Barbet—Rhum—Recent progress contributed to its Manufacture

Follow along: IG @birectifier

BARBET (E.) — Le rhum. Progrès récents apportés à sa fabrication. 3e éd. Paris, 1920.

[Compare this to Rafael Arroyo’s 1938 Circular 106, Rum Manufacture and marvel. This document was also widely referenced by Kervegant. Barbet, we learn in The Distillation of Whisky: Notes & Observations on its Historical and Practical Aspects, 1927-1931, was still alive and kept working into his 80’s. TDOW contains an in depth look at Barbet’s innovations. Something to note is that pH is not yet a thing. Cane growing regions are faced with mountains of molasses piling up that they do not have the infrastructure to process; infrastructure as simple as paved roads! Barbet to save the day! In this era, distiller’s are still very conscious of the idea that ferments need to sour, very much like Scotch described in TDOW. There is a tension between the idea that alcoholic fermentation should be even but long duration and the idea that alcoholic fermentation should be short & safe, but souring long. There is also industry wide fear of higher abv washes. For heavy product, higher abv than roughly 5% interferes, to a degree, with souring and for column product, the stills were not yet adapted to higher ABV. Barbet presents many interesting ideas with time stamps to understand an era that saw a shift to continuous distillation.

Something that is contemporarily very useful is the protocol for fermentation test to understand the sugar content of your molasses. Purely chemical methods still clarify molasses with lead which I’m sure many would want to avoid.]

Recent progress contributed to its Manufacture
E. Barbet & Fils & Cie


Total sterilization of musts. – Pure yeast starters.
Special rum yeast breeds. – Distillation. – Conservation and transport of molasses.
Production of potash salt works, etc.

First Chapter

Current state of rum manufacturing.

The most important source of alcohol production in hot countries is sugar cane, either through the direct processing of vesou (cane juice), or especially through cane molasses fermentation.

Fermentation of vesou gives more specifically what we call cane eau-de-vie, or sometimes rhum d’habitant. This product is poorly known in Europe.

Cane molasses distillation provides a raw product called Tafia in the Antilles, Caña in the Argentine Republic, Cachaça in Brazil, etc. Finally, the name rhum is generally reserved for the liquid prepared for consumption, that is to say improved by aging or by certain blends.

Needless to say, tafias can be rectified like all raw alcohols, and thus provide refined alcohols of very good quality.

Distillation of Vesou.

Although vesou distillation is relatively infrequently practiced, we will say a few words about this special work.

Cane juice is extracted by single or double pressing in the mill, under the same conditions as for sugar production. We generally just strain it to retain the bagasse twigs, and send it to fermentation.

This is spontaneous, like grape juice, but often much more active; it is easily finished in ten and even eight hours. The natural fermenting agent is found in large quantities on the waxy outer surface of the cane; these are yeasts which are not of the same nature as those which exist in Europe. In particular, their temperature preference is quite high.

“The alcoholic yeasts of Venezuela,” says Mr. Delafond, “are much smaller than the species known in Europe; they have a shape very close to an octagon, especially the protoplasm, and they need temperatures of 30° C. minimum to develop and propagate in good conditions, otherwise fermentations take place painfully and very slowly; falls almost never occur at zero; vicious and secondary fermentations take over very easily. While at 35° C., the temperature most suitable for this alcoholic ferment, fermentations take place very quickly and in good conditions with drops to zero.” [“Les chutes”–a fall, likely implies completion of alcoholic fermentation at zero sugar left so these ferments get stuck and bacteria takes over.]

“Alcoholic fermentation supports very strong measures of acidity expressed as SO4H2; European ferments would find themselves paralyzed in this acidic environment, where that of Venezuela behaves very well. Alcoholic yeast reproduction takes place between 30 and 36 degrees centigrade” (1). [pH is not yet a concept. These ferments have high titrateable acidity, but their pH is much higher than European wine ferments because lactic acid and acetic have less influence on pH than malic or tartaric.]

(1) Bulletin de l’Association des Chimistes de sucrerie et de distillerie, septembre 1898, page 323.

There is a small reservation to be made on the topic of acidity; our European yeasts do not like mineral acidity, but they resist certain organic acids well. Baker’s yeast ferments with great activity in lactic sourdoughs corresponding to 6 and 7 grams of SO4H2 acid per liter, and wine yeasts ferment grape juices which, in certain years, have a tartaric acidity which exceeds 9 and even 11 grams per liter, also in SO4H2.

The hotter a must is, the more it generally needs the assistance of a certain acidity (preferably organic, except volatile fatty acids), to protect itself against vicious fermentations. This is an even more true law for vesou than for the sweet musts of our countries, because unacidified vesou almost infallibly presents serious symptoms of bacterial diseases. Sometimes it is a veil of mycodermas which forms on the surface of the tank, a veil so felt-like and consistent that enormous pieces can be removed without them breaking. Sometimes it is viscous fermentation. Against these diseases, there is no other remedy than strong acidity, provided however that it does not reach a degree harmful to the yeast itself. [It gets a little tricky here where Barbet is talking about vicious ferments that are simply bad, and viscous that are also distinctly bad but famous for turning the ferment into a blob of jelly.]

“In vesou, spinning like oil”, says M. Delafond (1), “I searched under the microscope for the viscous ferment, and found it in large quantities forming a compact mucilaginous mass. Diluted in distilled water it forms rosaries, and we find it two by two when it is detached.” [I have seen this personally and it is kind of wild.]

(1) Bulletin de l’Association des Chimistes, octobre 1898, page 360.

From experiments on this ferment, Mr. Delafond concludes that the viscous ferment does not develop in a strongly acidic medium, that the addition of sulfuric acid stops viscous fermentation in active evolution, and finally that this ferment only propagates when alcoholic fermentation has begun its chemical work of splitting sucrose, because it feeds on levulose.

We can fight against acetic fermentation, on the one hand, by avoiding heating the vesou, and on the other hand by carrying out an anaerobic fermentation as much as possible.

For other bacteria, it is mainly the acidity of the must that acts best.

How can we provide musts with acidity in hot countries that are not too costly? The most practical and rational means in this sense is the use, to the greatest extent possible, of the dunder resulting from previous fermentations.

Between the first and second pressing of the cane, the bagasse is usually treated with water; it will be more rational, according to what we have just said, to re-hydrate the bagasse with dunder; it’s a question of dosage and trial and error.

It must also be remembered that there are other means of curbing the excessive heat of fermentation, namely reasonable use of certain antiseptics, refrigeration of the musts, their prior sterilization, etc.

Refrigeration of vesou is often not practical due to lack of fresh water. As for sterilization, we will talk about it later for cane molasses musts. Everything we say about molasses distillation can also be applied to vesou; the same methods and same devices can be used; therefore we will move immediately to molasses which is much more important for the distillery than cane juice.

Cane Molasses.

To realize the enormous production of alcohol which should result from the processing of cane molasses, it is enough to show at what annual figure we must estimate available molasses.

Around one million tons of cane molasses are produced per year, which should make it possible to make six million hectoliters of tafia at 60°! We do not do a sixth as much: because the average yield is barely half of the theoretical yield that we should have; because, except in a few special colonies, tafia is produced, through ignorance, so badly that it cannot find a buyer in Europe. These countries give up fermenting their molasses and throw it into the sea or even into rivers.

In Martinique and Guadeloupe, on the contrary, a lot of rum is produced, all the molasses from sugar production being used. However, current yields in both fermentation and distillation could be greatly improved.

We can easily see the fermentation losses by the following figures:

Martinique molasses have an average content of 70 p. 100 fermentable sugars.

At 140 kilos per 100 liters of molasses we have 140 X 70 / 100 = 98 kilos of fermentable sugars which, at a yield of 60 p. 100 which we obtain by perfected and rational methods, should give 98 liters of Rum at 60°. Instead, the output of many factories is generally 72-73 liters.

That’s a loss of 25-26 liters of rum per hectoliter of molasses!

But rum sells for 48 to 50 francs delivered at the dock in France, a loss of 25 x 0.48 = 12 francs per hectoliter of molasses! If we therefore consider the production of a distillery processing 50 hectoliters of molasses per day–and there are more significant ones in Martinique–the loss for this factory, counting the net proceeds from the sale of rum at 0 fr .30 per liter, is 375 francs. As long as the rum factory operates for 120 days, that makes a total of 45,000 francs that are neglected.

Quality.–If we now consider the current state of the rum industry from the point of view of the quality and aroma of the products, we note the following:

Among the most popular on the French market are those from Demerara (English Guyana) and Jamaica.

However, in recent years “Martinique” has managed to compete with the success of Demerara by providing as much coverage and perfumes under better conditions. In reality these tafias are far from being pleasant to the taste, but they are so rich in flavor that they allow almost unlimited duplication with industrial alcohol. All these very strong tafias (Demerara, Jamaica, Martinique) are obtained by special work whose basis is the reuse of dunder from the first distillation to dilute the molasses to be put into fermentation. Thanks to this dunder, the fermentations are better protected against putridity, and at the same time, the original odor is enhanced.

[I think this is translated correctly but Barbet is speaking a little poetically of grand arôme rums. Duplication may actually be about stretching the rums by blending. He defines coverage later in the document as stretchability. You almost think he is talking about artificial flavours but then the paragraph ends by promoting dunder.]

The aroma of a tafia results from several different sources.

1° First of all, there is the original odor pre-existing any fermentation. If you dilute molasses with water and boil it, without it having fermented, you obtain an odorous distillate. The smell varies both according to the vintage of the cane and according to its species, just as there are differences in vintage and terroir for eaux-de-vie. In addition, the fineness of the smell depends on the working method that was practiced in the sugar factory. All the colonies which adopt the work of acid juices for sugar extraction produce better molasses than the countries where slightly alkaline work is preferred (like Cuba). Alkalinity, by partially destroying certain reducing sugars, causes unpleasant odors. [This is a very interesting statement because we know rum oil relates to treatment with lime.]

On the other hand, experimenters have found that when the grappe (or fermented must) is very slightly acidic, or even almost alkaline, the original aromas which distill are more pleasant and more intense than when the must is very acidic. [Another very interesting statement.]

2° Secondly, we have the odors generated by fermentation. They depend on the yeast breed used, fermentation temperature, and original material concentration of the must. We can therefore, by a reasoned study of yeast and by the use of a rational, pure fermentation, based on the manufacture of appropriate pure starters, excite fragrant secretions, and improve them.

3° Finally, there are combinations which are made after fermentation and even after distillation. After fermentation, if we wait 24 hours or even more before distilling, several phenomena occur:

On the one hand, bacteria which had been stifled by the vigor of pure fermentation can subsequently begin to evolve, now that the yeast is without strength and have abandoned their place. These bacteria are not greedy like yeast, which only accepts selected sugars as food, maltose, glucose, levulose. The other sugars raffinose, mannose, glutose, caramel, etc., have been disdained although they can serve as food for bacteria.

In spontaneous fermentation, which is almost the only rule in hot countries, bacteria develop from the start of fermentation, because it is poorly protected and is sluggish. They evolve so well that the acidity of the must gradually rises to 6, 8 and 10 grams as sulfuric acidity per liter; on the other hand the yeast, increasingly hampered by this increasing acidity, ends up dying, well before all the fermentable sugar has disappeared. Hence the very poor yield of alcohol, both by cessation of fermentation before complete decomposition of the sugar, and at the same time by the fact that part of the good sugar, instead of making alcohol, was transformed into organic acids of all kinds.

Many people believe that these parasitic acids have their useful role, and that during distillation they combine with alcohol to form fragrant ethers characteristic of good rum.

Granted! but it is still necessary to make it impossible for them to harm the alcohol yield; and this is why, thanks to the method that we recommend, we begin by ensuring the transformation as complete as possible of the sugars into alcohol by fermentation as pure and as active as possible. Our fermentations do not last 48 hours, while so-called spontaneous fermentation sometimes lasts 6 and 7 days. Once the alcohol is produced with its yield close to the theoretical yield, it can no longer disappear, and from now on we can let the acid bacteria go free. They can no longer harm; they can only manufacture, by means of food materials disdained by yeast, the acids useful for etherification during distillation.

We therefore see that, thanks to this method, we have reconciled two things which seemed irreconcilable, the high alcoholic yield and an intense aroma.

The fermented must being rich both in alcohol and in acids favorable to etherification, it will be necessary to use to carry out the distillation, devices combined in a very particular way in order to force the ethers to be produced in abundance. This is a most important point to which we will devote a special chapter later.

After distillation, there are still reciprocal reactions between the components of tafia; it continues to form ethers; some of them combine with aldehydes to form acetals, the aroma of which is generally very pleasant.

Finally, there is oxidation of certain natural aromas and evaporation of some bad-smelling gases.

If we carry out a complete chemical analysis of a tafia, we find that apart from variously flavored products, it contains a certain quantity of characterized impurities, completely identical to those found in phlegm from beets or grains. Once isolated, these impurities have a bad odor and a bad taste just like industrial phlegm. We cannot therefore admit that these products are necessary and favorable to rum; all we can say is that their presence may not prevent a rum from being appreciated if there are also enough good flavors to cover and mask these imperfections. We will say that these rums are good, although impure, but not at all “because impure”. Consequently, if we manage to fractionate these higher alcohols or other bad quality bodies, without sacrificing the good aromas, these will stand out more because they will no longer have anything to mask. [Phlegm here denotes low-wines or feints.]

These are the preliminary explanations that it seemed useful to summarize because they will quickly make it clear the usefulness and importance of various improvements that we are going to describe.

Chapter Two

Improved performance.

If we want to appreciate the progress that can be made from the point of view of alcoholic yield, it is good to first give some brief overviews of the composition of cane molasses.

Cane molasses analysis is quite delicate because of the coexistence of a fairly large number of sugars and various reducing agents, sucrose, dextrose, levulose, mannose, glutose, raffinose, caramel, etc. For these kinds of analyses, we must refer to the work of Prinsen Geerligs and Pellet, in particular to the Special Memoir published by the latter in the Bulletin de l’Association des chimistes (1).

(1) Bulletin of November 1897 and following.

The choice of defecation reagent and its proportion are of special importance for the accuracy of the results.

The table below gives a certain number of analyzes of molasses from various sources:

On the other hand, it must be noted that from the point of view of alcohol yield, the analytical determination of the various sugars in molasses does not provide a basis for assessment that is far from certain. There is no other acceptable method for this purpose than the direct comparative testing of molasses by fermentation in the laboratory. [I have read about the fermentation test for sugar content elsewhere but it is seldom discussed.]

Here, for example, are two molasses, No. 1 and No. 2, both from Egypt:

It would seem, upon inspection of these analyses, that No. 1 must give more yield than No. 2 while the direct test provides rather unexpected results; No. 1 gives 27 cc. 59 of absolute alcohol per 100 grams of molasses, while No. 2 gives 31.03, or 3 1/2 p. 100 difference.

The direct fermentation test is therefore essential and we believe it is appropriate to recall the operating method that we have previously recommended.

We start by determining the acidity of the molasses, in order to be able to calculate how much sulfuric acid we will have to add to have a must of 2 grams of acidity per liter before fermentation.

Then, we take a 2 liter, flat bottomed, well sterilized boiling flask. The tare is determined; We weigh exactly 200 grams of molasses, add approximately 800 grams of water in order to carry out the trial fermentation on the basis of approximately 200 grams of molasses per liter. Sulfuric acid calculated to reach 2 grams in total is added, then the liquid is brought to a boil for ten or fifteen minutes. We thus carry out both desulfitation, sterilization, and a small complementary inversion of the sucrose.

The flask opening is closed with a non-hydrophilic cotton ball, and it is allowed to cool as quickly as possible to 30°. During this time, dilute 10 grams of good fresh baker’s yeast in a little sterilized water, and inoculate it in the flask. With a little rinsing water we arrive at approximately 1,050 grams of liquid, which is quite exactly one liter, since the initial density is approximately 1,050. We weigh the flask exactly, including the buffer, and place it in the incubator at 28, 29°. We monitor the rate of fermentation by weighing every six hours, for example. When fermentation is complete, the weight remains constant. It is then necessary to determine the real volume of the liquid, in order to know what exact weight of molasses will correspond to the volumes that will be taken for distillation, polarization, titration of acidity, etc. Suppose that, once the carbonic acid has been blown out, we find 1.020 grams of liquid. Let us cool the liquid to 15°and take the density with a hydrometer exactly at 15° or 1.018.

The real weight, in a vacuum, of one liter of this fermented liquid is 1.018 X 0.99916, the latter coefficient being the weight of water at 15°.

In a vacuum, our 1,020 grams of liquid weigh 1,021 grams. 1 (air displaced by the liquid, minus the air displaced by the brass weights).

The true volume of wine and V = P/d = 1021 gr. 1 / 1018 X 0.99916 = 1021.1 / 1071.1 = 1004 cc.

 The 500 cc. that we take for distillation therefore correspond to 200 grams x 500 / 1004 = 99 gr. 601 of molasses.

For the accuracy of the alcohol content, it is necessary to saturate exactly 500 cc. of wine, because otherwise the distilled volatile fatty acids cause a slight error in the alcoholometer. Push the distillation up to 200 cc. of distillation.

As an example, the two Egyptian molasses above gave the following results after fermentation in the laboratory.

What we obtain in terms of output in the laboratory, it is by no means impossible to produce it industrially, provided that the fermentations are pure. It is also necessary, in industry, to prepare the molasses ridding it of impurities harmful to fermentation, as well as infections of bad yeasts or bacteria.


In most Colonies, in almost all cane molasses distilleries, fermentation is completely left to itself; this is why we obtain such poor yields there, and often such detestable tafias!

Fermentation is spontaneous; in other words, since we know that there is no spontaneous fermentation, it is any germs in the air which seed themselves.

At the start of a season, as yeast spores are rare in the air, we often resort to empirical procedures which the foremen sometimes justify with singular theories. So, we add sometimes sulfuric acid, sometimes lime, sometimes ammonia salt. These chemicals can be useful in modifying the composition of the must, but they do not provide the ferment.

Other times, baker’s sourdough is used, which provides more lactic and foreign germs than real yeast.

The best practice is, without a doubt, to borrow cane ferments. For this, we take very fresh bagasse, fresh from the mill; we put it in a vat and fill it with diluted molasses. To make the operation even more correct, it is necessary, beforehand, to sterilize the vat and the diluted must. Fermentation does not take long to begin and we thus have a starter in which the yeast is a local yeast, that is to say suitable both for high temperatures and for the composition of the culture broth.

Once the first vat is fermenting, we take a little of its liquid to inoculate the next vat, and so on.

Although better than so-called spontaneous fermentation, this practice still leaves much to be desired, because the progressive infection of the must increases little by little.

It is necessary to point out in particular a special place where the infection develops as if at pleasure: it is the molasses diluting tank. We are used to making very large diluting tanks, both to better average the degree and composition of the must, and then to build up a sufficient reserve of it for the night. It is the antechamber of fermentation, and there, the must, left to its own devices, becomes even more deeply contaminated as the vat is rarely emptied completely and even more rarely. washed and sanitized.

On the contrary, we should hurry to send the must to fermentation as soon as it has just been diluted. It is not difficult to imagine very simple and very small combinations operating instantly and continuously to intimate the water and molasses mixture, and to adjust the two inlet taps so as to obtain the density that we desire.


Despite its concentration, cane molasses often undergoes sufficient decomposition in barrels to blow up the bungs and even the bottoms of barrels, releasing bad-smelling gasses. Such molasses are often so acidic that when diluted they produce a grappe which already has nearly 3 grams of acidity per liter. However, all volatile fatty acids are antiseptic and contrary to the start of fermentation; they can even stop the fermentation to make way for vicious secondary fermentations, acetic, butyric, etc… The best is therefore to treat these molasses with heat, and even with a little sulfuric acid, and to bubble it with the air in the boiling mass to remove the volatile acids. It is no longer a denitration, as they say in beet molasses distilleries, but a sterilization, using tools completely identical to the denitration of beet molasses.

In many cases this operation can be called desulfitation, because today sulfurous acid is used extensively in the manufacture of sugar. However, sulphites are antiseptic, and it is necessary to decompose them and remove them. They have another disadvantage, that of providing, at the end of fermentation, a fairly abundant release of hydrogen sulfide. The genesis of this gas was explained by Ruy-Pailhade, who attributed it to a reducing diastase of yeast, called by him, philothion, precisely because of its action on sulphites, and sometimes on the sulphates themselves.

Hydrogen sulfide is a yeast poison; it therefore hinders a good alcohol yield. But it is even more fatal from the point of view of aroma, because in contact with high alcohol content it generates sulfhydric ethers called “Mercaptan” whose nauseating and repulsive odor has a considerable malodorous power.

There cannot be good rum where mercaptan is formed. We must therefore go back to the source of this defect; to prevent it from taking root, it is necessary to radically remove the sulfurous acid from the molasses before sending them for fermentation.

The most radical and economical process consists of using the continuous sterilizer-desulfiter, E. Barbet system, of which we give an image opposite.

The first operation consists of diluting the molasses to approximately 28° Baumé and adding, if necessary, a small dose of sulfuric acid, in order to release the fatty acids and sulfurous acid. For this, two wooden vats equipped with an automatic mixer-diluter will be used, the dilution liquid consisting of boiling dunder.

The mixture, having a density which will correspond approximately to 26-28° Baumé at 15-20° temperature, will be pumped and sent to a charging tank on the continuous sterilizer. This works like a distillation column, and has all the appropriate regulation and monitoring organs. It is good to inject air at the same time as steam into the base in order to facilitate the expulsion of volatile acids and bad odors.

The syrup coming out of the bottom of the sterilizer needs to be diluted and cooled before going to fermentation. We divide it into two unequal lots. The largest batch passes through a refrigerant, then through an automatic diluter; The dilution liquid is water, but you can add a little more dunder if you want.

As for the second batch, it will be used, as we will explain, to prepare pure yeast starters and intermediate yeast starters.

The vapors which emerge from the sterilizer can be taken into a tubular refrigerant where they condense, providing an acidic liquid, which has a more or less pleasant odor. When the molasses is of good vintage and good quality, and without sulfurous acid, the acidic water scent will not be unpleasant; Also, we will keep it to mix either with the fermented grappe, or with the raw tafia from the first distillation.

If the odor is bad, there will be no need to condense the vapors, and we will immediately see the practical usefulness of continuous sterilization, through the expulsion of bad odors. Without this operation, all these products, remaining in the grappe, would have been released during distillation, further depreciating the quality of the tafia.

Pure Yeast. 

Various chemists have worked for several years to propagate the use of pure yeast in industry; firstly, studies were carried out to select breeds of yeasts acclimatized and appropriate to the nature of the musts to be fermented.

But it is not enough to have a good breed of yeast, you must have it in full vigor, and you must, at the same time, provide the vats with a sufficient number of active cells so that the fermentation of the vats takes place in a fairly short time, around 36 to 40 hours for molasses musts.

The stumbling block of almost all previous attempts has been the preparation of pure starters to be produced with the pure yeast culture supplied by the laboratories.

We started with three or four liters of yeast with which a first starter of 50 to 70 liters was prepared at the factory. This starter was transferred into a second 300 or 400 liter basin, then into a 20 or 30 hectoliter tank, and finally it was this starter which served as the “stock” for the large fermentation vats.

In this whole series of very long operations, contamination was impossible to avoid, so that in the final analysis the starter was neither purer nor more active than the pressed brewer’s yeast previously used. In addition, these preparations were very complicated.

Our tools are much more practical, and make it possible to obtain in a single operation numerous, abundant starters, loaded with a mass of vigorous cells.

Pure starter apparatus.

The pure starter apparatus is constructed either of tinned copper on the inside or of enameled cast iron. On the side is a smaller device that serves as a seeder; both are closed, and before putting them into service they are sterilized by putting steam into them for a sufficient time. They must also be washed in very hot acidulated water to dissolve the oxidation of the metal.

The seeder has an upper loading orifice, and it looks quite similar to a small device for saccharifying grains by acid. It is used, in fact–as much as possible–as a saccharifier of corn flour, a corn mash thus saccharified being the best thing to wake up the yeast and give it its favorite foods. After saccharification, the acid is saturated with a little wet chalk, the apparatus is closed, it is cooled by watering its surface, and when it is at 32° it is inoculated with the necessary precautions.

When this starter is very active, we load the large yeasting machine with a little sterilized juice, and we pass the starter through it by air pressure.

The yeasting machine is equipped with a coil through which cold or hot water can be passed as needed: it also has a sterilized air bubbler and a liquid emulsion system.

Pasteur showed that cultivation on a very aerated surface gave the yeast a very particular vigor, and even went so far as to modify the shape of the cells. At the same time, it exalts the property of yeast to sporulate. In a word, when a yeast is tired, we revive it by cultivating it on the surface, in a must to which we give as little depth as possible. Pasteur carried out this program very simply by putting very little liquid in a large flat-bottomed glass flask.

For our industrial yeast culture, we followed the Master’s precepts, which Mr. Calmette had taken care to remind us of, by installing in the upper part of the apparatus a series of very shallow copper trays, on which we continually force the must to rise and spread. We are thus doing aerobiosis.

We borrow from sterilized compressed air the driving force necessary to generate the incessant circulation of the must and this by means of emulsion tubes, the number of which depends on the power of the starter apparatus.

This “emulsion” system is widely used to raise acids, without the use of pumps.

We applied it to our starter apparatus and this in a way that was all the more advantageous as the expenditure of compressed air was already obligatory in any case. Instead of sending air to the bottom of the must through a suitable bubbler, which always uses the oxygen in the air incompletely, we use it to produce the continual rise of the must on the upper plates, and there it spreads widely on contact with the liquid, dissolving there and invigorating the yeast cells.

The starter apparatus has a dial thermometer, a water coil, a level tube, a safety valve and other control devices. It is also equipped with a special carbon dioxide exhaust vessel acting as a hydraulic shutter, and making it possible to judge the intensity of the fermentation in the device by the size and speed of the gas bubbles which bubble through the liquid.

When the starter is ripe, only two-thirds or more of it is taken for the fermentation vats. The rest is used to perpetuate the fermentation in the device which is filled again with sterilized and cooled must. Every eight or ten hours we can take a new starter for the vat room.

This is the process. All accessories are designed to prevent any internal contamination, and in this way the starter can be kept pure for months. Only occasionally do we empty the device, clean it and sterilize it again to begin with new, pure seed.

Molasses alone would not be an excellent culture broth for yeast. It can be improved by the addition of various substances, ammonia salts, phosphates, etc., or better by adding a small proportion of corn mash saccharified with acid, or a little malt peptone.

The lateral device can be used to make a small batch of corn flour every day. Or we will cook brewer’s yeast in this device, or even yeast collected at the bottom of the fermentation tanks. This provides excellent food for new generations of yeast.

Our starter apparatus is so effective in maintaining the purity of its fermentation that we can, if necessary, dispense with using pure seeds and simply purify the native yeasts. This is particularly valuable in sugarcane countries, because yeast seeds are likely to arrive contaminated or almost dead. The owners of cane molasses distilleries set up by us have often told us that, rather than buying seed in Europe which is not acclimatized to hot countries, they prefer to inoculate the device with a local ferment. After 4 to 5 generations in the apparatus with intensive aerobiosis, the leaven has purified and transformed. And at least we thus have a breed of yeast well adapted to the raw material and the climate. 

Continuous fermentation.

Rum distilleries are generally equipped with a very large number of small fermentation tanks. Large masses of liquid are avoided because the fermentation temperature rises too high there.

But the multiplicity of vats is such that we cannot have a pure starter for each of them, the yeasting machine cannot provide more than 2 to 3 starters per 24 hours. This situation is remedied by the use of intermediate tanks, with open air, but equipped with powerful air bubblers. These vats are a sort of extension of the yeasting apparatus, and like it, they are fed with sterilized juice. Their fermentation is still almost pure, and they provide a much more copious seed base.

We cannot establish an invariable rule for the number and size of these intermediate tanks; this essentially depends on the number and capacity of the fermentation tanks and you must consult us for each particular case. In any case we will remember that in these intermediate tanks we can no longer carry out indefinite fermentation as with the yeasting apparatus, because the tanks are open and exposed to contamination. We must therefore ask each intermediary to provide a very limited number of vats, and then liquidate it to start a new operation with a starter from the pure yeasting apparatus.

Our system produces very active and industrially pure fermentations. It only takes 36 to 40 hours for each fermentation tank, instead of these endless 5 to 6 day operations which don’t produce anything good.

Then, as we explained above, the alcohol yield is assured. It is now a matter of intensifying the aroma as much as possible.

The vats being established for long fermentations, these vats find themselves exaggeratedly powerful when, through our pure and active fermentation, we arrive at the fall in 36-40 hours. But we take advantage of this power to make the vat wait a long time before distilling it. During this period acidification bacteria develop to the detriment of the nitrogenous and hydrocarbon materials from which the yeast has not been able to make its nourishment. If necessary, we carry out a coupage with an old vat, in order to immediately seed it with acidifying bacteria.

Use of dunder for diluting molasses

Another factor of great importance to enhance the fragrant richness of the tafias is the reuse of dunder to dilute the molasses before fermentation.

Dunder acts directly and also indirectly on the quality of the distilled products.

Direct action is easy to understand. It results from the fact that the aromas take a very long time to come out through distillation, so that dunder, when they leave the column or the still, are far from having lost all their natural perfume. By their return to work, we recover in the next distillation part of what had escaped the first. 

This practice is in constant use in Martinique, and also in other colonies renowned for the quality of their rums, English Guyana, Jamaica, etc.

It also acts indirectly by improving fermentation and regulating it.

On the one hand, in fact, the boiling in the distillation column brought about a certain dissolution in the dunder of the constituent materials of the yeast. It is especially the nitrogenous materials which are peptonized, and which thus constitute an excellent culture broth for the subsequent fermentation. This is how in laboratories it is customary to boil brewer’s yeast in water to form the best culture broth. Industrial dunder is found in the same conditions, since the wine boils in the column in the presence of all its yeast.

On the other hand, dunder is acidic, and its use will provide free acidity to the must, while mineral acids are always very expensive for the colonies.

You should not be afraid of having a fairly high initial acidity in the must, which can reach 3 and even 4 grams as sulfuric acid per liter. As long as it is an organic acidity and is not due to volatile fatty acids, it does not upset the yeast, but only the bacteria.

Premature development of acids, lactic, acetic, butyric, etc., is the great bane of exotic fermentations. In a work published by M. Pairault, chief pharmacist of the colonies (1), this author showed that the acidity often reached 10 grams per liter, which infallibly stops fermentation. He also found that the yield was often less than half of the theoretical yield. The initial acidity alone can ensure relative purity of the fermentation, in the absence of the Pasteurian methods that we described above.

(1) Annales de Distillerie et de Brasserie, du Dr Fernbach, 1re année.

In our methods themselves, dunder acidity plays a beneficial and protective role, because we must not forget that fermentation, well underway in our pure starter apparatus, ends in the open air.

Cane molasses salts.

Since aerobic yeasts are very vigorous, we can allow a significant proportion of dunder to enter the musts in the vat room, without having to fear that the accumulation of salts and organic matter will stop fermentation.

As dunder becomes richer in salts, we can profitably extract potash salts by evaporation and incineration, as we are going to prove. Now potash is a product of a certain value, because almost all hot countries are obliged to import their alkalis from Europe and the United States.

Let us assume a daily work of 25,000 kilos of molasses, providing approximately 1,000 hectoliters of dunder. All the potash from the 25,000 kilos of molasses is dissolved in the thousand hectoliters of residue.

But let’s admit that, every day, 500 hectoliters of dunder are taken to help dilute the 25,000 kilos of new molasses; the daily rotation being established in this way, the result will be that every day only 500 hectoliters of dunder will be sent to the potash, and that these 500 hectoliters will contain exactly the potash of the 25,000 kilos of molasses. This is mandatory so that the potash output is equal to the work input. Therefore the dunder to be burned daily is reduced to half the volume of what was there by the previous work, hence saving half the fuel.

It remains to be verified whether this fermentation work at high density and high content of impurities of all kinds should not hinder fermentation and yield.

Let us return to the two Egyptian molasses No. 1 and No. 2, which we spoke about on p. 11. In the distilled dunder, let’s dilute a new quantity of No. 1 and No. 2 respectively in a proportion such that we again have 250 grams of fresh molasses per liter, that is to say the largest dose used industrially. This gives enormous initial densities, 1.109 and 1.102, or approximately 14° Baumé, because of the inflow of dunder.

Let’s add the same dose of yeast as the first time, 5 grams p. 100 grams of molasses.

We see the fermentation complete in 42 hours and give us good results in alcohol yield:

We should not attach importance to the figure of unfermented sugar, because there is an accumulation unfermented from the two successive fermentations. What is important is the alcohol yield. For the first molasses which is of poor quality, the yield was lowered by 0.89 p. 100, but for molasses No. 2 which is good, the alcoholic yield was on the contrary very slightly better, i.e. 0.27 p. 100 more. Therefore, we can claim on average the same industrial yield.

However, beet molasses dunder, such as is sent for incineration, contains only about 3 p. 100 ashes; we see that in the above case we obtained a liquid much richer in salts, and consequently having to give certain benefits through evaporation and incineration.

In the memoir of H. Pellet, cited above, we see many examples of industrial salt analyzes made with cane molasses dunder. In general, there is a fairly large proportion of insoluble matter, composed mainly of carbonate of lime, silica and coal; this comes from the abuse that we make of lime during fermentation. We sometimes see the insoluble reach 50 p. 100.

Carbonate of potash content fluctuates due to the insoluble content, from 10 to 35 and even 40 p. 100; on average 20 to 25.

If we consider that with the slightest triple effect concentration bringing the dunder to 11° Baumé, incineration is done without any expenditure of coal, we see that the extraction of the salt will certainly give a big profit, whatever the cost of fuel in the colonies.

The simplest and easiest to clean triple-effect system is the Barbet system: instead of directly heating the distillation column with high-pressure steam, we install either a small triple-effect, or even a double-effect working under pressure. Live steam boils the dunder at a low pressure; the steam produced in this device heats the base of the column, either by tubular, coil or double bottom.

Dunder thus concentrated is self-evaporating in the oven, that is to say that the combustion of the organic materials, during incineration, is sufficient to complete the evaporation of the water without expending fuel, except at the beginning for lighting the potash furnace.

Neither in the triple-effect nor in the oven do we have to spend coal; the salts are therefore obtained free of charge.

Instead of incineration, even greater profits can be obtained from dunder.

To do this, dunder will be concentrated to 41/42° Baumé, and glycerine from fermentation will be extracted using the Barbet processes. Glycerin has a very high value (from 150 to 200 francs per 100 kilos) and we will not forget that a factory which produces 20 hectoliters of alcohol per day therefore generates more than 100 kilos of glycerin per day. It is therefore a very valuable by-product. Equipment for extracting this glycerin quickly pays for itself.

Finally, the residue itself from the extraction of glycerin is in the form of a dry, granulated fertilizer which contains all the nitrogen and all the potash of the molasses.

Molasses which provides 20 hectoliters of alcohol at 100° contains 85 to 90 kilos of nitrogen, which are worth 1 fr. 50 per degree, therefore a new daily value of 125 to 135 francs. Finally, this fertilizer contains all the potash which is very useful to return to the earth, since cane requires quite large quantities.

We particularly draw attention to this rational and integral use of by-products from cane molasses distillation. It is at the same time a hygienic solution because we will avoid poisoning the waterways with dunder, as is too often done.


Distillation with purification and aging.–When the molasses are of good origin, and when the pure fermentation that we have just described has been carried out, you will still have to worry about choosing a good distillation device.

Father Labat’s device is a simple Charentais-style still. The rums it produces are good, but this device is very bulky and very expensive in fuel due to the fact that the distillation is done in isolated batches. Finally, it does not lend itself to large productions.

Continuous distillation devices have major disadvantages against them. Criticisms are in fact justified when we claim to use common columns, such as those found in industrial distilleries in Europe.

But the Barbet apparatus represented on p. 24, far from deserving distrust, on the contrary produces exceptionally fine rums with a very correct and very intense aroma.

Here is the application of the device as it works at the Vauclin rhummerie (Martinique).

The fermented wine enters continuously; it is regulated by a sensitive tap and passes successively into the refrigerant H and the wine heater G. From there it arrives at D in a first small column intended to expel the bad fermentation gases. Boiling occurs by means of a little alcoholic vapor taken from the top of the distillation column C C’ by the valve S. The mixture of gases and light vapors goes into the water condenser E. The gases and a few acrid aldehyde vapors escape into the atmosphere while the remainder condenses and enters D.

The wine, purified of its bad gases, descends into the exhaustion trays C’C, and from there passes into the copper boiler B, which is heated over an open fire (hearth A). You can also heat using a steam coil.

The boiler has a large enough capacity so that the dunder can undergo a long cooking process, which brings out all the flavors. [Essentially flavoured steam.]

The alcoholic and acid vapor passes to the base of the second column with plates F, which receives in the opposite direction the reflux from the tubulars G and H. On these plates the combination of the acid vapors and the alcohol takes place, this is that is to say etherification, source of most of the flavors of rum.

Reboiled, pasteurized, softened rum is extracted in the liquid state, and cooled in the refrigerant I. It goes from there to the test tube P. At the same time a very small proportion (1 or even 1/2 p. 100) of aldehydes which cool in a second coil I, and flow from there to the test tube T.

This device is very easy to operate. It ensures perfect exhaustion at the same time as great fragrance power.

It was further improved recently by the adaptation of the heat-resistant cooling tray system in our house. These trays are particularly effective in retaining tail impurities, caproic or others, which give a bad aftertaste to rums and brandies. They also reduce the fuel costs of the appliances to a minimum.

Head and tail products separated by the Barbet apparatus can sometimes be brought together either to undergo a new recycling with the apparatus, or even to constitute a sort of “essence of rum”, possessing to the highest degree the qualities of coverage requested by the trade (1).

(1) We say that a tafia has a lot of “coverage” when this tafia has a strong enough aroma to well cover the quantity of industrial alcohol that is incorporated into it.

It is worth pointing out that not all molasses are equally suitable for the production of fine rum. There are vintages for rum as well as for brandy; these two industries having the greatest similarity, the same principles of fermentation and distillation are applicable to them.

Generally speaking, we will choose molasses from acid sugar processing as we said above (page 7).

The method of extraction of the vesou must also have a certain influence on the quality of the molasses, because diffusion often brings gummy and waxy products into the juice in greater quantities than the work of the mills. It would be interesting to make special observations in this vein, to know whether the diffusion is favorable or unfavorable.

Rectification of the tafia. 

In many colonies there is an interest in producing neutral rectified alcohol, either for the fortification of wines from Europe or for the manufacture of liqueurs.

Tafia can be rectified as well as other raw alcohols made from beets or molasses, but it is obvious that this operation is all the easier as the tafia itself is less impure and less odorant.

So our pure fermentation methods are even more necessary in this case than for rum making. However, we will be careful not to overuse dunder and we will distill the vats as soon as they have fallen.

In the case of manufacturing neutral alcohol, it would be wise to resort to the use of direct continuous rectification of wines, which has all the advantages of simplicity of installation and operation, as well as economy and fuel cost.

The operation continuously splits head, tail and extra-tail products. Just as we explained earlier, these three batches of impurities which contain all the odorous products that the tafia contained, form by their combination a sort of rum essence which can be recycled or sold as is, according to commercial conditions.

Storage and transport of molasses.

We said at the start that cane molasses should make it possible to make six million hectoliters of tafia at 60° per year.

We are very far from it, and this is mainly due to the difficulties of transporting and storing molasses. In Cuba, Java, Louisiana, more than two-thirds of molasses is certainly thrown away; some factories are trying to burn them.

Sugar refineries give up distilling them on site as they have poor yields in terms of quantity and quality; moreover, having no outlet locally, they would have to transport the barrels of tafia, even though there is no road, and barrels are extremely expensive. Tropical climates are not at all suitable for transporting and keeping empty barrels in good condition; the sun quickly disintegrates them, and despite rebuilding and refilling, there is considerable transport waste. And then returning empty barrels is a big expense.

In three-quarters of cases, the value of the barrel is at least double that of the molasses it contains.

Molasses distilleries would have to be installed in the ports to make it easy to ship the tafias, then it would be necessary to manage to economically transport the molasses to these distilleries, and to allow easy warehousing and storage.

The solution to this problem is quite simple and practical: it consists of cooking the molasses into candy, in a vacuum if possible, until the so-called “broken” proof and pouring it either into sugar loaf shapes , or in any parallel sided shapes having relief.

All sugar mills have apparatus for cooking molasses, so all you have to do is obtain a certain supply of molasses loaf forms. To reduce the necessary supply, the “loaf pans” can be placed in sheet metal trays where cold water is circulated; We thus hasten the setting of the molasses loaf and the form becomes more quickly available.

Before pouring the molasses, we line the inside of the shape with a sheet of wrapping paper. We pour, and when the setting is made by cooling, we remove. Molasses bread is found all wrapped in paper.

This paper prevents the loaves from sticking together during transport; it also protects the molasses against humidity in the air, because it is a little hygroscopic.

We understand that the transport of these molasses loaves becomes extremely simple; the handling is identical to that of sugar loaves, and there is no dead weight to transport, except the paper which is negligible compared to the dead weight of the barrel. There is no longer any runoff or road waste, which often exceeds 2 p. 100.

Molasses concentration itself provides a very significant saving in transport, because it has removed 10 to 12 p. 100 water; there are only 88 to 90 left p 100 of the original weight.

For colonies deprived of roads, and this is the majority, transport can now be done on the back of a mule; moreover, we will keep the molasses loaves in ordinary stores, instead of having to build huge sheet metal tanks.

Finally, we no longer have to fear these frequent alterations of the molasses which make it unfermentable or which even go so far as to cause barrels to burst. Given the concrete state and the total solidification, there is no longer any possible alteration, especially since the mass has been sterilized by cooking which brought it to solidification. Storability is almost indefinite.


We have successively reviewed all the operations of the rum industry and in each of them we have shown how many defects there were in ancient errors.

All the improvements we have described are easy to apply.

Pure starters free the colonial distiller from the worry of supplying brewer’s yeast, which is almost always impossible, and they give them a yeast completely suited to their industry. Because when you want to make rum you wouldn’t believe how important the breed of yeast is. Even among yeasts resulting from the fermentations of the various colonies, or removed from the different varieties of sugar cane, there is a quantity which gives almost no aroma.

Among a few good species it was still necessary to choose those which simultaneously gave a good attenuation of the must, that is to say the yield.

Pure yeast starter apparatus, intermediaries for the continuous initiation of fermentation, are only additions of material, which leave the ordinary tools of open vats in place.

Continuous molasses sterilization and aseptic diluting with the help of dunder have such an influence on the increase in yield that installation costs, very modest in fact, are quickly amortized.

Dunder reuse in a large proportion provides great services for the purity of fermentations; it is even more precious for certain factories which have very little fresh water at their disposal. It is only possible at high doses thanks to the exceptional vigor of our aerobic starters, and then it provides the distiller almost free of charge with a by-product which is not lacking in value: potash salts.

Finally, the problem of rum distillation and that of molasses storage and transport receive, through our processes, eminently elegant and practical solutions.

The combination of all these improvements makes rum manufacture truly profitable for the colonies, and at the same time will restore to it an outlet in Europe and other countries which tends to diminish.

Could it be otherwise, when we see the qualities of rums that are currently in circulation! Three quarters of these are adulterated rums which present a real danger to public health, while it is so easy (we believe we have demonstrated this) to provide products of excellent quality at prices much lower than current sophistications.

In so many colonies, molasses has so little value that it must easily outweigh all other alcoholic materials. To do this, it is enough to adopt the paths of progress, and in particular Pasteurian methods, since they have already proven themselves victoriously in the industrial distillery of Europe. The fortune of the colonial distillery is this price.

Another interesting outlet is the manufacture of alcohol for lighting or industrial needs. In many regions, lighting alcohol will be produced at a much lower price than oil; it is a new resource that the colonies will do well to exploit for their own needs.

June 1913. 

1 thought on “Barbet—Rhum—Recent progress contributed to its Manufacture

Comments are closed.

Discover more from Boston Apothecary

Subscribe now to keep reading and get access to the full archive.

Continue reading

search previous next tag category expand menu location phone mail time cart zoom edit close