CHAPTER XXXII

THE CLASS OSTRACOPHORI[152]

=Ostracophores.=--Among the earliest vertebrates actually recognized as fossils belongs the group known as _Ostracophori_ (~ostrakos~, a box; ~phoreô~, to bear). These are most extraordinary creatures, jawless, apparently limbless, and enveloped in most cases anteriorly in a coat of mail. In typical forms the head is very broad, bony, and horseshoe-shaped, attached to a slender body, often scaly, with small fins and ending in a heterocercal tail. What the mouth was like can only be guessed, but no trace of jaws has yet been found in connection with it. The most remarkable distinctive character is found in the absence of jaws and limbs in connection with the bony armature. The latter is, however, sometimes obsolete. The back-bone, as usual in primitive fishes, is developed as a persistent notochord imperfectly segmented. The entire absence of jaw structures, as well as the character of the armature, at once separates them widely from the mailed _Arthrodires_ of a later period. But it is by no means certain that these structures were not represented by soft cartilage, of which no traces have been preserved in the specimens known.

=Nature of the Ostracophores.=--The Ostracophores are found in the Ordovician or Lower Silurian rocks, in the Upper Silurian, and in the Devonian. After the latter period they disappear. The species are very numerous and varied. Their real affinities have been much disputed. Zittel leaves them with the Ganoids, where Agassiz early placed them, but they show little homology in structure with the true Ganoids. Some have regarded them as aberrant Teleosts, possibly as freakish catfishes. Cope saw in them a huge mailed group of archaic Tunicates, while Patten has soberly and with considerable plausibility urged their affinity[153] to the group of spiders, especially to the horseshoe-crabs (_Limulus_) and their palæozoic ancestors, the _Eurypteridæ_ and _Merostomata_.

The best guess as to the affinities of the Ostracophores is perhaps that given by Dr. Ramsey H. Traquair ("Fossil Fishes of the Silurian Rocks of the South of Scotland," 1899). Traquair regards them as highly aberrant sharks, or, more exactly, as being derived, like the Chimæras, from a primitive Elasmobranch stock. In favor of this view is the character of their armature, the bony plates themselves to be regarded as formed by the fusion of shagreen grains or scales. According to Traquair: "Specialization from the most specialized form, _Lanarkia_, has been accompanied by (1) fusion of the spinelets (_Lanarkia_) or shagreen grains (_Thelodus_) into plates, scutes, and rhombic scales, supported by hard matter developed in a deeper layer of skin, and (2) alterations in the pectoral fin-flaps, which, becoming covered up by the postero-lateral plates in _Drepanaspis_, are finally no longer recognizable in the _Pteraspidæ_."

Woodward leaves their exact relationship undefined, while others have regarded them as mailed lampreys, at any rate to be excluded from the _Gnathostomi_, or jaw-bearing series. The apparent absence of true jaws, true limbs, and limb-girdles certainly seems to separate them widely from true fishes, but these characters are negative only, perhaps due to degeneration, and at any rate they are not yet absolutely determined. Certainly they offer no positive proof of affinity with the modern Cyclostomes.

Dr. Traquair regards the _Heterostraci_ or most primitive _Ostracophores_ as most certainly derived from the Elasmobranchs. Other writers have attacked the integrity of the group of Ostracophores, questioning the mutual relationship of its component parts. Reiss, for example, regards the association of the _Osteostraci_ with the _Heterostraci_ as "unbegründet" and "unheilvoll," while Ray Lankester, as quoted by Traquair, affirms that "there is absolutely no reason for regarding _Cephalaspis_ as allied to _Pteraspis_ beyond that the two genera occur in the same rocks, and still less for concluding that either has any connection with _Pterichthys_." Elsewhere Lankester states that the _Heterostraci_ are associated at present with the _Osteostraci_, "because they have, like _Cephalaspis_, a large head-shield, and because there is nothing else with which to associate them." Patten, on the other hand, seems inclined to deny the rank of _Heterostraci_ and _Osteostraci_ as even separate orders, regarding them as very closely related to each other as also to their supposed spider-like ancestors.

But the consensus of opinion favors the belief that the four orders usually included under this head are distinct and at the same time are really related one to another. For our purposes, then, we may regard the _Ostracophori_ as a distinct class of vertebrates. By placing it after the Elasmobranchs we may indicate its probable descent from a primitive shark-like stock.

On this subject Dr. Dean remarks: "The entire problem of the homology of the dermal plates and 'scales' in the Ostracophores and Arthrognaths is to the writer by no means as clear as previous writers have conceded. From the histological standpoint, admitting the craniote nature of the vasodentine and cancellous layers in the dermal plates, it nevertheless does not follow that they have been derived from the actual conditions of the dermal denticles of the ancestral Gnathostome, as were unquestionably the dermal plates of Teleostomes and Dipnoans. It seems equally if not more probable, on the other hand, that the dermal armoring of the distinct groups may have had an altogether different mode of origin, the product of a crude evolution which aimed to strengthen the skin by a general deposition of calcareous matter throughout its entire thickness. The tuberculation of plates thus acquired might have become an important step in the development of a more superficial type of armoring which is most preferably represented by the dermal denticles of Selachians. Nor, in passing, need the presence of a mucus-canal system in the early plated forms be of greater morphological importance than a foreshadowing of the conditions of Gnathostomes, for this system of organs might serve as well as evidence, in a general way, of relationship with Marsipobranchs. Nor is this evidence the more conclusive when we reflect that _no known type of Gnathostome, recent or fossil, possesses open sensory grooves in distinct dermal plates_. The presence, furthermore, of a dorsal fin and a 'truly piscine heterocercal tail,' as noted by Traquair, is by no means as Gnathostome-like as these structures at first glimpse appear. For they lack not merely the characteristic radial supports of fishes, but even actinotrichia. Their mode of support, on the other hand, as Smith Woodward points out, is of a more generalized nature, bent scales, homologous with those of the adjacent body region, taking the place of the piscine external supports." The actual position in the system to be finally assigned to the Ostracophores is therefore still uncertain.

=Orders of Ostracophores.=--Four orders of _Ostracophori_ are now usually recognized, known in the systems of Woodward and Traquair as _Heterostraci_, _Osteostraci_, _Antiarcha_, and _Anaspida_. The former is the most primitive and perhaps the most nearly allied to the sharks, the second is not very remote from it, the last two aberrant in very different directions. Hay places the _Antiarcha_ with the _Arthrodira_ under the superorder of _Placodermi_.

=Order Heterostraci.=--The _Heterostraci_ (~heteros~], different; ~ostrakos~, box) have no bone-corpuscles in the coat of mail. This typically consists of a few pieces above, firmly united and traversed by dermal sense-organs or "lateral lines." The ventral shield is simple. Four families are recognized by Traquair as constituting the _Heterostraci_, these forming a continuous series from shark-like forms to the carapace-covered _Pteraspis_. In the most primitive family, the _Thelodontidæ_,[154] the head and trunk are covered with small scales or tubercles of dentine and not fused into large plates. The tail is slender and heterocercal, the caudal fin deeply forked. Until lately these tubercles were regarded as belonging to sharks, and they are still regarded by Traquair as evidence of the affinity of the _Heterostraci_ with the _Acanthodei_. Dr. Traquair thinks that a flap or lappet-like projection behind the head may be a pectoral fin. The three known genera are _Thelodus_, _Lanarkia_, and _Ateleaspis_. In _Thelodus_ the scales consist of a base and a crown separated by a constriction or neck. _Thelodus scoticus_, _Thelodus pagei_, and _Thelodus planus_ are found in the Silurian rocks of Scotland. Other species, as _Thelodus tulensis_ of Russia, extend to the Upper Devonian.

In _Lanarkia_ the large sharp scales have an expanded base like the mouth of a trumpet. _Lanarkia horrida_ and _L. spinulosa_ are found in the shire of Lanark in Scotland. In _Ateleaspis_ (_tesselatus_) the skin is covered with small polygonal plates. The lateral flaps or possibly fins take the form of flat rhombic sculptured scales. In this genus the eyes seem to be on the top of the head.

In the _Psammosteidæ_ of the Devonian the head is covered with large plates which are not penetrated by the sense-organs. These plates are covered with minute, close-set tubercles, covered with brilliant ganoid enamel and with finely crimped edges. According to Dr. Traquair, these tubercles are shagreen granules which have coalesced and become united to plates formed in a deeper layer of the skin, as in _Ateleaspis_ the minute scales have run together into polygonal plates. These creatures have been considered as "armored sharks," and Dr. Traquair regards them as really related to the acanthodean sharks. Nevertheless they are not really sharks at all, and they find their place with the _Pteraspis_ and other longer known Heterostracans.

The family of _Drepanaspidæ_ consists of a single recently known species, _Drepanaspis gmundenensis_, found in a pyritized condition in purple roofing-slate in Gmünden, Germany. This fish, which reaches a length of about two feet, has a broad head, with eyes on its outer margin, with a slender body and heterocercal tail. The head has a broad median plate and smaller polygonal ones. The flaps, supposed to represent the pectoral fins, are here cased in immovable bone. No trace of internal skeleton is found by Traquair, who has given the restoration of this species, but the mouth has been outlined.

The best known of the Heterostracan families is that of _Pteraspidæ_. In this family the plates of the head are coalesced in a large carpace, the upper part originally formed of seven coalesced pieces. A stout dorsal spine fits into a notch of the carapace. The slender body is covered with small scales and ends in a heterocercal tail. The dermal sense-organs are well developed. _Pteraspis rostrata_ occurs in the Lower Devonian. Other genera are _Palæaspis_ and _Cyrthaspis_.

=Order Osteostraci.=--The Osteostraci (~osteon~, bone; ~ostrakos~, box) (called _Aspidocephali_ by Rohon) have bone-corpuscles in the shields, and the shield of the back is in one piece without lateral-line channels or sense-organs. Ventral shield single. The order includes three families. The _Cephalaspidæ_ have the shields tuberculate, the one between the eyes fixed, and the anterior body-shields are not fused into a continuous plate. The best known of the numerous species is _Cephalaspis lyelli_ from the Lower Devonian of England. _Hemicyclaspis murchisoni_ occurs in the Upper Silurian of England, and the extraordinary _Cephalaspis dawsoni_ in the Lower Devonian of Gaspé, Canada. _Eukeraspis pustulifera_ has the head-shield very slender and armed with prickles. In the _Thyestidæ_ the anterior body-scales are fused into a continuous plate. _Thyestis_ and _Didymaspis_ are genera of this type. The _Odontotodontidæ_ (_Tremataspidæ_) have the shield truncate behind, its surface finely punctate, and the piece between the eyes not fixed. _Odontotodus[155] schrenki_ is found in the Upper Silurian of the Island of Oesel in company with species of _Thyestes_. The _Euphaneropidæ_ are represented in the Devonian of Quebec.

=Order Antiarcha.=--The Antiarcha (~anti~, opposite; ~archos~, anus) have also bone-corpuscles in the plates, which are also enameled. The sense-organs occupy open grooves, and the dorsal and ventral shields are of many pieces. The head is jointed on the trunk, and jointed to the head are paddle-like appendages, covered with bony plates and resembling limbs. There is no evidence that these erectile plates are real limbs. They seem to be rather jointed appendages of the head-plate, erectile on a hinge like a pectoral spine. There are traces of ear-cavities, gill-arches, and other fish-like structures, but nothing suggestive of mouth or limbs.

This group contains one family, the _Asterolepidæ_, with numerous species, mostly from Devonian rocks. The best known genus is _Pterichthyodes_,[156] in which the anterior median plate of the back is overlapped by the posterior dorso-lateral. _Pterichthyodes milleri_ from the Lower Devonian, named by Agassiz for Hugh Miller, is the best known species, although numerous others, mostly from Scottish quarries, are in the British Museum. _Asterolepis maximus_ is a very large species from the same region, known from a single plate. _Bothriolepis canadensis_ is from the Upper Devonian of Scaumenac Bay near Quebec, numerous specimens and fragments finely preserved having been found.

_Microbrachium dicki_ with the pectoral appendages small occurs in the Devonian of Scotland.

The earliest remains of _Ostracophori_ are found in Ordovician or Lower Silurian rocks of the Trenton horizon at Cañon City, Colorado. These consist of enormous numbers of small fragments of bones mixed with sand. With these is a portion of the head carapace of a small Ostracophore which has been named by Dr. Walcott _Asteraspis desiderata_ and referred provisionally to the family of _Asterolepidæ_, which belongs otherwise to the Lower Devonian.

With these remains are found also scales possibly belonging to a Crossopterygian fish (_Eriptychius_). These remains make it evident that the beginning of the fish series lies far earlier than the rocks called Silurian, although fishes in numbers are not elsewhere known from rocks earlier than the Ludlow shales of the Upper Silurian, corresponding nearly to the Niagara period in America.

In the Ludlow shales we find the next appearance of the Ostracophores, two families, _Thelodontidæ_ and _Birkeniidæ_, being there represented.

=Order Anaspida.=--Recently a fourth order, _Anaspida_ (~a~, without; ~aspis~, shield), has been added to the _Ostracophori_ through the researches of Dr. Traquair. This group occurs in the Upper Silurian in the south of Scotland. It includes the single family _Birkeniidæ_, characterized by the fusiform body, bluntly rounded head, bilobate, heterocercal tail, and a median row of hooked spinous scales along the ventral margin. No trace of jaws, teeth, limbs, or internal skeleton has been found. Unlike other Ostracophores, _Birkenia_ has no cranial buckler with orbits on the top, nor have the scales and tubercles the microscopic structure found in other Ostracophores. In the genus _Birkenia_ the head and body are completely covered by tubercular scutes. The gill-openings seem to be represented by a series of small perforations on the sides. A dorsal fin is present. _Birkenia elegans_ is from the Ludlow and Downstonian rocks of southern Scotland. _Lasianius problematicus_ from the same rocks is very similar, but is scaleless. It has a row of ventral plates like those of _Birkenia_, the only other hard parts it possesses being a number of parallel rods behind the head, homologous with the lateral series of _Birkenia_. _Lasianius_ is therefore a specialized and degenerate representation of _Birkenia_, differing somewhat as "the nearly naked _Phanerosteon_ differs from other _Palæoniscidæ_ whose bodies are covered with osseous scales."

FOOTNOTES:

[152] This group was first called by Cope _Ostracodermi_--a name preoccupied for the group of bony trunkfishes, _Ostracidæ_. The still earlier name of _Placodermi_, chosen by McCoy (1848), was intended to include Arthrodires as well as Ostracophores. Rohon (1892) calls the group _Protocephali_, and to the two orders he assigns the names _Aspidorhini_ and _Aspidocephali_. These groups correspond to _Heterostraci_ and _Osteostraci_ of Woodward. Another name of early date is that of _Aspidoganoidei_, given by Professor Gill in 1876, but not defined until 1896. These fishes are, however, not "Ganoids" and the name _Ostracophori_ seems to receive general preference. The group _Peltacephalata_ of Patten corresponds essentially to _Ostracophori_, as does also the order _Hypostomata_ of Gadow.

[153] According to Professor Patten's view, the close resemblance of the shields of _Pteraspis_ to those of contemporaneous _Eurypterids_ indicates real affinity. But the _Eurypterids_ are related to the spiders and to _Limulus_. The only reason for thinking that _Pteraspis_ is a fish at all lies in its resemblance to _Cephalaspis_, which is in several ways fish-like, although its head shield is much like that of _Limulus_. All these resemblances in Patten's view indicate real affinity. Patten considers the _Pteraspids_ as derived from primitive arachnid or spider-like forms having a bony carapace as _Limulus_ has. From _Pteraspis_ he derives the other Ostracophores, and from these the sharks and other vertebrates, all of which appear later in time than the earliest Ostracophores. This view of the origin of vertebrates is recently urged with much force by Professor Patten (Amer. Nat., 1904, 1827). Most naturalists regard such resemblances in specialized structures on the outside of an animal as parallelisms due to likeness in conditions of life. The external structure in forms of really different nature is often similarly modified. Thus certain catfishes, pipefishes, sea-moths, and agonoid fishes are all provided with bony plates not unlike those of ganoid fishes, although indicative of no real affinity with them. Commonly the ancestry of vertebrates is traced through enteropneustans to soft-bodied worms which have left no trace in the rocks.

In the same connection, Professor Patten suggests that the lateral fold from which many writers have supposed that the limbs or paired fins of vertebrates is evolved is itself a resultant of the fusion of the fringing appendages on the sides of the body. Such appendages are found in the primitive mailed arachnoids and in _Limulus_. They are shown very plainly in Patten's restoration of _Cephalaspis_. About thirty of them of a bony nature and jointed to the body occur on either side between the gill opening and the vent.

[154] Called _Coelolepidæ_ by Pander and Traquair, but _Coelolepis_ is a later synonym of _Thelodus_.

[155] This name, inappropriate or meaningless, is older than _Tremataspis_.

[156] The earlier name of _Pterichthys_ has been already used for a genus of living fishes.