576 MORRELL AN b BURGEN :LVIIL-Tlze Polymerisation of Cyunamide.By GEORGE FRANCIS MORRELL and PETER BURGEN.As is well known, cyanamide undergoes, according to conditions,more or less rapid polymerisation into dicyanodiamide, yet thekinetics of the reaction, and the precise conditions under whichit occurs, do not seem to have been investigated. Since a t thepresent time cyanamide and its salts are of importance as fertilisers,a still greater interest attaches t o this subject, and this investigaTHZ POLYMERISATfON OF CYANAMIDE. 577tion was undertaken with a view of obtaining quantitative data as t othe course of the polymerisation, and tlie'influence of catalysts on it.The formation of dicyanodiamide during tlie evaporation ofaqueous solutions of cyanamide was first recorded by Beilstein andGeuther (Annalen, 1858, 108, 99; 1862, 123, 241), and later byDreschel (J.pr. Chem., 1875, [ii], 11, 298), and the influence ofammonia in accelerating the change by Haag (Annalen, 1862, 122,22). Beilstein and Geuther (Zoc. tit.) also make the statementthat a specimen of cyanamide which had been kept f o r severalmonths no longer gave the characteristic reactions for this sub-stance. Baumann (Ber., 1873, 6, 1373), studying the formationof carbamide by the action of mineral acids on cyanamide, observedthat dicyanodiamide was produced as a by-product, and that themore dilute the acid the greater the proportion of dicyanodiamideobtained. He also made the observation that alkali hydroxidesaccelerated the polymerisation, as Haag had previously observed inthe case of ammonia.Whilst the present investigation has in the main confirmed thequalitative statements of these early investigators, it has in itsquantitative aspect brought out the fact, which from statements inbooks of reference one would scarcely have expected, that the poly-nierisation of cyanamide in aqueous solution of even 5137-concentra-tion, and a t elevated temperatures, proceeds with considerablesluggishness, and that, on the other hand, an enormous acceleratinginfluence is exerted by acids and alkalis, even if present inextremely minute quantities.It was found, moreover, that in nocase did the polymerisation proceed as a bimolecular reaction asexpressed by the equation:and evidence was forthcoming that the change resulting in theformation of dicyanodiamide is ionic in character, although a singleexplanation, which wiI1 cover the acceIerating action of both acidsand bases, cannot be put forward.NHZ-CN + NH2*CN = CzH4N4,The Deter m i t i at io n of Cyaiza m ide.I n order to follow quantitatively the course of the polymerisa-tion, it was first of all necessary to find some trustworthy, and a tthe same time rapid, method of determining the amount ofcyanamide present in the solution from time to time.A numberof processes are to be found in the literature,* all involving as their* The original papers dealing with the metbods of deterniination used in thiswork are as folloms : Perotti, Cazzetta, 1905, 35, ii, 228 ; Curo, Zeitsch. angew.Chem., 1910, 23, 2407; Monnier, Cham.&it., 1911, 35, 601 ; Stutzer, Chem. Zeit.,1911, 35, 694 ; and Kappen, Cl2eut. Zeit., 1911. 35, 950578 MORRELL AND BURGEN :main principle the preciRitation of silver cyanamide by means ofan ammoniacal silver solution. With regard, however, to the con-ditions of precipitation and the method t o be adopted for the sub-sequent determination of the cyanamide, much diversity of opinionexists and conflicting results have been put forward, but it appearscertain that concordant results cannot be obtained without strictadherence to specified conditions, owing t o the variable compositionof the silver cyanamide precipitated in varying circumstances. Aconcensus of opinion favours a Kjeldahl determination of thenitrogen in the silver cyanamide precipitate as being, if not abso-lutely accurate, a t least more free from error than titration methods,and uninfluenced by the variability in composition of the precipi-tate.I n view of the absence of any definite information as to themagnitude of possible error involved by the much more easilycarried out titration methods, comparative experiments have beenmade with solutions of partly polymerised cyanamide of variousstrengths to ascertain the conditions necessary for obtaining con-cordant results, and the degree of accuracy attainable.The determinations were first carried out using solutions whichvaried in cyanamide content from normal to half-normal, and con-taining more and more dicyanodiamide as the, percentage of cyan-amide sank.Five C.C. of the solution were measured out, dilutedto 50 c.c., and 20 C.C. of tllis diluted solution used for each deter-mination. Caro’s method, namely, precipitation with aminoniacalsilver acetate, and determination of nitrogen in the precipitate byKjeldahl’s method, was compared with that finally adopted byKappen, namely, the addition of a known excess of ammoniacalsilver solution and titration of the filtrate, and washings withammoilium thiocyanate. The results are shown in the followingtable, the figures referring to 20 C.C. of the diluted solution, or2 C.C. of the original:AT/ 10-HydrochloricN/10-Ammoniacal X/lO-Thiocyanatc N/lO-Silver used acid requiredsilver added. for back titration.for NK-CN. for Kjeldahl.1. 40 C.C. 20.9 C.C. 19-1 C.C. 19.2 C.C.21-3 ,) 18.7 ), 18.7 ,,2. {2 :: 17-8 ,) 18.2 ), 18.6 ,,3.4* {:; :: 21.6 ,, 10-4 ), 10.7 ,,22-5 )) 17-5 ,, -29.2 ,, 10.8 ,, 10.8 ,,- 140 9 ) 134 9 ) 17-0 ,) 17.0 ))The table shows a t once the effect on the titration values of theaddition of a greater o r less excess of silver, the precipitate pro-duced in the former case containing a larger proportion of silverto cyanamide radicle than in the latter. This, according to Stutzer(Zoc. c i f . ) , is due to the formation of double and basic compoundsTHE POLYMERISATION OF CYANAMIDE. 579The Kjeldahl values, however, appear to be practically unaffected,and hence more trustworthy, This is in agreement with Car0 (Eoc.cit.), who states that although the precipitate may vary in com-position, it contains all the nitrogen of the cyanamide, which ityields as ammonia by the Kjeldahl process.Under the above conditions, then, the titration method admitsof considerable error; but it was found by further experiment thatif the precipitation was conducted in extremely dilute solution, avariation in the excess of silver within reasonable limits producedscarcely any measurable difference in the titration results.Themethod finally adopted was therefore as follows: The solutionunder test was rapidly cooled, 5 C.C. were measured out with apipette and diluted to 50 c.c., of which 20 C.C. were taken for eachdetermination. Before precipitation, however, it was again dilutedwith about 100 C.C.of water, and a measured excess of AT/lO-silversolution, also preferably somewlat diluted, and containing about2 per cent. of ammonia, was then added. After remaining forabout thirty minutes, for the very finely divided precipitate t ocoagulate, it was filtered through special '( barium sulphate " filterpaper, and washed first with weak ammonia water, and then withpure water until the washings were free from silver and ammonia.The excess of silver in the acidified filtrate and washings was thendetermined by titration with standard ammonium thiocyanate.The appended table shows the results obtained by the titrationscompared, where necessary, with the Kjeldahl determination on theprecipitate :N/10-Ammoniacalsilver added.{ ;: c;p'{;: ::{:: ::{E ::40 ,, {:; fiN/lO-HydrochloricN/1 O-Thiocyanate N/10- Silver used acid requiredfor back titration.for cyanamide. for Kjddahl.20.60 C.C. 19.40 C.C. 19.30 C.C.10.80 ,, 19.20 ,, 19-30 ,,11.15 ,, 18-85 ,,21-15 ,) 18-85 ,, -20-75 ,, 19-25 ,, -10.80 ,, 19.20 ,, -24.30 ,, 15.70 ,, -10.35 ,, 15.65 )) -29-00 ,, 11-00 ,, 10.50 ,,29.10 ,, 10.90 ,, -13-50 ,, 10.50 ,, 10.40 ,,10.40 ,) - 13.60 ,,30-35 ,) 9.65 ,, 9.35 ),10.70 ,, 9.30 ,, 9-20 ,,-From the results of 1, 2, 3, and 4 it is evident that the titrationmethod can be relied on if carried out as above described. Deter-minations 5 and 6 show, however, that if the excess of silver addedis very great, that ig of the order of four times the amoGnt theo-VOL.cv. Q 580 MOliRELL AND BURGEN :retically required for the precipitation, then the results begin tovary; but a determination of nitrogen in the precipitate still givesconcordant results, agreeing well with the lower titration reading,where only a moderate excess of silver was employed. Hence, inthe experiments about to be described, in every case where theinitial and final values of cyanamide were widely divergent, a p r eliminary series of experiments was made, precipitating throughoutwith the same volume of silver solution. A second series was thenconducted, readings being taken a t the same intervals as before,and a volume of silver added which was just about double theamount found to be required for the precipitation in the pre-liminary series.I n this way the addition of an excessive proportionof silver was avoided.The Polymerisation of CyanQmide in Aqueous Solution.The figures given refer, as before, to 2 C.C. of the original un-(a) Solution maintained a t looo in a boiling-water bath:diluted solution.N / 10-SilverprecipitatedTime in hours. N/lO-Silver added. by cyanamide.0 40 C.C. 19.30 C.C.1 40 ,, 19-20 y y4 40 Y Y 18-85 ),14 26 ,Y 13.00 ,y19 24 YY 10.45 ,)( b ) Solution gently boiled over wire-gauze :Time in hours,0123591523N/10-Silver added.40 C.C.40 Y Y40 Y Y40 Y Y40 9 940 Y Y40 Y Y40 y7N / 1 O-Silverprecipitatedby cyanamide.19.40 C.C.19.20 ,)18-30 y,17-30 ,,15.35 y y12.10 ,)8-60 ,)5.55 ,)( c ) Solution gently boiled over wire-gauze :Time in hours.026101216204N/lO-Silver added.40 C.C.30 y y30 9)20 ¶ Y16 Y Y40 y740 y yNI10-Silverprecipitatedby cyanamide.19-35 C.C.19.00 y y15-50 ,)12-15 y y10.50 ),9.00 ,,7.35 ,,K = l/t log, a/a - x.0.002e0-003e0.016e0.0 19eK = l/t log, a/a - x.0.0045e0-020e0.024e0-026e0-026e0.024e0.024eK = l/t log, ala - x.0-004e0.022e0-026e0-028e0-018e0.019THE POLYMERISATION OE‘ CYANAXIDE.581(d) Solution maintained a t 7 5 O in a thermostat:N / 1 O-SilverprecipitatedTime in hours. N/lO-Silver added. by cyanamide. K =zit.0 160 C.C. 90.60 C.C.10 160 ,) 81-25 ,, 0.02325 160 ,. 67.40 ,, 0.02015 160 9 , 76-35 ,, 0-02420 160 Y 9 71-40 ,, 0.025An examination of these results brings to light several remark-able facts.A t the commencement of the heating, in every case butlittle effect on the titration value was produced, even after theexpiration of from one to two hours, and in the case of series (a),where the solution was not actually boiled, but a water-bath wasused to effect the heating, the period of comparative stabiIity con-tinued f o r some four hours. During the initial stages of the secondperiod, the reaction is characterised by attaining its maximumobserved velocity, a t which it proceeds almost constant for aboutten hours in series (b), ( c ) , and (d), and still longer in series (a).Equal amounts of cyanamide are then polymerised in equalintervals of time, as may be seen a t a glance from the second andthree following readings in each table, and also from the ascendingvalues of “K,” calculated for a unimolecular reaction.A distinctretardation is then observable, but the figures obtained do not agreewell with the velocity of a unimolecular reaction, and are certainlyin even worse agreement with a bimolecular reaction.Since the results of each series of experiments are substantiallythe same, and the method of determination has been shown to betrustworthy, an explanation of the apparently erratic figures mustbe sought in the complicated nature of the reaction itself. Theinitial stability of cyanamide in pure water must be conceded asbeing characteristic of the pure substance.I n order firmly t oestablish this, a further series of experiments was undertaken, usingcarefully recrystallised cyanamide dissolved in distilled water, andheated a t looo in a water-bath in a flask of hard resistance glassprovided with an efficient reflux condenser. The cyanamide wasfound even a t this high temperature to be quite stable in itstitration-value during three hours’ heating :Time in hours. N! 10-Silver nitrate required.0 19-85 C.C.1 29-85 ,,2 19.95 ,,3 19.80 ,,The subsequent acceleration in the rate of polymerisationsuggests, in the light of our knowledge regarding the acceleratinginfluence of bases, the gradual contamination of the solution withQ Q '582 MORRELL AND BURGEN:basic constituents dissolved out of the glass.Such contaminationwould naturally be more rapid in solutions which were actuallyboiled, but the maximum of contamination would eventually bethe same in every case a t the same temperature. The experimentshave confirmed this, for in series ( b ) and (c) the period of stabilityis shorter than in series (a). Yet the final velocity-constantsattained are all' of the same order, series (a) gradually rising to0.019 after nineteen hours, ( 6 ) and (c) being a t that time 0.024and 0.019 respectively. This, furthermore, suggests a reason whythe observed rate of polymerisation in the first stages does notdecrease, the loss in reacting substances being accidentally justcompensated by the additional small quantity of catalyst still enter-ing the solution a t that stage.The linear character of certainperiods of the reaction would on this hypothesis be explained away,but the absence of the expected bimolecular reaction,NH,*CN + NH,*CN = C2H4N4,and the general indication of a unimolecular reaction, still demandsexplanation.An ionic theory of the polymerisation suggests itself as feasible,an ion maintained in the solution in constant concentration reactring with some other constituent to form dicyanodiamide ordicyanodiamide ion. Further information on the catalytic actionof bases was now found to be necessary in order to put a theory ofthis nature to the test. First of all, however, a series of experi-ments was made to demonstrate the influence of the solvent. Ifone or both the reacting substances were ions, either derived fromthe dissociation of cyanamide as an acid, or generated by thepresence of a base, then the use of a less powerfully ionisingsolvent, such as ethyl alcohol, would be expected t o depress therate of polymerisation.This expectation was borne out by experi-ment, as the following table shows:Polymerisation in Alcoholic Solution.Approximately 5N-alcoholic solution maintained a t 75O in athermostat:Time in hours.0510152026N/lO-Silver nitrate required for 0.5 O.C.22-50 C.C.22.45 ,,22-35 ,,22-15 ,,22-16 ,,22-15 ,,A comparison of this series with series (d) above, which wascarried out in aqueous solution of about the same strength underotherwise absolutely similar conditions, shows that whereas iTHE POLYMERISATION OF CYANAMIDE. 583alcohol the cyanamide was almost stable for twenty-five hours, inaqueous solution it had polymerised in the same period to theextent of 25 per cent. Only two explanations appear to hepossible, both are in accord with the ionic theory of the reaction,and both are probably factors in producing the final result.First,the hot alcohol has probably extracted less catalytic matter fromthe glass than the hot water, and, secondly, that which has beenextracted has generated fewer ions in alcoholic solution fromcyanamide, less cyanamidion, that is, than the same amount inaqueous solution.Polymerisation in the Presence of Bases.In the first series of experiments the accelerating action ofammonia was studied and found to be exceedingly pronounced,even when added in minute quantities.So, for example, in theexperiments given in the first table below, where cyanamide washeated a t looo in N/70-ammonia; the 50 C.C. of solution used con-tained only one drop of concentrated solution of ammonia, but yeta t the end of four hours 20 per cent. of the cyanamide had poly-merised, whereas in pure aqueous solution the amount changed inthat time was inappreciable.When an ammonia solution of ten times the above strength wasemployed, polymerisation was almost complete in one and a-halfhours, and a qualitative test made after three hours entirely failedto give the reaction of cyanamide a t all, showing, therefore, thatno condition of equilibrium is attained, but that the formation ofdicyanodiamide proceeds to completion.( a ) Cyanamide heated in a water-bath a t looo in N/70-ammonia :N / 10-Silver nitrate pre-cipitated by 2 C.C.Time in hours.of solution. K = l(t log, aJa -x.0.0 17.5 C.C.0.5 16.0 ,, 0.078e1.0 14.9 ,, 0.062e1-5 14.0 ,, 0.054e2.5 12.6 :, 0.046e4.5 10.6 ,, 0.038e8.5 8.4 9 9 0.025e( b ) Cyanamide heated in a water-bath a t looo in N/7-ammonia:N 10-Silver nitrate precipitatedby 2 C.C. of solution. Time in hours.0.0 18.30 C.C.0.5 6-75 ,,1.0 3-65 ,,1-6 1.40 ,,3.0 0.00 ,,Ammonia is undoubtedly lost during the prolonged heating in-volved in these experiments. The rate of reaction, therefore584 MORRELL AND RURGEN :diminishes more rapidly than it would if the ammonia content couldbe kept constant throughout, and K , which has been calculatedfor series (a) on the unimolecular formula, eventually sinks t o avalue 0.025e, of the same order as that obtained with pure aqueoussolutions of cyanamide.These results led us to study the actionof a non-volatile alkali, and for this purpose sodium hydroxide wasselected. Cyanamide was dissolved in solutions of the base ofwidely varying strengths, from .ii/800 (1 in 20,000) to 2iV (6 in!12-5), heated a t looc, and determinations made in the usual way,with the following results:( c ) Cyanamide heated in N/800-sodium hydroxide solution (1 in20,000) a t looo in a water-bath:Time in hours.0123456N/lO-Silver nitrate pre-cipitated by 2 C.C.of eolution.28-30 C.C.K = l/t log, ala-25.00 ,, 0.054e22-30 ,, 0.049e18.40 ,, 0.038e16.90 ,, 0.037e15.65 ,, 0.033e20.10 ,, 0-045e-X.( d ) Cyanamide heated in N/400-sodium hydroxide solution (1 in10,000) a t looo in a water-bath:N / 10- Silver nitrateN / 1 O-Silver precipitatedTime in hours.nitrate added. by 2 C.C. K = l/t log, a/a - x.1 18 Y , 7.10 ,, 0-115e2 16 9 , 5.70 ?, 0.095e4.25 ,, 0.053e 4 12 ? f0 20 C.C. 9.25 C.C.3 14 ?? 4.80 ,, 0-076~( e ) Cyanamide heated in N / 70-sodium hydroxide solution (1 in1750) a t looo.(1) Preliminary experiment precipitating throughout with 40 C.C.of N / 10-ammoniacal silver nitrate :Time in minutes.0153060Nt10-Silver nitrate precipitatedby 2 C.C. of solution.19-95 C.C.14.50 ,,10-80 ,,6.30 ?,(2) Confirmatory experiment precipitating with decreasingquantities of silver nitrate proportional t’o the amount of unchangedcyanamide present, as explained in the section on the “Determina-tion of Cyanamide ” THE POLYMERIS ATION OF CYANAMIDE.585N/ 10- SilverTime in minutes. nitrate precipitated. K = l/t log, a/a -x.15 12.8 ,) 0-621e0.593e 30 9-1 9 , 45 6.6 9 9 0.558e60 5.0 ,, 0.482e0 18.3 C.C.( f ) Cyanamide heated in N/4-sodium hydroxide solution a t looo.The solution a t the commencement was approximately normal withrespect to cyanamide.(1) Preliminary experiment as above.Time in minutes. N/lO-Silver nitratme precipitated.0 18-3 C.C.15 8.3 1;30 6.1 1 ,45 4.7 9 )60 3.7 s,(2) Confirmatory experiment as above.N/lO-Silver nitrateTime in minutes.precipitated. K = l/t loge a/a-x.15 6-70 ,, 1.736e30 4-75 ,, 0.603e46 3.70 ,, 0.434e60 3-00 ,, 0.364e0 18-25 C.C.(9) Cyanamide (approximately normal solution) heated in(Approximately the com- N / 2-sodium hydroxide solution a t looo.pound NaHN.CN.)(1) Preliminary experiment.Time in minutes.0163060N/lO-Silver nitrate precipitated.18.7 C.C.13.0 ,,9.3 9,5.2 9 ,(2) Confirmatory experiment.N / 10-SilverTime in minutes. nitrRte precipitated.30 8.5 #,K = l/t loge ala - x.0 18-4 C.O.15 12.1 ,, 0.728e0- 6 14e46 6.1 ,* 0-576e60 4.5 #9 0.528e(h) Cyanamide (approximately normal solution) heated inN-sodium hydroxide solution a t looo586 MORRELL AND BURGEX :(1) Preliminary experiment.Time in minutes.0153060N/lO-Silver nitrate precipitated.18.95 C.C.13-60 ,,10.10 ,,6.20 ,.(2) Confirmatory experiment.N/lO-SilverTime in minutes.nitrate precipitated.15 13-60 ,,K = l/t log, a/a - x.0 18.40 C.C.0.525e30 9.90 ,, 0.551e45 7.25 ,, 0.541e60 5.25 ,, 0.560eFrom the above six series of experiments several important factsare a t once obvious. Using equivalent quantities of ammonia andsodium hydroxide (series a and e), the velocity is vastly greaterin the latter case than in the former. On the ionic theory of thereaction, this follows a t once from the fact that the stronger basegenerates a greater concentration of cyanamidion than the equiva-lent amount of the weaker base ammonia.Approximately normal solutions of cyanamide being taken inevery case, the accelerating influence of the base has increased asits concentration rose from N / S O O to N / 4 , and, moreover, the in-crease in the more dilute solutions has been roughly proportionalt o the amount of base present.This will be seen by a comparisonof the initial velocity-constants :Concentration of sodium hydroxide. Velocity-constant .1 in 20,000 0-054e1 ,, 10,000 0-115e1 ,, 1,760 3.621eThe increase in concentration of cyanamidion in dilute sodiumhydroxide solution will also be approximately proportional to theincrease in concentration of hydroxide, according to the equilibriaH,N*CN + NaOH NaHN-CN + H,OI! Na- + H”*CNThen, with regard to the actual courses of the reactions, it willbe seen that although in the three cases just mentioned “R” onthe unimolecular formula is not constant, but gradually decreases,yet the values approximate much more closely on this formula thanon the bimolecular equation. For a purely unimolecular reactionwe should haveH,N*CN + H’N*CN = C,N,H’,THE POLYMERISATION OF CYANAMIDE.587the ion H’N-CN being assumed of constant concentration. I nactual fact?, hwever, cyanamidion is not present in quite constantconcentration, because although the small amount of sodiumhydroxide is constant, yet as the reaction proceeds more and moreof this base is shared by the dicyanodiamide produced. The ionH’N-CN does not, however, decrease a t anything like the same rateas does cyanamide ; if it did, the reaction should appear bimolecular.Apart altogether from the minor question of differences in thedegree of hydrolysis of sodium cyanamide and sodium dicyano-diamide, it is obvious that since the molecular concentration ofdicyanodiamide is only one-half that of the molecular concentrationof the cyanamide from which it was produced, and since both aredissociated primarily as monobasic acids, a larger proportion ofsodium hydroxide will be available for the generation of cyan-amidion than if each dicyanodiamide molecule had demanded asmuch as did the two molecules of cyanamide of which it is com-posed. Hence thk net result is a reaction the rate of whichdecreases more rapidly than a unimolecular, and more slowly thana bimolecular, reaction, which condition is fulfilled by the figuresobtained.The next feature to be considered is the remarkable fact thatwhen the concentration of sodium hydroxide has been increased toN / 4 , the reaction has its maximum observed velocity, the furtherincrease in concentration producing a retarding effect.At thisconcentration half the cyanamide will have been converted intoits sodium compound, which for simplicity’s sake we may imagineas being completely dissociated into Na’ and NH’*CN. The otherhalf of the cyanamide will remain undissociated, and this is pre-cisely the conditions, according to the theory, for the maximumvelocity, each of the reacting masses a t the initial stage havingequal concentrations,NH,*CN + NH’*CN = C2NkH’3.This, of course, can only momentarily be the condition, for as thereaction proceeds more and more base is set free to react with non-ionised cyanamide, until in the final stages there should be nearlysufficient sodium hydroxide to ionise both the dicyanodiamideformed and the small amount of cyanamide remaining.Thisremoval of undissociated substance corresponds with the rapid fallin velocity shown in series (f).Passing on to the final series, (h), with 8-sodium hydroxide, itwill be seen that the reaction now gives almost constant valuesfor K on the unimolecular formula, and, moreover, that it is analmost exactly similar reaction to series (e), where N / 70-sodiumhydroxide was used. The velocity-constants are almost identica588 MORRELL AND BUHGEN :in value, and if the curves are plotted they run almost exactlyparallel t o one another throughout their length.The conditionshere are much more complicated than when very dilute sodiumhydroxide was used. The solution during the reaction containsever-varying amounts of sodium cyanamide, sodium dicyanodiamide,and free sodium hydroxide, and, as the dilution is not great, thenature and exbent of dissociation and hydrolysis must also be factorsin the case. Comparing with series (e), it is probable that herematters are reversed, cyanamidion present initially in large amountcorresponding with undissociated cyanamide in that experiment,and undissociated sodium cyanamide, or cyanamide produced byhydrolysis, present in very small amount, and by compensatinginfluences held a t fairly constant concentration, corresponding withthe cyanamidion in series (e).Polymerisation in Presence of Acids.An ionic theory appears to explain the accelerating influence ofbases, but it is difficult t o see how the e'qually powerful acceleratingeffect of acids can be brought into harmony with the same theory.The addition of a strong acid to cyanamide solutions would resultin the practical removal of even the small amount of cyanamidionoriginally present.Yet this undissociated substance in the presenceof dilute sulphuric acid rapidly polymerises, when warmed, t odicyanodiamide. Hydrolytic side reactions are, however, in thiscase possible, the more concentrated the acid the greater the amountof carbamide produced (Baumann, Zoc.cit.), and in all circum-stances the dicyanodiamide is also hydrolysed to dicyanodiamidine(compare Lidholm, Ber., 1913, 46, 156). This latter substance isa fairly strong base, so that when only a small quantity of acid wasused the solution became after a time quite neutral, and theaccelerating influence of the free acid on the polymerisation therebyceased. When a larger quantity of sulphuric acid was employed( N / 10-solution) the velocity of polymerisation became so muchgreater than that of the subsequent hydrolysis that a t the end ofhalf an hour the solution was still acid, and gave but the slighestreaction for cyanamide with ammoniacal silver nitrate.Cyanamide heated in N / 70-sulphuric acid a t looo :0 19-00 c c.Time in minutes.N/lO-Silver nitrate precipitated by 2 C.C.15 16.45 ,,30 15-10 ,,60 14.40 ,,120 12-65 ,,A t the expiration of 120 minutes the solution had become neutralto litmus paperTHE POLYMERISATLON OF CYANAMIDE. 589Experiments on the Stability of Cyanamide at the OrdinaryTempera tu(r e.Accurate data as to the keeping powers of cyanamide a t theordinary temperature do not seem to be available in the previousliterature. There is only the statement of Beilstein and Geuther(Zoc. cit.), who assert that a sample kept for several months failedto give the reactions of cyanamide. Experiments were thereforemade with two 10 per cent. solutions, one kept in the dark, andthe other in ordinary diffused daylight a t summer temperatures,and also with a sample of the pure solid substance kept for thecomplete exclusion of moisture in a desiccator.The analyticalresults obtained showed that the admission or exclusion of lighthad no appreciable influence on the polymerisation; that the solidwas more stable than the dissolved substance, but that the stabilityof both was of an order quite a t variance with Beilstein andGeuther’s observation. After keeping for six months, in fact,only 8.9 per cent. of the solid was found to have changed, whilstin 10 per cent. solution nearly 35 per cent. of the dissolved cyan-amide had disappeared. The actual figures were as follows:N/lO-Ammoniacal silver nitrate requiredFor 0.5 C.C. For 0-5 C.C. For 0.1 gramof solution of solution of solid0 41.0 V.C. 40-8 C.C. 47-0 C.C.1 40.1 ,, 39.8 ,, 46.5 ,,6 26.9 ,, 27.0 ,, 42.8 ,,f n \Time in months. in daylight. in the dark. subatanw.The stability of the sodium salt a t the ordinary temperature hasnot been accurately investigated, but it has been noticed duringthe preparation of cyanamide from the commercial sodium cyan-amide, supplied by Kahlbaum, that very different yields have con-sistently been obtained from different samples, 50 grams of thecommercial product yielding in the poorest samples as little as5 grams of cyanamide, and as much as 14 grams from the best.This variation is quite possibly due t o the greater or less facilitiesfor polymerisation which have been allowed during the process ofmanufacture, for since the sodium salt in solution polymerises muchmore rapidly than the free cyanamide itself, any undue heat appliedduring the concentration of the mother liquors must inevitablyresult in an extensive amount of polymerisation.NoTE.-since the above paper was read a communication byGrube and Kriiger (Zeitsch. physikal. Qhem., 1913, 86, 65) hasappeared, which confirms from a somewhat different point of viewsome of the results arrived a t in this investigation.THR SIR JOHN CAS@ TECHNICAL INflTITUTIF,LONDON, E.C