308 DROMALIA ALEXANDRI. 



in a straight line. At this level, however, the series turns abruptly to the left 

 (as viewed from the side on which they are situated, i. e. the ventral face), and 

 from here onward they run in a helical spiral until they reach the basal extremity 

 of the siphosome. The spiral arrangement is often obscured by the closely 

 crowded cormidia, but it is clearly shown in views of the basal surface of the 

 corm with the distal portions of the cormidia torn ofT (Plate 23, fig. 11). A 

 similar arrangement has already been noted by Haeckel, who says ('88b, p. 288) 

 of Rhodaliidae in general, that "the cormidia . . . are arranged in regular 

 circles or spiral coils." In the "Siboga" specimens of Archangelopsis the corm- 

 idia were so crowded and contracted that Lens and Van Riemsdijk were unable 

 to make out their arrangement. But in the single "Albatross" example of that 

 genus a similar spiral arrangement appears to occur, though, owing to the con- 

 traction, it is not so clear as it is in Dromalia. 



The adult cormidia very closely resemble those of Rhodalia and of Steph- 

 alia, (Haeckel '88b). They are situated on conical gelatinous prominences each 

 traversed by a canal connecting with the general vascular system of the corm 

 (Plate 23, fig. 9). Near the region where the young cormidia are formed, each 

 one stands alone on its prominence. But with advancing age, additional cor- 

 midia are formed from buds which develop from the prominences near the 

 bases of the existing siphons, the result being that near the base of the corm 

 each prominence bears two or three complete cormidia just as in Stephalia 

 (Haeckel, '88b, pi. 6, fig. 35). 



Each cormidium consists of the usual parts, siphon, gonodendron, and 

 tentacle. The siphons show no features of special interest. Each gonodendron, 

 as in Rhodalia, consists of a gelatinous stalk with several terminal branches, 

 which bear the numerous gonophores and also from two to four long, thin-walled 

 palpons. Unfortunately all of the older gonophores were lost, only the J'oung 

 stages remaining attached. For this reason it is impossible to determine whether 

 there are both 9 and cf cormidia on the same corm in Dromalia, as Haeckel 

 thought was the case in Rhodalia, or whether gonophores of only one sex are 

 present as Brooks and Conklin have more recently ('91) maintained for a speci- 

 men probably belonging to the latter genus. For an account of the complicated 

 structure of the 9 gonophore in this family see Brooks and Conklin ('91). 



Tentacles. These have a well-developed suspensorial membrane (Plate 23, 

 fig. 9) and bear tentilla just as in Rhodalia. The structure of the tentilla is 

 one of the points of difference between Dromalia and Rhodalia, for while in 

 Rhodalia they have no involucre and only a single terminal filament (Haeckel, 



