98 



CLIMATIC CYCLES AND TREE-GROWTH 



correlation with tree-growth. So, remembering that the torrential 

 summer rains do not greatly benefit the trees, the year was divided, 

 as it is naturally, into winter and summer precipitation, the former 

 from November 1 to June 30 and the latter from July 1 to October 31. 

 It was immediately evident that this removed the unexpected dis- 

 agreement, for the winter values closely resemble the tree-growth, while 

 the summer rains (averaging 10 out of an annual total of 23 inches) 

 show no relation to the growth. This is shown in figure 8. Though 

 the length of record is not great enough to test satisfactorily any for- 

 mula for reducing rainfall to tree-growth, or the reverse, the evidence 

 indicates that the same principle of accumulated moisture used in the 



2J0O\ 



S '-co 



S 



o 



30 



■ 



o 

 a 

 i-t 



10 



F rescott tree gr owth 



2and3 



— 



30- 



20- 



Prescott r linfall calculate J from tree gro vth 



r 



Presc stt rainfall Nov 



to Nov. I 



0' 



I860 



1670 1880 1890 1900 1910 



Fig. 7 — Prescott rainfall and tree-growth 



Prescott correlation (Volume I, p. 66) applies here. The accumulated 

 moisture curve for the winter precipitation at Flagstaff is shown in 

 curve 4 of the figure. 



Flagstaff and Prescott difference — In the correlation between 

 rainfall and tree-growth at Prescott, it was not necessary to segregate 

 the winter rains for the purpose, because the correlation was apparent 

 when using the annual total. But in the Flagstaff area the winter 

 precipitation only can be used. Without doubt this difference arises 

 from the topography of the country. Prescott is situated in the lap of 

 the Bradshaw Mountains opening to the north and protected from the 

 southerly summer winds, while the Flagstaff area is mostly on the 

 south side of the lofty San Francisco Mountains, about which summer 



