402 FRANCIS B. SUMNER 



tain of these races are compared. Taking the four coefficients 

 for our two crosses, in the Fi generation (males and females 

 being dealt with separately), we find two positive and two nega- 

 tive figures, the mean being +0.036, which can hardly be re- 

 garded as significant under the circumstances. The condition of 

 the F2 lots is doubtless a fairer test of this question. Of the four 

 figures for this generation, three are positive and one negative, 

 the mean of the four being +0.105. That the latter fact can 

 hardly be regarded as furnishing evidence for an actual positive 

 correlation between these characters is evident when we consider 

 that in the Carlotta-Calistoga cross the mean figure for the two 

 sexes is negative, this negative condition being outweighed by a 

 positive one in the Eureka- Vic torville series. It does not seem 

 likely, that the two crosses would be found to differ in this re- 

 gard if adequate material were available. 



Another pair of characters for which I have computed correla- 

 tion coefficients in these series of hybrids are tail stripe and foot 

 pigmentation.-^ Of the eight coefficients here computed, four 

 (the Eureka-Victorville ones) were positive, the other four (the 

 Carlotta-Calistoga ones) being negative. In this case, however 

 the negative figures outweigh the positive ones, both in the Fi 

 and the F2 series, the means for the two generations being —0.184 

 and —0.035, respectively. It must be pointed out, too, that the 

 most nearly 'significant' figures, taken singly, are those of the 

 Carlotta-Calistoga cross, whereas it is doubtful whether these two 

 races differ appreciably in their mean foot pigmentation. This 

 fact renders it probable that all of these coefficients are 'acci- 

 dental;' that is to say, that they have no bearing upon the special 

 problem under discussion. 



"* The pigmentation of the sole of the (left) foot has been rated according to 

 an arbitrary scale of five grades. The feet of the mice were preserved in 70 per 

 cent alcohol and later transferred to glycerin. 



