GREGORY: FISH SKULLS 123 



The bony scutes of the sturgeons, including those in the rostrum, seem to me to be 

 obviously secondary characters developed after the loss of the ganoine, analogous to the 

 ossified exoskeleton of certain highly specialized teleosts, or to the fragmented rostrum of 

 certain pycnodonts. 



In brief, I fail to see in the embryo strugeon any especially elasmobranch characters 

 not shown in other fish embryos, or any which would imply the derivation of the sturgeons 

 and spoonbills from some post-elasmobranch type that stood widely apart from the lines run- 

 ning to the palseoniscids and their allies. 



Moreover, many of the peculiar characters of the sturgeons are foreshadowed by the 

 Jurassic Chondrosteus (Fig. 195), which on the other hand retains features that are clearly 

 inherited from a palseoniscoid stock, as well noted by A. S. Woodward (1895, p. viii). 

 Watson (1925, p. 831) has already shown the annectant character of the Chondrosteidse be- 

 tween the palseoniscids and the sturgeons and concludes as follows: 



"Thus such new information as I can add only emphasizes that resemblance between 

 Chondrosteus and the Palseoniscids which Traquair long ago pointed out, and shows how 

 untenable is the view of Bridge, adopted by Sewertzoff, that the Acipenseroides are the 

 most primitive of bony fish and owe many of their peculiarities to a persistence of Elasmo- 

 branch structures." 



Protospondyli (Semionotids, Pycnodonts, Garpikes, Macrosemiids) 



Long ago (1895) Smith Woodward, developing the results of Agassiz, Traquair and 

 others but adding greatly thereto from his own observations, traced the morphological 

 evolution of the vertebrae, scales, caudal fin, skull, mandible, etc., of the Paleozoic and 

 Mesozoic ganoids. Except in a few instances all subsequent palaeontological investigations 

 have tended to confirm his conclusions. He showed also that while the known fossils 

 often afford a good morphological series indicating the evolution of the different organs and 

 parts, they do not give us a direct ancestral series, and that there are great breaks in the 

 record as we pass from a lower major group to the next higher one. Thus, as we have seen 

 above, the chondrostean series, very primitive at first, radiates in many directions, most of 

 the branches becoming extinct during the Mesozoic era but a few of them giving rise to the 

 modern spoonbills and sturgeons. The catopterid branch (Fig. 2\A), while progressing 

 toward the next higher order (the Protospondyli or Holostei), never attained that goal 

 but became extinct long after the early members of that group had been derived from some 

 still undiscovered stock. Hence we cannot yet aflSrm positively either that the small- 

 mouthed Semionotidse, which are the oldest known branch of the Protospondyli, were 

 derived from large-mouthed palseoniscoid ancestors, or that they again branched into 

 large and small-mouthed descendants; although that still seems to be by far the most 

 probable view. 



The inference that mouths that were once large became small and that some of the latter 

 again became large, does not necessarily involve a violation of the supposed law of "irreversi- 

 bility of evolution," for the new large mouths probably became so by a different method 

 from that which first led from small to large; that is, different elements were involved in 

 different ways, as will presently be shown. 



We need not be surprised that there are still many gaps in our records if we realize 

 how relatively few fish-bearing horizons are known throughout the five hundred million 



