106 BKIDGMAN. 



Cadmium. The material used for the experiments was " chemically 

 pure " cadmium from Eimer and Amend. I have no analysis. In all, 

 eight sets of readings were made, four by the radial flow method, and 

 four by the longitudinal flow. In general the results were not good. 

 A number of smooth curves were obtained by the radial flow method, 

 but the numerical values for the different specimens varied greatly, 

 the extremes being 5.0 and 31.5%. One of the radial flow specimens 

 had a central silver tube to contain the heating element; the results 

 with this were very irregular. The three other specimens had a 

 0.007 inch central hole, and the results with them were not so irregular. 

 The best mean values with the four radial flow specimens were re- 

 spectively 5.8, 26.7, 4.2, and 14.8%. The mean of these is 12.9%, 

 or discarding the highest as differing too much from the others, the 

 mean of the three remaining is 8.3%. Doubtless much of the irregu- 

 larity in the results by the radial flow method was due to imperfections 

 in the castings. 



The individual readings for any one specimen by the longitudinal 

 flow method were widely scattered, and would not give much confi- 

 dence in the correctness of the result if it were not for the fact that the 

 different specimens gave results in agreement. The longitudinal flow 

 specimens were two of them cast and two extruded. One of the 

 castings was annealed at 130° for 30 minutes, the other not. The 

 extruded specimens were seasoned by a preliminary application of 

 12000 kg. These differences in treatment seemed not to affect the 

 results. The last of the longitudinal flow specimens did not give 

 results of value; the other three specimens gave changes under 12000 

 kg. of 9.0, 8.7, and 8.7% respectively. The magnitude of the correc- 

 tion for the pressure effect on the petroleum ether was 3.8%. The 

 best mean coefficient from these measurements is taken to be 8.9%. 



The pressure coefficient of thermal conductivity given by the 

 above mean value is +O.O574. For the mean pressure coefficient of 

 electrical resistance between and 12000 kg. I found 0.05932. The 

 results for the effect of pressure on thermal conductivity were not 

 sufficiently accurate to establish any departure from linearity with 

 pressure. The effect of pressure on electrical resistance becomes less 

 at the higher pressures; I found the initial coefficient to be O.O4IO85, 

 and the mean O.O5912. These results may be compared with those of 

 Lussana. Up to 3000 he found the effect of pressure on both thermal 

 and electrical conductivity to be linear; the coefficient of thermal 

 conductivity he found to be O.O4I22, and of electrical conductivity 

 O.O592. The first is higher than my value, and the second lower. 

 Lussana's value for the pressure coefficient of thermal conductivity 



