452 



A. deF. Palmer, Jr. — Ajiparatusfor 



The lower connection is made by a thick copper wire soldered 

 to the bottom of A (not shown).' The space between A and D 

 is filled with heavy cylinder oil and the top of the cavity is 

 closed by a tinned' steel screw 13, five-eighths of an inch' in 

 diameter and three inches long, actuated by a large tap wrench. 

 The whole apparatus is surrounded by a' tin can, KK, filled 

 with melting ice to keep the temperature constantly at 0° C. 



The method used for measuring the changes in the electrical 

 resistance of the mercury thread is illustrated diagramatically 

 in fig. 2. B and G are the battery and galvanometer respec- 

 tively, P Q equal resistances, R a known resistance, S the mer- 

 cury thread, and C D a bridge wire. K is made nearly equal 

 to the largest value of S and balance is produced by moving 

 the contact A along the wire. Then if x and x denote two 

 positions of A corresponding to different values of S and ft 

 represents the resistance of unit length of the wire, we have 

 for dS, the variation of S between the two settings, 



dS = (3(x —x) 



The sensitiveness and accuracy of the method depend largely 

 on the sharpness of the contact A and the condition of the sur- 

 face of the wire, but with reasonable care these conditions can 

 be kept constant for a long time. 



In a previous communication"* I showed that, up to two 

 thousand atmospheres, the relation between the electrical 

 resistance and pressure of mercury is given by the linear equa- 

 tion 



R=R (l-aP) 



where R is the initial resistance, R the resistance at the pres- 

 sure P, and a a constant coefficient. Transposing, we have 



p = R o- R 



aR 



*This Journal, IV, iv, p. 1, 1897. 



