REPORT ON THE HEXACTINLE 7 



numerous and extend quite up to the endodermal surface of the 

 mesogloea, even the processes found on that surface containing 

 numerous cavities. In Fig. 3, finally another arrangement is 

 shown. The preparation from which the drawing was made 

 was from a brown specimen which measured 3 cm. in height, 

 and the peculiarity which it presents lies in the band of mesogloea 

 destitute of cavities which traverses the muscle longitudinally, 

 dividing it into an outer and an inner portion, the latter being 

 separated from the endodermal surface by a distinct interval as 

 in Fig. I. This last condition recalls the arrangement figured 

 by Carlgren ('93) for the European M. dianthus and may be 

 termed the " layered" condition. 



The occurrence of a band of mesogloea, destitute of muscle 

 cavities, between the inner surface of the sphincters and the en- 

 doderm is probably to be regarded as leading to a layered con- 

 dition of the muscle. If, in a specimen similar to that from 

 which Fig. i was taken, the inclusion of muscle fibers within 

 the mesogloea were to occur again with the continued growth 

 of the individual, an arrangement would be found similar to that 

 seen in Fig. 3, and an alternation of periods of growth during 

 which inclusion went on with periods in which it ceased, would 

 result in the arrangement figured by Carlgren. It may be pre- 

 sumed that in the larger specimens of M. diantlnis a layered ar- 

 rangement of the sphincter will be the most frequent, but cases 

 like that represented in Fig. 2 show that it is not an invariable 

 arrangement for the species. 



I have examined the structure of the cinclides in one speci- 

 men of the collection and find that it agrees with what Carlgren 

 has described for the European individuals, the canals being 

 lined by endodermal cells, so that the cinclides may be regarded 

 as endodermal evaginations. The same is true with regard to a 

 specimen from our eastern coast (the M. margiiiaUini Auct.), 

 and this mode of formation may probably be regarded as typical 

 for Metridiiim dianthics, though further observations are neces- 

 sary to determine whether it can be regarded as characteristic of 

 the entire subfamily. 



The longitudinal musculature of the tentacles, and the radial 



