470 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1955 



350 years. Maximum ages are generally less than 600 years north- 

 ward from the 39°-40° belt to the Canadian border, are less than 500 

 years in southern British Columbia and at Banff, and may be no 

 more than about 400 years near the northern limits for the species in 

 central British Columbia and at Jasper Park. 



Much less complete sampling of species other than Douglas fir 

 does not allow more than a suggestion of the age distribution pattern. 

 One limber pine maximum seems to occur near 44° in eastern Idaho, 

 the piny on pine in the 39°-40'' belt, like the Douglas fir maximum, 

 and the ponderosa pine in the 37°-38° belt in southern Utah; all 

 species seem to show a marked decline in maximum ages southward 

 and to a lesser degree northward of the respective belts of maxima. 



Climatic histories. — In the Scandinavian Arctic, ring indices up 

 to five centuries in length and with correlation coefficients of the order 

 of +0.7 against observed growing-season temperature are obtainable, 

 particularly at or near the northern tree limit; in the drier areas 

 of the western United States and southwestern Canada, ring indices 

 up to 1,000 years in length and with coefficients of +0.7 to +0.8 

 against total yearly rainfall ending in June or July are obtainable, 

 particularly at the lower, or dry, forest margin. The significance of 

 the ring record in upper timberline trees of midlatitudes is not yet 

 entirely clear. For more stress-free areas, such as the eastern United 

 States and central Europe, variant conclusions have been reported 

 as to the climatic significance of ring growth — e. g., no relation, fair 

 relation to the annual number of rainy days, pronounced effects in 

 years of physiological drought, and moderate relation to rainfall of 

 certain months. However, C. J. Lyon and others have shown that 

 in such mesophytic areas the basis for significant ring histories does 

 exist in the fair degree of cross-dating which may be found in selected 

 species and trees. 



Wlien properly analyzed, the sequences of ring width in over-age 

 drought conifers may provide long and highly significant extensions 

 into the past of the gage records of rainfall and runoff. 



It must be emphasized, however, that the correlation observed for 

 recent decades between growth and rainfall may be applied only with 

 diminishing assurance to successively earlier centuries of growth index. 

 It has already been noted that the effect of hypothetical secular trends 

 in climate, such as a steady decrease in mean rainfall by as much as 5 

 percent during the life of a tree many centuries old, cannot at present 

 be separated from the tree's decreased radial growth as a function of 

 age. Other centuries-long, nonclimatic effects on growth rate, such as 

 might conceivably result from radical long-term fluctuations in the 

 activity of the soil micro-organisms, must remain as a source of uncer- 

 tainty which only the most widespread sampling can in part reduce. 



