NO. 4 TEMPERATURE VARIATIONS IN THE NORTH ATLANTIC 277 



These examples may suffice to show that, as a general rule, it 

 depends on the situation of a station, with regard to the centers of 

 action of the atinosphere, whether the correlation factor of the sta- 

 tion be positive or negative, i. e., whether the temperature at the 

 place varies in the same direction or in the opposite direction of the 

 solar activity. This proves that the fluctuations in temperature at 

 the earth's surface depend greatly on the changes in the circulation 

 of the atmosphere, while the latter are affected more directly by 

 the changes in solar activity. 



According to what has been stated above the shape of the normal 

 isothermal lines may be expected to demonstrate the distribution of 

 the positive or negative correlation between fluctuations in solar 

 activity and in terrestrial temperature. 



In figure 98 we have drawn some isothermal lines showing the 

 mean temperatures for August, September, and October, compiled 

 from Buchan's maps [1889], and have also introduced the maximum 

 values of the correlation co'efficients computed by Dr. Clayton for 

 his thirty stations. 



This map shows clearly that the said correlation is positive where 

 the shape of the isothermal lines indicates comparatively high nor- 

 mal temperatures {e. g., at Pilar in Argentina, at Zungeru in Nige- 

 ria, at Zomha in South Africa, at San Isidro, Philippines, at Ja- 

 cohshavn in Greenland, at Dawson in Alaska, at Laurie Island in 

 60° 44' S., 44° 39' W., nay, even at St. Johns, N. B., where the 

 isothermal line has a small bend towards the north) . But this cor- 

 relation is negative where the shape of the isothermal lines indicates 

 comparatively low normal temperatures (e. g., at Sacramento and 

 San Diego on the west coast of U. S. A., at Chicago, at Bathurst 

 in Gambia, at Punta Arenas, on Mauritius, on the Fiji Islands, at 

 Hongkong) . 



It should be kept in view that Dr. Clayton, taking it for granted 

 that the fluctuations in temperature at the earth's surface, at least 

 in the tropical regions, are directly affected by the fluctuations in 

 solar radiation, has used for his investigations the daily maximum 

 temperatures at the various stations. 



The maximum temperatures depend very much on the cloudiness 

 of the season, and do not give a trustworthy indication of the mean 

 daily temperature, which it would be of importance to know when 

 we wish to examine the effect of the circulation of the atmosphere. 

 The mean daily temperatures would probably have shown still bet- 

 ter agreement with the fluctuations in the solar radiation. 



