andelaslic Fluids, and onthe Measurement of Temper ai?ires. 447 



The march of the mercury in the ordinary thermometer mea- 

 siU'es the excess of the dilatation of this Hqxiid over that of the 

 substance which serves it for an envelope : when this substance 

 varies, the apparent dilatations of the mercury change ; but they 

 «-i11 all follow th.e same law, if the bodies of which the envelopes 

 are fohned have themselves a similar law of dilatation. M.Biot 

 has already made a fortunate application of this truth to the de- 

 termination of the real maximum of density of water, and to the 

 inquiry into the laws of dilatation of various liquids. 



It was by resting on the same principle that we endeavoured 

 to compore the dilatations of the glass and iron at various tem- 

 peratures. In order to know the apj)arent dilatation of the 

 mercury in the glass, it is sufficient to observe the thermometer 

 itself: in order to measure this same dilatation in the iron, it is 

 sufficient to cotistrutt a thermometer the reservoir and tube of 

 which are of iron. We cannot, it is true, estimate directly in 

 this case the iticrease of volume ; but by substituting the mea- 

 surement of the weights for that of the volumes, the determina- 

 tions are not less exact, when we operate on a large mass. 



The instrument which we made use of consisted of a cylin- 

 <lrical reservoir of hammered iron, very homogeneous, which 

 might contain about 135 grammes of mercury. It is surmounted 

 by a cone taken in the same mass, and terminated by a small 

 hollow steel tube, the interior diameter of which does not exceed 

 a demi-millimetre. We may screw to the upper part of this tube 

 another reservoir of the same metal, the capacity of which is 

 double that of the lower: when the two pieces are joined, the 

 extremity of the tube enters two or three millimetres into the in- 

 terior of this cylinder. We see by this arrangement that we 

 mav fill the reservoir and tlie tube like a thermometer, by boil- 

 ing at different times the mercurv, so as to avoid all suspicion of 

 the air or humidity having any eifect. 



When tlie vessel has assumed the desired temperature, we un- 

 screw tlie upper reservoir, and the lower will be found perfectly 

 filled with mercury. On submitting it aftervvards to a higher 

 temperature, the mercury which issues from the thermometer 

 of iion may be received into the upper reservoir. By weighing 

 precisely the empt\ vessel, and this same vessel containing the 

 mass of mercury which filled its capacity at the different tem- 

 peratures observed, we may sec by a vcr)' simple calculation, 

 if the supposition of a siitiiiar law of dilatation in the iron and 

 the glass will make the results of the experiments agree with each 

 Other. 



We could not expect to find in the interval of the first hnn- 

 dred thcrmomctrical degrees a very sensible variation in the ratio 

 of the dilatations of the two subslances which wc compared j 



but 



