NO. 10 SOLAR RADIATION AND WEATHER STUDIES ABBOT 29 



D. Dependence of Phase on Sun-spot Activity 



In caption 8, evidence was presented showing that the phases of 

 the 1 1 -month and other periodicities observed in temperature de- 

 partures of Bismarck, N. Dak., altered as a function of the sun-spot 

 activity. Referring now to figure 13, the data for BerHn are not wholly 

 consistent with that conclusion. As not all of the 11 -year sun-spot 

 periods show high Wolf numbers, let us restrict our inquiry to the 

 periods culminating about 1837, 1847, 1871, 1893, and 1918. Con- 

 sider first the 1 1 -month periodicity. In table 5 are given the months 

 within the 11 -month period when maxima prevail. The results cover 

 times of low, medium, and high Wolf numbers. The shift of maxima 

 for medium and high Wolf numbers is indicated in the fourth column. 

 In the last column are given without details the corresponding shifts 

 found for the 8-month periodicity data. 



Table 5. — Shift of Phase, Berlin Temperatures, Attending Sun-spot Activity 



Years Wolf Months ii-month S-montli 



covered numbers of maxima shifts shifts 



18311834 Below 40 II to 4 0.0 0.0 



1834-35 and 1839-41 40 to 80 I to 5 +1.0 — i.o 



183S-1839 Above 80 2 to 4 +1.0 —4.0 



1841-1844 Below 40 9 to 2 0.0 0.0 



1844-46 and 1849-53 40 to 80 6 to 11 —2.5 —2.0 



1846-1849 Above 80 8 to 9 — 3-0 — i-S 



1865-1868 Below 40 10 to II 0.0 0.0 



1868-69 and 1873-75 40 to 80 5 to 8 —4.0 +2.0 



1869-1873 Above 80 8 to 10 — 1-5 0.0 



1886-1890 Below 40 II to 2 0.0 0.0 



1890-92 and 1895-97 40 to 80 10 to 2 — 0.5 — o.i 



1892-189S Above 80 9 to II —2.0 —1.0 



1910-1914 Below 40 3 to 5 0.0 0.0 



1914-17 and 1919-21 40 to 80 I to 5 — i.o — 3-o 



1917-1919 Above 40 8 to II 



lor +5.5 J 



There appears a prevailing tendency for the phase to be earlier with 

 higher sun-spot activity, but it is not as conspicuous or regularly 

 progressive a tendency as appeared in the Bismarck data. In fact 

 the evidence seems to show that though there is a small change of 

 phase toward earlier dates within the cycles, when Wolf numbers 

 increase, yet this efifect is small compared with changes of phase which, 

 as we are about to point out, occur at integral multiples of ii| or of 23 

 years, counted from January 1819. Such changes of phase will next 

 be demonstrated. 



E. Dependence of Phase on Epoch Counted from 1819 



It was desired to present this phenomenon apart from changes of 

 phase accompanying variations of sun-spot activity. Hence the data 



