276 



ANGIOSPERMAE—DICOTYLEDONES 



comes into contact with the insect, getting cross-pollinated if this has previously 

 visited another flower of the species. The first flower visited by an insect will be 

 self-pollinated as the visitor backs out of it. Should insect-visits fail, automatic 

 self-pollination of unexploded flowers is possible, and may be effective under certain 

 conditions (vide infra). 



BurkiU (Proc. Phil. Soc, Cambridge, viii, 1894) aptly describes the basal pro- 

 cesses of alae and carina as two triggers, by which the flower is, so to speak, fired 

 off. He states that the upper surfaces of the alae are beset with papillae, serving 

 as foot-holds to insect visitors. There is also a marginal row of papillae on either 

 side of the vexillum's inner surface, to which long-legged insects would appear to 

 cling. The stigma remains unreceptive until its papillae have been subjected 

 to friction. By covering a number of inflorescences with nets, in order to keep 

 away insects, BurkiU was able to confirm the conclusion at which Urban had 



previously arrived (Verb. bot. Ver., 



Berlin, xv, 1873), i.e. that unex- 



1^ I . \\ %.\ ^ th f \ni "^^^ ploded flowers do not set fruits, 



although their stigmas are sur- 

 rounded by pollen. He succeeded, 

 however, in inducing such flowers 

 to fruit by ( I ) squeezing the stigma 

 through the tip of the carina, (2) 

 piercing the carina with a needle 

 and scratching the stigma, (3) 

 cutting off the tip of the carina and 

 rubbing the stigma with a paint- 

 brush. 



Visitors. — These are bees 

 and Lepidoptera. The slender 

 proboscis of the latter is undoubted- 

 ly able to bring about explosion, 

 providing it is thrust down to the 

 nectar from the middle of the flower, 

 while, on the other hand, explosion 

 is not effected by insects that suck 

 laterally, as e.g. the honey-bee, which inserts its proboscis into the base of the 

 flower from the side of the vexillum. 



Herm. MuUer (H. M.), Buddeberg (Budd.), and myself (Kn.) have observed the 

 following. — 



A. Hymenoptera. {a) Apidae: i. Apis mellifica L. $, very freq., skg. (H. M., 

 Kn.); 2. Bombus agrorum L. 5 and 5, skg. (H. M., Kn.); 3. B. terrester Z., do. 

 (Kn.^; 4. Cihssa leporina Pz. S, do. (H. M.) ; 5. Coelioxys rufescens Lep. J, do. 

 (Kn.); 6. CoUetes sp. S, do. (H. M.); 7. Halictus morio i^. 5, do. (Budd.); 8. Mega- 

 chile argentata F. 5 and S, do. (H. M., Budd.); 9. M. pyrina Lep., do. (H. M.); 

 10. M. willughbiella K. S, do. (H. M.); 1 1. Osmia aenea L. 5, freq., skg. and po-cltg'. 

 (H. M.); 12. O. rufa L. ?, skg. (H. M.); 13. Rhophites canus Ev. S, do. (H. M.); 

 14. Xylocopa violacea L. $, do. (H. M.). {b) Sphegidae: 15. Bembex rostrata Z., 

 skg. (H. M.). B. Lepidoptera. (a) Kociuidae: 16. Plusia gamma Z. {b) Rho- 



FlG. 89. Midicago saliva. L. (after Herm. Miiller). (i) 

 Unvisited flower, seen from below. {3) The same, after re- 

 moval of the vexillum and upper half of the calyx ; seen from 

 above. (3) Right ala seen from inside. (4) Carina, seen 

 from above and from the right. (5) Exploded flower, seen 

 from above and from the left, after removal of the vexillum 

 and upper half of the calyx ( x 3i). a, calyx ; b, vexillum ; c. ala ; 

 d. alar claw ; <r, anterior alar process ; f. cavity of this process ; 

 g. posterior fing^er-shaped alar process; /;, carina; i', carinal 

 process into which the anterior alar process fits ; k, the cohering; 

 filaments ; /, the free uppermost filament ; m. nectar-passages ; 

 n. anther ; o. stigma. 



