ALCOHOLISM AND BEHAVIOR OF WHITE RATS 



257 



(2.8 and 2.7 times). There are, then, six days on which the 

 differences are probably significant, indicating that the tests 

 went further than the controls, and there are no significant 

 differences indicating the reverse. 



Moreover that eleven out of twelve days should have the same 

 sign has a significance quite apart from the probable errors on 

 the individual days. Departures from equality as great as 

 this may be expected in one out of 157 times if there is no real 



TABLE 10 

 Showing the significance of the differences between the distance averages, based on 

 the total distance covered by each rat on each day. These differences are taken 

 from the averages for sexes and strains together in table 9. The plus sign indicates 

 that the average of the test is higher than that of the control. In the first half of 

 training the differences are either close to three times their probable errors or higher, 

 so they may be considered significant 



Training. 



Retention , 



difference between the tests and controls; but this is a smaller 

 probabiHty than when a difference is over four times its probable 

 error. 



A better idea of the differences between the 'plus' and 'minus' 

 ratios in table 9 is given by the graphs in figure 11. In these 

 curves the frequency distribution of the 'plus' ratios is given in 

 the solid line, of the 'minus' ratios in the broken line. These 

 distributions include only those ratios that do not involve 

 repetition of any primary data, namely, for males and females 



