258 THE FERTILISATION OF FLOWERS. [PART IIL. 
differ so much from each other. In all these forms the stigmatic 
papille consist of long cells, whose ends are cut off as round or 
elliptical knobs by a narrow neck. In the long-styled flowers these _ 
cells are 49-9 mm. long, and as the part below the constriction is — 
about twice as thick as the knob above, these knobs stand at a 
distance from each other equal at least to their own diameter. In~ 
the mid-styled form these papillze are only *f,%° mm., and in the 
short- styled 25-45 mm. long; and in both, the knobs stand very near 
together, since the part of the cell below the constriction is little 
thicker than that above. 
Now it was Darwin’s great service to prove beyond question, | 
by a long series of patient experiments, that in Lythrum Salicaria — 
reproductive organs, when of different length, behave to one another 
like different species of the same genus in regard both to direct 
productiveness and the character of the olipeind: and that con- — 
sequently mutual barrenness, which was once thought conclusive 
proof of difference of species, is worthless as such, and the last — 
barrier that was raised between species and varieties is broken — 
down. | 
Now of the eighteen possible ways of fertilisation —applyail 
the six different kinds of pollen to the three kinds of stigmas,— 
those six only are fully productive in which a stigma receives 
pollen from a stamen of equal height. The probability that the 
large pollen-grains from the long stamens are only suited to the 
long, widely separate papille on the long styles (and so forth in che 
other cases) is thus fully confirmed by experiment. But such ¢ 
adaptation, such a differentiation of the sexual parts, is only 
intelligible if it can be shown that under natural conditions insect- 
fertilisation causes the distribution of pollen to the proper stigmasy, 
and it is clearly proved by observation that this is so, 
The honey of Lythrum Salicaria, inclosed at the base of a tube 
5 to 7 mm. long, is accessible (1) to insects which have a proboscis 
5 to 7 mm. long; and (2), since the entrance to the tube is 23 to 
3 mm. wide, to insects whose head is small enough to be thrust 
partly or entirely into the tube ; and (3) to insects small enough to 
creep down to the base of the flower. 4 
Almost every insect, whose body along with its proboscis is at 
least 12 to 15 mm. long, must in sucking the honey touch the 
shortest organs with its proboscis, or, if this is less than 5 to 7 mm. - 
long, with the under side of its head ; the next series with a part of 
its ventral surface 4 to 5 mm. Gorkher back, and the third seri 
with another part of its ventral surface the same distance behind 
