may be divided into two classes: first, local errors of identity in small groups of rings in a few individual trees, which simply flatten the curve without affecting the final count; second, cases in which a given ring, in spite of attempts at cross-identification, is still in doubt, showing clearly in perhaps half of the trees and not in the other half. Such cases affect the final count, but do not flatten the curve. They leave a question of one year in the dating of all the earlier portions of the curve. Only two cases of this latter kind have been noted. One was the year 1822 in the Flagstaff pines (of which there is very little doubt) and the other is the ring 1580 in the sequoias, which was finally decided by material gathered in the special trip of 1919.
Apart from care in measuring the rings, the details of which will be given in Chapter IV, the most fundamental and essential feature of the method of studying tree-growth is the cross-identification of rings among a group of trees. The ease and accuracy with which this can be done in a fairly homogeneous forest is remarkable. A group of 13 tree sections collected along a distance of a quarter of a mile in the forest of Eberswalde, near Berlin, show almost identical records. Two to ten rings in every decade have enough individuality to make them recognizable in every tree. A group of 12 sections from Central Sweden show such agreement that there is not a single questionable ring in the last 100 years or more. Especially marked combinations of rings can occasionally be traced across Europe between the groups hereafter mentioned. In Arizona the identification across 70 miles of country is unquestioned, and even at 200 miles the resemblance is apparent.
The value and accuracy of cross-identification was first observed in 1911 in connection with the Prescott trees. After measuring the first 18 sections, it became apparent that much the same succession of rings was occurring in each; therefore the other sections were examined and the appearance of some 60 or 70 rings memorized. All the sections were then reviewed and pinpricks placed in each against certain rings which had characteristics common to all. For example, the red ring of 1896 was nearly always double, while the rings of 1884 and 1885 were wider than their neighbors. In the 60 years investigated several obvious details in each decade appeared in every tree. After this success it was evident that the process should be applied to the Flagstaff trees which had been previously collected. Of the 25, however, only 19 had been preserved. A minute comparison was made between these with complete satisfaction. Since then this process has been applied with great care to every group.
After the Flagstaff set was finished, it was compared with the Prescott group. It was interesting to find that the Flagstaff ring records