238 N. L. Bowen — The Ternary System : 



rial in any such experiment should itself be at equilibrium. If 

 it is desired, for example, to determine the temperature of begin- 

 ning of melting of a mixture of composition P, the heterogeneous 

 material prepared by rapid crystallization will be found to 

 begin to melt at the temperature of the eutectic between 

 diopside and tridymite. On the other hand, if the homogeneous 

 pyroxene P is the initial material, melting will not be observed 

 until the temperature of the point JS" (fig. 17) is reached. Now 

 the behavior of the heterogeneous material can be interpreted 

 after the true equilibrium has been studied throughout the 

 system but the converse is not true, that is, equilibrium in the 

 system can not be determined from the behavior of this hetero- 

 geneous material. 



It has already been stated that any mixture, say the mixture 

 P, can not in practice be cooled slowly enough to ensure perfect 

 equilibrium between the liquid and the solid phases, and a homo- 

 geneous pyroxene P cannot therefore be prepared by slow cool- 

 ing. Since it is necessary to prepare the homogeneous pyroxene 

 of composition P not only in order to obtain material to use in 

 studying equilibrium, but also to prove the very existence of this 

 pyroxene, some other method of making it must be resorted to. 

 The method is simple and consists in quenching a small charge 

 of composition P, from the liquid state, thereby obtaining a 

 glass, and then crystallizing the glass at a temperature below 

 that of the point N. Similarly the mixture M must be 

 crystallized from glass at a temperature below N", if a mixture 

 to be used in studying equilibrium is desired. 



It was in this manner that all the pyroxenes between diopside 

 and clino-enstatite were prepared and the fact of complete 

 solid solution between the two compounds, already indicated 

 by the shape of the fusion surfaces, was thereby demonstrated 

 beyond doubt. The heterogeneous preparations prepared by 

 crystallization during cooling from the liquid state create an 

 entirely false impression as to the extent of solid solution, when 

 examined under the microscope, unless the manner of crystal- 

 lizing already described is previously suspected. The prepar- 

 ation contains both pyroxene close to clino-enstatite and 

 pyroxene close to diopside and the natural conclusion is that 

 there is only limited solid solution close to each compound. 



Confirmation of the above statements concerning the mode 

 of crystallization of the cooling liquid comes from the work of 

 Zinke on the diopsicle-clino-enstatite series." The mixtures 

 were crystallized by cooling their melts, and Zinke noted, 

 besides diopside and clino-enstatite, pyroxene of small (and 

 variable) optic axial angle. The small optic axial angle is, 



*G. Zinke, Neues Jahrb. Min., 1911, II, p. 13& 



