130 



In the tables annexed table I gives the mean monthly 

 readings of the thermometers for each year, with the addi- 

 tional readings of the black bulb freely exposed, and also 

 that of the maximum thermometer in the shade. 



Looking at the yearly means, the black bulb in vacuo 

 gives the highest mean reading, the one with carbonic acid 

 gas comes next, followed by the condensed air one, that 

 filled with hydrogen giving the lowest temperature. 



Examination of the monthly values shows that the maxi- 

 mum for all the thermometers occurs in July, and the 

 minimum in January. The minima of the enclosed ther- 

 mometers read nearly all alike; with the maxima the vacuo 

 and carbonic acid ones are nearly equal, and the same remark 

 applies to the hydrogen one and the one filled with com- 

 pressed air ; the latter agrees with what Tyndal points out, 

 that hydrogen and atmospheric air absorb heat equally. 



Table 3 gives the differences of each monthly mean 

 referred to the reading of a freely exposed black bulb 

 thermometer. 



In volume 5, page 169, of Symons's " Meteorological 

 Magazine," there is a paper by Mr. Francis Nunes, giving 

 comparisons of carefully made black bulb thermometers by 

 Pastorelli, showing a considerable difference between the 

 thermometer in vacuo and the one partially exhausted ; his 

 observations were made in October, and show a difference of 

 1'2° to 11*5, the vacuo thermometer being the highest of the 

 two. Mr. Nunes also states that an enclosed thermometer 

 without any exliaustion reads still lower, being from 0*8° to 

 12 '8 below the vacuum thermometer. 



From my observations the difference between the vacuo 

 and condensed air thermometers is never very large, 

 amounting rarely in individual cases to 5'0° to 6'0°, but in 

 July, 18C5, reached occasionally 10-0° : the mean difference 

 in July only reaching 4*3°. 



I am not aware of any similar series of observations to be 

 found anywhere else, and thought it might be desirable to 

 tabulate the values for comparison with any subsequent 

 series that may be made. 



