54 ASCOMYCETES [CH. n 



here fairly elaborate; the asci are regularly arranged. The Erysiphaceae 

 may possibly have given rise to the Laboulbeniales, the only other group in 

 which a single daughter cell of the oogonium is responsible for the asci, and 

 perhaps to the lower Pyrenomycetes also; these, like the Erysiphaceae, have 

 regularly arranged asci, and in Chaetomium fimete the perithecium is without 

 an ostiole. 



(2) Gymnoasciis and its allies; thegametangiaare not much differentiated; 

 both, like the mycelial cells, are multinucleate (it is difficult to gauge the 

 significance of this character in Fungi) and the oogonium, or an outgrowth 

 from it, becomes septate after fertilization and gives rise to ascogenous 

 hyphae. Among the Endomycetaceae, Dipodascus suggests itself as the 

 most probable representative of the ancestor of these forms. The asci in the 

 Gymnoascaceae are irregularly arranged and the sheath is rudimentary. 

 Clearly such a form may have given rise to the Aspergillaceae and to the 

 rest of the higher Plectascales but perhaps to no other group. 



(3) Ascodesmis is a third type which might be derived either directly 

 or through the erysiphaceous type from an endomycetous ancestor ; the 

 antheridium and oogonium are but little differentiated, but the latter is 

 furnished with a trichogyne and becomes septate after fertilization; the 

 ascogenous hyphae are few and the sheath simple. Massee indeed places 

 this genus near Gytnnoascus although the asci are regularly arranged. It 

 cannot be far removed from the ancestor of the other Discomycetes. 



The table on page 52 may perhaps serve to elucidate some of these 

 hypotheses. 



