No. 444.] PTERASPIDjE AND CEPHALASPIDsE. 855 



Assuming that to be the case, we may form some idea of their 

 probable mode of development by a comparison with those of 

 Limulus. Miss Hazen and the writer have shown that here 

 each abdominal appendage arises first as a fold of ectoderm, 

 into which grows a muscle bud that soon divides into two princi- 

 pal parts to form the adductor and abductor muscles. Mean- 

 time the nerve to the appendage appears and an axial core of 

 cartilage is formed which grows from the basal mesoderm 

 through the middle of the muscle cells, toward the apex of the 

 appendage. 



As these processes agree in ever)' essential particular with 

 those known to occur in a segment of the lateral fold of verte- 

 brates, there can be no serious objection, from an embryological 

 standpoint, to the interpretation of the lateral fold as a series of 

 fused abdominal appendages. Assuming then that the lateral 

 fold is formed, phyllogenetically, by the fusion of a series of seg- 

 mentally arranged, and independently movable structures, such 

 as the fringing processes of the ostracoderms, it is clear that 

 the oar-like cephalic appendages of the ostracoderms cannot he 

 regarded as specializations of either a lateral fold, or of gill arches 

 in the (iegenbaurian sense. On the contrary we must consider 

 the paired cephalic appendages, gill arches and fringing processes 

 as various modifications of one set of serially homologous st rue- 

 such appendages in the trunk region. 



I assume, therefore, that the highly specialized condition of 

 the visceral arches and appendicular structures of modern fishes 



Even in the more remote ancestors, such as the Phyllopoda, 

 Trilobita, Phyllocarida and Merostommata. These appendages 

 varied greatly in form and function in different animals, and in 

 different regions of the body in the same animal. 



In the ostracoderm type, we may assume that certain ones of 

 the anterior cephalo-thoracic appendages were retained as oar- 

 like swimming appendages. Two or three pairs were retained 

 about the mouth followed by several pairs of respiratory append- 

 ages of an unknown character. 



