202 PROCEEDINGS OF THE AMERICAN ACADEMY. 



Tammann 1 on one occasion reached 5000 kgm. per sq. cm., and 

 Carnazzi, 2 working with Lussana's 3 apparatus, has also attained 

 5000 kgm. Both of these observers measured the pressure with a 

 secondary gauge involving directly the compressibility of water as 

 found by Amagat. But because Amagat's data run to only 3000 kgm., 

 the pressure measurements of both Tammann and Carnazzi must be 

 uncertain at these higher pressures. Tammann had to content him- 

 self with an extrapolation beyond 3000, and Carnazzi does not give 

 any results beyond 3000. 



The purpose of this paper is to provide data which shall enable 

 others to work, if they desire, beyond Amagat's pressure range with a 

 reasonable degree of confidence in the accuracy of the pressure meas- 

 urements. It seems that the most feasible way of doing this is to de- 

 termine, under conditions that may be reproduced with accuracy, the 

 variation with pressure of some easily measurable physical property. 

 Compressibility does not seem to be the best secondary property for 

 this purpose, for it cannot be measured with much accuracy con- 

 veniently. In this paper, advantage has been taken of a suggestion of 

 de Forest Palmer's, 4 and the variation of the electrical resistance of 

 pure mercury under pressure has been determined. The secondary 

 gauge, involving the variation of the resistance of mercury, has proved 

 itself trustworthy and accurate. 



This matter of a secondary standard is discussed in the second part 

 of this paper. The first part is occupied with a discussion of the 

 slightly novel form of gauge with which the fundamental direct meas- 

 urements of pressure were made. Amagat's manometre a pistons 

 libres is not well adapted for high pressures. Amagat himself was 

 accustomed to use it to only 3000 kgm. per sq. cm., and others follow- 

 ing him have not been so high ; thus Barus found that above 2000 kgm. 

 the leak became troublesome. In this paper a gauge is described with 

 which, by modifying the design and decreasing the dimensions, it has 

 been found possible to reach higher pressures than Amagat, frequently 

 without perceptible leak. In fact the primary gauge proved itself so 

 manageable, and is so simple to construct, that if it were not for the 

 greater convenience of the secondary gauge, the primary gauge could 

 be used directly in any high pressure investigation. This paper gives 

 results that have been obtained with this gauge up to 6800 kgm. per 



1 Tammann, Kristallisieren und Schmelzen, p. 201 (Leipzig, Barth, 1903). 



2 Carnazzi, Nuov. Cim., (5), 5, 180-189 (1903). 



3 Lussana, Nuov. Cim., (5), 4, 371-389 (1902). 



4 de Forest Palmer, Amer. Jour. Sci., 6, 451-454 (1898). 



