'.'/hen the raw data for the first-year height were analyzed, the mean squares 

 associated with areas were highly significant. This was true for both comparisons, 

 for example, between Clearwater and Trestle where the broadest elevational range was 

 sampled, and among all areas but utilizing only the 915 m, 1,065 m, and 1,220 m eleva- 

 tion zones (those common to all areas). The early germinating Trestle seedlings were 

 taller than the later germinating seedlings from the Clearwater and Kaniksu areas. 

 However, when first-year heights were adjusted for seed weight and germination time 

 the variability related to areas was much reduced. The comparison of all areas using 

 the middle el6vational zones showed no significant differences, but the Trestle Creek 

 and Clearwater areas differed at the 5 percent level when all elevation zones were 

 considered. These adjusted means differed by only 2 mm- -approximately 4.5 percent of 

 the mean. 



Subsequent height measurements taken in the second and third years followed a pat- 

 tern similar to first-year height. Analyses of the raw data showed significant differ- 

 ences among trees within stands and among areas. There were few or no significant dif- 

 ferences among elevations within areas or stands within elevations. When the data were 

 adjusted to take into account first-year height differences, second- and third-year 

 height differences among areas were not significant. In other words, the Trestle Creek 

 seedlings, which germinated early and were tallest at age 1, were still tallest at age 3 

 but relative to their height at age 1 they had not grown as much as seedlings from the 

 Clearwater and Kaniksu areas. 



Whether such adjustments should be made depends on the cause of the germination 

 time difference. If it is the result of genetic differences, no adjustment should be 

 made. If the difference resulted from cone collecting or seed processing timing or 

 techniques, then the adjustments would be appropriate. In the absence of information 

 to indicate genetic differentiation of germination time, I favor use of the adjusted 

 data. 



If the range of 120- to 180-m elevational zones is combined across areas to form 

 three broader zones called low, mid, and high (<1,000 m, 1,000-1,250 m, >1,250 m) like 

 the classification in the following studies, then seedlings of the low zone were tallest 

 and those of high zone were shortest. At age 3, heights by zone were: low - 220 mm, 

 mid - 211 mm, high - 207 mm. These means did not differ significantly. 



Secondary leaf length (mature third-year leaves) was the only other trait in which 

 differences among areas were significant. Within the Trestle Creek and Clearwater 

 areas, differences between trees within stands were significant but for the other areas 

 they were not. Stand-to-stand differences within an elevational zone were sometimes 

 large but seemed to occur at random. Significant cotyledon dif f erences--number and 

 length--were confined to trees within stands and stands within elevational zones. 



Bud burst at the start of the third year did not differ significantly among stands, 

 elevations, or areas (table 3) . The highest elevation source from Trestle Creek was 

 the first to break bud but the next to highest source was last. All buds broke within 

 a 2-week period and most within a 5-day period. The newly emerging leaves were rapidly 

 separated and in from 10 to 14 days the tip of the new bud was distinguishable. At 

 that time the seedlings were classified as having set buds even though stem elongation 

 and external bud development continued for nearly a month. No significant differences 

 in bud set time were observed except among trees within stands. 



The seedlings were 7 years old and had been in the field plots for 4 years when 

 they were last measured. Overall survival of seedlings of Clearwater origin was about 

 10 percent less than that of Trestle seedlings with the differences greatest at the 

 Canyon series of test sites. However, survival differences so far do not clearly indi- 

 cate differences related to area or elevation of origin or their interactions with test 

 site conditions. Average height ranged from 54 cm at the best site, low PREF, to 34 cm 

 at the poorest, high PREF (table 4). 



8 



