But it may be said, granting that changed conditions act on the sexual elements, How can two or more plants growing close together, either in their native country or in a garden, be differently acted on, inasmuch as they appear to be exposed to exactly the same conditions?

EXPERIMENTS IN CROSSING.

In my experiments with Digitalis purPage 447. purea, some flowers on a wild plant were selffertilized, and others were crossed with pollen from another plant growing within two or three feet distance. The crossed and self-fertilized plants raised from the seeds thus obtained produced flower-stems in number as 100 to 47, and in average height as 100 to 70. Therefore, the cross between these two plants was highly beneficial; but how could their sexual elements have been differentiated by exposure to different conditions? If the progenitors of the two plants had lived on the same spot during the last score of generations, and had never been crossed with any plant beyond the distance of a few feet, in all probability their offspring would have been reduced to the same state as some of the plants in my experiments -such as the intercrossed plants of the ninth generation of Ipomoea, or the self-fertilized plants of the eighth generation of Mimulus, or the offspring from flowers on the same plant; and in this case a cross between the two plants of Digitalis would have done no good. But seeds are often widely dispersed by natural means, and one of the above two plants, or one of their ancestors, may have come from a distance, from a more shady or sunny, dry or moist place, or from a different kind of soil containing other organic seeds or inorganic matter.

THE STRUGGLE FOR EXISTENCE AMONG SEEDS.

Seeds often lie dormant for several years Page 449. in the ground, and germinate when brought near the surface by any means, as by burrowing animals. They would probably be affected by the mere circumstance of having long lain dormant; for gardeners believe that the production of double flowers, and of fruit, is thus influenced. Seeds, moreover, which were matured during different seasons will have been subjected during the whole course of their development to different degrees of heat and moisture.

It has been shown that pollen is often carried by insects to a considerable distance from plant to plant. Therefore, one of the parents or ancestors of our two plants of Digitalis may have been crossed by a distant plant growing under somewhat different conditions. Plants thus crossed often produce an unusually large number of seeds; a striking instance of this fact is afforded by the Bignonia, which was fertilized by Fritz Müller with pollen from some adjoining plants and set hardly any seed, but, when fertilized with pollen from a distant plant, was highly fertile. Seedlings from a cross of this kind grow with great vigor, and transmit their vigor to their descendants. These, therefore, in the struggle for life, will generally beat and exterminate the seedlings from plants which have long grown near together under the same conditions, and will thus tend to spread.

PRACTICAL APPLICATION OF THESE VIEWS.

Page 458.

Under a practical point of view, agriculturists and horticulturists may learn something from the conclusions at which we have arrived. Firstly,

we see that the injury from the close breeding of animals and from the self-fertilization of plants does not necessarily depend on any tendency to disease or weakness of constitution common to the related parents, and only indirectly on their relationship, in so far as they are apt to resemble each other in all respects, including their sexual nature. And, secondly, that the advantages of crossfertilization depend on the sexual elements of the parents having become in some degree differentiated by the exposure of their progenitors to different conditions, or from their having intercrossed with individuals thus exposed; or, lastly, from what we call in our ignorance spontaneous variation. He therefore who wishes to pair closely related animals ought to keep them under conditions as different as possible.

As some kinds of plants suffer much more Page 459. from self-fertilization than do others, so it probably is with animals from too close interbreeding. The effects of close interbreeding on animals, judging again from plants, would be deterioration in general vigor, including fertility, with no necessary loss of excellence of form; and this seems to be the usual result.

It is a common practice with horticulturists to obtain seeds from another place having a very different soil, so as to avoid raising plants for a long succession of generations under the same conditions; but, with all the species which freely intercross by the aid of insects or the wind, it would be an incomparably better plan to obtain seeds of the required variety, which had been raised for some generations under as different conditions as possible, and sow them in alternate rows with seeds matured in the old garden. The two stocks would then intercross, with a thorough blending of their whole organizations, and with

no loss of purity to the variety; and this would yield far more favorable results than a mere exchange of seeds. We have seen in my experiments how wonderfully the offspring profited in height, weight, hardiness, and fertility, by crosses of this kind. For instance, plants of Ipomoea thus crossed were to the intercrossed plants of the same stock, with which they grew in competition, as 100 to 78 in height, and as 100 to 51 in fertility; and plants of Eschscholtzia similarly compared were as 100 to 45 in fertility. In comparison with self-fertilized plants the results are still more striking; thus cabbages derived from a cross with a fresh stock were to the self-fertilized as 100 to 22 in weight.

Florists may learn, from the four cases which have been fully described, that they have the power of fixing each fleeting variety of color, if they will fertilize the flowers of the desired kind with their own pollen for half a dozen generations, and grow the seedlings under the same conditions. But a cross with any other individual of the same variety must be carefully prevented, as each has its own peculiar constitution. After a dozen generations of self-fertilization, it is probable that the new variety would remain constant even if grown under somewhat different conditions; and there would no longer be any necessity to guard against intercrosses between the individuals of the same variety.

MARRIAGES OF FIRST COUSINS.

With respect to mankind, my son George Page 460. has endeavored to discover by a statistical investigation whether the marriages of first cousins are at all injurious, although this is a degree of relationship which would not be objected to in our domestic animals; and he has come to the conclusion from his own re

searches, and those of Dr. Mitchell, that the evidence as to any evil thus caused is conflicting, but on the whole points to its being very small. From the facts given in this volume we may infer that with mankind the marriages of nearly related persons, some of whose parents and ancestors had lived under very different conditions, would be much less injurious than that of persons who had always lived in the same place and followed the same habits of life. Nor can I see reason to doubt that the widely different habits of life of men and women in civilized nations, especially among the upper classes, would tend to counterbalance any evil from marriages between healthy and somewhat closely related persons.

DEVELOPMENT OF THE TWO SEXES IN PLANTS.

Under a theoretical point of view it is some Page 461. gain to science to know that numberless structures in hermaphrodite plants, and probably in hermaphrodite animals, are special adaptations for securing an occasional cross between two individuals; and that the advantages from such a cross depend altogether on the beings which are united, or their progenitors, having had their sexual elements somewhat differentiated, so that the embyro is benefited in the same manner as is a mature plant or animal by a slight change in its conditions of life, although in a much higher degree.

Another and more important result may be deduced from my observations. Eggs and seeds are highly serviceable as a means of dissemination, but we now know that fertile eggs can be produced without the aid of the male. There are also many other methods by which organisms can be propagated asexually. Why then have the two sexes been developed, and why do males exist