In the closely allied genus, Campanula, in which Specularia was formerly included, the anthers shed at an early period their pollen, and this adheres to the collecting hairs which surround the pistil beneath the stigma; so that without some mechanical aid the flowers cannot be fertilised. For instance, I covered up a plant of Campanula carpathica, and it did not produce a single capsule, whilst the surrounding uncovered plants seeded profusely. On the other hand, the present species of Specularia appears to set almost as many capsules when covered up, as when left to the visits of the Diptera, which, as far as I have seen, are the only insects that frequent the flowers. (5/20. It has long been known that another species of the genus, Specularia perfoliata, produces cleistogene as well as perfect flowers, and the former are of course self-fertile.) I did not ascertain whether the naturally crossed and spontaneously self-fertilised capsules contained an equal number of seeds, but a comparison of artificially crossed and self-fertilised flowers, showed that the former were probably the most productive. It appears that this plant is capable of producing a large number of self-fertilised capsules owing to the petals closing at night, as well as during cold weather. In the act of closing, the margins of the petals become reflexed, and their inwardly projecting midribs then pass between the clefts of the stigma, and in doing so push the pollen from the outside of the pistil on to the stigmatic surfaces. (5/21. Mr. Meehan has lately shown ‘Proceedings of the Academy of Natural Science Philadelphia’ May 16, 1876 page 84, that the closing of the flowers of Claytonia virginica and Ranunculus bulbosus during the night causes their self-fertilisation.) Twenty flowers were fertilised by me with their own pollen, but owing to the bad season, only six capsules were produced; they contained on an average 21.7 seeds, with a maximum of forty-eight in one. Fourteen flowers were crossed with pollen from another plant, and these produced twelve capsules, containing on an average 30 seeds, with a maximum in one of fifty-seven seeds; so that the crossed seeds were to the self-fertilised from an equal number of capsules as 100 to 72. The former were also heavier than an equal number of self-fertilised seeds, in the ratio of 100 to 86. Thus, whether we judge by the number of capsules produced from an equal number of flowers, or by the average number of the contained seeds, or the maximum number in any one capsule, or by their weight, crossing does great good in comparison with self-fertilisation. The two lots of seeds were sown on the opposite sides of four pots; but the seedlings were not sufficiently thinned. Only the tallest plant on each side was measured, when fully grown. The measurements are given in Table 5/67. In all four pots the crossed plants flowered first. When the seedlings were only about an inch and a half in height both lots were equal.