46 ANGIOSPERMAEDICOTYLEDONES 



erect themselves in that part of the hollow cone lying immediately behind its entrance. 

 Here they inevitably come into contact with the under-surface of the proboscis or 

 head of a humble-bee engaged in sucking", so that cross-pollination is effected. 



The stamens in their first immature condition are directed downwards, and 

 become erect as their anthers dehisce, thus putting themselves in the way of the 

 head of any humble-bee that may be sucking nectar. They bend down again after 

 withering, so as to make room for the styles, which are now turning up, and carrying 

 with them the newly-matured stigmas. Automatic self-pollination is thus excluded, 

 and cross-pollination is necessary for fertilization. Darwin states that artificial 

 self-pollination is ineffective. 



Visitors. Only humble-bees with a proboscis 19-22 mm. long can suck out 

 all the nectar in the legitimate way as described. Of the bees native to Central and 

 North Germany, there are but two with such a proboscis, i. e. Anthophora pilipes F. 

 (= Podalirius acervorum Z.) and Bombus hortorum Z. The former of these, 

 however, is no longer on the wing when Delphinium elatum comes into flower, 

 so that the garden humble-bee alone remains as the regular pollinator of this 

 larkspur. As a matter of fact, Hermann Muller and myself making observations 

 in the Lippstadt and Kiel Gardens respectively have seen hardly any other insect 

 visitor to Delphinium elatum sucking nectar in the legitimate way, and consequently 

 effecting cross-pollination. Some of our other native humble-bees, however, possess 

 a proboscis long enough to reach at least some of the nectar, e. g. B. agrorum F. 

 (proboscis 10-15 mm - l n g)> anc ^ B. senilis Sm. (proboscis 14-15 mm. long). 

 Hermann Muller also observed Anthophora personata III. 5, visiting the flowers 

 at Strassburg. In the Riesengebirge, Schulz noticed holes in the spurs that had 

 no doubt been made by thieving short-tongued bees. 



97. D. Staphisagria L. Hildebrand states that this agrees in the main with 

 the last species as regards both structure and visitors. 



98. D. Consolida L. (Herm. Muller, 'Fertilisation,' pp. 85-6, ' Weit. Beob.,' 

 I, p. 322; Kirchner, 'Flora v. Stuttgart,' p. 274; Schulz, ' Beitrage,' II, p. 204; 

 Knuth, ' Bliitenbiol. Herbstbeob.,' ' Weit. Beob.,' p. 231.) As first explained by Her- 

 mann Muller, the flowers of this species chiefly differ in structure from those of the 

 preceding in the union of the four petals. The posterior prolongations of the upper 

 petals are fused into a single spur 15 mm. long, which secretes and conceals nectar 

 in its end. The anterior lobes of the four petals form in addition a tube 7 mm. wide 

 at its commencement, which readily admits the head of a humble-bee. As it is only 

 open below, the pollen comes into contact with the under-side of a humble-bee's 

 head at the beginning of anthesis, while the stigmas do so later, for stamens and 

 carpels develop in the same order as in the case of D. elatum. Cross-pollination 

 is consequently secured in D. consolida when its nectar is sucked legitimately. 

 Automatic self-pollination is excluded : artificial self-pollination is only moderately 

 effective. 



A proboscis 15 mm. long would be able to reach the tip of the spur, which 

 is 15 mm. in length exclusive of the entrance, that measures 7 mm. 



Visitors. Among the humble-bees native to Central and North Germany, there 

 are once more neglecting Anthophora pilipes F. with a proboscis 1 9-2 1 mm. long, 





