S.F. Ashby: The Study of Fermentations in the Manufacture of Jamaica Rum (1909)

Follow along: IG @birectifier

[For international readers, I made the other works of the experiment station available here.]

This paper is a reprinting from the Report on the Experimental Work published by the Jamaica Sugar Experiment Station run by S.F. Ashby and H.H. Cousins. One thing that we see immediately is a shift of attention from bacteria as defining Jamaica rum to also including yeast. Pretty much no other industrial scale beverage but rum uses a fission yeast. They are special. They are also the ideal partner for bacteria because, as Ashby shows us, they have remarkable tolerance for fatty acids that halt a budding yeast in its tracks. There is, however, lots of potential for the “unremarkable”. At one time, regarding his experimental ferments, Ashby remarks: “The rum could hardly be called by that name, and it showed the same character for all ten yeasts; in no case was any characteristic flavour produced.” These yeasts can produce very clean ferments and thus light rums. Not all examples are suited for heavy rum.

Pombe yeasts may be remarkable because of what you can built on top of them. But, they are also special in their own right. The most well selected examples may be able to produce the most valuable components within rum oil which are rose ketones, derived from carotene in the cane. They also produce significantly less fusel oil than budding yeasts. This fact can dramatically change the way a spirit is distilled (much lower ABV and/or deeper into the tails) further justifying pot distillation over column.

One of the big mysteries left in heavy rum production is whether (and if so, how exactly?) fission yeasts perform a large percentage of the aroma creation we project on to bacteria. Not many people ask: whats with all the vinegar at Hampden? If vinegar is a flaw and liability in all other fermentations, how is it a feature of heavy rum production? Pombe yeasts, under very specific conditions that we currently do not understand well, may be able to convert the short chain acetic acid of vinegar into higher value longer chain fatty acids (certain bacteria can do this as well). Arroyo did not go this route and nothing in his work touches upon it.

There are other minor mysteries as well. Ashby notes that top fermenting fission yeast ferments may produce significantly more yeast biomass. The microbiology literature notes that at every instance of cell division a portion of the cell wall dissolves into the ferment made of long chain fatty acids that can become esters. Open fermenters may be key to providing the oxygen needed for successful top colonization and that increase in biomass that can benefit aroma. A question emerges: Does increased biomass, that mostly spins its wheels and does not accelerate fermentation come with increased enzyme activity that can form rum oil? Karl Micko at about the same time was finding very characteristic un-saponifiable aroma in Jamaica rum using methods that would eventually become the birectifier.

[We have clarified things here. There is a lot of evidence in symbiotic ferments of fission yeasts and bacteria for chain elongation by both yeast and bacteria. We also demonstrated that rum oil (carotene based aroma) can be created by bacteria. We have repeatedly observed a fission yeast top ferment in a closed fermentor. Other factors besides oxygen may relate to top fermenting behavior and we think we know what they are.]

Arroyo used a high pH method of heavy rum production that relied on pure bacteria cultures as well as neurotically controlled conditions. Vinegar process Jamaica rum may counter intuitively rely more on yeast than bacteria and tap into mysterious seldom seen features of yeast metabolism to create aroma. This does not mean that aroma beneficial bacteria is not present, but high acid fermentation in open fermenters may work differently than is commonly believed.

[So far we’ve done one control ferment that saw appreciable acetic acid but no infected dunder. It likely had a common LAB infection but not a more significant mixed culture biome capable of elongation. Symbiosis may be whats its all about. These rums celebrate the fitment of many components.]

We have so far been experimenting with Arroyo style ferments, but it may be valuable to have first hand experience with the basic work Ashby carried out, remarkable results or not.

[We have pursued high acid ferments and started to see remarkable results.]

The International sugar journal. v.11 (no.121-132, 1909).

By S. F. ASHBY, B.Sc. (Fermentation Chemist).

*From the report of the Jamaica Sugar Experiment Station, 1907.

It had been established during the three years that my predecessor Mr. Charles Allan, B.Sc., had worked on the manufacture of Jamaica rum, that flavour was mainly due to the compound ethers. These bodies were considered as produced by chemical combination of alcohol with various volatile fatty acids during and after fermentation of the wash, and particularly during distillation. The alcohol was the product of the action of yeasts on the sugar in the wash, but the acids were the work of bacteria, being partly pre-formed in the materials for setting up the wash, and partly produced in the wash during and after the yeast fermentation. The following acids were found, acetic, propionic, butyric, capryllic, capric, lauric, all of which yielded ethers with alcohol capable of giving varied flavours to rum. Acetic ether was shown to constitute about 98% of all ethers in rum, but contributed little flavour and owing to its volatility was very transient. Butyric ether was found to be more valuable, but the ethers of the higher acids, capryllic, capric, and lauric, were held to be of special importance for giving both body and characteristic flavour.

As the yeasts were considered to be only alcohol producers, attention was mainly directed to the study of bacteria producing the valuable acids. One such bacterium was isolated and the conditions under which it works determined (Report, 1906. pages 136-137). A microscopical examination of washes showed the presence of two yeast types, distinguished by very different modes of multiplying: to the one type belonged the oval and sausage shaped forms which multiplied by budding (Saccharomycetes) whereas the other type reproduced by division through the middle of the cell, that is by “fission” Schizosaccharomycetes). The oval budding forms were alone seen in cane juice washes, but the fission type was found to be the characteristic fermenting yeast of both common clean and flavoured rum washes. The latter kind could not be isolated, and indeed no systematic experiments appear to have been made with any of the yeasts.

Mr. Percival H. Greig, of Westmoreland, was the first to isolate a number of Jamaica distillery yeasts, and to study their action on washes in a state of pure culture. In molasses and dunder which he took to Jorgensen’s Laboratory in Copenhagen in 1893, the fission type of yeast was discovered and studied for the first time. Greig continued to work with these yeast in Jamaica till 1896 and published part of his results in the Bulletin of the Botanical Department March, August, and September, 1895, and January, 1996). He observed marked differences in the time required for fermentation, amount of attenuation, and alcohol-yield with different yeasts and drew particular attention to a slow working top fermenting fission form which alone was able to produce an agreeable flavour in washes. He recognised the importance for flavouring of fruit ether in rum, but appeared to think that these bodies, in so far as they were not contained in the original juice of the cane, could be produced at will by pitching the wash with a suitable flavour engendering yeast. On these grounds he strongly advocated the employment of pure yeast cultures in Jamaican distilleries, and insisted that the distiller should strive to suppress the action of bacteria.

As previously indicated Mr. Allan took up the precisely opposite view, pushing the yeasts into a subordinate position and devoted his attention mainly to the search after flavour producing bacteria.

As the yeasts must always be the central factors in fermentations for the production of spirits, it appeared to me natural to devote first attention to them, and to observe in particular whether some are really able to engender flavours of value in Jamaica rum.


Early in the year I isolated and obtained in pure culture a number of the oval budding yeasts from washes in the laboratory distillery which were set up from a mixture of fresh cane juice and dunder, and about the same time some fission yeasts were secured from a dead wash sent in from the country. As the result of some preliminary fermentation experiments it was observed that the oval cane juice yeasts worked more rapidly in washes of low acidity, but with an acidity of nearly 1% the oval yeasts showed very sluggish fermentation, while the fission type worked as well at the high acidity as at the low.

It seemed desirable to study the effect of three common distillery acids, lactic, acetic, and butyric on the two types of yeast, and accordingly a number of fermentations were set going in cane juice and dunder washes to which varying quantities of the single acids were added before putting in the yeast. The oval budding yeasts all showed bottom fermentation phenomena, but the fission yeasts all showed strong top fermentation with the production of an abundant fatty head. A vigorous yeast of each kind was selected for the experiment with the acids.

The amounts of the different pure acids added are expressed also as sulphuric acid by weight % of the wash by volume. The amount of the yeast added was a far as possible the same for both types, except in the butyric acid series, which was carried out at a later date with a larger amount of yeast. The results are set out in
Table I., which shows the amounts of sugar fermented at the end of each day to the sixth day. The figures were obtained by daily weighings, multiplying the loss of weight by two and calculating the resulting numbers on the total amount of sugar originally present.

[It is very clever that he weights the entire ferments to measure the escaped CO2 and thus how much sugar was fermented.]

With regard to acetic acid the results show that the budding yeast is much more susceptible to it than the fission yeast. In the presence of a half % of this acid the budding yeast show greatly reduced fermentation during the first three days, whereas the fission yeast was but slightly affected. One % completely prevented the activity of the budding type, but again only slightly reduced the fission yeast fermentation. Both yeasts are very resistant against lactic acid, but even here 0.7% showed an injurious influence on the budding yeast, whereas 1.4% hardly reduced fermentation by the fission yeast. Butyric acid proved to be very poisonous for both yeasts, but whereas 0.15% wholly prevented the budding yeast from fermenting it caused the period of fermentation to be increased by only one day with the fission type. Even 0.4% did not completely suppress the latter’s activity, but 0.5% prevented all fermentation. The conclusion to be drawn from those results is that the budding yeasts are suitable only for the fermentation of weakly acid washes, whereas the fission type is at home in washes of high acidity. A notable point which the figures bring out is that where the acidity is low the budding yeasts get to work far more rapidly than the fission yeast. This is particularly well shown in the case where no acid was added. Although both yeasts completed the fermentation in five days, the budding yeast multiplied and fermented much stronger in the two first days. The ability of the budding type to multiply and ferment more rapidly from the outset in the weaker acid liquors, like cane juice washes and fresh skimmings, explains why this is the only kind found in such liquor the acidity of which is generally under 0.5%. In the usual estate washes containing dunder, molasses, acid skimmings, and frequently specially added acid, the budding yeast is largely suppressed, but the more slowly developing and very acid resistant fission type takes possession, and is practically the only form found in washes the acidity of which is 1.0% and over.

[“specially added acid” here would be cane vinegar.]


In March I collected samples of fermenting washes, dead washes, skimmings, dunder, acid and rum, from several estates in Westmoreland and St. James, and from the washes was able to gain pure culture of many fission yeasts. These cultures were started from a single cell according to the method of Hansen in order to prevent the possibility of any of the growths consisting of mixtures. With ten of these derived from four estates a fermentation series was set going in a wash of the composition:—

The Brix was 17.4, the acidity 0.48%, and the total sugar present 14.5%.

The yeasts 3 and 9, although pure fission forms, showed a totally different kind of fermentation to most of the others, the yeast gathering mostly into a coherent mass at the bottom of the vessels, the bubbles breaking on the surface being glassy clear and containing practically no cells. This fermentation was evidently strictly of the bottom kind. Yeast 5 showed mainly bottom fermentation phenomena, but produced also a slight yeasty head. All the other yeasts formed a strong glistening brownish white head at the surface and the bubbles were thickly cloudy, these yeasts were accordingly strongly top fermenting. Under the microscope the two forms could be distinguished easily, the bottom type showing isolated and paired cells, but never more than two together, whereas the top yeasts showed long chains of four or more cells interlaced and apparently branched. Yeast 5 showed no chains but the cells were often united mechanically into flocks.

[This is quite fascinating and I will have to start observing the bubbles as best I can.]

The bottom yeast 3 und 9 completed the fermentation in two days less than the top forms, yeast 5 occupying an intermediate position. This character of the bottom yeast to ferment more vigorously than the top kind has preserved itself in all subsequent experiments. The Increase of acidity due to the yeast alone, all bacteria having been excluded, amounts to only about 0.1%. The attenuation was very much the same in all cases, but the highest amounts of proof spirit were obtained from the bottom yeast 9 and the mainly bottom yeast 5. The yield of proof spirit per degree attention was good, in four cases exceeding unity. The distillation was effected from glass apparatus with one retort, the liquor being divided into two parts, the first yielding high wines of 20 O.P. and the second portion giving rum of 40 O.P. with the high wines in the retort. The rum could hardly be called by that name, and it showed the same character for all ten yeasts; in no case was any characteristic flavour produced.

[Notice that Ashby has no cane vinegar in his wash. He also is making a single batch and not recycling dunder enough to accumulate time under heat. Arroyo also emphasized that the characteristic odor of rum oil was developed during molasses pre-treatment. These may be reasons he produced a bland rum.]

In another experiment with dunder, molasses, and water, a much larger amount of dunder was used, namely one half the bulk of the wash.

The Brix of the mixture was 18.6, acidity 0.7%.

The bottom yeast here showed a gain of two to three days in the fermentation period. The yield of proof spirit was very high. The rum obtained was very light, and gave no difference in flavour with the different yeasts.

[Note, another light rum even though he used more dunder.]

In another fermentation series with the yeasts 2 and 9 pure volatile acids were added to the molasses and dunder wash before pitching with the yeast. The Brix was 18.6, the natural acidity of the wash 0.46. Acetic acid was added equal to 0.5 acidity, and butyric acid equal to 0.1 acidity, so that the total acidity before fermentation amounted to 1.06.

The large amount of volatile acid added had a marked effect in slowing fermentation, the time required, as compared with the previous experiments, being 10 days as against six days with the bottom form, and 16 days as against nine with top yeast. The rum showed an improvement in flavour, and with the top yeast contained more than twice as much ether. This was due to the much longer period during which alcohol and volatile acids could react chemically to produce ethers in the wash containing the top yeast.

[Acids and esters make the rum heavier, but the rum quality may also benefit from “radiance” produced from the perceptual effect of rum oil on the esters. An extra long ferment under increasingly acidic conditions may also benefit from the “staling” mechanism of developing aroma related to carotene.]

The conclusion to be drawn from these experiments is that, whereas none of the fission yeast isolated from the estate washes was able to produce flavour on its own account, the top yeast owing to its slower fermentation admitted a greater amount of chemical ether production in a wash originally high in volatile acids. The latter result is in accordance with distillers’ experience as they consider that a wash showing a strong fatty head due to the top fermenting fission yeast yields the best flavoured rum.


It is well known that the alcohol accumulating during fermentation has beyond a certain concentration, different with different yeasts, a marked slowing effect on fermentation and finally stops it altogether. In order to test the maximum amount of alcohol endured by the Jamaica fission yeasts, it was necessary to set up a wash of very high gravity. In a first experiment with the yeasts 2 and 9, a wash consisting of 4000 dunder, 1600 molasses, and 700 water was set up at 30 Brix. This was practically completely fermented, so that the alcohol formed was below the maximum which the yeasts could endure. In a second series, a wash of 30° Brix was set up with molasses and an extract of yeast, and after some days a further quantity of molasses was added. In this case both yeasts stopped fermenting due to the action of the alcohol, while there was still abundant sugar left in the wash. The data and results of the Experiments are given in Table IV.

The first experiments show a complete fermentation by both yeasts, the bottom for taking 5 days less than the top yeast. The bottom yeast also shows higher yield of proof spirit. The influence of the accumulating alcohol on fermentation is very marked, for whereas the bottom yeast had produced 16.5% proof spirit in 7 days, only 7% more spirit was produced in the following 11 days.

[I have been making similar firsthand observations. What should be noted from this table is that Ashby is getting pretty incredible alcohol levels but he does not have the volatile acidity seen in grand arôme rums. Galion does not achieve an alcohol level above 5% and I’ve been told for Jamaica’s heaviest ferments, if you get above 5% you are doing really good.]

In the second experiment the maximum yield of alcohol which prevented all further fermentation was just under 25% with the bottom yeast and just over 23% with the top form; while the bottom yeast yielded 17% proof spirit in 7 days, only 7.7% more was produced in the following 12 days. A similar effect of the alcohol is shown by the top yeast. The top yeast showed a rather sudden falling off in fermentation with about 18.5% proof spirit present, but the top yeast gave a more gradual falling off; it appeared however, to be susceptible at about 16.5% proof spirit. The mixture of the two yeasts showed throughout intermediate results.

It is evident from these results that the fission yeasts which work the estate washes are capable of yielding very large amounts of alcohol in pure culture with abundant time at their disposal. Fermentation is rapid and uniform for 7-9 days, during which 16-18% of proof spirit is yielded. This means that a wash containing about 16% of sugar can be fermented in a reasonable time. Above this amount the loss often becomes serious owing to sluggish fermentation. This fact has been recognised in practical distillery work, so that estate washes are rarely set up with more than 16% of sugar and usually with less.


This experiment was devised with a view to observing the effect of varying the amount of sugar in the wash, on time, attenuation, and yield of proof spirit. The washes all contained the same proportion of dunder, namely, three-fifths, the gravity being varied by means of the molasses. The results were as follows:—

Here as usual the bottom yeast is the most rapid worker, showing a gain of three days. The time required is least with the lowest gravity, but there is a difference of two days between the 25 and 20 settings and of only one day between the 20 and 15 settings. This difference hardly shows itself during the period of the main fermentation. After five days the relative amounts of sugar fermented by the bottom yeast were 35, 51. and 68.

As there was a half more sugar in the 20 setting than is in the 15, and twice a much in the 25 setting, these figures indicate that the activity of fermentation was proportional to the amount of sugar present, i.e., in a given time twice as much sugar was fermented in the 25 setting as in the 15 setting, the 20 setting coming half way between. The difference, however, was shown by the time taken by the wash to die off after the main fermentation was over. The 25 setting took three days to die, the 20 setting one day, and the 15 setting only a few hours. The yield of proof spirit was as high for the highest gravity as for the lowest, and the bottom yeast gave as usual the best results.

On the other hand there was markedly more sugar left unfermented in the highest setting than in the other two, and the bottom yeast in all three cases left more than the top yeast. The dunder employed in this series was a light cane juice product having a Brix of 9 and an acidity of only 1.2. The amount which had to be used (1/3 of the wash) to secure a normal acidity was more than is usual in practical operations, where the dunder has an acidity of over 2%. The result was that the relative amount of sugar in the wash was low, and the attenuation and yield of proof spirit low also.

(To be continued.)

By S. F. ASHBY, B.Sc. (Fermentation Chemist).
(Continued from page 251.)

*From the Report of the Jamaica Sugar Experiment Station, 1907.


The yeast produced in some of the fermentations of the last experiment was collected, dried in the air and weighed. The results are shown in pounds for 1000 gallons of wash.

One pound of air dried yeast ferments sugar in pounds :—

The top yeast produces a half more yeast substance than the bottom yeast, consequently a pound of the bottom yeast is able to ferment a much greater amount of sugar. The amount of yeast produced by the top variety falls away with the reduction in gravity of the wash, so that only one half as much yeast is produced in a 15 Brix setting as in one at 25 Brix. The amount of yeast produced is proportional to the amount of fermentable sugar present for washes from 25 to 15 Brix, but at 30 Brix relatively less yeast is produced, so that the ratio to sugar fermented is wider.

At first sight it seems inconsistent that the top yeast should often attenuate more than the bottom yeast and leave less sugar unfermented, yet give a lower yield of proof spirit. The above results, show however, that it removes more sugar to build up its substance than the bottom yeast, and owing to its habit of gathering at the surface of the wash in intimate contact with the air, respiration is more active, causing a greater loss of sugar by combustion into water and carbonic acid. The bottom yeast is consequently a more economical worker.

[The observation may prove very key to heavy rum production. The top fermenting yeasts are sort of spinning their wheels. Every incidence of cell division increases aroma and their may be aroma beneficial enzyme activity the whole time.]

Stability of the Two Varieties.

Distillers often observe that during the advance of the season their fermentations which were at first of the bottom type, tend more and more to top characters, suggesting either a conversion of the bottom yeast into the top or else a gradual displacement of the former by the latter due to some change in the composition of the wash which favours the top yeasts. That top and bottom fermentation may proceed in the same wash was evident from the fact that both forms were in several cases isolated from tho same material.

Some observations have made it seem probable to me that at any rate one of the varieties is not stable. The fission yeast No. 3 when freshly isolated showed wholly bottom fermentation phenomena, and agreed entirely with the other bottom yeasts. It was allowed to lie for two months under a fermented cane juice wash, and was then freshened up again. I was surprised to find that it no longer showed bottom fermentation, but gave a strongly marked top fermentation. On comparing its behaviour with that of yeasts which had always been top fermentation, it was found that it gave quite similar results, with an equally slow fermentation and a lower yield of alcohol than the bottom yeasts. Under the microscope it was also identical with the top form. The view which remained for many years unchallenged in Europe was, that the top and bottom yeasts were distinct types, the one never passing into tho other. Quite recently Hansen has shown, however, that there is always a tendency to vary, and has actually obtained the one form from the other in the case of a number of brewery and winery budding yeasts. There appears to be much greater tendency for bottom yeasts to go over into the top form than vice versa. Further observations must show whether the fission yeasts are particularly liable to vary in this way, and whether the change so often seen in distilleries in Jamaica from bottom to top fermentation is due to a variation of the yeast.

[This is quite fascinating and makes it trickier to evaluate fission yeasts. In our project, to narrow down the field, the first thing Cory did was single out the top fermenting yeasts. I have no real idea how they could switch besides the possibility of mutation as they became better adapted to their environment.]

Conclusions with regard to the Two Varieties of Fission Yeast.

1. The bottom yeast is a characteristically more rapid worker than the top yeast giving a gain of 2 to 3 days in the fermentation period.
2. The bottom yeast forms less substance and consequently makes a smaller claim on the amount of food stuff in the wash.
3. The bottom yeast gives a rapid and uniform fermentation during the main period, but the wash dies slowly. The top yeast ferments very uniformly throughout, and shows no sharp transition to the final stage.
4. The yeasts attenuate about equally, but the bottom yeast gives a better yield of alcohol.
5. The top yeast leaves less unfermented sugar in the wash.
6. The bottom yeast gives a higher maximum yield of alcohol, namely 25% as against 23% with the top variety.
7. The bottom yeast shows the injurious effect of alcohol at a higher concentration than the top yeast, viz., 18 and 16 respectively.
8. Owing to its slower fermentation the top yeast admits of more ethers being produced in the wash than the bottom yeast where volatile acids are present. The rum is consequently better.


Owing to insufficient distillery space or small still capacity, it often happens that molasses have to be stored for weeks, during which period they undergo a rather active fermentation. This involves a loss of sugar, so that it seemed desirable to make some experiments with a view to (1) determining the amount of loss arising from the cause, (2) separating and studying the properties of the yeast causing the trouble, (3) finding a remedy for it.

Three yeasts were secured in pure culture from a fermenting molasses, all of which were able to set up fermentation in a liquor of very high gravity.

Yeast (a).—This was a budding form of the pastorianus type which formed spores on the gypsum block at the air temperature in under 18 hours. Transferred to mixtures of molasses and water of increasing gravity it fermented actively at 45 Brix, feebly at 60 Brix, and showed no fermentation in molasses alone of 90 Brix. It was therefore not the kind active in the stored material.

Yeast (b).—This was a fruit ether producing yeast forming a dry wrinkled friable skin on ordinary washes. It was a small budding yeast which formed hat-shaped spores on the gypsum block in 24 hours. It fermented strongly in molasses and water of 45 Brix, more weakly at 60 Brix, and not at all in molasses alone. It was also therefore not the form desired.

Yeast (c).—This was a small spherical or oval budding form characterised by the production of branched chains of cells in weakly acid washes, and a very abundant multiplication. It formed no spores and no skin on cane juice, but merely a yeast ring. It appeared therefore to be no true yeast, but a “torula.” This kind fermented actively in molasses and water of 45 and 60 Brix, and also in pure molasses of 90 Brix. It corresponded to the form most abundantly present in the original material, and was evidently the true agent.

As an alkaline medium acts very unfavourably on yeast fermentation lime suggested itself as the first substance to try as a remedy. In one experiment the molasses were allowed to ferment spontaneously without the addition of lime, and with the additions of 6, 12, and 18 lbs. of dry line to every 100 gallons of molasses, the lime being added as fresh milk of lime and well stirred in. The same experiment was repeated with sterile molasses into which a pure culture of yeast (c) had been introduced, but here only 3 and 6 lbs. of lime were used. The fresh molasses had a Brix of 90 and contained nearly 70% of sugars. After six weeks the Brix was determined and found to be as follows:—

The molasses alone fermented strongly with crude and pure yeast from the outset. With 6 lbs. of lime there was no fermentation for nearly three weeks, when it started, but was much stronger in the pure yeast culture. 3 lbs, of lime in the pure yeast culture did not prevent fermentation from starting within a few days. With 12 lbs. of lime in the crude culture fermentation had only just started between the fifth and sixth week. With 18 lbs. of lime there was no growth of yeast and no fermentation. In the crude there was a maximum loss equal to 13% of the total sugar, and in the pure culture this loss exceeded 21% Lime in small amount was therefore capable of checking this fermentation for a time, 6 lbs. to 100 gallons being sufficient to preserve the molasses for nearly three weeks. As the lime gradually loses its alkalinity and goes into the neutral carbonate the fermentation starts afresh. As it is very undesirable to bring an alkaline molasses into a distillery wash, as small an amount as possible should be used to check the foaming ; 6 lbs. of lime to 100 gallons molasses should be used first, the lime being freshly stirred up into a milk with a few gallons of water, but only enough of the latter to admit of a thorough stirring into the molasses. If after a time foaming shows evidence of beginning again a further smaller amount of lime milk must be stirred in.

The yeast or “torula” (c) ferments very sluggishly in a dilute molasses wash, and hardly at all in cane juice. Judging from the experiments with the molasses, it is able to produce about 14% of proof spirit. It cannot invert cane sugar, and hence the feeble fermentation in cane juice, but only attacks the ready formed invert sugar in molasses.


As this yeast in pure culture gave a very marked flavour to washes in which it was fermenting, some preliminary experiments were made with it in different media, the rum distilled off and the ethers determined therein. It was grown in three washes:—

(1) Molasses and water, Brix 15, Acidity 0.10.
(2) Molasses, half dunder and water, Brix 15. Acidity 0.34.
(3) Tempered cane juice and one-sixth dunder, Brix 15, Acidity 0.20.

The yeast formed the dry wrinkled surface skin in a couple of days in all the washes, and multiplied abundantly, at the same time the fruity odour was very perceptible. Fermentation was very slow, the time required for the washes to die was:—

(1) 24 days. (2) 27 days. (3) 17 days.

The acidity of the washes was:—

The ethers found in the rum were for 100,000 alcohol by volume:—
(1) 18,000. (2) 15,000. (3) 12,700.

In spite of the very high ether content the rum had a pleasant fruity flavour with no trace of “pepperiness”. These results were obtained by a simple distillation without any treatment of lees. The ethers consisted mainly of acetic ether, so that the yeast is able to produce both alcohol and acetic acid. There was no increase of ether production during distillation as a portion of (1) was neutralized before distilling and gave the same amount of ether as the unneutralized part, namely 18,000.

[“pepperiness” would be from acrolein which comes from a bacterial infection.]

The increase of acidity during fermentation was inconsiderable, a result which taken from the preceding one makes it highly probable that ether formation does not occur by a merely chemical reaction in the wash, but takes place in intimate relation with the actively working yeast cell.

Further work is being done on this yeast with a view to its introduction into distillery practice.

[Hopefully we can find more papers by Ashby specifically about Torula yeasts.]


Two perfectly different species of acetic acid bacteria were isolated from acid skimmings and dead washes.

I. A form which appears quickly on dead wash both of low and high acidity. At first a delicate blue dry friable film which becomes white when strongly developed, but is always easily broken up. In a glass vessel the film climbs up the sides high above the surface of the liquid. It consists of short rather plump rods which stain yellow or yellowish brown with iodine, but never blue, and forms only short chains. It resembles Bacterium kutzingeanum of Hansen except in its inability to turn blue with iodine.

II. A bacterium, which forms a very tenacious cartilaginous skin in skimmings and dead washes, consisting of long narrow rods. The skin turns blue with iodine and sulphuric acid, and is in all respects similar to Bacterium xylinum of A. Brown.

In order to observe the highest concentration of alcohol which admits of a development of acetic bacteria a dead wash holding 23% of proof spirit was exposed to the air. For six weeks there was no sign of an acetic film, and there was no rise in the acidity. Between the sixth and seventh week a film began to form, and at this stage the liquor contained 14% proof spirit, 9% having evaporated away from the wash.

In another experiment a dead wash containing 24.7% of proof spirit was diluted with water in varying amounts and seeded with a pure culture of acetic bacterium I. The progress of acidification is shown in the following table, the figure representing the increase of acidity expressed as sulphuric acid %.

More alcohol was added to c, d, and e, after three weeks, and the acid rose to 6.2, 5.8, and 5.3 respectively in another week, but showed no further increase. The greatest amount of acid produced was therefore equal to about 7.5% of pure acetic acid, the largest quantity which the bacterium could endure. The organism could not grow and work in 24.7% of proof spirit, and showed only feeble activity in 16.5%, but in 12.3% it worked strongly. The evidence shows, therefore, the amount of alcohol which can undergo vigorous acidification is between 12 and 16% proof spirit, which agrees with the result of the first observation.

The theoretical maximum amount of acetic acid which could be formed from the alcohol in cultures c, d, and e, is 7.3, 5.9, and 4.9%. The actual amounts formed in 20 days were 5.7, 5.2, and 4.6 so that:

The lower the amount of alcohol in a liquor, the more completely therefore it is oxidized to acetic acid. For practical purposes the highest acidity was reached in a fortnight at about 4%. Bacterium II. proved to be unable to grow and produce acid in a dead wash containing 12% proof spirit, but gave over 3% acid in a liquor with 8% proof spirit. This bacterium also makes greater claims upon the nitrogenous foodstuff in the liquor than bacteriun I. Bacterium I. is therefore the characteristic acetic acid producer in all liquors containing 10% and more of proof spirit, such as ordinary dead washes, while bacterium II. works best in liquors like fermented skimmings and fermented rum cane juice.

[Rum canes were likely rat eaten canes and thus a vector for butyric acid bacteria. To my knowledge know one currently singles them out for rum production.]

The following table shows the amounts of total and volatile acid (mostly acetic acid) and the relative amounts of volatile acid to total acid in some distillery liquors. Of special interest are the quantities of volatile acid in such materials as acid skimmings, and flavour, because in these liquors an attempt is made to produce as much volatile acid as possible. The volatile acid shows an average percentage of the total acid of from 22 to 27, or only about one quarter of the acid present is volatile. As the fresh skimmings, which come down from the boiling house, are practically neutral, the great part of the acid produced in the cisterns is the work of bacteria. Although the skimmings readily undergo fermentation, this is not entirely due to yeast, as the liquor is heavily contaminated by bacteria which produce fixed acids such as lactic from sugar. A number of such bacteria have been separated from the skimmings. They include the well-known rice grain bacterium, which can nearly always be found in skimmings. It forms large rounded gelatinous masses when strongly developed, consisting of enormous numbers of hand shaped colonies, the rod shaped bacteria being embedded at the ends of finger like processes of the jelly. This bacterium produces lactic acid and forms its jelly at the expense of the sugar present. Another rod shaped organism often develops in fresh cane juice contaminated by dirt from the mill or by soil, at a great rate, and converts the liquor in one day into a thick viscous mass in which yeast can only work very sluggishly. Gas and lactic acid are produced, the viscous substance being formed at the expense of the sugar. The presence of such objectionable organisms accounts for the poor yield of alcohol in skimmings, and the small amounts of volatile acid. Acetic acid bacteria are wholly dependent upon oxygen for their work of converting alcohol to acetic acid, and require therefore that the liquor in which they are working should expose as great a surface as possible to the air. This is only being imperfectly attained in distilleries even in the trash cisterns. It is proposed, therefore, to experiment on a practical scale with a view to the more rapid and more abundant production of acetic acid from alcoholic liquors.

2 thoughts on “S.F. Ashby: The Study of Fermentations in the Manufacture of Jamaica Rum (1909)

  1. “[It is very clever that he weights the entire ferments to measure the escaped CO2 and thus how much sugar was fermented.]”

    It’s an interesting idea. But I calculated this and the result was not very good unfortunately.
    Perhaps you know those simple formulas to calculate the abv by the spez. gravity before and after fermentation: %abv=(originalspecificgravity-finalspecificgravity)*139
    Sometimes you can read 135 instead of 139.
    Although this formula is very simple, it is pretty accurate, even with molasses washes, which have high gravities both before and after fermentation.

    Two calculated examples:
    From 1.100 g/l fermented to 1.000 g/l results in 13.9%abv with the simple calculation and only 13%abv with the calculation by CO² loss.
    From 1.06 to 1.0 results in 8.34%abv with the simple calc and 7.86%abv with the CO² calc.
    If you want, I could post more details about the calculation or I could post a link to a html with an online calculation.

    Because of practical experience (comparision of the calc with distillation results) I am sure, the CO² calc calculates too low abv numbers, the simple calc matches reality much better.
    But why? Calculation with molar masses and so on must be true, right?
    I think one reason is the in the wash captured CO². But considering the maximum CO²-capacity of water, it makes up only around 1/4 of the difference between those two calculation methods.
    Perhaps the main reason is the by yeast consumed O².
    Both the consumed O² and the captured CO² rise the spec. gravity and therefore it looks like less CO² (and alcohol) was produced.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

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