CYLINDROSPORIUM ON STONE FRUITS 
163 
only subepidermal). They are distinct in the shape of the ascocarp, 
its position in the leaf tissue, and in the shape and size of the asci. 
Those having the subepidermal ascocarps fall into two distinct 
groups. On the plums (P. domestica and P. insititia) the fruit bodies 
are coal black with a very decided tendency to be aggregated in clus- 
ters, the asci are smaller, more slender than in the other forms, and are 
almost filled by the spores which are more slender than in either of 
the other forms. 
Contrasted with this we find on P. virginiana, P. serotina, and P. 
mahaleh the light (yellowish brown to dull orange) colored fruit 
bodies, larger asci, and apothecial conidia much larger than in the other 
forms. The covering of the fruit body is composed of more delicate, 
thinner-walled cells. 
After this grouping was made it was found that the hosts of each 
group fall in the same natural group of the genus Prunus, with the 
possible exception of P. mahaleh. The hosts of the first group are all 
in the subgenus Cerasus of Engler and Prantl, with flowers in umbels. 
Those of the second group are plums which are placed in the subgenus 
Prunophora by the same authors, while those of the third, with the 
exception of P. mahaleh, are placed by them in the subgenus Padus 
which has the flowers in elongated racemes. P. mahaleh has its 
flowers in a short raceme, and is placed by them in the subgenus 
Cerasus but in a group separate from the other members of the sub- 
genus. Britton places it in a group with P. virginiana and P. serotina 
all having flower clusters terminating branches of the present year's 
growth. 
The results of cross inoculations show a correlation of hosts with 
characters of the ascogenous stage. In no case has it been found 
possible to transfer, by inoculations, the fungus from hosts of one group 
to hosts of another group. 
The fact that the hymenium of the ascocarp is surrounded — until 
nearly mature — by a wall of tissue which finally opens in a stellate 
manner places all forms of the fungus in the Phacidiales. The relation 
of the ascocarps to the leaf tissue (covered by the adherent host epi- 
dermis), the shape of the asci, and the elongated colorless spores 
grouped in a fascicle show a distinct relation to Coccomyces in which 
genus therefore all forms are included. Had only the light-colored 
form on the Padus group been found there might have been some 
question as to its relation to this genus, since Coccomyces is said to 
