2 SMITIISOMAiX M ISCEI.LAX F.OTS fOLLF.CTIOXS VOL. IO3 



For comparison, the most powerful mclhod of rittack seemed to be 

 to add together 3-year maps at solar niininuim and 3-ycar maps at 

 solar maximum and compare them with each other. This was done 

 for five solar cycles, with the result that a uniform pattern appeared. 

 In figure i I present one such cycle, perhaps the most typical of the 

 five. Plus figures indicate the excess in years of solar maximum 

 over years of solar minimum ; minus figures, the reverse. In the 

 north there was a strong increase along the main storm track, but 

 also a southerly projection. I have drawn a heavy dotted line through 

 the main area of excess at times of maximum and also through the 

 projection ; at both sides of the projection there are areas of decrease, 

 i.e., more storms at years of minimum than at maximum. In figure 2 

 I have assembled the dotted lines through the area of excess and the 

 projection. It will be noted that in the fifth period there was an un- 

 expected move to the south and that the southerly projection had 

 shifted even farther west than in the second period. Because of a 

 certain orderly progression in the shifts of the southerly projection 

 I ventured a hypothetical position for the sixth solar cycle, and was 

 correct in assuming an easterly motion of the southerly projection. 

 I now present in figure 3 the results for the last solar cycle, showing 

 an unexpected still greater shift to the south of the main area of 

 excess at solar maximum. There was some doubt as to the year 

 of maximum, 1937 or 1938, and consequently I made the maps for 

 both periods, figure 4. Except for slight variation in figures, the 

 maps are identical. 



Now what does the continued southerly shift of the pattern signify, 

 with the large increases at minimum in western Canada? The in- 

 terpretation of a relation between sunspot latitudes and latitudes of 

 earthly storms seemed to be justified liy the first four periods, but 

 contradicted by the fifth and sixth j^eriods. There remained one 

 way to test the relationship. In figure 5, taken from Clayton's 

 "World Weather," we have Maunder's chart of sunspot latitudes. 

 All these years I have been comparing 3 years at solar minimum 

 with 3 years at solar maximum, but evidently the latitude etifect 

 would be intensified if I compared 3 years at the beginning of the 

 cycle with 3 years at the end of the cycle. There is a noticeable 

 overlapping of the new^ cycle with the old; in order to avoid this 

 source of error I would eliminate by inspection the year of maximum 

 overlapping, as indicated in the chart, and proceed to test the latitude 

 efifect by comparing 3 years at the beginning of a cycle with 3 years 

 at the end of a cycle. I began with the first cycle available, figure 6, 

 and the results were a succession of surprises. Gone was all thought 



