[Top] [Prev] [Next] [Bottom]

Chapter 9. Filtration of Sake and Yield of Alcohol

At the end of the fermentation the mash is very thin and consists mainly of alcohol and water with a small quantity of the unaltered rice grains suspended in the liquid. The subsequent processes are essentially the same everywhere and it will not be necessary to refer in detail to the methods followed in different breweries. The separation of the liquid from the suspended matter is effected by the use of wooden press called fune, a sketch of which is given in Figure 1 on page 3.

It consists of a wooden box covered on the top by a wooden plate of rather smaller size which is pressed down upon the mass beneath by means of a long lever weighted at the free end with about 12 to 18 hundred pounds, and hinged at the other end to a post firmly dug into the ground. At the bottom of the front part of the press there is an aperture through which the filtered liquid escapes, flowing thence down a gently inclined surface into a receptacle placed below.

The mash (moromi) is put into long, hempen bags which have been strengthened by being soaked in kaki-no-shibu, the juice of the unripe persimmon.1

Each bag is filled about two-thirds full and then contains about 3-1/2 sho; the open end is folded over and tied, and from 300 to 500 bags are piled up in the press according to its size. At Itami there are four presses in use, two of which hold 400 bags, and the other two 342. At Nishinomiya, the press holds 500 bags. At first the weight put upon the lever is very small, otherwise the liquid would run through turbid, but afterwards the weights are increased to 12 or 18 hundred pounds. The pressure is kept up for 12 hours after which the weights are removed, the bags turned over, the pressure renewed for 12 hours longer. The filtrate is slightly turbid and, before use, requires clearing.

At the Tokyo brewery one half of the whole quantity of liquid was filtered on the twenty-seventh, and the remainder before the thirty-second day. A sample of the filtered liquid taken on that day had the following composition:



Glycerin, resin, and albumenoids


Fixed acid


Volatile acid


Water (by difference)



Specific gravity


Compared with the mash on the twenty-eighth day it will be observed that the percentage of alcohol is considerably less, a difference caused by the addition to the mash before filtering of the water used by the brewer for the purpose of rinsing out the tuns.

The composition of the pressed residue (kasu) was found to be:

Soluble solid matter


Starch and cellulose









The alcohol which is unavoidably left in the residue is extracted at a later period by a process of distillation which will be described on a subsequent page.

The amount of sake obtained by the brewer for the quantities given above for one moto was 6.86 koku of specific gravity 0.99, therefore weighing 326 kuwamme, and the weight of the residue was 58 kw. We are now in possession of all the data required to calculate the efficiency of the brewing process as regards the conversion of the starch used into alcohol.

As the sake contained 11.14% alcohol by weight, the total weight of absolute alcohol contained in 326 kw. was 36.32 kw. The 58 kw. of residue, also, contained 6%, amounting altogether to 3.48 kw.; the total quantity of alcohol, therefore, which the brewer obtained was 39.8 kw.

We have already seen that the materials used for one moto amounted to:

Dry rice

175.1 kw.


329.03 kw

504.13 kw.

The dry rice contains on average 84% starch, which, if it were completely converted into alcohol would furnish 80 kw. As the amount actually obtained was only 39.8 kw. we see that the yield is not quite one half of that which is theoretically obtainable. In accurate numbers it is 49.75%. That the loss of material during the preparation of the sake is considerable will be evident when the number of transferences from one vessel to another is considered. Two other sources of loss also are very important, the loss of matter by the rice first, during the process of cleaning and washing, and secondly, during the filtration.

The following calculation will furnish us with some guide to the quantity of material lost in these operations. Allowance is made for the carbonic acid evolved by assuming that it amounts to 98% of the alcohol formed. This number is the result of experiments made by many former observers upon the ratio of carbonic acid to alcohol formed during ordinary fermentation. Any difference between it and the truth will be too small to affect the conclusions:2

Total weight of sake obtained

326 kw.

Total weight of residue

58 kw.

Weight of carbonic acid lost

39 kw.

423 kw

As 504.13 kw. of dry rice and water were used at starting, the total quantity accounted for is only 84%. The weight of dry rice given above was corrected for the loss of weight during the conversion of a part of it into koji, so that the loss of 16% is over and above that experienced during the formation of koji. And indeed, this is not the whole loss because no account is taken of the additional water used in cleaning the vessels, amounting to about 18 kuwamme, which would raise the loss to 19%.

The yield of alcohol obtained at Itami is rather higher than that found in the Tokyo brewery. The quantity of sake obtained is 13.32 koku, which will contain 75.9 kw. of alcohol. 75 kw of residue are also obtained containing 3.8 kw. of alcohol, which altogether amounts to 79.7 kw. The weight of dry rice used we have seen to be 310.77 kw., containing 260.4 kw. of dry starch and ought to produce 140.3 kw. of alcohol. The actual yield is, therefore, 56.8 of that which theory indicates.

At Nishinomiya, the weight of dry rice used is 310.1 kw. and it ought to produce, as at Itami, 140 kw. of alcohol. The yield of sake for one moto is 14.1 koku, which, together with 80 kw. of residue would contain 77.7 kw. of alcohol, and the actual percentage of alcohol obtained is thus 55.5% of that theoretically possible.

There is a very general agreement between the actual yield of alcohol in the three breweries mentioned; although that found by myself as the result of the brewing operation in Tokyo is less than that calculated from the numbers given to me at Itami and Nishinomiya. We may assume that the percentages obtained at Itami and Nishinomiya are the best results, as they ought to be considering the long experience which the brewers of those districts have had. The operations at Tokyo on the other hand are conducted on a much smaller scale and it is scarcely to be expected that the brewers will possess the same skill as those in the great centers of sake production.

Mr. Korschelt, in the paper on sake already referred to3, has mentioned that the actual yield of alcohol according to information from one brewer is only 50% of that theoretically possible, and he expresses the opinion that in any case it is too little, and that the production must reach nearly 100%, because the conversion of starch into sugar is so complete.

I do not consider that the process followed at the Tokyo brewery is a very satisfactory one, but that practised at Itami may be regarded as the one which is carried out with the greatest degree of skill, and yet even there the yield is not more than 57% of that which might be obtained. The case in which Mr. Korschelt says he obtained 80.5% must be exceptional, and I am inclined to think that he has overrated the percentages of alcohol contained in the sake produced. At Itami the strongest sake does not contain, even before dilution, more than 14% alcohol, and it is not probably that the percentage in a Tokyo brew will be greater. In the process which Mr. Korschelt examined in Tokyo, and of which he gives details, the actual yield of sake is 67% of the theoretical yield. The mash consists of the following:

This mash contained 475.4 kw. of starch and ought to have yielded 256 kw. of alcohol. The mash just before filtering measured, according to Mr. Korschelt, 25 koku and contained 14.5% alcohol. If we assume that the specific gravity of the mash was 0.99 (as I found in a similar brew), the total weight of the mash would be 1187.5 kw. and would contain 172 kw. of alcohol, that is 67% of the theoretical yield. This yield is certainly greater than the average yield in other breweries, and may have been the result of special precautions on the part of the brewer, but even in this case, only two thirds of the alcohol was obtained.

In an earlier part of his paper Mr. Korschelt has calculated the theoretical composition of the moto, and also of the mash at the end of the principal fermentation, comparing it with the amounts of extract and alcohol actually found. He arrives at the conclusion that the whole of the starch used enters into solution, at any rate, in one of the examples he brings forward. In the case of moto he gives the theoretical percentage of extract as 35.46, while in one batch of moto he finds 36.86%. In none of the other examples does the percentage arrive at such a high point, being as a rule from 26 to 28%.

The method of water contained in freshly made koji, as used by the brewer, varies from 25 to 30% and never falls so low as 15%, which Mr. Korschelt assumes it to contain. The correction for this will cause an increase in the amount of water given in his paper (loc. cit. p. 250) from 2.925 kw. to 6.32 kw. Again, acting on the assumption that the sugar present in the mash is maltose, the weight of water taken up by the starch in conversion to sugar is calculated only as 1/18, whereas, dextrose being present, as I have shown, it should be twice as much, that is instead of being 2.6 kw. it will really be 5.2 kw. This correction acts in the opposite direction in two ways, first by adding to the weight of the extract, and by taking away from the weight of the water.

Further, the assumption is made that the matter other than starch dissolved from the rice will amount only to 2% of the rice, but in reality at least 12% is dissolved. I have found that the presence of the diastatic ferment of koji has the property of rendering the insoluble albumenoids of the rice soluble, and Messrs. Brown and Heron4 have shown that in the case of malt a certain proportion of the cellulose is held in solution. This will, therefore, add greatly to the concentration of the mash, and finally, the percentage of extract is increased by the removal of water and of carbonic acid during the fermentation. Mr. Korschelt allows 2% for the former, but he omits all correction for the latter. As we have seen however, the weight of carbonic acid evolved is about 98% of the total weight of alcohol formed, in consequence of which the total weight of the mash is diminished by that amount. Hence if the composition of the mash calculated on the supposition that the starch is completely converted into sugar is compared with the actual quantity of extract calculated from solid matter in solution and from alcohol, it is evident that the former will appear too low, and that therefore, the apparent solution of the starch will appear too favorable. This makes a very important item in the calculations, and its non-correction diminishes greatly the accuracy of the results obtained by Mr. Korschelt. The following method of calculating the results avoids the errors which have been pointed out, and shows that the whole of the starch is not brought into solution as Mr. Korschelt supposes.

The composition of the mash was given on page page 73, and we saw that it contained 475.4 kw. of pure starch. The weight of the whole brew before filtering was 1187.5 kw. This contained 172 kw. of alcohol, which is equivalent to 1.856 x 172 = 319.6 kw. of dry starch. The mash also contained 6.5% of extract, which we may assume to be entirely dextrin (although this assumption is in favor of the perfection of the method) and would thus weight 0.065 x 1187.5 = 77.2 kw. The sum of the two numbers, 319.6 + 77.2 = 396.8 kw., is the total weight of starch which has been brought into solution. We see, therefore, that only 83.5% of the total starch used has been dissolved.

So far, therefore, from being able to agree with Mr. Korschelt that the "process of sake brewing is so complete, that important improvements cannot be made in it, unless we would alter the ultimate product to such an extent that it would no longer be sake"5 we ought to conclude from the evidence given in his own paper that it is still capable of being much improved. And this conclusion is born out by all the evidence as to yield which I have been able to obtain, even from the oldest and best managed breweries.

[Top] [Prev] [Next] [Bottom]

1 For an explanation of the action of this liquid upon cloth and paper, see Ishikawa. Chem. News, Dec. 3, 1880. Transactions of the Asiatic Soc. of Japan, IX, 36.

2 The escaping carbonic acid must be saturated with water, and will also cause the evaporation of more or less of the alcohol formed, but it is not possible to estimate the amount of this loss with any approach to accuracy, and it is therefore included in the total loss of 16%.

3 Mittheilungen der deutsches Gessellschaft. 16tes Heft. p. 256.

4 loc. cit. p. 627.

5 "Das Verfahren beim Sake Brauen ist an sich so vollkommen, dass bedeutends Verbesserungen darin nicht gemacht werden konnen, wenn man nicht das schliessliche Product so dadurch verandern will, dass so eben nicht mehr Sake ist." (loc. cit., p. 257, English translation from Japan Mail, August 1878).