12 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. 



descent 1 It may be well at this point to inquire into the question of 

 homologies a little more fully. 



Various Arthrodiran genera may be selected for comparison with 

 Neoceratodus, and the cranial structure of the two types be examined 

 in most critical light ; it will be found that intimate correspondence 

 exists throughout. Inasmuch, therefore, as the skull of typical Arthro- 

 dires was constructed upon essentially the same model as in I^eocerato- 

 dus, the latter becomes a standard for interpi'eting certain minor details 

 which have hitherto been misunderstood. These we shall have occasion 

 to refer to presently. In comparing the cranial pattern of the Cera- 

 todout and Arthrodiran type, it will not do to confine our attention to 

 any one genus of the latter ; we must consider the range of variation 

 exhibited by the group as a whole. Thus, at first sight, it would seem 

 almost impossible to coordinate the median series of plates in Neocera- 

 todus (Fig. B) with those of Dinichthys (Fig. A), although the lateral 



Fig. ^. — Dinichthys ptistulosus Eastman. Middle Devonian; Iowa. Restoration of 

 the headshield, dorsal aspect, x i- C, central; EO, external occipital; M, marginal; 

 MO, median occipital; P, pineal; PO, preorbital; PtO, postorbital; R, rostral or meseth- 

 moid. Sensory canals represented by double dotted lines. Suborbital omitted. 



series stand in sensible agreement. The equation is readily solved, 

 however, by substituting Macropetalichthys in place of Dinichthys as an 

 intermediate term of comparison, on the principle that things equal to 

 the same thing are equal to each other. Both in Macropetalichthys and 



