106 ELEMENTARY STUDIES IN" INSECT LIFE 



ing with it the carina and altv. If in another unexploded 

 flower we carefully cut through one of the digitiform 

 processes of the ala 1 , the parts remain motionless ; but on 

 cutting the processes of the other side, explosion at once 

 follows. The pouched processes of the carina (Fig. 

 85, 1)3 and cS} are thus sufficient to hold the column 

 down without the aid of the processes of the alir (c.J) ; 

 the alii' alone are not sufficient to hold the column down 

 when the carina has been cut. Explosion can therefore 

 be effected equally well by separating the anterior 

 pouches, by separating the digitiform processes, or, 

 filially, by depressing the a la? 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 ala? and thrusts 

 its head in the middle line 



FIG. 86. Pollination of alfalfa flower lm dei' the Vexilluill, ill 

 by bee. See fig. 64 for honey-collecting 



tools of bee. either case explosion fol- 



lows. The stigma (c5) 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, dusting it with fresh pollen. 

 The first flower that the insect visits is, of course, not 

 cross-fertilized, but as the bee withdraws from the flower, 

 self-fertilization inevitably occurs. Self-fertilization is 

 undoubtedly efficient, for Hildebrand has shown that 

 flowers which wither unexploded when insects are ex- 

 cluded produce seed by self-fertilization. The same au- 



