Roubo and Watin: The Sweet Scent of Early Modern Varnishes

By Érika Wicky

Fig. 1: Nécessaire, wood and leather, ca 1760, 8,4 x 6 cm (closed), Grasse. Courtesy of Musée international de la parfumerie.

If the decorative arts of the eighteenth century shine so brightly, it is thanks to the innovation and mastering of techniques such as gilding and varnishing.[i] While the latter may have given surfaces the shiny appearance that was particularly desirable at the time, both were above all essential to protect wood, especially against insects. Varnishes were used not only for furniture, but also for small painted objects such as snuffboxes, fans, and boîtes à mouches or fly boxes (Fig. 1). Varnishes’ appealing visual characteristics and protective properties were marred, however, by what was generally perceived as their unpleasant smell. The recipes for some of these varnishes have been passed down to us, and with them, the possibility to retrieve their odors. Some of these recipes continue to be in use, such as the varnish described by French cabinetmaker André-Jacob Roubo in L’Art du menuisier-ébéniste in 1774 (Fig. 2), whose use remains frequent in restoration[ii] so that the same recipe is regularly remade at the Center for Research and Restoration of Museums of France. While current olfactory accounts of these products are much more nuanced, they do persist, which is a sign that olfactory sensitivity has certainly evolved, but also that the smell of the varnish remains a major part of the experience of its manufacture and use. What are the substances that have given varnishes such a characteristic olfactive identity through the centuries? And how has the meaning attributed to these smells evolved?

Fig. 2 : André J. Roubo, L’Art du menuisier-ébéniste, Board 11 (Paris: 1774). Courtesy of WikiMedia Commons.

There are a number of recipes for varnishes found in eighteenth-century texts, and all essentially contain aromatic plant resins. For instance, the varnish used by Jean Félix Watin (1773) was prepared as follows: “Two ounces of mastic in tears and half a pound of sandarac in a pint of wine spirit, when the materials are well dissolved together, incorporate four ounces of Venice turpentine.”[iii] Similarly, Roubo’s (1774) was “composed of one pint or two pounds of wine spirit, five ounces of sandarac, two ounces of mastic in tears, one ounce of gum elemi and one ounce of oil of aspic.”[iv] Essentially, the process involved Venice turpentine (the name given to the resin of Melese[v]), mastic in tears, sandarac and other more precious resins such as copal, or of lesser quality such as galipot, which were generally dissolved in alcohol (wine spirit), and then heated. This gave off such a strong and unpleasant odor that the practice was forbidden within city walls.[vi] Indeed, this smell could be significant during the process of preparation, and Jean Riffault, for example, recommended that cooking copal should be stopped when it gave off an empyreumatic or burnt smell [vii] (Fig. 3).

Fig. 3: Hendrik van Reede tot Drakestein, India Copal (Vateria indica L.), colored line. Courtesy of the Wellcome Collection.

Although the smell of varnish released during firing is still considered strong and dangerous today, our current perceptions seem to accord less significance overall to the smell of varnishes. For example, the sensory-aware description of a varnish recreation at the V&A mentions that the smell of lavender dominated that of turpentine.[viii] That the smell of turpentine could be camouflaged was well known in the eighteenth century, when aspic oil was often adulterated with turpentine, which was much less expensive. To ensure the quality of the aspic oil, Watin recommended approaching a soaked cloth to a fire, which dissipates the lavender oil quickly and allows the detection of the odor of any turpentine that may have been added.[ix] Since most of these raw materials were subject to adulteration, the artisans could, in fact, evaluate their quality thanks to their sense of smell. For example, according to Baudeau’s Encyclopédie méthodique (1783-4), an unpleasant smell from gum elemi (which is usually pleasant) can reveal that it was substituted with galipot.[x] The olfactory consensus here seems to rest on the fact that the artisan knew enough about the agreeable smell of gum elemi not to be tricked. Similarly, olfactory references in professional treatises, such as Roret’s guide, often rely on experience, invoking “characteristic” or “particular” odors rather than defining them by analogy, like when elemi gum is described as smelling like fennel.[xi]

Far from being a mere nuisance, the smell of the raw materials thus made it possible to identify them, in the same way that wood, whose different species were characterized by their smell, was sometimes considered when choosing materials. Odour could be so defining that, as Cheng He has shown, the concept of lacquer was long associated with a range of odorous resins rather than with a finished object.[xii] For example, objects decorated with Martin varnish [xiii] were simultaneously prized in the eighteenth century and dreaded for their bad odor. They were distinguished in particular by the use of copal, a plant resin whose use as an incense in Mexico is often mentioned in technical treatises.[xiv]

The manufacture of varnishes thus intersects with other types of olfactory knowledge. The resins used for the varnishes are found in many other recipes, in particular for perfumes. Louis Peyron’s 1986 compilation of recipes for perfumes to be burned, which had been recommended as a way to defeat the plague in the 1720s, reveals that the components of varnishes were all found in these recipes and that their strong odor was believed to have the prophylactic virtue of keeping miasmas away.[xv] The connection between varnish and perfume can be further explored in eighteenth-century perfumery treatises: in 1761, the royal perfumer proposed a “varnish” for the complexion made with sandarac spirit, whose protective and preserving virtues aligned with those of wood varnishes.[xvi]

The negative connotations associated with the smell of varnish seem to have faded with time; tear mastic was used in perfumery into the nineteenth century, [xvii] while, according to perfumer Lucile Lefranc-Gallo, gum elemi is now experiencing renewed interest in perfumery, alongside a range of spicy raw materials. The scent of varnishes produced according to eighteenth-century recipes, such as Roubo’s or Watin’s, takes us back to a time when smells had specific connotations and were part of a network of knowledge provided by olfaction, particularly with regard to the identification of materials. It is not only the sensitivity that has evolved, but also the entire olfactory culture.


[i] My warm thanks to Marc-André Paulin (C2RMF), Lucille Franc-Gallo and Ersy Contogouris for their advice and comments. This text forms part of the work conducted for the Marie Sklodowska-Curie research project PaintOdor (845788).

[ii] Nathalie Balcar, Frédéric Leblanc et Marc-André Paulin, “La protection de surface pour les meubles en marqueterie de métal du musée du Louvre : étude d’un vernis, entre formulations anciennes et expérimentations actuelles,” Technè, 49 (2020).

[iii] Jean Félix Watin, L’art du peintre, doreur, vernisseur (Paris: 1773), 229. Watin claimed to have discovered the secret of an odorless varnish, but he does not give its recipe in his book.

[iv] André J. Roubo, L’Art du menuisier-ébéniste (Paris: 1774).

[v] Pierre Chomel, Abrégé de l’histoire des plantes usuelles (Paris: 1712), 190.

[vi] Watin, 224.

[vii] Jean Riffault, Manuel théorique et pratique du peintre en bâtimens, du doreur et du vernisseur (Paris: Roret, 1825), 254.

[viii] “This time, despite both ingredients being very strong smelling, the scent of the lavender oil completely overpowered that of the turpentine.” Simona Valeriani, “Recreating Sixteenth-Century Varnishes on the History of Design Course,” June 23, 2017,

[ix] Watin, 202.

[x] Nicolas Baudeau, Encyclopédie méthodique (Liège: 1783-1784), 63.

[xi] Riffault, 229.

[xii] Cheng He, “Understanding the Fragrance of Lacquer in Early Modern Europe,” University of Toronto Art Journal 9.1 (2021): 68-76.

[xiii] Anne-Solenn Le Hô, Céline Daher, Ludovic Bellot-Gurlet, Yannick Vandenberghe, Jean Bleton, Myrtho Bonnin, Léa Drieu, Juliette Langlois, Céline Paris, Marc-André Paulin, “French lacquers of the 18th century and vernis Martin,” ICOM-CC 17th Triennial Conference, September 2014, Melbourne, Australia. hal-01279161

[xiv] Chomel, 512.

[xv] Louis Peyron, Odeurs, Parfums et parfumeurs lors des grandes épidémies méridionales de peste Arles 1721, Talk presented at the Arles Académie on March 23, 1986.

[xvi] Le parfumeur royal (Paris: 1761), 115.

[xvii] Auguste Debay, Nouveau manuel du parfumeur-chimiste (Paris: E Dentu, 1856), 37.




Novel Liquids: Brandy, Shrub, and Early Modern “Cocktails”

A receipt 'To make Shrub' from a ‘Book of Receipts for Cookery and Pastry 1732 & c.’, attributed to Sarah Tully. (Wellcome Collection, MS.8687, p. 131).
A receipt ‘To make Shrub’ from a ‘Book of Receipts for Cookery and Pastry 1732 & c.’, attributed to Sarah Tully. (Wellcome Collection, MS.8687, p. 131).

Tyler Rainford

Everyone has a favourite drink. Whether it’s a pint of pale ale, a smooth glass of merlot, or a certain cocktail for when we’re feeling fancy, we know what we like and what we’d rather avoid. We’re also very particular about how we drink it. Hold the ice. Shaken not stirred. Deviations from the norm can be seen as a form of alcoholic sacrilege, as today’s debates over the merits of the pre-mixed or pre-batched cocktail might indicate. Everyone has their preferences. In many ways, early modern Britons were similarly pedantic, and recipe books from the period indicate an intense interest in what went into any one drink. This was particularly clear following the emergence of distilled alcohols, which we commonly refer to as “spirits”.

Spirits had long played a vital role in the maintenance of household health and wellbeing in the early modern period. Featuring prominently in English receipt books alongside a host of physicks, syrups, and salves, these curious cordials were designed to alleviate the ailing body and bring a physical respite to the suffering drinker. However, as the seventeenth century progressed, their medical function became increasingly opaque. By the early eighteenth century, distilled spirits were also consumed for their apparent intoxicating effects, fiery flavours, and socially lubricating potential. Although not all contemporaries were enthused by the emergence of these new tipples, recipe book writers demonstrated an acute awareness of how these liquors could be crafted to serve individual tastes and personal pleasures.

This newfound taste for distilled spirits was most evident in the case of flavoured brandies, and recipes abound for brandies made from, or infused with, a host of fruits, sugars, and spices. Cherries, raspberries, oranges, and lemons were enthusiastically squashed and squeezed into this heady mix. An anonymous receipt book, possibly kept at Worth House, near Tiverton in Devon between 1714 and 1773, even contained a recipe for rhubarb brandy, combining rhubarb, cardamom seeds, saffron and nutmeg. The author considered it an ‘an excellent receipt’ but suggested no obvious medical benefits.

Another popular beverage, infused with the juices and rinds of citrus fruits, was shrub. Although relatively time consuming, this fruity liquor was easy to make and did not require the use of a still.  An example from the receipt book of Rebecca Tallamy, likely kept between 1735 and 1738, dictated:

To a Gallon of good Rum put a Quart of Juce fresh squees’d & strain’d, two pounds of good Loaf sugar, take half the Lemon rinds & six of Oranges & steep them one night in the Juce & Rum then strain it through a Coarse Cloth or Bag into a vessel or Bottle, Shake it three or four Times a day for Fourteen Days then let stand to settle till it is as water, then draw it off in Bottles Cork them well & hosen them down, besure not to put in any decay’d fruit nor a Sweet one.

(Wellcome Collection, MS.4759, fo. 173v.)

On the one hand, Tallamy’s receipt for shrub was straightforward. It contained three key ingredients – citrus fruits, sugar, and a form of liquor – that were to be mixed, strained, and shaken to produce what we might define as an early modern cocktail. However, beyond these three basic ingredients, there was no one prescriptive formula to be followed. Contemporaries either made do with what they had or adapted receipts to suit their individual tastes and preferences. The only restriction was to avoid using overripe or rotten fruit. Beyond this, the choice was theirs to make.

Another receipt book, supposedly belonging to Sarah Tully (c. 1708/9 – 1736), contained two recipes for shrub, each written in a different hand. One specified it should be made with brandy and three pints of boiling milk, whereas another suggested the reader could substitute the brandy for rum, ‘if you think Proper’, but made no mention of milk. Despite being penned only a few pages apart; these two receipts were considerably different, suggesting there was no one ‘proper’ way to make shrub. Personal preferences were paramount and could change over time. Another receipt book, likely composed by Anne Lisle in 1748, used the juice from ripe currants alongside a gallon of rum, brandy, or arrack, suggesting the choice of liquor was subject to taste. Similarly, a receipt book belonging to Anne Talbot of Lacock Abbey in Wiltshire claimed shrub could be mixed with white wine, cider, or brandy. The reader could choose ‘which [they] please’.

Clearly, shrub was a liqueur drunk for pleasure. But how was it consumed? Although some contemporaries might have drunk shrub as it was, it was frequently mixed with water to make punch: an immensely popular beverage with fascinating maritime origins. This choice is understandable. Undiluted, shrub was likely very strong, and very tart. Its zing needed to be tamed and receipt book authors made this clear. Sarah Tully’s receipt referenced above suggested that the liquor could be mixed with water to create an ‘Excellent punch at once’. Another receipt book, attributed to Mary Bent, contained a receipt for shrub with the addendum: ‘In Makeing the punch put two pints of Watter to one of Shrubb’. Shrub could be quickly transformed at a moment’s notice, suggesting it was a highly adaptable beverage.

In this respect, rather than being served as a drink in and of itself, shrub could be compared to an alcoholic squash, or a premixed cocktail. A point that might give some snobbish mixologists today pause for thought. It was bottled, sealed, and stored in keen anticipation of revels to come. Preparation was a fundamental part of this process. The prevalence of shrub in contemporary receipt books provides but one example of how individualised and adaptable the landscape of drink could be in early eighteenth-century England.


Tyler Rainford is a second year PhD student at the University of Bristol, funded by the SWW DTP. His research explores the role of intoxicants in early modern England, with a specific focus on how distilled spirits informed ideas about the self and society over the course of this period. More broadly, he is interested in consumption, work, and identity in the early modern Atlantic world, c. 1600 – 1800.

Cherries Galore in a Cesspit

By Merit Hondelink

As an archaeobotanist, an archaeologist specialised in studying plant remains found in archaeological excavations, I aim to reconstruct and interpret the relationships between humans and plants in the past. Archaeological plant remains, also known as subfossil plant remains, help us to reconstruct the former landscape and inform us how humans exploited it and even transformed the vegetation. Archaeobotanists do not necessarily study one time period, nor a specific region or topic. They can study plant remains from the Palaeolithic or the 20th century, and everything in between. They can focus on one specific site, work across the country or continent, and even work worldwide. They can delve into topics such as natural vegetation, forestation, domestication, trade, food consumption and much more. The one thing that all of this has in common is the link between humans and plants. But most archaeobotanists do specialize, most notably in the plant parts they study, such as fruits, seeds, pollen, wood or phytoliths. And most archaeobotanists have a beloved time period, favourite region or topic that they find most intriguing. In my case my research focuses on early modern Dutch urban food consumption.

I study what people ate in early modern Dutch cities, and how this changed through time. The best way to study what people ate in the past, is to look at their excrement and kitchen refuse, both of which can be found in the archaeologists treasure trove: the cesspit. These latrines were used to empty one’s bowels, but also served as a place to discard kitchen refuse and household waste. The content of a cesspit consists of organic remains from plants and animals, inorganic (culinary) material culture such as earthenware, glassware and ceramics, but also wooden cups and plates, as well as (decorative) objects, personal belongings and much, much more.

Figure 1: A selection of faunal and floral items found in a late medieval cesspit sample from Groningen. Photo: Dirk Fennema.

The content of an archaeobotanical cesspit sample consists of, among others, floral remains in different shapes and sizes (Figure 1). The items are sorted with the use of a microscope (Figure 2) and identified on a species level (and sometimes even on the level of species variety) by using a reference collection (Figure 3). The Groningen Institute of Archaeology offers a wonderful digital, open access, reference collection, see

Figure 2: A peek through the microscope. Visible is a fragment of text and different seeds and fruits, taken from an early modern Delft cesspit sample. Photo: Merit Hondelink.

When the content of a cesspit sample is analysed, sorted and identified, the interpretation begins. What can these plant remains tell us about past human-plant relationships? Most plant species are interpreted in a standardized way: wild plants inform us about the vegetation composition, make-up of soils and hydrology, whilst agricultural weeds in particular inform us about the crops grown and their local, regional, international or even global provenance. Wild but poisonous or toxic plants inform us about potential medicinal applications. A majority of plant species found in cesspits are classified as economic plants, grown as a food crop or cultivated for other useful purposes, such as fibres for textiles or seeds for oil. Identifying edible plants helps us better understand what plants people used for food and which parts people consumed. It also helps us better understand how food was prepared in the past, as preparation marks can be left behind on seeds and fruits.

Some preparation marks are easier to identify than others: nuts need to be cracked to get to the seed; apple seeds may be sliced when cutting up an apple, cereals can be ground, resulting into fragmented bran. But sometimes the archaeobotanist finds fragmented plant parts that, at a first glance, do not make sense.

Figure 3: A small selection of the tubes from the archaeobotany reference collection housed at the Groningen Institute of Archaeology (GIA) at the University of Groningen. Photo via GIA.

I have come across dozens and sometimes hundreds (or even more) cherry stones and plum stones in a single cesspit sample. No surprise there, cherries and plums were grown in local orchards, sold in the market and consumed with gusto. Most of these stones will have been discarded in the cesspit as a result from eating the fruits and spitting out the stones, or after de-pitting the fruits for dinner preparation. Only a small percentage is assumed to have been accidentally swallowed and secreted as excrement. Still, archaeobotanists find many fragments of cherry and plum stones (Figure 4). This is something that raises questions when you think about it. Why would these sturdy fruit stones be fragmented? A more pressing question when you are aware that the Rosaceae family, among others also including almond, peach, and even apple, contains – to varying degrees – hydrocyanic acid, also known as hydrogen cyanide and sometimes called prussic acid. The seed coat and fruit wall protects the consumer from digesting this acid, which can be poisonous when consumed. So why would someone break the stones of these fruits?

Figure 4: Two fragments of cherry stones found in an early modern cesspit in Vlissingen. Photo: Merit Hondelink.

To test the assumption that cherry stones were fragmented intentionally, and not through, for instance, pressure, an experiment was devised. Cherries were bought at the farmer’s market and taken to a physics lab to measure the pressure required to fragment the stones. After a number of tests, the calculated force to fragment a cherry stone averaged 23,9 kg or 239 Newton (Graph 1). This makes it more plausible that the stones were intentionally fragmented, as opposed to – for instance – fragmentation due to soil pressure.

Graph 1: Force needed to fragment a cherry stone. On the vertical axis the force (N), on the horizontal axis the elongation (μm). The point where the line falls is the moment the cherry stone breaks (max. force – max. elongation).

Consulting early modern cookbooks provided me with a list of recipes requiring the cook to de-stone cherries for the preparation of jams, sauces, syrups and tarts. Delicious experiments ensued, but I did not manage to fragment cherry stones whilst cutting and de-stoning, pressing through a cloth or colander, or by just baking the fruit with stones in a tart in the oven. Working a batch of cherries with a mortar and pestle did the job, though. But than you would have to pick the fragmented stones from the mushy cherries: not ideal at all. Picking up the eighteenth century encyclopaedia compiled by Noël Chomel gave me the hint I needed. In the Dutch version of his Dictionnaire œconomique (Algemeen huishoudelijk-, natuur-, zedekundig- en konst- woordenboek), he mentions different recipes for preparing cherries. Two recipes for cherry liquor instruct the reader to fragment the cherry stones by using a mortar and pestle (Figure 5). The fragmented fruits, including the stones and (I assume) the seeds are added to the brandy (Dutch: brandewijn) and, after closing the bottle, the mixture is put in the sun to infuse. Adding spices such as cinnamon, cloves and sugar is optional, according to the author.

Figure 5: How to make a pleasant cherry liquor (Noël Chomel, 1778).

So, it is plausible that the fragmented cherry stones found in early modern cesspits are the result of the domestic production of cherry liquor. Other fruits, such as plums and peaches, are also used to make a fruity liquor according to Chomel’s encyclopedia. However, what happens to the acid contained in the seeds? That requires further research. It might be that the prescribed infusing in the sunlight helps denature the acid into harmless molecules, leaving only the (bitter) taste behind. This line of research will be undertaken come summer with the aid of a brewer and some chemical analysis. In the meantime, a cherry and cinnamon flavoured lemonade is my poison of choice. Bottoms up!

A Roman Vegetarian Substitute for Fish Sauce

By Edith Evans

Roman cookery has been one of my research interests since the 1980s; I’ve accumulated a large repertoire of ancient recipes and usually do at least one live demonstration a year.  Most of the recipes include garum or liquamen – fish sauce – as a taste enhancer, providing salt and umami. Whilst finding fish sauce is fairly easy nowadays in Britain (the Romans used the same techniques to make it as the modern Thai and Vietnamese), using it at demonstrations disappoints vegetarians who would otherwise like to sample the plant-based dishes.

I found the answer to this problem in a Late Antique agricultural treatise:

Liquamen from pears: Ritually pure liquamen (liquamen castimoniale) from pears is made like this: Very ripe pears are trodden with salt that has not been crushed. When their flesh has broken down, store it either in small casks or in earthenware vessels lined with pitch. When it is hung up [to drain] after the third month without being pressed on, the flesh of the pears discharges a liquid with a delicious taste but a pastel colour. To counter this, mix in a proportion of dark-coloured wine when you salt the pears.
– Palladius: Opus Agriculturae 3.25.12

Liquamen castimoniale must have been required for people observing certain religious strictures (castimoniale means ‘to do with religious ceremonies’). Why would ordinary liquamen have been thought unsuitable? Was it the fish? (Pliny the Elder writes of a special fish sauce for Jews (Natural History 31.95) that he calls garum castimoniarum, although he’s obviously got the wrong end of the stick when it comes to Jewish food laws because he says it’s made using fish without scales). Alternatively, was it because liquamen was the product of fermentation? Fermentation was often considered a form of decomposition, which might have led it to be regarded as ritually unclean.

This has a bearing on how we interpret the recipe. Although Palladius tells us the ingredients to use (whole pears and salt, plus optional red wine) he does not give any information about the relative proportions. This leaves us with two possible techniques. Either you use a high proportion of salt and effectively create a brine utilising the juice of the pears, or you use a low proportion and promote a lactic fermentation by incubating the mix a suitable temperature (although Palladius doesn’t mention this). When used to flavour food, the product of the first method adds a strong taste of salt but no umami. The second would add some umami but also acidity, but a much lower amount of salt. However, if the problem was the fermentation itself, the second method would have been as unacceptable as standard fish sauce.

I’ve had a go at the lactic fermentation method, using 2% of the weight of the pears in salt, but when I tried it, the mix went mouldy before fermentation had a chance to take hold.  I’ve had much more success with the first method and have repeated it enough times to get a consistent product. The best pears to use are juicy varieties with very tannic skins, like Williams (also known as Bartlett) and Comice. I mash up the pears – stalks, skins, cores and all – mix them with 25% – 50% of their weight in coarse sea salt (I don’t bother with the wine), and leave them at the back of the fridge in a glass jar with the lid only lightly screwed on. At the end of two months (unlike us, the Romans counted inclusively), the pulp has started to separate out. The heavier elements form a pale layer at the bottom of the jar, whilst the top part of the mixture is more liquid and is a pale pinkish-brown. When drained through a nylon sieve, the colour of the resulting liquid is a very pale version of the colour of fish sauce. 

I’ve tried various proportions of salt, and found that, if you use 50%, you seem to get more liquid, probably because the mixture doesn’t draw in moisture from the air to the same extent.  But a smaller percentage of salt allows more of the delightful pear flavour comes through – I find it much more difficult to detect in the 50% version. Stored in a clean bottle it will keep for months without refrigeration.

Figure 1: The pear liqumen is in the flask with dark blue trim

I’ve only had a problem once, when spots of mould had appeared on the surface of a batch six months after I’d made it. As I was due to give a Roman cookery demonstration in a few days’ time I had to quickly rustle up something I could use, so I cored and cut up a pear, boiled it with 25% salt and a little water, removed the peel and pulped the flesh in the blender. It was much too pale, but the taste was the same and I decided it would be a useful method for someone who couldn’t wait two months – in fact that’s what I recommend for my Roman Cookery School videos ( and