height, vigour, and fertility of their offspring.  The same remark applies to crossed and self-fertilised seedlings when these are used as the parent-plants.  This want of correspondence probably depends, at least in part, on the number of seeds produced being chiefly determined by the number of the pollen-tubes which reach the ovules, and this will be governed by the reaction between the pollen and the stigmatic secretion or tissues; whereas the growth and constitutional vigour of the offspring will be chiefly determined, not only by the number of pollen-tubes reaching the ovules, but by the nature of the reaction between the contents of the pollen-grains and ovules.

There are two other important conclusions which may be deduced from my observations:  firstly, that the advantages of cross-fertilisation do not follow from some mysterious virtue in the mere union of two distinct individuals, but from such individuals having been subjected during previous generations to different conditions, or to their having varied in a manner commonly called spontaneous, so that in either case their sexual elements have been in some degree differentiated.  And secondly, that the injury from self-fertilisation follows from the want of such differentiation in the sexual elements.  These two propositions are fully established by my experiments.  Thus, when plants of the Ipomoea and of the Mimulus, which had been self-fertilised for the seven previous generations and had been kept all the time under the same conditions, were intercrossed one with another, the offspring did not profit in the least by the cross.  Mimulus offers another instructive case, showing that the benefit of a cross depends on the previous treatment of the progenitors:  plants which had been self-fertilised for the eight previous generations were crossed with plants which had been intercrossed for the same number of generations, all having been kept under the same conditions as far as possible; seedlings from this cross were grown in competition with others derived from the same self-fertilised mother-plant crossed by a fresh stock; and the latter seedlings were to the former in height as 100 to 52, and in fertility as 100 to 4.  An exactly parallel experiment was tried on Dianthus, with this difference, that the plants had been self-fertilised only for the three previous generations, and the result was similar though not so strongly marked.  The foregoing two cases of the offspring of Ipomoea and Eschscholtzia, derived from a cross with a fresh stock, being as much superior to the intercrossed plants of the old stock, as these latter were to the self-fertilised offspring, strongly supports the same conclusion.  A cross with a fresh stock or with another variety seems to be always highly beneficial, whether or not the mother-plants have been intercrossed or self-fertilised for several previous generations.  The fact that a cross between two flowers on the same plant does no good or very