- 6 - 



When results of all 4 years of data were assessed, it was apparent that 

 this approach was reasonably accurate when only low scores (which indicate 

 excellent storage potential) and high scores (which indicate very poor 

 storage potential) were considered. However, it did not provide useful 

 information about samples which had intermediate scores. 



Our second approach was to produce equations that described the rela- 

 tionships between mineral composition and keeping quality in mathematical 

 terms. Initial indications were that this was the better approach ( Proc 

 MFGA, cited above). When the 4 years of data were fully evaluated, this 

 initial assessment was reaffirmed. The equation that we developed was a 

 more accurate predictor than was the scoring method we used. Furthermore, 

 it could be used to predict keeping quality from any mineral composition, 

 not from just the very high and very low ranges. 



Our results showed that because Ca was the only element that was 

 generally affecting keeping quality of the apples, a mineral analysis was 

 needed only for Ca to make a prediction. The other elements in our survey-- 

 N, P, K, and Mg--did not need to be measured for our fruit. 



In a separate study we tested the value of including measures of fruit 

 maturity and fruit size in a predictive equation, since storage life is 

 reduced as apples become large and are harvested at a later maturity. This 

 study indicated that inclusion of a score for starch concentration in the 

 apple and inclusion of a measure of fruit size at harvest both could improve 

 the accuracy of a prediction based on Ca concentration in the fruit ? weeks 

 before harvest. 



To be able to predict potential keeping quality of commercial lots of 

 apples at harvest from mineral analyses, there must be the capability for 

 gathering, preparing, and analyzing representative fruit samples rapidly. 

 Our studies are all based on sampling apples 2 weeks before harvest. If 

 fruit samples could be taken earlier in the growing season it would allow 

 more time for analysis, but research in England has shown that the earlier 

 fruit samples are taken, the less likely they are to represent the mineral 

 content at harvest. Leaf analyses cannot substitute for fruit analyses. 

 While they are good measures of mineral status of trees, leaf analyses are 

 of little value in portraying the mineral status of fruit, especially that 

 of Ca . Thus, use of prediction capabilities requires the ability to obtain 

 results from a testing laboratory in a very short period of time--within 2 

 weeks after taking the samples from a tree. At present no laboratory in 

 Massachusetts is capable of doing this. 



This has presented a dilemma to us. On one hand, we know that we can 



predict with acceptable accuracy how well Mcintosh will keep after harvest, 



but we do not have the facility for producing the needed analyses. On the 



other hand, we are working with a rather simple problem: Ca deficiency is 



the basis for the prediction. Should our efforts be directed to 



establishing a facility for analyzing fruit and predicting storage life, or 



would it be better for us to continue to pursue methods of avoiding and 



correcting Ca -deficiency? To this point we have chosen the latter course, 



and have continued to try to improve preharvest and postharvest treatments 



to ensure that fruit possess sufficient Ca at harvest for maximum posthar- 

 vest life. '^ 



