138 



gray: temperature uniformity 



Greater uniformity of temperature in the central region than 

 had been secured by other experimenters after repeated attempts 

 to adjust properly the crowding of windings was regularly ob- 

 tained at widely differing temperatures even when heat losses 

 through the ends were reduced merely by the action of the double 

 plugs without supplying any heating current whatever to the 

 coils in the end blocks. Adjustment of the temperature gradient 

 was considerably facilitated by the installation of the flat coils 

 covering the ends of the outer heater tube described above. 

 This change also made it rather easy to produce fair uniformity 

 for the entire length of the heater tubes — 32 cm. longer than the 

 bars the furnace was designed to heat. 



That no unusual effort is required to obtain moderate uni- 

 formity of temperature is shown by the table below, which rep- 

 resents sixteen consecutive runs made while determining the 

 expansivity of a bar of invar. Alongside the temperature ob- 

 served at the center is recorded 



Temperature Distribution in Consecutive Trials of an Electric 



Furnace 



the correction that must be added to obtain the mean temperature 

 throughout the 30 cm. occupied by the bar. In many of the runs 

 the heating currents required to maintain the desired tempera- 

 ture were estimated and switched on late in the afternoon. The 

 furnace was then left to itself over night. The temperature 

 distribution found the following morning was explored and used. 

 Reference to the table shows that in only one case (and that 

 under known unfavorable conditions) did the mean temperature 

 differ by more than 0.16° from that at the center. 



