Dear Sir,

Having been much interested in your late work on the variation of plants & animals¹ I venture to address you (trusting that you will not think me impertinent in so doing) to suggest that the variation of plants may often be caused by a variation in the form of the membrane covering the seed, which variation would enable the seed to absorb food from the soil in a form different to what its progenitor did so (for instance a ses-qui-carbonate or bi-carbonate instead of a carbonate or sub-carbonate)² & this food being in a different crystalline form, the cells built up from it and the new plant itself would be in a different form to its progenitor.

Such variation in the covering of the seed need not arise by chance or accident, but be caused by the mere fact of exhaustion of the soil by the parent plant.

Suppose, for instance, that the parent plants fed on magnesia, which existed in the soil in the form of a subcarbonate: by gradual exhaustion of the magnesia what remained might be converted into carbonate: this however would be in a different crystalline form to the subcarbonate & would not “diffuse” into the roots of the plants, (which could absorb only the sub-carbonate,) & they would begin to die from starvation: some of them however are able to produce seed but their magnesia, being deficient in quantity, is more attenuated in their leaves & seed vessels than in those of healthy plants, and in consequence it gets acted on to a greater extent by the carbonic acid of the atmosphere and is converted from a subcarbonate into a carbonate; and the seeds are formed with this new crystal in their tissues or so as to allow the new crystal to pass through the membrane covering them: consequently these seeds when set can obtain their requisite supply of magnesia from the carbonate in the soil whereas the old plants can obtain none: & what you call “natural selection” would take place or in other words the old plants would die from starvation while the new plants would take their place and a new genus or variation of species would have taken place—

This idea (of course it is only a very crude idea) is somewhat supported I think by a fact which I have frequently noticed in Bengal in connection with the Balsam— It is that the seed of the imported double balsam will, if picked & preserved in the house until the next season, then produce plants & flowers, precisely the same as their immediate ancestors; but if the seed is not preserved in the house but allowed to fall on the ground & set itself the plants & flowers which come from it will not be anything like their immediate ancestors but will revert to the form of the common indian single balsam. There may be some other way of accounting for this fact but I attribute it to some change effected by drying the seed which causes the seed to shrink & thereby as I presume causes it to absorb its food in a different form to what it would do if it remained of its normal size & shape.³

I believe many seeds will revert in a similar manner in England, the cauliflower for instance, but I cannot state this as a fact.

Yours faithfully | George Rogers

4^th June 1868

Cha^s. Darwin Esq. M.A. &c