Turning the Sky Blue and Turning on Contrast Detection in Olfaction with Language

The sky isn’t blue until it is.  If you didn’t already catch this show on WNYC’s Radio Lab I’d start there.

The hosts along side Guy Deutscher, author of Through the Language Glass,  look at Homer’s the Iliad and Odyssey and investigate why the color blue is never mentioned. They even take it back to 19th century British prime minister & Homer-ophile William Gladstone who made the first inquiry.  Gladstone was intrigued by Homer’s use of strange colors for simple objects which he thought to be at odds with Homer’s acutely perceptive style. There was the wine dark sea, wine colored oxen, violet colored sheep in the cyclops caves, iron was also violet.  Honey and faces pale with fear were both green.  Homer’s visual world was very different from Gladstone’s.

Later in the 19th century antiquarian text specialist Lazerus Geiger took a stab at the mystery and came to the conclusion from witnessing similar color phenomenons in the ancient texts of nearly all other languages that even though blue is a primary color, languages do not tend to create a word for a color until they can produce the color.  In the development of language blue comes last because blue is so rare in nature and blue dyes are hard to create.  The Egyptians with their indigo were an exception.

The show then turns to Jules Davidoff, a professor of neuropsychology.  He shows twelve colored squares to a Namibian tribe (his stand in for Homer)  with no word for blue and asks “which one is different?”  Eleven of the tiles were green and only one was distinctly blue.  They had remarkable trouble figuring out which one is different. I see this as validating Marshal McLuhan’s idea in Understanding Media that literacy (of which there are many types and many subsections) facilitates fragmentation and allows us to break things down and deconstruct them. Fragmentation also leads to acting without reacting (detachement) which I’ve discussed a little bit in relation to flavor literacy. Being more literate in flavor may help us enjoy eating things that are acquired tastes and typically healthier as well as resist cravings.

Naturally this Radio Lab program turns to explaining the color of the sky which to Homer was not blue.  The program turns back to Guy Deutscher who experimented on his young daughter who was just learning to speak.  Deutscher teaches her all the colors of everything around her but never the sky. When asked what color it is in the beginning she cannot say. When eventually she can, she says it is white.  Finally she says it is blue but then flip flops for a while between white. After a while she settles on blue but it takes a while and it was never innate and obvious.

In my adventures with sensory science I’ve noticed that all the senses more or less follow the same rules. If vision and color are subject to this phenomenon where language is almost requisite for parsing sensations so too is olfaction and probably in exactly the same way.  Gordon M. Sheperd validates many of these parallels between the senses in Neurogastronomy. What is funny about olfaction is that for many people they wake up middle aged, they finally develop an interest in wine (or actually smelling whatever else they drink) but find their nose is in the state of that child that cannot determine if the sky is blue or white or just a colorless void.  They might even be like the Namibian tribe that cannot detect contrast in the green and blue tiles.  They have no words for what they smell so they say “all wine just tastes the same to me” (but maybe it does?).  The path and development of their other senses has been completely taken for granted.

In Homer’s world, fragmentation of color was poor relative to today, but some people believe their olfactory sense was more acute.  They hadn’t yet sanitized the world of its natural odors and the use of fragrance was rampant.  To get along in that intense olfactory world of good and bad aromas it would take detachment only possible by a well cultivated language.  A thousand years after Homer, in the Aeneid, the willow was acrid (a cross-modal euphonic metaphor!), probably referring to the subtle aroma of willow honey while today we only apply the acrid (Ngram) descriptor to extremes like bleach and ammonia.  Going back just a few hundred years, olfaction used to be refined enough that the aroma of cloves and cumin were bitter while now we can only apply such labels to extremes like wormwood and quinine.  Our currently atrophied olfactory world, which is widely constrained by our lack of language and categories to organize aromas likely parallels the constraints on Homer’s ability to parse and recognize color.  McLuhan said adding new media (the extensions of man!) would change our sense ratios.

In the last post I hypothesize that many aromas we perceive are in part illusions and I still stand by that. Completing gestalts with recollection is definitely a layer of aroma perception.  Another significant layer is the importance of language on our ability to parse aromas even at the simplest levels.

Many new wine drinkers have trouble recognizing oak. It takes a highly abstracted (over oaked) example to teach the typical new wine drinker what oak is so that they have a category for it.  Speaking of categories, when use of the wormwood hand sanitizer changed the threshold of perception of aromas in the wine it may parallel the twelve colored tiles experiment.  Every time I applied the hand sanitizer I thought “wormwood!” and had a well practiced category that was easily retrievable.  I may have noticed the aroma disproportionately because the category was at the fingertips of my recollection.  Funny enough, I did not smell wormwood, I smelt something more like yarrow flowers and recollection did not complete the aroma like I said can happen in the last post.  I do have unique experience with yarrow as well and other aromas in the olfactory-bitter category after experimenting widely in absinthe making (the recipe is out of date, I settled on all yerba mate!).  I wonder how other people would experience these things that do not have my unique stance.

Wine scholars put big emphasis on turning wine into words but no one really pin points why.  When you start to put big emphasis on language people sometimes shoot you down and say its not necessary, but do they smell the blue tile!?

And what categories can help me detect contrast and smell as many different tiles as possible? I champion categorizing olfaction in terms of gustation with the olfactory-sweet, olfactory-umami, olfactory-bitter, etc.  Those five categories can replace an entire unabridged dictionary of object comparisons (cherry, cassis, goose berry, eucalyptus).

Barb Stuckey emphasized language in her excellent primer on new flavor theory called Taste What You’re Missing.  Really nailing the edutainment I can’t seem to master, Stuckey explains the individual facets of the multisensory perception of flavor to help develop categories for beginning tasters.  Her idea is awareness of categories will help you “taste what you’re missing” and I whole heartedly agree.

Another method of developing categories that facilitate contrast detection is Anne Noble’s Wine Aroma Wheel.  Taken directly from the why use it section of her website: “Novice tasters often complain that they ‘cannot smell anything’ or can’t think of a way to describe the aroma of wine. They don’t have the words! Fortunately, it is very easy to train our noses and brains to associate descriptive terms with specific aroma notes in wine.”  If an organized wheel of categories is placed in front of someone it may help them smell the color blue.

All this contrast enhancement by language supported category formation is making me contemplate the auditory sense.  Is there more to be heard? It feels like there is no way I’m not hearing the color blue.  A great edutaining book on the neuroscience of music I read long ago is This is Your Brain on Music.  The author is a sound engineer turned neuroscientist and at one point I remember him telling the tale of how he educated and developed his ear to hear the unique things few other people can hear. Sound engineers have sound boards with tons of controls which are essentially categories. The more categories they are aware of, and can abstract (through turning a dial) to reinforce the properties of, the more contrast they can likely detect.  We function fairly well in this world with an auditory sense that in most cases gets no special training, but there is still always an amazing potential for enhancement.

Next time you encounter a glass of a wine ponder what beautiful and extraordinary aromas might lurk just beyond the reach of your words.

Olfactory Phantoms and Illustrations of the Dynamics of Perception

The other day I got an interesting incoming link. Something I wrote was selected to be part of a hypothetical college curriculum. It is on a curriculum sharing website but I don’t think anyone is actually using it yet. Students in a sensory science class get to learn about illusion in the various senses and because there is not a lot written about olfactory illusion I was selected probably by default (default! woohoo!). I covered another phenomenon years ago I called the maraschino blackberry illusion where texture (haptic heft) dramatically changed the threshold of perception of an aroma. I think some bar programs are finally getting around to using it. I did it again in a beautiful context with fernet (and kirschwasser) aromatized cherries and I’m doing it yet again using the reflux de-aeration technique (to preserve the lemon juice) to perfectly place a sidecar inside a golden raspberry. The aroma and color were leached out of the golden raspberry so I’m basically just using its perfect body of cells as a vehicle to harness the influence of haptic heft on aroma.

We are slowly getting to the punchline but we need to cover a few more strange phenomenons.

Whoever selected the paper for the curriculum was interested in the “illusion” where wormwood hand sanitizers I made were used to train someone (myself) on an aroma. I found with the unique training I experienced similar aromas in certain wines (I found two wines: Cos Cerasuolo and an Anjou blanc) that were bizarrely amplified; like 10X amplification. Sometimes the same phenomenon exists with the aroma of dead mice and those in the pest control business that inadvertently end up with unique olfactory training can notice the aroma in a room when no one else can.

Wine pairings are part illusory and reading Gordon M. Sheperd’s Neurogastronomy confirmed some of my suspicions about wine and food interaction that I covered in an incomplete post called: contrast enhancement (in space and time) for food & wine interaction. The idea didn’t exactly catch on but I am right so I guess I’ll just give people a few more years to adopt it (mach bands! & nutritional preference comparisons!).

Now for the punchline: I’m setting out to explore how many of the aromas we “perceive” under certain circumstances are to varying degrees actually phantoms.  They are induced by incoming sensory stimulus (or by words which are symbols) but they are ultimately just recollections.  This is different than just making loose comparisons of incoming stimlui to known things, in what I’m describing you eventually generate whatever percentage of the known thing that doesn’t exist in the incoming stimuli.

Maybe we can start with the simplest olfactory phantom I’ve been able to generate and then build some background around it. To make the champagne bottle manifold I started a small plastic foundry. To develop the skills I needed to make the manifold part I started making reproductions of 19th century door knobs; giant lion heads and rococo stuff.  Some I cast in a translucent red plastic to generate some Joris Karl Huysmans style artifice and decadence. Anyhow, I had to drill this red plastic. Well, every time I did, I started to smell cherries. After a while I knew I was going to and I still did. The black plastic smelled of licorice.

Compounds called phthalates in the plastic have an aroma that notoriously converges with gustatory-sweetness. This form of sweetness coupled with the color is enough to trigger a phantom aroma and illustrate some of the dynamics of perception.

Perception is a tricky thing because all sorts of facets seamlessly join together. In the past I’ve talked obsessively about the sensory and symbolic world being glued together by the theory of cognitive dissonance and becoming the mechanism by which we acquire acquired tastes.  Perception involves the meeting of sensory inputs with recollections.  Improperly using recollections to complete incoming experience is the basis of many optical illusions.  So perception is going to be (by varying degrees) divided by fresh incoming sensory experiences and a sort of filling in the blanks via recollection. Most of the time recollection will correctly fill in the missing pieces.  Incentives exist to use completion to save resources.  Apparently it is more efficient than processing everything from scratch. Olfaction being the sense most closely tied to memory and thought to be particularly resource intensive might be subject to more completion by recollection than any other sense, though I’m only speculating (this whole post is a giant speculation).

As the distribution of perception slides around we may be subject to more or less illusory completion.  I’ve seen sensory science researchers hint at this distribution by outlining different perception strategies such as an active or a passive strategy.  An active strategy may tip the scales to processing new sensory data while a passive strategy may tip the scales towards completion by recollection.

The distribution could then be subject to other influences such as reward mechanisms.  Certain degrees of salt in our food, such as on a tomato, lowers the threshold of perception of an aroma.  The co-experience of the aroma and and salt may etch themselves in our mind so we can use those aromas to predict the presence of salt in the future, salt being something we need.  Sugar under certain circumstances has a similar effect, often referred to as “flavor enhancement”, that can lower the threshold of perception of an aroma.  Our rewards systems might jostle the distribution of perception any which way.  The sweetness in the plastic did not make me pay attention to an actual incoming aroma sensation but rather created one based on the influence of the color.  In my red plastic example the cherry aroma was more or less 100% generated by recollection but in many other cases aroma fragments that are being sense are being added to.

Many spirits researchers talk about pattern recognition or gestalts being important to distillates.  The famous spirits consultant Dr. Jim Swan has mentioned that when coming up with a new scotch whiskey blend it needs to have enough elements to be recognizable as a scotch whiskey (a gestalt) but then enough to set it apart.  The other big time spirits guru John R. Piggot in his amazing paper “Origins of Flavour in Whiskies and a Revised Flavour Wheel: a Review” starts by saying “Improved congener analyses have not yielded greater understanding of whisky flavour: a dynamic interaction between individuals and flavour components.”  The dynamic interaction mentioned by Piggot is another way of acknowledging the distribution of perception.  He then goes on to mention holistic patterns and gestalts as part of perceiving flavor. In memory, the cherry is a gestalt of sweetness and color, and aroma.  When only two of the three are present recollection may complete the third.

Piggot’s paper gets very dense in the chemistry but his introduction is very accessible and pretty amazing. He demonstrates an astounding understanding of neuroscience to go along with his second to none understanding of every reaction that generates every congener in every step of the whisky making process.  Piggot notes the interaction of the sensory and symbolic world but doesn’t exactly use the words I’ve adopted so it is probably best to quote him. According to Piggot, “In consumers, causality interactions (slaving effects) exist between perceptual and sensation levels, dictated by cues (Fig. 1): the human mind influences the brain.”  There is something vague about saying “the human mind influences the brain” but I’d like to think it parallels my language.  I do not agree with the way he uses the term “perception” because I think you can stumble into “which taste do you mean” territory.  When he says “perception” I think it could be better named “recollection” and perception instead is the sliding summation of incoming sensation and recollection. Sensory scientists have needed to get away from the term “taste” for flavor perception and maybe the same needs to be done to deconstruct perception and consciousness.

Robert Léauté in his 1989 James Guymon lecture very vaguely mentions fatty acid esters being “fixitives” for other aroma compounds.  His idea of a fixative might relate to the building of overlapping incomplete gestalts that recollection might complete in beautiful ways but we’ll touch upon that in a bit.

I should probably mention that infamous incident were a group of wine experts were served room temp white wine dyed red and nearly all were fooled into thinking they were drinking a proper red wine.  They described the aromas of the wine using object comparisons attributed to red wines and not whites. Well they were drinking a red-wine and if a degree of the distinct things we think we smell are phantoms, they were perceiving everything exactly as they should.  Their advanced library of recollections may have even made them more vulnerable than an amateur taster.

Another known phenomenon in wine trickery is simply tasting a wine after being baited with an aroma suggestion.  The suggested aromas can appear vividly.  I do not think this is widely studied or acknowledged because no one wants to be the victim of it.  We feel as though it shouldn’t happen so we never submit to curiosity and explore experiencing it.  A aroma suggestion isn’t a sensory input like gustatory-acidity or tannin. It starts that way but becomes a symbol or stand in for a value that triggers recollection. Some symbols thrown around in wine-speak, like goose berry, are not good bait because for many people the word is not a stand in for any recollection. To this day I’ve still never eaten a goose berry.

Robert Parker, who is famous for his wild tasting notes, might perceive the world with some unique distribution abnormally skewed towards recollection almost like a form of mild autism.  The gestalts he encounters make him hallucinate wildly. Every wine Parker encounters, which remains incomplete to the rest of us, he can complete apparently with great pleasure.  This might be a bit of a stretch because the language used by wine critics has no real responsibility to describing the wine. They just have to use harmonic language and typically more extraordinary language for more extraordinary and rare sensory values.  If we accept the fact that wine makes Parker hallucinate, to be honest I’d like to join him and learn his technique because illusion or not, it seems like fun.

To get back to spirits, most whiskey’s are colored with caramel to converge with their aromas.  Accepting that it is important to the quality of our hallucinations we should probably thank and recognize producers that do a good job of it.  An award could be given at the San Francisco spirits festival like “best dye job”.  I used to scoff at coloring but now I’m warming up to it.

A realm I had fun exploring the influence of color was with amaretto.  Everyone knows the darkly colored Disaronno brand and other darkly colored generics but I was really taken by a Portuguese almond (benzaldehyde) liqueur where the color was interpreted as a much lighter raw almond shade.  I was so taken by it I re-distilled my own version and left it uncolored so the aroma would diverge from the crystal clarity.  The results were captivating.  There was a dramatic divergence from expectation.  The crystal clarity and the light it captured made the aroma glow but it was just the Portuguese version re-distilled, re-cut, and re-sugared.  Benzaldehyde based liqueurs are so easy to distill the real artistry probably comes in when coloring them.

In cocktails we may sometimes experience perception dominated by recollection.  Many cocktails that are simultaneously tart and bitter come across as distinctly grapefruity though there is no grapefruit in them.  True these drinks, which often feature citrus juice, have a few aroma compounds in common with grapefruit, but the loose gestalt is enough to trigger a very distinct association even when the chemicals and all there various ratios do not remotely line up.

Now to go back to Robert Léauté, the idea of partially illusory aromas might explain the importance of very generic aroma compounds like ethyl acetate and acetaldehyde in spirits.  In large quantities these compounds are considered flaws, but sub threshold Léauté describes them as his “fixatives”.  Because these compounds are so abundant in fruits they may be an integral part of gestalts. Ethyl acetate could function like the sweet smelling phthalates found in the plastic and when coupled with one other sensation like a color we would have enough to hallucinate marasca cherries like Robert Parker.

Léauté, who is a master cognac distiller, explains that ethyl acetate and acetaldehyde, should be kept under the threshold of perception but distillates should be cut to get as close to that line as possible.  It might make sense in this case to rename the “threshold of perception” to “threshold of attention”.  We seem to be able to perceive astoundingly small quantities of things.  Our own nose is more sensitive than any analytic tool we’ve been able to build.  According to Gordon M. Sheperd our nose can differentiate aroma compounds by one carbon atom.  So maybe something is not detectable to our attentional spotlight but it is somehow still detectable enough that it can interact with other components synergistically and influence pattern recognition.  D.W. Clutton’s paper from 1978, The Flavour Constituents of Gin finds all sorts of compounds in gin that are sub threshold yet they are somehow very important to defining the character of the product.  Thresholds of perception may work very differently than is commonly thought.

Recollection may have some strange bearing on wine pairings.  In some pairings when food and wine “match” an aroma from the food can be reflected back into focus.  I had thought previously that a change in contrast detection was experienced and likened the phenomenon to the black art theater of the magician Omar Pasha.  An alternate explanation could possibly be that with the next experience (which is the wine after the food) a similar pattern confuses the mind and triggers recollection of the food.  Many of these types of pairings happen when the perceived acidity of the food and the wine match.

I don’t want to leave people thinking that everything we smell is an illusion.  We obviously need intense libraries of recollection to generate phantoms from.  It is probably safe to say that we mostly always are actively smelling when we think we are, but where is the dividing line and small details are we adding in to what were are really smelling? And if alcoholic beverages like wine and spirits remind us of so many things as seen in so many tasting notes, could they be the hub of olfactory illusion?

Do some people not taste wine well because they have no language to fragment and parse the experience or because they have no library of recollections to generate the illusions?

Juniper Report: A Blog-Quality Survey of Academic Gin Literature

Feel free to skip to the very end if you get bored.

the first paper I found on Juniper was Controlling Gin Flavor by Herman Willkie and the team at Hiram Walker in 1937. Hiram Walker, I just noticed, was located in Peoria Illinois which is the same city as H. Shufeldt & Co. whom was an American producer of maraschino liqueur that was unearthed in the last post covering benzaldehyde in maraschino liqueurs from 1912.

Willkie explains new methods of standardizing the botanical charge to account for the varying essential oil contents of the botanicals.  He covers numerous testing procedures.  Most interestingly Willkie gives an introduction to the terroir of juniper and how it varies significantly by latitude.  He even gives some opinions.  Juniper expressions that Willkie did not enjoy I really enjoy.  This also led me to wonder if when recreating historic forms of gin such as Old Tom or Genever that we must consider their juniper sourcing.

In the limoncello article roundup where it was revealed that limoncello goes through no terpene separation and that terpenes and their unique distribution are the terroir of the product; what its all about. Terpenes got me thinking about juniper again and I wanted to see if there were any other great papers out there that would help connoisseurs understand gin and maybe even help new producers make the product of their dreams.

Numerous papers exist. Here is a bullet point run down of whats going on:

Characterization of Volatiles in Different Dry Gins from the J. Agric. Food Chem 2005, 53, 10154-10160 written by a Spanish team.

**they look at london dry gins and other “gins with geographic denominations”

**”When the production process takes place within a specific geographical area and fulfills certain  requirements concerning elaboration, composition, and quality, the gins can receive the denomination of geographical indication, as in the case of Plymouth gin (U.K.) and Mahon gin (spain).” I cannot not wait for “west coast” styles gins that match the ethic of west coast style I.P.A.s.

**G1-G4 are the top london dry brands while G5 is plymouth and G6 is Mahon. maybe we can guess the london dry brands by looking at some of their data. I bet they have tanqueray, beafeater, bombay or saphire? and I’m out of touch on what the fourth would likely be.

**they basically spend all there time proving that a method can detect compounds and how you need overlapping methods to get accuracy.

**”The highest contents of juniper characteristic monoterpenes were found in samples with geographic denomination G6, whereas sample with the geographic denomination G5 showed the highest concentrations of limonene and γ-terpinene. This is probably due to the use of citric species during gin aromatization.” so basically the spanish gin is my style and plymouth is boring and too citrusy.

**”In all of the samples, except G6, linalool was the most abundant among these compounds. Linalool is present in traces in juniper berries, whereas it is the major compound in the essential oil of coriander, in which it may represent >60%. Coriander seeds are well-known ingredients in gin aromatization, and linalool concentration may indicate the proportion of coriander employed for this operation.” … “The highest concentration of this compound was present in the London Dry Gin samples of the G4 group.” maybe that hint will elude to which is G4?

** there is a chart that quantifies the amounts of 66 different components in the six samples but I couldn’t get any great sense of the flavor from the numbers.

Comparison of a Novel Distillation Method versus a Traditional Distillation Method in a Model Gin System Using Liquid/Liquid Extraction from the J. Agric. Food Chem. 2008, 56, 9030-9036 by a team from Bacardi-Martini product development in collaboration with Clemson University in South Carolina

[Edited to add: It was explained to me by a well published, well patented mentor of mine that this study is likely a red herring. I wondered what incentive Bacardi had to publish it and it turns out it is likely just to support their patent application which they mention at the end.  The results are manipulated to build false novelty to justify a patent.  Things are presented in a way to throw people off. The model gins are not fractioned and the monoterpene level of the novel gin is never compared to other conventional commercial brands so you never get a true enough sense of the novelty on a chemical level.  They also never analyze what is left in their big ice chunk that represented the non-volatile fraction. The researchers are not naive and likely know all the finer points they should be pursuing.  The research went on to become the Oxley gin brand and I’m not sure if a patent was granted but I hope not. The process is not novel and spirits have been vacuum distilled for decades. Hopefully they cannot do anything to prevent smaller distilleries from experimenting with vacuum distillation. I should probably look into this more.]

**the novel distillation method was high vacuum distillation

**they studies a four botanical model gin composed of juniper, coriander, angelica, and lemon peel.

**”This research demonstrates the benefit of distilling botanical extract (particularly for the manufacture of gin) at temperatures below 0°C, which retains natural flavor of the botanicals better than under the traditional conditions, thereby producing a superior gin.”  this stupid statement sets the tone for the paper. they use two horrible oversimplfications: “better” and “superior”.

**in regards to traditional gin, “Some makers will apply a slight vacuum to get the distillation to take place at or around 60°C.” as opposed to 70-80°C.

**”The technical literature from the past 15-20 years has shown that vacuum distillation effectively circumvents high temperatures and reduces monoterpene formation in the final product” whatever that means.

**”Using a digital balance, 37.2 g of juniper berries, 52.8 g of coriander seeds, 18.0 g of angelica root, 10.8 g of dry lemon peel, 4270 g of 95% ABV GNS, and 687 g of deionized water were weighed and placed into a stainless steel pot and allowed to steep at room temperature (23°C) for 24 h.”

**for the atmospheric distillation: “and distilled until the thermometer reached 95°C. This insured that most of the alcohol had distilled at that temperature point.”

**for the vacuum distillation: “The distillation continued until the contents of the kettle froze, signifying that the majority of ethanol had been distilled.” cool!

**they acknowledge the significance of terroir on juniper berries when making comparison to other studies.

**vacuum distillation had a lower recovery rate than atmospheric distillation due to quite a few variables.

**they acknowledge “‘blow-by’, or vaporized alcohol that failed to be recondensed by the coldfinger and was passed out of the system through the vacuum pump.”

**”Coriander seeds (which are actually fruits) are normally the largest ingredient by weight in most gins.” is this really correct? their model gin had more coriander than juniper.

**linalool concentration decreased in the vacuum distilled gin post distillation relative to their un-distilled infusion of botanicals.

**vacuum distilled gin is a product with “less nasal pungency, more floral, less spicy aroma” So basically it is less of an acquired taste. Monoterpenes probably represent the most angular of junipers aroma components. To me, gin is supposed to be an acquired taste.

**they mention a patent application and that “proof of concept testing on a commercial scale is ongoing.” Bacardi who co-authored this paper owns Oxley Gin so this is apparently their feasibility study and Oxley is the product that emerged from the work. Why they would need to make this research public or team up with a university is beyond me. You would think they would have the capability and resources to do all of this privately.

Clutton, D.W. The Flavour Constituents of Gin, journal of chromatography, 167 (1978) 409-419

**”Dutch gin resembles the original gin produced in the 17th century, in that its flavour, reminiscent of almonds, is derived from the botanical ingredients and the source of the spirit used to make it.” I’m not sure how he comes up with “almonds” but Clutton is a big name in distillation research.

**”‘London’ relates to the method of production and not to the geographical location of the distillery; ‘Dry’ means that the flavour level is low.”

**their five samples varied and they only seemed to like one. others seemed to have flaws. who knows if they used famous brands.

**”UV analysis provides information on botanical flavour levels since juniper oil absorbs between 200 and 240 nm and coriander oil between 200 and 225 nm.” I’ve never seen UV analysis used and I’m wondering if there is anything low enough involvement about it that small distilleries could use it. An ultraviolet-visible spectrophotometer is only $1500 on ebay but knows how much time the testing takes to administer. They provide a chart which is pretty cool:

sample                   dilution       ppm juniper      ppm coriander     ppm cassia
1 london dry gin         1:1              45.8                   18.1                       —
2 london dry gin         1:1              27.7                     6.7                      0.5
3 london dry gin         1:1              37.5                   19.6                       —
4 plymouth gin           1:1              37.5                   24.6                       —
5 geneva gin              5:1              87.6                   26.2                       —

“The results show that the concentration of juniper oil in commercial gin samples varies from 25 to 50 ppm and for coriander oil from 5 to 25 ppm. The results obtained for Geneva gin (sample 5) must be regarded with caution since this product contains other species such as aldehydes, esters, etc. formed during fermentation.” …. “UV cannot differentiate between gin containing orange oil and those not containing this botanical, since limonene, the principal component, absorbs at 200 nm coincident with the absorption of coriander oil.”

“UV analysis only provides an indication of ‘total flavour level’ as ‘Juniper’ or ‘Coriander’. This is because the oils used for standardisation are steam distilled products, whereas gin is distilled in ethanol and part of the botanical flavour components are rejected as
‘feints’.”

So the technique has limitations. Here he acknowledges that steam distilled oils are un-fractioned unlike essential oils distilled with ethanol. In my distillation text I raised the same point in differentiating between distilled gins and compounded gins. One is fractioned and one is not. But from the above vacuum distillation study we also know that they may also differ significantly in degradation products (mono-terpenes) from time under heat. A steam distilled oil likely sees more time under head and higher heat than co-distillation with ethanol.

**There was some interesting stuff about odor thresholds and what constituents likely define the aromas but I’ll spare you. What is interesting is a chart they give that shows all the compounds then acknowledges what is at or above the threshold limit. This sort of proposes an active ingredient. I suspect though there are some flaws to the idea and there are all sort of synergies and interactions that change the threshold of perception. If compounds weren’t perceivable below the threshold they wouldn’t matter and it would be easy to compound things from only the “active ingredients” and I think that idea floated around in the artificial flavor business decades ago but the situation turned out to be more complex than that.

**A lot of their effort isn’t so much getting somewhere with the gin but getting somewhere with their analysis techniques. This was probably a pioneering a paper that opened up new analysis techniques to studying spirits.

**They spent a little time looking at how compounds like mono-terpenes accumulate in the heads and gradually decline. These compounds were the flaws of the vacuum distilled gin study. That study might have been biased if they never fractioned them in their model gin. So yes they are created but maybe you can remove them… The fractioning and selective
separation of this class of aroma component might be why big London Dry distillers expend so much effort tuning their stills.

**”GLC analysis shows that the early fraction of gin distillates are principally composed of juniper components. Coriander components distill over after a strength of approximately 75% ethanol is reached. Indeed a large quantity of flavouring components are run to waste in many typical distillations.” what I think he means is that fractioning is a big part of the gin distillation process but he is cheap and throwing away things annoys the engineer in him.

I wonder if before this era of analysis that even the big London Dry’s had a fair degree of inconsistency. This kind of research might have created an era of precision sculpted products that could meet global demand.

Sensory Characterization of Dry Gins with Different Volatile Profiles by a Spanish Team from Barcelona

**This study looks at 6 gins, four of which are London Dry and two geographic gins which are Plymouth and Mahon from Spain.

**They try and develop a sensory vocabulary and see if it can successfully be matched to chemical composition.

**The descriptors they narrow gin down to are juniper, citric, aniseed, spice, and licorice. These are all object comparisons and it would be interesting to see if the study could be done again with cross modal descriptors like looking at all the volatile components in terms of the gustatory division they converge with.

**They reference Clutton’s definition of gin from 1978.

**They mention other geographic indications I’ve never heard of: Ostfriesischer Korngenever, Genievre Flandres Artois, Hasseltse Jenever, Balegemse Jenever, Peket de Wallonie, Steinhager, plus the usual Plymouth Gin, and Gin de Mahon.

**They identify the sales of the brands which would help anyone positively identify them if they really wanted to.

**They make reference standard solutions of very specific volatile components that I would love to check out some time.

**They make other reference standards of things like paprika powder by simple infusion. Their choices were all things alleged to be in gin formulas.

**The initial list of descriptive terms identified by the panel in 3 gins during the session of vocabulary development. In bolt letters are the 10 preliminary selected attributes.

angelica root, aniseed, aniseed/fennel, aromatic plant, cane, cardamom, chili, citric peel, clean, coriander, cumin, detergent, eucalyptus, fennel, fertile, floral, fresh, fresh spice, freshener, fruity, herb louisa, juniper, lemon, lemon balm, lime, licorice, licorice root, orange, oregano, painting, polish, resin, rose grapefruit, rosemary, seaweed, seed, soil, solvent, spice, tangerine, thyme, varnish, vegetable spice, wood.

I see angelica root as converging with the same gustatory division as juniper which is olfactory-acid hence all those fresh and clean descriptors. the two combine with the intention of forming an overtone that is extraordinary as opposed to ordinary. citrus peel, fruity, aniseed/fennel, and licorice all contribute olfactory sweetness, but not with overtones so much as intervals when the citrus peels meet the anise. if rendered on an imaginary spatial scale, citrus and anise would be at opposite ends which is why experiencing the two together gives a sensation of depth and enlarged space. coriander is unique because it has qualities in common with the olfactory-sweet and the olfactory-acid so it influences overtones in both directions. spice can often be olfactory-bitter or olfactory-piquant. eucalyptus and licorice can also be olfactory-camphorous which is similar to piquancy and probably operates via the trigeminal nerve.

**descriptors were reduced to the five by a combination of examining frequency quotation and relative intensity. relative importance was then considered and then redundancy.

**”The results of the triangle test (Table 2) showed that the panelists could easily distinguish the gins with geographic indications (G5 and G6) and the London Dry Gin G1 (P<0.05). Gins G3 and G4, however, could not always be differentiated from the other samples.”

**they start to use some awesome spider graphs that I think should really be applied to cocktails. they give an intuitive look at how the gins different with some limitations. I should probably copy in the charts. the basically make it look like juniper does not dominate. the panel ends up searching for the other components and somehow makes them have out-sized intensity.

**”The geographic indication G5 [Plymouth] showed the highest intensity of citric attribute (Figure 3 and 4, Table 4). The intensity of the citric note in this brand was justified by its high levels of limonene and g-terpinene, together with linalool, which could contribute to the citric note (Table 5). On the other hand, G5 presented concentrations of juniper monoterpenes similar to those of G6 [Mahon] but showed the lowest intensity of juniper note by orthonasal perception. The high intensity of citric note in this gin brand could have masked the other attributes. The flavor perception did not depend only on 1 or 2 compounds on a complete volatile composition. Some interactions could take place between some compounds and also some compound could inhibit another, providing a different note.” I wish I remembered these gins better but I gave up on them years ago and now only drink and sell the local stuff.

**”The London Dry Gin G1 obtained the lowest punctuation for the juniper attribute by retronasal perception and was characterized by spices and aniseed notes. This gin contained the lowest levels of juniper characteristics compounds, while it presented significantly higher amounts of d-3-carene (P < 0.05). Also, the multiple regression analysis was performed. The spice note could be explained by a model with sabinene, d-3-carene, and p-cymene + ….; these compounds presented pepper and resinous notes, respectively (Table 5).

A paper I read won’t won’t quote is:

Effect of Latitude and Altitude on the Terpenoid and Soluble Phenolic Composition of Juniper (Juniperus communis) Needles and Evaluation of Their Antibacterial Activity in the Boreal Zone, J. Agric. Food Chem. 2009, 57, 9575-9584 by a Finnish team

The paper looks at the diversity of composition of juniper needles which are harvested for their essential oil which has a gin-like fragrance but also antibacterial properties.  The subject is not juniper berries but I suspect their properties parallel the needles.

As is mentioned by Willkie in the first paper, the character of the berries differs significantly with latitude. This paper adds altitude to the equation but find latitude is more significant.  Other papers I’ve seen add coastal proximity to the equation but I haven’t read them.  The most interesting gin I’ve ever made was with juniper from Cape Cod.

This paper and papers like it can add to gin connoisseurship.  We can now ask where the juniper has been sourced from and we can now ask to have gins made to reflect certain terroirs such as the northern most most juniper example.  Beer brewers already give us this with hops.  We also now know from the vacuum distilled gin paper that juniper Distillation involves the creation of many new aroma compounds through the degradation of precursor compounds.  Attention to detail with the cuts can refine the juniper expression.  Their paper might have been biased because they didn’t make heads cuts on their model gin distilled at atmospheric pressure.

A vacuum distilled gin isn’t without appeal, but I don’t think it should be touted as superior unless they are trying to give us a glimpse of a particular named juniper source.  Really interesting is the acknowledgement of partial vacuum gins which perhaps keep the temperature under a threshold. no brand was acknowledge. If I guessed I’d say Hendricks because they are the newest to the market and possible use new equipment that could handle such technology.  Bombay states the use of Tuscan juniper which is awesome but that isn’t going to mean much unless it is relative to other expressions that we know about.    Cascade Mountain Gin (the name might have changed), which sources their juniper from the world’s largest juniper forest gives us a three-fold unique expression. First its a wild foraged juniper from a distinct location. Second it is infused rather than distilled (or only partially?) so it was never degraded by heat making some of its aroma purity comparable to a vacuum distilled gin. And thirdly it was never fractioned so the outward lying juniper components that often get cut are in tact.  Other things like gustatory-acidity as still in tact which is why I think it is so enjoyable to drink neat. We end up with the pleasure of pondering; is its wild character due to being un-fractioned or due to unique sourcing? Lately it is my favorite gin.

Adventures in Aftermarket Maraschino

In my far from complete text on distillation I have a recipe for making maraschino liqueur which is a sweetened distillate of fermented sour cherries and a percentage of their pits which contribute the distinct aroma of bitter almonds. The really simple recipe is supposed to be a solution to a problem for people that live in areas that cannot access maraschino liqueur or in some cases cannot afford it. The recipe is constructed from hiram walker’s kirschwasser and a re-distilled amaretto. The alcoholic proof and sugar content are shaped to fit historically derived proportions. Because the recipe uses very small volumes, different ratios of cherry to stone pit aroma (benzaldehyde) are created using a gram scale to maintain accuracy when blending the two volumes to your own aesthetic. The recipe is simple and fun and makes a mean aviation. The distilled amaretto is astounding on its own. It is crystal clear and the divergence of the aroma from the color is relentlessly amusing. If any producer starts making a crystal clear amaretto, they will be met with instant success.

[I used this maraschino cheater in my nine round TKO and Ciroc vodka was directly inspired and has just released clear amaretto.]

The other day I just unearthed some more information about the role of benzaldehyde in maraschino liqueur in a paper titled: The Determination of Benzaldehyde in Maraschino Cherries and Maraschino Liqueur by A. G. Woodman and Lewis Davis. It was published in the Journal of Industrial and Engineering Chemistry, August 1912. I think the research was done right down the road at M.I.T. in Boston.

The paper applies new testing methods to maraschino that can determine the quantity of small amounts of benzaldehyde (5mg) as opposed to previous methods that could only handle comparatively large amounts (0.1g). Newly refined methods of this sort were constantly being applied to product analysis to protect consumers from adulteration and fraud. I’m in no position to put the chemistry to use (some day!) but these papers sometimes give us a great glimpse of what was on the market and what people were drinking.  Their method is easy enough to follow that someone very ambitious or a small scale commercial producer could easily apply it. Even more refined methods have likely been created over the years and it should be known that there is a correction to one of their reagent figures in another article.

Okay so here is the good stuff; brands with numbers:

brand                                   alcohol content     benzaldehyde (Mg/100cc.)

G. Luxardo                                 32.60               3.57

Richelieux                                  25.94               17.02

Marie Brizard & Roger               28.97               0.0

Cusenier (cherry liqueur)           32.63              12.01

Cusenier (maraschino)             19.00               0.0

Nuyens et Cie                            24.78              1.78

H. Shufeldt & Co., Peoria Ill.       30.60              41.3

According to the authors: “Genuine Maraschino, apparently, has a very low content of benzaldehyde.” The sugar and alcohol content I used for my faux rendering match luxardo’s from 1920, though I don’t think they have ever changed.

So what does this all mean? For starters I see these products as being all over the map, some have far more almond aroma than others and were all these brands really available in Boston in 1912? Without knowing exact numbers, I think the ratio of benzaldehyde to cherry I put in my faux maraschino was probably far higher than the genuine stuff and might have compared to the very last one in the list. I don’t really know which maraschinos in this list were premium brands at the time. I think at least luxardo and brizard but not cusenier because of their very low alcohol content.

If limited to small concentration but significant to the aroma of maraschino, I think it would be cool to have a known concentration of benzaldehyde so that renderings of “genuine maraschino” could accurately be assembled from great american kirschwassers like clear creek’s. To my knowledge they do not offer a maraschino. Who knows, if some bartenders create a market for a product that doesn’t really exist, maybe they will start making one and they will have a deep enough level of involvement that they can hit all the historically accurate numbers we now know exist.

Closing thoughts: You gotta earn that mustache and suspenders. Make your Ensslin era Aviations with historically accurate levels of benzaldehyde in your maraschino.

Advanced Limoncello Basics

If you enjoy this site, check out the Houghton Street Foundry, my fine arts workshop and follow @b_apothecary

Recently on egullet I attempted to help a few people troubleshoot their limoncello. They were experiencing cloudiness and did not know what to attribute it to. I suspected pectin while others thought terpenes. The conversation turned to terpene removal which made me look for other research out there.

As it turns out, limoncello is economically significant enough that a few PhDs have looked at it in depth. Some of the studies try to characterize limoncello chemically so that consumers can be protected from fraud and adulteration. Though some aspects of the studies are intimidating, there is still great practical stuff that home producers can learn from these studies.

I found a few papers and I’ll give a bullet point run down of the good bits.

“Analysis of volatiles in Limoncello Liqueur and Aging Study wiht Sensory” by Neil C. Da Costa and Theodore J. Anastasiou

**”it certainly does not contain any lemon juice”

**”but must be transferred to the freezer before serving” These guys posit the idea that limoncello is fairly perishable. Homemade is best because it can be freshest. Their study wonders how limoncello changes as it ages.

**They keep it in the freezer to minimize reactions of the extracted oils with water and ethanol as well as with air and light.

**Due to shipping, warehousing, etc. “there is a higher likelihood of commercially available Limoncellos having less of a true flavor profile than their homemade counterparts”. This is their hypothesis anyhow which they will test by looking at how significantly limoncello changes as it ages.

**They present a very long list of compounds they were going to cover but one I was concerned about was pectin which I couldn’t find in their list.

**They prepare their own limoncello recipe by following a Washington Post article: Grisco, J. Limoncello Recipe. The Washington Post, December 7, 2009 (this link was missing for a while but provided by an awesome commenter. thanks KM!)

**”seventeen large store-bought organic lemons (approx. 175g. each, approx 3000 g total) were washed, dried, and carefully peeled as to minimize the amount of pith. the peels were submerged in 1500ml of grain alcohol (95% abv, 190 proof). This mixture was allowed to stand at room temperature for 14 days with brief stirring every other day. After 14 days the spent lemon peels were removed by filtration leaving a dark yellow solution. 1700g of a simple syrup solution consisting of equal parts (1350g) water and granulated sugar were then added to the extract, resulting in an opaque yellow emulsion. this emulsion was allowed to sit at room temperature for an additional 21 days, again stirring every other day. …. the approximate alcohol concentration was calculated to be 50% (100 proof).”

**They didn’t use any terpene separation!

**One cool part of the technical stuff is the descriptors that accompany the analysis of “key volatile components”.

**After aging, some of the terpenes hydrolized/oxidized

**Aldedydes in such a high alcohol content formed acetals (which also happens to distillates when they age)

**There was ester formation (again this happens to distillates! and of course it is going to inspire another blog post!)

**”organoleptically, the fresh limoncello concentrated extract was described as strongly citral, fresh, lemon curd. the aged concentrate extract was more oxidized lemon, less fresh, heavy lemon, missing lower volatiles, lower citral. … the biggest difference detected was a reduction in the concentration of the highly volatile monoterpenes as they became oxidized, which gave rise to the loss of fresh citrusy notes. in addition the harsh gasoline, oxidized terpene note of p-cymene was increased, which had a negative impact on the aged sample.” … “the various ethyl ethers formed were not perceived as giving significantly negative notes to the aged sample.”

**They present two sensory “spider graphs” which are pretty cool and really intuitive. Hopefully I can paste the spider graphs in here because they are a tool we should probably see more of.

**They preferred the fresh but they didn’t find the aged stuff too terrible.

“Analysis of Some Italian Lemon Liqueurs (Limoncello)” by Versari Andrea, Natali Nadia, Russo Maria Teresa, and Antonelli Andrea

**For starters, those names sound like people that would know limoncello…

**”a total production of 15 million liters of Limoncello per year is estimated.” wow.

**When they detail the process they state that peels are infused in 95% ethanol for 2-7 days. Then it is diluted with syrup to an average alcohol content of 32%.

**”…Limoncello contains several volatile and nonvolatile minor compounds (ca. 2%), which are fundamental for its sensory characteristics. The former are terpenic compounds, which form the essential oil, and the latter include several classes of nonvolatile compounds with potential health-related properties, such as flavanoids, coumarins, and psoralens.”

**The presence of these nonvolatile constituents are sometimes used to tell whether an essential oil was cold pressed or steam distilled the latter being seen as inferior.

**They compare 12 samples by a variety of criteria

**The sugar contents were a surprise: 182, 186, 186, 185, 202, 277, 265, 223, 264, 199 g/L so there seems to be two different styles; a low and high sugar content.

**The samples also had notable citric acid contents that were all over the map: 2044, 438, 190, 209, 475, 152, 83, 1059, 172, 395, 846, 301 mg/L. I suspect the acid is added as a powder as opposed to being from fresh lemon juice. They are listed in the same order so the style that has a high sugar content in most cases also has a high acid content. Unlike the previous paper that sees limoncello as not being sour, quite a few commercial producers interpret it as a sour liqueur. I have no opinion personally.

**They also cover the ethanol content, acetaldehyde, ethyl acetate, methanol, propanol, and i-butanol content. For ethanol: 32, 31, 31, 31, 30, 30, 29, 40, 30, 27, 29, 28.

**A lower alcohol content than 31-32% might affect stability and lead to essential oil separation (those terpenes again!) while “higher values might not fit the consumer’s preference”. “in fact, Limoncello is considered a beverage, not a spirit.” On egullet someone mention 30% as being a key number for terpenes coming out of solution. This data makes me wonder if they are correct, and what would happen to the stability of the product that had 27%. Is that low number the producers intention or an accident?

**Speaking of accidents.. one sample had a declared ethanol content of 35% but 40% was found…

**”methanol showed values below the legal limit, whereas acetaldehyde content was above the regulatory limit for neutral ethanol which is set at 5 mg/L of anhydrous alcohol.”

**They find strange congener values that make them wonder if low quality grain alcohol was used or if the syrup in some cases had started to ferment. The occurrence of glycerol and high acetaldehyde level plus low methanol levels supports the hypothesis that high quality ethanol was used but some of the syrups likely started to ferment. Very weird. Kind of gross.  It sounds like some producers do not have control of their product.

**from their conclusion: “Citric acid content indicates the addition of lemon juice.” I do not think this is conclusive. Citric acid powder is easy enough to buy and the juice has all sort of oxygen sensitive compounds. Granted they have a lot of juice laying around…

**”Ethyl acetate, acetaldehyde, 2-methyl-1-propanol, and glycerol are most probably related to the occurence of microbiological activity in the sugar syrup used in the limoncello formulation” -Gross

**

Conclusion. Terpene separation is not a part of limoncello production. But the pectin issue wasn’t resolved. They might even be using a pectic enzyme, but in my understanding that might produce methanol that would make the authors note the phenomenon so maybe they are not? In the mean time just peel it thin! And please don’t let your syrup start to ferment.

1989 James F. Guymon Lecture: Distillation in Alambic by Robert Léauté

[So it turns out the reason Robert Léauté got selected to give the lecture was because of the Remy Martin-Scramsburg brandy collaboration that I had never heard of. In 1982 they invested a lot of money to build a world class Alembic distillery in Carneros.

A great piece from 1985. This great articles from 1986 shows how silver tongued and well spoken Robert Léauté is. Scramsberg had backed out of the partnership in 1987. And this Etude brandy might be a product of that distillery. I suspect they are still laying around liquor stores. This article from 2002 tells of the close and says glowing stuff about they product they were making. I think they were ahead of their time. The cocktail movement and spirits renaissance hadn’t happened yet.]

The James F. Guymon lecture series is pretty cool and there are a few other lectures from the series I want to track down. For the 40th annual meeting of the American Society for Enology and Viticulture Robert Léauté gave a talk about alambic distillation.  Léauté had been with Remy Martin since 1973 and at the time was head Cognac master, research and development manager, and RMS vineyeards technical adviser.

The presentation which I read as a paper is unique because of how articulate and concise it is.  Léauté covers things other people gloss over which makes the paper invaluable to beginning distillers trying to sort out some curiosities not found in other texts.  Most texts about distillation are either too simple or far too advanced. It has been hard to find explanations for practices geared towards intermediate levels of involvement.

Léauté doesn’t cover everything I’ve been curious about but I’m hot on the trail of explanations to those phenomenons.  Many of the things I’ve written need to now be reconsidered in the light of these new concise explanations.  Thanks Robert!

**The shape of the Cognac still head which is approximately 10% the size of the boiler provides a small amount of reflux thus influencing the degree of equilibrium (Equilibrium influences how volatile components distribute themselves).

**Léauté mentions the hydrometer port and how it is used to filter the distillation. This he unfortunately glosses over.  Germain Robin also mentions filtering the distillate at this point and claims to use unbleached toilet paper.  This kind of filter might be unique to Cognac distillers.  The filter probably doesn’t separate a solid but rather an insoluble liquid that floats to the top at some point of the distillation.  Léauté and Robin attribute it to fatty acids.  The toilet paper is likely used to soak up the small amount of liquid like a sponge as opposed to liquid flowing through it like a membrane but the process is never detailed.  The fact that toilet paper is enough of a sponge makes one think the oil might be fairly insignificant.  Léauté and Robin also fail to comment on when the insoluble liquid appears.  Fusel oils are often separated through a similar process during continuous column distillation because they are far less soluble in water than alcohol.  A fraction high in fusel oil is collected then enters a fusel oil separator where it is diluted with water until the oil comes out of solution and can be separated. The non oil fraction is then sent back to the still to be re-distilled.

**Simple easy to follow charts show how particular volatile components present themselves during the heads, hearts, and tails such as:

acetaldehyde & ethyl acetate

ethyl caproate, isoamyl acetate, ethyl caprylate, ethyl caprate & ethyl laurate

methanol, isobutanol, methyl-2-butanol & methyl-3-butanol

acetic acid, 2-phenyl ethanol, ethyl lactate & diethyl succinate

and finally furfural

The curves and their simplicity illustrates the significance and priority of separating heads and tails.

**A gem is Léauté’s explanation of how volatile compounds distill. Previous explanations I have read from Maynard Amerine seemed more complicated and harder to grasp.

“Each volatile component will distill following three criteria: boiling point, relationship with alcohol or water, and the variation of alcohol content in the vapor during the distillation.  With respect to the relationship with alcohol or water, there are several possibilities: (1) the component is completely or partially soluble in alcohol and will distill when the vapor is rich in alcohol; (2) the component is soluble in water and will distill when the vapor is low in alcohol; (3) the component is soluble in both alcohol and water and will distill throughout the entire distillation; or (4) the component is not soluble in water, but the water vapor will carry over this component (hydrodistillation).”

This starts to explain the phenomenon where components with fairly high boiling points come across at the beginning of distillation. Boiling points are not the whole story.

Type 1 components are acetaldehyde and ethyl acetate because they have low boiling points and are completely soluble in ethanol so they end up in the heads.

Type 2 components which also end up in the heads have fairly high boiling points (above that of water) but are completely or partially soluble in ethanol.  Fatty acids and fatty esters are in this category. This might also be where Germain Robin’s toilet paper filter comes in.

Type 3 components which are in the heads and the hearts “have a low boiling point (not above 200°C)” are soluble in alcohol and are completely or partially soluble in water. Examples are methanol and some of the higher alcohols. Methanol is harder to separate than people think.

Type 4 components which start during the middle of the heart have a boiling point above that of water and are soluble or partially soluble in water such as acetic acid or ethyl lactate.

Type 5 components (5 wasn’t yet mentioned but these are defined as appearing during the distillation) have a high boiling point and are very soluble in water. They start during the middle of the heart. Furfural is a type 5.

**Léauté mentions that “higher heat is favorable for the less volatile components, as increased heat [applied to the boiler] will allow them to distill earlier and to be present in the first fractions of the distillation in higher concentration.”  what isn’t clear is if Cognac distillers use this idea at any point in the run or if they only keep consistent heating programs.  I haven’t found an answer to the question whether applying significant amounts of energy to the boiler changes volatility through super-heated hot spots or through a slight pressure build up.  Amerine references one paper which might be the precedent for the observation in a journal article (a 1940 Australian brewing journal) and I’ve requested it through inter library loan but it hasn’t turned up yet.  I suspect this idea could be used to capture as much aroma as possible when re-distilling commercially produced spirits such as modifying chartreuse or a adding aromas to whiskeys, etc.

**Unlike other papers this one acknowledges three distillation “processes”. The first two are the usual but because in Cognac they take a “heads 2” fraction from the second distillation the third process is a unique distillation where the heads 2 fraction gets recycled and reprocessed.

**When describing each distillation phase, with the first phase Léauté mentions that “Each fraction is obtained at a temperature below 60°F. In this case, they are removed by filtration combinations between sulfury components and copper and a part of fatty acids and copper (Fig. 22).” The reference to figure 22 points to the furfural chart so I suspect it may be a typo and he meant to reference figure 21 which covers acetic acid. It isn’t clear what exactly Léauté means.  Germain Robin also references a specific condensing temperature for different fractions which I have never seen described in whiskey or rum production. I cannot figure out if Robin’s toilet paper filter is coming into play or if this is some sort of catalytic converter at play (the copper!) which he is referring to as “filtration” (English as a second language maybe?).

For the second distillation “heads and Cognac are obtained between 62°F and 66°F and the secondes below 60°F like the brouillis during the first distillation (Fig. 25).”

The slight change in temperature at the end of the run may help certain components come out of solution so they can be separated by the “filter”. It is probably only necessary when the distillate is rich in those components. Though the temperatures seem close together, the span might be all that you can practically change such a large volume of water by in a fairly short period of time. It is never said whether they use glycol chillers to keep the temperature precise. I suspect that the procedure and its high “involvement” (a term I like to use to point out varying levels of attention to detail) does not make or break the product. They do it because they can. They have mastered everything else and still have capacity to spare.  Whiskey or rum distillers may not follow the same procedures because they either do not encounter the same amounts of volatile components or they do not have the same degree of involvement.

**A gem is the comment “the purpose is not to make a table wine but to make the best possible raw materials to distill and produce a quality brandy.” Wines made to distill are different. It would be interesting to hear if and how grapes destined for distillation are treated are treated differently in Vineyard management practices.

**Léauté points to a 1978 study by Onishi et al. that systematically tried using the Cognac method on numerous varietals found in California.

**He then points to another Guymon memorial lecture by Elie Skofis in 1983 that gives suggestions for making Cognac style brandies in the U.S.  The tone of these papers is unique because they are intermediate as opposed to too advanced or too simple. They also make suggestions and give guidelines instead of just making cryptic observations using mass spec. or chromatography.

**Apparently Skofis’ paper gave recommendations for production practices. Léauté clarifies a few: “slight changes can be made: recommendation (4), no SO2 or no more than 20ppm to avoid having high quantities of acetaldehyde in brandy and recommendation (5), fermentation temperature between 68°F to 77°F. This is mainly done to reduce acetaldehyde and ethyl acetate by evaporation.

The very last comment strikes me as very interesting because in the book “Food Flavorings” by Joseph Merory a technique is mentioned for making fruit brandies where the unfermented juice is bolstered with spirits then distilled (unfermented!) to capture aroma, the then alcohol-free/aroma-free wash in then fermented to produce aroma for the next batch. Merory claimed that this was done to prevent aromas being lost with escaping CO2 during fermentation.  Merory might never have used his technique and it might have just been a fun potential experiment. Léauté makes it seem like evaporation or compounds escaping with fermentation gases can benefit the spirit by removing a portion of the two most volatile components you want to significantly reduce anyway.

**As for aroma created in the still (besides furfural)  Léauté says: “In addition, during distillation in alambic, the wine and the brouillis are cooked; many reactions occur between the compounds, and this phase generates delicate aromas.”

**Léauté mentions the use of “microdistillation” where glass lab-ware is used with copper shavings placed in the boiler to a give a glimpse of flaws that cannot be noticed by tasting the wine alone. He mentions it used to detect:

High acetaldehyde; too much ethyl acetate; acrolein; butyric smell; oxidization characteristics; high level of volatile phenolic compounds; and pollution by sulfury compounds, hydrocarbons, etc. So if a flawed 600 gallon batch of wine needs special treatment why not go into distilling it having an idea of what extra nasties you are going to encounter. I’m not aware if new, lower involvement American distillers are using microdistillation procedures to evaluate every new batch of distilling material. Most are understaffed and there is only 24 hours in the day.

**”Reaction between compounds during distillation: The first distillation last around 10 hours and the second distillation approximately 14 hours. When it is operating, many types of reactions occur between the compounds of mixtures which are boiling. The boiler can be compared to a reactor.”

“Regarding the future characteristics of the Cognac (or brandy), the reactions which occur during the first distillation are the most important. These reactions are functions of: the characteristics of the wine; the use of lees; pH and acidity; the size of the alambic; the temperature generated by the gas burner under the boiler; the duration of distillation; and the cleaning of the alambic. (Note: In addition, the wine extracts compounds from lees because of the heat. These components also can react with others during distillation.)”

pH might be a very important variable because it catalyzes reactions and could explain the importance of grapes high in total acidity, significance of the sour mash process, the use of sulfuric acid in rum production, or the freeze concentration of distilling material before distillation (to concentrate acidity!). Because the wine can extract a lot components from the lees, some of which are negatives, they are often separated while they wait for distillation and then are re-mixed before they are distilled. Whiskey mashes are made every week so it doesn’t sit around but wine is made only once in the season so it can sit around for considerable time while it waits its turn in the still. A still can get pretty grimy. It makes me wonder if any laissez faire cleanliness can add complexity.

“The yeasts represent around 60% to 70% of the lees. Their use in distillation wines give more fatty esters (like ethyl caprylate, ethyl caprate, ethyl laurate, esters C14 to C18), more fatty acids, and nitrogren compounds (like amino acids). The fatty esters give fruitiness to the Cognac; the fatty acids give body and are like fixatives for many other aromatic components; amino-acids are involved in the thermic break down reactions.”

Many distillers use the term fruitiness which implies olfactory-sweet aromas but I suspect these compounds also stimulate the umami which a lot of researchers still do not acknowledge. The term “fixative” might not refer to any chemical reaction but something at the level of perception. For example in cocktail construction a simple daiquiri can be tart and limey but when the non aromatic sugar is swapped for an orange liqueur the result seems to conjure grapefruit.  Compounds associated with the lees might be the orange liqueur in the example around which other more complex associations like grapefruit revolve.

“Types of reactions: The reactions are numerous: that is the reason that the double distillation technique in alambic is unique. Volatile components already in wine, may decrease or increase in concentration, depending on the types of reactions they are involved in. New volatile components can appear and generally are important for the aromas of Cognac.”

One thing that makes all of the reactions happen is the long time under heat that double distillation has relative to column distillation.  Esters can also form differently in column distillation because fatty acids that could latch onto ethanol often latch onto another higher alcohol instead that has accumulated in the column.  Some higher alcohol esters are seen as flaws and the phenomenon most significantly in continuous column distillation (as opposed to batch) because there is the opportunity for lots of higher alcohols to accumulate and congregate in one place on the various plates.  One component type that decreases are fatty acids and one type that increases as esters as one turns into the other through esterification.  Furfural also increases.  Copper reactions also decrease sulphury components and some fatty acid components.

“Examples of reactions: Some reactions have been known for years, such as hydrolysis, esterification, acetalization, reactions with copper, and furfural production. Generally, the above produce, in relatively large quantities (more than 2mg/L, constituents which are easily detected by gas chromatography. Other reactions produce, in a very small small quantities (less than 1 mg/L), constituents which are detected by glass capillary gas chromatogrpahy. During the last decade, many authors published articles about these reactions and the components they produced in spirits…”

“For instance, by hydrolysis, thermal break down and rearrangement may be generated: monoterpenes (linalool and alpha terpineol, <1 mg/L); ketones (alpha-ionone and beta-ionone, <0.01 mg/L); and others (vitispirane and T.D.N., < 0.1 mg/L).”

“Maillard reactions (reactions between a sugar and amino compounds) can also take place. The Maillard reaction is the main source of heterocyclic compounds such as furans, pyridines, and pyrazines…”

I think furans here refers to furfural.

**Léauté goes on to describe experimenting with Cognac production in California and how new varietals were explored.  They used a 6.5 gallons baby alambic and found each varietal had to have its own cuts made to maximize its varietal character.

**Some California brandies are analyzed chemically and Léauté makes some interesting observations based on the numbers. Two examples have unusually high amounts of hexanol which might indicate they were crushed and de-stemmed at a too high a temperature. These are considerations that can be made probably only as involvement deepens. He notes that some of the brandies are higher in ethyl acetate than others but below the threshold of perception.  Apparently you want to be as close to line without going over. Seeing the numerical data for spirits so that it can be compared to that of successful products or other data such as threshold of perception helps distillers optimize their cuts.

One of the distillates is rich in Isoamyl acetate which contributes a banana smell. This ester of a higher alcohol might be the result of column distillation as opposed to alambic.

**Léauté recommended chilling the experimental California wines to thwart malolactic fermentation but Cognac wines might go through malolactic fermentation as evidence by the ethyl lactate esters. His recommendation might have been made so the unique varietal character of the grapes could be explored. Malolactic fermentation produces generic aromas that though fun and potential “fixatives” for aromas, can rob a wine or spirit of its sense of place.

A great paper for beginning distillers. It can also be purchased properly for $10 from the American Journal of Enology and Viticulture.