424 



THE AMERICAN XATURALIST 



[Vol. XLV 



tive, is of great value not only for the interesting data regard- 

 ing variation which it presents, but also in indicating clearly the 

 rich reward which may be expected to follow a combined experi- 

 mental and biometric attack upon the fundamental biological 

 problem of the effect of stock on scion. 



In the papers so far discussed there has I n in every case some 



attempt at biometric analysis of the raw statistical data. There 

 are constantly appearing in agricultural literature papers in 

 which a great mass of first-class statistical material on variation 

 and correlation in agricultural plants is presented but not 

 analyzed biometrically, or only incompletely so. Examples of 

 this are found (to mention but two) in the interesting papers of 

 Kohler on potatoes and Westgate on alfalfa. A conspicuous 

 instance of failure to make profitable use of elementary bio- 

 metrical methods is seen in the paper of Stockberger and 

 Thompson on hops. These authors put their data in form for 

 calculating variation and correlation constants (e. g., they give 

 a correlation table for the correlation between number of vines 

 to the hill and yield per hill) but do not determine the constants. 



It is evident from what has preceded that biometrical methods 

 are rapidly gaining a place among the agricultural investigator's 

 working tools. Keeping always in mind the caution expressed at 

 the beginning of this article that biometric zeal be not allowed 

 to outrun biological discretion this movement merits only com- 

 mendation and further encouragement. The agricultural investi- 

 gator has an almost unique opportunity to make significant and 

 profitable application of biometric methods of research. 



Raymond Pearl. 



