435 



their distinctness. Each facet is, in the younger forms, slightly 

 raised in the centre and contains either a slight tubercle or a shallow 

 depression. Just below the vitreous layer the canaliculi are slightly 

 enlarged and very densely packed together, so that they produce a 

 rather thin layer, black in transmitted, and white in reflected light. 

 The remainder of the cuticula layer is longitudinally laminated and, 

 when its minute parallel canaliculi are filled with air, is very sug- 

 gestive of dentine. 



2) The middle layer is crossed in various directions by the 

 chitenous trabeculae containing the lacunae. The interspaces are 

 filled with loose connective tissue through which ramify nerves or 

 bloodvessels. The outer portion of the layer contains a varying 

 number of small chambers and canals — so it may be called the 

 reticulated layer 1 ) — to distinguish it from the inner, or cancel- 

 lated portion where the large cavities, or sinuses, prevail. 



3) The third, or inner, layer is composed of trabeculae arranged 

 parallel to the outer surface. It is horizontally laminated and is 

 pierced with large irregular openings through which nerves and blood- 

 vessels pass to the cancellated layer. 



Here then we have a most remarkable dermal structure, for noth- 

 ing, bearing the slightest resemblance to it, in coarse or minute 

 structure, is known in any other invertebrate. I venture to predict, 

 however, in view of their close relationship to Limulus, that similar 

 conditions will be formed in the Trilobites and Merostommata when 

 they are properly investigated. Sections have been made of some of 

 these forms, with negative results; but unless made through the 

 right places they can not be expected to show the structures in 

 question. 



With these possible exceptions then, the only animals known to 

 show such an exoskeleton as Limulus are some of the remarkable 

 fishes known as the Cephalaspidae. 



In Pteraspis, as described by Huxley and Lankester, we 

 have the same divisions of the exoskeleton, each with its special 

 characters very similar to those that I have just described in Limulus. 

 I shall return to this subject in more detail very soon. Meantime, 

 I will content myself with the assertion that no other animal now 

 known, not belonging to the Cephalaspidae, vertebrate or invertebrate, 

 resembles Pteraspis so closely in the structure of its exoskeleton 

 as Limulus. In fact it is safe to say that if nothing more was 



1) Not well shown in this particular section. 



