This species offers a somewhat perplexing case. In the first generation the self-fertilised plants, though few in number, greatly exceeded the crossed in height; whilst in the second generation, when the trial was made on a much larger scale, the crossed beat the self-fertilised plants. As this species is generally propagated by off-sets, some seedlings were first raised, in order to have distinct plants. On one of these plants several flowers were fertilised with their own pollen; and as the pollen is mature and shed long before the stigma of the same flower is ready for fertilisation, it was necessary to number each flower and keep its pollen in paper with a corresponding number. By this means well-matured pollen was used for self-fertilisation. Several flowers on the same plant were crossed with pollen from a distinct individual, and to obtain this the conjoined anthers of young flowers were roughly squeezed, and as it is naturally protruded very slowly by the growth of the pistil, it is probable that the pollen used by me was hardly mature, certainly less mature than that employed for self-fertilisation. I did not at the time think of this source of error, but I now suspect that the growth of the crossed plants was thus injured. Anyhow the trial was not perfectly fair. Opposed to the belief that the pollen used in crossing was not in so good a state as that used for self-fertilisation, is the fact that a greater proportional number of the crossed than of the self-fertilised flowers produced capsules; but there was no marked difference in the amount of seed contained in the capsules of the two lots. (5/24. Gartner has shown that certain plants of Lobelia fulgens are quite sterile with pollen from the same plant, though this pollen is efficient on any other individual; but none of the plants on which I experimented, which were kept in the greenhouse, were in this peculiar condition.) As the seeds obtained by the above two methods would not germinate when left on bare sand, they were sown on the opposite sides of four pots; but I succeeded in raising only a single pair of seedlings of the same age in each pot. The self-fertilised seedlings, when only a few inches in height, were in most of the pots taller than their opponents; and they flowered so much earlier in all the pots, that the height of the flower-stems could be fairly compared only in Pots 1 and 2. TABLE 5/70. Lobelia fulgens (First Generation). Heights of flower-stems measured in inches. Column 1: Number (Name) of Pot. Column 2: Height of Flower-stems on the Crossed Plants. Column 3: Height of Flower-stems on the Self-fertilised Plants. Pot 1 : 33 : 50. Pot 2 : 36 4/8 : 38 4/8. Pot 3 : 21* : 43. Pot 4 : 12* : 35 6/8. *Not in full flower.