Royal physical society of old cat.

Proceedings online

. (page 19 of 44)
Online LibraryRoyal physical society of old catProceedings → online text (page 19 of 44)
Font size
QR-code for this ebook

largely aquatic in their habits, or four, if CriodrUus be

It may be, perhaps, a mere coincidence that in two at
any rate out of the four aquatic forms mentioned in the
present note, there is a certain approximation in structure
to limicoline genera, although this approximation is not in the
direction of any special adaptation to an aquatic life — at any
rate not as far as we can see at present.

The progress of recent investigation into the structure of
OligochflBta has broken down all the distinction between the
"Terricolte" and "Limicolae" except one, which has been,
on the contrary, confirmed. That distinction is seen in the
ova. In the Naidomorpha, Enchytrseidaj, Lumbriculidje,
Tubificidffi, etc., the ova are few and large, the large size being
due to the increased quantity of yolk stored up in the ovum;
in earthworms, on the contrary, the ova are minute with
a very small quantity of yolk. To a certain degree AUurus
and AcaifUhodrilits Dalei ^ are intermediate in character ; the
ova are much larger than those of earthworms in general,
but are much smaller than the ova — ^heavily laden with yolk
— of such forms as Tuhifex. Without attempting — for the
present — to decide whether the large yolk ovum is or is not
the original condition in the OligochsBta, it is clear that as
both kinds of ova occur in allied forms, one must have been
derived from the other during the evolution of the group;
and it is therefore permissible to regard the large ova of
AUurus and especially of AcanthodrUus as intermediate in

' Named after Dr Dale of the Falkland Islands Company.


ized by Google

On the Structure of Coccosteus decipiens, Agassiz, 211

XXII. On the 'Structure 0/ Coccosteus decipiens, Agassiz.
By Dr R H. Tkaquair, F.E.S., F.G.S. [Plate XI.]

(Read 18th December 1889.)

In a paper (13) on Homosteus published in the Proceedings
of this Society for Session 1888-89, I entered into the
structure of Coccosteus so far as was necessary for the purpose
of instituting a comparison between the two genera. In the
present communication I propose to consider the structure of
Coccosteus in greater detail

The figure which I gave in that paper of the cranial shield
is reproduced in PL XI., Fig. 2, with the addition of the dorsal
cuirass. It is, I believe, accurate, and represents the result
of a close study of a very great number of heads. Compara-
tively few specimens are, however, available for the purpose,
those especially from Lethen and most of those from Orkney
being ill-adapted for following the sutures separating the
plates, while Cromarty and Edderton furnish those in which
the surface is most perfectly preserved, thus affording the
best opportunity for accurately distinguishing the true sutures
from those superficial grooves which in past times have been
so often confounded with theuL Quite recently, however,
the Edinburgh Museum has acquired a small collection of
Coccostev^'TemsLins from Stromness, in Orkney, in which the
details of the surface of the cranial plates are most beauti-
fully shown, and are entirely corroborative of the sketch
which I published a year ago.

As I have previously stated (12, p. 511), I retain only two
species of Coccostetts from the Scottish Lower Old Bed Sand-
stone, namely C. decipiens, Ag., and C. minor, H. Miller, the
difTerences which have led to the separation of "oblongtcs"
Ag., ** ciLspidcUus*' Ag., microspondylus, trigonaspis, and
pusillus, M*Coy, and Milleri, Egert., being dependent either
upon the mode of preservation or upon trivial variations in
the shape of certain plates, which are extremely common up
to certain limits. That which I find especially difficult to
understand is how Prof, von Koenen (10) should propose to


ized by Google

212 Proceedings of the Boycd Physical Society,

remove C. Milleri, Egert, and C. pusUlns, M*Coy, from
CoccostetLS altogether, placing them in Brachydeirus, the Cact
being that they are simply synonyms of decipiens, Ag. C,
minor, H. Miller, once mixed up with C. pusUlus, M*Coy,
may possibly have to be put into a new genus on account
of the structure of the vertebral column, which presents an
appearance as if possessed of ossified centra ; ^ but I can see
no reason for associating this species with v. Koenen's

The following description of the structure of the bony
skeleton of Coccostevs is therefore based upon an examination
of the common and well-known species G, d^dpiens, Agassiz.

Head, — In PL XL, Fig. 2,*the bones forming the cranial
shield are sketched, as well as the ramifications of the lateral-
line grooves. These bones are : — one median occipital (m. a),
two external occipitals {e. o.), two central plates (c), two
marginals (m,), two post-orbitals {pt o.), two pre-orbitaJs (p, a),
one posterior ethmoidal {p, e.), and one anterior ethmoidal
(a. e.), between which last and the premaxill» (p. ttwx) the
nasal openings (n.) are observable. I have already (13, p. 52)
explained that I have applied those names without the inten-
tion of considering any of the bones exact equivalents of
bones similaiLy named in ordinary fishes.

The orbit, the upper margin of which is formed by the
excavated outer edges of the post- and pre-orbital buckler-
plates, is completed below by the superior maxillary bone
(mx., Fig. 1), which strongly resembles in shape that of
typical PalseoniscidsB in being broadly expanded behind,
where it covers the cheek, and suddenly excavated to form
a tapering process directed forwards under the eye to the
premaxilla. To the posterior margin of the maxilla is fixed
the jugal or post-maxilla, a triangular plate with posteriorly
directed apex, which fills up the space between the maxilla
and the lateral part of the body-cuirass.

So far as I can see, the maxilla of Coecosteus decipiens does
not seem to have borne any teeth. But in a specimen from
Gamrie in the Edinburgh Museum there is distinct evidence

* This is apparently the species **with a true bony vertebra" referred
to by Murcbison in '*Siluria," 8rd ed., p. 604.


ized by Google

On the Structure of Coccosteus decipiens, Agctssiz. 213

of the presence of both vomerine and palatal teeth. The
specimen lies on its back, giving a beautiful view of the
ventral cuirass, in front of which are the two rami of the
mandible converging to meet each other anteriorly, while
external to and in front of them the upper margin of the oral
cleft is seen formed by the maxillae and premaxillae. No
teeth are, as usual, seen on the maxillae, but internal to them
and between them and the contiguous mandibular ramus is
seen a row of conical teeth, evidently placed on the edge of a
palatal or palato-pterygoid bone, which I have not yet seen
in its entirety. Also in front of the meeting of the mandi-
bular rami and behind the premaxillary and ethmoidal region
is a clump of five conical teeth, clearly vomerine in position
at all events. It is also clear that the whole of the dentition
of the front of the mouth is not here exposed, as the clump
referred to is on the left side of the middle line, and the
corresponding space on the right side is covered by the
anterior extremity of the corresponding mandible.

The bone representing the mandible is well known from
the description of Hugh Miller. It is an elongated, vertically-
flattened plate (Fig. 1, mn), broader behind than in front,
with rounded posterior extremity, slightly sigmoid contour
when seen from the side, and near the anterior extremity
sharply bent inwards towards its fellow. It is remarkable
for having two sets of conical teeth, one consisting of a row
of about half-a-dozen being situated about the middle of the
upper margin of the bone, while another row of about the
same number occupies the vertical anterior margin, which
would otherwise be symphysiaL This is certainly a very
curious circumstance, and one is simply at a loss to imagine
of what use teeth could be in such a 3ituation, or how they
worked. It was indeed the position of these peculiar
symphysial teeth that led Hugh Miller originally to compare
the working of the jaws of Coccosteus with those of an
Arthropod (2, Ist ed., p. 57; see also footnote in 4th and
subsequent editions).

There are no traces of ossified internal cranial bones, of
hyoid or of branchial arches ; consequently these parts must
have been entirely cartilaginous. I may mention that the


ized by Google

21-i Proceedings of the Roycd Physical Society,

bones figured by Huxley as " the chief parts of the hyoidean
arch " are in reality the ventral rami of the dermal plates
which I have termed " interlateraL"

Body-Cuirass, — The front part of the body behind the
head is encircled by a girdle of osseous dermal plates, some-
what comparable to a shoulder-girdle, expanded backwards
dorsally and ventrally, while at the lower part of the sides
the cuirass is so deeply cut in that the dorsal and ventral
expansions were long considered to have no connection with
each other. Most of the osseous plates which form this
cuirass are well known from the writings of Pander, H.
Miller, and Sir P. Egerton, but nevertheless some correction
is still necessary.

The great median dorsal plate (Fig. 2, m. d) is of an
elongated pentagonal figure, its short base articulating with
the median and lateral occipitals, its acute apex and elon-
gated sides articulating with the two dorso-lateral plates,
which it extensively overlaps. Its under surface shows the
well-known median longitudinal ridge, ending behind in the
"nail-head" prominence, as in the corresponding plate in
Hamosteus. The anterior dorso-lateral plate (a. d. I,), the os
articulare dorsi of Pander, is of a somewhat rectangular form
when detached, though in situ it appears irregularly trape-
zoidal owing to its upper and lower margins being obliquely
overlapped by the median dorsal and by the antero-lateral
respectively; its anterior margin shows a small articular
process by which it is joined to the external occipital
Immediately behind it is placed the posteinor dorso-laieral
(p. d. /.), or the os triangvlare of Pander, a triangular plate
which also articulates with the median dorsal above and the
postero-lateral below, while its oblique hinder border is

The antero-latei*al plate (a. /.), being the os marginale of
Pander, occupies a position below and in front at the
narrowest part of the lateral portion of the cuirass. It is
peculiarly trapezoidal in shape, or it might be described as
triangular, with the downward and forwardly directed apex
obliquely truncated. Its anterior border, gently convex in
the middle, forms part of the anterior margin of the cuirass.


ized by Google

On the Structure of Coccosteus decipiens, Agassiz. 215

though it is for the most part shut out from that by the
anterior dorso-Iateral above and the interlateral below; its
postero-superior margin, somewhat wavy or zigzagged, over-
laps the anterior dorso-lateral besides articulating with the
small postero-lateraL The postero-inferior margin is free
and slopes obliquely downwards and forwards; the short
anterior margin is fitted on to the interlateral This antero-
lateral plate is the one lettered " c " by Huxley (8, p. 30)
and "3" by Hugh Miller (7, p. 133, fig. 6), though he has
represented the very same plate on the preceding page
(p. 132, fig. 5, zz) Bs forming a part of the ventral cuirass.

The postero-lateral plate (p. I.) is a small one situated at
the posterior angulated margin of the lateral part of the
cuirass, and articulates with the antero-lateral, the anterior
dorso-lateral, and the posterior dorso-lateral, its posterior
margin being free. This plate is not noticed by Pander
or Huxley, but it is lettered 2 by Hugh Miller (7, p. 133,
fig. 6).

The interlateral plate {i. L) is one of great interest, as its
form and relations have not yet been properly recognised.
It consists of two parts, lateral and ventral, united at a con-
siderable angle to each other when uncompressed, which,
however, is very rarely the case. The lateral portion, seen
in Fig. 1, forms a sort of fork, on which the short inferior
margin of the antero-lateral plate articulates, and thus is
formed that connection between the dorso-lateral and ventral
portions of the cuirass which was unknown to Pander and
Huxley, and which, so far as I am aware, has not previously
been demonstrated. The lower limb of the fork forms a
conspicuous rounded lower margin, tuberculated like the
other plates, and bears a most suspicious resemblance to
the part represented by Prof. v. Koenen as a pectoral spine
in C. Bickensis (10, pi. il, fig. 2). In C, decipiens it is, how-
ever, very much shorter than the part alluded to in C.
Bidcensis; however, in C, minor it attains a very considerable
proportional length (13, pL iii., fig. 3, i. I), The ventral
portion (see Fig. 3), devoid of tubercular ornament, is
elongated in shape, and, passing inwards and slightly
forwards to meet its fellow of the opposite side, forms


ized by Google

216 ProcetdingB of the Royal Physical Society.

the anterior margin of the ventral portion of the body-
cuirass ; to it posteriorly are articulated the anterior ventro-
lateral and the anterior median plates. This part of the
bone was known to Pander, and is represented in two of his
figures (6, pL ii., fig. 2, and pL v., fig. 1, x), though in the
text he compared it with the jugular plate in PolypUrus or
Osleolepis, Huxley, on the other hand (8, p. 35, fig. 21, a),
considered the bone to be hyoidean in its nature, as we have
already noticed.

Neither Pander nor Huxley seems to have recognised the
lateral portion of this bone, which serves to articulate the
dorso-lateral portion of the cuiniss with the ventral ; indeed,
Huxley remarks (8, p. 32) that " the ventral shield appears
to me to have had no connection with the dorsal/' But of
the connection of the two in the manner I have described
there cannot be the slightest doubt. See also my figiure of
the parts in C. minor (13, pL iii, fig. 3).

The plates forming the expanse of the ventral shield are
already so well known from the figures of Pander and Hugh
Miller, that I need hardly enter into detail regarding them,
especially as I have in PL XI., Fig. 3, accurately given their
respective shape and mode of overlap. They are six in
number : — anterior median ventral (a. m, u), posterior median
ventral (p, m. v,), two anterior ventro4aterals (a. v, L), and
two posterior ventro-laterais {p, v, /.). I may, however,
mention that^ judging from the course of the lateral-line
groove on the anterior ventro-lateral plate, Pander has
reversed its position, putting the front end behind and vice
versd; for we shall presently see that on this plate the
sensory canal occurs on the anterior and not on the posterior
part of its surface.

DistribiUion of tJie Laterai-line Grooves, — The course of
the lateral sensory canal is indicated on certain of the
dermal bones by conspicuous grooves, which, as in the case
of Pterichthys and Bothriolepis, have often been mistaken for
sutures. There is, however, no difficulty in distinguishing
them from sutures, when one by experience really comes to
know the characteristic appearance of the latter.

On the anterior half of each anterior ventro-lateral plate is


ized by Google

On the Siritdure of Coccosteus decipiens, Agcussiz. 217

seen a curved groove, starting from near the middle of the
anterior margin and then curving sharply round to proceed
to the inner border close behind the antero-internal angle.
On the median dorsal plate a groove is seen of a V-shaped
contour, the apex being in the middle line somewhat in front
of the posterior extremity of the bone, the limbs diverging
forwards towards the superior margin of the posterior dorso-
lateral plate. On the anterior dorso-lateral plate a continua-
tion of this groove runs forwards to the postero-intemal
angle of the external occipital, near which it is met by a
branch coming diagonally upwards and forwards from the
postero-inferior angle of the plate. The side-canal thus
formed passes now on to the cranial shield at the point
indicated, and there at once gives ofif a branch running
forwards and slightly inwards, parallel with and close to the
outer margin of the median occipital, becoming lost on the
posterior margin of the central. The main groove then runs
forwards and outwards parallel with the outer margin of the
shield, giving off first a branch psissinc^ to the external pro-
jecting angle of the marginal plate, then turning forwards
and inwards still parallel to the shield-margin it passes on
to the post-orbital plate, where it gives off another branch to
the post-orbital angle. Here it bends sharply backwards and
inwards at an acute angle, runs on to the central plate,
approaching its fellow of the opposite side, and near the
middle of this plate it again turns sharply forwards, passes
on to the anterior part of the pre-orbital, and ends near the
small nasal opening in front In some specimens a cross
commissural branch is seen on the central plates, connecting
tlie two main trunks at the conspicuous angles which they
make in that place.

A groove is also observable on the maxilla, apparently
continued from the second external branch of the main
groove on the post-orbital, and passing along as a sub-orbital
branch close to the hoUowed-out orbital margin of the bona
It gives off behind the eye another branch, which passes in
a curved manner downwards and backwards towards the
margin of the bone posteriorly.

Sclerotic Ring, — ^A specimen from Gamrie, in the Edinburgh


ized by Google

218 Proceedings of the Royal Physical Society,

Museum, shows evidence of a sclerotic ring such as has been
figured by v. Koenen (10, pi. ii., fig. 2 ; pi. iv., fig. 2).

Internal Skeleton. — In the typical Coccosteus decipiens, Ag.,
there is no trace of vertebral centra, the space occupied by
the persistent notochord being always empty in the fossils.
Agassiz in his restored figure (1, pi. vi., fig. 3) has repre-
sented on both dorsal and ventral aspects of the notochordal
space a continuous row of distally-pointed neurapophyses and
hsemapophyses, also a dorsal and anal fin situated opposite
each other, each supported in Teleostean fashion by a series
of proximally-pointed interspinous bones, dipping down
between the neurapophyses, the supposed fin-rays being,
according to the same idea, pointed at their extremities.
Pander (6, pi. iv., fig. 1) still retains the two median fins,
with the long hsemapophyses in front of the anal, though he
was more correct in making the interspinous bones articulate
end to end with the neurapophyses by expanded extremities.
But though M'Coy had previously (5, p. 602) strongly
doubted the existence of an anal fin in Coccosteics, Pander's
figure has been copied into almost every text-book; Pro£
von Koenen has transferred the body-skeleton and fins as
there represented to his restoration of the allied genus
Pra^hydeirtis, while the anal fin is also mentioned as present
by Zittel in his handbook (14, p. 160). M'Coy was, however,
correct — there is no anal fin in Coccosteus ; but besides this
Pander's figure is incorrect in other points, which I shall now

It is not possible to trace the vertebral column to its com-
mencement, owing to its obscuration by the dorso-lateral
cuirass ; where it first becomes visible is about the middle of
the length of the great median dorsal plate. There we find
short broad neural pieces continued obliquely backwards and
upwards into neural spines, which gradually lengthen until
we come to the dorsal fin, which commences a little beyond
the apex of the plate just mentioned. Here we have two sets
of interspinous bones articulated end to end with each other
and with the neural spines, which latter are truncated and
not pointed. In a very good specimen in the British Museum
I count about fifteen ossicles in the proximal set and twelve


ized by Google

On the Structure of Coccosteus decipiens, Agassiz. 219

in the distal, though probably the numbers were equal in the
perfect state, and in both sets they have the same form,
namely, they are slender, elongated, and expanded at both
extremities. It is evident from the last-mentioned circum-
stance that the ossicles of the second row are not dermal
fin-rays, but belong to the same category as those of the first ;
two rows of interspinous bones being, in fact, of constant
occurrence in the primitive Ganoids.

Beyond the dorsal fin the neural spines become very short
as well as less oblique in their direction.

On the haemal aspect of the vertebral axis no such elon-
gated apophyses occur anteriorly, as depicted in the restora-
tions of Agassiz and Pander. Immediately behind the lateral
plates of the cuirass we find small, nearly circular, haemal
pieces without spines, then spines are added which, gradually
lengthening, become longest in the region opposite the dorsal
tin, whence they again diminish towards the extremity of the
tail. It is this peculiar lengthening of the haemapophyses
under the dorsal, a fact also noticed by M'Coy, which has
evidently given rise to the old idea of the presence of an
anal fin.

. In all specimens of Coccosteus where the internal skeleton
is well preserved there is found a pair of peculiar slender
bones (x), each of which is pointed at both ends and bent
below the middle at an obtuse angle in somewhat L-shaped
fashion, the long limb pointing upwards towards the verte-
bral axis, the short one forwards. These bones were noticed
by Pander (6, p. 73), who, though extremely doubtful as to
their nature, supposed that they " vielleicht den Extremitiiten
ale Stiitzen der weichen Flossen angehortea" Their position
is certainly suggestive of their having had something to do
with pelvic limbs — more I cannot say.

Mr A. Smith Woodward has pointed out to me that in
several specimens in the British Museum a small oval or
somewhat rhombic bony plate (y) is seen lying in a position
posterior to the last-mentioned bones. I have not observed
it in any other specimens than those ; but its presence in a
similar position in more than one example would seem to
indicate that it was a scute placed in the ventral mesial line.
VOL. x. Q


ized by Google

220 Proceedings of the Royal Physical Society .

Were pectoral memi>ers prese^U f — I have now examined
with the utmost care a very great number of specimens of
Coccosteics decipiens in all conditions of preservation and from
all the beds and localities of the Scottish Old Bed Sandstone
which have yielded such remains, including the collections in
the British Museum, in the Museum of Practical Geology, in
the Edinburgh Museum of Science and Art, the Gordon-
Cumming collection at Forres, and many others, but without
meeting with any other parts either of endo- or exoskeleton
than those I have described. And, in particular, I have not
seen the smallest evidence of the presence of any pectoral
limb, nor any trace of an articular surface on any of the
bones to which such a limb could have been articulated. It
can scarcely be believed that had such a limb been present it
would either have escaped preservation or observation in so
large a number of specimens. Nevertheless, more than one
author has been disposed to believe in the presence of such a
limb in Coccosteus,

In the restored figure of Coccosteus given by Hugh Miller
in the first edition of the " Old Red Sandstone " (2, pi. iii.),
no limb is represented, and its absence is positively affirmed
in the text. But in subsequent editions, and also in Duffs
"Geology of Moray" (3, pL viii., fig. 1). a peculiar
" paddle - shaped " body is represented appended to tlie
head. However, Hugh Miller, in a footnote, explains that
he has ascertained that the supposed arms "were simply
plates of a peculiar form." Of course there is not the
smallest doubt that the idea of this limb owed its origin to a
displaced maxillary bone.

But more recently, in connection with what appear un-
doubtedly to be fragments of a large and peculiar form of
Coccosteus, Trautschold (9 and 11) has described and figured
from the Old Red Sandstone of Russia certain peculiar
bodies, which he considers, though not without doubt, to
appertain to supposed large arms or " Ruderorgane " belong-
ing to that species, which he accordingly names Coccosteus
megalopteryx. What the fragments are to which he applies

Online LibraryRoyal physical society of old catProceedings → online text (page 19 of 44)