A STUDY OF CORRELATIONS AMOXG TERRESTRIAL TEMPERATURES. 383 



influenced by solar emanations. -Still, the question whether there is any relation 

 between magnetic storms, which afford us the best available evidence of the emana- 

 tions in question, and thunder storms or other exhibitions of movements of atmos- 

 pheric electricity, is an interesting one, well worthy of investigation by rigorous 

 statistical methods, and offering no difficulty. The main point to be enforced in the 

 present connection is that our investigation includes the effect of all cosmical causes 

 affecting the terrestrial temperature, and therefore of the emanations in question so far 

 as they produce any thermal effect. 



Dropping the consideration of magnetic, electric or radio-active emanations as 

 belonging to another branch of the subject, because they do not cause appreciable fluc- 

 tuation in terrestrial temperatures, we return to the main question now under consid- 

 eration — that of the relation between fluctuations in the sun's thermal radiation and 

 tlie corresponding changes in temperature. Accepting the fourth-power law of radia- 

 tion, fluctuations in the general temperature of the globe of 0°.2 C. on each side of the 

 mean would produce corresponding changes of 0.3 of one per cent, in the radiation of 

 heat by the earth into space. We have found that the fluctuations of world-tempera- 

 ture, if any at all occur, which is doubtful, do not exceed ± 0°.20 G. We may there- 

 fore assign three tenths of one per cent, as the ordinary limit of fluctuation of the 

 sun's radiation in lower periods. But the lag of temperature behind insolation is to 

 be considered in the case of short periods. 



Speaking in a general way, it is an observed fact that the maxima and minima 

 of temperature in the temperate regions do not occur until about a month after the 

 maxima and minima of radiation. But, admitting that a month will be required to 

 produce the completed effect through the entire atmosphere and on the surface of the 

 ground and ocean, it does not follow that the effect would be negligible in a shorter 

 period. It is also an observed fact in regions of middle latitude that the rays of 

 the sun between its rising and 2 p. m. elevate the temperature of the air at the earth's 

 surface as read by the thermometer, by an amount ranging from 8° to 10° C. every 

 day. Now, to fix the ideas, suppose that the sun's thermal radiance were increased 

 by one per cent, of its whole amount through ten consecutive days. The result would 

 be that the daily rise would be increased by an amount between 0°.06 and 0°.10. 

 This rise would be in part lost during the night by increased radiation and transmis- 

 sion to the earth and upper air. But, as the earth and air grew wai-mer day after day 

 the loss would be smaller and smaller, while the gain would continually accumulate. 

 It follows that we should not have to wait more than a week for the change of one per 

 cent, in the sun's energy to produce an effect exceeding that which our study of tem- 

 peratures shows can be actually found in the world-temperature. But this does not 



