12 



Records of the Indian Museum. [Vol. XXIII, 



The ravv material in the form of the actual measurements, 

 has been included in Appendix II. 



" Tables," throughout the present paper, have reference to 

 the indispensable volume edited by Karl Pearson, " Tables for 

 Statisticians and Biometricians ' ' (Cambridge University Press, 

 1914). 



Note on "Bias" in recording measurements. 



It is well known that different observers are affected with 

 different 'peisonal bias' in taking measurements. In the present 

 case the crude data showed an overwhelming preponderance of 

 " even " readings as against ( ' odd " measurements. 



In the case of Stature, we find no less than 193 " even ' 

 reading as against only 7 <( odd." We have no reason to believe 

 that Nature has any special preference for " even'' number of 

 millimeters, hence, apart from personal bias and fluctuations due to 

 random sampling we should have had 100 "even' and "odd' 

 readings each. Instead of this, we actually get 193 and 7. 



The presence of "bias" is obvious, but I have calculated 

 the ' ' Contingency ' ' ' for the whole group of the above seven 

 measurements. 



Table i. 



Contingency for <l bias." 



Organ. 



Theoretical 

 value. 



Stature 

 Head Length 

 Head Breadth 

 Nasal Length 

 Xasal Breadth 

 Zyg. Breadth 

 T'pper Face Length 



;/ =7 



m — m'\'* 



I 



\ m 



86-49 

 5476 

 65-61 



I-2I 



0'49 



3I-36 



0*25 



x*= 240-17 



The probability that "random sampling" would lead to as 

 large or larger deviation between theory and observation is given by 



P = e' \ i + — + — 



2-4 J 



f v 2 x 1 

 log /'= - \x l . log l0 « + log 1 1 + - + - ■ . J 



' Karl Pearson: Phil. Mag. Vol. I., pp. 157-175. 1 <>><>. 



