132 HEREDITARY CHARACTERS 



generally have to go to entirely different districts where 

 different conditions prevail, and often to districts which 

 have been cut off from each other geographically for a very 

 long period of time. 



We have also very good evidence to show that as soon as 

 a structure or character ceases to be useful to the race it 

 tends to disappear. 1 This means that when a character 

 ceases to be the subject of selection it will be eliminated, 

 the time during which it will be retained apparently being 

 proportionate to the period during which it has existed 

 among the ancestors of the organisms. 



We meet with statements that races of domestic animals 

 reproduce themselves with great uniformity if inbred, but 

 that " the moment one mixed up two different races, strains, 

 or breeds, one did something that was difficult to put in 

 words, but the result was what has been best described as 

 an ' epidemic ' of variations." 2 To begin with, variations 

 occur so frequently in inbred races of domestic animals, that 

 it is difficult for the standard to be kept up in many cases. 

 It is only maintained by rigid selection. Of course when 

 two different breeds are crossed, the offspring vary from 

 each other and from their parents enormously. This is what 

 one would expect, for the newly acquired characters of the 

 two parents are peculiarly liable to regression, and when two 

 different sets of newly acquired characters are mixed, it is 

 likely that many will disappear, and any characters between 

 those of the common ancestor of the two parents and those 

 of the two parents themselves may appear in the offspring. 

 There is, however, nothing to suggest the appearance of an 

 unusual number of new variations. 



Lastly, we have the fact that variations occur in par- 

 thenogenetic and in unicellular forms, where bi-parental 

 reproduction does not occur. 3 



1 See p. 66. 



2 Cossar Ewart, "Discussion on Heredity in Disease," in Scottish Medical 

 and Surgical Journal, 1900, iv. p. 308. 



3 See p. 60. 



