PART III. | THE MECHANISMS OF FLOWERS. 473 
hermaphrodite ones; his theory rests, if I understand it rightly, 
upon the two following arguments. Since no stigmas are mature 
at the time when in the first flowers of a proterandrous plant the 
stamens dehisce, these stamens are of no use to the plant, and since 
all economy of useless organs is advantageous to the plant, the 
stamens of the first flowers of proterandrous plants can be abolished 
by Natural Selection. This reasoning is undeniably correct, but it 
applies only to the first flowers of proterandrous plants, and not to 
Fig. 161, 
1—3.—Thymus Serpyllum, Fr. 
1.—Hermaphrodite flower, in first{(male) stage. 
| 2.—Ditto, in second (female) stage. 
3.—Female flower. 
4—6.—T. vulgaris, L. 
4.—Female flower. ; 
5.—Pistil of hermaphrodite flower in first stage. 
6.—Ditto, in second stage. ov, ovary; n, nectary. (xX 7.) 
the female stocks of Thymus which bloom all summer beside the 
hermaphrodite ones. This in my opinion is only to be explained in 
the way that I have suggested in the case of Nepeta (p. 484). The 
variability in size, which my theory presupposes, is present in our 
two species of Thymus, as in Nepeta, to such a degree that, while 
the largest hermaphrodite flowers (Fig. 161, 1, 2) are several times as 
large as the smallest female (Fig. 161, 3), the smallest hermaphrodite 
and the largest female flowers are nearly equal in size. The flowers 
