274 NEW JERSEY STATE AGRICULTURAL 



A METHOD FOR STUDYING CORRELATIONS OF PEAS. 



The method used in the study of correlations among peas was 

 as follows : 



1. A fair number of established commercial varieties was 

 grown. 



2. A number of characteristics not common to all, or not of 

 the same degree in all, were tabulated. (See Biometric list of 

 Commercial peas). 



3. The taljulated characteristics were placed together on a 

 separate card for each variety. 



4. The cards were collected into groups according to natural or 

 artificial divisions of a characteristic, as for instance, colors, 

 range in height, etc. (It is necessary to emphasize the importance 

 of having each division represented by a fair number of the varie- 

 ties, depending upon the characteristic and the total number of 

 varieties under observation). With one or two exceptions, no 

 division in the peas contained less than one-fifth the total num- 

 ber of varieties. 



5. For each division of a characteristic a table of the extremes 

 of variability was made out for each of the remaining charac- 

 teristics in that division (See Correlation table). For example, 

 a character, as height of mature plant, is divided into four sizes, 

 namely, (i) 1-2I/I feet; (2) 3 feet; (3) 3y2-4 feet, and (4) 

 4V2-5V2 feet. Under each of these four divisions are given the 

 extremes of the remaining characters associated with that size. 



6. After the divisions of all the characteristics are thus tabu- 

 lated, any existing correlation among them can be found by com- 

 paring the extremes of any characteristic with the corresponding 

 divisions of another charact-eristic ; thus, the first statement in 

 correlations, that range in height increases with the maturity of 

 plant but decreases towards the end, was drawn from the com- 

 parison of the extremes in heights corresponding to the divisions 

 of maturity period of plant. 



