METHOD OF ESTIMATING RAINFALL BY GROWTH OF TREES. 



117 



the change in growth with advancing age. Long, slow variations in the rate of growth 

 may not cause any divergence from the line, but brief periodic departures are manifest 

 and may be taken as truly representing the extent to which the cHmate departs from the 

 mean. The departures, however, are not all on the same scale, for where the growth 

 is large, the departures are large, and vice versa. Hence, they must all be reduced to the 

 same scale. This reduction has been effected by dividing the numbers used in figure 19 

 by the reading of the long, straight line in millimeters for the corresponding year. By 

 this process a series of values of tree growth is obtained such as the trees would give if they 

 grew on the average a millimeter per year and did not change with age. 



1.00 



1400 l.iflO ICOU ITiKi ISIHI 19UIJ 



Fig. 20. — oOO-^ear Curve of Tree Growth — 20-year Means. 



CLIMATIC CYCLES. 



In the corrected curve of growth thus obtained the minor deviations obscure the larger 

 features. Accordingly, in figure 20 the curve has been condensed into a 20-year smoothed 

 curve. This particular length of time was chosen because a 21-year variation is evident 

 in most of figui'e 19, and this, as well as smaller variations, must be removed in order to 

 leave larger variations unaffected. A 20-year smoothed mean accomplishes the desired 

 result and is easier to calculate than is a 21-year mean. Inspection of the curve of figure 20 

 shows a long and pronounced maximum of tree-growth between 1530 and 1620, a lesser 

 maximum shortly after 1700, and a still shorter one at about 1860. Strong minima occur 

 between 1505 and 1530, 1630 and 1675, here and there between 1740 and 1830, and again 

 between 1870 and 1900. Manifestly pulsations of some sort take place. They may or 

 may not be permanent. Perhaps they are nothing more enduring than a series of simul- 

 taneous wave systems on a water surface. Yet for the navigator a knowledge of the 

 existing system is important ; and so for the purpose of weather prediction we need to know 

 the nature of the pulsations now existing, and each one should be minutely studied. The 

 slow changes shown in the curve seem to have a somewhat regular periodicity. In order 

 to bring this out, a wavy line representing a cycle of 150 years has been placed above the 

 curve of growth. This cycle in the growth of the trees is fairly well evident, and it is 

 represented again in figure 21, where the three cycles shown in the main line of figure 20 



1.25 mm. 

 1.00 mm. 



.75 mm. 

 .50 mm. 

 .25 mm. 

 .00 mm. 

 Fig. 21. — A Possible 150-year Period. 



5 10 15 2021 



Fig. 22. — Mean Curve of the 21-year Cycle. 



