1 1 



I '. I ' C R D A N U R EDUCTION 



Notes to tin -preceding Abstract of the Temperature Record. 



Jul}', 18-37. The column headed "mean" contains the mean daily temperature 

 derived from six equidistant observations; the figures in the next column of "de- 

 duced mean" were obtained as follows: Suppose the mean temperature of July 3d 

 be required from the observations at 8 A. M. and 8 P. M., the observations at each 

 of these hours in the full series were compared with their respective mean, as 

 given in the preceding column ; thus, from 23 values, we find the correction to the 

 8 A. M. reading, to obtain the mean reading of the day, +0°.8, and in a similar 

 manner, for the 8 P.M. reading, +0°.2. Applying these corrections to 57°. and 

 57' .5 respectively, and taking the mean, we find for July 3d the mean temperature 

 57°. 7. The following table contains these corrections to each observing hour in 

 1 he month of July, in order to produce the mean of six readings a day, viz : — 



The means require a further small correction to refer them to what they would 

 be if hourly observations had been made. For this purpose, I have made use of 

 the tables of hourly corrections for periodic variations for Boothia Felix and Dront- 

 heim, as given in the Smithsonian collection of meteorological and physical tables 

 A. Guyot, and also of a similar table given in the discussion of Dr. Kane's 

 meteorological observations for Van Rensselaer Harbor, in Vol. XL of the Smith- 

 sonian Contributions to Knowledge. For these localities, to which has been added 

 Leith, we have, for the month of July, the correction to the mean of six observa- 



