THE DERMAL FIN-RAYS OP FISHES. 475 



an appearauce. Beaudelot (2), Hertwig (17), Ryder (32), and 

 especially Harrison (13), have given such complete accounts 

 of their development that there remains little to discuss 

 beyond their relation to the basement membrane, and the 

 mode of origin of the articulations. 



The lepidotrichia are first seen in transverse section as 

 thickenings of the basement membrane laid down over an 

 area of considerable width by the underlying mesoblastic 

 scleroblasts (figs. 24, 20, and 21). At this early stage in the 

 development of the ray, the thickening is directly continuous 

 with the membrane extending in unbroken line between 

 adjacent lepidotrichia. As the thickening becomes more pro- 

 nounced, it begins to bend with its convexity outwards, and 

 the edges become more and more sharply marked off from the 

 ordinary basement membrane. The scleroblasts now begin to 

 insinuate themselves round the edge of the growing lepido- 

 trich as it becomes, so to speak, peeled off from the 

 membrane (fig. 22). During the whole process of the 

 separation of the dermal ray from the basement membrane, 

 the latter is never broken. New cuticular substance is 

 deposited as fast as the mesoblastic cells grow round on to 

 the outer surface of the ray. For a long time, perhaps 

 thi"Oughout life, strands of cuticular substance extend from 

 the basement membrane to the lepidotrichia^ which never 

 sink very far from the epidermis, except quite at their 

 proximal ends (fig. 23). 



The young lepidotrichia are flattened in section, and grow 

 chiefly by the addition of new layers of bony substance on 

 their outer and inner surfaces. At first, as is generally the 

 case with Teleostean bones, Ave find only a thin layer enclosing 

 no bone- cells; but as the ray grows older and thicker, cells 

 become included, and it acquires the normal structure of 

 Teleostean bone. 



The branching of the lepidotrichia is not due to the proximal 

 fusion of originally distinct rays, as Hertwig states (17), but, 

 on the contrary, to the repeated subdivision of the groAving 

 distal end of an originally single ray. That this is the case in 



