83 



reaches a length of about 1214^. The carpogonium is conical 

 and passes evenly into the long trichogyne. 



After the fertilization (comp. Fig. 90) the sporogenous fila- 

 ments grow out. These are repeatedly branched and form to- 

 gether a spherical rather compact body. In the summit of the 

 filaments the carpospores are developed. Apparently even before 

 fertilization, or in any case before any visible cell division has 

 taken place in the carpogonium, filaments begin to grow out 

 from the cell underneath that upon which the carpogonial branch 

 is placed and these filaments are often considerably developed 

 before the division of the carpogonium has begun (Fig. 90 a, b). 

 In the more developed carpogonium they grow longer being 

 several times forked and bent inward and they more or less 

 surround the young cystocarp (Fig. 90 c); 

 in the mature cystocarp on the other hand 

 they are present like a whirl of prominent 

 filaments at its base, and have the appear- 

 ance of a collar thus giving the cystocarp a 

 very characteristic aspect (Fig. 90 d). 



The antheridia are developed in the ends 

 of the assimilating filaments as quite small 

 oblong cells; they are 2 3// thick. In the 

 material at hand they were only present in 

 very small numbers. I have not found an- 

 theridia and cystocarps together in the same 

 plant. 



The above description is based upon material in spirit sent 

 to me by Mr. 0. HANSEN GANNESKOV and collected at Rust up 

 Twist on the north side of St. Croix. 



Other specimens which I also refer to this species show 

 a few differences. They were gathered in Lime Tree Bay on the 

 south side of the same island where I found them growing epi- 

 phytic upon Udotea flabellata in shallow water. 



Fig. 91 shows that the plant was a little more robust than 

 the first described and the proliferations more regularly forked. 

 The assimilating filaments (Fig. 92 6, c) had nearly the same size and 

 shape, but the uppermost cells in the filaments were more sphe- 

 rical often nearly globular. The few carpogonial branches (Fig. 

 92 d) found were curved in the same way, but they contained 

 only two cells besides the carpogonium which was very long. 

 Neither more developed carpogonia nor cystocarps were present 



in the plant. 



6* 



Fig. 91. Liagora pul- 



verulenta C. Ag. Part 



of a plant. (About 



3:1). 



