Cultural stratigraphy at Mezhirich, an Upper Palaeolithic site in Ukraine with multiple occupations.

Soffer, Olga ; Adovasio, James M. ; Kornietz, Ninelj L. 等

The later Palaeolithic sites on the East European plain are celebrated for their solid buildings constructed of mammoth bones. Were these permanent settlements, occupied all the year round? Or were they seasonally occupied, in a land where winters are harsh? Stratigraphic explorations at Mezhirich, and excavation of the empty space between the buildings, leads to a decisive interpretation.

The Upper Palaeolithic record of Eastern Europe - replete with spectacular mammoth-bone dwellings and multiple features, including storage pits - is well known to professional and lay audiences [ILLUSTRATION FOR FIGURE 1 OMITTED]. It includes at least 13 sites such as Dobranichevka, Eliseevichi, Gontsy, Mezin, Mezhirich, and Yudinovo, which cluster near the valleys of the Dniepr and Desna rivers and their tributaries, also such outliers as Milovice near the Dyje in Moravia, and Anosovka II and Kostenki II on the Don in Russia (Soffer 1985; Svoboda et al. 1996). Radiocarbon assays from these sites, predominantly on burnt mammoth bone and teeth - a less than optimal medium - date their occupation to a period between some 22,000 and 12,000 b.p. (Svezhentsev 1993; Svoboda et al. 1996). Although they do show some regional differences, their lithic inventories are all assigned to the poorly defined Eastern Gravettian techno-complex.

The sheer size of some of these sites, together with the abundance of their features and the wealth of their inventories has led many scholars to see them as fully settled Palaeolithic villages occupied by sizeable communities on a year-round basis (e.g. Bibikov 1969; Klima 1983; Pidoplichko 1969; Shovkoplyas 1965). Pidoplichko (1969: 148), using meat weight estimates, postulated that Dobranichevka was occupied for 8 years, Gontsy for 9, Mezin for 8, and Mezhirich for 20 years. Other researchers, citing the presence of particular taxa at the sites that were presumably harvested in different seasons, have argued for year-round sedentism as well (e.g. Kornietz et al. 1981). These conclusions use untested assumptions including that:

* all of the dwellings were occupied at the same time

* all the osteological remains resulted from active hunting which gave the hunters access to the animal protein.

They also disregard the fact that the storage of food by freezing in dug-in pits erases the direct tie between the season of procurement and the season of consumption (Soffer 1989).

Finally, seeing these sites as sedentary villages occupied for a decade or two also stands in stark contrast to the thinness of the cultural layers reported from these purportedly single-layer sites where their thickness averages c. 10 cm (for full discussion see Soffer 1985).

In this paper we report the results of our macro- and micro-stratigraphic studies at one of these sites - Mezhirich - which clearly documents repeated seasonal re-occupation rather than year-round sedentism.

The site of Mezhirich

Location

The site, found in the centre of the Mezhirich village (49 [degrees] 38 [minutes] N, 31 [degrees] 24 [minutes] E), is located in the Kanev raion, Cherkassy oblast' in Ukraine, some 160 km (100 miles) directly south of the capital city of Kyiv. A suite of radiocarbon dates on osteological remains indicates that Mezhirich was occupied around 15,000 years ago (Svezhentsev 1993).

Cultural remains here are found in calcareous loess at the depth of 2.7-3.4 m. below the present-day surface, which, in turn, lies at the height of c. 98 m above the level of the Baltic sea (Gladkikh & Kornietz 1978). Geomorphological studies indicate that at the time of occupation the site sat on a promontory of the second terrace which sloped gently towards the confluence of the near-by Ros' and Rossava rivers. Macrostratigraphic work, which shows cultural remains associated with embryonic soil formation, coupled with palaeoenvironmental reconstructions, shows that people occupied the site during one of the milder interstadials, possibly the Trubchevskij, which followed the Last Glacial Maximum (Velichko et al. 1994).

History of research

Mezhirich is quite well known in the literature. Its fame rests on its spectacular features, which include mammoth-bone dwellings made of at least 149 animals, numerous external storage pits, internal and external hearths filled with burnt mammoth bones, external dump areas, as well as extensive faunal and artefactual inventories (Gladkikh et al. 1984; Kornietz et al. 1981; Pidoplichko 1976; Soffer 1985).

The site was discovered in 1965 when a local farmer, Mr Novitsky, expanding a food-storage cellar outside of his house, encountered large bones of what turned out to be a mammoth. Subsequent excavations, from 1966 intermittently to the present, have yielded four round or oval dwellings, ranging in area between c. 12 and 24 sq. m. These dwellings, located 10-24 m from each other, lie in a V-shaped pattern fanning out in the direction of the rivers [ILLUSTRATION FOR FIGURES 2, 3 OMITTED].

As at other like sites, the Mezhirich dwellings contained interior hearths and work-areas, exterior work-areas, and what appear to be garbage-dump areas found close to the dwellings.

Although extensive coring showed the presence of cultural remains over an area some 10, 000 sq. m in size, this area was not investigated continuously or contiguously; rather, a series of level-unconnected excavations were designed to uncover, investigate, and remove a particular mammoth-bone dwelling and associated materials in its immediate vicinity (Pidoplichko 1976; Soffer 1985; Soffer et al. 1994). These excavations of limited size, while producing a plethora of information, focused on studying individual components of the site rather than of the entire settlement; they left the central area between the dwellings totally un-investigated.

Site type

There is much which we already know about Mezhirich. The documented spread of cultural materials over an area of some 10,000 sq. m, if occupied simultaneously, represents one of the largest - if not the largest - Upper Palaeolithic settlement on record. The nature and abundance of the features and the diversity of the inventories suggests a residential site. The relative thinness of what has been reported as a single cultural layer (but see Kornietz et al. 1981) suggests a limited duration of occupation. Inventories for the three fully excavated dwellings indicate their use as habitation structures where such utilitarian activities as tool-manufacturing and repair and food preparation and cooking took place. Contextualizing Mezhirich in a regional settlement matrix argues against its being a seasonal aggregation location. At the same time, the site-wide planning and design evident in the construction of the mammoth-bone dwellings hints at Mezhirich being more than a simple base camp (Soffer 1985). The faunal inventory, taken with data on the making and use of in-ground storage pits and hearths, has been interpreted by the senior author of this paper as indicating occupation during the cold season or seasons (Soffer 1985; Soffer et al. 1994).

Recent research at Mezhirich

Our research at Mezhirich centres around the excavation of c. 200 sq. m that include Dwelling 4, the features that surround it, and the area between this dwelling and Dwellings 1 and 2 uncovered and removed in the 1960s and 1970s. It is a multidisciplinary collaborative project involving scholars from Ukraine, Russia, Great Britain, and the United States. As done by our predecessors, our basic site grid is of 2x2-m units, sub-divided into four metre-square quadrants sequentially lettered A, B, V, G. In addition to the scientific agenda, our work is motivated by the decree of the Presidium of Ukrainian Academy of Sciences to build a research museum at the site, which necessitates investigating the area on which it will be built.

In the 1992-1996 seasons we concentrated on excavating the central part of the site - the area lying between Dwellings 4, 2, and 1, including a microstratigraphic excavation of Pit 4, some 2 m east of Dwelling 4 - and a trench running across Dwelling 4 [ILLUSTRATION FOR FIGURE 4 OMITTED].

Pit 4

Of five pits which surrounded Dwelling 4, three (Pits 1-3) were fully excavated in the 1980s, while we were able to study the partially excavated pits numbered 4 and 5.

These two pits, south and east of the dwelling, are significantly different from each other. Our previous research showed Pit 5 to be a typical garbage-dump, packed with burnt bone and bone ash, fragmented burnt and unburnt bone, lithic debris, small pieces of amber and ochre, and carnivore and leporid paws in anatomical order (Adovasio et al. 1994). This pit contained very few large bones which primarily bordered it. The north - south stratigraphic profile across the centre of this pit, which has not been fully excavated to date, shows a clean sterile sand lens at the depth of 429 to 435 cm below a permanent wall-mark that separates the overlying pit fill from what lies below. The stratigraphic connection between this pit fill and the adjacent remains of the cultural layer indicates that, although most of the fill came from purposeful dumping episodes, some material flowed into the pit after the site was abandoned.

Pit 4 tells a different story. Partially excavated at various times in the 1970s and 1980s, enough remained of Pit 4 to allow reconstruction of its east - west profile and probable dimensions on this axis (c. 2.5 m) as well as its north - south profile and probable depth (c. 1 m). Fortunately, though the southern edge of the pit is incomplete, the eastern one-half to one-third was essentially intact from its surface to very near its base[ILLUSTRATION FOR FIGURE 5, 6 OMITTED]. Due to the well-preserved condition of this part of Pit 4, it was decided to excavate it by micro-stratigraphic levels and to extract a continuous sediment and geochemical sampling column.

The microstratigraphic excavation of Pit 4 produced very few cultural materials and indicates a distinctly different use and life-history for the pit. As FIGURE 6 demonstrates, Pit 4 is excavated from the surface of a thin (c. 1520 cm) depositional unit designated in 1993 - 1994 as Stratum A (Velichko et al. 1994) upon and into which Dwelling 4 is also excavated. Stratum A is a light yellowish brown (2.5Y6/ 4) loess-like silt with a noticeable abundance of very pale brown (10YR8/3) carbonates, many in the form of concretions which strongly resemble root casts. Numerous intercalated discontinuous sand lenses occur in Stratum A, some of which are plano-convex to biconvex in cross-section with diameters ranging from 10 to 20 cm. Geochemical analyses of Stratum A document an increase in humus and CaC[O.sub.3] content; it is reasonable to consider this unit a weakly or poorly developed palaeosol or incipient A horizon. The origin of the discontinuous sand lenses has been postulated as resulting from rain or melting snow accumulating as a series of puddles which washed the clay-sized particles out of the matrix (Velichko et al. 1994).

Pit 4 extends through and truncates a package of microstrata underlying Stratum A (Field Designation B-G, Z) as well as still deeper strata designated H and I. Units B-G and Z, cryogenically deformed, probably represent a period of locally severe climate while H may be another and older palaeosol. The base of Pit 4 penetrates into and partially truncates depositional units previously identified as Strata J, K, and L which appear to be a series of fluvial sands (Velichko et al. 1994: 14).

The base of Pit 4 is lined with a thin (c. 3 cm) layer of sterile sand (Sand 0) which in turn is superimposed by a level of loess fill (Field Designation F4a). This package is in turn overlain by another thin sterile sand lens (Sand 1). Above Sand 1 are three more loess fill units - the uppermost of which is subdivided into two parts. These are designated from lowermost to uppermost as F4b, F4c, F4d and F4d.1. All but the surface of the uppermost loess unit are veneered with sterile sands, labelled in ascending order, Sands 1, 2, 3 and 3a.

The large mammoth bones, with the exception of cranial elements, were found at the very bottom of the pit; they likely represent those meat stores which were not used prior to first abandonment of the pit. The different positions of the cranial elements in this pit - some near the outside edges, others diagonally slanting towards the centre - suggests they originally served as structural elements either on the sides of the pit or positioned on top surrounding the pit. The use of mammoth cranial elements to delimit this pit was previously reported in relation to a mammoth tusk stuck vertically at the western edge of this pit (Soffer et al. 1994). Similar structuring of pits with mammoth crania has been observed by Grekhova (1985) at the site of Eliseevichi as well.

Detailed particle size analysis (PSA) on each of the loess and sand units within Pit 4 are graphically displayed in FIGURES 5 - 8. As FIGURE 6 clearly indicates, sands 2, 3 and 3a exhibit very similar grain-distribution patterns which are interpreted as representing a single source. Sands 0 and 1, quite different from each other and from that 'set', are interpreted as two other sand sources [ILLUSTRATION FOR FIGURES 6 & 9 OMITTED]. The minor silt distributions with the sands are congruent with and verify this tri-fold grouping.

Significantly, the PSA profiles for Sands 0 - 3a are quite different from those cited in Velichko et al. (1994) for the fluvial sand units in the stratigraphic sequence beneath Dwelling 4; further, they lack any evidence of laminations or structures reflective of a fluvial origin. Indeed, rather than a fluvial or any other natural transport mechanism or origin, we interpret the sands within Pit 4 to result from deliberate planned anthropogenic activity reflecting at least three chronologically distinct use-episodes in the life of the pit. Pollen spectra recovered from the analogous sterile sand lens in Pit 5 have been identified as Palaeogene in age while those recovered from culture-bearing Stratum A are assignable to Late Pleistocene taxa. Sands of Palaeogene age are reported in the vicinity of the site. Extending these observations to Pit 4, we suggest that an anthropogenic transport mechanism offers the most parsimonious explanation for their occurrence in this pit.

Pit unit 4d, which 'caps' Pit 4, exhibits a fining upward pattern which suggests that after final abandonment, this feature was gradually infilled by loess; it may have held standing water. A leaching zone just above Sand 3a supports this [ILLUSTRATION FOR FIGURES 6, 10 OMITTED].

The hypothesis, to be tested first through biochemical assay, is that Pit 4 served on at least three separate occasions to store meat on the bone. This use of pits is documented at other similar sites on the Russian or East European Plain where they were excavated into permafrost to the depth of the summer thaw layer, and used to store meat in the naturally created ice-cellars (Pidoplichko 1969; 1976; Shovkoplyas 1965; Soffer 1985).

Soffer (1985; 1989) has argued that two types of pits found at Mezhirich - and yet another type of pit completely filled with large-sized bones, antlers, and tusks found at Anosovka II and Gontsy - resulted from the sequential use of the pits throughout the residential season at the sites: first to store meat on the bone, then to store bones to keep them fresh for tool-manufacturing and burning in hearths, and finally as garbage-dumps. Our excavations of the two of the three types at Mezhirich - one probably used for meat storage and the other for garbage disposal - lends partial support for this hypothesis but argues against it as well.

First, the interstratification of loess with anthropogenic sand lenses in Pit 4 suggests this pit served the same function during different occupations. The presence of a sterile sand lens in Pit 5 argues for a similar conclusion there. The functional transformation of one type of pit into another posited by Soffer (1985) probably did not occur during one long season of occupation; it may have taken place over a longer span of time and through a number of occupation episodes.

Central part of the site

Our excavation of the central part of the site between Dwellings 1, 2, and 4 addressed the relationship between different settlement components and the question of whether the different components at Mezhirich were occupied synchronically. Pidoplichko (1976) concluded that bones of the same individual mammoths in constructing different dwellings made their contemporaneity likely. Softer (1985) augmented this argument by noting the site-wide design evident in the construction of the dwellings. Testing this hypothesis through excavations revealed a more ambiguous scenario.

Specifically, the excavation of the area between three of the dwellings showed an absence of a continuous cultural layer between them and a great sparseness of cultural remains. Away from the dwellings and features directly associated with them, cultural remains - lithic and osteological fragments - occurred as very sparse discontinuous small patches of ephemeral thickness (to 1-1.5 cm). Circular matrix discolouration as well as the presence of almost vertically standing artefacts in these patches suggest disturbance by burrowing animals.

Work near the previously excavated Dwellings 1 and 2 revealed more abundant cultural remains and additional features. Two pits of uncertain origin lie just to the west of Dwelling 1 (numbers 7 & 8 in [ILLUSTRATION FOR FIGURE 4 OMITTED]). These pits, except for three large mammoth bones, had no cultural remains surrounding them. These features were identified in plan through the spatial distribution of concentric 'rings' of sand lenses which probably reflect the different episodes of pit fill. The periphery of pit no. 7 contained a mammoth mandible and a long bone which lay gently sloping towards the pit. A few parallel sand lenses, running north - south between these two possible pits, suggest their association with an already extant depression or gully whose relationship to the pits or depressions remains under study.

Another fragment of a mammoth long bone marked the western edge of the southern pit or depression (no. 8, [ILLUSTRATION FOR FIGURE 4 OMITTED]). The area between this pit and the backfill from the 1966 excavation of Dwelling 1, some 6 m to the NW, contained fairly abundant cultural remains - burned and unburned bones and lithics. These materials are likely related to the hearth Pidoplichko reported excavating to the south of Dwelling 1, and constitute a part of its periphery. Both of these features will be excavated during future field seasons.

A third possible pit was identified just south of the previous excavations around Dwelling 2 (keyed as number 6 in [ILLUSTRATION FOR FIGURE 4 OMITTED]). This feature was also seen in plan view through the presence of concentric sand lenses and three small fragments of mammoth bones on its periphery. As with the two pits near Dwelling 1, the periphery of this pit contained no cultural remains.

The only part of the central area between the dwellings which did contain a continuous, dark-coloured cultural layer was uncovered in squares 318 and 321 where it formed an oval continuous deposit c. 2 m long and up to 2 cm thick. Consisting of fragments of burned and unburned bone, lithics and ochre, it is similar in composition to the toptalishche or dump area we previously excavated south of Dwelling 1 (see Adovasio et al. 1994 with references). We interpret this feature, whose northern and western parts are truncated by previous excavations, as representing the remnants of a continuous midden deposit extending southward from Dwelling 2 to the periphery of Dwelling 4. Previous research has reported parts of this feature in squares around Dwelling 4 where it was reported to be thinner and less intensely coloured (Gladkikh & Kornietz 1978; 1982), and in squares around Dwelling 2, where it appears to have been considerably thicker (Pidoplichko 1976; Gladkikh 1977).

This feature between Dwellings 1 and 2 ties them chronologically and suggests that at least these two may have been occupied simultaneously. Tying patches excavated at different times into a continuous layer/midden deposit and arguing for contemporaneity of occupation is hypothetical; this hypothesis, however, is currently being tested with conjoining studies.

The midden deposit is thicker in the contiguous squares of excavations around Dwelling 2 (Gladkikh & Kornietz 1978) suggesting the bulk of it was associated with Dwelling 2. Its location south of Dwelling 2, when combined with our excavations of a thick midden deposit south of Dwelling 4 (Adovasio et al. 1994) permits us to suggest a non-random placement of midden deposits at the site. The dumping was done mostly south of the dwellings. Because we have yet to understand how these dwellings were entered, it remains unclear if this location simply reflects the positioning of entrances, or also conforms to cultural proscriptions or environmental considerations, such as prevailing wind direction.

Dwelling 4

The 1994 and 1995 field seasons also studied Dwelling 4, unearthed in 1978, made of the bones of at least 29 mammoths and containing an interior hearth. After its opening and prior to our excavation, the dwelling had its more fragile surficial tusks - about 40 - removed. Periodic cleaning and conservation have inadvertently removed some cultural materials from inside it.

The work here cleaned, stabilized and conserved the remaining bones, and investigated its occupation in a manner least destructive to a feature slated for in situ preservation and potential exhibition. A 1-mx5-m trench was partially hand-excavated across the dwelling in a west-to-east direction [ILLUSTRATION FOR FIGURES 11, 12 OMITTED], the orientation so selected because it involved the removal of the fewest remaining mammoth bones. The discussion of this dwelling is necessarily tentative in nature.

The distribution of the cultural materials in the excavated 5 sq. m showed that the dwelling was artificially emplaced into Stratum A rather than built upon its surface. This is congruent with previous speculation based on conductivity studies within the dwelling (Strong et al. 1994). Cultural remains from the outer squares, east and west in the trench, were repeatedly at higher elevations than in the centre. The range of materials recovered - lithic materials at all stages of the reduction and retouch sequence - from nuclei to minute debitage - was distributed in both the eastern and western parts of the trench; both primary flaking and retouch and rejuvenation of tools took place inside the dwelling. These production and maintenance activities were augmented by the use of colourants - red and yellow ochre, and biotite and muscovite mica - as well as by the use of bone and antler tools, e.g. an intact eyed ivory needle and a hammer made of reindeer antler. Unburned, complete skeletons of Arctic fox or weasel-sized animals in the trench suggest that skinning of fur-bearing carnivores also took place inside the dwelling; and burned and calcined osteological remains of herbivores, as well as hares, show the same of food preparation and consumption.

Our trench also contained a portion of a hearth, which was simply cleaned and opened, but not excavated; filled with burned bone charcoal, it was concave in profile - as little as 0.5 cm thick at its northern boundary, but deepening to 3 cm at the southern boundary. The northern, southern and western edges of this burned mass were delimited by large sections of mammoth tusks placed horizontally, which possibly curbed the feature. People occupying this dwelling made fires inside it which they probably used for light and for cooking. Microscopic analyses of flotation samples taken in cleaning the top of this feature show minuscule fragments of both hard and soft woods, possibly used to kindle the bones which served as the primary fuel. This analysis of plant remains, by Mason et al. (1995), also shows the microscopic presence of berries and seed plants - all potential plant nutrients used to augment the diet of animal protein postulated for the inhabitants of the site.

The microstratigraphy of the trench, and horizontal and vertical distributions of the artefacts, shows at least two discrete occupation horizons, each or both of which may have been associated with cleaning or sweeping episodes. Two layers of superimposed horizontally distributed artefacts are clearly present. The majority of the cultural remains, including an ochre-stained grinding stone, antler hammer, mica colourant, and a complete skeleton of a fur-bearing carnivore came from depths between -396.5 and -392; below is a sterile sand layer, then other cultural remains at the depth of-404 cm. Similarly, near the east end, many finds came from depths of -399 to -403 cm, a horizon separated by a clear sterile layer of sand up to 5 cm thick from another cultural horizon; at the depth of -406 cm to -411 cm. It contained more lithics, an ochre-stained grinding stone, and an articulated skeleton of a fur-bearing carnivore. A third, and the clearest indication of a second occupation layer, came from a small 10 cmx30 cm area in the northern part of the trench [ILLUSTRATION FOR FIGURE 13 OMITTED]. In it a hearth at the depth of -405 cm was underlain by a sterile layer and then a horizontally-lying unburnt reindeer antler together with a few lithics at the depth of -411.3 cm. Our on-going study of the vertical distribution of the lithic artifacts across the dwelling centre, controlling for the size and weight of the pieces, as well as con-joining studies, indicate that cultural remains in the dwelling did not experience any post-depositional disturbance due to seasonal freeze - thaw cycles (Suntsov 1996). At least two intact and chronologically separate occupation floors are recognized in this dwelling.

The sand separating the two layers of cultural remains in the trench differ in texture from the sand found elsewhere on the site. Particle size analysis of the Dwelling 4 sand lens indicates that it is the same sand or derives from the same source as Sand 0 as the bottom of Pit 4 [ILLUSTRATION FOR FIGURE 7, 9 OMITTED]. This, in turn, suggests that the excavation of Pit 4 is synchronous with the initiation of the second occupation of Dwelling 4 and, further, that all three pit-refurbishing events are part of the second use of the dwelling.

A smaller number of artefacts were found in both the western and eastern parts of the trench which lay above the two delimited occupation horizons. In sterile loess not coloured by inclusions of burned matter or colourants, these were stratified between very thin sand lenses. Their depths in the western part of the trench ranged from -388 cm to -393 cm and in the eastern part from -392 cm to -396 cm. We think these possibly result from regular sweep or clean episodes associated with one or both of the two main occupation horizons documented for the dwelling.

In sum, the data from the trench inside Dwelling 4 mirror those from Pit 4 associated with it, and suggest that the site was occupied during at least two occasions. We hypothesize that the site's occupants, upon each return to the site, cleaned and stabilized the surface of the features they intended to use by throwing down clean sand obtained from sources near the site, practices amply documented in the ethnographic and archaeological literature.

Interpretations and conclusions

Our excavations at Mezhirich during the 1994 and 1995 field seasons have shown the following:

Dwelling 4 served as a residential feature where people conducted a full range of activities, including processing and cooking of foods, tool manufacture and repair, the sewing and repair of clothing, and the skinning and processing of fur-bearers for hides or pelts. The diverse inventory inside the dwelling, the sparseness of cultural remains outside it, together with tentative evidence for the periodic cleaning and/ or sweeping of the dwelling and the dumping of garbage in outside middens, argue that processing and maintenance activities were carried out indoors. A concentration of diverse human activities in roofed-over space adds further credence to the hypothesis that these were cold-weather occupations in this land of harsh winters.

The presence of at least two occupation horizons inside Dwelling 4, mirrored in the stratification observed in Pits 4 and 5, argue that the site was occupied not by fully sedentary groups but by mobile groups who periodically returned here.

The sterile anthropogenic sand layers in the two pits and in Dwelling 4 suggests that groups returning to the site may have prepared and stabilized their living surfaces by lining them with clean sand obtained near-by.

The relative thinness of cultural horizons inside Dwelling 4 suggests that stays lasted a matter of months rather than years.

The patterned location of garbage disposal areas outside the dwellings, identified south of Dwellings 2 and 4, suggests some permanence in site use - something also reflected in the reuse of pits for the same functions. Periodic returns to the site, we think, occurred during a short time-span congruent with people's memories of it - as would be the case with seasonal returns.

The absence of a continuous cultural layer in the central part of the site between the dwellings can be interpreted in a number of ways. It can be used to argue against the synchronic occupation of the dwellings. But the explanation we favour is that since the site was probably occupied during the bitter cold seasons, a minimum of activities would have been performed outdoors - resulting in a minimum of archaeological remains deposited. One can predict that materials found outside would primarily consist of garbage dumped through periodic cleaning of the dwellings - as amply documented at Mezhirich.

Our data also suggest a probability that at least two of the dwellings at Mezhirich, numbers 2 and 4, were occupied simultaneously. This is the first empirical evidence linking two habitation complexes at the site behaviourally to each other and one being evaluated through on-going conjoining studies.

The distribution of the storage and refuse pits, around each dwelling rather than in a central communal location, tends to argue for autonomous economic behaviour of each household rather than for some central pooling of resources, evidence in good accord with egalitarian economic relationships postulated for coeval sites on the East European Plain (Soffer 1985).

Acknowledgements. The 1990-1995 field work and attendant laboratory analyses of Mezhirich, Ukraine were funded by the National Geographic Society, the International Research and Exchange Board (IREX), the Wenner Gren Foundation for Archaeological Research, Mercyhurst College, the University of Illinois, and the Tsentral'nyj Prirodovedcheskij Muzej, NAN Ukrainy. FIGURES 2, 4 & 11 were drafted by S. Holland and FIGURE 6 is by D. Pedler.

References

ADOVASIO, J. M., O. SOFFER, K.W. CARR, C.L. PEDLER, D.C. KIRKMAAT, D.R. PEDLER, R. DAVIS, N. KORNIETZ & M.R. BUYCE. 1994. Pits and middens at Mezhirich: the third installment. Paper presented at the SAA annual meeting, Anaheim (CA).

BIBIKOV, S.N. 1969. Nekotoriye aspekti paleoekonomicheskogo modelirovaniya paleolita, Sovetskaia Arkheologiia 4: 5-23.

GLADKIKH, M.I. 1977. Otchet o raskopkakh Mezhirichskogo pozdnepaleoliticheskogo poselenia v 1976 godu. Report on file at the Institute of Archaeology, Ukrainian Academy of Sciences, Kyiv.

GLADKIKH, M.I. & N.L. KORNIETZ. 1978. Otchet o raskopkakh Mezhirichskogo pozdnepaleoliticheskogo poselenia v 1978 godu. Report on file at the Institute of Archaeology, Ukrainian Academy of Sciences, Kyiv.

1982. Otchet o raskopkakh Mezhirichskogo pozdnepaleoliticheskogo poselenia v 1979-1981 godakh. Report on file at the Institute of Archaeology, Ukrainian Academy of Sciences, Kyiv.

GLADKIKH, M.I., N.L. KORNIETZ & O. SOFFER. 1984. Mammoth bone dwellings on the Russian Plain, Scientific American (November): 164-75.

GREKHOVA, L.V. 1985. Kostno-zemlianye konstruktsii na pozdnepaleoliticheskoj stoianke Eliseevichi, Trudy Gosudarstvennogo storicheskogo Muzeia 60: 5-33.

KLIMA, B. 1983. Delhi Vestonice. Praha: Academia.

KORNIETZ, N.L., M.I. GLADKIKH, A.A. VELICHKO, G.V. ANTONOVA, Y.N. GRIBCHENKO, E.M. ZELIKSON, E.I. KURENKOVA, T.A. KHALCHEVA & A.L. CHEPALYGA. 1981. Mezhirich, in A.A. Velichko (ed.), Arkheologiia i paleogeografiia pozdnego Paleolita Russkoj ravniny: 106-119. Moscow: Nauka.

MASON, S., J. HATHER & G. HILLMAN. 1996. Preliminary report on the plant remains from Mezhirich 1995 excavations, in Otchet o raskopkakh pozdnepaleoliticheskoi stoianki Mezhirich 1995 godu. Report on file at the Institute of Archaeology, Ukrainian Academy of Sciences, Kyiv.

PIDOPLICHKO, I.G. 1969. Pozdnepaleoliticheskiie zhilishcha iz kostei mamonta. Kyiv: Naukova Dumka.

1976. Mezhirichskie zhilishcha iz kostei mamonta. Kyiv: Naukova Dumka.

SHOVKOPLYAS, I.G. 1965. Mezin. Kyiv: Naukova Dumka.

SOFFER, O. 1985. The Upper Paleolithic of the Central Russian Plain. San Diego (CA): Academic Press.

1989. Storage, sedentism, and the Eurasian Paleolithic record, Antiquity 63: 719-32.

SOFFER, O., J.M. ADOVASIO & N.L. KORNIETZ. 1994. Mezhirich ca. 15,000 B.P. Paper presented at the SAA annual meeting, Anaheim (CA).

STRONG, D.E. & M??J??.M. ADOVASIO. 1994. Conductivity studies at Mezhirich, 1993. Paper presented at the SAA annual meeting, Anaheim (CA).

SUNTSOV, V. Yu. 1996. Evidence of multiperiod occupations from mammoth-bone dwelling at Mezhirich, Ukraine. Paper presented at the 30th International Symposium on Archaeometry, Urbana (IL).

SVEZHENTSEV, Y.S. 1993. Radiocarbon chronology for the Upper Paleolithic Sites on the East European Plain, in O. Soffer & N.D. Praslov (ed.), From Kostenki to Clovis: 2330. New York (NY): Plenum.

SVOBODA, J., V. LOZEK & E. VLCEK. 1996. Hunters between East and West: the Paleolithic of Moravia. New York (NY): Plenum.

VELICHKO, A.A., M.R. BUYCE, J.M. ADOVASIO & G.A. COOKE. 1994. Geology and geomorphology at the Upper Paleolithic site of Mezhirich: macro- and micro-Perspectives. Paper presented at the SAA Annual Meeting, Anaheim (CA).