April, '08] JOURNAL OF ECONOMIC ENTOMOLOGY 133 



between the check trees and the balance of the orchard, which was un- 

 sprayed, and the sprayed portion, the moths of the second brood mi- 

 grated to the sprayed trees and those plots nearest the unsprayed por- 

 tion showed very little benefit to the second brood, whereas the benefit 

 increased with the distance from the unsprayed trees. The effect of 

 these factors was very noticeable in studying the results. We then 

 happened to remember Dr. Forbes' suggestive paper given before this 

 association a few years ago, in which he showed the way in which the 

 plum curculio migrated from the untreated part of the orchard into 

 the part treated, and the necessity for leaving out of consideration a 

 few rows of trees between the untreated and treated part of the or- 

 chard. We therefore decided that in making future experiments we 

 would leave one end of the orchard unsprayed for checks, spray sev- 

 eral rows across the orchard next to the checks in the best possible 

 manner, calling this portion of the orchard the barrier plot, and would 

 then lay off our plots at right angles to this barrier plot so that the 

 influence of its effect upon the sprayed plots would be equal in all of 

 them. Happily at this time Prof. Quaintance and the writer met to 

 discuss methods of work upon this subject, and it is to him that I am 

 indebted for the suggestion that we make our plots three rows wide 

 and count only the middle row, thus having 15 trees in each plot, the 

 outer rows of which tend to reduce the influence of one plot on an- 

 other. Our work this season has shown not only the absolute necessity 

 for such an arrangement, but that it would be wise to go even further 

 and have the plots contain 35 trees each, 5 rows wide and 7 rows long, 

 and count the central five trees so as to better reduce the influence of 

 the neighboring plots. It is of the utmost importance in making an 

 experiment to give any exact results on the codling moth, that the trees 

 be of a uniform size, fruit evenly, and have borne approximately the 

 same the previous year. A few trees scattered thru an orchard which 

 have borne the previous year when the balance of the orchard did not, 

 wiU seriously affect the results of the work the following season. From 

 careful study of our records, it seems to me that too much importance 

 cannot be placed upon the ground plan of such an experiment, and I 

 cannot but feel that experiments based on individual trees scattered 

 thru an orchard are of little value in trying to determine the amount 

 or nature of the effect of spraying upon the different broods of the 

 codling moth. Furthermore at least five trees must be counted in each 



Fig. 3. Diagram of orchard of Albert DeMerit, Durham, N. H., used in 

 experiments of 1906. Circles represent trees. Circles with crosses are trees 

 which bore in 1905, but bore practically no crop in 1906. Solid lines show 

 boundaries of sprayed plots; remainder of orchard unsprayed. 



