i6 



SMITHSONIAN M iSCKLLANEOUS COLLECTIONS VOL. 8l 



and secondarily spreads to other tissues. The older locomotion theory 

 was defective in that it attributed the formation of segments to the 

 mechanical stress of movement. 



At the completion of metamerism, a segmented animal has attainel 

 a generalized structural stage in which it consists of a segmented body 

 part coextensive with the length of the alimentary canal (fig. ii), 

 and of a prostomial region (Pst) anterior to the mouth (Mth). Since 

 the mouth in annelids and arthropods marks the site of the original 

 anterior extremity of the blastopore on the ventral surface of the 

 body (figs. 4 B, 5 A, Bp), it is evident that nicsodcniuil segments can 

 not be formed morphologically anterior to the mouth, and therefore, 

 that the preoral region is never truly segmented. The common idea, 

 then, that the arthropod mouth lies behind the first head segment, or, 

 as proposed by some writers, behind the second or even the third seg- 



Stom 



Ment 



Proc 



VNC 

 -Diagram of the structure of a theoretically generalized segmented 



Fig. II.— 

 animal. 



An, anus; Arc, archicerebrum : Mcut, mesenteron; Mth, mouth; Ppt, peri- 

 proct; Proc, proctodeum; Pst, prostomium ; Stoiii, stomcdeum; VNC, ventral 

 nerve cord. 



ment, disregards the fundamental relation between the endodermal and 

 mesodermal layers. Segmentation can not exceed the extent of the 

 mesoderm, and the primitive extent of this layer in the annelids and 

 arthropods is defined by the positions of the mouth and the anus. The 

 blastopore never extends quite to the true cephalic extremity. The 

 stomodeal invagination, which gives rise to the definitive mouth, is 

 thus preceded by an unsegmented prostomium. The closed posterior 

 end of the blastopore, however, is at the posterior extremity of the 

 body, where the blastopore and endoderm originated, and the later 

 formed ]n-octodeum, therefore, opens terminally in the periproct. In 

 some arthropods a median lobe, or suranal plate, grows out over the 

 anus from the periproct, and simulates the prostomial lobe at the 

 anterior end of the body. Likewise, there may be lateral and subanal 

 lobes of the periproct. 



In as much as the most important evidence of the segmentation of 

 the arthropod head is derived from a study of the cephalic nerve 



