flowers were marked and allowed to fertilise themselves spontaneously under a net, and they yielded eight capsules, containing on an average 2.38 seeds, with a maximum of three seeds in one.  So that the crossed and self-fertilised capsules from the perfect flowers yielded seeds in the proportion of 100 to 65.  Fifty-three capsules produced by the cleistogene flowers contained on an average 4.1 seeds, so that these were the most productive of all; and the seeds themselves looked finer even than those from the crossed perfect flowers.

The seeds from the crossed perfect flowers and from the self-fertilised cleistogene flowers were allowed to germinate on sand; but unfortunately only two pairs germinated at the same time.  These were planted on the opposite sides of the same pot, which was kept in the greenhouse.  In the summer of the same year, when the seedlings were about 4 1/2 inches in height, the two lots were equal.  In the autumn of the following year (1868) the two crossed plants were of exactly the same height, namely, 11 4/8 inches, and the two self-fertilised plants 12 6/8 and 7 2/8 inches; so that one of the self-fertilised exceeded considerably in height all the others.  By the autumn of 1869 the two crossed plants had acquired the supremacy; their height being 16 4/8 and 15 1/8, whilst that of the two self-fertilised plants was 14 5/8 and 11 4/8 inches.

By the autumn of 1870, the heights were as follows:—­

Table 5/61.  Ononis minutissima.

Heights of plants measured in inches.

Column 1:  Crossed Plants.

Column 2:  Self-fertilised Plants.

       :  20 3/8 :  17 4/8.
       :  19 2/8 :  17 2/8.

Total :  39.63 :  34.75.

So that the mean height of the two crossed plants was 19.81, and that of the two self-fertilised 17.37 inches; or as 100 to 88.  It should be remembered that the two lots were at first equal in height; that one of the self-fertilised plants then had the advantage, the two crossed plants being at last victorious.]

SUMMARY ON THE LEGUMINOSAE.

Six genera in this family were experimented on, and the results are in some respects remarkable.  The crossed plants of the two species of Lupinus were conspicuously superior to the self-fertilised plants in height and fertility; and when grown under very unfavourable conditions, in vigour.  The scarlet-runner (Phaseolus multiflorus) is partially sterile if the visits of bees are prevented, and there is reason to believe that varieties growing near one another intercross.  The five crossed plants, however, exceeded in height the five self-fertilised only by a little.  Phaseolus vulgaris is perfectly self-sterile; nevertheless, varieties growing in the same garden sometimes intercross largely.  The varieties of Lathyrus odoratus, on the other hand, appear never to intercross in this country; and though the flowers are not often visited