1907] CHRISTMAN— MORPHOLOGY OF THE RUSTS 97 



derived type. With such a first crop of temporary spores once 

 formed, it is easy to see how they might be formed in successive 

 crops — the secondary uredospores — before the teleutospores are 

 finally formed and the sporophytic generation closed. 



A modification of the stalk cell with loss of essential function would 

 result in the formation of the true aecidium. The short intercalary 

 cells of the aecidium give us a decided argument for the belief that the 

 aecidiospores are not the primitive spore type, since I am of the 

 opinion that we can only regard the intercalary cell as practically 

 a functionless cell. The argument that the intercalary cell is for 

 the purpose of disjoining the spores is hardly tenable, since we have 

 many fungi in w^hich, as in Cystopus, chains of conidia are formed, the 

 spores of which separate readily without alternating sterile cells. 

 Looked upon as a rudimentary stalk, we can hardly assume that it is 

 in process of progressive evolution, since from that point of view w^e 

 would have the curious case of a functionless structure being first 

 developed, to be modified later into a structure of some service to the 

 plant. It is relatively common, on the other hand, to find that func- 

 tional structures degenerate on becoming useless, becoming smaller 

 and finally disappearing. The degeneration of the uredo stalk cell is 

 doubtless correlated with the development of the spore pustule, result- 

 ing in the formation of the spores in chains instead of in a horizontal 

 series. The spore-initial cells, which have in other cases divided to 

 produce spores and stalk cells, still divide, forming the spore and the 

 small intercalary cell, the degenerated representative of the former 

 stalk cell. With the change from the stalked primary uredospores to 

 the chains of aecidiospores, we have formed the typical autoccious 

 eu- type of rust. If now^ the sporidia find a more favorable host in 

 some other plant than that bearing the uredospores and teleuto 

 spores, a later change of host for the sporophyte generation would 

 explain heteroecism. It is certainly conceivable that the change of 

 host should occur at the germination of the first-formed binucleated 

 spore. Certainly the sporidia and the gamctophytic mycclia produced 

 by them are sufficiently different from the sporophytic aecidiospores 

 and uredospores and the mycelia which they produce to explain such 



a change in some measure. 



From either point of view it w^ould appear that a heteroecious 



