1835, and only the earlier part, ending in 1834, was used. Extrapolated values for the missing part were derived in the usual way. A few short, apparently doubtful, regions of rings required careful study and it was found that well-adjusted illumination of the rubbings was very necessary to their correct reading. When the ring impressions were deep in the paper, the end of the rubbing showing the tree center was held toward the source of light in order that the elevation corresponding to the beginning of the spring growth might be brightly illuminated. When the impressions were shallow and faint, it was noted that the rings became very distinct if the rubbing was held between the eyes and the light, thus giving a very faint and perfectly even illumination. If this did not bring out the individual rings, the rubbing was not used.
The location in which these trees grew was visited in 1918 and general contours were noted. The hills are low and comparatively flat-topped, with disintegrated rocks showing in railroad cuttings. The sides of the hills are steep, and the valley bottom is narrow and usually has a wash near its center. In general the drainage is toward the east, but there is no high and sharp ridge between this region and the ocean on the west. The situation is far enough north to have a good snowfall in winter. It is about 800 feet above sea-level.
The tabular matter giving the results of the measures on the 17 Douglas firs of Oregon will be found on page 117. The plotted values appear in figure 11.
THE SEQUOIA GROUP.
In 1911, after examining the writer's results obtained on the yellow pines, Huntington made an extensive series of measurements on the big tree, Sequoia gigantea. He did this work on the stumps themselves by direct counting from the outside. This introduced errors of beginning due to removal or injury of outer rings, and errors of omission which of course could not be checked. In order to correct for large errors of omission, he worked out an approximate correction on the grounds of probability which depended upon a comparison between two or more radii of the tree, and in that way many errors were compensated. In the vast majority of cases, his measures were not of individual rings but of successive groups of ten. I have collected seven of his trees, and after complete cross-identification verify his centers as shown in table 4.
But Huntington's method of working directly on the stump enabled him to get data from a very large number of trees, some 450, in a way that served his purpose very admirably. He was searching for general effects, and accuracy to a year or two was less essential. He wished to approximate absolute values of rainfall in past climates, in contrast with which my chief aim is to get relative and periodic values. These