MAGNETIC ROTATION OF SOME POLYHYDRIC AT~COHOLS. 17’7 XIX-The Magnetic Rotation of some Polyhydric By W, H. PERKIN, sen., Ph.D., F.R.S. THE remarkable changes in optical activity which many carbohydrates show when in solution in water have engaged the attention of several observers for a long period. To take an example, a freshly prepared solution of glucose has a rotation of [ u ] ~ + 105*16’, but this gradually diminishes and finally becomes constant after about six hours, the rotation being then [ a ] , + 52-49’ (Parcus and Tollens, Annalen, 1890, 257, 160). This phenomenon has been called bi-, multi-, or muta-rotation, and i t has been suggested by Tanret (Compt. rend., 1895, 120, 1060) that the first form of glucose should be called a-glucose and the second P-glucose ; this method of distinguishing the two modifications will be used in the present paper, not only in the case of glucose, but in all cases where birotation has been observed.A re‘surne‘ of the views which have been entertained in reference to birotation is given in a paper by Horace Brown and S, U, Pickering “On the thermal changes attending change of rotatory power of carbohydrates” (Trans., 1897, 71, 769). From this, it is seen that the earlier attempts to explain the phenomenon of bi- or multi-rotation were based on physical considerations. Subsequently, the probable chemical aspect of the matter came to be more fully discussed ; E. Fischer, for example, has suggested that the remarkable birotation shown by glucose may be due t o the gradual assimilation of water and conversion into the heptahydric alcohol, C6HI4O7, This view has latterly found considerable favour, and Brown and Pickering think that the results of the heat determinations made by them are con- sistent with it.As the study OP the magnetic rotations of the sugars might possibly throw some light on this difficult subject, it was thought desirable t o undertake the examination of some of the more important of these substances. Until lately, however, the measurements could not be Alcohols, Hexoses, and Saccharobioses. VOL. LXXXI. N178 PERKIN: THE MAGNETIC ROTATION OF SOME made with any degree of accuracy, because strong solutions of these sugars rotate the plane of polarisation through such large angles that, as is well known, the impurities in the sodium light seriously affect the appearance of the half-shadow disc of the polarimeter, causing the two sides to be very unequally tinted, so that useful numbers oannot be obtained.Thus, a 50 per cent, solution of fructose in a 100 mm. tube has an optical rotation of about 50°, and this is the point at which the magnetic rotation commences, Fortunately, after many attempts, I have succeeded in finding a simple spectroscopic arrange- ment by which this difficulty can be overcome, so that very large angles may now be measured with considerable accuracy, and with this new arrangement I have found it possible to determine accurately the magnetic rotations of a number of carbohydrates. I n a future com- munication, I hope t o give an account of this improvement and also of the new apparatus which I am at present using for the deter- mination of magnetic rotations, Besides the sugars themselves, two of the polyhydrio alcohols have been measured, so that the magnetic rotations of this class of com- pounds from the mono- to the hoxa-hydric are now known, with the exception of that of the pentahydric alcohol, C5H1,05, which, however, can be easily estimated, The examination of this series of alcohols was important in order that a basis might be obtained from which to calculate the probable rotation of the various sugars.The numbers obtained for the magnetic rotation of this group of alcohols may be briefly summarised as follows : Mol. mag. rot. Methyl alcohol,. ............. H,(CH*OH) 1.640 Glycol ........................ H,(CH*OH), 2.943 Glycerol ..................... H,(CH*OH), 4.1 11 Erythritol ..................H,(CH*OH), 5.230 Pentitol (missing) ......... H,(CH*OH), 6.300 est. Mannitol .................... H,(CH*OH), 7.351 If the magnetic rotations of the alcohols actually examined be plotted out, they form a regular curve, from which the rotation of the missing pentahydric compound may be calculated; also if the curve be carried further, the rotations of the heptahydric and other higher alcohols may be estimated, doubtless with considerable accuracy (see diagram), From this curve, it will be at once seen that the successive CH*OH groups have a smaller and smaller value as they are repeated; this, however, is not due to the group CH*OH as a whole, but to the hydroxyl group which it contains, since in the homologous series of paraffins, aliphatic acids, monohydric alcohols, and esters, it has been conclusively proved that the value of each CH,, even in com- pounds containing eighteen carbon atoms, is constant, namely, 1 *023.POLYHYDRIC ALCOHOLS, HEXOSES, AND SACCHAROBIOSES.1'19 Attention has previously been directed to the diminishing influence caused by successive displacements of hydrogen by hydroxyl (Trans., 1884, 45, 559); this diminishing influence is more clearly seen by subtracting from the value of the polyhydric alcohol that of the corre- sponding alcohol containing one hydroxyl less in its molecule. I n these cases, in which the magnetic rotation of the latter has not been 0.6 0.7 0.8 0-9 1 '0 1 '1 1-2 1.3 1.4 The m a p t i c rotations are found by adding the ordinates to the cudon num6ers of the abscGsoe.directly determined, it can be obtained by the addition of the value of CH, to that of the next lower alcohol, thus : Diff. for O H Mol. mag. rot. disp. H. ............ Glycol C?2H,tOH), 2'g43} 0.163 C,H,(OH), 4*111 ] 0.145 Glycerol Ethyl alcohol ... C,H,(OH) 2 9 3 0 ......... Less ......... CH, + C,H,(OH), 3.966 N 2180 PERKIN: THE MAGNETIC ROTATION OF SOME Diff. for OH Mol. mag. rot. disp. H. ......... CH, + C,H5(0H), 5.230 5.134 I O'Og6 7'351 } 0.028 Erythritol . . , . , . C,H,(O-H), Pentitol C5HZ (OH), 6'300} 0.047 Mannitol C,H,(OH), Less ......... Less ......... CH, + C,H6(OH), 6.253 Less ......... CH,+C,H,(OH), 7.323 ......... The influence of the hydroxyl group displacing hydrogen must, there- fore, evidently become practically nil when tbe substitution has been repeated seven or eight times.The results exhibited in the above tables will be found to be very important in the calculation of the probable rotations of glucose, fructose, &c. Glucose is known to be an aldehyde. Now the difference between the molecular magnetic rotations of an aldehyde and an alcohol, for example, between those of heptyl alcohol and heptyl aldehyde, is 0.438, so that the calculated rotation of glucose caii be obtained by subtract- ing this amount from that of mannitol." Mannitol .............................. '7.351 Less .............................. 0,438 Glucose ................................ 6.913 Fructose is known to be a ketone. The difference betweon the magnetic rotation of a ketone and an alcohol, for example, between that of sec.octy1 alcohol and of methyl hexyl ketone, is 0.495; this subtracted from the value for mannitol should give the rotation of fructose.Mannitol .............................. '7-351 Less .............................. 0.495 Fructose.. ............................... 6.856 - The actual determinations of the magnetic rotations of glucose and fructose in aqueous solution have given almost identical numbers in both cases, but the results are considerably lower than those calculated above. Glucose calc .......... 6.913 Fructose calc. ......... 6.855 Found ............ 6.723 Found ............... 6.729 Diff. .............. 0.190 Diff ................... 0-126 - - * The actual comparison should, of course, be between glucose and sorbitol, but the change of m e asymmetric carbon atom in passing from sorbitol to mannit01 would have, if any, so little effect on the magnetic rotation that it may be neglected.POLYHYDRIC ALCOHOLS, HEXOSXS~ AND SACCHAROBIOSES. 18 1 The question then arises : Why are the actual magnetic rotations of these sugars, debermined in solutions which have undergone the usual maximum change in optical rotation, lower than those calculated '2 IS this due to the assimilation of water and the formation of a heptahydric alcohol, C,H,(OH),, or must some other explanation be found '1 From the experimental part of the paper, it is seen that the magnetic rotation of glucose in aqueous solution, obtained by subtracting the value of 11H,O from that of a solution of the composition C,H,,O,,llH,O, is found to be 6.723, and the same number is found in a similar way from solutions of other concentrations.If, however, the glucose had assimilated 1 mol. of water from the solution to form the heptahydric alcohol C,H?(OH),, the rotation of this compound will be obtained by subtracting the value of only 10H,O from the result of the determin- ation, that is to say, it will be 7.723. From the examination of the curve (p. 179), it is clear that the rotation of the alcohol C,H,(OH)V will be 8-380 ; if from this we deduct the value for CH, (1.023), we obtain 7.347 as the value of the alcohol C,H,(OH)7, a number which is very different from that actually found, namely, 7.723. This evidence therefore seems to show that glucose in solution is the anhydrous substance C,HI2O,, and is not combined with water to form the hepta- hydric alcohol C,H,( OH),.Lowry (Trans., 1899, 75, 215), when referring t o the subject of birotation, suggests that the difference between glucose in the anhydrous condition and in solution, after all change has taken place and the optical rotation become constant, may be due simply to isomeric change, the aldehydic form I in the following table passing partly into one of the isomeric modifications I1 or 111, Of these expressions, formula I1 was first proposed by Tollens (Bey., 1883, 16, 923), and afterwards considered by E. Fischer as possibly, although not probably, repre- senting the constitution of anhydrous glucose. YHO $X€*OH $H*OH QH'OH ,+H~OH ?*OH YH*OH O\YH~OH VH*OH $?H*OH $?H*OH QH*OH ?€€*OH CH,*OH CH,*OH CH,*OH I.11. 111. If, however, formula 111 be examined, it will be seen that it re- presents an unsaturated compound, and this, according to the mag- netic rotation, cannot be correct. The introduction of an ethylene linking into the molecule of a saturated substance is known to raise FH-OH YH182 PERKIN: THE MAGNETIC ROTATION OF SOME the magnetic rotation by 1*620*, and the value of glucose (calculated from mannitol) would thus become 6.913 + 1.620 = 8.533, which is far higher than the value actually found (6.723). It has already been pointed out (p. 180) that the value for glucose in solution (6.723) is lower by Oa190 than it should be if the substance were an aldehyde, and the question then arises whether a compound of the formula I1 would have a lower rotation than one having the aldehydic formula I.That this will be the case can be shown from the following comparisons between the values found for glucose in solution, and those of ethylene oxide and the lactones, that is, of substances which are constituted somewhat similarly to formula 11. CH The value of ethylene oxide O<bHf, calculated from that of gly- Y col (2.943) by taking away 0.751 for the loss of the elements of water (see p. 184) is 2.192, the value found was 1.935, making a difference of 0.247 (Trans., 1893, 63, 490). I n the case of the lactones which have been examined, namely, butyrolactone and valerolactone, Butyrolactone. Valerolactone. the following are the differences between values found and calculated in a similar way: for butyrolactone - 0.230, and for valerolactone - 0.195, average, 0.212.Now the constitution represented by formula I1 agrees best with that of the lactones, inasmuch as it contains a chain of four carbon atoms closed by oxygen. If then glucose, when dissolved in water, assumes to a greater or less extent this constitution, there is good reason for believing that its rotation would be lower than that of the aldehydic form, 1, by about 0.2. This, it will be seen, agrees nearly exactly with the number actually found, and there is therefore strong support for the contention that, in solution, glucose has the constitu- tion represented by formula 11, or exists in some form analogous t o this.The solution would probably also contain a small quantity of glucose in its ordinary aldehydic condition ; it is therefore possible that the rotation of the /3-form in the pure state may be a little lower still than that found. The value for ordinary unsaturation with loss of H, is 1.112, but as no hydrogen is lost in this case, the value for unsaturation will be lq112+0*508, the value of H,.POLYHYDRIC ALCOHOLS, HEXOSES, AND SACCHAROBIOSES. 183 If now formula I1 be slightly modified, an expression for the POS- sible condition of fructose insolutionmay, in a similar way, be obtained which will be as follows : $.?H,*OH C*OH / I What has been rJaid about glucose applies equalIy well to fructose; the rotation is in bobh cases the same, and is lower than the calculated value, although not quite to the same extent ; it is therefore probable that fructose exists in solution, not as a ketone, but chiefly in a state represented by the above formula, or by some other formula similar to this, If we now consider the relationship between the calculated mag- netic rotation for glucose in its aldehydic form and that found for galactose in solution, we have the following numbers : GIUCOS~, calc........................... . 6 *9 13 Galactose, found ....................... .6'887 0.026 - I n considering these numbers, it should be noted that in optically active compounds, difference in configuration only does not appear to in- fluence magnetic rotation ; it is therefore probable that the magnetic rotation of galactose as an aldehyde is the same as that of glucose as an aldehyde, If, then, galactose in aqueous solution had been present entirely in its aldehydic form, the number found should have been 6*913, and the slight lowering observed in the value, namely, 0.026, appears to show that, whilst present for the most part in its aldehydic form, galac- tose has to some extent been converted into a modification similar to that represented by formula I1 in the case of glucose in solution.It is, however, remarkable that this small change appears to be accompanied by so large an alteration in the optical rotation, since galactose, which shows a rotation of approximately [a], + 134.5' in freshly prepared solutions, has a value of only [a]D + 84*2O, when the solution has been left to stand until the rotation has become constant, the formation of the small amount of the substance of formula I1 being accompanied by a fall in the optical rotation of no less than 50.3'.Thereis, however, no evidence to show what the optical rotation of substances of the type represented by formula I1 would be in the case of glucose,184 PERKIN: THE MAGNETIC ROTATION OF SOME fructose, or galactose. It is quite possible that such forms of the sugars, although similar in general character, might have very widely different optical rotations, and this is evidently the case, since fructose is lsvorotatory in solution, whilst glucose and galactose are dextro- rotatory in different degrees. Quite possibly a dextrorotatory alde- hydic sugar might yield a laevorotatory substance of the type repre- sented by formula I1 on going into solution, and this might be so in the case of galactose when it is entirely converted into its isomeric form.We have in nitrocamphor a remarkable instance of this kind of change, only of the reverse order; a-nitrocamphor, which is lsvo- rotatory, when changed into the isomeric $-nitrocamphor, becoming enormously dextrorotatory. Again, r-bromonitrocamphor in its normal condition has a rotation of [ a ] , -3S0, but in its pseudo-form has [a*]D + 188' (Lowry, Trans., 1899, '75, 223). The birotation of galactose is also much increased in amount by the addition of lead acetate to its solution, the rotation falling by 53 per cent. (Hanno Svoboda, Zeit. Ver. Rubenxucker.-Ind. Deut. Reichs, 1896, 46, A!t$t. 481, 29 pages; also Abstr., 1896, i, 406) I find also that a cold solution of caustic alkali reduces the rotation very considerably, As sucrose represents glucose and fructose less 1 mol.of water, its magnetic rotation can be easily calculated. The decrease in magnetic rotation caused by the loss of the elements of water when alcohol is converted into ethyl ether, acetic and propionic acids into their anhydrides, &c., averages about 0.752 (Trans., 1886, 40, 787), being in some cases a little less, and in others a little more than this ; therefore when this value is subtracted from those of the two sugars, the difference should approximately give the magnetic rotation of sucrose thus : a-Glucose + a-fructose, calc.. ....... .13*768 less H,O.. ............... .0.752 Sucrose calc............... .13416 found ..................... 12.586 - 0.430 From this it is seen that the experimental number is very much lower than the calculated. If, however, the experimental numbers of glucose and fructose in solution as @modifications be taken instead of those calculated for the magnetic rotation of the anhydrous or a-sugars, the following result is obtained :POLYHYDRIC ALCOHOLS, HEXOSES, AND SACCHAROBIOSES. 185 @Glucose + /I-fructose found ..... .13*452 less H,O ............... 0.752 Sucrose ..................... 12.700 found .................. 12.586 - 0.114 As the difference between the numbers actually found and those calculated in the above way is so small," it would seem that sucrose is apparently built up of the isomeric or /?-forms of glucose and fructose, and not of the aldehydic and ketonic forms. If, then, sucrose is built up of the isomeric forms of glucose and fructose, it will probably have the formula : $?H*OH 'VH CH,*OH CH,*OH and its constitution in the dry state and in solution will most likely be the same, since it does not exhibit the phenomenon of birotation.The above formula for sucrose has already been proposed by E. Fischer (Bey., 1893, 26, 2405); it is a modification of that suggested by Tollens (Ber., 1883,16,923), and clearly shows that when sucrose is hydrolysed it should at first be resolved into the isomeric or /I-modifi- cations of glucose and fructose : THOOH YH,*OH /?El *OH C*OH o<,YH*oH + o( ~ H - O H I FH*OH Q" FH \FH*OH CH,-OH CH,*OH &Glucose. B-Fructose. Naltose and Lactose.These sugars differ i n a marked manner from sucrose in that they possess birotatory and cupric reducing powers ; there can therefore be no doubt that they must have a structure essentially different from that of sucrose. * If, as supposed, the numbers found for these &compounds are a trifle high, on account of the solution containing a little of the a-compounds (see p. 182), this difference would be still smaller.186 PERKIN: THE MAGNETIC ROTATION OF SOME I n order to account for this difference, E. Fischer (Zoc. cit.) suggests TH,*OH YHO QH*OH FH*OH THOOH $?H*OH 0Q:::: $?H*OH CH--O----CH, Galactose radicle. Glucose radicle. I n this, the galactose radicle is represented as in the p- and the glucose radicle in the a-condition, whilst if this formula be applied t o maltose, one glucose radicle will be in the p- and the other in the a-condition.On investigating this matter, i t was a t first thought that the view of the difference in constitution between maltose and lactose on the one hand, and of sucrose on the other, received some immediate confirmation from the results of the magnetic rotations of the former, which are rather higher than the value obtained for sucrose; no doubt this has a bearing on the subject, but it is doubtful whether any great importance can be attached to this difference, From the fact, however, that these carbohydrates contain a glucose instead of a fructose radicle, their magnetic rotations should be about 0.05'7 higher than that of sucrose, the following formula for lactose : QH The rotations are as follows : Maltose, found ..................... ...12a690 Sucrose ,, ........................12.586 Lactose, found ...................... ..12.714 Sucrose ,, ....................... ,12586 L- + 0.104 - + 0.128 I f maltose be first considered, its magnetic rotation, on the assump- tion that its constitution is represented by the above formula, may be calculated thus : p-Glucose, found .......................... .6.723 a-Glucose, calc. ......................... .6*913 Less H,O.. .............................. .04'52 -- 13.636 Calculated mol. mag. rot. of maltose., .12*884 Found ................................... .12*690 Diff.. ................................... 0.1 94POLYHYDRIC ALCOHOLS, HEXOSES, AND SACCHAROBIOSES.187 This difference is almost exactly the same as that observed between a-glucose and P-glucose, 0.190 (p. 180), and points t o the probability that the second or a-glucose radicle in maltose also undergoes con- version, either entirely or in part, into the P-modification when the sugar is dissolved in water, and that the constitution of dissolved maltose is : QH,*OH CH*OH QH*OH /#?H*OH CH O < \ p * O H /CH~OH FH YH*OH '< CH*OH \b€€ 0 CH, The rotation, assuming that in a solution of maltose both glucose radicles are in the P-modification, may be calculated as follows : Mag. rot. of 2 mols. P-glucose ........... ,13.446 Less H,O .............................. 0.752 12.694 Found ................................ ,12*690 Diff. .................................0.004 The magnetic rotation of hcto8e~ as already stated, was found to be 12.714, and if this value be examined, it will be seen it aIso indicates that lactose in solution contains both the galactose and glucose radicles in the P-condition. It has been seen that galactose when in solution is chiefly in the a-condition ; if, however, it were principally in the P-condition,its rotation, no doubt, would be similar to that of @-glucose, so that the rotation of lactose should be the same, or nearly so, as that of maltose, and this is found to be the case, the difference being only + 0.020. In the dry state, it probably has the formula proposed by E. Fischer, and this is, of course, equally true of maltose. Very prob- ably these two carbohydrates, when in solution, always contain a little Qf the glucose radicle in the a- or aldehydic condition.EX PER IME NT AL. Erythritol, C4Hlo0,. The solution examined was supersaturated, containing 32.62 per cent. of erythritol, it being found possible to measure its rotation before crystallisation set in ; the composition of the solution was C4HIoO, + 1 4H20. This substance was purified by recrystallisation from water. Density, d 1So/lS0, 1.1043; d 20"/20°, 1.1033.188 PERKIN: THE MAGNETIC ROTATION OF SOME The average of three sets of determinations of the magnetic rotation made a t different times was : t. Sp. rot. Mol. rot. of sol. Mol. rot. of C4HI0O1' 15' 1.0220 19.230 5.230 Mannitol, C6H1,06. As in the case of erythritol, a supersaturated solution was employed ; it contained 18.176 per cent.of mannitol, its composition being C6H140, + 40H,O. This was recrystallised from water before use. Density, d 15'/15', 1.0752 ; d 2Oo/2O0, 1.0746. The average of four sets of measurements of the magnetic rotation made at different times gave : t. Sp. rot. Mol. rot. of sol. Mol. rot. of C,H,,O,. 17 -5' 1.0154 47.351 7.351 Ghcose, C6H1206. Two specimens of this substance were examined, one obtained from Kahlbaum, and the other, a very pure preparation, for which I am indebted to Dr. Horace Brown. With the former, four sets of measurements were made on different occasions and with solutions of various strengths, the most dilute being represented by C61E1206 + 20H,O, and with that from Dr. Horace Brown also four measurements were made, but with only one strength, represented by C6H120, + 11H20 and containing 47,619 per cent.of C6HI2O,. The products used were the monohydrate dried over sulphuric acid in a vacuum : The density of the solution C6H1206 + 11H20 was d 15'/15', 1.2147 ; d 2Oo/2O0, 1.2135. Magnetic rotation : t. Sp. rot. Mol. rot. of sol. Mol. rot. of C,HI,O,. 15' 1.0261 17.723 6.723 The average of the measurements made with Kahlbaum's specimen was 6.715, which is very close to the above. The permanent optical rotation of the solution containing 47.61 9 per cent, of C6H1,06 was [.ID 66-22' at 16.9'. This is a little higher than that given for weak solutions. If the magnetic rotations be calculated on the assumption that the glucose has assimilated a mol. of water and thus become a heptahydric alcohol, the solution will then have the composition C6H1407 + 10H,O.The calculation will be the same as the above, only the value of 10 instead of 11 mols. of water will have to be subtracted from the mole-POLYHYDRIC ALCOHOLS, HEXOSES, AND SACCHAROBIOSES. 189 cular rotation of the solution, and the rotation of the alcohol will thus become 7.723. Fmxose, C6Hl2OS. It was dried over sulphuric acid and its composition checked by a combustion ; it gave C, 39.8, and H, 6.8, the formula C,H1,O, requiring C, 40.0, and H, 6.7 per cent. Its solution was examined in one strength only, containing 50 per cent, of fructose and represented by C,Hl2O6 + 1 OH,O. This was prepared from inulin and obtai-ned from Kahlbaum. Density, d 15'/15O, 1-2226 ; d 2Oo/2O0, 1.2211. The average of five sets of measurements of the magnetic rotation, made on different occasions, gave : t.sp. rot. MOl. rot. Of Sol. MOl. rot. Of C,3H,&,3. 15' 1.0227 16*729 6.729 Optical rotation [.ID 96.19' a t 15'. Galactose, C,H1,06. This substance was examined in a very supersaturated solution, from which it does not crystallise very quickly. It contained 50 per cent. of the sugar, its composition being represented by C,H,,O, + 1 OH,O. Density, d 15"/15", 1.2311 ; d 20°/200, 1.2299. The average of four sets of measurements of the magnetic rotation, made on different occasions, gave t. Sp. rot. Mol. rot. of sol. Mol. rot. of C6H120B. 15' 1.0396 16.887 6.887 Optical rotation [a], 84.23' at 14.6". Sucrose, CI,H22011. The specimen used was ordinary sugar recrystallised from alcohol The composition of the solution used was represented (75 per cent.).by C12H,,011 + 19H,O, and contained 50 per cent. sucrose. Density, d 15'/15', 1.2327 ; d 2Oo/2O0, 1-2316. The magnetic rotation, determined on four different occasions, was : t. Sp. rot. MoI. rot. of sol. Mol. rot. of C12HaOll. 15' 1 *0247 31.586 12.586 Optical rotation [a ID 66.51' at) 17". VOL. LSXXI. 0190 MAGNETIC ROTATION OF SOME POLYHYDRTC ALCOHOLS. Maltose, C1,H,,O1l. For a very pure specimen of this compound, I am indebted to Dr. The solution employed contained 47% per cent. of the Horace Brown. sugar, its composition being represented by Cl2H2,OI1 + 20H,O. Density, d 15"/15', 1.2214 ; d 20°/200, 1.2205. The average of three sets of measurements of the magnetic rotation, made on different occasions, gave : t.Sp. rot. Mol. rot. of sol. Mol. rot. of CI2HnOl1. 15' 1.0288 33*690 12,690 Optical rotation [a]= 137.0' a t 16.7'. Lactose, C,, H,,O,,. This was obtained from Kahlbaum and was purified by fractional crystallisation, the crop deposited during the first 12 hours being rejected. The solution used was a supersaturated one containing 33,333 per cent. of the sugar, and had the composition C,,H,,O,, + 41H,O. Density, d 15"/i5O, 1.1413 ; d 20°/200, 1.1406. The magnetic rotation, as determined four times on different occa- It was dried in a vacuum over sulphuric acid. sions, was : t. Sp. rot. Mol. rot. of sol. Mol. rot. of Cl2H2?OI1. 18.4' 1.0213 53.714 12.714 Optical rotation [a]D 52.6' at 18'. Summary. The chief results obtained in this investigation go to show : (1) That the influence of successive hydroxyl groups in polyhydric alcohols on the magnetic rotations diminishes as they increase in number, until about the seventh is reached, when it becomes almost nil. (2) That solutions of glucose and fructose, after all change has taken place, give magnetic rotations which indicate that birotation is not due to hydration, but that it is caused by a change in the constitution of these substances. (3) That galactose, when in solution, does not undergo isomeric change to so large an extent as glucose. (4) That sucrose is built up of the isomeric or @forms of glucose and fructose by the elimination of the elements of a mol, of water.HYDROCYANIC, CYANIC, AND CYANURIC ACIDS. 191 ( 5 ) That maltose is formed from 1 molecule of glucose in the aldehydic or a-condition and 1 molecule in the isomeric or @-condition by the elimination of the elements of a mol. of water and that lactose is similarly derived from 1 molecule of a-glucose and 1 of @galactose, both being constituted in a similar manner t o that proposed by E, Fischer for lactose, also that when in solution these sugars undergo isomeric change, the a-portion becoming transformed, more or less, into the @-condition. This change accounts for the birotation and cupric reducing power of the two sugars.