the greatest ease, using the “hairs as footholds while sucking the honey; but the smaller bees are impeded by them, and when, having at length struggled through them, they reach the slippery precipice above, they are completely baffled.”  Mr. Belt says that he watched many flowers during a whole season in North Wales, and “only once saw a small bee reach the nectary, though many were seen trying in vain to do so.” (3/5.  ‘The Naturalist in Nicaragua’ 1874 page 132.  But it appears from H. Muller ‘Die Befruchtung der Blumen’ 1873 page 285, that small insects sometimes succeed in entering the flowers.)

I covered a plant growing in its native soil in North Wales with a net, and fertilised six flowers each with its own pollen, and six others with pollen from a distinct plant growing within the distance of a few feet.  The covered plant was occasionally shaken with violence, so as to imitate the effects of a gale of wind, and thus to facilitate as far as possible self-fertilisation.  It bore ninety-two flowers (besides the dozen artificially fertilised), and of these only twenty-four produced capsules; whereas almost all the flowers on the surrounding uncovered plants were fruitful.  Of the twenty-four spontaneously self-fertilised capsules, only two contained their full complement of seed; six contained a moderate supply; and the remaining sixteen extremely few seeds.  A little pollen adhering to the anthers after they had dehisced, and accidentally falling on the stigma when mature, must have been the means by which the above twenty-four flowers were partially self-fertilised; for the margins of the corolla in withering do not curl inwards, nor do the flowers in dropping off turn round on their axes, so as to bring the pollen-covered hairs, with which the lower surface is clothed, into contact with the stigma—­by either of which means self-fertilisation might be effected.

Seeds from the above crossed and self-fertilised capsules, after germinating on bare sand, were planted in pairs on the opposite sides of five moderately-sized pots, which were kept in the greenhouse.  The plants after a time appeared starved, and were therefore, without being disturbed, turned out of their pots, and planted in the open ground in two close parallel rows.  They were thus subjected to tolerably severe competition with one another; but not nearly so severe as if they had been left in the pots.  At the time when they were turned out, their leaves were between 5 and 8 inches in length, and the longest leaf on the finest plant on each side of each pot was measured, with the result that the leaves of the crossed plants exceeded, on an average, those of the self-fertilised plants by .4 of an inch.

In the following summer the tallest flower-stem on each plant, when fully grown, was measured.  There were seventeen crossed plants; but one did not produce a flower-stem.  There were also, originally, seventeen self-fertilised plants, but these had such poor constitutions that no less than nine died in the course of the winter and spring, leaving only eight to be measured, as in Table 3/23.