propagated by layering or by cuttings, of course, only continue the individual character of the mother-plant or stock." The accompanying plate shows various forms of grape flowers. [Illustration: FLOWERS OF VITIS] He further says in a foot-note: "These fertile plants, however, are of two kinds; some are _perfect hermaphrodites_, with long and straight stamens around the pistil, the others bear smaller stamens, shorter than the pistil which soon bend downward and curve under it; these may be called _imperfect hermaphrodites_, approaching females, and they do not seem to be as fruitful as the perfect hermaphrodites, unless otherwise fertilized." Beach[122] tested many of our cultivated varieties by sacking the clusters at blooming time and thus determined their capacity to fertilize themselves. From the data thus secured he divides them into four classes: 1st. Those that are able to fertilize themselves so that the clusters are perfect or varying from perfect to somewhat loose. 2d. Those in which the clusters are marketable, varying from moderately compact to loose. 3d. Clusters so loose as to be unmarketable. 4th. Those which are self-sterile or showing no fruit on covered clusters. Of 169 varieties tested, he found 38 belonging to the first class, 66 to the second class, 28 to the third, and 37 to the fourth. Later it was found that the reason why certain varieties were self-sterile was on account of impotent and abortive pollen, the percentage of abortive pollen grains varying with different varieties and this percentage determining the degree of self-sterility. The upright or depressed stamen is not an invariable criterion of the condition of the pollen although it is usually. There are a few instances in which upright stamens bear impotent pollen but these are very exceptional. Munson made similar tests of vines of twenty-two[123] American species of vines secured from their habitats. In every case he found that they showed only two forms, the staminate vines and the self-sterile hermaphrodite, no perfect hermaphrodites being found. While of some of the species the number of vines tested was a half dozen or less, in most instances many vines were tested from different places. This is particularly interesting in that it becomes a puzzle as to where our perfectly hermaphrodite cultivated forms could have come from if such forms are not present in the wild vines of our woods and prairies.