PART II. } THE MECHANISMS OF FLOWERS. 177 
f 
carina, as in Sarothamnus and Genista, which prevent the column 
_ from springing up in the young flower, though these margins are 
coherent here also, but we may separate them completely from one | 
another without explosion taking place. The structures which 
: "prevent it are two pairs of projections developed on the petals of 
the carina and ale, which lie upon the column. 
Two are directed forwards, and consist of deep pouches in the 
‘superior basal angles of the carina (7, 4, 5) which meet in the 
middle line above the column. Into these fit still deeper pouched 
“processes of the ale (¢, 2, 5), which both connect the ale firmly 
( eh the carina, and also help to strengthen the hold upon the 
ome. These anterior pouches of the ale (e, 2,5) are only a 
urther development: of those shallow depressions by means of which 
Bie alee are united to the carina in Trifolium and Melilotus. 
| Besides this anterior pouch, each ala possesses at the base of its 
upper border a long finger-shaped process, which is directed back- 
wards, and is a further development of that which we shall see 
“appearing in Melilotus. The two processes (g, 2, Fig. 56) curve 
upwards and inwards so that their ends almost meet in the middle 
line above the staminal column. 
___ If in a young flower we cut through the claw of the carina, the 
_ column springs upwards to some extent, carrying with it the carina 
and ale. If in another unexploded flower we carefully cut through 
i one of the digitiform processes of the ale, the parts remain motion- 
less; but on cutting the process of the other side explosion at 
1 Bice follows. The pouched processes of the carina (7) are thus 
' not sufficient to hold the column down without the aid of the 
| processes (g) of the alse, nor are the latter sufficient after section 
of the carina. Explosion can therefore be effected equally well by 
| separating the anterior pouches (e, 2), by separating the digitiform 
_ processes (g, 2), or, finally, by depressing the alz and carina. 
If an insect inserts its proboscis in the middle line between the 
| anterior pouches and the digitiform processes, or if it stands upon 
the alze and thrusts its head in the middle line under the vex- 
| illum, in either case explosion follows. The stigma (0, 5) projects 
beyond the anthers, and, therefore, is the first to strike the under 
) surface of the bee’s body or proboscis; an instant later the anthers _ 
' come in contact with an area close around the spot that the stigma 
| touched, dusting it with fresh pollen. The first flower that the 
i asect visits is, of course, not cross-fertilised, but as the bee with- 
draws from the flower self-fertilisation inevitably occurs. Self- 
rtilisation is undoubtedly efficient, for Hildebrand has shown that 
N 
