79 



In view of the suspicious decrease of the thermal conductivity with 

 the thickness of the prism experimented on, it is well to say that some 

 manufacturers apparently use rubbers of different compositions for thick 

 and for thin plates. Much cheap thin rubber is to be had in the market, 

 and the specific electrical resistance of some of this is so low as to make 

 the material useless for insulating purposes. Of the specimens mentioned 

 in the table, G and II are pieces of expensive rubber made especially for 

 use in induction coils of high grade. K, which is nearly of the same 

 thickness as II, is an excellent rubber of high electrical resistance, but is 

 not quite equal in this respect to II, which has a very different thermal 

 conductivity. Extended experimentation seemed to show that the indica- 

 tions of the thermal elements might be trusted whether the plates were 

 thick or thin. A redetermination of the conductivities of N and O with 

 different arrangements of the prisms gave 0.000215 and 0.000199 respec- 

 tively. When a disk of the standard rubber was turned down thin, and 

 its specific conductivity redetermined, it was found to be almost the same 

 as before ; but when disks F and I were made slightly thinner by turning 

 off the polished surface on each side, the conductivity of the thinner disks 

 was increased by five per cent and three per cent respectively. This 

 seems to show that in the cases of some sheets of hard rubber the thermal 

 conductivity of the skin near the polished surfaces is somewhat less than 

 that of the mass of the material in the sheet. This difference, however, 

 is too slight to accouut for the differences of conductivity shown by the 

 table. It is well to emphasize the fact that the specific heat between 

 25 C. and 100° C. of the standard rubber is less (0.339) than that of 

 the thinnest sheet rubber (0.372) that I have experimented upon. 



