490 Prof. H. H. Swinnerton — Classification of Trilobites. 



absence of a facial suture in members of this last-named family is not 

 regarded as barring them from admission to the Opisthoparia, or as 

 justifying their reference to the Hypoparia. 



The facts given above together with the evident relationship of 

 Dionide i to the Trinucleidae on the one hand and to the Harpedidse 

 on the other creates a strong presumption in favour of the view that 

 the ocelli of the latter are not of different origin from but are vestiges 

 of compound eyes. 2 If this be so they may mark the position of an 

 opisthoparian type of facial suture which has disappeared under 

 similar influences to those which have acted on Trinucleus. 3 



Lake, 4 discussing the systematic position of Shnmardia, says : " The 

 genus Shumardia has by some writers been placed in the family 

 Agnostidse. Hoberg, however, in 1890 pointed out that it presents 

 scarcely any resemblance to Agnostus, except in the absence of eyes 

 and facial sutures, and he concluded that it belongs rather to the 

 Olenidse. At a later date Pompeckj considered it to be most closely 

 related to Conocoryphe, while Keed, like Moberg, included it in the 

 Olenidae." Allowing for the absence of eyes and sutures Shumardia 

 resembles Ellipsocephalus 5 much more closely than it resembles either 

 Agnostus, Olenus, or Conocoryphe. This comparison is analogous to 

 that of Trinucleus with Orometopus. The difference in time of 

 appearance of Shumardia and Ellipsocephalus, however, makes one 

 hesitate to place them near together. 



In discussing the Agnostidse Beecher claims to have found " under 

 favourable conditions of preservation " a distinct plate separated by 

 a suture which " can be compared only with free cheeks". 6 So far 

 as literature is known to me this is the only case in which Beecher 

 actually saw hypoparian free cheeks. It is interesting to note, 

 therefore, that Jaekel, 7 whose material seems to have been excep- 

 tionally good, definitely states that sutures are absent. From the 

 fact that some workers seem to have found vestiges of eyes on the 

 cheeks he expresses the opinion that the Agnostidse are forms in 

 which the eyes have degenerated and the free cheeks have fused 

 with the fixed cheeks. He shows, 8 moreover, that the habit these 

 Trilobites had of shutting themselves up with the edge of the 

 pygidium fitted closely against that of the cephalon necessitates 

 absence of eyes and, correlatively, of facial sutures on the ventral side. 



1 F. E. C. Seed, Geol. Mag., 1912, p. 202. 



2 Cf. F. E. C. Seed, Geol. Mag., 1898, p. 446 et seqq. 



3 Some time after this paper was finished Mr. H. H. Thomas kindly called 

 my attention to an article on Harpes, by Eudolf Eichter, in the Zoologischer 

 Anzeiger for December, 1914, a copy of which had reached England. This 

 author has made a very careful study of the minute structure of this genus. 

 He shows that the ' eyespots ' consist of biconvex lenses like those of the 

 normal Trilobite eye, and he regards them as vestiges of a once well-developed 

 ■compound eye. He also shows that the marginal suture cannot be regarded as 

 homologous with the true facial suture. 



4 Monogr. Palaeont. Soc, 1907, p. 40. 



5 Vide infra. 



6 Amer. Jowrn. Sci., 1897, p. 183. 



7 Op. cit., p. 387 ; vide also Woods, op. cit., p. 225. 



8 Op. cit., p. 388. 



