This extraordinary diversity in the means for favouring or preventing cross- and self-fertilisation in closely allied forms, probably depends on the results of both processes being highly beneficial to the species, but directly opposed in many ways to one another and dependent on variable conditions.  Self-fertilisation assures the production of a large supply of seeds; and the necessity or advantage of this will be determined by the average length of life of the plant, which largely depends on the amount of destruction suffered by the seeds and seedlings.  This destruction follows from the most various and variable causes, such as the presence of animals of several kinds, and the growth of surrounding plants.  The possibility of cross-fertilisation depends mainly on the presence and number of certain insects, often of insects belonging to special groups, and on the degree to which they are attracted to the flowers of any particular species in preference to other flowers,—­all circumstances likely to change.  Moreover, the advantages which follow from cross-fertilisation differ much in different plants, so that it is probable that allied plants would often profit in different degrees by cross-fertilisation.  Under these extremely complex and fluctuating conditions, with two somewhat opposed ends to be gained, namely, the safe propagation of the species and the production of cross-fertilised, vigorous offspring, it is not surprising that allied forms should exhibit an extreme diversity in the means which favour either end.  If, as there is reason to suspect, self-fertilisation is in some respects beneficial, although more than counterbalanced by the advantages derived from a cross with a fresh stock, the problem becomes still more complicated.

As I only twice experimented on more than a single species in a genus, I cannot say whether the crossed offspring of the several species within the same genus differ in their degree of superiority over their self-fertilised brethren; but I should expect that this would often prove to be the case from what was observed with the two species of Lobelia and with the individuals of the same species of Nicotiana.  The species belonging to distinct genera in the same family certainly differ in this respect.  The effects of cross- and self-fertilisation may be confined either to the growth or to the fertility of the offspring, but generally extends to both qualities.  There does not seem to exist any close correspondence between the degree to which their offspring profit by this process; but we may easily err on this head, as there are two means for ensuring cross-fertilisation which are not externally perceptible, namely, self-sterility and the prepotent fertilising influence of pollen from another individual.  Lastly, it has been shown in a former chapter that the effect produced by cross and self-fertilisation on the fertility of the parent-plants does not always correspond with that produced on the