What we have here is a remarkable document from 1913 in an era before full on chemical control where many distilleries are vacillating between budding yeasts and our hero Schizosaccharomyces Pombe. Budding yeasts, at this time, appear to be a problem as they work too fast and too hot.
It is hard to completely get a grip on what is going on. If none of these estates are making molasses, where are they getting their skimmings? Do they get it when they juice a portion of their own cane and temper?
This appears to be before the sulfuric acid era where, because they temper with lime, they have to add significant amounts of vinegar to return to proper pH. This is complicated by the fact that it is also before the era of pH! Back then, it was safe to err on the side of too much because too much acid apparently led to Pombe dominance and too much sweets also favored Pombe. Sugar in the dunder from an incomplete fermentation could also be re-fermented. So,—all roads led to Pombe! Unless of course you were overrun with budding yeasts…
What we’re missing is any correlation to sensory quality. It looks like they are pursuing efficiency by the same means we would pursue big flavors today, so it is all very confusing.
Many producers today, likely use a Pombe yeast for their pot still products, but may not be aware because their operations are so consistent they haven’t dragged out a microscope in quite a while. Pombe yeasts may be due for a revival because, besides growing consumer interest and association with terroir, they may have special properties that make them more resilient to the deteriorating quality of molasses.
REPORT ON THE INSPECTION OF CERTAIN RUM DISTILLERIES IN ST.
JAMES, WESTMORELAND AND TRELAWNY.
BY S. F. ASHBY, B.Sc., Microbiologist.
A visit was first made to the distilleries of an estate in Westmoreland, at the request of the proprietor, who complained of unsatisfactory returns from two of his distilleries.
Distillery A.: Sugar is no longer made at the factory. Molasses are brought from Estate B. Canes are crushed by a water mill, the juice tempered roughly and run into a cistern on which a thick crust of course and fine trash is maintained. The mixed juice and skimmings undergoes slight souring and very little if any yeast fermentation.
[There is immediately a lot going on here that probably does not resemble anything produced outside of Hampden. They’ve cart over molasses and they are crushing canes and tempering them with lime and heat. For some reason they run that into a cistern with cane trash that no doubt creates vinegar. I suspect that a lot of this acetic acid gets salted out. The question is: is anything redeeming produced? Also, where do their skimmings come from? Could it be their tempered juice?]
A sample of juice fresh from the tempering cistern was acid to phenolphthalein but distinctly alkaline to litmus paper. The wash is set up at 30-32 Arn. from 500 dunder, 200 juice, ten cents molasses, and 200 water and fermented in ground cisterns of 1000-1200 galls. capacity. The still is old double and (2000 galls.) direct fired with one retort, low wines going into the stills and the lees rejected.
[I think the two different measures that show it simultaneously acid and alkaline imply that it has a significant pH buffer. Arn is the Arnaboldi scale which is a measure of density but calibrated for Tropical heat instead of European room temp.]
The coil was found leaky recently and after repairs returns rose 5 per cent. The cooling is apparently deficient as rum is collected at well over 100° F.
The washes are set up at varying temperatures depending on the coolness of the available dunder or juice. They attenuate in from 3-5 days to 11-12 Arn. (losing about 19-20 points) or from 22–23 Brix to 8.5–9 Brix. This would indicate a yield of 70-80 gallons rum per 1000 wash; the actual yield from the two stills rarely exceeds 100 galls. or about 50 gallons per 1000.
[Attenuate implies the reduction of density due to fermentation. These days we would look at the difference between the starting gravity and the final gravity and be able to imply alcohol creation. Everything is a bit wacky here because significant aroma is also created that compromises alcohol. They don’t really know their fermentation efficiency until they run it through the still but then things are compounded by distillation efficiency.]
If the washes are set up with cool materials they start to work after 24 hours. In the next 24 period the temperature may rise to 98° and the density fall from 32–17 Arnaboldi; ten hours later the temperature reaches its maximum of 102°. At the end of the second 24 hours of actual working the wash is practically dead at 12 Arn. (9 Brix) with a temperature of 97°. This is the most rapid type or 72 hour wash.
[Keep in mind, the wash is still as high as 9 Brix when it dies because it has so much besides sugar creating a high density. Brix in this case doesn’t just imply sugar like it does in wine production.]
Next to this cistern was a setting which heated up to 102° in 72 hours and was practically dead in 96 hours at 96°. Beyond this cistern was one which heated to 100° in 72 hours and was hardly dead in 120 hours at 96°. All the washes threw up yellowish brown moist heads consisting mainly of fission yeast flocks mixed with granular matter. In the 4-5 day fermentations these heads persist intact and got covered by a white dry film of Mycodorma while dying or when dead. The more rapid fermentations often partially break up the heads but I did not see a case where the head had more than half fallen. Samples of plunged washes showed both the oval budding cane yeast and fission yeast, the latter probably of the intermediate bottom type as the chains typical of the top form were never seen. The proportions of the budding and fission cells varied greatly. In the 72 hour wash the budding yeast was in great excess during the whole period; in the 96 hour wash it was in pronounced excess throughout and was evidently the main agent. In the 120 hour wash (then not quite dead) the fission yeast had the monopoly. The table shows the behavior of the three washes:—
[We get our answer to the difference between the ferments of the first two sentences in Ashby’s last two sentences. The wash dominated by more fission yeasts took longer to ferments. The Mycodorma is likely the onset of a vinegar culture, but in the Arrack literature they differentiate between two different Mycodorma which forms after they transfer the ferment to clay pots to rest and one type can actually get rejected. Ashby appears aware of the difference between top fermenting and bottom fermenting fission yeast. This is before the era of chemical control so they likely have a great deal of trouble standardizing their wash.]
The tendency of the budding yeast is towards intense fermentation, high maximum temperature, low increase of acidity and rapid dying off.
The fission yeast gives a lower and cooler fermentation, dies more slowly and produces a greater increase of acidity.
Samples of dead washes to be forwarded will show which fermentation has been the more economical.
[The cooler fermentation could be due to the longer fermentation time. Economics get interesting here, is it cheaper to have more vats to house slower cooler ferments or to have less vats that have temperature control, but can work fast? Schizosaccharomyces Pombe Fissions yeasts may increasingly make sense as they improve sensory quality, likely require less cooling infrastructure, and possibly will better handle the inferior blackstrap molasses that is becoming more common.]
A characteristic feature in all washes 72 hours or more old was the practical absence of living budding yeast, the cells being shriveled and coloured by the wash. Living cells of the the fission yeast were numerous even in dead washes. While both yeasts were always found in washes the relative ascendency of one species could not be foreseen, but might be inferred from the run of the fermentation. A fermentation of 4-5 days with a maximum temperature of 100° or less is what the distiller aims at; not the extremes of 3 days or less on the one hand and 6 days or more on the other. This result can be consistently attained only by encouraging one yeast at the expense of the other. Some of the fermentations at these distilleries indicated a condition of equilibrium between the activities of the two yeasts which seemed to secure the end desired; this relation much be, however, quite temporary and accidental and beyond the power of any distiller to produce at will. The budding yeast can be encouraged by employing untempered juice (entirely or in part) or by undertempering (insufficient lime and heat); in this case untempered juice naturally started to work before it could be mixed into the wash and caused a rapid and hot fermentation with very bad returns; this was due to the one sided activity of the budding yeast off the cane. The factors favouring especially a fission yeast fermentation have not been worked out by experiment in the estate distillery. It is known to resist acid and multiply rapidly only at high temperatures. It causes sweet dunder to work and dominates washes were much ‘acid’ is made. I believe that by recording the densities, acidities and temperatures of materials and washes fortified by microscopic examinations, the conditions determining a fission yeast fermentation could be established in a couple of months and the distiller put in control of them.
[Very cool. For starters, under tempering either means insufficiently limed during pasteurization or not heated long enough or hot enough. Fresh sugarcane juice rapidly bursts into fermentation. If it starts before the molasses is added, a budding yeast would have a significant footing. When he says “it causes sweet dunder to work”, I’m pretty sure this means a fermentation was stuck, they distilled it anyhow so the dunder now has a significant amount of sugar yet to ferment and the fission yeast is able to do it. I do not know of any report on establishing fission yeast fermentations. Eventually they cultured them and sent them around to all the estates.]
Poor returns at this distillery are not due to any one dominant factor but to a sum of small losses:
1. Old defective still.
2 Bad circulation of water in cooling tank.
3. Frequent overheating in fermentation; high temperature means less alcohol and more fusel and other products.
Distillery B.: Skimmings of normal thick consistency are used rapidly in setting up washes in the ground cisterns; otherwise the settings are as at Distillery A. The dunder is drained less. Washes set at 30 Arn. attenuate to about 9. dropping some 21 points; this is about two points better than at Distillery A; the new still yields 60 gallons rum per 1000, being 20 per cent. better. The difference may be due in part to more sweets in the washes used in this distillery. The washes carry a similar head for 48 hours which breaks up and falls entirely leaving a clean fermentation. The distiller inclines to relate this to the recent use of chemicals in the factory (sulphur, bloomer, etc.) and also the fact that fermentation is now more rapid by at least 24 hours than formerly. The washes are practically dead in 96-120 hours, rarely in less time. The highest temperature observed was 98°. The early heads contained much fission yeast. Plunged washes showed a mixture of both yeasts, in one case very marked excess of budding yeast and in another wash equally striking dominance by the fission type. Fermentation though mixed was quieter, more uniform and cooler than at Distillery A indicating conditions making for better returns. The fermentation in vats under an iron roof (in another building) is frequently unsatisfactory. The washes heated up rapidly, die off and begin to work again after cooling: live dunder is a common occurrence. The main cause of the trouble is the high air temperature under the roof; good ventilation is a simple and obvious remedy.
[Live dunder here may imply a stuck fermentation? I was not able to find what bloomer is. It appears that sulfuric acid may not have been common yet. I always suspected that was why they made so much vinegar and that sulfuric acid eventually made it obsolete. That however may be wrong as demonstrated by what Hampden currently practices and the incredible results they get.]
Distillery C.: Sugar is no longer made at this factory. The canes are ground by water power, the roughly tempered juice mixed into washes with little delay in ground cisterns. The setting here is lower, 24-26 Arn, and the attenuation to 10–11 Arn. with a fall of 14-15 points. The returns from the double direct fired still are poor, usually 20 per cent. below the 4 gallons standard. Fermentation has recently become more rapid than usual, 72 hours or at most 96 hours, heating up to 102 and probably more. The heads of the early periods show abundance of both yeast but usually break up and fall later. Plunged washes in all stages showed great predominance of the budding yeast. The distillery was evidently over-run by this yeast, probably quite a recent event coincident with the newer and better type of fermentation. Enquiry showed that the juice was being under tempered and was evidently introducing living cells of the cane yeast abundantly into the washes. The remedy is stronger and more uniform tempering of the juice, a clean up of the juice and wash cisterns with hot lime and reintroduction with the well stirred 3-4 days old wash from Distillery B.
[Overrun by an infestation of budding yeast!]
Washes are set up in ground cisterns and pumped into 3000 gallon vats. The still is of 1500 gallons capacity with steam heat and two retorts. There is plenty of space for materials to cool. At present there is a shortage of dunder which is being replaced by water.
The skimmings tested showed a low acidity of 0.2 per cent. Washes are only frequently set up at 24 or lower and attenuate to 11-12 Arn. dropping only 12-13 points. This is evidently due to insufficient sweets in the wash. The returns from the still, 60 gallons from 1500 wash, are low even for the attenuation. The washes take long to die off in spite of the low percentage of sugar—from 5-6 days.
The acidity of the dunder was high in relation to its gravity 1.71 p.c. against a density of only 11.3; this density should correspond with an acidity of about 1.30. The acidity of a freshly set wash was 0.095. Fermenting washes were examined with acidities of 1.50 and 1.78 while a dead wash showed 1.82 p.c. The increases during fermentations were three or four times the normal which should not exceed 0.25 p.c. The skimmings showed only slight acidity but may have brought infection into the washes where the acid bacteria were able to multiply rapidly in the presence of a weak yeast. Water is also a likely source of infection. The yeast present was a mixture of fission and budding forms. A fission yeast sent from Hope the previous week was being worked up and looked like strengthening the fermentation.
[What Ashby is using here to troubleshoot is the Δ acidity concept I’m always promoting. He knows his baseline and what is normal and is seeing a deviation from it. Vivian Wisdom mentioned that Hampden uses surface water and they think it is key to their process but here it may be a problem because they are not in control of their fermentation kinetics.]
Conclusion.—The settings were too low and should be increased about 3 points Arn. to 26–27 to give an attenuation of 15-16 point and a still yield at least 50 per cent. higher. Skimmings boxes should be cleaned out with hot lime or ½ per cent. formalin. To see if infection by acid bacteria is conveyed by water two vats should be set up under like conditions, one with boiled water and the other with the usual supply. The distiller should try to set up at an initial temperature of about 85° F. and see that the maximum during fermentation does not exceed 100° F. He should see that the quantity and strength of the charges are uniform.
[Ashby proposes a great simple experiment. The Δ of acidity relative to the control would be a great addition. These kinds of experiments may help some productions walk a little more on the wild side like Hampden and reap positive gains.]
The fermentation is carried out in vats under a well ventilated iron roof.
In theory the materials are set up without delay so that the skimmings may not lie in the clarifying vats for days; in practice the supply of skimmings and molasses is in excess of what can be worked off with the single still. The washes are set up with 400 dunder, 400 skimmings and molasses and water to raise the density a further 10 point Arn. The large amount of skimmings in the traditional wash implies an abundant supply with a high ratio of juice to dirt. Most of the washes were dead at the time of my visit so that every stage of fermentation could not be observed.
[Wow extra skimmings! That second to last sentence is the most rum thing every said.]
The washes are set up at 28 Arn. (19-28 Brix) and attenuated to 10 (7 Br.) under favourable conditions; dead washes were examined which had fallen to 14 (10.5 Br.) only. Fermentation is at present slow the washes dying off rather sluggishly in 5-7 days.
Skimmings showed an acidity of 0.8 per cent. and wash was being set up at an acidity of 0.9-1.0 per cent. During fermentation the acidity rises to 1.35 and in one case was found at 1.56 per cent. The Westmoreland washes had shown an increase of acidity scarcely exceeding 0.2 per cent. so that the increase at this Estate had doubled the normal and in one case was three times as high. This is attributable to heavy inoculation with acid producing bacteria in the very sour skimmings and also accounts for the slow dying off of the washes and the variable attenuation. Microscopic examination of plunged washes showed a fission yeast predominant due to suppression of the budding yeast by the acid skimmings and acid washes.
The high tempering of the juice for production of white sugars encourages the souring of skimmings. An experiment has been planned and is now in progress at this estate to test the value of formalin for controlling souring in the skimmings. In the past the most usual trouble in this distillery has been due to rapid, hot fermentation caused by the budding yeast.
[Formalin is a formaldehyde based cleaning agent that was invented around the 1890’s.]
A fortnight earlier samples of ‘live’ and fermented dunder were received and examined.
At the end of March the dunder worked off about 4° Brix (6 Arn.) and when dead was free from fermentable sugar but contained over 5 per cent. proof spirits. On the 16th April it was of lower density and had worked off only 2° Brix indicating an reduction of sugar to one half. Unfortunately on the 16th the distillery showed dead washes only and of considerable age, 168-180 hours.
[I think the complaint here is that it took 7 days to die which generally isn’t reasonable.]
The washes are set up at about 27 Arn. from 300 dunder, 400 skimmings, molasses and water to 1000. At first attenuation was about 7 Arn. and 20 points fall. After a few weeks fermentation was very rapid for 24-36 hours but soon stopped after attenuating about 15 points and ‘live’ dunder began to show. The course of fermentation now appears to be frequently rapid and hot for 48 hours and later a gradual dying off in 6-7 days with great increase of acidity from about 0.8 to 1.88 and 2.1 per cent. The plunged dead washes showed dead cells of the budding yeast in the majority and many living cells of the fission type. The inference from this finding is that the washes are often set up at a high temperature with a vigorous infection of budding yeast; heat up to many degrees above 100 F. the budding yeast dies out and fermentation ceases or after cooling starts again from a development of the more resistant fission yeast which had survived from the early head. In a wash which is dying off with considerable sugar unfermented acid bacteria are very active and account for the abnormal acidities of the dead washes.
[This is quite fascinating and probably does not correspond to anyone’s modern day experiences with a stuck fermentation. The difference here is they were dealing with a spontaneous fermentation that likely had a lot of yeast diversity. Fission yeasts also likely populate their vats and the surrounding building. Modern day ferments are so clean there is no alternative culture to rise up and unstick them.]
The root of the difficulty is to be found in the traditional composition of the wash with the high proportion of 4/10 skimmings. The space allotted to the skimmings in the cooling boxes is insufficient so that even after 12 hours the temperature stands at 120° and over; if the other materials stand at 78° the mixture would still show 95° in the freshly set wash causing a rapid, hot and incomplete fermentation. Satisfactory returns will be uncertain till the skimmings can be cooled to 96° or under and the wash set about 86°, if the proportion of skimmings much remain at 4/10.
S.F. Ashby, Microbiologist.
21 April 1913.