INTERPRETATION OF THE CURVE OF THE SEQUOIA. 



163 



The groups are as follows : 



Group I. Sections showing 60 rings or less; trees of rapid growth: (o) From dry local- 

 ities, 23 sections, (ft) From damp localities, 14 sections. 



Group II. Sections showing from 60 to 100 rings; trees of moderate growth: (a) From 

 dry localities, 31 sections. (6) From damp localities, 18 sections. 



Group III. Sections showing over 100 rings; trees of slow growth: (a) From dry local- 

 ities, 25 sections, (b) From damp localities, no sections. (Trees in damp places 

 almost universally grow rapidly.) 



The uncorrected curves of growth derived from these five groups are shown in figure 43. 

 The solid lines represent the trees living in damp locahties, while the dotted lines represent 

 those from dry locaUties. The trees in the swamps grow the fastest, as would naturally 

 be expected, but in times of adversity they appear to be the first to suffer and are on the 

 whole more affected than are those in the drier locations. Hence we conclude that, for a 

 curve of growth possessing the highest degree of accuracy, vigorous, rapidly growing trees 

 located in small swamps which are easily dried up are the most advantageous. The differ- 

 ence in the form of the various curves, however, is comparatively sUght, except that the 

 rate of growth in the successive groups is slower and slower. In general the curves show 

 periods of maximum or increasing growth about 1850, 1854-55, 1862-64, 1868-70, 1876, 

 1882, 1886, 1894-96, 1902, and 1908, while periods of minimum growth are almost as 

 markedly in agreement. If a time unit of 5 or 10 years were used instead of one year, and 

 if the proper corrections were apphed to eliminate the various errors due to age, longevity, 

 and the hke, the five curves would be practically identical. This supports the conclusions 

 of Professor Douglass as to the possibihty of obtaining fairly reUable records from a small 

 number of trees. Nevertheless, there can be Uttle doubt that much more accurate results 

 are obtained where a large mmiber is employed. 



1850 Date 



1870 



1890 



1900 



1910 



4.00 



Rainfall curves for 

 Fresno-San Francisco 



3-year means plotted 



in the last year of 



the three 



10 Actual rainfall 



Fig. 44. — Growth of Trees at Hume, and Rainfall at Fresno. 

 (See Table I, pp. 328-329.) 



We have now to determine how far the synchronous fluctuations in the rate of growth 

 of these five groups of trees are due to variations in rainfall. The trees of all the groups are 

 scattered over an area of about a mile square. No fires appear to have occurred in the 

 region for many years, certainly not during the last 30 or 40, and probably not for centuries. 

 The trees were all strong and vigorous at the time when the sections were cut, and there 

 was no sign that they were influenced by any special diseases or parasites. They were 

 scattered in all sorts of locations, from places where swamps or perennial brooks bathed 

 their roots to dry, rocky hillsides subject to constant drought. Unless the variations in 

 the rate of growth are due to cUmate, there seems to be no adequate explanation of their 

 existence. Nevertheless, when the combined curve of the five groups is placed beside the 

 curve of rainfall at Fresno and, before 1882, San Francisco, as is done in figure 44, the 

 degree of agreement is scarcely so great as one would expect. The combined curve of the 

 five groups is obtained by using all of the 111 trees as far back as 1884. At that point the 



