PART ITI. | THE MECHANISMS OF FLOWERS. 553 
sat upon one flower while sucking another, cross-fertilisation 
(owing to the prominent situation of the stigma) was often effected 
by pollen adhering to the under surface of their bodies. In absence 
of insects, self-fertilisation can only take place in flowers directed 
more or less downwards. 
Visitors : A. Hymenoptera—Apide : (1) Apis mellifica, L. $,s. and f.p. 
B. Diptera—Syrphide : (2) Merodon eneus, Mgn.,s. andf.p. C. Lepidoptera 
—Rhopalocera: (3) Papilio Machaon, L., s.; (4) Meliteea Athalia, Esp., s. 
Visited also by many other insects. See No. 590, 1. 
Anthericum Liliago, L.—The flower resembles that of the 
former species (590, I). 
392. ALLIUM uRSINUM, L.—When the flower first opens the 
style is only from one-third to one-half of its ultimate length ; 
the papille are not yet developed on the stigma, and the 
anthers are still allimmature. First of all the three inner anthers 
dehisce at slow intervals, one after the other; at this stage 
the style has reached three-quarters or more of its full length 
(43 to 5 mm.). Then the three outer anthers dehisce one after 
another; the style has now reached its full length (6. mm.), and 
the papille on the stigma become developed. The flowers are thus 
imperfectly proterandrous. The anthers dehisce introrsely, and 
then turn that side which is covered with pollen more or less 
upwards. In specimens flowering in my reom I found some flowers 
in which the style was so bent during the last stage that the stigma 
touched one of the anthers which still retained some of its pollen. 
In this way self-fertilisation was effected to a limited extent when 
no insect visited the flower. In a wood at Stromberg Schloss- 
berg on the 16th May, 1868, I saw Bombus pratorum, 9, flying 
quickly from flower to flower of A. w7sinum ; she thrust her proboscis 
into each flower in search of honey, and after scarcely two seconds 
hurried on. The honey is secreted by the ovary, in the three 
notches between the carpels, and fills the space between these 
notches and the bases of the three inner stamens, Hence the bee 
when sucking must touch the stigma with one side of its head and 
the anthers with the opposite side, and, in old flowers, can only 
lead to cross-fertilisation. 
393. ALLIUM CEPA, L.—The honey in this species is placed in 
the same position as in the preceding one; the anthers and stigma 
are also similarly situated. 
