GEOPHYSICAL LABORATORY. 137 



group for the alums. The manner of arrangement of the atoms within the 

 unit cell is outlined, though no attempt is made to locate those atoms having 

 variable parameters. It is pointed out that the 12 water molecules fall into 

 two sets of 6 each. The hydrogen atoms in the ammonium groups of the 

 ammonium alums present an interesting problem in the impossibility of 

 arranging them into a chemically plausible radical which will possess a 

 symmetry in keeping with that of the rest of the crystal. 



These spectrographic observations clearly show the practical inability 

 of unaided spectrometer measurements to furnish reUable data for the deter- 

 mination of crystal structures. 



(483) On the hypothesis of constant atomic radii. Ralph W. G. Wyckoff. Proc. Nat. 

 Acad. Sci., 9, 33-38. 1923. 



It is shown that the existing crystal structure data are not in agreement 

 with the hypothesis of constant atomic radii. They conform, however, to 

 the rule that in isomorphous crystals composed of only two kinds of atoms 

 the interatomic distances have additive properties which can be illustrated 

 through a summing up of "atomic radii." The data also show that for 

 compounds of different crj^stal structures in which the atomic environments 

 are different the interatomic distances likewise are changed. Where the 

 changes of environment are relatively small, this change in interatomic dis- 

 tance may be almost negligible; in other cases it amounts to several tenths of 

 an Angstrom unit. 



( 484) The compressibiHty of minerals and rocks at high pressures. Leason H. Adams and 

 Erskine D. WilUamson. J. Franklin Inst., 195, 475-592. 1923. 



This paper concludes the presentation of the results for the compressi- 

 bility of 40 solids, including 14 minerals and 10 rocks, at hydrostatic pressures 

 up to 12,000 megabars, corresponding to a depth of 40 kilometers below the 

 surface of the earth. The method used was that previously described and 

 shown to yield consistent results.^ According to this method, the specimen, 

 completely surrounded by a liquid, is subjected to pressure in a thick-walled 

 steel bomb, and the decrease in volume determined by the piston-displace- 

 ment. 



The compressibility of the minerals usually falls off slightly as the pressure 

 is increased. For the less compressible minerals, however, the change in 

 compressibility is so small as to escape detection by the present method,' the 

 precision of which corresponds to about O.OIXIO"^ per megabar, that is, 

 to 1 per cent of the total compressibility of the less compressible minerals. 

 The absolute accuracy, of course, is not so high. 



In connection with the compressibility of rocks a complication is intro- 

 duced by reason of their porosity, which even in the case of igneous rocks is 

 often enough to affect the compressibility. In order to determine the effect 

 of porosity at pressures within the range of experiment— 1,000 to 12,000 

 megabars — the porous rocks were covered with a thin jacket of pure tin, 

 which served to prevent the liquid from entering the pores and thus allowed 

 the closing-in of the pores to contribute to the decrease in volume of the 

 sample as a whole. It was found that at pressures above 2,000 megabars 

 porosity has very little effect on the compressibility; but a comparison with 

 the results of F. D. Adams and E. G. Coker shows that certain rocks, notably 

 the more porous ones, may have at low pressures an abnormally high com- 

 pressibility. 



The results show that, except for very low pressures, the compressibility 

 of a rock may be calculated directly from the known compressibility of the 



1 See Annual Report of the Director, 1919; in Year Book No. 18, Carnegie Inst. Wash., p. 160. 



