1876.] of the Daily Range of Atmospheric Temperature. 



157 



But, on the other hand, and against this evidence, there is a temperature- 

 fluctuation between 1859 and 1866 as great, or nearly as great, as any 

 which apparently corresponds to sun-spot period. This temperature- 

 oscillation may perhaps be identified with a subsidiary sun-spot fluctuation 

 as exhibited in the curves of Messrs. DeLa Rue, Stewart, and Loewy, but 

 it is out of proportion to it in relative magnitude. 



If we still regard it as most likely, though not pro^"en, both from the 

 evidence of the paper and from collateral considerations, that there is some 

 connexion between the daily temperature-range and the state of the sun 

 with regard to spots, then we may perhaps suppose that this redundant 

 temperature-oscillation is a local phenomenon. There is, however, another 

 possible explanation which will be afterwards alluded to. 



It is then endeavoured to ascertain whether the temperature-range has 

 any reference to the relative position of the sun and moon. Tor this 

 purpose the whole period of 21 years has been portioned out into luna- 

 tions, each lunation being divided into 8 parts : — (0), (1), (2), (3), (4), 

 (5), (6), (7) — (0) corresponding to new and (4) to full moon. 



Trom the whole series of lunations the following result is obtained : — 



Phase of lunation (0) (1) (2) (3) (4) (5) (6) (7) 

 Temp.-range 1408 14-20 14-29 1405 13-95 13-83 1404 14-17 



which presents the appearance of a curve of double period superposed 

 upon one of single period. The range, however, is not great. 



If we now make use of the lunations corresponding to the six winter 

 months (October to March), we obtain : — 



Pkase of lunation (0) (1) (2) (3) (4) (5) (6) (7) 



Temp.-raBge 11-18 11-37 11-32 10-88 10-52 10-49 10-79 11-05 



Treating in the same way the limations for the six summer months, we 

 obtain : — 



Phase of lunation (0) (1) (2) (3) (4) (5) (6) (7) 

 Temp.-range 1696 17-02 17-23 17-22 17-35 17-15 17-24 17-27 



It is then noticed how large and persistent the winter lunar variation 

 is, and how the series of observations may be split into two parts, each of 

 which represents it. Thus we obtain : — 



Phase of lunation (0) (1) (2) (3) (4) (5) (6) (7) 



Winter variation (1855-65) 11-08 11-30 11-24 10-77 10-50 10-70 10-89 10-93 

 (1866-75) 11-29 11-44 11-40 ITOO 10-57 1017 10-69 11-18 



Confining ourselves to the winter variation, which it thus appears is 

 very prominent, it is then attempted to show that this variation (as far as 

 we can gather from the Kew observations) appears to vary with the sun- 

 spot period, being greatest at times of maximum spots and least at times 

 of minimum spots, very nearly in the proportion of two to one. 



Allusion is then made to certain recent researches of Mr. J. A. Broun, 

 in which he shows that the sun appears to be one-sided as far as his action 

 of certain kinds upon the earth is concerned. From this, and from the 

 fact that while the moon appears to be concerned in the temperature-range, 



