p s t e I s i a 253 



large specimens which have not yet begun to 

 differentiate the sorus: one of these (Plate XVI., 

 Fig. 11) measures 23x6 centimeters with a 

 holdfast 13 miUimeters. There seems to be 

 scarcely any localization of the sorus into a 

 definite region of the lamina as in most kelps, 

 but almost any portion may develop gonidangia. 

 Sometimes the sorus occupies a small oval 

 patch with fairly definite margins, near the 

 base of the lamina; again it may cover almost 

 the whole lamina leaving only a narrow mar- 

 gin all around. But more often it is irregular 

 in shape especially at the distal end. Very 

 often instead of bearing a single fruiting area 

 the lamina has a series of parallel stripes of 

 fructiferous tissue separated by similar sterile 

 stripes; these may number as many as five. 

 Both sides of the lamina are fertile but there 

 is no close correspondence between the sori 

 of the two as may be seen from Plate XVI., Figs. 

 3,5,6. I have a single specimen (Plate XVI., 

 Fig. 6) in which there are two sori, one distal 

 to the other, as in Nereocystis. The lower is 

 a large oval area, 27x5 centimeters, then comes 



