DODGE: STUDIES IN THE GENUS GYMNOSPORANGIUM 137 



two lying further back, the other two occupying separate branches. 

 A cell wall is finally laid down, cutting off the outer portion of the 

 germ tube containing the nuclei and most of the cytoplasm (Figs. 

 8-10). This curious method of germination is in no sense similar 

 to the development of a promycelium, though four nuclei are pro- 

 duced in each case. It may be that it is fairly common among the 

 rusts as Sappin-Trouffy has pointed out.^ Whether or not it is pos- 

 sible to find an appropriate artificial medium for the development of 

 the mycelium of a rust in artificial cultures, it would seem that such 

 cases as these afford at least a starting point. Four-nucleated germ 

 tubes are the rule in these cultures, but fully developed tubes with 

 only two nuclei are not difficult to find (Fig. 11). 



In some of my cultures in which the petiole of a leaf had been 

 infected at the junction with the blade, it was found that the winter 

 bud was larger than usual. The mycelium must have run down the 

 petiole and become established in the bud. When such plants were 

 put in the cold frame over winter and taken out in the spring, these 

 buds developed small leaves which at once became evenly covered 

 with spermogonia and later were transformed into large galls from 

 which aecidia developed quite normally. 



In some cases the mycelium seems to penetrate into the tissues of 

 the stem where a spindle-shaped swelling or burl is formed. In the 

 following spring a green gall bursts out through the cork, forming a 

 nodular swelling outside and from this spermogonia and aecidia are 

 produced. I have had several cases in which Roestelia transformans 

 has survived the winter and developed aecidia the following spring 

 The same is true in my cultures of R. Botryapites. In October, 1915, 

 winter buds of six Amelanchiers showed signs of being infected. All 

 of these survived the winter and developed spermogonia and ripened 

 aecidiospores in the month of June, which is several months earlier 

 than they can be found in nature. These are not cases where the 

 formation of an aecidium has simply been delayed. On the con- 

 trary, an entirely new crop of spermogonia arises from newly formed 

 tissue, new gall growth, and we find the aecidia developing as in 

 normal cases of infection with sporidia. 



GYMNOSPORANGIUM FRATERNUM 



The buffer cells in the teleutospore sori of G. fraternum are much 

 more striking in appearance, forming as they do a perfectly even 

 palisade layer that frequently extends entirely across the sorus without 

 interruption (Text-fig. 4). This is a very characteristic feature of 



^ Sappin-Trouffy, P. Rocherches histologiques sur la famille Ics Uredinees. 

 Le Botaniste 5: 59-244./. 1-69. i D 1896. 



