152 ANNUAL RECORD OF SCIENCE AND INDUSTRY. 



other. In order that the meteorologist may at any time 

 test the perfection of the vacuum within his tube, Mr. Hicks 

 lias very ingeniously inserted two wires into the sides of 

 the bulb in such a way that a galvanic current applied to 

 the wires will, by the nature of the light that is spread 

 through the vacuum bulb, show with considerable accuracy 

 what proportion of gas, and especially of watery vapor, is 

 there present. A pressure within the vacuum bulb exceed- 

 ing one tenth of an inch of mercury is not admissible if an 

 accurate instrument is desired, and the vacuum can be easily 

 brought to within one fiftieth of an inch, in which condition 

 the radiation solar thermometers will prove strictly compa- 

 rable. Especially is it important that the bulb should be 

 filled with dry gas, and that not the slightest trace of moist- 

 ure should exist. Mr. Hicks said that, although he had made 

 hundreds of tubes with Torricellian vacua, he never knew 

 one to fail showing stratification and white light when the 

 tube w T as thoroughly clean and free from moisture. Quar. 

 Jour. Meteor. Soc. y II., April, 1874. 



THE THERMAL CONDUCTIVITY OF MERCURY. 



Herwig has been continuing the inquiry previously insti- 

 tuted as to whether the thermal conductivity of mercury 

 varies with the temperature a question of much moment in 

 connection with the reliability of the indications of the mer- 

 curial thermometer at different temperatures. He finds that 

 between 40 and 160 Centigrade the heat-conducting power 

 of pure mercury is perfectly constant. He is now occupied 

 in a series of experiments to show how far solid metals differ 

 in their behavior from mercury. 13 A, Feb. 27, 1875, 222. 



A NEW SOURCE OF ERROR WITH THE MERCURIAL THER- 

 MOMETER. 



Mi*. J. M. Morgan, in employing a mercurial thermometer 

 in the operation of distillation, the instrument being in- 

 serted into the apparatus to such a depth that the whole of 

 the quicksilver thread was surrounded by heated vapors, 

 observed after the operation had continued for several days 

 that the temperature registered was too low by 3. An ex- 

 amination showed that this error was due to the fact that 

 a portion of the mercury had vaporized and condensed in 



