those humble places, will do a few days' weeding, prick out some rows of cabbages, feed up a few score of any variable larva, he will not wait long before he learns the truth about variation. If he go further and breed two or three generations of almost any controllable form, he will obtain immediately facts as to the course of heredity which obviate the need for much laborious imagining. If strictly trained, with faith in the omnipotence of selection, he will not proceed far before he encounters disquieting facts. Upon whatever character the attention be fixed, whether size, number, form of the whole or of the parts, proportion, distribution of differentiation, sexual characters, fertility, precocity or lateness, color, susceptibility to cold or to disease—in short, all the kinds of characters which we think of as best exemplifying specific difference, we are certain to find illustrations of the occurrence of departures from normality, presenting exactly the same definiteness elsewhere characteristic of normality itself. Again and again the circumstances of their occurrence render it impossible to suppose that these striking differences are the product of continued selection, or, indeed, that they represent the results of a gradual transformation of any kind. Whenever by any collocation of favoring circumstance such definite novelties possess a superior viability, supplanting their 'normal' relatives, it is obvious that new types will be created.
The earliest statement of this simple inference is, I believe, that of Marchant, who in 1719, commenting on certain plants of Mercurialis with laciniated and hair-like leaves, which for a time established themselves in his garden, suggested that species may arise in like manner. Though the same conclusion has appeared inevitable to many, including authorities of very diverse experience, such as Huxley, Virchow, F. Galton, it has been strenuously resisted by the bulk of scientific opinion, especially in England. Lately, however, the belief in mutation, as De Vries has taught us to call it, has made notable progress, owing to the publication of his splendid collection of observations and experiments, which must surely carry conviction of the reality and abundance of mutation to the minds of all whose judgments can be affected by evidence.
That the dread test of natural selection must be passed by every aspirant to existence, however brief, is a truism which needs no special
- Marchant, Mém. Ac. roy. des sci. for 1719; 1721, p. 59, Pls. 6-7. I owe this reference to Coutagne, L'hérédité chez les vers à soie (Bull. sci. Fr. Belg., 1902).
- This progress threatens to be rapid indeed. Since these lines were written Professor Hubrecht, in an admirable exposition (Pop. Sci. Monthly, July, 1904) of De Vries's 'Mutations-theorie,' has even blamed me for having ten years ago attached any importance to continuous variation. Nevertheless, when the unit of segregation is small, something mistakably like continuous evolution must surely exist. (Cp. Johannsen, 'Ueb. Erblichkeit in Populationen und in reinen Linien,' 1903.)