IV. 



Note on the Evolution of the Scales 

 of Fishes. 



MUCH has been written by those who have studied the minute 

 structure of fish-scales, concerning their primitive nature and 

 mode of origin. It is generally admitted that the earliest hard parts 

 appearing in the skin of fishes, were merely isolated calcified points or 

 papillae, such as those in which the tubercles of sharks and skates 

 arise ; and that all the later types of scales and plates are only the 

 result of the fusion and elaboration of these fundamental structures. 

 Palaeontology, too, specially favours this idea ; and an elaborate 

 memoir just published by Dr. Rohon (3) shows more clearly than ever 

 how large a majority of the fragments of dermal armour of chordate 

 animals in the Upper Silurian rocks are almost identical in structure 

 with the comparatively simple skin tubercles of the sharks and skates. 



There is, however, another aspect of the subject which seems to 

 have been less studied, and to which it is now possible to make one 

 more small contribution. We refer to the question of the original 

 character and subsequent phases of evolution of the typical over- 

 lapping scales of the trunk of ordinary fishes. Professor Ryder has 

 inferred (4), from theoretical considerations, that these scales must 

 have been originally disposed round the body in a series of rings 

 corresponding to the successive plates of muscle ; and he attempts to 

 show, by an argument from mechanics, that they must have been at 

 first rhombic in shape, in consequence of the direction of the strains 

 arising from the working of the muscles during motion. It is obvious 

 that the only proof of the theory can be obtained from Palaeontology ; 

 and we propos2 in this brief note to summarise a few new or little- 

 known facts bearing upon the subject. 



Perhaps the simplest form of squamation met with in any fish- 

 like organism is that of the Lower Palaeozoic Cephalaspis and its 

 allies, in which (as Professor Lankester first remarked) each segment 

 of muscle seems to be surrounded by a distinct ring of six or seven 

 scales ; this ring being overlapped by the one immediately in front 

 and overlapping that next behind (2, 5). Towards the extremity 

 of the tail in Cephalaspis, however, where the motion is much greater 

 than in the rest of the trunk, each of these rings is subdivided into a 

 set of comparatively small rhomboidal scales— an arrangement 



