The Millet-Eaters of the Roman Empire

Just a few days ago, only the second isotope study of millet consumption in the Roman Empire was published, by Pollard and colleagues in the American Journal of Physical Anthropology.  In a small Romano-British cemetery in Kent (late 3rd-early 4th century AD), a salvage archaeology project uncovered a dozen burials that were simple in nature: only coffin nails and hobnails from boots were found in most graves.  Among these simple farmers, though, was an individual with a surprisingly high carbon isotope value, so Pollard and colleagues undertook a dietary (C/N) and migration (Sr/O) study of the individuals.

The anomalous partially complete skeleton was that of a male over the age of 45 buried wearing hobnail boots. The individual's nitrogen isotope ratio was a bit high (11.2 permil), indicating aquatic resource consumption, but was not higher than average for Roman Britain.  His carbon isotope ratio from collagen, however, came in at -15.2 permil, in stark comparison to the average of the other individuals of -19.8 permil (see below).  This difference may not seem dramatic until you factor in the standard deviation - variation within the d13C ratios of the others from the site was only 0.3!  This person was therefore eating a whole bunch of C4 resources - millet, sorghum, or animals foddered on those grains.

Figure 3 from Pollard et al. showing the anomalous individual (SK12671)
compared with other Romano-British sites and the two anomalous individuals
published in Muldner et al. 2011.

Evidence of C4 plant consumption is surprisingly absent from the archaeological record of the Roman world, even though authors like Pliny note that millet and beans were frequently eaten together by people in rural Italy.  As far as I know, only one bioarchaeological study has been done on skeletons from Italy looking at C4 resource use (Tafuri et al. 2009).  Researchers found evidence of millet consumption in the elevated d13C ratios of people from northern Italy in the Bronze Age compared with people in southern Italy.  Another Romano-British cemetery yielded two individuals with a mixed C3-C4 diet, where carbon isotope values ranged from -16.8 permil to -15.8 permil.  So this new person from Kent provides the highest d13C ratio obtained so far from bone collagen in the Roman period.  Below is a graph of the Bronze Age millet-eaters and the Romano-British people from Pollard and colleagues' study:

Figure 4 from Pollard et al. 2011 comparing Romano-British
samples with Bronze Age north Italian samples

Curiously, Pollard and colleagues didn't look at carbon values from bone apatite, but they did look at the carbon isotope ratio of the dental apatite, which in this individual was -7.2 permil, also significantly higher than the values from the others, which range from -13.8 to -11.5 permil.  This likely means that his C4 resource use was in the form of direct consumption of millet rather than from consuming protein from animals that were foddered on millet.

Finally, they investigated the individual's strontium and oxygen isotope ratios to see if he perhaps immigrated to Britain from an area with more evidence of millet production and consumption, like Italy.  This is where the paper gets interesting - the strontium ratio is .708826 and the oxygen (from carbonate) is 26.1 permil.  These values are within the range of expectation for someone from southern Britain, so the authors could not rule out a local origin for the man.  However, my dissertation work (Killgrove 2010) showed that those values are equally likely to occur in or near Rome - my local strontium range for Rome is .7079-.7102, and the local oxygen range (drawn from Prowse et al. 2007) is 24.9-27.1 permil.  Pollard and colleagues suggest that this man may have come from northern Italy, where growing millet was common, but I am not convinced because his strontium isotope ratio of .7088 is far too low for the older geology of northern Italy, unless he was located near the east coast (and then his oxygen ratio should be lower).  Rome itself is around .7090, and .7088 - if we assume a western Italian origin - is more like Naples.  Granted, it is extraordinarily difficult to pinpoint homeland, and part of this article addresses the problems with identifying immigrants through just Sr and O isotope analyses.  As I have started to write up my Sr/O study for publication, it's something I'm keeping in mind.  Interestingly, the authors suggest that the inclusion of hobnailed boots in this man's burial may signify that he was "walking back" from Britain to his true homeland.

But the publication of this article - in AJPA no less - makes me excited because I'm sending off my C/N isotope article tomorrow to the Journal of Archaeological Science.  And in that article, I have a section on individual ET20, a male in his 30s from the site of Castellaccio Europarco, in the Roman suburbs.  ET20 has an astoundingly high d13C ratio: -12.5 permil.  This is on par with the isotope ratio of millet itself, and carbon ratios this high tend only to be found in populations that ate maize (corn).  However, the d13C ratio from ET20's bone apatite is only -8.6 permil, which is not dramatically higher than the rest of the population, suggesting that this individual was consuming his C4 resources in the form of animals who were foddered on millet.  His d15N ratio is 8.3 permil, which is a bit lower than expected from the population, so perhaps he was eating beans along with his millet or millet-fed animals like Pliny suggests.  I did do Sr/O on this individual, and they came back at .709631 and 25.3 permil, respectively.  Both of these are within my admittedly broad "local" range of Rome, but no one else among the locals has such a high d13C ratio.  I suggest in my dissertation (Killgrove 2010) that he may have come from northern Italy - the strontium ratio is higher than expected from Rome, indicating a childhood spent on slightly older geology.  I also found with ET20 that his d13C ratio from enamel apatite was -4.0 permil - so he changed his diet between the time he was born and the time he died at Rome.  Here's a quick graph from the forthcoming paper showing just how far to the right (C4 use) ET20 is in comparison with others from Castellaccio and Casal Bertone (compare with the graphs above, where no one reaches the high carbon value that ET20 does):

From Killgrove & Tykot, n.d.

At any rate, the person that Pollard and colleagues found (and the two people found by Muldner et al. 2011) show that we have a lot left to learn about C4 resource use in the Roman Empire.  Millet may have been considered a substandard grain by many authors, the kind of food that rural or poor people eat, but there is growing evidence that many people were consuming at least a C3-C4 mixed diet and several people were eating quite a bit of millet or animals foddered on the grain.  Isotopes are letting us tease out differences in diet at the levels of the individual and the population - especially in Rome, as I've blogged about before here.  Although the overall diet mostly tracks with historical and artistic records from the Roman world, the diversity in the lower-class diet is surprising and intriguing, and I think will eventually be able to tell us more about things like status.  Watch this space for more on the diet of my Romans as I work through the process of submitting and revising my C/N isotope article this week!


Killgrove, K. (2010).  Migration and mobility in Imperial Rome.  PhD dissertation, University of North Carolina at Chapel Hill. [PDF]

Killgrove, K. & Tykot, R. (n.d.) Investigating the diets of the lower classes in Imperial Rome through carbon and nitrogen isotope analyses.  Manuscript in submission.

Muldner, G, Chenery, C, & Eckardt, H (2011). The "headless Romans": multi-isotope investigations of an unusual burial ground from Roman Britain Journal of Archaeological Science, 38, 280-290

Pollard AM, Ditchfield P, McCullagh JS, Allen TG, Gibson M, Boston C, Clough S, Marquez-Grant N, & Nicholson RA (2011). "These boots were made for walking": The isotopic analysis of a C4 Roman inhumation from Gravesend, Kent, UK. American Journal of Physical Anthropology PMID: 21959970

Prowse TL, Schwarcz HP, Garnsey P, Knyf M, Macchiarelli R, & Bondioli L (2007). Isotopic evidence for age-related immigration to imperial Rome. American Journal of Physical Anthropology, 132 (4), 510-9 PMID: 17205550

Tafuri MA, Craig OE, & Canci A (2009). Stable isotope evidence for the consumption of millet and other plants in Bronze Age Italy. American Journal of Physical Anthropology, 139 (2), 146-53 PMID: 19051259


Popular Posts