1914] FROMME—AECIDIUM CUP 27 



until a complete layer is formed over the entire mass. The subse- 

 quent enlargement of the peridium is accomplished by the basipetal 

 growth of the lateral chains of peridium initial cells. 



The "buffer" cells of the caeoma are homologous with the 

 pseudoparenchyma cells of the cup, and represent a scanty pro- 

 duction of pseudoparenchyma. The production of a peridium 

 seems to be dependent on the presence of a considerable pseudo- 

 parenchyma, and this is in turn the result of the deep location of 

 the primordium in the host tissue, and the deep location of the 

 gametes in the primordium. 



As I have elsewhere stated (12), in the caeoma of Mel. Lini the 

 gametophoric hyphae are 3 or 4 cells long. Two sterile cells are 

 normally produced, and though, as a rule, but one gamete is formed 

 in each hypha, two are occasionally found. In the production of 

 more than one sterile cell and the tendency to produce more than 

 one gamete there is seemingly a transition toward the conditions 

 found in the cup. The cup of Ur. Caladii, as found on Arisaema, 

 is very much like a caeoma in appearance before the formation of 

 the peridium. The gametophoric hyphae are 6 or 7 cells in length, 

 and 4 or 5 of the apical cells are sterilized. If, as I have suggested, 

 the production of a peridium is correlated with the amount of 

 pseudoparenchyma produced, the line of separation between the 

 aecidia without and those with a peridium should be found at some 

 point between the caeoma of Mel. Lini with 2 sterile cells, and the 

 cup of Ur. Caladii with 4 or 5. A careful study of the more deeply 

 seated caeomas may perhaps reveal a more or less marked differen- 

 tiation of the outermost spore layer. Those species of cupulate 

 aecidia with evanescent peridia are perhaps most nearly like the 

 deep caeomas. 



The study of the origin of the peridium that is found in 

 the uredosorus of certain genera should prove of considerable 

 interest, but apparently no such study has ever been made 

 and no data are available. According to Arthur's treatment 

 of the Uredinales in the North American flora (2), 6 genera of the 

 family Uredinaceae have uredosori with peridia: Piicciniastrum y 

 Melampsoridiiim, Hyalopsora, Uredinopsis, Melampsoropsis, and 

 Cronartium. Melampsoropsis has catenulate uredospores, while 



