July 10, 1914] 



SCIENCE 



73 



matter. This is slow, tedious and furnishes much 

 opportunity for loss or contamination. The 

 Eeinseh test (1841) is rapid and simple, using the 

 minimum of added reagents. In its present form 

 its results are uncertain and seldom removed from 

 the region of doubt. It does not require the pre- 

 liminary removal of organic matter. The present 

 work is believed to have made the detection of 

 arsenic much more certain by eliminating entirely 

 the destruction of organic matter. The arsenic is 

 secured on copper strips as in the usual Eeinseh 

 procedure. These strips are introduced into a ' ' du- 

 plex" Marsh apparatus which has been devised in 

 this work. By the "duplex" feature (two hydro- 

 gen generators) no arsenic is lost while displacing 

 air from the generator containing the copper strips 

 which possibly contain arsenic. The use of two gen- 

 erators can be dispensed with by introducing an ex- 

 tra reagent to dissolve the arsenic from the copper. 

 Either procedure greatly reduces the time neces- 

 sary for the detection of arsenic with all the pre- 

 cision of Marsh's method. The new procedure 

 has already been tried with thorough satisfaction 

 in two poisoning cases where the presence of ar- 

 senic was proven finally to be present by the older 

 procedures. The new procedure consumes but an 

 hour or two where the old ones consumed usually 

 one or more days. 



The Decomposition Voltages of Salts in Liquid 



Ammonia. I. The Ammonium Salts: H. P. 



Cady and C. a. Nash. 

 Adsorption and Stabilization: J. C. Bluchee and 



E. F. Farnau. 



Further experimental facts are adduced to sub- 

 stantiate Bancroft's stabilization theory of dyeing. 

 These include examples of adsorption of dyestuffs 

 and inorganic compounds on colloidal hydrous 

 aluminium-, copper- and cobalt-oxides. 



The Ideal Diffusion Coefficient and a New Funda- 

 mental Law of Diffusion: G. McP. Smith. 



Further Observations on the Preparation of Selenio 

 Acid and Selen-ates: Philip L. Bltjmenthal. 



A Burette Calibrating Pipette: E. C. Foulk. 



Preparation of a Standard Magnesium Salt Solu- 

 tion: E. 0. Foulk and O. E. Sweeney. 



Concerning the Atomic Weights of Carbon and 

 Sulphur: Theodore W. Eichaeds and C. E. 

 Hoover. 

 In order to verify the silver-halogen standard of 



atomic weights by reference to a ratio entirely 



different, a precise quantitative comparison was 



made between sodium carbonate and silver. The 



purest sodium carbonate was fused in a stream of 

 carbon dioxide. It was then with all possible care 

 analyzed exactly with very pure hydrobromic acid, 

 and the amount of silver needed to precipitate the 

 bromine was determined as well as the weight of 

 silver bromide. For every 10.59950 grams of 

 sodium carbonate 21.5760 grams of silver were 

 needed. Hence carbon according to the Interna- 

 tional Standard of Atomic Weights became 12.005, 

 sodium being 22.995. If silver is taken as 107.871, 

 carbon became exactly 12. These results are com- 

 pletely concordant with the usually accepted values 

 concerning carbon and silver. The agreement is 

 striking and affords a much-needed and very wel- 

 come confirmation of the whole fabric of our 

 table of atomic weights. The investigation was 

 continued by converting weighed amounts of the 

 purest sodium carbonate into sodium sulphate. 

 The results were concordant among themselves, but 

 pointed to a somewhat smaller atomic weight of 

 sulphur than that usually recognized, namely, 

 32.055, if silver is taken as 107.88. This research 

 verifies in a striking way that published by one 

 of the authors, twenty-four years ago. The tech- 

 nique of this work will be of great value to any 

 one desiring to ma.ke exact acidimetric or alkali- 

 metric analyses. 



The Critical Point and the Significance of the 

 Quantity 6 in the Equation of van der Wools: 

 Theodore "W. Eichaeds. 



In this paper many results (especially those of 

 Kamerlingh-Onnes) were quoted to show that the 

 apparent bulk of the molecules of gases must be 

 supposed to change according to circumstances. 

 It was pointed out that the magnitude and direc- 

 tion of this change is such as would be expected 

 if the molecules and atoms are compressible, but, 

 if this is the case, the reasoning of van der Waals, 

 which infers that the bulk of the molecules is only 

 one quarter of b, is no longer sound, for this 

 reasoning assumes the ineompressibility of the 

 molecules. The present argument shows rather that 

 the actual bulk of the molecules when uncom- 

 pressed by collision or by the compressing effect of 

 affinity must be much larger than has been sup- 

 posed, indeed larger than the actual bulk of the 

 liquid under ordinary conditions, and perhaps that 

 assumed at the critical point. It was pointed out 

 that the continuity between the liquid and the 

 gaseous states may be supposed to exist, if at all, 

 only at the critical point, and that the application 

 of the equation of van der Waals to liquids is of 

 doubtful significance. The critical temperature is 

 defined by supposing that it is the point where 



