430 University of California Puhlioations in Botany [Vol. 6 



upon the shapes that the cells of each species have assumed in the 

 course of their development. 



The genera, Hyella Bornet and Flahault (1888), and Radaisia 

 Sauvageau (1895) as originally described, are not so clearly distinct 

 as to make it a simple matter to assign new species to them with much 

 degree of certainty. Hyella caespitosa, the type of the genus as de- 

 scribed by Bornet and Flahault, has branched prostrate multicellular 

 filaments, presumabl}^ arising from single cells on the surface of the 

 host shell. The cells of these filaments, at least in part {loc. cit., 

 fig. 9), divide in all planes creating rows of cells within the original 

 cell wall or sheath, similar to the formation of Stigonema filaments. 

 From those prostrate filaments numerous erect filaments arise by 

 divisions of certain cells of the prostrate filaments, or basal filaments 

 as I shall call them. These erect filaments dissolve the shell and pene- 

 trate into it, and are very numerous. Gonidangia are formed near 

 the surface of the shell on short branches arising from the cells of 

 the basal filaments, and grow toward the interior of the shell or are 

 sessile on the basal filament. Figures 2 and 3 of Bornet {loc. cit.) 

 show that the filamentous nature of the basal filaments is nearly or 

 quite destroyed when the erect filaments are nearly mature. They 

 appear as separate groups of cells when viewed from above. Forti 

 (1907) recognizes four other species of Hyella, all but one boring 

 into calcareous material, Hyella endophytica B0rgesen (1902) pene- 

 trating into Chondrus crispiis. Recently Howe (1914) described 

 Hyella infestans inhabiting Leptocladia peruviana. The descriptions 

 of these five species fail to distinguish between the basal filaments and 

 the erect filaments, also nothing seems to have been learned about 

 the early stages of development. The superficial cells of H. fontana, 

 H. Balani and H. endophytioa may develop into gonidangia, in this 

 respect differing from H. caespitosa. Howe did not observe any 

 gonidangia in H. infestans and it is left indefinite as to which cells are 

 transformed into gonidangia in //. jurana. Howe places his species 

 in the genus Hyella on account of its close resemblance to H. endo- 

 phytica of B0rgesen. If basal filaments creeping at the surface of the 

 host are to be deemed essential to the genus Hyella, as first diagnosed, 

 it may yet be shown, when the early stages of development are made 

 known, that none of the species subsequently described belong to 

 that genus. 



Howe and B0rgesen are being followed in placing the species 

 described in this paper, although the gonidangia have not been 



