104 



CHEMISTRY. (CHEMICAL ANALYSIS.) 



current year's table were those of calcium from 

 40 to 40.1, iron from 56 to 55.9, and tellurium 

 from 127 to 127. r,. 



The results of a redetermination of the atomic 

 weight of uranium have been published by Prof. 

 T. W. Richards and Mr. Merigold. Of previous 

 determinations, that of Zimmermann, who in 1886 

 found the value 239.59, was regarded by the 

 authors as most worthy of consideration. Zim- 

 mermann's method, which was based upon the 

 preparation of pure UO 2 and its conversion into 

 U,O 8 , seemed likely to give too high numbers, 

 because of the difficulty of obtaining the lower 

 oxid free from occluded gases, and of oxidizing it 

 completely. The authors chose for the basis of 

 their research the analysis of uranous bromid. 

 The analysis was effected by oxidizing uranous 

 bromid to uranic bromid by means of hy- 

 drogen per oxid and then precipitating the bro-. 

 min by means of silver nitrate. The results 

 showed, satisfactory concordance, and led to 

 a conclusion expressed by the authors as fol- 

 lows : " If O be taken as 16 and bromin as 

 79.955, the atomic weight of uranium appears to 

 be not far from 238.53." It is remarked that 

 although this number differs by more than a 

 unit from that given by Zimmermann, the per- 

 centage difference (0.45) is smaller than many 

 which have often been passed unheeded in the 

 case of elements of smaller atomic weight. It is, 

 however, a noteworthy difference, and the prob- 

 ability seems to the authors to be that Zimmer- 

 mann's number was too high. 



Through concentrating by fractional crystal- 

 lization the greater part of the radiferous barium 

 at her disposal, Mme. Curie succeeded in obtain- 

 ing about 1 decigram of perfectly pure radium 

 chlorid. This enabled her to determine the 

 atomic weight of radium. She found it to be 

 taking chlorin = 35.4 and silver =107.8 radi- 

 um = 225, with a probable uncertainty of not 

 more than 1 unit radium being considered a 

 bivalent element. Pure anhydrous radium chlo- 

 rid is described by Mme. Curie as being spon- 

 taneously luminous. From its chemical proper- 

 ties, radium appears to be an element of the 

 alkaline earthy series, and is the proper homo- 

 logue of barium. According to its atomic weight, 

 it should be placed in Mendeleef s table below 

 barium in the alkaline earthy series, and on the 

 line with thorium and uranium. 



It has been often observed that when calcu- 

 lated to the standard O= 16, many of the atomic 

 weights approach whole numbers in a much larger 

 proportion than they theoretically should accord- 

 ing to the theory of probabilities. This subject 

 is referred to by Arthur Marshall, in the Chemi- 

 ker Zeitung (July 19), in a paper in which at- 

 tention is called to some very remarkable rela- 

 tionships which appear to exist in many cases 

 between the atomic weights of allied elements. 

 Taking from the tables accepted by the German 

 Chemical Society the 18 values calculated to two 

 places of decimals, the chances against their 

 approaching whole numbers are as 4.120 to 1. 

 If the atomic weights are referred to H = 1, little 

 or no tendency appears to approach whole num- 

 bers. The most striking relationships appear 

 when certain of the atomic weights are referred 

 to entirely different standards. Thus, the atomic 

 weights of the halogen elements and silver are ex- 

 actly in the ratio Cl; Br; Ag; 1 = 90; 203; 274; 

 322. In the case of the alkali metals the propor- 

 tions are even simpler: L; NH 4 ; Na; K; Rb = 

 7; 18; 23; 39; 85. Then in the horizontal series, 

 V; Cr; Mn; Fe; Ni; Cu; Zn = 54; 55; 58; 59; 

 62; 67; 69. Other analogies are shown in Mr. 



Marshall's paper. It is not yet time, the author 

 observes, to work out relationships for all the 

 elements ;. for there is still too great uncertainty 

 about many of the atomic weights; but those- 

 referred to above appear to be thoroughly well 

 established. 



In experiments for determining the atomic 

 weight of arsenic, W. Clarence Brough proceeded 

 by four methods, namely, conversion of silver 

 arseniate with silver chlorid and reduction of 

 this to metallic silver; conversion of silver ar- 

 seniate into silver bromid; conversion of lead 

 arseniate into lead chlorid; and conversion of 

 lead arseniate into lead bromid. The mean of 

 results obtained in 26 experiments was 75.008, 

 with a probable error of 0.006. While this 

 result is not regarded as conclusive, it is con- 

 sidered as in many respects " certainly confirma- 

 tory." 



Prof. T. W. Richards has determined the 

 atomic weight of caesium at about 132,879, with a 

 range of from 132,873 to 132,882. 



The same subject of the most suitable standard 

 for the calculation of atomic weights is discussed 

 by Mr. Cecil Hollins, who shows that when oxy- 

 gen was taken instead of hydrogen, and the value 

 of its atomic weight was raised from 15.88 to 16, 

 the unit was lowered to 0.9925 of the atomic 

 weight of hydrogen. In this scale the atomic 

 weights of the elements exhibit a considerable 

 tendency to agree with Prout's whole-number 

 hypothesis 18 of them being integral against 6 

 when H = 1 is used as the standard, and the 

 sum of the deviations from unity being only 

 10.298 as against 15.861. So, by further lowering 

 the unit, he might expect that the atomic weight* 

 of the elements would still more closely approach 

 to integers. By dividing each of the atomic 

 weights into the nearest whole numbers, the au- 

 thor obtained factors ranging between 1.0101 for 

 helium and 0.9859 for silicon, the mean of which 

 was 0.99877. If this be taken as the unit, the 

 atomic weights of the elements show a further 

 tendency to become integral; and by treating^ 

 these in the same way the total weight is again 

 reduced, the mean factor now being = 0.99867. 

 The results upon 17 elements on whose atomic 

 weights authorities are agreed, namely, hydro- 

 gen, helium, lithium, sodium, potassium, oxygen,, 

 sulfur, selenium, nitrogen, fluorin, chlorin, car- 

 bon, chromium, beryllium, calcium, aluminum, 

 and boron, were encouraging. Thus, while 3 of 

 the 17 atomic weights are integral when the 

 standard is H = l, and 6 when it is O = 16, 10 

 are so when the factor 0.99877 is taken, and 15 

 when the factor is 0.99867. The total deviations 

 under these different units, taking them in the 

 same order, are 2.87, 1.448, 1.064, 0.729. The 

 author's paper is only preliminary, and his re- 

 search is to be continued. 



Chemical Analysis. The thirty-seventh, 

 volume of the proceedings of the American 

 Academy of Arts and Sciences contains the re- 

 sults of an investigation of the decomposition of 

 mercurous chlorid by dissolved chlorids. the 

 work of Messrs. T. W. Richards and E. H. Archi- 

 bald. This decomposition is shown to be consid- 

 erable if the chlorid solutions are fairly concen- 

 trated a point of considerable importance in 

 the analytical determination of mercury as mer- 

 curous chlorid. The action is not of a catalytical 

 nature, but a definite condition of equilibrium 

 is set up, the dissolved mercury being supposed 

 to exist in the form of a complex ion represented 

 by the formula HgCl 4 in the solution. 



A method for the extraction of bromin pro- 

 posed by Anson G. Betts is based on a reaction* 



