Bateson insists that the mutational variation represents a change of loss. “Almost all that we have seen,” he says, “are variations in which we recognize that elements have been lost.” (_Science_, Jan. 20, 1922, p. 59.) In his Address to the British Association (1914), he cites numerous examples tending to show that mutant characters are but diminutions or intensifications of characters pre-existent in the wild or normal stock, all of which are explicable as effects of the loss (total or partial) of either positive, or inhibitive (epistatic) hereditary factors (genes). One of these instances illustrating the subtractive nature of the factorial mutation is that of the Primula “Coral King,” a salmon-colored mutant, which was suddenly given off by a red variety of Primula called “Crimson King.” Such a mutation is obviously based on the loss of a germinal factor for color. The loss, however, is sometimes partial rather than total, as instanced in the case of the purple-edged Picotee sweet pea, which arose from the wholly purple wild variety by fractionation of the genetic factor for purple pigment. Even where the mutational variation appears to be one of gain, as happens when a positive character appears _de novo_ in the phenotype, or when a dilute parental character is intensified in the offspring, it is, nevertheless, interpretable as a result of germinal loss, the loss, namely, total or partial, of a genetic inhibitor. Such inhibitive genes or factors are known to exist. Bateson has shown, for example, that the whiteness of White Leghorn chickens is due, not to the absence of color-factors, but to the presence of a genetic inhibitor—“The white of White Leghorns,” he says, “is not, as white in nature often is, due to the loss of the color elements, but to the action of something which inhibits their expression.” (Address to the Brit. Ass’n., Smithson. Inst. Rpt. for 1915, p. 368.) Thus the sudden appearance in the offspring of a character not visibly represented in the parents may be due, not to germinal acquisition, but the loss of an inhibitory gene, whose elimination allows the somatic character previously suppressed by it to appear. Hence Bateson concludes: “In spite of seeming perversity, therefore, we have to admit that there is no evolutionary change which in the present state of our knowledge we can positively declare to be not due to loss.” (_Loc. cit._, p. 375.)