SO • BOTANICAL GAZETTE [July 



must have been derived from the pedicellate, and the primary uredo 

 is therefore more primitive than the aecidium. For, as he argues, 

 the intercalary cell is thus seen to be in reality a stalk cell that has 

 been necessarily shortened by the catenulate method of spore- 

 production and persists as a vestigial cell. If the aecidium, how- 

 ever, is considered the primitive sorus and the uredo the derived 

 form, the intercalary cell must be considered a functionless" structure 

 that persists and is later modified into a structure of value to the 

 organism in the stalk of the uredospore. 



In view of this convincing explanation of the nature of the inter- 

 calary cell, one can scarcely agree with Grove (13) that Endo. 

 Sempervivi is to be considered as a representative of the primitive 

 type of rusts. The fact that the spores of this species are borne in 

 a complex aecidium-like fructification, surrounded by a peridium 

 and accompanied by intercalary cells, makes its acceptance as a 

 primitive type extremely difficult. To all morphological appear- 

 ances the spores are aecidiospores and the sorus an aecidium. The 

 spores, however, function as teleutospores, as Hoffman has con- 



m * 



clusively shown. Grove's acceptance of Endo phyllum as a primi- 

 tive type and his relegation of the micro-forms to the position of 

 reduced types seem inconsistent when the relative complexity ot 

 the two types of sori are considered. It certainly seems more 

 logical to consider Endo phyllum a reduced form, as the gametophy- 

 tic generation of a former eu-heteroecious or autoecious species that 

 has dropped the uredo and teleuto stages, with the assumption of a 

 teleutosporic method of germination by the aecidiospore, and 

 to regard the micro-forms as the more primitive and ancestral. 

 The assumption of the teleutosporic method of germination by 

 an aecidiospore is not a difficult conception, since nothing more 

 than the fusion of nuclei in the spore should be necessary for its 

 accomplishment. 



It is interesting to note that Ki nkel (17) has recently discov- 

 ered a companion form to Endo. Sempervivi in the common orange 

 rust of the blackberry, Caeotna nit ens. The life history of the two 

 species is practically identical, seemingly, but the sorus of the latter 

 is a caeoma, while that of the former is an aecidium cup. C aroma 

 miens, therefore, should offer a much better ancestral type for 



