38 RECORD AND DISCUSSION OF TEMPERATURES. 



vation of temperature by the nearly total obscurity of the atmosphere during the 

 six days in question. 



The mean amplitude of the wave in the winter 1853-54: is 19.4, and in the 

 winter 1854-55, 20. 1, or 19. 8 from an average during these two winters. 



It has been remarked by an eminent astronomer that, if the moon emits any 

 sensible heat, it is probably expended and becomes apparent by a tendency to dis 

 appearance of clouds under the full moon. Supposing this to be a fact, it would 

 seem that the powerful radiating force of the earth s surface, called into activity 

 under a clear sky, produces, as a secondary effect, the phenomenon of greatest cold 

 at the time of full moon. The process is going on gradually, and, when combined 

 with the tendency of a fall of snow about the period of new moon, would favor 

 the production of the caloric waves observed during the winter season. These 

 waves could not be explained either in range, duration, or regularity, by the effect 

 of various winds and calms, since their total effect could only amount in maxima 

 to Gi, according to the previous investigation. 



A maximum cold will be produced, as stated by Dr. Kane, by a concurrence of 

 the time of full moon with a perfect calm and a great transparency of the atmo 

 sphere, during the middle of the winter season. The opposite effect requires, for 

 its full development, a concurrence of the time of new moon with a continued fall 

 of snow, a generally obscured atmosphere, and winds from a direction between 

 N. N. E. and E. S. E. (true). 



Enough has been shown to make these alternations of relative cold and warm 

 periods in winter an interesting and instructive subject for further study, specially 

 with a view of tracing out and confirming the apparent connection of the concur 

 rence of the two principal lunar phases, with a tendency to obscurity and tran 

 sparency of the atmosphere. 



In accordance with Prof. Dove s investigations of the return of cold about the 

 llth of May, the mean daily temperature on May 13th (1854) of +2.S was 9.3 

 lower than the computed (by preceding formula) temperature. 1 



Hourly Corrections for Periodic Variations. The following table for reducing the 

 mean of observations taken at any hour of the day to the true mean temperature 

 of the day, has a similar arrangement, and was prepared for the same use, as those 

 given for other stations in the Smithsonian Miscellaneous Collections of Meteorological 

 and Physical Tables, by Prof. A. Guyot (2d edition, Washington, 1858). The 

 figures necessarily present some anomalies, since they are derived directly from a 

 series of hourly observations extending over seventeen months; they present, 

 therefore, only the differences between the hourly and the true means. 



1 While this paper was going through the press, I received the March number (1859) of the London, 

 Edinburgh, and Dublin Philosophical Magazine, containing J. Park Harrison s article on the &quot;Lunar 

 Influence on Temperature as connected with Serenity of the Sky.&quot; He states that, from 20 years of 

 observations at Greenwich, the mean temperature is above the average at the period of the new moon 

 (also at first quarter and before last quarter), it is below the average before and after full moon (also 

 between new moon and first quarter, and at and after last quarter). 



