444 MR. G. MASSEE ON THE 
swollen fertile cells forming the characteristic basidium. This 
complete transition connecting the extreme poles—conidiophore 
and basidium —proves that Juel's suggestion to separate those 
species presenting the typical basidium from the remainder still 
having the conidiophore form of basidium, cannot be followed, 
and further proves that the suggestion resulted from an examina- 
tion of only a limited number of species. 
As a rule those species of Stilbum, &c., known to be the coni- 
dial form of ascigerous fungi have the most primitive basidia, that 
is basidia of the conidiophore type; whereas the independent 
species more frequently have typical Protobasidiomycete basidia. 
To this rule, however, there are marked exceptions, as already 
shown in the ease of the conidial condition of Spherostilbe 
microspora, where the basidia are short-celled and clavate. 
The above discovery further indicates that the Protobasidio- 
mycetes as a group are derived from the conidial phase in the 
life-cycle of ascigerous fungi; the evolution is effected by the 
disappearance of the ascigerous form of reproduction, whereby 
the conidial stage assumes the standard of a species: this change 
being contemporaneous with the gradual conversion of the so- 
called conidiophore to the typical basidium or spore-bearing 
organ. 
The earlier realization of this fact was probably retarded to a 
certain extent by the Friesian conception of a Basidiomycete, 
whieh required above all things the presence of a compact, con- 
tinuous hymenial surface. This idea held good until corrected 
by the researches of De Bary, Brefeld, and Miller. 
As already stated, Brefeld supposes the Autobasidiomycetes 
to be descended from the Protobasidiomycetes through the dis- 
appearance of the transverse septa in the basidium, and the gradual 
concentration of the spores in a definite number at its apex. 
To those who can accept an unproved assumption, the basidium 
of Tulostoma, a cylindrieal organ without septa, and having three 
or four spores scattered at intervals throughout its length, looks 
very much like the basidium of a Protobasidiomycete in which 
the transverse septa have been arrested; but unfortunately we 
possess no evidence of the gradual disappearance of septa in the 
basidia of any known Protobasidiomycete, whereas, on the other 
hand, it can be demonstrated that the basidia of Tulostoma agree 
in all essentials with so-called conidiophores that pass directly 
into typical Autobasidiomycete basidia. 
