LII— LIII] 



Introduction 



Ixxix 



Illustration (ii). Consider the fourfold Table below and discuss the relative 

 probabilities that it has arisen from a population which shews 0, 1, 2, 3, etc. indi- 





A Not-J 



Totals 



B 



Not-5 ... 



127-5 1 

 863-5 ;' 87 



127-5 

 9.50-5 





991 



87 



1078 



viduals for this size of sample in the cell B, not-.4. On the assumption that 

 is really the population of this cell, the probability is unity. Hence we have the 

 following result. 



Population) 

 of cell S 







1 



1 



2 



S 



^ 



5 



6 



7 



s 







10 



11 



13 



13 k 

 over 



Probability) 



ofO [ 



occurring ) 



•36788 



-13534 



-04979 



-01832 



-00674 



-00248 



•00091 



-00034 



-00012 



-00005 



-00002 



■00001 



-00000 



Sum = 1-58200. 



Whence taking the a priori probabilities proportional to the probability of 

 occurring on the separate possibilities we have : 



Prohahilities that the Table arose from a population 

 u'ith X in the B, not-A cell. 



X 



Probability 



X 



Probability 







•632,110 



7 



-000,575 



1 



■232,541 



S 



■000,215 



3 



-085,5.50 



9 



-000,076 



s 



-031,473 



10 



•000,032 



k 



-011,580 



11 



■000,013 



B 



-004,260 



12 



-000,006 



6 



-001,568 



13 and over 



■000,000 



The " association " of such a Table cannot therefore be considered " perfect," for 

 in 37 °/j, of cases it would arise from a Table with a imit or more in the B, not- J. 

 cell. The above is actually a Table of the correlation of stature in father and son. 

 Grave caution is therefore needful in discussing such "perfect association" tables. 



Table LIII (p. 12.5) 



Angles, Arcs and Decimals of Degrees. (Based on Button's Mathematical 

 Tables.) 



This Table gives degrees in radians for the first two quadrants; it then gives 

 minutes and seconds from 1 to 60 in fractions of a degree and in radians. The 



