226 BOTANICAL GAZETTE [September 



In the mature opened condition shown in fig. 12, the thin- walled 

 cells of the upper part of the apothecium still persist on the margin 

 of the fruit body. The opened perithecia present in surface view 

 the appearance shown in fig. 5. The folding back of the shield 

 is shown in section in fig. 15. The perithecia develop independently 

 of the acervuli, as would be expected from the origin of the two. 

 In fig. 13 is shown an old acervulus by the side of a perithecium. 

 In none of the material which had wintered could acervuli be found 

 which were bearing conidia. 



1 



The epidermal cells of the host persist for a long time, so that 

 the ascogenous layer and shield are separated. They may become 

 entirely destroyed as the asci elongate and the peritheciurgi becomes 

 mature (fig. 14), or they may persist on the margin of the mature 

 perithecium (fig. 12). The perithecia vary in shape from spherical 

 to discoid. One of the large discoid perithecia is represented in 

 section in fig. 14. The septate, knobbed paraphyses extend 

 between and beyond the asci until the time when the spores are 

 nearly mature. Asci in many stages of development occur within 

 each perithecium. Mature asci extend slightly beyond the para- 

 physes and the spores are discharged from an apical pore (fig. 16) 

 formed by the rupture of the wall. The asci are oblong or subcla- 

 vate, tapering above rather bluntly, and are 70-80X15 /*. 



Apparently the spores are not discharged with violence. Agar 

 plates were inverted above rose leaves in moist chambers, the 

 surface of the agar coming nearly in contact with the leaf. No 

 spores were observed to have lodged on the surface of the agar, 

 as would be expected if they were projected forcibly from the 

 ascus. As far as I have been able to observe, they merely pile 

 up in a whitish heap in the opened perithecium. The spores 

 are 20-25X5-6 /*, varying extremely in form (fig. 17), as do the 

 conidia. They resemble the conidia very much except that they 

 are not so strongly constricted at the septum. They are hyaline 

 and bicellular. Usually large granules and several guttulae are 

 present in each cell. The cells are generally unequal in size, the 

 upper one being broader. 



