Actiniae. 301 



111 botli the species from the Antarctic that I have described tlie 

 broud chambers are practically identical, diliering only in minor 

 details, so that the description here given must be understood to 

 apply to both species, unless otherwise noted. Specimens possessing 

 brood chambers in these two species may be at once recognised, in 

 whatever stage of development they may be in, by the presence of a 

 constriction or groove more or less marked, in the body-wall, running 

 completely around the animal at a distance from the parapet of 

 about one-third in the total height of the column. In addition the 

 body- wall above the constriction is more or less wrinkled, and the 

 verrucas become confused with vermiculate furrows, especially in 

 XJrtidna sulcata, where also the body -wall is thinner than in the 

 lower part of the body. In this last-named species, specimens 

 having well-developed embryos in the brood chambers have this 

 upper thin-walled portion of the column, as well as the oral 

 disc and tentacles, withdrawn in the contracted condition to 

 the level of the top of the brood chambers. The thinning of the 

 body-wall above the invagination constriction is not so noticeable 

 in U. carlgreni. 



The earliest phase in the development of the brood chambers that 

 I find among these Antarctic specimens shows a series of invagina- 

 tions of the body-wall along the line of this constricted zone. The 

 invaginations are already considerably advanced and form distinct 

 sacs from 2 to 3 mm. deep, projecting on the inside of the body- 

 wall into the ccelenteron (fig. 11). They appear to have arisen by 

 invaginations of the body-wall corresponding to the mesenteric 

 spaces, for, when looked at from below, the lower ends of the inva- 

 ginations are seen projecting between the mesenteries. Horizontal 

 sections, however, show that the partition walls between adjacent 

 invaginations have broken down and the cavities are more or less 

 continuous, but extend deeper in the mesenteric spaces. This last 

 feature — the sending of prolongations down between the mesen- 

 teries — is characteristic of all stages in the development of the 

 chambers that I have examined, and suggests their method of 

 gradual enlargement. As seen from the figure, which is taken from 

 a specimen of U. carhjrcni (fig. 11), tlie chambers are absolutely 

 empty and open freely directly to the outside. The walls consist 

 of invaginated ectoderm, inesoglcea, and endoderm, which three 

 layers are fairly thick, except at the innermost parts, where 

 both mesoglaa and endoderm are thinned out somewhat, especially 

 the former. 



The junction of adjacent cavities appears to go on until there 



