48 University of California Puhiixxitions in Botany [Vol.8 



cells soon become more or less freed and independent of each other. 

 Horizontal divisions begin and short filaments, three to five cells long, 

 are foraied above the surface of the host. Meanwhile it seems povssible 

 that the prostrate layer has begun to dissolve the cuticle and the under- 

 lying cortical cells, and division of the basal cells may have begun 

 forming filaments which push into the host, completely dissolving the 

 host cells as they penetrate (pi. 3, fig. 12). 



The plants upon which this species is founded are scarcely mature, 

 the gonidangia being \ery rare. The depth to which they may pene- 

 trate the host is thus far rather uncertain. It is not at all unlikely 

 that they may penetrate to a greater depth than that reported in the 

 diagnosis above. It certainly seems that the whole of the cells of the 

 host is actually absorbed, or at least destroyed, as far as the parasite 

 travels. 



Radaisia suhimmersa may be looked upon as being on the border 

 line between Radaisia and Hyella so far as its relation to the host is 

 concerned, being partially internal and partiall}^ external, and pos- 

 sibly growing in both directions from the original basal layer. The 

 filaments, however, being more or less parallel, and having the goni- 

 dangia on the outer free ends, are characters which have led us to place 

 it in the genus Radaisia. 



3. Radaisia clavata S. and Q. 



Plate 3, figs. 17, 18 



Plants forming microscopic, deep blue-green cushions on the surface 

 of the host, up to 100/a diam., more or less fan-shaped in median sec- 

 tion; filaments very closely compact, 70-100/a long, sparsely branched 

 near the outer ends ; cell walls 4-5/a diam. at the base, 7-8/x above, 

 3-4/i, long, cell divisions often irregularly^ oblique, cell walls thin, 

 hj^aline; protoplast homogeneous; gonidangia 8-9/x diam., terminal, 

 hemispherical ; gonidia angular, 1-1. 5/x diam., formed by simultaneous 

 division. 



Growing on Gymnog&tigrus linearis, in the lower littoral belt. 

 Lands End, San Francisco, California. This is the only known locality 

 in which this plant grows. The host plant is common along the Cali- 

 fornia coast and extends as far north as the Straits of Juan de Fuca. 

 It is not at all unlikely that it may be found on the same host in 

 other localities. 



Setchell and Gardner, in Gardner, New Pac. Coast Alg. II, 1918, 

 p. 445, pi. 37, figs. 17, 18. 



