174 ANNUAL FLUCTUATION OF THE TEMPERATUHE. 



To exhibit the annual fluctuation in a concise form, suitable for comparisons and 

 further deductions, a number of characteristic stations have been selected, repre- 

 senting various climatological features, and for which the numerical values of the 

 several quantities entering in the expression — ■ 



r := ^ + L\ sin (0 + Ci) -f- B. sin (20 + a) + ^3 sin (39 + Q 



have been computed and tabulated. In preference, stations having long and 

 reliable series of observations have been selected, and they comprise with some 

 rough approximation to uniformity of distribution, the area of the United States, 

 with a few representative stations in Arctic and British North America. The 

 results are based on the monthly means presented in the general table of tempera- 

 tures (Section I), they were first corrected for daily fluctuation^ according to the 

 ■hours of observation, whenever needed, those depending on 7^, 2^ 9^ ms receiving 

 no correction. They were next corrected for inequaliti/ in length of months and 

 for curvature, as explained. It was deemed sufficient for the present purpose to 

 stop at the term involving i?., C3, considering that this and any subsequent term 

 represent rather local peculiarities and, moreover, are subject to considerable changes 

 with the use of additional observations. The days of average epochs of maxima 

 and minima were computed by the formula — 



Q=B, cos (0 + C) + 2B., cos (29 + Q + 3^ cos (30 + C,) 



resulting from putting '~- = 



da 



The 46 stations are given in five groups, each arranged according to latitude. 



* Excepting the results for Fort Franklin, to which no corrections whatever have been applied, it 

 is a series of less than two years. The expressions for the Arctic stations. Van Rensselaer Harbor, 

 Port Foulke, and Port Kennedy, were taken from ray discussion of the Physical Observations in the 

 Arctic Seas by Dr. I. I. Hayes; Smithsonian Contributions to Knowledge, No. 196, Washington, 

 June, 1867, p. 180. To these a fourth term has now been added, and the parameters have been 

 corrected for curvature. [On p. 180 B^ for Van Reasselaer Harbor should have been 35.39.] 



