Minerals of the Com/position M~gSi0 3 . 417 



or 50 grams) is then quickly dropped through, the tube into the 

 crucible, after which temperature readings are made at half- 

 miuute intervals until the temperature of the crucible again 

 becomes practically constant. 



In a second experiment under the same conditions, the 

 product of this first change is introduced into the furnace as 

 before, and a second temperature curve determined under 

 exactly the same conditions. A comparison of the two curves 

 plainly indicates the direction of the heat change. If an evo- 

 lution of heat has taken place, the temperature rate is accel- 

 erated and the curve of the unstable body lies to the left and 

 above that of the stable form, as shown in curve I of fig. 11. 

 On the other hand, if there has been an absorption of heat, 

 the temperature rate is retarded and the curve of the unstable 

 form lies to the right and below that of the stable form. 



In order to test the method, substances were first tried in 

 which the direction of heat change was known, e. g., wollas- 

 tonite, which passes into pseudo-wollastonite with absorption 

 of heat ; wollastonite glass, which crystallizes with evolution 

 of heat ; and albite, which melts with absorption of heat, — 

 all very slow changes. Fig. 11 contains the results obtained 

 with wollastonite and wollastonite glass. The number of 

 curves in a single figure and their general similarity of form 

 have made it inadvisable to confuse the figures by attempting 

 to show the observed points on each curve. The observed 

 electromotive forces in microvolts are therefore tabulated 

 separately (Table YIII). The corresponding temperatures are 

 not important, as the point which we desire to establish depends 

 merely upon the relative displacement of that portion of the 

 curve in which a change of state may be expected to occur. 

 The curves II are heating curves for pseudo-wollastonite, 

 which is the stable form at the temperatures of the experi- 

 ment, and therefore suffers no change of state. Curve I, the 

 corresponding curve for wollastonite glass, lies to the left and 

 above the curves for pseudo-wollastonite, and plainly indicates 

 an evolution of heat, since all other thermal conditions are 

 identical in the two cases. The curves for wollastonite, III, 

 lag behind those for pseudo-wollastonite, and indicate heat 

 absorption. The results for albite and albite glass are shown 

 in fig. 12 (Table IX). It is interesting to compare these 

 results for albite with those obtained for orthoclase (which 

 resembles albite very closely in its thermal properties), in a 

 previous investigation in this laboratory by the Frankenheim 

 method (fig. 13).* 



* Arthur L. Day and E. T. Allen, "The Isomorphism and Thermal 

 Properties of the Feldspars," this Journal, series 4, xix, 98, 1905. 



Am. Jour. Sci.— Fourth Series, Vol. XXI, No. 131. — November, 1906. 

 29 



