l8o MAINE AGRICULTURAL EXPERIMENT STATION. I914. 



Pearl and Miner" have recently pointed out that a deviation of 

 three times its probable error may be expected to occur, on the 

 basis of chance alone, more than 4 times out of a hundred trials. 

 So that a deviation of even three times the probable error can- 

 not be regarded as "certainly significant." 



Consequently it is not possible to say of any individual variety 

 that it is certainly more or less variable in the four year test 

 than in the one year test. However, the fact that all but two 

 varieties show a larger variability in the four year test and the 

 fact that the sum of the plus differences is more than twice the 

 sum of the minus dififerences tends to show that the results of 

 tests made in different seasons are on the whole more variable 

 than the results of tests made in a single year. As pointed out 

 above this is what might be expected. Very probably a larger 

 number of tests of each variety would show a significant dif- 

 ference in the coefficients. 



3. The Victor which gave such a small variation in the four 

 year test shows a very large variability in the 1913 plots. The 

 difference between these two tests is 7.81 percent and the prob- 

 able error of this difference is ±2.89 percent. The difference is 

 not quite three times its probable error and cannot be considered 

 as certainly significant. 



In the case of the Senator the variation in the 1913 plots is 

 again greater than that of the yearly production. In this case 

 the difference is 11.57^6.94 percent. Here, while the differ- 

 ence is large, it is less than twice its probable error and cannot 

 be considered as statistically significant. It is quite possible that 

 the plots of these two varieties were distributed in very unequal 

 soil in 1913 and hence that their variation is relatively much 

 larger than it ought to be. The results of future tests will help 

 to settle this question. 



4. The Imported Scotch shows practically the same varia- 

 bility in both cases. The difference between the two coefficients 

 is much less than its probable error. On the whole the Imported 

 Scotch has been much less affected by environmental conditions, 

 whether inter- or intra-seasonal, than any of the other varieties. 



"Pearl, R., and Miner, J. R. A Table for Estimating the Probable 

 Significance of Statistical Constants. Me. Agr. Exp. Sta. Ann. Rept. 

 1914, pp. 85-88. (Bulletin 226). 



