BRACHIOPODA. 583 



remains an entirely closed sac. Attachment is effected by the hind 

 (caudal) end (Fig. 464, e, /), and as soon as it has occurred the mantle- 

 lobes turn forward and envelop the first segment. 



The larva now assumes the adult condition ; the shell valves are 

 developed on the mantle-folds, the provisional setae are replaced by 

 the permanent ones, and a mouth is formed at the front end. The 

 exact relation of the mouth to the umbrella-like head segment is 

 obscure. It has been suggested that the mouth is formed upon 

 the ventral side of it, and that the head segment (Fig. 465) is really 

 a preoral lobe, like that of Phoronis which soon disappears or is 

 reduced to the epistome (lip of the buccal groove). The tentacles of 

 the arms appear on the inner side of the dorsal mantle lobe (Fig. 

 465, C). 



Nothing is known of the embryonic development of the Ecardines. 

 Their larvae differ from those of the Testicardines in possessing a 

 shell while free-swimming. A larva of Discina radiata with a large 

 process overhanging the mouth has been described by F. Miiller 

 (Mutter's Arch., 1860), and we owe an account of the larva of Lingula 

 to Brooks (Chesapeake Zool. Lab. Kesults, 1878). 



A consideration of the above facts does not lead us very far in deciding the 

 question of the affinities of the Brachiopoda. If we adopt the suggestion 

 made above, that the mouth is formed on 

 the ventral surface which, it must be 

 remembered, is opposed to Kowalevsky's /. 



statement that it is at the anterior end 

 we may regard the larva as a trochosphere 

 with a preoral ring of cilia and preoral sense- 

 organs. 



The presence of setae secreted in cutaneou 

 sacs, the indications of segmentation, and 

 the relations of the coelom in the adult, 

 further suggest Annelidan affinities. Finally, 

 the form of the central nervous system is 

 not opposed to this view of the relationship 

 of the group. If we adopt this view we 

 should regard the BncMvfa as Annelids 

 with three segments, marked in the embryo D alimentary canal ; Af anus ; L 

 by the annular constrictions of the integu- liver ; St rudiment of peduncle, 

 ment, and in the adult by the imperfect 



septa represented by the gastro- and ilioparietal bands, and by the two pairs 

 of nephridia in Rhynchonella, which have relations to these bands. 



But it must be remembered that according to present accounts there are no 

 traces of mesoblastic segments in the embryo, the coelomic sacs being con- 

 tinuous and not divided up by septa. This, however, may be due to imperfect 

 observation, a possibility which becomes the more likely when we remember 

 that the larvae are extremely minute, the observations very difficult to make, 



