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 (https://m.youtube.com/user/GGATArchaeology and https://en-gb.facebook.com/GGATarchaeology/).

Observing Textures in Recipes

By Elaine Leong

I have held a long fascination with how textures are represented in recipes. As we all know, then as now, producing medicines and food often involves a multi-step process, and careful observation of changes in textures is often the key to success.

Classic White Sauce

Take, for example, the classic white sauce. It all seems simple enough – we mix and heat together butter and flour and then add milk (hot or cold, depending on where you stand on this issue), simmer and whisk away and, voilà, we should have a silky-smooth sauce, ready for some posh mac and cheese, or baked endive, and much more. Now readers, I know what you’re thinking. It sounds so easy on paper but, if we were honest, we all have stories of failed batches of béchamel. The sauce can taste raw (classic mistake of not cooking the flour enough) or be lumpy (the blender trick never works for me) or end up too thin or thick.

Mixing Flour and Butter

A few years ago, I finally found the perfect recipe for me from Annie Bell’s In my Kitchen. However, though Annie provides the perfect ingredient proportions for my family’s taste of white sauce, for the crucial step – cooking the butter and flour together – I rely on Martha Schulman’s description. The mixture needs to be heat for around 5 minutes until it looks like ‘wet sand.’ The monitoring and observing of textures, particularly any changes, is key to making the perfect white sauce and many other dishes besides.

The early modern recipe archive is also filled with similar sets of instructions where changes in texture were used as markers for the completion of a particular step in production. In my recent book Recipes and Everyday Knowledge, I discuss some of these examples in a chapter on “Recipes Trials.” Due to the generosity of friends and colleagues and the enthusiasm of groups such as the Early Modern Recipes Online Collective (EMROC) and Before Farm to Table, new examples emerge all the time. Today, I wanted to share a particularly intriguing recipe, which came to light at the EMROC transcribathon last fall.

Dawson’s Recipe for Lemon Wafers

The recipe is for making lemon wafers and is part of the recipe collection of a seventeenth-century English gentlewoman named Jane Dawson. The instructions are brief but detailed. We are told to dry, sift and beat double-refined sugar and mix with the juice of a lemon until it becomes the consistency of honey.[1] Then, scooping some of the mixture in a spoon, we should heat the spoon over a chafing dish of hot coals until the surface of the mixture touching the spoon is “crisp” – that is (according to the OED) rippling, folding or wrinkling. Taking care that the mixture does not boil, we should then spread the melted mixture onto a square piece of paper, pinning the two corners of the paper together in order to curl or bend the wafer and let it dry in this configuration. When we are ready to eat or serve the lemon wafer, we should wet the “wrong” side of the paper with water to release the candy.

As with making béchamel, key to this recipe are the practices of observing and interpreting changes in texture. Two points are of particular importance here – ensuring that the sugar and lemon juice mix achieves the ‘consistency of honey’ and that the mixture heats until it crisps or ripples on the hot spoon.

After the transcribathon, some EMROC members were so intrigued by this recipe that they tried their hands at re-creating it. Lisa Smith, Maggie Simon, and their various assistants spent afternoons mixing and tasting lemony sugar syrup and heating it using a variety of methods from plate warmers to electric hobs. I’ll leave you to read about their adventures here and here, but it is telling that both ended their posts with a reflection about the assumed knowledge required for this recipe. One particular texture was picked up for comment – the consistency of honey. Both Lisa and Maggie were stymied by the instructions to mix fine sugar and lemon juice to the “right” consistency of honey. After all, as a natural product, honey can come in many guises. Our intrepid makers tried to reproduce the thickness of raw honey, runny honey, and crystalized honey and each resulted in a different product with varying degrees of success.

Maggie’s Sugar and Lemon Juice Mixture (Photo taken by Maggie Simon)

Observing textures or changes in textures is clearly a key part of following recipes. Yet, it turns out that it is hard to convey hands-on experiential knowledge on paper, particularly across time and space. Often times, descriptions of textures are made using analogies (e.g. consistency of honey) or metaphors (e.g. wet sand) requiring the recipe writer and reader to work within a similar frame of reference. Further focus on reading or interpreting representations of “textures” in past and present thus seems a fruitful way to shed light on histories of observation, sensorial and experiential knowledge.


[1] Folger v.b. 14, p. 47.