256 PRESIDENTIAL ADDRESS. 



Is there any evidence wliicli will help ns to decide 

 whether correlated variations are in all cases the result 

 of natural selection through elimination? When we con- 

 sider the evidence which is afforded by domesticated animals, 

 the impression left on the mind (at any rate in my own 

 case) is that the variations from which selection has been 

 made have been determinate, and that if the tendency in 

 particular directions is the result of natural selection there 

 is not much direct evidence of the fact. In many cases 

 the new varieties have originated in " sports " — that is to 

 say, in determinate and somewhat marked aberrations from 

 the type. The races of domestic pigeons so fully considered 

 by Darwin, and summarized by Mr. Wallace, have probably 

 originated to a greater extent from particular " sports," 

 the characters of which have been intensified by selective 

 breeding, than by the slow accumulation of small deviations 

 from the original type. Here it would seem that there 

 lies open a profitable field for experimental observation, 

 with the object of throwing light on the direction or 

 directions of variation in particular cases, and the dimen- 

 sions so to speak of the steps of variation. Are the steps 

 small and uniform, or does nature stride and even leap 

 sometimes, and if so to what extent? The way in which 

 pigeons tumble back in reversion is most remarkable and 

 perhaps instructive. Darwin crossed white fantails with 

 black barbs, and obtained black, brown, or mottled mongrels. 

 He crossed a barb with a spot, and obtained dusky and 

 mottled mongrels. He then crossed the two sets of mongrels ; 

 and forth stepped the wild rock-pigeon (or its counterpart) 

 with blue colour, barred and white-edged tail, and double- 

 banded wings. In reversion there are very considerable 

 jumps. 



Darwin, as is well known, thought it probable that varia- 



