104 



BEIDGMAN. 



some reason it was not possible to make such good contact between 

 silver and tin as between siher and lead. It may be that the greater 

 ease of alloying between silver and tin resulted in inhomogeneity of 

 the casting in the neighborhood of the central tube. Fairly consistent 

 results were obtained, however, from those castings in which the 

 heating element was placed in an axial hole 0.007 inch in diameter 

 cast with the specimen. The results with the poor specimens which 

 were discarded varied from 9 to 45% (the latter was for the split 

 sheath which showed a progressive change with successive applica- 

 tions of pressure). The results with the better specimens are shown 

 in the table. 



TABLE II. 

 Summary of Results for Tix. 



It is to be noticed that the results with the longitudinal flow speci- 

 mens are lower than those for the radial flow specimens. The indi- 

 vidual readings were not so good by the longitudinal method. It 

 would perhaps have been bettcF to have used a piece of extruded tin 

 wire rather than the casting, in order that the specimen might be more 

 homogeneous. The correction for lateral flow in the longitudinal 

 method was 4.3 %, about a quarter of the total eifect as measured. It 

 is further to be noticed that the readings on the radial flow specimens 

 which were repeated agree with each other more closely than they do 

 with the mean, so that apparently the difl^erences between the speci- 

 mens are real. This may perhaps be partly ascribed to the effect of 

 crystalline structure. Tin does not crystallize in the regular system, 

 and is distinctly more crystalline in character than lead, for example, 

 so that it is not surprising that there should be differences. 



