INTERPRETATION OF THE CURVE OF THE SEQUOIA. 165 



and find causes for a large number of the divergencies ]:)etween the curves of growth and 

 precipitation. 



The seasonal distribution of rainfall, however, is only one of the two chief factors which 

 cause the curve of the trees to disagree with that of the rain. The other is what Professor 

 Douglass has called the conversation factor. The extent to which the curve of growth 

 lags behind that of precipitation, even when the 3-year mean rainfall is plotted in the last 

 of the 3 years of any given group, suggests that this factor plays a more important part 

 here than among the pines in Arizona. It apparently depends not only upon the amount 

 of water stored in the soil, but upon the amount of reserve strength which the plant has 

 been able to acquire by enlarging its root system or by the growth of branches and buds. 



Figure 46 represents the result of the most satisfactory of five or six methods by which 

 an attempt has been made to gage the effect of the i-ainfall of preceding j^ears upon the 

 growth of the sequoias at Hume during any particular year. From the 63 years for which 

 records of rainfall at San Francisco and Fresno are availalile (that is, from 1849-50 to 1911- 

 12) I have selected two groups. One group consists of 15 years, during which the trees at 

 Hume not only formed rings having more than the average thickness of 3.5 mm. according 

 to the corrected figures used in plotting the curve, but also grew faster than during the 

 Ijreceding year by an amount of 0.25 nun. or more. The other group consists of 14 years, 

 during which the trees not only grew less than the average amount, but also grew less 

 rapidly than during the preceding years by an amount of 0.23 mm. or more. This last 

 figure was selected instead of 0.25 mm. simply in order to make the two groups as nearly 

 equal as possible. In the selection of these two groups it is clear that two chief criteria 

 are employed, the absolute amoimt of growth and the relative amount. A glance at the 

 curve of growth in figure 44, the upper solid line, will show how the two criteria are applied. 

 According to the first criterion (absolute growth) the years are divided into two classes. 

 One class comprises all those which grew more than 3.5 mm. and whose position in the 

 diagram is above the median horizontal line, while the other class comprises all which grew 

 less than 3.5 mm. and whose position is below the Une. From the class of rapidly growing 

 trees a smaller class was selected by means of the second criterion, relative growth. All the 

 years in which the growth was more than 0.25 mm. in excess of the preceding j^ear, or in 

 other words all the years which are preceded by a rapidly rising portion of the curve in 

 figure 44, were selected and the rest rejected. In the same way, among the slow-growing 

 trees selection was made of all which show a growth of 0.23 mm. or more in deficiency of 

 the preceding year — that is, which are preceded by a rapidly falling jjortion of the curve. 



For these two groups of years of rajiid and increasing growth on the one hand, and of 

 slow, decreasing growth on the other hand, I have calculated the average rainfall, first 

 during the season preceding the period of growth, then during the two seasons preceding it, 

 and so on until 5 years have been included. The results appear in figure 46. In the 

 figure the upper curve represents the rainfall of what we may call the progressive years, 

 and the lower of the reactionary, while the solid horizontal line indicates the mean rainfall 

 for the entire period of 62 years, which amounts to 10.67 inches for Fresno. The meaning 

 of the curves is plain. During the years immediately preceding times of rapid and increas- 

 ing growth the average rainfall was 12.58 inches; for the period of 2 years preceding such 

 times it was 12.02 inches; for 3 years 11.86 inches; for 4 years 11.30 inches; and for 5 years 

 11.17 inches, a series of figures which increases steadily as the times of rapid growth are 

 approached. In the case of the slow-growing trees, on the contrary, the figures are for 

 1 year 9.98 inches, 2 years 9.51, 3 years 9.64, 4 years 10.08, and 5 years 10.53, a series which, 

 in general, decreases as the times of slow growth are approached. If the two curves 

 were carried back a few years farther they would coalesce. Figure 47 illustrates the same 

 thing as 46, except that the mean rainfall for the first, second, third, fourth, and fifth years 

 preceding the years of growth has been plotted instead of the means for periods of 1, 2, 3, 



