I915-] THE HEREDITARY MATERIAL. 151 



long. These correspond to the two grandparents. The two smaller 

 classes are white long and red miniature. 



We can account for this result if we assume first that the two 

 factors that went in together in the same chromosome tend to hold 

 together. This would account for the two larger classes. Second 

 that the two smaller classes are due to interchanges between the 

 two X chromosomes. Such interchange would here take place only 

 once in three times. 



We can test this conclusion by planning the experiment in such 

 a way that wdiite and miniature now go in from opposite sides, — 

 w^hite from one parent, and miniature from the other. When we do 

 this we find that the large classes in the second (back cross) 

 generation will be red miniature and white long and that the small 

 classes will now be red long and white miniature. The ratio of the 

 large to the small classes will be exactly the same as in the first 

 case. In other words the interchange between the X chromosomes 

 is the same regardless of what factors each contains. 



If one admits that the chromosomes are the bearers of the 

 hereditary factors he is forced to admit that experiments like these 

 prove that somehow interchange of factors in homologous chromo- 

 somes must occur. 



If one thinks of the factors as lying in a linear series in the 

 chromosome (and there is certain evidence that I can not consider 

 here that makes this view imperative) then the chance of a crossing 

 over taking place somewhere in the region between two pairs of 

 factors would be greater the farther apart the factors lie. The 

 percentage of times that crossing over takes place becomes then a 

 measure of the distance apart of the factors in question. If we 

 make this assumption we find that we can give a consistent explana- 

 tion of everything that we have found in the inheritance of linked 

 factors in Drosophila. Not only this, but a far more important 

 fact comes to light. If we determine, on the aforesaid basis, the 

 relation to each other of all the known factors in each of the four 

 groups, then, when a new factor appears, we need only determine 

 its group and its relation to two factors in that group. With this 

 information we can predict its relation to all other members of that 

 group. In other words we can predict what the numerical relation 



