290 S a 1 m o n , On Specialization of Parasitisixi in the Erysiphaceae. 



toucli T. repens. In the next experiment (Exper. no. 37) the Oidium 

 infected two pots of seedlings of T. pratense, but failed to touch 

 VOUng" seedling'S of T. repens, Lupinus luteus, and Pisum sativum. 

 In the last experiment with this Oidium (Exper. no. 58), the fimgus 

 was sown on young seedlings of T. prafensc, T. repens, T. incar- 

 naium, T. hybridum, T. filiforme, T. medium, and T. montanum, 

 with the result that the three pots of T. pratense were infected, 

 while all the other species remained untouched. 



It seems clear from these experiments, (which are summarized 

 at Table 13), that the Oidium on Trifolium pratense is a ,biologic 

 form' specially adapted to this species, and incapable of growiiig not 

 only on species belonging to other genera of Leguminosae, but also 

 on numerous species, — if not on all other species — within the 

 genus Trifolium itself. Observations made in the field certainly 

 gave strong confirmative evidence that this Oidium on T. pratense 

 is incapable of touching T. repefis. An oat - field sown with T. 

 pratense and T. repens occuri'ed near Cambridge. During the late 

 Summer months the seedling plants of T. pratense became thickly 

 covered — in patches here and there over the whole field — with 

 tlie Oidium, while the plants of T. repens growing in about equal 

 quantity and often closely intermixed remained perfectly free. This 

 State of things contined throughout the whole season, terminating 

 with the production of perithecia on T. pratense. 



Neger, it must be noted, who experimented with the Oidium 

 on T. incarnatum, states positively (4, p. 254) that this Oidium 

 cannot infect T. repens. We have then the existence clearly proved 

 of two ,biologic forms' within the genus Trifolium, and as E. Poly- 

 goni has been recorded (1) (2) as well on T. agrarium, T. alpestre, 

 2\ arvense, T. filiforme, T. hyhridum, T. involacratum, T. longipes, 

 T. Lupinaster, T. medium, T. minus, T. monatithutn, T. montanum, 

 T. moranfhum, T. paucißorum, T. procumhens, T. repens, T. rubens 

 and T. spadiceum it seems very probable that with respect to the 

 present fungus on the species of this genus there exists as complete 

 a specialization of parasitism as the experiments given above have 

 proved to be the case with E. Graminis on Bromus. 



A Single experiment (Exper. no. 43) was made with the 

 Oidium on Pisum sativum. This was found to infect fully P. 

 arvense, but failed to touch Lupinus luteus, Colutea arborescens, 

 Onobrychis satica, and Trifolium pratense. (See Table 14). As the 

 Oidium on T. pratense proved likewise imable to infect Pisum 

 sativum, the present Oidium on Pisum sativum must rank as an- 

 other ,biologic form.' 



Some experiments were made on the subject of the production 

 of ascospores in Erysiphe Graminis. As is well known, the asci 

 of this species contain as a rule, at the end of the growing season, 

 in apparently ripe perithecia on the living host-plant, only a mass 

 of granulär protoplasm with no trace of ascospores. Wolff's ex- 

 periment of inducing the production of ascospores by completely 

 immersing the perithecia in water was repeated; at the same time 

 a few results differing from those obtained by this author were 

 noted. Wolff (10, p. 34) states; , Lassen wir dami die Peri- 

 thecien der E. Grami?iis, in deren Ascis jetzt ebensowenig wie zui' 



