1875.] 



417 



[Chase. 



each map. I then divided the total rain-fall upon each day of the lunar 

 month by the total number of stations reporting for the corresponding 

 day, and took successive differences between the resulting averages, by 

 Airy's method. The normals thus deduced are given in the accompanying 

 table, together with the normals for various local curves. The curve de- 

 duced from 43 years' observations at Philadelphia, covers a longer period 

 than any other to which I have had access in the United States, and its 

 striking resemblance to the Signal Service curve is shown by the diagram. 

 The resemblance is the more significant in view of the fact that the 

 periods represented by the two curves are entirely independent. The flex- 

 ures in the Philadelphia curve average about If days earlier than those of 

 the general curve. On the hypothesis of cumulative tidal undulations, 

 this would represent a daily difference of Iff hours, or 221°, a difference 

 correspoading to disturbances originating in our Western territories. 



Occasional breaks in my series of weather maps, the interference of 

 storms with the transmission of reports, and other causes, combine to 

 render these results imperfect, but their indications are of such a char- 

 acter as to convince me that a careful study of the full returns, which 

 are forwarded thrice a day to the Signal Service Bureau, would lead to 

 the discovery of important laws governing the lunar influence at various 

 seasons of the year, at various periods of the day, and in various sections 

 of the country. 



Lunar-Monthly Bain-fall, from Observations of Signal Service Bureau, 

 and at Local Stations. 



