NATURAL CLASSIFICATION OF THE TRILOBITES 117 



Morphology of the Cephalon. 



The structure of the trilobite head suggests homologies 

 which should be noticed here, and if these correlations are 

 based upon true structural likenesses, they serve not only to 

 emphasize the primitive character of the trilobite, but also 

 aid in interpreting certain organs and structures in the higher 

 Crustacea. 



The five fused somites of the crustacean head are generally 

 believed to correspond to the third, fourth, fifth, sixth, and 

 seventh neuromeres, leaving the first and second unrepre- 

 sented either by distinct segments or appendages. These 

 two neuromeres commonly constitute most of the cerebral 

 mass above the oesophagus, and innervate the ocelli and paired 

 eyes. In some the antennae are innervated from supra- 

 cesophageal ganglia, while in other forms their ganglia are 

 infra-cesophageal. It was formerly supposed that the stalked 

 eyes of the higher Crustacea represented appendages of a 

 head segment, but this belief has been abandoned on account 

 of the derivation of stalked out of sessile organs, as in 

 Peneus, and also because the eyes do not always have a 

 relatively fixed position, but may pertain to the first, second, 

 or third head segments. Their structural position in the 

 trilobites, however, is invariable, and it seems probable that 

 in some families of higher Crustacea the eyes are in exact 

 correlation, and may be similarly interpreted. 



The writer 5 has previously discussed this question, and 

 adduced reasons for considering the free-cheeks in trilobites 

 as "the pleura of an oculiferous head segment." In many 

 species (Dalmanites, j?Eglina, etc.) the free-cheeks are con- 

 tinuous, forming one piece extending around the front of the 

 head, between the cranidium and the hypostoma, while in 

 others there is a separate piece, the rostral plate, between the 

 proximal ends of the cheek pieces holding a like position. 

 These structures occupy the exact position of a true segment, 

 and since, upon theoretical grounds, additional head seg- 

 ments are to be accounted for, the only satisfactory correlation 



