Richards — Transition Temperature of Sodic Sulphate. 205 



While it seemed probable that this recrystallization had 

 freed the salt from anything which could affect its melting 

 point, further proof seemed proper. Hence a kilogram or so 

 of pure sodic hydric carbonate was thoroughly washed with 

 pure water, and was then dissolved in an excess of pure dilute 

 sulphuric acid. After six crystallizations the salt was still per- 

 ceptibly acid to methyl orange, and hence after testing the 

 melting point the salt was recrystallized again, collected on a 

 large Gooch crucible and washed. Although now free from 

 any trace of acid, it was precipitated as anhydrous salt by heat- 

 ing to 100° (the mother liquor being rejected), redissolved at 

 33°, and recrystallized by cooling. These final operations 

 were conducted with the purest water and all the precautions 

 necessary in the most refined work. A portion of the later 

 mother liquors was neutralized with pure soda and twice 

 recrystallized in order to obtain a third sample. 



Another thermometer was used in testing the melting points 

 of these three samples. This thermometer, admirably made by 

 Bandin of Paris in 1880, is graduated to the fiftieth part of a 

 degree, and can be read by means of a telescope and microme- 

 ter to within a thousandth. Its scale is about forty centimeters 

 long, and covering the twelve degrees between ^1° and 33°, 

 including the zero point below a small auxiliary bulb. This 

 small bulb was always in the thermostat, leaving only about 10° 

 of the column to be corrected for the lower temperature of the 

 room. Following are the uncorrected results obtained with 

 this thermometer : 



The slightly acid salt melted at 32*520 



The neutralized and recrystallized salt melted at .32*560 

 The purest, nine times recrystallized salt melted at. 32*561 

 The twice recrystallized salt first made melted at. 32*560 



The last observation was made in order to compare this 

 series with the other which has been made with the Beck man n 

 thermometer. As the pressure had increased since the other 

 readings were taken, the arbitrary thermometer itself could not 

 be satisfactorily compared with that of French manufacture. 



Evidently it is a matter of no great difficulty to obtain sodic 

 sulphate of constant "melting point." Even two recrystalliza- 

 tions of the purest Glauber's salt of commerce carried on in 

 porcelain vessels seemed to be- enough to eliminate any serious 

 impurity. In order to obtain certain results, the purification 

 should be continued until the melting point is constant. The 

 repeated recrystallization of the slightly acidified salt is 

 especially instructive in this regard, for it is safe to conclude 

 that when the acid has gone other accidental impurities would 

 have disappeared also ; and the methyl orange test is one of 

 extreme delicacy if the solution is not too concentrated. 



