584 JV. L. Bowen — Melting Phenomena of the 



The limits of error of the- figures given for the liquidus 

 (do 2°) are very accurately determined by ascertaining the 

 temperature of the final disappearance of crystals. The check 

 on these values by the reverse method of finding the tempera- 

 ture at which crystals first form in the glass was perfect, within 

 approximately the same limits of error, in all the artificial mix- 

 tures except AbgAn,. It is possible to hold glass of this com- 

 position at a temperature a few degrees below the liquidus for 

 several hours without the formation of crystals. No check on 

 the liquidus point as determined by the reverse method was 

 therefore possible in this case. That the point is sufficiently 

 well established by the former method alone and with an accu- 

 racy practically equal to that in the more calcic members, is 

 apparent, however, from the following. Beginning with half- 

 crystalline material and holding it at 1263° for an hour and a 

 half, it was found that a few rare crystals still persisted, 

 whereas the same half-crystalline material held at 1268° (only 

 5° higher) for only half an hour gave a clear glass. It is obvi- 

 ous from the latter that the persistence of crystals at 1263° for 

 an hour and a half could not have been due to insufficient 

 time of exposure. In spite of the readiness with which under- 

 cooling occurs, persistent superheating does not occur with 

 this method of procedure. The importance of fixing the point 

 on the liquidus corresponding to Ab 8 An, is considerable 

 because it determines the position and slope of the liquidus at 

 a composition fairly close to albite itself. 



Use of Natural Feldspars. 



It was found impossible to completely crystallize pure arti- 

 ficial mixtures richer in albite than A^An^ The partially 

 crystallized preparations as rich as Ab.Aiij, which were easily 

 obtained, served perfectly well for fixing the liquidus, but, of 

 course, were of no service in locating the solidus. A very 

 pure natural oligoclase from Bakersville, North Carolina, of 

 composition Ab 77 5 An 225 was therefore used in an effort to fix 

 the corresponding point on the solidus. A small portion of 

 this material, carefully selected under the microscope, was 

 finely ground and subjected to successive quenchings as before. 

 The temperature of the beginning of melting was found to be 

 1158°±5°. (See Table II.) 



For the same reason pure, natural material was also used to 

 determine the melting point of albite. This Amelia County 

 albite has been carefully analyzed* and found to contain about 

 2 per cent of anorthite and 1 per cent of orthoclase. What 

 was really determined in this case also was the temperature of 



* Day and Allen, 1. c, p. 48. 



