04 mains; agriculturai, kxpe;rime;nt station. 1909. 



If, on the average, the exceptionally good mother produced 

 the exceptionally good daughters, and the exceptionally poor 

 mother the exceptionally poor daughters it would be expected 

 that the two lines in each of these two figures would run, in the 

 main, parallel to each other. No such a parallelism is apparent, 

 however. The two lines in each of the figures zigzag up and 

 down in a manner quite independent of each other. The 

 daughters' line is as liable to go up when the mothers' line is 

 going down as to do the opposite. These diagrams clearly do 

 not give any evidence that egg producing ability was inherited 

 in this experiment. 



Resort must be had, however, to a more exact appreciation 

 of the correlation between mothers and daughters in respect to 

 egg production. Such correlation may be accurately measured 

 by proper mathematical methods. It is not necessary to enter 

 into a discussion of these methods at this place.* 



In general the procedure is as follows : A statistical table 

 known as a "correlation table" is made showing the egg produc- 

 tion of each individual daughter and each individual mother. 

 From this table is calculated the so-called "coefficient of corre- 

 lation" which measures the average degree of association 

 between mother and daughter in respect to the character under 

 discussion. This coefficient of correlation is of such character 

 that it may take values only between the limits of zero and i. 

 When the correlation coefficient is equal to zero it means that 

 there is absolutely no association or correlation between the 

 characters under discussion. That is to say, in the present case 

 a correlation coefficient of zero would mean that an exception- 



* Descriptions of modern biometrical methods may be found in any 

 of the following works : 



1. Pearson, K. Grammar of Science. Second edition. London (A. 



& C. Black) 1900. Pp. xviii-l-548. 



2. Elderton, W. P. Frequency Curves and Correlation. London 



(Layton) 1907. Pp. xiii-)-i72. 



3. Davenport, C. B. Statistical Methods with Special Reference to 



Biological Variation. Second edition. New York (J. Wiley & 

 Sons) 1904. Pp. viii+223. 



4. Davenport, E. The Principles of Breeding. Boston (Ginn & Co.) 



1907. Pp. xiii-f 727. 



5. Davenport, E. and Rietz, H. L. Type and Variability in Corn. 



111. Agric. Exp. Sta. Bulletin No. 119, pp. 1-38. 1907. 



