6 George Thomas Hargitt 



men in which the distal end was in the sand a new stolon-hke 

 growth formed from it and pushed up through the sand. 



ExrERiMENT II. — This experiment with Tiibitlaria crocea was 

 conducted in the same manner as experiment 5. The pieces, 

 however, were all placed with the distal ends in the sand. In 24 

 hours a hydranth had formed on the end of a branch (fig. 5). 

 In 36 hours more three more hydranths had formed, all on the 

 ends of branches, some extending laterally, others downward ; 24 

 hours later four hydranths had developed on other branches, one 

 extending laterally and three pointing downward. Two hy- 

 dranths had regenerated at the end of the main stem, i. e., at the 

 proximal end. After this the pieces Ijegan to degenerate and 

 soon died. Thus, in this experiment, covering a period of about 

 80 hours, ten hydranths regenerated. Of these, eight developed 

 at the distal ends of branches and two at the proximal ends of 

 stems, which would seem to indicate a rather marked polarity. 

 The formation of hydranths at the proximal ends of stems was 

 always much slower in this species than tlie development at the 

 distal ends. 



Femiaria was found growing under entirely different condi- 

 tions in dift'erent habitat, though the morphological differences 

 Avere slight. The most abundant was that growing on the piles of 

 docks in the harbor. The other form was attached to floating eel 

 grass and matured later than the previous form (cf. Hargitt. 

 1901, p. 224). Experiments to determine regeneration were tried 

 with both forms. 



ExPERniENT 13. — Pcnuana from the piles of docks. — The 

 pieces were simply laid flat on the bottom of the dish. Regener- 

 ation was rather slow. After some time the coenosarc emerged 

 from the perisarc. If a hydranth was to regenerate the coenosarc 

 enlarged into a knob from which tentacles budded and the hy- 

 dranth formed. If, on emergence fromlhe perisarc, the coenosarc 

 came in contact with the bottom or tlie side of the dish it flat- 

 tened out to form holdfasts (fig. 10). The behavior of these 

 pieces resembles very closely that of Mar gel is noted by Loeb 

 (1893). Every branch or stem which came into contact with the 

 dish formed holdfasts, and hydranths formed only from the free 



280 



