328 AXXALS XEW YORK ACADEMY OF SCIEXCES 



forward nearly to the tip of the rostrum: in others it is much shorter; 

 none have the separate nasals which appear in higher types ; on the other 

 hand, the transverse occipital row of small plates (see below, p. 329) seen 

 in Osteolepis and Gyroptycliius, are barely indicated in Tristichopterus 

 and had not split off in Rhizodopsis or Megalichthys. 



As the conical teeth of Crossopterygii were on the outer margins of the 

 jaws and the movements of the jaws were chiefly vertical, the suggestion 

 may perhaps be hazarded that the transverse stresses conditioned the 

 evolution of the longitudinal median sutures, while the anteroposterior 

 stresses resulted in transverse sutures, those separating the rostro-frontal 

 and postorbito-parietal segments being in Osteolepis particularly deep. 

 At any rate the resulting skull pattern includes many more or less rect- 

 angular plates, as also in the Actinopterygii. In the primitive Dipnoi, 

 on the other hand, the teeth were clustered in fan-like ridges on the roof 

 of the mouth and the movements of the jaw were more oblique: this 

 possibly may have partly conditioned the oblique character of the skull- 

 roof sutures and the polygonal form of the elements. In the Stegocephali 

 both rectangular (e. g., Branchiosaurus) and oblique (e. g., Stegops 

 divaricata Moodie) types occur. 



After the primary sutures had been established and the continuous 

 dermocranium fragmented into small plates, there usually followed an 

 enlargement of certain plates, and sometimes a coalescence of adjacent 

 plates, as in the Dipnoi and other groups of fishes : such enlargement and 

 coalescence following ex hypofhese not only upon the shifting of lines of 

 greatest mobility or the diminution of mobility between elements, but 

 also from the circumstance that as a general rule in evolutionary series, 

 after individual structures have been differentiated out of a former con- 

 tinuum they appear to acquire a certain evolutionary initiative, so that 

 they may enlarge or decrease or shift their position, or crowd apart their 

 fellows in a manner strongly hereditary, but often without any assignable, 

 immediately adaptive purpose. 



The position and apparent homologies of the elements of the skull- 

 roof in the Ehipidistia, in comparison with other primitive fishes and 

 with the Stegocephali, are as follows : 



The orbits in primitive Crossopterygii are small and sometimes far 

 forward (Osteolepis, Fig. 2), quite near the front of the nasal rostrum, 

 which is extremely short and obtuse : the same is also true of the most 

 primitive Actinopten-gii (Palaeoniscidae). The roof of the nasal rostrum 

 (ethmoid) is undivided and there are no paired nasal bones. The short 

 frontals, lying between the orbits, ahnost form part of the rostrum. 

 Between them in some genera (Osteolepis. Ghiptopowus, Diplopterus) 



