APPLICATION OF EQUILIBRIUM LAW TO SEPARATION OF CRYSTALS. 281 



The order in which separation occurs is at once given by reference to 

 a vapour pressure diagram constructed, e.g., by inserting the 'vapour- 

 pressure difference ' at each of the various transition points. 



The Evaporation of Sea Wate?-. 



On concentrating sea water — disregarding Calcium sulphate on account 

 of the small quantity present — the first salt to crystallise out is Sodium 

 chloride. When deposition of this salt sets in, the solution has the com- 

 position : 



l.OOOHsO 47Na2Cl2 l-OSK^Cla 7-36MgCl2 S-STMgSO^, 



Following the rules previously given, it is obvious that the position in 

 space of the point n, which represents a solution of this composition, will 

 be 3-57 — 1-03 = 2-54 units on the ox axis to the left of the origin, 7-36 

 units above the origin on the OY axis, and l-03 + 7-36 + 3-57 = ll-96 

 units above the plane. 



As long as only Sodium chloride is deposited, the relative proportions 

 of the Potassium and Magnesium salts remain unchanged, and only the 

 amount of these salts present relatively to the water increases. Such a 

 change is expressed in a model constructed in the manner previously 

 described by motion along a line joining the origin to a, away from o. To 

 ascertain what salt will separate next, the point at which this line ulti- 

 mately cuts the upper surface of the model must be determined. When 

 this is established with the aid of the model, it is found to lie in the 

 Magnesium sulphate (MgS0.,.7H. O) field. Hence it follows that further 

 concentration ultimately causes the separation of Magnesium sulphate 

 together with Sodium chloride, and the course followed on evaporation 

 will be across the Magnesium sulphate field, away from the hypothetical 

 point representing the solution saturated only with Magnesium sulphate 

 and Sodium chloride. This point must be on the Magnesium sulphate 

 axis as well as on the line kj (representing the change in composition of 

 a solution saturated with Magnesium sulphate and Sodium chloride as the 

 amount of Magnesium chloride varies), and will obviously fall at their 

 point of intersection, J^ Supposing the Magne.sium sulphate field to 

 have been cut at a point /S, the path followed on concentrating the solu- 

 tion will be along j'/3 produced, until the next field is entered. In a 

 similar manner, the subsequent course is traceable until the point w is 

 reached. As a matter of fact, some uncertainty exists as to the exact 

 course of crystallisation, as the investigation of Leonite, Kainite, and 

 Kiaserite is not yet complete. 



The order in which the salts are deposited is probably as follows : — 



(1) NaCl; (2) NaCl and MgS04.7H20 ; (3) NaCl and Leonite; 

 (4) NaCl, Leonite, and KCl, or NaCl and Kainite ; (-5) NaCl, Kieserite, 

 and Carnallite ; (6) NaCl, Kieserite, Carnallite, MgCL.6H.,0, the solution 

 then drying up without further change. 



Not only does the succession thus indicated agree with that actually 

 found experimentally on evaporating sea water at 25°, but also very fairly 

 with the geological succession as observed at Stassfurt. Thus the lowest 

 deposits of rock salt represent stage 1, the overlying Kieserite and 

 Kainite beds stages 2, 3, and 4, and the uppermost Carnallite region 

 stages 5 and 6. 



But although it is clear fropa the general agreement of the results 



