the late Dr. Joule's Thermometers, 483 



In the paper communicated to the French 1 



Academy in 1846 J 



In the paper read before the Eoyal Society 1 



in June 1849 j 



one 



division = 



.•042772 





)» 



•042897 





>> 



>> 



•042902 

 012791 



In the paper read before the Eoyal Society "1 



in Jan. 1878 J 



In the last mentioned paper, calculated 1 



from the actually observed boiling-point J 



It will be seen that the last value nearly agrees with the 

 first; the point to be explained therefore is the high value of 

 the second number. It is possible that a correction similar 

 to that of the last paper w r as already then applied, and as we 

 have reason to believe that a great part of the change of zero 

 took place in the first four years, we may account for certainly 

 half the difference in this way. The point is not now of great 

 importance, because the scale-value of A must be obtained 

 quite independently of Joule's assumed interval for his 

 standard. Joule's own value depended not only on that 

 interval but also on the correct calibration of his standard. 

 From the method of calibration employed, an error quite as 

 large as that caused by the wrong value of the boiling-point 

 might easily be introduced. Nevertheless a re-determination 

 of the distance between boiling- and freezing-points seemed 

 to me to be of interest, especially as the depression of the 

 zero might give some indication as to the nature of the glass 

 of which the thermometer is made. 



When the thermometer came into my possession there 

 was a large bubble of air in the bulb, and the mercury in 

 the stem broke into pieces when attempts were made to 

 drive the bubble into the upper reservoir. I finally suc- 

 ceeded, however, in removing it, but the experiments on the 

 boiling-point were always a little difficult as the mercury 

 when placed in steam had a great tendency to distil into the 

 upper parts of the stem. In order to see that the thread was 

 continuous the thermometer had to be inverted occasionally, 

 the mercury running into the reservoir, and small pellets 

 sometimes remained there on re-inverting, so that no value 

 is to be attached to the actual position of the zero points 

 observed ; though the quantity of mercury separated was 

 always so small that the distance between the freezing- and 

 boiling-points could not be affected. Table II. gives the 

 observations made ; — 



