THE AMERICAN NATURALIST. [Vol. XXXVII. 



Since writing the above description of the pectoral append- 

 ages of Tremataspis, a new specimen of the distal joint has come 

 into my possession, Fig. 6 A. It is unquestionably of the same 

 character as that shown in Fig. 3. It has the same general form 

 and texture but differs from it in certain details. It is consider- 

 ably smaller, measuring only 5.66 mm. in greatest length by 

 1.65 mm. in greatest width. There are six or seven very prom- 

 inent marginal serrations or spines. The curvature of the dor so- 

 median surface indicates the presence of the peculiar shoulder 

 seen in Fig. 3, but the surface is broken along the edge where 

 such a shoulder should be. The smooth anterior margin of the 

 shining surface layer appears to have been straight instead of 

 scalloped. There is the same peculiar wrinkled and lustreless 

 surface on its median side, extending not quite to the distal end. 



At the proximal end, this layer forms an irregular prominent 

 surface with two or three round openings for the passage of 

 blood vessels, but without any trace of an articular condyle, or 

 articulating surface, such as was seen in the first specimen. 

 The surface at this point is intact, showing that an articulating 

 surface could not originally have been present. 



While there can be no doubt that the two structures in ques- 

 tion are of the same nature, and that both belong to Tremataspis, 

 the differences between them, in the above mentioned details, 

 are so great that we can hardly regard them as identical. We 

 are left then to the conclusions that ( 1 ) they are either homolo- 

 gous structures from different species of Tremataspis, against 

 which it may be urged that the surface texture is the same in 

 both and like that known in T. schmidtii only. (2) or they 

 may be serially homologous structures from the same species. 

 This conclusion harmonizes with the fact that both structures 

 agree in surface texture but not in form, and with the fact that 

 there is a series of marginal openings on the ventral surface 

 serially homologous with the anterior ones to which the pectoral 

 appendages are attached. I have already suggested that these 

 openings indicated the presence of other appendages of a smaller 

 size than the pectorals, although there was no evidence at that 

 time available that furnished any hint as to what the character 

 of these appendages might be. This fact we have always borne 



