[Reprinted from American Chemical Journal, Vol. VI, No. 5.] 
DETERMINATIONS OF LACTOSE/IN MILKS BY 
’ OPTICAL METHODS. 
By Harvey W. Wiley. 
The usual method of determining milk sugar by evaporating the 
sample to dryness, and extracting the sugar with alcohol after 
exhausting with ether, requires a great deal of time and labor. If 
some reliable optical method could be devised, the determination 
of the lactose in milk would be the work of only a few minutes. 
The difficulties which are encountered in seeking for such a method 
are numerous and serious, so much so that little credit has hereto- 
fore been given to any of the processes of optical analysis in use. 
In the following paper I will give the results which have been 
obtained under my direction in the study of this problem. The 
optical work collected in the following tables was done chiefly by 
my assistants, Messrs. G. L. Spencer and C. A. Crampton ; the 
alcohol extractions were made by Mr. A. E. Knorr, and the nitro- 
gen estimation by Messrs. Fuller and Trescott. 
Specific Rotatory Power of Milk Sugar. 
Crystallised milk sugar when first dissolved possesses a higher 
rotatory power than it has in the milk from which it was derived. 
This increased optical activity may be compared with the original 
by the ratio 8 : 5, nearly. After the solution has stood for twelve 
to twenty hours, or immediately on boiling it, this extra rotatory 2 
power is lost. In estimating the specific rotatory power of milk 
sugar the numbers given always refer to the constant and not the 
transient gyratory property. 
Among the earliest numbers assigned to the rotation of lactose 
are those of Poggiale (a)D — 54.2 and Erdmann (^a)D =51.5, 
(Sucrose (a)D — 66.5). Biot1 places this number for lactose at 
60.23, and Berthelot2 at 59.3 for the transition tint (a)j. Hoppe- 
Seyler, in his “ Handbuch der physiologisch-chemischen Analyse,” 
gives this number at (a)} — 58.2. Since the ratio of (a)D to (a)} is 
1 : 1.1306, the above numbers become for Biot (a)D = 53.27, for 
Berthelot (a)D — 52.47, and Hoppe-Seyler (a)D — 51.48. Hesse3 
observed the rotation number to be (a)D = 52.67, when the solu- 
tion contained 12 g. per 100 cc. and the temperature was 150. On 
the other hand, when the concentration is only 2 g. per 100 cc. the 
number assigned is (a)D — 53.63. It appears from this that the 
specific rotation power of a solution of milk sugar diminishes with 
the increase of its concentration, and this view is adopted by 
Landolt, Tollens and Schmidt. 
The following general formula4 is used to correct the reading of 
the polariscope for concentration of solution : 
(a)D = 54.54 — .5575 -f .05475 ** — .001774 c3, 
in which c — number grams sugar in 100 cc. solution. These 
observations are contradicted by the work of Schmoeger,5 who, in 
an elaborate series of experiments, using instruments of different 
construction and observing all necessary precautions, found the 
rotation number of lactose sensibly constant for all degrees of con- 
centration up to the saturation point. In 32 series of investigations, 
in which the degree of concentration gradually increases from c — 
2.3554 to c — 36.0776, and in which a constant temperature of 20° 
was maintained, the variations in the numbers obtained were 
always within the limits of error of observation. The mean of all 
these numbers fixes the value of (a)D at 52.53. 
According to Schmoeger, variations in temperature have far 
more to do with changes in rotatory power than differences of 
concentration. The value of (a)D falls as the temperature rises. 
1 Compt. rend. 42, 349. 
2 Diet, de Chem. par Ad. Wurtz 2, 2d part, p. 188. 
3 Annal. Chem. u. Pharm. 176, 98. 
4 Tucker, Sugar Analysis, 91. 
5 Berichte der deutsch. chem. Gesell. 1880, 1922 et seq. 3 
Under 20° the disturbing influence of temperature is greater than 
above 20°. At the latter degree (a)D varies inversely about .075 
for each i° change of temperature. Pellet and Biard,1 as a result 
of their observations, fix the rotatory power of milk sugar at 58.94 
for (a), ((«)* = 52.12). 
After a careful review of the methods used in the above resume 
and the numbers determined by them, I am inclined to accept the 
mean obtained by Schmoeger as the one entitled to the greatest 
credit. It also has the advantage of being almost the mean of all 
the various numbers which have been assigned as the specific 
rotating power of lactose, viz. 
Poggiale, . . 54.2 
Erdmann, . . 51.5 
Biot, . . . 53.27 
Berthelot, . . 52.47 
Hoppe-Seyler, . 51.48 
Hesse, . . . 52.67 
“ ... 53-63 
Schmoeger, . . 52.53 
Pellet and Biard, . 52.12 
Mean, 52.65 
In the present state of our knowledge, therefore, the specific 
rotatory power of milk sugar should be taken at (a)D —52.5. I 
propose, at an early date, to make a careful study of this subject, in 
order to fix, if possible, an exact number for the expression of the 
rotating power, and to examine the conflicting evidence respecting 
the influence of the degree of concentration on the same. The esti- 
mation of lactose in milk by the polariscope is rendered difficult also 
by the presence in milk of various albumens—all of which turn the 
plane of polarisation to the left. As will be seen by the data given 
further along, the ordinary method of removing these albumens, 
viz. by a solution of basic lead acetate, is far from being perfect. 
If, therefore, a portion of the albumen be left in the liquid sub- 
mitted to polarisation, the rotation to the right will be diminished 
by its presence. 
Hoppe-Seyler1 assigns as the rotation power of egg albumen 
(a)D— — 35.5, and for serum albumen (a)D =— 56. Both acids 
1 Bull, de 1’Assoc, des Chimistes 1, 5, 171 et seq. 2 Wiirtz, Dictionnaire de Chimie 1, 91. 4 
and alkalies seem to increase the rotating power, which may with 
acetic acid reach (a)D —— 71. 
Fredericq1 gives the rotation number for blood serum for the 
rabbit, cow and horse at (a)Dz=z — 57.3, and for the dog at — 44. 
Paraglobulin, according to the same author, has a rotation number 
(«)„= — 47.8. 
Milk albumen3 has the following numbers assigned to it: 
Dissolved in MgSC>4 sol. (a)D —— 80 
“ “ dil.HCl. “ =—87 
“ “ dil. NaOH sol. “ =—76 
“ “ strong KOH sol. “ = — 91 
The hydrates of albumen3 have rotation powers which vary from 
(a)D =— 7140 to(a)D = — 79-05. From the chaotic state of know- 
ledge concerning the specific rotating power of the various albu- 
mens, it is impossible to assign any number which will bear the test 
of criticism. For the purposes of this paper, however, this num- 
ber may be fixed at (a)D — — 70 for the albumens which remain in 
solution in the liquids polarised for milk sugar. 
The phenomenon of “ birotation ” in milk sugar has already 
been noticed. The problem of analysis of this sugar is, however, 
still further complicated by the facts pointed out by Schmoeger4 
and Erdmann6 that when milk is rapidly evaporated in a plain 
dish the sugar is left in the anhydrous state, and that this sugar in 
fresh solutions exhibits the phenomenon of “ half rotation.” When 
such sugar is extracted w’ith alcohol and re-evaporated, it, doubt- 
less, is still anhydrous. But in the calculation of results this sugar 
is generally estimated as containing water of crystallisation, and 
thus an error, which Schmoeger reckons at as much as .2 per cent., 
is introduced into the results. This fact, not well recognised, com- 
bined with the knowledge that in the process of evaporation many 
particles of sugar must be occluded by the hardening caseine, tends 
to throw doubt upon the accuracy of estimating the sugar by the 
extraction method. 
The work which I undertook had for its object the determina- 
tion of the best method of preparing the milk-sugar solution for 
1 Compt. rendus, 93-465. 
* Hoppe-Seyler in Handbook of the Polariscope, Landolt, p. 248. 
* Kiihne and Chittenden, Am. Chem. Journal 6, 45. 
* Ber. d. deutsch. chem. Gesell. 1880, 1915 et seq.; 1881, 212 et seq. 
8 Ber. d. deutsch. chem. Gesell, 1880,2180 et seq. 5 
the polariscope, and a comparison of the numbers obtained by 
this instrument with those given by the ordinary process of 
extraction. 
The reagents used for removing the albumens were : 
(1) Saturated solution basic lead acetate, sp. gr. 1.97. 
(2) Nitric acid solution of mercuric nitrate diluted with an equal 
volume of water. 
(3) Acetic acid, sp. gr. 1.040, containing 29 per cent. HC2H3O2. 
(4) Nitric acid, sp. gr. 1.197, containing 30 per cent. HNO3. 
(5) Sulphuric acid, sp. gr. 1.255, containing 31 per cent. H2SO4. 
(6) Saturated solution sodium chloride. 
(7) Saturated solution magnesium sulphate. 
(8) Solution of mercuric iodide in acetic acid; formula1 KI, 
33.2 g. HgCL, 13.5 g. 
Strong HC2H3O2, 20.0 cc. 
Water, 640 cc. 
Alcohol, ether and many solutions of mineral salts, hydrochloric 
and other acids were also tried as precipitants for albumen, but 
none of them presented any advantages which would make a 
detailed account of the experiments of any interest. 
Table No. I. 
This table contains a record of the experiments which led to the 
adoption of 1 cc. acetate of lead solution, or 1 cc. acid mercuric 
nitrate, as the best amount of each for 50 cc. of milk. 
Nearly all the polarisations were made in a 400 mm. tube. 
From two to four observations were made with each sample. An 
average of these readings was taken for each determination. In 
the calculations the value of (a)D was taken at 53 instead of 52.5, 
the number which subsequent investigations have led me to believe 
more exact. The instrument employed was a “ Laurent Large 
Model” polariscope. 
In all cases the volume of the solution was corrected for the 
volume of the precipitated caseine. This volume was assumed 
to occupy 2 cc. for each 50 cc. milk. 
Since in the Laurent instrument the weight of sucrose in 100 cc. 
to read evendegrees on the scale is i6.i9g. [(a)D =66.67],it follows 
that the weight of lactose in 100 cc. to read one degree on the 
1 Journal de Pharmacie et de Chimie, Aug. 1884, 108. 6 
Table I. 
Reagents Employed 
in Precipitating Albumens. 
• rt'o 
Pb 1 cc. 
Pb 2 cc. 
Pb 3 cc. 
Pb 4 cc. Pb 5 cc. 
A 5 cc. 
Other Reagents. 
o DjJ 
o> a 
Per cent. Lactose. 
«i 0 >> 
1 
I 
4-57 
( 10 cc. 
3-67 
4-23 
2 
4-52 
4.48 
\ 2.45 
3-57 
4.14 
3 
4.46 
3-57 
4 
3-92 
4.19 
3-35 
5 
4-35 
3-55 
4-32 
b 
3.7. 
4.01 
3.00 
4-63 
( 4-44 
7 
4.101 
4.96 H 
X 4.68 H 
8 
4.16 { 
4.29 
4-33 
l 3-8o 
( 3.67 H 
9 
00 
4-59 
4.58 H 
f 4.04 
l 4.12 H 
H28Q4 
10 
4.10 
4.12 H 
3-47 H 
\ 
ii 
4.80 
4.87 H 
(4 cc.) 
4.44 H 
4 cc. 
5 cc. 
6 cc. 
8 cc. 
12 
4-77 
5.02 H 
4.82 H 
4.50 H 
4-31 H 
4.70 H 
4.76 H 
3 cc. 
4.76 H 
4.78 H 
13 
4-25 
4-25 
3-75 
3-38 
3-38 
3-97 
3-89 
HNOa 
14 
4.22 
4.90 H 
4.58 H 
IS 
„ T, S 4-4° 
4.68 H 
4.66 H 
4.50 H 
3,14 U-32 H 
16 
3-3° 
4-43 H 
3-98 
i7 
4.72 
4-43 
4.18 
3-87 
3-65 
3.26 
3-96 
f 3-98 
{ 3.88 H 
3.88 
18 
00 
00 
4.87 
4.87 H 
4-37 
4.25 
(5 cc.) 
4.27 
4.43 H2804 
i9 
4-31 
4.71 
f 4-43 
4-51 
4-59 
4.86 H 
{ 4-43 H 
4.79 H 
4.59 H 
20 
4.39 
4.11 
21 
4.70 
4.17 
f 4-69 
X 4.67 H 
22 
4.961 
4-93 
4.97 H 
( 4-45 
[ 4-43 H 
23 
4.601 
4.41 
f 3-86 
l 3-90 
f 3-94 
4-45 
13-86 
l 3-9° 
( 4-io 
24 
4-74 | 
4.41 
S 4-21 
f 4-32 
f 4.32 
4.45 H 
i 4-35 H 
( 4-44 H 
(4.55H 
25 
4-59 { 
4-33 
( 3-94 
S 4-03 
f 3-98 
4-37 H 
f NaCl 
1 cc. MR 
( 3-93 H 
( 4.10 H 
( 4.10 H 
26 
4-39 
\ 4-29 H 
4.29 H 
27 
4.60 
4.18 H 
4.66 
28 
4.26 
3.67 H 
4.09 7 
scale for each per cent, lactose present would be 16.19, 53; 
66.67, x— 20.37. 
If 52.5 be taken as the value of (a)D for lactose, then x = 20.56. 
In table No. I, A indicates acetic acid, Pb basic acetate of lead, 
MR acid mercuric nitrate, etc. The letters C and H indicate the 
temperature ; C denoting the ordinary temperature of the room, 
and H that the sample was heated to ioo° and cooled before 
filtering. 
The numbers obtained by extraction with alcohol are taken as 
the basis of comparison, not because I believe them to be more 
reliable, but because that method is the one generally employed in 
the estimation of milk sugar. 
In the alcohol extraction the milk was evaporated to dryness in 
a thin glass capsule, the dish and dried residue pulverised in a 
mortar, washed with ether into a continuous extraction apparatus, 
exhausted with ether, and then with 80 per cent, alcohol for ten 
hours. 
Duplicate analyses are indicated in the table by the small 
brackets. 
Remarks on Table I. 
The results obtained by using various other reagents for the 
precipitation of the caseine, viz. MgSCh, CuSO-i, HC1, etc., have not 
been entered in the table. In none of these cases was there 
sufficient encouragement to warrant an extended trial. In most 
cases the precipitation was slow or imperfect, and the filtration 
difficult. 
One important fact should not be overlooked, viz. that any 
excess of basic plumbic acetate causes a rapid decrease in the 
rotatory power of the solution; whether this decrease is due to 
precipitation of the sugar or solution of the albumens does not 
clearly appear. Illustrations of this decrease are seen in analyses 
2, 12, 13 and 17. 
It seems to make little difference whether the precipitation is 
made hot or cold. The question of temperature is set forth in 
greater detail in the next table. From all the experiments made it 
clearly appeared that the best optical results are obtained by 
the use of a minimum quantity of basic lead acetate, or of either the 
acid mercuric nitrate or iodide. For 50 to 60 cc. of milk, 1 cc. of 
the lead acetate or mercuric nitrate solution of the strength noted, 8 
and 25 cc. of the mercuric iodide solution are the proper quan- 
tities. It makes no difference, however, if a large excess of the 
two latter reagents is employed. Of the three the last is to be 
preferred. 
Table No. II. 
In this table will be found the results of the comparative deter- 
minations of milk sugar by extraction with alcohol, by precipitation 
with i cc. basic lead acetate, and the same with i cc. acid mercuric 
nitrate, hot and cold, to each 60 cc. of milk. 
In many of the analyses the large differences in results by the 
three methods show a fault of manipulation, but all the results 
have been given without selection. 
Table II. 
Reagents Employed. 
No. 
Extract 
c. 
H. 
C. 
H. 
by Alcohol. Pb 1 cc. 
Pb 1 cc. MR 1 cc. 
MR 1 cc. 
Percentage of Milk Sugar. 
I 
4-55 
4-74 
4.92 
2 
4.10 
4.22 
4-5° 
3 
4- Si 
4-54 
4.22 
4.68 
4.62 
4 
4-36 
4-55 
4-53 
4.89 
4.90 
5 
4.05 
4.14 
4.09 
4.48 
4-39 
6 
3-84 
3-84 
3-98 
3-98 
7 
4.52 
4.67 
4-73 
5.01 
5.00 
8 
4.25 
4.21 
4.26 
4.51 
9 
4-45 
4.61 
4-54 
4.87 
4.87 
IO 
4.92 
5.20 
5.22 
5-43 
5*47 
ii 
3-84 
3-72 
4.00 
3-96 
12 
4-53 
4.61 
4.64 
4.87 
4.85 
4-57 
4-54 
4-55 
4-91 
4.84 
14 
4.66 
4.29 
4-45 
4-63 
IS 
4.17 
3-65 
3-75 
3-95 
3-87 
16 
5.02 
4.66 
4.64 
4.86 
4.86 
17 
4.68 
4-03 
3-94 
4-39 
4-37 
18 
4.23 
3.82 
3-89 
4.08 
4.02 
i9 
4.96 
4.70 
4.84 
5.04 
5.04 
20 
4.85 
4-39 
4.41 
4-53 
4-65 
21 
4-63 
4-47 
4-47 
4.69 
4.67 
22 
4-47 
4-39 
4-45 
4.67 
4.71 
23 
4.46 
4-23 
4-3i 
4-65 
4-63 
24 
4-47 
4-59 
4.67 
5.01 
4-95 9 
Extract 
No. by Alcohol. 
c. 
Pb i cc. 
H. 
Pb 1 cc. 
C. 
MR 1 cc. 
H. 
MR 1 cc. 
25 
4.40 
4.41 
4-55 
4-45 
26 
4.85 
4.67 
4-73 
4-97 
27 
4-45 
4.21 
4-33 
4-57 
28 
4.44 
3.98 
4.10 
4.28 
29 
4.10 
4.21 
4-55 
3° 
4.38 
5-57 
4.69 
4.89 
31 
4.20 
4.21 
4-37 
4-57 
32 
4.69 
4-59 
4.67 
4.89 
33 
4.52 
4.27 
4.41 
4.41 
34 
4-37 
4-65 
4-93 
4-93 
35 
4.52 
4.27 
4.41 
4.56 
36 
4.88 
4-83 
4-93 
5-i7 
37 
4.61 
4-30 
4-43 
4-57 
38 
4-79 
4-59 
4.67 
4-91 
39 
4.67 
4.26 
4.41 
4-51 
40 
4-79 
4.64 
4-74 
4.94 
4i 
3-95 
4.10 
4.26 
4-38 
42 
4.00 
4.61 
4.61 
4-77 
43 
4-63 
4.24 
4-37 
4-57 
44 
4-77 
4.64 
4.70 
4.94 
45 
4.85 
4-53 
4-73 
46 
4-71 
. 
4.67 
4-93 
47 
4-34 
4.06 
4.12 
4.40 
48 
4.05 
4.67 
4-77 
4-83 
49 
3-67 
4.12 
4.18 
4-36 
5° 
3-78 
4.58 
4.62 
4.82 
5i 
4.19 
4.27 
4-57 
4-53 
S2 
3-83 
4.68 
4.78 
4-97 
53 
3.86 
3-97 
4.07 
4.21 
54 
4-59 
4-59 
4.61 
4-83 
55 
4.02 
4.26 
4-36 
4.40 
56 
4-36 
4.62 
4.76 
4-94 
57 
4.20 
4.18 
4.28 
4.48 
58 
4.09 
4-52 
4.56 
4-74 
59 
4.09 
4.28 
4.46 
60 
4.12 
4.49 
4.81 
61 
4.20 
4-33 
4.41 
62 
4-45 
4-25 
4-77 
63 
4-33 
4.09 
4-37 
64 
4.62 
4-33 
4-99 
Mean, 
4-33 
4-34 
4-38 
4-58 
4.63 
In the following table will be found the percentage of milk sugar 
obtained by using varying quantities of the mercuric iodide reagent, 10 
and a comparison of the results obtained with those given by the 
use of acid mercuric nitrate and basic plumbic acetate. 
Reagents Employed. 
Mercuric Iodide. 
No. 
PbA 
MR 
20 CC. 
26 CC. 
30 CC. 
35 cc. 
Percentage of Milk Sugar. 
I 
4.28 
4.48 
4.56 
4-56 
2 
4.46 
4-57 
4.62 
4.66 
3 
4-37 
4-6S 
4.63 
4-63 
4.60 
4.65 
4 
4-37 
4-53 
4.60 
4-53 
4-63 
4.60 
c 
4-38 
4-63 
4.50 
4-53 
4-53 
4-59 
6 
4-33 
4.67 
4-43 
4-53 
4.60 
4.66 
7 
4-3° 
4.67 
4.67 
4.67 
4-59 
4-57 
8 
4-33 
4-59 
4.50 
4-53 
4-50 
4-59 
9 
4.27 
4.60 
4-63 
4.60 
4.66 
4.66 
4-34 
4.60 
4-57 
4-58 
4.61 
4.62 
Table III. 
Albumen remaining in Filtrate from 'Lead, Acetate and Mer- 
curic Iodide Solutions. 
From the fact that the polariscopic readings show that solutions 
of milk prepared with lead acetate have a lower rotating power 
than those prepared with mercury salts, it is to be inferred that the 
lead reagent either leaves certain soluble and transparent kinds of 
albumen in solution, or else dissolves a portion of those which are 
at first precipitated. To test the accuracy of this supposition a few 
analyses were made to determine the amount of albumen left in the 
filtrate from the lead and mercury reagents. At the same time 
different quantities of the mercuric iodide solution were used, in 
order to determine the amount which would give the best results. 
For 60 cc. milk the quantity of mercuric iodide to be used should 
be 25 to 30 cc. 
In the following table will be found the percentages of albumen in 
the whey after precipitating with the reagents noted and filtering. 
Ten cc. of the filtrate were evaporated to dryness in a thin glass 
dish, and the dried residue (with the glass) burned with soda lime. 
The calculated nitrogen was then multiplied by 6.25 and the 
product taken as the percentage of albumen. 11 
Table IV. 
Per cent. 
Per cent. 
Per cent- 
Per cent. 
Per cent. 
Per cent. 
albumen 
albumen 
albumen 
albumen 
albumen 
albumen 
in filtrate 
in filtrate 
in filtrate 
in filtrate 
in filtrate 
in filtrate 
from PbA. 
from Hgl,. 
from Hgl,. 
from Hgl,. 
from Hgl2. 
from Hgl,. 
15 cc. 
20 CC. 
25 CC. 
30 CC. 
35 cc. 
.0865 
.195 
.0865 
.0865 
.0562 
.0865 
•”3 
.0674 
.0865 
.0865 
.0865 
.0865 
•”3 
.0674 
.0562 
•”3 
.0562 
.0312 
.0865 
.0674 
.0562 
•”3 
.0562 
.0562 
•II3 
.0674 
.0562 
.0312 
.0865 
O.562 
•II3 
.009 
.0865 
•03 
.0862 
.1412 
•”3 
.1412 
•”3 
.1412 
.0562 
.195 
.1412 
•TI3 
.1412 
.1412 
•”3 
.1412 
.086$ 
.1412 
.0865 
.1412 
•IT3 
•”3 
.1362 
.125 
.009 
.009 
.1412 
.009 
•”3 
.0865 
Mean, .1182 
.0789 
.0888 
.0839 
.0964 
.0828 
In table No. V will be found percentages of albumen remaining 
in filtrate from lead acetate precipitation of forty-two samples taken 
from those represented in table No. II. From these two tables it 
is at once seen that the quantity of laevo-rotatory matter remain- 
ing in milk after treatment with basic lead acetate is much greater 
than in those samples treated with the two mercuric salts. This 
explains at once the higher per cent, of milk sugar obtained by 
using the last-named reagents, and shows that the use of lead 
acetate as a clarifying agent must be abandoned. 
Per cent. Albumen after Precipitation 
by Lead Acetate. 
No. 
Pr ct. 
No. 
Pr ct. 
No. 
Pr ct. 
I 
.250 
15 
.271 
29 
•35° 
2 
.306 
16 
•237 
3° 
•374 
3 
•135 
17 
.237 
31 
•329 
4 
.272 
18 
.169 
32 
•3°5 
5 
i-34 
*9 
.103 
33 
•3°5 
6 
•239 
20 
.271 
34 
• 237 
7 
.301 
21 
.237 
35 
•305 
8 
•3°5 
22 
.271 
36 
•339 
9 
•237 
23 
•235 
37 
•237 
IO 
•339 
24 
.271 
38 
•374 
11 
.271 
25 
•237 
39 
•203 
12 
•3°5 
26 
•237 
40 
•373 
*3 
.267 
27 
.271 
41 
•305 
.237 
28 
•339 
42 
•339 
Mean, 
.278 12 
Comparison of Results obtained by Extraction with Alcohol and 
Polarisation. 
By consulting table II it will be seen that the percentage of 
sugar obtained by extraction with alcohol is practically the same as 
that got by polarisation of the lead acetate filtrate. 
Thus, the mean percentage of sugar by alcohol (65 analyses) is 
4.32; by lead acetate, cold (53 analyses) is 4.34; by lead acetate, 
hot (64 analyses) is 4.38 ; by mercuric nitrate, cold (61 analyses) 
is 4.58; by mercuric nitrate, hot (24 analyses) is 4.63. 
If now the milk sugar, as has already been intimated, exists in 
an anhydrous state after extraction with alcohol, the percentage of 
it after the addition of the molecule of water would be increased. 
Thus molecular weight of anhydrous milk sugar 342 : mol. wt. of 
the hydrous from 360 = 4.38: .r, whence the value of = 4.61. 
This agrees very nearly with the number obtained by acid mer- 
curic nitrate. 
By a study of table V it is found that mercuric iodide gives 
nearly the same rotatory power as mercuric nitrate, and also that 
by combustion the filtrates from the milks clarified by lead acetate 
contain more albumen than those prepared with mercuric iodide. 
There is, therefore, every reason for believing that the numbers 
given by the mercury salts are nearer the truth than those from 
the lead. 
It may be urged that the increased rotatory power observed by 
the mercury salts is due to the conversion, by the dilute acids, of a 
part of the lactose into galactose, which has a rotatory power 
greater than that of milk sugar. But when it is remembered that 
the quantity of acid introduced is extremely minute, that the 
samples need not be warmed, that they can be filtered and polar- 
ised within a few minutes of the time of the introduction of the 
reagents, the suggestion is seen to be of no force. 
For example, in the acid mercuric nitrate it was found that the 
percentage of sugar was the same whether one, five or ten cc. of 
the reagent were employed, and whether it was polarised immedi- 
ately or after heating and cooling. It is evident that one cc. of the 
reagent, containing less than a half cc. of nitric acid and diluted in 
100 cc. of liquid, could not exert any notable effect on the rotatory 
power of the solution. 
In the mercuric iodide solution 20 cc. of acetic acid are used for 
every 660 cc. of the reagent. 13 
Thirty cc. of this reagent contain, therefore, about one cc. of 
acid. This in 100 cc. of liquid, immediately filtered and polarised, 
could not affect in any marked degree the rotatory power. 
Since combustion with soda lime shows that the filtrate from the 
mercuric iodide sample is practically free from albumen, it is evi- 
dent that the numbers obtained in this way must be a near approx- 
imation to the truth. 
The Process of Analysis. 
The reagents, apparatus and manipulation necessary to give the 
most reliable results in milk sugar estimation are as follows : 
Reagents.—1. Basic phimbic acetate, sp. gr. 1.97. Boil a 
saturated solution of sugar of lead with an excess of litharge, 
and make it of the strength indicated above. One cubic centi- 
metre of this will precipitate the albumens in 50 to 60 cc. of milk. 
2. Acid 7nercuric nitrate dissolves mercury in double its 
weight of nitric acid, sp. gr. 1.42. Add to the solution an equal 
volume of water. One cubic centimetre of this reagent is suffi- 
cient for the quantity of milk mentioned above. Larger quantities 
can be used without affecting the results of polarisation. 
3. Mercuric iodide with acetic acid (composition already given). 
Apparatus. 
1. Pipettes marked at 59.5, 60 and 60.5 cc. 
2. Sugar flasks marked at 102.4 cc. 
3. Filters, observation tubes and polariscope. 
4. Sp. gr. spindle and cylinder. 
5. Thermometers. 
Manipulation. 
1. The room and milk should be kept at a constant temperature. 
It is not important that the temperature should be any given 
degree. The work can be carried on equally well at 150, 20° or 
250. The slight variations in rotatory power within the above 
limits will not affect the result for analytical purposes. The tem- 
perature selected should be the one which is most easily kept con- 
stant. 
2. The sp. gr. of milk is determined. For general work this is 
done by a delicate sp. gr. spindle. Where greater accuracy is 
required use sp. gr. flask. 14 
3. If the sp. gr. be 1.026 or nearly so, measure out 60.5 cc. into 
the sugar flask. Add 1 cc. of mercuric nitrate solution or 30 cc. 
mercuric iodide solution and fill to 102.4 cc- mark. The pre- 
cipitated albumen occupies a volume of about 2.4 cc. Hence the 
solution is really 100 cc. If the sp. gr. is 1.030 use 60 cc. of milk. 
If sp. gr. is 1.034 use 59*5 cc* milk. 
4. Fill up to mark in 102.4 cc. flask, shake well, filter and 
polarise. 
Notes. 
In the above method of analysis the specific rotatory power of 
milk sugar is taken at 52.5, and the weight of it in 100 cc. solution 
to read 100 degrees in the cane sugar scale at 20.56 grams. This 
is for instruments requiring 16.19 grams sucrose to produce a 
rotation of 100 sugar degrees. It will be easy to calculate the 
number for milk sugar whatever instrument is employed. 
Since the quantity of milk taken is three times 20.56 grams, the 
polariscopic readings divided by 3 give at once the percentage of 
milk sugar when a 200 mm. tube is used. 
If a 400 mm. tube is employed, divide reading by 6 ; if a 500 
mm. tube is used, divide by 7.5. 
Since it requires but little more time, it is advisable to make the 
analysis in duplicate, and take four readings for each tube. By 
following this method gross errors of observation are detected and 
avoided. 
By using a flask graduated at 102.4 f°r 60 cc. no correction for 
volume of precipitated caseine need be made. In no case is it 
necessary to heat the sample before polarising.