124 THE CLIMATIC FACTOIl AS ILLUSTRATED IN ARID AMERICA. 



ating importance, but as a matter of fact they play by no means so great a role as would 

 l)e expected, for, as Professor Douglass has shown, two or three trees, or even a single 

 tree, under exceptionally favorable conditions, gives a fairly accurate climatic record but 

 little disturbed by accidents. Finally, the fourth reason for variation is the changing 

 conditions of weather and climate which prevail from year to year. It is these which we 

 wish to determine, and this can be done only by eliminating variations due to the other 

 three causes. Let us therefore turn to the problem of how this elimination is to be accom- 

 plished. As accidents are the matter which critics of the method here discussed are most 

 likely to emphasize, let us take them up first. 



The elimination of the effect of non-cUmatic accidents upon the rate of growth of trees 

 is accomplished largely by the process of averaging. If a sufficient number of trees is 

 used, and if the trees are distributed over a wide and varied area, purely individual accidents 

 will disappear by the law of averages. Wliere a large group of trees is concerned, each year 

 — and still more each decade — is characterized by about the same number of cases of the 

 slipping of the soil, the crashing of one tree into another, the eating of roots or buds by 

 rodents, and the many other Uttle accidents which are continually checking and some- 

 times stimulating the growth of vegetation. The combined effect of all these accidents is 

 a nearly constant quantity. Assurance that this quantity is actually constant can be 

 obtained only by using a sufficiently large number of measurements distributed over a 

 sufficiently wide area. Another type of accidents, such as the ravages of insects and of fire, 

 can not be gotten rid of quite so easily, since they are more widespread and prevail much 

 more abundantly in some years than in others. Even with these, however, the effect can 

 be reduced to a small amount by means of abundant and widely scattered measurements. 

 Moreover, when the ravages either of insects or of fire are unusually widespread and 

 become regional instead of purely local phenomena, the cause is almost always found in 

 unpropitious conditions of climate. Hence, where the effects of such ravages can not be 

 eliminated, they will only rarely be found to mask the effects of cUmate or prove opposed 

 to them. As a rule they merely intensify the retardation of growth which normally 

 accompanies times of unfavorable conditions of climate. The case is even stronger than 

 here appears, but I shall defer further discussion of it until we have some actual curves 

 before us and can discuss the matter in relation to them. 



The elimination of the differences in rate of growth due to the fact that young trees 

 grow more rapidly than older ones is easily made. A glance at the curves prepared by 

 Professor Douglass and presented in the preceding chapter will show that the earlier 

 portions are much higher than the later parts. This, as is ah-eady apparent, does not 

 represent a climatic difference, but is merely due to the fact that trees grow rapidly in 

 their youth. Manifestly allowance must be made for this varying rate of growth. I have 

 called this allowance the "corrective factor for age." The method of obtaining it is 

 illustrated in the accompanying diagram, figure 26. Let the horizontal line represent the 

 course of time as indicated in years by the figures 10, 20, 30, 40, etc. Let the vertical 

 distance indicate the average thickness of the ring of wood added each year. Suppose 

 that we have 100 trees varying in age from 50 to 200 years. Let us suppose further that 

 we have averaged up the rate of growth of all these trees during the first year of their 

 lives and find that it amounts to one-tenth of an inch. In the same way we find that the 

 growth during the tenth year amounts to 0.15 of an inch; during the twentieth year 0.175; 

 during the thirtieth, 0.19; and the fortieth 0.20. After the fortieth year the rate of growth 

 begins to diminish until at the one-hundredth year it has fallen to a figure no larger than 

 that of the first, while at the two-hundredth it has fallen still lower, to 0.05 of an inch. 

 Manifestly it is an easy matter to plot a curve from these figures. The curve will rise 

 rapidly at first, as appears in figure 26, and then will fall more and more slowly. Such a 

 curve, when plotted, will not be perfectly regular, but will be somewhat wavy, as shown 



