Sally Grainger answers questions about The Story of Garum: Roman Fish Sauce in a Modern Context
Sally Grainger walks us through her experimentations with garum and how to differentiate ancient sauces (including rare Mediterranean survivals, namely colatura de alici and pissalat) from the modern forms in the east.
Editors note: This publication contains the video of the talk from the Fermentology webinar series, as well as a lightly edited transcript of the lecture. The format has been modified slightly to aid in readability.
Garum, an ancient Roman staple, was made by fermenting ungutted fish in the hot sun with salt. However, this notorious ingredient was transformed by Roman cooks and home-bakers with honey, herbs, and wines, and has many parallels with luxurious fish sauces used in five-star restaurants today. In this presentation, Sally Grainger will talk through her experimentations with garum and differentiate ancient sauces (including rare Mediterranean survivals, namely colatura de alici and pissalat) from the modern forms in the east.
Sally is a Roman food historian and experimental archaeologist. She has authored several books, including Cooking Apicus: Roman Recipes for Today and her latest The Story of Garum: Fermented Fish Sauce and Salted Fish in the Ancient World.
Roman fish sauce is not Roman fish sauce. It's Greek: it started out as a Greek sauce. It's a complex evolution of numerous different types of sauce, made with all the parts of fish, different components of fish, and salt.
The earliest evidence comes from Athens from the fifth century B.C.E, and the latest evidence comes from Byzantine texts from Constantinople. So it's a huge period of time that is constantly evolving. It can be very difficult to pin it down, and this is part of the problem.
Garum had a reputation then, and it still does now. Many archaeologists and historians will say that it's stinky and disgusting. “That smelly Roman fish sauce, bleh,” they will say. So there's disgust; there's confusion. And there's also [AUDIO OUT] because the ancients are in awe of it as well. If you are from the culinary world, you understand what umami is. You know fish sauce is a magical ingredient, but the Romans didn't. They were confused. A quote from the Roman Senator Seneca captures the feeling of disgust:
"What? Do you not think that so-called garum sociorum, the costly extract of poisonous fish, burns up the stomach with its salted putrefaction?"1
But the point is, this is a special kind of fish sauce. This is the most expensive, the elite fish sauce. And it's made with sanies. I've translated this from the Latin as extract, but actually sanies is an exudate from wounds. It is the viscera and the blood from specific fish; this is where the oral disgust comes from.
A proper understanding of ancient fish sauce draws on a variety of sources:
The didactic literature, i.e. the cookbooks and recipe books themselves;
Zooarchaeology to identify fish bones and understand the size of the fish you are examining;
The archaeology of amphorae;
The archaeology of the production sites themselves, where the fish sauces were made;
Marine archaeology: for instance, shipwrecks have been found in the Mediterranean with amphorae containing fish sauce. In some shipwrecks the amphorae were sealed with the fish sauce still inside; amphorae with labels have even been found, allowing us to understand fish sauce in a very real time;
The archaeology of consumption sites (i.e., discarded amphorae in towns and cities);
Epigraphy, the labels written on bottles and jar of amphorae, which can be very complex;
Literature, like satirical plays, that address fish sauce in a very different way than the didactic cookbooks;
An understanding of modern fish sauce.
The aim here is to outline the original ancient recipes themselves, compare them with modern techniques in South East Asia, and also compare them with some of the ingredients some of the fish sauces that are now available today around the world. Colatura di alici, for instance, is a modern sauce from southern Italy. Pissalat is a fish paste from Nice, France. Flor de garum comes from Cadiz, Spain. And Worcester sauce even draws on some elements of Roman fish sauce.
As mentioned above, Roman fish sauce is Greek fish sauce. There are two types in the literature. You get what's called garos, which is the Greek word for it, which we understand as liquamen. This is the one that is equivalent to modern fish sauce: small and medium-sized fish and mixtures of all kinds of different things. And if they're larger fish, they are cut open so that the viscera is exposed to the salt.
When lots of tiny fish are used, they likely dissolve on their own. If you add a mixture of small and medium-sized fish, you add a bit of extra viscera from another supply of fish to help to dissolve them. And if you use more larger fish, they are cut open and the viscera is exposed to the whole thing, which digests together. There is some suggestion that extra viscera formed a fundamental part of the recipe, but it didn't. It was just a matter of adding some form of digestive enzyme to the whole fish.
Ancient recipes salt the fish at 15%. In comparison, it's at least 25% and sometimes up to 40% salt in modern fish sauces. And they only salt them, they only ferment them, for up to three months. They still end up with an amber liquid. But that amber liquid is much less salty, and as a result can actually be used in larger quantities. It's not a broth or stock, not at all. But a greater volume can be used.
So for instance, there is a recipe for a dressing, which is fish sauce and oil and wine blended together, using equal quantities of all three. This can’t be done with a modern fish sauce.
The other fish sauce, as discussed in the quote from Seneca earlier, is this unique black and bloody sauce. It is called haimation from bloody and melan from black. It's the blood and viscera from either mullet, mackerel, and later tuna, and it's fermented with salt, although we do not know how much salt as the recipe doesn't state. This mixture is only left to ferment for two months.
This is a distinctly dark, black sauce, and it tastes of strongly of iron. It is used as a table sauce, so it's poured onto cooked food. It's never used in the kitchen in the cooking process. But it's also blended with oil and wine to make dressings, and not remotely a stock or broth.
As shown above, liquamen is the standard fish sauce, and garum is the black and bloody one. But some writers use the terms interchangeably, and therefore, there is confusion. Because you want to talk about garum as the standard fish sauce, but actually it's not: it's liquamen.
Ancient recipes survive in two texts. One is called the Geoponica, and it's a Greek agricultural manual from the Byzantine period. But the information that it contains spans the entire Roman period. So the recipe cannot be dated to the eleventh century C.E., but the book itself was compiled at that time. The other recipe is from Gargilius Martialis, his text is written in the second century C.E., but it is almost certainly also Byzantine, and the recipe is added to his text later.
However, both of these recipes, or these texts with recipes in them, are a long way from the first century C.E. That is the period where Roman fish sauce was at its was most popular, and the trade in it was most extensive, and the epigraphic and literary evidence is strongest. So these literary resources are quite contradictory, and they're very difficult to interpret.
One kind of fish sauce is open to the sun in large cetariae, open in cement-lined tanks. These are the tanks that are found in Baelo Claudia near Cadiz, one of the famous fish sauce factories in Spain. Their fish sauce is also made in enclosed containers, and this could mean small pots and jars with a sealed lid, or, wine dolia may have been used. But it has to have a sealed lid.
The open to the sun method involves additional brine and/or wine as the fish dissolve, to counter the evaporation. That you add wine is specifically stated in the recipe. After three months, the fish have disintegrated into a thick emulsion, with particles of muscle tissue suspended in brine.
But there are also many pieces of fish, as I learned with my fish sauce, from sardines of 10 to 20 centimeters in size. And after taking my creamy emulsion off, half the fish were still intact and undissolved. So it's always possible to rebrine the residue, and particularly so if you're only fermenting for three months, because you do not dissolve all of it.
During this process, the brine that the particles of muscle tissue are floating in becomes saturated in protein and salt. Then it cannot absorb any more protein or salt. And the whole process is suspended until more liquid is added. So that liquor is taken off.
And there it is. This is the liquor that you saw in that other picture, standing in a fish tank with all the bones removed. And you can see it floating. You can see the individual particles of muscle tissue, and you can see the sauce forming around it.
The agitation of removing it from the fish tank encourages liquefaction, and also gentle heat, but not any kind of cooking temperature. If you place your amphora of fish sauce in a warm room, it will encourage it to dissolve, but you don't want to add heat actually to the vessel.
I believe that fish sauce in the ancient world was stored and marketed in this unfiltered state. And therefore, protein could continue to dissolve into the liquid as it is stored. Otherwise the fish paste, these particles, can be taken off.
It actually floats on the surface over time, and so it can be scooped off, and you have a completely separate fish paste to use on bread as a relish, and you have the clear, sparklingly clear sauce. As you can see, it's possible to filter it quite easily.
So this kind of sauce-- liquamen flos it was called-- was marketed in various kinds of vessels, including amphorae and little jugs. And you find labels for it, for example: Liquamen flos scombri excelense. I would translate that as the finest fish sauce: that's the excelense. It's either a cream or unfiltered for the flos, and it's made from mackerel. And these labels are found particularly in Pompeii, but also all over the Roman world.
So I have actually tried to make a black and bloody garum. And believe me, it's not easy to bleed mackerel. It was a long and complicated process. But I made a small amount, a really dark liquid and a thick, gloopy paste of dissolved viscera, which had no culinary value at all. This fish paste looks like the same fish paste as a whole fish sauce but actually tasted quite bitter and unpleasant.
So this type of sauce was probably made in much smaller quantities and only consumed by elites. We're fairly certain about that. The dominant product in terms of the trade and the economy was the other kind of fish sauce that was called liquamen.
There is great confusion in the late empire, as, in fact, the blood viscera garum ceases to be made. It's much less common. And ancient writers at that time start to revert to using garum to mean the original fish sauce, i.e. liquamen. Sometimes we do not know which one they meant when they're talking about fish sauce.
The other production method in the ancient world is the enclosed, completely airtight container. Same mix of fish, but it's layered between huge quantities of herbs and also even more excessive salt, which leads to a very dry environment. This is estimated at two parts fish to one part salt, which is very excessive.
Below is a quote from the Gargilius recipe.
“Then the second layer is laid down using fish - whole if they are small, cut in pieces if they are larger - over this is added the third layer of salt two fingers deep, and the vessel is to be filled right to the top in this, with successive triple layers of herbs, fish, and salt. It should then be closed up with a little fitted and put aside as it is for seven days.”2
That's all the recipe says. The mixture is then stirred. And then you leave it up to one month in total, which is completely different from the three months and completely contrary to Asian recipes that use eighteen months, so it's very quick sauce.
Most of the salt, in fact, cannot dissolve. It stays [AUDIO OUT] It is very difficult to stir, and the yield of sauce is very small. And in fact, the protein is exceptionally low, because we do know that too much salt impedes enzyme activity. But clearly the container is airtight, and so there is no loss of volume.
But this is not the same as using a small container that is closed but using the open to the sun method. And it is difficult to know how much of that enclosed recipe was ever used in the Roman world. Because the image of the fish sauce shop in Pompeii that you can see-- that wonderful sort of doorway into a courtyard. And this is a shop that was designed for the manufacture of fish sauce and the marketing of fish sauce. So the amphorae are stacked at the back through the little doorway.
And in front you can see the dolia. Now the dolia - which is a big pot - the dolia are covered with a roof tile. So it's not a lid designed for the pot. It's simply a tile from the roof placed on top to stop debris, and let vermin in, and so on. There would be a considerable amount of evaporation.
So I think the open to the sun technique was the standard technique. And I'm not sure that the enclosed technique was used that much, but it is difficult to know for sure. The people who make the sauce do not speak about their methods at all. It's only the people who consume it that talk to us about their interest in it and what they like about it.
When comparing ancient and modern recipes, there are two important things to take away. Modern recipes take eighteen months, while ancient recipes take three months. The shorter time will result in considerable difference in the rate of disintegration and liquefaction of the fish, which leads to the systematic rebrine. But the ancient sources are silent about that rebrine. We just have to assume they did it.
The low salt in an ancient recipe makes for a much more desirable sauce, because we know that too much salt interferes with the enzyme activity. But that low salt increases the presence of lactobacillus bacteria. I wasn't able to identify all the different bacteria that were present in my ancient fish sauce, but there was a high content of lactobacillus bacteria. And modern fish oils does not contain any, as far as I'm told.
So this lactobacillus bacteria increases the fermentation, increases enzyme activity. It increases umami and flavor. But it also contributes to the degradation of the fish sauce thereafter. So it has a shorter shelf life. And this is partly because there's a certain amount of conversion of the protein into ammonia. This is quite natural, and it happens naturally with fish sauce.
But the higher the protein content, the more likely the sauce is to degrade, which is rather sad, because it means that it would not be storable for more than, say, about six months. It is difficult to judge. Fish sauce is not aged at all, according to the recipes and according to [AUDIO OUT] and so it was probably consumed quite quickly.
There is one other fish sauce in the ancient world that is equally as popular but much more difficult to understand, because it is quite rarely described. And this is the brine derived from salted fish. So it is necessarily cleaned and sanitized sauce, because the viscera and the blood have been removed from the fish, particularly tuna and mackerel. It can be aged for up to four years. And it was used in exactly the same way as liquamen.
And in fact, in the later Roman period when Christianity takes over, many communities follow a blood prohibition. So you were not allowed to consume any kind of animal blood, whether it be meat or fish. And so they start to consume muria and call it garum, and think of it as garum, and think of it as their standard fish sauce.
And this whole issue about blood prohibition is the obvious reason why the black and bloody garum sociorum ceases to be made, because the Roman world becomes the Christian world. I was in Athens once preparing an ancient Greek banquet, and we were going to use salted tuna. So I salted my tuna for two days, and then I soaked it in fresh water for two days and used it as a freshly salted tuna.
The color of the brine ended up being quite unusual, because ever since, I have I have tried to duplicate this brine and never could. And I think it's something to do with the bleeding of the tuna. The tuna that I was given in Athens had not been fully bled out, so the flesh was quite pink.
And so obviously, the brine becomes pink. Ever since I have had fresh tuna given to me in kitchens all over the world - in America, in London, in France - and it's grayish-brown. It doesn't have an exciting color to it at all. But this is potentially what a muria in the dining room as a desirable source would have looked like.
The color of sauces varied, so identifying your sauce required a knowledge of what shade they are, how they blended. Because often I think the diner in the Roman dining room would not have seen neat fish sauce necessarily. They would have seen it blended with oil and wine.
So it's very difficult to understand how they understood and engaged with fish sauce, when it was always going to be mixed in with other ingredients before they saw it. And of course, they never saw it in the kitchen. That's the most important thing. The diner never enters the kitchen and sees the cooking process.
The first modern comparison will be with South East Asia, because they have perfected the techniques for making fish sauce. They know what they're doing. They use anchovy, but also many other types of fish, mainly under 12 centimeters. The fish are salted fresh at a ratio of about 25% to 40% salt by weight. They are put into big tanks, which are covered just to stop debris, but otherwise evaporation is quite a common occurrence.
Fish disintegrate in the presence of digestive enzymes. So they break apart, and they break apart into particles of muscle tissue. This is stirred on a regular basis. And you have to deal with the concept of the-- because you lose volume all the time, and it is often replaced. Fish sauce is made at an air temperature of about 30 degrees, give or take. But the liquid temperature of the sauces themselves never reach higher than 25 degrees.
The temperature is an important point, because many people tried to make sauces quickly by using heat, and it doesn't work, often spoiling the flavor. And it's left to ferment for between twelve and eighteen months, at which point the residue is re-brined in situ in the tanks. And the second quality sauces are blended with the first quality, and you get a standard by mixing them all together. It's then transferred into tankers and bottled.
Producing these sauces is a very long, static process. Eighteen months is an awful long time to invest in fish sauce. The end result is an amber gold liquid with a relatively high protein level. Relatively high, because the standard is 10 milligrams per liter, although it is possible to go much higher than that.
The other common way to make fish sauce in South East Asia is to use the enclosed wooden barrel drum. These are actually sealed, not open to the air at all. Same type of fish, same salt levels. These are bamboo barrels, and they're filled with layers of fish and salt, which are compressed. And then more fish and salt are added until it is absolutely full. It is then left to ferment and dissolve in the same way for up to eighteen months. This method also calls for occasional stirring, but not as often as in the open tanks.
There are also some techniques in which some of the liquid is removed from the bottom and poured in the top, essentially filtering it through. These tanks have the filter mechanism integrated into the barrels. The residue is re-brined in situ or often traded for second quality sources. Again, it's a long, static process, and the resulting sauce is amber gold.
As a result, South East Asian sauces are very salty. They use a lot more salt than they need to. But the taste, the traditional taste, is for a very salty liquid, and it's used in relatively small amounts. It's not used, for instance, to add volume to a sauce at all. It's used in relatively small quantities. You can find [INAUDIBLE] for export.
This type of fish sauce, when found in supermarkets, tends to be quite low quality, possibly even less than 10 milligrams per liter in terms of protein. And they can be colored, when extra sugar is added. But it is quite a clean and sanitized production method, and it works very well. This sauce is a great option if what you want is an intense umami kick, but also a strong salt taste. However, there is very little fish taste to the sauce.
Another kind of fish sauce that you can find in South East Asia is bagoong and patis. Now this is a Philippine sauce made from anchovy, krill, and prawns. It is salted in the same way, three parts fish to one part salt, again very high in salt. The mix is then left for anywhere from one week to six months. As a result, there is much less fermentation and much less umami.
This method results in a thick suspension of fish paste (bagoong), which is almost more desirable than the patis (the resulting liquid). In fact, the results of this method closely resemble ancient liquamen and allec: it is manufactured similarly, used as a paste, and the liquor is used as a seasoning.
Colatura di alici is a modern fish sauce made in Campania. The original meaning of this phrase is the filterings of anchovy, and it is a fish brine. The fish (the anchovy) are beheaded and deboned and eviscerated in one action. You snap the head off, you pull the spine out, and the viscera comes away as well.
The remaining portion of the fish is then salted to remove the blood, and that first liquor that comes out of the fish is thrown away. It is then layered with fresh salt and compressed for anything up to six months, but sometimes a good deal longer.
And the liquid that came out of that process was either discarded initially - anecdotally used as a seasoning - and the anchovy was meant to be consumed as anchovy. But they decided to change the method. And it's difficult to know when this happened. We're talking probably about twenty, twenty-five years ago now.
Producers decided to discard the anchovy as anchovy, and to let the sauce form into a paste, and then strained it fully to market it as a cooking liquor. It's an artisanal product, and it's manufactured using habitual techniques. There are films on YouTube of people making it.
The salt levels are unspecified, because they don't weigh it. They simply sprinkle it on between the layers of fish, and you never know what actually the ratio is. There's no viscera, so there's no digestive enzymes, which means the protein levels are very low. But there is a huge perception of quality, and it is very expensive. However, compared to Roman fish sauce it is very salty and quite low in protein.
Another modern sauce is the result of a collaboration between scientists and archaeologists at the University of Cadiz. They follow the Gargilius Martialis recipe completely. So they use all the herbs, only two fingers of salt, and they leave it to mature for about six months.
Spanish chefs have been know to rave about it, but they do admit it's a little too salty, which it is by Roman fish sauce standards. The herbs add a very distinctive quality to it, giving it a very interesting flavor. This sauce is exceptionally expensive: 30 euros for 100 milliliters.
I only just discovered this fish sauce. I only discovered it yesterday, and it's made by a chef in Catalonia. And it's been reformulated for the modern palate according to the blurb. And it's made with water, salt, and anchovy.
But of course, I don't know the ratio of salt and how long he leaves it. All these issues are very important, of course. This sauce is about 5 pounds 50 for 120 milliliters. It's not too bad, not too expensive.
Pissalat is an interesting fish paste. It's very similar to the fish paste that's scooped off the top of the fish sauce in the ancient world. They behead and debone and eviscerate in exactly the same way as colatura di alici, and they throw away the first liquid. The remainder is then softened in the brine. Instead of keeping the brine and having a sauce, you discard the brine, and you end up with a fish paste. It can be aged for anything up to six months, but is usually only left for a few months. Obviously there's no enzyme in the pissalat, so there is a limited umami taste. This particular sauce is used quite a lot in Nicoise cuisine, and it is quite tasty.
This is a rich and thick sauce. It is fabulous on bread, on toast. Very thinly spread, I have to say. I wouldn't like to eat an awful lot of it, but interesting, very interesting. And of course, very similar to patum piperium, which is that wonderful anchovy paste that you can buy, which is a cooked anchovy paste, but still very similar.
Another popular modern fish sauce is Red Boat. It's anchovy prepared in bamboo barrels, in the same way. It's 25% salt when it's made, to make sure it's stable and there's not too much bacteria. But the salt is reduced after production to 15%. The mixture is then left to mature and ferment for twelve months. It has the highest protein levels of any sauce. You remember 10 milligrams per milliliter is the standard. Red Boat is 40 milligrams per milliliter, and it can go up to 50.
This sauce has exceptional flavor. On the tongue it explodes, and it packs a kaleidoscope of flavor. Very sticky and thick, so it's almost like a boiled stock, a concentrated stock. And so if you want to, if people ask me what kind of sauce should I use to reproduce Roman recipes, I would always say Red Boat.
Now all sauces darken with age. A fresh bottle of Red Boat will have that golden amber color. Very high nutrition means that it doesn't age well. It becomes quite pungent and much less desirable. A really fresh bottle of Red Boat is more desirable. Also, Family Reserve is 50 milligrams per liter, so even more protein. Exceptional quality, and you'd only need a tiny amount.
Lastly is Worcester sauce. It has some hints associated with Roman sauces. The exact recipe is a secret: there is a lot of vinegar, and sugar, salt, and cooked anchovy, so very different to what we've been discussing.
In the late Roman/early medieval period, we find recipes very, very similar to - virtually the same as Worcester sauce - that are called garum and are used as a seasoning in the same way as fish sauce was used. And they become ketchup, and they turn into Lea & Perrins. So there's continuity all the way through.