174 



FISHBERG 



true average lies within the limits of i o. 1 1 3 3 ; that the chances 

 are four to one that the true average lies within twice these 

 limits, and nineteen to one that it lies within thrice these limits. 

 From the nature of the method used in obtaining the probable 

 error, it will be observed that the probable error is less, i. e., 

 the average is more accurate, the greater the number of observa- 



15 































i 11 11 1 1 1 1 1 1 

























1 









STATURE 



435 JFWFSSFS 



— 



14 























/ 





























/ 









13 

























, 



1 





1 



s. 













































1 









/ 



\ 



— 



V 























12 

























1 





/ 







\ 

















































/ 







\ 























11 

























1 





/ 













































/ 











/ 





























10 



















/ 









f 











































N 



/ 









j 































9 













I 





\ 



' 









Y 























































/ 









c 



nl 





















I— 



uj8 

 



























/ 



\ 







f,\ 













































f 



\ 









-\ 





















UJ-, 

 Q. ' 













1 













/ 





\ 











^ 









































/ 































6 











CO 











^ 



i\ 





































r^ 











uj\ 



%\ 







































UJ 

















^ 





























5 











UJ 













































4 















































































































3 

































































^ 











1 





































2 





















/ 















/ 



\ 









V. 





























y 



f 













^ 





\ 











\ 











1 







^ 











/ 

































\ 













u 



y^ 





— - 



/ 











[— 















L_ 











V. 



•^ 



_ 



134 138 142 146 150 154 158 162 166 170 174 178 182 136 



STATURE 



Fig. I. 



tions. This accuracy does not increase, however, in the same 

 proportion as the number of persons measured, but as the square 

 root of this number. As the standard deviation decreases, the 

 probable error of the mean decreases. 



As our study of the Jews is mainly concerned with their racial 

 characteristics, if any such can be discerned, we can not rely on 

 the average alone to determine the type. This is only a con- 

 venient, but very vague, means to express a complex group of 



