4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IO3 



barometric depressions almost equals the high of 1937 with 57 

 storms. We note that in 1895 there was a sudden drop from 56 to 

 17 and in a few years to 2. Furthermore, we notice that in 1940 

 there was a sudden drop from a high of 57 to 29 and that the interval 

 between the two sudden drops was 45 years, exactly four solar cycles. 



1 think there can be little question that the years of minimum, 1941- 

 43, will show a decrease in eastern Canada. With the plus area in 

 eastern Canada assured, we shall have to wait 3 or 4 years for the 

 possible completion of the lower part of the script T pattern. 



How shall we interpret these findings? We naturally first think 

 of the alternations in solar activity as shown in the accompanying table, 

 figure 13, and of Clayton's northerly shift of the centers of action 

 with increased solar activity, but the figures resist attempts at cor- 

 relation with the observed reversal. It seems, therefore, necessary 

 to consider the theory that solar activity may possibly be completed 

 in four cycles, of which the third and fourth are in some mysterious 

 way the exact reversal of the first and second. We have become ac- 

 customed to the idea of a reversal of solar activity through Hale's 

 discovery of a reversal of magnetic polarity with each new solar 

 cycle. But in what way could such a Hale cycle itself be reversed ? 



Finally a word concerning the years selected for comparison. It 

 will be remembered that for the first three and a half periods I 

 selected by inspection from Maunder's chart of sunspot latitudes 

 the years of maximum overlapping and used 3 years on either side 

 of such a year. Maunder's chart does not include the minima of 

 1923 and 1933 ; consequently I selected these years as boundaries. 

 Figure 14 summarizes the official years of minimum and maximum 

 solar activity, together with my periods selected as described. I 

 would call attention to the fact that the first and third periods begin 



2 years after the official minimum ; the second period, i year after 

 minimum, and the fourth period begins with the minimum. There 

 are very slight differences in the Wolfer sunspot numbers which 

 determine the solar minimum, and it seemed worth while to test the 

 matter by making a series of maps using the official minimum year 

 as the boundary ; for the fifth test period, which was originally so 

 bounded, I included the minimum year and have already presented it 

 as the superior map. In figure 15 I present the first period. The 

 central plus band is strengthened, but the southern minus band is 

 weakened. My original period seems to be the best for the three- 

 banded effect. In the second period, figure 16, the plus band in the 

 north is weakened, uncovering, as it were, a narrow band of minus 

 which crosses the map. Also the southern curved minus area is 



