Abstract
Many differences in the ring-width growth within a tree may be attributed to changing supplies of food and hormones. In moist sites or during periods of favorable climate, there may be sufficient food for the production of wide rings throughout the tree. But in dry sites or during years of low moisture and high temperatures, food competition within the tree is likely to be greater and the cambium at the base of the stem is likely to receive a limited food supply and may produce narrow rings. The cambium at the stem base depends upon the entire crown for food, hence ring growth reflects the tree's ambient climate. But, the cambium in the top of the tree or in the upper branches depends upon a more restricted portion of the crown for its food and hormone supplies. The rings produced by the cambium vary greatly from branch to branch and are less reliable indicators of the climate surrounding the entire tree than rings at the tree base. Therefore, ring series at the base of trees in semiarid sites provide the most reliable, as well as the longest, record of macroclimatic variation. Tree-ring widths in certain coniferous species growing in semiarid sites appear to represent the integrated effect of climate on food-making and food accumulation in the crown throughout the 14 to 15 months previous to and including the period of growth. Trees in warm, low-elevation sites may utilize winter moisture most efficiently; trees in cool, high-elevation or more northern sites may utilize early summer and early autumn moisture most efficiently. But even with these differences, a significant amount of variance is found to be common among tree-ring series from a wide range of sites, species, and geographical areas in western North America (14, 19, 34), emphasizing a common dependence of ring widths on the gross regional patterns of precipitation and temperature. The remaining variance, which is not correlated among sites, may be attributed to local en-vironmental and climatic differences, to variability among and within trees, and to compounding effects of occasional fires, insect or other infestations, and recurring years of high seed production. It is evident that a large portion of the variability in ring-width patterns from semiarid sites in western North America does reflect differences in climate from year to year. If ring chro- nologies are derived from a number of trees in semiarid sites and if adequate corrections for age and trend are made, these chronologies may be used to reconstruct a first approximation of annual, or somewhat longer period, climatic fluctuations in the past (19).