584 THE FERTILISATION OF FLOWERS. [PART IY. 
rather more numerous bees can do so. The pollen is most com- 
pletely protected in those flowers whose stamens are completely 
shut up within the corolla (Fumariaceew, Papilionacee, Rhinanthus, 
Melampyrum, Pedicularis, Salwia), or whose pollen is united in 
masses and enclosed in pouches (Asclepiadew, Orchidew). These 
plants are no longer able to have their pollen transported by 
miscellaneous insects, but they have developed adaptations of the 
most perfect kind for cross-fertilisation by certain special insects. 
The mechanism is so perfect and so effectual in these flowers 
that cross-fertilisation is thoroughly insured, though some of them 
offer only pollen to their visitors (Genista, Sairothanwinns) others 
pollen and sap (Cytisus), others sap only (Orchis). 
The contrivances that we have just reviewed give to two 
divisions of anthophilous insects exclusive or almost exclusive 
access to the food materials contained in certain flowers ; namely, 
in most cases to bees, which, however, must usually share their 
honey with the Lepidoptera; and in relatively few cases to 
Lepidoptera, when, if the pollen lies exposed, flies may devour 
it and bees collect it, playing a subsidiary part in the work of 
fertilisation. But certain small Diptera which resort to small dark 
cavities have a much more complete monopoly of those flowers 
which furnish them with such temporary abodes, and which have 
become adapted for cross-fertilisation by them (Aristolochia 
Clematitis, Arum maculatum); for the narrow entrance usually 
excludes all other insects. 
Limitation of Insect-visits by Time of Flowering and by Place 
of Growth. 
It is obvious that every species of flower can only be visited and 
fertilised by those insects which occur at the time when the plant 
is in flower, and in the stations where it grows. The insect- 
visitors of a plant are therefore limited by the season and by the 
time of day when it flowers, by its geographical distribution, and 
by the nature of its habitat. 
We may assume, furthermore, with Delpino (175 A), that the 
geographical range of many flowers finds its limit where insects 
suitable for their fertilisation begin to fail. But the particular 
cases which Delpino adduces to support this view are, in part, 
based on insufficient observations.' Our stock of observations is 
1 Thus Delpino supposes that the geographical range of Roses is determined by 
the occurrence of Cetonie and Glaphyride (175 A, p. 18), and that Lpilobiwm and 
Mvyosotis are exclusively fertilised by bees (p. 19). My lists contradict both of these 
assumptions, 
