METEOROLOGY. 



333 



The equatorial curve will be observed to have two maxima at the 

 equinoxes in March and September, and two minima at the solstices 

 in June and December. Since the earth is nearer the sun in March 

 than in September, the curve shows a greater intensity in the former 

 mouth, other things being equal. 



In the latitude of 10° the sun will not be vertical at the summer 

 solstice^ but only when the declination is 10° N., which happens 

 twice in the year. The curve corresponds in every particular with the 

 known course of the sun. Above the latitude of 23° 28' the tropical 

 flexure entirely disappears ; and there is only a single maximum at 

 midsummer. 



For comparison with the curves of intensity, I have also traced curves 

 o{ temperature observed at Calcutta, in lat. 22° 33' N.; at New Orleans, 

 in lat. 29° 57'; and at Philadelphia, in lat. 39° 57'. The curve for 

 Philadelphia is adjusted from the daily observations made at the 

 Girard College Observatory from 1840 to 1845, under the direction of 

 Prof. Bache. The rest are interpolated graphically from the mean 

 monthly temperatures. 



Retardation of the effect. — In the temperate zone the temperatures 

 will be seen to attain their maximum about one month later than the 

 sun's intensity would indicate. At Stockholm it is somewhat more 

 than a month ; and, during this interval the earth must receive dur- 

 ing the day more heat than it loses at night ; and, conversely, after 

 the winter solstice it loses more heat during the night than it receives 

 by day. In illustration of this point, and to approximately verify 

 the formula, I here insert a former computation of the sun's intensity 

 for the 15th day of each month, on the latitude of Mendon, Mass., and 

 the results are found to agree very nearly with those observed at that 

 place about one month later, as follows : (The observed values are 

 taken from the American Almanac for 1849, and are derived from fif- 

 teen years' observations.) 



It may be proper to observe that the formula was divided by sin L, 

 a constant factor ; and the numbers in the second column were then 

 successively computed: their sum, divided by twelve, gave 10163 as 

 the mean, to be compared with 47°. 1, the observed mean at Mendon. 

 Then as 10163 : 47°. 1 : : 5040 : 20°. 3, Jan. 15, &c. Let it also be 



