compn'ss'ibUity of Water to Practical Purposes. 37 



plied to by a friend to suggest a form of guage for measuring 

 the pressure of condensed gas intended to be used for a furnace. 

 I had recently been making experiments with the very conve- 

 nient compression apparatus of Oersted, in which the changes of 

 volume of water and air are exhibited at once, under any pres- 

 sure, that of the water being sensibly uniform for equal incre- 

 ments of pressure, whilst that of air rapidly diminishes. 



It is the very trifling compressibility of water (or any other 

 liquid) which gives the value to this application, and which 

 seems to have been practically overlooked. 



The reason is obvious. The changes of volume produced by 

 a pressure of only one or two atmospheres, in the case of air, 

 are quantities very large in proportion to the primitive volume, 

 so that, in the consideration of an additional change, we are 

 obliged to take into account not merely the effect upon the 

 primitive volume, but upon the volume affected by the first 

 unit of pressure. In other words, we are not at liberty to ne- 

 glect quantities of the second order, which we may safely do in 

 the case of any known liquid. In the case of water, for in- 

 stance, the variation of volume for one atmosphere does not ex- 

 ceed ggjooo of the whole ; so that the variation of the variation 

 is necessarily insignificant. All that we know of the constitu- 

 tion of liquids, would lead us to infer, that such would be the 

 case, and upon this circumstance depends the linearity of the 

 expression, which connects the volume of a liquid such as wa- 

 ter, and the pressure to which it is subjected. 



Within ordinary practical limits, we may confidently antici- 

 pate the sensible proportionality of pressure and change of vo- 

 lume ; and this is fully borne out by a comparison of the best 

 experiments on the compressibility of water made within great 

 and within narrow limits. 



I did not hesitate, therefore, to recommend the trial of a ma- 

 nometer of water instead of air, for measuring the elasticity of 

 gas up to 40 atmospheres of pressure. 



The construction of such an instrument being almost like that 

 of the common thermometer, is incomparably simpler than that 

 of the other instruments above mentioned ; and almost the only 

 practical difficulty is common to all these, namely, the accurate 

 flttermination of the temperature of the fluid employed, 



