My Dear Darwin

It was through pure forgetfulness at the moment of writing my letter last August, that I did not endeavor to supply to you such instances as I could recollect (for I have no memoranda on the subject) of what D^r Torrey and I long ago (in Fl. N. Amer. 2, p. 38, 39) called diœcio-dimorphism.¹

In Rubiaceæ I think this dimorphism was first somewhere mentioned by D. Don, but I have no reference.² We are most familiar with it in Houstonia (a genus which I have proposed to restore, in which it occurs in all our N. American species, and, as you will see in Wight & Arnott, Prodr. Fl. Pen. Ind. 1. p. 408, in some Indian ones.³ In the living plants I have formerly noticed it in H. purpurea, and it is every year noticed familiarly in our most abundant H. cærulea. Of this I shall try to send you some tufts—seedlings of this year—which you can blossom in early spring, or in your house in winte⁠⟨⁠r.⁠⟩⁠ We must run the chance of your getting ⁠⟨⁠both⁠⟩⁠ subsexes. Wight & Arnott (see p. 439) ob⁠⟨⁠served⁠⟩⁠ it in Knoxia. Another familiar case is furnished by our pretty little Mitchella repens; this is recorded by D^r. Torrey in Flora of State of New York, 1. p. 315.⁴ This would be a good plant for you to experiment upon, if I had only thought of marking plants the two sorts.

Diodia Virginica (and I suppose D. teres) also shows it.

The dimorphism may perhaps occur in many Rubiaceæ, but I have reason to think it that the far greater part show no such tendency. And included or exsert stamens are pretty good and constant generic characters in much of the order⁵   Mitchella is an interesting instance for you from its relationship (with Nertera) to Coprosma, one of the few diœcious genera of Rubiaceæ, and in which the stamens are long-exserted in the male flowers, the styles or stigmas, in the female.

In these and in all genuine cases of the kind, the two sorts of flowers are always borne on different roots.

Upon consideration, I can say nothing of Labiatæ or Borragineæ from my own observation. But I believe that D^r. Torrey can tell us about the latter. ⁠⟨⁠I⁠⟩⁠ have just written to ask him.⁶

⁠⟨⁠In⁠⟩⁠ Labiatæ, what dimorphism I have observed—and without particular examination—is merely such as in Nepeta Glechoma, where the stamens of the earlier flowers appear to abort. I know not whether they do, or whether it is a case of precocious fertilization, such as I will presently call your attention to: but as the flowers are otherwise normal, I suppose it is not a case of this.

A case in point, which will interest you, occurs in Rhamnus—a genus which has both truly hermaphrodite and polygamo-diœcious species.

Rhamnus lanceolatus, of our Southern & Middle & Western States, bears, on different trees, two kinds of flowers which differ in the pistil only;—i.e. as far as appeared to ordinary investigation the stamens are the same in both; but in one the style is short and included; in the other long and exserted. In the latter the flowers are subsolitary in the axils and fruitful; in the former the flowers are more numerous and clustered, and not so fruitful. Yet they do mature some fruit. Here you have an initial state of dimorphism, apparently affecting only the female organs. (See Gray, Genera Am. Bor. Illustr 2, p. 180, f. 168.)⁷ Your close observation might likely enough find a difference in the pollen.

Next take certain portions of the genus Plantago,—upon which, instead of writing details, I refer you to my Manual p. 269,⁸ and to my observations in Pacific Rail-Road Surveys, vol. 4. p. 117,⁹ or, in the Extra copies of Botanical Report on D^r. Bigelow’s collection, p. (117) 61.)¹⁰ Hooker possesses both the volume and the separate paper. I only add that, in Plantago Virginica I have since seen in herbaria, the staminate sort maturing seed, but generally it has gone to dioicism, while in P. Patagonica the two sorts are fruitful, perhaps almost equally so. (If you have difficulty in referring to the memoir above-mentioned, I will have the foot-note copied for you.)

Ilex opaca, the analogue of your Holly, I find in dried specimen, that the female flowers have stamens with filaments as long and anthers as large as those of male flowers, but in a flower-bud examined the anthers have no pollen. The male flowers have no pistil.¹¹

Our Hollies (Ilex & Prinos) must be more particularly examined, to see whether anthers of any fertile flowers bear pollen. I should say they did, from general impression, and from my description in Manual, which was condensed from MSS, prepared long ago for Flora of N. America. But we descriptive botanists have not been careful nor exact enough for your purposes.

In our loose observations we never should have noticed, in Primula, Houstonia, &c—any difference in the pollen of the two sorts of flowers. I should much like to have you tell me what the difference is, in the pollen, and I will make observations next spring, upon Houstonia. The discovery that the pollen of one is good for the pistil of the other, but not for the pistil of its own flower, is most important.— I should rather expect this, but I want to know more about the fact—how you made it out, &c &c

You should next turn your attention to a very different sort of dimorphism, which is almost equally common, perhaps,—one which looks to close- instead of cross fertilization. I allude to such cases as that occuring in the European Impatiens nolitangere, where it was discovered by Weddell, & published by Ad^r. Jussieu (Malphigiaceæ, p. 85),¹² and in our American two species, by D^r. Torrey. See about this, Gray Genera Illustr. 2, p. 132, f. 152, 153.¹³

Here all the fruit in the early part of the season comes from pistils fertilized by their own pollen precociously, in the young bud. But at or after midsummer, many of the conspicuous full-grown flowers are fertile,—their ovary lengthening enough to push out the stigma beyond the connivent scales, so that pollen of the same flower, or of other flowers brought by insects, can get access to it.

Malpighiaceæ, Violaceæ, Leguminosæ, &c—furnish very numerous instances of such precociously-fertilized flowers.— these flowers always far more fertile than others.— And a similar case in Specularia perfoliata has very long been known.

I did not know of diœcio-dimorphism in Linum. I have suspected it in Oxalis, from the differences in relative length of the stamens & styles of some species.