Researches in Vegetable Physiology. 235 



the nectar, touch the back of the keel ; the latter throws itself 

 briskly backward ; the insect receives a few grains of pollen, 

 and transports them to the neighbouring flower. Without 

 this intervention, often not a single seed is produced. 



The family Fumariaceee, lately studied by M. Hildebrandt*, 

 presents us with a perfectly analogous example. The stamens 

 and the pistil are narrowly enclosed between the two petals, 

 and appear to be removed from all exterior action. But the 

 base of the petals, which is produced into a spur, offers an 

 abundant provision of nectar. To reach it the insect must 

 pass between the two petals, the upper part of which, borne 

 upon a sort of hinge, separates easily. It thus loads itself 

 with pollen. 



Lastly, some flowers are polymorphic. By this name we 

 designate the species in which the stigma and the stamens, 

 which are placed at different heights in the corolla, do not 

 always occupy the same relative positions. In some individuals 

 the stigma, borne upon a long pistil, passes the corolla more 

 or less, whilst tlie stamens remain very short ; in others the 

 stamens advance and the pistil remains short. 



Mr. Darwin t was the first to study this peculiarity in 

 Primula and Lythrum. M. Hildebrandt has since observed a 

 great number of polymorphic flowers. Better than any others 

 they show the necessity of crossings. In fact a pistil is fer- 

 tilized only by the stamens which are developed at the same 

 height with itself relatively to the corolla, and consequently of 

 necessity in another flower. Numbers in connexion with this 

 subject are more eloquent than any thing else. In experiment- 

 ing upon a trimorphic Oxalis, M. Hildebrandt J obtained the 

 following results : — 



28 flowers with long styles, fecundated with pollen from 

 flowers with long stamens, produced 28 capsules, each contain- 

 ing on an average 11 '9 fertile seeds. 



23 flowers with long styles, fecundated with pollen from 

 median stamens, produced 2 capsules, which, together, only 

 furnished a single seed. 



14 flowers with long styles, fecundated with pollen from 

 short stamens, produced no capsule at all. 



38 flowers with median styles, fecundated with pollen from 

 median stamens, produced 38 capsules, containing on an ave- 

 rage 11*3 seeds. 



* " Bestaubungsvorrichtungen bei den Fumariaceen,"' Pringsheim's 

 Jalirb. vii. p. 423. 



t " Dimorphism in Primulu" Linn. Soc. Joum. vi. 1862. " H^t^ro- 

 moi-pbisme et aes consequences," Ann. Sci. Nat. 1863, tome xix. 



X Bot. Zeitung, 1871, Nos, 26 and 27. 



