The Latero-Sensory Canals and Related Bones in Fishes. 403 



of the Muraenidae. In every such recorded case, however, the section 

 of canal so enclosed is not included, between, and thus so defined and 

 limited by, two primary tubes of the canal to which it belongs; and 

 hence is not a section of that canal similar to the, in principle, pri- 

 marily independent sections that, in Amia [^], fuse to form a contin- 

 uous canal. Such primarily independent sections of a continuous canal 

 normally each contain, and are developed in relation to, a single sense 

 organ; and a latero-sensory ossicle develops in relation to, and appa- 

 rently for the protection of that organ. This was fully set forth in 

 my description of the development of the canals in Amia; and it is 

 those ossicles that develop in relation to a single sense organ, and 

 those ossicles only, that are units of the latero-sensory skeleton. 

 Where two or more organs are enclosed in a single bone, that bone 

 contains a corresponding number of fused latero-sensory units. Where 

 there is no enclosed organ, the canal has either been secondarily en- 

 closed in an underlying and primarily independent bone, or it is en- 

 closed in what is quite certainlj^ simply a detached portion of a next 

 adjoining latero-sensory unit. The latero-sensory skeletal units, al- 

 though each forming a morphological whole, are certainly not alwaj^s, 

 from their earliest beginnings, an actual and continuous whole deve- 

 loped centrifugally from a single centre. On the contrary, spicules 

 of bone appear along the growing edges of a central portion, at first 

 wholly independent of that portion but soon absorbed by it. Mecha- 

 nical causes, simply, might prevent certain of these fusions and so 

 give rise to tAvo or more portions of what is morphologically a single 

 latero-sensory unit. Furthermore, the conditions found in Chimaera, 

 where each organ is said to be protected by a series of half rings 

 of a bony material, and the conditions described b}" Schleip in em- 

 bryos of Salmo, where the extrascapulars are said to first appear as 

 "a large number of disconnected little plates of bone", would both 

 seem to indicate that the latero-sensory units, in the earliest phylo- 

 genetic stages of their development, were each normali}'' represented 

 by several independent portions. But each such unit is always deve- 

 loped for the protection of, and hence in some direct relation to a 



single sense organ, the only apparent exception to this rule, that I 



26* 



