94 
FLOWERS AND INSECTS 
vegetative reproduction, or to have shorter tubes enabling the hemitro- 
pous visitors to obtain the honey. Thus the insect fauna and its 
distribution are important factors in determining the geographical 
distribution of plants. Usually the number of species visiting any one 
flower that is regularly insect-pollinated is at least two or three and 
often very many ; if this flower be studied in different countries it will 
be found visited by different species, but of the same biological class, 
as regards length of tongue, &c. ( e.g . short-tongued bees may be 
replaced by long-tongued flies). There are however a number of cases 
of mutual adaptation of one particular flower and one particular insect, 
and in these cases the distribution of the latter regulates that of the 
former. The best known case is that of Bombus and Aconitum (. q.v ., 
and see Bryonia, Angraecum, Yucca, &c.). 
Floral Mechanisms etc. It is evident, that if a 
visiting insect is to be of any use, that part of its body 
which touches the pollen must also be the part to touch 
the stigma. In a declinous plant this is easily enough 
managed, but difficulties arise when the flower is herma- 
phrodite. So long as insects merely sprawled about flowers, 
feeding on pollen, it would not, perhaps, be much disad- 
vantage to have the stigmas at a distance from the anthers, 
as insects would probably touch them sooner or later ; there 
would however probably be at least as much self- as cross- 
pollination, even with insect visits. When nectaries and 
perianth were more fully evolved, the track taken by the 
insect visitors would be more definite ; the same species of 
insect would visit the same species of flower always in the 
same manner. It would now become a necessity to place 
anthers and stigmas so as to touch the same part of the 
insect, i.e. they must generally be close together, and so 
cause difficulty in avoiding pure autogamy. The ways in 
which this is effected present a bewildering variety, but all 
have the same general underlying principle, to ensure cross- 
pollination as far as may be done without seriously affecting 
the certainty of setting seed. 
The most certain mode of avoiding autogamy, while 
keeping anthers and stigmas in positions to touch the same 
portion of an insect visitor, is of course declinism . This was 
perhaps common among the early flowering plants, but is 
not so now; it is found in a number of cases, however, 
most of them apparently derived from hermaphrodite an- 
cestry (see p. 68). 
