PART III. | THE MECHANISMS OF FLOWERS. 545 
petaloid style, which is the functional stigma; a little further on 
it rubs against the pollen-covered surface of the anther, which lies 
close to the petaloid style. In this way, Sprengel declared that 
two stigmas in each flower were regularly fertilised by means of 
pollen from the previously-visited anthers of the same flower, and 
that the stigma first visited in each flower was always cross- 
fertilised by means of pollen from the preceding flower: thus self- 
fertilisation took place twice as frequently as cross-fertilisation. 
Though Sprengel was not aware of the advantage of cross-fertilisa- 
tion, he was repeatedly led by his observations to the conclusion 
“that nature is unwilling that any hermaphrodite flower be 
fertilised by means of its own pollen.” In order to adhere to this 
theory in the ease of Iris, which he abandoned again and again in 
other cases, Sprengel chose to Jook upon each third of the flower 
in Iris as a separate hermaphrodite flower, which thus was 
regularly cross-fertilised. 
My frequent observations of insect-visits differ materially from 
Sprengel’s view of the case. It is true that humble-bees visit 
the flower of Jris pseudacorus, but they, as a rule, act in a way 
very different from that which Sprengel described; and a very 
much more frequent visitor is, in my experience, the long-tongued 
hover-fly, Rhingia rostrata. 
In many flowers of Jris pseudacorus the petaloid styles stand 
6 to 10 mm. above the corresponding outer perianth-segments. In 
others each petaloid style lies so close to the perianth-segments 
that only a very small passage is left, which, owing to the con- 
vexity ofthe style, lies immediately below the stigma. These two 
forms of flowers have adapted themselves to different fertilising 
agents. 
In flowers of the first variety the fly (Ahingia) crawls along 
the outer perianth-segment to the nectaries without touching 
either stigma or anther; it inserts its proboscis (11 mm. long) 
into one nectary after the other, and then moves a few steps 
backwards to feed on the pollen. When under the anther it 
raises its head, stretches out its tongue, and so reaches the pollen. 
Then it flies to another perianth-segment of the same flower, or to 
another flower. Its visits to this variety are thus not only useless 
but hurtful, leading to waste of pollen. 
In flowers of the other form this fly creeps through the small 
passage beneath the true stigma, and touches with its back first the 
stigma and then the anther before reaching the nectaries. After 
sucking, it creeps backwards without staying to eat pollen, until it 
NWN 
