fertility even increased in the later generations, as if they were becoming habituated to illegitimate fertilisation.  Plants of the first illegitimate generation when taken from the greenhouse and planted in moderately good soil out of doors grew well and were healthy; but when those of the two last illegitimate generations were thus treated they became excessively sterile and dwarfed, and remained so during the following year, by which time they ought to have become accustomed to growing out of doors, so that they must have possessed a weak constitution.

Under these circumstances, it seemed advisable to ascertain what would be the effect of legitimately crossing long-styled plants of the fourth illegitimate generation with pollen taken from non-related short-styled plants, growing under different conditions.  Accordingly several flowers on plants of the fourth illegitimate generation (i.e., great-great-grandchildren of plants which had been legitimately fertilised), growing vigorously in pots in the greenhouse, were legitimately fertilised with pollen from an almost wild short-styled cowslip, and these flowers yielded some fine capsules.  Thirty other flowers on the same illegitimate plants were fertilised with their own pollen, and these yielded seventeen capsules, containing on an average thirty-two seeds.  This is a high degree of fertility; higher, I believe, than that which generally obtains with illegitimately fertilised long-styled plants growing out of doors, and higher than that of the previous illegitimate generations, although their flowers were fertilised with pollen taken from a distinct plant of the same form.

These two lots of seeds were sown (for they will not germinate well when placed on bare sand) on the opposite sides of four pots, and the seedlings were thinned, so that an equal number were left on the two sides.  For some time there was no marked difference in height between the two lots; and in Pot 3, Table 6/93, the self-fertilised plants were rather the tallest.  But by the time that they had thrown up young flower-stems, the legitimately crossed plants revealed much the finest, and had greener and larger leaves.  The breadth of the largest leaf on each plant was measured, and those on the crossed plants were on an average a quarter of an inch (exactly .28 of an inch) broader than those on the self-fertilised plants.  The plants, from being too much crowded, produced poor and short flower-stems.  The two finest on each side were measured; the eight on the legitimately crossed plants averaged 4.08, and the eight on the illegitimately self-fertilised plants averaged 2.93 inches in height; or as 100 to 72.

These plants after they had flowered were turned out of their pots, and planted in fairly good soil in the open ground.  In the following year (1870), when in full flower, the two tallest flower-stems on each side were again measured, as shown in Table 6/93, which likewise gives the number of flower-stems produced on both sides of all the pots.