88 BOTANICAL GAZETTE [august 



the caeomas just before the conjugation. From the closely built 

 structure of the aecidium cup it would not necessarily follow that 

 the structure is developed from a single organ^ since we have uredo- 

 spore pustules in the Coleosporiums, which, without having a perid- 

 ium, are quite as compact as are the aecidium cups of such a form 

 as Uromyces caladiij and no sexual process can be assumed to have 

 occurred in their origin. 



As a foundation for a correct morphological interpretation of 

 these various spore forms, I have undertaken a careful study of 

 the development of those spore types which do not originate in a 

 fusion cell. 



UREDOSPORES 



The secondary uredospores of Phragmidium potentillae canadensis 

 are especially interesting because of their general resemblance to the 

 primary uredospores which arise from a fusion cell as I have described 



r 



(6). The secondary uredospores also arise from a large basal cell 

 which contains two large nuclei. The basal cell here is borne upon 

 a mycelium of binucleated cells {fig. i). No fusion occurs in its 

 formation, but it is plainly the equivalent, in its mature condition, of 

 the so-called ^'basidium^' of the aecidium cup which arises from a 

 fusion cell. The two nuclei of the basal cell divide by conjugate 

 division, and the pair of daughter nuclei lying in the distal portion of 

 the basal cell are separated off by a cross wall, thus forming the first 

 spore-initial cell. A further conjugate nuclear division, followed by 

 a cell division, at once separates this initial cell into a spore and a 

 smaller cell below, which elongates to form a stalk upon which the 

 spore is borne (fig. 2). About the time the first spore is formed, a 

 bud appears on the basal cell beside the stalk of the first uredospore 

 (fig. 3). A second simultaneous division of the nuclei of the basal 

 cell provides this bud with two nuclei, leaving two in the basal cell. 



A wall now separates off this second uredospore-initial cell (fig. 4). 

 It in turn divides, forming a second spore and stalk in the manner 

 described for the first (fig. 5). In the same manner a third spore is 



X 



produced (fig. 21). 



On comparing these figures with the ones given in my descrip- 

 tion of spore-formation in the primary uredo (6) of this same rust, 

 the striking likeness is at once apparent. The only difference lies 



