118 THE FERTILISATION OF FLOWERS. [PART III. 
which cross-fertilisation is assisted in Viola tricolor, which had been 
overlooked by Sprengel. I have observed (No. 570) that this lip, 
which insures cross-fertilisation, is only developed in the large- 
flowered variegated variety of V. tricolor, while in the small- 
flowered, whitish-yellow variety, V. arvensis, the orifice of the 
stigma is devoid of a lip and is curved inwards; in this form the 
flower fertilises itself spontaneously very soon after opening. 
- In all our violets the spurs of the lower anthers act as honey- 
glands, and the hollow spur of the lower petal which contains 
them forms a receptacle for the honey. To reach the honey in 
Viola tricolor, the insect must thrust its proboscis into the flower 
close under the globular head of the stigma. This lies in the 
anterior part of a groove, fringed with hairs, on the inferior petal. 
The anthers, which together form a cone, shed their pollen inte 
this groove, either of themselves or when the pistil is shaken by — 
the insertion of the bee’s proboscis. The proboscis, passing down 
this groove to the spur, becomes dusted with pollen; as it is 
drawn back it presses up the lip-like valve so that no pollen can 
enter the hollow stigmatic chamber; but as it enters the next 
flower it leaves some pollen on the upper surface of the valve, and 
thus cross-fertilisation is effected. 
I have observed very numerous bees visiting a handsome, 
large-flowered. variety which is not uncommon on sandy soil at 
Lippstadt, growing along with V. arvensis. Apis mellifica, L. %, 
always sucks in a reversed position with its head downwards ; 
it often alights in the usual position and then turns round 
Humble-bees usually suck hanging back downwards to the 
flowers, which are bent down by their weight. Bombus terrestris. 
(7 to 9) and B. lapidarius (12 to 14) usually turn round, so as to 
introduce their proboscis in the same way as the honey-bee. 
B. hortorum (18 to 21) and Anthophora pilipes (19 to 21) grasp the 
flower from behind with their forefeet, and insert their tongues 
from below in an opposite direction to the honey-bee (590, IL). 
Andrena albicans, K. 6, made efforts to suck honey, thrusting 
its proboscis repeatedly into the same flower. The attempt must 
have been fruitless, as the proboscis of this bee is only 2 to 2} mm. 
long, while the pollen-collecting groove described above is 3 mm, 
long, and the spur extends for another 3 mm.; in this case, the 
repeated insertion of the proboscis into the same flower must have 
led to self-fertilisation. I repeatedly noticed a common small fly, 
Syritta pipiens, L., eating pollen in the groove and on the anthers 
of V. tricolor; as it often touched the stigma also with its 
