Regressions of freezing tolerance for phase two were also significant (1 percent 

 level) for each population. Values of i?^ ranged from 0.45 to 0.81. Figure 4 illustrates 

 the variety of observed responses; but, only the response curves of maximum contrast 

 differ significantly (5 percent level) for either r or b. 



20 40 60 80 100 



SEVERITY OF FREEZING TREATMENT (PERCENT) 



Figure 4. --Response curves of freezing damage according to severity of the freezing 

 treatment for three populations during the second phase of cold acclimation. 



The curves in figures 3 and 4 express a greater range in response and a consequent 

 higher variance among populations for cold tolerance during phase one than during 

 phase two. Previous studies [Rehfeldt 1979) have also suggested that population 

 differentiation is readily detected for acclimation during autumn, but that relatively 

 little genetic variance exists among populations for hardiness during deep winter 

 dormancy. In concurrence with previous studies (Rehfeldt 1979) , the curves in both 

 figures that illustrate maximum tolerance to freezing depict populations from cool 

 subalpine environments at relatively high elevations (above 1,300 m) . 



7 



