The honeys, from white to black

Where it is practiced

Costa Rica is the country of origin of the color nomenclature of honey (proceso miel, in Spanish). It was born in the micro-mills, small processing units set up on the farm, which multiplied from the early 2000s. Until then, the small Costa Rican producer delivered his cherry to a cooperative mill and had no control over its processing. In 2006, Juan Ramón Alvarado, of the Brumas del Zurquí micro-mill, won the first two places in the national Cosecha de Oro competition with honey lots. The process then spread across the country in the late 2000s. Two pressures were added: the specialty market, and a water restriction that followed an earthquake. At the Las Lajas micro-mill, Oscar and Francisca Chacón say they switched to honey after losing access to water for several weeks in the middle of the harvest. El Salvador is the second hearth, with farms such as Los Pirineos and producers such as Jorge Raúl Rivera, who mostly work black honey on Pacamara and Bourbon varieties. The process is now practiced well beyond: Panama, Guatemala, Peru, Uganda, Vietnam, Myanmar, and Yunnan in China, where the most recent scientific work comes from. No census says what share of the world harvest goes through this processing.

The principle

Honey is not a single process. It is a family of settings along a single axis.

The starting point is common to all of them. The cherry is pulped, that is, stripped of its skin and outer flesh. What remains is the parchment, the rigid envelope around the bean. And on that parchment remains the mucilage, the sticky layer of flesh that adheres to it, rich in pectins, sugars and water. In a washed coffee, it is removed, by fermentation and then washing, or by mechanical scrubbing. In a honey, all or part of it is left, and the coffee is put to dry with it.

This gesture sets two processes against each other on the same lot. Drying removes the water. Fermentation consumes the sugars. The two compete for the same mucilage, and it is drying that decides. As long as water activity stays high, yeasts and bacteria keep working. They produce acids, alcohol and esters, aromatic molecules that carry the fruitiness. When the water drops, microbial activity stops. The lot freezes in the state it was in.

Hence the real lever of the process. It is not the mucilage in itself, it is the length of time the coffee stays moist enough to ferment. Two things govern that length.

The first is the quantity of mucilage left. Mucilage is hygroscopic, that is, it retains water. The more there is, the more the layer slows evaporation, and the longer the window stays open.

The second is the management of drying. Shading, the thickness of the layer and the frequency of turning all act on the same speed of evaporation. A thick, little-stirred layer dries slowly, even with little mucilage.

The colors name the position on this axis. White is the fastest and leanest setting, black the slowest and the most loaded. Yellow and red lie between the two. The name is descriptive in the literal sense: it is the shade the parchment takes once dry, from pale blond to an almost black brown.

We must now state the central point of this sheet, and state it plainly. No standard sets these limits. No body publishes numerical thresholds for white, yellow, red and black. Neither ICAFE, the Costa Rican coffee institute. Nor the Specialty Coffee Association. Nor the International Coffee Organization. Since two different levers lead to the same result, the contradiction is structural. One producer reaches his black by leaving one hundred percent of the mucilage. Another reaches it by drying a less loaded lot in the shade. Both use the word correctly. Neither describes what the other describes.

Why a honey has more body: the explanation is not the quantity of material

The fuller body of honeys is usually explained by the sheer quantity of mucilage left on the parchment: more of it remains, so more of it passes into the cup. The explanation is incomplete, and Hackler and Särchen give a second, more interesting one, which comes down to metabolic regime and not to mass.

They write that the pulp remaining after pulping allows a fuller body and higher sweetness compared with washed coffees, and they add the reason: because the coffee is in open air and not under water, conditions are aerobic, so less sugar is broken down and more reducing sugars remain.

The reasoning can only be understood with the energy balance set out in the plate on fermentation. In the presence of oxygen, a microorganism draws about twenty-seven units of energy from each sugar molecule, against only two when oxygen is lacking. It therefore has to consume thirteen times less sugar to do the same work. A mucilage drying in the air, that is, in aerobiosis, is a mucilage where the organisms present burn little to live a lot. A mucilage immersed in a closed tank, on the contrary, is a medium where an enormous amount of raw material must be burned for the same growth. The difference in residual sugars between a honey and a washed coffee thus does not come only from having left more of it: it comes from consuming proportionally less of it.

This mechanism also sheds light on a counterintuitive observation reported by the same authors, that of Knopp, Bytof and Selmar, who find more residual glucose and fructose under aerobic conditions. Aerobiosis spares the sugar instead of speeding up its loss, and that is exactly what a honey exploits.

Two reservations go with this contribution. The first is that the authors give no figures: neither a comparison of reducing sugar content between honey and washed, nor a body score from a panel. It is a plausible mechanism, well founded in biochemistry; it is not a measurement. The second is that the drying of a honey is not uniformly aerobic: the interior of a thick layer of mucilage, especially in the first hours and on a loaded bed, can become anoxic, and that is precisely the risk the management of the process tries to avoid through stirring and layer thickness. The aerobiosis rule holds for the surface, and keeping it up at depth is work, not a given.

A sensory description by color usefully completes this sheet, with the caution that it comes from an unsourced brochure and must be presented as a reading from a cupping table. According to Kornman, white honeys can be almost indistinguishable from traditional washed coffees; black honeys echo naturals and taste of raisin and fresh grape; yellows and reds often have a pulpy character of ripe cherry or plum, without being as openly fruity as the darkest ones. That these descriptors line up continuously with the quantity of mucilage left is consistent with the mechanism above, without proving it.

Sources for this passage: Hackler and Särchen, chapter 6, p. 93; energy balance, chapter 3, p. 25 and 26; residual sugars in aerobiosis, chapter 3, p. 23, after Knopp, Bytof and Selmar, 2006. Descriptors by color: Kornman, p. 15.

The sequence

1. Harvest and sorting. Ripe cherries only. This is a heavier constraint here than in a washed process. Washing, in a washed coffee, carries away part of the sugars and masks differences in ripeness. Here nothing is washed, so everything that enters the layer of mucilage stays on the bean to the end.

2. Reception and flotation. The cherries are poured into water. Those that float are hollow, dry or damaged, and leave. This sorting sets the baseline level of the lot, because dry and defective fruits are the ones that carry the most fungal contamination.

3. Pulping, without water or with very little water. This is the water saving of the process, and it is the argument that led to its adoption in Costa Rica after the water restriction of the late 2000s.

4. Setting the proportion of mucilage left. Two schools coexist, and this is where the nomenclature breaks down.

• The mechanical school sets a mechanical demucilager, a machine that rubs the parchment to scrub off a share of the mucilage. The color is decided by this setting.
• The drying school leaves one hundred percent of the mucilage on all its lots and decides the color later, by the management of drying alone. This is what Las Lajas does in Costa Rica.

Neither school measures the proportion obtained. The announced percentage is a machine position or an estimate, not a weighing.

5. Starting the drying. Layer thickness is the first parameter. Cenicafé, the national coffee research center of Colombia, gives a range of 2 to 10 centimeters for parchment in solar drying. A honey is laid thin at the start of drying, because the mucilage glues the beans together and a thick layer sets into blocks.

6. Turning. It governs two things at once, the uniformity of moisture and the detachment of agglomerated beans. The frequencies reported by farms range from hourly turning on the first day to a few passes a day. At the other extreme, the black honeys of Las Lajas are described as barely stirred at all, which deliberately prolongs fermentation under the layer.

7. Shading. This is the lever specific to the dark colors. Yellow honey gets the most sun. Black honey is dried in the shade, under netting or under cover, and sometimes under a tarp.

8. Duration. It follows from everything above, and it varies widely.

• Yellow honey in Costa Rica: about 8 days, according to the official communication of the Costa Rican sector.
• Black honey at Los Pirineos, in El Salvador: about 25 days in a thin layer on shaded raised beds.
• In a controlled trial in Yunnan, at 25 degrees and 75 percent relative humidity: 15 days for red, 22 days for black.

9. End of drying. The usual target is 10 to 12 percent bean moisture. The Chinese trial stops at 11.0 percent, plus or minus 1.0. This is not a comfort figure: at 12 percent moisture, water activity falls to around 0.60, the threshold below which molds no longer develop.

10. Resting in parchment, then hulling and sorting of the green coffee.

What it gives in the cup

Here we must separate what has been measured from what the trade asserts.

What is measured. Three recent studies, all Chinese and all on Yunnan arabica, compared mucilage levels under controlled conditions.

The first, published in Foods, compares five treatments: dried cherry, washed process, red honey, black honey and anaerobic fermentation. Black honey and anaerobic come out ahead on mouthfeel, balance and overall score. The numerical tasting values appear only in the appendices of the article, and this sheet could not read them.

The second, published in Food Chemistry: X, compares five mucilage retention rates for their effect on volatile compounds and microbial flora. The group designated as red honey, defined there by a retention of 75 to 80 percent, gives the highest relative content of volatile compounds. Thirteen aroma compounds there exceed the threshold of olfactory contribution: four esters, three aromatic compounds, three terpenoids, two heterocycles and one ketone. The associated descriptors are mellowness and fruitiness.

The third, a targeted metabolomics study, identifies 1,682 metabolites in the green bean across five mucilage removal rates. The removal rate significantly changes the profile of non-volatile compounds. The families that move the most are amino acids and their derivatives, organic acids, phenolic acids and sugars.

One limit runs through these three studies, and it is a serious one. They measure an effect of the proportion of mucilage, not an effect of the colors. Since the colors are not defined the same way from one article to the next, their results cannot be carried over as they are to a bag labeled red honey.

What is claimed. The trade announces four things for honeys:

• a sweetness intermediate between washed and natural;
• a rounder body than a washed coffee;
• a less lively acidity than a washed coffee;
• a simple proportionality between color and the intensity of fruit. None of these statements rests on a trial comparing the four colors on the same lot, the same farm and the same year, with a panel. No trial of this kind was found.

What goes wrong

Molds and ochratoxin A are the underlying risk of the process, and it is physical in origin, not caused by some exotic microbe. The fungi that produce ochratoxin A belong to the sections Circumdati and Nigri of the genus Aspergillus. They need a high water activity to produce the toxin: from 0.96 to 0.98 for Aspergillus carbonarius, up to 0.99 for Aspergillus westerdijkiae and Aspergillus steynii. The risk is therefore a matter of time spent above the threshold. A 2024 review describes drying in three phases. The first lasts 6 to 8 days, around 0.95 water activity. The transition lasts about 6 days, around 0.90, and this is where toxin production becomes possible. Finally comes the arrival at 12 percent moisture, that is, 0.60. A washed coffee dries in 7 days, a cherry coffee in 12. A honey falls between the two, and a black honey exceeds both. Lengthening the drying means lengthening the stay in the dangerous window.

Over-fermentation gives the named defects of the trade. Fermented and sour. Acescence, that is, the turn to vinegar, when acetic bacteria oxidize ethanol into acetic acid. Stinkers, caused by butyric acid, which smells of rancid butter, and propionic acid, which smells of onion. The phenol defect over long durations.

Agglomeration is the risk specific to mucilage left on. The beans stick to one another and form blocks. Inside the block, water no longer escapes, fermentation continues and mold sets in. This is what the frequency of turning in the first days is meant to prevent.

Rain during drying is the seasonal risk. Rewetting the lot raises water activity again, reopens the window, and cancels the progress made. The process requires a reliable dry season, which restricts the terroirs where it is reasonable.

The named defects found when sorting the green coffee that increase with drying time are black beans and sour beans. A bean that has turned entirely black is the terminal stage of uncontrolled fermentation.

One commercial risk remains, which is not a defect but costs just as much. The lack of a definition makes purchasing non-reproducible. Two bags labeled red honey, from two farms, do not describe the same processing. A buyer who reorders on color alone is buying a word.

Benchmarks

• Target final bean moisture, and the corresponding water activity10 to 12 percent, or 0.60
• Water activity needed by Aspergillus carbonarius to produce ochratoxin A0.96 to 0.98
• Layer thickness in solar drying of parchment (Cenicafé)2 to 10 centimeters
• Drying of a yellow honey in Costa Ricaabout 8 days
• Drying of a black honey at Los Pirineos, El Salvadorabout 25 days
• Mucilage retention attributed to red honey according to the source50 to 60 percent in Foods 2026, 75 to 80 percent in Food Chemistry: X 2025