this evaluation, with numbers of harvests and 

 samples per harvest varjdng among years. 



An important point illustrated in our results is 

 that individual factors are correlated highly with 

 each other, i.e., are interwoven. For example, 

 changes in scald susceptibility associated with varia- 

 tions in average temperature from year to year are 

 correlated highly with changes associated with days 

 below 50°F, days above 86°F, rainfall, and sunshine. 

 Therefore, if we determine just the effect of aver- 

 age temperature, the value obtained will include 

 some of the effects of all the other variables, which 

 means that our results will exaggerate the effects 

 of average temperature. Conversely, if we deter- 

 mine the effects of average temperature and thenthe 

 effects of rainfall, the results may understate the 

 actual effects of both variables since some of the 

 affect of each is taken out through its correlations 

 with the other. Therefore, the statistical approach 

 cannot actually quantify the effects of variables, it 

 can only tell you how much variation in scald sus- 

 ceptibility can be accounted for by a series of mea- 

 surements. 



Overall, scald susceptibility among samples 

 varied from 0% to 100%, averaging 35% of fruit in a 

 harvest developing scald after 20 to 25 weeks in 

 32°F air plus seven days at 70°F. Some years had 

 greater scald susceptibility than others, and of 

 course, scald decreased with later harvest in a year. 



Table 1 shows the overall effects of low tempera- 

 ture on scald susceptibility of Massachusetts Deli- 

 cious apples. "R^" values can be interpreted as the 

 proportion of the scald variability among samples 

 that could be accounted for by a given series of 

 measurements. Average temperature after August 

 1 among samples accounted for 16%> (R^ = 0.16) of 

 scald variation, with lower temperatures decreas- 



ing scald. When we then took into account the num- 

 ber of days from August 1 to harvest in which the 

 temperature fell to 6°C (43°F), we accounted for an 

 additional 37% of scald variability, bringing the to- 

 tal to 53% (R^ - 0.53). If we made the temperature 

 cut-off 8°C (46°F) or 10°C (50°F) we accounted for 

 slightly less of the variability, but if we made the 

 cut-off at 12°C (54°F) we accounted for very little 

 variability beyond that picked up by average tem- 

 perature. 



What this means is that not only does low tem- 

 perature reduce scald susceptibility, but also that 

 specific low temperature events (days when it 

 dipped to 6°C, 8°C, or 10°C) had extra benefit in 

 making the fruit less scald susceptible. In the past 

 we have measured low temperature as "hours be- 

 low 50°F" but in our analyses here we found that 

 just counting days on which this happened gave 

 better results. The results in Table 1 show that it 

 does not make a great deal of difference whether 

 the cut-off is 43, 46, or 50°F. 



When we measured fruit maturity at harvest 

 using a starch-iodine test, we accounted for very 

 little extra scald susceptibility (Table 1). This does 

 not mean that maturity is unimportant, because 

 we know it is. What it does mean is that in Massa- 

 chusetts, the effect of maturity is very closely linked 

 to low temperature events, so that when you mea- 

 sure temperature effects, you are also including the 

 effects of maturity. Apparently, in Massachusetts 

 the typical decline in temperature during the Fall 

 coincides so closely with Delicious maturation that 

 you cannot statistically separate the effects of fruit 

 maturity (starch score) on scald susceptibility. 



In Table 2, the effects of rainfall after August 

 24 and sunshine after September 3 are added to 

 those shown in Table 1. These dates were chosen 



Fruit Notes, Fall, 1995 



