438 



could no more shed its dermal bones than a vertebrate could shed 

 its cartilaginous cranium or its vertebral column. This fact is of 

 great importance in explaining how dermal bones of vertebrates 

 could be derived from those of arthropods. 



In addition to its dermal skeleton, Limulus is provided with 

 three other kinds of skeletal structures of very obvious significance, 

 namely: 1) the cartilaginous gill bars, histologically very much like 

 the cartilage of Petromyzon; 2) the segmentally arranged carti- 

 lages partially surrounding the ventral cord, and 3) the cartilaginous 

 cranium. The latter is not a simple plate as described by Lankester 

 and others, bat, as I have determined recently, it has a distinct 

 roof in the occipital region similar to that I described in Scorpions. 

 The roof is partly membranous and partly cartilagenous and is sup- 

 ported by two pairs of slender cartilages that have heretofore escaped 

 notice. They form the boundaries of a large occipital foramen 

 for the exit of the spinal cord and several pairs of nerves. I 

 have also found two small foramina on the floor of the cranium 

 (posterior haemal margin) for the exit of another pair of nerves. The 

 whole forms a perfect picture of a simple cartilaginous cranium such 

 as we might expect to find in some primitive vertebrate. 



I will say in conclusion that the facts I have here presented are 

 easily demonstrated and their signification obvious. They cannot 

 be explained away as due to imperfect methods of any kind. They 

 must be met fairly and squarely for we must regard them, either as 

 the most extraordinary examples known of what, to hide our 

 ignorance, we sometimes called analogy (as though that was in itself 

 a sufficient explanation), or else as proof positive that these forms, 

 which agree so completely in the general appearance and in the 

 minute structure of their various independent sets of organs, are 

 genetically related. It may sound paradoxical to some, but in my 

 opinion, to accept the latter conclusion is the more conservative 

 course. 



If we adopt the former we destroy the very foundations of the 

 science of morphology, for we then confess that a most thorough- 

 going similarity of structure does not indicate genetic relationship, 

 and, furthermore, that all speculations based on the assumption that 

 they do are worthless. 



