Richards — Transition Temperature of Sodic Sulphate. 207 



In order to decide more definitely the exact value, Professor 

 Sabine was so kind as to bring from the Jefferson Physical 

 Laboratory of Harvard University two of the admirable stand- 

 ards in use at that institution. These two instruments, the 

 thermometers Xos. 11,142 and 11,143, of Tonnelot, have been 

 subject to extremely minute scrutiny and calibration. They 

 are of course accompanied by a very detailed report 

 from the Bureau Internationale des Poids et des Mesures. 

 and upon this are based all the corrections recorded 

 below, except the last. This one, necessitated by the 

 fact that the thermometers were not wholly immersed in the 

 constant mixture, was calculated from the relative cubic expan- 

 sion of glass and mercury. The length of the column exposed 

 to the temperature of the room (25*5°) was estimated carefully 

 with the help of another thermometer, and was found to be 

 about 22° allowing for the conductivity of the glass and mer- 

 cury. The correction, 0*024°, agrees closely with Rimbach's* 

 empirical values, although his were not intended for such small 

 corrections ; probably it is not more than 0*003° in error. 



Both thermometers are about six decimeters long, and have 

 their hundred degrees divided into tenths. With the help of 

 the accurate Geneva cathetometer employed, they could prob- 

 ably be read to within jl^°. The readings* were made with 

 the cooperation of Professor W. C. Sabine. 



The Tkaxsitiox Temprrati're of Sodic Sulphate. 



Thermometer Thermometer 



11,142 11,143 



Tonnelot. Tonnelot. 



Reading at transition point +32*400 + 32-4f»> 



Reading in melting ice + 0-120 + 0-101 



Uncorrected difference. ._ +32-280 +32-397 



Correction for calibration + 0-147 + 0-028 



interior press. (32°) + 0-0330 + 00328 



(0°) — 0-0076 — 0-0075 



" exterior press. (32°) — 0*0006 — 01006 



" " " (0°) ...+ 0-0007 + 0-0007 



" size of divisions + 0-0104 + 0*0059 



" for projecting column + 0*024 + 0'024 



u 

 u 



Transition point according to mean Paris- 

 ian mercury thermometer .+32*487 +32*480 



The purest sodic sulphate, and ice from water which had 

 just been twice distilled in platinum (from permanganate, and 

 then from a trace of acid potassic sulphate) were used, and the 

 usual precautions of all kinds were taken. The readings were 

 made with the thermometers vertical: they are reduced to the 

 horizontal position above. The corrected barometer stood at 

 761 mm during the higher readings, and at 762*5 when the ther- 



* Rimbach, loc. cit. 



