THE FLOWERING OF APPLE TREES 143 



extent in days of the blossoming period, and D, the average date 

 to which the different varieties lagged behind the earliest blos- 

 somer, the results show considerable differences in different 

 seasons, notably so in 1908 and 1909, and it was found impossible 

 to connect these differences with any of the special meteorolo- 

 gical conditions of the seasons in question, so that they must 

 be ascribed to irregularities in the behaviour of the varieties 

 themselves. 



It is a question of practical importance, as well as of interest, 

 how far the relative earliness or lateness of flowering of any 

 particular variety is a constant quantity, and how far it varies 

 according to the season. That some varieties are habitually 

 early or late flowerers is accepted as a matter of general knowledge, 

 but it has never yet been shown how far such properties may be 

 relied upon as constant, or to what extent the bulk of varieties 

 exhibit distinctive characteristics in this respect. 



According to the analysis on p. 142, the Woburn records indicate 

 a considerable amount of variation in the behaviour of different 

 varieties in different years, and these records may be further 

 examined from this point of view, for they supply, in the case of 

 each variety, ten different pairs of years in which the relative 

 dates of flowering can be compared. The following values were 

 obtained for the average difference between the relative dates of 

 flowering of individual varieties in any two years : 



A verage 



Difference. (Min. and Max.) 



Early . . . .2*5 (o'8 4-6) 



Mid-Season . . . 2'8 (o'8 4*4) 



Late . . . .2-4 (0-6 1 6'6) 



Average . . .2*55 (0*6 6*6) 



The values did not seem to be dependent at all on the class of 

 apple examined, whether dessert or cooking, late or early, so that 

 the average results need alone be considered, and these show that, 

 if any variety flowers on a given relative date in one season, the 

 probability is that it will flower within 2*55 days of that date (either 

 earlier or later) in any other season : the difference, of course, 

 may be much less, or much greater, the actual differences found 

 with individuals in pairs of seasons ranging from o to 13 days, 

 but rarely exceeding 8 or 9 days. 



1 One variety in this case gave a difference of o, and another a 

 difference of 0-5, but as only two seasons of flowering were available 

 in those cases, they have been omitted. 



