250 E. E. FREE 



series are to be compared. For the partial escape of these 

 difficulties two methods are in vogue. The first is the method 

 of classification, by which the plants are placed in certain roughly 

 defined classes, as for instance, " excellent, " "good," "fair," 

 "poor," etc. The second is the method of ranking, by which 

 the plants are arranged in order of goodness, the best at the top 

 of the list, the poorest at the bottom. An outgrowth of this 

 second method is the assignment to each plant of a number 

 (usually a percentage of "perfect" goodness) representing its 

 relative position in the series. In certain cases, as, for instance, 

 in scoring plants in agronomic judging, this number is made up 

 of certain numerical units assigned to easily determinable ele- 

 ments of condition according to an arbitrary scale. 



Both the method of classes and the method of ranking suffer 

 from an important inherent difficulty, — their dependence for 

 accuracy upon the observer retaining constantly in mind the 

 character and conditionof all plants of the series at the same time. 

 With series of a few plants only this is not difficult but with scores 

 or hundreds to compare it becomes practically impossible. 

 Even the agronomic method of numerical scores for arbitrary 

 unit features, suffers from this difficulty when the system of 

 scores is first decided upon. It is practically useful, therefore, 

 only when comparisons continue to be made on substantially 

 the same basis and under substantially the same conditions. 

 Because of the initial work necessary in the preparation of the 

 score card the method is of no service in the numerous cases of 

 comparisons on new or changing criteria which one is constantly 

 called upon to make in the course of much biological research. 



For nearly ten years the writer has been accustomed to meet 

 this difficulty by a method derived from that used by the psychol- 

 ogists in the investigation of affections and which enables the 

 formalization and partial quantitative expression of compara- 

 tive judgments formed upon any criteria whatsoever. This 

 method has proven so useful to the writer and associates in so 

 many and so varied researches and seems so little known to 

 investigators in experimental biology that it is here presented, 



