140 JOURNAL OP ECONOMIC ENTOMOLOGY [Vol. 2 



and ask us to give credence to their conclusions. Our recent experi- 

 ments have even more strikingly sho\^Ti this, and in his report the 

 writer has taken space to give the complete records of all the plots _ 

 sprayed the last three years. The variation between trees side by side 

 in the center of large plots is so considerable that the utter futility of 

 experiments with but one or two trees in a plot is at once apparent. 

 This is strikingly shown in the work of Lloyd (/. c, 388, 390), in 

 which for two years eleven trees were treated in as many ways and 

 the percentage of wormy fruit was greater where two heavy applica- 

 tions were given than but one application, in both years. 



Our records also show the necessity of comparing results in percent- 

 ages rather than by the number of wormy apples, unless the trees are 

 all of uniform size, evenly fruited, regularly placed, and bearing the 

 same year. 



The arrangement of plots must also receive more careful considera- 

 tion. Unsprayed check trees must be isolated and yet of sufficient 

 number to represent the orchard. The plots must be large enough 

 so that the central trees in each will not be materially influenced by 

 the next plot. In our own experiments we have endeavored to de- 

 termine the proportion of larvae killed in the calyx by eating the side 

 of the apple and by eating the foliage, and we have secured consid- 

 erable evidence on these points. But we have been forced to the con- 

 clusions that the orchards available are too small and are not uniform 

 enough to secure comparable results. Our records will be found to 

 be filled with inexplicable inconsistencies and contradictions, which 

 can be due to nothing else than the lack of uniformity in the trees, 

 as regards the year bearing, number of fruits borne and position. To 

 determine such points or the influence of any given spray on the 

 first and second brood, a large orchard is needed, all the trees of the 

 same variety, or of but few similar varieties, with plots of each variety 

 treated similarly, so that their similarity may be determined by the 

 comparison of those treated the same. The trees should all bear uni- 

 formly as to number of fruit and the year bearing. The plots should 

 contain 21, or preferably 35, trees and the fruit from five central trees 

 should be counted, both windfalls and picked fruit. As orchards of- 

 fering such conditions are not available in New Hampshire, we have 

 been compelled to forego further investigation of these problems. 



If we are to secure any results upon problems of this kind which 

 are at all conclusive and which Mill not be as readily contradicted, we 

 must arrange our experiments on a larger scale and with greater ex- 

 actions and care. The conditions surrounding such biological prob- 

 lems are so varied that too great care cannot be taken in arranging 



