158 
CHARLES B. PLOWRIGHT. 
grew in the intervening part were quite free from the Puccinia. 
Neither did I observe anywhere in it a single pustule of P. 
phragmitis. In the spring of 1884 I from time to time 
visited this ditch and carefully examined the Pumices and 
Ranunculi growing on its banks, for I felt certain, from the 
localised profusion with which P. Magnusiana occun’ed, that 
I should meet with its iEcidium at both ends, but not in the 
central part. This surmise was found to be correct, and to 
confirm the conclusion I had already, from a series of ex- 
perimental cultures, arrived at. At both ends of the ditch 
I found the Rumices free from iEcidia, but the plants of 
Ranunculus repens abundantly affected with iEcidia, 
while in the middle neither one or other of these plants 
had any iEcidium upon them. Germinating spores of P. Mag- 
nusiana were placed upon Ranunculus repens with the 
uniform result of giving rise to the iEcidium (315, 335, 358). 
It is obvious that if Ranunculus repens is the host plant to 
two specifically distinct iEcidia, the possibility of other Ranun- 
culi also being hosts of one or other of them must be considered. 
P. Magnusiana was, therefore, applied to R. acris (360)> 
R. ficaria (361), R. auricomus (359), but without any 
result. When, however, R. bulbosus was infected it always 
developed the iEcidium (393, 394, 395, 396, 397), hence P. 
Magnusiana has its iEcidium upon both R. repens and R. 
bulbosus. Conversely, the secidiospores of P. Magnusiana 
which had been artificially produced upon R. re pens (369) 
and R. bulbosus (422), were placed on Phragmitis, where in 
due time they gave rise to the orange Uredo with paraphyses. 
To make more sure, a part of the same spores from R. repens, 
which, when placed on Phragmitis (369) gave origin to the 
Uredo of P. Magnusiana, were applied to Poa trivialis 
(370), but they gave rise to no Uredo ; and in like manner a 
part of the spores from the iEcidium R. bulbosus (423) were 
placed upon Dactylis glomerata, but they gave rise to no 
Uredo. The two iEcidia on R. repens were most carefully 
examined side by side, but no difference could be detected by 
me. As these two iEcidia occur in a state of nature, however, 
