452 CHARACTERS OF PALEOZOIC FISHES COPE. 



the i)osterior lateral anj;les of the head-shield, and there are two lines 

 which penetrate the matrix more deeply than the rest. 1 can not make 

 out that any canal radiates from this point except the one which reaches 

 to the center of the supratemporal. This region corresponds to that 

 of the intercalary of the fishes, but its boundaries I can not make out. 



This arrangement of cranial elements may be compared with those 

 of Coccosteus and Iloinostens. It differs from the former in tlie pres- 

 ence of a malar bone bounding the orbit below, and in the presence of 

 the "squamosal" behind it. In Homosteus, elements which occupy 

 the position of the two mentioned are present (Traquair, I. c), but they 

 are called by Traquair postorbital and marginal, names which he ap- 

 lilies to my possible postorbital and supratemporal. I think the ele- 

 ments described by Traquair are homologous with the malar and 

 squamosal of Macropetalichthys, so that the " postorbital " (my post- 

 frontal) and '' marginal " (my supratemporal) must be sought for else- 

 where in Homosteus. Traquair's "central" appears, from its position, 

 to include my postfrontal, while the supratemporal mny be embraced 

 in the anterior part of Traquair's " external occipital." This question 

 can, however, only be settled by the discovery of intermediate types. 

 In any case, a general affinity to the Arthrodira is indicated by the 

 segmental structure of the skull, as well as by the character of the tubes 

 of the lateral line system. 



The inferior surface of the skull presents the following characters. 

 This is important, as I do not know of any description of this region in 

 an Arthrodire, excepting in the cases of the Dinichthys and Titanich- 

 thys described by Newberry. (Fig. 6.) 



In the first place the posterior part of the head-shield, the " median 

 occi[)ital" region of Traquair, is produced very far posteriorly, as in 

 Homosteus. This region <loes not seem to have i>rotected the brain, 

 but rather the anterior part of the vertebral axis, and seems to have been 

 a nuchal plate. In the specimen I am now describing, the posterior 

 extremity of this element is broken away for a short distance on both 

 sides of the middle line, revealing a cast of its interior. This is I)i]o- 

 bate, by reason of a vertical constriction at the middle line. That 

 this is not a cast of the cranial cavity is proven not only by its form, 

 but by the fact that there is no cast representing a medulla oblongata or 

 a foramen magnum. The chamber was absolutely closed posteriorly. 

 The lateroposterior angle of this cavity is exposed l)y the loss of the 

 external wall. It is obtusely angular. Turning now to the inferior 

 asi)ect of the skull, we observed, at the middle line of the inferior-pos- 

 terior border, a wide, upward excavation, looking backwards and down- 

 wards. It rapidly contracts into a groove with an angular superior 

 middle line. Whetlier this groove is part of a tube can not be ascer- 

 tained, owing to the loss of the bony tissue on each side and below, but 

 it may be only the apical angle of a roof-shaped si)ace, whose lateral 

 slopes arc produced on each side, sloping well downwards and out- 



