[THIS paper reviews the fundamental laws underlying atomic weight determinations. A sketch is given of the work of Dalton, Berzelius and Stas upon these constants of nature. A description is given of the schools of chemists which have carried out atomic weight determinations at the University of Pennsylvania under E. F. Smith, and at Harvard University under T. W. Richards.] At the beginning of the nineteenth century, John Dalton1 applied the ancient Greek philosophical theory of atoms—the theory of Leucippus and Democritos— as an explanation of his own laws of definite and multiple proportions, and thus founded the chemical atomic theory. Dalton then determined the “relative” or atomic weights of the elements; in his “New System of Chemical Philosophy,” Dalton took hydrogen as the unit with which he compared the other elements. He did not distinguish between atoms and molecules, hence his table contains atomic weights for water, ammonia and other compounds. Moreover Dalton had no means for ascertaining the number of atoms in a molecule, so he assumed that the atom (our molecule) of water consists of one atom of hydrogen and one atom of oxygen, whence the atomic weight of oxygen is given as 7, less than one-half of our value 16. The study of atomic weights was aided greatly by the discovery of certain laws. In 1808 Gay Lussac and Humboldt deduced the law of combination by volume of gases. “When two or more gaseous substances combine to form a gaseous compound, the volumes of the individual constituents as well as their sum bear a simple relation to the volume of the compound,”2 provided that the gases be under the same conditions of temperature and pressure. This law and the laws of Boyle and Charles have been used by Avogadro and Ampere,3 and, later on, by Canizzaro as the basis of the molecular theory. In 1819 Dulong and Petit announced the law of specific heats; the product of the specific heat of an element times its atomic weight always gives approximately the same constant 6.25, i. e., all the elements have approximately the same atomic heat. In 1831 Neumann and Regnault discovered that the molecular heat of a compound is a multiple of the atomic heat directly proportional to the number of atoms in the molecule, e. g., lead chloride PbCI, contains 3 atoms in its molecule, and has a molecular heat of 18.45, about three times 6.25. M'itscherlich deduced the law of isomorphism in 1819. “The same number of atoms combined in the same manner produce the same crystalline form, the latter being independent of the chemical nature of the atoms, and determined solely by their number and arrangement.”* However barium permanganate and sodium sulphate are isomorphous in the sense that they possess the same crystalline form, yet they differ from each other in respect to the number of atoms in the molecule. Kopp° has found that compounds which have the same number of atoms, arranged in the same manner in the molecule and possessing the same crystalline form—isomorphous compounds according to Mit-scherlich—are characterized by the property of forming overgrowths, e. g., a crystal of one alum will grow in the solution of another alum. Within recent years, the Periodic Law of Lothar Meyer and Mendelejeff” has been helpful in determining the correct values of atomic weights. This law states that the properties of the elements are periodic functions of their atomic weights. Berzelius devoted his life to atomic weight work; he made use of the laws of isomorphism, specific heats, and combination of gases by volume, and carried out analyses and syntheses without number; his work on atomic weights extended to many elements, including metals, non-metals and rare earths. As a standard atomic weight, Berzelius took oxygen equal to 100. His first table of atomic weights appeared in 1818; a revised table, which took into consideration the newly discovered laws, was issued in 1826. Many of the determinations of Berzelius, calculated to oxygen equals 16, compare favorably with the results of recent investigators. The next great investigator to devote his life to * Journal of the Franklm Institute. 1 Ostwald's Klassiker der exakten Wissenschaften Nr. 3, Die Grundlagen der Atomtheorie, Abhandlungen von J. Dalton, 1-20. 2 Remsen: Principles of Theoretical Chemistry, Fifth Edition. 34. :i Ostwald's Klassiker der exakten Wissenschaften Nr. 8, Die Grundlagen der Molekulartheorie, Abhandlungen von Avogadro nnd Ampere. 4 Wnrtz: Elements of Modern Chemistry. Fifth American Edition by Greene pnd Keller, 47. 5 Kopp: Periehte der deutschen chemisehen Gnsellsenaft, 1879. xii. 909. “Ostwald's Klassiker der exakten Wissenschaften Nr. 68, Das natiirliehe system der chemisehen Elemente, Abhandlungen von LothM Meyer trod D, Mendelejeff, atomic weight research was Jean Servais Stas, whose first work on the atomic weight of carbon appeared in 1841. His work on silver, sodium, potassium, lithium, lead, chlorine, bromine, iodine, sulphur, nitrogen and oxygen, is classic. Stas was exceedingly careful in all his work, yet was guilty of errors in manipulation; for instance, he would drop dry sodium chloride into silver nitrate solution, and yet expect to obtain a precipitate of pure silver chloride free from occluded or included sodium chloride. In his' earlier work, he neglected the solubility of silver chloride in water. Jn his investigations, Stas referred the halogens to silver. The task was undertaken to test the truth of Prout's hypothesis that the atomic weights of all the elements are simple multiples of hydrogen equal to one. Stas decided that this hypothesis is without foundation. Stas died in 1890; the work on atomic weights has been continued in Europe by such men as Guye and Gutbier and in America by Keiser, J. P. Cooke, Mallet, E. Fc Smith, and T. W. Richards. During recent years a conflict has waged among chemists concerning the standards for atomic weights. A few chemists led by Lothar Meyer have taken hydrogen equal to 1. The greater number of chemists have used oxygen equal to 16. Few of the elements form compounds with hydrogen and the ratio of an element A to hydrogen can usually be determined only by multiplying the ratio of A to a second element B, which most frequently is oxygen, by the ratio of B to hydrogen. Now the ratio of oxygen to hydrogen has been determined by various investigators and cannot be regarded as a fixed value. All the elements save fluorine and the members of the argon-helium group form oxides and hence allow a direct comparison between oxygen and the elements. Moreover if the ratio of an element to some element other than oxygen (say to silver or chlorine) be determined, the ratio of this second element to oxygen has been definitely determined. Oxygen equal to 16.00 has been accepted by the International Committee on Atomic Weights as the standard, and the column referred to hydrogen equal to 1 has been dropped from their reports since 1906.7 However the United States Pharmacopoeia still retains hydrogen equal to 1,000 as the standard.8 t A new method for the determination of atomic weights was introduced by Smith and Maas” in a paper “Uber das Atomgewicht von Molybdan.” They heated a known weight of anhydrous sodium molybdate in a stream of hydrogen chloride gas, weighed the residual sodium chloride and calculated the atomic weight of molybdenum. This work has been continued by E. F. Smith and his pupils at the University of Pennsylvania. Hibbs1” passed hydrogen chloride gas over heated potassium nitrate, weighed the residual potassium chloride and obtained the atomic weight of nitrogen. Sodium nitrate was treated in the same way for the same purpose. Sodium pyroarseniate, treated in this way, gave a value for the atomic weight of arsenic. Leriher11 treated silver selenite with hydrogen chloride gas, weighed the silver chloride formed, reduced the latter in hydrogen, weighed the metallic silver and thus obtained two values for the atomic weight of selenium. Ebaugh12 treated silver orthoarseniate with hydrogen chloride gas, weighed the silver chloride and then re.duced the latter to metallic silver, which was weighed. This gave two values for the atomic weight of arsenic. Two more values were obtained by passing hydrogen chloride gas and hydrogen bromide gas over lead orthoarseniate, and weighing the residual lead chloride and lead bromide. Friend and Smith” determined the atomic weight of antimony by treating potassium antimonyl tartrate with gaseous hydrogen chloride, and weighing the residual potassium chloride. Smith and his pupils have devoted much time to the determination of the atomic weight of tungsten. In 1899 Hardin” stated that “so far as known, there is no perfectly reliable method for the determination of this constant. The method of reduction and oxidation is probably more accurate than any of the other methods which have been employed. The results 7 Report of the International Committee on Atomic Weights, Journal of the American Chemical Society, 190(i, xxviii, 1; 1907, xxix, 107; 1908, xxx, 1; 1909, xxxi, 1- 1910, xxxii, 1. 8 Pharmacopoeia of the United States of America, Eighth Decennial Revision, 596. '?> Smith and Maas: Zeitschrift f iir anorganische Chemie, 1893, v. 280. 10 Hibbs: University of Pennsylvania Thesis 1896. 11 Lenher: University of Pennsylvania Thesis 1 898. 12 Ebaugh: University of Pennsylvania Thesis 1901. Journal of the American Chemical Society, 1902, xxiv, 489. 13 Friend and Smith: Journal of the American Chemical Society, 1901, xxiii, 502. 14 Hardin: Journal of the American Chemical Society, 1899, xxl, 1026. obtained by it vary about onevunit, and even more in exceptional cases.” The story of the determination of the atomic weight of this element is a tale of the gradual elimination of impurities. Smith and Pennington took precautions to remove molybdic acid from their tungstic acid.” Schneider had noticed a very slight residue when purified tungstic acid was dissolved in aqueous potassium hydroxide, but neglected it. Taylor10 found that purified tungstic acid, when treated with aqueous sodium carbonate, gave a white flocculent residue, which turned reddish-brown after standing in contact with the sodium carbonate solution for several hours. Exner and Smith17 prepared ammonium paratungstate from volframite of Lawrence County, South Dakota. In the mother liquor they isolated ammonium vanadico-phosphotungstate.18 Hardin':l had proved that tungstic acid obtained by ignition of ammonium tungstate contains nitrogen. Wyman20 had experienced great difficulty in obtaining complete solution of tungstic acid in ammonia. Exner and Smith now examined the insoluble residues of Wyman and found them to contain ammonia and chlorine. To eliminate these impurities from tungstic acid, dry ammonium tungstate was thrown into pure boiling nitric acid, to which a small quantity of pure hydrochloric acid was added from time to time, and was digested for several hours. The tungstic acid thus obtained was thoroughly washed, suspended in water and dissolved by means of ammonia gas, in order to obtain ammonium paratungstate. This salt was submitted to fractional crystallization, and a product was finally obtained which did not give a precipitate on standing over night with sodium carbonate solution. Therefore this salt was free from those impurities which the investigators had striven to eliminate. The pure ammonium paratungstate was treated in porcelain vessels with nitric acid and a little hydrochloric acid, evaporated to dryness and ignited in order to obtain tungstic acid, which was reduced to metallic tungsten by the method of Hardin.21 The tungsten metal was heated in chlorine gas to obtain tungsten hexachloride. A weighed quantity of the hexachloride was converted into the trioxide by means of water, aided by a very slight quantity of nitric acid. From the ratio of the trioxide and hexact sride, the atomic weight 184.04 was obtained for tungsten. Hardin has applied electro-analysis in the determination of the atomic weights of mercury, silver and cadmium.22 A paper on atomic weights would be incomplete without reference to the work of T. W. Richards of Harvard, and his pupils. Their work, which is found chiefly in the Proceedings of the American Academy of Arts and Sciences, covers the atomic weights of copper, barium, strontium, zinc, magnesium, nickel, cobalt, iron, uranium, calcium, csesium, potassium, sodium, iodine, chlorine, sulphur, arsenic, nitrogen, silver and chromium. Richards and the Harvard school have worked chiefly on the halides of the metals, changing them to silver halide which is weighed. In connection with this work., the nephelometer was invented to measure the opalescence of silver chloride solution. A perfected form of nephelometer has been described by Richards and Wells.2” Elaborate apparatus for fusing and bottling pure salts and metallic silver for atomic weight work had been devised by Richards and Parker.24 “A Revision of the Atomic Weights of Sodium aad Chlorine,”20 by Richards and Wells, shows in detail the manner in which Richards and his pupils attack a problem of this kind. We thus see that the masters of the science for over a century have devoted their best efforts to determining these constants of nature, With the introduction of new and more accurate methods of analysis and of new methods for the purification of reagents, the old rule of Berzelius, which calls for the use of pure reagents and a simple method of analysis, is found co give accurate and concordant results. 15 Pennington and Smith: Proceedings of the American Philosophical Society 1894. xxxiii, 332. 16 Taylor: University of Pennsylvania Thesis 1901. 17 Smith and Exner: Proceedings of the American Philosophical Society, 1904, xliii. 12.°>. 18 Smith and Exner: Journal of the American Chemical Society, 1902, xxiv, 573. 19 Hardin: Journal of the American Chemical Society, 1807, xix, 675. '.= ' 20 Wyman: University of Pennsylvania Thesis 1902. 21 Hardin: Journal of the American Chemical Society, 1897, xix, 667. 22 Hardin: University of Pennsylvania Thesis 1896. 23 Richards and Wells: American Chemical Journal, 1904, xxxi, 235; Wells: American Chemical Journal, 1906, xxxv, 99. 24 Richards and Parker: Proceedings of the American Academy of Arts and Sciences. 1896. xxxii, 59. 63. 25 Richards and Wells: A Revision of the Atomic Weights of Sodium and Chlorine, The Carnegie Institution of Washington, Publication No. 28, 1905.