56 smithsonian miscellaneous collections vol. 79 



Pressure. 



All the observations of pressure were made with mercurial 

 barometers which were checked with the standard barometer 

 at the Central Observatory from time to time by inspectors. 



The pressure was reduced to o° C. and to gravity at 45^ Lat. 



The mean pressures were the means of daily observations at 

 7*% 13'^ and 21^^ divided by three. 

 Temperature. 



The means of temperature were the means of observations of 

 7'\ 13^ and 21^ by the formula ^(7^'' + i3'' + 2i'i). These 

 means were reduced to the means of 24 hours by means 

 of corrections obtained from the treatise of Wild on " The 

 Temperature Conditions of the Russian Empire." 



The mean temperatures at each station were reduced to a standard 

 level for that station by assuming a decrease of —0.45° C 

 for each 100 m. 



The data for 191 6 to 1920 are not yet completed. 



SWEDEN 

 HAPARANDA 



Hours of Observation. 



At Haparanda the observation hours were 8'\ 14'^ and 2i'\ The 

 temperature means were computed by Ekholm's formula 

 (Nils Ekholm : Calcul de la temperature moyenne mensuelle 

 d I'air aux stations meteorologiques Suedoises, Appendix to 

 the Observations Meteorologiques Suedoises, vol. 56, 1914). 



Pressure. 



The pressure means are the direct mean values of the three ob- 

 servations daily at 8'\ 14'' and 2i'\ Since November 1864 

 the height of the barometer above sea level has been 9.2 m., 

 and the whole of the pressure data is reduced to that height. 

 The earlier height was 16.2 m. For the years 1860-1895 the 

 values of the pressure are taken from the treatise of H. E. 

 liamberg on "' La pression atmospherique moyenne en 

 Suede " (Kungl. Svenska Vetenskapsakademiens handlingar 

 I'd. 31, No. i). Hamberg's figures were sea-level values and 

 have been reduced back to 9.2 m. level. We have also made a 

 correction for gravity which according to the determination 

 of P. G. Rosens is some 0.05 mm. greater than that used by 

 Hamberg. In other respects the data of H. E. Plamberg is 

 good and correct ; although we must consider that of the 



