THE FERTILISATION OF CERTAIN PLANTS. 53 
the long persistence of the corolla when fertilisation is pre- 
vented. I was informed, also, by a large horticulturist, that 
when the bees get at his flowers in the greenhouses the 
plants become almost worthless for sale, owing to the speedy 
fall of the blossom. The teleological significance of this is of 
course apparent. The corolla is of no further use when ferti- 
lisation has once occurred, and its maintenance is a useless 
burden. It is therefore cast off. But how this is brought 
about is not so obvious. I presume that the fertilisation causes 
a larger flow of nutriment to the stimulated ovary, and that 
this flow takes place at the expense of the corolla, which then 
perishes from starvation. 
The last exceptions I shall consider are wild thyme (T. ser- 
pyllum) and marjoram (Origanum). 
In each of these we have a didynamous flower, which cer- 
tainly is fertilised by insects, for its anthers ripen and shed 
their pollen before the stigma is mature. The four anthers, 
however, are not placed longitudinally, one pair above the 
other, for the stamens diverge laterally and project beyond 
the corolla in a fan-shape. The breadth which the anthers thus 
cover is much greater than can come within the range of the 
centrally placed style. The difficulty, however, vanishes when, 
instead of examining a single flower, we consider the general 
mode of growth and the inflorescence. Thyme, as a rule, grows 
in patches. The small flowers are crowded together on the 
surface of the patch, forming a continuous carpet of blocm, 
from which project upwards the ripe anthers of the younger 
and the ripe stigmas of the older flowers. Over this carpet 
crawl the bees, their bodies so large in proportion to the indi- 
vidual blossoms as to be in contact with many at a time. 
The crowded heads of marjoram form a similar carpet. Thus 
all parts of the under surface and sides of the bees get dusted 
with pollen, and similarly all parts come sooner or later into con- 
tact with stigmas. There is thus no occasion for that definite 
arrangement of the anthers which is profitable in the large- 
flowered didynamia. What is required is that the anthers 
should project beyond the corolla, and this they do. 
It would clearly also be disadvantageous that the dehiscent 
surface of the anthers should be turned, as in most other didy- 
namia, forwards. To ensure the readiest contact with the insect 
as it crawls over the flowers, this surface should face directly 
upwards. Now with the longer stamens, which are entirely 
free of the corolla, this is the case. The lower pair of anthers 
only, as a rule, half overtop the corolla’s lip. The dehiscent side 
of their upper free lobe is turned upwards ; that of the lower 
lobe faces forwards. Thus always the dehiscent surface assumes 
the most advantageous position. That of each free lobe is in 
