446 BOTANICAL GAZETTE [june 



with the third cell. In many cases it lies very close to the second 

 cell also. Thus the structure of the carpogonial branch suggests 

 that the fusion nucleus passes from the carpogonium into the sec- 

 ond or third cell. This evidently does occur, although satisfactory 

 stages showing the process have not been found. Such figures as 

 42 and 43 show that the sporogenous filaments originate from either 

 the second or third cells of the carpogonial branch. Since the 

 actual passage of the fusion nucleus into the cell producing the 

 sporogenous filaments has not been observed, there will naturally 

 arise a question concerning the origin of the nuclei in these filaments. 

 It cannot positively be stated that the nuclei in the sporogenous 

 filaments are descended from the fusion nucleus of the carpogonium, 

 but most of the evidence leads to this conclusion. Hundreds of car- 

 pogonial branches which have not been fertilized have been exam- 

 ined, and in only two or three cases is there any evidence that the 

 third cell is binucleate. The second cell has never been observed 

 to contain more than one nucleus. Oltmanns (9) states that in 

 Dudresnaya coccinea the cell of the carpogonial branch with which 

 the sporogenous filament fuses is often binucleate, but that these 

 nuclei never move out into the sporogenous filaments. Spermatia 

 are found fused to trichogynes projecting beyond the surface of the 

 thallus. Sporogenous filaments are found arising from cells of 

 carpogonial branches* whose trichogynes probably had projected 

 beyond the surface of the thallus (fig. 42). The cells which produce 

 the sporogenous filaments are those which in other carpogonial 

 branches are always close to or in contact with the carpogonium. 

 Considering these facts it seems highly probable in Dumontia 

 filiformis, as in Dudresnaya purpurifera and D. coccinea, that the 

 nuclei in the sporogenous filaments are derived from the fusion 

 nucleus in the carpogonium. All the cells in the carpogonial 

 branches stain very faintly at the time of the formation of the 

 sporogenous filaments. The cytoplasm in all the cells, particularly 

 the terminal ones, becomes very thin (fig. 42) and in some cases 

 practically nothing but the cell walls is visible. The cytoplasm 

 in these cells is disintegrating, but not in the same manner as it 

 does in the trichogynes and some of the other cells. The failure 

 to find the carpogonium may be due to the fact that it disintegrates 



