ELECTRICAL RESISTANCE VNDER PRESSURE. 93 



The wire was wound as usual bare on a bone core, and connections 

 made with spring chps, using a protective coating of gold foil at the 

 point of contact. 



Runs were made on two samples and at five temperatures, 0°, 25°, 

 oO°, 75° and 96°. The runs on the first sample were terminated by the 

 failure of the insulation of the three-terminal plug. The transmitting 

 medium was Xujol and petroleum ether. ( 'hemical action was never 

 entirelv absent, but as in the case of sodium, it was mostlv confined 

 to low pressures. It increases rapidly with rising temperature, and 

 at 96° was so rapid that 1500 kg. was the lowest pressure at which 

 readings were attempted. Unlike sodium and lithium, calcium shows 

 pronounced seasoning effects of pressure, and the runs at 25° and 75°, 

 which were those of the initial application of pressure to the two 

 samples respectively, were much less regular than the subsequent 

 runs. Because of the necessity of obtaining readings rapidly because 

 of chemical action, preliminary seasoning applications of pressure were 

 omitted, and the initial runs were included in the final results. At 25° 

 the total zero shift (initial application) was 4.2% of the total pressure 

 effect; at 0° 0.5%; at 50° 3.9%; at 75° (initial application) 6.9%; 

 and at 96° (zero taken from 1500 kg.) 3.3%. At 25° and 75° the indi- 

 vidual points lie Aery closely on two smooth curves, different for in- 

 creasing and decreasing pressure. The incomplete seasoning shows 

 itself in a sequence of readings like that of an open hysteresis loop. 

 The open end of this loop at atmospheric pressure has the width given 

 above by the zero displacements. At the other temperatures the 

 readings also show a tendency to hysteresis effects, but the departure 

 from the mean is much less. At 0° the maximum departure of any 

 single point from the smooth curve representing the mean of the 

 points with increasing and decreasing pressure is 1.2% of the total 

 effect; a^ 50°, 1.5%; and at 96°, 1.5%. 



The temperature coefficient of resistance at atmospheric pressure 

 was determined by the same method and at the same time as the 

 readings for sodium and lithiimi. The "observed" resistances, which 

 are the mean of points with ascending and descending temperature, 

 at four temperatures (0°, 25°, 50°, and 75°) all lie within the limits of 

 error (one part in 7000) on a second degree curve. The total shift of 

 the zero during the run, presumably due to chemical action, was 1.6% 

 of the total temperature effect. The \alue given in the table for the 

 resistance at 100° was obtained from the second degree curve by 

 extrapolation. The mean coefhcient between 0° and 100° determined 

 in this wav is 0.003327. 



