‘parrut] THE MECHANISMS OF FLOWERS. 311 
ae if these female plants flowered earlier than the hermaphrodites, 
i ‘so that their stigmas were mature cotemporaneously with the 
anthers of the first hermaphrodites, then their origin might be 
explained on the principle of economy, as a saving of the first 
stamens which have no stigmas to fertilise. Ifthe female capitula 
_ were less conspicuous than the others, the same explanation might 
__ be applied in this case as I have put forward in describing Thymus 
and Glechoma. But here neither the one fact nor the other is true ; 
but the female heads are just as conspicuous as the others at: 
develop cotemporaneously with them. I can only account for 
_ their existence by the very great difference between the duration 
of the stigmas in the hermaphrodite capitula and the length of 
time that the anthers are covered with pollen. 
It is of special importance for Scabiosa arvensis that the anthers 
‘on the same head should develop gradually throughout several 
days, and that the stigmas should ripen all on the same day, almost 
simultaneously. For owing to the first character there is never a 
, lack of pollen, and owing to the second, as soon as a few sunny 
hours occur and bring out the insects, in a very short time all the 
‘stigmas on a capitulum are fertilised with pollen from another. 
_ Both characters together are therefore of use to the plant in the 
changeable and often continuously rainy weather of our summers. 
But when a spell of sunny weather appears the number of herma- 
_ phrodite flowers in the male condition must always be much 
eater than in the female condition; the pollen of many heads 
must be wasted, and it must be an advantaxe to the plant if in 
- some cases the stamens are aborted and the stigmas come so much 
ra the quicker to maturity. As every character of advantage for the 
maintenance of the species, which appears accidentally as an 
abnormality, can and must be retained by natural selection, so in 
| Scabiosa arvensis, abortion of the stamens, occurring accidentally 
in certain plants, must have been perpetuated and intensified. 
_ The power of self-fertilisation is not quite lost in the herma- 
phrodite plants; for in capitula allowed to flower in the house and 
LL eft untouched, many of the stigmas as they grow up may be seen 
_ to come in contact with anthers still dusted with pollen. But as 
» a rule insect-visits are so numerous that self-fertilisation can 
_ only come into action in long-continued bad weather. 
) : Visitors : A. Hymenoptera—(a) Apide: (1) Apis mellifica, L. § (6), ab., 
Sy more rarely c.p. ; (2) Bombus hortorum, L. g 9 9 (17—21) ; (8) B. terres 
-: L. 2 ¢ (7—9) ; (4) B. lapidarius, L. § (10—12) ; (5) B. pratorum, L. 
