TEE RELATlVE STRENGTHS OF THE ALKALIKE HYDROXIDES. 121 XIV.- The Reltetive Strengths of the Alkaline Hydr- oxides and of Ammonia as &Ieasuwd by theilr Action on Cotarnine. By JAMES JOHNSTON DOBBIE, M.A., D.Sc., ALEXANDER LAUDER, B.Sc, and CHARLES KENNETH TINKLER, Research Student of the University of Edinburgh. IN a paper recently communicated to the Society (Trans., 1903, 83, 598), it was shown that the spectra of cotmarnine in ethereal or chloro- form solution, of cyanohydrocotarnine and of ethoxyhydrocotarnine are identical or nearly so with those of hydrocotarnine and its salts, whilst the spectra of dilute aqueous or alcoholic golutions are identical with those of the cotarnine salts. From this, it was argued that the substances in the first group are all constituted like hydrocotarnine, those in the second being like the cotarnine salts. This con- clusion is supported by the fact that, whilst the substances in the first group are colourless, those in the second are yellow.It was further shown that an ethereal or chloroform solution of cotarnine becomes yellow when treated with alcohol and gives absorption spectra which approach more and more nearly to those of the cotarnine salts as the quantity of alcohol is increased. On the other hand, when an aqueous solution is treated with sodium hydroxide or other soluble base, the reverse change takes place and the yellow colour gradually disappears ; the proportion of cotarnine which undergoes this change depending on the amount of basic hydroxide present. It was suggested that a fuller study of these reactions might throw light on the con- ditions of isomeric and tautomeric changes generally, and that it might be possible to use the action of the bases on cotarnine as a means of comparing their strengths.I n the present paper, we propose to give the results of some preliminary experiments on the latter subjeci;. I n our former paper, an account was given of the absorption curves of the two forms of cotarnine (Figs. I and 11) : and photographs of the spectra mere reproduced in Plates I and I1 (carbinol and ammonium forms respectively), which accompany the paper. Reference to these photographs will &ow thst the spectra of the two forms are highly characteristic. The change which cotarnine in aqueous solution uudergoes when122 DOBBIE, LAUDER, AND TINKLER : RELATIVE STRENGTHS acted on by sodium hydroxide depends, as already stated, on the quantity of this reagent present.After the addition of any given quantity of the hydroxide, a state of equilibrium between the two forms of cotarnine is established almost instantaneously and, provided that the temperature is kept constant, no further change takes place, a t any rate in moderately dilute solutions, even after a lapse of several hours. Each additional quantity of sodium hydroxide causes a further change until the ammonium form is all converted into the carbinol form. This is practically the case when a milligram-molecule of cotarnine is dissolved in one litre of a normal solution of sodium hydroxide. By photographing the spectra of the solution after each addition of the alkali hydroxide, the change from the one form to the other can be followed through all its phases.Figs. 17-24 of Plate 111, accom- panying the paper already quoted, show generally how the spectra of the ammonium form change into those of the carbinol form as the quantity of sodium hydroxide is increased. The spectra thus pbtained are in fact the spectra of mixtures of the two forms and can all be reproduced exactly by mixing together hydrocotarnine hydrochloride and cotarnine hydrochloride in the proper proportions. By preparing a series of such mixtures containing the two substances in known pro- portions, it is possible, from a comparison of their spectra with the spectra of an aqueous solution of cotarnine, which has been acted on by sodium hydroxide, to determine the amount of the ammonium form which has been changed into the oarbinol form.By using other soluble bases in place of sodium hydroxide, the data are obtained which are required for a comparison of their strengths. As already stated, the action of bases on cotarnine is so rapid t h a t no comparison of their strengths, founded on the amount of change produced by equimolecdar quantities in a given time, is possible. Temperature has a consider- able influence on the action, and consistent results can only be obtained when the experiments are carried out at the same tem- perature. * It mill be seen from the tabulated comparison of the results obtained a t 1 2 O and 1 8 O respectively (Table I) that the effect of heating is to diminish the action of the alkalis, a given quantity a t the higher temperature producing a smaller effect than the same quantity a t the lower temperature.The results are represented graphically in the accompanying curve (p. 123). * The results given in our previous paper were obtained a t the ordinary tempera- ture, without precautions being taken to ensure that the temperature was always the same. They therefore differ somewhat from those given in the present paper.123 0 Percentaye of the ammonium form of cotarni.ne converled into the carbinol form. TABLE I. Strength of Percentage of carbinol form produced. 94 93 89 87.5 N/20 82.5 80 N/50 70 65 sodium hydroxide. At 129 At 18". N/4 N/10 Nf 100 52.5 50124 DOBBIE, LAUDER, AND TINRLER : RELATIVE STRENGTHS I n Table IT, the results obtained at 12' by the action of equi- molecular solutions of the hydroxides of the alkali and alkaline earth metals and ammonia are given, and t h e results are represented graphically in the accompanying curve.The percentages give the amount of ammonium form which is converted into the carbinol form. 0 Perccntngc of the amnzonizcm form of cotarnine converted into the cnrhinol formOF TEE ALKALINE HYDROXIDES AND OF AMMONIA. 125 Streiigth of base. N A?/ 3 AT/ 4 N/2 ;j; n;/8 A y o N/14 N/16 A720 N/25 N / 3 2 AT/5 0 AT/70 N/100 N / Z O O ivf400 N/600 KO H. ~-___ 98.5 97.5 96'5 95 93 92.5 91 89 88 83 7 7 5 67 *5 62'5 52-5 42 *5 25 20 - - LiOH. - -- 98.5 97'5 96'5 94 93 92.5 90 88 85 82.5 77.5 70 65 57 -5 42 '5 30 20 __ I TABLE 11.NaO If. 98.5 97 95.5 94 93 92.5 91 89 87.5 82.5 80 77.5 70 60 52 5 40 25 20 - The hydroxides of lithium, sodium, and potassium give nearly identical curve$, showing that, so far as this reaction is concerned, they all have nearly the same strength. It is not possible t o get a sufficient quantity of calcium or barium hydroxide into solution in one litre of water t o convert a milligram-molecule of cotarnine entirely into the carbinol form, but that portion of the curve which can be drawn for these bases approximates closely to the corresponding portion of the curve f o r the hydroxides of the alkali metals, the slight difference between the curves indicating that the hydroxides of the alkali metals are the stronger. Some experiments mere made with thallium hydroxide, but, as solutions of this substance themselves possess considerable absorptive power, trustworthy results could only be obtained with dilute solutions.These data indicate that thallium hydroxide is a weaker base than the alkali hydroxides, but very much stronger than ammonia. A very concentrated solution of ammonia (more than nine times the strength of a normal solution) is required to convert the ammonium form of cotarnine entirely into the carbinol form. We propose to discuss the case of ammonia more fully in con- nection with investigations on the alkylated ammonias. The general result of our experiments is to show that in their action on cotarnine, the bases have the same relative strengths as in other reactions. The curves representing the relation between the amount126 DOBBIE, LAUDER, AND TINRLER : RELATIVE STRENGTHS of change and the quantity of base producing it are apparently hyper- bolic, but we have so far been unable to find an equation for them. The action of sodium hydroxide on an aqueous solution of cotarnine is capable of explanation in terms of the dissociation theory of electro- lytes.Cotarnine in dilute aqueous solution is a strong electrolyte (Hantzsch and Kalb, Ber., 1899, 23, 3109). It may, therefore, be assumed that the solution contains a mixture in equilibrium of the undissociated ammonium form together with the hydroxyl and other ion resulting from its dissociation, with practically none of the carbinol form. By the addition of sodium hydroxide the active mass of the hydroxyl ions is increased and the dissociation of the ammonium form is diminished. The ammonium form then passes into the carbinol form, in which dissociation is at a minimum, until the equilibrium is restored.The further addition of sodium hydroxide leads t o a repetition of these changes until, when the solution is normal, the ammonium form is practically all converted into the carbinol form. A very small quantity of sodium hydroxide produces no appreciable effect on an aqueous solution of cotarnine, but we have not found it possible to determine with accuracy the exact point a t which the solutions are isohydric. Experimental Details. The determinations of the percentages of the ammonium form of cotarnine converted into the carbinol form, which are given in this paper, were made by comparing photographs of the absorption spectra * of cotarnine in solution in sodium hydroxide, or other soluble base, with standard photographs of mixtures containing hydrocotarnine hydrochloride and cotarnine hydrochloride in known proportions.I n the preparation of the standard series of mixtutes used for this purpose, a solution containing 1 milligram-molecule of hydroco tarnine hydrochloride dissolved in one litre of water mas taken as the starting point. This was mixed with a solution of cotitrnine hydrochloride of the same equivalent strength in the proportions required to give a solution containing 97.5 per cent. of hydrocotarnine hydrochloride and * The method af photographing absorption spectra which we employ has already been sufficiently explained in former papers, see especially Trans., 1889, 55, 649, and the papers by Hartley therein cited.The reference lines employed were those given by an alloy of cadmium, tin, and lead. The photographs of the spectra, from which the data discussed in this paper have been derived, have not been re- produced as illustrations on account of the difficulty of obtaining reproductions sufficiently delicate to show the details of the spectra on which their accurate comparison is based. So far as photographs are necessary for the elucidation of the paper, this purpose is served by those already published (Dobbie, Laudcr, and Tin kler, Eoc. eit. ).OF THE ALKALINE HI~DROXIDES AND OF AMMONIA. 127 2.5 per cent, of cotarnine hydrochloride. The proportion of cotarnine bydrochloride to hydroco tarnine hydrochloride was increased by 2.5 per cent, a t a time until a solution containing 97.5 per cent.of the former and 2.5 per cent. of the latter was reached. The absorption spectra of these mixtures form a graduated series between the spectra characteristic of the two forms of cotarnine. The Reries of standard photographs having been prepared, i t is easy to estimate the amount of change produced in an aqueous solution of cotarnine by any given quantity of a soluble base, by photographing the spectra of the solution and comparing tho photographs with those of the reference series. If the photograph to be determined does not coincide exactly with any in the reference series, its position between two successive mixtures can always be found. A number of intermediate mixtures are then photographed until a photograph is obtained which exactly corresponds with the one under examination, the exact composition of the mixture in question being thus determined.In preparing the series of reference photographs, hydrocotarnine hydrochloride wds preferred t o the cyanohydrocotarnine formerly employed, on account of its greater stability and because of its solubility in water. These two substances have the same spectra. The spectra of mixtures of hydrocotarnine hydrochloride with cotarnine hydrochloride are identical with those obtained by passing the light through an equal thickness of the two solutions contained in separate cells. Thus the spectrum obtained by passing the light through a 20 mm.layer of a mixture of hydrocotarnine hydrochloride and cotarnine hydrochloride solutions, in equal proportions, is identical with that obtained by passing i t in succession through two cells each containing a layer 10 mm. thick of one of the solutions. The same result is obtained by using a 10 mm. cell filled with a n aqueous solution of cotarnine and another cell of equal thickness containing a concentrated sodium hydroxide solution of cotarnine. These results were confirmed by numerous experiments, in which layers of various thicknesses mere employed. In these experiments, the thickness of quartz employed was the same; when this precaution was not observed, the accuracy of the comparison was interfered with by the slight absorption due to the quartz plates. We hope to render the foregoing method more accurate by a modifi- cation of our apparatus, and by using hydrastinine in place of cotarnine. With cotarnine, i t is quite easy, in some parts of the series of mixtures of the two forms, to detect differences of 0.5 per cent. with certainty, but in mixtures containing a large proportion of the ammonium form it is not possible to detect with accuracy, differences of less than 2.5-5.0 per cent, The spectra of the ammonium form hydrastinine are more complicated than those of the corresponding128 PERKIN AND THQRPE: form of cotarnine, and admit of much smaller differences being readily distinguished. We have to express our thanks to Professor Crum Brown and to the authorities of the University of Edinburgh for kindly affording us facilities for conducting part of this work i n the University laboratory; and t o Professor W. N. Hartley for the use of the apparatus with which the experiments were carried out. MUSEUM OF SCIEKCE BSD AILT, ED INBU iw H