PREHAR\'EST APPLE DROP 13 



tween per cent drop and yield ; therefore we cannot give an estimate of 

 the number of bushels dropped unless regression for weight is known. 

 A lack of correlation or a correlation coefficient of zero between 

 per cent drops and yield could mean either that the per cent drop was 

 the same for all trees or else that it was high for both heavy-and light- 

 yielding trees and lower for intermediate yielders, or vice versa. This 

 is further evidence of the greater value of actual weight rather than 

 a percentage relationship in evaluating the change in amount of 

 drops with change in tree yield. 



Likewise, with Mcintosh located in the part of the orchard 

 known as the BF block, there was no correlation betw^een per cent 

 drop and yield from 1935 to 1941, yet there was a correlation of .953 

 between weight of drops and yield. This is nearly a perfect correla- 

 tion and. estimated from the regression equation, Mcintosh trees 

 bearing 1000 pounds of fruit would drop 360 pounds and those yield- 

 ing 3000 pounds would drop 869 pounds of fruit. On this basis it may 

 be seen that each tree dropped the equal of one year's crop during a 

 period of 4 years. From the grow-er's point of view, such data, ob- 

 tained over a long period of time, would be of considerable 

 significance. 



The correlation between per cent drops and yield on different 

 harvest dates for individual trees and in individual years is shown in 

 Table 9. The correlations in general are higher than those based on 

 the total crop harvested during the season. There is less variability 

 in the correlation between per cent drops and yield on the later-pick- 

 ing dates. There is also a high correlation between per cent drop and 

 yield on the trees harvested on the later dates. This indicates that 

 heavy yield becomes a more important factor in dropping as the sea- 

 son advances, that it is highly important to consider the maturity of 

 the fruit and tree in studying the relation of 3'ield drop, and that the 

 individuality of different trees must not be overlooked. 



Table 9. Correlation Between Yield and Percentage Drop on 

 Different Harvest Dates 



* Significant ** Highly Significant 



It should be understood, therefore, that with increased yield there 

 would be more dropped fruit to handle, and increased expense in 

 handling the crop. Therefore, the results of these calculations are 

 not to be interpreted as an urge to disregard drop when there is no 

 correlation or when there is a negative one between per cent drop and 

 vield. 



