The purity of each of these combinations was carefully determined by photography of the visible and ultraviolet regions of the absorption spectrum with a Féry quartz spectrograph. The solutions were prepared by dissolving 2 grm. portions of the ignited oxides in a slight excess of nitric acid and diluting the solution to a volume of about IO CC. The absorbing layer was 10 cm. long. The less soluble fractions contained small amounts of samarium and the more soluble ones praseodymium. In order to determine the proportions of these impurities, to a solution of a 2 grm. portion of the oxide of Fraction 2605-6, which was free from samarium although it contained praseodymium, were added known amounts of a standard solution of pure samarium nitrate, the absorption spectrum of the solution being photographed after each addition of samarium. (The samarium material was very kindly furnished by Prof. C. James, of New Hampshire College). By comparing photographs of the different fractions with those of the material containing known amounts of samarium, it was possible to estimate with considerable exactness the proportion of samarium (through the band A 401) in the former. The percentage of praseodymium (through the band A 444) in the more soluble fractions was Estimated in a similar fashion by comparing the photographs of the fractions analysed with those of a solution of Fraction 2590-1-2, to which known amounts of praseodymium were added. The results of these comparisons are given below :

Since cerium and lanthanum nitrates are both more soluble in concentrated nitric acid than praseodymium nitrate the quantities of thess impurities remaining must have been very small indeed.

Although the proportions of impurity are small, yet because the atomic weights of both samarium and praseodymium are considerably different from that of neodymium it is not surprising that the extreme crystal fraction yielded a result somewhat higher than the average and that the more soluble fractions apparently possessed an atomic weight slightly lower than the average. However, since the amounts of impurity were determined with some accuracy corrections could be applied to the final result.

Preparation of Materials.

The Preparation of Neodymium Chloride.-In order to change the nitrate to chloride essentially the same processes were employed as in the previous investigation. Neodymium oxalate was first precipitated by adding to the dilute solution of the nitrate a considerable excess of oxalic acid. The oxalate was collected upon a disc of filter-paper in a porcelain Gooch crucible, and after being dried was ignited to oxide in a platinum boat in an electrically heated porcelain tube, pains being taken to avoid heating the boat to a temperature at which platinum might vaporise into the contents (see Baxter and Chapin, Journ. Am. Chem. Soc., 1911, xxxiii., 16). The resulting oxide was dissolved in redistilled nitric acid in a quartz dish and after the solution had been diluted the oxalate was reprecipitated by means of a dilute solution of either recrystallised oxalic acid or recrystallised ammonium oxalate. The oxalate was washed, dried, and ignited as before; then it was dissolved in a quartz dish in hydrochloric acid which had been distilled through a quartz condenser, and the chloride was at least three times crystallised in quartz dishes by saturating the aqueous solution with hydrochloric acid gas at a low temperature. Centrifugal drainage of the crystals was always employed. The product was preserved in quartz in a desiccator containing fused sodium hydroxide.

Reagents.-Pure silver, water, and reagents were prepared exactly_as_recently described in the account of similar work by Baxter and Stewart on praseodymium chloride (Fourn. Am. Chem. Soc., 1915, xxxvii., 524). The Drying of Neodymium Chloride.-In the earlier work by Baxter and Chapin the attempt was made to prepare the salt for analysis by drying it carefully in a current of hydrochloric acid gas and eventually fusing the salt. In the first experiments the salt dried in this way invariably yielded a considerable amount of insoluble material, the exact nature of which was not discovered until later. Hence the expedient was adopted of drying the salt for analysis as carefully as possible below the fusing temperature and then determining the proportion of water retained. Ultimately it was found possible to obtain fused salt which would yield a perfectly clear solution, but the necessary information as to the proper treatment was obtained too late to be of service. It was partly because of the slight uncertainty involved in applying a correction for the residual water that the present investigation was undertaken. In the recent research by Baxter and Stewart upon praseodymium chloride a similar difficulty was met (Fourn. Am. Chem. Soc., 1915, xxxvii., 527), and in the latter research it was shown that the insoluble material is the oxychloride, and that its formation can be wholly or almost completely avoided by drying the salt as completely as possible previous to fusion and then fusing the salt as rapidly as possible. No matter how carefully the preliminary drying is carried out, and no matter what precautions are taken in the drying of the hydrochloric acid gas in which the salt is fused, prolonged fusion invariably yields a very considerable proportion of insoluble matter, probably owing to a small amount of air contained in the hydrochloric acid gas. In these respects the neodymium chloride resembles the praseodymium salt exactly. A considerable amount of the insoluble neodymium compound was prepared by fusing some of the carefully dried chloride for an hour. The insoluble residue was collected and weighed and its content of neodymium and chloride was determined. The results of these experiments indicate conclusively that the insoluble matter is the oxychloride. The following table contains the results of these experiments :

Weight of insoluble material

AgCl found

AgCl calculated from Ndoci Nd2O3 found

Nd203 calculated from NdOCI

By following the same procedure in the drying of the neodymium chloride that was used in the drying of the praseodymium chloride we were able to prepare salt which yielded immediately a clear solution. The details of this procedure are as follows:-A platinum boat containing the powdered crystals of the hydrated salt was placed in a quartz tube forming part of a Richards "bottling apparatus." The bottling apparatus, which contained the weighing bottle in which the boat had initially been weighed, was connected with an apparatus for delivering dry hydrochloric acid gas, nitrogen, and air. This apparatus is described in the paper on praseodymium chloride (Ibid., 1915, xxxvii., 526). The boat was then gradually heated to a temperature slightly above 100° but consider ably lower than the transition temperature of the salt, 124°, until nearly all of the first five molecules of water of crystallisation had been expelled by efflorescence. The temperature was then raised to 180° or thereabouts, where the sixth molecule of water evaporates, and finally the salt was heated to about 350° for several hours. aluminium block oven which had been used for producing uniform temperature up to this point was now replaced by an electrically heated sleeve, and the salt was brought to the fusing-point, 785°, as quickly as possible (Matignon, Comptes Rendus, 1901, cxxxiii., 289; 1905, cxl., 1340). Then it was allowed to cool rapidly, and after the acid gas had been displaced by nitrogen and finally by air the boat

The

and contents were transferred to the weighing bottle without exposure to moisture and weighed. In a few cases a trace of insoluble salt was visible when the chloride was dissolved in water, but the proportions in these cases, judging from earlier experience, were less than o'r mgrm. By allowing the solution to stand for a day or two the basic salt dissolved completely. In Analyses 1, 2, 3, 5, 11, 13, 15, 16, 17, 18, 19, 21, 27, 29, 31, and 32 the solution of the chloride was clear at the outset.

Since the basic salt seems to form through the action of the air in the hydrochloric acid upon the fused salt more readily at higher than at lower temperatures, in three experiments the salt, after the usual careful preliminary drying, was heated almost but not quite to the fusingpoint for some time. In this way all but negligible amounts of water must have been expelled. In fact these analyses (Nos. 4, 7, 9, 20, 23, and 25) show a slightly greater percentage of chlorine rather than smaller than the specimens which were actually fused. In these cases also the salt yielded a perfectly clear solution.

The Method of Analysis.

The method of analysis was like that previously used with neodymium and praseodymium and other chlorides. The salt was dissolved in water and the solution was diluted to a volume of 1000-1500 cc. in a glass-stoppered precipitating flask. Pure metallic silver equivalent to the chloride within a very few tenths of a mgrm. was weighed out, dissolved in nitric acid, and diluted to about the same volume. (The silver used had already been tested in the The silver solution was then praseodymium work). added to the chloride solution in small portions with frequent agitation. After standing at room temperature for some time the analysis was cooled to o'o° in order to reduce the solubility of silver chloride (Richards and Willard, Journ. Am. Chem. Soc., 1910, xxxii., 32), and the clear solution was tested in a nephelometer for excess of chloride or silver. The estimated deficiency of either was added in the form of hundredth normal solution, and the solution was again thoroughly shaken, allowed to clarify and tested as before, and the process was repeated until exactly equivalent quantities of silver and chloride had been used.

In the analyses by Mr. Whitcomb, after the end-point of the comparison had been reached an excess of o'05 grm. of silver nitrate was added for each litre of solution, and the analyses were allowed to stand some time longer at o'o. Then the silver chloride was washed several times with ice-cold silver nitrate solution containing o'05 grm. per litre and many times with ice-cold water before being collected on a weighed platinum-sponge Gooch crucible. The chloride was dried in an electrically heated air-bath at 190° for at least eighteen hours and weighed. Residual moisture was determined by the loss in weight when the main bulk of the precipitate was fused in a porcelain crucible. The weight of silver chloride dissolved in one litre of the filtrate and silver nitrate washings was assumed to be 0.000004 grm. per litre (calculated from the solubility product of silver chloride at o'0° as found by Kohlrausch, 9x10-12-Zeit. Phys. Chem., 1908, lxiv., 167). Chloride dissolved in the aqueous washings together with that obtained from the precipitating flask was estimated by nephelometric comparison with standards.

In the analyses by Mr. Stewart, after a similar excess of silver nitrate had been added, the system was allowed to stand at room temperature for several days before filtration. In computing the correction for silver chloride dissolved in the filtrate and silver nitrate washings, which were not chilled, the assumption is made that the solutions were saturated at 25°, since the experiments were carried out during the summer months. Using Kohlrausch's determination of the solubility product of silver chloride at 25°, 17 X 10-10, the solubility in o'0003 normal silver nitrate is o'00008 grm. The aqueous washings were chilled to diminish the solubility of the precipitate and were analysed nephelometrically.

The main mass of silver chloride was dried at 155° and the moisture retained was found from the loss on fusion.

In Mr. Whitcomb's experiments the portions of the original solution removed for nephelometric comparisons were returned and a correction applied for the silver chloride thus introduced. In Mr. Stewart's experiments the test portions were rejected and a correction of 0'00005 grm. was added for each 100 cc. of solution removed.

All objects were weighed by substitution for similar counterpoises, a No. 10 Troemner balance being used for the purpose. Weights were standardised to hundredths of a mgrm. by the Richards substitution method (Journ. Am. Chem. Soc., 1900, xxii., 144). The following vacuum corrections were applied :

The concordance of the results with each fraction is

satisfactory. It is interesting to note that with the ununcorrected results a perceptible difference exists between the head and tail of the series, and that this difference largely disappears when corrections are applied for rare earth impurity. The uncorrected results with the purest fractions and the corrected result with all the fractions differ very little from Baxter and Chapin's earlier result, 144 275 (Ag=107.880).

In Table IV. the ratio of silver used to silver chloride obtained in the same experiment is given for all complete pairs of analyses. Since the average ratio is essentially identical with that found by Richards and Wells (Publ. Carnegie Inst., 1905, xxviii.; Journ. Am. Chem. Soc.,

1908, xxviii., 456), 0.752634, it can reasonably be concluded, that errors from occlusion by the silver chloride or from loss of silver chloride are absent. This is in accord with earlier experience of the same kind.

Summary.

Neodymium nitrate was purified by fractional crystallisation of the nitrate from concentrated nitric acid. Chloride, prepared from the final fractions of nitrate, was analysed by comparison with silver. The atomic weight of neodymium was thus found to be 144 261. The average of this result and that found earlier by Baxter and Chapin, 144 275, is 144.268. The rounded-off figure 144 27 (Ag=107.88) seems to represent fairly the final outcome of both researches.

We are greatly indebted to the Carnegie Institution of Washington for generous pecuniary assistance, as well as to Dr. H. S. Miner, of the Welsbach Light Company, for the neodymium material.-Journal of the American Chemical Society, xxxviii., No. 2.

THE CONTENT, METHOD, AND RESULTS

OF THE

HIGH SCHOOL COURSE IN CHEMISTRY.* By ALEXANDER SMITH, Columbia University, New York City (Concluded from p. 8).

The Method.

In striving for the desired results, success depends entirely on the methods used. We have time for a single illustration only. In live chemistry, is it necessary to pay attention to physics? Many texts suggest an answer in the negative. In their introductions they labour to distinguish between chemistry and physics, and leave the impression that, when the distinction has been formulated, the physics will be identified wherever found, and will be eliminated. Let us analyse the situation in practical chemistry. How is chemical change brought about? Some people say by chemical means. One recent author says that mere mixing does not produce chemical change, as if mixing solutions did not produce chemical actions in hundreds of cases, and as if mixing suitable substances with chlorine was not almost always sufficient to start the reactions. Rubbing salt in a mortar is a physical operation, but tapping fulminating mercury and tickling iodide of nitrogen with a feather are apparently chemical means! Boiling water is physical, but boiling liquid nitrogen trioxide is chemical. Precipitation is due to physical im miscibility, and is physical. Heating increases the speed of motion of the molecules and is a physical, and, indeed, literally, a mechanical means of producing chemical change. Even electrolysis consists in shooting a stream of electrons at the solution, and is now mechanical. All the means we use are physical.

Again, heating a platinum wire is a physical phenomenon. When we decompose mercuric oxide by heating, what sort of a phenomenon is it? A phenomenon is something seen, or more generally something perceived by the senses. We collect the gas given off, and we see the mercury accumulating on the wall of the tube. Certainly, the phenomena are physical here, also. We even reason physically, for the mercury cannot come from nothing, so we conclude that it is condensing from an invisible vapour. During the process the mass of the oxide becomes less and less. All that we can perceive by the senses is physical.

A certain group of these physical phenomena is not only physical but also chemical. The sub-group is easily defined when the time comes. If one substance

* Address before the Science Section of the New York State Teachers' Association. From School Science and Mathematics, 1916, xvi.

gives two or two give one, or two give two, all the substances concerned being different, then the phenomenon, in addition to being a physical is also a chemical change; or, if the composition of a material is such that the elements are present in proportions that can be represented by multiples of the atomic weights by whole numbers, then it is a chemical compound (e.g., a hydrate); if the proportions are not of this nature (e.g., a solution), then it is a physical aggregate.

Thus we use physical means to produce chemical change because they are the only means available. We study the resulting physical phenomena to learn what the change was because they are the only phenomena; we reason in physical terms about the phenomenon, in order to understand and to classify it, because physical data are the only data available; finally we draw the conclusion that it was a chemical change, altering certain substances into certain other substances.

Many teachers of chemistry do not realise how far physics dominates the experimental work in chemistry. On distributing standard texts to the members of a class made up largely of teachers of chemistry, and asking them to report as to certain specific features of the text assigned to each, I have frequently instructed them to consider whether the physics required in practical chemistry was adequately represented. The almost invariable report was, "Yes; Boyle's and Charles' laws are in the book!" Suppose we were rigidly to exclude all physics, including physical properties and phenomena, from the course in chemistry, what would happen? We could not describe common salt, because the solid state, whiteness, and cubical crystalline form would be lacking.

It is evident that a very specific and ready knowledge of many parts of physics will be acquired in connection with a course in live chemistry. The physics need not be called physics, nor need it be forced, for it comes in quite naturally. What physical property of oxygen enables us to utilise the oxygen we draw into our lungs? Evidently its solubility in the moisture on the inner surface of the lungs. When we burn phosphorus in oxygen the pentoxide form an immense cloud of particles and nothing remains in the spoon, but when we burn iron wire in oxygen no such smoke is formed and the oxide remains in globules. Why this difference? In classes of freshmen in college, all of whom have admission credit in chemistry, I have never received a sensible answer to this question. The question is asked of course to show them that even from experiments they have nearly all seen before there may still be something of an obvious nature that they had never considered. The few who venture a reply say that the pentoxide must be lighter than air, an explanation disproved by the fact that the substance in a bottle on the table is in the bottom of the bottle.

Is this real chemistry? The problems of the chemical industries seldom concern the chemical reaction itself. The difficulties almost all arise because of the physical properties of the crude materials and of the products. In the contact process for sulphuric acid the difficulty did not lie in the reaction, which had long been known, but in the poisoning of the catalyst by foreign substances transported mechanically by the gases. Again, a novice would expect the sulphur trioxide to dissolve easily in water, but it was found that the excess of oxygen used in the process carried a large part of the oxide right through the water and out of the absorbing units. Investigation showed that 97 to 98 per cent sulphuric acid absorbed it much more completely, and so the acid itself is employed for the purpose.

Is this good education? Well, it teaches closer observation whenever the real facts are wanted; it substitutes simple reason for memorising; it teaches the need of alertness and penalises superficiality; it calls attention to facts rather than words as the subject of study; it is not physical chemistry, which is a theoretical subject, but practical chemistry. It would seem, therefore, that we should give attention to the physical means used to produce chemical changes, that we should show how a