496 E. HIXTINGTON SOLAR HYPOTHESIS OF CIJMATIC CHANGES 



t-uefficients for each month of the year, one set being ihe relation of tem- 

 perature and tree growth and the other rainfall and tree growth. A 

 period of 50 years. 1851 to 1900. was used for rainfall, and 42 years, 1866 

 to 190T, for temperature, these years being determined merely by the 

 rec-ords which happened to be at hand. The coefficients show that rapid 

 growth depends chiefly on two conditions. The first is relatively high 

 temperature in the late winter and early spring. The sec-ond is abundant 

 rainfall from April to Xoveml>er. The amount of snow in the winter is 

 of comparatively little imix)rtance except where it prevents the ground 

 from warming up in the spring. The summer temperature also makes 

 relatively little difference. These facts suggest that the most favorable 

 conditions for the growth of trees in Germany are periods when rainfall 

 is abundant during the summer and when the winters are comparatively 

 free from storms and snow, so that the trees can l^egin their growth early 

 in the season. The amount of rainfall in climates like that of northern 

 Germany depends largely on the number of cyclonic storms, although in 

 summer it is c-onsiderably influenced by minor showers which are not ac- 

 companied by pronounced barometric disturbances. In order to illustrate 

 the relation of storms to the growth of the trees, I have added to figure 2 

 a little upper curve showing the number of storms from April to Xoveni- 

 ber, inclusive, during the period from 18T6 to 1891. the only period dur- 

 ing which such data are available. This curve is l>ased on figures which 

 have been kindly supplied by Prof. C. J. Kullmer. They include all the 

 storms whose centers passed within 2i^° north or south of El>erswalde. 

 The curve has heen smoothed to correspond with the smoothing of the 

 other curves. Its earlier portion probably falls too low; for, as Professor 

 Kullmer points out. the records of storms in Euro|>e during the early 

 years of the present series are not particularly reliable. The curve thus 

 obtained agrees in its main features with the curve of the trees. The 

 agreement is not perfect, however : for other factors, such as the tempera- 

 ture of the early spring, play a part in determining the rate of growth. 

 Xevertheless it is so close that we are possibly justified in concluding that 

 the curve of tree growth shown in figure 2 represents the main fluctua- 

 tions in the cyclonic storms of summer. 



The curve of tree growth represents not only the number of summer 

 storms, but also the conditions of the winter, as has already been said. 

 AMiere the curve is higli the months of February- and March appear to 

 have been fairly warm and dr^-. It will be tioticed that these conditions — 

 that is. abundant stnnus and rain in summer and an early heating up of 

 the ground in the spring — are tvpical of continental climates. In oceanic 

 Himates, although the winters are on the whole warm and wet, the springs 



