340 ANGIOSPERMAE—DICOTYLEDONES 



forwards. Insects can only obtain nectar legitimately by alighting upon the left ala 

 and pushing their proboscis under the opening on the right side of the carinal tip. 

 To prevent access in any other way, there is here an upwardly and forwardly directed 

 scale- like appendage. 



Only large humble-bees are able to work the flower mechanism. When the 

 carina is depressed the end of the style with its pollen-covered brush springs out of 

 the opening in the carina, and a narrow canal appears, beginning just below the 

 carinal opening, and running past the end of the style, along the right margin of 

 the groove in the staminal column as far as the base of the nectary. The upper 

 free stamen retains its position, while the nine united ones are bent downwards. 

 As the stigma is touched by the proboscis of an insect visitor before the pollen, it 

 follows that cross-pollination is regularly effected by visits. Automatic self-pollination 

 is excluded : unvisited flowers remain infertile, as already stated. 



Visitors. — I saw, at Kiel, Bombus hortorum L. 5, skg. legitimately. In spite 

 of constant watching, I have but seldom observed insect-visits, and automatic self- 

 pollination takes place in the large majority of cases, though Darwin was of the 

 contrary opinion {vide supra). Some humble-bees obtain the nectar by perforation : 

 I observed Bombus terrester L. as a nectar-thief at Kiel. 



Alfken saw a bee (Megachile maritima K. 5) at Bremen, and another (Osmia 

 maritima Friese 5) in Juist. Leege observed (Juist) a bee (Osmia maritima Friese 5), 

 freq., skg. and po-cltg., a hawk-moth (Deilephila galii Roll.), and a Noctuid (Chari- 

 clea umbra H/n.) very freq. 



810. P. multiflorus Willd. (Herm. Rliiller, 'Fertilisation,' p. 217; Kirchner, 

 ' Flora v. Stuttgart,' pp. 575-6 ; Knuth, ' BloemenbioL Bijdragen.') — The flower 

 mechanism of this species agrees completely with that of P. vulgaris, and was first 

 described by Farrer (Ann. Mag. Nat. Hist., London, Ser. 4, ii, 1868, pp. 256-60). 



According to Ogle (Pop. Sci. Rev., London, ix, 1870, p. 166), the flowers are 

 infertile when bees are excluded, but Kirchner says that they can produce perfect 

 fruits by self-fertilization. The honey-bee and other small bees which are too feeble 

 to depress the carina, use the holes made by Bombus terrester Z. in the calyx for 

 stealing nectar. Hermann Miiller states that more powerful bees, with a sufliciently 

 long proboscis, alight upon the left ala and first touch the stigma with the base of 

 their proboscis, when they thrust it into the base of the flower, and thus effect cross- 

 pollination as in P. vulgaris. As the interlocked alae and carina are pressed down- 

 wards, the spirally twisted end of the style protrudes from the tip of the carina, which 

 is also coiled into a spiral of two turns. The stigma faces downwards and to the left, 

 while the pollen-covered stylar brush rubs against the base of the bee's proboscis, 

 and dusts it afresh. When insects visit the flowers cross-pollination is therefore 

 assured, and self-pollination prevented. Hermann Miiller states that automatic 

 self-pollination is impossible, as the stigma projects from the tip of the carina, while 

 the pollen is enclosed within it. 



Visitors. — I several times observed (at Kiel) Bombus hortorum L. 5 as an 

 invited guest. Schletterer records Eucera longicornis L. for the Tyrol. 



