198 Phillips. — The Development of the Cystocarp 
absent in the genus Delesseria. Brannon (’ 97 ) found moreover 
that no such fusion occurs in Grinnellia. 
7. The pericentral cell is described as playing the part of 
the auxiliary. I have already given reasons for believing that 
the auxiliary is an anterior cell cut off from the pericentral 
cell. An auxiliary cell so derived occurs in most Rhodome- 
laceae and in all Ceramiaceae. 
An interesting feature in the development of the cystocarp 
in the species here under consideration is the fact that, when 
the cells adjacent to the central cell in a fertilized procarp 
become charged with nutriment prior to the formation of the 
gonimoblast-filaments, they also become multinucleate, as 
many as eight or ten nuclei at times occurring in one cell. 
I have found that elsewhere in the thallus of D. sangninea the 
greatly elongated axial cells contain more than one nucleus. 
With regard, finally, to the systematic position of the 
Delesseriaceae, I have no hesitation, on the ground of the 
remarkable correspondence in the process of development of 
the cystocarp, in placing them close to the Rhodomelaceae. 
There is the same invariably 4-celled carpogonial branch, 
the auxiliary cell is derived anteriorly from the pericentral 
cell, and there are always found, in the early cystocarp, two 
sterile filaments which degenerate later into mucilage. Indeed, 
the observation of the way in which these sterile filaments 
arise in Delesseriaceae affords a clue as to their origin in 
Rhodomelaceae which would otherwise be wanting. In 
Rhodomelaceae they would seem to be vestigial structures, 
and the cylindrical Rhodomelaceae would seem to have been 
derived from forms with a flattened thallus like Delesseriaceae. 
The two families form one alliance ; the simplest forms being 
represented by Nitophyllum , and the most complex by the 
polysiphonous Rhodomelaceae. 
