1917] Gardner: New Pacific Coast Marine Algae I 401 



one in forminfr the cortex remain short and divide several times, 

 nsually dichotomously, but not infrequently trichotomously, finally 

 ending- in very small cells in which tlie spermatia develop, one in each 

 terminal or antheridial cell (pi. 32, fig. 1). The anteridia are usually 

 distributed over the whole frond, except on the stipe-like portion. 



The carpogonial branches arise close to the growing point, usually 

 on the second cell from the joint-like cell. Each carpogonial branch 

 is one of the dichotomy which seems to have been suppressed in growth 

 while the cell which gave rise to it continues to grow. This in time 

 makes the carpogonium seem to have arisen laterally. Usually the 

 carpogonial branch becomes three-celled, the terminal cell becoming 

 the carpogonium (pi. 32, fig. 3). Plants have been observed with three 

 and even four sterile cells composing the carpogonial branch, but 

 even on these plants the two-celled branch predominates. The carpo- 

 gonium has a long, slender trichogyne which extends straight out 

 beyond the surface of the plant (pi. 32, fig. 3). 



The characteristic feature in the development of the cystocarp 

 is the direct origin of the gonimoblastic filaments from the fertilized 

 egg without its previous division several times as in Nemalion. Sev- 

 eral of these arise successiveh' from the carpogonium by a pushing 

 out of its cell wall at definite points (pi. 32, fig. 2). Tlie protoplasm 

 in these voung gonimoblasts is soon cut off from the mother cell bv 

 a cross-wall, after which they continue to grow out more or less 

 horizontally or obliquely among the cortical filaments, becoming more 

 or less contorted and frequently branched, tlie branches constantly 

 tending obli(iuely outward and terminating about one cell beneath 

 the surface of the phiiit. The end cells of the gonimoblasts become 

 the carpospores, which are considerably larger tlian the neigliboring 

 cortical cells though very similar in shape. There is no sterile tissue 

 surrounding the cystocarp and no carpostome, the spores escaping 

 from the point of their origin. The cystocarps are thus more or less 

 diffuse, irregular in outline, do not produce a protuberance and hence 

 are inconspicuous (pi. 32, fig. 4). They are quite numerous and are 

 developed both on tlie branches and on the main frond except near 

 the base. After fertilization the two or three cells of the branch 

 on which the carpogonium develops become much enlarged and the 

 protoplasm very dense, staining deeply with acid fuchsin. 



With this understanding of the structure of Cumagloia Ander- 

 sonii as revealed by this investigation, it now becomes necessarj' to 

 endeavor to jioint out its relationship to other members of the Rhodo- 



