The South Basilica at Polis

We’ve officially renamed the EF2 basilica. It is now “the South Basilica” and we’ve officially moved from pondering ceramics and architecture on the ground to writing. 

And this is our first effort to bring together the results of over three years of study (and many years of field work by many people before then)!

We’ve been working on lovely new illustrations of the site plan and our basilica, and they’re almost presentable now:

Figure 1 WRC3


And here’s our most recent text:

Analyzing Residual Ceramics in a Fill Deposit

I know this isn’t one of my best titles ever, but it describes what I am trying to do pretty accurately. As we move toward a comprehensive preliminary study of the basilica in area EF2 at the site of Polis-Chrysochous on Cyprus, we want to be able to say something about some of the ceramic assemblages associated with the buildings. In particular, we have a massive fill level that we have tentatively associated with the second phase in the construction of the basilica.

The fill is dated by both coins and ceramics to the middle decades of the 7th century an consisted over 3000 pieces of broken pottery. Because the excavators did not save all pottery and tended to discard less apparently diagnostic artifacts, the fill consisted over over 40% fine ware (compared to 9% transport amphora sherds, 32% utility wares (primarily storage vessels and other less diagnostic coarse and medium coarse fabric sherds), and 15% kitchen wares). While it may be possible to reconstruct what they discarded into the pottery dump, for now we think that the assemblage is more or less representative of what was excavated.

imageCRS Sherds in the R09 Fill at Polis

The majority of fineware in this assemblage is Cypriot Red Slip (or, what we maybe should call Late Roman D Ware). The chronology of the assemblage represents almost the full chronological range of CRS production which began some time in the 4th century and perhaps continued as late as the 8th. Like many significant assemblages of CRS, CRS9 and its variants (largely identified by Henryk Meyza and his work at nearby Paphos) make up the largest percentage of our material. The particularly production long life of CRS9 (beginning in 400 and continuing with some variation to almost the end of the 7th century) might account for its preponderance in our assemblage.

Unlike many other sites on the island the Polis R09 Fill had an impressive quantity of CRS8 an CRS11 sherds.


These likely date to the final century (or 150 years) of CRS production. The size of CRS11 vessels and the distinct folded-over shape of the rim of many of the CRS11 basins found at our site may suggest a local production center, although the rim of the vessel shown above is similar and comes from Anemurium in Asia Minor.

Our percentages are similar to those produced by the nearby Canadian Palaiopaphos Survey Project, but they identified fewer CRS11 sherds.

imageCPSP CRS Assemblage

A nearly contemporary assemblage associated with two basilica churches excavated by Marcus Rautman at Kopetra on Cyprus also produced a similar distribution of forms.

imageCRS from Kopetra

The main difference between the CSPS and Kopetra assemblages are that they probably represent a wide range of depositional processes from discard to (perhaps) use. Our assemblage from trench R09 at Polis is probably all secondary discard and two steps removed from its primary use context. It was deposited as a single event, but the material likely derived from a range of domestic discards.

UPDATE: One more example, here are the different forms documented by the Pyla-Koutsopetria Archaeological Project.


End of the Season Varia

At the end of a study season, I’m always left with various things that I’m all excited about, but I don’t have sorted out for a blog post.

For example, I usually have a satisfying photograph of the end from the last day at the apotheke (or storeroom) like this:

IMG 0663

The table is empty, and that’s a good thing, because it was usually filled with pottery under study or being catalogued.

I also have a few photographs that try to capture the range of activities during the season in a single shot. So, I have a photograph that shows off Brandon Olson’s illustrations of Late Roman fine ware like this. Scott and Brandon are looking at two chronologically contemporary, but physically distinct areas of the excavation for joins.


And I since we did some 3D modeling of parts of the basilica using Agisoft Photoscan, I invariably have some cool screen shots like those below. The first one is southwest corner of the narthex. If you look carefully you can see the lines of the original arched opening which was latter walled up with less well-sorted (and weight bearing) rubble walls.

Screenshot 062513 031844 PM

Here is another showing the buttresses in the north aisle of the church. You can see clearly how the eastern apses do not bond with the main wall of the north aisle. 

Screenshot 062513 032224 PM

One of the most useful things about modeling architecture using Agisoft is that we can show parts of the basilica at almost impossible angles without having to get a crane and reshooting photographs.

I also have a little gaggle of photographs that I like, but don’t really know what to do with. So I have this one of the “super moon” over the plataea of Polis at night. I like it because it looks a bit like a painting.

IMG 0660

Then I always have ridiculously beautiful scenes like this:

IMG 0637

Or like this:

IMG 0638

I don’t recall whether these two photos appeared on the blog.

It’s ok if you find this kind of thing empty and self-indulgent. I promise that I’ll  get back to more substantive blog posts over the next couple of days. I have some writing and thinking time in Larnaka before I head to Greece to check out an area where I hope to do some fieldwork next year.

Discovery in Pieces

Yesterday was one of my best days ever as an archaeologist. I didn’t discover some amazing new site. And I didn’t find some amazing or valuable object.  And my day didn’t involve out witting Nazis or mummies (or zombies).

Over the last few weeks, we have worked to establish dates and architectural relationships for the various parts of the basilica at EF2 at Polis. We figured out that the narthex and the south portico of the church were added to the building most likely date to the middle years of the 7th century. We found particular types of late Cypriot Red Slip pottery in a foundation trench associated with the narthex, and we know that the south portico – a long porch build along the south side of the church – had to date to after (or the same time as) the narthex. The south portico appears to have been cut into a massive rubble fill the extended north from the south wall of the church. (We’ve tentatively argued that this rubble fill was a response to a local drainage issue.

The other phase of the church that we’ve been studying involves the building of  buttresses along the walls along the walls of the main nave. The ceramics from deposits associated with these piers also date to the 7th century, but we had not been able to associate their construction with the building of the narthex and south portico. Until yesterday.

Yesterday, Scott Moore discovered a join between two pieces of a stamped Cypriot Red Slip plate. One piece came from the massive rubble course south of the basilica that narrowly pre-dates the building of the portico. The other piece of the same plate came from a foundation deposit associated with the construction of a buttress on the north wall of the nave. This sherd, then, connects the two major changes to the basilica: the addition of a narthex and portico and the reconstruction of the main nave.


The two pieces of pottery came from trenches about 15 m apart. The trench to the south of the south portico was excavated in 1985 and the trench in the nave in 1990. One sherd was inventoried as a find (which means that it stood out as something with intrinsic value) and one sherd languished in the boxes of context pottery. (We had to look through over 20,000 artifacts in the context pottery boxes to find this little guy!).

Bringing together the context pottery and the inventoried finds, two different episodes of excavation, and the narthex, portico, and buttresses of the main nave has unlocked the chronology of the church. It was a pretty good day.

Residuality and Fill Levels at the EF2 Basilica at Polis

One of the challenges that we faced working at the EF2 basilica at Polis-Chrysochous on Cyprus was that so much of the ceramic evidence came from various kinds of fill levels at the site. In other words, we had very little material from traditional use contexts and an unbelievable quantity of pottery associated with either construction deposits or the massive rubble fill level that extends south from the basilica. While analysis of ceramics from areas of such relatively undifferentiated contexts has not always been the rule on archaeological project, recently decades have shown how the study of material in these fills levels can produce high-resolution snap shots of the certain components of a communities material culture.

R09FillNote the “Rubble Layer” in the scarp drawing

As we looked at the pottery from these levels we began to think about how to approach assemblages of ceramic artifacts produced by activities completely unrelated to the original to the original purposes of the objects. The artifacts present in the leveling and construction fills, for example, represent past activities at the site, habits of discard, and construction practices. They also provide chronological “type fossils” that allow us to date architectural features associated with the levels. 

Gavin Lucas in his new book, Understanding the Archaeological Record, puts it nicely:

“If we think about the archaeological record in terms of the residuals of assemblages, we must consider such residues as possessing the memory of the assemblage itself, insofar as the organization of the residue captures, however faintly, the organization of the parent. It is the residue of this organization that is being sought, not simply the elements or objects which were part of it.” (p. 211).

Over the last couple of weeks, we’ve worked out way through the material in the fills that extend south from the basilica and have paid particular attention to the very common Cypriot fine ware (or table ware) called Cypriot Red Slip as well as contemporary imported pottery (see my post yesterday for more on this). Our intent is to analyze the residual ceramics in these fills much like we’d interpret survey data. In fact, we intend to compare the assemblage produced by this fill with assemblages from both similar contexts (especially those associated with the nearly contemporary basilicas at the site of Kalavassos-Kopetra) as well as the results of survey work in the larger Paphos and Polis area.

Our goal is to be able to speak to and from the architecture as well as the assemblage in our analysis of activity in the area of EF2 at Polis.

More Pottery, More Problems

About two weeks ago, I was feeling pretty good about the date our our basilica at the site of Polis. We dated the church on the basis of five or six fairly secure deposits associated with the construction or modification of the church. The pottery in these contexts is largely the locally(ish) produced fine ware, Cypriot Red Slip.

The more pottery we have, however, the more problems it creates. And here’s how it goes.

First, we have to identify the major wares present and the make an effort to distinguish the different shapes. That often means spending hours looking at sheets of rim profiles and reading fiddly descriptions of fabric. Because these pots were not made on a production line, any sherd we find does not really line up precisely with the object in our books so we have to wiggle it to fit a category (and, moreover, the potters were not sitting around discussing how to produce Cypriot Red Slip Form 9!). It’s like getting some kind of polyhedron to pass through a round or square hole in a child’s game.

Screenshot 6 19 13 3 22 PMScreenshot 6 19 13 3 25 PM

Then, once we are satisfied that we have fit our sherd into the typology, we can begin attempting to date our shapes on the basis of stratified examples of these vessels elsewhere. Most scholars who contribute to the typologies we use to identify the sherds also make an effort to date the pottery. Unfortunately, the bewildering array of shapes and sub-types can devolve into equally bewildering chronological arguments. I had a bit of a “down-melt” this morning when confronted with several possible for a type ranging from 580/600-700 to early 5th to 7th century. That’s a big difference and 580/600 is not a secure date but TWO different dates separated by a slash. In terms of normal humans living in  normal time, this is meaningless. I was not born in 1972/1988. 

IMG 0598

Finally, once we get some dates on some pots, we have to reconcile the chronologies of various vessels within the deposit with one another. This always involves dating the deposit to after the date of the more recent object. Once we have the terminus post quem (that the date after which) for the deposit, we can begin to attempt to understand how earlier material made its way into the collection of pot sherds deposited as a single event. Since most of our deposits are associated with the construction of the basilica, it is easy enough to understand the various earlier sherds as being part of the debris used to backfill a foundation trench or pack a floor. In fact, from a use standpoint the latest and earliest sherd in the deposit functioned essentially the same way. They were all residual and probably all cast aside some time earlier in either a dump or in some kind of local destruction.

The problem is, of course, the more pottery there is, the more complicated the chronological relationships are. For each deposit, we have to sort out both the very local chronology of material, but also the relationship between it and others at our site which may not have the same types (or sub-types) or pottery, but may have a similar date. As a great man once said, mo’ pottery, mo’ problems

The PKAP-Polis Project’s Season of Work

This weekend is the annual CAARI (Cyprus-American Archaeological Research Institute) Workshop. This meeting attracts archaeologists from all over the Republic of Cyprus to present their work often as their field or study seasons are underway. At its best, it is a great way to catch up with both old friends and professional news. 

Typically, my project, the Pyla-Koutsopetria Archaeological Project presents some of their research, but, alas, this year, we were not invited to participate. Rumor has it that we were not invited because we made all the other projects look bad, and this was bad for morale in the archaeological community. Apparently our reports on what we had accomplished on the island in such a short time brought some very important senior archaeologists to tears at the relative insignificance of their own achievements

Despite the situation being as it is, Scott Moore and I have opted to soldier on. Instead of leaving the archaeological community in awe of our achievements through a direct presentation of our genius, we have decided to contribute a brief report on our work at Polis in their larger work. 

We hope that it will be seen as sufficiently modest to get invited back to the CAARI workshop again in the future:

A Brief Report on the PKAP-Polis Team’s 2012 and 2013 Work

Over the course of the 2012 and 2013 season, we have continued to study the stratigraphy, architecture, and finds from the Christian basilica style church in E.F2. To facilitate this work, we have prepared a comprehensive GIS-based site plan of the church, transcribed close to 40 excavation notebooks from the area, and created an relational database integrating digitized notebooks, analyzed context pottery, and registered finds. These tools and the study of over 20,000 artifacts from a fills, collapse, discard areas, and use levels has allowed us to begin the process of dating the major phases of this basilica and locating it in the history of the busy area of EF2.

The most immediate significance of this work is that we can now date the basilica’s construction to the 6th century AD with substantial modification over the next century including the addition of a narthex and south portico and its transformation from a wood-roofed to barrel-vaulted church. The ceramic assemblages associated with the various construction phases contained a wide range of well-attested pottery in the southwestern Cyprus including local fine wares (Cypriot Red Slip) and imports (African Red Slips and Phocaean Wares), Late Roman Amphoras, and various Late Roman kitchen and cooking wares. It is worth noting that this assemblage is rather distinct from assemblages along the south and eastern sides of the island which feature far more imported fine wares and more numerous LR1 amphoras than we have currently recognized at Polis.

While our primary focus has been on the Late Antique and Early Byzantine levels at E.F2, we plan to expand our work to include the systematic study of the Hellenistic, Roman and later Medieval remains in this area. Our intial study of material related to these earlier periods in the area has revealed the existence of a well-defined 1st c. BC/1st c. AD horizon characterized by Cypriot Sigillata and imported Eastern Sigilatta A table wares and a range of cooking vessels in recognizable Roman fabrics. Amphora and utility wares are far less common with the exception of the ocassional example of John Leonard’s infamous “pinch-handled” amphoras.

In 2013, we also conducted a campaign of high resolution laser scanning of the area of EF2 collection over 50 million individual data points with a Leica ScanStation C10. The result of this work not only complemented the more fanciful 3D reconstructions accompanying the City of Gold exhibit, but also provided detailed visual support for the study of notebooks and ceramics. The laser scans will allow the research team to document architectural relationships during the offseason, to produce vertical elevations, and to supplement and revise the existing plans of the site and its buildings.

Water and our Basilica

While helping our ceramicist, Scott Moore, make his way through thousands of pot sherds excavated from the basilica at EF2 at Polis, I’ve been working on writing up some of the preliminary observations on the architecture of the church. Among the most interesting aspects of the the church, is how the builders managed water at the site. The position of the church perpendicular to the north slope of a hill exposed it to apparently significant flow of water. Moreover, the entire area of E.F2 seems to be riddled with well, drains, and water pipes suggesting that water management was more than just an issue for the builders of the church.

P1020936The exposed foundation wall of the south aisle of the basilica. The exposed walls running under the basilica are beneath the rubble drainage layer.

Here is what I penned in the gaps between batches of pottery of the last few days on issues of water and architecture at the basilica at EF2. It’s all provisional and a work in progress, but it’s what I’ve been thinking about the last couple of weeks here at Polis.

From as early as the Hellenistic period there is evidence for concerns about water at the area of E.F2. There are numerous wells in the area associated with the workshops to the south and west of the basilica in the Hellenistic period. The Roman period saw the construction of complex systems of water pipes associated with the paved roads and what appear to be settling basins and drains. White most of these features likely contributed to water supply for various industrial and domestic activities in the city of Polis, it is possible that they also served the important role of water management in the area of EF2. The location of EF2 on the slop of a hill likely exposed the site to the risk of season flooding especially in the event of torrential Mediterranean winter rains.

Several unusual features in the architecture of the the basilica appear designed to protect the foundations of the basilica from the flow of water south to north across the site. On the foundations, below the level of visible walls, a plaster lip protected ran along the roughly mortared foundation of field stones of both the eastern apse and the south side of the basilica. The plaster lip or rim was best preserved along the foundation of the eastern apse where it extended for approximately 15 cm. The purpose of this rim appears to be to prevent water from running down along the foundation through the less densely packed earth associated with the foundation cut. Elsewhere along the line of the foundation excavations revealed sections of foundation wall covered with moist green clay (S06.1991.8). In other places in EF2, similar clay was associated with roof fall, and the water proof character of this clay has led to its continued use to seal roofs even until relatively recent times (e.g. H10.1997.11,4 (vol 1., 55). It seems, then, that the builders of the basilica made an effort to seal the foundations of the church against both water run off from the roof of the building or the surface and the seepage of ground water.

The south side of the basilica saw a more substantial effort to manage the flow of water downslope in the area. The continued presence of a paved road along the upsloap, south side of the church and the probable existence of an open courtyard immediately to the south of the building exposed the southern foundation wall and the piers supporting the south portico to the corrosive effects of water run off. In an effort to counter this risk of water destabilizing the south foundations of the church, the builders designed the courtyard to act as a massive drain. Beneath a level of limey, packed earth which probably represented the ground surface of the courtyard, a loose level rubble which in some places exceeded a meter in depth may have functioned as a massive French drain designed to prevent water from pooling against the south wall of the church and running down running directly down the soft foundation cuts of for the walls. Instead, the porous character of the rubble level served to slow the flow of water south and perhaps even allow it to drain away prior to reaching the vulnerable south wall of the basilica.

The rubble layer is most likely contemporary with the first phase of the basilica and extends almost to the depth of the basilica foundation. Later burials have probably disturbed the integrity of the limey, packed, floor, but there nevertheless appears to be no pottery in the packing that is later than the 7th century with Cypriot Red Slip Form 9 being the latest present (in R09.1986.6,1-2). The massive leveling course of rubble below the floor packing was, in turn, cut by the foundation of the piers of the south portico. In levels associated with the foundations of the the south protico the latest material dates to between 600 and 700 and includes well-document Cypriot Red Slip Form 10. Below the level of the foundations, however, the material is slightly earlier, in general perhaps representing at late 6th to early 7th century date. This rubble level appears to sit immediately atop early Roman deposits dating to the 1st century BC to first AD and even earlier level of Hellenistic date. The diverse assemblage of fine wares, kitchen wares, and transport and utilities wares present in the massive rubble leveling course indicates that it was not only the product of a well-provisions and connected community, but that the rubble course was at least partially associated with discard from other locations in the community. 

P1020976Roman period water pipe.

Doing Work at Polis

For the third straight year, I’ve sequestered myself for a few weeks in the lovely village of Polis-Chrysochous to commune with the notebooks from the Princeton Polis Expedition. These notebooks detail the excavations at the site of E.F2 on the Princeton grid. This site dates from the Hellenistic to Medieval period and the most conspicuous feature is an Early Christian to Medieval basilica style church. This church and the great group of colleagues working at Polis drew me to the site initially.

Since 2009, I’ve been working on producing a database from the notebooks, assisting Scott Moore and Brandon Olson in analyzing the context pottery, and integrating their work with the notebooks and the existing registry of finds.  This means getting three databases to talk to each other. Two of the three – one designed to accommodate our notebooks and one designed to accommodate the new readings of the context pottery – meld together smoothly. The database accommodating the registered finds is a different matter. It was built over 20 years and is not normalized. It can only link to the other databases through a series of concordances. This is tedious stuff to develop and test.

The greatest challenge, however, is to understand the notebooks. Polis was one of the last large-scale Mediterranean excavations not to be excavated stratigraphically. Instead, excavators defined “Levels” which could be stratigraphic or simply spatial and then made “Passes” through these levels which could also be stratigraphic or simply spatial or just arbitrary. What I’ve tried to do is to superimpose a stratigraphic system on top of the existing system of levels and passes in order to understand the depositional processes that formed the archaeological record.

This is both a nightmare and a rush. Whereas some people love archaeology for the thrill of discovery, I have to admit to getting my rush in the problem solving aspects of the discipline. I love reconstructing the spatial relationships through the irregular lens of the Polis notebooks. This is a process of course. Here are the steps:

1. Read the notebooks and transcript the Level and Pass descriptions. Nothing works better than transcribing to study the details of excavation. This practice also allows me to organize the levels and passes which tended to appear almost randomly throughout the notebooks as the trench supervisors often had multiple contexts open at once.

0009Polis Notebook Page

2. Once we have the notebooks transcribed and analyzed, I build an informal pseudo-Harris Matrix (sometimes I call them a Franco Harris Matrix). I used Tuft University’s VUE program to attempt to illustrate the relationships between various levels. 

3. This allows me to identify sensitive contexts that might be able to inform architecture or activity areas. In most cases, we can simple identify a handful of contexts that must be earlier or later than each other. Inevitably numerous contexts are lost to contamination, irregular or obscure excavation decisions, or ambiguous depositional relationships.

4. The ceramics from these contexts are read in the museum by Scott Moore or Brandon Olson, and we draw in the registered finds (typically more distinct or diagnostic objects) from that database to produce a comprehensive dataset of the finds from the level and pass.

5. Finally, at the end of the year, I bring together the read pottery, the stratigraphy, the architecture, and the finds to try to make arguments for the history of the site. We’ve been targeting specific areas of E.F2 each summer and will hopefully have the entire basilica documented by the end of this field seasons. We’re really close.

SCottatWorkScott Moore watching a movie from Netflix
when he should be analyzing pottery.
(Actually, he’s looking up a form in a scanned pottery volume on his iPad.)
The photo is with Camera Noir on my iPhone 5.