_ PART IIL. | THE MECHANISMS OF FLOWERS. 219 
Four different types of structure may be distinguished in 
_ Papilionacew, according to the manner in which the pollen is 
q applied to the bee. These distinctions were first drawn by 
| Delpino (172, 178, 360), and transitions are not wanting from one 
_ to another: 
1. Papilionaceze in which the stamens and stigma emerge from 
_ the carina and again return within it. They admit repeated visits. 
 (Melilotus, Trifolium, Onobrychis, Cytisus). 
| __- 2. -Papilionaceze whose essential organs are confined under 
_ tension and explode (Medicago, Genista, Surothamnus). In these 
_ only one insect’s visit is effective, sometimes under certain 
conditions two (Sarothannus). 
3. Papilionaceze with a piston-mechanism, which squeezes the 
pollen in small quantities out of the apex of the carina, and not 
_ only permits but requires numerous insect-visits. (Lotus, Anthyllis, 
— Ononis, Lupinus). 
4, Papilionaceze with a brush of hairs upon the style which 
"sweeps the pollen in small portions out of the apex of the carina. 
y ' They for the most part require repeated insect-visits. (Lathyrus, 
_ Pisum, Vicia, Phaseolus). 
In all these groups, the stigma and the pollen are applied to 
the under side of the bee. ‘The pollen can therefore as a rule be 
_ collected quickest and most conveniently by bees with abdominal 
\ _ brushes ; and so we find Lotus, Ononis, and Genista tinctoria visited 
especially by these forms. In Sarothaimnus both the upper and 
lower surfaces of the bee are dusted with pollen and come in 
contact with the stigma. 
In those Papilionaceze whose reproductive organs either simply 
' emerge or spring out with an explosion, cross-fertilisation is ensured 
_ bythe stigma projecting beyond the anthers, and coming first in con- 
tact with the bee. In those forms in which the pollen is squeezed 
' or swept out bit by bit, the stigma is at first coated with its own 
_ pollen, which has probably no action upon the stigma and is 
_ rubbed away by the first visitors; and the stigma only becomes 
adhesive and so capable of fertilisation after its papillae have been 
exposed to friction. In absence of insects, self-fertilisation seems 
take place on a large scale in very few Papilionacesze (Piswm) ; 
} ™ several it occurs to a small extent (Z'rifolium repens, Vicia 
. and in many it never occurs (Phaseolus, Onobrychis, 
) Sarothamnus). In cases where self-fertilisation is impossible in 
the ordinary flowers, cleistogamic flowers which regularly fertilise 
| themselves probably compensate. 
