148 MENDEL'S LAWS 



magna^ Dr E. Warren found " on measuring the offspring it was 

 obvious that the children of the same brood exhibited considerable 

 variability." 1 A more extended inquiry on Aphides led him to 

 suppose, however, that variability in asexual reproduction, though 

 considerable, is less than when offspring have two parents. We 

 shall see that this relative paucity of variations is due, probably, 

 to a comparative lack of retrogressive variations. Lastly, it is a 

 well-established fact that parthenogenetic species, for example 

 dandelions and hawkweeds, are particularly rich in varieties, a fact 

 implying abundant variations. 2 



245. Variability, then, can occur apart from conjugation. We 

 saw that the evidence is strong that spontaneous variability is an 

 adaptation controlled and maintained by Natural Selection, on 

 which account every species and organ tends to be variable in the 

 right degree. 3 The least variable types and structures are those 

 which are closely adapted to the environment by stringent selection. 

 The more variable types and structures are those which are less 

 closely adapted and in which therefore variability is encouraged by 

 Natural Selection, or those which are not stringently selected for 

 the reason that very close adaptation is not necessary. Thus men 

 and garden plants are exceptionally variable because they are so 

 circumstanced that individuals that have varied largely, for 

 example in colour, have survived in considerable numbers and 

 have transmitted their increased variability to descendants. Thus 

 again the squirrel's tail is more variable than its head and forefoot 

 because there is a greater need for close adaptation in the latter. 

 Since, then, spontaneous variations occur in parthenogenetic 

 reproduction, the materials are present for the work of Natural 

 Selection, and that agency is able to increase or decrease the 

 variability of the species to any useful extent. Bi-parental repro- 

 duction is not necessary for the purpose ; and it is, at the very 

 least, most improbable that species can have been burdened with 

 sex to produce that which was already present, and which, 

 apparently, is kept within bounds only by selection. Undoubtedly, 



1 Proc. Roy. Soc. London, 1899, vol. Ixii. p. 154. 



* See De Vries, Species and Varieties, pp. 59-61. " Thousands of forms (of 

 dandelion and hawkweed) may be cultivated side by side in the Botanical Gardens 

 and exhibit slight but undoubted differentiating features, and reproduce them- 

 selves truly." I follow De Vries in supposing that dandelions and hawkweeds 

 are parthenogenetic. I know nothing of the matter personally. There seems, 

 however, to be some doubt (see Bateson, Mendel's Principles of Heredity, p. 247). 

 At any rate it appears clear that these plants, if they do conjugate, do so com- 

 paratively rarely. 



3 See 163. 



