skip to content

Darwin Correspondence Project

To William Herbert1   [c. 1 April 1839]

(1st.) In describing what can be effected by crossing different species of plants,2 Mr. Herbert speaks, almost in the same terms, used by breeders of animals in regard to the almost unlimited effects of judgment and skill in uniting and opposing the different qualities of the parents. But breeders of animals always insist upon the great importance of choosing an animal, which not only possesses the requisite qualities, but which has possessed them for several generations, whereby they become fixed in the breed. Does Mr. Herbert consider a character, produced by crossing, or more especially if arising from simple variation, which has passed through some generations, very much more likely to continue in the offspring than one which has appeared for the first time. If one flower of a plant were impregnated by the pollen of another species, retaining its normal character, and a second flower, by the pollen of the same species but after it had undergone considerable variation, would the seedlings in the first case be impressed more persistently, (that is in a greater number of successive generations) with the character of the plant from which the pollen had been derived in its wild or old than in its new or cultivated state ? Or again if two flowers, say of a Rhododendron, were impregnated by the pollen of two Azaleas, of the same variety, but of which two, one had been raised many times successively by seed, and the other had only lately appeared with its new character, has Mr. Herbert reason for prefer the pollen of the plants of successive generations to produce the desired effect, than that of the other. In short do species (using the word in its common sense) impress their character more persistently (though perhaps not more on the first generation?) than varieties. Would the stain on the successive seedlings of a cauliflower from the pollen of the wild cabbage be more indelible, than that from some other artificial variety? Would a simple variety, produced by accident, or cultivation, be much more likely to appear from a seed, the parent of which had possessed the desired peculiarity for some generations backwards, than from one, in which it had just appeared; the seeds being sown in each case under precisely the same circumstances. One is urged to think that a condition, namely high-breeding which is of so much consequence in the propagation of animals, cannot be without effect in the breeding of plants.

(2d) As an inference from the above questions, it may be asked does the tendency to sport with respect to any particular character (see p. 348 of Mr. Herbert’s work) become less in a cross, after that character shall have been permanet through several successive generations?

(—3d) Do the seedlings of a cross (to the parents of which, there is no reason to suppose any pollen has had access excepting that of the one desired) entirely resemble each other; or do some retain more of the character of the mother and others of the father plant, as frequently happens with a litter of mongrel puppies?— If this be so, does this difference in the offspring one with another happen more commonly where species or where varieties have been crossed? Does it depend upon the amount of difference between the parents, or whether they be species or varieties? Mr Knight3 seems to have found this variation in mongrel apples &c &c common, whereas Mr Herbert makes in the Hort. Transacts. an opposite assertion with regard to the hybrid Criniums.—4

(4th)— In Mr. Herberts work (p. 372) it is said some genera are much more difficult to hybridise than others. As a general fact are not such genera, slow in sporting, or yielding many varieties?— Is the converse true: that is, is there any case of tolerably allied species, which have given rise to many varieties, which will not hybridise each other, and generally produce fertile offspring?— In short is there any relation between a facility in varying (a capability which seems very different in different animals) and a facility in giving hybrids, and especially fertile hybrids.

(5th.)— Is it generally more easy to make hybrids between two cultivated plants, or between one cultivated plant and a wild one (or a cultivated plant which has retained its original character) than between two wild ones? In short, does the variation of the original character, which generally follows cultivation, facilitate hybridisation? For example (although, perhaps, from the certainty in either case, the example is a bad one) is it more easy to hybridise an apple with the pollen of a crab, than the crab with the pollen of an apple? Or can no difference be perceived in such cases?

(6th) Do those genera of which the species are difficult to be hybridised, stand in general rather isolated in the series of the vegetable kingdom? Will Mr Herbert name the principal genera of this description which he may have tried unsuccessfully ?

(7th.)— It has been said (apparently with litte foundation) that amongst birds, species originally coming from distant parts of the world, are more likely to breed together, than those from nearer countries.—5 Has Mr. Herbert observed anything of this kind in plants?

(8th.)— In Mr. Herbert’s great experience, has he ever known a character appear for the first time in a plant, and be afterwards transmitted to the second generation, or grandchild and not to the child? There being, however, always this great source of doubt to guard against, that the same cause, which produced the character in the grandfather, may again act in the grandchild. A case, in example would be supposing a man should be born with six toes, or hare lip, or be an albino &c, the first in the family known to have been so, and he should marry a perfect woman, and have perfect children, and yet that some of the grandchildren should suffer under such peculiarities.

(9th.) Mr. Herbert (p. 336) says that owing to the hybridising process, being somewhat inimical to fertility, it is less likely that useful fruit trees should be produced, than flowers. As this remark, is made in allusion to Mr. Knight’s experiments, which were chiefly made on varieties, does Mr. Herbert believe that the crossing of even very distant varieties is inimical in any degree to fertility?— In some other passages Mr. Herbert seems of opinion that the pollen of a slightly different variety is more favourable to the fecundity of a plant than that of its own.

(10th.)— Has Mr. Herbert ever seen an hybrid Cryptogamic plant, either ferns or mushrooms?— The foreign journals have lately mentioned some cases of hybrid ferns.6

CDarwin 12 Upper Gower St.—

CD annotations

Top of first page: ‘(1)’7 brown crayon


The draft, in Syms Covington’s hand, originally had a heading ‘Questions for Mr Herbert’. It was deleted by CD in ink similar to that of the other alterations he made throughout the draft (see Manuscript Alterations and Comments). At the end of the draft CD added in pencil: Covington Copy these questions on a sheet of paper but leave a narrow margin, & make each paragraph with the numbers begin a fresh line.— Will you let me have them by ten oclock tomorrow.— This final copy has not been found. Herbert, Rector of Spofforth, Yorkshire, was a leading authority on plant hybridisation. His Amaryllidaceæ (1837) is cited frequently in CD’s Notebook E during the first half of 1839. Herbert concluded from his experiments that ‘In fact, there is no real or natural line of difference between species and permanent or descendible variety’ (W. Herbert 1837, p. 341).
W. Herbert 1837, p. 346. CD’s copy (Darwin Library–CUL) is heavily scored and annotated in the sections headed ‘Preliminary Treatise’ and ‘On crosses and hybrid intermixtures in vegetables’. Several of the annotations deal with topics repeated in CD’s questions here, see especially pp. 348, 366. CD’s indebtedness to Herbert’s work is apparent in the many references in Natural selection.
Thomas Andrew Knight. His experiments on crossing fruit and vegetables are reported throughout the Transactions of the Horticultural Society, of which he was President (1811–38). Notebook E has frequent notes on his articles.
‘Twenty-four mule seeds [of Crinum cross] were produced in one capsule; the young plants are all exactly like each other, and differ in the same points from the parents, and I am confident that the same botanical distinctions will be perpetuated in the seedlings that may be raised from them. A new species of Crinum … has been thus created’ (W. Herbert 1820, p. 190).
A view put forward in Eyton 1837, p. 359 and Eyton 1838, p. 1. CD also mentions Eyton’s views in Notebook D: 25.
CD refers to a notice of Martin Martens’ work on hybrid ferns in Annals of Natural History 2 (1839): 236.
CD kept this draft with the letter from William Herbert to J. S. Henslow, 5 April 1839, a letter to Herbert, 26 June 1839, and one from Herbert, [c. 27 June 1839]. They were numbered 1, 2, 3, and 4 in brown crayon. CD also included a memorandum, numbered 2a, which has been transcribed following the letter from Herbert, [c. 27 June 1839].


Eyton, Thomas Campbell. 1838. A monograph on the Anatidae, or duck tribe. London: Longman, Orme, Brown, Green & Longman.

Herbert, William. 1820. Instructions for the treatment of the Amaryllis longifolia, as a hardy aquatic, with some observations on the production of hybrid plants, and the treatment of the bulbs of the genera crinum and amaryllis. Transactions of the Horticultural Society of London 3: 187-96.

Herbert, William. 1837. Amaryllidaceæ; preceded by an attempt to arrange the monocotyledonous orders, and followed by a treatise on cross-bred vegetables, and supplement. London: James Ridgway & Sons.

Notebook D. See de Beer 1960; de Beer and Rowlands 1961; de Beer, Rowlands, and Skramovsky 1967; Notebooks.

Notebook E. See De Beer 1960; De Beer and Rowlands 1961; De Beer, Rowlands, and Skramovsky 1967; Theoretical notebooks.


Questions on breeding of plants: variation in established versus new varieties; predominance of wild species and old varieties when crossed with newer forms; predominance of males versus females; correlations between ease of hybridisation and tendency to vary and undergo cultivation; reversion; correlations between hybridisation and geographic distribution.

In WH’s Amaryllidaceae [1837], does he intend to say crossing is inimical to fertility?

[Sent via J. S. Henslow; note to amanuensis Syms Covington.]

Letter details

Letter no.
Charles Robert Darwin
William Herbert, dean of Manchester
Sent from
London, Upper Gower St, 12
Source of text
DAR 185: 62
Physical description
A 3pp †

Please cite as

Darwin Correspondence Project, “Letter no. 502,” accessed on 8 December 2022,

Also published in The Correspondence of Charles Darwin, vol. 2