GREGORY, PRESENT STATUS OF ORIGIN OF TETRAPODA 331 



"Along the posterior margin of the cranial shield are three plates {s. t, 

 figs. 1 and 2), one mesial, somewhat poh^gonal in form, and two lateral, each 

 apparently of a triangular shape. These are obviously the representatives of 

 the three plates, which occur in a similar position in Osteolepis, Glyptolwmus, 

 Megalichthys, &c., and of which different interpretations have been given by 

 different authors. In Professor Huxley's description of Glyptolcemus,^ the 

 mesial one is marked 'supra-occipital,' the two lateral 'epiotic' Mr. Parker ' 

 has, however, pointed out that they are dermal bones, and not to be considered 

 homologous with those other deeper ossifications of the cranial cartilage. By 

 Pander ^ they are in Osteolepis simply designated 'Hautknochen' and consid- 

 ered to be equivalent to the five little plates, which in the recent Polypterus 

 occur immediately behind the transverse row of supra-temporals, and between 

 the pair of upper supra-claviculars (supra-scapulars) , being in reality the first 

 scales of the back. On the other hand, he considered the transverse chain of 

 small plates (supra-temporal) which lie immediately behind the parietals of 

 Polypterus, to be representel in Osteolepis microlepidotus hj the narrow por- 

 tion of the cranial shield, which in that species is marked off near the hinder 

 margin by a more or less interrupted superficial transA^erse groove. I am my- 

 self very much inclined to the belief that the three dermal bones in question 

 are in reality equivalent to the transverse supra-temporal chain in Polypterus 

 and Lepidosteus, and which have their representatives as well in the am- 

 phibian Labyrinthodonta as in most Teleostean fishes ; the transverse grooving 

 across the posterior part of the cranial shield in many Saurodipterines being 

 probably only of the nature of superficial markings.'" 



In brief, the skull-roof of Rhipidistia may be regarded as consisting of 

 the following regions or segments : first a rostrof rontal segment, com- 

 prising the parts which later became separated into tlie premaxillas, 

 ethmoid, nasals, frontals and postfrontals. In Osteolepis this is sep- 

 arated from the next segment by a sharply defined transverse suture. 

 The second or parietal segment of the skull-roof includes only the elon- 

 gate parietals and the wrongly named "squamosals'^ (pterotics). The 

 third or occipital segment which is rarely distinct from the preceding 

 (Osteolepis, etc.) consists of a transverse row of small elements, which, 

 in spite of their resemblance to the dermoccipitals and tabularia of Stego- 

 cephali, are probably not homologous with them. The fourth or nuchal 

 segment, which originally pertained to the pectoral girdle rather than to 

 the slmll, is sharply separated from the true occipital segment ; it usually 

 consists of a single median and a pair of lateral elements; all three of 

 these are usually named "supratemporal," but they are probably homol- 

 ogous with the paired dermosupraoccipitals and tabularia of Stegocephali. 



Returning to the region above and behind the orbits we find not only 



3 Dec. Geol. Surve.v, x. p. 2. 

 * Shoulder Girdle and Sternum, p. 19. 



^Ueber die Saurodipterinen, &c., des Devonischen Systems, p. 11-12. St. Petersburg, 

 1860. 



