The whole case appears to me a very curious one.
It may be observed in Tables 4.23 to 4.25 that some of the illegitimate unions yielded during neither year a single seed; but, judging from the long-styled plants, it is probable, if such unions were to be effected repeatedly by the aid of insects under the most favourable conditions, some few seeds would be produced in every case. Anyhow, it is certain that in all twelve illegitimate unions the pollen-tubes penetrated the stigma in the course of eighteen hours. At first I thought that two kinds of pollen placed together on the same stigma would perhaps yield more seed than one kind by itself; but we have seen that this is not so with each form's own two kinds of pollen; nor is it probable in any case, as I occasionally got, by the use of a single kind of pollen, fully as many seeds as a capsule naturally fertilised ever produces. Moreover the pollen from a single anther is far more than sufficient to fertilise fully a stigma; hence, in this as with so many other plants, more than twelve times as much of each kind of pollen is produced as is necessary to ensure the full fertilisation of each form. From the dusted condition of the bodies of the bees which I caught on the flowers, it is probable that pollen of various kinds is often deposited on all three stigmas; but from the facts already given with respect to the two forms of Primula, there can hardly be a doubt that pollen from the stamens of corresponding length placed on a stigma would be prepotent over any other kind of pollen and obliterate its effects,--even if the latter had been placed on the stigma some hours previously.
Finally, it has now been shown that Lythrum salicaria presents the extraordinary case of the same species bearing three females, different in structure and function, and three or even five sets (if minor differences are considered) of males; each set consisting of half-a-dozen, which likewise differ from one another in structure and function.
[Lythrum Graefferi.
I have examined numerous dried flowers of this species, each from a separate plant, sent me from Kew. Like L. salicaria, it is trimorphic, and the three forms apparently occur in about equal numbers. In the long-styled form the pistil projects about one-third of the length of the calyx beyond its mouth, and is therefore relatively much shorter than in L. salicaria; the globose and hirsute stigma is larger than that of the other two forms; the six mid-length stamens, which are graduated in length, have their anthers standing close above and close beneath the mouth of the calyx; the six shortest stamens rise rather above the middle of the calyx. In the mid-styled form the stigma projects just above the mouth of the calyx, and stands almost on a level with the mid-length stamens of the long and short-styled forms; its own longest stamens project well above the mouth of the calyx, and stand a little above the level of the stigma of the long-styled form. In short, without entering on further details, there is a close general correspondence in structure between this species and L. salicaria, but with some differences in the proportional lengths of the parts. The fact of each of the three pistils having two sets of stamens of corresponding lengths, borne by the two other forms, comes out conspicuously. In the mid-styled form the pollen-grains from the longest stamens are nearly double the diameter of those from the shortest stamens; so that there is a greater difference in this respect than in L. salicaria. In the long-styled form, also, the difference in diameter between the pollen-grains of the mid-length and shortest stamens is greater than in L. salicaria. These comparisons, however, must be received with caution, as they were made on specimens soaked in water after having been long kept dry.
Lythrum thymifolia.
This form, according to Vaucher, is dimorphic, like Primula, and therefore presents only two forms. (4/8. 'Hist. Phys. des Plantes d'Europe' tome 2 1841 pages 369, 371.) I received two dried flowers from Kew, which consisted of the two forms; in one the stigma projected far beyond the calyx, in the other it was included within the calyx; in this latter form the style was only one-fourth of the length of that in the other form.