CYCLONIC STORMS IN TEMPERATURE LATITUDES 517 



moiiy with the sun-spots. Taking the eruptions as a whole, it is hard to 

 see that they have had any special influence. The only possible generali- 

 zation seems to be that as a rule the volcanoes are active at times when 

 the number of storms is somewhat lower than would be expected, but 

 1886 and 1890 are exceptions. The logical conclusion seems to be that 

 while volcanic eruptions may have a temporary effect on the general tem- 

 perature of the earth, the changes of temperature thus caused do not 

 perceptibly influence cyclonic storms. This supports the idea that storms 

 are not due primarily to conditions of temperature. It thus strengthens 

 our previous conclusion that although volcanic eruptions may be of im- 

 portance, they are nevertheless only a minor factor in causing climatic 

 changes. It also strengthens the conclusion that storms are not due pri- 

 marily to conditions of temperature. It does not, however, shed any 

 light on the reason for the occasional disagreements between the curves 

 of sun-spots and of storms. That problem must for the present be left 

 unsolved. 



Shifting of the storm track in Europe. — Having reached this conclu- 

 sion, we must next ascertain whether the phenomena of Europe are sim- 

 ilar to those of America. Because of the scarcity of data for European 

 storm tracks and because of the lack of homogeneity in the 16 years for 

 which records are available, we can not as yet determine whether the total 

 storminess is greater in periods of many sun-spots than in periods of few. 

 Fortunately, this does not prevent us from comparing the distribution of 

 storms under different circumstances.^^ In Europe only one period of 

 maximum spots is available, the years 1882-1884, but there are two mini- 

 mum periods, 1877-1879 and 1888-1890. Therefore each of these has 

 been compared with the same maximum. 



The results of this comparison are shown in figures 10, 11, and 12. 

 The first two show the relative number of storms in periods of minimum 

 sun-spots, 1877 to 1879 and 1888 to 1890, respectively, as compared w4th 

 the intervening maximum of 1882 to 1884. The third diagram, figure 

 12, shows the other two combined. Figure 1 1 covers a much larger area 

 than the other two maps, since the region where observations were in- 

 cluded after 1882 was much larger than in the preceding period. In 

 spite of certain places where there is a good deal of doubt as to the relia- 



'^ We can not compare the figures directly, as was done in America. By the use of 

 percentages, however, we can obtain a fairly satisfactory result. The method Is merely 

 to find the average storminess for all squares during each three-year period of maximum 

 or minimum spots, and then to call this average 100 per cent. On this basis the value 

 of each square for each period is computed and the differences are determined. Pro- 

 vided that all parts of the country were mapped in the same way during a given period, 

 the results are the same as if direct comparisons were instituted according to the method 

 employed in America. 



