22 THE METHOD OF DEVELOPMENT 



nutritional characters from those which result from other stimuli. 1 

 Small variations which grade into one another are termed * con- 

 tinuous ' : large variations (' sports/ abnormalities, congenital 

 deformities, ' mutations ') are termed ' discontinuous.' 2 All 

 variations, all differences between parent and child due to differ- 

 ences in germ-plasms, may be placed in one or other of two 

 categories : either they are progressive, or they are retrogressive. On 

 progressive variations is founded progressive evolution, as when 

 the wing of a species of bird undergoes increase in size or power- 

 On retrogressive variations is founded retrogressive evolution, as 

 when the wing, owing to lessened utility, undergoes subsequent 

 decrease. 3 An extra digit on the hand of a child of normal parents 

 is an example of progressive variation ; a missing digit, if founded 

 on a germinal peculiarity, not on a mutilation, is an example of 

 retrogressive variation. Of course any given variation may be com- 

 pound, consisting of both progressive and retrogressive variations. 

 Thus a congenitally malformed digit may display both additions 

 and subtractions when compared to the structure in the parent. 



40. Obviously, since the child treads in the developmental foot- 

 steps of the parent, since, like the parent, it is in turn germ, embryo, 

 foetus, infant, child, youth, and adult, a progressive variation 

 implies a complete recapitulation of the parental development (as 

 regards the particular structure or character that has varied), plus 

 an additional step ; whereas a retrogressive variation implies an 

 incomplete recapitulation, one or more of the steps of the parental 

 development being omitted. The line of thought on which we 

 are now engaged is so very essential to a clear understanding, both 

 of evolution and heredity, that it will be well worth the reader's 

 time to close this volume for a space, and to endeavour to think of 

 a variation which is neither progressive nor retrogressive, but 

 something else. He will find it impossible. If he then tries to 



1 The reader must constantly bear in mind the real, or at least the more 

 useful distinction between variations and acquired differences. Obviously, to 

 take an example, there is, between two men one of whom is ' by nature ' lean, 

 who cannot get fat no matter what he eats, and the other who is thin merely 

 because he is starved, a difference more radical, more innate than that between 

 a fat, well-fed man and one who has been starved, but would get fat if afforded 

 the opportunity. 



2 See 249, 286. 



3 Sir Ray Lankester has objected to the expression ' retrogressive evolution ' 

 on the ground that retrogression implies an exact and orderly reversal of the 

 antecedent progression. But if we term the sum of progressive variations ' pro- 

 gression,' it appears right to term the sum of retrogressive variations ' retrogres- 

 sion.' At any rate I cannot find a better term and, at least, my meaning is clear. 



