1918] Gardner: New Pacific Coast Marine Algae II 435 



host it is continuous for several millimeters. This, however, may be 

 due to coalescence of several colonies. Later cell divisions in the 

 horizontal plane increase the mass in thickness, uj) to lOOft or more. 

 At first the cells are arranged in vertical rows, but soon this arrange- 

 ment is destroyed by false branching or by divisions in other planes. 

 The cells are mostly quadrate in this mass becoming somewhat spher- 

 ical in its outer portion. Sooner or later at various points from the 

 under side of the horizontal layer, certain cells are able to penetrate 

 through the cuticle and make their way among the cortical filaments 

 of the host. These penetrating filaments branch, and have a terminal 

 cell two or more times as long as the other cells. Growth in length 

 of these filaments is apical. Enlargement of the cells back of the 

 growing points and divisions in other planes soon produce groups of 

 cells which, encroaching on one another, form a solid mass, the cortical 

 cells of the host at times completely disappearing from the area which 

 they occupy. Plants penetrate to a depth of 200;u,. 



Chlorogloea lutea is very closely related in some ways, particularly 

 in its method of development within the host, to Hyella socialis found 

 growing with it, and also described in this paper. The method of 

 behavior of both within the host is the same. I have not, however, 

 seen H. socialis extending beyond the surface of the host as C. lutea 

 does. H. socialis thus appears to be wholly endophytic, although on 

 account of the absence of the gonidia, I consider the plants in this 

 collection to be immature, and I am unable to say positively what 

 might have developed later. 



The habitat and method of development of Hyella soci^alis are so 

 similar to those of Hyella endophytica B0rgesen (1902, p. 525), which 

 forms gonidangia near the surface of the host, that, notwithstanding 

 the absence of gonidangia, there can be but little doubt that it belongs 

 to the genus Hyella. On the other hand, the genus Chlorogloe-a never 

 has gonidangia and the species lutea assigned to the genus here, is so 

 assigned because of the absence of gonidangia and because of its 

 vesemhlanGe to C. tuherculma (Hansg.) Wille (1900), notwithstanding 

 the close similarity in its method of development to H. socialis with 

 which it is so intimatel}- associated. The size, color, shape, and early 

 arrangement of the cells in a colony very closely approximate to those 

 of Chlorogloea tuherculosa, as described by Wille, but that species is 

 wholly endophytic, and does not have radiating, branching filaments. 

 Furthermore WiUc states that cell division is in one plane, "nach 

 einer Richtung des Raumes" (1900, p. 4. pi. 1, figs. 4-6). This 



