Chance selection and sample size appear to contribute much to the conflicting 

 results of the various studies. The vigor-quality study generally shows a marked con- 

 trast between low- and high-elevation stands, but it mainly samples only two stands and 

 only four trees in the high one. Subsequent results from the natural selection study 

 where 50 trees were sampled per stand confirm the differences among the low and high 

 stands represented in the vigor-quality study, but also show that other choices of single 

 low and high stands could lead to the conclusion that low and high stands do not differ. 



On the other hand, the elevational study sampled a much broader range of stands but 

 again with few trees per stand. Consequently, the lack of differences among low and high 

 stands may be too conservative. For example, two 5-tree collections were obtained from 

 a stand at Bertha Hill at 1,585 m in the Clearwater area for the elevational study. 

 Seedlings from one of these collections were the tallest of the Clearwater collections 

 but those from the other were about average. The mean height of the two was only 

 slightly less than that for the best elevational zone. However, the results from the 

 50-tree collection made for the natural selection study indicate that although seedlings 

 from this stand are among the fastest growing of the high-elevation stands they are 

 slower growing than those of five of the six low-elevation stands. Among the low- 

 elevation stands in the natural selection study, even a 50-tree sample indicated only 

 that seedlings of the poorest stand differed significantly from the best ones. 



The results of this group of studies also differ from those found in other western 

 conifers. Callaham and Liddicoet (1961) found that mid-elevation ponderosa pine families 

 were tallest at low-, mid-, and high-elevation sites during the first 20 years after 

 outplanting. However, although the high-elevation trees were doing very poorly at the 

 low- and mid-elevation sites at age 20, at the high site they were taller than the low 

 trees and nearly as tall as the mid trees. This pattern continued through age 29 

 (Conkle 1973). Conkle also found that the variance component related to elevation zones 

 was much larger than that for families within zones at the low and mid sites but not at 

 the high site. In our studies the family component was always much larger. 



In working with Douglas-fir, Rehfeldt (1974) found that differences between low, 

 or warm, habitat type trees and high, or cool, habitat type trees could be detected with 

 seedlings from as few as 7 trees per stand. He also found growth differences between 

 stands on the east and west sides of the mountain ranges along the border between Idaho 

 and Montana. 



In summary, the results from this series of tests indicate that, at least within 

 the north Idaho portion of its interior range, western white pine has a different varia- 

 tion pattern than that of most other conifers that have been intensively studied. 

 Western white pine is highly variable, but most of the variation is related to differ- 

 ences among the offspring of a single tree or among the trees in a stand. Although 

 differences among stands, elevation zones, or geographic areas were sometimes signifi- 

 cant, the proportion of the variance attributable to these sources was usually smaller 

 than that for trees within stands. 



All of the seedlings involved in the studies reported here were started under pro- 

 tected and irrigated nursery conditions. They escaped the variety of environmental 

 conditions that might lead to ecotypic differentiation during the critical germination 

 and early establishment growth phases often found when reproduction occurs naturally. 

 Squillace and Bingham's (1958) germination test results may indicate such a differenti- 

 ation, but the results reported here do not show ecotype formation for other traits. 



19 



