UREDINEAE AND HYMENOMYCETES 531 



i.e. at the outer end of the basidium : but here only one sterigma 

 arises terminally whilst the other three arise laterally, one from each 

 of the three lower cells. In order to convert the basidium of an 

 Auricularia or of a Coleosporium into a basidium of the Agaricus 

 type, all that would be necessary would be the removal of the septa 

 and the sliding upwards of the place of origin of the sterigmata on to 

 the top of the basidium. The essential arrangements of the basidium 

 of an Agaricus, of an Auricularia, and of a Coleosporium for the 

 shooting away of the spores, therefore, appear to be identical and 

 stand in contrast to the essential arrangement of the basidium of a 

 Puccinia. The basidium with a straight axis and with the mature 

 spores occupying a position at the free end of the basidium is evi- 

 dently the only suitable kind of basidium for a compact hymenium. 

 On the other hand, the Puccinia type of basidium is excellently 

 adapted for spore-discharge under conditions in which the formation 

 of a compact hymenium is rendered impossible owing to the fact 

 that the basidia must develop independently of one another from 

 teleutospores packed irregularly. 



Parasitism has been responsible for the origin of the teleuto- 

 spore, the teleutospore for the isolation of the basidia and the non- 

 production of a hymenium, and the isolation of the basidia for the 

 curvature of the basidial shafts, curved shafts with sterigmata on 

 the convex sides being for the discharge of the spores under the 

 conditions provided for their development as efficient as, or possibly 

 more efficient than, straight shafts with terminal sterigmata. 



I regard the Agaricus type of basidium as having been evolved 

 along with the formation of a compact hymenium, and do not think 

 this type primitive. It is difficult to imagine how the Basidio- 

 mycetes were evolved and I shall therefore venture to speculate 

 upon some of the steps in their phylogenetic development as follows. 

 The first basidia of the primitive ancestral Basidiomycetes, prob- 

 ably, were produced on the mycelium as isolated structures, were 

 transversely septate, and were provided with lateral sterigmata, so 

 that at first each basidium was somewhat like the conidiophore of 

 a Nectria, although, of course, from the nuclear point of view, 

 profoundly different. We must assume that violent discharge of 

 the basidiospores took place from the first. The most primitive 



