New York Agricultural Experiment Station. 



129 



imsprayed rows as in this experiment they have little influence 

 on the results. The difference between the average of the 

 five unsprayed rows and the average yield of the five rows sprayed 

 five times may be safely taken as the result of the spraying. 



Yield by series. — The five rows sprayed three times constitute 

 Series I and the average yield of these five rows make the yield of 

 Series I. The yields given for Series II and III have been com- 

 puted in the same manner. The yield by series is shown in the 

 following table : — 



Table IV. — Yield by Series at Geneva. 



Increase in yield due to spraying three times, 88 bushels per acre. 

 Increase in yield due to spraying five times, 118 bushels per acre. 



The difference in the appearance of the foliage on Series I and II, 

 although somewhat greater than in 1902, was, nevertheless, quite 

 small and seemed insufficient for a difference in yield of 30 bushels 

 per acre. 



The impossibility of correctly estimating differences in yield by 

 the eye is shown even more strikingly in Plate IX, which is from 

 a photograph of Rows 11, 12 and 13 as the tubers lay on the ground 

 after digging. The average observer would say that the yield on 

 the row sprayed every two weeks was certainly somewhat greater 

 than that on the unsprayed row, and that, probably, the row sprayed 

 three times would outyield the unsprayed row by a few bushels per 

 acre ; but only one having much experience in making such estimates 

 would expect to find the differences anything like as great as they 

 really were. As a matter of fact the row sprayed every two weeks 

 outyielded the unsprayed row by 123 bushels per acre; and the row 



