152 SCIENCE PROGRESS 



of two kinds, which when separated by a process of picking 

 gave two sodium-ammonium tartrates, identical in all respects, 

 except that whilst one was dextro-rotatory like the salt of 

 ordinary tartaric acid, the other was laevo-rotatory to exactly 

 the same amount. The failure of subsequent investigators to 

 produce in crystallisation anything but the double racemate led 

 to a thorough investigation of the whole subject, 1 and to the 

 establishment of the fact that above 28 the double racemate 

 crystallises out, whilst crystallisation below 28 yields a mixture 

 of the two tartrates. Hence it has been said that Pasteur owed 

 his discovery to the happy chance 2 that he carried out his 

 evaporation below 28 . 



5. The Camallite Problem. — This is a case of technical 

 importance. The potassium chloride used as the starting-point 

 in the manufacture of the many potassium compounds of com- 

 mercial importance (carbonate, hydrate, cyanide, nitrate, etc., 

 etc.) is found native as the mineral sylvite, but by far the larger 

 amount of that required is obtained from the mineral camallite, 

 KCl.MgCl 2 .6H 2 0. This salt has its transition point at 168 , 

 when it breaks up into KC1, MgCl 2 .2H 2 and 4H 2 0. Owing 

 to the great difference in solubility between the constituent 

 chlorides, the transition interval is so extended that at ordinary 

 pressure there is no temperature at which carnallite can be 

 treated with water without partial decomposition and separation 

 of the less soluble KC1 as a solid phase. This is the principle 

 underlying the old empirical process of the separation of the 

 KC1, which is carried out as follows : 



Treatment with water until a solution is obtained, which at 

 25 is saturated with carnallite, results in the separation of 

 about five-sixths of the KC1 in the solid form. The solution 

 saturated for the double salt is evaporated, when carnallite 

 crystallises out, and a solution is produced which for 1,000 

 molecules of H 2 contains 105 of MgCl 2 and 2 of KC1, and 

 which therefore, for practical purposes, may be considered a 

 solution of pure magnesium chloride. This solution is disposed 



1 Van't Hofif, loc. at., p. 82. 



2 Considering the popular meaning of the term "chance," one would wish 

 to protest against its use in a connection such as the above, and to add some 

 such parenthesis as : " chance, which in the matter of great scientific discoveries 

 with unfailing discrimination and unbroken regularity selects for the recipient of 

 its gifts only the truly great." 



