8 

 / 



half gallons of water. The arse^iic solution was made by boiling 

 one ounce of arsenic in one quart of water, and adding this solu- 

 tion to twenty gallons of cold water. The method of procedure 

 was precisely as last year, the trees being thoroughly sprayed with 

 a hand force-pump, and with the Deflector Spray and Solid Jet- 

 Bose Nozzle manufactured by the Lowell Faucet Company, Low- 

 ell, Massachusetts. The fluids were thrown in a fine mist-like spray, 

 applied until the leaves began to drip. 



The character of the season was the precise opposite of that of 

 the year preceding. 'Then a scant crop followed an abundaxit one; 

 now an extraordinarily abundant crop followed a very scanty one. 

 The ratio of insects to the apples exposed to their attacks must 

 have been, consequently, doubly diminished as compared with that 

 of 1885. Further, our experiments were last year limited to 

 early apples, no late fruit being available; while this year only 

 fall and winter varieties were used. In these, finally, an addi- 

 tional brood of the codling moth probably developed, making 

 three in all for the season; while last year we certainly had but 

 two up to the time of harvesting our experimental fruit. 



After spraying, the apples which fell, both from the poisoned 

 trees and the untreated checks, were regularly picked up at inter- 

 vals of about two weeks, until October 4. These lots were each 

 examined for codling-moth injuries, as were also the picked apples 

 gathered from the trees October 4. The entire number of apples 

 examined from all the trees was 38,688. 



From the data thus recorded tables were prepared, such as I 

 published in my last year's report; and upon these tables the 

 present discussion is based. As this year's experiments are really 

 continuous with those of the year preceding, I have brought my 

 last year's results into this discussion, pi ef erring to present the 

 whole subject in a complete and finished form, even at the expense 

 of some slight repetition of matter already published. 



I have made principal use of the figures in the column marked 

 "Corrected" — a matter requiring a somewhat elaborate explanation. 

 Ratios are comparable, of course, only when reckoned on the 

 same base. For the check trees, the ratios of damage to the fallen 

 apples were calculated on the total number fallen, and for the 

 picked apples the basis of the reckoning was the number remain- 

 ing on the trees at the end of the season. Since the treatment 

 of the experimental trees kept from falling a certain number of 

 the apples which would otherwise have fallen, it is evident that 

 ratios computed on the fallen and picked apples from these trees 

 are not comparable, without correction, with those obtained from 

 the checks. The number of fallen apples being relatively smaller, 

 the ratios of injury calculated on that number will be too large; 

 while witli the picked apples the case will be reversed. The ratios 

 were consecpiently corrected throughout by the application of for- 

 muhe printed in our Bulletin 1, but which it is not necessary to 

 repeat here. 



