Range of Atmospheric Temperature. 



591 



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

 Winter lunar ef- 1 



feet, whole se- I ii'iS ii"37 n'S^ io'88 10*52 io'49 i°79 ii"05 



Do. 1866-75 (M) ... I1-29 ii"44 ii"4o ii'oo 10-57 10-17 10-69 n'lS 

 Do. (min. years) (N) 10-75 10-86 10-99 10-58 10-13 10-33 10-56 10-83 

 Do. (max.years) (P) 11-76 12-05 ii'75 ii'29 ii-o6 10-69 ii'i2 11-34 



Judging from these results, we would remark that the great and per- 

 sistent effect o£ the lunar winter variation renders it difficult to entertain 

 any doubt of its existence. 



17. It will now be desirable to indicate one result of the supposition 

 that the lunar effect is somehow connected with the state of our lu- 

 minary *. 



Mr. J. A. Broun, in a series of very interesting investigations, has in- 

 dicated the possibility that the sun is one-sided in his action on the earth, 

 in other words, that a certain heliocentric longitude or section of the sun 

 is especially influential. From the results of this paper it may likewise 

 be imagined that a particular configuration of the sun and moon is espe- 

 cially influential in a meteorological point of view. 



Suppose now, that, although not a year of maximum sun-spots, there 

 are yet a great number, and that the influential solar longitude is turned 

 towards the earth when the moon is in her most favourable position. 

 This favourable conjunction may occur several times in succession, inas- 

 much as the period of the sun's rotation does not differ greatly from the 

 lunar period, and thus we might have a meteorological effect produced 

 even rivalling that of the solar maximum sun-spot period. 



In fine, without confining ourselves to a particular hypothesis, it is 

 evident that if the sun and moon have both to be taken into account 

 as regards meteorological action, the meteorological influence may not 

 be always proportional to the sun-spot area. It may perhaps be 

 thought to militate against the hypothesis of a connexion between the 

 lunar effect and the sun period, when we see, as above, that the first 

 eleven years of the series, taken consecutively, exhibit a less lunar effect 

 than the last ten years. To this, however, it may be replied that un- 

 doubtedly the last ten years, embracing the maximum of 1870, represent 

 greater solar activity than the first eleven years. 



Again, if the range of temperature for different years were merely a 

 solar effect, having no connexion with the moon, we might expect 

 the summer range and the winter range to be increased and dimi- 

 nished in the same or nearly the same proportion. Now, if in Table I. 

 we regard the years 1858-61, 1868-72 as representing both years of 

 solar maximum and years of maximum temperatare-range, we shall 

 find that the sum of the mean ranges for the six winter months of 



* This is the hypothesis assumed by Mr. J. A. Broun to explain the facts of terres- 

 trial magnetism. 



