546 THE FERTILISATION OF FLOWERS. [PART III. 
has passed from beneath the petaloid style, when it flies away to 
another flower or another perianth-segment of the same. If the fly 
be caught as it is creeping backwards from the nectary, many pollen- 
grains may always be found among the hairs on its back ; some of 
these are without fail applied to the stigmatic surface next visited. 
Thus the flowers in which the petaloid-styles stand close to the peri- 
anth-segments are better adapted for fertilisation by Rhingia in two 
respects than the flowers of the other form: (1) the fly is compelled 
to carry pollen to the stigmatic surface, often to that of another 
flower, and (2) it is hindered from eating pollen; whereas, in 
the flowers of the other form, Rhingia passes in and out without 
accomplishing fertilisation, and moreover uses up pollen as well 
as honey. 
In relation to bees’ visits the conditions are exactly reversed. 
When the petaloid style lies close upon the outer perianth- 
segment, the opening, which is sufficient for Rhingia, is much 
too small to admit a humble-bee. On May 25th, 1868, I watched 
a large parasitic humble-bee (Psithyrus vestalis, Foure. 2, 25 mm. 
long, and 10 mm. broad) creeping in various directions over the 
flowers; finally, placing its head above the base of the free part of 
an outer perianth-segment, it inserted its tongue sideways into one 
of the nectaries, and so reached the honey without touching either 
the stigma or anther. 
In flowers of the other form, I have observed frequently in the 
case of Bombus agrorum, B. hortorum, and B. Rajellus, that humble- 
bees alight on one of the outer perianth-segments, and creep under 
the petaloid style to the honey-passages, touching on their way 
both stigma and anther with their backs. Instead of crawling 
backwards along the path they came by, they climb sideways on 
to one of the other outer perianth-segments, and then crawl up it 
to enter beneath the stigma, as before. After completing the 
circuit of the flower and exhausting all its honey, the bee flies to 
another, and so effects cross-fertilisation regularly, 
Occasionally, but only in those few flowers where the position 
of the petaloid styles was intermediate between the forms that I 
have described as typical, I have seen humble-bees creeping back- 
wards from under cover of the style, as Sprengel describes them. 
The flowers of Iris pseudacorus thus present the interesting 
peculiarity that although originally, as their dimensions show, they 
were adapted simply for bees, they have since to a great extent 
become adapted for fertilisation by flies (Rhingia). Each of the 
two extreme forms has its own advantages and disadvantages. 
