566 TRANSMISSION 



produced only 40 performers and 8 performing sires, but Geo. 

 Wilkes was one of these, and the average of 21.7 (174-5-8) tells 

 but a small part of the story of these 8 performing sires of this 

 remarkable progenitor of speed. His 142 non-performing sires 

 produced speed at an average of 17.7 each. It would be an 

 interesting study to determine how the descendants of these 

 142 non-performing sires compared with those of the 8 perform- 

 ing sires, a study that the writer has left to others 1 or to a 

 future time. 



A conservative conclusion from these data would be that per- 

 formance is not an invariable index of breeding powers but that 

 on the average the performers are much more likely to get speed 

 than are non-performers of the same breeding. 



This difference, if it really exists, is without doubt inherent; 

 indeed, it is not difficult to find instances in which that which 

 seems to be the general principle is reversed, so that the non- 

 performers are the better breeders (see lines I, 44, 62, 79, 130, 

 and 155).; all of which shows that while good negative testimony 

 may be found in a single instance, positive statements must be 

 based upon a comprehensive study of large numbers. And so 

 we need to go through our records and our experience carefully, 

 hunting for the things that constitute ground on which pre- 

 potency may safely be predicated. Without doubt purity of 

 blood, in the sense of the highest possible percentage of diameters 

 favorable to the purpose desired, unalloyed by disturbing factors, 

 will be found to constitute the real basis of prepotency. When 

 discussing the mathematics of breeding it was found that, no 

 matter what the combinations, a few individuals will always 

 remain pure. By the same process of reasoning, when we mix 

 the elements of desirable characters, diluting them as little as 

 possible with " wild blood," we shall, by the same law of prob- 

 abilities, once in a while effect a phenomenal combination. Such 

 a one is produced by methods not under our control, except as 

 we increase the probability by increasing the intensity of breeding. 

 This is the very heart and soul of "line breeding," and means 

 that the best-bred animal is the most likely to be prepotent. In 



1 Work of this kind runs into days, weeks, and months, at a surprising rate. 

 The data given here represent many months of laborious work. 



