46 THE WATER 



short side-branch ending blindly. On reversal the 

 mercury leaves this branch first, and then breaks at 

 the point where it joins the main thread. 



The temperature read off is not as a rule accurate, 

 but a correction must be applied to it in addition 

 to those shown on the certificate. If we cool a re- 

 versing thermometer — say, in water at o° — reverse it, 

 and then bring it into water at a higher temperature, 

 it will be seen that the temperature reading continues 

 to rise slowly for some time. This is due to the ex- 

 pansion of the mercury in the capillary ; in fact, the 

 capillary and the small bulb are acting like an ordinary 

 thermometer. 



To calculate the correction we require to know 

 (a) the temperature of the thermometer at the moment 

 of reading ; (b) the kind of glass of which it is made ; 

 and (c) the volume, expressed in degrees of the stem, 

 of the secondary bulb, and the portion of the stem 

 below the o° graduation. This information is given in 

 the case of the thermometers made by Richter. The 

 temperature at the moment of reading is shown by a 

 small auxiliary thermometer sealed up in the same 

 outer tube with the reversing thermometer. The 

 name or number of the glass is engraved on the stem, 

 thus : " Jenaer 59 III " or " Jenaer 16 III," and the 

 third quantity is also marked on the stem thus : 

 'Vol. °=ioo°," or whatever the "volume at 0° " 

 happens to be. The apparent dilatation of mercury 

 in the glass 59 III is ^ttru> in l6 ni ^sW The cor " 

 rection to be added to the reading of the reversing 

 thermometer is, in the case of 59 III, if vol. °=ioo°, 

 ~~F ioo (T/) ' wnere T is the temperature shown by 

 the reverser, and t that shown by the auxiliary. If 

 the air is warmer than the water, t will be greater than 

 T, and the correction will be negative. 





