November 13, 1902] 



NATURE 



41 



slightly different ; the first raises the arm to the eighth row of 

 contacts and the second rotates it to the third contact in that 

 row. The subscriber is now connected to No. 56S3 and can 

 ring him up if he is not engaged ; the signal that he is engaged 

 is given through another set of contacts on the connector 

 switch, an interrupted current being sent along the caller's line 

 and causing his receiver to hum. When the conversation is 

 finished and the caller hangs up his receiver, all the switches 

 which he has been using return to their normal position of rest. 

 The exchange is run on the central-battery system and 

 metallic circuits are used throughout. It will be noticed that 

 the subscriber's connections are duplicated at the exchange, one 

 pair of wires running to the first selector and one to the con- 

 tacts corresponding to his number on the connectors. It is also 

 obvious that the number of connections in one group of 

 hundreds or thousands which can be made at once is limited 

 by the number of trunk lines ; ten of these per hundred sub- 

 scribers have been found to be practically sufficient, but the 

 number could, of course, be increased without limit if it was 

 found desirable. Once two subscribers are connected through, 

 their conversation cannot be interrupted, since any attempt to 

 call either up results merely in the caller receiving the busy 

 signal, and any calling up between other subscribers does not 

 affect the lines they are using. This alone is a very great con- 

 venience from the user's point of view ; in addition, the gain in 

 time in getting connected up, the impossibility of getting on to 

 a wrong number except by the subscriber's own fault and the 

 secrecy of the system must be reckoned to its advantage. So 

 far as the exchange is concerned, the chief advantage lies in 

 doing away with the exchange girl ; the cost of maintenance is 

 said to be no more than in a manually operated system, the 

 floor space required for connections and switches about the 

 same ; there is, therefore, a slight saving in room, since none 

 of the resting rooms which the strain upon the operators now 

 renders necessary is required. M. S. 



INSTANTANEOUS CHEMICAL REACTIONS 



AND THE THEORY OF ELECTROLYTIC 



DISSOCIA TION 1 



TT is generally held that instantaneous chemical reaction, if 

 not all chemical action, is dependent upon ions ; in other 

 words, that such reactions take place between electrolytes. In 

 order to test this point, the author has attempted precipitation by 

 double decomposition (like the reaction between silver nitrate and 

 hydrochloric acid) in solutions that are excellent insulators. As 

 a solvent benzene was chosen, though it seems that petroleum 

 ether or toluene would have been equally good. The benzene 

 used was the best that is made by Kahlbaum, free from thio- 

 phen. It was allowed to stand for days over phosphorus 

 pentoxide, from which it was distilled, and was finally kept 

 standing over metallic sodium. The conductivity was tested by 

 comparison with that of air. For this purpose an Arrhenius 

 resistance cell, with plates less than a millimetre apart, was 

 placed in series with a sensitive galvanometer, and a dynamo 

 giving a pressure of no volts. When the cell contained air a 

 slight movement of the needle could be seen on closing the 

 circuit, and on replacing the air by benzene the deflection was 

 somewhat less. The insulating properties were therefore 

 good. 



Some difficulty was found in obtaining suitable solutes owing 

 to the general insolubility of salts in hydrocarbons. Certain 

 oleates, however, are soluble, and those of copper, nickel and 

 cobalt were used. These were prepared by heating pure oleic 

 acid with the calculated quantity of standard solution of sodium 

 hydroxide and then adding to this sodium oleate solution a 

 slight excess of the sulphate of the heavy metal. The precipitate 

 was thoroughly washed with water and finally dried at 1 10°. 

 The salts so obtained were analysed by reduction in hydrogen. 



These oleates are readily soluable in benzene, even in the 

 cold, and give colours similar to those of salts of the corre- 

 sponding metals in aqueous solutions. On heating the dark red 

 solution of cobalt oleate in toluene it turns blue, and on cooling 

 it again becomes red, in the same way as cobaltous salts change 

 colour in aqueous solutions. It was found that 5 per cent, 

 solutions did not conduct any better than pure benzene. 

 Metallic sodium does not cause any precipitation, and was, in 



1 Abstract of a paper in the Journal of Pliysical Chemistry, vol. vi. 

 pp. 1-14, 1902, by L. Kahlenberg. 



NO. 



1724, VOL. 67] 



fact, used as a desiccating reagent ; the only change that ever 

 took place was the usual slight pinkish coloration that freshly 

 cut surfaces always assume after a time. Magnesium, aluminium 

 and zinc have been kept in a copper oleate solution for weeks 

 without in the least changing their appearance and lustre. It is 

 therefore abundantly proved that these oleates in benzene are 

 not ionised. Cryoscopic determination of the molecular weight 

 of copper oleate in benzene gave figures about 2400 and the 

 boiling-point method about 2650, whereas the theoretical figure 

 is 625'6, so that, according to the usual idea, the copper oleate 

 would appear to be polymerised. 



A solution of dry hydrochloric acid gas in benzene was next 

 prepared. The gas was obtained by dropping the aqueous 

 solution into concentrated sulphuric acid and further drying by 

 sulphuric acid and phosphoric anhydride. The conductivity of 

 this solution was no higher than that of the benzene itself. 

 It does not attack the carbonates of sodium, calcium and barium, 

 or bright magnesium ribbon. Zinc, however, is attacked, 

 whether amalgamated or not, but platinum in contact with it 

 makes no difference, the hydrogen being evolved from the zinc 

 alone. Thus voltaic action is absent. A dilute aqueous solution 

 of the acid seemed to act rather less readily on amalgamated 

 zinc than did the solution in benzene. Similarly, contact with 

 platinum or other metals does not cause magnesium to be acted 

 on by the acid. Iron, nickel, cobalt, copper and cadmium are 

 not attacked ; tin and aluminium are slightly acted upon, and 

 lead very slightly. This is so whether the metals are by them- 

 selves or in contact with others. Metallic sodium is fairly 

 rapidly attacked. The chlorides of the metals acted upon are 

 practically insoluble in benzene. 



In all the experiments, great precautions were taken against 

 moisture, the generators and other apparatus being connected 

 to suitable drying trains. The flask containing the benzene and 

 substance to be tested was fitted with a doubly perforated rubber 

 stopper, and was connected with the drying train of the hydro- 

 chloric acid generator and also to a large tower filled with 

 pumice and phosphoric anhydride. Before introducing the 

 benzene and substance, the flask, stopper and connecting tubes 

 were heated to drive off moisture, and while still hot, the 

 benzene and substance to be tested were quickly introduced and 

 the whole at once connected with the train. The air was then 

 displaced with dry hjdrogen, which passed through the hydro- 

 chloric acid generator, and finally the acid was slowly evolved 

 until the train was saturated. 



When dry hydrochloric acid gas is passed into a solution of 

 copper oleate in benzene, there is formed instantly a heavy brown 

 precipitate which is cupric chloride. We have here, then, a case 

 of instantaneous precipitation by double decomposition which is 

 perfectly comparable with that of the formation of silver chloride 

 in aqueous solutions, when silver nitrate solution is treated with 

 hydrochloric acid. Yet the benzene solutions conduct no better 

 than benzene itself, nor is there the least perceptible increase 

 of conductivity at the instant of the formation of the pre- 

 cipitate. The oleates of nickel and cobalt, when treated in 

 benzene solutions with dry hydrochloric acid, react in a perfectly 

 analogous manner. Analysis showed the precipitation to be 

 complete. 



It was found that the conductivity of two samples of anhydrous 

 stannic chloride is no better than that of air. This salt mixes 

 with benzene in all proportions, giving mixtures which are 

 equally non-conductors. Yet when such a solution is poured 

 into a benzene solution of copper oleate, there forms instantly a 

 heavy brown precipitate which is principally anhydrous cupric 

 chloride. The precipitate takes down some of the stannic oleate 

 which is formed with it and is difficult to manipulate, but 

 analysis shows that the reaction is in the main a simple double 

 decomposition. 



Anhydrous phosphorus trichloride, arsenic trichloride and 

 silicon tetrachloride are miscible in all proportions with benzene 

 and give solutions which are insulators, like the solution of 

 stannic chloride. In each case, when a solution of copper oleate 

 in benzene is treated with a solution of PC1 3 , AsCl 3 or SiCI 4 in 

 the same solvent, copper is precipitated as a dark brown pre- 

 cipitate. This is essentially cupric chloride, but is in each case 

 contaminated with some of the oleate. 



We see, then, that HC1, SnCl 4 , PC!..,, AsCI 3 and SiCl 4 each 

 precipitate cupric chloride from benzene solutions of copper 

 oleate. There is, then, apparently double decomposition by 

 means of ions, and yet the solutions are non-conductors, showing 

 that ions are not present. 



