DAY AND SOSMAN: NITROGEN THERMOMETER SCALE 171 



ards was utilized for the most part. To this design certain modi- 

 fications suggested by the unpublished work of Prof. G. A. Hulett 

 of Princeton were added by way of rendering the determination as 

 far as practicable independent of particular experimental condi- 

 tions emplo3 T ed. 



Heat was supplied electrically from a coil of high resistance 

 wire about the sulfur tube, the coil ending about 2 cm. below the 

 surface of the liquid sulfur (Bureau of Standards usage). An 

 independent coil surrounded the vapor region, separated from it 

 by an annular air space of about 1 cm. (Hulett I. The bulb was 

 surrounded first by a shield of sheet aluminum ( Bureau of Stand- 

 ards) with holes near the top and bottom to permit the free circu- 

 lation of the sulfur vapor, and a hole in the center of the bottom 

 to permit the escape of liquid sulfur which chanced to condense on 

 the shield. The shield afforded protection against any direct 

 interchange of radiation with the furnace or with the boiling liquid, 

 and its steep conical roof diverted the condensing liquid sulfur 

 away from the bulb. Subsequently, with the purpose of varying 

 these conditions, the aluminum shield was replaced by another 

 of similar form but of glass (Hulett) and of somewhat smaller 

 diameter. This was suspended from the conical aluminum roof 

 of the first shield, which now overhung the side walls by several 

 millimeters, with the effect that liquid sulfur condensing upon 

 the cone could drip from the overhang instead of running down 

 the side wall past the bulb. The radiation conditions were also 

 radically altered by the substitution of glass for aluminum both 

 around the bulb and below it. 



Further variation was provided by changing the current in 

 the two heating coils. Variations of some 40 per cent in the main 

 coil about the boiling liquid were tried and the upper coil was 

 varied from zero (Bureau of Standards usage) to nearly 40 per 

 cent of the current in the main coil. Or, in other terms, the 1 cm. 

 air jacket about the vapor was varied in temperature from the 

 normal gradient (without heat in the upper coil) to a temperature 

 equal to that of the sulfur vapor itself. 



None of these changes produced any measurable change in the 

 temperature of the sulfur vapor as recorded by the gas ther- 



