INFLUENCE OF SOLVENTS ON ROTATORY POWER. 971 CI1I.-The In,jhence of Solvents on the Rotatory Powers o f Ethereal Dimet~~oxyszcccinates and Tai-trates. By THOMAS PURDIE, F.R.S., and WILLIAM BARBOUR, B.Sc., Berry Scholar in Science. THE observations recorded in this paper were made with the view of comparing the effect of solvents on the rotations of ethereal tartrates and their alkylated derivatives, the ethereal dimethoxysuccinates. The facts already known concerning the influence of solvents on optical activity, more particularly the extended observations of Freundler on the ethereal diacyl tartrates (Ann. Chim. PAYS., 1895, [vii], 4, 235), and the recent systematic investigation by Patterson (Trans., 1301, 79, 167, 477) of the influence of various solvents on ethyl tartrate, suffice to show that the action in question, even in the case of non- electrolytes, is very complex, and that at present no general theory on the subject can be formulated.The phenomena presented by the esters of tartaric and other hydroxy- acids are probably much complicated by the presence of the hydroxyl group, which is well known to exercise a marked influence on physical properties in general, and it appears from Freundler’s researches that, even when the hydroxyl is acylated, variations of optical activity occur under the action of solvents, for which it is difficult to discover a satis- factory explanation. Freundler, for instance, was led to conclude that the effect of benzene on the rotations of the diacyltartrates was con- nected with a peculiar form of dissociation, in which the acyl groups were split off from the compounds. It seems probable that the alkyl- ated esters of hydroxy-acids 5hould present phenomena of n less com- plex kind, and that they should be less subject, at all events, to the possible disturbing effects of molecular association than the parent esters.We purpose, therefore, investigating the action of solvents on the rotations of mono- and di-alkyloxysuccinic and alkyloxypropionic esters, and in the present preliminary paper we record observations on solutions of the three dimethoxysuccinic esters described in the pre- ceding communication, in water, methyl alcohol and benzene, We972 PURDIE AND BARBOUR: INFLUENCE OF SOLVENTS ON have also made some observations on solutions of the corresponding tartrates, but have refrained from proceeding further with these compounds, as we understand that Mr.Patterson's investigation, alluded to above, will include the whole series of tartaric esters, Conversion of d-Dirnethoxpwccinic Acid into d-Tartar ic Acid. For a comparison of the rotations of the tartaric and dialkyloxy- succinic esters, it is essential that the relationship of the two classes of compounds, with respect to configuration, should be known with certainty, As indicated in the preceding paper, i t is very improbable that in the process of alkylating the esters of optically active hydroxy- acids with alkyl iodide and silver oxide, any inversion of configuration occurs, but we thought it well, nevertheless, to set the question at rest by reconverting dimethoxysuccinic acid into tartaric acid.5 *5 grams of d-dimethoxysuccinic acid were accordingly heated with ten times the weight of fuming hjdriodic acid in a sealed tube at looo for 8 hours. The production of an oily layer, smellingof alkyliodide, indicated that the desired reaction bad occurred. The aqueous layer, after most of the hydriodic acid had been removed from it by heating in a vacuum, deposited a crystalline acid on standing in a desiccator over potash ; this, after washing with ether to remove iodine and re- crystallising from water, melted at 167'. The melting point of tartaric acid is 167-170'. A determination of the specific rotation in aqueous solution at 20° gave c=4-097, Z=2, U = +1-24', [ u ] y = +15-13', According to Landolt (Bey., 1873, 6,1075), the specific rotation of tar- taric acid under these conditions is 14-52', The sparingly soluble potassium hydrogen salt of the acid gave in aqueous solution a t 20' c=O.615, Z=4, U = + 0 * 5 6 O , [~.t]r= +22-76', the value quoted by Landolt for potassium hydrogen tartrate at the same concentration being +22-6l0. On analysis, the salt gave C = 25.66 ; H = 2.80 ; K = 20.47, 20.84.C,H,O,K requires C = 25.51 ; H = 2.66 ; K = 20.81 per cent. d-Dimethoxysuccinic acid yields d-tartaric acid by the treatment described, and as the esters in both cases are active in the same sense as their acids, it follows that ethereal dimethoxysuccinates and tar- trates of similar sign of rotation have the same configuration. I n the tables below, we give the results of our polarimetic observa- tions.The dimethoxysuccinic esters used were prepared in the same manner as the specimens described in the preceding paper, and, as will be seen from the numbers quoted, they showed practically the same specific rotations as the previous preparations. The tartrates used were obtained from Kahlbaum, and purified by fractional dis-BOTATORY POWER. 973 Water ................. tillation. I n calculating the speci6c rotation of these esters, the specific gravities quoted by Yictet mere used. The observations were taken in a 2 dcm. tube at 20'. The specific gravity of the benzene used in the solutions mas d 20°/4*=0*8785. The methyl alcohol employed was Kahlbaum's, dried by distillation from calcium oxide and from sodium; the specimen used in the observations on methyl dimethoxysuccinate had the specific gravity d 2Oo/4O = 0.1927, but that used in the other observations, having accidentally absorbed moisture, had the specific gravity 0.7946.The concentration, c (grams of substance in 100 C.C. of solution), was determined in a few instances by making up the solution of a weighed quantity of substance to a known volume, but in most cases i t was found from p (grams of substance in 100 grams of solution), and the specific gravity of the solution d 20°/40. Met ?h y 1 D inze t hoxy szc c cin CL t e. It was found impossible to take observations on the pure ester in the superfused state, but the following approximate data may be quoted, which are calculated from observations at 60' on the assump- tion that the changes with temperature are the same as in the case of the ethyl ester.19 9988 I I c. Solvent. I -___ ...............I 11.1694 Me&yl alcohol ....,.I 23'0151 ...... I 12'0806 " . * * * * I 9 ) ,, 9 3 1 6'2598 d 20"/4". - - - - - - 0.9033 0.8591 0.8276 0.8102 ~~ [a]?. + 78.71" 78q45 78'50 101-63 104'66 105.47 104 '42 78.90 76'32 81'04 - [ M ]iO". + 162.1 161-6 161 *7 209.4 215'6 217'3 215.1 162-5 157'2 166.9 Mol. sol. vol. - - - - - - 176.9 172.6 169'2 170.2974 PURDIE AND BARBOUR: INFLUENCE OF SOLVENTS ON [M1~OO- &'thy I dimethoxysuccinate. [a]:= -t89*7", [MI:= +209'9", d 20"/4"=1*0976, ill --2213*2. - d Mol. sol. vol. Solvent. I c. 1 d 20"/4". Water .................. Methyl alcohol ...... Y t ), ...... Benzene ...............) ) , .............. ,, ............... 2 ) ) ) ...... 5'3752 - P. 19.0407 0.8407 9'7365 0.8178 6.0569 0.8089 19.3137 0.9142 10.1117 0'8969 5'3130 0.8879 [a 1:. + 6.12" 6.29 6.75 + 89.1 1" 87.27 87'41 87-66 102.65 104'14 104'93 + 12%" 13'0 13.9 Benzene ............... ) ) ............... ,, ............... I 21 -0627 10.7339 5.6479 + 208 '5" 204'2 204-5 205.1 240.2 243.7 245.5 - 209.7 208.7 208-5 212% 21 2-3 213.3 Solvent. Methyl alcohol ...... Benzene ............... 2 ) )) 2 ) , , ...... ...... , , ............... ,) ............... p. 1 d 20"/4". 23.7085 12 *3697 6.6571 21'561 9 11.4730 5'6932 0,8479 0.8213 0.8086 0.9136 0.8966 0.8873 + 85-79" 84.50 84'99 99'24 101*00 101 '26 I + 224.8" 221 '4 222.7 260.0 264.6 265.3 242.3 243.1 244'0 245'1 245.7 246.3 Methyl tartrate.[a]:oo= +2'14", [M]iO"= +3*8", d 20"/4"=1-3284, '11=134-0. d Solvent. Water .................. 1 5'0231 1 1.0131 1 +20'04" I +35*7" 1 126.3 Ethyl tartrate. M d [a]?= +7%2", [MI?= +15'7", d 20"/4"=1~2059, -=170.8. Solvent. 1 p . 1 d 20"/4". I I 0.9308 0.9038 0.8913 i 1 Mol. sol. vol. 171.9 173.4 174'9ROTATORY POWER. Water .................. Benzene ............... ,, ............... ,, ............... ,, .............. 975 4.8206 1 *0071 22.2112 0'9257 11'1266 0'8993 5.4685 0.8890 5-6205 o m 8 6 Propyl tuvtrate. [a]?= + 12-31", [ M]to"= + 28'81", d 20"/4"= 1-1344, ?!= 206'3. d Solvent. Solvent. -1 p. 1 d 20°/4". Tartrate. Dimethoxysuccinate. I I Water ........................ ,, ........................... ,, ........................... Benzene ..........................Methyl alcohol ................. Methyl + 31 -9" - 17'5" Ethyl $38'2 * - 1'4 Propyl -1-33-6 slightly affected $ Ethyl -k 7.9 * - 4.8" Methyl - 19'5 j- +- 38'1 [ u1p. + 26'67' '18'31 20.34 20.78 19.62 [ M + 62.4" 42 *8 47% 48'6 45'9 Mol. sol. vol. 191.7 205'2 211'0 208*8* 212'53 * This number for molecular solution-volume is probably too low. f The specific rotation founcl by Freundler in about a 5 per cent. solution was 3- 20 -1". So far as the available data go, it may be said that, in general, the rotations of the dimethoxysuccinates are less affected by solvents than those of the corresponding tartrates, and that the action is in most cases oppositely directed. The following table in illustration of the statement shows the changes produced on the molecular rotations by the solvents mentioned at concentrations of about 5 per cent., a rise or fall of dextro-rotation being indicated by + or - respectively.I 1 * Patterson (Zoc. cit.). I- Freundler (Zoc. cit.). $ The ester is very slightly soluble in water, and the conclusion is based on an ob- servation made in aqueous methyl alcohol. Attention may be drawn to the following facts concerning the dimethoxysuccinates. Tbe three esters examined undergo a consider- able rise of rotation when dissolved in benzene, and the rotations increase with dilution, the lowest member of the series being most affected by the initial action of the solvent, and also most affected by change of concentration. The greater influence of the benzene on the fist member of the series results in the extinction of the maximum specific rotation at the ethyl term, which is shown by the pure esters,076 PURDIE AND BARBOUR: INFLUENCE OF SOLVENTS ON The effect of water and of methyl alcohol on the rotations of the three esters is, in general, in the opposite sense to that of benzene, and the first member of the series is again most affected.The rota- tion in w@ter, unlike that in benzene, is scarcely affected by ohange of concentration, at least from 20 per cent. downwards. The observa- tions on these esters in methyl alcohol present two points which are worthy of notice. First, the rotation-concentration curves of the methyl and propyl compounds exhibit a minimum, more pronounced in the curve of the former, between the concentrations of 20 and 5 per cent., a peculiar phenomenon of which Patterson (Zoc.cit.) has found striking examples in certain solutions of ethyl tartrate. The absence of a corresponding minimum in our ethyl ester is probably due to the range of concentration examined, in the case of this compound being less than in that of its two homologues. Secondly, it will be seen that whilst methyl alcohol lowers the rotations of the methyl and ethyl esters at all the concentrations examined, it slightly raises that of the propyl ester a t a concentration of 20 per cent, This peculiarity is referred to below. Finally, it may be remarked that, owing to the greater lowering effect of methyl alcohol on the first member of the series, the specific rotations of the three compounds in this solvent exhibit a more pronounced maximum at the ethyl term than those of the liquid esters.Turning to the tartrates, Freundler (Zoo. cit.) has shown that benzene in 5 per cent. solutions lowers the specific dextro-rotation of the methyl ester from + 2*14O to - 8*S0, but raises that of the propyl compound from + 12-44O to + 20*lo. H e remarks, with regard to these effects, that they are entirely irregular, It seemed to us very remarkable that the rotations of two closelyrelated members of the samehomo- IogouB series should be influenced to such a marked degree in oppo- site directions by the same solvent, and we therefore examined the effect of benzene on the intermediate ethyl ester, and also made further observations.on the prop91 ester in the game solvent. Our results with the latter compound confirm those of Freundler, and show besides that the rotation increases somewhat on dilution.We find that, with respect to the effect of benzene, the behaviour of the ethyl ester is intermediate between that of its two adjoining homologues. The specific rotation is only about 1' less than l-hat of the pure compound, and is but little affected by concentration. The effect of methyl alcohol on the dimethoxysuccinates, referred to above, gives indications of a similar reversal of the influence of the solvent on ascending a homologous series, and we are led to think that the phenomenon, to which,so far as we know, attention has not hitherto been drawn, will be found to be one of general occurrence, Freundler's data (Zoc. cit.) furnish a number of examples,ROTATORY POWER.977 of which the propyl diacetyl-, dibutyryl-, and dihexoyl-tartrates may be mentioned. Dihexoyl- Diacetpl-, Dibntyryl-, tartrate, [ €2 I”. c .I,. c U l D . Pure ester ..................... + 13.4’ + 5.2’ + 2.2’ Solution in methyl alcohol. , 12.1 9.3 5.4 ,, acetone ......... 10.4 7.2 5.3 ,j ethyl alcohol ... 9.6 6.3 3-6 Similarly, for the diacetyltattrates, alcohol and acetone lower the rotations of the ethyl and propyl and raise that of the butyl ester; also for the dipropiohyltartrates, the same solvents lower the dextro- rotatidn of the methyl ester, and raise that of the ethyl ester. Other available data indicate that, on ascending a homologous series, the effect of the solvent on the rotations tends, at least, towards a minimum, and would probably become reversed at a higher point in the series; Patterson’s observations on ethyl tartrate, and our own on the methyl and propyl esters in 5 per cent.aqueous solution at Z O O , afford an example, thus : Methyl, Ethyl, ProPYl, [ a ID’ [QID’ [ a I”. Pure ester ..................... + 2.1” 7.6” 12.3” Solution in water ............ 20*0 26.2 26.7 The cause of the phenomenon referred to is probably related to that which determines the occurrence of a maximum rotation in a homolo- gous series of liquid compounds. It might be expected that the rela- tion between the effects produced on the initial members of a series by the action of a solvent and by the addition of CH, to the active molecule of the pure liquid would be retained throughout the series, and that the effect of the solvent, like that of the addition of CK,, would therefore be reversed about the point of maximum rotation, I n the case of the tartrates, the reversal of the effect of benzene and the maximum rotation occur, as a fact, a t the same, namely, the propyl term.In general, however, the points of reversal of the influence of the solvent and of maximum rotation do not coincide. Thus, if the effect of water on the rotations of the tartrates under- goes reversal, it must be at a term higher than that of the maximum rotation; the same is possibly the case in the series of propyl diacyl- tartrates quoted above, where, as the rotations decrease from the diacetyl term upwards, i t is perhaps admissible to conceive that the point of maximum rotation lies a t or below this term. Although the facts mentioned seem to us to indicate a connection between the effect of solvents and that of the addition of mass on the rotations of the auccessive members of a homologous series, it does not follow that the978 PURDIE AND BARBOUR: INFLUENCE OF SOLVENTS ON Water .................) ) .................. ,, .................. Benzene ............... y y ............... Eth);lene dibromide. ,? ,, . ..... : ........ Y ) ,) . mechanism of the actions in altering the state of dissymmetry of the active molecules is the same, and i t is therefore not to be expected that the points of maximum rotation and reversal of the effect of the solvent should coincide. We give below the results of zt number of molecular weight deter- minations by the freezing point method in water and benzene, made with the view of discovering, if possible, a connection between the influence of the solvents on rotation and the molecular association of the dissolved active substance.S denotes the number of grams of substance in 100 grams of solvent. 3.6050 7.7713 12,7460 0'8853 24970 4.6370 0.6788 1-5300 2.3630 Solvent. 1p-l- M. Methyl dinzethoxyslcccinate, M= 206. Benzene ............ , ............ ) ) ............ ) ) ............ 1.6377 3 -4079 7'2093 7'8837 190-9 193'8 188'4 193.7 195'4 203'8 202-8 213-2 216'8 175.3 176'4 1734 Water .................. ,, .................. ), .................. 2554 296'2 34 2 *8 6'6313 5.1511 9.9887 Ethyl dimethoxysuccinate. M= 234. Benzene ...............,, ............... ,) ............... Benzene ........... ), ........... ) ) ........... 1.8194 3'8953 7'5564 3.3735 5.0373 9'0774 215.8 219-4 225'0 Propyl dimethoxyncecinate. M= 262. Benzene ............ ) ) ............ ) , ............ 1,0649 3,7210 4.9748 231 -9 236-9 241 -3 244.3 220-2 281'4 299'8 Freundler, it is well known, found that in a number of substances normal molecular weight in solution was accompanied by normal rota- tion, abnormal molecular weight by abnormal rotation. Patterson (Zoc. cit.) has shown, however, in the case of ethyl tartrate in various solvents, that normal molecular weight may coexist with abnormal rotation, and our observations furnish various instances of the same kind. The rotations of the three dimethoxysuccinates are consider- ably raised by benzene, and that of the methyl ester lowered by water, but the molecular weights are normal in these solvents so farROTATORY POWER.979 as the freezing point method can detect. Water increases the rotation of methyl tartrate by about ten times the value of the constant of the pure compound, but the molecular weight is nevertheless normal. The change here, it is true, may be due to depolymerisation, but the similar effect produced by water on the rotation of propyl tartrate cannot be thus explained, as there is no reason t o believe that this ester in the pure state is associated. Ethyl tartrate in a 5 per cent. benzene solution presents an exception to Freundler’s rule of another kind, for nearly normal rotation is here associated with distinctly abnormd molecular weight.The molecular weight determinations given above, and those by Freundler (Zoc. cit.), show that the tartrates, owing, no doubt, to the presence of the hydroxyl group, are subject t o extensive molecular association, even in comparatively dilute benzene solutions, and %hat in the alkylated esters this tendency to association entirely disappears. If molecular association is a predominant factor in the changes of rotation produced by solution, it should be clearly manifested in the rotations of the two closely-related classes of esters in the solvent mentioned. Such, however,-is not the case. The rotations of the dimethoxysuccinates, as already mentioned, are considerably increased by benzene, and those of the tartrates are modified in a manner that cannot be accounted for by the simple aggregation of active molecules as they exist in the pure liquid compounds, Methyl tartrate in benz- ene (Freundler, Zoc.cit.) exhibits extensive association accompanied by depression of rotation, propyl tartrate (see our tables) association accompanied by rise of rotation, and ethyl tartrate association, increasing rapidly with concentration, accompanied by only slightly changed rotation, which varies little with concentration. The coalescence of the active molecules of the ethyl ester evidently pro- duces little or no effect on their state of dissymmetry; it is natural to conclude that the same holds true for the molecules of its two homo- logues, and that the reversed effect of benzene on the rotations of the methyl and propyl esters is due mainly to an initial specific action of the solvent on the state of dissymmetry of the simple molecules of the compounds. Patterson, in the suggestive paper already referred to, has attempted, with some considerable success, to trace a connection between the influence cjf solvents on rotation and the internal pressure of the liquids, and more particularly between rotation and molecular solution- volume, which, on the assumption that the change of volume on solu- tion is suffered entirely by the dissolved substance, may be regarded as a measure of the pressure in question.Our observations are not extensive enough, and so far as the densities of the more dilute sola- tions are concerned, probably not accurate enough, to be employed as980 PURDIE AND BARBOUR: INFLUENCE OF SOLVENTS ON in any sense a crucial test of this theory; we have, however, calcu- lated the molecular solution volumes by the usual formula, when densities were available (see the Tables, pp. 973-975), to ascertain if the results gave any support or otherwise t o his views. For ethyl tartrate, to which so far his observations have been confined, Patterson finds that the values of the rotations in aqueous and different alcoholic solvents at infinite dilution stand in the inverse order of the molecular solution-volumes ; further, that solvents, such as methyl alcohol, which raise the rotation on solution, lower t8he molecular volume below that of the pure compound, whilst octyl alcohol, which causes a distinct fall of the rotation, raises the molecular volume.2limethoxysuccinutes.-The rotations of the methyl and ethyl esters are both lowered by methyl alcohol, as already stated, but the former decidedly more than the latter, while that of the propyl ester is only slightly affected. The molecular solution-volumes in this solvent are less than the molecular volumes of the pure esters. Benzene raises the rotations of all three esters considerably, but influences the mole- cular volumes very slightly. As exact data for the pure methyl ester at 20" are not available, in order to ascertain if any connection can be traced between rotation and molecular solution-volume in this group, we have compared the effects produced on the two constants by methyl alcohol and benzene respectively.We give below the differences be- tween the molecular rotation and molecular solution-volume of each ester in the two solvents in about 10 per cent. solutions a t 20". Diff. [M ];O". Diff. M.S,V. Methyl ester ............... 579" 7.7" Ethyl ,, ............... 39.2 3.6 Propyl ,, ............... 43.2 2.6 These figures certainly indicate a connection between the constants. The methyl ester, which shows the greatest difference of rotation, also gives the greatest difference of molecular solution-volume. The differ- ences of molecular solution-volume exhibited by the ethyl and propyl esters, it is true, do not stand in the same numerical order as the rota- tions, but the respective data for each ester approximate closely to each other, and the apparent discrepancy may well be due to slight error in determining the densities.Our observations on the methyl ester in methyl alcohol at varying concentrations appear also to indicate a connection between the con- stants in question. The rotations and molecular solution-volumes vary in the same order, each showing a minimum at the 10 per cent. solu- tion, but me do not attach great significance to this, as the range of concentration in the solutions is not sufficiently great, and the propyl ester does not show a corresponding regularity.ROTATORY POWER. 981 l'urtrates.-We have refrained from proceeding further with observ- ations on tliese substances, as we wished to avoid trespassing on Mr. Patterson's field of work. The following conclusions may, however, be drawn from the observations made, The effects of solution in water on the rotations and molecular solution-volumes of methyl and propyl tartrates correspond with those found by Patterson for ethyl tartrate, namely, a notable increase of rotation and decrease' of volume, but as will be seen from the numbers below, which refer to 5 per cent.solutions a t 20°, no connection between the two classes of effects is apparent. Increase of [ M 11. Decrease of M. V. Methyl tartrate ............ 31.9" 7.7" Ethyl ,, ............ 35.2 11.9 Propyl ,, ............ 33.6 14.6 The rotation of ethyl tartrate is slightly lowered by benzene, that of propyl tartrate much raised, but no corresponding change of molecular volume is evident ; as in the case of ethyl tartrate in octyl alcohol, the molecular solution volumes are greater than the molecular volumes of the pure liquids, and increase with dilution.Conzpcwison of Etlql Tcwlrccte and EthyZ Dimethoxysucciizccte. Below, we give the changes of rotation and molecular volume pro- duced by solution of these esters in methyl alcohol and in benzene at a concentration of about 5 per cent. and a t 20'; numerical rise and fall are indicated by the signs + and - rcspectively. Change of [MI:*". Change of 1I.V. Ethyl tartrate in methyl alcohol *... ......... + 7.9" - 11.5' ,, dimethoxysuccinate in methyl alcohol - 4.9 - 4.7 Ethyl tartrate in benzene ..................... - 1.8 + 4.1 ,, dimethoxysuccinate in benzene ...... + 35.6 unaltered The rotation of ethyl tartrate is more raised by methyl alcohol than that of the dimethoxysuccinate is lowered, and in agreement with this the tartrate shows a greater change of molecular volume.I n benzene, the tartrate also shows the greater change of molecular volume, but the effects on rotation, it will be seen, are in entire disagreement with this. The facts given above furnish no clear evidence of a definite connec- tion between rotation and molecular solution-volume, and we doubt if any such connection will be found to prevail generally. Patterson has pointed out several disturbing factors, which might account for the * Pattersoil (Zoc. dt.). trot. LXXIX. 3 x982 THORPE AND HOLMES: THE OCCURRENCE OF apparently abnormal relations between rotation and molecular solution- volume which he eccountered in some of his solutions. Another dis- turbing phenomenon suggests itself t o us, which may be of frequent occurrence. Patterson’s theory is based on the idea that a progressive change in the volume of an asymmetric molecule will be accompanied by a corresponding progressive change of shape, and therefore also of rotation. According t o Guye and Crum Brown, a continuous change in the value of one of the four coefficients of the groups attached to the asymmetric carbon atom, such as occurs in ascending a homologous series, may cause the rotation to oscillate between maxima on either side of the zero point. When the volume: and therefore the shape, of an asymmetric molecule is altered by the action of a solvent, no doubt the value of all four co- efficients will undergo change, but one of these will probably be more subject to the action than the others. A progressive change of mole- cular volume, whether caused by change of concentration or by the action of a succession of different solvents, might therefore result in a periodic change of rotation. This may possibly account for what seems to be a common phenomenon, the occurrence, namely, of a point of minimum rotation in the concentration-curves of solutions of active compounds, We doubt if this will be generally the case. UNITED COLLEGE OF ST. SALVATOR AND ST. LEONARD, UNIVERSITY OF ST. ANDREWS.