19 1 7l DUNN—DUMONTIA 443 



it is evident that this is actually its course. The failure to obtain 

 a satisfactory section was due to the varying and indirect course 

 of the trichogyne. Though most of the sections examined were 

 12 /x thick, the trichogyne nearly always passed out of the section 

 and it was very difficult to locate it in the adjoining sections. The 

 trichogyne, just beyond its point of attachment to the carpogo- 

 nium, is often much coiled (figs. 34, 37, 40). The trichogyne is 

 always surrounded by a fairly thick gelatinous wall which is a 

 continuation of that of the carpogonium (figs. 34, 37). The 

 granular cytoplasmic content of the trichogyne stains with the 

 same intensity as does that of the terminal cells of the carpogoniai 

 branch. No structure was seen in any trichogyne which could 

 positively be identified as a nucleus. In a few cases a body was 

 seen which appeared to be similar to a nucleolus (fig. 39). This 

 body is surrounded by a light area, but not by a definite membrane, 

 and is therefore not thought to be a nucleus. There are present in 

 some of the trichogynes (fig. 37) 2 or 3 masses which, with hema- 

 toxylin, stain like chromatin. The question of the presence of a 

 nucleus in the trichogyne of the Florideae is still unsettled. 

 Svedelius reports that the trichogyne nucleus in Delesseria san- 

 guined disintegrates before fertilization and the chromatin granules 

 pass out into the cytoplasm. It is possible that some of the granules 

 seen in Dumontia and other forms are chromatin granules of similar 

 origin. 



There are two types of mature carpogoniai branches. A cell 

 is sometimes formed as a lateral outgrowth of the basal cell of the 

 carpogoniai branch (fig. 35). The cell thus formed is a super- 

 numerary cell and will not be numbered, as it is not always present 

 and has no special function. This supernumerary cell has 1 

 been observed in a carpogoniai branch which is not mature, 

 basal cell of the carpogoniai branch is often found to be binucleate 

 (figs. 35, 38, 40) and sometimes contains as many as 3 nuclei (fig. 

 36). The basal cell is sometimes binucleate after having cut off 

 the supernumerary cell (fig. 35). It thus appears that there is a 

 tendencv of the basal cell to form a lateral branch. Also the third 



The 



some of the carpo 



(figs- 33, 38). It is difficult to determine whether the nucleus in 



