222 E. A. ANDREWS. 



The radiating vertical canals are very long in the deep fin beneath 

 the posterior myotomes and present slight proximal swellings in 

 the region just posterior to the anus ; swellings that are like those 

 more posterior ones of the dorsal fin represented in Fig. 25. 



Anterior to the anus, however, these spaces are not continued as 

 fin-ray spaces similar to those of the dorsal fin ; in fact there are no 

 representatives of such spaces and the lymph canals cease to form 

 parallel, vertical channels and anastomose in an irregular way. 



There are thus no double fin-rays in the region between the 

 anus and atriopore though such are believed to exist in all other 

 Acraniata : a striking and important difference. 



At the extreme anterior tip of the animal the ventral fin has 

 the same structure as the dorsal, with which it is continuous around 

 the anterior end of the notochord (Fig. 10). 



In this region the ventral fin has also a large median lymph 

 space that anteriorly is somewhat sub-divided near the tip of the 

 notochord, but posteriorly expands vertically as a large median 

 space that penetrates some distance into the ventral fin and then 

 rises again to end abruptly anterior to the pre-oral chamber oppo- 

 site the posterior end of the second myotome (Fig. 10). In trans- 

 verse section, near the tip of the notochord, this ventral space 

 (Fig. 11) is smaller than the dorsal space, but in sections posterior 

 to the brain it is in every way larger than the dorsal spaces above 

 the nerve cord. 



Between this anterior end of the median fin and the posterior 

 part in the caudal process important modifications of the median 

 fin, in fact its entire disappearance as a median fin, are brought 

 about in connection with the openings of the digestive and 

 branchial cavities. These modifications are the departure of the 

 median fin from a true median position, and the substitution of the 

 paired rnetapleural folds in place of a median fin throughout the 

 length of the branchial region. 



The appearances seen in the youngest specimens favor that ex- 

 planation of the formation of the atrium given by Willey (26) so 

 that we may adopt the terms sub-atrial ridges for the horizontal 

 outgrowths of the metapleural folds and abandon the older idea of 

 extensive epipleural downgrowths. The complete metapleura have 

 the same structure and relationship as in the common amphioxus, 



