presumably the original
scale covering of the anterior part of the body. The pores of the
lateral line canals (fig. 65) are easily distinguishable and show that
the pattern of the sensory canals is similar to that of Tolypelepis
(Stensio, 1958, fig. 215) except that the lateral and medial dorsal
lines appear to be continuous. Parts of four transverse commissures
can be identified. Ptomaspis canadensis is one of the largest of the
Cyathaspididae, and is exceeded in size only by Allocryptaspis and
some species of Poraspis. The relative width of the type specimen
may have been increased somewhat by flattening.

Dikenaspis,^ new genus

Type species. — D. yukonensis.

Diagnosis. — The dorsal shield is moderately broad, the orbits are
placed relatively far back, the rostrum has a strong median lobe on
its anterior margin, and the postbranchial lobes are deep and trun-
cated anteriorly. The dentine ridges are fine. The dorsal shield is
divided by the ridge pattern into rostral, central, and paired lateral
epitega. The ridges are mostly transverse on the rostral epitegum,
and converge at the midhne anteriorly on the central epitegum and

^ From Slktiv, after the manner of; and aa-ins, shield.

DENISON: SILURIAN HETEROSTRACI 117

in the pineal area. The lateral line canals are branched in a complex
fashion. There are two pairs of lateral lines on the rostrum.

Dikenaspis yukonensis, new species

Type. — PU 17088, a nearly complete dorsal shield, and the pos-
terior part of a ventral shield (figs. 66, 67) .

Referred specimens. — PU 17089, the anterior half of a dorsal shield ;
PU 17377, a ventral shield.

Diagnosis. — The length of the dorsal shield is about 30 mm.; its
proportions are moderately broad, with width/length= .63. There
are approximately 9 to 10 dentine ridges per millimeter.

Discussion. — Dikenaspis shows so many similarities to Dinaspi-
della and Irregulareaspis that a relationship is probable. In all three
genera the dentine ridges are fine and the deep postbranchial lobes
of the dorsal shield are abruptly truncated anteriorly. They all have
the orbits placed relatively far posteriorly: orbital length/total length
is .17 in Dikenaspis and Irregulareaspis, and .20 in Dinaspidella; it
is less than .17 in other Cyathaspididae, except in some species of
Poraspis. The branching of the lateral line system in Dikenaspis
resembles that of Irregulareaspis (Kiaer and Heintz, 1932, fig. 9).

On the other hand, the presence of clearly distinct epitega in
Dikenaspis is presumably a primitive cyathaspid characteristic that
has disappeared in Dinaspidella and Irregulareaspis. On the rostral
epitegum (fig. 67, rep) the dentine ridges are convex posteriorly at
the sides and anteriorly in the middle; one or two ridges are parallel
to the anterior edge. The central epitegum has the ridges arranged
in a strongly elliptical fashion anteriorly. The pineal triangle is not
distinct, and this area is included in the elliptical central pattern.
The rostral region of the dorsal shield is narrow compared to its
branchial region. The anterior edge of the rostrum has a distinctive
shape, largely because of the strongly developed median lobe, which
is even more pronounced than in Vernonaspis. On either side, the
median lobe is separated by a deep notch from the powerfully de-
veloped preorbital processes. The orbital notches are quite deep.

The pores of the lateral line system are large and obvious. In
the type dorsal shield some of the canals themselves can be seen be-
cause of the transparency or removal of the superficial layer of the
shield (fig. 67, cd, Idl, soc). Where visible the canals have branched
so as to form a network approaching that of Irregulareaspis in com-
plexity. A striking feature of Dikenaspis is the presence of two pairs

Fig. 66. Dikenaspis yukonensis, type, dorsal shield and part of ventral shield,
PU 17088 (X 2).

118

DENISON: SILURIAN HETEROSTRACI

119

of sensory lines in the rostral region, instead of the single pair of
supraorbital canals of other Cyathaspididae. A possible explanation,

Fig. 67. Dikenaspis yukonensis, dorsal shield of type, PU 17088 (X 4).
cd, branched transverse sensory commissure; Idl, lateral dorsal sensory line; mdl,
pores of median dorsal sensory line; or, orbital notch; pc, pore of ? profundus
sensory line ; pm, pineal area ; rep, rostral epitegum ; soc, ?supraorbital sensory canal.

suggested by Stensio (1958, p. 401), is that the medial pair (fig. 67,
soc) represents the supraorbital canals of modern fishes, and that the
lateral pair (fig. 67, pc), which passes immediately above the orbits,
is innervated by the profundus nerves.

The ventral shield, PU 17377, agrees with Dikenaspis yukonensis
in size (length=26.1 mm.), proportions (width/length=.67), and fine

120 FIELDIANA: GEOLOGY, VOLUME 14

dentine ridges (8 or more per mm.). The anterior border is concave,
and the posterior border is convex with a rounded median point.
The lateral borders do not show deep incisions for branchial plates,
such as occur in Dinaspidella and Irregulareaspis (Kiaer and Heintz,
1932, p. 16). The shield is deeply arched posteriorly, but anteriorly
it is nearly flat. Posteriorly the dentine ridges are arranged longi-
tudinally, but in the anterior half they converge toward the midline.
The most anterior and lateral ridges are, except for occasional irregu-
larities, continuous around the anterior and lateral parts of the shield.
They are concentric around the inner part of the shield, rather than
conforming to the outer contour. This arrangement is reminiscent
of that in Pteraspididae.

Ariaspis,^ new genus

Type species. — A. ornata.

Diagnosis. — The dorsal shield is relatively broad and weakly
vaulted. There is no median process on the rostral margin, but the
preorbital processes are strongly developed. Narrow lateral laminae
extend from the orbits to the branchial openings and are separated
from the rest of the shield by sharp downward angulations. There
are no postbranchial lobes, and the postero-lateral corners of the
shield are small points. The posterior edge of the dorsal shield is
nearly transverse except for the incorporation of a small, median,
projecting scale. No separate epitega are distinguishable.

Ariaspis ornata, new species

Type.—I'U 17103, a complete dorsal shield (figs. 68, 69).

Diagnosis. — This is a small species, with the length of the dorsal
shield 21.5 mm. The ratio of width/length is .72. The dentine
ridges are fine, 7.5 to 8 per millimeter, and have sharply or roundly
convex crests.

Discussion. — The single but well-preserved specimen referred here
represents a genus that is clearly distinct from other Cyathaspididae.
Comparable lateral laminae are otherwise known only on the dorsal
shields of Allocryptaspis and Ctenaspis. In Ariaspis these laminae
form very narrow downwardly directed rims, and can hardly rep-
resent fused branchial plates, as appears to be the case in Allo-
cryptaspis. As in Ariaspis, postbranchial lobes are also reduced
or absent in Anglaspis and Ctenaspis. Because of the absence of

1 From api, denoting excellence; and acrwis, shield.

DENISON: SILURIAN HETEROSTRACI

121

these lobes and branchial notches, it is impossible to determine in
Ariaspis the precise position of the branchial openings, but presum-

ably they were close behind the terminations of the lateral laminae.
The median scale that is incorporated in the posterior edge of the
shield is probably unique in Cyathaspididae; it is comparable to
the dorsal spine of Pteraspididae.

In addition to these specializations, the absence of distinct epitega
shows that this is an advanced genus. The ridge pattern is compar-
able to that in Homalaspidella and some Poraspis. A number of
ridges are parallel to the anterior edge. Irregular ridges lie between
these and the orbits. The pineal macula (fig. 69, pm) has a prominent,
oval ridge pattern. The rest of the ridges are arranged antero-poste-
riorly near the midline, and pass into an elHptical pattern toward the
lateral margins. These ridges are long and continuous except near

122

FIELDIANA: GEOLOGY, VOLUME 14

the posterior margin, where a distinct band appears to represent a
separately formed part of the shield.

Fig. 69. Ariaspis ornata, type, dorsal shield, PU 17103 (X 6). ms, median
dorsal scale; or, orbital notch; pm, pineal macula.

The pores of the lateral line system are small, and the pattern of
the canals has been determined only so far as it has been made vis-
ible by immersion in alcohol and aniseed oil. The medial dorsal
sensory lines are rather widely spaced and appear to be broken into
short lengths. The anterior three transverse commissures are repre-
sented by short medial and lateral canals; the fourth was not visible.

DENISON: SILURIAN HETEROSTRACI 123

Apparently there are no striking peculiarities of the lateral lines in
Ariaspis.

Homalaspidella Strand 1934

This genus, originally described as Homalaspis by Kiaer and
Heintz (1932, p. 14), has hitherto been known only from the Early
Devonian, Red Bay series, of Spitsbergen. It includes small forms
whose dorsal and ventral shields are moderately narrow, rather uni-
form in breadth, and arched relatively little. The rostral edge lacks
a median process, the orbital notches are shallow, and the preorbital
length is short. The dentine ridges are flat-topped, except perhaps
near the lateral margins. Epitega and hypotega are hardly or not
at all distinct. On the dorsal shield, there is a belt of nearly con-
tinuous ridges parallel to the anterior and lateral edges; central ridges
are more or less longitudinal and may continue through the pineal
area. Dorsal transverse commissures of the sensory canal system
are poorly developed.

Homalaspidella borealis, new species

Type. — PU 17101, a dorsal shield, incomplete posteriorly (figs.
70-72).

Referred specimens. — PU 17102, an imperfect but articulated spec-
imen, showing an incomplete dorsal shield, a branchial plate, and
some scales of the tail; PU 17378-81, incomplete dorsal shields;
PU 17092 (fig. 73), 17382, ventral shields, essentially complete;
PU 17383, the posterior part of a ventral shield.

Diagnosis. — The dorsal shield has a length of 28 to 30 mm., and
a ratio of width/length of .54. The rostral edge is nearly transverse.
The preorbital processes and orbital notches are very slightly de-
veloped, and the orbits are far anterior. The ratio of postbranchial
length to total length is .33-.37. The posterior edge of the dorsal
shield is gently convex and lacks a median lobe. The dentine ridges
are moderately coarse, 5.2-6.2 per millimeter on the dorsal shield,
and 5.0-5.4 per millimeter on the ventral shield.

Discussion. — Homalaspidella borealis is presumably much older
than the type and only other species, H. nitida. However, the dif-
ferences, most of which are indicated in the diagnosis, hardly warrant
a generic distinction. The ratio of pineal length to total length is
.25-.26, about as in the type species. The ratio of orbital length to
total length is .10-. 12, slightly less than in the type species, where it
is .14; this difference may be accounted for largely by the greater

124

FIELDIANA: GEOLOGY, VOLUME 14

convexity of the rostral margin in the type species. The preorbital
processes and orbital notches are distinct though small in the type

Fig. 70. Homalaspidella borealis, type, incomplete dorsal shield, PU 17101 (X 4).

species, but are barely indicated in H. borealis. The posterior part
of the dorsal shield is differently shaped in the two species : in H. ni-
tida the postbranchial lobe is short, but the posterior edge is extended
by a pronounced lobe; in H. borealis the postbranchial lobe is longer,
but the postero-median edge is only slightly convex. No other cya-
thaspids, except Allocryptaspis and Ctenaspis, have gone as far in
reducing all traces of distinct epitega. The dentine ridges of H. bore-
alis (fig. 71) tend to be sinuous and discontinuous with many inter-
calations. PU 17379 (and to a less extent other specimens) has

Fig. 71. Homalaspidella borealis, type, incomplete dorsal shield, PU 17101
(X 6). or, position of orbit.

125

126

FIELD lANA: GEOLOGY, VOLUME 14

most lateral line pores set in very short ridges, one pore per ridge.
These peculiarities of ridge arrangement occur to some extent in

Fig. 72. Homalaspidella borealis, type, incomplete dorsal shield, PU 17101
(X 4). cd Y_3, transverse commissures of sensory system; Idl, mdl, lateral and
medial dorsal sensory canals; pm, pineal area; soc, supraorbital sensory canal.

H. nitida. Of the dorsal sensory canals of H. borealis (fig. 72), not
only the transverse commissures but also the median longitudinal
canals are divided into short lengths.

The ventral shield of H. borealis (fig. 73) has a length of 25 to
29 mm. and a ratio of width/length of .58. The anterior edge is
transverse or concave, and the posterior edge is convex with a
rounded median point similar to that figured for H. nitida by Kiaer
and Heintz (1935, fig. 55, C). The ridge pattern of the ventral
shield is similar to the dorsal pattern, with a few anterior transverse
ridges that continue around the lateral margins, and with longi-

DENISON: SILURIAN HETEROSTRACI

127

tudinal central ridges that become irregular antero-medially behind
the transverse ridges. Distinct hypotega are not indicated.

Fig. 73. Homalaspidella borealis, ventral shield, PU 17092 (X 4).

Undetermined Cyathaspididae

Two incomplete dorsal shields do not belong to the genera de-
scribed above. PU 17384 (fig. 74) preserves the left side to or nearly
to the midline, from the posterior edge anterior as far as the orbit.
Its total length is estimated to be more than 60 mm., making this
one of the largest Cyathaspididae. It has about six flat-topped den-
tine ridges per millimeter, arranged in an essentially longitudinal
pattern, or somewhat diagonally on the lateral epitegum. The ridges
tend to be short and slightly irregular, with many intercalations,
especially posteriorly. The posterior edge is nearly transverse, prob-
ably with a slight median lobe. The postbranchial lobe is well de-
veloped and sharply demarked; its dentine ridges terminate against
its anterior edge rather than running parallel to it. This specimen

128

FIELDIANA: GEOLOGY, VOLUME 14

is of interest in suggesting a subdivision of the shield, indicated by-
shallow grooves on the surface, and by breaks or irregularities in the
dentine ridge pattern. The elements are: (a) a central disc (fig. 74,

Fig. 74. Cyathaspid indet., part
of dorsal shield, viewed dorso-later-
ally, PU 17384 (X 2); only enough
ridges are shown to indicate the pat-
tern, bro, branchial notch; ca, cen-
tral area; cd, pores of transverse sen-
sory commissures; la, lateral area; Idl,
mdl, pores of lateral and medial dorsal
sensory canals; or, posterior part of
orbital notch; pbr, postbranchial lobe;
pla, postero-lateral area; sa, scale-like
posterior band.

CO); (6) a lateral area corresponding to the posterior part of the
pteraspid orbital plate (fig. 74, la) ; (c) a postero-lateral area, extend-
ing between a and b, bounding b posteriorly, and forming the post-
branchial lobe (fig. 74, pla); and (d) a scale-like band along the
posterior border (fig. 74, sa).

PU 17385 is a dorsal shield lacking much of the central and right
antero-lateral region. Its estimated total length is 39 mm. It is dis-
tinguished by its exceptional relative breadth; its ratio of width/
length is about .77. It has 7}/^ to 8 dentine ridges per millimeter,
and these are sinuously longitudinal postero-medially and denticu-
late on the rostral epitegum.

DENISON: SILURIAN HETEROSTRACI

129

There are four undetermined cyathaspid ventral shields. An im-
perfectly preserved anterior end of a ventral shield, PU 17094, is
highly arched and has a narrow concavity on its anterior margin.

Fig. 75. Cyathaspid indet., ventral shield, PU 17387 (X 2).

There are anterior and lateral marginal areas, separated from the
central area both by their dentine ridge pattern and by a surface
swelling that runs parallel to the edge at a distance of about 3.5 mm.
On the anterior marginal area the dentine ridges are parallel to
the concave margin anteriorly, but posteriorly they form rows of
denticles that converge postero-medially. On the anterior ends
of the lateral marginal areas the ridges are parallel to the margin
of the shield. These ridges are very fine, about 8 per millimeter.
On the central area the ridges are somewhat coarser, averaging about
6 per millimeter and are arranged elliptically. They are broken into
short lengths or even into small denticles antero-medially, and in-
clude some slightly coarser and higher ridges among the others.

Fig. 76. Cyathaspid indet., ventral shield, PU 17387 (X 4). cvl, cvm, pores
of lateral and medial ventral transverse commissures; Ivl, pores of lateral ventral
sensory canal; pol, pores of postoral canal.

130

DENISON: SILURIAN HETEROSTRACI 131

PU 17093 is a ventral shield lacking the posterior edge and much
of one side. It was associated with PU 17088, Dikenaspis yukonen-
sis, but does not belong to Dikenaspis. Its estimated length is
34 mm. The dentine ridges, of which there are about 5 per milli-
meter, are arranged longitudinally with a slight flaring anteriorly,

PU 17387 (figs. 75, 76) is a ventral shield that is extremely inter-
esting in having a ridge pattern similar to that of Pteraspididae.
In the central part the arrangement is elliptical, with the ridges con-
verging toward the midline at either end. The more anterior and
lateral ridges are continuous around the central ellipse except at the
posterior edge, which they meet at a steep angle. The course of
these outer ridges is determined by the ridges medial to them, and
the outer ridges are not necessarily parallel to the outer edge of the
shield. Such a ridge pattern is typical of Pteraspididae, though it
is approached in Cyathaspididae by Dikenaspis and Cyathaspis. The
lateral line pattern, however, is typical of Cyathaspididae. It is in-
dicated by the pores (fig. 76), which are arranged in a series of V's,
pointed posteriorly in the midline. These may be identified as the
paired postoral canals (fig. 76, pol), and six pairs of median ven-
tral transverse commissures (fig. 76, cvm) . The paired lateral ventral
canals (fig. 76, Ivl), and a few lateral segments of the ventral trans-
verse commissures (fig. 76, cvl) are also indicated. Though some-
what crushed, this shield was originally highly vaulted posteriorly.
Its dimensions are: length in midline=37.5 mm.; maximum width=
26.5 mm. The anterior edge is concave; the posterior edge is gently
convex. The dentine ridges are flat-topped, closely spaced, and 5-
5M per millimeter.

PU 17386 is a ventral shield lacking only the most anterior part.
Its length is estimated to have been 30 mm. when complete. The
posterior end is convex with a broadly rounded median lobe. The
dentine ridges are 6-7 per millimeter, or still finer laterally. Their
pattern is essentially longitudinal, and the ridges are broken into
relatively longer lengths anteriorly and shorter lengths postero-medi-
ally. Many short ridges are intercalated. The pores of the lateral
line system are relatively large and indicate a canal pattern similar
to that of Poraspis. There is nothing to exclude this ventral shield
from Poraspis, but a definite assignment does not seem warranted.

PU 17099 (fig. 77) is identified as a cyathaspid branchial plate.
It is complete except for one corner and has a length of 25.0 mm.
and a maximum depth of 6.3 mm. Its top and bottom edges are
nearly parallel, though it tapers slightly toward each end from its

132

FIELDIANA: GEOLOGY, VOLUME 14

greatest depth near the middle. One end, probably anterior, is
truncated slightly obliquely; the other end is bluntly pointed. The
dentine ridges are about 5 per mm., and are arranged mostly longi-
tudinally, although some are diagonal on the half with the truncated

Fig. 77. Cyathaspid indet., branchial plate, PU 17099 (X 4).

end. This resembles the branchial plate of Archegonaspis cf. lind-
stromi figured by Heintz (1933, fig. 3), except that there is no bran-
chial notch. It might be the branchial plate of Vernonaspis bamheri.
A number of isolated scales are referable to the Cyathaspididae
but are not generically identifiable. Most of them are the long dorso-
lateral scales, and they agree in general with such scales in Poraspis,
as described by Kiaer and Heintz (1935, pp. 112-116). In detail,
however, they show many variations. A number of them have, as
is usual in Poraspis, several long ridges arranged dorso-ventrally or
parallel to the anterior border; commonly some of these ridges curve
and pass into the longitudinal, or antero-posteriorly arranged ridges.
In other cases the dorso-ventral ridges are separated from the longi-
tudinal ridges by an area of short ridges. In other dorso-lateral
scales, the dorso-ventral ridges are absent and the anterior border
is denticulate; this condition is also known in Poraspis. One scale
has an anterior finely denticulate band, then a band of short ridges;
the longitudinal ridges are limited to the posterior third of this scale.
A few ventro-lateral and caudal scales show nothing remarkable.
Two specimens of Homalaspidella borealis (PU 17102, 17383), for-
merly attached to each other, show remains of two articulated tails.
Unfortunately, the scales are incompletely and poorly preserved, and
their arrangement and structure cannot be determined.

Family TRAQUAIRASPIDIDAE

PTraquairaspis Kiaer 1932

The traquairaspid from the Yukon possibly belongs to a new
genus, but the one specimen is not adequate to distinguish it.

DENISON: SILURIAN HETEROSTRACI

133

Fig. 78. ?Traquairaspis angusta, type, ventral disc, PU 17388 (X 3/2).

PTraquairaspis angusta, new species

Type.~F\J 17388, a ventral disc (figs. 78, 79). This specimen
was obtained in two pieces without any contact. They surely be-
long together, and have been united with the spacing determined
approximately by the contour of edges and surfaces.

Diagnosis. — The dimensions of the ventral disc are: estimated
length= 80-85 mm.; maximum width=35.5 mm. The disc is rela-
tively narrow, with the estimated ratio of width/length= .42-.44.

134 FIELDIANA: GEOLOGY, VOLUME 14

The smooth central area covers nearly all of the ventral disc, and its
ornamented margin is narrow.

Discussion. — The ventral disc of fTraquairaspis angusta is longer
than those of T. pococki and T. campbelli, and somewhat shorter
than that of T. symondsi. It is relatively much narrower than in
the other species, in which the ratio of width/length ranges from
.58 to .70. The ornamented margin covers only a minor part of the
disc of ?T. angusta, while it is broad in T. pococki and T. symondsi;
T. campbelli is intermediate in this respect.

The anterior edge of the ventral disc is gently rounded. Its lat-
eral edges are slightly concave anteriorly and then are convex, reach-
ing the maximum width just in front of mid-length, and from there
tapering gradually posteriorly. The posterior edge is biconvex with
a median notch.

The central area is smooth except for pores of the lateral line
system, which also occur in T. campbelli, and a number of grooves.
One pair of irregular grooves (fig. 79, gr) extends from the posterior
edge forward for two-thirds the length of the disc; these grooves have
a number of short median and lateral branches. In the anterior
part, a small circular area is marked by a groove, and small areas
are separated from the main smooth area by grooves antero-laterally
and postero-laterally. These grooves do not appear to be related to
the lateral line canals. Their scalloped edges resemble the margins
of the smooth area and of the tubercles of the brim. They may be
considered as demarcations of separately formed areas of the super-
ficial layer.

The ornamented brim is relatively narrow. Anteriorly it is 2.5-
3.5 mm. wide and continues the plane of the adjacent smooth area.
Laterally it is 5 mm. or less in width and is turned dorsally at a
sharp angle from the smooth area. Posteriorly the brim is absent.
On the anterior brim the tubercles are crowded and short, elongate
parallel to the margin, and have many prominent side processes.
There are no ridges between the tubercles here. Around each antero-
lateral corner, these tubercles pass into a band marked by numerous
low, fine, close ridges, elongated parallel to the edge; some, but not
necessarily all of these ridges are crowned with dentine. A single
row of high tubercles extends backward in this band from the antero-
lateral corner parallel to the edge of the smooth area; the individual
segments are short anteriorly and very long posteriorly; the side
processes of these tubercles are relatively short. This combination
of high tubercles and low ridges is reminiscent of the ornamentation
of T. pococki and T. campbelli.

DENISON: SILURIAN HETEROSTRACI

135

The lateral line system is indicated by a number of relatively
large pores. A group of pores arranged in a V on the anterior part

cvm

Fig. 79. ?Traquairaspis angusta,
type, ventral disc, PU 17388 (X 3/2).
cvl, cvm, pores of lateral and medial
parts of transverse sensory commis-
sures; gr, groove in smooth central
area of disc; Ivl, pores of lateral ven-
tral sensory canal; pol, pore of post-
oral canal.

of the smooth area represents the paired postoral canals (fig. 79, pol) .
Behind them are pores of the medial parts of three ventral transverse
commissures (fig. 79, cvm), and three more posterior pores may indi-
cate the fourth and fifth commissures. Lateral parts of three trans-
verse commissures (fig. 79, cvl) are indicated by one or two pores
each. Pores piercing the separated areas at the antero-lateral cor-
ners of the smooth area indicate the anterior parts of the lateral
ventral canals (fig. 79, Ivl) . No pores are to be seen on the posterior
part of the disc. The pattern is much as in T. campbelli (White, 1946,
figs. 36-38).

136 FIELDIANA: GEOLOGY, VOLUME 14

HETEROSTRACI OF UNDETERMINED FAMILY

In addition to ?Traquairaspis, there are a number of specimens
with coarse ornament which are referred to Heterostraci, but which
cannot be related with certainty to previously known families or gen-
era. Based on the ornament, these may be grouped into four types.

Fig. 80. Heterostraci indet.
Type A, marginal plate, PU 17390
(X 1). A, probable dorsal face; B,
probable ventral face.

Heterostraci indet., Type A, has tubercles whose margins are
crenulated or provided with side processes. These tubercles may be
long (to 21 mm.) and slender (.5 mm.), in which case they are widely
spaced (figs. 80, A, 81, B) ; or they may be short and broad (to
1.5 mm.), in which case they are crowded closely together (figs.
80, B, 81, A). All intermediates occur between these tubercle types,
and both types may occur on the same specimen (fig. 80). In the
spaces between the narrower tubercles there are one to several fine,
low ridges, arranged parallel to the tubercle margins (fig. 82, A).
Only between the broadest, most closely packed tubercles are these
interstitial ridges lacking; elsewhere there may be one to five of them.
They are irregular, and adjacent to the tubercles they are usually
zigzagged, with side processes at the angulations.

This type of ornament is comparable to that on the lateral brim
of the ventral disc of fTraquairaspis angusta. It differs in having
more closely spaced side processes on the tubercles, and perhaps in
the character of the interstitial ridges, which are not well preserved
in the type of ?T. angusta. Three of the specimens (PU 17390-92)
are fragments of plates of the lateral margin, and in one, PU 17390
(fig. 80), the plate has two laminae, representing parts of the dorsal

D

Fig. 81. Heterostraci indet. (X 1). A, Type A, PU 17095; B, Type A,
PU 17098; C, Type B, PU 17097; D, Type D, PU 17096.

137

138 FIELDIANA: GEOLOGY, VOLUME 14

and ventral shields. The margins of PU 17391 and 17392, and the
narrow lamina of PU 17390 have short, broad tubercles, while the
rest of the plates have long, narrow, widely spaced tubercles. These
plates cannot be identified with any plates of Traquairaspis, as re-
stored by White (1946, figs. 40-44). PU 17098 (figs. 81, B, 82, A) is
a fragment with long, slender tubercles, and with one row of short,
forked tubercles; some of the latter are notched by pores of the lat-
eral line system. PU 17095 (fig. 81, A) is a small, symmetrical,
presumably median plate; its tubercles are short and broad, and
usually there are interstitial ridges in the interspaces.

Heterostraci indet.. Type B, represented by a shield fragment,
PU 17097 (fig. 81, C), is possibly a variant of Type A. Near one
edge it has long, narrow tubercles, as little as .5 mm. in width, and
provided with small side processes. These tubercles are separated
by interspaces as wide as 1.2 mm., in which there are as many as 9
fine, low ridges. Toward the other edge of the plate the tubercles
become wider and the interspaces narrower, until finally the elongate,
flat-topped tubercles are as much as 2 mm. wide, and the interspaces
are very fine grooves. Except for its coarseness, the latter type of
ornament resembles that of some Cyathaspididae. PU 17100, an
incomplete, narrow ridge scale, may belong to this type.

Heterostraci indet., Type C, is represented by a single incom-
plete plate, PU 17389 (fig. 82, B). Its tubercles are elongate (to
9 mm.) and slender (.3-.7 mm.) and, in contrast to those of the other
types, have smooth margins without crenulations or lateral processes.
A few of the tubercles are short and forked. The tubercles are sep-
arated by interspaces of 1.0-2.5 mm. in which are numerous (4-10)
low, simple, relatively straight ridges. Pores of two lateral line
canals open between the fine ridges, and in some cases notch the
tubercles.

Heterostraci indet., Type D, is typified by PU 17096 (figs. 81, D,
82, C) . This fragment is covered for the most part with short, broad,
ovoid tubercles, but these grade toward one edge into an ornament
of narrow tubercles with very prominent side processes (fig. 82, C),
similar to those of Traquairaspis symondsi or T. plana (Orvig, 1961,
figs. 2-3). The spaces between the tubercles completely lack the
fine ridges of the other types. Similar tubercles without interstitial
ridges are found in the marginal plates of Cardipeltis and in the frag-
ment from the Early Devonian of Podolia figured by Orvig (1961,
fig. 4) as "gen. et sp., indet." A few scales (PU 17394-95), and a
small plate (PU 17393) are also referable to this type.

li^ltaL^!^

Fig. 82. Heterostraci indet., ornamentation (X 5). A, Type A, PU 17098-
B, Type C, PU 17389; C, Type D, PU 17096.

139

140 FIELDIANA: GEOLOGY, VOLUME 14

GEOLOGIC AGE

Vernonaspis, the commonest form in the Yukon occurrence, is
known otherwise only in the Vernon Shale of New York State, a
formation of the Salina Group of the Late Silurian. A similar genus,
Archegonaspis, occurs in the nearly equivalent Graptolithengestein
and the Hemse Group of northern Europe, both belonging to the
Monograptus nilssoni zone of the Early Ludlow. Ptomaspis, in re-
gard to the specialization of its shield, is intermediate between To-
lypelepis and Archegonaspis. The former occurs in the Late Silurian
of the Baltic region, and in the Early Wenlockian of the Canadian
arctic. Dikenaspis is most closely related to Dinaspidella and Irregu-
lareaspis, both Early Devonian (Gedinnian) genera from Spitsbergen
and Podolia. However, it is beheved to be more primitive than the
Devonian genera. Ariaspis is a specialized cyathaspid and, though
not nearly related to other known genera, is suggestive of a De-
vonian rather than Silurian age. Homalaspidella is otherwise known
only from the Ben Nevis Group of the Red Bay Series of Spits-
bergen; the latter is correlated with the Gedinnian. Among the
undetermined Cyathaspididae there is a dorsal shield that is clearly
specialized in its large size, and a ventral shield specialized in its
pteraspid-like ridge pattern; either is suggestive of Early Devonian
rather than Silurian age. Traquairaspididae are otherwise known
from Downtonian beds at the base of the Devonian in Great Britain,
and from strata of equivalent age in Spitsbergen and Germany.

In summary, the vertebrates indicate either a Late Silurian or
an Early Devonian age. The invertebrates identified by Lenz (p. 108)
suggest an earlier dating. The most significant of them for correla-
tion is the Monograptus dubius from beds overlying the vertebrate
horizon. This species is a long-lived one in Europe, ranging from
the Wenlock (zone 27 of the British section) into the Middle Ludlow
(zone 35 or 36). The end of its known range overlaps the suggested
vertebrate dating, and so Late Silurian, either Early or Middle Lud-
low, may be taken tentatively as the probable age of this fauna, as
estimated from the vertebrates. If this is correct it will extend
downward the range of Homalaspidella and the Traquairaspididae,
and suggest earlier specializations in certain Cyathaspididae.

ECOLOGY

Very little information is available upon which to base an eco-
logical interpretation of this occurrence. The vertebrates are de-

DENISON: SILURIAN HETEROSTRACI 141

scribed as occurring in dolomites that are interbedded with lime-
stones and graptolitic shales. The graptolitic shales are marine, and
the dolomites contain, in association with the vertebrates, articulate
brachiopods, cryptostomatous bryozoans, crinoids, scolecodonts, and
doubtless other invertebrates as yet undetermined. There can be
no question as to the marine deposition of these dolomites. It is
important to note that the vertebrates are not rare, but abundant
and varied, at least in the few small samples from which the present
collection was obtained. It seems to be an unavoidable conclusion
that these Heterostraci were also marine forms. They are another
example of the so-called "Cyathaspidinae-Acanthodii" assemblage
which, during the Silurian, was largely or entirely restricted to ma-
rine deposits (Denison, 1956, pp. 376-387). The rarity of Hetero-
straci in most Silurian marine sediments is probably to be explained
in large part by their special and restricted habitats at this time.
What these habitats were cannot be determined now, but it is pos-
sible that, when better known, the present occurrence in the Yukon
will give some clue.

REFERENCES

Denison, Robert H.

1956. A review of the habitat of the earliest vertebrates. Fieldiana: Geol., 11,
no. 8, pp. 359-457.

Flower, R. H., and Wayland-Smith, R.

1952. Cyathaspid fishes from the Vernon Shale of New York. Bull. Mus.
Comp. Zool., 107, pp. 355-387, figs. 1-2, pis. 1-8.

Heintz, Anatol

1933. Neuer Fund von Archegonaspis in einem obersilurischen Geschiebe.
Zeitschr. Geschiebefors., 9, Heft 3, pp. 123-131, figs. 1-5.

KiAER, J., and Heintz, A.

1932. The Downtonian and Devonian vertebrates of Spitsbergen. IV. Sub-
order Cyathaspida. Skr. om Svalbard og Ishavet, no. 52, pp. 1-26, figs. 1-12,
pis. 1-11.

1935. The Downtonian and Devonian vertebrates of Spitsbergen. V. Sub-
order Cyathaspida. Part I. Tribe Poraspidei, Family Poraspidae Kiaer.
Skr. om Svalbard og Ishavet, no. 40, pp. 1-138, figs. 1-57, pis. 1-40.

Orvig, 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.

Stensio, Erik

1958. Les cyclostomes fossiles ou ostracodermes. Traite Zool., 13, fasc. 1,
pp. 173-425, figs. 107-219.

White, Errol I.

1946. The genus Phialaspis and the 'Psammosteus Limestones.' Quart. Jour.
Geol. Soc. London, 101, pp. 207-242, figs. 1-60, pis. 12-13.