November 18, 1921] 



SCIENCE 



497 



volumes. A present-day book for beginners 

 should contain little more than the basal prin- 

 ciples and the more striking and interesting 

 facts and illustrations. For advanced work 

 special treatises on separate branches of the 

 science are desirable. Already the economic or 

 industrial geology has been divorced from gen- 

 eral study. The same is true for earth forms 

 or physiography; and partially for paleon- 

 tology. Further differentiation may cover 

 dynamics and geophysics; surficial processes; 

 sedimentation and structure ; meteorologic and 

 glacial geology; with perhaps later division of 

 the historical. 



Grabau is now in China, as professor of 

 paleontology in the University of Peking, 

 and Paleontologist to the Chinese Geological 

 Survey, and we may anticipate further en- 

 richment of geologic literature from his proli- 

 fic and facile pen. 



H. L. Fairchild 



SPECIAL ARTICLES 



A PRECISION DETERMINATION OF THE 



DIMENSIONS OF THE UNIT CRYSTAL 



OF ROCK SALT 



All measurements of X-ray wave-lengths 

 and of crystal structures depend upon the solu- 

 tion of the atomic marshalling of some crystal 

 and a calculation of the dimensions of the 

 fundamental unit of that crystal in terms of its 

 mass and density. The crystal most used in 

 this connection is rock salt (NaCl). It is the 

 purpose of this note to give the side of the unit 

 cube of NaCl in terms of the most accurate 

 data available. 



The NaCl crystal was early shown by Bragg 

 to be a cube, alternate corners of which are 

 occupied by Na, the remaining corners being 

 occupied by CI. Since one half of one Na and 

 one half of one CI are each associated with one 

 unit cube, the mass of the unit must be 



l/2[ilNa4-4Cl]m, 



where Anh is the atomic weight of Na, 

 Aci is the atomic weight of 01, 

 m is the mass in grams associated with 

 one unit of atomic weight. 

 The 1919 International Table of Atomic 

 Weights gives 



^Na = 23.00 

 Aci = 35.46 



If these values should be wrong by .01 the 

 error would be less than .05 per cent, in each 

 case. 



j» is most easily found as e/F where e is the 

 charge on the electron, F is the Faraday 

 constant in electrolysis. Millikan^ gives e as 

 4.774 X 10-" Abs. E. S. units of charge with a 

 maximum error of .1 per cent. 



This gives e = log-i 19.20176 



= 1.591X10-" absolute 

 coulombs. 



Vinal and Bates,- of the Bureau of Stand- 

 ards give 



F (Iodine) =96,515 

 (Silver) =96,494 



Mean =; 96,505 iaternational coulombs. 



They have determined the absolute coulomb 

 as being .004 per cent, greater than the inter- 

 national coulomb, and recommend the value in 

 absolute coulombs, 



ii' = 96,500 



The maximum error is .01 per cent. 

 From the above 



■m = e/F — log-i 24.21723 



= 1.649 X 10-2* gms. 



The density of NaCl is given by Zehnder 

 (1886) as 2.188, by Retgers (1890) as 1.167, by 

 Krickmeyer (1896) as 2.174 and by Gossner 

 (1904) as 2.173. Gossner's work^ seems to have 

 been done with special care. He measured 

 eleven artificial crystals of NaCl, obtaining 

 densities ranging from 2.171 to 2.175. His 

 measurements on natural crystals gave 2.173. 

 Taking these results in connection with those 

 of Krickmeyer, we may assign to NaCl a den- 

 sity of 2.173 =t .002, thus giving a maximum 



1 B. A. Millikan, ' ' A new determination of E, N, 

 and related constants, ' ' Phil. Mag., 34, 1917. 



2 G. W. Vinal and S. J. Bates, ' ' Comparison of 

 the silver and iodine voltameters, and the determi- 

 nation of the value of the Faraday, ' ' Bull. Bureau 

 of Standards, 10, 425, 1914. 



3 B. Gossner, ' ' Untersuchung polymorpher Kor- 

 per, ' ' Zeit. f. Kryst., 38, 132, 1904 



