USHER : RADIUM EMANATION. 389XLIV.-The In$uence of Rudiurn Emanation onEpuilibriuni in a Gaseous System.By FRANCIS LAWRY USHER.SOME interesting deductions concerning the nature of chemical changeinduced by radium emanation have lately been recorded by Cameronand Ramsay (Trans., 1908, 93, 966) as the result of quantitativeexperiments on the decomposition of water, ammonia, hydrogenD D 390 USHER: THE INFLUENCE OF RADIUM EMANATION ONchloride, and the oxides of carbOIi, and on the combination ofhydrogen and oxygen and nitrogen and hydrogen in presenceof theemanation.The principal conclusions drawn by these authors are (1) that thechanges observed are due almost entirely to the a-particles, and (2)that each particle of ernanation in disintegrating produces, ceterisparibus, the same amount of change.The experiments described areregarded by the authors as preliminary, and the results havequalitative rather than quantitative significance. At the suggestionof Sir William Ramsay, the investigation t o be described in this paperwas undertaken with the object of obtaining a more definite knowledgeof the mechanism of chemical change produced by the emanation,FIG. 1.based on an accurate study of the course of some simple reaction. Forthis purpose, five series of observations have been made, three withpure dry ammonia, and two with a mixture of hydrogen and nitrogenof the composition 3H, + N,.EXPERIMENTAL .It will be convenient to describe in detail the method of procedurein the two cases.Fig. 1 represents the apparatus used for theexperiments with ammonia. I n the Erst place the tap B was closed,arid the whole apparatus exhausted by means of a small mercurypump. Ammonia, prepared from pure ammonium chloride and sodiumhydroxide, was than introduced through B and condensed in thEQUILIBRlUM IN A GASEOUS SYSTEM. 391vessel D, which was surrounded with liquid air. B mas again closed,and the apparatus once more exhausted in order to remove traces ofair. The liquid air was removed from D, which now contained pureammonia, and a convenient quantity of this (about 1 c.c.) was pumpedoff and collected over mercury in a carefully dried gas-tube. D wasonce more cooled with liquid air, so as t o condense the ammoniaremaining in the apparatus, the tap E was closed, rtnd the system onthe pump side of 3 thoroughly exhausted.Radium emanation, accumulated duringfour or five days from a solution contain-ing 0.21 11 gram of metallic radium, andmixed with about 0.5 C.C.of hydrogen,was now introduced through the capillarysyphon H, and E was then surroundedwith liquid air. After about fifteenminutes, in which time all the emanationhad condensed, the hydrogen was removedthrough the pump, and the requiredquantity of ammonia, prepared in theway described, was introduced throughH and frozen in Pon top of the condensedemanation, If the pump was worked atthis stage, traces of gas continued to passover indefinitely, and an analysis of thegas thus collected showed it to consistsolely of hydrogen and nitrogen, so thatit appears that solid ammonia is decom-posed by the emanation, even at - 190'.The drying tubes, C and K, containedlime freshly prepared from marble.Themixture of ammonia and emanation masnext introduced into the apparatus shownin Fig. 2. This consisted essentially ofa short length (about 5 cm.) of glasstubing of 1 cm. bore, containing an opaqueglass point sealed in so as t o form aconstant-volume gas chamber, B, theFIG. 2.Lvolume contained between a mercury surFace set to the point andR mark, a, OLI the capillary stem being previously accurately deter-mined by calibration with mercury. The constant-volume chamberterminated above in a capillary syphon, S, and at its lower endwas sealed to a piece of narrower glass tubing about 80 cms.long,including a stopcock, T, the only one used in the apparatus, whichwas permanently below the mercury surface and never came int392 USHER: THE INFLUENCE OF RADIUM EMANATION ONcontact with the gas. A mercury reservoir, R, of the samediameter as the chamber B, mas connected with the apparatus bya length of rubber pressure tubing. The greatest care was taken todry the inner glass surface thoroughly, and for this purpose, beforethe stop-cock was greased, the entire apparatus was placed in a largeair-oven and kept at a high temperature for several hours while acurrent of dried air was passed through it. I n order to introduce thegas, the apparatus was filled with pure dry mercury, and the tubecontaining the gas was brought over the end of the syphon, 8, ina mercury trough.By lowering the reservoir, R, the gas wasadmitted, and the end of the mercury thread which iollowed it was setto the mark a on the capillary tubing. The thread was then frozenin the horizontal portion of the capillary a t 6 by means of a paper-cupcontaining solid carbon dioxide, and the tip of the syphon was thensealed with a small blowpipe flame. Finally, the apparatus was fixedup against a glass scale ruled in millimetres, and frequent readingswere taken of the pressure exerted by the gas when the mercurysurface was set exactly to the point. During the interval betweeneach successive reading, the tap T was closed, so that the reactionproceeded at constant volume.It happened, on a few occasions, thatafter the capillary tip had been sealed and the mercury threadhad thawed, the latter was no longer set exactly to the mark a, andin such cases the distance between the two was measured and acorrection on the volume was made, as the capillary had previouslybeen calibrated by weighing out mercury. I n making a reading thetemperature of the gas and of the mercury column was carefully noted,and the barometric height was read a t the same time,The above description refers to the experiments with ammonia, butthose with nitrogen and hydrogen were carried out in exactly thesame manner. The gases were obtained by sparking pure ammoniaover mercury in a glass tube. The undecomposed ammonia wasremoved with a few drops of phosphoric acid, the residual mixtureof nitrogen and hydrogen was carefully dried, and a convenientquantity was collected in a tube in the same way as the ammonia, thecalcium oxide, however, being replaced by phosphoric oxide.Thesubsequent procedure differed slightly from that employed in theexperiments with ammonia. The apparatus shown in Pig. 1 wasmodified to some extent, but it will suffice here to say that, afterhhorough exhaustion of the apparatus through t h e mercury pump, thepreviously dried emanation, accompanied by its excess hydrogen, wastaken in through a capillary syphon, the emanation was frozen withliquid air, and the hydrogen removed through the pump. Finally, theliquid air was removed, and the sample of nitrogen and hydrogencollected for the experiment was admitted and allowed to mix witEQUILIBRIUM IN A GASEOUS SYSTEM.393the emanation, the mixture being then pumped off and transferred t oone of the reaction vessels already described.It will be obvious that throughout the whole of the operations justdescribed there is no possibility of serious contamination. It is truethat during the process of purification, gaseous emanation was broughtinto contact with the stop-cocks H, G, and E (Fig. l), but the totaltime of contact between tap grease and emanation was certainly lessthan thirty minutes, and it was proved by a blank experiment withemanation and some of the same rubber tap grease that if oxygen isexcluded, the only gaseous product of the action is pure hydrogen,the amount of which produced in half an hour would be quitenegligible.Each experiment was allowed to proceed for at Ieast four weeks, a tthe end of which period the amount of emanation still present wasinsignificant.During the first two days, readings were taken everyfew hours, and afterwards at the rate of about one every twenty-fourhours.At the conclusion of each experiment, the gas was removed fromthe reaction vessel and analy sed, the ammonia, nitrogen and hydrogen,and gases absorbable by potassium hydroxide being determined.Reference will be made to those analyses in tbe discussion of theresults.Experiment 1.-Volume o€ reaction chamber : 2.1 187 C.C. Initialvolume of ammonia at 0' and 760 mm. = 0.4514 C.C. About half thequantity of emanation taken for this experiment was accidentallylost, so that the proportion of ammonia to emanation is not known :CorrectedPOI. of gas.Time i n days.0.0 0'45140'56 0.48010.77 0'49201 '56 0.51831'83 0.52352 '58 0'54364'54 0.57655 5 4 0'58387'56 0'59569 -67 0'602012-58 0.604840-00 0-627Volumeincrement.0.00.0290-0410'0670.0720.0920-1250.1320'1440-1510.1530.176l/Aog. Yo/ v,.-0,05160.05380.04480.04130.03840-03010'02720'02210.01830'01430.0054l/E,tlog Yo/ vt. -0.05710'06200-05940.05710.06150.07070.0740.0860.1030.137 394 USHER: THE 1NFLUENCE OF RADIUM EMANATION ONExperiment 11.-Volume of reaction chamber : 3.1655 C.C. InitialInitial pressure = volume of ammonia at Oo and 760 mm.= 1.843 C.C.474 mm. :Time in days.0.00-0420'0830'1040.1350.1910-8651-0311-19s1.8542.042tL.185'L.8403'2303.8403'9585 -8406.8967'8408.84012.88636-0009 -8.10Correctedvol. of gas.1.8431,8541.8711'874l.8811-8912.0682'1042.1452.2622.3022.3232.4162'4552'5122.5232,6522.7032'7242.7572.7752.8042-871Volumcincrement.0.00.0110.0280.0310-0380.0480-2250-2610'3020.4190.4590 ~4800-5730.6120.6690.6800.8090.8600.8810.9140'9320.9611 -028l / t log. Yo/ vt.I0.06190.08010.07080.06700.06010.06540.06430'06490.06050.06100.06000.05700.05430*051@0.05050.04300.03960.03500.03370'03110.024 80.00981 /El,( 1 og.Yo/ Yt.-0.06230'08120.07200.06650'06210.07640.07770.08080.0540.0580.0890 -0950.0970.1020.1030.1230.1360.1470'1650-1830.253 -Expeviment 111.-Volume of reaction chamber : 2.406 C.C. InitialInitial pressure volume of ammonia at Oo and 760 mtn. = 0.909 C.C.= 306 mm. :Time i n days.0'00.0310'0730-761-081-752.082.783.754 '756-757-168.759-7510.7611-7613-7714-71?15-7632.00Correctedvol. of gas.0-9090'9140.9231.0371 '0831'1541 '1 881.2371'2911'3251.3771 '3871.3971 '4011.4081'4191'419[1'440]1.4191.433Voluiiieincxenien t.0.00.0050.0140.1280.1740.24502790'3280'3820.4160'4680.4780.4880 4920'4990'5100'5100.5100'524[O .53 11116I0g. Vo/ V*.-0.07720.09230.08680.08550-07790.0766O * O i O O0'06320.05590.04660.04180.03820'03470'03210 03040'02600'02270'0117I O ~ O ' L ~ S ]11 E~tlog. Fro/ Vt.-0.07760-09340 '1 0000.1030.1080.1100.1160'1240.1320.1560.1680.1840 '2020.2210.2530.310[ 0,35910.392EQUILIBRIUM IN A GASEOUS SYSTEM. 395Expe&ne)tt I \'.--Nitrogen and hydrogen. Same t u b e as in Exp. I.Initial Initial volume of mixed gases a t 0' and 760 mm. = 1.602 C.C.pressure = 615 mm. :Time in days.0.00.1350.698273982,9903.6784-0104.69811.8031 *89Corrccted vol. of gases.1,6021.5851.5401-4621 -4601'4421'4411.4251.4031.362Volnmc incrcnient.0 '0- 0'017- 0.062- 0.140- 0'142- 0'160- 0.161- 0.177- 0.199- 0'2.10Experi.rnent V.-Nitrogen and hydrogen.Same tube as in Expt. 11.Initial Initial volume of mixed gases a t 0' and 760 inm. =2-323 C.C.pressure =594 mm. :Time in days.0.00.71.i2.74 -75-76.730'0Correctcd vol. of gases.2.3232.2732'2472.2302'2412.2282.2252*18lVolume iii cwiiien t.0.0- 0.050 - 0.076 - 0 '093 - 0.082- 0.095 - 0 '098 - 0'142The analysis of the gases at the conclusion of each experiment wascarried out by means of a small glass burette, provided with a stop-cock and capillary syphon, and containing six opaque glass points.The volume between a mercury surface set to each of these pointsand a mark on the capillary fitem mas accurately determined bycalibration with mercury, and measurements were made by observing,against a glass scale, the difference between the level of the mercuryin a reservoir connected with the burette and that of the mercury setexactly to one or other of the points.The measurements are in allcases correct to within 0.003 C.C. The gas was always measiired dry,and was, if necessary, for example, after explosion or treatment witha wet reagent, pumped through a small tube of phosphoric oxide.Ammonia was cietermined by absorption with a few drops of phos-phoric acid, hydrogen by explosion with a measured excess of oxygen,and carbon dioxide by absorption with a lump of fused and moistenedpotassium hydroxide." The residual ga3, after removal of excessoxygen by phosphorus, was measured and cousidered to be nitrogen.The treatment of the gmes with liquid and solid reagents took placein small gas tubes, the gas being completely freed from the reagent* Any other acid gases, oxides of nitrogen, etc., arc consequently called ''CO2.396 USHER: THE INFLUENCE OF RADIUM EMANATION ONand dried before introduction into the measuring burette.Thefollowing table gives t8he results of the analyses of the gases at thetermination of the experiments :I. 11. 111. IV. V.c. c. c. c. c. c. C.C. c. c.NH, ...... ........ 0,173 0.78 0.312 0'006 0*010H, ... . . . . . . . , . . . . . . . 0,327 1 *56 0.814 0'980 1'321N, ..............I . . . 0.121 0.56 0'298 0'356 0'669 co, . . . . . . . . . . . . . . * 0 -002 0.014 0'000 0'019 0-185CO .............. ... 0.004 0.004 0'009 0*001 0'003Discussion of Besults.It is interesting to compare these results with the figures given byCameron and Ramsay, and for this purpose it is convenieut t ocalculate the values of the expression: Q = 100 -'L -'' and thecorresponding logarithms, as' is done by these authors. These valuesare given in the following table for Expts. I and 111; the reasonsfor omitting Expts. 11, IV, and V will be given presently.'a -KJEXPERIMENT I.Time in days.0.000.560 -771 '561-832.554 -545 '547'569.6712'58Time in days.0.000.761 '081.752'082-783.754.756.757 -768 -759.7510.7611-7613.77YoO - Yt.0.1760.1470.1350.1090.1040-0840.0510.0440.0320.0250'023Yo0 - Yt &=loo- va - Yo'100'083 '676.862-059.147-729'025.018'214.213-1EXPERIMENT 111.0.5240.3960-3500.2790-2450.1960-1420-1080-0560-0460.0360'0320-0250'0140.014100.075.666.853.346'837-427 -120'610.78 -86'96.14-82.72'7Log.Q.2'0001 '9221.8851-7921.7721-6791 -4621.3981-2601-1521'117Log. Q.2.0001.8781-8251 *7271-6701.5731'4331.3141 '0290'9440.8390.7850-6810'4310'431Log. Q/100/t = - k.-1,3921'4941'3341.2451-2431.1851.0860.9790.8770'702Log. Q / l O O / t = - k.-0'1610.1620.1560.1590.1540.1510.1440'1440.1360'1330'1250-1230.1 330.11EQUILIBRIUM IN A GASEOUS SYSTEM.397The figures in the last column of the preceding tables represent theconstant in the equation 'd = e - k t , or, rather, the constantcalculated with common instead of Napierian logarithms. It will benoticed that the value of - k diminishes fairly regularly with time,and that the underlying assumption, which would require it to remainreally constant, does not, therefore, strictly represent the facts.The time of half action is in Expt. I, 2.4 days, and in Expt. 111,1.9 days, both considerably less than the half-life period of theemanation, which is 3.86 days (Sackur). It seems, therefore, thatthe simple hypothesis that each atom of emanation in decaying pro-duces the same amount of change, that, in fact, the effect is at anytime proportional to the amount of emanation present, although i tmay be true under certain conditions, requires some modification'to make it agree with the experiments here recorded.Let us assume that the velocity of reaction a t any time, t , isproportional, both t o the amount of emanation and of ammoniapresent at that time.Now, during avery small space of time, Et is constant, so that the above expressioncan be integrated by keeping KEt as the constant, and the resultingexpression can be subsequently corrected for the variation of Et. We'03 -'od V Then -x =kEtk'Vt= HEtVt.get then : llog yo = KEt.t P+For the sake of comparison, we may consider that the velocity ofdecomposition depends only on the amount of ammonia present, andis not influenced by the decay of the emanation.We should then1 To find that -.log- was constant. The reaction has been treated as ant vtirreversible unimolecular one, since it is obvious from the analysis ofthe gases a t the end of Expts. IV and V that recombination takesplace to an almost inappreciable extent. Whereas in Expt. 111, 65.7per cent. of the ammonia put in was decomposed, in Expt. IV,starting with nitrogen and hydrogen, only 0.75 per cent. of themixture was recombined.1 1t vt E:tt v, The values of - log 5 and of -. log 5 have already been tabulatedfor Expts. I to I11 on pp. 393, 394, and it is interesting to note thatwhile the constant becomes smaller with time when no correction forthe decay of the emanation is applied, it becomes larger when thecorrection is introduced.Obviously, it is unreasonable to omit thecorrection for decay of the emanation; nevertheless, when the fullcorrection is put in, the constant changes in the opposite direction,although at the same time a distinct improvement is noticeable. I398 USHER: THE INFLUENCE OF RADIUM EMANATION ONis here suggested that a third factor, namely, the eficiency of theemanation, is required in order to explain the increase of the velocit1constant with time. Since an a-particle is effective over a range ofabout 8 cm. in ammonia gas under the pressures employed i n theseexperiments, a large proportion of its energy must be wasted when i tis enclosed in a tube of 1 cm.bore, although this waste need not beproportionately greater at one time than another; but on0 atom ofemanation is capable, as will be shown later, of decomposing at least134,000 molecules of ammonia, It is, therefore, highly probable thanwhen, as in Expt. 111, the emanation is present in the proportionof 1 atom to 10,850 molecules of ammonia, the efficiency of an a-particlewill be greater, as its chance of colliding with a larger number ofmolecules increases; in other words, each a-pnrticle will do morework when there is mor0 work to do.It is possible to make an approximate correction by assuming thatthe efficiency is proportional to the ratio of the number of emanationmolecules to ammonia molecules at any time, although of course thiscannot be expected to hold over an extended period.We may assume t h a t - - dv = KEtV&, where Pt is the efficiency.dt1 VEt' Vft v+ Then if Ptm ' we get -.log 0 = constant. This expression givesa much better value for k over a" period of six or seven days, startingat one day from the commencement of the experiment, but it after-wards becomes smaller again, a result which is to be expected for tworeasons: first, because the assumed correction is the most drasticpossible, and can only be strictly valid over a very short range ; andsecondly, because as the reaction proceeds, the energy of the emana-tion is more and more used tip in useless work, namely, in impartingincreased velocity to the accumulating products of decomposition.W ecan, t,herefore, make the further assumption that the efficiency isproportional, not only t o the ratio of the amounts of emanation andammonia, but also to theactualquantity of ammonia present. I n this case,pt = -.Vt, vt and the velocity constant becomes R= 1 log 5.Et &Vt. t" vtThe constants calculated in the two ways suggested are tabulatedon p. 399 for Expt. 111.The correction appears to be rather too great in the first case, andslightly too small in the second, but on the whole both sets of constantsare much better than when no correction for change in efficiency isintroduced. It would doubtless be possible by suitably compromisingbetween the two methods t o obtain a still more constant value of K ,but is it probably not worth while to attempt this, because there arEQUILIHHIUM IN A GASEOUS SYSTEM.399Time in days. K = l/Yttlog. Yo/ Yt. K =0'0730.761 -081 *762-082-753.754 -756.757 -768.759.7510'7611.760.1030.1110.1160-1170'1210.1200.1200.1130.1060.0970.0910,0830,0780.0761- - Iog. Yo/ vt. J7m. t0.0980.1060.1090.1110.1160-1180'1210'1210.1290-1270-1270'1250.1340'138slight complications in all the experiments, and these will nowconsidered.It will have been noticed in Expt. IV that if the amountbeofnitrogen and hydrogen recombined is calculated from the observedchange of pressure, there should be 0.240 C.C. of ammonia at theconclusion of the experiment.As a matter of fact, the analysis showsthat only 0.006 C.C. was formed. The gases and the apparatus wereboth very carefully dried, there was no contamination by air, and nopossibility of leakage during the course of the experiment; the gaswas under considerably reduced pressure the whole time. Clearly,then, nearly a quarter of C.C. of gas had ceased to exert anypressure. We can make up a balance sheet with respect to the totalquantity of hydrogen and nitrogen put in a t the commencement, andfound a t the termination, of two typical experiments : Nos. 111 andIV, reckoning as hydrogen and nitrogen these gases in combinationas well as free. This balance sheet gives the clue to the observeddiscrepancies.Hydrogen Nitrogenput in at put in at Hydrogen Nitrogencomniencement. eommencemcnt, found at end.found a t end.c. c. c. c. c. c. c. c.Espt. I11 ...... 1.363 0'454 1 '282 0.454Hydrogen lost in Expt. 111 ............... 0.081 C.C.Nitrogen ,, ) ) I V ............... 0.041 ),I n Expt. 111 a small quantity of nitrogen mas probably lost, forthe &st reading was made as soon as possible after, but not atprecisely the same moment as, the emanation mas mixed with theammonia. The initial volume therefore refers to a mixture ofammonia with a trace of its decomposition products, and not to pureammonia, as is assumed for the purpose of the above calculation.)) IV ...... 1.201 0'400 0'989 0-359Nitrogen ), ) ) 111 ............... 0'000 ,)Hydrogen ), I V ............... 0'212 ,400 USHER: THE INFLUENCE OF RADIUM EMANATION ONIn Expt.IV we find, as one would expect, that a larger proportionof hydrogen and nitrogen is missing, because the partial pressure ofthese gases is considerably higher than in the preceding experiment,in which no free hydrogen or nitrogen was introduced initially.Now this missing gas can only have disappeared in three mays : (1)it may have reacted chemically with the glass of which the apparatusmas made; (2) it may have been driven into the walls of the vesseland remained there, or (3) it may have gone completely through theglass. The first possibility is very unlikely, because nitrogen waslost as well as hydrogen, and the glass did not present the appearanceof having been attacked ; only the usual brownish-violet coloration wasobserved,In order to settle this question definitely, it was decided to carryout a blank experiment with pure hydrogen and emanation, arrangedso that any loss of gas could be observed and measured, and so as todetect any gas which might pass right through the glass.In themeantime, the three reaction chambers employed in the fiveexperiments already described were coarsely powdered, placed in apiece of Jena glass tubing, first exhausted cold by a Topler pump,and finally heated to redness and again exhausted; nearly 2 C.C.of gas were pumped out of the heated glass, and its composition wasas follows. The measured total volume was 1.817 C.C. :CO, 1.416 C.C.CO 0.340 ,,H, 0.096 ,,N, 0-066 ,,Total ... 1.818 C.C.The experiment was rather unsatisfactory, as the powdered glasswas not treated with chromic acid to remove traces of grease,dust, etc., before being heated; nevertheless, nearly 0.1 C.C.ofhydrogen was recovered, and a rather smaller quantity of nitrogen.The apparatus used for the blank experiment was made of glass ofabout the same thickness as that used in the previous experiments.It is diagramatically sketched in Fig. 3. The constant volumechamber, C, containing an opaque glass point, was itself sealed intoa wider piece of glass tubing, A, which was drawn out at the top andconnected, through a small phosphoric oxide tube, with a Toplerpump, no taps being used. The space between the reaction chamberand this outer tube was at the commencement of the experiment verythoroughly evacuated, and the pump with which it was connected masworked from time to time during the experiment in order to collectany gas which might be driven througn from the reaction chamberEQUILIBRZUM IN A QASEOUS SYSTEM.401The latter was sealed, immediately below the inserted join, to a pieceof narrower glass tubing about 80 crns. long, the lower end of whichwas connected through an air-catch with a length of rubber pressuretubing attached at its distal end to a mercury reservoir, H. A smallcapillary syphon, 8, was sealed on shortly below the inserted join, andwas used for taking in the hydrogen and emanation. Another pieceof tubing was sealed onabout 4 crns. below thesyphon, and was con-nected through the tap 17with a second Toplerpump*Mercury was firstpoured into the reservoir,and the rubber tubingwas clipped when themercury stood a shortdistance below the T-piece carrying the tap T,The end of the capillarysyphon was sealed, andthe apparatus was thenvery thoroughly ex-hausted through T, Thelatter was then sbut, andt h e reservoir was raiseduntil the mercury stoodin the tubing betweenthe lower T-piece andthe capillary syphon.Theend of the latter was thenscratched with a glassknife, and the point wasbroken off inside a smallFIG. 3.gas tube containing the emanation mixed with about half a C.C. ofhydrogen, carefully purified and dried. The gas entered the apparatus,and the reservoir mas again raised, until the hydrogen and emanationwere forced up into the reaction chamber, and the mercury in thereservoir was level with the tip of the capillary syphon ; the latter wasthen sealed in a blow-pipe flame.The entire apparatus was fixed up infront of a glass scale, and frequent readings were taken of the pressureof the gas when the mercury was set to the point. The volumeof the reaction chamber was subsequently determined by measuringthe pressure of Y quantity of dry air introduced into it, and afterward402 USHER: THE INFLUENCE OF RADIUM EMANATION ONremoved and re-measured in a constan t-volume point burette alreadycalibrated.Throughout this experiment no gas passed through the wa!lsof the reaction cbamber into the surrounding vacuous space, whichremained quite empty.The amount of hydrogen in the tube decreased,however, from 0.610 C.C. to 0*4S7 C.C. The following table gives thereadings :Correctedvol. of hydrogen. Time in days.0.00 0-5731-15 0.5301'81 0'5142.81 0.5073-81 05104.81 0.5056-81 0'4958.81 0'49415-81 0'49025-82 0.489Volumedecrease.0'0430'0.590'0680'063O'OGS0.0780.0i90'0830-084-K= l/Et. t log. Yo/ Yt.-0.03630'03550.03150.02610.02iO0 03130 03600 07110.267The results are not sufficiently regular to admit of their applica-tion as quantitative corrections to tho experiments with ammonia andwith nitrogen and hydrogen. It is even possible that aEter some timethe glass walls of the containing vessel become so pitted by thebombardment that the surface is appreciably altered, and, in anycase, the problem is probably much more complicated than it a t firstappears.A t the conclusion of this last experiment, the reaction chamberwas powdered, and the powder was carefully cleared with hot chromicacid, washed, dried, and put in a clean Jena-glass tube andexhausted cold, It was then exhausted at, a red heat, and inthis way 0.076 C.C.of gas was extracted. Its composition mas :CO, 0.014 C.C.H2 0.061 ,,--Total 0.075 ,,There can therefore be no doubt that hydrogen, and, to a smallerextent, nitrogen, is driven into the glass walls of its containingvessel when mixed with radium emanation, The greater part ofsuch gas can be recovered when the glass is strongly heated,I n calculating the velocity constants and the values of theexpression &, it was stated on p.396 that the reasons for omit-ting Expts. 11, IV, and V would be given later. 1V and V wereof course omitted because the observed pressure changes do notreally iudicate recombination of nitrogen and hydrogen, as hits justbeen shown. It will be seen on referring to the analysis of thegas from Expt. 11 that, although there is an apparent loss oEQUILIBRIUM IN A GASEOUS SYSTEM. 403hydrogen, there is more nitrogen at the end than there was at thebeginuing, and the same is true for Expt. V. There was noordinary leakage during either of these experiments, but the sameapparatus was used for each, and this curious result may be explainedon the hypothesis that there were bubbles in the glass vessel at itsjunction with the capillary tubing at its upper end, and that the glasswas so thin in the region of these bubbles as to become perforated bythe a-particles, and so allowed a slow diffusion of air from without.Unfortunately, the discrepancy was not discovered until after thetubes had been ground up.Chemical E’ciency of the Emnncction.It has been pointed out that in all the experiments hithertodescribed, the emanation probably brings about only a fraction of theamount of decomposition which it could effect under more favourableconditions.I n Expt. 111 the total volume of ammonia decom-posed was 0.5‘37 c.c., and the emanation which mas mixed with itwas the product of six days’ accumulation, and therefore, accordingto the recent work of Gray and Ramsay (Trans., 1909, 93, 1073),0-000081 C.C.I n this experiment, then, the ratio of the volumes ofemanation and ammonia was 1 to 7380, or, in other words, eachatom of emanation decomposed on the average 7380 molecules ofammonia.In order t o get some idea of the amount of chemical work whichcould be done by the emanation under favourable conditions, an ex-periment similar t o those. described above was carried out on a muchlarger scale. A large round-bottomed flask of 2 litres capacity wasfilled with ammonia a t about 260 mm. pressure, andjmixed with sixdays’ accumulation of emanation. The course of the reaction couldnot, of course, be followed, but, at the end of a month, the gases werepumped out and the quantity of nitrogen and hydrogen produced wasroughly estimated.It was found that 10.9 C.C. of ammonia had beendecomposed. This is, in every sense, a minimum value, for no accountis taken of the fact that that portion of the emanation wbich wasnear the walls of the flask was not entirely used up in decomposingammonia, and no carrection is introduced for gas driven into theglasP.There is, however, no doubt that the conditions of this experimentwere extremely favourable as compared with those of the precedingones, and were probably such as to secure a t least 90 per cent. of themaximum amount of decomposition. In this case, one molecule ofemanation decomposed 134,300 molecules of ammonia.If we take this as an approximate measure of the chemicalefficiencyVOL. XCVII.E 404 USHER : RADIUM EMANATION.OF the emanation, we can calculate the fraction of the total energyof the emanation which is used in effecting chemical decomposition.It is reasonable to assume that the a-particles are mainly respon-sible for the effects observed, and, further, that their power to decom-pose ceases when they no longer produce any other physical effect,that is, when their velocity is reduced to 64 per cent. of their averageinitial velocity of projection.On this hypothesis, the total kinetic energy of one atom of emana-tion available for chemical work will be that of three a-particles, sincethe atoms of radium-A, -23, and -C are projected with less than thecritical velocity.Now, the total kinetic energy of one a-particle is 6 x 10-6 ergs.,hence the energy available for chemical work = (6 x 10-6) x (0.64)2 =about 2.5 x 10-6 ergs.Therefore one atom of emanation produces about 7.5 x 10-6 ergs.available energy.Taking the mechanical equivalent of heat as4.182 x lo7 ergs. per calorie, this amounts toNow, 134,300 molecules of ammonia (which are decomposed by oneatom of emanation) require for complete decomposition about2-02 x lO--15 calories, hence the chemical efficiency of the a-particle inthis exneriment isor a little more than 1 per cent.As regards its influence on equilibrium in the system ammonia-hydrogen-nitrogen, it can only be said that if any definite state ofequilibrium were reached under ideal conditions, it would be onecorresponding with a very high temperature.Under ordinary con-ditions there is no true equilibrium, but only a state of rest dependingon the proportions of ammonia and emanation, surface, and possiblyother factors as well. The emanation cannot be called a catalyst inany sense, and the effects produced are probably mechanical or electricalin origin.The principal conclusions arrived at in the course of this investigationmay be summarised as follows :(1) Ammonia is decomposed by radium emanation at the ordinarytemperature, and the decomposition is nearly irreversible.(2) Recombination was not observed to take place to a greaterextent than 0.86 per cent,(3) Decomposition of solidified ammonia by solidified emanationproceeds with appreciable velocity a t - 190'ATTEMPTED RESOLUTION OF IlACEMIC ALDEHYDES. 405(4) The decomposition a t the ordinary temperature follows approxi-mately the course of a unimolecular homogeneous reaction whencorrecting factors for the decay of the emanation and alteration of itsefficiency with time are introduced.(5) If t h e ratio of ammonia to emanation molecules does notexceed 10,000 to 1, the statement that each atom in disintegratingproduces the same effect is not strictly true, on account of the wasteinvolved when t h e system is rich in emanation.(6) The largest effect observed mas the decomposition of 134,300molecules of ammonia per atom of emanation,(7) The energy required to produce the largest effect observed wasabout 1 per cent. of the energy actually expended during the productionof that effect.(8) All experiments with gases in glass vessels in presence of theemanation are complicated by the fact that gas is driven into theglass, and can only be recovered by heating strongly.(9) Hydrogen is driven into glass t o a greater extent than nitrogen,and as much as 0.24 C.C. of the former gas has been thus lost duringa single experiment.I wish, in conclusion, to express my indebtedness to Sir WilliamRamsay, who kindly placed at my disposal thethe experiments, and whom I have also to thankcriticism.UNIVERSITY COLLEGE,LONDON.emanation used infor his advice an