A STUDY OF CORRELATIONS AMONG TERRESTRIAL TEMPERATURES. 379 



this is really the case we may estimate the change in question as affecting terrestrial 

 temperatures by two- or three-hundredths of a degree. A more exhaustive inquiry 

 nto this question certainly seems of scientific interest, but I must, as with the contin- 

 uation of the present work generally, leave this in other hands. The main point 

 reached is that the influence of any such inequality in the sun upon meteorological 

 phenomena is so nearly evanescent that it can be brought out only by the most 

 refined methods of investigation, and cannot be of practical import. 



CHAPTER VII. 

 Discussion of Results. 

 § 16. Summary of Conclusions. 



The general results of the preceding discussion, so far as concerns fluctuations in 

 the sun's radiant energy, may be summed up in the following propositions. 



J. A study of the annual departures of temperature over many regions of the 

 globe in equatorial and middle latitudes shows consistently a fluctuation correspond- 

 ing in period with that of the solar spots. The maximum fluctuation in the general 

 ■ average is 0°.13 C. on each side of the mean for the tropical regions. The entire 

 amplitude of the change is therefore" 0°.26 C, or somewhat less than half a degree of 

 the Fahrenheit scale. As this fluctuation has ample time to produce its entire effect 

 on the earth, we conclude from it that the corresponding fluctuation in the sun's 

 radiation is 0.2 of one per cent, on each side of the mean. 



2. Additional to this periodic fluctuation there is some rather inconclusive evidence 

 of changes requiring generally about six years to go through their period, which can 

 be most plausibly attributed to corresponding changes in the sun's radiation. The 

 phenomena may be expressed in the briefest way by saying that, during the years 

 1871-1904, there seem to have been periods of two, three or four years warmer than 

 the normal, followed by similar periods which were cooler than the normal. But 

 although the general tendency is toward changes in this period of about six years, 

 they show no such correspondence with the solar spots as justified their being attributed 

 to the sun-spot period. Moreover, they do not appear in any marked way before 1871. 

 The average departure from the mean being less than 0°.10 C. prevents a more exact 

 statement of their law, and still leaves open the question whether they are real. This 

 can be settled only by a more complete discussion of meteorological data than the 

 writer has attempted to make. 



3. Apart from this regular fluctuation with the solar spots, and this possible more 

 or less irregular fluctuation in a period of a few j'ears, the sun's radiation is subject Jo 



