could be identified at once in terms of the rings at Prescott; the narrow ring of 1851 was seen to correspond to one in the Prescott series. The compressed series from 1879 to 1885 likewise had its counterpart at Prescott and formed the portion of the sections which gave the most difficulties in identification. On the whole, so far as can be judged without minute study, the Prescott trees from relatively high elevations approximating the elevation at Flagstaff have a considerably closer resemblance to the Flagstaff sections than do those growing at lower altitudes.
Cross-identification and climate. — The process of cross-identification appears to be applicable to areas far removed from one another, but as the distances increase the resemblances between tree-growth records decrease, due to climatic differences. The correspondence between trees in different regions thus becomes a test of climate and we note a possible field for the application of this process in the delineation of similar climatic areas or meteorological districts. It seems to the author that in this way the growth of vegetation may easily be made of fundamental value in practical meteorology.
MONTH OF BEGINNING ANNUAL MEANS.
It is evident that it must take some time for the transmutation of rain into an important part of the organic tissue. There is evidence, as will be shown later, that the summer rains often have a prompt effect. The winter precipitation, however, is necessarily more remote in its action. Much of the first growth in the spring must come from the melting of the autumn and winter snows. It seems reasonable, therefore, to consider any snowfall as applying to the following yearly ring. At Flagstaff the precipitation of November is almost always in the form of snow, and therefore that month should certainly be considered as falling after the arboreal New Year of that locality. In view of the uncertainty as to the exact month when the precipitation begins to have an influence upon the growth of the following season, and of probable variations in different years, it seemed wise to test the matter by a purely empirical method. The annual rainfall was ascertained for yearly periods beginning (1) with July 1 of the preceding year, (2) with August 1, and so on to (9) with March 1 of the current year. Another method involved a separating of the summer rains, one-half to apply on each adjacent winter, while a final method involved a similar division of the winter rains. This was done for 12 years at Flagstaff and 43 at Prescott. Part of the Flagstaff curves are given in the lower portion of figure 4, where the rainfall can be compared with the growth of the trees. The curves plotted from these tests were found to have substantial disagreements, although of course the smoothed curves of all of them would be practically identical. A comparison of the growth of the tree with these various curves showed