208 Richards — Transition Temperature of Sodic Sulphate. 



mometers were in ice. The middles of the bulbs were about 

 45 mm below the surface of the inverting mixture and about 

 3Y mm under the ice water. The pressures corresponding are 

 about 4*5 ,nm and 5*7"" n of mercury respectively. The sum 

 total of the corrections for exterior pressure is wholly insig- 

 nificant; they are included above only for the sake of com- 

 pleteness. One division of No. 11,142 equals 1*000320 standard 

 mercury degrees, while one of No. 11,143 equals 1-000181 

 degrees. 



The two results (32*487° and 32*480°) are probably as nearly 

 equal as could be expected, considering that three of the four 

 readings unfortunately fell on dividing lines. 



The agreement of the average 32*484 with the result 32*481 

 obtained from Baudin's uncorrected instruments is notable ; 

 evidently the empirical graduation of the older day included a 

 close approximation to most of the corrections. The result of 

 Lowenherz, 32*39°, is the only previous value with which it is 

 worth while to compare this, but his was meant only for rela- 

 tive work and was referred merely to " a normal thermometer " 

 of uncertain origin. 



While it would be unsafe to predict the precise limit of 

 error of the final value 32*484 (or 32*379 referred to the hydro- 

 gen thermometer) the result is certainly near enough to the 

 truth to be a great boon to those who have not a standard 

 instrument at their disposal. If, as is usually the case, it 

 is only desired to attain a result accurate to within T ^ 

 degree, the thermostat is unnecessary. Surprisingly accurate 

 results may be obtained in the ordinary Beckmann freezing- 

 point apparatus, if the outer bath is a degree or two above the 

 point desired.* In fact, it is really an easier point to use than 

 either of the old points, for one can keep neither ice nor steam 

 always at hand in a bottle. Probably the inversion tempera- 

 ture might be of great use as a means of maintaining large 

 vessels at a perfectly constant temperature : this point deserves 

 consideration. Moreover, while sodic sulphate has several 

 advantages as a standard, there are undoubtedly many other 

 substances which would also serve at other temperatures. One 

 of these, calcic nitrate, inverting at about 42*7°, is at present 

 under investigation at this laboratory, and it is hoped that a 

 complete temperature-scale of standards may be established. 

 It is also hoped that with an admirable new air-thermometer, 

 at present nearly finished in the Jefferson Physical Laboratory, 

 a direct determination of the transition temperature of sodic 

 sulphate may be made in terms of the hydrogen standard. 



In brief, this preliminary paper shows that sodic sulphate 

 " melts" at almost exactly 32*48° according to the mean mercury 

 thermometer, and that this temperature is so easily obtained 

 and so constant as to be of great use in the future of thermom- 

 etry and thermostatics. 



Chemical Laboratory of Harvard College, June, 1898. 



* For practical details, pp. 203-206 should be consulted. 



