THE GENETICS OF SPONTANEOUS TUMOR FORMATION 257 



chromosomal influence has, therefore, largely disappeared. When crosses 

 \V, X, Y and Z which lacked the extra-chromosomal high tumor influence are 

 compared with these, there are found to be 4 tumors when 122 were expected. 

 The percentage is similar to the previous crosses but the original extra- 

 chromosomal influence was different. 



Having eliminated the extra-chromosomal influence, we may next com- 

 pare the various crosses as regards their chromosomal composition. Crosses 

 S and U should resemble the original low cancer strain. Actually this was 

 the case as no mammary tumors were recorded in them. 



Table 6 



Crosses T, \, W and Y had roughly the same formulae as the original 

 outcross BdFi and BdFo. The number of animals observed should have 

 given a total of 24 mammary tumors if the tendency to form mammar}- 

 tumors had been transmitted through the chromosomes. Actually 7 or 2^% 

 of that number were observed. 



Crosses X and Z should be comparable to virgin females of the "high 

 tumor" strain. There should have been 144 mammary tumors. Actuallv 

 there were 3. Even if the extra-chromosomal influence was ruled out, there 

 should have been approximately 14 mammary' tumors formed. 



There has, therefore, been a ver>^ clear decrease in cancer incidence which 

 requires further study. 



Evidence that at least part of the extra-chromosomal influence can be 

 transferred by foster nursing. — There are two important wa>'s in which 

 evidence of the importance of foster nursing in determining the incidence of 

 mammarv tumors in mice can be obtained. 



