Meteorological Observations. 125 



the zero point, and the difficulty of ascertaining its amount. Now, 

 while the annual normals of Table i have not been corrected for 

 this source of error in any way, chiefly on account of the importance 

 of introducing no mental bias into the harmonic formulae, it seemed 

 advisable to allow for it in the monthly and annual corrections of 

 Tables 3, 4, 5 ; also an occasional libera manii correction has been 

 inserted where it seemed wanted. Nevertheless, there still remains 

 some little outstanding uncertainty in the corrections proper to dew- 

 point and humidity, particularly during the hours before dawn in 

 the winter. The three Tables in question, together with Table 2, 

 give the corrections to be applied to the mean monthly or annual 

 values of the respective elements of temperature of the air, tem- 

 perature of the dew-point, ratio of humidity, and barometric pres- 

 sure, for any hour in order to obtain the mean. A few examples 

 will make their use clear. 



At Philippolis the mean pressure at VIII. for the month of 

 Februarv, as deduced from observations made during a number of 

 years, was found to be 25-537 inches; the correction to be applied 

 to Februarv observations at VIII.. in order to obtain the mean 

 hourlv pressure of the month, is given in Table 2, as - '045 inch ; 



•■• 25-537 — ■o45 = -5'49- inches, 

 is the mean pressure required. 



At Kimberlev the mean pressure at XX. for the month of 

 October, for certain years, was determined according to the register 

 of the late G. J. Lee, to be 26'o55 inches; the correction to be 

 applied to October observations at XX.. in order to obtain the mean 

 of the month, is given in Table 2 as 4- '005 inch. 

 .'. 25*055 -i- '005 = 26'o6o inches, 



is the mean pressure required. 



At Aliwal North the mean temperature of the air at VIII. for 

 the month of November, as determined by five years' obser^-ations, 

 is 64°' I : Table 3 gives the correction to be applied to November 

 obser\ations at VIII.. in order to obtain the mean of the month, as 

 + o°-9; 



.-. 64°-i+oO-9 = 650-o. 



is the mean temperature required for the period in question. 



At Benaauwdheidsfontein the dew-point at XIV. for the year 

 was found to be 44°'4 (Glaisher's Tables being used). Our Table 4 

 gives the correction to the temperature of the dew-point at XIV., in 

 order to obtain the mean dew-point of the year, as — o°"7 ; 

 .'. 44°.4 ■ — o°'7 = 43°'7> is the dew-point required. 



The corresponding humidity-ratio was found to be 343% ; 

 the proper correction given in Table 5 is -1-19*7%; 



••- 34'3+i8'7 = 54"o%. 

 is the humidity required. 



Various combinations of hours are adopted in different coun- 

 tries for the purpose of deducing the mean temperature from second 

 order observations : 



