BROWN, MORRIS AND MILTAAR : THE SOLUTION-DENSITY, ETC. 275 XX I I. - The So 1 yc t i o j ~ - dens it y a jzcl Ciip~ic- yedir c in 9 Power of Dextrose, Lcvulose, and Ikce rt-suguir-. By HORACE T. BROWN, F.R.S., G. HARRIS MORRIS, Ph.D., and J. €I. MILLAR. The Xolutio./2-clensities. IN a previous communication (Trans., 1897, '71, p. 72), we have entered fully into the necessity for the accurate determination of the solution- densities of the carbohydrates, in order that the correct factors for the estimation of the solid matter in carbohydrate solutions may be de- duced. I n the paper above referred to, we gave the results of the accurate determination of the solution-densities and divisors for various concentrations of maltose, soluble-starch, and the products of the con- version of starch by diastase.We have now carried out a series of similar determinations for dext,rose, levulose, and invert-sugar. The apparatus used for drying the respective sugars was that described in our former paper (this vol., p. 76). I n this, crystallised dextrose can be rendered perfectly dry in 4 hours at 105-106" without the slightest decomposition or coloration taking place. Crystallised levulose can be dried still more readily, but the operation requires greater care, on account of the low melting point of this sugar. The temperature of the bath must not be allowed to rise27G BROWN, MORRIS, AND MILLAR : THE SOLUTION-DENSITY AND above 75" for the first three or four hours, but after that it may rise to 85-90'. The latter temperature must not be exceeded, or the substance fuses and begins to colour.A total period of 6 hours a t the given temperatures is amply sufficient to drive off all moisture without the slightest discoloration taking place. The sugars used in the experiments were prepared with the greatest possible care by the most approved methods, and were repeatedly recrystallised from alcohol u n ti1 their properties were absolutely con- stant. The dextrose was prepared from three sources, namely, from pure maltose by acid hydrolysis ; from cane-sugar by inversion with acid, and from a crystallised sample of dextrose obtained from Kahl- baum. The levulose mas prepared in two mays, the one from inulin in the manner described by Wohl (Bey., 1890, 23, 2107), and the other from a sample of Schering's crystallised product.Invert-sugar was obtained by the inversion of pure, crystallised cane-sugar with yeast, asd in some experiments, by mixing equal quantities of perfectly pure dextrose and levulose. I n the results given in the following tables, Column a gives the weight of dry substance taken, Column b gives the total weight of solution. Column c gives the specific gravity of the solution at 15.5", referred Column cl gives grams of sugar per 100 C.C. (reputed)". Column e gives the divisor for the determination of grams per 100 to water at 15.5". C.U. (reputed) from the specific gravity. Fuble I.-Solution-density of dnhydrous Dextrose. I a. - ' I 1 '2988 1-6324 3'2659 3.2751 8'3380 6'4414 9'0925 6. C. 49.8518 30'5690 51'2638 31'1712 53'7774 36 '5604 8 9 '9196 1010*12 1020.95 1025.07 1041 -94 1062.93 1072.03 1094 -6 6 d.2.6317 5-4516 6'5306 10 -9 470 16 '4800 18.8870 24'9330 1 3'845 3'842 3.839 3.831 3.818 3.813 3'796 The above results are expressed graphically in the dextrose curve of Plate I. (p. 278),* and the true divisor for any pat ticular concen- tration can be obtained by inspection of this curve, or from the following equation, in which D is the required divisor, and G the specific gravity of the solution when water = 1000. D = 3.848 - 0*00028 (G - 1000) - 0.0000028 (G - 1000)2. * For the meaning of this expression see this vol., 1,. 77, footnote,CUPRIC-REDUCING POWER OF DEXTROSE, LEVULOSE, ETC. 2'7'7 The divisors for dextrose deduced from the foregoing experiments agree very closely with those recently given by C.O'Sullivan and Stern (Trans., 1896, 69, 1695). Y'a6Ze II.-SoZution-density of Anhydrous Levdose. - 1 2 3 4 5 6 7 8 - n. 1'2326 3.1436 2.5962 2'0306 6'5677 4'8289 5'0967 7'0191 1). 50 %868 51.3029 40'7573 27 *5 29 4 54.0685 38 '6 4 76 32 *5304 34 '0 505 1009'72 1024.67 1025'65 1029'67 1049 -85 1051 -35 1065'06 1 OSi '10 d . 2.4676 6.2784 6.5330 7'5576 12.7530 13'1360 16.6870 22'4000 e. 3'939 3'929 3'926 3'925 3'909 3'909 3-898 3 '886 The foregoing results are embodied in the levdose curve on Plate I., the equation for which is D =' 3.946 - 0.00068 (G - 1000) - 0~0000007 (G - 1000)2. We have already fully discussed (Zoc. cit.) the relation of the divisor we have hitherto employed (3.86) to the true divisors for the carbo- hydrates, and it is, therefore, unnecessary t o again refer to the matter, We may, however, point out that here, as in the cases we have pre- viously given, the divisor decreases as the concentration of the solution increases.So far as levulose is concerned, this is in direct opposition to the results obtained by Hijnig and Jesser (Moncctshefte, 1888,9,562), from which it appears that the divisor increases with the concentration. Since it has been conclusively shown that invert-sugar is a mixture of equal amounts of dextrose and levulose, we calculated from the fore- going results the divisor to be applied to solutions of invert-sugar. This was done by taking the mean of the divisors for dextrose and levulose for solutions of one-half the gravity of that of the invert-sugar solu- tions; thus, the divisors for dextrose and levulose for a solution of sp.gr. 1010 are 3.845 and 3.939 respectively; the mean of these is 3.892, which was taken as the divisor for a solution of invert-sugar of sp. gr. 1020. W e are aware that this is not quite correct, as solutions of the same specific gravity do not contain exactly equal amounts of dextrose and levulose, and, moreover, no allowance is made for the lower solution- density of the solution of double strength; but the error thus introduced is so small, being in the fourth place of decimals, that it fa119 well within the limits of experimental error. In this way, we obtained the following table.278 RROWN, MORRIS, AND NILLAR : THE SOLUTION-DENSITY AND Gravity. I/ Gravity. 1 Divisor. Divisor. 1 I I I 1010 1020 1030 1040 1050 1060 3.894 3.892 3.889 3.886 3-883 3-8813 1070 1080 1090 1100 1110 1120 3.817 3.874 3.871 3.868 3'865 3.862 The foregoing values mere checked at several points by determinations made with pure invert-sugar, and were found to be correct.They are expressed graphically in the invert-sugar curve of Plate I, the equation for which is D = 3.897 - 0*00025 (G - 1000) - 0*0000004 (G - 1000)'. The Czqwic-yeducing Powem. I n the paper already mentioned (Zoc. cit., p. 95, et sep.), we discussed, a t considerable length, the conditions requisite for the accurate and concordant determination of the cupric-reduction of the carbohydrates, but it mill be advisable to briefly recapitulate them here. They are (1) the use of a Fehling's solution of constant composition; (2) the maintenance of the same degree of dilution in all experiments ; (3) the precipitation of an amount of copper oxide which shall fall between certain limits ; and (4) an invariable method of determination, both as regards mode and time of heating.The composition of the Fehling's solution me use is Recrystallised copper sulphate . . . Rochelle salt ... . .. ... .. . ... .. . .. . ., . 173.0 ,, 9 ) Anhydrous sodium hydroxide . . . 65.0 ), ? ? 34.6 grams per litre. Fifty C.C. of this solution, which should be freshly mixed, are placed in a beaker of about 250 C.C. capacity and having a diameter of 7.5 centi- metres. This is placed in a boiling water bath, and when the solution has attained the temperature of the water, the accurately weighed or measured volume of the sugar solution is added, and the whole made up to 100 C.C.with boiling distilled water. The beaker, which is covered with a clock glass, is then returned to the water bath, and the heating continued for exactly 12 minutes. The precipitated cuprous oxide is now rapidly filtered off through a Soxhlet tube, mashed first with hot water, then with alcohol and ether, and finally dried. When dry, the cuprous oxide is reduced to metallic copper by gentle heating in a stream of hydrogen, and weighed. The weight of copper obtained must be cor- rected for that due t o the slight spontaneous reduction which alwaysBROWN, MORRIS, AND MILLAR. Journ. Chern. SOC. Xarch, 1897. PLATE IBROWN MORRIS AND MILLAR. J o u m , Cham, sbc, Naroh, 1897CUPRLC-REDUCING POWER OF DEXTROSE, LEVULOSE, ETC.279 takes place, and which should be determined for each batch of Fehling’s solution. Working in this way, we have determined with great care the cupric- reducing power of dextrose, levulose, and invert-sugar, taking in each series of experiments gradually increasing amounts of the sugar, and determining the copper reduced by each amount. The results are em- bodied in the following tables, which are given in the same form as those in our previous paper, to which reference has been made. The sugars used were prepared as indicated in the first part of this paper, and were most carefully and thoroughly purified. I n the dextrose table, the source from which the sugar mas obtained is indicated in the first column ; the determinat’ions marked cc were made with dextrose prepared from Kahlbaum’s product ; 6, from maltose, by acid hydrolysis; and c, from cane-sugar, by inversion with acid.Tcc 6 Ze I V.-C~~~”’ic-1.ecluctiosa of Dextisose. - Sonrcc. b CL c C b CL C c 6 C6 c 0 CL c 6 Dextrose by 3-86 divisor. 0-0457 0.0474 0’0491 0.0736 0’0884 0’0950 0.0980 0 1180 0-1338 0’1425 0.1555 0.1786 0‘1911 0’1968 0’2105 Dextrose tlbsolute. 0,0459 0.0476 0.0493 0.0739 0.0888 0.0954 0-09P4 0.1185 0.1344 0.1431 0.1592 0.1793 0.1919 0.1976 0.2114 CU weighed. 0’0938 0.0983 0’1024 0.1513 0.1809 0.1963 0’1996 0’2369 0.2648 0.2836 0’3089 03397 0.3635 0.3732 0.3920 CUO )er gram of 3-88 lexc,rose. 2572 2599 2’614 2 5 7 7 2-564 2 5 7 7 2.553 2’516 2‘480 2.494 2.442 2.385 2.384 2.377 2.334 CUO ier gram of absolute lextrose.2.562 2.589 2’604 2.567 2,554 2.567 2543 2507 2.470 2.485 2-433 2.375 2.375 2.365 2.325 116’6 117’9 118’5 116’8 116’3 116‘8 115-8 114.1 112.4 113‘1 110.7 108’1 308’1 107.8 105.8 - fCnbnnll1 tr. 116’2 118‘1 116’3 115.8 116‘4 115.3 113.7 112’0 112.7 110’3 107.7 107.7 107.4 105.4 11 7.4 The above results are expressed graphically in curve I of Plate I1 (p. 280), which shows that the copper oxide reduced by 1 gram of dex- trose appreciably decreases as the amount of copper reduced increases,280 BROWN, MORRIS, AND MILLAR : THE SOLUTION-DENSITY AND Levulose by 3-86 divisor. 0-0403 0.0443 0.0446 0-0473 0'0507 0-0964 0.0964 0.0964 0.0964 0'1455 0-1443 0'1438 0.1890 0'2011 0.2022 0'2043 Fable V.-Cup.ic-reduction of Levulose. Levulose absolute, 0.0396 0'0466 0'0500 0'0949 0'0949 0'0949 0.'09 4 9 0'1432 0'1421 0'1416 0'1861 0'1980 0'1991 0'0435 0'0438 0.2012 - cu weighed.0.0750 0.0822 0.0831 0.0887 0.0939 0.1765 0-1765 0.1770 0.1787 0.2612 0'2601 0'2585 0'3290 0.3507 0.3526 0'3576 - - cue per gram of 3.86 levulose. 2.329 2'323 2.332 2'323 2.318 2.295 2 9 9 5 2.301 2'324 2.251 2.260 2-254 2.182 2.186 6.186 2.194 - cue per grain of absolute levnlose. 2.375 2'369 2'379 2.360 2 -355 2'332 2-332 2'338 2'361 2-287 2.295 2.289 2'217 2.221 2'220 2.228 - 105'6 105.3 105'7 105-4 105-2 104'1 104.1 104.4 105'3 102.1 102.4 10272 99 .o 99 -2 99.2 99.5 Kabsolote. 107.7 107'4 107.9 107'0 10673 105.7 105.7 106.0 107.0 103'7 104.0 103 '8 100.5 100-7 100-7 101.0 7 From the foregoing numbers, curve I11 of Plate I1 has been con- structed. It will be seen that, throughout the series, any given amount of levulose reduces appreciably less copper than the same weight of dextrose.Tu, 6 Ze V1.- Ctcpic-reduction of Invei*t -sugar. Invert- sugar by 3.86 divisor. 0 *0506 0.0860 0.1095 0-1303 0'1582 0'1847 0.2167 Invei.t-sugar absolute. c 11 weighed. 0,0502 0-0854 0.1087 0 1293 0.1570 0.1833 0.2152 0,0979 0.1681 0.2085 0.2511 0.2965 0.3386 0.3890 ' CUO per grain of 3'86 invert- sugar. 2'426 2'449 2'387 2.416 2'350 2.298 2'250 CUO per gram if absolute invert- sugar. 2.445 2-468 2.405 2.435 2-368 2.316 2'267 110.0 110.9 111'0 111'9 108-2 109.0 109'5 110'4 106'5 107'4 104.2 105'0 102.0 102.8 The above results are expressed graphically in the invert-sugar curve I1 of Plate 11, which falls midway between the curves for dextrose and levulose.From the curves obtained when the foregoing experimental results are plotted on a system of rectangular co-ordinates, we have con- structed the folIowing table showing the reducing powers of the threeCUPRIC-REDUCING POWER OF DEXTROSE, LEVULOSE, ETC. 281 I 2 3 I rR - 5c 5E: 6C 6: 7c 75 8C 85 90 95 1 oc 105 115 120 125 130 135 140 145 150 155 160 165 170 176 180 185 190 195 200 205 i i a - sugars in question. We have given, in the first column, the quantities of sugar from 50 to 205 milligrams, at intervals of 5 milligrams, and, opposite to these, the amounts of Cu and CuO precipitated by each quantity, together with the weight of CiiO corresponding to 1 gram of each sugar, when the respective quantities are oxidiaed under the conditions of our method.Cn grams. 0'103C 0.1134 0 '1 23€ 0 '1 341 0'1543 0'1644 0.1740 0'1834 0.193a 0'2123 0.2218 0'2313 0.2404 0.2496 0.2585 0.2675 0.2762 0.2850 0.2934 0.3020 0.3103 0.3187 0.3268 0.3350 0.3431 0-3508 0 '3590 0'3668 0'3745 0.3822 0'1443 0'2027 - Dextrose. CLlO grams. 0'1289 0'1422 0.1552 0'1682 0.1809 0-1935 0.2061 0'2187 0.2299 0.2420 0.2538 0'2662 0.2781 0.2900 0'3014 0,3130 0'3241 0'3354 0'3463 0.3573 0'3673 0.3787 0.3891 0.3996 0'4098 0'4200 0'4302 0'4399 0.4501 0.4599 0'4689 0'4792 Ell a . GE 9 @ u 2 0 o+ 2'578 2'585 2'587 2.589 2.585 2'580 2.577 2'572 2'555 2'547 2'538 2'535 2'528 2'522 2'512 2'504 2'493 2'484 2'473 2'464 2'448 2'443 2'432 2'428 2'410 2-400 2'390 2.377 2.369 2.358 2.344 2.338 c Ll grams. 0'0923 0.1027 0.1122 0'1216 0'1312 0.1405 0.1500 0.1590 0'1686 0,1774 0'1862 0'1952 0.2040 0.2129 0 *2215 0'2303 0'2390 0'2477 0'2559 0'2641 0'2723 0.2805 0,2889 0'2972 0'3053 0.3134 0.3216 0.3297 0.3377 0'3457 0.3539 0'3616 Levnlose.cuo grams. 0.1155 0.1287 0.1407 0 -1 524 0.1645 0'1761 0.1881 0'1993 0.2114 0 '2224 0.2331 02447 0.2558 0.2669 0-2777 0.2887 0.2997 0.3106 0.3209 0.3311 0.3409 0.3517 0.3622 0.3726 0.3828 0.3930 0.4032 0.4134 0.4234 0 ~ ~ 4 3 3 5 0'4431 0 '4534 2.310 2'341 2.345 2'346 2-350 2 *349 2.351 2'346 2 '349 2.341 2.331 2-331 2.325 2'321 2'314 2'310 2.305 2'300 2'292 2'284 2.273 2 '269 2'264 2'258 2-252 2-245 2'240 2-234 2'228 2'223 2.216 2-211 - c 11 grams. 0.0975 0'1076 0.1176 0.1275 0-1373 0 '1 468 0-1566 0.1662 0.1755 0.1848 0'1941 0.2034 0'2128 0,2220 0.2311 0'2400 0.2489 0-2578 0'2663 0.2750 0.2832 0.2915 0.3002 0'3086 0'3167 0-3251 0.3331 0'3410 0'3490 0.3570 0.3650 0,3726 - Invert -sugar.CUO grams. 0'1221 0.1349 0'1474 0'1598 0'1721 0'1840 0'1963 0'2084 0'2200 0'2317 0.2430 0'2550 0 '2668 0.2783 0'2898 0.3009 0.3121 0.3232 0-3339 0.3448 0.3546 0.3665 0.3764 0-3869 0'3971 0.4076 0.4177 0.4276 0.4376 0.4476 3'4570 3'4672 - 2 '442 2'453 2'457 2.469 2.459 2'454 2.454 2.451 2'445 2'439 2 '430 2'429 2'425 2'420 2'415 2'407 2-400 2 '394 2.385 2'378 2'364 2.358 2.352 2.345 2'336 2.329 2.320 2'311 2.303 2.295 2'285 2.279 - Before commenting on our results, it will be necessary to refer briefly to the method at present in use of expressing the cupric-reducing power of the sugars. I n 1876, C. O'Sullivan (Trans., 1876, ii, 125),282 BROWN, MORRIS, AND MILLAH : THE SOLUTION-DENSITY AND defined this ‘‘ to be the amount of cupric oxide, calculated as dextrose, which 100 parts reduce.” The cupric rednction of dextrose mould then be 100, and 1 gram of this sugar was considered to reduce 2.205 grams of copper oxide.O’Sullivan expressed this value by K, and the symbol K had previously been used by one of us and Heron (Trans., 1879, 35, 607) with the same significance. It has long been known that the 2.205 value for dextrose was not quite correct, being, in fact, too low, and consequently that the true cnpric-reduction of dextrose was distinctly above 100 when referred to this standard. As we have, however, already pointed out a t length, i n the paper previously referred to (this vol., p.97), this is immaterial if it is clearly and definitely understood to what basis the cupric-reducing power is referred. From the foregoing experiments with dextrose, it will be seen that we have obtained considerably higher values than those mentioned above. We find that the amount of copper oxide reduced by 1 gram of the sugar ranges from 2.562 to 2,325 grams, according to the extent to which reduction of the Fehling’s solution is carried; and on the- 2.205 basis, these numbers correspond to a K of 116.2 and 105.4 respectively. Recently, and after our experiments were finished, Heron (Journal Federated Irzstitutes Bvewing, 1896, 2, 443) and C. O’Sullivan and Stern (Trans., 1896,69, 1691) pointed out the higher reducing power of dextrose. The former states that 1 gram reduces 2.26 grams of cupric oxide, but he does not mention the precise conditions under which this value was obtained.The latter authors give 2,306 grams of CuO as the equivalent of 1 gram of sugar, and the mean reducing value as K = 104.6 ; they use 30 C.C. of Fehling’s solu- tion diluted to a t least 90 c.c., and take from 0.12 t o 0.13 gram of sugar. Under these conditions, me should expect that their results mould be somewhat lower than ours. The method employed by Kjeldahl in his investigations on the cupric- reduction of the sugars (R6mnzti du Compte-rerzdzc des travaw du bboratoire du Carlsberg, 4’”” vol., 1’”. livr., 1895), very closely resembles ours, with the exception that the Fehling’s solution is heated for 20 minutes, and the reduction carried on in an atmosphere of hydrogen.As would be expectedfrom this variation in the conditions, his values for the cupric-reduction of dextrose are uniformly higher than ours ; but the difference is fairly constant throughout the series. This, we have already shown, was also the case with maltose (loc. cit.). It has always been generally held that the reducing powers of levulose and invert-sugar were less than that of dextrose. This fact is well shown in the foregoing results, and also in Table VII., which gives the amount of copper oxide reduced by 1 gram of each sugar at different stages of reduction. It has been stated by J. O’SullivanCYUPRIC-REDUCING POWER O F DEXTROSE, LEVULOSE, ETC. 283 (Trans., 1892, 61, 408) and by Heron (Zoc. cit.) that the reducing power of invert-sugar is the same as that of dextrose, but this is nega- tived by our results, which are, on this point, in accord with those of Kjeldahl (Zoc.cit,), Ost (Ber., 1850, 23, 3003), Honig and Jesser (Moi~atsliefte, 1888, 9, 562), Soxhlet (J. pr. Chem., 1880, 21, 227), and other workers. The results obtained by these workers, although not exactly the same as ours, owing to the differing conditions of experiment, yet show that levulose has an appreciably lower reducing power than dextrose, and that the cupric-reduction of invert-sugar stands inter- mediate between the two. If we take the results expressed in Table VII, and calculate from them, at certain points, the cupric-reducing power, K, expressed on the old basis of 2.205, we get the following values for the three sugars.Milligraiiis Dextrose Levulose Invert-sugar Sugar. K. K. K. 50 116-9 104.7 110.7 100 115.1 105.7 110.2 150 111.0 103.1 107.2 200 106.3 100.1 103.6 If, however, we take the value of dextrose a t each of the above points as 100, and express the values of levulose and invert-sugar as percentages on this number, we get the following results. Milligrams Sugar. Dextrose. Levulose. Invert-sugar, 50 100 89-60 94-72 100 100 01.s4 05.74 150 100 92.s5 96.56 200 100 04.1 1 97-05 We see, then, from these numbers, that if the reducing power of dextrose be taken as 100, when what may be regarded as the usual amount of copper is reduced (150 to 200 milligrams), the values to be assigned to levdose and invert-sugar closely approximate to those which have been usually taken for these sugars.I n our former paper, t o which reference has repeatedly been made, we showed the great influence which the amount and nature of the alkali in the Fehling’s solution exercised on the quantity of copper reduced by a given weight of maltose, or of the starch-transformation products. With dextrose and levulose, the influence is far less. Kjeldahl has shown (loc. cit.), and we have confirmed the observation, that the amount of sodium hydroxide per litre may be varied within fairly-wide limits, without producing any considerable alteration in the amount of copper reduced by a given weight of dextrose; whilst284 BROWN, MORRIS, AND MILLAR : THE SOLUTION-DENSITY, ETC. - 105.0 104.1 101.9 101.0 - I Glendinning has proved (Trans., 1895, 6’7, 999) that an equivalent amount of potassium hydroxide may be substituted for the sodium compound without causing any alteration in the reducing power. The variant which has the greatest influence in the case of dextrose and levulose is the state of dilution of the Fehling’s solution.If the 50 C.C. of Fehling’s solution is diluted with 100 c.c., 150 c.c., or 200 C.C. of water, instead of with the 50 C.C. of our standard method, the reducing power is appreciably lower at all stages of reduction, and the greater the dilution, the greater the difference. This is well seen in the following table, in which the results are given in terms of K absolute. Thble QII1.--Reducing Poww of Dextvose and Levulose at di$ere?Lt Degrees of Dilution. - - 102.6 102.9 102-2 101 ‘9 100‘9 99 -8 99.7 98 *9 - - - - Sugar grams. I Dextrose. Dilution 1 :2. 0.0725 0.0846 0‘1120 0.12’15 0.1 697 0.1830 0.1901 116.7 114.6 - - - 108.1 - Dilution 1 :3. 115.2 113’9 107’8 - - - - Dilution 1 :4. 113.0 111.8 - - - 107.2 - Levulose, Dilution Dilution Dilution 1 : 2 . I 1 : 3 . 1 1:4. I The degree of dilution has, however, a much greater influence on the reducing power if the experiments are made in a different manner, namely, by keeping the total volume of solution constant, and decreas- ing the amount of Fehling’s solution in such volume, Kjeldahl made a series of experiments in this way, using 15, 30, 50, and 75 C.C. respectively of Fehling’s solution and making up the volume in each case to 100 C.C. When nearly the maximum amount of copper was precipitated from the more dilute solutions, much lower values than those given above were obtained for the cupric-reducing powers, and we have confirmed these observations by direct experiment. In our former paper (this vol., p. 106), w e fully discussed the limits of error of the methods employed for the determination of solution density and cupric-reducing power, and they, therefore, need not be recapitulated here,