220 C. Barns — Effect of Pressure on the 



If 6 be the symbol of temperature, the following approxi- 

 mate results apply, isopiestically, at ordinary temperatures 

 and pressures: dR'/R'= 800x10" 6 dd : do/v =180x10-° 3d. 

 Hence dR'/R' = 4 - 4, dv/v, where R' refers to electrical resist- 

 ance considered in its thermal relations. 



2. Again by subjecting a concentrated solution of zinc sul- 

 phate to pressures between 10 atm. and 150 atm., isothermally, 

 I found less accurately, —dR/R=oOXlO' 6 dP. The other 

 relations corresponding to the above must be estimated: 

 -dv/v=5uxl0~ 6 8P, -dR'/R' =-04, dd, and &0>=2OO dd. 

 The chief magnitudes are here different in order and even in 

 sign from those applying for mercury. For this reason the 

 estimate made is sufficient for the following remarks. 



3. The liquids were compressed in capillaiw glass tubes, and 

 allowance made for the volume changes of glass. In case of 

 mercury I used a tubular steel piezometer of special construc- 

 tion, containing filamentary glass tabes. In most of the ex- 

 periments the steel tube was surrounded by a jacket of circu- 

 lating cold water ; but this precaution was not found, essential. 



To save space I will lump my results in a graphic diagram. 

 With mercury I made eight series of measurements, using two 

 different Bourdon gauges for pressure measurement. The first 

 of these was graduated between zero and 300 atm., and the 

 other between zero and 1000 atm. The chart, in which the 

 observations corresponding to the different series are num- 

 bered, shows the gauges to have been in satisfactory accord. 

 Otherwise there would be some obvious divergence between 

 the data of series 1 to 6 made with the first gauge, and those 

 of series 7 to 8, made with the other gauge. 



The chart is easily intelligible. The abscissas denote either 

 pressures in atmospheres, or volume decrements per unit of 

 volume. The ordinates are the corresponding decrements or 

 increments of electrical resistance per unit of resistance. The 

 curves for compression are in full lines and may be coordin- 

 ated either with pressure, or with volume decrement. The 

 curves for thermal changes of resistance (dR'/R') are given in 

 broken lines and can only be coordinated with volume decre- 

 ment. All the loci are nearly linear, seeing that the pressure 

 interval is less than 400 atmospheres. 



4. An inspection of the chart shows at once, that to bring 

 the compression loci into coincidence with the thermal loci, 

 the former must be rotated around the origin in a direction 

 contrary to the hands of a watch. The angle of rotation is con- 

 siderably greater for zinc sulphate solution, than it is for mer- 

 cury. From this follows the remarkable result, that both in 

 the case of the metal and of the electrolyte, the effect of 

 isothermal compression is a decrement of resistance nearly pro- 



