METHOD OF ESTIMATING RAINFALL BY GROWTH OF TREES. 121 



The lowest curve is an inverted sun-spot curve for 125 years, 1771 to 1896. There 

 appears to be a marked similarity between this and the temperature curve. Even the 

 subordinate crest, which sometimes shows in the sun-spot descent from maximum to 

 minimum, matches this suppressed second crest of temperature and its following faint 

 minimum. This would seem impossible in the absence of a real relationship between them. 



The relation between tree growth and sun-spots here shown, however it comes about, 

 does not stand alone. A series of measures on 13 tree sections from the forest of Ebers- 

 walde, near Berlin, Germany, the first of a number of series to be made on North European 

 pine trees, discloses a striking time relation of the same character. The 13 trees were 

 divided into two subordinate groups, given in curves 1 and 2 of figure 2.5. These show 

 most satisfactory agreement. The third curve gives the growth of the whole 13. The 

 fourth curve is the same as the third, but corrected for age. The fifth curve shows the 

 tree growth smoothed in overlapping groups of three, and below it is the sun-spot curve. 

 The similarity may be traced without further comment. This gives very strong support 

 to the view here entertained that there is a relationship between the tree growth and the 

 sun-spot activity through the mediation of the weather. 



CONCLUSION. 



In the foregoing investigation it has been shown: (1) that the variations in the annual 

 rings of individual pine trees in the diy regions of northern Ai'izona exhibit such uniformity 

 that the rings of one tree can be identified in others over large areas and the date of their 

 formation established with practical certainty; (2) that the ring thicknesses are propor- 

 tional to the rainfall with an accuracy of 70 to 82 per cent in recent years and that this 

 accuracy presumably extends over centuries ; (3) that the tree year for such records begins 

 in the autumn; (4) that double rings are caused by spring drought and are indicative of the 

 distribution of rainfall throughout the year; and (5) that an empirical formula can be 

 made to express the relationship between tree growth and rainfall. The ring record at 

 Flagstaff, Arizona, has been traced back 500 years and various cycles found in it. An 

 approximate 33-year cycle shows in the last 200 years. A 21-year cycle shows in 400 

 years and an 11-year cycle displays a similar duration. The 11-year cycle shows marked 

 relationship to the California coast rainfall and temperature and to the sun-spot curve. 

 In corroboration, attention is called to the still more remarkable agreement between tree 

 growth in northern Germany and the sun-spot curve. All of this confirms the idea that 

 observation of tree growth may be a powerful help in studying the climate of the past. 



Further research will probably show other and perhaps still more importarit relation- 

 ships between the growth of vegetation, meteorological elements, and changes in the sun. 

 Meanwhile, the methods of computing rainfall from tree growth must be still further per- 

 fected. Already, however, the original purpose of the work here outUned has been accom- 

 pUshed. Its most important part, I hope, has been the estabhshment of a method of 

 estimating rainfall, capable of extension to other regions, and for the benefit of other 

 branches of science. 



