8!)() JOURNAL OF KORKSTRY 



the coefficient, which even in cases where the two kinds of values are 

 entirely independent, there will 1)e a certain amount of accidental 

 deviation of the coefficient from 0. The value of the coefficient (r) 

 is, therefore, ordinarily judged from its probable error (Er) which is 



found bv the formula $ = -6,745 1-r^ King (1914) gives the 



V 2n 



following rules for interpreting (r) : 



(1) If r is less than the probable error, there is no evidence what- 

 ever of correlation. 



(3) If r is more than six times the size of the probable error, the 

 existence of correlation is a practical certainty. 



There might be added to the above statements that in those cases 

 in which the probable error is relatively small: 



(1) If (r) is less than .30 the correlation cannot be considered at 

 all marked. 



(2) If (r) is above .50 there is decided correlation. Common sense 

 as well as the arithmetical relation between the coefficient and its 

 probable error must, of course, be used 'in drawing conclusions from 

 correlations." 



From the figures in Table 3 the coefficients were calculated as 

 follows : 



(a) Humboldt seed Ages classes M to M + 6 



r = -0.124 ± .372 



(b) Monterey seed Age classes M to M + 6 



r = -0.135 ± .18 



(c) Humboldt seed All age classes 



r = 0.65 ± .185 



In each case r being much less than the probable error there is evi- 

 dently no correlation between age of seed and germination per cent. 

 Calculations were made in the case of Humboldt seed for ages M to 

 M + 6 in order to eliminate weakness of the data in the older age 

 classes; also to base the comparison between the two lots of seed on 

 the same age classes. 



When the straight line averages for the points in Table 3 are calcu- 

 lated by the method of least squares, lack of correlation is again indi- 

 cated by the absence of decided slope in the line (fig. 1). 



The indicated decrease in total germination per cent with increasing 

 age of seed is, according to such calculations, as follows : 



Monterey seed .9 of one per cent per year 



Humboldt seed 0.19 of one per cent per year 

 (Ages M to M-l-6) 



Humboldt seed 0.43 of one per cent per year 

 (All ages) 



