THE CAUSE OF RETROGRESSION in 



characters (i.e. variations) tend to disappear in the next generation 

 (e.g. birth-marks, moles, etc.). 1 



182. Since, then, the tendency to retrogression is greater in the 

 case of new characters than of old-established traits, the stringency 

 of selection by which the former are maintained, or by which fresh 

 progression may be caused in them, is proportionately intense. 2 

 In other words, offspring frequently display progressive variations 

 in the case of old-established characters, whereas in the case of 

 new traits the great mass of variations is retrogressive. Thus, if 

 we try to improve the long-established characters of a compara- 

 tively old race, for example ordinary horses, our task is much less 

 difficult than if we operate on the new characters which have 

 recently been evolved under stringent selection, for example the 

 special traits of race-horses. 3 In the latter case the tendency to 

 retrogression is sometimes so great that our utmost efforts may do 

 no more than prevent retrogression. In other words, stringent 

 selection no longer causes progression. 



183. Since even old-established characters retrogress on cessa- 

 tion of selection, it follows that the frequency, or the magnitude, 

 or both? of retrogressive variations tends to be greater than that of 

 progressive variations, and that the difference is most considerable 

 in the case of newly evolved characters. The fact that retrogres- 

 sive variations tend to exceed progressive variations in magnitude 

 is well shown by the fact that while many generations of selection 

 may be needed to evolve a prize breed, the offspring of a pure-bred 

 prize pair may in a single generation display a loss of the special 

 characters of the parents and revert to the ancestral type. This 

 is not uncommon in the case of prize animals, but it is even more 



1 Certain variations tend, apparently, to possess a greater degree of 

 stability than others. We shall deal with them when we discuss Mendelism 

 and the now popular Mutation theory of evolution. See chapters vii., viii., 

 and ix. 



2 Some Mutationists maintain that antiquity (i.e. long-continued selection) 

 has nothing to do with the stability of characters. They insist that all true 

 variations are absolutely permanent from the first ; whereas ' fluctuating 

 variations ' depend, not on change in the hereditary tendencies, but on differences 

 of nutrition and the like on differences in stimuli. According to them evolution 

 is founded entirely on the former class of variations. We shall consider this view 

 also later. 



3 " Two years ago thirty-two yearlings were sold for 51,250 guineas. These 

 thirty-two yearlings are represented by two winners of five races, Florio Rubatino 

 and La Reine, who have contributed 2000 to the total cost ; and there is not, so 

 far as can be known, a single one of the thirty with any prospect of making a race- 

 horse." (Quoted from the Times, December 27th, 1897, by Sir Walter Gilbey, 

 Race-horses, p. 6.) 



