Diopside — Forsterite — Silica. 241 



In the second place, even when all heating-curve breaks have 

 been obtained, it is not always easy to decide what the nature 

 of the changes which have given the breaks may be. It is 

 always possible that the materials on which a heating-curve is 

 run is not itself an equilibrium mixture and the temperatures 

 at which breaks occur are therefore, possibly, not the tempera- 

 tures of equilibrium changes. When the attempt is then made 

 to decide what change of phase has taken place at the tempera- 

 tures of the breaks, the conclusions may be erroneous. 



With the method of quenching, however, when supplemented 

 by microscopic examination of the product, there is no room 

 for doubt as to the nature of the phases present at equilibrium 

 and the temperatures of all changes can be definitely fixed, 

 even when there are two, or even more, at temperatures little 

 removed from each other, and whatever may be the relative 

 values of the heat-effects involved. The treatment can, more- 

 over, always be made such that both the material started with 

 and the final product represent equilibrium. 



The above statements are not intended as an argument for 

 the general superiority of the quenching method over the 

 method of heating curves. The existence of a promptly 

 reversible change, such as the inversion a- ft quartz, can be 

 easily missed when the quenching method is employed but 

 may be readily detected by the method of heating curves. No 

 fault can be found with the accuracy of the heating-curve break 

 itself, but there is some danger of misinterpreting the breaks 

 when the method is employed by itself in the study of equili- 

 brium in a complex system. 



Both the factors mentioned, viz.: the overlooking of certain 

 changes and the obtaining of breaks which do not represent 

 equilibrium changes , are involved in the results of the earlier 

 work on the system CaSiO,-MgSiO,. 



The material on which the heating-curves were made was 

 prepared in most cases by cooling a mixture from a tempera- 

 ture at which it was liquid to room temperature, crystallization 

 taking place freely. It has been shown, in the discussion of 

 the course of crystallization, that, when mixtures are allowed 

 to crystallize in this manner, the resultant product consists, for 

 the more calcic mixtures, 100-75 per cent diopside, of homo- 

 geneous pyroxene. By virtue of undercooling, somewhat more 

 magnesian mixtures, extending perhaps as far as 65 per cent 

 diopside, may also consist of homogeneous pyroxene if the 

 cooling has been at the proper rate. In the earlier work, 

 therefore, on finding that all mixtures extending from diopside 

 to Di 65 En 35 gave a homogeneous product it was concluded that 

 solid solution extended as far as Di 65 En 35 . 



The more magnesian preparations, however, crystallize to a 

 mixture of forsterite, pyroxene of variable composition and 



