Despite their discrepancies in area, Pylos, Ordines and Beylerbey are the only PRAP sites where numbers of potsherds are in triple digits during the Mycenaean period; other sites became or remained smaller still, and those immediately to the east of Mt. Aigaleon, which we have followed Chadwick in recognizing as the provincial boundary of the polity, are virtually abandoned in this period (Davis et al. 1997: 423-24). If the identifications of Ordines and Beylerbey with taxation centres are valid, one could speculate that Pylos, as the controlling administrative authority, was able to encourage certain sites to flourish at the expense of others. The incentive would be to create a Network of Preferred Providers, as American insurance companies call it, a network we see in the tablets as the nine 'taxation centres' of the so-called Hither Province. The continuing growth of Ordines might indicate that it was one of the settlements that emerged in the Mycenaean period as beneficiaries of the palatial administrative system. Beylerbey, though larger than Ordines in area, is reduced to a comparable density of ceramics per ha. This too could have been a direct result of the concentration of authority at Pylos, and perhaps even a desire on the part of the centre to limit the strength of its former competitor and current subordinate.
An important implication of this view of sites as 'taxation centres' is that the references on Linear B documents, particularly to the nine places of the Hither Province, are not to districts, but to place-names representative of districts. It is tempting to see these as the chief settlements within the Pylos polity. However, if we look behind the specific documentary mentions, then some inconsistencies emerge. In the first instance, not all such place-names are of equal status. Some, notably in the Further Province list of seven or eight places appear to be explicitly district rather than settlement names (e. g. a-te-re-wi-ja, ‘Atrewia!, 'the land of Atreus': PY Ma 244 [see Ruiperez and Melena 1990: 115; Melena 1994-95: 274]). Equally, some names appear to be interchangeable, as in the case of ro-u-so (PY }n 829.10) which alternates with e-ra-to (PY Cn 608.9; Vn 20.9), or the name *e-ro (PY Jn 829.19) which is apparently 'replaced' by the district names a-te-re-wi-ja (Ma 244) and e-sa-re-wi-ja - also based on a name, either a title or a personal name (Ma 330). More significantly, those texts that deal with 'taxation centres' deal only with a selective range of commodities, no doubt those generally available across the polity and not subject to micro-regional variation, notably in the standard suite of commodities, including a basic cloth type {*146) and animal skins {*152), assessed in the Ma tax documents.
It is interesting to compare the mentions made of some place-names in contexts beyond their listing in the standard range of assessment or contribution texts. The 'residue' of references are not equivalent and might tell us something about different functions or different relationships between them and the palatial centre. For example, pe-to-no only occurs in assessment or contribution texts, while a-ke-re-wa enjoys a range of additional mentions (including personnel, land, and livestock). One can argue that these references are consistent with the identifications suggested above relating to the sites of Ordines and Beylerbey: Ordines (if it is Linear B pe-to-no) appears to acquire significance only within the LHlIl period, whereas Beylerbey (if it is Linear B a-ke-re-wa) appears to have a more complex history and therefore relationship with the palatial centre. A similarly complicated situation appears to arise in the case of the place-name pa-ki-ja-ne, clearly an important cult centre, probably close to the palace, perhaps in the vicinity of Hora Volimidia (e. g. Chadwick 1972:109), but also a 'taxation centre' in the assessment and contribution texts. It is also not clear that 'taxation centres' are always significant settlements. For example, in the case of one such place-name, ro-n-so, we have the document that gives the total assessment for its district (PY Ma 365) and another giving the breakdown of contributions of the type of cloth denoted by the logogram *146 for it and the eight other place-names in its district (PY Mn 456; Killen 1996). ro-u-so contributes a smaller amount of *146 than others within its area, something that goes against our expectations, if such 'taxation centres' are major places within their districts. We should note, however, that ro-u-so clearly has an importance in the palace-organized manufacture of textiles from its mention in the Pylos Aa / Ad records, which suggest a female workforce of 86, second only to Pylos itself in the Hither Province (Chadwick 1988: 85).
If we are reading the variation in archaeological prominence of sites correctly, then the shifting pattern of relations with the centre that it implies and the possibility suggested by the Linear B data that sites of apparently lesser status become the focal points for tax collection, then a possible implication is that relations between the centre at Englianos and other sites were mediated through individuals. These would either be anonymous officials (such as the ko-re-te and po-ro-ko-re-te explicitly identified as responsible for acquiring the bronze mobilized in PY Jn 829) or named members of the elite. In this interpretation, therefore, Pylian expansion would be less about bounded territorial gain than about the negotiation of relations with key individuals at sites that came to have a strategic importance in the centre's strategy of mobilizing commodities to support its political economy.
In the broader context of urbanism in the Aegean Bronze Age, we hope to have suggested ways in which detailed archaeological data can be brought into dialogue with textual data to outline the overall development of power relations among settlements through time in the Pylos region. Just as it seems that not all sites within tht; LHIIIB Pylian polity had identical histories, we would like to suggest that comparative research into the particular histories of other regions dominated by Mycenaean centres might well reveal differences that have implications for our conception of the Mycenaean world as a unity.
World History
