Bioarchaeology of Roman Seafood Consumption

As more Americans are developing an interest in the local food movement, preferring to purchase and consume perishable foods closer to home in order to cut down on economic and ecological costs of food production, we're also learning more about how the ancient Romans were decidedly not localvores.

Imported Seafood?

It's no revelation that the Romans were importing food to the center of the Empire to feed their growing population: grain from North Africa, to distribute to the Roman poor as a grain dole (Garnsey 1988); olive oil from Greece; and wine from France.  But today's Nature news brings a summary of an article in the International Journal of Nautical Archaeology by Beltrame, Gaddi, and Parizzi, who argue that a Roman shipwreck discovered with a lead pipe in its hull was transporting live fish around the Mediterranean.

Lead tube from the keel (photo by D. Gaddi)

The 2nd century AD ship was found off the coast of Aquileia in the northern Adriatic and recovered in 1999. Cargo aboard consisted of around 600 amphorae containing oil from Africa, whole and processed sardines, salted mackerel, and garum (a fermented fish sauce the Romans loved).  While the cargo isn't unusual, the presence of a lead pipe in the hull of the ship is - based on the position of the pipe, Beltrame and colleagues think seawater could have flowed into that area of the hull.  The pipe is not part of a bilge-pump, the authors argue:

Lead tube and lead sheet
(fig. 7 in Beltrame et al. 2011)

No seaman would have drilled a hole in the keel, creating a potential way for water to enter the hull, unless there was a very powerful reason to do so. At most he would drill a bilge-hole in the top-side shell, well above the waterline, so to avoid as the entrance of water inside the planking. Beyond the potential danger of water infiltration inside the keel, discharging below the waterline created another technical problem—counter-pressure at discharging, which would have involved extra power from the pump to expel reflux fluids. Finally, this solution would have required the installation of a non-return valve on the feed-well to prevent water entering the pipes when the pump was not functioning.

While the purpose of a bilge-pump is to send water out of a ship, a suction-pump is responsible for bringing water onto a ship.  The authors favor this latter interpretation and further explain:

If we posit the presence of a suction-pump capable of taking in from the sea at least 252 litres per minute—a flow that could have been almost doubled when applying a larger force— the discussion over its use remains open. [... A] possibility is that a large quantity of water could be used to fill basins for the transportation of live fish.

Their idea is interesting but not without drawbacks, most notably the difficult relationship Romans had with the sea (Purcell 1995) and the general lack of evidence for a market for live seafood.  Beltrame and colleagues do tackle these issues, though:

The transport of live fish in the ancient world is a controversial topic: on one hand some scholars, using an anthropological and deductive approach, invoke the apparent technical limitations faced by the Romans in preserving perishable foods to argue that the consumption or marketing of fish would have happened in the immediate vicinity of the port where they were unloaded, or the vivarium (fish-pond); they consider that fishing was mainly a subsistence activity, and that in few places was there a real market.

[...] If critically read, the written sources do not appear to provide substantial evidence supporting the theory of transport by sea, at least over long distances, of live fish in Roman times, except in special cases and perhaps for small quantities (as quoted by Pliny). But Macrobius in the late Imperial period does offer the description of vivariae naves, apparently identifying a specialized vessel, and Athenaeus describes watertight tanks placed on board— not compartments within pierced hulls.

[...] We like to suggest that our ship was used for transporting fish alive from the Istria vivaria to the rich tables in Aquileia; we must not forget, however, that during the spring, even the shallow waters of the lagoon of Grado and Marano could be a large natural fish-pond which could feed a burgeoning market for young live fish in other locations.

Stable Isotopes and Marine Resources in Rome

The question remains, though, who would have been consuming these fish?  Although we think of the Mediterranean diet today as including a lot of seafood, it's unclear how much of the average Roman diet was made up of freshwater fish, seafood, shellfish, and garum... and it's unclear if there was such a thing as the average Roman diet, since dietary practices varied based on age, sex, social class, and occupation. For information beyond the historical and the archaeological, we can turn to the bioarchaeological data and look at patterns of marine resource consumption in Rome and surrounding sites in the Imperial period.

First, a bit of background.  The human diet comes from a variety of sources, but mainly from foodstuffs containing protein and carbohydrates. For decades, bioarchaeologists have been measuring the isotope ratios of carbon and nitrogen from human skeletal remains to get an overall picture of a person's diet in the years before his death (Katzenberg, 2008).

Carbon enters the food chain through photosynthesis, so a human’s carbon isotope ratio is largely affected by the kind of plants he ate.  The two major photosynthetic pathways are called C3 and C4 – C3 plants are found in temperate climates and include wheat and barley.  C4 plants are more tropical grasses and include corn, millet, and sorghum.  An individual who ate more C3 plants will have a more negative C isotope ratio, and someone who ate a lot of C4 plants will have a more positive C isotope ratio.

Nitrogen, on the other hand, is obtained primarily through consumption of other organisms - both plants and animals - so understanding N isotopes means understanding an organism’s trophic position, or position in the food chain: what it eats and what eats it.  Plants such as legumes directly process atmospheric N and therefore have an N isotope value close to zero.  Terrestrial plants are slightly higher, then there are herbivores that eat those plants, followed by carnivores that eat herbivore meat.  A person who ate mostly marine resources would have a higher N isotope range, as the trophic chain in the marine environment is much longer than on land.  It's also interesting to note that a breastfeeding infant would also have a very high N isotope ratio, as the infant is consuming his mother’s dissolved tissues.

For part of my dissertation research (Killgrove 2010), I did a palaeodietary study on 48 individuals from two Imperial-period cemeteries in Rome with help from isotope chemist Rob Tykot.  We looked at the perimortem diets of men, women, and children and saw a diverse range of foods consumed:

Carbon and nitrogen isotope analysis from two Imperial Roman cemeteries

Based primarily on historical sources and previous isotope studies of the Roman diet, we may expect that the lower-class individuals of Rome would have consumed a diet consisting of cereals (wheat, barley, and/or millet), olive oil, wine, and legumes, with some possible contributions from meat and fish. And that is, more or less, what the chart above shows me. The people we analyzed from Rome consumed a diet based largely on plants such as wheat and barley, and meat from terrestrial animals, with variable but minor input from legumes, marine resources, and millet/sorghum.

There are two exceptions in the graph above: a male has a very high C isotope value, indicative of a diet high in C4 plants. This aberrant diet could indicate that he was eating a large quantity of millet and beans, which Pliny notes were often eaten together by people in rural Italy. And a 2.5-year-old child with a very high N isotope value was either still nursing or had been recently weaned.

Diversity in the Imperial Roman Diet

Map of sites with palaeodiet information.  Blue = Casal Bertone;
Red = Callixtus; Green = Castellaccio Europarco; Yellow =
ANAS; Purple = Portus; Not shown: Velia (south of Naples)

Only a handful of palaeodiet studies have been done on populations from Imperial Italy. The best comparative study for individuals from Casal Bertone and Castellaccio Europarco is Prowse and colleagues’ analysis of skeletons from the Imperial necropolis at Portus Romae (Prowse, 2001; Prowse et al., 2004, 2005) as well as a small analysis of some garum samples. Recently, a dietary analysis was published from the Christian (3rd-5th century AD) necropolis of St. Callixtus, which was located about 3 km from the city walls of Rome and less than 5 km from Casal Bertone (Rutgers et al., 2009). Although this site is as likely to represent a periurban population as Casal Bertone is, the fact that these individuals were early Christians raises the issue of differences in diet due to asceticism. Prowse (2001) published additional sample data from the cemetery known simply as ANAS, which was located halfway between Portus and Rome. Further south on the Italian peninsula, a dietary study of people from early Imperial Velia (modern Elea) was undertaken by Craig and colleagues (2009).

The graph below shows the mean and 1 sigma standard deviations (following removal of outliers) for the data published from Casal Bertone, Castellaccio Europarco, St. Callixtus, Portus Romae, ANAS, and Velia.  This is a rather crude comparison but does show that the average diet at each of these sites was quite different:

Palaeodiet comparisons at sites along the Tyrrhenian coast of Italy

People in different places were eating different types and amounts of carbohydrates (the carbon axis) as well as different amounts or types of marine resources.  Prowse and colleagues (2004) interpret the ANAS isotope values as indicative of a terrestrial diet. At Portus Romae, on the other hand, individuals had comparatively higher nitrogen isotopes, not unexpected since Portus was located on the Tyrrhenian coast. The Roman Christians buried in the necropolis at Callixtus have average nitrogen values midway between Portus Romae and ANAS, indicating more marine protein consumption than the latter but less than the former. This necropolis was located about 3 km from the Tiber River, meaning people living in the area could have had access to marine resources. Rutgers and colleagues (2009) note the comparatively low carbon values from individuals in the Callixtus necropolis and interpret them as evidence of consumption of freshwater fish. Finally, at Velia, a site in Campania on the Tyrrhenian sea, Craig and colleagues (2009) concluded that their average isotope values were consistent with a largely terrestrial diet of cereals with a small contribution from meat and marine resources. These researchers further note, however, that there was substantial variation within the population at Velia, suggesting that the diet of individuals buried there was not uniform.

Comparisons between the diets of individuals at Casal Bertone and Castellaccio Europarco with Imperial-period sites from the Italian peninsula show that there was no singular Roman diet. To a base of cereals, olives, and wine were added terrestrial meat, legumes, fish, and millet in different proportions and from different sources. Although copious amounts of food were imported from various areas of the Empire, the diet of the average lower-class Roman was likely contingent on foodstuffs available in the immediate area.  However, there is clear variation in the diet of the common people in the Italian peninsula. 

Importance of Imports

It's honestly hard to say at this point whether imported food significantly impacted the diet of the average Roman.  With this palaeodiet study for my dissertation, I learned that there is a great deal of variation in diet within the Italian peninsula - some of this may be due to imported foodstuffs, particularly grain from Africa, and some of it may be due to the particular composition of the population in terms of age, sex, and ethnicity or original homeland.  If the average Roman was eating a significant amount of imported foods, this could also affect the person's strontium and oxygen isotope ratios, those measurements that tell us whether someone was born in Rome or grew up elsewhere.

Roman skeletons have a lot to teach us and can help us answer questions about diet, immigration, and disease in the Empire.  Much more stable isotope research is needed on populations from Rome and Imperial Italy, in addition to zooarchaeological and palaeobotanical studies, in order to more fully understand the variety of natural resources available for both human and animal consumption in this important historical time period.  This research needs to be integrated with historical literature and archaeological finds as well, in order to present the fullest view possible of what life was like in Rome - were people eating food they sourced locally, or were they importing exotic or low-cost foodstuffs?  The answer is probably that they were doing both - but the task for us bioarchaeologists is to figure out along what lines the diet varied.

C. Beltrame, D. Gaddi, & S. Parizzi (2011). A presumed hydraulic apparatus for the transport of live fish, found on the Roman wreck at Grado, Italy International Journal of Nautical Archaeology : 10.1111/j.1095-9270.2011.00317.x

Additional References:

• Craig O, Biazzo M, O’Connell T, Garnsey P, Martinez-Labarga C, Lelli R, Salvadei L, et al. 2009. Stable isotopic evidence for diet at the Imperial Roman coastal site of Velia (1st and 2nd centuries AD) in southern Italy. American Journal of Physical Anthropology 139(4):572–583.
• Garnsey P. 1988. Famine and food supply in the Graeco-Roman world. Cambridge: Cambridge University Press.
• Katzenberg M. 2008. Stable isotope analysis: a tool for studying past diet, demography, and life history. In: Katzenberg M, Saunders S, editors. Biological anthropology of the human skeleton. New York: Wiley-Liss. p 413–442.
• Killgrove K. 2010.  Migration and mobility in Imperial Rome.  PhD dissertation, University of North Carolina at Chapel Hill.
• Prowse T. 2001. Isotopic and dental evidence for diet from the necropolis of Isola Sacra (1st-3rd centuries AD), Italy. PhD dissertation, McMaster University.
• Prowse T, Schwarcz H, Saunders S, Macchiarelli R, Bondioli L. 2004. Isotopic paleodiet studies of skeletons from the Imperial Roman-age cemetery of Isola Sacra, Rome, Italy. Journal of Archaeological Science 31:259–272.
• Prowse T, Schwarcz H, Saunders S, Macchiarelli R, Bondioli L. 2005. Isotopic evidence for age-related variation in diet from Isola Sacra, Italy. American Journal of Physical Anthropology 128:2–13.
• Purcell N. 1995. Eating fish: the paradoxes of seafood. In: Wilkins J, Harvey D, Dobson M, editors. Food in antiquity. Exeter: University of Exeter Press. p 132–149.
• Rutgers L, van Strydonck M, Boudin M, van der Linde C. 2009. Stable isotope data from the early Christian catacombs of ancient Rome: new insights into the dietary habits of Rome’s early Christians. Journal of Archaeological Science 36(5):1127–1134.

Posted in: Archaeology, Diet, Food, Isotope Analysis, Physical Anthropology, Romans