578 University of California Public at ions in Botany [Vol. 8 



No species of CoUodesme has as yet been seen with gametangia and 

 tliis suggests the extreme possibility of a microscopic gametophyte. 

 This combination of characters leads us to suggest the new family 

 Coilodesmaceae of the order Chordariales. 



Key to the Genera 



1. Fronds solid, ligulate 38. Phaeostrophion (p. 585) 



1. Fronds hollow, eomplanate or saccate 37. Coilodesme (p. 578) 



37. Coilodesme Stroem. 



Fronds saccate, eomplanate from the beginning or more or less 

 flattened in age, linear to ovate or obovate in outline, rounded above, 

 tapering more or less at the base into a longer or shorter solid stipe, 

 with small disk-shaped holdfast or with penetrating rhizoids epiphytic, 

 parasitic or saxicolous, composed of two tissues, the inner consisting of 

 large, irregular, thick-walled cells, with few chromatophores, the 

 outer cortical portion consisting of a few to many layers of smaller 

 color-bearing assimilating cells arranged more or less in radial rows; 

 hairs absent ; unilocular zoosporangia developed within the cortex, 

 scattered over the whole surface of the frond except the stipe and 

 extreme basal portion ; paraphyses absent ; gametangia unknown. 



Stroemfelt, Meeresalg. Isl., 1886a, p. 173. 



The type species of the genus is C. bidlhgera, of Icelandic origin. 



The species of CoUodesme seem to divide themselves fairly defi- 

 nitely into two groups, the one saxicolous, the other epiphytic. Of 

 the epiphytic plants in the collections available to us, all are from the 

 western coast of North America or the northeastern coast of Asia and 

 seem to differ from one another essentially according to their host 

 plant, only one host plant having more than one species associated 

 with it. This is not surprising, since Rathbone (Journ. Linn. Soc, 

 1904, p. 671, pi. 24, figs. 5, 6) has shown the actual penetration of the 

 host and our own investigations have demonstrated an intimate meta- 

 bolic relation. The parasitism of all the epiphytic species of our coast 

 is also striking in that the hosts are all members of the related genera, 

 Halidrys, Cystoseira, and CystopkyUum, each of which seems to have 

 one or more species of Coilodesme peculiar to it. Okamura (Icones 

 Jap. Algae, vol. IV, no. 3, 1918, p. 55, pi. 144, figs. 10-13) figures 

 and describes a Coilodesme Cystoseirae growing on Cystophyllum 

 hakorfntense at Akkeslie, Japan. From the data presented, we are 

 inclined to question the reference of this plant to C. Cystoseirae as 

 interpreted by us, but careful comparison is necessary to determine 



