The most important of all the experiments on Mimulus are those in which flowers on plants of the eighth self-fertilised generation were again self-fertilised; other flowers on distinct plants of the same lot were intercrossed; and others were crossed with a new stock of plants from Chelsea.  The Chelsea-crossed seedlings were to the intercrossed in height as 100 to 56, and in fertility as 100 to 4; and they were to the self-fertilised plants, in height as 100 to 52, and in fertility as 100 to 3.  These Chelsea-crossed plants were also much more hardy than the plants of the other two lots; so that altogether the gain from the cross with a fresh stock was wonderfully great.

Lastly, seedlings raised from a cross between flowers on the same plant were not superior to those from flowers fertilised with their own pollen; but this result cannot be absolutely trusted, owing to some previous observations, which, however, were made under very unfavourable circumstances.

[Digitalis purpurea.

The flowers of the common Foxglove are proterandrous; that is, the pollen is mature and mostly shed before the stigma of the same flower is ready for fertilisation.  This is effected by the larger humble-bees, which, whilst in search of nectar, carry pollen from flower to flower.  The two upper and longer stamens shed their pollen before the two lower and shorter ones.  The meaning of this fact probably is, as Dr. Ogle remarks, that the anthers of the longer stamens stand near to the stigma, so that they would be the most likely to fertilise it (3/3.  ‘Popular Science Review’ January 1870 page 50.); and as it is an advantage to avoid self-fertilisation, they shed their pollen first, thus lessening the chance.  There is, however, but little danger of self-fertilisation until the bifid stigma opens; for Hildebrand found that pollen placed on the stigma before it had opened produced no effect. (3/4.  ‘Geschlechter-Vertheilung bei den Pflanzen’ 1867 page 20.) The anthers, which are large, stand at first transversely with respect to the tubular corolla, and if they were to dehisce in this position they would, as Dr. Ogle also remarks, smear with pollen the whole back and sides of an entering humble-bee in a useless manner; but the anthers twist round and place themselves longitudinally before they dehisce.  The lower and inner side of the mouth of the corolla is thickly clothed with hairs, and these collect so much of the fallen pollen that I have seen the under surface of a humble-bee thickly dusted with it; but this can never be applied to the stigma, as the bees in retreating do not turn their under surfaces upwards.  I was therefore puzzled whether these hairs were of any use; but Mr. Belt has, I think, explained their use:  the smaller kinds of bees are not fitted to fertilise the flowers, and if they were allowed to enter easily they would steal much nectar, and fewer large bees would haunt the flowers.  Humble-bees can crawl into the dependent flowers with