Temperature and the Serenity of the Sky. 161 



part obtained from obsei'vations of an earlier series of years, and 

 at stations far distant from each other, follow the curve of twenty 

 years' mean temperature at Greenwich, and point to a maximum 

 of effects after first quarter, and a minimum at last quarter. 



Other evidence is to be found in Howard, who appears to 

 have failed in establishing the fact of lunar influence (in which, 

 from long-continued observations, he had the most perfect be- 

 lief) mainly from the unfortunate way in which effects, which he 

 nevertheless noticed and commented on, were neutralized by 

 mixing up the means of days, which a very slight glance at the 

 curves will shov,^ required to be kept separate. 



It should be explained that the vertical lines in the Plate of 

 Curves represent the different days, and the horizontal lines de- 

 grees and tenths of degrees of temperature, and also tenths and 

 units of cloud. The dark line intersecting each curve is the 

 mean of the period. 



P.S. In the second volume of Admiral Smyth's translation of 

 Arago's 'Popular Astronomy' (p. 313), I And the following 

 passage referring to Sir John Herschel's explanation of the moon's 

 influence on the clouds, which it will be seen he entirely adopts: — 

 " In a word, provided we do not lose sight of the fact that the 

 rays which dissipate the clouds are quite different from those 

 whose calorific qualities we have been endeavouring to estimate 

 at the instant when they reach the surface of the earth, the fact 

 which I previously called a prejudice will no longer be contrary 

 to physical laws; and we shall obtain an additional illustration 

 of the remark, that popular opinion ought not to be rejected 

 without examination." 



I find also in page 318, observations of rain arc alluded to, 

 which appear to have been made under M. Arago's personal su- 

 perintendence : — 



" The discussion of the observations made at Paris led to the 

 following conclusions : — 



" The maximum number of rainy days is found to lie between 

 the first quarter and the full moon, the minimum between the 

 last quarter and the new moon ; and the latter number is to the 

 former as 100 is to 126. The accordance exhibited between the 

 German observations and those made at Paris is, as we have 

 seen, very striking." 



M. Arago adds, however, that his assistant, M. Gasparin, 

 found at Orange that the minimum of rainy days occurs between 

 the full moon and the last quarter; but it docs not appear 

 whether this was early or late in the quadrant, or what was the 

 duration of the observations. At Montpellier, in 1777, Poitevin 

 also arrived at diffcrcut results during ten years' observations. 



