350 EWINS: T H E MUTUAL SOLURILITY O F FORMIC ACID ANDXXXVIIL-The Mutual Solubility of Formic Acid andBenxenp, and the System : Benxme-Formic Acid- Water.By ARTHUR JAMES EWINS.THE work which has hitherto been carried out on the mutual solu-bility of various pairs of liquids has been summarised by Rothmund( I ‘ Loslichkeit und Loslichkeitsbeeinflussung,” 1907, pp. 66-78).From the data there brought together i t appears that of about iortypairs of liquids studied, some twenty-five to thirty consist of pairsof liquids of which one component is water. This liquid, as is wellknown, possesses a very high dielectric constant. Methyl alcohol,which also has a high dielectric constant, is also a Component cftwo other systems. It seemed probable, therefore, that some otherliquid possessing a high dielectric constant might also be partlymiscible with certain organic liquids.A few preliminary experi-ments with formic acid * showed that this liquid was, indeed, partlymiscible with a number of liquids, among which were the following :Ethylene dibromideMethyl iodideEthyl iodideAcetylene tetrachlorideCarbon tetrachlorideAmylene (Trimeth ylethylene)%soAmy1 etherisoAmyl chlorideBenzeneBromobenzeneTolueneXyleneSafroleisosafrolePhenetoleAnisoleAll these mixtures showed maximal critical solution temperatawes.Further investigation would, in all probability, considerably acidto their number.Of these binary mixtures, that consisting of formic acid andbenzene was considered to be most advantageous for the purposeof a detailed investigation.The critical solution temperature iswell within the limits of experimental determination, but, a t thesame time, admits of the mutual solubility of the liquids beingdetermined over a fairly considerable range of temperature. Therewas no likelihood of any interaction between the two components,and both liquids are readily obtainable in a fairly pure condition.Additional interest attached t o the investigation, since the deter-mination of the critical solution temperature should afford the mostreliable method of ascertaining the purity of both liquids and, con-sequently, of establishing for these liquids such physical constantsas melting points, boiling points, density, etc., concerning which a* AccorJing to Thwiiig (Zcitsch.physikal. CILC~L, 1894, 14, 293), the Faluo ofthe dielectric coiistant of formic acid is 62’0, whilht Drude (ibid., 1897, 23, 308)gives 67’0BENZENE, AND THE SYSTEM : BENZENE-FORMIC ACID-WATER. 351considerable amount of uncertainty a t present exists, especiallywith regard to formic acid. The selection of benzene and formicacid was further influenced by the fact that the addition of waterto this binary system gives a tertiary system which is of interestowing to the peculiar character of the system so formed. Formicacid is completely miscible with water under all ordinary conditions.Benzene and water are, however, practically immiscible, their mutualsolubility being extremely small. These three liquids, theref ore,form an extreme case of a system of three components which canform two pairs of partly-miscible liquids.Systems of this classhave been previously studied by Schreinemakers (Zeitsch. physikal.Chern., 1898, 27, 95; 1899, 29, 577), the most closely analogoussystem being that of which the components are phenol, water, andaniline. It must be noted, however, that the analogy is representedthus :A phenol formic acidB water benzeneC aniline waterwhere A and A', and B and C are partly miscible liquids, A and Cbeing, in each case, completely miscible. Owing, however, to thevery slight mutual solubility of benzene and water, the region ofheterogeneous systems is very much more extended in the systemunder investigation, as is seen from the curve shown later (Fig.4).EXPERIMENTAL.The Preparation of Pure Formic Acid.Examination of the literature shows that there is a very con-siderable variation in the melting and boiling points which havebeen recorded for formic acid from time to time. This is seen fromthe following table :Melting Boilingpoint8*6O8.43'7.8'8.2"----7.0'5.6"8.35'point. -100.5"100.8"100.4'/759 mm101.0"/760 mm.101.3"/760 111111.100~5-100~8"/759 mm.100- 8 - 1 oi.00/760 -.101~O099.7'/741 mm.50.0°/120 mm. }Berthelot (Annalen, 1855, 98, 139)Pettersen and Eckstrand (Ber, 1880,13,1191)Zander (Annalen, 1884, 224, 59)Schmidt (Zeitsch. physikal. Chem., 1891, 7 ,Kahlbaum (Ber., 1883, 16, 2480)R,ichardson (T., 1886, 49, 763)Schiff (Annalen, 1885, 234.324)Kahlbaum (Zeitsch. physikal. Chem., 1898,Jones and Murray (Amer. Chem. J., 1903,Beckmann (Zeitsch. physikal. Chem., 1906,Garner, Saxton,'and Parker.(Amer. Chem. J . ,446)26, 591)30, 93)57, 129)1911, 46, 236352 EWINS: THE MUTUAL soLuBIrmy OF FORMIC ACID ANDThe preparation of the anhydrous acid is a matter of considerabledifficulty, owing t o the fact that the usual desiccating agents suchas phosphoric oxide, sulphuric acid, sodium, &c., all react with theacid. Further, the latter is itself extremely hygroscopic, so thatthe greatest care is necessary in manipulation in order to avoidreabsorption of water.The most recent attempt to prepare anhydrous formic acid wasmade by Garner, Saxton, and Parker (Zoc.cit.), who employed dis-tillation under diminished pressure from anhydrous copper sulphatefor the purpose. They obtained an acid melting a t 13.35~ andboiling a t 99*7O/741 mm., but it is probable, as will be seen later(p. 353), that this acid still contains a small amount of water.The following method of preparing the pure acid was adoptedand gave consistently good resul-ts.Kahlbaum ” formic acid was fractionally distilled through aYoung-Thomas stillhead (three compartments) to the outlet tubeof which was fused a small Liebig’s water-cooled condenser. Athermometer graduated in tenths of a degree was placed in positiona t the top of the fractionating column. I n order to prevent theacid from coming into contact with the cork through which thethermometer passed, the stem of the column above the outlet tubewas lengthened and surrounded by a glass jacket through whichwater was passed.The vapour of the acid was thus condensed justabove the sidc-tube and flowed back into the column. Formicacid “ Kahlbaum ” distilled in this way yielded an acid boiling, forthe most part, from 1 0 0 . 3 O to 100*8O, under a pressure of 760 mm.The lower and higher boiling portions of the acid were neglected.On refractionation, the main bulk of the acid was obtained boilinga t 100~4-100~6°/ 760 mm. Further fractional distillation wasfound to be useless as a means of purification, since no lowering ofthe critical solution temperature with pure benzene could be de-tected. The purified acid was next submitted to fractional re-crystallisation in the following manner.The acid was placed in awell-stoppered bottle and frozen solid in a cold room (temperaturebelow 0.). The bottle was then inverted over another bottle (at-mospheric moisture being carefully excluded) and the acid allowedt o thaw partly, at a temperature of not more than one or twodegrees above its melting point. When about one-eighth of theacid had slowly drained away, the remainder was liquefied a t asomewhat higher temperature, and it5 melting point determinedaccurately by means of a Beckmann thermometer. The process wasrepeated until, after three o r four recrystallisations, an acid wasobtained the melting point of which remained constant to 1/100thof a degree and gave a constant critical solution temperature witBENZENE, AND THE SYSTEM : BENZENE-FORMIC ACID-WATER,.353pure benzene. The amount of pure acid obtained in this mannerfrom 2 litres of Byfurther manipulation it was possible to obtain still more of the pureacid from the residual fraction.The course of the purification as indicated by the lowering of thecritical solution temperature * is shown by the following table :Kahlbaum " formic acid was about 300 C.C.Formic acid. M. p. R. p." Kahlbaum " 7.4"After fractionation . . . . - ( w ) 100*6-100- 7'/760After recrystallisationAfter recrystallisation99 > 9 - ( b ) 100~5-100~55"/760of purer fraction ( 6 ) - -until constant ...... 8.39" -Critical solutiontemperature (purebenzene).82"77"76"74.2'73.2"The Melting Point of Pure Formic Acid.The melting point was determined with a Beckmann apparatus.Since formic acid is very hygroscopic, the apparatus was fitted withthe sulphuric acid trap recommended by Beckmann for use in suchcases.That this arrangement was efficient was shown by the factthat three determinations of the melting point carried out on thesame sample of acid did not vary by more than O s 0 l o .The Beckmann thermometer was then immediately standardisedby comparison with a standard Kew thermometer and the freezingpoint of the acid so found to be 8 ' 3 9 O . with a probable error ofThe purity of the formic acid obtained was very strikingly shownby its behaviour on determining the melting point.With ordinarypreparations, as is well known, the amount of super-cooiing neces-sary to induce crystallisation is somewhat large, being, even withfairly pure preparations, fro= 2 O to 5 O . The acid obtained byGarner, Saxton, and Parker, for example, melting a t 8-35O and,therefore, of a high degree of purity, is stated by these authors tohave shown a super-cooling of 2 ' 5 5 O , even with vigorous stirring.I n the present instance, however, crystallisation commenced withan amount of super-cooling which was extremely small, the valuesactually obtained being 0'35O, 0'4O and 0 ' 1 5 O respectively in threeconsecutive determinations.& 0'02O.(b.) The Boiling Point of Pure Formic Acid.The boiling point of the acid cannot be taken as a criterion ofpurity, since very small amounts of water distil with the acid, the* For this method of controlling the purity of organic liqnids, compare Crismcr(Bull.A d , Boy. Belg., 1895, 30, 97), Flaschner (T., 1908, 93, 1000) andothers354 EWINS: THE MUTUAL SOLUBILITY O F FORMIC ACID ANDboiling points of the two liquids being very nearly the same a t theatmospheric pressure. The following values for the nearly anhydr-ous acid were obtained a t various times during the progress of thework :from which b. p./760 inrri.= 1 100-47O.100.3 -100*35O/ 754 mm.100.4 -100*45O/758 mm.100*55-100.6°/ 764 mm.Garner, Saxton, and Parker (Zoc. cit.) give 99'7O/741 mm., in goodagreement with the values tabulated above.The Density of Pure Formic Acid.The density of the pure acid obtained as described, was deter-mined by means of a Sprengel pattern pyknometer of nearly 30 C.C.capacity, the ends of which were provided with ground glassstoppers.As the result of three different determinations, the followingvalues were obtained for the density:D18 = (a) 1.2258( b ) 1.2260( c ) 1.2258the mean value being 1.2259 ( j- O*OOOl).This result is in good agreement with that obtained by Garner,Saxton, and Parker (Zoc.cit.), nameIy, 1.2260 a t 15O and 1.2200 a t26O, but is slightly higher, again pointing t o the probable presenceof traces of water in tke acid obtained by these authors.The Prepration of Pure Benzene.I n spite of the apparent ease with which benzene can be obtainedin the pure condition, no very good agreement has been reachedwith regard t o one of the most important crit'eria of purity-themelting point., as is shown by the following table of recorded values :5.17" R.Abegg (Zeitsch. physikal. Chem., 1894, 15, 213)5.44-5.445' E. Beckmann ( ,, 9 , ,, 1886, 2, 715)5.5" 9 9 ( ?9 1890, 6, 438)5-42" Lachowicz (Rer., 1 888,"21, 22dk)5.7" Schrvder (Zeitsch. physikal. Chem., 1893,11, 457)5.4" Lineberger (Amer. Chem. J., 1896,18, 437)5.54" Hansen (Zeitsch. physikal. Chem., 1904, 48, 595)The method of purification adopted in all cases is that of frac-tional crystallisation or distillation over sodium or a combinationof these two. The most recent determination, that of Hansen, wascarried out on a sample of benzene which had been purified, as faras possible, in the usual manner, and then, just previously to makingthe determination, had been boiled for a short time t o expel thBENZENE, AND THE SYSTEM : BENZENE-FORMIC ACID-WATER.355last traces of water. The melting point was, in this way, raisedfrom 5.470 to the value noted above, 5‘54O.The benzene employed in the present investigation was preparedfrom Kahlbaum’s benzene “ pure for analysis and molecularweight determination.” After one recrystallisation and subsequentdistillation from sodium in an all glass apparatus into a receiverfitted with a drying tube, benzene was obtained which gave a criticalsolution temperature with a sample of “ Kahlbaum ” formic acidof 81.9O. Further treatment of the benzene either by recrystal-lisation, distillation over phosph oric oxide, or metallic soliumfailed t o lower this critical solution temperature.The benzene was,therefore, considered to be as pure as could be obtained and em-ployed in the present investigation. The melting point of thebenzene was found to be (after careful standardisa-tion of the thermometer) 5.58O. This value is ingood agreement with that of Hansen (Zoc. cit.). Thevalue 5 ’ 7 O given by Schroder (Zoc. c i t . ) (whose methodof purification was simply recrystallisation) is probablytoo high.The Mutual Solubility of Formic Acid and Benzene.I n determining the mutual solubility of these twoliquids the synthetic method wae adopted.The actual procedure was as follows. A number ofbulb tubes of the shape shown (BC, Fig.1) were madefrom ordinary good quality glass tubing. The capacityof the bulb C was from 2 to 3 c.c., the length of thecapillary B about 3 cm., and the internal diameter ofthe capillary approximately 2 mm. The tubes werethoroughly cleaned by boiling with distilled water,and, after being compietely dried, were ready for use.I n order to fill the tube with known weighb of thetwo liquids the bulb tube was first weighed accurately to milligrams,The liquids were contained separately in two burettes, which wereprovided with drying tubes at the top so that excess of moisture wasprevented. The liquids were then run into the bulb by means offine capillary funnels ( A ) , which were so made that the wider por-tion of the funnel roughly fitted over the jet of the burette.Thecapillary stem of the funnel passed through the narrow portion ofthe bulb tube B projecting into the bulb C. On opening the tapof the burette the liquid passed directly into the bulb C withoutcontact with the surrounding atmosphere and without wetting thewalls of the narrow neck (Bj of the tube. By rapidly withdrhwingthe capillary A , wetting of the walls of the tube B was agai356 EWINS: THE MUTVAL SOLUBILITY OF FORMIC ACID ANDavoided. The composition of the mixture was accurately obtainedby weighing the tnbe after addition of each liquid, the amountsbeing roughly controlled by the readings on the burette. Owingto the hygroscopic nature of the pure formic acid and also in aless degree of benzene, i t was found necessary always to discardany liquid remaining in the jet of the burette after each filling,otherwise inaccurate results were obtained.For a similar reasonboth the bulb tube and capillary funnel were always carefullydried just before use. When both liquids had thus been introducedand weighed, the bulb tube was sealed off a t the top of the narrowportion. It was found that the possible loss of liquid during thisoperation, owing to vaporisation, was negligible. The amount ofliquid taken was so arranged that the bulb of the tube was abouttwo-thirds filled in order to allow of complete mixing of the liquidsby subsequent shaking.The temperature of complete miscibility of the mixture was thendetermined.The tube was immersed in water which was kept wellstirred by means of a mechanical stirrer, and the temperature wasrecorded by a standardised thermometer; this was graduated infifths of a degree, and was of such a range that, for all temperatures,the thread of mercury was completely immersed in the bath. Thetemperature was raised by passing steam into the water, a methodwhich was found t o be far more satisfactory, both in rapidity andin minimising the risk of breakage of the beaker, than the usualmethod of heating on a sand-bath by gas burners. The bulb tubewas placed as near t o the bulb of the thermometer as convenient,and the temperature of the bath gradually raised. The contentsof the tube (which was attached to a glass or wire holder) werefrequently shaken, until finally a temperature was reached at whichthe two layers disappeared and a perfectly clear homogeneous liquidwas obtained.This temperature was noted and the bath thenallowed to cool very slowly. The clear liquid gradually assumed thepeculiar blue fluorescent appearance characteristic of liquids nearthe critical solution temperature (see Konovalov, J . Russ. Phys.Chem. SOC., 1902, 34, 738; and Donnan, Chem. News, 1904, 90,139) ; this suddenly gives place to a true opalescence followedalmost immediately by a separation into two layers. The temper-ature a t which the opalescence appeared was noted and the tem-perature of tbe bath again slowly raised, and the point a t whichcomplete solution occurred again noted.By repeating theseobservations a few times the temperature of complete miscibilitywas easily obtained accurate t o 0.2O. The results obtained witha sel.ies of mixtures in which the concentrations of formic acidranged from 10 t o 90 per cent. are tabulated belowBENZENE, AND THE SYSTEM : BENZENE-FORMIC ACID-WATER. 357No. of Weight of Weight of Percentage of Solution1 0.082 0.814 9.2 21*0°2 0.250 1.862 12.8 39.13 0.191 1-152 14.2 44.24 0.248 1-165 17.5 51.45 0.210 0.855 19-7 56.C8 0.207 0.778 21-0 58.47 0.430 1.511 22.2 59.98 0-409 1.218 25.1 64.29 0.41 1 0.901 31.3 70.110 0.368 0.627 36.9 72.511 0.755 1.091 40.9 73.012 0.745 0.991 43.0 73.213 0.865 1.050 45.2 73.014 0.690 0.722 48.9 73-215 1.209 1.124 51.8 73.216 0-593 0.500 54.2 73-417 1.191 0.892 57.2 72.318 1.444 0.794 64.5 70.219 1.321 0-579 69.5 66.220 1.209 0.390 75.6 57.7(a) 0.794 0.177 81.8 41.20.546 81.9 41.0 * 22 (a) 1.211 0.201 85.8 25.2(b) 0.656 0.108 85.8 25.423 1.821 0.207 89.8 3.8* The second tube ( b ) iu each of these cases was made up with a different sampleof formic acid prepared separately at an interval of about twelve months, andaffords striking evidence of the trust to be placed in the method of purificationemployed.The curve obtained (Fig.2) from these experimentd data showsno marked variation from the type usually obtained from pairsof liquids showing maximal critical solution temperatures. Theportion of the curve in the neighbourhood of the critical solutiontemperature is very flat, a phenomenon which is, however, shownin many other cases (for example, in aniline and water).Theproportion of formic acid, therefore, in the mixtures which have asolution temperature corresponding with that of the critical tem-perature is seen to vary from about 40 to 53 per cent. From thepractical point of view this is an advantage, since in estimating thepurity of a sample of formic acid by this method, the mixture withbenzene can be made up volumetrically (having due regard to thedifferent densisies of the solutions) a t about 45 per cent. of formicacid without fear of falling beyond the limits of the temperatureof complete miscibility. On either side of the flat portion, thecurve falls very steeply, that is to say, the temperature-coefficient ofthe mutual solubility of the two liquids is comparatively small untilthe neighbourhood of the critical temperature is reached.This isseen experimentally on cooling a mixture containing from 40 to 50per cent. of formic acid, which has been heated above the criticalsolution temperature. On cooling, the clear liquid first becomestube. formic acid. benzene. formic acid. temperature*21 (b) 2.47358 EWINS: THE MTJTUAI, SOLUBILITY OF FORMIC ACID ANDfluorescent a t a little above the critical solution temperature, nextopalescent a t that temperature, and then almost immediately separ-ates into two well-defined layers almost equal in volume.As will be seen from Fig. 2 the generalisation first put forwardby Rothmund (Zoc. cit.) for this type of mixture of liquids holdsgood in this case also.The points bisecting the ordinates lyingbetween the arms of the curve lie on a straight line. The point a twhich this straight line cuts the mutual solubility curve determinesFIG. 2.Formic acid-Benzene mutual solubility curve.8 0"7060201010 20 30 40 50 60 70 80 90 100Percentage of formic acid.the true critical concentration of the system, which is thus seen t obe formic acid, 48 per cent., and benzene, 52 per cent.(1.) T h e Influence of Water on t h e Freezing Point of Formic Acidand on the Critical Solution Temperature with Beruene.The main difficulty in the preparation of pure formic acid un-doubtedly lies in the elimination of water from the extremelyhygroscopic acid.The effect of the presence of water in formicacid, both on the freezing point of the acid and on the formic acid-benzene critical solution temperature was, therefore, considered tobe of sufficient interest to warrant a quantitative study.Weighed quantities of pure formic acid were diluted by additionof weighed quantities of water. The melting points of the mixturesobtained were then determined by means of the Beckmann ap-paratus. The same care was taken in these experiments to avoidabsorption of moisture from the atmosphere as was exercised preBENZENE, AND THE SYSTEM : BENZENE-FORMIC ACID-WATER. 359viously. The melting points were determined one after anotherwithin a short space of time in order to avoid any error due t opossible change of zero of the Beckmann thermometer.The maxi-mal complete solution temperatures of these solutions with purebenzene were then determined as already described. The resultsobtained are tabulated below :Percentage ofwat,er.00.0530.0990.2240.5020.9842.04Depression of freezingpoint of formic acid. -0.07"0.16"0.305"0.70"1.35"2.67'Critical solution tempera-ture with pure benzene.73.2"74.9"75.8"76.7"79.3O84.2"92.8"From these results it follows that the effect of water on thefreezing point of the acid is additive. The values deduced for themolecular weight of water lie between 17 and 21, in good agreementwith the values obtained by Jones and Murray (Amer. Chem. J.,1903, 30, 193). The results tabulated above show also that, ex-cept in the case of very small amounts of water, where the effectappears to be somewhat greater, the rise in the critical solution tem-perature is proportional to the amount of water added.It is wellknown that the addition of small quantities of a third substanceto a binary system of partly miscible liquids has a very markedinfluence on the critical solution temperature which is either raisedor lowered according to the liquid added and the composition ofthe system. This is the case, for example, with methyl ethylketone and water, where the addition of a very small quantity ofethyl alcohol so altered the mutual solubility relations that a closedring was obtained (Bruni, Atti R. Accad. Lincei., 1899, [v], 8, ii,141).I n the latter case i t is to be noted that the third componentof the system is a liquid completely miscible with each of the twooriginal components. The addition of water t o the system formicacid-benzene as before indicated is different, for, in this case,whilst water is miscible in all proportions with formic acid, itssolubility in benzene is extremely small, so that practically t h e twoliquids may be considered to be immiscible. I n such a system,therefore, it would seem that the effect on the maximal solutiontemperature is proportional to the amount of the added thirdcomponent.Incidentally it will be seen how much more readily the presenceof water in formic acid can be detected by the critical solutiontemperature method.The rise in the critical solution temperatureoccasioned by the presence of 1 per cent. of water in the mixtureis some loo, whereas the fall in melting point produced by the sam360 EWINS: THE MUTUAL SOLUBILITY OF FORMIC! ACID ANDamount of water is only 1*3*, so that unless very special precautionsare taken in the determination of the melting point much greatertrust can be placed in the former method.The Ternary System, Formic Acid Benzewe-Water.As already explained, this system is of interest as being an ex-treme case of a type which has already been studied by Schreine-makers (Zeitsch. physikal. Chem., 1898, 27, 95; 1899, 29, 577),namely, a system of three components in which two pairs of partlymiscible liquids can be formed. Its strongly-marked character-FIG.3.The influence of water on the mutual solubiiily offomnic acid and hcnzene.120"1101 co60504010 20 30 40 50 60 70 80 90 100Percenlage of fomnic acid.istics are due t o the very small mutual solubility of water andbenzene.I n order t o depict graphically the complete conditions obtainingin the system, it is necessary t o obtain a series of isothermal curvesshowing the variations of composition of the system which can occurwhen the temperature (and pressure, which under the experimentalconditions is a negligible factor) remains constant. Such curveswere obtained experimentally as follows. A series of mixtures offormic acid and water containing different percentages of waterwas made up and the temperatures of complete miscibility of vary-ing proportions of these dilute formic acid solutions with benzenewere determined in the manner already described. A series oBENZENE, AND THE SYSTEM : BENZENE-FORMIC ACID-WATER.361curves (Fig. 3) or, more correctly, portions of curves, were thusobtained, showing the effect on the mutual solubility of benzeneand formic acid brought about by the presence of water, the pro-portion of water to formic acid being constant for each curve.The following are the experimental data from which these curveswere obtained. The percentage of water in the dilute formic acidmixture was varied from 5 to 40 per cent. by weight.Formic Acid containing 5 per cent. of Water.Dilute formic acidper cent.3.75.610.214.875.380.081.587.7Benzeneper cent.96.394-489-885.224.720.018.512.5T*.57.5"779511294.580.57851Formic Acid containing 10 per cent.of Water.3-65.17.979.681.485.596.494.992.120.418.614.570'82-5111105.59985Formic Acid containiny 15 per cent. of Water.2-53.44- 085-790.093-097.596.696.014.310.07.071"87101100.58146Formic Acid containing 25 per cent. of Water.88.091.594.012.08.56.0122"97.574Formic Acid containing 40 per cent. of Water.94- 0 6.0 105"96.2 3.8 8297.0 3.0 76* T=Temperature of complete miscibility.From the curves obtained, it will be seen that the addition ofwater t o mixtures containing a comparatively large proportion ofbenzene has the effect of very greatly raising the temperature ofcomplete miscibility, so much so that this portion of the curveVOL.cv. B 362 EWINS: THE MUTUAL SOLUBILITY OF FORMIC ACID ANDbecomes almost vertical. With mixtures containing a relativelylarge percentage of formic acid the effect is not so inconvenientlygreat, and the effect of relatively large differences of concentrationof water in the system as a whole can be more readily observed.This can only be due t o the fact that the presence of water in thesystem very greatly diminishes the solubility of benzene in formicacid. This is the reason also why the critical maximal solutiontemperature of formic acid and benzene is so greatly raised by thepresence of traces of water in either of the components.I n order to obtain the data for the required isothermal curvesmentioned above it is now only necessary to draw through thesecurves lines parallel to the horizontal axis from points on thevertical axis representing definite temperatures.At the points ofintersection of any one of these lines with the various curves wehave mixtures of definite composition showing complete miscibilitya t that temperature.I n this way the following figure8 were obtained:Formic acid.1.92.253-0416.079-383.182.178.970-968.21. Isothermal at 50°.Composition of System.Benzene97.897.696-884.020-712.68.87.26.53.02.2-133.154.131.663-079.460.177.970.768.0Isothermal at 70°.97.696.695.768.637.016.611 [8.46.83.32.76.473.1676.0676.069.267-47-83. Isothermal at 90°.96.894.091.823.016-211.87- 74.3Water.0.30.250-16004.49.113.923.638.80.40-360.2004.18.913.723.638-70-60.60.83.868.4613.223.138.BENZENE, AND THE SYSTEM : BENZENE-FORMIC ACID-WATER.363Formic acid.4.47.213.664-669-370.167.556.84. Zsothermal at llOo.Benzene.94.892.085.822.023.017.610.07.0Water.0-80.80.73.47-712422.637.2With the help of the figures thus obtained the system benzene-formic acid-water can be represented graphically by the methoddue t o Roozeboom (Zeitsch. physikal. Chem., 1894, 15, 147) andFIG.4.The terwrg system : benzeae-formic acid-water.Isothermals at SO", 70", go", and 110".employed by Schreinemakers (Zoc. cit.). This consists in plottingthe compositions of the various mixtures obtained as describedabove within an equilateral triangle. In the figure shown (Fig. a),the amounts of formic acid are measured in the direction B.F., andof benzene in the direction P.B. The series of isothermal curvesshown in the diagram is thus obtained. The areas enclosed bythese curves within the triangle represent the range within which,for the particular temperature for which the isothermal is drawn,heterogeneous mixtures will be formed. In other words, any mix-ture of the three components, benzene, formic acid, and water, thecomposition of which is represented by a point within such an araa,will separate into two layers. The figure obtained shows, in a veryB B 364 HEWITT, JOHNSON AND POPE: THE ABSORPTIONstriking manner, the extreme nature of the system under considera-tion. Even a t comparatively high temperatures heterogeneousmixtures are included in almost the whole area of the triangle. Itcan be seen a t once, for example, that below a temperature of 70°,only mixtures containing a relatively very small percentage ofbenzene are completely miscible’ with formic acid containing anappreciable amount of water. With comparatively large percent+ages of benzene the amount of water which can be present in thesystem is so very small as to be hardly capable of representationin the figure given.The influence of temperature on the system is seen to be asfollows :With mixtures containing a relatively large proportion of formicacid, increase of temperature permits of a somewhat wide variationin the amounts of benzene or water present in the homogeneousmixture. Where, however, a large proportion of benzene is con-tained in the mixture, the effect of temperature on the variationof the other two components is very small indeed.I n the analogous case already mentioned, of aniline, phenol, andwater, all three components are capable of variation over muchwider limits.The expenses of this investigation were, in part, defrayed by agrant from the Research Fund Committee of the Chemical Society,for which the author wishes to make grateful acknowledgment.The author’s t.hanks are due t o Dr. G. Barger for suggesting thesubject of this research and for his interest and advice throughoutthe work.THE GOLDSMITHS’ COLLEGE,NEW CROSS, S.E