On the structure and classification of the Tremataspidae. 27 



The Tremataspidae resemble the Arthropods in : (a) their general arthropod appear- 

 ance ; b) in the minute structure of the shell ; c) in the structure and arrangement of the 

 openings for the eyes and olfactory organs ; d) in the character and arrangement of the oral 

 plates; e) in the absence of vertebrate jaws and mouth and in the presence of a small cen- 

 trally placed oral opening similar to that of an Arachnid ; f) in the presence of numerous 

 pairs of jointed appendages. 



The presence of segmented appendages in Tremataspis and related forms is indicated 

 by the following evidence : 1) the presence of a pair of oar-likc and jointed appendages in 

 Pterichthys, Bothriolejris, Cyathaspis, Pteraspis, Tolypaspis and Tremataspis, unlike those of 

 any true vertebrate; 2) the presence of a fringe of jointed and movable appendages (23 — 

 30 pairs) along the ventral margin of the trunk of Cephalaspis. 3) the presence of a pair of 

 crushing mandibles like those of an Arthropod in the head of Cephalaspis ; 4) the presence 

 in Pterichthys and Bothriolepis and Tremataspis of oral plates that appear to be movable 

 laterally, like the jaws of an Arthropod ; 5) the presence in Tremataspis of a series of eight 

 other pairs of openings like the ones to which the large swimming appendages are attached ; 

 6) the presence in Tremataspis of two pairs of entapophyses which, like those in Limulus 

 apparently serve for the attachment of muscles moving several pairs of appendages ; 7) the 

 presence in Cyathaspis and Pteraspis of from four to six radiating grooves on the inner sur- 

 face of the dorsal shield, which like similar markings in Limulus indicate the points of at- 

 tachment of dorso-ventral muscles moving several pairs of appendages. 



The concurreut testimony, from so many different and independent sources, to the fund- 

 amental similarity between the Tremataspidae and the Arthropods establishes beyond 

 question that the resemblance between them it due, not to any incidental parallelism, or t 

 mimicry, but to genetic relationship and community of origin. It could not have been a re 

 mote relationship, signifying merely a common origin of both types from some auneliden an- 

 cestor, because the resemblance consists in the common possession of highly specialized 

 structures characteristic of the fully established Arthropod and Vertebrate types. 



The genetic relationship, therefore, between the Tremataspidae and the Arthropods 

 can mean mothing less than the derivation, through changes in structure and function, of 

 one type from the other. 



But we must not too hastily conclude that the Tremataspidae are true Arthropods, and 

 by transferring them from one group to the other avoid the real problem at issue. 



We cannot so easily ignore the profound significence their structure has for the origin 

 of Vertebrates. Their vertebrate affinities are too obvious to be denied. 



It is quite out of the question to separate the Tremataspidae very far from the 

 Cephalaspidae, as is conclusively shown by the similarity in the structure of their orbits 

 or from the Pteraspidae and Asterolepidae, as is shown by the appendages and the lateral 

 line markings ; or from either, as is shown by the microscopic structure of the shell. The 

 whole group must be kept together. But, as it is quite impossible to include in the 



4* 



