Iebruaby 1, 1918] 



SCIENCE 



123 



The Equilihrium of Tortugas Sea Water 

 with Calcite and Aragonite: J. F. McClen- 

 don, Department of Physiology, University of 

 Minnesota and Tortugas Laboratory, Carnegie 

 Institution of Washington. The surface 

 water of the sea is the supersaturated solution 

 of CaCO, and it is only necessary to introduce 

 calcite crystals in order to cause precipitation 

 of this substance. 



An Oenothera-Lihe Case in Drosophila: 

 Herman J. Muller, The Rice Institute, Hous- 

 ton. Report of an extended series of experi- 

 ments showing that it will not do to accept 

 evidence apparently in favor of factor incon- 

 stancy without the support of highly rigorous 

 factorial analysis. 



Is Death from High Temperature due to the 

 Accumulation of Acid in the Tissues? Al- 

 fred Goldsborough Mayer, Department of Ma- 

 rine Biology, Carnegie Institution of Wash- 

 ington. Death is probably due rather to the 

 formation of acid than to coagulation of pro- 

 teid substances. 



National Research Council: Meetings of 

 the Executive Committee. 



Edwin Bidwtell Wilson 

 Mass. Inst, of Technology 



SPECIAL ARTICLES 



THE DETERMINATION OF ATOMIC WEIGHTS 

 BY MEANS OF X-RAYS 



It does not seem to be generally realized 

 that the recent developments in the study of 

 crystal structure by the use of X-rays afford a 

 method of determining atomic weights which 

 may be of considerable value. 



From the spectra obtained by expiosing crys- 

 tals of two different substances to X-rays of 

 the same wave-length, the ratio of the dis- 

 tances between adjacent layers of atoms in the 

 two substances can be easily determined, as is 

 well known. If the relative distances are de- 

 termined in the direction of each crystal axis, 

 these results, together with the inclination of 

 the axes to each other in each crystal, enable 

 one to calculate the ratio of the volumes of 

 the elementary parallelepipeds of each crystal. 

 From this ratio and the ratio of the densities 

 we can easily calculate the ratio of the molec- 



ular weights. From the ratios of molecular 

 weights, atomic weights can be calculated in 

 the usual manner. 



This method requires the preparation of 

 elements or compounds in a state of purity; 

 the production of crystals of practically per- 

 fect internal structure, though not necessarily 

 with perfect faces, or of large size; and the 

 measurement of densities and spectral angles 

 and, except when the crystal axes are mutually 

 perpendicular, the measurement of the angles 

 between axes. All of these measurements can 

 be made with considerable accuracy. 



Most of the measurements of the angles of 

 X-ray spectra that have been made hitherto 

 have not been highly accurate, for such meas- 

 urements have been used principally for the 

 determination of the relative positions of 

 atoms in crystals, and for this purpose great 

 accuracy is not required. It appears, how- 

 ever, that sufficient accuracy might be ob- 

 tained to permit the determination of atomic 

 weights with greater accuracy than that of 

 most of the chemical determinations, and 

 perhaps such accuracy has been obtained with 

 recent spectrometric apparatus. 



We are accustomed to think of the density 

 of a substance as being a rather variable 

 quality, but very few density measurements 

 have been made upon perfectly pure material 

 in the form of crystals of perfect internal 

 structure. Very many materials, including 

 metals, are handled commonly in the form of 

 masses composed of a great nimiber of small 

 crystals, which, even if they consist of pure 

 material, are likely to be very much distorted, 

 and at the surface between two crystals there 

 must be a layer of atoms many atoms deep 

 which are not located exactly according to 

 the space lattice of either crystal. There ap- 

 pears to be no reason why the density of a 

 flawless crystal of pure material should not 

 be quite definite, except that the surface 

 forces might cause a difference in density near 

 the surface, as the result of which the mean 

 density might depend somewhat upon the size 

 of the crystal. 



After suitable apparatus had once been set 

 up this method should permit the determina- 



