534 STEVENS: DIOECISM IN THE TRAILING ARBUTUS 
EVIDENCE OF HETEROSTYLY 
The claims as to heterostyly in Epigaea are based entirely 
upon the variation in the length of the pistils As Darwin* 
(p. 3) emphasized, however, in his original discussion of heter- 
ostylous plants, anatomical characters alone do not furnish con- 
clusive evidence of heterostyly. The essential differences between 
the forms of a heterostylous species are physiological differences 
in the pistils and pollen. Unless it be proven that one form is 
wholly fertile only when pollinated with pollen from another 
form, we do not have conclusive evidence that the species is 
heterostyled. 
In Epigaea, moreover, the anatomical evidence is anything 
but conclusive. Gray (loc. cit. 75), to be sure, divides the large 
stigma kind into two groups distinguished by different lengths of 
~ style, but there are all gradations from the longest style to the 
shortest (Wilson, loc. cit. 59). The small stigma form shows even 
less evidence of heterostyly, for both styles and stamens show all 
gradations in length and the stamens vary within much narrower 
limits than do the pistils (Gray, loc. cit. 75). And in place of the 
correlation of pistils of a given length with stamens of a different 
length characteristic of heterostylous flowers, we find in Epigaea 
any length of stamen associated with any length of pistil (Wilson, 
loc. cit. 59, and Miss Langdon, loc. cit. 11); and flowers with 
stigmas at the anther level are common (Halsted, loc. cit. 249). 
Moreover, as Halsted points out, the pollen grains from the 
stamens of different lengths show no difference in size; and this, 
while by no means conclusive, lessens the probability that the 
species is truly heterostylous. All these observations have been 
confirmed by the writer. 
Experiments, however, show still more conclusively that 
Epigaea cannot be considered a heterostylous plant. The writer 
transplanted several wild plants with large stigma, each of which 
bore a dozen or more buds, and cultivated them out of doors. 
The plants apparently suffered no injury and all blossomed readily. 
While the pistils showed a wide variation in length they may be 
roughly classified as long-styled, mid-styled, and short-styled 
forms. Flowers of each of these forms were pollinated with pollen 
from long and mid-length stamens. No flower with stamens 
