576 HOURLY CORRECTIONS FOR PERIODIC VARIATIONS. 



V Observatoire Physique Central de Russie. The observations made at Prague, Mu- 

 nich, Geneva, with those at St. Bernard, Makerstoun, Greenwich, Brussels, and Wash- 

 ington, were published by their respective Observatories ; those of Utrecht, by Buys- 

 Ballot ; of Dublin, by Lloyd, in his Notes on the Meteorology of Ireland ; those of 

 Schwerin were communicated in manuscript by Dippe ; the observations at Melville 

 Island are published in No. 42 of the Parliamentary papers for 1854 ; and those at 

 Bossekop, by Martins and Bravais, in the Voyage de la Commission Scientifique du 

 Nord. 



The tables of this second series being mostly deduced from longer series of obser- 

 vations than those in the first, when the same station is found in both, the table in the 

 second is generally to be preferred. 



Glaisher's table for Greenwich has been taken from the Greenwich Observations. 

 Captain Lefroy kindly furnished the tables for Toronto and Lake Athabasca. To 

 him the author is also indebted for the observations made at Montreal by Mr. McCord, 

 from which Table X. was computed. Table III., for Philadelphia, was deduced by 

 the writer from the observations made at Girard College under the direction of Prof. 

 A. D. Bache. 



In order to facilitate the selection of the tables, they are marked in the table of 

 contents with capitals, which have the following signification : 



A and B mean that the tables have been derived from hourly and bi-hourly obser- 

 vations, and have been computed by BessePs formula ; C, that the tables contain 

 values obtained by interpolation. 



A', B', and C' indicate the tables based respectively on hourly and bi-hourly or 

 partly interpolated observations, which give simply the differences between the hourly 

 and the true means. ,. 



The figures added to the letters indicate the number of years during which the 

 observations used in forming the table were carried on. The stations are arranged, 

 in each continent, in the order of their latitude. 



USE OF THE TABLES. 



In order to reduce meteorological means obtained from any set of hours to the true 

 means, the table best suited to the purpose must first be selected. The diurnal vari- 

 ation changing with the seasons, the latitude, the altitude, and the distance from the 

 sea-shore, the station which comes nearest, in all these respects, to the station the 

 observations of which are to be corrected, must be adopted. 



Suppose the thermometer has been observed at Baltimore, during the month of 

 January, at 7 A. M., 1 P. M., and 7 P. M., and the monthly means of these hours to 

 be respectively 27, 35, and 31 Fahrenheit. We take Table III., Philadelphia, it 

 being the nearest in latitude and climatic situation. We find the correction for the 

 hours 7, 1, and 7, and we have 



Observed Means. Corrections. True Means. 



For 7 A. M. 27 + 3.63 = 30.63 



For 1 P. M. 35 3.87 = 31M3 



For 7 P. M. 31 1.13 = 29.87 



Sums, ~9& 1.37 = "91^63 



Means, 31 0.46 = 30.54 True Mean for January. 



It is obvious that the corrections can be applied, either separately to each hour, as 

 is done above, or collectively, in taking the mean of the three hourly corrections and 

 applying it to the mean of the three observations, as in the last line, which is the more 

 convenient method. Therefore, in order to find the correction for any set of hours, 

 it suffices to take the mean of the corrections given in the table for the hours compos- 

 ing the set. The true daily means can be found in the same way, and the true yearly 

 means can be derived from the corrected monthly means, or by applying the correc- 

 tions given in the last column. 



F 12 



