164 



THE CLIMATIC FACTOR AS ILLUSTRATED IN ARID AMERICA. 



37 trees of group I are drcjpped and a correction is applied, as explained in a preceding- 

 chapter, in order to compensate foi- the difference between the average rate of growth of 

 all the trees and that of the 74 trees which still remain. This has no effect whatever 

 upon the form of the sinuosities of the curve; it merely serves to prevent it from dwindhng 

 away in its earlier portions, and to prevent the dropping of trees from causing apparent 

 sinuosities where none exist. From 1883 back to 1850 the curve is based on 74 trees. 

 Then group II is dropped, and the curve is based on 25 trees from 1849 to 1812. Finally, 

 14 of the trees of group III are dropped, leaving only 9 which carry it back to 1766. The 

 earlier portions of this curve do not appear in figure 44, but are used in later discussions. 

 In figure 44 a certain degree of agreement can be seen between the curve of growth 

 and the curve of precipitation. This becomes clearer when the simple rainfall curve is 

 replaced by the smoothed curve shown in the middle fine. This smoothed curve is made 

 by taking the mean of three years' rain and plotting it in the thiid year of eacli group 

 of three. The reason for plotting it in the third year instead of the middle year is that the 

 effect of a rainy season can not possibly be felt before it occurs, but is felt in the years 

 succeeding its occurrence. Between the smoothed curve of precipitation and the curve of 

 growth a considerable degree of agreement is manifest. For instance, the rainfall maxima 

 of 1862, 1868, 1876, 1886, 1895-97, 1901, and 1907 are all accompanied or closely followed by 

 arboreal maxima. Marked disagreement, however, is evident in such years as 1878, 1882, 

 and 1904. The explanation of these discrepancies seems to be found largely in the seasonal 

 distribution of the precipitation, as is shown in figure 45. For example, the smoothed 



rainfall curve rises irregularly 



Inches 



If July Aug. Sept. Oct. Nov. Dec. Jan. Feb. Mar. Apr. May June 



Inches 



0" 



I^ro. 45. — Mean Monthly Distribution of Rainfall Compared with Di.sl ri- 

 bution in Exceptional Years. San Franciseo-rresno. 



from 1871 to 1880, and judging 

 from the 10 years before this 

 period and the 20 or more years 

 following it, we should expect the 

 tree curve to do likewise. Up to 

 1877 the two do agree very well, 

 but then comes a marked discrep- 

 ancy lasting till 1882. Figure 

 45 suggests the probable cause 

 of this. It shows the average 

 seasonal distribution of precipitation since 1850 at San Francisco in the dotted line, while 

 the other lines show the distribution for special years, either there or at Fresno. The year 

 1876-77 was one of the worst on record. Not only was the rainfall scanty, but its distri- 

 bution was perhaps the worst since records have been kept. October was rainy, but precipi- 

 tation in that month is of little use, since it either takes the form of rain, or else having 

 fallen as snow it melts off unless promptly covered by fresh supplies. No new snow came 

 in either November or December. January had a good supply, although not quite the 

 average amount, and each of the next four months received only a third of the normal 

 supply. In May the ground must have been as dry as it usually is in August. The 

 growth of vegetation must have been checked almost as soon as it began, and the trees 

 must have suffered sadly. Apparently they were so injured that they could not make a 

 good growth the next year in spite of abundant precipitation, well distributed; then they 

 began to recover, and by 1882 were in such a condition that they grew well in spite of scanty 

 precipitation. That year, however, was very different from 1877. The fall and early 

 winter were dry, but February, March, and April equaled or exceeded the normal, and 

 those are apparently the most important months. In 1904 a similar ca.se occurs: The 

 precipitation remains at a low point, but the growth of the trees is accelerated. The cause 

 seems to be plainly evident in the unusually large fall of snow or rain dining February, 

 March, and Ai)ril, as shown in figure 45. Thus we might go on to analyze year after year. 



