5. W. Caraher, R. S. Moore, J.S. Noller, D. K.

Pettegrew, The PylaKoutsopetria Archaeological Project: Second Preliminary Report (2005-2006 Seasons), RDAC (2007).

The Pyla-Koutsopetria Archaeological Project: Second Preliminary Report (2005-2006 Seasons)*

William R. Caraher, R. Scott Moore, Jay S. Noller and David K. Pettegrew

INTRODUCTION
The Pyla-Koutsopetria Archaeological Project has now completed its fourth season of archaeological work in the coastal region of the village of Pyla. Our goal has been to document the extensive array of cultural material in the area and to ascertain the place of the site in both local and interregional social, economic, and political networks through time. Pyla-Koutsopetria is particularly significant because it is one of a handful of mid-sized sites larger than a rural village yet smaller than a polis centre to receive systematic investigation in Cyprus.1 It has recently been argued that such mid-sized sites formed an important part of the skeleton of ancient Mediterranean connectivity.2 By investigating the material signature of such a site, our goal has been to establish the economic, political, and cultural character of the places which formed vital links in inter- and intra-regional Mediterranean-wide trade networks. The site of Pyla-Koutsopetria is located about 10km. to the east of the centre of Larnaka, at the western fringe of the Dekeleia Cantonment. Maria Hadjicosti conducted excavations at Koutsopetria over two short campaigns in the 1990s, bringing to light parts of an Early Christian basilica.3 In 2004 and 2005, we conducted an intensive pedestrian survey of the area, fulfilling a call by John Leonard, among others, for a systematic investigation of this substantial coastal site.4 As we have described in some detail elsewhere,5 we surveyed the highest density area of the site using a 40 40m. grid system in which we sampled 20% of the surface area for density and collected

unique sherds using a recording system known as the Chronotype system. In order to identify the sites borders, we also laid a number of larger survey units over a broader area of low artefact densities. We contextualized our artefact densities through the comprehensive geological investigation of the area, including subsurface prospecting. And we conducted experiments to
____________ * The Pyla-Koutsopetria Archaeological Project thanks Director Dr Pavlos Flourentzos, Director of the Department of Antiquities, for the generous permission to work in the area, and deeply appreciates the cooperation of Dr Maria Hadjicosti, Dr Tom Davis at CAARI, and Mr Marinos Avraam and his staff at the Larnaka District Archaeological Museum. We have received funding and technical support from the Kress Foundation, INSTAP, ASOR, Indiana University of Pennsylvania, The University of North Dakota, The Ohio State University Excavations at Isthmia, Greece, and several private donors. An earlier version of this article was given at the 24th Annual CAARI Archaeological Workshop, June 2006 (cf. W.R. Caraher, R.S. Moore and D.K. Pettegrew, The PylaKoutsopetria Archaeological Project: A Third Preliminary Report, paper delivered at the 24th Annual CAARI Archaeological Workshop, Nicosia, Cyprus, June 2006). 1. The best known mid-sized site on Cyprus is Pegeia-Ag. Georgios. Cf. C. Bakirtzis, The Role of Cyprus in the Grain Supply of Constantinople in the Early Christian Period in V. Karageorghis and D. Michaelides (eds), Proceedings of the International Symposium, Cyprus and the Sea (Nicosia 1995), 247-53; J. Leonard, Roman Cyprus: Harbors, Hinterlands, and Hidden Powers (Unpublished Ph.D. Dissertation, State University of New York at Buffalo) (2005), 614-18. 2. P. Horden and N. Purcell, The Corrupting Sea (Oxford 2000). 3. S. Hadjisavvas, Annual Report of the Department of Antiquities, Cyprus (Nicosia 1993), 70-72; idem, Chronique des fouilles et dcouvertes archologiques Chypre en 1999, BCH 124 (2000), 693. 4. J. Leonard 2005, 431. 5. For a discussion of our methods, cf. W. Caraher, R.S. Moore, J.S. Noller and D.K. Pettegrew, The Pyla-Koutsopetria Archaeological Project: First Preliminary Report (2003-2004 Seasons) in RDAC 2005, 250-56.

WILLIAM R. CARAHER, R. SCOTT MOORE, JAY S. NOLLER, DAVID K. PETTEGREW

reveal the limits of our artefact sampling strategy, including the resurvey of twenty grid squares using a more intensive sampling component total collection of all the artefacts found on the surface of the soil using hands-and-knees artefact counts. This report will present the preliminary findings from this survey, providing a quantitative analysis of the material recovered, an evaluation of the archaeological experiments, and a summary discussion of the ceramic finds. The complete publication of finds and geological analysis is currently under preparation.

THE SITE IN CONTEXT The site at Koutsopetria consists of a robust scatter of cultural material on the narrow coastal plain at the base of a series of prominent coastal heights. The western limit of the site is below the height of Vigla and its eastern boundary below the height of Pyla-Kokkinokremmos with its wellknown Late Bronze Age fortification (Fig. 1).6 The highest-density scatter of ceramic material is below the ridge of Vigla and covers over 10ha. A second area of moderate artefact density continues for some 30ha. below the slopes of Kokkinokremmos. This latter area is bounded to the south by a stretch of coastal plain almost entirely devoid of cultural material that suggests an infilled harbour of ancient date (see below).

Over the course of four seasons of archaeological fieldwork in the Koutsopetria region, our teams surveyed 252 40 40m. grid squares and 49 larger units, covering a total area of 63.6ha. In the course of survey, we counted over 20,000 artefacts and collected, read, and recorded around 8,500 artefacts using the Chronotype system. Additionally, we collected another 3,000 artefacts using alternate survey techniques such as total collection circles (see below), bringing the total number of artefacts analyzed and collected to 11,500 with a weight of nearly a half a ton. Of the artefacts collected from the site, we have inventoried over a thousand with photographs, drawings, and basic catalogue descriptions. In addition to the ceramic artefacts, we have also documented over 430 worked stone features found throughout the survey areas including an array of agricultural processing equipment, walls, local cut stone, elaborately carved gypsum, and marble fragments.

EXPERIMENTAL DATA PKAP was committed from the start to a lowimpact survey method as is increasingly common for survey projects in the eastern Mediterranean.7 Low-impact surveys, in contrast to the methods typically employed in total coverage surveys, have sought to collect as much data as possible from an area while preserving in situ as much of the archaeological record as possible. PKAP embraced this goal by using a two-tiered sampling strategy.8 A team of four field walkers

____________ 6. V. Karageorghis and M. Demas, Pyla Kokkinokremos: A Late 13th Century Fortified Settlement in Cyprus (Nicosia 1984). 7. For a recent discussion of this concept, cf. T.E. Gregory, Less is Better: The Quality of Ceramic Evidence from Archaeological Survey and the Practical Proposals for LowImpact Survey in a Mediterranean Context in E. Athanassopoulos and L. Wandsnider (eds), Mediterranean Archaeological Landscapes: Current Issues (Philadelphia 2004), 1536. 8. Caraher et al. 2005, 250-56. For the most recent discussion of the Chronotype system, cf. R.S. Moore, A Decade Later: the Chronotype System Revisited in W.R. Caraher, L.J. Hall and

Fig. 1. Survey units and local topograpy for the Pyla-Koutsopetria Archaeological Project (W.R. Caraher).

THE PYLAKOUTSOPETRIA ARCHAEOLOGICAL PROJECT

spaced at 10m. intervals and examining twometre wide swaths counted artefacts visible on 20% of the surface of the unit. As they walked their swath of the unit, each walker collected one example of each unique type of artefact present. This sampling method, called the Chronotype system, ensured that we collected a sample of all objects present in our 20% sample of the unit. This minimalist collection strategy produced a collected assemblage from each unit that was around 65% of the total number of artefacts counted or less than 15% of the estimated total number of artefacts in a unit (an estimate based on our 20% sample). This method produced a record of the types of artefacts present in a unit, while leaving over 85% of the material on the ground in situ. In recent years, Mediterranean survey archaeology has become increasingly reflexive, assessing the interaction between method, data production, interpretation, and conclusion.9 In light of these current issues, PKAP sought to understand how our sampling and collection strategy influenced our analysis and interpretation of the archaeological material. We conducted research in two areas. First, we implemented a reflexive study of the data collected through identical field methods elsewhere in the Mediterranean, the results of which we have published elsewhere.10 Our second approach was to conduct experiments designed to assess the effects of the intensification of collection strategies both on overall artefact density and on the typological and chronological character of the collected assemblage. The results are summarized below. During the 2004 and 2006 survey season, we surveyed 21 units using more intensive survey techniques (Fig. 2). In 2004, we collected a 5% (approximately 80sq.m.) sample of the surface of 10 grid squares through more rigourous vacuuming or hoovering collection using total collection circles. In these total collection circles, we did not use our standard Chronotype collection method but rather collected all artefacts found within the bounds of a circle with a radius

Fig. 2. Experimental units are shown with heavy black borders and unit numbers. Shading in all other units represents artefact densities according to a Jenks/K-Means gradient (W.R. Caraher).

of 5.1m. In 2006, we used the same technique to sample an additional eleven units for density, but did not collect artefacts. This brought the total number of experimental circles to 21, or about 8% of the total number (n=252) of grid squares. We identified units suitable for our experimental hyper-intensive survey through a number of methods. We sought to re-sample units with a

____________ R.S. Moore (eds), Methods and Meaning in Medieval and Post Medieval Greece: A Tribute to Timothy E. Gregory (under review); and W.R. Caraher, D. Nakassis and D.K. Pettegrew, Siteless Survey and Intensive Data Collection in an Artefact-rich Environment: Case Studies from the Eastern Corinthia, Greece in Journal of Mediterranean Archaeology 19.1 (2006), 11-13. 9. For recent work in experimental archaeology, cf. R. Schon, Experiments in Archaeological Survey Methodology (Ph.D. Dissertation, Bryn Mawr College, 2002); idem On a Site and Out of Sight: Where have our data gone? in JMA 13 (2000), 107-11; E. Gibson, Seeding experiments in M. Given et al., Troodos Archaeological Project Environmental Project: First Preliminary Report (June-July 2000), RDAC (2001), 428-29; N. Meyer and R. Schon, Experimental Data in M. Given and A.B. Knapp (eds), The Sydney Cyprus Survey Project: Social Approaches to Regional Archaeological Survey (Los Angeles 2003), 52-56; and T.F. Tartaron et al., The Eastern Korinthia Archaeological Survey: Integrated Methods for a Dynamic Landscape in Hesperia (2007), forthcoming. 10. Caraher, Nakassis and Pettegrew 2006.

WILLIAM R. CARAHER, R. SCOTT MOORE, JAY S. NOLLER, DAVID K. PETTEGREW

wide range of densities and, consequently, selected units that roughly corresponded to the quartile breaks in overall artefact densities for the survey area as a whole. In particular, we were concerned with sampling from the lowest and highest density units because our analysis of finds data from other surveys suggested that our sampling strategy tended to produce the most distorted picture of the artefacts actually on the surface in these environments.11 We also sought to re-sample units distributed throughout the entire area surveyed during the 2004 and 2005 field seasons (Fig. 2). Although these 21 experimental units were not spaced at equal intervals, we did resurvey units from each area of our site, especially those areas associated with significant features, such as the well-built, stepped, check dam (Unit 154) and the remains of the so-called Venetian castle (Unit 53). Finally, in addition to testing our sampling strategy, these experiments were also designed to produce additional chronological data. Hence we chose a number of units in lowerdensity areas where the small number of artefacts collected from the unit provided only a limited basis for chronological analysis. For each unit selected for experimental resurvey, we employed a similar technique. We documented all the artefacts from a circle with a radius of 5.1m. (approximately 80sq.m.), to produce a 5% sample of the total area of each unit (1,600sq.m.). A team of two to four field walkers took as long as was necessary to document the artefacts present in the unit marking their progress with a rope attached to a central point. The person-time per unit ranged from thirty minutes to four hours, depending upon the density of material in the unit. In the 2004 and 2006 seasons, we counted all the pottery, tile, and other classes of artefacts within the circumference of the circle, carefully vacuuming (removing) all material until no artefacts remained visible within the circle. In the 2004 experiments, we also collected the artefacts, placed them in labelled bags, and took them back to the Larnaka District Museum for storage. We are now able to make preliminary conclu-

sions on how sampling strategies affect overall artefact density. While we understand that overall density of artefacts cannot stand alone as an indicator of any specific form of cultural activity, it does present a rough indication of the extent of activity areas diachronically in a particular area. Moreover, as Mediterranean archaeologists have begun to compare artefact densities between different regions, it is imperative that archaeologists understand how different methods relate to and create artefact densities. Table 1 compares the densities produced by our standard field walking technique and those produced by our experimental total collection circles. The table compares the total number of artefacts collected from our 5% total collection circles (Circle Total), to the maximum counted by any one walker over the course of our typical technique (5% Pedestrian Sample) and to the total number of artefacts counted with our typical survey technique (20% Pedestrian Sample). By using the maximum count of a field walker during our standard survey, we can compare two 5% samples and isolate the effectiveness of the two methods employed by field walkers. From these figures we have produced an estimated unit total (Estimated Total (Pedestrian) and Estimated Total (Circle)) to which we have applied a rather simple calculation to account for differences in surface visibility.12 These calculations allow us to

____________ 11. Caraher, Nakassis and Pettegrew 2006. 12. For example, for a 5% hoovering circle of a unit with 50% visibility, we would compute the Putative Total by multiplying the circle total density 20 (i.e., a 5% sample = 1/20 of the area of the unit) 2 (i.e., 50% visibility indicates only half the area of the unit was visible). Another example: for a 20% pedestrian survey with 100% visibility, we would compute Putative Total by multiplying the total density 5 (i.e., a 20% sample = 1/5 of the unit) 1 (i.e., 100% visibility indicates the entire surface of the unit was examined). Archaeologists have suggested more sophisticated adjustments for surface visibility based on experience elsewhere, and these more sophisticated models would certainly influence our findings. The relationship between visibility and artefact density is not fully linear (for discussion, cf. Schon 2000, 2002); nonetheless, calibrating for visibility differences does provide a systematic (if imperfect) correction and is of some value.

THE PYLAKOUTSOPETRIA ARCHAEOLOGICAL PROJECT

compare an estimated total count based on two methods of counting. The differences between an estimated total count based on our 5% experimental circles and 5% maximum pedestrian walker appears in our Percentage Difference column. The most obvious conclusion to draw from the comparison is that more intensive scrutiny of the surface in our total collection circles produces radically higher artefact densities than does our standard field walking methods. The average percentage increase between Estimated TotalPedestrian and Estimated TotalCircle averages over 400%. In 17 of the 21 units, in fact, the total collection circles produced more artefacts than the 20% sample of the surface produced through our standard survey. If we had compared the estimated total based on our 20% sample (not shown in Table 1) to the estimated total based on our 5% circle, we find an even more remarkable 800% increase in total number of artefacts. In general the difference between our 5% pedestrian survey and a 5% total collection circle tends to decrease as overall density increases and this is consistent with a general trend of intra-unit variation decreasing as overall density increases. Nevertheless, the general pattern at the 5% level suggests that more intensive collection strategies like hoovering do greatly increase artefact density. Only two units are exceptional. The Estimated TotalCircle from Units 66 and 187 fell below the Estimated TotalPedestrian. There are several possible explanations for this anomaly. It is possible, on the one hand, that artefact densities were highly variable within the units and that the hoovering circle sample fell outside the highestdensity part of the unit. For the moderately dense Unit 66, in particular, the Estimated Total (Pedestrian) is based upon only 20% of the surface being visible in our original survey, which equates to a 1% sample of the surface; such a small sample may be particularly prone to distortion. The other exception is the very low-density Unit 226, which produced only one artefact

through normal pedestrian methods and zero artefacts through hoovering circles. This grid square occurred at the part of the site where densities trailed off toward zero, and where the finding of artefacts may be fortuitous. In any case, neither of these exceptions negates the general pattern from our survey data: sampling the surface of the soil more intensively improves our overall artefact density for the grid square. Whether the additional quantity of pottery counted through more intensive and time-consuming methods like hoovering is ultimately justified depends upon a closer analysis of the chronological and functional attributes of the finds. The data for this analysis derives from the ten 5% total sampling units surveyed in 2004. In these units, we collected all artefacts visible in the surface matrix and then documented the artefacts from these units consistent with how we documented artefacts from standard grid squares. This experiment, then, allows us to ascertain the effectiveness of the Chronotype sampling method for producing an accurate sample of the material in a unit, and especially whether our Chronotype collections represent the diversity of artefacts from a typological and functional perspective.13 The following analysis compares the assemblages produced by the two methods (20% pedestrian survey sample using Chronotype collection vs. 5% hoovering sample using total collection) to determine their relative effectiveness for representing the artefacts present in a unit. First, the experiments suggest that more intensive hoovering circles do not contribute in consistent ways to our understanding of chronology in the survey units. For the ten experimental units from 2004, our pedestrian grid square collection produced an assemblage representing on average 5.4 total periods per unit, whereas total
____________ 13. We have argued elsewhere that Chronotype collection strategies are as effective for determining the chronology of a unit as more intensive regimes that tend to produce redundant chronological data. Cf. Caraher, Nakassis and Pettegrew 2006.

WILLIAM R. CARAHER, R. SCOTT MOORE, JAY S. NOLLER, DAVID K. PETTEGREW

collection circles showed a meagre increase to an average of 5.9 periods per unit. While this would suggest that total collection circles produce a slightly more chronologically refined assemblage, it is notable that half of the units (n=5) produced the same number of chronological periods as the normal collection. Moreover, the additional periods represented in total collection in some cases had very broad chronological ranges such as Post-Prehistoric or Ancient. In other cases, a single example of Archaic, ClassicalHellenistic, Hellenistic and Early Roman appeared through total collection experiments; there were also instances, however, of single examples of Archaic, Bronze Age, Classical, Hellenistic, Early Roman, and Late Medieval in our standard pedestrian collections. This analysis would seem to show that the two main influences on this experiment, the variable intensity of scrutiny and extent of sampling, may in fact cancel out in the production of chronological diversity from a unit. Chronological diversity, however, is only one measure of diversity within the survey unit. Functional and typological diversity are particularly important, especially for a project interested in issues of trade and inter- and intra-regional connectivity, as well as the functional variation within an area dominated by a single chronological period (Late Roman). Higher intensity survey methods produce more varied assemblages, including the number of chronotypes/unit. On average, grid square collections produced 11.9 chronotypes per unit or about 2.1 chronotypes per chronological period; our total collection circles produced 15.6 chronotypes per unit for a marginally improved 2.5 chronotypes per period. For the Late Roman period, especially, there was substantial improvement in the number of chronotypes per collection unit: grid square collection produced 49 chronotypes, whereas total collection circles produced 69 chronotypes. Two units in particular, 40 and 71, registered significant gains in the number of Late Roman chronotypes present, and Unit 71 recorded a greater number of Roman artefacts.

Finally, it is perhaps unsurprising to note that total collection circles have tended to produce a significantly greater number of artefacts. On average, the total collection circles, representing 5% of the total surface of the unit, produced 240 more artefacts per unit than the 20% Grid Square sample. While this number was certainly exaggerated by two very high-density units (71 and 40 which produced 1008 and 354 additional artefacts, respectively), the trend was consistent across all the units in our experiments. The consequence of this vast increase in quantity of artefacts is a substantial increase in the time needed to process and analyze artefacts and a dramatic decrease in efficiency, as total collection circles produce significant quantities of redundant data. Resisting such inefficiency is responsible behaviour in a world of increasing time and budget restraints, and also recognizes the increased strain on storage facilities available to many projects. A major survey can produce quantities of artefacts rivalling the quantities produced by an excavation; the expense of storing, cataloguing, and collecting large quantities of artefacts in terms of time, money, and resources must be carefully weighed against the gains in knowledge. The results of our 2004 experiments allow us to draw several conclusions. First, for large scale, intensive surveys focusing on ascertaining the chronological diversity present in a given unit of space, Chronotype collection provides an efficient means of gathering chronological data from the soil matrix. Yet, the less intensive Chronotype collection strategy also under represents the functional and typological diversity of artefacts present in a unit. In general our experiments reinforced the delicate balance that exists between efficiency and precision of analysis in a survey environment.

DENSITY While our experimental data suggests that our method under represented artefact density compared to methods emphasizing a greater scrutiny of the soil matrix, it is broadly successful in rep-

THE PYLAKOUTSOPETRIA ARCHAEOLOGICAL PROJECT

resenting relative density between units on a general scale. Despite our small sample of experimental units, relative densities produced through both sampling regimes remained stable for the entire range of artefact densities, although the rank of units with more middling densities (from 300 artefacts / ha. to 3,000 artefacts / ha.) varied slightly between the different sampling methods. This, then, confirms the methodological validity of our highest density areas, which will form the centre of the following analysis. Artefact densities indicated one central area of exceptional density which we designated Zone 1. Zone 1 includes the immediate vicinity of the excavated basilica and extends for ca 500m. to the east. The borders of Zone 1 are marked by declining densities, visually evident in this map by a K-Means or Jenks analysis which shows natural breaks in statistical arrays. Zone 1, the highest density area from our site, produced overall densities in excess of 6,000 artefacts per hectare for an area of 11ha., which is well above the typical density threshold of 3,000-5,000 artefacts / ha. for defining sites in the Eastern Mediterranean.14 Moreover, this area produced a large quantity of architectural material including partially exposed walls, carefully prepared gypsum blocks, and cut stone. Several fragments of agricultural processing equipment also derived from Zone 1. Perhaps a traditional survey would have designated this high density area a site and the surrounding lower- to moderate-density units off-site. Despite the statistically significant drop off in density, PKAP continued to extend our grid squares to the north and east in large part because reconnaissance survey had suggested higher artefact densities in the fields at the base of Kokkinokremmos. As we extended our grid to the north, we were able to identify not a single highdensity area, but rather, several relatively isolated high-density areas stretching loosely along the base of the Mavrospilos and Kokkinokremmos ridgeline. These areas combined to form Zone 2, an area of approximately seven hectares, with an

Fig. 3. Zone 1 and Zone 2. Shading in all other units represents artefact densities according to a Jenks/K-Means gradient (W.R. Caraher).

overall artefact density of approximately 2,000 artefacts per ha. While this area did not have densities nearly as high as Zone 1, it clearly appeared to be distinct from its surrounding areas. The division between Zone 1 and Zone 2 was related to the previous installation of a water treatment plant for the British base, which involved considerable excavation and disruption of the top soil.

CERAMIC FINDS
At present we are able to offer some general observations regarding the chronology and function of our site. Of the artefacts collected over the course of the survey, only 53% can, at present, be dated to a period of 1,000 years or less, which is rather typical for surface material; most of these diagnostic artefacts are Late Roman in date, but there are significant traces of earlier material as well. Despite the presence of several significant Bronze Age sites in the immediate vicinity of Pyla-Koutsopetria, we have identified very little Bronze Age material at Koutsopetria and none of
____________ 14. S.E. Alcock, J.F. Cherry and J.L. Davis, Intensive survey, agricultural practice, and the classical landscape of Greece in I. Morris (ed.), Classical Greece: Ancient Histories and Modern Archaeologies (Cambridge 1994), 138.

WILLIAM R. CARAHER, R. SCOTT MOORE, JAY S. NOLLER, DAVID K. PETTEGREW

the kinds of high-quality fine ware found at PylaKokkinokremmos or Vergies.15 It may be that the site of Steno, located just to the northeast of our survey area and shielded by the fortified acropolis of Kokkinokremmos, was preferred for concentrated settlement over the more exposed coastal plain. The Cypro-Archaic to Hellenistic period witnessed a graduate increase in number of artefacts while the total number of artefacts remained small, forming less than 3% of all the pottery datable to 1,000 years or less. The finds from the Cypro-Geometric to Cypro-Archaic period tend to be clustered at the base of the coastal ridge of Vigla. They consist primarily of Archaic basket handles in multiple fabrics and shapes and painted fine ware sherds from both the Geometric and Archaic periods.16 A sample of artefacts from the height of Vigla itself produced artefacts from the Archaic period as well, but we did not recover any earlier, particularly Geometric, material. This material may well provide important context for the 7th century B.C. Phoenician inscription found in the area of Vigla or Paliokastro and published by Lacau.17 This text suggested the presence of a Sanctuary of Bes in the area. While our finds cannot confirm the existence of a sanctuary in particular, they do suggest substantial activity on and about this prominent coastal height as early as the 8th and 7th century B.C. From the Classical to Hellenistic periods, finds in the area gradually increase and extend further to the south into the coastal plain. Judging by the number of lead sling pellets extracted over the years from its height and the inscribed game board, it is clear that the height of Vigla housed some form of military presence during the Hellenistic period.18 Our sample of artefacts from Vigla produced fine wares (imported black glaze, Hellenistic colour-coated wares), cooking pots, and amphoras, although we recovered no additional sling pellets. The large, inscribed stone basin discovered in 1951 at the base of Vigla and datable to the 3rd century B.C. indicates a sanctuary in the area perhaps associated with olive oil

production; Hadjisavvas has identified this basin as a settling basin for an olive press.19 Classical to Hellenistic material also extended well to the east of Vigla and was particularly concentrated at the base of Kokkinokremmos, in the area that we have denoted as Zone 2. While amphoras, fine wares, and medium coarse fabric utility wares are represented generally in many units with Classical to Hellenistic material, it is notable that there were very few kitchen or cooking wares found below Vigla compared to our sample of material from the top of Vigla or Zone 2. Perhaps this reflects the presence of a sanctuary where storage vessels and fine wares might be expected to outnumber the kitchen wares more commonly associated with domestic assemblages. Roman pottery (100 B.C. - 749 A.D.) makes up the dominant chronological component at Koutsopetria. While the distribution of Roman artefacts is similar to material from earlier periods, Roman material extends further south toward the shore. The overlap in the Early Roman and the Classical-Hellenistic material presumably indicates continuity of occupation at this site. The most prevalent artefact types from the Early Roman period are highly diagnostic
____________ 15. A.H.S Megaw, Archaeology in Cyprus, 1952, JHS 73 (1950), 135; O. Masson, Kypriaka II: Recherches sur les antiquits de la rgion de Pyla, BCH 90 (1966), 7-8; Karageorghis and Demas 1984, 50-53. 16. S. Hadjisavvas, Olive oil processing in Cyprus from the Bronze Age to the Byzantine Period (SIMA 99) (Gteborg 1992), 30, 78; J. Leonard, The Anchorage at Kioni in J. Fefjer (ed.), Ancient Akamas: Settlement and Environs (Aarhus 1995), 141, 155; A.G. Sagona, Levantine Storage Jars of the 13th to 4th Century in Opuscula Atheniensia 14 (1982), 336-62. 17. P. Lacau, Une inscription phnicienne de Chypre, BIFAO 2 (1902), 207-11; P.-L. Couchoud, BCH 90 (1966), 8; A. Caquot and O. Masson, Deux Inscriptions Phniciennes de Cypre, Syria (1968), 295-301. 18. I. Michaelidou-Nicolaou, Table a jeu de Dhekelia (Chypre), BCH 89 (1965), 122-27; I. Nicolaou, Inscriptiones Cypriae Alphabeticae XVI, 1976, RDAC 1977, 20916; I. Nicolaou, Inscriptiones Cypriae Alphabeticae XVIII, 1978, RDAC 1979, 344-51; I. Nicolaou, Inscriptiones Cypriae Alphabeticae XIX, 1979, RDAC 1980, 261-62. 19. S. Hadjisavvas 1992, 60, 76.

THE PYLAKOUTSOPETRIA ARCHAEOLOGICAL PROJECT

fine wares such as Cypriot Sigillata and Eastern Sigillata A. Unlike the pattern evinced by the Classical to Hellenistic periods, kitchen and cooking wares seem to be distributed rather evenly throughout the units with Early Roman and Roman material. Many of the common transport and utility wares, such as Koan and Rhodian type amphoras, were lacking from the site, although less diagnostic body sherds certainly represented additional amphoras. Only one example of the so-called pinched handled amphoras common to the many Roman sites on the western part of the island emerged from our assemblage.20 Over 80% of the material at our site dates to the Late Roman period. Material of this date occurs in almost 90% of the units investigated, extending southward to the shore and eastward along the base of the Kokkinokremmos ridge. The quantity and distribution of this material is augmented by the abundance of Late Roman roof tile present at the site, which represents 35% of the Late Roman material collected over the course of the survey. The majority of roof tiles come from the field immediately adjacent to the excavated basilica. The most distinctive shape are the large, heavy Corinthian style tiles which appear in a variety of fabrics, including one of a yellow colour which has been associated with production sites on the Mesoria plain.21 This type of tile is not uncommon on other sites on the island and in shipwrecks off the coast, suggesting that both local production centres (e.g., the kiln at Zygi) and trade supplied the Late Roman building boom on the island. In addition to roof tile, we collected a robust assemblage of fine wares. While Cypriot Red Slip makes up the majority of Late Roman fine wares, we have detected some patterning in the distribution of both local and imported Late Roman Red Slips. In Zone 1, for example, CRS makes up 55% of the assemblage followed by ARS at 28% and PHW at 17%. In Zone 2, CRS represents a more substantial majority: 76% with PHW following at 15% and ARS at a mere 9%. Compared to other surveys particularly those

of large sites at Kopetra and Maroni-Petrera we have noted a far higher proportion of ARS and lower proportion of PHW.22 It is also interesting to note that our PHW is rather concentrated at the site, appearing in areas where ARS is less common, as for example in the eastern half of Zone 1 and the core units of Zone 2. Most of our Late Roman fine wares date to the 6th and 7th centuries, the most commonly encountered forms being ARS 105, CRS 9, and PHW 10. Only a few Late Roman forms, in fact, date to before the middle of the 6th century. Amphoras from all periods make up approximately 15% of our total quantity of pottery from Pyla-Koutsopetria. LR1 amphora was the largest category of Late Roman amphora, accounting for 30% of the total amphoras from all periods. While there is still debate over the contents of LR1 amphora,23 the large numbers of LR1 at Pyla-Koutsopetria indicate the importance of the site as a centre of exchange for the local area. We have recently begun a project to group our LR1 amphora into categories based on fabric; we hope that soon we may be able to offer more nuanced typological identifications based on recent research conducted on LR1 amphoras.24 Densities decline markedly in the postRoman period. The eastern part of Zone 1, however, contains a small amount of Medieval material. The most interesting concentration of post____________ 20. Leonard 1995, 144-46; Leonard 2005, 889-905. 21. M. Rautman, A Cypriot Village of Late Antiquity: Kalavassos-Kopetra in the Vasilikos Valley (Portsmouth, RI 2003), 53-55, 213. 22. Sturt Manning et al., The Late Roman Church at Maroni Petrera: Survey and Salvage Excavations 1990-1997, and other traces of Roman remains in the lower Maroni Valley, Cyprus, 42-43; and Rautman 2003, 168-76. 23. Hugh Elton, The Economy of Southern Asia Minor and LR1 Amphorae in Esparraguerra, Garrigs, and Ontiveros (eds), LRCW I: Late Roman Coarse Wares. Cooking Wares and Amphorae in the Mediterranean (Oxford 2005), 691-92. 24. S. Demesticha, Amphora Production on Cyprus During the Late Roman Period in Actes de VIIe Congrs Intenational sur la Cramique Mdivale en Mditerrane; Thessaloniki 11-16 octobre 1999 (Athens 2003), 469-76.

10

WILLIAM R. CARAHER, R. SCOTT MOORE, JAY S. NOLLER, DAVID K. PETTEGREW

Roman artefacts comes from the southeastern corner of Zone 1 near a ruined wall that remains visible to the north of the Larnaka-Dekeleia road. This wall has been alternately identified as a church or as a Medieval / Venetian-Ottoman fortification. It seems likely that this fortification is the reason why this stretch of coastline was identified as Paliokastro on 19th and early 20th century maps of the island. Recent plowing has cut through a floor in this area and exposed sub-floor packing, some of which is Late Roman in date. While it is difficult to associate this floor conclusively with the ruined wall visible from the Larnaka-Dekeleia road, it does establish a Late Roman terminus post quem for a building in this area. The scatter of Medieval and Ottoman / Venetian pottery suggests that this area may have remained in use from the 13th to the 19th century. More modern artefacts are also present in some quantities in the area. Most, however, are associated with debris from the modern road or the ongoing agricultural use of the area.

that is now in-filled. We have also documented at least two basins that are presumably settling basins for drains or other water systems; comparanda at Kourion and elsewhere suggest a Roman or Late Roman date. There are several features, consistent with olive processing on the site, which are also compatible with a Late Roman or Early Byzantine date. We have documented, for example, several spouts that may have belonged to an olive press bed. These finds, along with the large settling basin from the Hellenistic period, suggests a long-term commitment to olive oil production in Koutsopetria. This agricultural aspect of the site would not, of course, be incompatible with the areas important religious element; certainly, churches in Cyprus and elsewhere in the Eastern Mediterranean were tied to agricultural production.

ARCHITECTURAL FEATURES In addition to recording the distribution of pottery, PKAP has also sought to document the features at the site. Numerous walls can be traced throughout, although few for more than 10 metres; one well-preserved wall shows several courses of Late Roman brick with gypsum mortar. Fragments of mortar and cobble are common at the site and probably indicate the widespread presence of buildings. The ubiquity of large gypsum blocks (some exceeding a metre in length) throughout Zone 1, and more humble building material, like coarse local limestone, would suggest the same conclusion. Some of this limestone appears to have been quarried locally from either Vigla or other wave eroded bedrock outcrops. Quarry cuts on Vigla are joined by cut stairs, perhaps designed to provide access from below or for quarry operations. Amidst the building material, there are other features compatible with both agricultural production and habitations. There is at least one well

CONCLUSIONS AND FUTURE DIRECTIONS Koutsopetria is a site that continues to reveal new information about its past. Recent deep plowing in previously uncultivated areas has exposed a broad swath of cultural material, bringing to light new kinds of artefacts: Late Roman kitchen wares, an Archaic-Hellenistic figurine, forms of combed ware that date to the 7th or 8th century, marble basins, and marble moulded pieces. The pottery exposed by the plowing is consistent with the pottery collected during the field walking seasons, but with better quality pieces that indicate more discrete functional concentrations: one area filled with roof tiles, another with amphora handles and toes, and a third with both Late Roman fine wares and cooking wares. While this development has introduced new windows into the past use of the area, it has also come at the expense of the destruction of parts of the site and its cultural remains. Our plans for next season include finalizing our study of the finds and producing publications for the site. As part of this process, we will make our raw data accessible on-line by next summer. We also plan to continue limited fieldwork in the

THE PYLAKOUTSOPETRIA ARCHAEOLOGICAL PROJECT

11

vicinity of Koutsopetria, including a collaborative project with a team of researchers from the University of Edinburgh at the site of Kokkinokremmos. Given the widespread exposure of architecture on the surface of Koutsopetria, an exciting possibility is the implementation of geophysical investigation at the site. A final direction for further fieldwork includes an expanded

pedestrian survey of the broader micro-region. This would set the site of Pyla-Koutsopetria within its immediate regional environment and also contribute substantially to understanding a landscape that is daily disappearing in the modern development of the coastline of southeast Cyprus.

12

Grid Unit 1 1 1 2 3 3 5 4 6 10 10 24 38 43 54 39 65 78 76 174 225 437 537 243 218 160 90% 40% 30% 60% 70% 117 100% 106 100% 96 20% 81 40% 1900 4300 1080 780 1444.44 3900 5066.67 5800 6428.57 65 90% 533.333 20 80% 250 16 60% 333.333 70 18 86 203 63 31 435 236 477 1274 736 1096 14 100% 120 11 13 40% 200 37 8 70% 142.857 14 6 50% 120 32 90% 100% 40% 100% 60% 80% 90% 40% 100% 100% 100% 90% 100% 100% 60% 70% 3 40% 150 13 40% 3 60% 66.6667 18 100% 2 80% 25 4 90% 88.9 360 650 711.1 280 1850 220 2333.3 450 1911.1 10150 1260 620 8700 5244.4 9540 25480 24533.3 31314.3 1 90% 22.2222 0 100% 0 1 80% 25 8 90% 177.8

Density Walked per ha 711.20% 0.00% 355.60% 540.00% 433.33% 592.58% 196.00% 925.00% 183.33% 699.99% 180.00% 358.33% 534.21% 29.30% 57.41% 1115.38% 363.07% 244.62% 502.89% 422.99% 487.11%

5% 20% Pedestrian Pedestrian Sample Sample Visibility (#Artefacts) (# Artefacts) Original Estimated Circle Unit Total Total (# Visibility (Pedestrian)(# Artefacts) Resurvey Estimated Unit Total (Circle) Percentage Difference

151

31.25

226

31.25

169

62.5

34

93.75

85

93.75

102

187.5

173

250

141

406.25

191

437.5

45

500

53

625

120

2031.25

127

2531.25

66

3000

187

3312.5

40

3656.25

154

5000

WILLIAM R. CARAHER, R. SCOTT MOORE, JAY S. NOLLER, DAVID K. PETTEGREW

63

6812.5

61

7593.75

18

13656.3

71

16781.3

Table 1. Results of Experimental Total Collection Circles (units divided based on density quartiles).

THE PYLAKOUTSOPETRIA ARCHAEOLOGICAL PROJECT

13

REFERENCES
ALCOCK, S.E.; J.F. CHERRY and J.L. DAVIS 1994: Intensive survey, agricultural practice, and the classical landscape of Greece in I. Morris (ed.), Classical Greece: Ancient Histories and Modern Archaeologies (Cambridge), 137-70. BAKIRTZIS, C. 1995: The Role of Cyprus in the Grain Supply of Constantinople in the Early Christian Period in V. Karageorphis and D. Michaelides (eds), Proceedings of the International Symposium, Cyprus and the Sea (Nicosia), 247-53. CAQUOT, A. and O. MASSON 1968: Deux Inscriptions Phniciennes de Cypre, Syria, 295301. CARAHER, W.R.; R.S. MOORE and D.K. PETTEGREW: The Pyla-Koutsopetria Archaeological Project: A Third Preliminary Report (Paper delivered at the 24th Annual CAARI Archaeological Workshop, Nicosia, Cyprus, June 2006). CARAHER, W.R.; R.S. MOORE; J.S. NOLLER and D.K. PETTEGREW 2005: The Pyla-Koutsopetria Archaeological Project: First Preliminary Report (2003-2004 Seasons) in Report of the Department of Antiquities of Cyprus, 245-68. CARAHER, W.R.; D. NAKASSIS and D.K. PETTEGREW 2006: Siteless Survey and Intensive Data Collection in an Artefact-rich Environment: Case Studies from the Eastern Corinthia, Greece in Journal of Mediterranean Archaeology 19.1, 7-43. DEMESTICHA, S. 2003: Amphora Production on Cyprus During the Late Roman period in Actes de VIIe Congrs Intenational sur la Cramique Mdivale en Mditerrane; Thessaloniki 11-16 octobre 1999 (Athens), 469-76. ELTON, H. 2005: The Economy of Southern Asia Minor and LR1 Amphorae in Esparraguerra, Garrigs and Ontiveros (eds), LRCW I: Late Roman Coarse Wares. Cooking Wares and Amphorae in the Mediterranean (Oxford), 691-92. GIBSON, E. 2001: Seeding experiments in Michael Given et al., Troodos Archaeological Project Environmental Project: First Preliminary Report (June-July 2000), RDAC, 428-29. GREGORY, T. 2004: Less is Better: The Quality of Ceramic Evidence from Archaeological Survey and the Practical Proposals for Low-Impact Survey in a Mediterranean Context in E. Athanassopoulos and L. Wandsnider (eds), Mediterranean Archaeological Landscapes: Current Issues (Philadelphia), 15-36. HADJISAVVAS, S. 1992: Olive oil processing in Cyprus from the Bronze Age to the Byzantine Period (SIMA 99) (Gteborg). 1993: Annual Report of the Department of Antiquities, Cyprus (Nicosia), 70-72. 2000: Chronique des fouilles et dcouvertes archologiques Chypre en 1999, BCH 124, 693. HORDEN, P. and N. PURCELL 2000: The Corrupting Sea (Oxford). KARAGEORGHIS, V. and M. DEMAS 1984: Pyla Kokkinokremos: A Late 13th Century Fortified Settlement in Cyprus (Nicosia). LACAU, P. 1902: Une inscription phnicienne de Chypre, BIFAO 2, 207. LEONARD, J.R. 1995: The Anchorage at Kioni in J. Fefjer (ed.), Ancient Akamas: Settlement and Environs (Aarhus). 2005: Roman Cyprus: Harbors, Hinterlands, and Hidden Powers (Unpublished Ph.D. Dissertation, State University of New York at Buffalo). MASSON, O. 1966: Kypriaka II: Recherches sur les antiquits de la rgion de Pyla, in BCH 90, 7-8.

14

WILLIAM R. CARAHER, R. SCOTT MOORE, JAY S. NOLLER, DAVID K. PETTEGREW

MANNING, ST. et al. 2002: The Late Roman Church at Maroni Petrera: Survey and Salvage Excavations 1990-1997, and other traces of Roman remains in the lower Maroni Valley, Cyprus (Nicosia). MEGAW, A.H.S. 1950: Archaeology in Cyprus, 1952, JHS 73, 135. MEYER, N. and R. SCHON 2003: Experimental Data in Michael Given and A.B. Knapp (eds), The Sydney Cyprus Survey Project: Social Approaches to Regional Archaeological Survey (Los Angeles), 52-56. MICHAELIDOU-NICOLAOU, I. 1965: Table a jeu de Dhekelia (Chypre), BCH 89, 122-27. MOORE, R.S. (under review): A Decade Later: the Chronotype System Revisited in W.R. Caraher, L. Hall and R.S. Moore (eds), Methods and Meaning in Medieval and Post Medieval Greece: A Tribute to Timothy E. Gregory. Nicolaou, I. 1977: Inscriptiones Cypriae Alphabeticae XVI, 1976, RDAC, 209-16.

1979: Inscriptiones Cypriae Alphabeticae XVIII, 1978, RDAC, 344-51. 1980: Inscriptiones Cypriae Alphabeticae XIX, 1979, RDAC, 261-62. RAUTMAN, M. 2003: A Cypriot Village of Late Antiquity: Kalavasos-Kopetra in the Vasilikos Valley (Portsmouth, Rhode Island). SAGONA, A.G. 1982: Levantine Storage Jars of the 13th to 4th Century in Opuscula Atheniensia 14, 336-62. SCHON, R. 2000: On a Site and Out of Sight: Where have our data gone?, JMA 13, 10711. 2002: Experiments in Archaeological Survey Methodology (PhD. Dissertation, Bryn Mawr College, Philadelphia). TARTARON, T. et al. (forthcoming): The Eastern Korinthia Archaeological Survey: Integrated Methods for a Dynamic Landscape, Hesperia.