240 



THE CLIMATIC FACTOR AS ILLUSTRATED IN ARID AMERICA. 



of the many non-climatic accidents to which vegetation and prices are both hable. This 

 error can not be wholly eliminated. Its effect, however, is reduced by the fact that each 

 point in the unsmoothed curves, the dotted lines in figure 76, represents the average 

 condition at that particular phase of 13 cycles. Moreover, the values for each cycle are 

 in turn the average of a considerable number of records in the case of the price of wheat 

 and of 18 measurements along the radii of 11 trees in the other case. In order still further 

 to eliminate accidents I have drawn the solid lines which represent the results of smoothing 

 by the use of 3-year means. For instance, the year before minimum, the year of minimum. 



1510-1754. 



36 

 1.90 

 1.85 



2Vrs jlVeai 

 before before 



Sunspot Curve 



Price of wheat in 

 Enjldnd 1610-1754. 



3,40 

 I 3.30 



Growth of 13 Sequoia 

 trees in California 

 1610-1754. 



6.5 



iWrs.of 



increasing 



Spots 



IVear 

 befoi 



lYear!2Yrs.|ll4Yr3.of 

 after after decreasing 

 Ma» t.laM-1 Spots 



Sunspot Curve 

 1755-1901 



Growttiof 13 Sequoia 

 trees in California 

 1755-1899 



Volcanic eruptions 



p,(i. 70. — The Sun-siwt Cyrlo :inil TeiTPstria! Phcnom(>iia. 



and the year after minimum have been averaged, and the result plotted in the minimum 

 year. Thus in general each point in the smoothed curve of tree growth is the average of 

 18 by 13 by 3, or 702 measurements. With so large a basis of facts it seems probable that 

 accidental errors have been largely eliminated. The remaining departures of the curves 

 from straight lines would seem to indicate some permanently variable factor which varies 

 in harmony with the sun-spot cycles and for that reason will not disappear even when a 

 large body of data is averaged. Aside from climate there appears to be no known factor 

 which could vary in such a way as to have the same periodicity as sun-spots and yet cause 

 the rate of growth of vegetation to fluctuate. 



Taking the solid lines G, H, I, and I' and the dotted line G' in figure 76 we see that they 

 are in substantial agreement. The curve of prices is at a minimum at the time of the sun- 

 spot minimum, and reaches a maximum a year before that of the solar curve. Its highest 

 point is nearly 25 per cent higher than the lowest. The curve of the sequoias for the 

 earlier period is at a minimum a year before the sun-spots and at a maximum a year later 

 than the spots. The difference in this case is only 4 per cent, but the change from minimum 

 to maximum, even in the unsmoothed curve, is so regular that it seems as if it must be due 

 to a genuine difference in the amount of rain at different portions of the sun-spot cycle. 

 The fact that the sequoia curve for the second period, I', presents almost the same appear- 

 ance lends color to this conclusion, more especially as the difference between maximum 

 and minimum is here about 10 per cent. Such an agreement among phenomena occurring 

 in two distinct periods in regions far remote from one another is important, since it adds 

 another to the many lines of evidence which suggest, though they do not prove, a connection 

 between the solar cycle and terrestrial chmate. It also points to the probability that in 

 extra-tropical regions cHmatic records extending over a long period may disclose a cycle 

 which is completely masked by the minor variations which take place from year to year. 



Without attempting to press this point further, let us see exactly what our curves appear 

 to indicate. The curve of the price of wheat has its maximum a year before the solar 

 maximum. Inasmuch as the price of grain responds quickly to variations in the crop, it 



