1889.] BOTANICAL GAZETTE, 1 69 



views of previous experimenters. But "lacerata y" and •*#" 

 have not the slightest resemblance to R. cancellata. On the 

 contrary, as I pointed out in my previous paper, the Rocstelia 

 which I called " laceraia y" is closely allied to R. cornuta, 

 which has been referred bv Mr. Plowright, also in concur- 

 rence with previous opinion, to G. conicum. 



Turning for a moment to our birds'-nest Gymnosporan- 

 gium, the question of identities is still further confused by 

 my cultures of this species, made on three successive years 

 with identical results. The Roestelia obtained, as I have 

 mentioned above, was referred to R. cornuta, there being no 

 other alternative among described American Roestelias. It 

 seemed not unlikely that the differences between the culture 

 and R. cornuta might have been accidental ; but the subse- 

 quent cultures render this supposition very improbable. In 

 all cases we have the same rapidly-developing Rcestelia re- 

 sembling R. lacerata in its mode of growth. In nature the 

 form has probably been confused with R. lacerata, with 

 which it must occur simultaneously on Amelanchier. It 

 should be noted that upon Sorbus (the natural host of R. cor- 

 nuta) the most careful sowings of our " birds'-nest" Gym- 

 nosporangium have given no result whatever ; and also that, 

 although our species is certainly known to occur only on 

 Juniperus Virginiana, the European G. conicum is found 

 only on J. communis. Further careful cultures and observa- 

 tion, together with a more extended examination of Euro- 

 pean specimens than I have been able to make, will, of course, 

 be necessary to confirm this supposition ; but I am decidedly 

 of opinion that our " birds'-nest" species is a distinctly Amer- 

 ican form as yet unnamed. 



Unfortunately, however, we still have to account for what 

 appears to be the tvpical R. cornuta on Sorbus collected at 

 Eastport bv Prof. Farlow, and subsequently at Grand Me- 

 nan Island by Mr. Mivabe, and in the White Mountains 

 by Prof. Farlow and myself. In addition also to these, we 

 have unaccounted for a not very common form on Amelan- 

 chier and one on Pvrus arbutifolia found at Kittery, Maine— 

 both of which do not appear to differ essentially from R. cor- 

 nuta. Whether all these forms may not be referable to 

 "lacerata y," and consequents to G. globosum. I do not 

 feel at present prepared to say ; but the only remaining alter- 

 native would seem to be that the true G. conicum occurs in 

 this country on its proper host (J. communis), but has hith- 

 erto been overlooked. 



