GLADSTONE AND TRIBE ON THE ACTION, ETC. 139 XVIII. An, Inquiry ilzto the Action of the Copper-zinc Couple on Alka- line Ozy-salts. By J. H. GLADSTOKE, Pres. C.S., F.R.S., and ALFRED TRIBE, F.C.S., Lecturer on Chemistry in Dulwich College. PART I. IN 1873 Professor Thorpe established the fact that the copper-zinc couple in presence of nitre and water converts the whole of the nitrogen of the salt into ammonia. Some months prior t80 the publi- cation of his paper it had been noticed by ourselves that the couple quickly reduces an aqueous solution of this salt to nitrite of potassium. This reduction first to nitrite, then to ammonia, is of considerable interest, and appeared to us t o call for a more extended study. A close observation of the course of the chemical change we thought might reveal Dr.Diver’s hyponitrite about which so little is known, and perhaps lead to the real explanation of this and similar reactions, about which nothing definite has yet been ascertained. The couples used in this research were made by adding 100 C.C. of a 2 per cent. solution of copper sulphate twice to every meter of zinc foil 5 centimeters wide. Potassium Nitrate. We give the details of two experiments on the reduction of this salt. The nitrite ar;d ammonia were each estimated daily at about the same Y 2140 GLADSTONE AND TRIBE ON THE ACTION OF TEE ICNOZ. 0 '6'73 0'820 0.778 0.704 0.589 0'347 0.150 Jsil hour ; the ammonia by the Nessler method, the nitrite by potassium permanganate. In the first experiment 4 meters of foil were used, and 500 C.C. of a 1.2 per cent.solution of the salt; in the 2nd, 8 meters of foil and 640 C.C. of a 2.4'7 per cent. solution. The results are given below :- NH,. -- 0 -012 0.025 0.03 0-0425 0'0'72 0'1212 0.184 - Ezpt. I. ~~ Time. 1 day .......... 3 days.. ........ 4 ) , .......... 5 ,, .......... 6 ,, .......... 7 ,, .......... 7 days 12 hours.. 8 ,) .......... 4 hours ........ 1 day .......... days.. ........ 4 ............ 5 ............ 5 days 6 hours . . >> - . * . * * ' * ~~~~~ ~ remperature Centigrade. 15'-17" 1 , 9 ) 7) ), >, 9 9 )S 1'1 13 14 14 ' 5 14 5 15 -5 17 0 '526 1 -182 1,429 1 -587 1 *587 0-758 Nil - 0 -006 0 -022 0 '036 0.0456 0 -07 0.256 0 *40 Equal to KN03 reduced. --- 0.871 gram 1 *lo3 ), 1.132 ,, 1-24 ,, 1-123 ,, 1-09 ), 1.127 , l - 0.66 ,) 1.535' ), 1.912 ,l 2.157 ,, 2.302 )) 2-422 ), 2.376 ,, iVork done each day expressed n rnil1igra.m~ of Hy dr0qen.x 107 % 23 -3 6 .6 19 -5 51 -3 78 *6 215 - 583 '6 241 *7 74 *2 47 62 -5 342 *4 955 '2 During the progress of the experiments, hyponitrite was tested for with silver nitrate twice daily, but in every case with a negative re- sult. The amount of permanganate used up in the estimations during the latter part of the action, jointly with the ammonia formed, points not only to the non-formation of this hyponitrite, but also of any body requiring for its oxidation more permanganate than the nitrite. It starts some- what energetically, then diminishes considerably, again increases, and finally ends more rapidly than it began. Ammonia and its equivalent of potash increase but slowly from the commencement until the time when the maximum amount of nitrite is produced, when this salt rapidly gives way, accompanied of course by an increasing amount of The course of this chemical change is remarkable.* Calculated in accordance with equations :- (a.) E N 0 3 + H? = HyO + KNOa, (b.) KNO, + HB = KHO -t ZHSO + NH,.COPPER-ZINC COUPLE ON ALKALINE OXYLSALTS. 141 the alkalies. This rapid breaking down of the nitrite is coincident with the renewed activity of the couple, and it might therefore be thought that it is a consequence of the smaller stability of the nitrite ; but this can scarcely be the explanation, since we see that almost from the very commencement the couple elects to deoxidise the nitrate, when it has the opportunityof attacking the nitrite.'The question therefore arises, are the variations in the amount of the action, especially the acceleration, due to the ammonia and pot- ash produced P The following series of experiments show the in- fluence of these bodies on the amount of hydrogen set free. The first column represents the hydrogen from a couple immersed in water alone ; the second and third from similar couples immersed in similar volumes of aqueous solutions of NH, and KHO respectively. The strengths of the solutions of ammonia and potassium hydrate in each set of experiments approximate t o the equivalent proportions. The gas measurements were taken after two hours. In Expt. A, the strength of NH, used was 0*026 per cent. ; KHO, In Expt. B, the strength of NH, used was -0256 per cent.; KHO, I n Expt. C, the streugth of NH, used was 0.51 per cent.; RHO, I n Expt. D, the strength of NH, used was 20.0 per cent. ; KHO, 0.087 per cent. 0.846 per cent. 1.69 per cent. 65.8 per cent. Taking the action of the water couple in each set of experiments as unity, the numbers below express the relative quantities of hydrogen obtained :- Water. Ammonia. Potash. A ............ 1 1.8 2-0 B ............ 1 1.75 2-75 c ............ 1 0.91 2.27 D ............ 1 0-36 1-59 from which it appears that the two weakest ammonia solutions aug- ment the action, while the stronger undoubtedly diminish it. The influence of potash is always an accelerating one, but the effect is a t the maximum when the alkali is present in comparatively small quan- tity.These unexpected results were corroborated by the following series of experiments with different strengths of the alkalies. In I, ammonia, 0.21 per cent. was used; in 11, 0.41 per cent. ; and in 111, 0.78 per cent. I n IV, potash, 0.65 per cent. was used; in V, 1.27 per cent. ; and in VI, 2-12 per cent.142 GLADSTONE Ah'D TRIBE ON THE ACTION OF THE i Temperature. Time. 1 Hydrogen in C.C. reduced to 0" and 'I60 mm. ! ___ Ammonia. Potash. 4 hours.. 20 ), .. 68 ,, .. I. 12" 46 11*3* 175 12" 1 544 2 hours .. 14' 75 18 12 -5" 358 66 :: ::I 13" 1 1055 111. - - - 12 133 861 11. IT. v. ---- 55 60 160 175 455 522 36 28 205 210 645 817 --~-- 36 128 442 50 303 888 TI. 45 135 416 23 195 743 -- The ammonia produced in the reaction cannot, it is evident, be the cause of the diminution or of the acceleration already noticed, because where the diminution occurs, the ammonia present would increase the action, and where the increase occurs, it would occasion a de- crease.Potash does augment the production of hydrogen, but this throws no light on the cause of the diminution, and only partially accounts for the great amount of reduction a t t'he end of the reaction : for it appears from our experiments that the maximum amount of hydrogen obtainable per day from couples with water corresponding with those used, amounts to 24 milligrams. And leaving out of con- sideration the influence of the ammonia-which in general is a retard- ing one-the potash produced in the reaction might give 66 milli- grams of hydrogen, whereas thc reduction of the last day in Expt.1 is equal to 107.2 milligrams above that of the first day, whilst in that of Expt. I1 it is 185.8 milligrams for the same quantity of couple. This reduction to nitrite may be shown to an audience as fol- lows :- Pour a solution of nitre-about 10 per cent.-with enough copper sulphate to colour the liquid distinctly, on to some granulated zinc in a tap-funnel ; leave it for a few seconds, and then run off some of the liquid, when the green colour of copper nitrite will be evident. Drop the remainder of the liquid into some starch-solution with potassium iodide and acetic acid : the blue iodide of starch will be instantly formed. F o r this purposeadd to about 5 C.C. of solution 12 drops of copper sulplzate and four or five pieces of zinc foil (1 x 6 inch).Wait about three minutes, then pour the liquid, or a part of it, into about 5 C.C. of starch solu- tion containing a little potassium iodide and acetic acid. A blue coloration forms at once, or in a second or two. By inverting the test- tube containing the test before pouring off, the liquid will be seen to The reaction may also be utilised as a test for nitrates.COPPER-ZIKC COUPLE ON ALKALINE OXT-SALTS. 143 be green. Confirmation may be had by pouring some Nessler reagent into the completely decolorised solution. The reaction with starch can be readily and certainly obtained with 1 part of nitre in 500 of water, and the Nessler reaction with 1 part in 10,0!)0. PART 11. At least three views may be taken of the foregoing change :- I.It may be considered that the zinc, augmented in its activity by contact with the copper, combines with the oxygen of the nitrate. 11. That the zinc and copper electrolyse the water present, and that the nascent hydrogen set free effects the reduction in the vicinity of the negative metal. 111. That the two metals electrolyse the nitrate of potassium, with formation of nitrate of zinc, the reduction being effected a t the nega- tive pole through the agency of the potassium. The first view may be expressed by the following equation:- KNO, j- Zn = ZnO + KNO,, while the further action requires the intervention of water, with the following result :- KNOz + 3Zn + 5H,O = KHO + 3Zn(HQ)2 + NH,. Of course this may be brought about by the formation of interme- The second view may be represented thus- diate compounds.(..) Zn I OH, 1 Cu= ZnO 1 H, ] Cn, ( b . ) KNO, + H, ( c . ) KNOZ + He = HZ0 + NH, + KHO. HZO + KNOZ, And the third view thus- Zn I NO,K I NO,K 1 Cu = Zn(NOJ2 [ K, I Cu. J t may be conceived that this potassium is actually set free and at once acts upoii the water, thus- K, + 2H2Q = 2KH0 + H,; or other combinations of the elements present may be easily conceived to take place with the same final result. The KHQ and Zn(N03)2 produced would of course react, with formation of zinc hydrate and potassium nitrate, so that the sereral changes at each cycle produce a molecule of hydrogen or its equiva- lent, which acts on iiitre in accordance with equations b and c. With reference to the first view, it was found that powdered zinc144 GLADSTONE AND TRIBE ON THE ACTION OF THE and dry nitre, when heated together, detonate explosively, a quantity of gas-probably nitrogen-being produced, together with a solid, consisting of zinc oxide and potassium nitrite and oxide.And Schon- bein and Divers have shown that a solution of nitre is reduced by metals more positive than zinc to potassium nitrite and ammonia ; but in this case it must be borne in mind that hydrogen is simultaneously produced by the decomposition of water. Again, an experiment suggested itself which we hoped would assist in deciding between this and the second and third views. A boxwood cell was cut vertically into two equal parts, some pieces of parchment paper were placed between these, and the divisions of the cell held firmly together by a clamp.A solution of nitre was placed in each of the divisions, a strip of zinc being placed in one, a strip of platinum in the other. The strips were connected together by a metallic wire, and allowed to remain so for two or four days, the action being a feeble one. The general result of several experiments was, a little ammonia in each of the divisions ; free potassium hydrate in the platinurn one, none in the zinc ; and about ten times less nitrite in the platinum than in the zinc division. This great increase oE nitrite in the zinc division would appear to lend material aid to the first view, which requires the reduction to take place by, and in the immediate vicinity of, the zinc plate. At this stage it occurred to us that this reduction might, after all, be due to a couple action, the negative element of such couple being the impuri- ties in the zinc itself; and on trial, we found that similar zinc, when not connected with platinum, reduced nitre to almost the same extent as when metallically associated with that metal as described above.Moreover, we found that equal quantities of granulated redistilled zisc and commercial zinc placed in contact with equal volumes of nitre solution gave in equal times nitrite in the ratio 1 : 2.7. No deduction can therefore be drawn from the presence of the proportionately large amount of nitrite in the zinc division of the cell. That an electric current traverses the nitre solution in the cell ex- periment, from the zinc to the platinum in the liquid, was ascertained by including a galvanometer in the circuit, which, in conjunction with the appearance of potassium hydrate in appreciable amount along with an excess of unaltered nitre in the platinum division, lends, we think, material aid to hypothesis 111. Whether hypothesis I or 111 be the more tenable may not be con- sidered yet decided, but the results exhibited in the following table of a series of comparative experiments on the action of the couple on water, and on solutions of potassium nitrate and nitrite, certainly point to the untenability of 11.COPPER-ZISC COUPLE ON ALKALINE OXY-SALTS.145 Employed for No. 1 KNOr, 500 C.C. of -922 per cent. 9 9 9 , No. 2 7 7 481 7, 9 7 -956 7, 9 , ,, No. 1 KN03, 435 ,, ,, 1.25 ,, 9 , ,, No. 2 ,, 462 ,, ,, 1-18 ,, For comparison the total work done by the couples in the given times is expressed in milligrams of hydrogen. 1.31 -7 106 124 183 222 234 237 296 331 No obs 2. 34 -6 ---- 108 144 193 239 ervation 247 243 369 - KNO,. 28 29.4 45 46 93 87 101 93 117 108 141 132 148 136 165 153 KN03. 24 -4 39 73 84 96 116 120 134 1. ~~ 40 -2 118 200 209 273 293 - - - - 2. 33 -8 -_I 106 198 211 247 272 309 - - - It is evident that the nitre does not simply remain passive, and allow itself to be reduced by the hydrogen evolved from the direct decomposition of water by the couple, which hypothesis I1 requires : for were this the case, the reduction of the salt, a t least at the begin- ning of the action, would be equivalent only to the hydrogen set free from the water couple alone, whereas the oxygen actually removed is equivalent t o about 29 times that amount.It is noticeable that the first numbers given by the nitrite and the nitrate are nearly the same, and that in the respective columns of the salts they do not diverge much from one another for some time, which is of considerable in- terest, as showing that the ammonia and potassium hydrate, which must pour into the solution in the nitrite experiment from its very commencement, do not materially augment the action of the couple. The weight of evidence certainly inclines to hypothesis 111, which is to a great extent confirmed by the following experiments. ElectToZysis of Nitrate of Potassium. I. A V-shaped tube was used, the bend being well plugged with fine asbestos. About 25 C.C. of a 5 per cent.solution of nitre were poured into each of the limbs. An amalgamated zinc plate 1.5 centimeters wide, connected by a wire with the platinum end of 4 Grove’s cells, was immersed to the depth of one decimeter in one limb of the tube, a copper plate, of double the surface connected with the zinc end of the146 GLADSTOKE AND TRIBE ON THE ACTION OF THE battery, being placed in the other. The current was allowed to pass through the liquid for four hours. A trace only of gas escaped from the copper electrode, in the vicinity of which alkali was found imme- diately after making connection. A quantitative examination of the solutions in the respective limbs gave, in addition to the undecornposed 11 i tre- Zinc limb . . . . . . . . 0.3273 gmms Zn(N03):! 0.197 ,, KHO Copper limb .. . . . , 0.0765 ,, KNO, { 0-00375 ,, NH3. The amount of zinc nitrate was ascertained by precipitating the metal as carbonate, and calculating the oxide subsequently obtained as nitrate. Had zinc oxide been produced in this experiment and passed into solution, this method of analysis would give a quantity of nitrate greater than that actually present. The question therefore amrose, does zinc oxide form in the experiment? and, if so, does it dissolve in a solution of nitre op zinc nitrate ? We agitated zinc hydrate with nitre solution, also boiled the substances together. No zinc passed into solution. Small quantities of zinc hydrate were agitated with a solu- tion of zinc nitrate. None disappeared. Schindler also states (GnzeZirz, v, 34) that zinc oxide when boiled with the nitrate does not dissolve. Again, zinc oxide was looked for on the zinc plate, and in the solution in its neighbourhood, during the action in this and subsequent expcri- ments. Not a trace could be seen even when 20 Grove cells were employed.Every NH3, therefore, represents 4 of the ZII(NO,)~ estimated as described, as seen by the equation :- (u.) 4Zn I 8N03K I Cu = 4Zn(NO,), I K, 1 Cu ( b . ) 8K + GH,O + KNO, = 9KHO + NH,, and every KNO, represents one of zinc nitrate, thus :- (c.) Zn I 2N03K I Cu = Zn(PU’03)2 I I(, I Cu (d.) K2 + H,O + KNO3 = KNO2 + 2KHO. Calculating then the ammonia and nitrite found in the copper limb into zinc nitrate, according to these equations, we obtain 0.397 gram of that salt, which is 0.01 in excess of that found in the zinc limb by analysis, and equal to 0.0043 gram of zinc oxide-a quantity which might have formed on the zinc plate and yet have escaped detection.I n other respects it was similar to the last. 11. A 3 per cent. solution was used in this experiment. Found in zinc limb.. . , . .. . 0.2596 gram !&I(NO~)~ 0.1366 ,, KHO KNO, 0.00294 ,, NH,, Copper limb . . . . . . . . . . . . 0.05355 ,,COPPER-ZINC COUPLE ON ALEALINE OXY-SALTS. 147 Nitrate. Chlorate. 1st hour .... 1 : 1-15 3rd ,, . .. . 1 : 2.50 the two latter of which products give 0.2501 gram of Zn(N03)2, which is slightly less Bhan that found. 111. A 3 per cent. solution was used, the current passed for three hours. In other respects similar to I and 11. Found in zinc limb . .. .. . 0.1645 gram Zn(NO& 0.0542 ,, RHO Found in copper limb . . . . 0.03706 ,, KNO, { 0.00157 ,, NH3, the latter of which are equal to 0.1524 gram of Zn(N03)2, which is again below that actually found by analysis. We may therefore con- clude, that in these experiments nitre alone suffered electrolytic de- composition. One other point is worthy of notice here. I n the experiments just described not a trace of nitrite or ammonia could be detected in the zinc limb, which we t.hink finally disposes of hypothesis I, because, if the zinc electrode, when charged with positive electricity, fails to reduce nitre in its immediate vicinity, the zinc of the couple so charged might reasonably be expected not to act. Of course we do not contend that in these electrolytic experiments we have identically the same conditions as obtain in the couple decompositions, but that we have a similarity of condition no one can doubt.Nitrate. Clilorate. 5th hour .. .. 1 : 3.86 21st ,, . . . . 1 : 2.97148 GLADSTONE AND TRIBE ON THE ACTION OF THE electrolysed by an external battery, alkali may a t once be detected close to the negative electrode, but instead of the potassium chlorate being reduced to chloride, the greater part of the hydrogen escapes a.s gas. Employing the arrangement as used for the nitre solution, a current. of four cells gave in four hours 0.00244 gram of potassium chloride in the copper limb, and 0.4775 of zinc chlorate in the zinc limb. Calculating the equivalent of the potassium chloride found according to the equa- tions- ( U .) 3zn I 6C103K I C U = 3Zn(C10,)2 I K6 I CU. (b.) &, + 3820 + KCLO, = GKHO + KC1. we get only 0.0228 of zinc chlorate. Therefore 95.2 per cent. of the hydrogen formed in the experiment passed through the liquid. About four years ago it was pointed out by one of us (Chem. Soc. Jour., 1874, p. 415) that finely divided copper, immersed in acidulated water, agglomerated-that is, formed into more or less coherent lumps- when subjected to the action of nascent hydrogen : also that the finely divided particles of palladium and platinum-metals known to con- dense hydrogen, agglomerated, when similarly treated ; further, that the agglomerated copper, palladium, &c., deglomerated when treated with nascent oxygen:-from all which it was inferred that the copper of the couple was also capable in a slight degree of absorbing hydrogen gas.Graham has shown that occluded hydrogen is a somewhat powerful reducing agent ; and to us, having traced the reducing action of the couple in some way to hydrogen, it appeared of exceptional interest to ascertain whether hydrogen associated with the finely divided copper of the couple could reduce nitre t o nitrite. With this object some copper was precipitated by immersion of a zinc plate in a 2 per cent. solution of copper sulphate till decolorised. As has been already pointed out, the deposit thus obtained contains metallic zinc. To remove this and to charge the residual metal with hydrogen, dilute sulphuric acid was added, which immediately brought about a power- ful agglomeration of the copper.After standing with the acid for about an hour, with occasional shaking, the metal was well washed, and some nitre solution added, when almost immediately the whole deglomerated, shrinking in volume to about a quarter. The solu- tion contained a small quantity of both ammonia and potassium nitrite, which was found equal per 100 grams of copper to 4 milligrams of hydrogen. For the next trial copper was deposited as before, but from a 1 per cent. solution of the sulphate. The deposit was treated with succes- sive portions of copper sulphate for some hours. The residual metal was then divided into two portions. To one, some nitre solution was added; to the other some dilute sulphuric acid. This after shakingCOPPER-ZINC COUPLE OX' ALKALINE OXT-SALTS.149 was poured into a dish, at the bottom of which was a sheet of platinum in connection with the zinc end of 4 Grove's cells. As soon as the metal had settled, the positive electrode was dipped into the acidulated water near its surface. The copper slowly agglomerated without ma- terial change in colonr, and in about 30 minutes hydrogen was freely escaping from it. After washing this a few times, and neutralising with potassium hydrate, a quantity of nitre was added equal to that added to the first portion. Ammonia and nitrite were found in both. The portion not treated with hydrogen gave, per 100 grams of copper, 3.5 milligrams of hydrogen, the agglomerated giving, for a similar quantity of copper, 8 milligrams of hydrogen. Zinc was also found in small quantity in both, but the non-agglomerated contained at least four times the amount of that present in the hydrogenised. Another effort was made to get the finely dividedcopper free from zinc. Some deposit, obtained as in the last experiment, was digested, with occasional shaking, for three days with about 5 per cent.solution of copper sul- phate. It was then well washed and divided into two parts. To one some nitre was added, The other, previously to the addition of an equal quantity of this salt, was digested with dilute sulphuric acid for am hour, washed, again mixed with dilute acid, hydrogenised for two hours, and the adhering acid neutralised with potash. The non- hFdrogenised contained nitrite and ammonia, equal, per 100 grams of copper, to 2.7 milligrams of hydrogen.The hydrogenised portion, for the same amount of copper, contained 19.3 milligrams of hydrogen. Both portions still contained a small quantity of zinc-whether as nietal or oxide, or both, there is no means of determining. It is certain that some oxide existed in the portions not treated with acid, and some zinc may have existed as metal, but completely protected by a covering of copper. The reduction effected by the non-hydrogenised portions may also be due to hydrogen occluded during their preparation, but we cannot speak with certainty about it, in consequence of the possible presence of metallic zinc. There can be no doubt, however, that the finely divided copper of our couple does condense hydrogen, and when in this condition reduces nihe to nitrite and ammonia.Two facts which appeared difficult to reconcile, now appear intelli- gible enough: the one, that the couple reduces the chlorate in the cold without the least escape of hydrogen; the other, that in the ordinary electrolysis of the chlorate, nearly the whole of this gas escapes without reducing the salt. The reason is obvious. I n the first case the hydrogen is probably wholly occluded the moment it is set free, while in the other, only a small quantity of the gas is con- densed by the negative plate. Taking into consideration all the facts brought out by this inquiry, we consider it proved that-150 GLADSTONE: BND TRIBE ON THE ACTION, ETC. Time. NH3. 1 NH,N02. 1 Equil:lcnt PITHdKO3. a. The action of the copper-zinc couple on these oxy-salts is of an electrolytic nature.b. The negative radicle combines with the zinc, whilst the positive radicle, or its equivalent of hydrogen from decomposed water, is set free against the copper crystals. c . The reduction and hydrogenisation of the salt take place in the immediate vicinity of the negative metal. We also think it probable that hydrogen is actually set free against the copper, but is condensed by the finely divided metal, and in that condition does its work of reduction and hydrogenisation. Of course the resulting zinc compouiid and the alkaline hydrate decompose one another, producing the original salt and zinc hydrate. Work done each hour expressed in milligrams of IIy drogen. It may be assumed that this action of the couple is a general one, true not only of nitre, but of all nitrates containing metals which de- compose water a t the ordinary temperature.Animonium nitrate, though not strictly belonging to this class of bodies, should, according to what is known of the electrolysis of ammonium salts, and to the views just enunciated, give off a fourth of the hydrogen in its positive radicle when its solution is subjected to the action of the couple ; and since it is very probable that the real reducer in these actions is hydrogen, it naturally occurred to us that this salt ought not to form an exception to the above generalisation. But t o place the matter beyond doubt, we instituted a few additional experiments. A qualitative experiment with the couple showed at once that both nitrite and ammonia were produced, but not hyponitrite. We give the details of a quantitative trial:-410 C.C. of about 1.2 per cent. solution of ammonium nitrate were added to the usual quan- tity of couple (temp. 15" C.), and the ammonia and nitrate estimated. Subjoined are the results reckoned for 100 parts of solution :- 1 hour. ......... 4 hours ........ 11 . . . . . . . . . . 23 . . . . . . . . . . 0 -039 0 9209 0 -4445 0.114 1 0.185 0 -767 0 -179 0-159 1.04 I 0'2443 1 Xi1 1 -143 I---- -------- 100 45 * 5 16.4 7.8 The action of heat on a solution of ammonium nitrite resolves it, as is well known, into nitrogen and water, from which it was inferredWRIGHT AXD LUFF OS THE ALKALOIDS, ETC. 151 that were this reduction attempted a t or near the boiling point of the solution of the nitrate, the nitrite as quickly as produced would be decomposed in like manner. On trial, gas was evolved, but this proved to be mainly nitric oxide, which gas Thorpe had noticed on boiling solution of ammonium nitrate with the couple (Chenz,. Xoc. JournaZ, 187% p. 544). The amoitnt of this oxide of nitrogen we found in- creased with the strength of the nitrate solution. Thus with strengths of 20, 10 and 5 per cent. and an excess of couple, the nit,ric oxide was 8.5, 6 and 3 respectively. I n the cold the nitrate, even in solution of 20 per cent., was completely reduced to ammonia in about 24 hours without the escape of nitrogen free or combined. We might suggest therefore that, in estimating unlinown nibrates by T horpe’s process, it would be well to allow the couple to remain in contact with the nitrate solution for a t least 24 hours prior to distillation.