Interpreting Ceramic Assemblages from the Nemea Valley Archaeological Project

Chris Cloke concludes his three-part series today on patterns of settlement and land use in the Nemea Valley.  If you missed the first two, start by reading Part 1 and Part 2Part 1 defines “site” and “off-site” (or “tract”) in terms of NVAP procedure.

In today’s final post on the Nemea Valley Archaeological Project (NVAP), I’ll be looking at a final component of the finds – the variety of the types of ceramic vessels present in the sample – to assess what this can tell us about the differences between finds from documented sites and those from survey tracts.

A fuller picture begins to emerge when we look at the functional classes of pottery making up these assemblages. For the Classical period, as well as the Archaic and Hellenistic periods preceding and following, there is a clear and consistent functional shift between on-site and off-site ceramic finds:


While less well-represented functional classes – food preparation (various cooking pots) and table service (pouring and serving vessels) – are proportionally similar on and off sites (tracts), there is an inverse relationship between drinking and dining vessels (cups, plates, bowls, etc.) and household storage and utilitarian vessels (pithoi, jars, basins, mortaria, etc.). For the Archaic to Hellenistic periods, drinking and dining vessels made up easily the largest part (over 1/3) of site finds, while among off-site (tract) finds of the same periods this percentage sharply declines and is overtaken by a majority of household storage and utilitarian vessels (also over 1/3). Together with a higher percentage of amphoras among tract finds than at documented sites (a reversal of what one might expect), the predominance of household storage and utilitarian wares not only explains the heavier average weight of off-site sherds, but more importantly suggests that off-site material in these periods was not simply comprised of refuse from sites the survey located (which would produce comparable proportions of vessel types). Rather, this off-site tract material represents different types of activities going on away from places designated as “sites.”

In short, it seems that the tract finds for these periods are picking up traces of rural storage and domestic activities, indicative that many small rural foci of activity may have gone unrecognized by the survey during fieldwork. The idea that there are perhaps many unknown “sites” in the survey area is compatible with a picture of the Archaic, Classical and Hellenistic countryside(s) characterized by dispersed small farms.

The Late Roman pottery, broken down by function, looks rather different:


The first notable trend is that the off-site (tract) material contains a high percentage of undetermined or unknown sherds (those too fragmentary to make confident identification of their original shape possible), due in large part to the many small, fragmentary sherds belonging to this period (the very thing we should expect from manuring-derived scatters).

Focusing on identifiable vessels, however, the basic proportions among on-site and off-site finds appear very similar:


The correlation of site and tract pottery in terms of function suggests that both “assemblages” of sherds recovered are products of a similar range of activities. Although we should also expect that the weights of sherds found on sites and those found in tracts would be similar, they are not; tract finds are less than half as heavy, on average, as site finds. That the on- and off-site finds of this period are, as a whole, similar in function, yet different in weight suggests the tract finds were derived from sites but underwent different depositional and/or post-depositional processes to get to where they were found. Given also that this material is found over a wide area and that densities (in terms of sherds per hectare) remain low even in the face of a spike in total sherd count during the Late Roman period, manuring seems a strong explanatory model for the observed patterns.

Granted, this type of analysis ideally must also take into account geomorphology and formation processes I have not discussed here, but by-and-large, I believe there are important differences in the artifacts themselves that speak to changes in farming methods and land use around Nemea.

The general pattern observed here is that, in pre-Roman periods “tracts” are concealing numerous small rural areas of activity, while in the Roman period, the “tract” material seems to have been derived from “sites” themselves. Although documented “sites” in the Early to Late Roman periods were few, landowners seem to have been intensive with their farming, manuring to gain better yields, and in the process seeding the fields with small bits of broken pottery and other refuse. There seem, generally, to be fewer and larger activity areas in the Roman period, indicating control of larger plots of land by fewer individuals or organizations centered in towns and at larger rural villas. Moreover, agents in just such a system could enable manuring more easily through mobilization of labor (something that was always a problem in the Archaic, Classical, and Hellenistic periods). At the same time, changing policies of taxation, demanding, by the Late Roman period, payment in crops themselves encouraged farmers to get the most out of available land. Manuring and other practices aimed at increasing crop yields were therefore advantageous in such times.

While not every survey will be able to do something similar to this with data already collected (weighing sherds individually is not always standard practice, and is very time-consuming), David tells me that the Pyla-Koutsopetria Archaeological Project (PKAP) has recorded finds in such a way as to enable this type of analysis (although the Eastern Korinthia Archaeological Project, EKAS, has not). It will be interesting to see if similar or different patterns emerge in different locales. Manuring was not a strategy employed in all places at all times. Its implementation and success depended on local agrarian traditions, the soils and climates characterizing a locale, the types of crops being grown and their overall volume, and the ready availability of manure and other compostable waste (animal manure, for instance, was not easy to obtain in places where stock raising or transhumant pastoralism were not practiced).

I look forward to hearing David’s and others’ responses to this short case-study presented here this week. I hope also that there will be many future opportunities to synthesize and compare NVAP data with those of the Eastern Korinthia Archaeological Survey (EKAS) and other regional projects for the sake of enriching the picture of the Corinthia and its environs.


Categories: Agriculture, Archaeological Survey, Ceramics, Economy, Landscape, Southern Corinthia, Periods, Nemea, Territory


  1. nakassis

    This is great work. It’s wonderful that the NVAP survey material is getting out there, and even better that it is in such capable hands!

  2. Chris,

    This is really interesting. I wonder, though, whether the differences in collection method between on-site and off-site could be influencing (although certainly not explaining away) some of the differences.

    For example, if on-sight collections produced tons of “diagnostic” pottery, I wonder whether there was a greater tendency to overlook more “non-diagnostic” material than in the tract survey where every sherd might be the only evidence for human activities. Again, I am not suggesting that this would explain away these trends, but it might have shaped the results – particular the proportion of finewares to utility wares and the overall size of artifacts.

    Just a thought… great work. I’ll look at our Late Roman material from Pyla-Koutsopetria over the next few days to see if similar trends occur (although it is much more difficult to identify on-sight and off-site at PKAP:


  3. Chris Cloke


    Thanks for the comments and kind words – and good points all.

    I think, however, that the way NVAP did “on-site” collection actually would have had the opposite effect- by making very thorough collections in grid squares or by doing grabs of everything in a certain radius along transects, they were picking up all material within certain areas of these sites. Tract collection, on the other hand, tends to favor what can be seen at a glance, and so is less likely to catch every little scrap of body sherd than what they did on sites.

    While siteless survey has certainly become the norm and gets rid of hard and fast archaeological constructs like the “site,” it’s interesting that the sort of analysis done here can really only be done using data from a survey that clearly distinguished “sites” from “not sites.”

    It’ll be really interesting to see what sorts of trends emerge from PKAP data!


    • nakassis

      Chris, I don’t think it’s true that “the sort of analysis done here can really only be done using data from a survey that clearly distinguished “sites” from “not sites.”” A siteless survey would in my opinion be a better candidate for this analysis, because the collection method would be uniform across the landscape. It would just then be a matter of defining ‘sites’ in the lab (which is a far more secure procedure than the usual method of defining sites in the field, in my view) and then running the analysis. This is essentially what I tried to do in the JMA article that Bill, David and I published back in 2006 (

      • Chris Cloke


        Very good point- I certainly agree that siteless survey has numerous advantages (eliminating the pressure to define sites on the fly being one of the greatest). What I should have said to be clearer was simply that “sites” do at some point need to be distinguished – something which can be done, as you point out, after fieldwork. Once that’s done it’s a simple matter of approaching site and off-site material as distinct bodies of evidence to be compared and contrasted. EKAS’ methods of site definition make it a great candidate for doing comparative studies, although the impact of differences in sampling method on number-crunching weights will need to be considered.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s