446 THE FERTILISATION OF FLOWERS. [PART III. 
below the anthers, in the position h, 2, and in doing so it cannot fail 
to touch the downward-pointing ends of some of the anthers. The 
four anthers are held together posteriorly by matted hairs (e, 5), 
and anteriorly dehisce from the apex inwards (5, Fig. 152) by a 
wide fissure ; so that a single touch on the part of the bee is com- 
municated to all the anthers, and causes them all to shed a part of 
their dry, dusty pollen. Hairs directed downwards at the edges 
of the anthers (7, 5) prevent the pollen being scattered at the sides, 
so that great part falls on the bee’s proboscis, and is carried to the 
stigma of another flower. 
The development of the flower of Odontites serotina varies in a 
peculiar manner, and seems to depend upon the more or less 
sunny nature of the spot. The style bearing the fully-developed 
stigma usually projects far out of the bud shortly before it expands 
(1, Fig. 152), and the flower can therefore be cross-fertilised when 
it is scarcely open. In shady places, for instance among corn, the 
style lags behind while the corolla and stamens continue to grow, 
so that the stigma stands between the taller anthers and gets self- 
fertilised (2, Fig. 152). From the great productiveness of all the 
capsules in shady places where insect-visits are rare, we may con- 
clude that such self-fertilisation is efficient. In sunny spots the 
style continues to grow along with the corolla and stamens, and 
overtops the anthers to the last; it may either lie above the latter 
(3,. Fig. 152) or project at the side (4, Fig. 152). In this case, the 
stigma can only receive pollen by the aid of insects, and from other 
(either older or younger) flowers. Although this difference seems 
in general to depend upon the sunny or shady character of the 
spot, other unknown conditions (perhaps heredity) must co-operate ; 
for I have sometimes observed in the same spot, and even on the 
same plant, flowers which fertilised themselves, and others incapable 
of doing so. 
An imperfection in the flowers deserves special mention. The 
upper lip, as the figure shows, projects so slightly that the anthers 
in great part protrude beyond it and are unprotected. This is 
sometimes the case to such a degree that bees thrust their tongues 
into the flower above the stamens, and thus suck honey and shake 
out the pollen without leading to fertilisation. 
I have only seen bees visiting this plant. 
(1) Apis mellifica, L. 9, very ab., thrusting its proboscis into the flower 
usually below the stamens, but sometimes above them,—it inserts its proboscis 
even into young unexpanded flowers, with long exserted styles ; (2) Bombus 
lapidarius, L. 9 9, s.; (3) B. silvarum, L. 2 9, s. 
