PART III. | THE MECHANISMS OF FLOWERS. 527 
hitherto been protected from contact with the water, is now carried 
by currents to the stigmas of female flowers. Va/llisneria is, in a 
certain sense, a transition from the water-fertilised to the insect- 
fertilised flowers ; Hydrocharis is distinctly entomophilous (178, I1.). 
Orv. ORCHIDEL. 
This family is remarkable for the following characters, due to 
its wide distribution and to its enormous number of species: first, 
for great variety of habit and diversity of station; secondly, for its 
immense variety of peculiar and highly-specialised flowers; and 
thirdly, for the unusually large number of seeds produced in each 
capsule. The diversity of habit and place of growth must be ex- 
plained by supposing all the parts of the plant concerned in nutri- 
tion to have been very variable. I have discussed at full length 
their variability in our recent species, in a paper upon the genetic 
relations of Hpipactis viridifolia, EB. microphylla, and £. latifolia 
(565). 
The multiplicity of forms of flowers is not difficult to under- 
stand if we only suppose the ancestors of the family to have been 
as liable to variation in their flowers as our recent species are. Dr. 
Rossbach has discussed the variability of the flower in Orchis fusca, 
and I myself in the genus Habenaria.? 
I do not doubt that Orchids owe their extraordinarily perfect 
, adaptations to particular insects not only to the tendency of the 
parts of their flowers to variation, but also to the separation in 
time of the two stages in the act of impregnation. At the time 
of flowering the ovule is, as a rule, not yet developed; pollen is 
applied to the stigma, and the pollen-tubes grow out, but the ovule 
is only developed and impregnated weeks or months, sometimes six 
months, later. The extremely complete adaptations to cross-fertil- 
isation have in their turn resulted, in many cases, in the flowers 
becoming absolutely sterile to their own pollen. 
In regard to capacity for fertilisation by their own pollen, 
Orchids show the greatest possible differences, all of which, however, 
are linked together by intermediate conditions. We find in this 
order, cleistogamic flowers* and open flowers; flowers regularly 4 
1 Verhdl. des naturh. Vereins fur pr. Rheinl. und Westfal. p. 166, 1857. 
2 Tbid., pp. 36-47, 1868. 
3 Schomburgkia, Cattleya, and Epidendrwm (Criger, 149) ; Dendrobiwm (Ander- 
son, 5); Zhelymitra (Darwin, 167). ; 
4 Ophrys apifera (Darwin, 155); Neotinea intacta (Darwin, 159) ; Gymnadenia 
tridentata and Platanthera hyperborea (Asa Gray, 274); Epipactis viridifolia (H. 
Miiller, 565) ; Epidendrwm (Fritz Miiller, 553). 
