The purpose of this long, interim report is fourfold:

1. to provide a straightforward preliminary account of results of excavations in the Aşvan region in the years 1968–72;

2. to attempt an explanation of how and why the modern studies were carried out;

3. to offer a tentative demonstration of the relevance of the modern studies to archaeology (i.e. the excavation of objects and structures);

4. to outline, by using concrete examples, the general theory and methods, and thereby to demonstrate the feasibility of a co-ordinated approach.

The Keban Project was founded in 1968 under the aegis of the Middle East Technical University and with the co-operation of the Turkish Government. The stated aim of the Project was the rescue of historical materials from the area to be flooded by the Keban Dam, located just below the confluence of the Murat and the upper Euphrates (Fig. 1). The necessity of immediate action had been demonstrated largely through the efforts of the Faculty of Architecture and Restoration. In the area of the flood zone staff and students had carried out feasibility and recording surveys which then made clear both the nature and requirements of a rescue operation. Their results were published in Doomed by the Dam (METU, Ankara 1967). In 1968 a Committee (METU Keban Dam Area Salvage Project Executive Committee) was set up, under the chairmanship of the then Rector of METU, Dr. Kemal Kurdaş, to co-ordinate the efforts of those various teams, both Turkish and foreign, which had accepted the Committee's appeal for joint action in a rescue operation.

This site has not been tested by excavation. Such finds as are known were collected from the surface of the mound in the autumn of 1971. No attempt was made to collect systematically: the intention was simply to provide sherd material in order to date the site. The site was located during pre-excavation surveys by Whallon and Kantmann (map, METU 1968 Report: site N 52/3).

The mound of Kurupınar is situated on the west side of the upland valley, c. 2 km. west of the şose to Çemişgezek, along a track leading to Engüzek; a poor spring is found near the north foot of the site. Engüzek lies c. 2·5 km. away to the southwest.

Short preliminary notices of the work at Çayboyu have appeared in recent issues of this journal and the METU series of preliminary reports (Diamant 1971, 1972; Diamant & Aksoy 1972, 1973). No further excavation has taken place since the initial two short seasons (1970–1). What follows is thus only a slightly more detailed summary of the material recovered up to this time.

The mound of Çayboyu lies about one kilometre east of Aşvan village, on the edge of a small wadi. It is barely visible on the northwest side as one approaches from the village, since the change in elevation is very gradual. However, on the southeastern side the wadi and subsidiary irrigation channels have eroded away what may be as much as half of the original mound, leaving an exposed cross-section of it about three metres high. One of the modern irrigation channels runs periodically just at the base of this scarp, from which pottery and obsidian artifacts have been collected.

The work at Taşkun Mevkii consists to date of two seasons, 1970 and 1971. A third and final season is projected for the summer of 1973. Excavation in 1970 was restricted to a series of slots and one 4 × 4 m. square at the northern edge of the site in order to test the depth of occupation and to establish preliminary phasing (Fig. 1). On this basis the area was expanded in 1971 to expose structures, particularly of Phase 3 which had proved to be most rewarding in 1970. It was found that the first season's phasing was substantially accurate, requiring only minor adjustments of Phase 1.

Phase 1 (Fig. 2, plan; Fig. 8, section)

(a) In 1970 painted pottery appeared in topsoil, thus linking the top of Taşkun Mevkii with Aşvan Kale and other sites in the area. The painted designs found at Taşkun Mevkii are represented on whole pots from Aşvan Kale (see French and Helms, below, p. 157, Fig. 5). It is now clear that this final occupation exists only in pits cut from topsoil or above and hence no contemporary occupation surfaces could be found in the area excavated.

The purpose of this interim report is to give a description of the major architectural features of the site, the evidence for their dating, and a brief account of the artifactual remains from them. This work is necessarily incomplete: only the major excavated levels are distinguished and illustrated; a detailed division of the Medieval levels into subsidiary phases has not been attempted here; and before a complete and definitive account of the artifactual remains, especially the pottery, can be given, further work must be done on the material. This may throw new light on the problems of dating.

Fig. 1 shows the area excavated, with the exception of the step trenches dug down the north side of the mound, which are of no direct relevance to the architecture of the later periods considered here, and the trenches on the west side of the mound, dug ad hoc to investigate the later Medieval structures. These are shown in Figs. 11 and 12. The continuous thin lines in Fig. 1 indicate the internal sections which were drawn from the topsoil to the point where excavation was halted but which were later removed.

The pottery illustrated here is only a small proportion of the total recovered; the intention is to present a brief survey of the commoner fabrics and surface treatments.

The stratigraphic connections from trench to trench were, as discovered, not entirely satisfactory. Though the lowest two of the four trenches which were dug (1969) into settlement debris on the north scarp, viz. G1d and G2b, produced no substantial structural features, the sequence of surfaces, tip lines and aggradations was reasonably clear (Fig. 1). In the trenches above G2b, i.e. G2d and G3b, stratigraphic observations were limited. No lines of aggradation could be observed, the trenches apparently being taken up by the remains or debris of a thick mud brick wall or walls. Into this mass of mud-brick were dug pits. Above both features (mud-brick and pits) began a series of structures and associated deposits, datable (after the 1972 season) to the middle or second half of the first millennium B.C.

The pottery presented here comes from the lowest two trenches, G2b and G1d.

The medieval fortress and church of Taşkun Kale were partially excavated in 1970 and 1971. No excavations took place in 1972, as the Aşvan Project's endeavours were concentrated elsewhere.

Earlier reports in this journal (AS XXI (1971), 6–8, and XXII (1972), 14–15) concentrated on the architectural finds. This article consists of a more detailed account of the site and the remains, followed by some thoughts on Taşkun Kale's historical significance. It is an interim report in traditional style. A full and systematic use of the data gathered by Wagstaff, Payne, Hillman, etc., must await the completion of their studies.

The Site (Figs. 1 and 2)

Taşkun Kale lies about 4 km. SSE of the village of Aşvan on the edge of a rolling plain (“undulating upland basin” – Wagstaff p. 210). Eastwards it over looks a small valley at a point where the valley sides fall steeply to rushy flats, through which a perennial stream, the Kuru Çay, flows north towards the Murat.

The terrain is uneven. North and south of the site tributary wadis break the valley sides. The highest point is the kale proper, a flat topped höyük formed of occupational débris, below which there may well be a natural eminence which attracted the first settlers.

During the second season of excavations on Aşvan Kale (June – September 1969) an area of over two hundred square metres was cleared to two related late Hellenistic phases (“Hellenistic lla and b” –Mitchell, above, p. 123). On 28th September, during the removal of a pisé wall of the earlier phase, a crudely made pot with two pierced lug handles was found set in the pisé. In it was a hoard of forty-eight silver coins (see Figs. 2–3).

The coins are catalogued and registered as Aşv./69/326 nos.l –48. They are deposited in Elâziğ, and will be displayed in the new museum which is being built to house the finds from the various digs in the Keban region. All forty-eight coins are drachms; forty-seven are of Ariobarzanes 1 Philoromaios (96–63 B.C.) and one is of his rival Ariarathes IX Eusebes (101–87 B.C.). The coins had a green incrustation probably similar to that of the Cappadocian hoard discussed by Mørkholm, who thought it was caused by burial in a copper or bronze container. This is not so for the Aşvan hoard, where the incrustation must be caused by the corrosion of the copper content in the coins. It appears that the late silver coinage of Ariobarzanes I was far from pure.

This is a preliminary report. It aims simply to list the coins found in the excavations at Aşvan 1968–72 and Taşkun Kale 1970–1. Full details will be published after the excavations are completed.

The Aşvan Project grew out of the second aspect of the Aşvan excavations, viz. the recovery of archaeologically relevant information from the modern situation, and began (in 1969) with the botanical work of Gordon Hillman.

Very quickly, particularly in 1970, with the help of those who undertook to carry out area-studies, the outline of a scheme for collecting data was sketched and, within a general framework, a number of problems were suggested; of all the problems, those of subsistence and subsistence-patterns were selected as offering the widest base for developing a co-ordinated programme of investigation. The choice was dictated, firstly, by the nature of excavated materials (structures, tools, refuse of stored food remains, both plant and animal) and, secondly, by the availability of a defined geographical area (the concession area i.e. the Aşvan bölgesi) in which was found a number of ancient settlement sites. All these sites could be excavated, together or separately, not only as part of a larger programme, but also by the same methods, standards and techniques and to the same ends; together, the sites covered a time-span of c. 7000–8000 years. In other words the Aşvan region was seen as a unit with clear, recognizable potential for area-studies while at the same time a good, ethical reason for undertaking a regional survey was apparent — any information rescued would be valuable.

Physical conditions have an important and many-faceted influence upon the human occupation and use of an area. Their influence is not always direct, but depends upon the perception of the physical environment by a resident or immigrant group. Only a part of the perceived environment affects decisionmaking and influences behavioural patterns as expressed, for example, in the siting of settlements and the use of land for subsistence. Behaviour, in turn, not only affects the physical environment – by the clearance of forest, for example – but also influences the way in which the physical environment is perceived. Continuous reappraisal is thus integral to successful human occupation and use of an area. Accordingly, established patterns may change, perhaps almost imperceptibly but sometimes radically. The mere existence of an established pattern, however, shapes subsequent reappraisals and, to some extent, controls necessary modifications and changes. A model of geographical change through time is thus more spiral than linear.

Following Wagstaff's paper on physical geography and settlements, a brief expansion of the discussion of distribution of agricultural resources is in place. This topic provides the necessary background for studies conducted in the village itself as well as for the following paper on “Agricultural productivity and past population potential at Aşvan”.

My object here is firstly, to outline the spatial distribution of major land resources (this is achieved via Figs. 1 and 2), secondly to discuss a few of the factors which appear to have conditioned this distribution and, thirdly, to outline some problems involved in the location of ancient settlements relative to these land resources. Few conclusions can, however, be drawn until we have more data on sedimentation history (and thence ancient distribution of soils) in the region.

This paper is not a discussion of archaeological evidence for past population levels in our area, nor of likely sequences of change in population through time. Instead, it is a set of estimations of upper levels of population that would have been possible at recent and ancient agricultural settlements in the Aşvan region, assuming certain combinations of territory, land management systems and levels of productivity. This contrasts, therefore, with the approach more frequently adopted in estimating past population levels which uses settlement size and structural remains.

The points in time at which these hypothetical situations (with their corresponding population levels) were likely actually to have existed in the Aşvan region are not considered in detail here. We may be in a position to fix time correlates in some cases once we have more data on local erosion and sedimentation history and once we have assembled all data arising from the site materials, in particular the data from the very large quantities of plant and animal remains.

The full analysis of the initial and most exhaustive of our studies of present-day agriculture at Aşvan could not be completed in time for the present volume. It seemed relevant, however, to present a brief outline of some of the principles involved and the data collected.

Samples of vegetable remains represent a body of information which is concerned (in part, at least) with human manipulations of plant resources. These plants may have functioned as foods, fuels, building-timber, tools, dyes, drugs, cosmetics or as decorations. The compounded information may, further, suggest certain forms of ancient economy, though quantification of the data at a level representative of the settlement as a whole is rarely possible. Any such inferences can, however, only exist within the realms of one's own familiarity with equivalent modern situations (or, less consistently, historical situations). If our deductions here are to be repeatable, our present-day (or historical) models must be defined: it is not sufficient to have hazy analogues lurking in the data banks of our subconscious.