56. BUREAU OF AMERICAN ETHNOLOGY [bull. 52 



ments and methods which have been adopted in the study of any 

 particular problem, but in the present instance the methods applied 

 were not the usual methods and merit, therefore, a brief word of 

 description. In the petrographic-microscopic examinations the 

 methods employed were those specially adapted to the study of 

 fine-grained preparations. They have recently been described in 

 a paper by one of the writers ^ and need not be considered in detail 

 here. The specimens were examined both in the thin section and 

 in powder form. In the thin section the texture of a specimen 

 can be studied to best advantage, while in the powder the optical 

 properties of its several mineral components can be ascertained 

 most readily and accurately. 



The chemical analyses cited below were made by Mi\ J. G. Fair- 

 child, of the United States Geological Survey, and the methods used 

 in his determinations were the standard methods of the Geological 

 Survey laboratory. 



In the thermal work the standard thermo-electric apparatus and 

 methods of the Geophysical Laboratory ^ were adopted. The mate- 

 rial to be heated was placed in a platinum crucible and suspended 

 therein in an electric resistance furnace where it could be brought 

 to any desired temperature and held there for any length of time. 

 The temperature of the furnace was determined by means of a 

 platinum-rhodium thermoelement inserted into the furnace together 

 with the charge. This method for ascertaining high temperatures 

 has been thoroughl}'- tested in the Geophysical Laboratory and 

 furnishes temperature readings whose probable eri^r is not more than 

 2° C. Throughout this paper the temperatures are expressed in 

 degrees centigrade. 



In the present series of experiments the conditions were varied 

 considerably, but not to extremes which departed widely from 

 natural conditions. One point of difference, however, should be 

 noted. The amount of material used in an experiment was small, 

 only a few grams at most, while in nature kilograms and usually 

 tons are used and the resulting products are on a corresponding 

 scale. The smaU charge in the platinum crucible oxidizes readily 

 and the iron is converted almost entirely into ferric iron (hematite), 

 while in the natural products a lower state of oxidation (magnetite) 

 is present. The attempt was made to overcome this difficulty by 

 mixing a reducing agent, generally graphite, with the charge before 

 treatment, but this proved insufficient for the purpose. In another 



> Fred. E. Wright, The Methods of Petrographic-microscopic Research. Their Relative Accuracy and 

 Range of Application. Publication No. 158, Carnegie Inst, of Washington, 19n. 



2 A. L. Day, E. T. Allen, and J. P. Iddings, Publication No. 31. Carnegie Inst, of Washington; A. L. 

 Day, E. S. Shepherd, and F. E. Wright, in A7ner. Jour. Sci., 4th ser., xxn, 265-302, 1906; W. P. White.in 

 Phys. Rev., xxv, 334-352, 1907, and in Amer. Jour. Sci., 4th ser., xxvm, 453-489, 1909. 



