5 SALMON: SUPPLEMENTARY 
Wandzelle zum Peritheciendurchmesser durch folgende approxi- 2 
mative Zahlen ausdriicken: 1:5 resp. 1: 18.’’ Now the size of . 
the cells in the species of Uncinula varies from 10 to 15 wide, 
reaching rarely to 20; so that it is perfectly correct to represent 
the perithecial wall of some forms of ‘‘.S. Castagnez”’ as being 
of similar structure to that of species of Uncinula. As regards the 
proportion between the number of cells and the diameter of the — 
perithecium also, S. hamuli type agrees with species of Uncinula. 7 
Neger, it may here be noted, has raised De Bary’s section “‘ 771- 7 
chocladia’ of Erysiphe to generic rank, and places in the genus | 
Microsphaera astragali and Erysiphe tortilis, 1 consider this 
arrangement undesirable for the following reasons: The two 
genera Microsphacra and Erysiphe are undoubtedly very closely 
allied, and it seems to me to be more in accordance with the 
principles of a natural classification to include in these genera 
the species which on both sides approach each other, than to take 
these out and base a new genus on them. Further, if we follow 
Neger and place WW. astragali in the new genus 77ichocladia, it is 
clearly unnatural for 12 Baumleri, so closely allied, to be gener- 
ically separated, and we should find it impossible to draw a satis- 
factory line between “ 7richocladia” (including M. Baumleri) and 
Microsphaera. It still seems to me to be better to include JZ as- 
tragali, which although certainly approaching some forms of £. 
polygoni, yet on the other hand is closely allied to Microsphaera 
Baumleri and M. euonymi, in the genus Microsphaera, and to keep 
£. tortilis in Erysiphe on account of its relationship with certain 
forms of £. polygont. 
Neger (52, p. 337) referring to the fact that I included in my 
monograph under the name Frysiphe polygoni all plants which 
showed certain morphological characters, notwithstanding that 
they grew on the most diverse host-plants—no fewer than 190 
different species of plants (belonging to 89 genera) being given as 
hosts, observes: ‘ Freilich, ob diese Anschauung berechtigt ist, 
ist bis heute durch nichts bewiesen. Es liegen namlich nur ganz 
vereinzelte auf culturversuche begriindete Bestatigungen dieser 
gewohnlich in extenso angenommenen Voraussetzung vor, was 
um so mehr auffallen muss, als doch bei anderen Pilzfamilien, z- 
B. den Uredineen, trotz der dort bestehenden grésseren Schwie- 
