650 THE PHYLOGENY OF THE FUNGI 



in the case of Kunkelia mentioned above the secondary and tertiary spores 

 are similarly discharged and in the Tilletiaceae this seems to extend to 

 secondary sporidia and even to conidia in some species. So we might 

 imagine that this might have been a habit acquired for asexual spores 

 that has been extended to the sporidium and basidiospore as well. 



If the holobasidium was a primary step in the evolution from the 

 Ascocoriicium-like Ascomyceteae we can assume that in some cases the 

 normally slender sterigmata as in Corticium became stouter, bearing at 

 their tips the points on which the basidiospores perch and finally attaining 

 the separable condition found in Tulasnella. The basidium itself may be 

 considered to have formed vertical walls as in Tremella and finally trans- 

 versely septate walls would bring about a structure like the basidium of 

 Auricularia. From Ceratohasidium which has stout sterigmata, only two 

 in number in some species, it is an easy step to Dacrymyces. 



From the simple, resupinate forms like Corticium, Tomentella, etc. it 

 is easy to make surmises as to the modifications which increase the hy- 

 menial surface so as to derive the Hydnaceae, Clavariaceae, Polyporaceae, 

 and eventually Agaricaceae. 



In the immediately preceding paragraphs we have left the Uredinales 

 hanging without a connection. They cannot be derived from any Auricu- 

 lariales such as we now know, for well organized spermogonia and recep- 

 tive hyphae are lacking in the latter. We can only guess that the trans- 

 formation from ascus to holobasidium occurred deeper down in the 

 Ascomyceteae than the present Ascocorticium and that the development 

 of holobasidium to phragmobasidium took place perhaps at several levels 

 or when once accomplished proceeded in various directions. The earlier 

 Uredinales must have come from the newly developed holobasidiomycete 

 while the latter still retained the primitive characters of its ancestors, 

 including spermogonia, receptive cells, diplobiontic life cycle, croziers or 

 clamp connections, and probably well pronounced parasitism. We must 

 assume that at an early stage of development along this line thick-walled 

 resting probasidia began to develop. This would be the beginning of 

 teliospore formation. On resuming their growth these probasidia would 

 grow out as thin-walled promycelia except in such cases as Coleosporium 

 where the teliospore wall was thin enough to permit of stretching as a 

 whole, thus producing a so-called "internal promycelium" instead of the 

 usual external one characteristic of the order. We need not homologize the 

 separate cells of the promycelium whether external or internal with 

 ascospores but consider these cells as parts of a phragmobasidium. Under 

 this assumption the sporidia may be considered as ascospores in external 

 pockets not as secondary spores. It must be admitted that this whole 

 suggestion stands on shaky foundations and the whole matter needs 

 further study. (Fig. 209.) 



