138 THEORY OF EVOLUTION 



each side of that point, a to A and b to B, 

 should be protected, so to speak, from further 

 crossing over. This in fact we have found to 

 be the case. No other explanation so far ]3ro- 

 posed will account for this extraordinary 

 relation. 



Wliat advantage, may be asked, is there 

 in obtaining numerical data of this kind? 



\ 



Fig. 68. Scheme to indicate that when the members of a 

 pair of chromosomes cross (at a-b) the region on each side is 

 protected inversely to the distance from a-b. 



It is this: — whenever a new character appears 

 we need only determine in which of the four 

 groups it lies and its distance from two mem- 

 bers within that group. With this information 

 we can predict with a high degree of proba- 

 bility wliat results it will give with any other 

 member of any group. Thus we can do on 

 pa2)er wliat would require many months of la- 

 bor by making the actual experiment. In a 

 word we can predict what will liappen in a situ- 

 ation where prediction is impossible without 

 this numerical information. 



