METHOD OF ESTIMATING RAINFALL BY GROWTH OF TREES. 



109 



1911 agree with the 25 cuttings of 1904-06. The 7 came from about 12 miles southeast 

 of the town, while the 25 came from places from 6 to 12 miles farther west. The general 

 form of the two curves is strikingly similar, just as is the general form for the four groups 

 at Prescott, as shown in figure 9. This similarity indicates that even a small group of trees, 

 no more than seven in number, is sufficient to give results of considerable accuracy. Indeed, 

 we may go farther and say that a single tree may give results of moderate accuracy pro- 

 vided it grows fast enough, and provided allowance be made for the cumulative effect of 

 a series of good or bad years and for the vagaries due to the age or special position of the 



Year 



1S70 



1S80 



1890 



1900 



1910 



1900 



1910 



190O 



1910 



Year Year 



Fig. 11. — Annual Growth of Trees at Flagstaff, and Variations in Annual Rainfall 

 according to Month which is reckoned as the Beginning of the Year. 



tree. This is evident in figure 12, where the 7 sections from the last Flagstaff group are 

 plotted separately, the most rapid grower at the top, just below the rainfall cm"ve, and 

 the slowest grower at the bottom. All ahke rise because the conditions of rainfall in 

 1900-10 were more favorable than in 1890-1900, and all, but especially the curve of sec- 

 tion 4, show a more or less close relation to the curve of rainfall at Flagstaff, even though 

 that place was some 12 miles away. The great sinuosity of the curve of section 4 as com- 

 pared with section 5, at the bottom, is noteworthy, for section 4 was cut from a fast- 

 growing tree. This difference supports the conclusion already reached, that slow-growing 

 trees are of less value than rapidly growing ones in the study of the climate of the past. 



