EXPERIMENTAL STUDY OF HEAT LEAKAGE. 



757 



c. Radiation effects. 



In a further study of the question of heat-conduction down the walls 

 of the thermometer tube, undertaken about a year after the experi- 

 ments just described, some unexpected eifects due to radiation were 

 noticed. These experiments were made with a specially-constructed 

 thermometer, the coil of which is illustrated in Fig. 9. As is indicated, 



Figure 9. Coil of the experimental platinum resistance thermometer 



No. 8. 



the coil is wound on a mica cross supported by means of four quartz 

 tubes which also serve to insulate the leads. The containing tube, of 

 uniform bore throughout and large enough to permit the coil to slide 

 in or out easily, may be of either steel or glass. No insulating joints 

 were used. This thermometer is called No. 8 in what follows. 



The experiment consisted in inserting this thermometer, together 

 with a compai'ison thermometer (the one used in the experiments just 

 described, hereafter called No. 4) in the vapor of naphthalin boiling 

 in the hypsometer (see below, page 762) regularly used for comparing 

 thermometers. Several measurements, extending over twelve days, 

 were made with both the steel tube and the glass tube. The apparent 

 temperature by No. 8 was invariably higher when the steel tube was 

 used. Omitting one or two measurements, doubtful because some 

 naphthalin penetrated to the interior of the glass tube before a proper 

 method of packing it was discovered, the results were as follows, in 

 chronological order: 



* This is the temperature equivalent of R%~Ri on the .scale of 4. It would 

 be slightlj' different on the scale of 8, but the hundredths of degrees would 

 be the same on the two scales. It is of course immaterial which scale the 

 difference is referred to, as far as the comparison is concerned. 



