NOTICES OF PAPERS CONTAINED IN THE ‘‘ Philosophical Translations” and Foreign Journals. BY HENRYVI.T~~~~, B.A. F.C.S. On the Dimsion ef Liquids.* By Professor Graham F.R.S. F.C.S. The experiments described in the author’s former paper on this subject furnished strong grounds for believing that isomorphous salts possess a similar diffusibility. All the salts of potash and ammonia which were compared appeared to be equi-diffusive; so also were the salts of certain rnagnesian bases. A single preliminary observation on the nitrates of lead and barpta however opposed the general conclu- sion and demanded further inquiry. It is scarcely necessary to say that any new means of recognising the existence of the isomorphous relation between different substances must prove highly valuable.Let us inquire therefore how far liquid diffusion is available for that purpose. The salts were still diffused froin weak solutions that is from solu- tions containing from 1. to 8 per cent of salt; but now a measure of the solution equal to 100 grs. of water was made to contain 1 grain of the salt to form what is called the 1 per cent solution instead of 1 grain of salt being added to 100 grs. of water as before without reference to the condensation which generally occurs. The quantities 1 2 4 and 8 per cent thus indicate the parts of salt present in a constant volume of liquid,-as 10 20 40 and 80 grs. of the salt in 1000 water grain-measures of the solution. The same phials for the solution and jars for the external water-atmosphere continued to be used and the nianipulations were similar.It is believed however that the temperature of the liquids was maintained more uniform in the new experiments than the old partly by the better regulation of the temperature of the apartment and partly by placing the jars close together upon a table with upright ledges and covering the whole * Philosophical Transactions 1850 p. 805. G2 PROFESSOB GRAHAM ON THE over with sheets of paper during the continuance of an experiment. The mass of fluid in 80 or 100 jars which were employed at once and placed together made the small oscillations of temperature which might still occur slow and less injurious. The investigation is also extended to sever$ new substances such as hydrocyanic acid acetic acid sulphurous acid alcohol ammonia and salts of organic bases without reference to isomorphous relations.It is very necessary to have data which are minute and accurate respecting the diffusion of a considerable variety of substances. This it is the object of the present investigation to endeavour to supply leaving speculative deductions in general respecting the nature and laws of liquid diffusion for a future occasion. The density of all the solutions was observed at a constant tempera- ture namely 60' Fahr. 1. Hydrochloric acid-The period of diffusion arbitrarily chosen for this acid was five days. The di5sate or quantity of acid diffused was determined by precipitating the liquid of the external reservoirs with nitrate of silver and weighing the chloride of silver formed.In the 1and 2 per cent solutions the liquids of two jars were generally mixed and precipitated together. The number of cells diffused at once unless otherwise specified was always eight cells of the 1 and 2 per cent solutions and four cells of the 4 and 8 per cent solutions. In this abstract the means only of these experiments are given. The diffusates at the same temperature were found to be as follow Diffusion in five days at 51' Fahr.; two cells. Grs. Ratio. From 1per cent solution . . 7.41 0.97 From 2 per cent solution . . . 15.04 2.00 From 4 per cent solution . . . 30.72 4-08 From 8 per cent solution . . . 67-68 9.00 The 2 per cent solution is taken as the standard of comparison for the ratios instead of the 1 per cent solution from the greater accu- racy with which the diffusion of the former can be observed.The increasing diffusibility with the larger proportions of acid here observed is unusual at least in the degree exhibited by the 8 per cent solution. Other substances as will be immediately observed of nitric acid appear to lose proportionally in diffusibility as their solutions are concentrated. Hydrochloric acid belongs to the most diffusive class of substances known; it appears to exceed hydrate of potash at 53'05 as 7.56 to 6.12,or as 100 to 80.9*. * Phil. Trans. 1850 39. DIFFUSION OF LIQUIDS. The rapidity with which hydrochloric acid diffuses and the facility with which that substance may be estimated induced the author to examine the progression with which its diffusion takes place with increasing times in a minute manner.The 2 per cent solution was diffused for times increasing by six hours from twelve hours or 0.5 day to 4-75 days six cells being diffused for every period. Instead of determining the acid diffused separately in each jar or pair of jars the contents of the six jars of each experiment were mixed together and a definite proportion of the liquid precipitated by nitrate of silver and the chloride of silver weighed so as to obtain at once the mean result. Another observation for 5.75 days is added although made at a sen-sibly higher temperature. DIFFUSION OF HYDROCHLORIC ACID 2 PER CENT SOLUTION; ONE CELL.Time. Temperature. Diffusate in grains. Differences. days. 0-5 5i.75 0.909 0.75 53.75 1.312 *403 1 53.75 1 766 *454 1-25 53.75 2.353 *587 1.5 53-75 2.596 -243 I*75 53.58 3.178 -582 2 53.58 3.410 *232 2.25 53.42 3.967 *557 2.5 53.58 4.339 *372 2-75 53.50 4.618 0279 3 53-50 4.969 *351 3-25 53.50 5.304 -335 3.5 54-85 5.857 *553 3.75 54.85 6.254 -397 4 54.85 6.407 *153 4*25 54.85 6.795 -388 4.5 54.71 7.034 *239 4.75 5471 7.473 *339 5-75 5 6-46 8.363 The differences are evidently affected by accidental errors of obser- vation. The diffusion in 3.5 days is also increased by a rise of tempera-1' more than ofture in that and the following experiments. The diffmion always increases with the time but less rapidly according to a gradually diminishing progression.Hydriodic acid.-Time of diffusion five days as for hydrochloric acid. The acid diffused was estimated from the weight of iodide of silver which it gave when precipitated by nitrate of silver. PROFESSOR GRAHAM ON THE Diffusion from 2 per cent solutions at 51' Fahr. Hydrochloric acid . . 15.04 100 Hydriodic acid . . . 15-11 100.46 These experiments indicate a similarity of diffusion between the two isomorp hous substances h y droch lor1 c and h y driodic acids. Hydrobromic acid.-Time of diffusion five days. The diffusate was estimated from the bromide of silver. Diffusate from 2per cent solutions at 59O.7 Fahr. Hydrochloric acid . . 16.55 100 Hydrobromic acid .. . 16-58 100-18 Hydrobromic acid appears therefore to coincide in diffusibility with hydrochloric acid at this temperature. It mq be remarked that these three acids hydrochloric hydrobomic and hydriodic do not exhibit the same correspondence in another physical property namely the densities of their aqueous solutions containing the same proportion of acid. The densities of 2 per cent solutions of hydrochloric and hydriodic acids appear to be respectively 1*0104and 1.0143 at 6O0 Fahr. and that of hydrobromic acid is an intermediate number. The same acids are also known to differ considerably in the boil- ing-points of solutions containing the same proportion of acid. A considerable diversity of physical properties appears here to be compatible with equal diffusibility in substances which are isomor- phous.Bromine.--Pure water readily dissolves more than 1 per cent of this substance. The solution prepared however contained only 0.864 per cent of bromine as was ascertained by treating it with sulphurous acid and afterwards precipitating by nitrate of silver. Its density was 1.0070. It was evident from the slow appearance of the brown colour in the exterior cell that bromine diffuses less rapidly than hydro- bromic acid. The diffusion-time of bromine was made ten days or double the time of hydrobrornic acid. Two cells contained together a diffusate of 5.80 QFS. of bromine ; another two cells a diffusate of 5.88 grs. ;mean 5.84 grs. at 60'01 Fahr.; or 6.76 grs.for a 1 per cent solution. Doubling the last result we have 13.52 grs. for a 2 per cent solution which is still considerably under the diffusate of hydrobromic acid (16.58 grs.) in half the time. 3. Hydrocyanic acid-Time of diffusion five days. The acid diffused was estimated from the cyanide of silver which it gave with nitrate of silver. Hydrocyanic acid 1.766 per cent made up to a density of 1.0142 with sulphate of potash. Diffused at 64O.2 in six cells 11.40 11.86 11.80 ; niean 11.68 grs. for two cells. Calculated for 2 per cent 13.23 grs. at 64@*2in two cells or about 13.10 grs. at 62O.8 DIFFUSION OF LTQUICS. assuming this acid to be affected in the same way by temperature as hydrochloric acid. Hydrocyanic acid here appears less diffusive than hydrochloric acid at the same temperature 62O*8,as 13.10 to 16.40 or as 79.6 to 100 and not to belong therefore to the same class of diffusive sub- stances.4. Nitric acid.-Time of diffusion five days. The quantity of this acid diffused was always determined with great exactness by neutraliza- tion by means of a normal solution of carbonate of soda. The diffusion of the different proportions of this acid at one tem- perature is as follows Diffusion of nitrate of water in five days at 51O.2 ;two cells. Grs. Ratio. From 1per cent solution . . 6-99 0.95 From 2 per cent solution . . . 14.74 2 From 4 per cent solution . . . 28-76 3-90 From 8 per cent solution . 57-92 7.86 The usual approach to equality of diffusion between chlorides and nitrates is observable in hydrochloric and nitric acids at least in the f and 2 per cent solutions.Diffusion from 1 per cent solution at 53O.5. Hydrochloricacid . . . . 7.56 100 Nitrate of water . . . 7.28 96.3 Diffusion from 2 per cent solution. Hydrochloric acid at 51' . 15.04 100 Nitrate of water at 51'-2 . 14-74! 98.0 The 2 per cent solutions of both acids were also diffused at higher temperatures. Diffusion from 2 per cent solution. Hydrochloric acid at 62O.8 . 16.46 100 Nitrate of water at 63O.2 . . 16.76 101.8 Here the diffusibility of the two acids is as nearly as possible equal. Diffusion from 4 per cent solution. Hydrochloric acid at 51' . 30.72 100 Nitrate of water at 51°*2 . 28.76 93.7 Diffusion from 8 per cent solution.Hydrochloric acid at 51' . . 67.68 100 Nitrate of water at 51O.2 . 57.92 85.3 The Pride divergence between these two acids in the 8 per cent PROFESSOR GRAHAM ON THE solution is produced by the remarkably increased diffusion of hydro-chloric acid in that high proportion. 5. Sulphuric acid.-That time of diffusion arbitrarily chosen for this acid was ten days. The diffusate of this acid wits determined in the same manner as that of nitric acid. The diffusion of the different proportions of sulphuric acid is as follows :-Diffusion of sulphate of water in ten days at 49O.7; two cells. Grs. Ratio. From 1 per cent solution . . . 8.69 1.03 From 2 per cent solution . . . 16.91 2 From 4 per cent solution .. . 33.89 401 Froni.8 per cent solution . . . 68-96 8.16 The diffusibility of different strengths of this acid appears to be pretty uniform but with a slight tendency to increase in the higher proportions like hydrochloric acid. Sulphuric acid is inferior in velocity of diffusion to hydro- chloric acid but still appears to possess considerably more than half the diffusibility of the latter. 6. Chromic acid.-Time of diffusion ten days. The diffusates from four cells of the 2 per cent solution were mixed together and the quantity of chromic acid diffused for two cells reduced by means of hydrochloric acid and alcohol and weighed as oxide of chro-mium. 1.762 per cent of anhydrous chromic acid density 1*01404 dif-fused at 67'-3 gave 19.78grs.of chromic acid in two cells. Calculated for 2 per cent 22-43grs. of chromic acid in two cells at 67O.3. The diffusion of sulphuric acid at 63'05 was 19.73 grs. which would give about 21 grs. of that acid at 67O.3. 7. Acetic acid.-Time of diffusion ten days. This acid cannot be determined accurately by the acidimetiical method owiag to the ace- tates of potash and soda being essentially alkaline to test paper like the carbonates of the same bases although neutral in composition. The weight of carbonate of baryta dissolved by the acid was had re- course to. Diffusion of acetate of water in ten days at 48O.8 ; two ce11s. Grs. Ratio. From 2 per cent solution . . . 11.31 2 From 4 per cent solution . . . 22.02 3.83 From 8 per cent solution .. . 41.80 7.26 The diffusibility diminishes with the larger proportions of acid. This acid appears to be considerably less diffusive than sulphuric acid. I was led to over-estimate the diffusion of acetic acid in a prcliminary DIFFUSION OF L?QVIDS. observation of my former paper by trusting to the acidimetrical method of determination. Hydrochloric acid appears to diffuse about two and a half times more rapidly than acetate of water at the same temperature. 8. Subhurous acid.-The time of diffusion chosen for this acid was ten days for comparison with sulphuric acid. The usual number of eight cells of the 1 and 2 per cent solutions were Wused and four cells of the 4 and 8 per cent solutions The whole diffusates of each proportion were then mixed together and the proportional quantity of liquid representing two cells in the 1 and 2 per cent solutions and 1 cell in the 4 and 8 was converted into sulphuric acid by a slight excess of bromine and determined from the sulphate of baryta.Diffusion of sulphurous acid in ten days at 68O.1; two cells. Grs. Ratio. From 1 per cent solution . . . 8.09 0.954 From 2 per cent solution . . . 16.96 2 From 4 per cent solution . . . 33.00 3.821 From 8 per cent solution . . . 66.38 7.827 This substance appears to be less diffusive than sulphuric acid at the same temperature; the diffusion of sulphurous acid at 68O.1 con-siderably resembles that of sulphuric acid at 49O.7. 9. Ammonia.-The time of diffusion chosen was 4.041 days or that of hydrate of potash with chloride of sodium at seven days.The whole difisates of each proportion were mixed together and the quan- tity of ammonia diffused for two cells determined by an alkalimetrical experiment which was always repeated twice. It was necessary for diffusion to have the ammoniacal solution made denser than water which was effected by the addition of common salt. Diffusion in 404 days at 63q4; two cells. Grs. Ratio. From 1 per cent solution . . . 4-93 1*Of29 From 2 per cent solution . . . 9.59 2 From 4 per cent solution . . . 19-72 4-117 From 8 per cent solution . . . 41.22 8.605 Ammonia appears to have a diffusibility approaching to that of hydrate of potash. It appears somewhat less diffusive than hydro- cyanic acid at the same temperature in the proportion of 12 to 13 nearly ;or to possess about three-fourths of the diffusibility of hydro-chloric acid.10. AkohoE.-Tinze of diffusion ten days. The quantity of alcohol diffused was determined by careful distillation. PROFESSOR GRAHAM ON THE Diffusion in ten days at 48O.7; two cells. From 2 per cent solution . . . . . . . 8.62 From 4 per cent solution . . . . . . . 16.12 From 8 per cent solution . . . 35.50 It would be unsafe to draw any conclusion as to the proportionality of the diffusion of alcohol to the strength of the solution from these experiments. Alcohol does not appear to belong to the same class of diffusive substances as acetic acid which might be expected from their simi- larity of composition but possesses a considerably lower diffusibility.Diffusion from 2 per cent solutions in ten days. Acetate of water at 48O.8 . . 11.51 100 Alcohol at 48O.7 . . . . . 8.62 74.9 The diffusion of alcohol approaches to one-half of that of sulphate of water at nearly the same temperature. Alcohol may be substituted for water to dissolve certain salts and also as an atmosphere into which these salts may diffuse. From experiments which have been commenced on this subject it appears that the diffusion of hydrate of potash iodide of potassium chloride of calcium and others is about four times slower in alcohol of density 0.840 than in water. The salts likewise often exhibit the same relations in their diffusibility in alcohol as in water with some singular exceptions such as chloride of mercury.11. Nitrate of baryta.-Time of diffusion 11-43days.* The salt diffused was precipitated by sulphuric acid and calculated from the weight of the sulphate of baryta formed. Diffusion in 11.43 days at 64'01 ; two cells. Grs. Ratio. From 1per cent solution . . . 7.72 1.026 From 2 per cent solution . . . 15.04 2 From 4 per cent solution . . . 29.60 3.936 From 8 per cent solution . . . 54 50 7.247 12. Nitrate of strontia.-Time of diffusion 11-43days. Of anhy-drous nitrate of strontia 0.82 per cent; density 1.0063. Diffused at 51O.5 in eight cells 5-59 5.62 5.44 5.69; mean 5.59 grs. for two cells ; calculated for 1 per cent 6.79 grs. at 51O-5 for two cells. The diffusion of nitrate of strontia almost coincides with that of the isomorphous nitrate of baryta at the same temperature.* This time is to that of sulphate of magnesia (16*166 days) as the square root of 8 is to the square of 16; but does not appear to expiess the true relation between these salts. DIFFUSION OF LIQUIDS. Diffusion from 1 per cent solutions at 51O.5 in 11.43 days. Nitrate of baryta . . . . . 6.73 100 Nitrate of strontia . . . . . 6.79 100.89 13. Nitrate of lime.-Time of diffusion 11.43 days. The diffusate was evaporated to dryness with an excess of sulphuric acid and the nitrate of lime which is always supposed anhydrous was esti- mated from the sulphate of lime produced. Diffusion in 11.43 days at 64'01 ;two cells. Grs. Ratio. From 1 per cent solution .. . 7.66 1.021 From 2 per cent solution . . . 15.01 2 From 4 per cent solution . . 29.04 3.372 From 8 per cent solution . . . 55.10 7.334 The results throughout for this salt are almost identical with those of nitrate of baryta although these two salts differ greatly in solubility and in one being a hydrated and the other an anhy- drous salt. 14. Acetate of lead.-Diffused for 16.166 days j the time chosen formerly for sulphate of magnesia with seven days for chloride of sodium. The solution contained 0.965 per cent of anhydrous salt with the density 1.0080. As this solution of acetate of lead was found to be precipitated by pure water about 2 per cent of strong acetic acid was introduced into the solution and the same acid was added in a less proportion to the water jars.The salt of lead diffused was afterwards determined by means of sulphuric acid Dif-fused in eight cells at 53'01 7.45 7.29 7.46 and 8.07 grs.; mean 7.56; or 7.84 for 1 per cent in two cells. 15. Acetate of baryk-Diffused for 16.166 days. The solution contained 0977 per cent of anhydrous salt with the density 1.0073. The same addition of acetic acid was made to it as to the preceding acetate of lead in order that the circumstances of diffusion might be similar for both salts. The salt diffused was estimated also in the form of sixlphate. Diffusion of 1 per cent solutions in 16.166 days; two cells. Acetate of baryta at 53O.5 . . 7-50 100 Acetate of lead at 53O.1 . . . 7.84 104*53 Here of two isomorphous salts that of greatest atomic weight sensibly exceeds the other in diffusibility.16. Chloride rf barium.-Time of diffusion 11.43 days. The diffused salt was weighed as sulphate of baryta. PROFESSOR GRAHAM ON THE Diffusion in 8.57 days at 63O; two cells. Grs. Ratio. From 1per cent solution . . . 6.32 1*047 From 2 per cent solution . . . 12.07 From 4per cent solution . . . 23.96 3-970 From 8 per cent solution . . . 45.92 7.608 17. Chloride of strontium.-The diffused salt was weighed as sulphate of strontia. Diffusion in 8-57days at 63'; two cells. Grs. Ratio. From 1 per cent solution . . . 6.09 1*045 From 2 per cent solution . . . 11.66 2 From 4per cent solution . . . 23.56 4.041 From 8 per cent solution .. 44~46 7.626 The series of ratios in the preceding table will be found on com- parison to correspond closely with the ratios of chloride of barium. It may be useful to compare farther the amounts diffused from similar solutions of these two isomorphous compounds. Diffusion in 8-57days at 63' ; two cells. Chloride of barium 1 per cent . 6-32 100 Chloride of strontium 1 per cent 6.09 96.36 Chloride of barium 2 per cent . 12.07 100 Chloride of strontium 2 per cent 11.66 96-90 Chloride of barium 4 per cent . 23.96 100 Chloride of strontium 4per cent 23.56 99.16 45.92 100 Chloride of barium 8 per cent Chloride of strontium 8 per cent 4446 96.83 The near coincidence of the 4 per cent solutions probably arises from an accidental error of observation in the chloride of barium for the latter departs here from the progression of its ratios.We appear then to have a small but constant difference of about 34 per cent in the diffusion of these two isomorphous salts the chloride of barium which possesses the highest atomic weight having the advantage. The diffusion of the 1 per cent solution of the same salts for the longer period of 11.43 days gives 7.50 for chloride of barium at 50O-9 and 7'052 for chloride of strontium at 51° or nearly the same temperature. For the first time we have in the barytic salts a divergence between chlorides and nitrates for the nitrates of the same bases have a number about 6.8 only at the same temperature. I am led however to believe that this discrepancy becomes much DIFFUSION OF LIQUIDS.9s less at low temperatures by experiments which are at present in progress. 18. Chloride ofcalciuum.-Time of diffusion 11.43 days. The salt diffused was weighed as sulphate of lime Diffusion in 11.43 days at 63O.8; two cells. Grs. Ratio. From 1 per cent solution . . . 7.92 1.032 From 2 per cent solution . . . 15.35 2 From 4 per cent solution . . . 30.78 4.010 From 8 per cent solution . . . 61.56 8.021 We may now observe how far the diffusion of the chloride of calcium is analogous to that of nitrate of lime. At the inferior tem- peratures the results for the 1 per cent solution of these two salts were as follow Chloride of calcium at 50'9 . . 6.51 100 Nitrate of lime at 51O.5 . . . 6.54 100.46 While at the higher temperatures namely 63'48 for the chloride of calcium and 64'*1 for the nitrate of lime the results for the different proportions of salt are Chloride of calcium 1 per cent .. 7.92 100 Nitrate of lime 1 per cent . . . 7.66 96.72 Chloride of calcium 2 per cent . . 15.35 100 Nitrate of lime 2 per cent . . . 15.01 97.79 Chloride of calcium 4 per cent . . 30.78 100 Nitrate of lime 4 per cent . . . 29.04 9435 Chloride of calcium 8 per cent . 61-56 100 Nitrate of lime 8 per cent . . . 55.10 89.5 1 The correspondence between the 1 and 2 per cent solutions of chloride and nitrate is sufficiently close but in the 4 and 8 per cent the salts diverge as happens also with hydrochloric and nitric acids themselves The nitrate in both falls off while the chloride sustains throughout the high diffusibility of the lower proportions.19. Chloride of manganese.-Time of diffusion 11.43 days. The salt diffused was estimated by means of nitrate of silver. The 1 per cent solution of density 1*0085,gave at SO0& in eight cells 6-67 6.26 6.79 and 6.81 grs.; mean 6-63for two cells. 20. Nitrate of magnesia.-Time of diffusion 11-43 days. The salt diffused was estimated as sulphate. The 1 per cent solution of density 1.0073 gave at 50°*8 in eight cells 6.29 6-39 6.52 and 6.76 grs. ;mean 6.49 for two cells. 21. Nitrate of copper.-Time of diffusion 11.43 days. The salt diffused was estimated from the oxide of copper obtained by ignition. PROFESSOR GRAHAM ON THE The 1 per cent solution of density 1.0075 in eight cells at 5O0.S gave 6.52 6-36 6.18 and 6.70 grs.; mean 6.44 for two cells. Comparing the preceding salts with chloride of calcium diffused at the same temperature 50'43 we have the following results Chloride of calcium . . . . 6.52 100 Chloride of manganese . . . 6.63 101.15 Nitrate of magnesia . . . . 6.49 99-69 Nitrate of copper . . . 6.44 98.92 This group of salts belonging to the same isomorphous family of bases the magriesian again corresponds closely in diffusibility. The following additional magnesian chlorides were diffused all 1 per cent solutions either in six or in eight cells. The salt diffused was estimated by means of nitrate of silver. The results referred to chloride of calcium at nearly the same temperature 50O.8 are as follow Chloride of calcium .. . 6.51 100 Chloride of zinc . . . . . 6.29 96.61 Chloride of magnesium . . . 6.17 94.77 Chloride of copper . . . . 6.06 93.08 These salts present a greater latitude in their diffusibility if belonging to the same class than is usual. 22. Protochloride of iron.-A solution of this salt of 1.023 per cent was diffused at 53O.5 a somewhat higher temperature than the corresponding chlorides. It gave 6.45 6.48 6.48 and 6.28 grs. in two cells; mean 6.44 or 6*30 for 1 per cent in two cells. This salt appears therefore to belong to the last group. 23. Sespuichloride of iron.-A full series of observations was made upon the diffusion of the different proportions of this salt from 1 to 8 per cent but in all of them decomposition was determined by the diffusion with turbidity also in the solution-phial except in the 8 per cent solution.The mean diffusion from the 1 per cent solution in 11-43 days at 63O.3 was 4-13grs. of sesquichloride of iron with 1.28 gr of free hydrochloric acid in two cells. This result indicates that one-half nearly of the sesquichloride of iron is decomposed in the diffusion. The mean diffusion from the 8 per cent solution at 63O.3 was 55.88grs. of sesquichloride of iron with 6.66 grs. of free hydro- chloric acid in two cells. It appears from this experiment that perchloride of iron approaches the chloride of calcium in diffusibility. That the proto- and persalts of the niagnesian metals should have a similar rate of diffusion is not unlikely from other analogies which they exhibit.24. Subhate of magnesia.-The time chosen for the diffusion of this salt namely 16.166 days is a multiple by 2 of the time of DIFFUSION OF LIQUIDS. sulphate of potash and by 4 of the time of hydrate of potash. The diffusate was evaporated to dryness and weighed. Diffusion in 16.16 days at 65O.4; two cells. Grs. Ratio. From 1 per cent solution . . . From 2 per cent solution . . From 4 per cent solution . . . From 8 per cent solution . . . From 8 per cent solution at 62O.8 . From I6 per cent solution at 62'08 From 24 per cent solution at 62'08 7.31 12.79 23-46 42-82 42.66 75.06 102*04 l*lM 2 3.671 6.701 1 1.759 2.340 25. Su&hate of zinc.-Time of diffusion 16.166 days.The dif- fused salt was evaporated to dryness and weighed. Diffusion in 16-16 days at 65O.4; two cells. Grs. Ratio. From 1 per cent solution . . . 6.67 1.091 From 2 per cent solution . . . 12-22 2 From 4 per cent solution . . 23.12 3.784 From 8 per cent solution . . . 42.26 6.916 From 8 per cent solution at 62O.8 . 39.62 1 From 16 per cent solution at 62'3 74.40 1.878 From 24 per cent solution at 62O.8 101.42 2.560 It will be remarked that the diffusion of these two isornorphous salts sulphate of magnesia and sulphate of zinc differs so much in the 1 per cent solution as 7.31 to 6.67 that is as 100 to 91-25;or 8.75 per cent. This I have no doubt however is an accidental error the disturbances from changes of temperature and other causes of dispersion being in direct proportion to the duration of the experi- ment and therefore much increased with these long times; while the 1 per cent solution also appears to be generally the proportion most exposed to such errors.The sulphate of zinc appears to be the truest throughout in its diffusion of these two salts. The approach to equality becomes close in the 4 per cent and larger proportions of salt particularly with the unusually high proportions of 16 and 24 per cent which were observed in these salts. The diffusion of both salts falls off remarkably in the higher proportions. The result of the comparison of these two magnesian sulphates is no doubt favour-able to the similarity of diffusion of isomorphous salts. 26. Sdphate of alumina.-The time of diffusion chosen was 16.166 days or the same as that for sulphate of magnesia.The usual number of eight cells of the 1 and 2 per cent solutions were diffused and four cells of the 4 and 8 per cent solutions. The whole diffusates of each proportion were then mixed together and the quan- PROFESSOR GRAKAM ON THE tities of alumina and sulpburic acid diffused for two cells deter- mined separately. Diffusion in 16.166 days at 65O.4; two cells. Grs. Ratio. From 1 per cent solution . . . 5-48 1.074 From 2 per cent solution . . . 10.21 2 From 4 per cent solution . . . 19.28 3.780 From 8 per cent solution . . 33-52 6%72 The diffusion of sulphate of alumina it will be observed is very sensibly less than that of sulphate of zinc at the same temperature.27. Nitrate ofsilver.-Time of diffusion seveu days. The quan- tity of salt diffused was ascertained by precipitation with hydro- chloric acid and weighing the chloride of silver formed. Diffusion for seven days at 63'04; two cells. Grs. Ratio. From 2 per cent solution . . . 13.61 2 From 4 per cent solution . . . 26-34 3.87 From 8 per cent solution . . 51-88 7.62 28. Nitrate of soda.-Time of diffusion seven days. The quan- tity of salt diffused was ascertained by evaporation to dryness. Diffusion in seven days at 63'04; two cells. Grs. Ratio. From solution of 2 per cent . . 12.35 2 From solution of 4 per cent 23.56 3.82 From solution of 8 per cent . . 47.74 7.72 The ratios of the last column of the preceding Table are sensibly the same as those already obtained for nitrate of silver.But the diffusibilityof nitrate of soda appears to be increased less rapidly by temperature than nitrate of silver. Hence the diffusibility of these two salts appears more similar at low than high temperatures. Diffusion from 2 per cent solutions in seven days at 53O. Nitrate of silver . . . . 11.24 100 Nitrate of soda . . . . 10.81 96.17 Diffusion from 2 per cent solutions in seven days at 63O.4. Nitrate of silver . . . . 13.61 100 Nitrate of soda . . . . 12.35 90-74 29. Chloride of sodium-Time of diffusion seven days. The salt diffused was treated with nitrate of silver and the chloride of silver weighed. DIPFUSION OF LIQUlDS.97 Diffusion in seven days at 63O.J; two cells. Grs. Ratio. From 1per cent solution . . . 6.32 1-023 From 2 per cent solution . . . 12.37 2 From 4 per cent solution . . . 2496 4.036 From 8 per cent solutiorr . . . 48.44 7.832 These numbers resemble closely those obtained in the diffusion of chloride of barium during the longer period of 8.57 days. The chloride of sodium and nitrate of soda will be seen to exhibit the usual approach to parallelism between the chloride and nitrate of the same metal by the following comparison I)iffusioii of both at 63O.4. Chloride of sodium 2 per cent . . 12.37 100 Nitrate of soda 2 per cent . . . . 12.35 99.83 Chloride of sodium 4 per cent . . 24.96 100 Nitrate of soda 4 per cent . . . . 23.58 94.48 /I Chloride of sodium 8 per cent .48.44 100 Nitrate of soda 8 per cent . . . . 47.74 98-55 As usual the chloride is slightly more rapid in its diffusion tliaii the nitrate. 30. Chloride of potassium.-Time of diffusion 5.71 days. The salt diffused was treated with nitrate of silver and the chloride of silver weighed. Diffusion in 5.71 days at 62'; two cells. Grs. Ratio. From 1 per cent solution . . . 6.69 1*005 From 2 per cent solution . . . 13.32 2 From 4 per cent solution . . . 25-94 3.895 From 8 per cent solution . . . 53.64 8.054 The ratios are in remarkably close accordance with the proportions of salt diffused. The times 5.7'1 and seven days chosen for the chlorides of potas-sium and sodium it will be observed are as the square rootsof 2 and 3.A certain deviation from this ratio of the times of equal diffusion appears on comparing the experimental results obtained at present for these salts. Diffusion of chloride of potassium in 5.71 days at 62O and of chloride of sodium in 7 days at 63O.4. Chloride of potassium 1 per cent . . 6.69 100 Chloride of sodium 1 per cent . . . 6.33 94.47 Chloride of potassium 2 per cent . . 13.32 100 VOL. 1V.-NO. XIII. fi YltOPESSOR GRAHAM ON THE Chloride of sodium 2 per cent . . . 12.37 92.86 Chloride of potassium 4 per cent . . 25.94 100 Chloride of sodium 4 per cent . . . 24.96 96.23 Chloride of potassium 8 per cent . . 53.64 100 Chloride of sodium 8 per cent . . 48.44 90.30 The difference would be about 1 per cent greater if the diffusion of both salts were reduced to the same temperature.The chloride of potassium deviates of course from the nitrate of soda in a similar manner. But chloride of potassium corresponds more closely with nitrate of silver than with chloride of sodium and nitrate of soda at the temperature of the experiments. Diffusion of chloride of potassium for 5.71 days at 62O and of nitrate of silver for 7 days at 63O.4. Chloride of potassium 2 per cent . . 13.32 100 Nitrate of silver 2 per cent . . . . 13.61 102-18 Chloride of potassiixm 4 per cent . . 25.94 100 Nitrate of silver 4 per cent . . . . 26.34 10154 Chloride of potassiuni 8 per cent . . 53.64 100 Nitrate of silver 8 per cent . . . . 51-88 96.71 The coincidence in rate would appear even closer in the 2 and 4 per cent solutions if the diffusion of the nitrate of silver was dimi- nished about 1 per cent on account of its higher temperature.It might thus be supposed that the nitrate of silver followed the sodium rate more accurately than the nitrate of soda and chloride of sodium themselves do. A series of observations were made upon the diffusion of the 1per cent solution of chloride of potassium at a nearly constant tem- perature of 56* but for different times varying from five days to eight days and eighteen hours to discover the progression which proved to be pretty similar to that of the 2 per cent solution of hydrochloric acid. Six cells were diffused for each period of which the mean result is given the times advance by ten hours.Diffusion of 1 per cent solution; two cells. Time. Temperature. Diffusion in two cells. Differences. 5 days 55O.71 5.89 5 days 10 hours 55 90 6.25 0-36 5 days 20 hours 55 .79 6-55 0.30 6 days 6 hours 55 *79 6.71 0-16 6 days 16 hours 55 -90 6.95 0.24 7 days 2 hours 55 -9 7.48 0.53 7 dajs 12 hours 55 9 7.58 0.10 7 days 22 hours 56 03 8.08 0.50 8 days 8 hours 56 *28 8.34 0.26 8 days 18 hours 56 -15 8.60 0.26 DIFFUSION OF LIQUIDS. When the quantities of chloride of potassium are placed beside the same quantities of hydrochloric acid in the former table it is found that the times of diffusion of the salt and acid exhibit an approxi- mately constant ratio. The squares of these times of equal diffusion are as 1 to 2.04 for the shortest period of the chloride of potassium, and as 1 to 2.10 for the longest period but one.The variation in the differences towards the middle of the table is too great to be explained except I fear by some error of observation although no ordinary precaution was neglected in the execution of this laborious series of experiments. 31. Iodides and bromides of potassium and sodium. Iodide ofpotassi?cm.-Time of diffusion 5-716 days. The dlffu- sate was estimated by means of nitrate of silver. (1). Iodide of potassium 1.977 per cent; density 1*0146. Dif-fused at 53O.5 in eight cells 11*4~5,11~506,10~94!! and 11.062grs. ; mean 11.24 for two cells and 11-36for 2 per cent. Comparing this salt with the isornorphous chloride of potassium we have Diffusion of 2 per cent solutions in 5.716 days Chloride of potassium at 55O .. 11.48 100 Iodide of potassium at 53'5 . . 11.36 99.65 The diffusion of the iodide would slightly exceed that of the chlo- ride instead of falling below it as in the table if the temperatures were made equal. (2). Again iodide of potassium 1971 per cent ; observed density 1*01486. Diffused at 59O.8 in eight cells and the mean diffusate of the whole cells determined it gave 12.33 grs. of iodide of potassium for two cells; or 12.51 grs. for a 2 per cent solution. Bromide of potassium-Time of diffusion and mode of estimating diffusate as above. The solution contained 1.975 per cent of salt and had a density of 1.014850. Diffused at 59O.8 in eight cells it gave a mean diffusate of 12.30grs.for two cells ; or 12.46 grs for 2 per cent. For comparison a solution of chloride of potassium containing exactly 2 per cent of salt and having the density 1.0133,was diffusedl in the same circumstances of time and temperature as the two preceding salts. The mean diffusate of eight cells was 12.24grs. for two cells. Hence the following result of the diffusion of three isomorphous salts PROFESSOR GltAHAM ON THE Difl'usion of 2 per cent solutions in 5.7'16 days at 59O.8. Grs. Ratio. Chloride of potassiiim . . 12.24 100 Bromide of potassium . . . 12-46 101-80 Iodide of potassium . . . . 12.51 102.21 Mean . . . . 12.40 Iodide of sodium-Time of diffusion seven days temperature 59O.8.A solution of 2.011 per cent and density 1.01618 diffused in eight cells gave a mean diffusate of 12-34grs for two cells; that is, 12-18grs. for 2 per cent solution. Bromide oj' sodium.-Time of diffusion and tcniperature as above. A solution of 2.146 per cent of density 1.01726 diffused in eight cells gave a mean diffusate of 12-80grs. ; that is 11.93 grs. for 2 per cent. A comparative experiment was made with a solution of chloride of sodium containing 1.917 per cent of salt and of density 1,01376 in eight cells at 60° The diffusates for four pairs of cells were 11.65 11.75 11.63 and 11-47grs. ; mean 11-63grs. which gives by pro- portion 12-14grs. for a 2 per cent solution. As the present salt differs only OO.2 Fahr. in diffusion-temperature from the ti170 pre-ceding salts which is inadequate to produce an assignable difference of diffusion the three salts may be supposed to be diffused at the same temperature without sensible error.Diffusion of 2 per cent solutions for seven days. Grs. Ratio. Chloride of eodiuni at GOO . . 12*14 100 Bromide of sodium at 59O.8 . 11.93 98-27 Iodide of scdium at 5','*8 . . 12.78 100.33 Mean . . . . 12.08 In both these isomorphous groups of salts of potassium and sodium there is certainly a near approach to equality of diffusion. The times for the salts of the two bases being in the empirical proportion of the square roots of 2 and 3 the mean diffusates also approach pretty closely; namely 12.40grs. for the salts of potas-sium and 12.08grs.for the salts of sodium which are as 100 to 97-42 Here the members of each group are certainly very similar to each other in density and probably other physical properties which was not the case with the equidiffusive group containing the hydrogen acids of the same salt-radicals. 32. Cliloride of amr~~onium.-Time of diffusion 5.716 days. The salt diffused was estimated by means of nitrate of silver. Solution 0.988 per cent; density 1.0036. Diffused at 53O in DIFFUSION OF LIQUIDS. eight cellc 6.09 6-07 5.67 5.87; mean 5.92grs. and 5-99 for 1 per cent in two cells. This is somewhat more than 5.68 one-half of the diffusate of the 2 per cent solution of iodide of potassium at nearly the same temperature. The diffusion however of the small proportions of salts of ammonium such as the 1 per cent solution is apt to be given in excess from their low density.33. Dichloride of copper.-Time of diffiision seven days or that of chloride of sodium. The salt diffused was obtained by evaporation to dryness in an air-bath after treating the liquid with an excess of chlorine in the form of chloride from which the dichloride was calculated. It was an object of interest to discover whether the dichloride of copper (Chiz Cl) which should be isornorphous with the chloride of sodium may separate from the protochloride of copper and other magnesian salts and assume the high diffusibility of the salts of alkaline metals. But the salt in question is entirely insoluble in water. A solution however was obtained by dissolving an equivalent qiian- tity of the red suboxide of copper recently precipitated in hyclro-chloric acid of density 1.033 so as to give one grain of dichloride in every hundred water-grain measures of the solution.This acid solution did not precipitate by dilution with water. The salt was diffused into pure water at a mean temperature of 53O.2. 1. Dichloride of copper diffused 6.66 6-57 7.01 and 6.48 grs. ; mean 6.68 grs. in two cells. Chloride of sodium at 53O.4 nearly the same temperature gave 5.90grs. in thc same time. Reducing the result to the temperature of 51' by an approximative correction we should have 6.48 grs. of dichloride of copper for that tempera- ture at which chloride of calcium gave 6-51grs. in 11.43 days and protochloride of copper (Crx Cl) 6.06 grs.at nearly the same tempe- rature also in 11.43 days. So far as we can judge from an experiment at a single temperature it would appear that the diffusion of dichloiide of copper is more rapid than that of the chloride (CuCl) in a proportion which sup- poses the former compound to possess half the "sulution-density" of the latter the times of equal diffusion 7 and 11.43 days being when squared as 1 to 2. With the view of discovering whether the large proportion of hydrochloric acid amounting to 7 per cent present in the preceding solution of dichloride of copper modified the diffusion of the salt a portion of the same acid solution was treated with chlorine-gas to convert the copper-salt into chloride and diffused into water after the excess of chlorine was removed by agitation of the solution with air.The proportion of salt present was thus increased in weight from I ,to 1.36 per cent. The time of difl'usion was 11.43 days and the temperature 53'. 2. Chloride of copper diffused from a 1-36 per cent solution of the 102 PROFESSOR GRAHAM ON THE salt in hydrochioric acid 5-83 5.66 and 5.30 grs in two cells; mean 5.60 grs. The corresponding diffusion from a 1 per cent solution may be supposed to be less than 5*6grs. in the proportion of 1.36 to 1 without any great error. The results thus become chloride of copper diffused 3.98 3-85 and 3-58grs. ; mean 3.80 grs. in two cells. It hence appears that the diffusion of chloride of copper is much diminished by the presence of a great excess of hydrochloric acid in the same solution.Different causes suggest themselves for this result such as the possibility of a combination existing of chloride of copper with chloride of hydrogen in the acid solution; or the influence which must be admitted of the more soluble substance in a mixture of two similar substances in repressing the diffusion of the less soluble. The present result however is entirely opposed to the idea that the high diffusibility of the dichloride of copper observed before is due to the hydrochloride acid present. 3. The diffusion of chloride of sodium also appears to be repressed by contact with a large excess of hydrochloric acid. One per cent of chloride of sodium raised the density of dilute hydrochloric acid from 1.035 to 1.0408.Diffused into pure water for seven days at 52"*9 in eight cells the diffusates of chloride of sodium were 3.80 3.87 4-00and 3.86 grs. ; mean 3.88 for two cells. The diffusion of chlo- ride of sodium is thus reduced in a corresponding measure with that of chloride of copper by association with seven times its weight of hydrochloric acid. These results are interesting in a very different point of view I have always watched for the appearance of some absorbent or imbibing power on the part of the acids more analogous to an endosmotic attraction for water as usually conceived. If such an attraction existed it would complicate the phenomena of diffusion for the volume of water absorbed by the acid would displace and project a portion of the latter into the reservoir the phial not being extensible.The high diffusibility of hydrochloric and nitric acids would be thus explained. Rut by such a mechanical displacement the chloride of sodium would be thrown out in the preceding experiment as well as the hydrochloric acid which is not the case. 4.Even in hydrochloric acid of density 1.184 (25 per cent) the diffusion of 1 per cent of chloride of sodium for seven days at 56O.6 was found to amount to 4.7 grs. only in two cells and is less than from a solution in pure water. 5 In comparing the influence of nitric acid with that of hydro- chloric acid upon the diffusion of chloride of sodium it was found that in a 7 per cent solution of nitric acid the chloride of sodium (1 per cent) was entirely decomposed in the diffusive process at 5fiO.6 and gave hydrochloric acid in the full diffusive equivalent of that acid together with nitrate of soda.DIFFUSION OF LIQUIDS. 34. Bicarbonate of potash.-Time of diffusion 8.083 days or double that of hydrate of potash. The water of the jars was partially charged with carbonic acid gas to prevent the decomposition of this and the other bicarbonates in the act of diffusion. The whole diffusates of each proportion were mixed together and the quantity of bicarbonate of potash diffused for two cells converted into the chloride of potassium evaporated to dryness and weighed. Diffusion in 8.08 days at 68'02; two cells. Grs. Ratio From 1 per cent solution .. . 7-23 1.029 From 2 per cent solution . . . 14.05 2 From 4 per cent solution . . . 26.72 3.806 From 8 per cent solution . . . 52 01 7.408 35. Bicarbonate of ammonia.-Time of diffusion 8,083 days. The whole diffusates of each proportion were mixed together and the quantity of bicarbonate of ammonia diffused for two cells deter- mined by an alkalimetrical experiment which was always repeated twice. Diffusion in 8.08 days at 68'02 ; two cells. Grs Ratio. From 1 per cent solution . . . 6 91 1.013 From 2 per cent solution . . . 13.65 2 From 4 per cent solution . . . 27.00 3-959 From 8 per cent solution . . . 50.10 7.346 The amount and progression of the diffusion of this salt correspond well for all the proportions diffused with the preceding isomorphous bicarbonate of potash.36. Bicarbonate of soda.-Time of diffusion 9.875 days. The whole diffusates of each proportion were mixed together and the quantity of bicarbonate of soda diffused for two cells converted into chloride of sodium evaporated to diyness and weighed. Diffusion in 9-87 days at 68O.1 ; two cells. Grs. Ratio. From 1 per cent solution . . . 7*31 1-059 From 2 per cent solution . . . 13-81 2 From 4 per cent solution . . . 26.70 3.869 From 8 per cent solution . . . 52.38 7.590 A remarkable approach to equality in the diffusion of the bicarbo- nates of potash and soda in the times chosen is observed equally in all the proportions of salt from 1 to 8 per cent. The results for the three bicarbonates may be stated as follow PROFESSOR GRAHAM ON THE the diffusate of the 2 per cent solution of bicarbonate of potash being made equal to 200,as a standard of comparison.Diffusion of bicarbonates of potash and ammonia in 8.08 days at 68O.2,and of bicarbonate of soda in 94375 days at 68O.1 Bicarbonate Bicarbonate Bicarbonate of potash. of ammonia. of soda. From 1per cent solution . 102.9 98.3 104.0 From 2 per cent solution . 2OO*O 19413 1116.4 From 4 per cent solution . 380.6 384-3 380.0 From 3 per cent solution . 740.8 712.6 74t3.3 Or making the diffusate from each proportion of the bicarbonate of potash equal to 100 Bicarbonate Bicarbonate Bicarbonate of potash. of ammonia. of soda. From 1 per cent solution From 2 per cent solution From 4 per cent solution From 8 per cent solution .. . . 100 100 100 100 95.53 97.15 100.97 96.19 101.07 98.20 99 84 101.03 The bicarbonate of ammonia is slightly lower in general than the bicarbonate of potash possibly from a small loss of the former salt by evaporation in the different operations. The times chosen for these two bicarbonates is to that of the bicarbonate of soda as the square root of 2 to the square root of 3 and the remarkable agreement observed in the diffusion of these salts gives support therefore to that relation. In alluding to this relation however it is proper to add that the carbonates of potash and soda deviate from it in a sensible degree and the hydrates of potash and soda very conside- rably. If the relation therefore has a real foundation it must be masked in the salts last named by differences existing between them in certain properties the discovery and investigation of which is of the last importance for the theory of liquid diffusion.37. Hydrochlorate of morphine.-Time of diffusion 11.43 days. The crystallised salt was assumed to be of the cornposition C,,H,,N06. HCl+ 6H0 with the equivalent 3745. The quantity diffused was determined from the chlorine which was precipitated as chloride of silver in an acid solution. Hydrochlorate of morphine 1.88 per cent of the salt supposed anhydrous diffused at 64O.1 in six cells 11.03 10.72 11.01 ; mean 10.92grs. of the anhydrous salt for two cells. Calculated for 2 per cent 11.60 gre. at 64O.1 for two cells.38. Hydroeldorate of strychnine.-Time of diffusion 11*43 days. The crystallised salt was assumed to be of the composition C,,H,,N,O . H C1+ 3H0 with the equivalent 397.5. Hydrochlo- DIFFUSION OF LIQUIDS. rate of strychnine 2per cent density 1.0065 diffused at 64O.1 in six cclls 11-54 11.62 11.31 ; mean 11.49 grs. for two cells. The quantgies refer to anhydrous salt and were estimated from the chlorine as with hvdrochlorate of momhine. .I I These two analogous salts appear to approach very closely in diffusibility. Diffusion from 2 per cent solutions at 64O.l ; two cells. Hydrochlorate of niorphine . . 11.60 100 Hydrochlorate of strychnine . . 11.49 99.05 For a similar period of 11.43 days but at a lower temperature 53O.4 the 1 per cent solution of hydrochlorate of morphine gave a mean result of 5.49 grs.from two cells and the hydrochlorate of strychnine 5.77 grs. from two cells. But the weights of chlo-ride of silver from which these numbers are deduced were too small to admit of much precision. The diffiision of these salts of organic bases in 11.43 days is exceeded by the diffusion of chloride of ammonium or potassium in 5-71 days or half the former time. The vegeto-alkalies appear thus to be divided from ammonia and potash. The new observations of the present paper are favourable to the existence of a relation amounting to close similarity or equality in diffusibility between certain classes of substances. The chlorides and nitrates of the same metal generally exhibit this correspondence as in the chloride of calciurn and nitrate of lime the chloride of sodium and nitrate of soda and also in hydrochloric and nitric acids.Isomorphous salts exhibit the same relation as has been observed in the chlorides bromides and iodides of potassium sodium and hydrogen in various salts of baryta strontia and lead in numerous magnesian salts in the salts of silver soda and probably those of suboxide of copper and in several additional salts of potash and ammonia. Corresponding salts of two of the vegeto-alkalies are also found to be equidiffusive. Before discussing the relations between the different groups of equidiffusive substances which are thus formed it will be necessary to examine their diffusion at widely different temperatures a siibject attended with considerable difficulty.DR. WI1,LIAMSON ON THE 106 Theory of Etheriflcation. By A. W. Williamson Ph.D.* When sulphuric acid is brought in contact with alcohol under certain circumstances a new arrangement is effected in the elements of the alcohol which divide into two groups forming ether and water. Now it is well known that the process by which this change is effected may be represented in two ways the difference of which consists in their respectively selecting for starting-point a different view of the constitution of alcohol. According to the one view an atom of alcohol weighs 23 and is made up of C2 HG0; so that to form ether two atoms of it are needed one of which takes C9 H4 from the other setting free the water with which these elements were combined ;whereas according to the other view alcohol weighs $6 and contains ether and water.These are not the only points of difference which are urged ; but they are the most real and tangible and their consideration is sufficient for our present purpose. If by any direct fact we could decide which of these two expressions is the correct one the ground would be clear for an examination of the process of etherification itself. The following experiments were made with the view of obtaining new alcohols by substituting carburetted hydrogen for hydrogen in a known alcohol. With this view an expedient was adopted which may render valuable services on similar occasions. It consisted in replacing the hydrogen first by potassium and acting upon the compound thus formed by the chloride or iodide of the carburetted ' hydrogen which was to be iutroduced in the place of that hydrogen.The process was first tried with common alcohol which after careful purification was saturated with potassium and as soon as the action had ceased mixed with a portion of iodide of ethyl equivalent to the potassium used. Iodide of potassium was readily formed on the application of a gentle heat and the desired substitution was effected; but contrary to expectation the compound thus formed bad none of the properties of an alcohol-it was nothing else than common ether C* HIO 0. Now this result at once appeared to be inconsistent with the higher formula of alcohol; for if that body contained twice as many atoms of oxygen as are in ether the product ought clearly to have contained twice as much oxygen as ether does.The alternative was evident; for having obtained ether by substituting C2 H5 for H in alcohol the relative composition of the two bodies is represented by expressing that ,fact in our formula. Thus alcohol is C2H5 o, H Phil. Mag. [3] XSXVII 350. THEORY 01” ETHERIFICATXON. and the potassium compound is C2 0; and by acting upon this by iodide of ethyl we have Of course the proportion between the two bodies is the only point here considered and the same reasoning would be applicable to any multiple of the formuh assumed. Some chemists may perhaps prefer doubling them in order to avoid the use of atoms of hydrogen potassium &c.; but the author has not felt himself justified in doing so because that would involve doubling the usual formula for water ; for as will be presently shown water is formed in etherifiea- tion by replacing the carburetted hydrogen of alcohol by hydrogen which of course obliges us to assume the same unity of oxygen in both.Alcohol is therefore water in which half the hydrogen is replaced by carburetted hydrogen and ether is water in which both atoms of hydrogen are replaced by carburetted hydrogen thus go C2 H5 o C2 H5 H C2 H5 This formation of ether might however be explained after a fashion by the other theory-by supposing the potassium cdmpound to contain ether and potash which separate during the action of the iodide of ethyl; SO that half the ether obtained would have been contained in that compound and the other half formed by double decomposition between potash and iodide of ethyl thus ’* K2 0 + C4 Hl0 I2= 2 IK -+ 2 (C4 HI0 0).But although the insuficiency of this explanation becomes evident on a little reflection a further and more tangible method of arriving at a conclusion was devised. It consisted in acting upon the potassium compound by iodide of methyl in which case if that compound were ether and potash the resulting mixture should consist of ether and oxide of methyl; whereas in the contrary case a body of the composition C3 Hs 0 should be formed. Now this substance was actually obtained and neither ether nor oxide of methyl.In this experiment the two theories cross one another and must lead to different results; for it is evident that in the first-mentioned decomposition by which ether was formed the only difficulty in explaining the process decisively consisted in our inability to prove that the carburetted. hydrogen introduced instead of the hydrogen did not have in the- product an atom of oxygen to itself but that on the contrary it was coupled with the carburetted hydrogen already contained in the alcohol-the two in combination with One DR. WILLIAMSON ON THE atom of oxygen. It is clear that if alcohol contains ether and water and the carburetted hydrogen in the first experiment formed a second atom of ether by taking the place of the hydrogen of this water the process being the same in the second experiment we should then have obtained two ethers; whereas if the formation of ether from alcohol be effected by synthesis a new carburetted hydrogen being added to the one already contained in the alcohol we ought to obtain the new intermediate ether which was really produced.The boiling-point of this remarkable body is a little above 10' C. ; it has a very peculiar smell distinctly different froin that of common ether; and like that body it is only slightly soluble in water. It is not acted upon by the alkali-metals at the common atmospheric temperature. By acting upon the potassium-alcohol in like manner by iodide of amyl a similar substitution was cffected of the elements of that carburetted hydrogen in the place of the hydrogen of alcohol and an ether obtained which boiled at 111' C.arid had the composition C7 HI6 0. There is some reason to believe that this body is the same which Balard obtained by decomposition of chloride of amyl by an alcoholic solution of hydrated potash and which that dis- tinguished chemist took for oxide of amyl. From the perfect analogy of properties between the known terms of the alcoholic series it was to be expected that similar substitu- tions might be effected in the others; and this expectation has been verified by experiment. Of course the formuk of the other alcohols must be reduced to half fbr the same rcasons as that of common alcohol. Methylic alcohol is therefore expressed by the formula H3 0 as common alcohol is C2 g50; and in the same manner H amylic alcohol is c5 0 and the same of the higher ones.In conformity to this fact we must be able to obtain the same interme- diate ethers by replacing hydrogen in these alcohols (methylic and amylic) by the carburetted hydrogen of iodide of ethyl as by the inverse process described above. This has been verified in the case of the three-carbon ether which may be obtained indifferently by replacing one-fourth of the hydrogen of methylic alcohol by C2H5 or by replacing one-sixth of the hydrogen of common,alcohol by CH3. C2 H5 Its rational formula is therefore C H3 O. By acting upon the compound z30 by iodide of amyl a third C H3 ethereal compound was obtained of which the formula is c5Hll 0.This is evidently the only one of' the three new ethers which con-taining an even number of carbon atoais might he conceived to have TH EORY OF ETH ERI FIC.ATION. been formed from one alcohol; but when treated with monobasic acids as the hydrochloric it cannot be expected to act in the same manner as its homogeneous isomeric the ether c3 C3 H7 0 of the three-C3 H7 o. carbon alcohol The next thing to be done is to explain theprocess of etherification by the action of sulphuric acid (SO4 H2) upon alcohol; and in order to accomplish that we must show the connexion between those substances and the reagents used in the above-described experiments. With this view we have merely to add to the above facts the acknowledged analogy of the simple and compound radicals in their compounds.It niiist first be shown how a substance analogous to the iodide of ethyl is formed and then how by double decomposition with alcohol it produces ether. This is very easy; for sulphovinic acid is strictly analogous to iodide of ethyl plus iodide of hydrogen which we should obtain by replacing SO* in its formula by an equivalent of iodine ; and in order to represent the formation of this sulphovinic acid which is well known to precede that of ether the simplest mode is at the same tirile the one most free from hy-pothesis; it consists in stating the fact that sulphuric acid and alcohol are transformed into suhhovinic acid and water. bv half the hydrogen of the former changing places with the carbu&ted hy-drogen of the latter ; thus Now from this point it is clear that the process is the same as in the decompositions above described ; for by this sulphovinic acid coming in contact with an atom of alcohol it reacts exactly in the same manner as the iodide did forming of course sulphuric acid and ether so4 so4 C2 H5 --c2 H5 * C2H5 0 c2 135 0 The sulphuric acid thus reproduced comes again in contact with alcohol forming sulphovinic acid which re-acts as before ; and so the process goes on continuously as found in practice.We thus see that the formation of ether from alcohol is neither a process of simple separation nor one of mere synthesis; but that it consists in the substitution of one molecule for another and is effected by double decomposition between two compounds.This 110 DH. WILLIAMSON ON THE view of the matter is therefore consistent with the contact theory inasmuch as it acknowledges the circumstance of contact as a necessary condition of the reaction of the inolecules upon one another. By reducing the formula of the alcohols to one atom of oxygen it also retains the equality of volumes which the contact theory insists upon between the vapours of these bodies and their ethers so that ether truly contains the elements of olefiant gas in addition to those of alcohol in one atom. But on the other hand it attaches equal importance to all the essential facts of the chemical theory and rests the explanation of the process as much upon them as upon those of the contact theory; for one-sixth of the hydrogen in alcohol truly exhibits different reactions from the remaining five and must there- fore be contained in that compound in a different manner from them; and the alternate formation and decomposition of sulphovinic acid is in this as well as in the chemical theory the key to explaining the process of etherification.Innovations in science frequently gain ground only by displacing the conceptions which preceded them and which served more or less directly as their foundation ; but if the view here presented be con- sidered a step in our understanding of the subject the author begs leave to disclaim for it the title of innovation; for the conclusion here deduced consists in showing the compatibility of views which have hitherto been considered contrary ; and the best possible justification of the eminent philosophers who have advocated either one of the two conteiiding theories is thus afforded by reconciling their argu- ments with those of their equally illustrious oppo,ients.Let us now direct our attention to the transfer of homologous molecules in alternately opposite directions which as we have endeavoiired to show is the cause of the continuous action of sulphuric acid in this remarkable process. It may naturally be asked why do hydrogen and carburetted hydrogen thus continu- ously change places? It cannot be from any such Circumstance as superior affinity of one molecule over another for one moment sees reversed with a new molecule ' the transfer effected during the preceding one.Now in reflecting upon this remarkable fact it strikes the mind at once that the facility of interchange must be greater the more close the analogy between the molecules exchanged ; that if hydrogcbn and amyl can replace one another in a compound hydrogen and ethyl which are more nearly allied in composition and properties must be able to replace one another more easily in the same compound; and that the facility of interchange of hydrogen and methyl which are still more siniilar will be still greater. But if this be true must not the exchange of one molecule for another of identical propertics be the most easily effected of all ? Surely it must if there be any difference at all ;and if so the law of analogy forbids our imagining the fact to be peculiar to hydrogen among THEORY OF ETHERIFICATION.substanees resembling it in other respects. We are thus forced to admit that in an aggregate of molecules of any compound there is an exchange constantly going on between the elements which are contained in it. For instance a drop of hydrochloric acid being supposed to be made up of a great number of molecules of the composition C1 H tbe proposition at which we have just arrived would lead us to believe that each atom of hydrogen does not remain quietly in juxtaposition with the atom of' chlorine with which it first united but on the contrary is constantly changing places with other atoms of hydrogen or what is the same thing changing chlorine.Of course this change is not directly sensible to us because one atom of hydrochloric acid is like anothcr; but suppose we mix with the hydrochloric acid some sulphate of copper (of which the component atoms are undergoing a similar change of place) the basylous elements hydrogen and copper do not limit their change of place to the circle of the atoms with which they were at first combined,- the hydrogen does not rnerely move from one atom of chlorine to another but in its turn also replaces an atom of copper forming chloride of copper and sulphurie acid. Thus it is that at any moment of time in which we examine the mixture the bases are divided between the acids ; and in certain cases where the difference of properties of the analogous molecules is very great it is found that the stronger acid and stronger base remain almost entirely together leaving the weaker ones combined.This is well known in the case of a mixture of sulphuric acid and borax and is a confirrnation of our fundamental assumption that the greater the difference of pro- perties the more difficult is the alternate interchange of one molecule for another. But suppose now that instead of sulphate of copper we mixed sulphate of silver with our hydrochloric acid in aqueous solution and that a similar division of the bases between the acids established itself in the first moment forrriing four compounds SO4H2 SO*A@ ClH C1 Ag ;it is clear that this last-mentioned compound being insoluble in water must on its forination separate out and remove from the circle of decompositions which solubility established.BUt of course the three compounds remaining in solution continue the exchange of their component parts and give rise successively to new portions of chloride of silver until as much of that compound is precipitated as the liquid contained equivalents of its component parts a very small quantity remaining in solution and in the circle of decompositions. Such is the general process of chemical decomposition. Of course a compound is removed as effectually from the circle of decompo-sitions by assuming the gaseous form under the circumstances of the experiment or even by being a liquid insoluble in the men-struum. This explanation coincides in its second part with that 112 DR.ANDERSON ON THE CONSTITUl’ION OF which was proposed many years ago by Berthollet; but not making use of the atomic hypothesis upon which the preceding explanation is based that eminent philosopher went no farther back than the division of the acids between the bases on the mixture of salts a fact here deduced from the motion of atoms. It is well known that the general fact upon which Berthollet founded his view is denied by some eminent chemists of the present day; but the instances which they adduce are perhaps only apparent exceptions to the lam and will on further examination be found to afford additional con- firmation of the truth of the great Savoysien’s conception as already shown in the case of boracic and sulphuric acids.In using the atomic theory chemists have added to it of late years an unsafe and perhaps unwarrantable hypothesis namely that the atoms are in a state of rest. This hypothesis the author of the present paper discards and reasons upon the broader basis of atomic motion. On the Constitution of Codeine and its Products of Decompodtion. By T. Anderson M.D.* The composition of codeine has been variously stated by different chemists. According to Regnault the formula of anhydrous codeine is C, H, NO, and that of the crystallised base C, K, NO +2 HO. Gerhardt however objects to this formula because the numbers of equivalents of carbon and oxygen are uneven and the SUM of the equivalents of hydrogen and nitrogen is likewise indivisible by two.From his own analysis Gerhardt deduces the formula c, H, NO, and this formula is fully confirmed by the recent researches of Dr. Anderson. The codeine with which the author’s experiments were made was prepared as usual from the mother-liquor from which morphia had been precipitated by ammonia. This liquid was evaporated to crys-tallisation and the crystals pressed to separate the hy drochlorate of ammonia which is more soluble than the hydrochlorate of codeine these operations being repeated till the greater part of the sal-ammo- niac was got rid of. The crystals of hydrochlorate of codeine were then dissolved in boiling water and the solution treated with caustic potash whereby the codeine was precipitated as an oil which after- wards concreted into a solid mass and was partly deposited in crystals as the solution cooled.Another crop of crystals may be obtained by evaporating the solution and the mother-liquor when concentrated to a small bulk becomes * Transactions of the Royal Society of Edinburgh. CODEINE AND ITS PRODUCTS OF DECOMPOSITION. 113 filled on cooling with long silky needles of morphia which had been retained in solution by the excess of potash. The crystals of codeine are purified by solution in hydrochloric acid boiling with animal charcoal and reprecipitation by potash ;and the resulting precipitate is finally dissolved in hydrous ether free from alcohol to separate any morphia that may adhere to it Anhydrous ether dissolves codeine much lees easily; the solution when evaporated yields smalf crystals of anhydrous codeine.Codeine crystallised from water or hydrous ether forms crystals belonging to the right prismatic system and containing 2 equivalents of water which are given off at 212'. Codeine is an extremely powerful base rapidly restoring the blue colour of reddened litmus and precipitating the oxides of lead copper iron cobalt nickel &c. from their solutions. It is precipitated by potash from its salts and is generally stated to be insoluble in that al- kali; but this is true only of very highly concentrated solutions as a considerable quantity of strong potash may be added to a saturated solution of codeine in water without producing precipitation and even when a very large amount of potash is added acertain quantity of the base is still retained in solution.Codeine is soluble in ammonia but not more so than in water; 100 parts of a moderately strong solu- tion of ammonia at 60' dissolve 1-40parts of codeine ;and according to Robiquet 100parts of water at 59' dissolve 1.26 parts. Codeine is precipitated from all its solutions by ammonia; it does not how- ever fall immediately but is slowly deposited in small transparent cry st als. SALTS OF CODEINE. HydrochZorate.-C, H, NO .HCI. Obtained by saturating hot hydrochloric acid with pure codeine. A concentrated solution becomes nearly solid on cooling; but a more dilute solution yields radiated groups of short needles which under the microscope are found to be four-sided prisms with dihedral summits.The crystals are soluble in 20 parts of water at GO' and in less than their weight of boiling water. The salt dried in the air retains 4 equivs. of water ; one which goes off at 212' and the other three at 250'; the salt at the same time losing acid and aquiring an alkaline reaction. The salt is also sometimes deposited from its solution in anhydrous crystals. Hydriodate of Codeine.-C, H, NO,. HI. Obtained by dissolving codeine in hot hydriodic acid and leaving the solution to cool where-upon it is deposited in long slender needles which fill the whole liquid if it has been sufficiently corrcentrated. Soluble in 60 parts of cold water ;more soluble in boiling water. The crystals contain 2 atoms of water. Xulphate of Codeine.-C,G H, NO,.HO . SO,. Crystallises in radiated groups of long needles or by spontaneous evaporation in flattened four-sided prisms. Dissolves in 30 parts of cold water ; very VOL. 1V.-NO. XIIT. 1 DR. ANDERSON ON THE CONSTITUTEON OF soluble in hot water. Neutral to test paper when pure; but very apt to retain a small quantity of acid which can only be got rid of by repeated crystallisation. The crystaflised salt contains 5 equivalents of water which are given off at 212'; the formula of the crystallised salt is C, H, NO ,HO .SO +5 Aq. Nitrate of Codeine.-C,B HZlNO .HO .NO, Formed by slowly adding nitric acid of specihc gravity 1.060 to powdered codeine an excess of acid being carefully avoided.. Easily soluble in boiling water from which it is deposited in small prismatic crystals on cool- ing.When heated OD platinum it melts and on cooling concretes into a brown resinous mass; at a higher temperature it is rapidly decomposed leaving a bulky coal difficult of incineration. Phosphate of Codeine.-C, M, NO,. HO .2 HO PO,. When tri- basic phosphoric acid is saturated with powdered codeine and the solution mixed with strong spirit this salt is obtained in small scales or short thick prisms which are readily soluble in water. They con- tain 3 equivalents of water of crystallisation which are given off at 212O. Oxalate of Codeine.-C, H, NO .HO .C 0,. Deposited on cooling its hot saturated solution in short prisms or sometimes in scales. Soluble in 30 times its weight of water at 60' and about half its weight at 212'.The crystals heated to 212' gave off 3 equiv-alents of water. At 250' the salt turns brown andat a higher tem- perature is entirely decomposed. Hydrosulphocyanate of Codeine.-C, HzlNO .HC NS,. De-posited in beautiful radiated needles on mixing the solutions of hydrochlorate of codeine and sulphocyanide of potassium. The crystals contain 1 equivalent of water which they give off at 212'. Chloride of Platinum and Codeine,-A moderately concentrated solution of hydrochlorate of codeine mixed with bichloride of platinum deposits a pale-yellow pulverulent precipitate ; a more dilute solution yields no immediate precipitate but after a time deposits minute tufts of silky needles The salt is soluble in water and on cooling is deposited partly in grains partly as a powder ; not however without partial decomposition.By ebullition with excess of chloride of pla- tinum it is completely decomposed. To obtain it pure it niust be precipitated in the cold and without excess of chloride of platinum. The formula of the crystallised salt is C H, NO,. HC1+ PtCl,+ 4HO. At 212' it gives off 3 equivalents of water and the remainder is expelled at 250' the salt at the same time turning brown and being partially decomposed. Codeine forms many other crystallisable salts none of which how- ever have been completely examined. The chromate forms fine yellow needles. With solution of corrosive sublimate codeine forms a white precipitate soluble in boiling water and alcohol and deposited in stellate groups of crystals on cooling.With chloride of palladium CODEINE AND ITS PRODUCTS OF DECOMPOSITION. 115 a yellow precipitate is obtained which is decomposed on boiling and yields metallic palladium The tartrate and hydrocyanate of codeine are uncrystallisable. PRODUCTS OF DECOMPOSITION OF CODEINE. Actionof Sukhuric Acid.-Am orp h ous Codein e.-When codeine is dissolved in moderately strong sulphuric acid the mixture digested for a while on the sand-bath and then treated with carbonate of soda a grey precipitate is obtained consisting of codeine in ail amorphous state. The precipitate must be collected on a filter washed with water dissolved in alcohol and precipitated from the solution by water. It then forms a grey powder with more or less of a green shade insoluble in water readily soluble in alcohol and precipitated by ether from the alcoholic solution.Fuses at 212' to a black resinous mass Readily soluble in acids forming salts which are amorphous and dry up by evaporation into brown resins. Analysis shows it to be identical in composition with codeine in its ordinary state. The action of sulphuric acid upon codeine is indeed analogous to that which it produces upon quinine but the resulting amorphous codeine is not so stable a substance as quinoidine. Moreover the action does not stop at tbe point at which amorphous codeine is pro-duced; for if it be continued a deep-green substance is formed con- taining sulphur and analogous to the sulphomorphide described by Arppe and the corresponding sulphonarcotide of Lauren t and Ge rh ard t.Action of NitricAcid.-Nitrococteine-C,G H, NO,NO, When strong nitric acid is poured upon codeine and heat applied violent action takes place nitrous fumes are abundantly evolved and the solution acquires a red colour. If the fluid be evaporated over the water-bath a yellow resinous acid is left soluble with a red colour in ammonia and potash; but when dilute acid is used a nitro-base is obtained having the composition above given. The best mode of preparing it is to add finely powdered codeine to nitric acid of specific gravity 1*060,kept at a moderate heat in a flask small portions of'the liquid being taken out every now and then and tested with ammonia till the precipitate formed on neutralising the acid no longer increases in quantity.The liquid is then to be saturated with ammonia and stirred rapidly whereupon it will become filled with a bulky precipi- tate of nitrocodeine. The action is very rapid and great care must be taken not to let it go too far otherwise the resinous substance above-mentioned will be produced its formation being indicated by the escape of red fumes; it is best to stop the action before the whole of the codeine is decomposed; but even then it is impossible to avoid the formation of a small quantity of the resinous substance its I2 DR. ANDERSON ON THE CONSTITUTION OF' presence being indicated by the dark colour which the liquid assumeJs on the addition of ammonia The precipitate formed by ammonia is in the form of minute silvery plates with a very slight shade of yellow.It is purified by solution in hydrochloric acid boiling with animal charcoal and reprecipitating with ammonia and subsequently crys- tallised from a solution in alcohol or in a mixture of alcohol and ether. Nitrocodeine crystallised from alcohol is deposited in the form of slender silky needles which have a pale fawn colour and on drying mat together into a silky mass. From alcohol and ether it is obtained by spontaneous evaporation in small yellowish crystals which under the microscope are seen to consist of four-sided prisms terminated by dihedral summits. Nitrocodeine is sparingly soluble in boiling water from which it is deposited in minute crystals on cooling.It dissolves abundantly in boiling alcohol but sparingly in ether. It is soluble in acids forming salts which are neutral to test paper and yield the base in the forin of a crystalline powder on the addition of potash or ammonia. When heated carefully it melts into a yellow liquid which concretes on cooling into a highly crystalline mass. At a higher temperature it decomposes suddenly without flame leaving a bulky charcoal. Cryatallised nitrocodeine is anhydrous. Hpdrochlorate of Nitrocodeine is obtained by dissolving nitroco- deine in hydrochloric acid and evaporating the solution whereupon the salt is left in the form of a resinous mass which cannot be made to crystallise. Subhate of Nitrocodeine C, H, (NO,) NO,. HO .SO, forms radiated groups of short-pointed needles neutral to test-paper and very soluble in boiling water. Oxalate of Nitrocodeine crystalliaes in short prisms of a fine yellow colour readily soluble in water. Platinochloride of Nitrocodeine C, H, (NO,) NO,. HC1+ PtCl +4 HO is precipitated from the solution of the hydrochlorate in the form of a yellow powder insoluble in water and alcohol. By treating an alcoholic solution of nitrocodeine with hydrosulphate of ammonia a new base is obtained which from its analogy with other compounds similarly prepared is probably composed of c36 H, N O, and may be called Axocodeirae ;it has not yet how-ever been completely examined on account of the great difficulty of preparing and purifying it. Action of Bromine on Code~ne.-Bromocodeine-C, H, &NO,.When bromine-water is added in small successive portions to finely powdered codeine the base is rapidly dissolved the solution losing its colour of bromine and acquiring a peculiar and characteristic red shade. After a certain quantity of bromine has been added small crystals make their appearance which are hydrobromate of codeine;but these CODEINE AND ITS PRODUCTS OF DECOMPOSITION. 117 ye only observed if the bromine-water has been thoroughly saturated and are deposited in a small quantity only the remainder being retainedin solution. When the whole of the codeine has been dis- solved ammonia is added and bromocodeine is immediately thrown down as a silvery-white powder; in this state it contains a small quantity of unchanged codeine.It is collected on a filter washed several times with cold water and redissolved in hydrochloric acid from which it is precipitated by ammonia and finally crystallised from boiling spirit. Bromocodeine is scarcely soluble in cold water ;but by boiling a somewhat larger quantity is taken up and deposited again on cooling in minute prisms terminated by dihedral summits. It is readily soluble in alcohol particularly on boiling and is best cryetallised from spirit diluted with its own bulk of water. The crystals are always very small but brilliantly white. It is soluble in ether. When heated it melts into a colourless liquid which is destroyed at a temperature somewhat above its melting-point. It dissolves in cold sdphuric acid and the solution becomes dark- coloured when heated.It is attacked by nitric acid but much less rapidly than codeine itself. Bromocodeine forms two hydrates the first containing one and the second two equivalents of water. Hydrochlorate of bromocodeine forms radiated needles closely re- sembling those of hydrochlorate of codeine The hydro-bromate is sparingly soluble in cold water readily soluble in boiling water and is deposited from the solution in small prismatic crystals. It contains two equivalents of water which are not expelled at 212'. Tribromocodeine.-C, Hg8Br NO,. By continuing the addition of bromine-water beyond the point at which bromocodeine is formed a further action takes places and a bright-yellow precipitate makes its appearance which at first re-dissolves in the liquid but after a time becomes permanent and goes on gradually increasing till a very large quantity of bromine has been added when at length a point is reached at which no further precipitate is produced.If the solution be left till next day however bromine again causes a precipitate; and if it be added as long as anything falls and the solution be again left standing another precipitate is produced identical in all respects with that before obtained ;and this may be repeated day after day for a very considerable time. The yellow precipitate thus obtained is the hydrobromate of tribromocodeine. It is collected on a filter and washed with water in which it is very sparingly soluble. When this salt is dissolved in dilute hydrochloric acid and ammonia added the tribromocodeine is immediately precipitated as a flocculent powder which must be washed with water andpurified by solution in alcohol and precipitation by water.Tribrornocodeine as thus obtained is a bulky white precipitate perfectly amorphous and when dry more or less grey in its colour. It is insoluble in water and ether but readily soluble in alcohol. DR. ANDERSON ON THE CONSTITUTION OF Hydrochloric acid dissolves it sparingly in the cold but more readily on boiling; in this process however it appears to undergo a partial decomposition as a small quantity is always left insoluble. Heated on platinum-foil it becomes brown and is entirely decomposed at its melting-point leaving a coal difficult of incineration.In such cases as have been hitherto examined the substitution of three equivalents of bromine in a base has entireIy destroyed its basic properties; but tribromocodeine is a base though a very feeble one. Its salts are all sparingly soluble in water and amorphous. The hydro-chlorate is obtained by dissolving the base in hot dilute hydrochloric acid and is deposited on cooling as an amorphous powder. The hydrobromate is the substance deposited during the preparation of the base. It is a bright-yellow powder perfectly amorphous and very sparingly soluble in cold water; on boiling however a larger quan- tity is taken up and is again deposited on cooling. Its composition appears to be 2 (C36H, Br NO,) +H Br.The platinochloride is a brownish-yellow powder soluble in water and alcohol. The action of bromine on codeine does not terminate with the production of the base just described but the author has not pursued the investigation any further. There must also exist a dibromoco-deine C, H19Br NO, but it has not yet been actually obtained. Action of Chlorine on ~ode~ne.-Ch~orocodeine-~, H, C1NO,. The action of chlorine on codeine is more complex than that of bromine. When chlorine gas is passed into an aqueous solution of codeine or chlorine-water added to it theliquid turns brown becoming continually deeper in colour and on the addition of ammonia yields an amor- phous resinous base. As there were no means of determining when the action was complete the product was not further examined.A definite compound analogous to bromocodeine was however obtained by dissolving codeine in excess of dilute hydrochloric acid at about 150' or 160' Fahr. adding finely pounded chlorate of potash-agitating the liquid-continually testing with ammonia and adopting precautions similar to those required in the preparation of nitro- codeine and finally precipitating by ammonia. The reaction is as f0hWS 3 (C36H21NOs.HCl)+3 HCl+KO. ClO,=KC1+6 HO +3 (C36H, C1 NO,. HC1.) The chlorocodeine is precipitated in the form of a silvery crystalline powder closely resembling brornocodeine. It has generally a yellow- ish colour and retains a small portion of codeine from which it is purified by dissolving in hydrochloric acid boiling with animal charcoal and reprecipitating with ammonia.It is finally obtained in crystals from its solution in boiling spirit. Chlorocodeine is sparingly soluble in boiling water and the solution CODEINE AND ITS PRODUCTS OF DECOMPOSITION. 119 on cooling deposits minute crystals exactly similar to and isomor- phous with those of bromocodeine. It is readily soluble in strong alcohol especially with the aid of heat and sparingly soluble in ether. It dissolves in sdphuric acid in the cold without change but the solution is charred by heating. Nitric acid dissolves it and the solu- tion is decomposed by boiling but not by any means so readily as codeine. Red fumes are evolved together with a peculiar and exces- sively pungent vapour.The salts of chlorocodeine are exactly similar to those of bromo- codeine. The hydrochlorate crystallises in groups of needles readily soluble in water. The su&hnte c3 H, c1NO,. H[o SO,+ 4Aq is deposited from its hot solution in radiated groups of short prisms which dissolve abundantly in boiling water and alcohol. Theplatinochloride is obtained in the usual way in the form of a pale-yellow precipitate scarcely soluble in water. Action of Cyanogenon ~odeine.-~icyan~codeine-~~, Nzl No6. 2 C N.-When cyanogen is passed into a solution of codeine in the smallest possible quantity of alcohol the gas is rapidly absorbed the liquid acquiring first a yellow and by continued action a brown colour. If the solution be then left to itself for some time the smell of cyanogen disappears and is replaced by that of hydrocyanic acid and crystals are gradually deposited.If the cyanogen be passed through the solu- tion in a slow continuous current the deposit of crystals is very abundant. These are collected on a filter and washed with a small quantity of alcohol; and the filtrate on being again exposed to the action of cyanogen yields an additional qnantity of crystals but not so pure as the former. The product is purified by solution with the aid of heat in a mixture of alcohol and ether from which it is depo- sited in crystals which are colourlese or slightly yellow. Obtained in this way however they are apt to retain a small quantity of codeine and it is therefore advantageous to pass cyanogen into the mixture to be used for their solution by which means the last traces of codeine are converted into the new compound.Dicyanocodeine is soluble in boiling absolute alcohol or a mixture of alcohol and ether and is deposited on cooling in anhydrous thin six-sided plates having a brilliant lustre. It is difficultly soluble in water but is dissolved on the addition ot alcohol. Nothing however is deposited from the solution on standing and by evaporation it is decomposed and crystals of codeine are left behind. With hydro- chloric acid it is converted into a crystalline salt but decomposition take place immediately; for on the addition of potash to the liquid ammonia escapes and if it be left for twenty-four hours hydrocyanic acid is evolved. With sulphuric and oxalic acid dicyanocodeine likewise forms sparingly soluble compounds which decompose rapidly with evolution of ammonia and hydrocyanic acid.The instability of these compounds prevented their further examination. DR. ANDERSON ON THE CONSTITUTION OF Dicyanocodeine belongs to the same class of compounds as cyaniw line but differs from that substance in containing two equivalents of cyanogen. ACTION OF ALKALXS ON CODEINE. Codeine when treated at moderate temperatures with potash yields more than one volatile base according to the circumstances under which the experiment is made. Similar results are obtained by the use of hydrate of potash or of potash-lime or soda-lime prepared in the usual way. The method employed in the experiment was to mix codeine with four or five times its weight of potash-lime or soda-lime and introduce the mixture into a retort with a tubulated receiver having a doubly bent tube attached to its tubulature the end of which passed into a small flask containing hydrochloric acid in order to retain any of the very volatile base which might not be condensed in the receiver.The retort was introduced into an oil-bath and kept at a uniform temperature of 250' Fahr. As soon as this temperature is reached a slight peculiar odoiir is observed which soon becomes more powerful and a small quantity of water retaining the bases in solu- tion collects in the receiver. The decomposition at 250" however is excessively slow and even after many days bases are evolved ap- parently in undiminished quantity ; but the mixture was maintained steadily at this point in hopes of obtaining the product free from ammonia which the preliminary trials had shown to be produced at higher temperatures; but even with this low heat it was evolved always in appreciable and in some experiments even in considerable quantity.The temperature was therefore gradually raised to about 350° when a larger quantity of base was obtained; and after the heat had been sustained for some time small crystals made their appear- ance which deposited themselves in a line round the retort just above the level of the oil in the bath but soon rose into and collected in the neck of the retort. These crystals resemble beazoic acid in their external appearance and are at first perfectly colourless but soon acquire a brownish shade by exposure to light and air.They are a base and rapidly restore the colour of reddened litmus. They are sparingly soluble in water but readily in acids and give a precipitate with bichloride of platinum. The quantity of this substance obtained was excessively minute ; and though considerable quantities of codeine were operated upon all that was obtained served only to make the few qualitative experiments now detailed. The watery fluid which collected in the receiver possessed a pungent and peculiar smell ;it restored the colour of reddened litmus with great rapidity and gave 'abundant fumes with hydrochloric acid. On the addition of solid potash R highly volatile and pungent oily base CODEINE AND ITS PRODUCTS OF DECOMPOSITION.121 collected as a layer on the surface of the fluid and at the same time a gaseous base escaped along with ammonia. From the small quantity of these substances obtained it was impossible to obtain any of them in a pure state Their constitution was therefore deter-mined by the analysis of their platinum-salts which can be separated from one another thouFh not without difficulty. In order to prepare these salts the basic fluid was saturated with hydrochloric acid and evaporated to dryness in the water-bath when it left behind a beauti- fully crystalline mass highly soluble in water and deliquescent in moist air. This was dissolved in absolute alcohol to separate ammonia and the filtered solution mixed with an alcoholic solution of bichloride of platinum when the platinum-salts were immediately thrown down as a pale-yellow powder very sparingly soluble in absolute alcohol but readily dissolved on the addition of water.The separation of the two bases is best effected by heating the washed pre- cipitate with boiling absolute alcohol and adding water in small quantities until the whole is dissolved. The crystals which deposit on cooling consist of one of the salts in a state of purity if the process has been properly managed or at all events only require a repetition of the process to make them absolutely pure. The salt thus obtained is scarcely soluble in absolute alcohol or ether but is readily soluble in water and dilute spirit and is thrown down from the latter solution by ether in the form of fine yellow scales.Its analysis gave results agreeing with the formula C H N .H C1. Pt C1,. The base is conse-quently the methylamine of Wurtz with whose description of that substance and its platinum-salt it perfectly agrees. The preparation of the platinum-salt of theother base was attended with much greater difficulty; and it could not be obtained quite free from methylamine. In order to obtain it the fluid which had depo- sited the methylamine-salt must be evaporated to a small bulk the salt which separated filtered off and ether added to the mother-liquor. Immediately a precipitate is obtained generally in the form of minute yellow needles but sometimes in scales. It is sparingly soluble in alcohol and ether and highly soluble in water from which it crystal-lises in long needles and with such facility that a few drops evaporated on a watch-glass leave the salt they contain in the form of five or six needles crossing the whole space occupied by the solution.The quantity of this salt was too small to admit of carrying its purification by recrystallisation as far as was to be desired and consequently a small quantity of methylamine remained in those subjected to analysis. The results of the analysis approach most closely to the formula C,H,N. HC1. Pt Cl,; consequently that of the base itself C,H,N. The base then obviously belongs to the same series as methylamine and forms the term of the series corresponding to metacetonic acid and in accordance with the system of nomenclature adopted by Wurtz it VOL.IV.-NO. XIII. K 122 DR ANDERSON ON THE CONSTITUTION OF CODEINE. receives the name of metacetamine. No examination was made of the salts of this base as it was not obtained in sufficient quantity but the author takes the opportunity of stating that before obtaining it from codeine he bad ascertained its existence among the products of destruc-tive distillation of animal substances.* The residue in the retort after these bases have been evolved is dark-cinnamon-brown and slightly coherent ; it dissolves in water with a dark-brown almost black colour and gives with acids a floccu- lent brown precipitate of a humus-like substance and perfectly amor- phous which was not examined.It still contains nitrogen ; and by exposure to a heat gradually raised to low redness it gives an additional quantity of volatile bases among which ammonia becomes more and more abundant as the temperature rises. A non-basic oil also makes its appearance but only in very small quantity. Wertheimf- has lately examined the action of soda-lime on certain organic bases He has obtained metacetamine from narcotine and methylamine from morphia ; and considering these substances to be directly eliminated from the bases he expects to obtain the residual atoms in the form of a definite compound. Dr. Anderson enter- tained a similar idea with regard to codeine until he detected the formation of two different bases which seemed to him rather to indicate that these substances appear as the result of a true destructive distillation; and that possibly by varying the circumstances of the experiment other bases may be obtained.The author has also observed another remarkable decomposition of codeine by which volatile bases are obtained. The formation by the action of nitric acid of a resinous acid has already been mentioned. This acid which is insoluble in water dissolves readily in dilute potash with a red colour; and the solution on boiling evolves a volatile base in great abundance. The further examination of the circumstances under which this change takes place is reserved for a future commu- nication. Dr. Anderson has likewise examined the action of iodine on codeine which yields a magnificent crystalline compound presenting the phenomena of pleochroism in a remarkable manner.* The author has likewise conyinced himself that the petinine described by him two years since as existing in bone-oil is represented by the formula C HI,N and not by C3H, N which he then gave for it. He has also ascertained the existence of ethylamine 2nd methylamine in bone-oil. $-Ann. Ch Phartn. LXXIII 208.