284 BRIDGMAN. 



general three tables for each metal. The first table shows the total 

 e.m.f. of a circuit composed of uncompressed and compressed metal. 

 The lower jvmction is always at 0°; the entries in the table show the 

 e.m.f. for different temperatures of the upper junction, and for differ- 

 ent pressures of the compressed branch. The temperature interval of 

 this table is 10°, and the pressure interval 2000 kg. The e.m.f. 

 corresponding to 0° is, of course, zero for all pressures. The second 

 table gives the Peltier heat absorbed by unit quantity of positive 

 electricity in passing from uncompressed metal to metal compressed 

 to the pressure in the table, both uncompressed and compressed metal 

 being at the temperature indicated in the table. The temperature 

 interval is 20°, and the pressure interval 2000 kg. in the second table. 

 The third table shows how much greater the Thomson heat is in metal 

 compressed to the pressure of the table than in uncompressed metal, 

 both uncompressed and compressed metal being at the temperature 

 indicated in the table. The temperature interval of the third table 

 is also 20°, and the pressure interval 2000 kg. The diagrams give 

 graphically the results of the tables, plotted at constant pressures 

 against temperature. The scale of the diagrams of Peltier heat and 

 Thomson heat is half that of the e.m.f. diagrams. The smaller scale 

 for the two heats has been chosen because of their smaller accuracy. 

 The corresponding tables for all metals are given in the same sized 

 units, so as to facilitate direct comparison. 



Detailed Data. 



Tin. This was Kahlbaum's "K" tin, from the same small ingot 

 as that whose resistance under pressure was measured, but not the 

 identical piece of wire. It was made by extrusion at 100° into wire 

 0.02 inches in diameter, considerably larger than the resistance wire, 

 and was annealed in an electric furnace to 120° for several hours. It 

 was not subjected to a preliminary pressure seasoning for the runs; 

 this is unnecessary for soft metals. 



Tin was one of the most unsatisfactorv metals measured, because of 

 the minuteness and irregularity of the effect. Tin is known to form 

 other allotropic modifications below 20°, the reaction usually not run- 

 ning because of internal viscosity. During the experiment, therefore, 

 the lower end of the wire was probably always in the metastable 

 region, and the readings themselves gave evidence of internal insta- 

 bility. The readings were never steady, but continually flickered 



