NOTE ON SOME FISH-REMAINS FROM THE UPPER DEVONIAN ROCKS IN NEW YORK STATE. By H. S. WILLIAMS, of Ithaca, N. Y. [ABSTRACT]

THE author described several small fish bones from the Devonian rocks at Ithaca, N. Y. The exact horizon is in the Chemung, near the base, and below in the Ithaca shales, which form the transition between Portage and Chemung.

The form and peculiar markings were the interesting features in the specimens.

One was elongate oval, much like the elytron of a beetle, the other, more produced and pointed at the end, which was incurved, the convex margin of each being thickened and the opposite side running out into a blade-like expansion, and covered by a reticulation of veins. The veining on the elytron-like form is made up of several strong, longitudinal, nearly parallel, linear veins connected by finer cross veins.

In the larger, pointed specimen (14 inches long), the veins proceed from the broad, convex, axial margin, run along toward the point nearly parallel, diverge, and terminate upon the free thin margin, the longer veins terminating upon the serrate points at the margin. The cross veins are irregular.

Both specimens are unsymmetrical, concave on one side, which appears to be the underside, and convex on the other veined surface. At first glance they appear like insectean, or crustacean remains, but closer examination of the larger specimen, of which some of the substance of the bone is preserved, reveals the structure of fish bone.

The occurrence in the same slab with one of the latter of a distinct fish jaw with teeth (another was found in the same bed), of the Ceratodus, or Dipterus type, has furnished suggestion as to their nature.

The jaw consists of a thin, oblong plate produced into a lobelike expansion behind, the outer margin, gently curving and beset above by a single row of distinct teeth-like serrations, standing vertical upon the plate. There are five in the series, which curves with the outer margin of the plate; the longer is the more anterior, and they gradually diminish toward the posterior end. The whole jaw is about half an inch long, and the longest tooth is something

CLASSIFICATION OF PRODUCTI; BY S. H. TROWbridge.

193

over a line in height. The edges facing each other are trenchant. These teeth resemble Dipterus sherwoodi of the Catskill more nearly than any species I have seen described, although there appears to be but a single row of denticles, corresponding to the outer one in that species. I have therefore provisionally put it in this genus, under the name Dipterus Ithacensis.

A study of Ceratodus and allied fish makes it probable that no endoskeletal parts were ossified in these Devonian fish, and it is conjectured that the peculiar bones are probably the pectoral fin spines or covers of a fish with teeth (of a single row) like Dipterus, showing some relationship perhaps to Pterichthys.

The shorter elytron-like specimen is from the Chemung, three hundred feet above base, at Ithaca, associated with Mollusca. The larger pointed form is from Ithaca shales, at base of Chemung, associated with plant remains, a few frail Mollusca, and a large fish bone, probably a head plate of Dinichthys.

REMARKS ON THE CLASSIFICATION AND DISTRIBUTION OF PRODUCTI. By S. H. TROWBRIDGE, of Glasgow, Mo.

THE work of which this paper gives a brief notice was begun some years ago with a view to determining the differences, if any, between European and American species of the genus Productus.

But the species of the same name, according to different authorities or even to a single authority-were found to differ so widely that the question of geographical distribution, from named specimens, was involved in a maze of inextricable confusion. Before determining this point, therefore, it seemed necessary to decide what description of specimen should be considered as entitled to exclusive use of any given name. To determine this from recognized authorities was found to be almost equally confusing, from lack of harmony between the specimens and their descriptions and figures as given by even those who originally described them. Hence it became necessary to go farther back and inquire whether there is in existence a satisfactory classification of the Producti.

Having been compelled, from all accessible data, to answer this in the negative, the still more fundamental inquiry was forced upon me: can the species of Producti be satisfactorily classified? or must we conclude that the species so run into one another that the lines of distinction cannot be clearly made out, or must be drawn arbitrarily? The last two finally became the first and main questions to answer, and are so discussed here, the other questions mentioned receiving light only somewhat incidentally, but with sufficient clearness, it is hoped, to show good reasons for the conclusions reached.

In making comparisons concerning these forms I have had recourse to the descriptions and figures of both European and American authorities, giving especial attention to works containing original descriptions and figures, and, perhaps, to tons of specimens; including those from the Geological Surveys of Kentucky and Missouri, from years of private collecting in the Western states, and all those from every quarter of the globe, found in the fossil collections of the Museum of Comparative Zoology, many of which were the specimens originally described. To make the conclusions as satisfactory as possible, I have made separate comparisons respecting size, shape and general appearance, markings on the dorsal and ventral valves, spines, configuration of both valves, sinus, beaks, ears, hinge-line, etc., putting each in a separate column for more convenient reference, as will be scen by referring to tabulated descriptions, etc.1

The question of classification divides itself into two parts: 1, as to external peculiarities: 2, as to internal markings.

DeKoninck, in his text of Animal Fossils, bases a tabulated classification on: 1, presence or absence of longitudinal striæ and size; 2, continuity or bifurcation of striæ, and peculiarities of spines; 3, presence or absence of transverse plications; 4, of lobes. All authorities agree in basing classification, founded on the external view, on some or all of the following peculiarities, which will be noticed somewhat in detail, not however, in relation to all the so-called species of Productus, but, for the sake of brevity, to only what may be fairly considered representative species. 1. Size and shape. By studying carefully the tabulated data in the columns devoted to this point, the following facts, among

1The tables and drawings referred to in this paper were exhibited in the Section when the paper was read, but were not presented for publication.

many others, appear. In size, from verbal descriptions, the species are arranged, beginning with the largest, thus: costatus Sow., antiquatus, semireticulatus, martini, flemingii, var. burlingtonensis, longispinus DeK., spinosus, flemingii Sow., longispinus Sow., lobatus, and concinnus. From the figures the arrangement is somewhat different semireticulatus Mart., martini Sow., costatus Sow., antiquatus, flemingii, var. burlingtonensis, longispinus DeK., spinosus, lobatus Sow., longispinus Sow., flemingii Sow., concinnus. Since the size even of the same undoubted species is so variable, from age or conditions favorable or otherwise in different localities, it cannot be safely depended on to characterize species. It may, however, throw light on the number of broods in a given time, or the periods of growth in a species when a great number of specimens can be compared.

The expansion of the front and sides beyond the visceral part is used to distinguish certain species, especially semireticulatus, but it occurs also in martini, tenuicostus, lobatus, flemingii and others, and we must conclude that these are all one species or that this peculiarity is not characteristic.

The relation of the length to the width is very generally mentioned in describing species. Yet by referring to the tables we find that at least three species: semireticulatus, flemingii, var. burlingtonensis, and lobatus are variously represented as having the length equal to the width, less than width and greater than width. Among specimens of flemingii from Visè, I found several with a raised band on the edge where the front was short, or in the centre of the front when this was so produced as to be greater than the width of the shell; and in one of these which showed both valves could be distinctly seen the proof that the raised band indicates the place where the dorsal meets the ventral valve and follows it to the edge.

Most shells seem to be broken off at this meeting point, thus making a shell, originally longer than wide, to become wider than long. And one of Sowerby's figures of P. martini shows a large crack which nearly separates the long expanded front from the visceral part. This visceral part alone is indistinguishable from DeKoninck's P. flemingii from Visè. (See Mineral Conchology, Agassiz's Edition, 1854, Pl. 317, p. 210.) The presence of the band. on the front has been considered of sufficient importance to base

a new species upon it, the P. marginicincta; whereas in fact it is doubtless merely accidental.

2. Surface markings. Much stress has been laid upon the radiating costæ and concentric striæ or wrinkles on both the dorsal and ventral valves. In the descriptions tabulated, the dorsal valves of only two, costatus and semireticulatus, are found described and the same peculiarities of radiating costæ, concentric striæ and the nodulose appearance at their intersection, appear in each. Many more figures of dorsal valves are found, but there is a very monotonous, if not annoying, resemblance in them all. P. costatus is described by DeKon. as having, on ventral valve, radiating costæ which do not bifurcate, yet in specimens thus named by himself they do fork on both the dorsal and ventral valves, and this peculiarity belongs also to P. semireticulatus and P. flemingii, var. burlingtonensis. Costatus agrees with lobatus in the number of costa, and with longispinus in their coalescing on the front. Four species-costatus, semireticulatus, flemingii, var. burlingtonensis, and antiquatus-correspond in having concentric striæ on their upper half. De Koninck's description of P. martini gives it transverse folds on the beak, but Sowerby, in his original descriptions, separates martini from antiquatus, because the former is not corrugated and the latter is. (When doctors thus disagree who is to decide?) Again, flemingii and antiquatus are both described as having wrinkles on the ears, while the figures of not less than a half dozen other species show them, and the specimens of perhaps a half dozen more. These facts, with the statement of Mr. Meek, that costa of the same species are often exceedingly variable (which can be easily verified by examining specimens in any considerable quantity), lead to the conclusion that we do not find in surface markings, the key to all these difficulties of classification. 3. In regard to spines also, equal confusion occurs. Hall gives as one characteristic of P. costatus, var., its row of spines in a curve from beak to base. This is seen on a few, but on many undoubted P. costati, equally well preserved, it is not to be found. And Sowerby describes P. martini in about the same way. Some of De Koninck's figures represent a P. longispinus Sow., with four rows of spines close together, forming a wide band from ear to ear, across the front just below the umbo. But among all De Koninck's collections in Agassiz's Museum, I find only one marked at all in