February 19, 1904.] 



SCIENCE. 



313 



venture brief replies to Professor Metcalf's 

 questions, premising only that these sugges- 

 tions are incidental to ' A Kinetic Theory of 

 Evolution' outlined in Science of June 21, 

 1901, and in subsequent papers. 



"Are there mutations which are distinct 

 from fluctuating variations? Are fluctuating 

 variations restricted to rather narrow limits, 

 and are the larger variations which occur of a 

 different sort, estahlishing a new mean ahout 

 which a new series of fluctuating variations 

 cluster?" Yes; mutations or spots which ap- 

 pear among inbred domesticated plants and 

 animals differ from the ' fluctuating ' individ- 

 ual diversity of unsegregated wild types in the 

 amplitude or abruptness of the variation, and 

 in a more or less pronounced reproductive 

 debility. ' Muctiiations ' and mutations are 

 extremes of the same series of phenomena, 

 but their evolutionary significance is very dif- 

 ferent. New types are built up through the 

 interbreeding and accumulation of genetic 

 variations, but mutations which depend for 

 their existence on narrow segregation do not 

 contribute to the evolutionary progress of 

 species. 



" Are mutations {or variations') definite or 

 indefinite? Do they follow certain lines or 

 do they occur in all directions?" Variations 

 of both kinds occur in many directions. The 

 idea that they exactly offset each other and 

 thus maintain a stationary average has no 

 warrant of observation and is opposed to the 

 calculus of probabilities. Species tend to 

 move in some directions, but not in all direc- 

 tions (Darwin), nor in one particular direc- 

 tion (Naegeli). 



"If the direction of mutations {or varia- 

 tions) is wholly or in part predetermined, what 

 are these predetermining factors? Are they 

 internal {involved in the nature of the organ- 

 ism), or external {environmental), or hoth? " 

 They are internal, but not predetermined. 

 Organisms of the same descent under the same 

 conditions give diverse mutations. Of their 

 causes in detail nothing is knovsm; mutations 

 are, however, induced by persistent inbreeding. 

 The direction or the extent of variation may 

 ' upon external conditions. A vari- 



ation in the direction of larger size would not 

 be able to develop without adequate food. 



"Is there a tendency in mutants {or vari- 

 ants) to revert toward the condition of the 

 parent stock?" Normal genetic variations 

 are more vigorous and prepotent than their 

 immediate relatives, but mutations tend to 

 ' revert ' when the abnormal inbreeding is 

 remedied by crossing. 



" Are mutants {or variants) of one sort 

 more {or less) fertile or more {or less) vigor- 

 ous when bred together than when bred with 

 the parent stoch or with mutants {or variants) 

 of another sort?" Sustained vigor and fer- 

 tility, and evolutionary progress, as well, de- 

 pend on normal interbreeding (symbasis). 

 Mutative varieties are, in general, rendered 

 more vigorous by crossing with less inbred 

 stock, but often at the loss of their peculiar 

 characters. 



" Does mutation {or variation) cause 

 partial {or complete) segregation? " Muta- 

 tions are sometimes completely segregated by 

 sterility, perhaps also by cytological or other 

 malformations which prevent the resumption 

 of interbreeding, but such abnormalities have 

 no general evolutionary significance. 



"Are hybrids between mutants {or variants) 

 of different sorts or between mutants {or 

 variants) and the parent stoch intermediate in 

 character between the two parents, or do they 

 follow wholly or chiefly one parent? If the 

 latter, which parent is foUoived in the several 

 kinds of crosses?" Crosses between different 

 mutants or even between similar mutants of 

 different descent tend to ' revert ' to the 

 parental type. In crosses between mutants 

 and their immediate and equally inbred rela- 

 tives the mutant is prepotent, but individuals 

 of the parental type may be prepotent if of a 

 sufficiently remote line of descent. When the 

 divergence of descent is too great or too long- 

 standing to permit a return to the ancestral 

 form, or when the prepotency of the mutation 

 is balanced, as it were, by the prepotency at- 

 taching to smaller degree of inbreeding of the 

 form with which it is crossed, there result 

 disjunctive or ' Mendelian ' hybrids.* 



* Further confirmation came to hand after this 

 letter was sent in. Professor Davenport finda 



