[121] The nature of this mistake is now clear; for as stated above xenia is only likely to occur when the maternal seed-coat is pigmented. The violet coats in this experiment are themselves cases of xenia.
[122] Knight, it was seen, crossed round ♀ × indented ♂ and consequently got no change of form.
[123] Cotyledons seen through coat.
[124] Ordinary dominance of round.
[125] This is an extraordinary statement to be given as a general truth. There are sometimes indications of this kind, but certainly the facts are not usually as here stated.
[126] If we were obliged to suppose that this is a matured conclusion based on detailed observation it would of course constitute the most serious “exception” yet recorded. But it is clear that the five statements are not mutually consistent. We have dominance of round white in first cross.
In the second generation blue wrinkled give only blue wrinkled, and blue round give blue wrinkled and round, in accordance with general experience. But we are told that white round give only white round. This would be true of some white rounds, but not, according to general experience, of all. Lastly we are told white wrinkled give all four classes. If we had not been just told by Laxton that the first cross showed dominance of white round, and that blue wrinkled and blue round give the Mendelian result, I should hesitate in face of this positive statement, but as it is inconsistent with the rest of the story I think it is unquestionably an error of statement. The context, and the argument based on the maple crosses show clearly also what was in Laxton’s mind. He plainly expected the characters of the original pure varieties to separate out according to their original combinations, and this expectation confused his memory and general impressions. This, at least, until any such result is got by a fresh observer, using strict methods, is the only acceptable account.
Of the same nature is the statement given by the late Mr Masters to Darwin (Animals and Plants, I. p. 318) that blue round, white round, blue wrinkled, and white wrinkled, all reproduced all four sorts during successive years. Seeing that one sort would give all four, and two would give two kinds, without special counting such an impression might easily be produced. There are the further difficulties due to seed-coat colour, and the fact that the distinction between round and wrinkled may need some discrimination. The sorts are not named, and the case cannot be further tested.
[127] See later.
[128] The number in Haage and Schmidt’s list exceeds 200, counting colour-varieties.