REPORT ON THE HEXACTINI^E 21 



in the regions of the siphonoglyphes. That irregularities occur 

 the arrangement of the latter structures has already been noted. 

 In one of the three specimens which I sectioned I found three 

 siphonoglyphes, while in the second individual though only two 

 were present they were not opposite one another, thirty-seven 

 pairs of mesenteries intervening between the two on one side while 

 on the other side there were fifty-five pairs. In the third speci- 

 men but a single siphonoglyphe occurred. 



In connection with the siphonoglyphe a peculiarity of structure 

 exists. The endoderm over the general surface of the stomato- 

 daeum is comparatively low and inconspicuous, but in the region 

 of the siphonoglyphes it suddenly becomes very high and retains 

 this condition over the entire surface of the grooves (Fig. 12). 

 I speak of the endoderm being thickened, but the appearance is 

 rather as if the endodermal surface of the mesogloea were drawn 

 out into exceedingly fine fibrillae which anastomose with one 

 another to form, a reticulum with elongated meshes, the endo- 

 dermal cells being arranged on the terminal part of the fibrillae. 

 The appearance presented is very similar to that which I have 

 described as occurring in the mesenteries of Cerianthiis aineri- 

 canus ('90). 



The individual with three siphonoglyphes had three pairs of 

 directives, while the one with two siphonoglyphes had corre- 

 spondingly two pairs unequally spaced, and the third specimen 

 with one siphonoglyphe had but a single pair. As might be ex- 

 pected from the arrangement of the directives considerable irregu- 

 larity occurred in the mesenteries. In the upper part of the col- 

 umn each alternate pair seemed to be perfect as a rule, but lower 

 down seven imperfect pairs intervened in certain sectors between 

 successive perfect ones. It may be supposed from this that there 

 were typically four cycles of mesenteries, three of which were 

 perfect and one imperfect. The typical arrangement occurred 

 however only in a few sectors ; examples of a fifth incomplete cycle 

 occurred here and there and in addition both imperfect and per- 

 fect pairs were not infrequently intercalated so that it was impos- 

 sible to determine even whether the arrangement was hexamer- 

 ous or not, although it may be supposed that in one specimen at 



