FIG. 9.-ELECTRIC PLOUGHING ON THE DOUBLE WINDER SYSTEM IN PROGRESS. FIG. 8.-PORTABLE MOTOR WAGON, SHOWING CABLE DRUM. FIG. 10.-ELECTRIC PLOUGHING ON THE SINGLE WINDER SYSTEM, SHOWING ANCHORAGE WAGON. It has been shown by repeated measurements that they saleable corn resulting from electric drive amounts to from 1-1 per cent more than with the usual steam or other drive employed, due to the more uniform turning moment of the electric motor. It may also be added that there is no danger incurred by bringing the electric motor close to the fences or sheds, which is not the case with the steam locomotive, petrol engine, or gas engine, as the motor is fully enclosed, and in the larger sizes is carried with all its accessories in a small completely locked; wagon, which allows of easy transport and permits the motor to be placed as near as may be to fences, &c., and in any position in the sheds or yard. It can be placed in the barn itself, if necessary. In reply to a question when giving evidence before the Canadian Committee on Agriculture and Colonisation, a witness said, when referring to electric driving: "You can get a better quality of chaff, it absorbs the moisture from the grain when it is kept for a longer period before threshing; then again, your grain is a better product from leaving it longer in the head. In 1912, three-fourths of the grain in the country was heated because it was threshed too soon, and most of it was harvested wet." (Report of Select Standing Committee on Agriculture and Colonisation, Canada. App. No. 2, 1914). The cost of an equipment for electric threshing depends, of course, on the size of the threshing machine to be given, and on the accessories necessary. An approximate idea of the cost of the necessary motor wagon may be obtained from the figures already given; thus a motor wagon of 34 h.p. capacity, with accessories, should cost before the war approximately 2500 marks, and allowing 500 marks for the part cost of electrical conductors, &c., they would give a total first cost of 3000 marks exclusive of the cost of the transformer wagon if such were needed. The cost of this would, in any case, in part, have to be charged to other work. General. For operations on the farm other than plough. ing and threshing, portable motors of the type already described are generally employed. Stationary motors are also used where the work is of a more continuous nature. An illustration of a permanent motor driving a model dairy in Canada is given in Fig. 14, the electricity in this case being supplied from the mains of the Ontario HydroElectric Commission. The fellowing figures for power consumption on certain farming operations based on tests give some indication of the costs of performing these operations by electricity. Chaff Cutting. A motor of from 2-5 b.p. will cut in an hour up to 20 cwt. of chaff with a current consumption of O'I to 0.3 kw. hours per cwt. For purposes of comparison it may be said that it usually takes one man one hour to cut 1 cwt. of chaff, while two horses and a driver are capable of cutting 3 to 4 cwt. per hour. Turnip Cutting. The current consumption for this work varies from o'01 to 0.02 kw. hours per cwt. Oat Crushing. The current consumption for oat crushing is about 0.25 kw. hours per cwt. Milk Separating.-A motor of 1 h.p. can deal with an output of 2100 litres per hour with a current consumption for skimming, churning, and kneading the butter of 0.25 units per 100 litres. In forming an idea of the probable cost of current to be consumed by a motor of given size on farm work, it should be remembered that an agricultural load is a very fluctuating one, and for general purposes it is usual to assume that a motor is on full load only during one-third of the working time, and that during the remaining twothirds of the working time it is only on half load, and it is found in practice that working on this basis gives a fairly good idea of the probable consumption. In making a comparison between the working cost of driving by electric motor and by heat engines, in addition to the expenditure in coal, petrol, electric current, &c., allowance should be made for interest, depreciation, maintenance, lubrication, and attention, and if these factors are taken into account it will generally be found that with the smaller motors up to, say, 20 h.p. electrical driving is cheaper than any other form of driving with current at the usual rates obtainable in country districts, whereas for outputs of above 20 h.p. the running costs of other forms of power begin to be competitive if the cost of electricity is excessive, particularly if the machines are in continuous service. The low first cost of electrical plant and the discontinuous service required for motive power in farm work are, however, very much in favour of electrical driving. The following table gives approximate pre-war costs of the motor cradles, trucks, motor sledges, and motor wagons already described : Motor cradles up to 3 b.p., including 15 metres of flexible cable-without gear reductions .. Ditto, with gear reduction Motor trucks or sledges from 2 to 6 b.p., including 20 metres of flexible cable 240-550 marks. 350-700 marks. 900-950 marks. 1400-1500 marks. 2200-2400 mares. 2900 3550 marks. REPORT OF THE INTERNATIONAL COMMITTEE ON ATOMIC WEIGHTS FOR 1919-20. THE last regular report of this Committee, apart from an annual recommendation to continue the use of the table of atomic weights then presented, was published in 1916. The interruption in the series of reports was, of course, due to the world's war, which created difficulties of a serious kind among all international organisations. Cooperation with Germany became impossible, partly because of the difficulty of correspondence and partly because of the personal hostilities created by the conduct of the war. There was also an inevitable slackening of scientific activities, and this was well shown by the unusually small number of new researches in the field of atomic weights. Now that peace is in sight, it seems wise to resume the preparation of these reports, even though they may not be for some time quite so truly international as heretofore. The determination published since the preparation of our last report may now be summarised as follows: : Hydrogen.-A very thorough investigation by Burt and Edgar on the volumetric composition of water has given the volume ratio of hydrogen to oxygen as 2.00288: I. (Phil. Trans., 1916, 216 A, 393; this research was noted in the previous report for 1917; its review by Guye renders its repetition desirable here). From this value, taking the normal litre weights of oxygen and hydrogen as 1 42900 and 0.089873 grms. respectively, the atomic weight of hydrogen becomes 100772, or, rounded off, 10077. Guye (Fourn. Chim. Phys., 1917, xv., 208), from a discussion of Burt and Edgar's data, accepts this value as lying between the two extremes of 1'00767 and 1'00773. If, however, instead of trusting to the densities of the gases and their physical constants exclusively, we take into account the admirable researches of Morley, Noyes, and others, upon the synthesis and analysis of water, the most probable general mean for the atomic weight of hydrogen becomes 1'0078, which differs from the volumetric value by only 1/10000. (Computation by F. W. C). That is, the two distinct lines of attack upon the problem agree within the limits of actual uncertainty. For ordinary purposes the approximate value 1'008 is close enough. It must be remembered that the tables prepared by this committee are for the use, not so much for specialists, as for In this research sodium fluoride was compared not only with borax but also with the sulphate, and the 8 values found ranged from 19.002 to 19.008, in mean 19.005. The rounded-off value F 190 may be retained for all practical purposes. Working chemists in general; and too much refinement | affected by errors. The new value 10.900 should be will only lead to confusion. No determinations of these adopted as the most probable. or any other constants can be absolute and final. All are subject to errors which may be reduced nearly, but not quite, to insignificance, but never eliminated entirely. For example, in the determination of atomic weights from gaseous densities it is not possible to guarantee the absolute purity of the gases, or to avoid errors in weighing, in reductions to a vacuum, or in the values given to the physical constants that are used in the final computations. Some of these errors may be so small as to be negligible, and in the aggregate they may tend either to reinforce or to compensate one another but their extreme magnitude can be estimated with some approach to accuracy, and expressed by means of the usual sign. At present an accuracy to within 1/10000 is the best we can expect to obtain. (For an elaborate discussion of sources of error in atomic weight determinations, see Guye and his col. leagues-M. Germann, Moles, and Renard-in Journ. Chim. Phys., 1916, xiv., 25, 195, 204; 1917, xv., 60, 360, 405; 1918, xvi., 46). = 12'00. Carbon.-Two investigations on the atomic weight of carbon were reported from the Geneva laboratory in 1918. First, Stahrfoss (Fourn. Chim. Phys., 1918, xvi., 175) determined the density of acetylene, ethane, and ethylene. Acetylene proved to be unsatisfactory, because of its tendency to polymerise. From ethane he obtained the value C = 12'006 and from ethylene C = 12.004. On account of some uncertainties in the reduction, he prefers, provisionally, the value C Secondly, Batuecas (Journ. Chim. Phys., 1918, xvi., | 322) determined the density of ethane, and reduced bis observations by 3 methods, giving C 12'005, 11999, and 11.996. The last two being concordant he regards as preferable, and their mean, C 11'998, he adopts. It will be remembered that Richards and Hoover, by purely chemical methods, found C = 12'005; and a later combination of all determinations published before 1918 gave the chairman of the committee the mean value C = 12.0025. For ordinary purposes the rounded off value C = 12:00 may be used, and is so given in the table. = Bromine.-Three sets of determinations of the mole. cular weight of hydrogen bromide have been made in Guye's laboratory at Geneva, by Moles (Journ. Chim. Phys., 1916, xiv., 389; see review by Guye in the same number, p. 361), Reiman (Ibid., 1917, xv., 293), and Murray (Ibid., 1917, xv., 334; Reiman and Murray assume H=1'008; Moles prefers 10076). The acid used was prepared by several distinct methods, and all gave | concordant results, which may be summarised as follows, when H = 1'0078: Moles These values are wonderfully concordant and the variations are far within the allowable limits of experimental error. In a recent combination, by the chairman of this committee, of all the available data relative to the atomic weight of bromine, the value found was Br = 79'9228, in complete harmony with the Geneva determinations. For ordinary purposes the rounded-off figure 79 92 is enough. Boron and Fluorine.-In a very original investigation Smith and Van Haagen (Carnegie Inst. Pub., 1918, 267) bave simultaneously redetermined the atomie weights of boron and fluorine. Their starting point was anhydrous borax, Na2B4O7, and their chief difficulty was in insuring the complete dehydration of that compound. The salt was then converted, in a series of successive experiments, into sodium sulphate, carbonate, nitrate, chloride, and fluoride, which gave 8 independent values for boron rang. ing from 10.896 to 10'905, in mean 10'900. This value was computed with Na 22 997, CI 35'457, S = 32.064, N = 14'010, and C 12.005. The authors finally discuss all previous determinations and show wherein they were = = = = = Lead.-Oechsner de Coninck and Gérard (Comptes Rendus, 1916, clxiii., 415) have attempted to determine the atomic weight of lead by calcination of the nitrate, and find Po 206 98 when N2O5 = 108. This determination is evidently of no present value. With this exception the other recent researches relative to this constant have referred to isotopic lead, and the normal element is considered only in comparison with it. Richards and Wadsworth (Journ. Am. Chem. Soc., 1916, xxxviii., 2613), for instance, find for normal lead Po 107.183, and Richards and Hall (Ibid., 1917, xxxix., 537) give Pb 207 187, values slightly lower than the accepted 207 20 as determined by Baxter and Grover. Similar determinations by A. L. Davis (Fourn. Phys. Chem., 1918, xxii., 631) gave discordant results. As for isotopic lead its atomic weight is so variaable as to show that it is nearly, if not always, a mixture of isoiopes, and not a constant which can as yet be placed in the table. The values found have very great significance, but they are far from final. (For discussions regarding the atomic weight of isotopic lead see the Presidential address of Richards before the American Association for the Advancement of Science in December, 1918. Also F. W. Clarke, Proc. Nat. Acad. Sci., 1918, iv., 181). Gallium.-By the analysis of carefully purified gallium chloride, Richards, Craig, and Sameshima (Proc. Nat. Acad. Sci., 1918, iv., 387) find Ga 70 09 and 70'11. These determinations, however, are only preliminary, but they justify the provisional adoption of the value 70 10. The original values given by the determinations of Lecoq de Boisbaudran vary from 69 70 to 70°12, the last one being very near the new value. Zirconium.-From the ratio between zirconium chloride and silver, Venable and Bell (Journ. Am. Chem. Soc, 1917, xxxix., 1598) find Zr 91.76. Although this determination is regarded as preliminary, the authors, by pointing out sources of error in all previous values, believe the new one to be the most probable. It seems best, however, to await the complete investigation before changing the value heretofore accepted. Tin.-Baxter and Starkweather (Proc. Nat. Acad. Sci., 1916, ii., 718) by electrolyses of stannic chloride, find Sn = 118.703 when Cl = 35'457. This is in complete agreement with Briscoe's determination, Sn 118-698. The value 118 70 has already been adopted by the com mittee. = Tellurium.-Staehler and Tesche (Zeit. Anorg. Allgem. Chem., 1916, xcviii., 1) from careful syntheses of tel lurium dioxide, find Te = 127:51, which is confirmatory of the accepted value 127.5. International Atomic Weights. Aluminium 39'9 74796 Symbol. (1920). Atomic weight. 137'37 by it should be adopted. The other sulphate determinations are questionable. Samarium.-The atomic weight of samarium has been determined by Stewart and James (7ourn. Am. Chem. Soe., 1917, xxxix., 2605) from the ratio between the chloride and silver. The value found is 150'44, which is essentially that given in the table. No change is needed. Dysprosium.- Engle and Balke (Fourn. Am. Chem. Soc., 1917, xxxix, 67), by conversion of the oxide into the chloride, found Dy 164 228. Later, by the same method, Kremers, Hopkins, and Engle (Ibid., 1918, xl., 598) found Dy = 163.83. This discordance, like that already shown for yttrium, led the last named chemists to determine the ratio between dysprosium chloride and silver, which gave 162:52. The earlier method is discredited and the last value, rounded to 162 5, seems to be the one best entitled to acceptance. Thorium. In a long series of concordant analyses of thorium bromide, Hönigschmid (Zeit. Elektrochem., 1916, xxii., 18) finds Th 232 152 from the silver ratio and 232 150 from the silver bromide ratio when Br 79'916. The value Th= 232.15 should be adopted for general use. He also studied thoria from uranium ores, which contained ionium. For this mixture he obtained an atomic weight slightly in excess of 231 50. This may approximate to the unknown atomic weight of ionium. Uranium.-The latest series of determinations of the atomic weight of uranium by Hönigschmid and Horovitz (Monatsh., 1916, xxxvii., 185) was based like their earlier series upon analyses of the tetrabromide. Two sets of analyses were made, one upon a bromide which had been fused in bromine vapour, the other in nitrogen. The value obtained ranged from U= 238.04 to 238.16, the latter being in harmony with their former determinations. The rounded figure 238 2 is given in the table. Helium.-Taylor (Phys. Rev., 1917, x,, 653), using the microbalance for determining the density of helium, finds He 40008. Guye (Fourn. Chim. Phys., 1918, xvi., 46), in a recalculation of the data, finds He = 3.998. The value 4 should be retained. Argon. From the density and compressibility of argon Leduc (Comptes Rendus, 1918, clxvii.. 70) finds A = 39 91. He regards the second decimal as uncertain, and advises the adoption of the rounded value 39'9. In the above table of atomic weights proposed for 1920, few changes have been made from the values given in the last preceding table. The new values are A = 399; B = 109; Ga = 70'1; Th 232'15; and Yt = 89.33. In addition to these the atomic weight of nitrogen should be changed from 14:01 to the more precise value 14'008. The latter figure represents all the best determinations, and is probably correct to within I in the third decimal place. For so small a value the change is insignificant. 12. Electrolysis of Uranyl Acetate Solution. THE electrolysis of a neutral solution of uranyl acetate containing 10 grms. of metallic uranium per litre was carried on for one hour at 20° with a pressure of 34 volts and a current density of 195 amp. The cathode was rotated, and a porous cup was used as a diaphragm. The cathode was covered with a thin iridescent film that appeared metallic. The porous cup in the meantime had become partially filled with the yellow oxide of uranium. The catholyte was made acid with acetic acid and again electrolysed with 30 volts and a current density of 3.76 amp. for one hour at a temperature of 60°. The electrode became covered with a slight deposit of the yellow oxide mixed with a thin film quite metallic in appearance. The catholyte remained acid throughout, and no precipitate of the yellow oxide appeared under these conditions. A larger voltage, 85, and a current density of 18.2 amp. gave a deposit quite metallic in appearance, but contaminated with a small portion of the oxide. Only a little deposit adhered to the cathode, in quantity insufficient for analysis. The solution soon became heated to boiling, | and the larger portion of the uranium precipitated as a mixture of yellow and black oxides. Attempts to keep the solution cool were not successful when high current densities were employed. The temperature was kept below 60° by surrounding the cathode with a glass coil through which cold water could be circulated, and by keeping the electrolytic bath surrounded with a mixture of salt and ice. Pure metallic uranium could not be plated from an aqueous solution of uranyl acetate even with high current densities. The electrolysis under these conditions always produced a mixture of what seemed metallic uranium mixed with the black oxide. No analysis of the deposit could be obtained. 13. Electrolysis of Uranyl Salts in Acetone. A saturated solution of uranyl nitrate in acetone, when electrolysed with a voltage of 14 and a current density of o'77 amp. gave a deposit at the cathode of the yellow oxide. The saturated solution of uranyl chloride under the same voltage and current density gave likewise a deposit of the yellow oxide, which sloughed off and piled up below the electrode. When made acid with hydrochloric acid the deposit was the same in appearance, but it dissolved as soon as it fell off. Formaldehyde added to the solution lowered the conductivity but caused the deposit to adhere to the cathode. The deposit was smooth and bronze-like in appearance. On analysis the deposit yielded 46.50 per cent uranium. The bronze colour was due to some organic matter from the formalde byde and acetone, and the solution had a distinctly caramel odour. No pure metallic uranium was deposited. 14. Electrolysis of an Alkaline Solution of Uranyl Sulphate. A uranyl sulphate solution made alkaline with sodium carbonate was electrolysed with a rotating cathode using a voltage of 6 and a current density of 6 68 amp. at 58° C. The greater portion of the deposit dropped off the cathode, but there was left a small amount, very metallic in appearance, which did not dissolve readily in nitric acid. A portion of the deposit showed it to be the black oxide as usual. The deposit was washed with cold dilute nitric acid, and the insoluble portion dried and weighed. Its weight was very small. It was dissolved in hot nitric acid, * From the Journal of Physical Chemistry, xxxiii., No. 8. and the procedure repeated until the accumulated deposits were large enough for an analysis. The error introduced in this way is great, but this deposit was found to be richer in uranium than previous deposits analysed; 72 per cent of uranium was found. 15. Electrolysis of Double Potassium Uranyl Tartrate. An acid solution of potassium uranyl tartrate containing 5 per cent uranium was made by dissolving potassium uranate in tartaric acid. It gave a poorly conducting solution and deposited a smooth black coat on the cathode. This, like other deposits, was poorly adherent, and the greater portion fell off the electrode. The temperature at the time of electrolysis was 70° and the voltage 19.5, which gave a current density of 5 45 amp. The portion that adhered was dried and analysed. It contained 73.8 per cent uranium. A tartrate solution made alkaline with potassium carbonate was electrolysed with a potential of 15 volts and a current density of 19:4 amp. at boiling temperature. Under these conditions a more adherent deposit was formed which on analysis showed 73'5 per cent uranium. Another sample prepared under similar circumstance gave a uranium content of 76 4 per cent. In this run the electrolyte was prepared by dissolving sodium uranate in sodium acid tartrate. The solution was slightly acid during the run. This same solution when made just alkaline with sodium bydroxide and electrolysed with a potential of 12 volts and current density of 14 amp. at 76° gave a fairly adherent deposit. It was black with the exception of a few grains that were yellow. The deposit contained 81.5 per cent uranium. With a rotating cathode and a temperature of 80° a voltage of 9.5 and a current density of 75 amp., the deposit formed from this same solution was smooth and well adhering while wet. Its colour was black with a slight tinge of brown. When dried with alcohol and ether the brown portion scaled off easily. Analysis showed it to contain 70-8 per cent of uranium. The same solution electrolysed with a potential of 35 volts and a current density of 35 amp. at a boiling temperature gave a strongly adherent black deposit. This deposit contained 75.8 per cent uranium. The uranium content of deposits obtained from alkaline tartrate solutions approaches and sometimes exceeds the amount found in the later work with acetate solutions. The products are doubtless the same in general but vary in their uranium content, in the oxygen present, and in the bmount of hydration. Pure uranium was not obtained. 16. Electrolysis of Double Sodium Uranyl Citrate. A solution containing 5 per cent uranium was prepared by dissolving sodium uranate in citric acid in the molecular quantity for producing the double salt. It was electrolysed at 80° with a potential of 40 volts and a current density of 12'4 amp. The solution was slightly acid at the beginning but soon became basic. A dense poorly adherent black deposit formed at the cathode but dropped o rapidly. At the same time there was a precipitate of uranic oxide formed. No deposit was obtained uniform enough in appearance to justify analysis. The black deposit was always contaminated with the yellow oxide. During the electrolysis the anode became bronze in colour. On examination this proved to be carbon contaminated with some uranium salts. According to Smith (Am. Chem. Fourn., 1879, i., 336) a deposit does not form when a double aikalı tartrate or citrate is electrolysed, but this is contrary to my experience. 17. Electrolysis of Uranium Tetrachloride dissolved in Pyridine. Since the oxides of uranium are deposited from aqueous solutions of either usanyl or uranous salts it would seem possible to obtain metallic uranium from a solution which |