1889. J BOTANICAL GAZETTE. I 7 I 



satisfactorily to identify. At Cullowhee two or three small 

 cedars badly infested by G. macropus and G. globosum, 

 served to infect the region within a distance of a mile or more, 

 and during June and July the Crataegus and Pyrus coronaria, 

 common in the vicinity, were covered with spermogonia. 

 On leaving this locality (July 18) the spots on Pyrus were 

 developing a luxuriant growth of Roestelia pyrata, while 

 those on Crataegus showed no signs of secidia. This Cratae- 

 gus form was, however, collected and forwarded to me during 

 the following August by my friend Mr. T. B. Cox, and 

 proved to be "lacerata y" ; in other words, the secidium of 

 G. globosum, as was to be expected. Spermogonia were also 

 seen on Amelanchier at the falls of the Tuckaseegee river, 

 between Cullowhee and Highlands ; but no aecidia were pro- 

 curable. The virulence of R. pyrata on certain varieties of 

 cultivated apples in this region was remarkable. I noticed 

 several instances on the road between Sylva and Asheville, 

 where cedars had been planted near small orchards, some of 

 the trees in which were so infested by R. pyrata that the 

 bright color was striking at even a considerable distance, 

 while side by side with these were individual trees which 

 showed no sign of the fungus. 



In the mountains of East Tennessee, in other respects a 

 mycologist's paradise, no signs of any Koestelias were to be 

 found, although shortly after entering the mountains, I saw 

 abundant spermogonia on Amelanchier from the car window. 



SUMMARY OF CONCLUSIONS. 



Roe 



is known. The American forms thus named hitherto, in- 

 clude two distinct species, namely, R. pyrata, which is the 



aecidium of Gymnosporangium macropus 

 which is the jecidium of G. clavariicforme. 

 Roestelia lacerate, as it is generally kn< 

 and as it is distributed in Ellis' N. A. F., No. 1085, is incor- 

 rectly named and is the rccidium of G. globosum, to which 

 should also be referred the smaller form common on Pyrus 



malus. 



R 

 R 



R. cornute is not the a-cidium of the " birds'-nest Gym- 

 nosporangium previously referred to G. conicum and dis- 

 tributed as G. clavipes in Ellis' N. A. F., No. 1084 (b). The 



"birds"-nest" form with its Roestelia as obtained by cultures 

 is therefore probably unnamed, unless G. conicum is erro- 



