GEOPHYSICAL LABORATORY. 143 



(450) The crystal structure of silver oxide (Ag20). Ralph W. G. Wyckoff. Am. J. Sci., 



3, 184-188. 1922. 



Employing the method of powders, it is shown that silver oxide has the same 

 structure as that assigned to cuprous oxide. The length of the side of the 

 unit cube is determined to be 4.76^ a. u. 



(451) The reaction principle in petrogenesis. N. L. Bowen. J. Geol., 30, 177-198. 1922. 

 Petrogenic theory has passed beyond the stage where the conception of 



eutectics can longer be regarded as of any considerable service. Experimental 

 investigations and the study of the rocks themselves in the light of such 

 investigations have made it clear that the eutectic relation is unimportant, 

 but that another relation between liquid and crystal phases, here called the 

 reaction relation, is of fundamental significance. The ordinary solid solution 

 series, such as the plagioclases, may be regarded as a continuous reaction series, 

 because during crystallization each member is produced from an earlier 

 member by reaction with the liquid, the variation of composition being con- 

 tinuous. There are also discontinuous reaction series exhibiting related char- 

 acters but with discontinuous changes of composition. The series olivine- 

 pyroxene-amphibole-mica is a prominent example among the rock-forming 

 minerals. 



On the basis of these considerations the minerals making up the rocks of 

 an igneous sequence can be arranged as reaction series, and it is the existence 

 of such series that controls the crystallization and differentiation of the rocks 

 of the sequence. Even the graphic structure, often regarded as a eutectic 

 structure, is probably to be considered the result of reaction between the 

 phases. 



(452) The crystal structures of the alkali hahdes. I. Ralph W. G. Wyckoff. J. Wash. Acad. 



Sci., 11, 429-434. 1921. II. Eugen Posnjak and Ralph W. G. Wyckoff. 

 Ibid, 12, 248-251. 1922. 



These papers furnish an account of the study of the crystal structure 

 of some of the alkali halides and, with the exception of rubidium fluoride, 

 complete our knowledge of the structures of these compounds. The crystal- 

 structure data were obtained from the study of X-ray photographs of powders. 

 Except csesium chloride, bromide, and iodide, which have a body-centered 

 structure, all the other alkali halides have the sodium-chloride arrangement. 



(453) The melting of potash feldspar. G. W. Morey and N. L. Bowen. Am. J. Sci., 



4, 1-21. 1922. 



A pure synthetic orthoclase was prepared by crystallizing glass of the 

 composition KAlSisOg in a bomb with water- vapor. This material is partic- 

 ularly suitable for the determination of the melting-point of pure orthoclase 

 and was used for that purpose. The temperature ordinarily given as the 

 melting-point of orthoclase is about 1200° and has been determined on 

 natural crystals. When our artificial crystals were held at 1200° for a week, 

 they gave a product which had the appearance of a glass, megascopically , but 

 which, examined under the microscope, showed a structure described as a 

 very fine cross-lining. At higher temperatures this structure became more 

 distinct, taking the successive forms shown in figure 1, and finally becoming 

 typical leucite crystals. The point at about 1200° is therefore not the true 

 melting-point of orthoclase, but is the temperature at which it melts incon- 

 gruently, breaking up into liquid and leucite. The exact temperature of this 

 decomposition we have determined as somewhat lower than 1200°, namely, 

 about 1170°. The temperature of final disappearance of leucite is about 

 1530°, so that the interval of incongruent melting is remarkably large, viz, 

 360°. Three natural potash feldspars, microcline from North Carolina, sani- 



