A Preliminary Report on Excavations and Related Studies, 1992-1994
Part 1

Karl M. Petruso1
Muzafer Korkuti2
Lorenc Bejko2
Sytze Bottema3
Brooks B. Ellwood4
Julie M. Hansen5
Francis B. Harrold6
Nerissa Russell7

1 Program in Anthropology, University of Texas at Arlington, Arlington, Texas, USA; 2 Instituti Arkeologjik, Akademia e Shkencave, Tiranë, Albania; 3 Biologisch-Archaeologisch Instituut, Rijksuniversiteit, Gröningen, The Netherlands; 4 Department of Geology, Louisiana State University, Baton Rouge, Louisiana, USA; 5 Department of Archaeology, Boston University, Boston, Massachusetts, USA; 6 Office of the Dean of Liberal Arts and Sciences, University of Nebraska at Kearney, Kearney, Nebraska, USA; 7 Department of Anthropology, Cornell University, Ithaca, New York, USA

       [The following is a version of a report that appeared in Iliria, the journal of the Archaeological Institute of the Academy of Sciences of Albania, vol. 26, nos. 1-2 (1996), pp. 183-224]

The Site and its Excavation
Konispol I (Upper Paleolithic) and II (Mesolithic)
Konispol III (Neolithic)
Konispol IV (Eneolithic)
Konispol V (Bronze Age)
Konispol VI (Iron Age)


       The project described in this report was conceived as a binational collaborative endeavor during a visit by the American co-director (Petruso) to Albania in August 1991. During that visit, in the company of the Albanian co-director (Korkuti) several archaeological sites were visited and various topics and problems in the prehistory of Albania were discussed. It was decided that the district of Sarandë in southernmost Albania--which had demonstrated clear evidence of early occupation of the country, but which presented a number of problems that await resolution--would be the most fertile area for research. The co-directors settled on a cave north of the town of Konispol as the focus of the project (see maps, Figs. 1 and 2). Financial support was obtained for a brief pilot season of excavation in summer 1992; the results of the excavation and of related explorations in the southern part of the district were promising indeed (Petruso et al. 1992), and led to the planning of the first major field campaign in summer 1993. This report describes in summary form the major and typical finds of the first two seasons of excavation and related studies at the Konispol Cave.
       Of all the periods of Albanian antiquity, the Stone Age has been the least systematically explored and is the least well understood. Frano Prendi, in his fifty-page Cambridge Ancient History survey in English of Albanian prehistory (1982)--which is arguably the most widely-read summary in English of the topic--was unable to devote more than one page of text to the Paleolithic and Mesolithic periods combined; indeed, in the ten years since that chapter was written, virtually no new pre-Neolithic sites have come to light, and little analytical or synthetic work which might have illuminated the picture has been carried out. Three Paleolithic sites were mentioned by Prendi in the aforementioned survey. One of these, on Mount Dajti near Tiranë--discovered before World War II by German archaeologists--has never been relocated, and finds recovered from the site have long since disappeared. The other two sites--the cave of Shën Marinë and an open site near the village of Xarë--are both located in Sarandë District within easy walking distance of the Konispol Cave.
       The Mesolithic period is for all intents and purposes a total blank. Only one site (Vidhëz in south-central Albania), long known but only recently excavated, has been identified tentatively as dating to the Mesolithic (Luftin 1990; Korkuti and Petruso 1994:707). The excavator had suggested that trapezoidal flint tools of microlithic size, in a stratum which contains no pottery whatever, might be indicative of this transitional--and thus far elusive--stage.
       Because of the very sparse nature of the remains of the earliest period, the decision was made in 1991 to explore the known but controversial sites in Sarandë District, focusing on excavation in the Konispol Cave, in the hopes that we might once and for all resolve the ambiguities in the scholarly literature regarding the earliest occupation of this region.
       The potential of the Konispol Cave to provide significant evidence for our understanding of the Stone Age environment and economy in southern Albania was clear; it was agreed that the project would turn its attention to questions which had not been asked previously by Albanian archaeologists. Several unanticipated--and, as it turned out, irresolvable--logistical problems related primarily to the unavailability of supplies and equipment in the country in the period immediately following the change in government of 1990-1991 severely hampered our recovery design, but permitted us to plan concretely for the 1993 season. In 1993 we instituted a recovery regime that included screening (half-inch mesh) of all earth removed from the trenches. In addition, under the direction of Dr. Julie Hansen, a regime of water sieving was established.


       The Konispol Cave (known locally as Shpella e Kërçmoit) is a large solution cavity in a maquis-covered karstic limestone landscape. It lies at an altitude of 400+ m.a.s.l., near the top of the Saraqint ridge, approximately 20o10' E, 39o40' N (see map). Its main entrance is shielded from the cold north winds; it faces the southwest, from which there is a commanding view of the plain below and the strait and island of Corfu. The cave is approximately 50 m long, and at most 6 m deep and 6 m in interior height (plan, Fig. 4 and elevation, Fig. 5). "Windows" in the ceiling above the main entrance and at its western end provide some light and ventilation. Our main efforts in 1992 and 1993 focused on determining the depth of the geological and cultural deposits, extracting samples for radiocarbon dating, and collecting data relevant to reconstructing the economy of the cave during its prehistoric occupation. Preliminary excavations by the Albanian co-director during 1989 and 1990 suggested that the cave had a significant sequence of Neolithic occupation, with sporadic evidence of use in the Bronze Age, Early Iron Age, and classical period (Korkuti and Shabani 1989).
       During the campaigns of 1992-94 Trenches VIII, IX, XII, and XXI were excavated, for a total surface area of 19.36 m2. We reached a maximum depth of 4.20 m below the current floor of the cave (in Trench VIII). The stratigraphic levels represented in the cave have been designated provisionally as follows :

Konispol I      Upper Paleolithic
Konispol II      Mesolithic
Konispol III      Neolithic
Konispol IV      Eneolithic
Konispol V      Bronze Age
Konispol VI      Iron Age
Konispol VII      Archaic through Hellenistic   

       We have followed the tradition in Albanian archaeology of digging in 10-cm. passes, respecting the lithostratigraphy and anthropogenic features insofar as possible. In so doing we follow the system of "unit" nomenclature described by Jacobsen and Farrand (1987:7). In the descriptions that follow, the final period of use of the site (Konispol VII) will not be dealt with; our focus will be on the most intensive periods of use, namely the Stone and Bronze Ages.

KONISPOL I (Upper Paleolithic) and II (Mesolithic)


       The earliest evidence of human use of the Konispol Cave is represented by the lithic remains from Trenches VIII and IX/X. We have examined and entered into a computer database more than 3,000 fragments of chipped stone excavated during the 1992 and 1993 field seasons. Already several patterns and trends have appeared.
       At the broadest level, the lithic remains from the site can be divided into two groups, which we have labeled the Upper Ensemble and the Lower Ensemble. The Upper Ensemble is characterized by types of flint of various color and quality (mostly reds and browns) which are available locally, to judge from our unsystematic walkover of the area between Konispol and the Pavel River. In this component, retouched blades, truncated blades, and bladelets predominate; backed blades, endscrapers, and burins are less common. Retouch is often very fine, delicate, and subtle, and frequently discontinuous rather than continuous along a surface. This component covers the earlier Neolithic period, but begins somewhat earlier: in general, the technology and typology of the earliest pieces in this component are of the sort known from the later Upper Paleolithic and Mesolithic of the southern Balkans. Units 19 through 22 in Trench X produced geometric microliths (a trapeze and a triangle) as well as several truncated bladelets and small blades (Fig. 6). Another trapeze (from Unit 39 of Trench XXI) may well be contemporary with the aforementioned units of Trench X. In Greece (especially at the Franchthi Cave) and elsewhere such artifacts are indicative of the late Epipaleolithic (after ca. 13,000 BP) and the Mesolithic, although they continue into the Neolithic (Perlès 1990).
       Given the stratigraphic position of these artifacts, just below the ceramic-rich Early Neolithic units, and somewhat above the level in Trench VIII radiocarbon-dated to around 11,400 BP (q.v. below), a provisional Mesolithic designation for Units 19 and 20 seems justified. Parallels for the geometric microlithis and modified bladelets of the Konispol Upper Ensemble can be found in the rather meager Mesolithic archaeological record of the Balkans at Franchthi Cave, especially in Phase VIII (Perlès 1990) and at Odmut Cave in northwestern Montenegro (Srejovic 1978; Markovic 1978). The Konispol Cave microliths bear comparison with the only other microliths attested thus far from Albania, which were excavated in 1990 at an open site near the town of Vlushë in the south-central part of the country (q.v. above). More informative comparisons of the Konispol Cave Mesolithic, however, will need to await recovery of larger samples from elsewhere in the cave.
       The Lower Ensemble has a completely different character from that of the Upper. The lithics are dominated by a blocky white flint, generally badly weathered and desilicified, with poor flaking characteristics. The material, similar to flint nodules which crop out within the south wall of the cave and on the hillside below the cave terrace, shatters into splinters when struck. Indeed, it appears--to judge from the splinters recovered from sieving and water-sieving (more than 500 in one unit of Trench X alone)--that knapping of this material was being carried out within the cave. While the sheer number of fragments of white flint is impressive, recognizable tools are few, and of little analytical value. Only four retouched tools were recorded from this entire ensemble, but a number of unretouched blades, bladelets, flakes and cores were recovered. These pieces are sufficient to indicate that their makers could and did use prismatic core techniques to produce blades and bladelets when the quality of the raw material allowed. Thus the operative limitation was not a lack of certain flintknapping practices, but a lack of interest in acquiring suitable raw material.
       No specific cultural attribution is yet possible for this ensemble; assuming the accuracy of the 1992 radiocarbon date from Trench VIII, Unit 28, these levels probably fall within the later Upper Paleolithic era. Thus far, no fruitful comparison of assemblages from the final Paleolithic of the southern Balkans--e.g., Franchthi (Perlès 1987) or, closer to Konispol, in Greek Epirus, Asprochaliko, Kastritsa, and Klithi (Adam 1989; Bailey and Gamble 1990; Bailey, Gamble and Higgs 1983; Bailey et al. 1986)--is possible. The putative isolation of a Mesolithic and perhaps Upper Paleolithic at the Konispol Cave is of some importance for Epirus.

KONISPOL III (Neolithic)


       Ceramics are the most diagnostic artifact of the Albanian Neolithic. In general, the Konispol Cave has produced relatively sparse ceramic remains--by no means what one would expect from a living site--with respect both to the total amount of ceramics and the repertoire of shapes and functions. A brief descriptive summary of the prehistoric pottery by period follows. Complementary studies, including petrographic analysis, are under way at this writing; results will be described in future reports.

Early Neolithic (Fig. 7): Of the recognized subphases of the Neolithic in Albania, the Early Neolithic is the one most fully represented; the EN pot-bearing strata average nearly a meter in thickness. Nonetheless, the total amount of cultural remains from this period is low, evidenced mainly by animal bones and lithics. The pottery is thick in section. The fabric is characterized by generous sand inclusions. Some vessels are well-polished and have a slight luster. The tendency is toward a monochrome color, but this is not uniform over the individual vessel surfaces. Mottling in the course of firing is commonly attested. The most typical colors are red and reddish-brown. Also represented are vessels of grayish-black color. In general, the vessels seem to be competently fired.
       The vessels are of simple forms, mainly cups and wide-mouthed vessels. Carinations are few. The most typical are vessels with nipple-like perforated handles with flat bottoms.
       The ornamentation is represented by fragments with impressions made in two techniques: (1) impressions made by fingernails, in the style called Smilçiç and (2) impressions made by an instrument, in the manner described in the literature as Devollite. Both of these impresso techniques are attested at Early Neolithic sites elsewhere in northern Epirus. The lack of painted pottery is noteworthy; painted pottery is one of the main characteristics of the Early Neolithic of southeastern Albania (Podgorie I Culture), for Macedonia and Thessaly. The lack of painted pottery in the EN of southern Albania coincides with the situation in NW Greece, where painted pottery is also missing.
       This correlation of features--monochrome surfaces, simple forms, nipple-like perforated handles, and impressed Devollite decoration--suggests a cultural and chronological correlation between EN Konispol and Podgorie I, Velushka I/II and Borodin in Pelagonia, Presesklo in Thessaly, Starçevo in Serbia, and Smilçiç in Dalmatia.
       Pottery seems to appear some time after the occurrence of the first domesticated animals in the cave; whether we have here evidence of an aceramic Neolithic (see below under Zoöarchaeology) is not yet clear, however.

Middle Neolithic (Fig. 8): The Middle Neolithic period is represented by a much thinner stratum (averaging less than 0.50 m). The wares of this period are mostly thick-walled, as was the case in the EN, and contain sand as temper. The most common color of pottery in this period is gray or mottled gray-black. Thin-walled vessels appear at this time; these are of fine fabric and generally are characterized by smooth surfaces. The most typical profile in the thin-walled ware is S-shaped; typically these vessels have everted rims. There are in addition simple conical bowls and large vessels with cylindrical necks. The bottoms of the MN vessles are flat, slightly concave, and very rarely with high cylindrical feet. A thin-walled, lustrous, fine black burnished ware appears in this phase.
       Ornamentation in this period remains rare, and is mainly impressed, with relief beads and gray painting. The most common are vessels bearing various kinds of impressions, including Devollite, and Smilçiç. These impresso techniques continue a decorative tradition from the EN. Though few in number, of great interest are the fragments of Smilçiç-type impression which suggest contact with the Dalmatian coast cultures. Finally, a matt gray painted ware is represented for the first time in the Middle Neolithic.
       The ornaments with beads and lines in relief are made with the same known technique in the other cultures of MN Albania, e.g., Dunavec I, Cakran, and Kolsh II. The gray painted technique is rarely present in those cultures.
       The thin-walled pottery of black color, the S-shaped vessels, the high conical pots, the ornaments with beads in relief and the gray painted ware, as well as the existence of rhytons, suggest that the MN of Konispol is to be equated with Dunavec-Cakran and Blaz III. In contrast to the EN, when the culture of Konispol was distinct from EN cultures in SE and N Albania, during the MN there is evidently something of a unification over the entire territory of Albania, suggesting a cultural phenomenon of wide geographical spread. This is evident in the similarities and parallels with Dunavec and Cakran and with Velushka-Tumba III-IV in Macedonia, with Dimini I and II in Thessaly, with Proto-Kakanj in Bosnia, and Danilo I in Dalmatia.

Late Neolithic (Fig. 9): The Late Neolithic strata average some 30 cm in thickness. The most typical wares are a light-colored painted pottery in the style known as Maliq I and another ware represented by the pottery with thick walls, continuing from MN, which preserves the MN characteristics. The main form of the vessels are hemispherical cups with straight rims, cups with everted rims, and vessels with bulbous bodies and cylindrical necks.
       The painted pottery is of two categories: local and imported. The local pottery is again characterized by sand inclusions, is well-fired, and has a polished surface. The fabric of this pottery is usually red, often bearing broad painted lines creating triangle patterns; often these are filled with parallel hatched lines. By all indices, this pottery is clearly differentiated from the other painted pottery.
       The imported painted pottery is thin-walled, made from well-cleaned clays mixed with fine sand and thoroughly fired. The external surfaces of the pots are polished but non-lustrous. They are often buff or cream colored, and more rarely gray. The painting is of various shades of brown, always matt, and painted directly on the ground of the vessel. The main motifs consist of groups of parallel lines and barbed lines. The motifs are well-executed.
       The native painted pottery is similar to that of LN Katundas Cave, while the other painted pottery is similar to that of Maliq I, which is likely its source. These attributes relate the LN pottery of Konispol Cave technologically and chronologically with Maliq I in Albania and Classical Dimini (IV) in Thessaly.

KONISPOL IV (Eneolithic)

       The Eneolithic period is represented by a thin layer (ca. .20 m). The most typical ware is characterized by thick walls. It has sand inclusions and is in general carelessly fired. It is monochrome, but has no single typical color. The other ware, characterized by thin walls, is made of clean clay mixed with fine sand, with a visible polish. There are a few fragments of black pottery with luster.
       The profiles of these vessels include distinctive cups with straight rims; pots of S-profile, and bowls (Fig. 10). The handles are vertical and in the form of a chin.
       Ornamentation is scarce. A few fragments are attested with plastic ornaments in forms of ribs in relief, punctated bands, and, very rarely, small conical nipples, as well as incised decorations with gray painted patterns (the latter are also quite rare). The shapes of the Eneolithic pottery of Konispol Cave relate it to Eneolithic cultures attested elsewhere in Albania, represented especially in Maliq Phase II. There are in addition parallels with the Rakhmani phase in Thessaly, Armenohori in Pelagonia, and Bubanj-Hum in Serbia.

KONISPOL V (Bronze Age)

       The uppermost prehistoric strata in the Konispol Cave are thin and often disturbed. The Bronze Age pottery is represented mainly by rather thick-walled vessels (Fig. 11). In addition, a thin-walled black ware is characteristic; this ware appears to follow an Eneolithic tradition.
       One of the most characteristic features of the EBA pottery is short vertical handles near or slightly below the rim. This feature is attested at Maliq in Phases IIIa and IIIb. Also attested are vessels with well developed vertical handles; this feature continues to be common in the Late Bronze Age.
       Among the shapes, a type with everted rim known at Katundas and Bënjë is known. A very important element is the existence of barbotine decoration in low relief. These specimens appear on thick-walled pots, and suggest parallels with Maliq Phases IIIa and IIIb.
       The most typical technique of ornamentation in the EBA pottery at Konispol is incision. This decoration is haphazardly executed on well-made vessels, and incisions cover much of their surfaces. Here again a continuation of Eneolithic tradition may be noted. Finally, there are a few fragments decorated with bands in relief, with and without punctations.
       It should be pointed out that, aside from the observed similarity in vertical handles mentioned above in connection with the EBA ceramics, only general parallels can thus far be drawn between Konispol Cave and other sites in the later Bronze Age. This era is not well represented in the cave, and this is compounded by the fact that it is poorly defined stratigraphically. In general, it appears that human use of the Konispol Cave in the Bronze Age was sporadic and discontinuous; the ceramic sequence is by no means as complete as it was for the earlier periods.


       Brief mention may be made here of the ceramics of the first millennium B.C. In general, use of the cave seems to have been very sporadic during this period. The sequence has been studied most closely to date in Trench XII, which was opened in the west-central part of the cave at a high point on the cave floor, in the hopes of recovering strata of the latest period of use of the site. The pottery in the upper levels was generally mixed Iron Age through Hellenistic; at ca. 0.22 m below the surface, a hard clay deposit with very few cultural remains was reached, suggesting a hiatus of unknown duration in the use of the cave. The pottery of the uppermost levels is approximately 25% wheelmade, typically unguentaria, black glazed cups, and hydrias. Most diagnostic pieces date to the 3rd-2nd centuries B.C. The handmade pottery consists for the most part of polished and burnished monochrome vessels that are typical of the local Iron Age. Sherds are very small, and shapes are difficult to identify owing to the paucity of rims, bases and handles. In general, the assemblage would seem to be characterized by open utilitarian shapes, primarily for cooking and eating. These pieces and a few matt-painted sherds are paralleled at other sites in southern Albania, and particularly the Korçe Basin.

Proceed to Part 2