sign in sign up
<<
Home , Andania, Asine, Asine (Messenia), Cardamyle, Eira (Messenia), Filiatra

HESPERIA 74 (2005) REGIONAL Pages 147-209 ARCHAEOLOGICAL PROJECT, PART VII

Historical , Geometric through Late Roman

abstract

In this article, the authors explore patterns in regional activity inMessenia, corner the southwest of the Greek P?loponn?se, from the Geometric to the cen end of the Late Roman period (ca. eighth century b.c. to seventh on tury A.D.). The analysis is based extant historical evidence, the campaigns of the Minnesota Messenia Expedition, and?above all?the results of the Pylos Regional Archaeological Project. These three data sets have been inte as far as in order to trace in the grated, possible, long-term changes region a and to provide foundation for further work in this still underexplored por s tion of historical landscape.

a a Messenia possesses varied and fertile topography, unique and check a ered history, and rich, if increasingly threatened, archaeological heritage. The region today is perhaps most celebrated, by scholars and tourists alike, for its Age remains: the Palace of and its archive, the tholos tombs of Peristeria and (Fig. 1). But the centuries after the end of the Mycenaean palaces also pose interesting problems for the archaeology of Messenia. on assess The principal emphasis in the present study is presenting and to ing the contributions made the long-term history of Messenia by the work of the Pylos Regional Archaeological Project (PRAP).1 The goals, methodologies, personnel, and preliminary results of PRAP have been reviewed elsewhere, but two essential aspects of the project will be restated area a zone briefly here.2 First, the study comprised large (encompassing

1. This was as assistance study undertaken part support, and the received Stephen Hodkinson, Nino Luraghi, from other see Davis and for mate of the Pylos Regional Archaeological organizations, Graham Shipley sharing and as such owes thanks to all et al. 488. We also thank was Project, 1997, p. rial unpublished when this report the that have Robertson for her usual to organizations supported Rosemary written, and the project's codirectors that endeavor. for excellent work on and for comments on the Major funding illustrations; manuscript. Na Patrick 2. For PRAP has been provided by the Jennifer Gates, Livingood, project methodologies and tional Endowment for the Solemeto for et Humanities, and?especially?Julie results, see, especially, Davis al. 1997; et the National Geographic Society, and assistance with GIS analyses and data Zangger al. 1997; Davis 1998c. See the Institute for to Aegean Prehistory. management. We would like express also Bennet, Davis, and Zarinebaf For details of our to those institutions' gratitude Richard Catling, Shahr 2000; Lee 2001.

? The American School of Classical Studies at

American School of Classical Studies at Athens is collaborating with JSTOR to digitize, preserve, and extend access to Hesperia ® www.jstor.org 148 SUSAN E. ALCOCK ET AL.

Figure 1.Map ofMessenia, showing the principal places mentioned in the text. R. J. Robertson

ca. some 250 km2) roughly centered around the , of which were 40 km2 intensively surveyed, including the Englianos ridge (the site of the Palace of Nestor), coastal areas north of the Bay of Navarino, and was to valleys east of the Adgaleon range (Fig. 2). Second, PRAP designed consider all periods of the past in their changing environmental settings. on monuments no means Although work post-prehistoric sites and is by unknown inMessenia, the investigations conducted by PRAP allow, for an a on a the first time, analysis of lengthy historical time span the basis of rigorously explored sample of the landscape. It could be ruefully argued that the study of the historical past of Messenia in particular requires such archaeological contributions. While can few parts of Greece boast the documentary richness of detail available for Athens and , Messenias peculiar historical trajectory leaves it to most especially vulnerable misunderstanding and neglect. The well known "fact" about the region, of course, is its unprecedented domina as tion by Sparta inArchaic and Classical times. External observers, such us and , tell something of the fate of Messenia and many of its inhabitants, the famed of antiquity. It is difficult, however, to overstate the limitations and biases of these sources, and the complete HISTORICAL MESSENIA 149

flit I \ M03_ jmmmJ^s^" Mouzaki Mt Airtat&nn I n'^'vieT

^ Metamorfosi) \ Vlasis |L^ ^^^^^^^^^^^^^^J ^^^^^A. K\(Y7^/ ? r^^Hora

Settlement j o ,?k ^j0tfK???^ ?-'wL r~^X ^~X>Stenosia //^:::=:-^/Handrinou \

Figure 2. Region investigated by lack of independent Messenian testimony from this time period must be PRAP. Areas intensively surveyed Those who lived in the have become, in the famous and sites defined, 1991-1995. appreciated. region a without After liberation in 370/69 b.c. R. J. Robertson phrase, "people history."3 by and his alliance of anti-Spartan states, the Messenians be come an "visible" again, in outburst of civic foundation and monumental area? display?most famously atMessene, northeast of the PRAP study even and through the crafting and promulgation of local .4 But a the freed Messenia, minor player in the tangled world of Hellenistic never a politics, has attracted great deal of scholarly attention, and the same is true of Roman Messenia, part of the Roman empire along with the rest of Greece.

to seen 3.Wolf 1982. In the minds of many had really begun inquire into in the results of excavations in the the own and at exam historians, domination by Sparta their past" (Lazenby Hope Asklepieion ; see, for to s see amounted the "loss" of Messenia Simpson 1972, p. 81); also Pearson ple, Alcock 2002a, pp. 164-173, and history: "In effect, Messenia had ceased 1962; cf.Alcock 1999. further discussion in this article, below. to a own 4. is have history of her before the The monumental display best i5o SUSAN E. ALCOCK ET AL

Pylos / 11 Koryfasion? PRifUPfcgGION and areas walked

IONIAN SEA ! Methoni Figure 3.Map indicating the survey territories of UMME (with subre gions indicated), the Five Rivers survey, and PRAP. R. J. Robertson

can an as What intensive regional analysis, such PRAP's, offer to the are study of this unusual region? We in the strongest position to address not questions typically posed inMessenian historical scholarship, revolv as ing they do around life "on the ground" and not as viewed from an was external perspective. For example: What the nature of settlement and, by inference, of social organization and community structure in the re gion? How did alterations in political and administrative regimes affect where and how people lived?What economic change can be detected in the can to region, and how it be related the shifting political history? What contact are types of external visible in the regions material culture and what impact might such exchange have had upon local organization? What evidence emerges for ritual practice in this landscape? And how isMessenia or like, unlike, other parts of the Greek peninsula? answers to were The beginnings of these questions advanced by the University ofMinnesota Messenia Expedition (UMME).This multistage, an multidisciplinary project published, in 1972, the results of extensive surface of a to west exploration very large area, bounded the and south by to east the sea, the by the Mountains, and to the north by the an area ca. Alpheios valley andTriphylia?in all, of 3,800 km2 (Fig. 3).5 In later stages, the project moved on to excavation and more localized survey at the site of Nichoria and its hinterland (the Five Rivers survey). Re were a gional data presented and analyzed in fashion admirable for its day, 5.McDonald and Rapp 1972. HISTORICAL messenia ?51

ar and the project has proved highly influential in the field of classical were chaeology.6 Survey directors honest, however, about their ability to our deal with the historical periods: "the focus of this study and of earlier on publications has been consistently the earlier material"; "our surface concen search and the general emphasis throughout this book has been on trated the ."7 Moreover, UMME's extensive methods of reconnaissance left room to wonder whether the full spectrum of human to activity in the region?from major settlements small and isolated rural structures?could possibly have been observed.8 an a Perceiving PRAP as both "heir" to and as further stage in the a more work begun by UMME, PRAP investigators chose to follow in to tensive field methodology, and involve several ceramicists specializing to reassess in historical material, in order counter these problems and to cut UMME results. Potential benefits both ways, of course; compared with a more the wide-ranging earlier project, PRAP examined far restricted some territory, its intensively explored 40 km2 comprising only 1 percent of UMME's 3,800 km2 (see Fig. 3). Both investigations have something, if not same to as as the thing, offer, and throughout this report, far possible, UMME and PRAP results will be compared and integrated. we Before proceeding, should raise two important caveats about the present study. First, it should be admitted that knowledge of historical Messenian ceramic material?as with surface our always survey, principal source of evidence?remains limited, a circumstance due in part to the relative paucity of relevant excavated sites. Exacerbating this predicament, our were many of finds in poor condition, making identification difficult despite the best efforts of project ceramicists. Pieces that could be dated no more to or precisely than the "historical" phase, that remained simply our more "unknown," made up, in all, nearly 30 percent of than 20,000 or catalogued ceramic finds. The presence of large amounts of undatable, a to only broadly datable, pottery is, of course, problem familiar allMedi terranean not us surveys, and all of believe that these problems in dating entirely invalidate the results presented below. But the issue of chrono must logical imprecision be borne in mind when considering the data and when employing them in further investigations.9 Second, although PRAP shows promise of a more detailed examina tion of the regional landscape than has been previously possible, its spe not across zone cific results should be too readily extrapolated the entire we to as to refer today "Messenia." It is necessary keep in mind the par ticular, local characteristics of the survey territory: its location at the far western edge of the P?loponn?se (Fig. 1), its distance from the dominant Hellenistic and Roman city of Messene, and its more restricted fertility

6. For see major UMME publications, 1972, pp. 123,143, respectively. This reconnaissance methods, Davis et see Mc is less but not McDonald and Rapp 1972; marked, entirely al. 1997, p. 395; Cherry 1983, p. 393. I?III. in the a Donald 1984;Nichoria Spencer overcome, project's subsequent For comparison of UMME and sets in for Lukermann PRAP see (1998) the project historical work; see, example, methodologies, Bennet a recent see in context; for critique, Foti and Moody's interest later historical 1998. adis 1995. periods (1978, pp. 82, 84). 9. For more critical discussion of 7.McDonald and 8. For assessments of UMME's these see the Hope Simpson matters, Appendix, below. 152 SUSANE.ALCOCKETAL.

TABLE 1. CHRONOLOGICAL SCHEMAS EMPLOYED BY UMME AND PRAP

UMME (McDonald and Rapp 1972) Geometric Ca. 900-700 b.c.

Archaic Ca. 700-500 b.c. Classical Ca. 500-323 b.c.

Hellenistic Ca. 323-146 b.c.

Roman Ca. 146 b.c-a.d. 330 a.d. Byzantine Ca. 330-900

PRAP (Davis 1998c) Geometric Ca. 900-700 b.c.

Archaic Ca. 700-480 b.c.

Classical Ca. 480-323 b.c.

Hellenistic Ca. 323-31 b.c.

Roman Ca. 31 b.c.-a.d. 700 b.c.-a.d. Early Roman Ca. 31 400 Late Roman Ca. a.d. 400-700

as compared with that of other Messenian zones, such the River or as a ne valley the Stenyclarian plain.10 IfMessenia whole has been area come glected by historians and archaeologists, the of the survey has in even one to for less remark. M. N. Valmin, of the few explore it, observed in 1930:

La c?te ouest de laMess?nie, entre l'embouchure de laN?da et a a Mothon, ?t? toujours trait?e d'une fa?on superficielle. On pr?tendu que pendant la domination elle ?tait rest?e comme et d?serte, qu'elle avait m?me ?t? utilis?e p?turage qu'elle au vue et devait offrir peu d'int?r?t point de ?ologique topographique.11

For J. G. Frazer, following in the footsteps of , the western coast was "of extremely uniform and monotonous character.... On all this line not a more areas of coast Pausanias mentions single place."12 The inland of not even the PRAP study region do receive this level of commentary, making even more any possible archaeological testimony desirable. The time span under study in this article is divided into two parts in the following discussion: Geometric through Classical, and Hellenistic through Late Roman (see Table l).The first division represents the period after the end of Mycenaean palatial culture and immediately before and during the period of Spartan domination, the second the centuries that followed the liberation and its existence under Roman region's subsequent 10. On of specific characteristics control. This structure is but it allows the bipartite admittedly very crude, the PRAP survey area, see Davis et al. to to complicated process of landscape change be related the region's shift 1997, pp. 398-400; Davis 1998b, et al. ing political and economic circumstances. Both sections of the discussion pp. xxix-xxxiv; Zangger 1997, are a pp. 549-553. organized in the following way. Presented first is skeletal historical 11. Valmin 126. For a more of Messenia and of the PRAP area in the time 1930, p. summary study specified treatment s general of Valmin research to establish a framework for the reader and to period, designed prompt inMessenia, see Valmin 1938. relevant with to the evidence. An overview of the questions respect survey 12. Frazer 1965, vol. 3, p. 462; see as next also Alcock 146-149. regional picture, sketched by UMME, is reviewed; it should be 2001, pp. HISTORICAL MESSENIA 153

are a noted that the chronological categories employed by UMME not direct match for those of PRAP (for the distinctions, see Table 1). An overview of the PRAP data is then presented, highlighting artifact dis more tributions, settlement patterns, and other significant trends; spe are on cific details of many categories of basic information that presented are not cat the online PRAP site gazetteer repeated here.13 These three are so egories of evidence presented in turn that overlaps and contradic can are tions among them emerge, and other scholars encouraged to take own ar the elements provided here and evolve their reconstructions. The a ticle ends with consideration of the region's "sacred landscape" before and some after liberation and with brief remarks comparing the survey evi one dence for this part of Messenia to other, very relevant region of Greece? itself.

GEOMETRIC THROUGH CLASSICAL (CA. 900-323 B.C.)

Historical Background: Messenia

The story of Spartan expansion and Messenian subjugation has been told and retold many times.14 "Story" is the operative word, since the exact course of theMessenian Wars of Spartan conquest, and, indeed, their very a historicity, remain clouded issue, with all specific points hotly debated. causes were Various of the conflict between Sparta and Messenia adum brated in antiquity, from mutual accusations of impiety at the border sanc tuary of Limnatis to Spartan greed for the fertile lands of "broad Messene, Messene rich for plowing, rich for planting."15 Although dates to vary from authority authority, the is traditionally placed in the later eighth century (in archaeological terms, roughly the Late Geometric period), the Second in the first half of the seventh century accounts (the Early Archaic period). Much later textual tell of desperate names conflict, occasional treachery, and great heroism, invoking places and 13. http://river.blg.uc.edu/prap/ such as , , Aristocrates, , and the Battle of the "Great PRAP.html. The online gazetteer, terms Trench." In of territorial expansion, although certainty is impos which will be available in ultimately it seems most that extended control numerous cate sible, likely Sparta gradually, begin printed form, presents with the and then to gories of evidence, including site loca ning perhaps Stenyclarian plains, moving tion and col areas further west. With the land was divided into or description, dimensions, annexation, kleroi, lection intervisible and re strategies, sites, land allotments, for the convenient organization and manipulation of an artifact as well as access to sources summary, and of manpower. online databases of finds and of photo After final victory, some of the of the sites and their features. Sparta's proportion region's population graphs is said to have fled their exiles who none 14. source homeland, becoming long-term The principal ancient is it is never lost their sense of Messenian Those Paus. 4; commentaries by J. G. Frazer theless, claimed, identity. or (1965) and by D. Musti andM. Torelli who lacked the ability, the resources, the will to flee remained and were For see Kiechle (1991). discussions, transformed into the famous helots of antiquity, whose labor underpinned 1959; Huxley 1962; Oliva 1971, the militaristic Spartiate way of life. Helotage existed in both Laconia and pp. 102-114; Lazenby and Hope Simp Messenia, though it followed quite different trajectories in the two regions. son 1972, pp. 84-86; Cartledge [1979] The Messenian helots worked as Harrison and 1998. labor 2001; Spencer probably principally agricultural trans. in while their Laconian in closer to 15. , fr. 3, J. Bennet ers, counterparts, dwelling proximity Harrison and 151. their serve in more as Spencer 1998, p. masters, could other, direct fashions, such personal 154 SUSANE.ALCOCKETAL.

or as servants military support. The precise status of helots has been bit one was terly debated and remains deeply obscure, but ancient suggestion men to place them "between free and slaves."16 so com This review, far, has closely reflected the standard scholarly a munis opinio, belief inMessenian "wars" and helot enslavement through Spartan military conquest and annexation. This position has recently been a our challenged by Nino Luraghi, who argues that critical examination of admittedly poor sources signally fails to support any such scenario of mass enslavement of an indigenous united population. Instead, he suggests that means helots were created through the less violent, and less glorious, of their being descendants of slaves, bondsmen, and other dependent labor ers. was a Their "imaginary conquest," in this formulation, tradition in never vented much later and made entirely unitary.17 The debate contin our ues, and is unlikely to be assisted by particular regional perspective. not a What is challenged here, however, is the presence of dependent popu lation inMessenia inArchaic and Classical times, and the need to investi gate their local conditions of existence. some Our textual sources, despite their thinness, do offer observations a relevant to this inquiry. If helots existed in state of collective servitude to the Spartiate class, they also significantly outnumbered that group?which, more over notoriously, diminished still in size time?though perhaps not once to the remarkable degree envisioned.18 Helots, most would agree, some to appear to have possessed limited rights property?in particular, some cases links to specific plots of land, perhaps in plots that they had seem to some previously owned. They have lived in form of self-reproduc to at at 16. Poll. Onom. trans. ing family units. They could claim the right asylum, least the Sanc 3.83, P. 2001, 303). tuary of at Laconian Tainaron, and, under Spartan rule, they Cartledge ([1979] p. For reviews of helot status maintained certain cults and ritual observances.19 many would agree general Finally, and see Ducat or came some sense conditions, 1974,1978, that helots possessed, to possess, shared of origins and 1990; Cartledge [1979] 2001, pp. 160 of rise to a common sense of Messenian All heritage, giving identity.20 177, 347-356; 1987, pp. 170-177; a to these factors played part in shaping the helot response servitude and 2001; Hornblower 2000, pp. 68-73; Powell the ways inwhich they inhabited and exploited theMessenian landscape. 1988, pp. 248-252; Whitby 1994; Ste. Croix 2002. One intriguing aspect of the helot response, quite differ particularly 17. for was Luraghi 2002,2003a, 2003b; ent from that of chattel slaves, their apparent ability to organize resis an view, seeWees 2003. tance. uncertain to us in their opposing of Messenian rebellion, two recent of Episodes though 18. For appraisals and are in several ancient sources. The see frequency precise nature, reported helot demography, Figueira 2003; a best attested instance of rebellion is major revolt in the 460s B.c., when Scheidel 2003; cf. Cartledge 1987, a 37-40, 4.2. Messenians rose up, taking advantage of devastating Laconian earth pp. fig. on a 19. For further discussion of helot quake, and defiantly took refuge Mount Ithome.21 After siege?which, cults and ritual observances, see in one lasted for ten rebel Messenians were allowed to version, years?the in "Change the Ritual and Mortuary on condition never set foot in the Some leave, they P?loponn?se again. Landscape," below. from this wave of exiles were resettled the Athenians at on 20. 113 by Naupactus E.g., Hodkinson 2000, pp. 129. For a recent reconsideration of the Corinthian Gulf. Men from this community later returned to theMes Messenian see 1999. senian homeland as part of the Athenian campaign at Pylos (425 B.c.). identity, Figueira 21. Ancient references to the fifth unrest to During the , helot contributed Spartan century rebellion: Thuc. 1.101.2-103.3; and rose in rebellion?one final time?when troubles, they Epaminondas, Diod. Sic. 11.63-64; Plut. Cim. 16.4; after his at invaded Laconia and termi shortly victory (371 b.c.), Paus. 1.29.8,4.24.6, 7.25.3; Ael. VH nated Spartan control of Messenia. 6.7. HISTORICAL MESSENIA 155

was The effect of the "helot danger" noted by contemporary and later observers: (4.80.2) remarked that "most Spartan institutions a have always been designed with view to security against the Helots"; the were an helots "like enemy constantly sitting in wait for the disasters of to the Spartans" (Arist. Pol. 1269a37-39).22 Sources refer various Spartan tactics, initiated in response, that were intended to maintain order, to in to still fear, and sharply demarcate Spartiate from helot, including arbitrary murders and beatings, humiliation and abuse, and the wearing of distinctive were clothing. It has been debated how far such acts everyday occurrences, were more and it is likely that Laconian helots far vulnerable than those at a greater distance from Sparta. For all that, Sparta's heavy dependence upon helot agricultural labor suggests that the need to maintain order would terror are not a have been paramount; strategies of unlikely for ruling group to a attempting keep control of large and distant subject population. to to The Spartan need for helots farm and drudge provides, of course, reason a the for their having held such tiger by the tail. How helots met their obligations is nowhere detailed, but recent r??valuation, by Stephen a Hodkinson, of the sparse available data makes convincing case for a sys tem of sharecropping. Sharecropping, he argues, would have divided be tween the two parties, cultivator and landowner, the unavoidable risks of a a farming in Mediterranean climate, effecting balance that offered "the more secure was arrangement for the Spartiate-helot relationship which characterized by long-term mutual interdependence."23 This model chal lenges traditional thinking about the relationship between and on helots, which is usually viewed in entirely negative terms, marked the one on hand by Spartiate depersonalization and humiliation, and the other to by helot willingness "eat [the Spartiates]?even raw."24A further impli some cation of this model is the probability of degree of hierarchy in helot a society, given such system's need for "middlemen." Moreover, given the need for articulation and negotiation between helot and master, visions of as helot life entirely egalitarian?with everyone equally helpless?become more increasingly implausible. A nuanced picture, both of helot society can a and of Spartiate-helot relations, only emerge from broadly based 22. Trans. P. Cartledge ([1979] understanding of how the relationship between Spartiates and helots played 2001, pp. 299, 306). For a more mini out on the in terms of settlement see ground, patterns, malist view of the "helot danger," community organiza and ritual behavior. Roobaert 1977;Talbert 1989;Whitby tion, 107-111. must 1994, pp. While helots have constituted the majority of the inhabitants of 23. Hodkinson 131. 1992, p. See, Archaic and Classical Messenia, the presence of other elements should be for the broader discussion, Hodkinson more com noted. For example, perioikic communities, although much 2000, pp. 125-131. mon in Laconia, did appear further west. Perioikoi were a within the 24. Xen. Hell. 3.3.6, trans. P. Cart group of and not to the but at ledge ([1979] 2001, p. 306). body Lacedaemonians, separate equal Spartiates, same time a 25. Shipley 1997. For further dis the distinct from the oppressed helots. In recent, thorough cussion of the of Messenia and at poleis review, Graham Shipley identifies, most, five securely attested perioikic see 2004b. Laconia, Shipley 2004a, inMessenia: Aithaia, , Aulon, andThouria.25 26. see poleis Cardamyle, Only Onperioikoi, Shipley 1997; one of these?Thouria?definitely lay inland (see Fig. 1). In the majority also Roebuck 1941, pp. 28-31; Kiechle of instances, the remained to in 1959, pp. 68-71; Cartledge [1979] perioikoi loyal Spartan authority, although the ca. to 2001, pp. 178-193; Powell 1988, revolt of 460 b.c., certain communities chose join the underdog pp. 247-248. helots in rebellion.26 SUSANE.ALCOCKETAL. 156

To round out the picture of the regions population, the communities were of Asine and Methoni supposedly the result of Spartan resettlement at of refugees from the Argolid. Asine, least in the fourth century, also a to a held Spartan garrison, and Xenophon {Hell. 7.1.25) alludes Spartan were colony at Pherae. All these settlements coastal in orientation.27 From this very bare-bones sketch, certain problems emerge that clearly recommend archaeological examination. Although contemporary and later sources speak to the fate of Messenian exiles, they say almost nothing, apart from reporting rebellion, about those who remained inMessenia under Spartan rule. Yet moments of revolt hardly represent everything there is to understand about the organization of helot life; anthropologists as such Jean Comaroff and James Scott have warned about the dangers of on moments a s concentrating only violent in people ongoing struggle with can coercion and exploitation.28 Regional archaeology contribute further evidence about the day-to-day existence of these people: where and how they lived, what factors enabled communication, what rituals encouraged identity and solidarity. In other words, evidence from theMessenian land scape can outline the structures of life that encouraged helot accommoda tion and survival, and that allowed helot rebellion.

Historical Background: The PRAP Study Area

as a As has already been suggested above, if historical Messenia whole remains little known, the western side of the peninsula might be consid a more can ered the forgotten of the forgotten. Nevertheless, few details be added about the specific territory investigated by PRAP. From the Spar at tan perspective, this territory lay the far end of their Messenian hold some as crow ings, the modern town of Hora lying 65 km, the flies, from no conse the modern community of Sparta. This fact of distance doubt had we ac quences for this section of westernmost Messenia. For example, if wars area cept the historicity of the of conquest, absorption of this into areas to east. Spartan territory probably occurred later than that of the an or There arises, moreover, the question of just how close eye, how harsh a on zone. hand, the Spartiates could have kept this particular Distance, not to mention the impressive barrier of theTaygetus mountain range, has as a often been invoked general point to remember when assessing the logistics and practicalities of Spartan-Messenian relations.29 a most Distance from Sparta certainly played part in the dramatic and area: best-documented events in this particular the establishment in 425 B.c. of the Athenian base at Pylos (called Koryfasion by the Laconians), on 27. Asine andMethoni: Hall the entrapment of Spartan forces , and the subsequent guer 1995, was over to 581-584; 1997, 74-75, 77; rilla-style raiding from Pylos until the phrourion finally given pp. pp. Malkin 1994, pp. 83-89. Asine and the in 409/8 b.c. The arguments of the Athenian Demos Spartans general Pherae: and thenes in of this on theMessenian coast Lazenby Hope Simpson support fortifying particular point 1972, pp. 84, 86; Roebuck 1941, p. 30. line had met with a marked lack of enthusiasm: initially 28. Comaroff 1985, p. 261; Scott 1985. were The generals said that there plenty of deserted promontories 29. Valmin 1930, pp. 14-15; Huxley in the if he wanted to one and the to P?loponn?se, occupy put city 1962, pp. 59-60; Lazenby andHope but that this had expense; thought place important Simpson 1972, pp. 85-86; Cartledge was a 46. differences when compared with any other. There harbour [1979] 2001, p. HISTORICAL MESSENIA 157

next to it, and the Messenians had been natives of the land in the a old days: they spoke dialect similar to the Lacedaemonians', and a would be able to do them great deal of damage if they used Pylos a a as base?also, they would make reliable garrison for it.30

Demosthenes' predictions proved to be correct. nexus This of events raises interesting questions about the complex interactions among local inhabitants, anti-Spartan forces, and the Lacedae are monians themselves. Some helots said to have rebelled and joined the garrison in their harassing activities; others aided the trapped Spartans by to at bringing food the island, personal risk; and "those of the perioikoi came who lived nearest" to help in ejecting the intruder.31 Precisely where were near these perioikoi and these helots coming from is uncertain. The est definitely attested perioikic community isAulon, far to the north by the ca. River; the "refugee" settlement atMethoni lay 15 km to the south. Of Pylos, Thucydides remarks that "not only the place itself but also a was the neighboring country for considerable distance unoccupied (ere seem was mos)."32 It would that it this supposedly undeveloped state that so a on allowed the Athenians easily to gain toehold. Much has been built this observation of Thucydides'. Indeed, the presumed "emptiness" has led scholars to suggest that the western marches of Messenia were left as a or even pasture and hunting reserve, abandoned, left devastated, after the a Spartan conquest.33 While there may well have been falling off in the intensity of Spartan exploitation the further the distance from Laconia, more evidence is required before the PRAP study area can be securely written off as awild preserve or an empty quarter. After 409/8 b.c., Pylos/Koryfasion may well have remained in Spartan as a hands, possibly perioikic community, even for a short time after Mes are senia's general liberation. We told that Arcadians took the settlement, over and that they gave it to theMessenians in 365 b.c. (Diod. Sic. 15.77.4). Methoni may have remained outside Messenian control until even later? possibly until the Battle of Chaironeia (338 b.c.), when it became part of to Philip Ils territorial reward his Messenian allies.34 Amid this gradual cen and slightly patchy relinquishing of Spartan control, the mid-fourth must a tury have been time of significant social and economic upheaval for area as the within the territory investigated by PRAP, also for the rest of Messenia. An important part of this upheaval would have been the emer of as an never gence Pylos/Koryfasion independent polis?though not, and as a as again, notorious site its incarnation during the Peloponnesian War.35

30. trans. B. see see Thuc. 4.3.3, J. Wilson Pritchett 1965;Wilson 1979. margin (Thuc. 4.8.6); Wilson 1979, (1979, p. 7). 32. Thuc. 4.3.2-3, trans. C. F. Smith, pp. 52-53. 31. trans. 1930. Thuc. 4.8.1, J. B.Wil Cambridge, Mass., 34. On the actions of the Arcadians, son (1979, p. 9). Events at Pylos and 33. Valmin 1930, pp. 22-23 (noting see Roebuck 1941, p. 38 and n. 62; on are see Sphacteria narrated by Thucydides the prevailing view, though Valmin Methoni, Shipley 1997, pp. 242-243, on 265. In on see (book 4): helots aiding theAthe himself believed otherwise); Roebuck general, this period, Laz on helots the and nians, 4.41.2-3; aiding 1945, p. 151; Kiechle 1959, p. 55; Hux enby Hope Simpson 1972, pp. 89-90. on 35. For to Spartans, 4.26.5-9; theperioikoi, ley 1962, p. 60. Thucydides' grasp of guides Koryfasion that 4.8.1. For other ancient the is on evi review testimony, region's topography other archaeological finds in the area, see Xen. Hell. see 1.2.18; Diod. Sic. dence rather suspect; he misjudges the Baltas 1987, esp. fig. 2; 1997, 13.64.5. events of a On the of 425 b.c., length Sphacteria by considerable pp. 115-120. 158 SUSAN E. ALCOCK ET AL.

on It remains difficult, based these pieces of information alone, to answer the most fundamental of questions about this area of westernmost Messenia in the Geometric through Classical periods: who, if anyone, lived area in this at this time? From among the various possible answers, most quickly able to be discarded is that of Spartiate inhabitants. Neither has perioikic settlement been securely attested within the PRAP study region, at a except possibly Pylos/Koryfasion for brief period in the decades im to area was mediately prior Messenian liberation. The possibility that the to a as inhabited by satellite populations linked community such Cyparissia or Methoni (both of which may possibly have held perioikic status for part of this time span) isworth considering, but would presume the existence are means a of extremely extensive territories. We left, by of process of elimination that is admittedly less than ideal, with the likelihood that the was population in the PRAP study region of helot status. This suggestion can as perhaps be taken particularly likely for the earlier part of this time more span and for the inland parts of the region.

Overview of UMME Patterns

Certain areas within the UMME territory appear consistently utilized from as the Geometric through Hellenistic periods, such the Stenyclarian plain, the Pamisos valley, the Alpheios valley, and Triphylia. The especial fertility of these zones may well account for this persistent use, but investigatory con bias?for example, the attractive qualities of Olympia?must also be a sidered possible factor affecting this observation. area Evidence for the Geometric period (ca. 900-700 B.c.) within the comes of UMME investigation disproportionately from tombs (for the see numbers and locations of the Geometric sites identified by UMME, more rare: Figs. 4, 5).36 Settlement evidence is two of the Geometric loci area reported by UMME within the overall of the PRAP study appear to more be habitation sites.37 Closer investigation, through excavation and intensive work carried out UMME in the Nichoria likewise 4 by environs, 36. The data presented in Figures and 5 are derived from McDonald and testified to sparse occupation; the project directors proposed that Nichoria's a as a to 1972, pp. 264-321 ("RegisterA: position, within "buffer zone" to Sparta, acted disincentive activ Rapp on was Prehistoric Habitation Sites"; "Register ity in this area.38 Settlement the Nichoria ridge itself abandoned B: Habitation Sites"). the the of Post-Mycenaean before century, presence Late Geometric 4 mid-eighth although Figures and 5 include all sites, defi indicates that the had not been The ex graves territory entirely given up. nite and possible, assigned by UMME cavators a on to the dates indicated. In suggest that the positioning of Late Geometric burial Figure 5, a to east dots indicate that a the ridge, from which there is fine view both the and south, "may overlapping partic site more one a ular possessed than iden reflect conscious wish the living to enlist the dead hero's continued by tifiable function. across vigilance to help them ward off the incursions of the hated foe from 37. UMME 1 Chora: Ano Englia Such use of heroic associations have been nec Taygetus."39 symbolic may nos, investigated by PRAP as site B07 for?on the whole?the UMME evidence fails to Palace Nestor Lower essary, survey support (Hora of Town); a UMME 35 Koukoun?ra: Katar?chi. the picture, derived from later literary sources, of largeMessenian popu McDonald and 1972, lation rising up in opposition to invasion. Hope Simpson 144. The Archaic 700-500 a clear increase p. period (ca. b.c.) produced very 38.McDonald and Coulson 1983 in the number of discernible sites 4, 5). In the PRAP (Figs. study region, ("buffer zone," p. 328). two to area specifically, site numbers rise from five. In the broader of 39.McDonald and Coulson 1983, a as to 326. UMME investigation, greater proportion of settlements, opposed p. HISTORICAL MESSENIA 159

Habitation j

||fe| Cemetery

^| Shrine

J_ 75 E

Figure 4. UMME site numbers (habitation, cemetery, shrine), Geometric Archaic Classical Hellenistic Roman Geometric-Roman. R. J. Robertson

a across burial sites, is represented, and there is shift in their distribution the landscape. By contrast with sites of the Geometric period, 63 percent of which produced traces of earlier, prehistoric occupation, only 36 per cent of the Archaic sites show such indications. Establishment of new 40. UMME 504 Andhr?as: new Ayios sites in locations, coupled with the abandonment of older locales? Korythos; UMME 530 Ayios from whatever cause?is a measure, however crude, of settle Floros: Pamisos. resulting ment Another new feature in the is the 41. The decrease in size of settle disruption. landscape appearance ments was to have been of sanctuaries.40 suspected par recognizable sites in the are ticularly characteristic of Archaic settlements described by McDonald and Hope Simpson Pamisos zone, in contrast to as an valley being "rather small," expected result of the loss of freedom following other of the parts Messenia, including the ,41 but the characterization is not borne remote entirely "more areas" of Pylos and Cypa out by the data they themselves present for estimates of site size. Although rissia: Lazenby and Hope Simpson such estimates must be taken with a considerable of salt?for 1972, p. 94;McDonald and Hope grain they all surface scatters from that often a num Simpson 1972, pp. 144-145. lump together occupations span 42. This calculation is based on mean area ber of periods?the size for sites in the UMME study with McDonald and 1972, pp. 264 Rapp reported Archaic occupation is approximately 3 ha (median 2 ha;N= 13).42 and uses data for sites with a 321, only Such sizes would be as toward the existence definite attestation of Archaic material. normally interpreted pointing of often denoted as hamlets or rather than 43. For a recent discussion of the community dwellings, villages, see isolated structures inhabited one or two extended families. These latter classical farmstead, Pettegrew 2001. by or are 44. McDonald and Hope Simpson sites, which might be either permanent farmsteads seasonal shelters, 1972, 145. a p. usually assessed in size range of up to approximately 0.5 ha.43 45. These distinctions are based In their discussion, UMME chose to combine on to site synthetic investigators the periods assigned each the Classical and Hellenistic results (ca. 500-146 b.c.). As was noted above, in Registers A and B (McDonald cor were honest about their lack of confidence in and and Rapp 1972, pp. 264-321), they dating interpreting historical seem to have believed that the rected where possible against the later material; moreover, they "major Lukermann and work of Moody cultural and economic changes" resulting from Messenia's liberation (1978). The of the data presentation would only gradually have become apparent.44 And yet, despite these dis so is, however, telegraphic that this some claimers, the UMME gazetteer of sites and pocket maps do allow for of UMME results, mapping particular two and on that we be to differentiation between the basis, report, ly with regard site function, should periods, seen as than a number of distinctions between "Classical" and "Hellenistic" UMME be impressionistic rather low, exact. a data.45 To great extent, however, Messenia's liberation from Sparta and i6o SUSAN E. ALCOCK ET AL.

Geometric Archaic

Cemetery Cemetery A Shrine A Shrine

Classical Hellenistic

Figure 5. Distribution of UMME ' sites: Geometric, Archaic, Classical, Hellenistic. R. J. Robertson HISTORICAL MESSENIA l6l

TABLE 2. NUMBERS OF UMME SITES

Subregion PeriodSouthwest Southeast Center North Total

Geometric 3 2 9 5 19 Archaic 6 6 15 15 42 Classical/Hellenistic 26 4819 13643 Roman 22 31 14 22 89

re the founding of Messene and other poleis?major disjunctions in the gional long-term history?remain masked in the UMME results. attest a Data from the UMME project marked increase in the number area of sites in the in the Classical/, with identified habi more tation sites increasing than threefold over Archaic numbers. The on summary picture sketched by the UMME investigators is, the whole, cheerful: "The general impression is that, despite the known raids and was a military campaigns, this relatively settled and prosperous epoch."46 In early presentations of UMME results, the investigators appear by and or large unwilling to "see" "read" the fact of Spartan control in the land an scape, avoidance that is somewhat mitigated in the project's later pub zones to see lications. The richer agricultural of Messenia continue high a levels of activity, and the southwest district, in particular, presents "strik a an ing increase" in the number of sites, jump from six to twenty-six, over to increase of 4 1 (Table 2).47 In the PRAP study area, the number of more sites identified by UMME than triples from the Archaic to Classi cal/Hellenistic periods, increasing from five to sixteen, of which six from were the latter period intensively reinvestigated by PRAP.48 Apart from this general florescence in site numbers, three other trends can be noted in this Classical/Hellenistic patterning. First is an increasing more or preference for accessible, lower coastal locations, especially along more the western seaboard; stable political conditions and better regulated commerce are the explanations given for this phenomenon.49 The second a development is heavier utilization and occupation of the area around a Mount Ithome, the site of Messene. Third, in the Hellenistic period, across slight increase in site numbers has been observed the region at large and also in the specific PRAP study region, where the number of recog to nized sites expands from nine twelve (Figs. 4, 5).Work in the hinter an land of Nichoria likewise has suggested "increased tempo" of activity in to Hellenistic times ("no doubt owing the nationalistic euphoria following the expulsion of the Spartans"), although the acknowledged difficulties in

46.McDonald andHope Simpson 48. UMME 3 Korifasion: Portes Hope Simpson 1972, p. 94.While raw 1972, p. 145. (PRAP103); UMME 11Tragana: percentages of coastal sites do not 47.McDonald andHope Simpson Viglitsa (106);UMME 15 Gargaliani: alter all that much from the Geometric 145. Table 2 the num 1972, p. presents Kanalos (D01); UMME 400 Roma period onward (see Lukermann and nos bers reported byMcDonald andHope Viglitsa (104);UMME 403 Tra Moody 1978, p. 96, figs. 7-12), Mc which are at Rachi Donald and Simpson (1972, p. 146), gana:Tsopani (117);UMME Hope Simpson argue that from the data mined 406 there is a new on points discrepant Marathoupolis: Dhialiskari (G01). striking emphasis from the UMME gazetteer by the 49.McDonald andHope Simpson such locations in Classical/Hellenistic authors see of this article. 1972, p. 145; also Lazenby and times. l62 SUSAN E. ALCOCK ET AL.

distinguishing between Classical and Hellenistic surface sherds temper the claim.50 On historical grounds, all of these developments?coastal settle on of settlement?make better sense ment, emphasis Messene, expansion one more con for the period following liberation than for the before, but clusive independent archaeological confirmation remains desirable. Without directly approaching the question of where Messenian helots same as lived, UMME investigators leaned generally in the direction the majority of ancient historians who had considered this issue: the helots were one on dispersed, isolated from another their individual kleroi. This on a on picture is based handful of ambiguous textual references and the assumption that Sparta would have wished to "divide and conquer" their troublesome populations.51 The Five Rivers survey of the Nichoria hinter land, for example, stood firmly in this camp:

Given its fertility, the Five Rivers area was probably among those lands initially divided into helot farmsteads. Indeed, under the helot was to to system, it Spartas advantage keep the helot population even scattered throughout the countryside, perhaps restricting the an or size of helot communities to extended family often fifteen members; ideally this would inhibit the Messenian proclivity to resist and revolt against their Spartan overlords.52

Despite such hypotheses, the Five Rivers survey identified few sites of the none Archaic to Classical periods, and of the size of small farmsteads that might be correlated with helot families. In the Nichoria area, sites of such were size dated, though with caution expressed about both date and func tion, to the Hellenistic era.53 The sizes of Archaic sites reported by UMME point away from any model of profound rural dispersal. The evidence, however tentative, for larger groupings of people in the landscape continues in the Classical/ Hellenistic period, if less markedly (mean site size 2.2 ha, median 1.7 ha; = N 54).54 These results contrast with those of almost every other survey set data for Greece in the Archaic through Hellenistic periods, which fea numerous to ture the discovery of small farmsteads, if differing degrees and with differing temporal patterns.55 The extensive methodology and prehistoric focus of UMME, however, left open the question of whether a or Messenia represents genuinely different settlement trajectory, whether smaller sites had simply been overlooked.

in 50. Coulson andWilkie 1983, p. 95. Coulson andWilkie (1983, nitely occupied the Classical period, see state "the evidence the Hellenistic or both. Re p. 337; also Lukermann and Moody p. 337) that meager period, 1978, p. 96. from Nichoria and its environs indi ported sites specifically identified with cates times but not with Clas 51. Ancient citations include Thuc. that throughout Classical Hellenistic material, the local was still tend to be smaller 1.2 5.34, 8.5.4, and 34.27.9; population mainly sical, (mean ha, see with farmsteads and small shrines median 1 A^= than those re also Lazenby and Hope Simpson rural, ha; 13) the for the broader Classical/Hel 1972, p. 94; Lukermann and Moody scattered about countryside." ported an 1978, p. 95; Coulson andWilkie 1983, 53. Lukermann and Moody 1978, lenistic period, indication, observed n. Coulson and Wilkie also at that smaller sites p. 337. The possibility of helot aggre p. 105, 31; Nichoria, in this gation is envisioned by Lotze (1959, 1983, p. 332; Harrison and Spencer emerge only later period. p. 38). For a review of these arguments, 1998, p. 160. 55. See, e.g., Snodgrass 1990; Alcock, see on and Davis Harrison and Spencer 1998, 54. Based McDonald and Rapp Cherry, 1994, pp. 157-165; em van pp. 159-161. 1972, pp. 264-321, these figures Jameson, Runnels, and Andel 1994, and sites as 52. Lukermann Moody 1978, ploy data for identified defi pp. 248-257, 383-394; Pettegrew2001. HISTORICAL messenia 163

Overview of PRAP Patterns

an The figures in Tables 3 and 4 suggest increase in regional activity fol lowing the Geometric period, with relatively constant numbers of sherds definitely datable to the Archaic and to the Classical periods.56 The major occurs disjunction in this sequence between the Classical and Hellenistic area more periods, when definitely dated sherds within the PRAP survey than triple in number. The nature of that break will be discussed further area can no means as a below. While the by be characterized "deserted" was landscape (asThucydides had suggested true in the late fifth century), material of Archaic and Classical date is relatively limited in comparison with that of later historical periods. The nature of the ceramic evidence some may very well play part in this apparent change (see the Appendix, at below). Palynological evidence, least for the latter end of this time span, certainly suggests the conducting of active cultivation.57 areas The nine discrete intensively surveyed by PRAP include coastal or an near-coastal plains and low knolls (areas I, II, III, V, VI), upland a plateau (IV), ridge structure (VII), and inland valleys (VIII, IX) (Fig. 6). sur Conditions of bedrock and soil, which affect the stability of ancient faces and their ability to reflect distributions of artifacts of the past, vary significantly among these areas, ranging from highly degraded marl ridges to marine terraces that can still preserve Paleolithic artifacts in situ. In very general terms, the coastal transects I, II, III, and V and the inland and zones upland IV, VIII, and IX may be considered relatively stable surfaces for the historical periods, and any absence of Archaic-Classical and later areas cannot to material in these automatically be attributed geomorpho at on logical change. Area VI, least part of which lies Holocene alluvium, area are more and VII, which follows the Englianos ridge, problematic. zones area as a The marl of VII, which "have changed significantly result of natural landscape instability and anthropogenically accelerated erosion,"58 as may well impede recognition of small, possibly historical sites, such COI (Tt?ig2in.2iAlafinoremd) and C02 (Tragana^/?z?/z7z?z) (see Fig. 2), where on activity the ridge is indicated from Archaic through Hellenistic times but where few specifically diagnostic artifacts have been recovered.59 Definitely dated Geometric material is found in eight of the nine ar eas areas (Fig. 7), in greatest, and fairly equal, amounts in I, VI, VII (28, 24, and 29 percent, respectively).60 The subsequent Archaic and Classical centuries under control?show a more periods?the presumed Spartan distinct "clustering" of sherd scatters, and, by inference, of human settle ment and other activities. Area VI, which lies north of the headland of

a 56. InTables 3 and 4, sherd that is within the stated time span, and while (map of the bedrock geology of the counted as Archaic was iden evidence so "possibly" broadly assigned provides study region). tified the as little in certain it has 59. see by project's pottery experts help inquiries, For C02 (Tragana^/?z?/zW), or as to a more in out et "Archaic?" belonging proven useful sketching long Davis al. 1997, p. 459. indefinite term levels of 60. 7 chronological category (e.g., activity. Figure presents definitely Archaic-Hellenistic). Numbers given in 57. For further discussion of the dated material from both tract and site these in to tables for Early and Late Roman, palynological evidence relation collections. Raw numbers have been exclude material dated "Ro the PRAP see to account however, study area, "Pollen and "weighted" for differences in man." Material to PRAP below. extent assigned overlapping Patterns," the areal of the nine surveyed et zones. periods (e.g., Archaic/Hellenistic) 58. Zangger al. 1997, p. 568; could to or 554-576 and 4 belong any period periods see, further, pp. fig. 164 SUSAN E. ALCOCK ET AL.

TABLE 3. NUMBERS OF SHERDS RECOVERED IN PRAP TRACT WALKING

Period Definite Possible Total

Geometric 5 57 62 Archaic 23 150 173 Classical 28 304 332 Hellenistic 85 345 430 46 Early Roman 25 71 Late Roman 61 27 88 Roman 295 514 809

Geometric/Archaic 48 151 199 Geometric/Classical 109 301 410 Archaic/Classical 84 300 384 Archaic/Hellenistic 312 241 553 Classical/Hellenistic 193 308 501 Hellenistic/Roman 442 625 1067 Hellenistic/Early Roman 142 348 490 740 Roman/Byzantine 508 1248

TABLE 4. NUMBERS OF SHERDS RECOVERED IN PRAP "OFF-SITE" TRACT WALKING

Period Definite Possible Total

Geometric 0 10 10 Archaic 4 39 43 Classical 8 60 68 Hellenistic 15 64 79

Early Roman 9 1 10 Late Roman 6 4 10 Roman 37 122 159

Geometric/Archaic 8 41 49 Geometric/Classical 25 58 83 Archaic/Classical 21 56 77 Archaic/Hellenistic 76 30 106 Classical/Hellenistic 41 56 97 Hellenistic/Roman 62 157 219 Hellenistic/Early Roman 24 65 89 168 Roman/Byzantine 136 304

two ancient Pylos, is strikingly emphasized in these later periods, produc can ing 55 percent of Archaic and 62 percent of Classical sherds that be zone area definitely dated. For the Archaic period, another "active" is II. Low-level scatters of Archaic and Classical material were found in most of the other areas, but neither the coastal transect of area V nor the inland area a to or IX produced single sherd definitely dated either the Archaic areas Classical periods, when these may indeed have been "deserted." Of two relatively strongly represented Geometric sites, K01 (Gargaliani area area Ordines) in I and B07 (Hora Palace ofNestor Lower Town) in VII, the latter definitely and the former possibly continues in attenuated form HISTORICAL MESSENIA 165

%fH ^i;" 1 K-N?\?\\ y\Xxj ? va

nine areas Figure 6. Location of the intensively surveyed by PRAP. V X ^ R. J. Robertson

are into Archaic times, though both then largely abandoned in the Classi cal period (Tables 5,6; Fig. 8:a-c).61 Geometric through Archaic continu area ity is also visible in VI at 104 (Romanou Romanou) and possibly in area at area at II D01 (Gargaliani Kanalos) and in VI 101 (Koryfasio Beyler one as occu bey). Only site, 104 (Romanou Romanou), emerges definitely pied in both Archaic and Classical times; indeed, this is the only site where activity clearly spans the Geometric through Classical periods. There are as only three other definite Classical sites, each of which rates "possibly" as con occupied in Archaic times. All in all, Romanou emerges the most a tinuously active settlement in historical times (see Fig. 9 for breakdown of sherd material discovered at this site).62 The pattern of "clustering" observed in the distribution of individual area as finds by is distinct in the distribution of sites well (see Fig. 8:b-c). Of the five definite Archaic sites, evidence is particularly strong for occupation

term in a 61. The traditional "site" is or, exceptional cases, with only against background levels of surface no scatter used in the present article, in lieu of single sherd and other relevant and received further attention material B05 a additional collec "Place of Special Interest" (POSI) artifactual (e.g., Classical through variety of in PRAP Palace Nestor tion in are found other publications. [Hora of Chamber Tombs]). strategies; addition, all loca The authors fully acknowledge the While such decisions inevitably impose tions where past anthropogenic activity, concentra most contingent and relative character of site arbitrary cutoff points, these rather than natural processes, et out to definition (see, further, Davis al. tions stand quite clearly from the plausibly led artifact deposition. n. 62. a material 1997, p. 401, 27) and that site defi surrounding level of "background noise" For discussion of nition is a in at Romanou in PRAP sur always difficult, and usually comparison with overall densities found the a see et controversial, determination. In the and densities for particular historical vey, Davis al. 1997, pp. 462 sites less and more detailed following discussion, "definite" periods. Moreover, additional, pre 467, figs. 23,24, for are or more to those with three sherds cisely dated material is usually found descriptions of the individual sites to more arti securely datable the period indicated; accompany the diagnostic discussed below, their estimated size, sites are those with indirect to and a their ceramic "possible" fewer facts, lending support period catalogue of and or more see than three such sherds, with six identifications. All sites reported here other finds, the online site gazet to two were tract as concen teer. sherds definitely dated contigu noted in walking ous trations material that stood out periods (e.g., Archaic/Classical), of i66 SUSAN E. ALCOCK ET AL.

Geometric

Archaic

75 Classical H 13

Hellenistic D

25

Figure 7. Percentage of definitely dated sherds found in each surveyed j^n, PRAP area, Geometric-Hellenistic I II III IV V VI VII VIII IX R. J. Robertson

at the two sites in area VI, 104 (Romanou Romanou) and E01 (Romanou in Glyfadaki), which lie only 1.5 km from each other. Another pairing area Archaic times is visible in II with D01 (Gargaliani Kanalos) and D03 over m. (Gargaliani Megas Kambos [2]), separated by just 500 Such clus seen area tering is again in VI in Classical times, with the two most pro ductive sites being 104 (Romanou Romanou) and 101 (Koryfasio Beyler bey)?intervisible sites approximately only 1.5 km from each other. Although at this point, the site sizes employed in discussion by PRAP on still rely by and large "overall," multiperiod assessments, observations can be made about the relative proportion of individual sites. Like UMME, a PRAP did not locate plethora of small, "farmstead-size" sites. Romanou, sense for example, is notably large. A of the site's size in the Archaic and Classical periods is presented in Figure 10, where the "minimum" maps represent the distribution of material definitely dated to the specified pe riod and "maximum" includes all material that might date to the period.63 A size range from ca. 18 to 22 ha (minimum to maximum) emerges for the 63. "Maximum" thus Archaic 14 to 27 ha for the Classical?dimensions most includes, e.g., period, frequently for the Archaic sherds dated accorded the label of in collection period, "village" survey reports.64 "Large-site" Archaic?, Archaic to Classical, Geo such as the use of microtracts were strategies, (see Fig. 11), implemented metric to Archaic, and Archaic to Hel at un lenistic. Distribution of this this ancient settlement, part of which undoubtedly lies unexplored maps type at an for other PRAP sites will be made der the modern village; certainly the digging of graves on-site mod available online. ern cemetery church continues to up ancient material.65 Sites D01 bring 64. and and D03 Kambos have been Contiguous near-contiguous (Gargaliani Kanalos) (Gargaliani Megas [2]) tracts with relevant material were en assessed at 0.45 and 0.24 ha, and these estimates are area respectively, probably circled, and the resulting measured, a low, given that PRAP could only explore restricted part of the settlement to determine these sizes. at a 65. On collection Kanalos and combination of bulldozing and cultivation has severely "large-site" methods, see Davis et al. 1997, 405. damaged the sites atMegas Kambos. p. McDonald and Rapp (1972, pp. 310 The evidence of clustering of sites larger than small farmsteads begins 311) the at Romanou to an answer to report presence sketch the "Where did Messenian helots live?" an question of Archaic pithos burial, Classical For the PRAP in the Archaic the answer seems study area, period, prima and Hellenistic statuettes, Roman to at or and a medieval rily to lie the north Kanalos and Megas Kambos (2) in the vicinity graves, column. HISTORICAL MESSENIA 167

TABLE 5. DEFINITE AND POSSIBLE PRAP SITES

Early Late Geometric Archaic Classical Hellenistic Roman Roman Roman

Definite B07 B07 A06 A04 A04 A04 A04 101 D01 101 A06 D01 B01 B01 104 D03 104 D01 D03 D01 B07 K01 E01 M02 D03 G01 D03 D01 104 E01 104 G01 D02 G01 L07 G03 D03 G02 104 G01 G03 L07 G02 101 G03 104 101 106 104 117 K01 K01 L03 L02 L07 L07 M05 M04

Possible A02 A06 A04 B01 A03 A03 A03 A03 C01 B02 B07 A06 B02 A06 D01 C02 B07 COI B01 B07 B02 D02 G01 C01 C02 B07 COI COI D03 101 C02 D02 C04 D02 E01 E01 K01 D01 103 D02 E01 119 L02 L05 D03 L06 E01 G02 124 M02 M02 E01 M02 G02 101 K05 102 G03 103 L04 123 101 119 K01 103 K01 L05 K01 K05 M04 K05 L03 L02 L04 L03 M05 L04 M02 M05

TABLE 6. DISTRIBUTION OF PRAP SITES

Late PRAP Early Area Geometric Archaic Classical Hellenistic Roman Roman Roman

? ? I 1 1 1 ? ? II 2 2 3 III 1 IV 1 1 V 2 2 VI 5 2 VII 2 VIII 2 1 IX 2 2 Total 16 15 SUSAN E. ALCOCK ET AL.

Figure 8 (left and opposite). Distribu tion of PRAP sites: (a) Geometric; (b)Archaic; (c) Classical. R. J. Robertson

of Romanou, and habitation at Romanou seems to continue into the Clas as sical period well. A handful of other Archaic and Classical sites were, however, identified outside such clusters. Limited amounts of Archaic were at a material found B07 (Hora Palace ofNestor Lower Town), very heavily investigated portion of the study region, and in the southern end as a of K01 (Gargaliani Ordines), which ranks "possible" site. For the Clas area we sical period, outside of the dominant VI, find M02 (Gargaliani Kalantina [2]) and A06 (Metamorfosi Ay iosKons tad [2]), measuring an 0.75 and 0.56 ha, respectively. Kalantina (2) is unusual site that can, as a a on very tentatively, be proposed possible cult place, suggestion made a the basis of the limited range of finds (including fine wares, cups, and a krater of Classical date) and their concentration at the lip of natural on sinkhole (min. D. 2.5 m) the Gargaliani plateau.66 Ayios Konstadtinos cut a (2) also exhibits blocks and possible column fragment, but such finds accord better with the subsequent Hellenistic epoch, which boasts the prin ceramic on the site. cipal presence Some of the "definite" sites in the survey territory?such as Classical Ayios Konstadtinos, represented by the minimum number of closely dat able sherds?and the various "possible" Archaic and Classical sites might as or be explained the product of seasonal dwellings other forms of our "off-site" rural activity (see Table 4). This patterning contributes to sense of shifting levels of human activity in the Archaic and Classical land scape. Indeed, visible variations?in the number and distribution of indi vidual artifacts and of "definite" and "possible" sites?very clearly make the that southwest even under was no point Messenia, Spartan control, came static landscape. Nevertheless, itwould appear that, when it to choos 66. Davis 1998a, pp. 277-278. A their of the inhabitants of the PRAP ing residence, wares to principal places study very few fine belonging the area in Archaic and Classical times followed more oriented were communally Geometric/Archaic period found at this site. preferences. HISTORICAL MESSENIA l6o

a rescue Evidence derived from excavation in the village of , an in the northern Soulima valley (see Fig. 1), introduces additional piece of evidence for the question of where the Messenian helots lived. A large x to a on building (some 30 17 m) originally thought be Late Roman villa, on the basis of its size and sturdiness, has been reassigned, the basis of a b.c. sec ceramics, to period of occupation from the sixth century to the ond quarter of the fifth century. The excavator, Nikolaos Kaltsas, proposed was a that the site the substantial home of Spartan landlord supervising 170 SUSAN E. ALCOCK ET AL.

Figure 9.104 (Romanou Romanou). Sherd numbers, Geometric-Late

Roman. R. J. Robertson

750

250

-1 -j o < o -i ce ce ce < o o CL ce >>N r? t\ Jr ? x Jj LU < O X _i X * helots who lived in attendance, and the structure's violent destruction has been subsequently associated with the helot rebellion of the 460s.67Whether Kaltsas is correct about the identity of the householder, the image of helot can workers living in aggregation be related to the patterns of communal dwelling derived from survey data cited above. Results of the surveys by UMME and PRAP, and of the excavations at no a Kopanaki, provide sign, however, of general Messenian pattern of scattered helot residence on individual kleroi, and this lack of evidence for helot dispersal remains far less than ideal evidence for the phenomenon's absence. Our uncertain of the relevant ceramic understanding sequences, some to first and foremost, might leave hesitant accept this negative evi sure are we dence: how of recognizing the material world of presumably poor and relatively isolated helots? The indications from PRAP data, when on com coupled with the UMME data, do suggest, the other hand, that was at munal dwelling the preferred option under Spartan rule, least in more this distant portion of Messenia. an Why the Messenians would have preferred such arrangement is clear when one considers the vulnerability of helots in times of agricultural are shortage. Even though many parts of Messenia famously fertile, bad constant to years were inevitable, especially given the drain of foodstuffs or Laconia. Lacking civic structures, and lacking easy-to-hand friendly access to to Spartan patrons, Messenians would of necessity have had gen erate other systems of mutual support, and local networks?facilitated both across by aggregated living and by communication between settlements the region?would have become especially vital. Living in communal or in close to would also have allowed the foster For groups, proximity others, 67. the original publication of these see Kaltsas for a more ing of group identity, providing helots with potential opportunities and data, 1985; to own com recent assessment, see Harrison and means by which transmit their histories and memories. Such 1998, pp. 161-162 and 75. municative and mutual in turn, would have reinforced Spencer fig. sharing support, to set 68. Ritual complements this collective sentiments in all their force.68 are potentially incendiary tlement pattern further discussed in the would have allowed this course of which in Why Spartans action, "Change the Ritual and Mortuary a to is less how below. would have posed potential risk themselves, apparent. If, Landscape," HISTORICAL MESSENIA 171

Archaic Classical

Sherd density ?

Minimum Minimum

Maximum Maximum

Figure 10.104 (Romanou Romanou). Minimum and maximum areal ever, their "bottom line" stipulated the survival of helot workers and the to ac extent, Archaic and Classical. continued influx of helot-grown products, then they may have had Solemeto and R. Robertson J. J. cept nucleated residence and the local support networks that it sustained. The settlement patterns observed by UMME and by PRAP are, in this sense, somewhat akin in effect to Hodkinson's proposed system of share as cropping discussed above. The nature of sharecropping, he stressed, in volves shared risk between cultivator and landowner. Sharecropping maxi mized labor inputs, keeping the helots too busy to cause trouble, yet it also to kept them alive; it allowed Spartiates monitor agricultural activity, but itwould not have forced helots into revolt over rents. Communal settle ment a worked in similar fashion, allowing essential Spartan needs to be some measure adequately met, while also providing helot families with of 172 SUSAN E. ALCOCK ET AL.

Figure 11.104 (Romanou Romanou). Collection of microtract. PRAP Archive

to security. Community dwelling would also have contributed the estab as lishment of some form of helot authority, or hierarchy, much sharecrop or ping would have required the existence of "bailiffs" other middlemen by means of which helots could be controlled, to a great if not total extent, without requiring massive investments of Spartiate time and effort.69When our new nu coupled with regional evidence, these considerations provide ances to bitterly polarized perceptions of Spartan-Messenian interactions. Patterns of settlement and patterns of economic exploitation suggest that were a risks and benefits shared in balancing act that, periodically and ultimately, broke down in rejection and revolt?but the system worked for a very long time.

Pollen and PRAP Patterns

These observations about settlement and other traces of human activity can cor be brought into dialogue with results from detailed palynological ing in the Osmanaga Lagoon (see Fig. 2).70 A post-Bronze Age increase a to con in deciduous oak and decline in olive pollen led Sergei Yazvenko clude that "during the Early Iron Age the landscape experienced the least was re intensive human impact of the last 4,000 years."71 That decline versed ca. 800 b.c.:

a a The palynological data argue for dense population and high level of agricultural production during the Classical/Hellenistic period. seems Human control of the landscape to have reached its maxi mum a on at ca. at that time, with strong emphasis olive cultivation b.c. 500-100 (220-160 cm). The peak of olive pollen coincides with a to high C/N ratio, which indicates increased erosion due higher 69. Hodkinson n.d.; Alcock 2002b; on the pressure landscape.72 Hodkinson 2003. are were 70. For an of the The fossil data cited dominated by olive pollen; cereal and grapes explanation applic of the results of this research to also but are harder to trace these means. There ability undoubtedly important, by see the PRAP study region, Zangger is a lack of indicators, aswell, that would to herd pollen point widespread et al. 1997, pp. 576-585. or in the forms of economic that had been et ing grazing region, activity 71. Zangger al. 1997, p. 593. of theMessenian land 72. et al. 594. previous explanations for the presumed "emptiness" Zangger 1997, p. HISTORICAL MESSENIA 173

western scape. In general, assumptions about the desertion of Messenia in on a the years of Spartan domination, based largely single remark of are Thucydides, countered by the evidence of survey and pollen studies. are The date ranges of periods of cultivation that obtained through are this type of palynological analysis of necessity crude, and the tempta must tion to force them too neatly into existing historical periodizations be avoided. It would be tidy, perhaps overly tidy, to correlate the reclama tion of land by olive cultivation with the inception of Spartan control. The ca. em span of olive cultivation highlighted by Yazvenko, 500-100 b.c., as braces the Classical/Hellenistic break defined by PRAP, presented here as a period of limited settlement activity. The palynological argument for a seem "dense population and high level of agricultural production" might incompatible with that break. Two suggestions might help to reconcile the evidence. First, the rela tive lack of dispersed settlement should not be correlated automatically a with severely underexploited countryside; nucleated settlement makes a more onerous no means intensive cultivation task but by precludes it. Second, much of the palynologically attested increase in cultivation could be assigned to the Hellenistic period, when?as shall be seen?high levels are more of agricultural production readily compatible with the settled landscape. Supporting this hypothesis is the fact that the all-time peak in can an olive pollen be assigned absolute date of 230 b.c., though the date on is admittedly dependent regular sedimentation rates and radiocarbon error margins.73

Summary: Geometric through Classical

This regional review of Messenia in Geometric through Classical times, or, in other words, for the epoch encompassing the phenomenon of Spar tan control, yields various summary points. Some are relevant to the entire more area region, others to the specific of PRAP study. a areas 1. Far from presenting static landscape, studied by UMME and PRAP show clear signs of variation among Geometric, Archaic, and Classical settlements and patterns of land use. 2. The failure of UMME to locate numerous small and isolated sites was (potential "farmsteads") echoed in the PRAP results. a 3. In the PRAP study zone, there appears to be preference for area "clustering" of settlement, VI presenting the greatest traces of activity as measured by overall number of sherds and number of sites. 4. The dominant settlement of the PRAP in the Geo et study region 73. Zangger al. 1997, p. 594, metric through Classical periods is 104 (Romanou Romanou), where, reporting his calibration and an correction of the UMME cores the surface scatter for which commu pollen suggests agglomerated an these that suggested "olive spike" between nity throughout periods. 1100-700 b.c., Yazvenko redates the no means as was 5.While by "deserted," formerly hypothesized, the to the 540-140 b.c., thus spike period portion of southwest Messenia investigated by PRAP cannot be with PRAP results. coinciding characterized as 74. For further discussion of ritual densely populated. 6. Shrines were identified at several a in see UMME, and hint of activity these survey areas, places by in and ritual was observed PRAP atM02 "Change Ritual Mortuary Land activity by (Gargaliani scape," below. Kalantina [2]).74 174 SUSANE.ALCOCKETAL.

HELLENISTIC THROUGH LATE ROMAN (CA. 323 B.C.-A.D. 700)

Historical Background: Messenia

The end of the Spartan control of Messenia that, by the time of Leuctra, or more had been in place for three centuries must have led to profound, if largely unrecorded, transformations in all dimensions of Messenian life. were some The helots freed; at least portion of the Messenian diaspora con returned home; perioikic communities gradually shook off Spartan a new trol; existing systems of land tenure were terminated; and mixed was a new on population settled in and dominant city?Messene, the slopes of Mount Ithome. Piero Tr?ves rightly alludes to the situations "revolu

tionary novelty."75 to Messene would long remain the regions leading polis. According was Pausanias, the choice of Ithome for its setting oracularly inspired, but s was the city strategic role also obvious. Epaminondas founded Messene, serve as a together with Megalopolis, to bulwark against future Spartan was a aggression, and its remarkable wall circuit visible proclamation of new the Messenians' ability to defend themselves. At least in its descrip tion centuries later by Pausanias (4.27.5-9), the civic foundation celebrated a own the reemergence of the Messenians as people with their gods, their own heroes, and their own history. name Pausanias reported that "to the capital they gave the ofMessene, towns but they founded other also."76 By his time, the later second century A.D., the poleis ofMessenia included Pherae,Thouria, Korone, Abia, Colo some were nides, Asine, Methoni, Pylos, and Cyparissia, of which former accounts perioikic settlements. Since, in written from the early centuries as a of liberation, "theMessenians" tend to be referred to single unit, it is some generally assumed that they formed type of federal league under the saw en dominance of Messene.77 The Hellenistic period hoiMessenioi a gaged in variety of external alliances, almost inevitably with polities larger more in to and powerful than themselves and almost always opposition 75. Tr?ves 1944, p. 103. Roebuck 1941 and 1945 remain the most com Spartan interests. Among other bedfellows, the Messenians allied with treatments at prehensive of Hellenistic Philip II of Maced?n, remaining neutral the Battle of Chaironeia and Messenia; see also Alcock 1998. On being rewarded with grants of Spartan territory including Methoni and this from the perspec on period Spartan the Dentheliates the west of Fierce be see Ager slopes Taygetus. hostility tive, Cartledge and Spawforth 1989; tween Messenia and Laconia long endured, the competition often taking Kennell2003. most over 76. Paus. trans. G. Frazer the form of territorial disputes, famously the Ager Dentheliates 4.27.7, J. vol. and its border sanctuary of Artemis Limnatis.This competition continued (1965, 1). 77. Roebuck 1941, pp. 109-117; well into the Roman period, with rulings generally made in favor of the Lazenby andHope Simpson 1972, Messenians.78 p. 90. On the scale of the territory of Messenia remained somewhat isolated from much of the Although Messene, see Roebuck 1945, pp. 156 to more military activity of the age, its position adjacent Sparta, and par 157. not 78. On Artemis Limnatis and the ticularly the strategic location ofMount Ithome, did go unrecognized. the one boundary disputes surrounding For example, the mountain formed of the "horns" that Philip V of see Paus. Maced?n was advised to seize if he wished to control the "ox" of the P?lo sanctuary, 3.7.4, 4.4.2,4.31.3; Strabo 8.362; Tac. Ann. 4.43; Kolbe Conflict with V first the ponn?se (Polyb. 7.12.2-3). Philip brought region 1904; Valmin 1930, pp. 12-13,27-32; into contact with Roman in 191 would order a re authority, which, B.c., Cartledge and Spawforth 1989, a luctant Messenia to join the Achaian League. Playing minor part in the pp. 138-139; Malkin 1994, pp. 34-35. HISTORICAL MESSENIA !75

war League's with Rome in 146 b.c., Messenian troops remained at home sea. to guard against possible attacks from the In the aftermath of the war, Messenia regained its independence. It is unclear, however, whether Mes sene over reasserted political supremacy other communities. un Together with much of the rest of Greece, Messenia then passed der defacto Roman control. Roman officials became involved in local deci to to sion making, and inscriptions testify friendly contributions Roman most on coffers, famously the financial levy (octobolos eisphora) assessed cen Messene in the early first century A.D.79 In the civil strife of the, first as as tury b.c., Messenia had not suffered much other portions of the main land, but neither did it go untouched. Comments are recorded about col lections of booty in the P?loponn?se (Plut. Pomp. 24); Agrippa attacked Methoni during the Actium campaign (Strabo 8.359).80 Messenia sided with Antony in that particular conflict, paying the price in territorial losses soon to Sparta. It was, in all likelihood, after his victory that formally constituted the province of Achaia. From this point, Messenia one a formed just portion of larger provincial entity, and individualism? with their administrative and ritual functions?remained the principal political units in the peninsula.81 Epigraphical and archaeological evidence, together with the testimony of Pausanias, suggests that?despite undoubted vicissitudes?Messene and towns even other Messenian survived and thrived during Hellenistic and at Imperial times. Long-running excavations Messene, by far the best explored historical site in the region, demonstrate clear signs of civic pros most perity, notably in the construction of the Asklepieion in the first half b.c. of the second century This impressive complex has been closely linked to man Damophon of Messene, the entrusted with repairing the Zeus of at career can Pheidias Olympia. The of Damophori, whose family tree be one re traced inMessene's prosopography, offers just sign of the region's newed status in the cultural life of the Greek world.82 as a Strabo (8.362), describing the region whole, remarks thatMessenia was com deserted in his day, that is, around the time of Augustus. His can as a ment, however, be taken rhetorical exaggeration typical of much of Early Imperial literary production.83 In fact, in the Imperial period, Messenia seems to have been one of the areas Greece most for qi popular immigrants, including Italian businessmen (negotiatores) and Roman citi zens. The presence o? negotiatores inMessenia is first detected in the later as men Hellenistic period, and inscriptions, such the octobolos eisphora attest tioned above, that Rhomaioi owned considerable properties in the

IG see 79. V.l.1432-1433; Giovan 81. For discussion of the history pp. 36-63; Themelis 1993,1994a, in nini 1978 for the date and details of the of Messenia Hellenistic and Early 1994b, 1998,2001. The famous regu decree. Roman see and lations of the of times, Lazenby Hope Mysteries 80. On troubles in one most first-century Simpson 1972, pp. 92-93; Alcock (91 b.c.), of the venerable of see and all to exis Messenia, Lazenby Hope 2002a, pp. 164-174. Messenian rituals, speak the 92-93. 82. For ancient of tence of Simpson 1972, pp. Epigraphical praise Damophon wealthy local residents: Roe evidence from at see on the Augustan period of Messene, Paus. 4.31.10; his buck 1941, pp. 7-10, 35-36; Breuillot Messene to a of points program build prosopography, seeThemelis 1994b, 1985, pp. 799-802; Piolot 1999, 37. ing and restoration, presumably follow p. On the Asklepieion and other 83. Baladi? 1980, pp. 301-305; a see ing period of neglect and disrepair: features of the cityscape, Feiten Alcock 1993, pp. 24-32. see SEG XXIII 207;Migeotte 1985. 1983, pp. 84-93; Habicht 1985, 176 SUSAN E. ALCOCK ET AL.

area. no The fertility of the region doubt contributed to its attractiveness, a as but itswestern orientation may have been factor in its popularity well.84 were Special marks of imperial favor garnered by communities and indi re vidual families. Pausanias (4.35.3) tells us, for example, that Methoni rare one ceived the gift of freedom and tax immunity under Trajan, and a family from Messene boasted high priest of the provincial imperial cult a rare and, under Hadrian, reached senatorial status, relatively event for the Greek elite in the Early Empire.85 If the fertility of Messenia, coupled with the conditions of the pax some Romana, appear to have provided at least of the regions inhabitants more with prosperous times, this security began to be threatened in the A.D. later Empire, the fourth to seventh centuries By the late sLxth century, a re combination of barbarian invasion, plague, and natural disasters is events ported to have struck the land. Although the exact sequence of in this portion of the P?loponn?se remains unclear, and although such gen re eralized catastrophes manifestly need to be analyzed in their specific gional context, it has been argued that, toward the end of the Late Roman a period, the P?loponn?se "no doubt suffered real decline from its material or and demographic level in the 2nd 3rd century A.c."86 Another trend can over that be monitored these centuries is the region's conversion to Christianity, signaled by the appearance of Messenian bishops and the as construction of Christian monuments such the basilica at Ayia Kyriaki near .87 What happened next, in the ensuing Dark Age and Byzan tine are matters for another period, analysis.88

Historical Background: The PRAP Study Area

Specific details about the PRAP study region in the Hellenistic through are Late Roman periods relatively rare, but suggest that in these centuries, zone own course more too, this western followed its in comparison with central parts of Messenia. Messenia's entanglements with larger political are structures well illustrated by the Achaian League's incorporation of ca. B.c. Pylos in the later third century, 220 Despite Messene asking for Rome's intervention in the matter, Methoni and Cyparissia would also soon as (in 191 b.c.) join the League independent units, essentially cutting off the west coast, and several good ports, from the control of Messene.89 In the second Pausanias this ei century A.D., traveled through area, a or sea. ther following coastal route possibly moving by He says little about the area, treating only Pylos/Koryfasion in any detail (4.36.1-6). In

84. Hatzfeld [1919] 1974, p. 79; 87. For this basilica, built in the first D03,108, K03, K05, L02, M04 (possi see Wilson 1966, p. 150. On connections half of the sixth century A.D., Ger ble). For presentations of these sites western see stel Davis et al. and their see the online site between Greece and Italy, 1998, pp. 212-214; material, et Doukellis 1990; Alcock 1993, p. 143. 1997, pp. 474-475. gazetteer. See also Davis al. 1997, can 85. Paus. 4.32.2,4.35.3; Habicht 88. We report, however, 16 defi pp. 474-482; Gerstel 1998; and the nine sites: references in Davis 313. 1985, pp. 58-59; Arafat 1996, pp. 194 nite and possible Byzantine 1998c, p. 195. A03, A04, A06, B01, B02, COI, D01, 89. Polyb. 4.16.7,4.25.4, 9.38.8; 86. Topping 1972, p. 65; for similar E01, G01, G03,101,104, K01, L04, Livy 36.31.9; Roebuck 1941, p. 69, n. sentiments, see Gerstel 1998, p. 212. L07, M05 (definite); B03, B04, D02, 12. HISTORICAL MESSENIA IJJ

his remarks on that site, Pausanias pays greatest attention to the heroic past: to Neleus, Nestor, and their associated monuments, including the "Tomb of ," usually identified with the tholos tomb at Voido on a koilia. This focus distant past at the expense of the living present, while typical for Pausanias's treatment of Roman Greece, is especially pro nounced in book 4 (Messeniaka), in which his neglect of many aspects of not Messenian history, least the period under Spartan control, has done to course on much dictate the of subsequent scholarship the region.90 can or Where archaeological evidence illustrate amplify this outline of events? The astonishing nature of Epaminondas's foundation?especially in its implications for local patterns of demography and land use?should we have clear material correlates. By extension, should be able to explore on a more the degree to which Messenia, being liberated and assuming as "normal" political organization, returns to normal in other ways well, comes to and how far it resemble other parts of the Greek landscape. The a fertility and westward orientation of Messenia make it particularly in zone on teresting to monitor, and while archaeological perspectives the passage of Greece into the Roman empire have begun to be explored, nuance further detailed regional studies are needed to the current gener alized picture.91

Overview of UMME Patterns

Issues raised for the Hellenistic period by UMME findings have been discussed above (pp. 159-162). The Roman period, in UMME's chrono logical scheme, began with the defeat and temporary dissolution of the Achaian League in 146 b.c. and ended in the fourth century a.d. (see Ta 90. See Alcock 2002a, pp. 172-173. ble 1). In almost all instances, this five-century-long span is treated by On the "amnesia" of Pausanias, see Al UMME as a unit. Sites as Roman UMME cock 2001. single reported by (Fig. 12) are 91. On the "Romanization" of significantly fewer than those of the Classical and Hellenistic periods raw Greece, see, e.g., Alcock 1993; Hoff (Figs. 4,5). Weighting these numbers by length of period would make and Rotroff Salomies 2001. these even more marked.92 1997; discrepancies 92. For an of such example weight Within that overall pattern, however, there exists a degree of variabil see Alcock 11. ing, 1993, p. 41, fig. zone a ity.The northern of the UMME study area, for example, suffered 93. Frazer 1965, vol. 3, p. 447. For more dramatic site loss than other the other on (Table 2). contrast, observations the Roman peri subregions By Five a net od in the Five Rivers area, see Luker Rivers region around Nichoria experienced gain in settlement mann and a to to Moody 1978, pp. 97-98; during Roman times, circumstance attributed the area's fertility and Coulson and Wilkie 337. 1983, p. the proximity of the region to the town of Korone (modern Petalidhi), 94. Of the eleven sites within the where ancient foundations have been found, also marble PRAP area identified as Roman "many sarcopha study to a guses, pointing the existence of thriving and art-loving population in by UMME, PRAP systematically rein times."93 Yet a third can be traced in the southwestern vestigated three: UMME 15 Garga imperial trajectory liani:Kanalos (PRAP D01); UMME zone, where, in UMME's estimation, the number of sites drops only from 400 Romanou: and to Viglitsa (104); 26 22 (Table 2), and, in the overall PRAP study region, from sixteen to UMME 406 Dhia Marathoupolis: fifteen.94 McDonald and Hope Simpson connect this relative stability to liskari (G01). an external orientation, observing that "it was, of course, the southwest 95.McDonald andHope Simpson that retained closest contact with the outside world in medieval times."95 1972, p. 146. 96. Preference for coastal noted to have increased in earlier Lukermann and Moody 1978, locations, already 7-12. its in p. 96, figs. periods, hits peak Roman times.96 i78 SUSAN E. ALCOCK ET AL.

Habitation Cemetery A Shrine

Figure 12. Distribution of UMME sites: Roman. R. J. Robertson

Another development clearly indicated in both the UMME and Five was Rivers results the appearance of impressive structures appointed with baths, mosaics, marble elements, and other "villa-like" accoutrements. "Vil are some a " reported in number in Roman Messenia, reaching total of seven seen in the hinterland of Nichoria alone.97 This phenomenon, by no means everywhere in Greece, is observed also in Achaia, to the north, numerous recent where such complexes have been reported in work along the Gulf of .98 In the UMME view of Roman Messenia, coastal towns (e.g., Korone, Asine, Methoni, Cyparissia, and Pherae) were the main to a settlements, which villas and farms formed background. The common was a assumption that Greece in Imperial times quiet backwater not surprisingly colors their interpretations: "Pausanias' account shows that area a a although the may have been 'backwater,' it had romantic appeal to cultured Romans."99 on Distinguishing changes in settlement distribution the basis of use UMME reports is complicated by the project's of the term "Roman." Reexamination of Five Rivers sites suggests that many of them continued in use in the Late Roman and Early Byzantine periods (ca. A.D. 330-600); as so, J. Rosser andW. A. McDonald have noted, "it is probable that much of the 'Roman' occupation in the Nichoria environs lasted down into the 6th century, i.e., until the period of the Slavic invasion."100 Other UMME sites a "Roman" presumably followed similar course, but the data precludes greater precision.

97. On the villas in the hinter in see Valmin 469 99. and Messenia, 1938, pp. Lazenby Hope Simpson see land of Nichoria, Lukermann and 475; Lazenby and Hope Simpson 1972, 1972, p. 96. Moody 1978, pp. 99-100; Coulson and p. 96; Alcock 1993, pp. 63-71. 100. Rosser and McDonald 1983, Wilkie 1983, p. 338. For reports of 98. Petropoulos and Rizakis 1994; p. 354. other apparent Roman-period "villas" Rizakis 1997. messenia historical 179

Overview of PRAP Patterns

a In the landscape of the PRAP survey area, PRAP traced florescence of settlement and other signs of activity datable to the Hellenistic period. As noted above, between Classical and Hellenistic times, the number of defi more nitely dated sherds than triples (see Table 3), and it nearly doubles in the off-site locations (see Table 4). Sixteen "definite" sites of Hellenistic can date be defined, compared with four for the Geometric, five for the Archaic, and four for the Classical period (Table 5; Fig. 13:a). This "in creased tempo" of activity matches Classical/Hellenistic patterns reported by UMME.101 Expansion of such magnitude in both site numbers and more off-site data points toward intensive cultivation and exploitation of a this countryside inHellenistic times and, less directly, to larger resident can population. These observations be connected with the palynological ca. evidence, discussed above, which pinpointed the years 500-100 B.c.? a and, particularly, later years in that span?as period of "higher pressure on the landscape."102 now Area VI, containing several sites of different function, continues in the Hellenistic period to produce the majority of material (Figs. 7,14), as it had in the Archaic and Classical periods (Fig. 7), but, throughout the area of the PRAP survey, sites are much more scattered across the land scape (Fig. 13:a; Table 6). Settlement hierarchies also appear to become more complex in post-liberation times, with the development of "normal" poleis (noted above). In the PRAP region, sites ranged from the asty of to as Pylos/Koryfasion large, ancillary villages such Romanou, elite resi as dences such Glyfadaki (see below), and smaller, less impressive rural some habitations that, in cases, likely deserve the label "farmstead." an Examples of this last include two small sites in area IX, area newly inhabited in Hellenistic times: L02 (Maryeli Farfa ti Rahi [1]) and L07 measures ca. (Ayioi Apostoloi Palaiospitia). Site L02 0.09 ha, and lines of at to a walls the site, the longest preserved length of nearly 21 m, may or represent old field boundaries terrace walls; at L07, measuring 0.88 ha, en the Hellenistic and Early Roman components were confined almost to area. or tirely approximately one-third of the Whatever political insti tutional transformations accompanied the liberation of Messenia, there was a zone. decided impact upon the rural landscape in this southwestern on An increasingly visible emphasis coastal settlement is another spa tial development occurring in the area in Hellenistic times. Three new or coastal near-coastal sites appear in areas that were all but empty in his 101. Coulson and Wilkie 1983, torical times, areas III and V 7,14), sites G01 Dia 337. (Figs. (Marathopolis p. G02 S o and G03 extract liskari), {Vromoneri Ayia tiro), (Vromoneri (see 102. See the from Zangger Pigadid) et 1997 on 13:a). This accords with UMME's observation about an al. p. 172, above. The possi Fig. increasingly bility that this development inHelle coastal orientation in Classical/Hellenistic times, which they attributed to nistic Messenia reflects instead a more commerce. stable political conditions and better-regulated in ceramic on change usage, following area A closer examination of the highly productive VI, and its one "Mediterranean-wide shifts in com on possible and five definite Hellenistic sites, offers another perspective merce, travel, and/or production," is in Hellenistic Messenia. in Romanou (104) and (101) ap raised the Appendix, below; such change Beylerbey re as Romanou in its consid material changes would, of course, pear larger settlements, particular maintaining status flect altered and circumstances. erable size and increasing its volume of relevant material (see Fig. 9). By i8o SUSAN E. ALCOCK ET AL.

Figure 13 (left and opposite). Distri bution of PRAP sites: (a)Hellenistic; (b) Early Roman; (c) Late Roman; (d) Roman. R. J. Robertson

at some m contrast, research Glyfadaki (E01), which lies only 250 inland a from the present shoreline, reflects the existence of only single, impres a sive dwelling. Geophysical prospection at the site recovered linear anomaly, m cor some 45 in length with perpendicular offshoots, that suggests the a ner of sizable building with interior cross-walls. The weakness of the electromagnetic signal, together with the results of coring, suggests that the walls have been robbed out, leaving only foundation trenches. Found HISTORICAL MESSENIA l8l

was assem in close association with this anomaly "a recognizably domestic blage" of Hellenistic and, especially, later Hellenistic date.103 Among the were wares finds amphora and lamp fragments, loom weights, and fine as that included pottery?such moldmade bowls?reminiscent of examples ceram from Hellenistic Messene (see the Appendix, below). The density of ic finds had a distinct "fall for three 103. Davis et al. 1997, p. 461. off";104 example, adjacent grid squares 104. Davis et al. 1997, p. 461. held, respectively, 285 sherds and 395 tile fragments, 127 sherds and 362 tile l82 SUSAN E. ALCOCK ET AL.

Hellenistic

Roman |_j Early

Late Roman ^| ^| Roman

Figure 14. Percentage of definitely dated sherds found in each surveyed PRAP area, Hellenistic-Late

Roman. R. J. Robertson

one fragments, and 14 sherds and tile fragment, the higher densities of area material being located inside the delineated by the proposed lines of walls. Specifics of the site, including maps of the anomalies and ceramic we densities, have been presented elsewhere;105 what stress in the present context status is Glyfadaki's apparently elite and relatively short life. near area are Also of Hellenistic date both in and VI wealthy graves; for at in area example, the tumulus of 117 (Tragana TsopaniRahi) (Fig. 15) at near on VI and in the cemetery Divari Yialova the Bay of Navarino, some seen 5 km southeast of Glyfadaki (see Fig. 2).106 Little could be at site 117 by PRAP investigators, but excavations in the 1960s had recov a ered gold diadems and other jewelry, silver vase, glass bowls, coins, and much pottery. The tumulus lies less than two kilometers from Glyfadaki, was and intervisible with that complex and with the settlement at Romanou. we While had previously hypothesized some measure of differentiation et and hierarchy in helot society, this stratification only becomes manifest in 105. Davis al. 1997, pp. 459-465; as Harrison 1998; et al. 1997, the regional landscape in the Hellenistic period, witnessed by the ap Zangger pp. 596-601. For illustrations of the pearance of these rich tombs; this accords well with our historical under finds, see Davis et al. 1997, 23,24, of the social and economic transformations. this figs. standing region's Finally, and the online site gazetteer. area also of Hellenistic ritual at 106 displays signs activity, (Tragana Viglitsal 106. Divari (UMME 401);Tsopani Tragana Tholos Tombs) and 102 (Koryfasio Haratsari [Osmanaga TholosJ). Rahi (UMME 403): McDonald and 310-311. For more Both sites have yielded signs of votive offerings in the setting of prehis Rapp 1972, pp. information on the material from these toric tholos tombs, 106 evincing evidence of Hellenistic definitely activity, see 198-200. at tomb 2 Evidence at 102 is less and sites, below, pp. specifically Tragana (Fig. 16). clear, 107. It should be noted that the a the site is better considered have "possible" example.107 Tragana finds also been inter of these Hellenistic to and as a Many patterns appear endure, strengthen, preted Hellenistic habitation; for can as references on tomb see An in Roman times. Six sites be identified "definite" Early Roman sites, Tragana 2, as tonaccio 1995, pp. 79-80. On Osma eight "Late Roman," and fifteen as undifferentiated "Roman" (see Table see as are naga, Alcock 1991, p. 466; Anto 5). These sites, those of Hellenistic date, widely distributed, and, for naccio 1995, pp. 72-73. For further the first in at least one of the three each of time, divisions, discussion chronological of Messenian cult places, the nine areas is at least one site see in surveyed represented by (see Figs. 13:b-d, below, "Change the Ritual and as zone 14; Table 6). Area VI continues the most active of settlement, and Mortuary Landscape." HISTORICAL MESSENIA 183

Figure 15.117 (Tragana Tsopani Rahi) (atmodern fieldhouse). View over intensively cultivated fields (note irrigation pipes). Bay of Navarino and in the Sphacteria PRAP Archive background.

Figure 16.106 (Tragana Viglitsa/ Tholos tomb 2. Tragana Tombs), PRAP Archive

an coastal locations remain popular. Palynological evidence presents in stance of change, indicating "diminished human land use"?an increase in a deciduous oak and decrease in olive pollen?for the broad period from ca. b.c. to a.D. 100 1200.108 A possible explanation for this development is change in patterns of land tenure in favor of larger and less intensively a utilized estates, hypothesis supported by the appearance at G01 (Mara of a villa associated with a substantial es et thopolis Dialiskari) presumably 108. Zangger al. 1997, pp. 594 tate. in 595. The presence of villas elsewhere Messenia had been reported by 109. Dialiskari: UMME 406; UMME, as noted above. another candidate for Erana area was leading Dialiskari, which lies in the northern coastal III, known long is (UMME 408), north Ayia Kyriaki before its investigation PRAP, and has been as a area. by suggested possible of the PRAP study On the Erana candidate for the elusive site of Erana, which Strabo (8.348, 361) locates debate, see Dimakis 1938; Frazer 1965, somewhere the coast between and near vol. 3, p. 462; Liritzis 1969; and, for along Cyparissia Pylos/Koryfasion, the island of Proti.109 Strabo's remarks on Erana are but to early references, Valmin 1930, pp. 136 brief, appear 140. to some manner refer of community?perhaps o? polis status, though this SUSAN E. ALCOC K ET AL.

300

o 200

100

Figure 17. G01 (Marathopolis Dialiskari). Sherd numbers, Xuj-1 Hellenistic-Late Roman. < o5?--LUCQ I Jj ? x R. J. Robertson never is explicitly stated. Arguments about where to assign this toponym continue. PRAP investigations discourage its application to Dialiskari, find an ing that, although there is initial, delimited Hellenistic component to the site, by far the bulk of the material is dated Late Roman and Roman as a seems (see Fig. 17), when the site's identification villa complex clear. a The site of Dialiskari boasts significant number of imported and luxury several architectural and other remains. materials, standing components, a structure These elements include bath with hypocaust piers (Fig. 18), two a an bedrock quarries, probable fishpond (less likely evaporation tank a for salt production) (Fig. 19), black and white mosaic floor (long incor a stone a porated within modern house), and brick walls, polygonal struc ture of mortared brick, several column fragments and capitals of nonlocal a stone, and cemetery.110 Several of these features, notably the hypocaust, were revealed in bulldozing operations related to cultivation and house construction, and subsequent geophysical study has detected possible ad traces ditional of subterranean chambers of the hypocaust system. Although a some a map of all these various elements (Fig. 20:a), separated by dis tance m or to a of 100 more, fails reveal the presence of single, unified structure, the distribution of surface finds suggests that most should be a considered part of unified complex.111 Artifacts observed at the site, in wares am cluding imported fine (e.g., African and Phocaean Red Slip), phoras, mosaic tesserae, slag, and circular hypocaust tiles, map in close

110. The mosaic floor has based on from Greece recently comparanda figs. 16,19,21,73-77,79,81. to Mu on nature see et been transported the Olympia and Italy and the luxury of 111. On Dialiskari, Davis al. seum for it other at restoration; will be placed features the site, comprises 1997, pp. 469-474; Stone and Kampke on exhibit in public Messenia, quite beachfront rock-cut channels running 1998; for maps and illustrations of the in Kalamata. We thank Xeni from a 9 x 3 see et possibly rectangular basin (ca. m), features described here, Davis al. of the 7th one ca. 34 west Arapogianni, director (L. m) extending 1997, pp. 470-471, figs. 27,28; Stone of Prehistoric and Classical southwestward and a 13 Ephorate second (L. m) and Kampke 1998, pp. 194-196, for this information. The for com 87-90. Antiquities, extending northwestward; figs. On the geophysical work, structure identified as a see see et 601. fishpond, paranda, Higginbotham 1997, Zangger al. 1997, p. HISTORICAL MESSENIA 185

Figure 18. G01 (Marathopolis Dialiskari), exposed hypocaust chamber. PRAP Archive

Figure 19. G01 (Marathopolis Dialiskari), probable fishpond PRAP Archive complex. i86 SUSAN E. ALCOCK ET AL.

Figure 20. GOl (Marathopolis Dia liskari), "core zone": (a) location of rters architectural and structural I major components; (b) location of finds, by R. b category. J. Robertson HISTORICAL MESSENIA 187

Figure 21. G01 (Marathopolis Dia liskari). View over the beachfront "core zone." fenced The compound in front of the modern villa contains much bulldozed ancient debris. PRAP Archive

association with the majority of these structures (compare Fig. 20:a, b).112 This close correlation of the structures and artifacts suggests the existence of a "core zone" to the site, which has been estimated at some 6 ha in extent.113 This core zone is coastal in nature, and borders an unmodified natural harbor.

The impressive finds at Dialiskari, which clearly distinguish it from an other sites in the PRAP study area, perhaps attest intended conspicu ousness: the site's architectural complexity and multiple functions strongly at suggest elite ownership and consumption. Few elements the site could no or be explicitly associated with agricultural productivity; olive were presses, for example, found. Millstones were, however, and fragments on a of three rotary querns may even suggest grain processing reasonably we can large scale. On present evidence, say little about the impact of Dia on or liskari its surrounding hinterland, about its reflection of patterns of a land tenure, beyond arguing that such "power residence" surely indi cates the presence, in west coastal Messenia, of a substantial property recon holder?an observation that lends support to the textually derived as as our struction of trends in the area, well to palynological patterns. not to While its scenic coastal setting has made the site vulnerable, least the ongoing construction of modern beach villas (Fig. 21), Dialiskari as an has recently been registered archaeological site by the Hellenic Min istry of Culture.114

re across extent other cist 112. For details of the pottery erally the 35 ha of the remains; the small rock-cut at site site. For a of overall sherd densi of the had been com covered G01 (Marathopolis map graves cemetery Dialiskari) and at other sites in the ties at Dialiskari, see Stone and pletely denuded of finds, and neither see the it nor the can be PRAP survey area, Appendix, Kampke 1998, p. 196, fig. 91. The quarry precisely two dated. below, with Tables 7-9. outlying features, the cemetery 113. and one of 114. We thank Xeni Cooking wares, glass frag the quarries, presumably Arapogianni this information. ments, lithic finds, and nonhypocaust constitute subsidiary foci of activity for were more or date to a than the tile fragments found gen different period i88 SUSAN E. ALCOCK ET AL.

Summary: Hellenistic through Late Roman

can Various points be summed up from this regional review of Messenia in Hellenistic through Late Roman times, from liberation to reconquest. are to to more Some relevant the entire region, others the specific PRAP zone. study

1.The data from PRAP reveal significantly increased numbers of individual artifacts and sites during this time span. UMME results for the Roman period, apart from those of the more intensive reconnaissance of the Five Rivers survey, do not conclusively agree with that picture. 2. Across the PRAP study area, the distribution of individual artifacts and sites, characterized by "clustering" in the previous more epoch, is widespread. 3. A long-term trend toward coastal emphasis in site location now becomes markedly apparent. 4. There is a greater differentiation in site sizes and site types, to from large settlements (e.g., Romanou) small locales (e.g., to or Farfa ti Rahi [1]), individual elite dwellings wealthy complexes (e.g., Hellenistic Glyfadaki and Roman, especially Late Roman, Dialiskari); UMME had acknowledged this change in its enumeration of Roman "villa" structures in Messenia. Visible social stratification emerges for the first time in the regional landscape. 5. at Hellenistic ritual activity prehistoric tholos tombs?atTragana at tomb 2, and possibly Osmanaga?together with other as more developments, such visible sanctuary placement and ritual practice (discussed below, pp. 189-191), raise the possibil ity of related developments in the sacred landscape.

CHANGE IN THE RITUAL AND MORTUARY LANDSCAPE

was at one The possibility earlier raised that least PRAP site,M02 (Garga a liani Kalantina [2]), may have been rural shrine in the years of Spartan control. From both textual and archaeological evidence, it is clear that the inhabitants of Messenia, helots and perioikoi alike, maintained cult places era. in that Important locales include the cult of Zeus at Ithome, the newly at discovered Sanctuary Q-Q Messene, the Sanctuary of Apollo Korythos on the , and possibly the site of mysteries at Andania in the Inscribed votives from "the Messenians" have to Stenyclarian plain.115 115. For specific references these as a to cult see Alcock 143 been found, such spear butt dedicated Apollo Korythos, possibly in places, 2002a, pp. the wake of the helot rebellion of the 460s.116 146 and nn. 23-27. 116. Bauslaugh 1990; cf.Hall 2003. Other less cult have been identified, in impressive places tentatively 117. Nichoria and environs: two shrines (one Archaic, one Classical) noted in the Five Rivers cluding Coulson andWilkie 1983, pp. 332, work in the Nichoria Additional scenes of ritual survey region.117 activity 337; Lukermann andMoody 1978, as include ancient Bronze Age tombs where, in many parts of Greece, pp. 110-112. HISTORICAL MESSENIA 189

Figure 22. D01 (Gargaliani Kana View of the PRAP tos). "." Archive

Geometrie practice of tomb cult has been observed. More unusually, Mes seems at a senian tomb cult to have continued, if intermittently and low level, throughout the years of Spartan control, the best-explored example at being Tholos F Nichoria, published under the subtitle "the hero cult cen in the tholos" and dated to the end of the fifth and the early fourth tury b.c.118 Shrines and tombs provided meeting points in the landscape, much in the way of residential clustering, and may well have served the same communicative and supportive functions. All this activity, however, remained small-scale and unobtrusive in character relative to post-libera tion practice. are on The Messenia-wide results from UMME indecisive the point of the obtrusiveness of ritual activity in the Hellenistic period, when the more or or 118. "Hero cult in the tholos": number of unquestioned shrines, which held less steady (at 10 Coulson andWilkie 332 seven 1983, pp. 11) through Archaic and Classical times, decreases to (Fig. 4). Evi see also Alcock 339; 1991, pp. 460-461, dence from the hinterland of Nichoria is also modest; one shrine contin no. 3. On Messenian tomb see cult, ues from the Classical to be another inHellenistic times.119 Korres 1981-1982; Alcock 1991, period, joined by occurs more in the of tomb pp. 456,460-462,465-466; 2002a, Change definitely practice post-Classical cult, which becomes more 10 definite and 11 pp. 146-152; Antonaccio 1995, pp. 70 undeniably frequent?registering van der 1997. more 102; Kamp Antonaccio possible Hellenistic instances inMessenia?and its manifestations (1995, 92-94) to see the most pp. prefers visible. This development is perhaps witnessed spectacularly at the tomb at Nichoria reused as a shelter. "Tomb of at Voidokoilia, on the headland 119. On shrines in the hinterland Thrasymedes" lying prominent the center of and some 3.5 km from of Nichoria, see Coulson and Wilkie opposite polis Pylos/Koryfasion only in the areas atTra 1983, p. 337; Lukermann andMoody Romanou, and, specific intensively surveyed by PRAP, tomb 2 1978, pp. 109-110. gana (Fig. 16) and Osmanaga.120 120. On see Korres a more Voidokoilia, Other indications of public display of ritual behavior in the a recent 1988. For discussion of "post can on Hellenistic period also be posited the basis of PRAP investi Classical" tomb cult inMessenia, see gations. In their reconnaissance of the region, McDonald and Hope Simp Alcock 2002a, pp. 166-167, with a list son had at Kanalos (UMME 15) "several ham of cults on p. 166, n. 66. reported good squared, mer-dressed and thus the existence of a 121.McDonald andHope Simpson blocks," likely temple, 237. on one two 1961, p. of the flat-topped knolls of the site's "acropolis" (Fig. 22).121 E. ET AL. 190 SUSAN ALCOCK

($:?

Figure 23. M04 (Gargaliani_^yfo5 Konstadtinos). PRAP Archive

some or so an In our work at the site thirty years later, which included intensive gridded collection of surface material from the entire ridgetop, we recorded significant damage resulting from bulldozing and found three ashlar blocks. Near these blocks, and around a roughly circular mound of we numerous earth at the site's highest point, located tile fragments painted to black and red; these have been assigned the Hellenistic period, and some manner structure at seem to indicate the creation of of ridgetop cult that time. can Another Hellenistic cult construction be proposed, if with less on a over assurance, forM04, (Gargaliani ^iy/ayKonstadtinos), located ridge was looking the Gargaliani plateau (Fig. 23). In 1994, the site greatly dis a in team had rupted by the demolition of church, which PRAP members a new observed possibly ancient blocks, and by the construction of church two cut in its place. In subsequent investigations in 1995, ashlar limestone blocks, other possibly ancient building material, and Late Classical and wares were no Hellenistic fine found?though material indisputably of votive character.122 an structure at Finally, within the ruins of early modern site A06 (Meta of a morfosi Ayios Konstadtinos [2]), which also exhibits characteristics "spe an cial purpose" site, PRAP has reported apparent column fragment?of very battered limestone, with a dowel hole?and at least two well-cut lime stone blocks.123 The apparent column fragment is not necessarily in situ, use as an or and may have been transported for olive crusher threshing at no floor roller. As M04 (Gargaliani Ayios Konstadtinos), explicitly votive material was discovered, and the site's assemblage, which is dominantly a Hellenistic, possibly points toward domestic function. These observations of more elaborate religious structures inHellenis tic Messenia, at least in some cases at relatively elevated points, suggest no to that ritual visibility was longer be avoided. Indeed, the prominence sites have become asMessenian com of such may increasingly important 122. Davis 1998a, pp. 278-280. as a manner 123. Davis et al. 458-459. munities sought to mark territory their own, in familiar for 1997, pp. HISTORICAL MESSENIA I9I

9?W

wmh

Figure 24.104 (Romanou Romanov), architectural element (possible epi style block). PRAP SF 0574. PRAP Archive

more centuries in other parts of Greece.124 This "normal" patterning of can Greek ritual practice be observed at Romanou, for example, where a several carved blocks, including significant architectural piece that is an probably epistyle, have been found (Fig. 24);125 in the emergence of in civic temples inMessenia's Hellenistic and Roman poleis, known part through Pausanias; and, ultimately, in the appearance of shrines and dedi to at cations the imperial cult, most remarkably the Sebasteion Messene itself. we Although possess strong hints that the Messenian sacred land can scape underwent change after the Classical period, unfortunately little be said about contemporary developments inmortuary patterns. Rock-cut are cist and tile graves located by PRAP at various sites largely unable to at be assigned to specific historical periods. The sites which evidence of possibly Hellenistic mortuary activity has been observed, and the charac teristics ofthat evidence, are:

A01 (Metamorfosi Ayios Konstadtinos [1]): cist and tile graves (roughly Hellenistic-Byzantine) A03 (Metaxada Kalopsana [3]), Locus 2: human bone, found a with small Roman glass vessel B01 (Hora Kalianesi): worked slabs, apparently spolia from the Palace of Nestor, reused in graves of possibly Hellenistic Roman date (Fig. 25)126 D01 one tile 124. Polignac 1984; Alcock and (Gargaliani Kanalos): grave (uncertain date) (Fig. 26); Osborne 1994. three rock-cut graves (uncertain date, though Late Roman 125. PRAP SF now in the ware 0574; African Red Slip may be associated) Hora Museum. G01 (Marathopolis Dialiskari): rock-cut cist graves, completely 126. Davis et al. 481. 1997, p. denuded of finds 127.104 (UMME 400): seeMc G02 (Vromoneri So tira): sherds (Hellenistic and Roman) Donald and Rapp 1972, pp. 310-311, Ayia to own clustered about a catacomb or for references their and previous possible cistern/storage facility work at the site. 104 (Romanou Romanou): traces of graves127 IQ2 SUSAN E. ALCOCK ET AL.

Figure 25. B01 (Hora Kalianesi), collection. Grave cover slabs (center), apparently spolia from the Palace of Nestor. PRAP Archive

Figure 26. D01 (Gargaliani Kanatos), tile grave (with snapped long bones) eroding in bulldozed scarp. PRAP Archive

119 (Pyrgaki Tsouka): human bone and tiles, uncovered in bulldoz a ing of knoll (possibly Hellenistic and Roman, possibly Byzantine-Turkish) 123-27 (Lefki Kaldtamou [1-2, 4, 5, 3]): prehistoric tumuli; the accompanying graves cannot be securely dated128 an com L03 (Nl&ryzXiAyios Ioannis): enigmatic stone structure, a prised of three upright slabs, perhaps the remains of burial a monument; ceramic artifacts, found in nearby bulldozer cut; among these, a bowl (Hellenistic-Roman) to a The chronological indications provided by this evidence point late date (e.g., Hellenistic-Roman) for the mortuary activities described. That inference accords with the of visibil chronology, proposed above, greater 128.123-27 (UMME 14):Mc and in ritual and of social and economic ity prominence behavior, greater Donald and Rapp 1972, pp. 266-267T; differentiation in the landscape. Davis et al. 1997, pp. 485-488. HISTORICAL MESSENIA 193

CONCLUSION

some re This report has attempted to build baseline arguments about a a a gion that presents complicated and unusual post-prehistoric history, history that has contributed to its frequent archaeological neglect. How far our and contentions are an artifact of our under reasoning present standing will, of course, be tested by time and further research. At present, to Messenia appears display certain patterns of regional activity somewhat out of step with those observed in other parts of the Greek mainland and islands. The relatively late advent of small "farmstead" sites inMessenia, by comparison with the dates of their appearance in other regions reported one in survey results, is example of this difference, aligned here with the peculiar history of Messenia in Archaic and Classical times. Another is are re the apparent proliferation inMessenia of Roman villa sites, which ported in relatively few numbers outside of the regions of Messenia and near her neighbor, Achaia. Such observations underline the need for con stant comparison of regional data sets and for sensitivity to local trajecto as as more ries and idiosyncrasies, well to overarching, shared patterns of

development.129 we can a To conclude, briefly illustrate the usefulness such compara tive approach has had in studies of the difference in settlement patterns of Laconia and Messenia at the time of Spartan domination. The Laconia an area to a Survey, operating in adjacent Sparta, recovered pattern of dis persed, small-scale settlements where, it has been argued, helot cultivators a reverse dwelled in mixture of isolated farmsteads and hamlets?quite the of the Messenian picture. As Richard Catling has noted, comparing set tlement patterns of Laconia and Messenia with reference to these results, "a clear distinction begins to emerge in the ways in which these two were groups distributed in the landscape, and presumably the ways inwhich two were the regions farmed."130 Stephen Hodkinson takes the implica a tions of these findings step further: "In short, those helot farmers under masters were the closest degree of supervision by their Spartiate settled 129. The need for such comparative a in pattern of residence less conducive to collective coordination of agri as well as the difficulties approaches, cultural the helots themselves."131 The ramifications of are production by they entail, discussed in the papers these differences remain to be but their inAlcock and Cherry 2004. completely explored, implications for a obscure area in Greek and are 130. Catling 2002, p. 253. illuminating very history archaeology 131. Hodkinson 270. 2003, p. crystal-clear. APPENDIX CERAMIC OBSERVATIONS, HELLENISTIC THROUCH LATE ROMAN

Questions and Problems132

Pottery constitutes by far the lion's share of the abundant and wide-rang was ing data retrieved by PRAP. Pottery collected assiduously, in part for the information that sherds encode about their origins and functions? so and about the local and foreign connections and activities of the region's ancient inhabitants?but primarily for intrinsic chronological informa more a tion. It is than little disappointing, therefore, that the collected pottery of the Hellenistic and Roman periods?during which fundamen more tal political, religious, and economic shifts occurred?provides ques answers tions and problems than and insights. a one The first question is simple of quantity. Why is there relatively so more little pottery from the Hellenistic and Roman periods? Far pieces were of definitely dated pottery found dating to the Late Helladic period (ca. 1680-1060 b.c.) than to the roughly equivalent Hellenistic-Early b.c. a.D. Roman period (ca. 323 to 400).133 Two possible explanations may was be offered. One is that there significantly less settlement, at least within the PRAP survey area, during the later period. If true, this would be an important and interesting aspect of long-term patterns of settlement in the region. more Another, troubling explanation may also be offered, however: we that simply cannot recognize the vast majority of the region's locally produced Hellenistic and Roman pottery. A fundamental problem with pottery picked up in survey is that it is bereft of its single most vital chro a context. nological clue: specific depositional Without the surety of other

nature of the that can in collections can 132. The production of this article interpretations be discovered these be a team are not has been effort, but it should be drawn from it, shared by all the reliably dated to the Late Helladic that this is as to noted appendix the work of authors. period (565 sherds) the Helle two 133. Given the variation in collec individuals only: Andrea Berlin nistic and Early Roman periods tion (Hellenistic through Early Roman, methodologies among the sites, (142 sherds). The known presence of most to measure in submittedMarch 1999) and Sebastian the unbiased way Mycenaean remains the study area, ex on not was Heath (Later Roman). The views quantity is the basis of material least the Palace of Nestor, of at recovered in the initial tract course a reason pressed points here, concerning the collections. chief for the selection data set the four times as of area. quality of the survey and Roughly many sherds the survey HISTORICAL MESSENIA 195

a secure associated remains, it is impossible to assign and specific data or range to ambiguous unidentifiable sherds. Surveys in regions where there has been long study of local, stratified assemblages may successfully a are avoid this problem, because relatively small percentage of their finds unrecognizable.134 InMessenia, the Bronze Age constitutes such a well a studied period, and in fact only small percentage of the material picked areas up by PRAP in of known, heavy Bronze Age occupation (e.g., around the Palace of Nestor) was considered unidentifiable. That was not the case for material of the Hellenistic and Roman periods, however. Stratified pottery has only recently been the object of study at the region's largest site,Messene, and little has yet been published. Of the few or stratified narrowly dated assemblages from elsewhere in the P?loponn?se, however, one fact is quite clear: the pottery of the Hellenistic and Roman periods is very "local" in its varieties, styles, and decorative effects.135 Bailey's comments on at the Hellenistic and Roman pottery recovered Sparta ap ply equally well toMessenia: "as always in the early stages of the study of ... are locally produced pottery [t]he dates suggested all rather uncertain ... and comparatively wide; most of the pottery is little known archaeo logically."136 It is likely that many sherds currently identified in the PRAP as are databases Hellenistic and Roman not at all from those periods; and conversely, it is likely that significant quantities of unidentified sherds within the PRAP database derive in fact from Hellenistic or Roman times. a This fundamental aspect of the ceramic database necessitates heavy, an on almost exclusive, reliance securely identified and dated imported wares wares, especially fine table and amphoras.137 In fact, Hellenistic and/ or at area Roman-period occupation any site within the PRAP study could be attested only if such imports existed. This in turn leads to another cru or cial quandary: PRAP simply cannot recognize account for those times wares were or in which imported few, rare, absent. As it happens, such wares was times constitute the majority. Local production of table always more, rather than less, common, and the importation of wine and oil? carried in transport amphoras?was generally unnecessary given Messenia's nar agricultural productivity.138 There were, in fact, only three relatively row to periods of time in which Messenians turned outside producers.139 was One the later Hellenistic period (the second and first centuries b.c.); was the second the Early Imperial period (first century A.D.); and the third was the Late Roman period (fifth-seventh centuries A.D.).

answer or 134.The Sydney Cyprus Survey demand the yes' no,' and 135. Much of the pottery from is a sur ... Project good example of such dislike 'maybe,' perhaps.'... Hence Hellenistic {CorinthVII.3; et as are to veys (Given al. 1999), any of the pottery expert is expected iden Corinth XVBl.l; Romano 1994), the ware can grids surveyed under the auspices tify every of every period. We (Bruneau 1970), of if us of the Archaeological Survey Israel do that, of course, you don't ask (Rudolph 1978), and Sparta (Bailey Frankel and Getzov how we reach our As is to (e.g., 1997). conclusions.... 1993) unique each site. offers we is John Hayes general but apt know, survey work extremely 136. Bailey 1993, pp. 221,249. comments on the of for noses to problems "read good getting people's the 137. Bailey (1993, p. 221) followed, "As have but without exca an identical at ing" survey pottery: surveys ground, supporting procedure Sparta. become more so the vation we have no chance of methodical, classifying 138. VanWersch 1972. demands made on ... have our finds with the of 139. specialists necessary degree See "Imported Wares," below. more become exacting. Computers detail" (Hayes 2000, pp. 105,107). 196 SUSAN E. ALCOCK ET AL.

Only the last of these three periods, the Late Roman, corresponds a precisely to specific part of the PRAP chronological schema (Table 1). The first two?later Hellenistic and Early Imperial?fall within periods are more that defined by PRAP much broadly. The PRAP schema consid ers ca. Hellenistic settlement to represent roughly the years from 323-31 b.c. to A.D. b.c., and Early Roman to represent roughly the years from 31 400. This leads to the last large question that the ceramic data inspire: mean to what exactly does it talk about "Hellenistic" and/or "Early Ro man" settlement within the PRAP study area? are a PRAP's period dates clear and defensible from political and mili are are tary point of view; in other words, they "historical" dates. They not, are not however, "archaeological" dates, which is to say, they congruent with the known production dates of the imported pottery. Since it is the on so pottery that comprises the "facts the ground" and testifies to the to period of activity represented, it is important articulate precisely what are see the pottery dates (for which, Tables 7-9). Given this discrepancy in as chronological divisions, shifts in quantities of material?such the in crease to in numbers of sherds from the Classical Hellenistic period? should be understood to indicate, for example, "150 years of local produc tion (difficult to recognize and not closely datable) followed by the appearance of identifiable imports in the second century b.c." The shift an may indeed indicate "increased tempo" of activity, but may instead in dicate that only then did Mediterranean-wide shifts in commerce, travel, on and/or production leave traces theMessenian landscape.140 In sum, the recov quantitative and chronological ambiguities of the ceramic evidence are so amount ered by PRAP profound that its utility for identifying the and date of Hellenistic and Roman activity is, it would seem, very ques tionable.

Local Wares

account an Imports for extremely small percentage of the total number of sherds retrieved by PRAP. Most of the pottery represents "local" produc tion from a site, or sites, within Messenia, as yet unidentified. While spe or wares cific dates forms could rarely be assigned, three commonly ap pearing could be isolated. None of these three have been sampled by or neutron activation and so none can be as p?trographie analysis, firmly or signed to mineralogical chemical groups. It is notable, however, that wares these three PRAP find ready correspondence with the three Messe nian wares that Frederick Matson identified for UMME.141 We have termed

these as follows: Messenian Household Ware, Messenian Utility/Cooking Ware, and Messenian Coarse Ware.

Messenian Household Ware

This is the fabric most found for both and commonly slipped plain pottery 140. See below for discussion of of the Hellenistic-Late Roman The is dense and fine periods. paste fairly Late Hellenistic and Early Roman micaceous but otherwise with little visible It is grained, slightly temper. import patterns. not a at 141. Maison 201-203. quite soft, chalky in texture, and consequently does hold slip well 1972, pp. HISTORICAL MESSENIA 197

an to a all. The color ranges from even, light reddish yellow (5YR 7/8) more a warmly hued light orange (5YR 6/8). Pieces of paler light brown occur are occasionally (7.5YR 7/6-8/6). Vessels almost always fully fired. are Most forms are somewhat thick walled (0.4-0.6 cm), and generally even or treat finished simply, cursorily, with few extraneous details surface seem to ments. This fabric would correspond to UMME sherds with "fine texture, pale yellow, pink, and light red in color," and with "few visible inclusions except occasional bits of limestone and ochre, and minute flakes of muscovite," although the description is given in the context of Late Helladic ceramic remains.142

Messenian Utility/Cooking Ware

occurs This fabric most commonly among vessels intended for cooking as same as and kitchen tasks, such basins and jars. It is of the color Messenian a Household Ware, but differs in having profuse amount of sand inclu or sions. The sand may have occurred naturally within the clay have been a or deliberate addition intended to harden strengthen the vessel walls. ware This is fairly soft, despite the sand, and the surface has generally to are weathered away. The fabric corresponds UMME sherds that de as a scribed "coarse textured, made from naturally sandy clay," about which not as Matson said further, "it is possible to speak of this group intention ally tempered without better knowledge of the textural variations in the local clays."143

Messenian Coarse Ware

a coarse ware to This term is reserved for distinctive, exceptionally used some manufacture pithoi and large basins. It shares the same color range as the Household and Utility/Cooking wares, from light yellow brown are to (7.5YR 7/6) to light pink (5YR 7/6). Vessels very thick walled (up a core. seems 2 cm), and there is almost always light firing The paste a clearly to have had variety of added tempering agents, which appear as visually small, medium, and large, rounded and angular, red, white, gray, and brown inclusions. Most of the small and medium rounded in are ware as clusions probably sand. Matson identified this "intentionally some tempered pottery, e.g., for pithoi," and proposed possible identifica tions for the inclusions.144 The red may be red shale, which breaks into mm hard, angular grains up to 5 in length. The gray, white, and brown inclusions may be chert, and the small and medium rounded lighter brown may be mudstone.

Imported Wares

While the specific output and production dates of many Hellenistic and are Roman-period pottery manufactories known, imported pottery still a presents problems of identification, primarily due to tremendous amount of which black- and table ves 142. Matson 1972, p. 201. "copy-cat" production through red-slipped sels of similar were theMediterranean basin. 143. Matson 1972, p. 202. shapes produced throughout 144. Matson 1972, p. 203. A single apt exemplum illustrating the magnitude of these problems is the 198 SUSAN E. ALCOCK ET AL.

con bowl with incurved rim. Beginning in Early Hellenistic times, and at tinuing least through the fifth century A.D., almost every known fine table ware manufacturer in the central and eastern Mediterranean pro a duced similar version of this shape.145Meanwhile, the all-purpose utility of the form, perhaps coupled with the ubiquity of imported models, in as spired myriad "local" versions well.146 Regrettably, Messenia's alkaline cause over soils most imported pottery to lose its surface slip time, while the vessels manufactured in the friable local clays do not hold their origi nal slips well to begin with. Moreover, the interior clay color (that is, the a color of sherd that has lost its surface slip) of Messenian Household Ware is quite similar to the interior clay color of many different imports. As a consequence of these conditions, it is sometimes impossible to iden an an tify unslipped rim fragment of incurved rim bowl?one of the most common forms of table ware?in terms of or even specific ware, origin, narrows period. This circumstance further the field of informative survey sherds: not only to imports alone, but to those pieces among the imports are that at present sufficiently distinctive.

Late Hellenistic

wares A small but interesting array of imported fine and amphoras dating b.c. from the mid-second through the first century is attested from Messe a a a nia. PRAP recovered Campana A plate, Campana B bowl, and first b.c. as as century Eastern Sigillata A (ESA) bowl, well two fragments of Dressel (Dr.) 1A amphoras (see Table 7). Other finds from the immediate region fill out this picture considerably. From the Divari and Tsopani Rahi come cemeteries black-glazed moldmade and hemispherical bowls, gray ware (Attic?) ung?entar?a, white-ground lagynoi, and amphoras from Brin as as disi, well coins dating between 280 and 146 b.c.147 From Messene a itself comes tremendous amount of Hellenistic Eastern Sigillata A, in a cluding fishplate (the very earliest shape, dating 150-100 b.c.), gray ware Graeco-Italic and Dr. 1 and ung?entar?a, amphoras, Aegean lagynoi.148 at The imported pottery found Messene, Divari, and Tsopani Rahi, one and by PRAP at sites throughout the survey area, reflects point of outstanding interest: by the middle of the second century b.c., Messenia received table wares from both eastern and western producers, but practi

ware cate 145. Bowls with incurved rim ap century a.D.), African Red Slip Messene that surely fall into the in form late to as pear both early and late Eastern {LRP, 27, second early gory of "local production" well. Sigillata A (Hayes 1985, ESA form 20, third century a.D.), and Phocaean Red 147. Tomb contents of both the a.D. 80 ware to Rahi cemeteries 150-100 b.c., and form 65, Slip {LRP, form 1, late fourth Divari and Tsopani are in museum at 120), Cypriot Sigillata (Hayes 1985, late fifth century a.D.). displayed the Pylos. no. 148. The contexts of the Messene form P20, early first century a.D.), 146. Bailey 1993, p. 227, 33 are Eastern Sigillata B (Hayes 1985, ESB (Sparta);Rudolph 1978, pp. 228-229 finds pit XVI/2/XQPI 01 and form 14, mid-first century a.D.?), Pon (Mycenae); Bruneau 1970, pp. 522-523 the second of three successive floor tic Sigillata (Hayes 1985, form 4, late (Argos); Corinth VII.3, pp. 29-33, fills from XVI/2S5. We thank Petros ware In we Themelis for us this mate first century a.D.), early ?andarli pis. 2,43,44 (Corinth). fact, showing form do not know of site where "local and for us to discuss it (Hayes 1985, 16, second century any rial, allowing ware of this form is not here. a.d.?), later ?andarh (Hayes production" iden 1985, form H4, late second to third tified. We have examined vessels at HISTORICAL MESSENIA 199

TABLE 7. SELECTED LATE HELLENISTIC IMPORTS AND IMITATIONS RECOVERED BY PRAP

PRAPArtifact Ware Shape Formt TypeNo* Date or Site No. b.c. Campana A Plate Morel 2233b Ca. 250-200 E93-901312-06 B Bowl Campana E93-901122-01 ESA Bowl EAA form 22B Ca. 110 b.c 104 a.d. 10 ? E01 Amphora Dressel 1A Late second-mid- b.c. first century

ware are *Form and type numbers for Campana derived from Morel 1981; for Eastern SigillataA (ESA), fromHayes 1985 (=EAA).

no eastern Mediterranean commodities?for am cally example, transport a wares phoras. Taken in Mediterranean-wide context, the variety of table a is not remarkable; similar picture appears at, among other sites, Sabratha, Berenice (Benghazi), Athens, and .149 In neighboring Laconia, how are no at ever, there practically imports all in the Hellenistic period, nei wares nor ther table amphoras, and only in and after Augustan times does a there appear scattering of both eastern (Eastern Sigillata B and Pergamene western relief ware) and (Italian Sigillata) table wares, along with both eastern (Cretan Dr. 43) and western (Dr. 6) amphoras.150 The wide range wares of imported table attested inMessenia demonstrates, in contrast, at least some connections with the wider Mediterranean world.

Several scholars have suggested that the appearance of Italian table wares and amphoras at Athens and Delos reflects the latter's Athenian its economic in takeover, attendant transformation, and the consequent on volvement of Italian businessmen the island.151 Elizabeth Lyding Will has termed this development "economic Romanization."152 Comparable influences could account for the specific and various imports attested in Messenia, for the near absence of Late Hellenistic eastern Mediterranean amphoras, and for the few examples of western (Dr. 1) amphoras: Italians on traveling to and from the Aegean, who may well have visited Messenia on occasion, would have been more likely to convey commodities their over on way than their way back.153 The specific types of Late Hellenistic are more imports found inMessenia suggestive of such circumstances of commerce. casual acquisition than of deliberate

exem were 149. Fulford andTomber 1994, 1970, pp. 513-518), the Spartan Italian wine exports intended for arrived from there via to pp. 2-4 (Sabratha);Kenrick 1985 plum may have commercial redistribution Aegean (Berenice);Agora XXIX, pp. 221-223 local transit. natives, and not for resident Italian seems (Athens); Rotroff 1997, p. 98 (Athens); 151.Morel 1986, p. 488; Rotroff domestic consumption wanting." Morel 1986 (Delos). 1997, p. 99;Will 1997, pp. 120-129. 153. The Italian predilection for the wine in Dr. 1 am 150. Bailey 1993, pp. 221-222. The 152.Will 1997, p. 120. Nicholas Campanian carried vessel firm out is single nonlocal Hellenistic Rauh (1999, p. 171) points that in phoras well illustrated by the vessels identified at is a finds is not so Roman ly Sparta white-ground terpreting these simple: that surveyors and/or soldiers no. at at lagynos (Bailey 1993, p. 232, 79, "As with Campanian 'black glazed' left Corinth the end of the second a ware in evi b.c. fig. 8). Since, however, great many table [found the Aegean], century See Romano 1994, pp. 63, at of nos. have been found Argos (Bruneau dence capable demonstrating that 86-88, 63-69. 200 SUSAN E. ALCOCK ET AL.

TABLE 8. SELECTED EARLY ROMAN IMPORTS AND IMITATIONS RECOVERED BY PRAP

Ware No.* Date PRAP No. Shape Form/Type Artifact

Thin-walled Beaker Moevs LXVIII Augustan I93-9041611GR-05 ESA Bowl EAA form 42 10 B.C.-A.D. 20/30 I93-9040321GR-05 ESA Bowl ??Z/fform46 A.D. 1-25 I93-9041611VC-09 Imitation ESA Plate EAA ?oxm33 a.D. 1-30/50 I93-9040453GR-02 Thin-walled Beaker Moevs XLVII Tiberian-Claudian G94-9011021GR-02 A.D. Imitation Cypriot Bowl EAA form P5/6 Early-mid-first century I93-9040551GR-10 Sigillata Imitation ESA Plate EAA form 34 A.D. 40-70 I93-9040563GR-01 Imitation ESA Bowl ?^form48 a.D. 40-70 G94-9011021GR-03 Imitation ESA Bowl EAA form 49 a.D. 40-70 G94-9010991GR-04

ESB Plate EAA form 8 Mid-first century A.D. D93-901121-02 a.D. ESA Bowl with stamp: Mid-first century 192-027-09 XAPIC Imitation ESB Plate EAA form 17A Ca. a.D. 50-100 I93-9040453VC-04, G93-097-01 a.D. Imitation Cypriot Bowl ?/??/formP12 Ca. 50-150 I93-9040441GR-14 Sigillata ESA Plate EAA form 36 a.D. 60-100 I93-9040471GR-16 ESA Bowl EAA form 50 a.D. 60/70-100 192-057-02 ESA Bowl EAAioim5\ a.D. 70-120 G94-9010991VC-05 ESA Bowl EAA form 52 Hadrianic 192-045-08

ware are *Form and type numbers for thin-walled derived from Moevs 1973; for Eastern Sigillata A (ESA), Eastern Sigillata B (ESB), and Cypriot Sigillata, from Hayes 1985 (= EAA).

Early Roman

case As is the for PRAP's Late Hellenistic imported pottery, that of the cem Early Roman period is helpfully augmented by finds from the Divari etery and Messene itself. PRAP finds include many Eastern Sigillata A an bowls and plates and Eastern Sigillata B plate, imitations of Eastern ware Sigillata A and B and Cypriot Sigillata, and Italian thin-walled (see Table 8). The imitations imply knowledge of imported origins and, in fact, the Divari tombs did contain Augustan-period beakers of Italian thin an enormous a walled ware. At Messene, pit contained great deal of Au gustan- and Tiberian-period Eastern Sigillata A bowls, Italian cooking a.d. ware, and local versions of first-century thin-walled and Eastern are Sigillata B forms, all covered in black glaze.154 Notably absent Italian as as fine wares, well transport amphoras from both the eastern and west ern Mediterranean. wares Eastern Sigillata A table represent the majority of the region's a Early Roman imports, and this is surely meaningful manifestation. or Whether not the much-reduced number of Italian products inMessenia indicates reduced traffic from the west, the increase in eastern products more east. to reflects from the In order evaluate the context certainly activity sig 154. The of the Early nificance of this Messenia's should be from increase, import patterns compared Roman-period finds Messene a with those elsewhere in the P?loponn?se. The comparison reveals, again, is again pit XVI/2/XQP I 0r HISTORICAL MESSENIA 20I

dramatic difference between Messenia and Laconia?more specifically, a Sparta. Excavators at Sparta have identified few pieces of Italian Sigillata, a one couple of Eastern Sigillata B bowls, and, exceptionally, Pergamene a relief-ware bowl, but not sherd of Eastern Sigillata A.155 This evidence some indicates that Sparta acquired few luxury goods from Italy and from Minor, but not from the southeastern Mediterranean. on At Corinth, the other hand, Eastern Sigillata A comprised 24 per wares cent of the imported table in late Augustan and Tiberian times, while Italian Sigillata accounted for almost the entire remainder.156 The range and quantity of Italian Sigillata found at Corinth represents active and deliberate commerce, supplied specifically from Arezzo. The city's as a economic vitality during this period has been explained direct out new growth of its colonial status and of its favorable position astride the isthmus, which was able to accommodate traders from east and west.157 The significant disparities among the Early Roman imports attested at Corinth, at Sparta, and inMessenia reveal regional idiosyncrasies in are trade and in the acquisition of imported luxury goods that explicable according to the locales' specific situations and settings. In the Early Ro man period, for example, Corinth's political fortunes affected its economic case as ones; and it may be the that inland cities, such Sparta, received some "trickle-down" activities and benefits. Messenia, on the other hand, seems to cut have been off from the west?perhaps because traffic from Italy, which in Late Hellenistic times had sailed around and sometimes now stopped inMessenia, aimed instead at Corinth and sites north. While Italian traders ignored Messenia in the Early Roman period, the region's inhabitants received, in quantity, the products of the south eastern Mediterranean potteries. The spread of large quantities of Eastern Sigillata A into the central Mediterranean at this time iswell attested by even the finds from , Berenice, Sabratha, Carthage, and Pompeii.158 Explanation of this spread has been sought in the circumstances of the to to supply of Egyptian grain Rome.159 Shippers sailed north from Egypt was Phoenician and north Syrian ports, where Eastern Sigillata A pro duced, and thence to Puteoli.160 Redistribution from Italy may account for the notably privileged position of Eastern Sigillata A at sites in southern as with other eastern fine as well as for the unusual Italy compared wares, that now in North Africa?abundant Eastern 155. Bailey 1993, p. 222. import patterns appear 156. 193. ves Slanel987,p. Sigillata A alongside Italian-made amphoras, mortaria, and cooking 157. Slane 1987, 200; Corinth amount p. sels.161 InMessenia, the large of Augustan- and Tiberian-period XV??I.2;Wright 1980; Slane 1986. Eastern Sigillata A?along with the telling absence of other contempo 158. Sackett 1992, pp. 150-152 rary eastern fine wares?suggests that may have (Knossos); Kenrick 1987, pp. 141,144 Egyptian grain ships in the southwestern on their toward The (Berenice); Fulford andTomber 1994, stopped P?loponn?se way Italy. p. 3 (Sabratha);Hayes 1976, p. 77 contemporary absence from Messenia of the widely distributed Italian fine Pucci 1977 as so common ex (Carthage); (Pompeii). table wares?such Arretine, then at Corinth?may be 159. Fulford andTomber 1994, plained by Messenia's agricultural prosperity, which exempted her from pp. 3-4. the imperial redistribution of grain carried out, for example, at Carthage, 160. Rickman 1980, p. 129. On the source Sabratha, and Berenice. probable north Syrian of ESA, The contrast between the of the later Hel see Slane 1997, p. 272. admittedly scanty imports 161. 74-75. an Riley 1981a, pp. lenistic and Early Roman periods inMessenia presents ironic picture. 202 SUSAN E. ALCOCK ET AL.

seem In later Hellenistic times, Messenia's few foreign visitors likely come as to have from Italy, and insofar Messenia enjoyed international as or connections, they may be characterized western, at least central, Medi terranean. In Early Imperial times, however, Messenia became a region bypassed by Rome and linked instead, albeit ephemerally, to the east. a Augustus may have transformed the Mediterranean into "Roman sea," but Messenia's inhabitants seem not to have benefited thereby.

Later Roman

a.d. The second and early third centuries in the PRAP study area ofMes are senia not easily defined in terms of ceramics, not least owing to the absence of imported fine wares. In the western Mediterranean, this is the period when African Red Slip (ARS) becomes dominant; in the Aegean, a it is the period of the so-called "?andarh monopoly," term that high lights the dominance of ?andarh among red fine wares at many sites, including Athens and Knossos.162 ?andarh is also known at Berenice and area re Ostia.163 Its absence from the PRAP study further suggests that the was not were con gion served by the main routes of long-distance trade that wares to veying cheap, well-made fine tables around theMediterranean. on We are firmer ground with the evidence of later ARS (see Table 9). a Tract walking and site collection in the PRAP survey territory recovered ware. total of 94 sherds of this The "villa" at G01 (Marathopolis Dialiskari) and the large settlement at 104 (Romanou Romanou) produced 34 and 45 such sherds, respectively. The only other relatively large recording of ARS is that of 11 sherds at D01 (Gargaliani Kanalos), perhaps to be associated with rock-cut graves. Sites A04 (Metamorfosi Skarmingd) and G02 (Vro an areas moneri Ayia Sotird) each produced unclassified body sherd, and an VII and VIII each contained off-site LRP 53A rim (late fourth-early are fifth century A.D.). Romanou and Dialiskari the enduring features in access the Roman landscape, and their to these imports is presumably both a cause and a symptom of this stability. The range of ARS forms in the PRAP study area, while covering the main categories of the ware, is hardly comprehensive. Of the 94 sherds were common was recorded by PRAP, 40 diagnostic. The most form were the LRP 50 plate, of which eight rims and three possible bases found. can In the absence of stratified deposits, the form be difficult to date more to to re precisely than the early third early fifth centuries. Sherds covered by PRAP include extremely thin-walled and highly polished can an vessels that be of early-third-century date, and LRP 32/58 and an are LRP 5 8A, from Kanalos and Dialiskari, respectively, also broad a ly datable from the late third into the fifth century. Two LRP 67 and single LRP 104B complete the series of large plates found by the proj ect. The rest of the identified ARS from Romanou and Dialiskari con

sists mainly of LRP 80A and LRP 99 bowls of the fifth and sixth centu at 162. On see LRP. ries. The only additional piece of note, found Dialiskari, has exterior ARS, "?andarh near to monopoly": Coldstream, Eiring, and rouletting and is in form Bonifay type 76.164 In general, then, the Forster 2001, p. 145. range of ARS forms is relatively incomplete?with mainstays of the 163. Kenrick 1985, pp. 257-265 series such as LRP LRP and LRP 91 or not iden 61, 87, missing securely (Berenice);Martin and De Sena 2003, of sustained access to North African tifiable?but nonetheless indicative p. 44 (Ostia).

products. 164.Bonifay2004,p.203. HISTORICAL MESSENIA 203

TABLE 9. SELECTED LATE ROMAN IMPORTS RECOVERED BY PRAP

PRAP SiteNo. Formt No* Date Ware Shape Type (No. of Examples Recorded)

Deep ARSplate LRP 50 Early third-early fifth century a.d. D01 (2), G01 (2), 104 (4) Deep ARSplate LRP 32/58 Late third-early fifth century A.D. D01 (1) DeepARS plate LRP5SA Ca. a.d. 290-375 G01 (1) fifth a.d. Deep ARSplate LRP 53A Late fourth-early century Off-site (2) DeepARS plate LRP 67 Ca. a.d. 360-420 G01 (1),104 (1) ARSBowlLRP 80A Mid-late fifth century a.d. G01 (2) 76 A.D. Deep ARSplate Bonifay Mid-fifth-early sixth century G01 (1) FlangedARS bowl LRP 91? Mid-fifth-early sixth century A.D. 104 (1) (base only) LRP a.d. Deep ARSplate 104B Late fifth-mid-sixth century G01 (1) ARSLRPBowl 99 Late fifth-mid-seventh century a.d. G01 (5) LRP ID fifth a.d. Deep PRS plate Early-later century 104 (1) DeepPRS plate LRP 3C Ca. a.d. 460-490 G01 (4) DeepPRS plate LRP 3D Late fifth century a.d. G01 (1) LRP 3F A.D. Deep PRSplate Early-mid-sixth century G01 (1) DeepPRS plate LRP5B Ca. a.d. 500-550 104 (1) DeepPRS plate LRP 10 Ca. a.d. 570-660 A04 (1),G01 (3), 104 (1) ? Grande A.D. Amphora Africano Third-fourth century G01(lor2) ? LR a.d. Amphora 2 Fourth-sixth century D03 (1), G01 (9), G03 (1),104 (2) ? LR 4 a.d. Amphora Fourth-sixth century G01 (1) ? LR a.d. Amphora 5/6 Fourth-sixth century G01 (1)

*Formand type numbers forAfrican Red Slip (ARS) are derived fromLRP and Bonifay 2004; for Phocaean Red Slip (PRS), fromLRP; for amphoras, from Riley 1981b. ware area The other Late Roman fine found in the PRAP study is Phocaean Red Slip (PRS), of which surface collection recorded 16 rims two can and bases. The predominance of diagnostic pieces be explained by the fact that the thin slip of most PRS sherds may not have survived in the contrast acidic soils of the region, in to the sturdier slip of ARS that allows even to small body sherds be confidently identified. Comparison of the 18 a PRS diagnostics to the 40 ARS diagnostics is thus better indicator of the wares relative amounts of these than comparison of the total sherd counts for each. One also needs to account for the fact that the identified PRS forms?one LRP ID, amaximum of eleven LRP 3, one LRP 5B, and five LRP 10?range in date from the early fifth to the seventh century A.D. Given this chronology for PRS, any allocation of the broadly dated ARS forms to the fourth century will decrease the proportion of ARS to PRS was during the time when PRS imported. The imported amphoras suggest limited contact with the Mediter A.D. ranean-wide market for transported foodstuffs in the early centuries access but growing to imports in the fourth century and later. The third a century may be represented by so-called Forlimpopoli amphora165 and one or two by Africano Grande rims, which can, however, date substantially numerous as later. The discovery of combed body sherds, well as 13 rims, of Late Roman Amphora 2 (LR 2) from the Aegean basin indicates that 165. PRAP artifact no. D93 trade may have increased in the fourth In however, 901352-01; similar in form to Corinth century.166 general, no. the Late Roman is poor. A LR 5/6 Palestinian XVIII.2, p. 116, 252, fig. 29. amphora assemblage single sherd and a LR 4 Gaza rim indicate at least minimal access to these 166. For the LR typology, see Riley single 1981b. common relatively forms, but the absence of the easily recognizable LR 3, 204 SUSAN E. ALCOCK ET AL.

a on from southwestern Asia Minor, places firm limit the characterization of this region's integration into the Mediterranean economy. How does this Late Roman assemblage compare with that of other predominantly rural regions known from systematic surface collection? no wares The Laconia survey recorded almost Roman fine of any period, are ex though imports known from Sparta itself.167The publication of an to cavated late-fifth- early-sixth-century ARS bowl atMegalopolis fur ther confirms the availability of imports in urban areas.168 Published mate a rial from the Northern Keos survey provides selection of fifth- and to sixth-century ARS roughly comparable that recovered by PRAP, but its more fourth-century assemblage is substantial than theMessenian, and its a Pontic Sigillata, Late Roman Athenian fine ware, and LR 1 indicate 167.Lawsonl996,p.lll. more set of economic interactions.169 complex 168. Gans 1998. at from Looking comparanda further afield, the assemblage recovered 169. Sutton et al. 1991. Vroom assem in the Biferno valley survey in Italy generally matches the PRAP (2003) shows thatARS, PRS, and LR 2 are blage, though it contains the important addition of two sherds of late also present in rural . see to 170. Barker 1995; especially sites fifth- early-sixth-century Cypriot Red Slip.170These two centuries have A198 andA248. been as a time of active trade in the Mediterra recognized long-distance 171. For further discussion of trade nean. of the PRAP area in this network is certain? Participation study in the Late Roman Mediterranean, see even no means when with perhaps regular?but by impressive compared Reynolds 1995; Kingsley and Decker that of other regions.171 2001.

REFERENCES

XXIX = Hellenistic mies Agora S. Rotroff, ofAncient Greece, ed. P. Cart Athen ian and E. E. L. Pottery: Imported ledge, Cohen, and Foxhall, Wheelmade Table Ware and Related London, pp. 185-199. Material S. and F. eds. {AgoraXXIX), Princeton Alcock, E., J. Cherry, 2004. 1997. Side-By-Side Survey: Compar S. E. 1991. "Tomb ative Studies in Medi Alcock, Cult and Regional the the Post-Classical Polis,"AJA 95, terranean World, Oxford. 447-467. pp. Alcock, S. E., J. F. Cherry, and J. L. -. 1993. Graecia capta: The Land Davis. 1994. "Intensive Survey, Roman scapes of Greece, Cambridge. Agricultural Practice, and the -. 1998. "Liberation and Con Classical Landscape of Greece," quest: Hellenistic and Roman Mes in : Ancient Histo in Davis 179-191. ries and Modern senia," 1998c, pp. Archaeologies, -. 1999. "The Pseudo-History of ed. I.Morris, Cambridge, pp. 137 Messenia Unplugged," TAPA 129, 190. pp. 333-341. Alcock, S. E., and R. Osborne, eds. -. 2001. IV 1994. Gods: Sanctuaries "The Peculiar Book Placing the and the Problem of the Messenian and Sacred inAncient Space Greece, Past," in Pausanias: Travel and Oxford. in Roman ?d. S. E. C. M. 1995. An Memory Greece, Antonaccio, Archae and Ancestors: Tomb Cult and Alcock, J. Cherry, J. Eisner, ology of Hero in New York, pp. 142-153. Cult Early Greece, Lanham, -. 2002a. Archaeologies of the Md. Greek Past: K. 1996. Pausanias' Greece: Landscape, Monuments, Arafat, and Memories, Cambridge. Ancient Artists and Roman Rulers, -. 2002b. "A Simple Case of Cambridge. in M. 1993. at Exploitation," Money, Labour, Bailey, D. "Excavations and Land: to the Econo The Roman Approaches Sparta: Stoa, 1988-91, HISTORICAL MESSENIA 205

1 a Greek Rural The Laconia P. N. 1990. "Actia Preliminary Report, Part (b): Landscape: Doukellis, Nicopolis: 1: and structures ur Hellenistic and Roman Pottery," Survey Methodology Interpre Id?ologie imp?riale, et BSA 88, pp. 221-249. tation {BSA Suppl. 26), ed.W. Cav baines, d?veloppement r?gional," R. 1980. Le de R. W. V. 399-406. Baladi?, P?loponn?se anagh, J. Crouwel, Catling, JRA 3, pp. Strabon: Etude de and G. 151 1974. "Le des g?ographie Shipley, London, pp. Ducat, J. m?pris hilotes," Paris. 1451-1464. historique, 256. Ann?conSocCiv 29, pp. F. 1983. -. 1978. de Baltas, C. A. 1987. IIo?o?: Na?apivo, Cherry, J. "Frogs Round the "Aspects Fhilotisme," Pond: on Ar Ancient 5-46. Ni?xaoTpo, Av?xTopo N?aropo?, Perspectives Current Society 9, pp. in -. Athens. chaeological Survey Projects the 1990. Les hilotes {BCH Suppl. -. in 1997. n??o?: Na?apivo, Ni? Mediterranean Region," Archae 20), Athens. in Mediterranean F. 1983. oder xaozpo, ?vaxzopo N?aropo?, 2nd ological Survey the Feiten, "Heiligt?mer ed., Athens. Area {BAR-IS 155), ed. D. R. Keller M?rkte?" AntK26, pp. 84-105. ed. 1995. and D. W. 375 1999. "The Evolution of Barker, G., TheBiferno Valley Rupp, Oxford, pp. Figueira,T. J. The and Geo 416. the Messenian in Survey: Archaeological Identity," Sparta: London. L. and New ed. S. Hodkinson morphological Record, Coldstream, J. N., J. Eiring, Perspectives, 2001. Knossos and A. 211 Bauslaugh, R. A. 1990. "Messenian G. Forster. Pottery Powell, London, pp. Dialect and Dedications of the Handbook: Greek and Roman (British 244. 661 at -. 2003. "The of 'Methanioi,'" Hesperia 59, pp. School Athens Studies 7), Demography 668. London. the Spartan Helots," in Luraghi and Bennet, J. 1998. "The PRAP Survey's Comaroff, J. 1985. Body of Power, Spirit Alcock 2003, pp. 193-239. in Resistance: Culture and the Contribution," Davis 1998c, of The History Fotiadis, M. 1995. "Modernity and a Politics of Time pp. 134-138. of South African People, Chicago. Past-Still-Present: = Results in Bennet, J., J. L. Davis, and F. Zarine Corinth Corinth: of Excava the Birth of Regional Archaeo tions Conducted the American baf-Shahr. 2000. "Pylos Regional by logical Projects inGreece," AJA 99, III: School Classical Studies at Athens 59-78. Archaeological Project, Part of pp. = SirWilliam Gell's Itinerary in the VII.3 G. R. Edwards, Corin Frankel, R., and N. Getzov. 1997. in thian Hellenistic Princeton Israel. and Regional Landscapes Pottery, Archaeological Survey of Map the Morea in the Second Ottoman 1975. ofAkhziv (1),Map ofHanita (2), 343-380. XVIII.l = E. G. The Period," Hesperia 69, pp. Pemberton, Jerusalem. 1997. and Kore: The G. 1965. Pausaniass Berlin, A., and K. W. Slane. Sanctuary of Frazer, J. Descrip II. 1: Hellenistic and Greek Princeton 1989. tion 6 New York. TelAnafa The Pottery, of Greece, vols., = Roman Pottery {IRA Suppl. 10, XVIII.2 K.W.Slane,T^ Fulford, M., and R. Tomber, eds. 1994. Demeter and Kore: Excavations at 1948-1951 pt.II.l), Ann Arbor. Sanctuary of The Sabratha, sur c?ra Pt. Bonifay, M. 2004. Etudes la Roman Pottery and Lamps, Princeton II: The Finds, 2: The Finewares romaine tardive 1990. and mique dAfrique Lamps {SLSMonograph 3), {BAR-IS 1301), Oxford. Coulson, W. D. E., and N. Wilkie. London. et les 1983. to Roman Times: 1998. "Eine Breuillot, M. 1985. "L'eau dieux "Archaic Gans, U.-W. sp?tantike The Site and inNichoria aus Nordafrika de Mess?ne," Dialogues d'histoire Environs," Terra-Sigillataschale ancienne 11, pp. 789-804. III, pp. 332-350. inMegalopolis," AA1998, pp. 499 1970. 1998a. "From to Bruneau, P. "Tombes d'Argos," Davis, J. L. Pausanias 506. BCH 94, pp. 437-531. Present," in Davis 1998c, pp. 273 Gerstel, S. E. J. 1998. "Medieval Mes Cartledge, P. [1979] 2001. Sparta and 291. senia," in Davis 1998c, pp. 211 Lakonia: A c. 1300 -. 1998b. of Regional History, "Glimpses Messe 228. et 362 b.c., repr. Berkeley. nia Past," in Davis 1998c, pp. xxix Giovannini, A. 1978. Rome la circu -. 1987. and xliii. lation mon?taire en au IIe si?cle Agesilaos the Crisis of Gr?ce ed. 1998c. avant Sparta, Baltimore. -, Sandy Pylos.-An J?sus-Christ, Basel. -. 2001. "Rebels and sambos in Nestor to Archaeological History from Given, M., A. B. Knapp, N. Meyer, Classical Greece: A Comparative Navarino, Austin. T. E. Gregory, V. Kassianidou, in ed. S. E. L. N. and View," Spartan Refections, Davis, J. L., Alcock, J. Bennet, J. Noller, Wells, Urwin, P. Y. C. Cartledge, London, pp. 127-152. G. Lolos, and W. Shelmer H. Wright. 1999. "The Sydney 1989. Inter Cartledge, P., and A. Spawforth. dine. 1997. "The Pylos Regional Cyprus Survey Project: An Hellenistic and Roman A Tale Part I: of Sparta: Archaeological Project, disciplinary Investigation Long Two of Cities, London. Overview and the Archaeological Term Change in the North Central 2002. "The 391-494. 19 Catling, R. W. V. Survey Survey," Hesperia 66, pp. Troodos, Cyprus,"/^ 26, pp. to D. 1938. 39. Area from the Early Iron Age the Dimakis, "npc?TYj-AiocX.ax?pi," Classical Period (c. 1050-c. 300 in Meooiqviaxo "Ero? 1938, pp. 81 Habicht, C. 1985. Pausanias' Guide to in and in 93. Ancient B.c.)," Continuity Change Greece, Berkeley. 2o6 SUSAN E. ALCOCK ET AL.

M. 1995. "How the -. N.d. Hall, J. Argive Was "Subsistence, Patronage, Heraion?" 577 Argive' AJA 99, pp. and Communal Organisation" 613. (unpublished paper). -. 1997. in Greek Ethnic Identity Hoff, M., and S. Rotroff, eds. 1997. Romanization Antiquity, Cambridge. The ofAthens (Oxbow -. 2003. "The Dorianization of Monograph 94), Oxford. the in Messenians," Luraghi and Hornblower, S. 2000. "Sticks, Stones, Alcock 142-168. and The of 2003, pp. Spartans: Sociology M. and T. H. eds. in Hansen, H., Nielsen, Spartan Violence," War and Vio 2004. An and lence inAncient van Inventory ofArchaic Greece, ed. H. Classical An poleis: Investigation Wees, London, pp. 57-82. Conducted the Polis G. L. 1962. by Copenhagen Huxley, Early Sparta, Centre the Danish National for Cambridge, Mass. Research Foundation, Oxford. Jameson, M. H., C. N. Runnels, and Harrison, A. 1998. "Bouka," in Davis T. H. van Andel. 1994. A Greek 205-209. The Southern 1998c, pp. Countryside: Argolid and N. 1998. to Present Harrison, A., Spencer. from Prehistory the Day, "After the Palace: The Early 'His Stanford. of in oto tory' Messenia," Davis 1998c, Kaltsas, N. 1985. "CH ??yaiy?\ olx?a 147-162. Ko7iav?xL pp. TYJ?Msaorjv.a?," ArchEph HatzfeldJ. [1919] 19774.Les trafiquants 1983, pp. 207-237. italiens dans l'Orient N. hell?nique, repr. Kennell, 2003.uAgreste genus: Helots New York. in in Hellenistic Laconia," Luraghi Hayes,J.W. 1976."Pottery: Stratified and Alcock 2003, pp. 81-105. in Excava at Groups and Typology," Kenrick, P. M. 1985. Excavations Sidi tions at Conducted Carthage, 1975, by Khrebish, Benghazi (Berenice)ULI: the ed. H. University ofMichigan I, J. The Fine Pottery {LibAnt Suppl. 5), Ann 47-123. Humphrey, Arbor, pp. Tripoli. -. -. 1985. "Sigillate orientali," in 1987. "Patterns of Trade at dell'arte Enciclopedia antica, classica, Berenice: The Evidence of the Fine e orientale. Atlante delle cera Rei Cretariae Romanae forme Wares," Cer?mica romana nel miche II: fine FautorumActa 25-26, pp. 137-154. bacino mediterr?neo (tardo ellenismo Kiechle, F. 1959. Messenische Studien: 1-96. zur der eprimo impero), Rome, pp. Untersuchungen Geschichte -. 2000. "The Current State of Messen ischen und der Aus Kriege Roman Ceramic Studies inMedi der wanderung Messender, Kali terranean or m?nz. Survey, Handling Pot from in tery Surveys," Extracting Kingsley, S., and M. Decker, eds. 2001. Mean and in East ingfrom Ploughsoil Assemblages, Economy Exchange the ed. R. H. and Mediterranean Late Francovich, Patterson, during Antiquity, G. Barker, Oxford, pp. 105-109. Oxford. Higginbotham, J. A. 1997. Piscinae: Kolbe, W. 1904. "Die Grenzen Mes in Roman in der Artificial Fishponds Italy, seniens Ersten Kaiserzeit," Chapel Hill. ^M29, pp. 364-378. Hodkinson, S. 1992. "Sharecropping Korres, G. 1981-1982. "'H Tipo?Ar] and ?i? Sparta's Economic Exploitation uoctixy] ty]v ixsTayeveotepav of the Helots," in OtXoX?xcov: Lako "Xpr\oivtc5v MuX?]voc?x?)v x?cpwv nian Studies inHonour Hector of Meacrrjvia?," Antiquity andByzan ed. M. tion. Acts the Second International Catling, J. Sanders, London, of 123-134. pp. Congress ofPeloponnesian Studies, -. 2000. and in 25-31 1980 Property Wealth Patrae, May (Pelo Classical Sparta, London. ponnesiaka 8), Athens, pp. 363 -. 2003. "Spartiates, Helots, and 450. -. a the Direction of the Agrarian Econ 1988. "Evidence for Helle an omy: Towards Understanding of nistic Chthonian Cult in the Pre in at in Helotage Comparative Perspec historic Cemetery Vo?dokilia tive," in Luraghi and Alcock 2003, Pylos (Messenia)," Klio 70, pp. 311 pp. 248-285. 328. historical messenia 207

on in Lawson, J. 1996. "The Roman Pottery," Dow His Eightieth Birthday Settlement Greek Landscape in and in a Greek 189-209. Continuity Change {GRBSMonograph 10), Durham, ArchaeologyVM//14, pp. Rural The Laconia 185-191. L. 1999. "Pausanias et les Landscape: Survey N.C., pp. Piolot, Mys 2: W. and W. D. E. Coul t?res d'Andanie: Histoire d'une ArchaeologicalData {BSA Suppl. McDonald, A, W. son. 1983. "The Dark at Ni in Le Arch?olo 27), ed. Cavanagh, J. Crouwel, Age aporie," P?loponn?se: R. A in et ed. W. V. Catling, and G. Shipley, choria: Perspective," Nichoria gie histoire, J. Renard, Rennes, London, pp. 111-123. III, pp. 316-329. pp. 195-228. and R. F. 1984. La naissance de la Lazenby, J. F., and R. Hope Simpson. McDonald, W. A., Hope Polignac, de. 1961. cit? et 1972. "Greco-Roman Times: Simpson. "Prehistoric grecque: Cultes, espace, soci?t?, in P?lo Literary Tradition and Topographi Habitation Southwestern VIIF-Vir si?cles, Paris. in A. cal Commentary," McDonald ponn?se,"^/^ 65, pp. 221-260. Powell, 1988. Athens and Sparta: and 81-99. -. 1972. Greek Political and So Rapp 1972, pp. "Archaeological Explo Constructing E. 2001. in cial 47 8 Lee, W. "Pylos Regional ration," McDonald and Rapp History from b.c., London. Part 117-147. K. Studies inAn Archaeological Project, IV: 1972, pp. Pritchett, W. 1965. and the Human W. and G. R. cient Greek Change Landscape McDonald, A., Rapp Jr. Topography 1, Berkeley. a terre in Modern Greek Village inMes 1972. The Minnesota Messenia Ex Pucci, G. 1977. "Le sigillate 49-98. a Bronze e in senia," Hesperia 70, pp. pedition: Reconstructing Age italiche, galliche, orientali," 1969. "CH Einstrumentum domesticum di Erco Liritzis, S. ?pxatoc ttoAi? t?j? Regional Environment, Minneapolis. JJX? L. 1985. de lano e nella et? AuTixrj? MeaoYjvioc? "Epocva," Migeotte, "R?parations Pompei prima imp? Tcov monuments ? au 21, pp. 152-180. publics Mess?ne riale (Quaderni di cultura materiale xal 9 Lotze, D. 1959. MeraCo ??sud?pcov temps d'Auguste," BCH109, 1), ed. A. Carandini, Rome, pp. ?o??c?v: Studien zur 597-607. 21. Rechtsstellung pp. in Grie M. M. 1973. Roman N. K. 1999. unfreier Landbev?lkerungen Moevs, T. The Rauh, ", Rome, chenland bis zum 4. Jahrhundert Thin Walled Potteryfrom Cosa and the Eastern Mediterranean v. Chr., Berlin. (1948-1954) {MAAR 32), Rome. Wine Trade, 166-88 b.c.," inHel LRP = W. Late Roman Pot 1981. lenistic Rhodes: and J. Hayes, Morel, J.-P. C?ramique campa Politics, Culture, London 1972. ed. V. P. tery, nienne:Les formes {BEFAR 244), Society, Gabrielsen, Bilde, 1978. Rome. L. Lukermann, F., and J.Moody. T. Engberg-Pedersen, Hanne -. ? vernis "Nichoria and Vicinity: Settlements 1986. "C?ramiques stad, and J. Zahle, Aarhus, pp. 162 and Circulation," inNichoria I, noir d'Italie trouv?es ? D?los," BCH 186. 461-493. Trade in pp. 78-112. 110, pp. Reynolds, P. 1995. the Western N. M. 1991. Pau ad. Luraghi, 2002. "Becoming Messe Musti, D., and Torelli. Mediterranean, 400-700: The m<3,n,"JHS 122, pp. 45-69. sania, Guida d?lia Grecia. Libro IV: Ceramic Evidence, Oxford. -. 2003a. "Helotic Slavery Re La Messenia, Milan. Rickman, G. 1980. The Corn Supply of in = Excavations at in considered," Sparta: Beyond the Nichoria Nichoria , Oxford. Mirage, ed. S. Hodkinson and Southwest Greece, Minneapolis Riley, J. A. 1981a. "Italy and the = A. Powell, London, pp. 229-250. I G. Rapp Jr. and S. E. Asch Eastern Mediterranean in the -. and 2003b. "The Imaginary Con enbrenner, eds., Site, Environs, Hellenistic and Early Roman of the in and 1978. Periods: quest Helots," Luraghi Techniques, The Evidence of Coarse II =W. A. McDonald in and Italian Alcock 2003, pp. 109-141. and Pottery," Archaeology and S. E. eds. N. C. The Bronze and Luraghi, N., Alcock, Wilkie, eds., Age Society: Prehistoric, Roman, 2003. Helots and Masters in 1992. Their Occupation, Medieval Studies {BAR-IS 102), ed. Laconia and Messenia: III = A. W. D. E. Histories, W. McDonald, G. Barker and R. Hodges, Oxford, Structures Dark Ideologies, (Hellenic Coulson, and J. Rosser, eds., pp. 69-78. Studies and -. 4), Washington, D.C. Age Byzantine Occupation, 1983. 1981b. "The Pottery from Cis 1.1994. and in P. 1971. Her terns Malkin, Myth Territory Oliva, Sparta and Social 1977.1,1977.2, and 1977.3," the Cam Amsterdam. in Excavations at Spartan Mediterranean, Problems, Carthage, 1977, L. 1962. "The Conducted the bridge. Pearson, Pseudo-History by University ofMichi and E. C. De Sena. 2003. of Messenia and Its His Ann Martin, A., Authors," gan VI, ed. J. H. Humphrey, toria "Ostia: Overview of the Pottery," 11, pp. 397-426. Arbor, pp. 85-124. Re i Cretariae Romanae Fautorum Petropoulos, M., and A. D. Rizakis. Rizakis, A. D. 1997. "Roman Colonies Acta 38, pp. 43-49. 1994. "Settlement Patterns and in the Province of Achaia: Territo F 1972. "Ceramic in the Coastal Area of in Matson, Studies," Landscape ries, Land, and Population," The inMcDonald and Patras: Roman in the Rapp 1972, Preliminary Report,"//M 7, Early Empire East, pp. 200-224. pp. 183-207. ed. S. E. Alcock, Oxford, pp. 15-36. A. 1984. Min 2001. McDonald, W. "The Pettegrew, D. K. "Chasing the Roebuck, C. A. 1941. A History of nesota Messenia A Look Classical Farmstead: the Messenia 369 to 146 Survey: Assessing from b.c., in Studies Presented to Formation and of Rural Back," Sterling Signature Chicago. 208 SUSAN E. ALCOCK ET AL.

C. A. 1945. "A Note on A. M. 1990. Ar Roebuck, Snodgrass, "Survey Messenian and Economy Popula chaeology and the Rural Landscape tion," CP 40, pp. 149-165. of the Greek City," in The Greek I. B. 1994. "A From to Romano, Hellenistic City: Alexander, from Deposit Corinth: Evidence for ed. O. Murray and S. Price, Oxford, Interim Period Activity (146-44 pp. 113-136. 57-104. b.c.)," Hesperia 63, pp. Spencer, N. 1998. "The History of A. 1977. "Le hil?te?" in Roobaert, danger Archaeological Investigations Ktema 2, pp. 141-155. Messenia," in Davis 1998c, pp. 23 Rosser, J., and W. A. McDonald. 1983. 41. "The Byzantine Occupation: Intro-? Ste. Croix, G. E. M. de. 2002. "The in 353 Helot in duction," Nichoria III, pp. Threat," Sparta, ed. 356. M. Whitby, New York, pp. 190 Rotroff, S. 1997. "From Greek to 195. in in Roman Athenian Ceramics," Stone, D. L., and A. Kampke. 1998. Hoff and Rotroff 1997, pp. 97-116. "A Late Roman Villa on the Mes Rudolph, W. 1978. "Hellenistic Fine senian Coast," in Davis 1998c, Ware Pottery and Lamps from pp. 192-198. at above the House with the Idols Sutton, R. F, J. F. Cherry, J. L. Davis, Mycenae," BSA 73, pp. 213-234. and E. Mantzourani. 1991. "Gazet L. H. 1992. Roman Pot teer of in Sackett, "The Archaeological Sites," in Knossos: From Greek as tery," City Landscape Archaeology Long-Term to Roman Excavations at Keos in Colony. History: Northern the Cy Mansion II cladic Islands Earliest Settlement theUnexplored {BSA from L. H. Lon Suppl. 21), ed. Sackett, until Modern Times (Monumenta don, pp. 147-256. archaeologica 16), ed. J. F. Cherry, Salomies, O., ed. 2001. The Greek East J. L. Davis, and E. Mantzourani, in Roman Context. Los the Proceedings of Angeles, pp. 69-156. a the Fin R. 1989. "The Role of the Colloquium Organised by Talbert, Institute at 21 and in at nish Athens, May Helots the Class Struggle 22,1999, Helsinki. Sparta," Historia 38, pp. 22-40. W. 2003. "Helot Numbers: P. 1993. von Scheidel, Themelis, G. "Damophon A Simplified Model," in Luraghi Messene: Sein Werk im Lichte der and Alcock 2003, pp. 240-247. Neuen Ausgrabungen," AntK 36, C. 1985. the Weak: 24-40. Scott, J. Weapons of pp. Forms Peasant -. at Everyday of Resistance, 1994a. "Artemis Ortheia New Haven. Messene: The Epigraphical and 1997. "The Shipley, G. Other Lake Archaeological Evidence," in An Peri cient Cult Practice daimonians': The Dependent Greek from the oikic Poleis of Laconia Messe ed. R. and Epigraphical Evidence, H?gg, nia," in The Polis as an Urban Centre Stockholm, pp. 101-122. andas a Political -. Community (Acts 1994b. "Damophon of Mes of the Polis sene: New in Copenhagen Centre 4), Evidence," Archaeology ed. M. H. in New Excavations Hansen, Copenhagen, the P?loponn?se: pp. 189-281. and Research, ed. K. A. Sheedy, -. 2004a. "Messenia," in Hansen Oxford, pp. 1-37. and Nielsen 547-568. -. 1998. De 2004, pp. "The Sanctuary of -. 2004b. "Lakedaimon," in Han meter and the Dioscouri atMesse sen and Nielsen 2004, pp. 569-598. ne," in Cult Practice K. 1986. from Slane, W. "Two Deposits from the Archaeological Evidence, ed. R. 157 the Early Roman Cellar Building, H?gg, Stockholm, pp. 271-318. 186. Corinth," Hesperia 55, pp. -. Im -. 2001. "Roman The 1987. "Italian Sigillata Messene: to Rei Cretariae in ported Corinth," Gymnasium," Salomies 2001, Romanae Fautorum Acta 25?26, pp. 119-126. 1972. pp. 189-205. Topping, P. "The Post-Classical -. 1997. "The Fine Wares," in Documents," inMcDonald and

Berlin and Slane 1997, pp. 247-406. Rapp 1972, pp. 64-80. HISTORICAL MESSENIA 200,

a to A.c. A Case A. N. 1966. Tr?ves, P. 1944. "The Problem of 7th the 20th Century Wilson, J. Emigration from in the History of Messenia," JHS 64, Study from Boeotia, Italy Republican Age of Rome, pp. 102-106. (Archaeological Studies Leiden Manchester. B. 425 b.c.: A Valmin, M. N. 1930. ?tudes topogra University 10), Leiden. Wilson, J. 1979. Pylos, sur laMess?nie H. van. 2003. Historical and phiques ancienne, Wees, "Conquerors Topographical Study of and of and the Lund. Serfs: Wars Conquest ThucydidesAccount of Campaign, -. 1938. The Swedish Messenia Forced Labour in ," Warminster. Lund. in andAlcock E. R. 1982. and the Expedition, Luraghi 2003, Wolf, Europe People van S. 1997. 33-80. without der Kamp, J. "Anony pp. History, Berkeley. mous S. 1980. Tomb Cults inWestern Whitby, M. 1994. "Two Shadows: Wright, K. "ATiberian Pottery a in from Messenia: The Search for His Images of Spartans and Helots," Deposit Corinth," Hesperia 49, Pharos Shadow ed. A. Powell 135-177. torical Explanation," 5, The of Sparta, pp. E. B. pp. 63-88. and S. Hodkinson, London, pp. 87 Zangger, E., M. Timpson, S. Yaz H. 126. F. and Knauss. Van Wersch, J. 1972. "The Agricul venko, Kuhnke, J. inMcDonald and E. 1997. "The Archae tural Economy," Will, L. 1997. "Shipping Amphoras Pylos Regional as Rapp 1972, pp. 177-187. Indicators of Economic Romani ological Project, Part II: Landscape Ceram zation in in Hoff and Ro Evolution and Site Vroom, J. 2003. After Antiquity: Athens," Preservation," ics 117-134. 549-641. and Society in the Aegean from the troff 1997, pp. Hesperia 66, pp.

Susan E. Alcock Sebastian Heath

University of Michigan The American Numismatic Society

department of classical studies 96 fulton street

2l6o ANGELL HALL new york, new york ioo38

ANN ARBOR, MICHIGAN 48109-IOO3 [email protected] [email protected] Nigel Spencer Andrea M. Berlin Institute of Archaeology

beaumont street University of Minnesota 36 department of classical and oxford oxi 2pg

near eastern studies united kingdom

305 folwell hall, 9 pleasant ST. SE [email protected] minneapolis, minnesota 55455 -OI25

[email protected] David L. Stone

Florida State University Ann B. Harrison department of classics

florida The Getty Center tallahassee, 32306-1510 12oo getty center drive [email protected] los california angeles, 9oo49-1679