Late Mesolithic Individuals of the Danube Iron Gates Origin on the Dnipro River Rapids (Ukraine)? Archaeological and Bioarchaeological Records

6.1 The Ornaments Made of Pharyngeal Teeth of Fish

Personal ornaments made of unperforated pharyngeal teeth of fish are typical and abundant at the cemeteries of the Mariupol type in the Dnipro region and neighbouring territories. Telegin mentions their presence among grave goods from the Late Mesolithic Mar’ivka cemetery, and sub-Neolithic cemeteries of the Vilnianka, Yasynovate I, Lysa Hora, Mykilske II, Mariupol, Deriivka I, Vovnihy II, Chapli, Osypivka, Sobachky, as well in the burial within the Kam’iani Potoky site (Telegin, 1991). They were also found at the Sakarovka burial ground in the Prut-Dniester interfluve, in Moldova (Larina & Dergachev, 2003, p. 100). Some of these sites are radiocarbon dated (Hedges, Housley, Bronk Ramsey, & van Klinken, 1995b, p. 427; Kotova, 2003, Table 1, 3; 2018; Lillie, 1998, p. 186, Table 1; Telegin et al., 2000). Their date range spans approximately from the start of the 7th to the middle of the 5th millennium BC (without considering the freshwater reservoir effect). Therefore, all of them are younger than the earliest date of Vasylivka II.

According to the finds from the Vilnianka burial ground, it was established that the teeth of “vyrezub” and “carp” weighing from 6 to 20 kg were used for ornaments (Shpet, 1956). In the absence of a detailed species definition, the group name “carp” likely means the European carp (Cyprinus carpio carpio Linnaeus, 1759) here. As for the “vyrezub,” it may represent the Black Sea roach, or kutum (Rutilus frisii Nordman, 1840). This species lives in the basin of the Black, Azov and Marmara seas to this day. It is a typical anadromous fish that spends most of its life in brackish marginal sea areas and migrates up rivers in April–May to spawn among weeds and over a stony substrate in shallow waters with fast current, which characterise the Dnipro and Southern Buh River rapids regions (Bulakhov, Novitskyi, Pakhomov, & Khrystov, 2008, pp. 182–183). Genomic studies have shown that in the Mesolithic, fish of this species also reached the Iron Gates reach of the Danube. However, they do not occur there now (Živaljević, Popović, Snoj, & Marić, 2017). Since only mature, large individuals ascend into rivers, a vyrezub was of great subsistence importance, and its pharyngeal teeth were large enough for making the ornaments.

The precise numbers of pharyngeal teeth in the North-Pontic burial grounds and single burials were rarely calculated. For example, a total of 183 unperforated teeth of fish were found in the cemetery of Mar’ivka (Bodianskyi, unpublished, 1950, pp. 26–35). Also, information has been published that a total of 75 such teeth were found at Sakarovka (Larina & Dergachev, 2003, p. 100), 251 – at Mykilske II (Telegin, 1991, p. 55, Table 3), and 422 – at Vovnihy I (Rudynskyi, 1956, p. 155). Their number in certain burials varies from 1 to 2 teeth to “more than 100” in burial No. 20 of a man of 20–25 years old at the Vilnianka cemetery, “about 200” in burial No. 28 of a child at Osypivka (Telegin, 1991, pp. 61, 71), and 204 teeth in burial No. 15 at Vovnihy I (Rudynskyi, 1956, p. 155). However, unclear definitions “several,” “row,” or “handful” are used much more often.

Unlike unmodified pharyngeal teeth of fish, the perforated ones are rarer in Ukraine. Describing the grave goods from Vasylivka II, Bodianskyi (unpublished, 1953, p. 24) wrote that “these teeth have holes cut with a flint on their side, which is so far characteristic only for this burial ground.” Nadiia Kotova repeated this assertion word for word without questioning it (Kotova, 2010, p. 88). However, before the excavation of Vasylivka II, such items had already been found at the Zamil-Koba 1 site in the Crimea (Krainov, 1938). Stoliar (1959, p. 145) noted this fact, which was repeated by Dušan Borić and Emanuela Cristiani dozens of years later (Borić & Cristiani, 2016, p. 98).

Zamil-Koba 1 (44°37′19.25″ N, 33°43′51.99″ E) is a rock shelter site, situated 3.6 km west-southwest of the village of Krasnyi Mak (former Büyük Qaralez) of the Bakhchisarai district in the Autonomous Republic of Crimea, Ukraine (Figure 2). Dmitry Krainov discovered the site in 1935 and excavated an area of 132 m² between 1935 and 1937 (Krainov, 1938). There, he identified two cultural layers – the lower “Azilian” and the upper “Tardenoisian.” Today, the materials of the former reliably correlate with the Shan-Koba Final Paleolithic–Early Mesolithic culture and the second with the Murzak-Koba Late Mesolithic culture (Bibikov, Stanko, & Koen, 1994, pp. 147, 155).

According to Krainov, the burial of a human skull, of which only the occipital bone and a tooth survive, was situated under a rocky wall. These remains lay in a cone-shaped pit 62 cm in diameter, which penetrated 30 cm into the lower part of the lower cultural layer from the violet “hearth” horizon of this layer, 21 cm below the surface. In addition, “a large number of small, indefinable bones of animals and maybe a human,” charcoals, “flint tools and chunks,” as well as “two fish teeth with cuts, which were used as pendants” lay in the pit (Krainov, 1938, pp. 14, 23). On the published drawings of both the finds discussed (Figure 5(1 and 2)), horizontal incisions, which shaped the through holes on the front of their necks, are clearly visible.

Figure 5

Pendants of cyprinids pharyngeal teeth (1, 2, 4–22), and diagonally truncated flint blade (3) from: 1, 2 – Zamil-Koba 1 (after Krainov, 1938, Plate V, Figures 4 and 5), 3–22 – Skelia-Kamenolomnia, burial No. 2 (after Haskevych, 2020, p. 157, Figure 4: 3–22).

Although the filling of the pit with “dark soil” distinguished it from both the greyish-white loam and the violet “hearth” horizon with the Shan-Koba materials, Krainov correlated the burial with the “Azilian” period (Krainov, 1938, p. 14). It should be noted that there is only one pharyngeal tooth with a hole among all materials of the Shan-Koba culture. It was found in Layer 6 of the eponymous Shan-Koba site (Bibikov et al., 1994, p. 67, Plate XX: 12). According to a series of radiocarbon dates (Benecke, 2006; Biagi, Khlopachev, & Kiosak, 2014; Manko, 2010, Table 2), this layer was most probably deposited during the Allerød and Dryas III. However, drilling, not cutting a hole distinguishes this find from the Zamil-Koba 1 pendants. Conversely, some personal ornaments made of animal teeth with similar cuts are among the finds of the upper, “Tardenoisian” layer of the Zamil-Koba 1 site (Krainov, 1938, Plate VIII: 9, 10). It may indirectly indicate that the two pendants under discussion are of the Late Mesolithic age too. This is also evidenced by the abundant ornaments made of fish teeth with cuts from Vasylivka II, as well as by the analogous items from the Skelia-Kamenolomnia burial ground.

The cemetery of Skelia-Kamenolomnia (48°17′43.29″ N, 35°9′23.49″ E) was discovered and explored by Bodianskyi in the Dnipro Rapids region in 1954, i.e., a year after the large excavation at Vasylivka II. Like the latter, it was only mentioned in passing in several works of Telegin (1968, p. 175; 1991, p. 64, Figure 28: 14) and an anthropologist Tamara Konduktorova (1973, p. 35). They both cited the most complete source of information about the site, which is unpublished Bodianskyi’s field diary of 1954 housed in the Archives of IA NASU, Fund 64, Folder 1954/11, No. 2186 (Bodianskyi, unpublished 1954). According to this, the cemetery was located on the top of a 35 m high granite rock on the right bank of the Dnipro River between the villages of Voloske and Mayorka, Dnipro district, Dnipropetrovsk region, 12 km north of Vasylivka II (Figure 1). The granite quarry destroyed seven burials there. They were arranged in one row about 11 m in length running from northeast to southwest. The human remains lay at a depth of 1.20–1.45 m, in the lower part of the clay loam, where it contacted the weathered surface of the gneissic-granite rock. The quarry workers communicated that the skeletons occurred in the extended supine position and lay in pairs at a distance of 0.7–1.0 m from each other, with their heads oriented to the west, i.e., perpendicular to the river. Bodianskyi excavated only two burials (one largely destroyed – No. 1, and the other almost intact – No. 2), which were in situ.

Only bones of the left arm have survived from burial No. 1, which is one of the seven destroyed graves mentioned. There were no finds around them. Burial No. 2 (the eighth in the cemetery) was found 1 m to the south of burial No. 1. Probably, they were paired with each other. The skeleton lay on the bottom of a pit about 1.8 m × 0.7 m at a depth of 0.23 m, with inclined uneven walls. The deceased was stretched out on his back, head to the west, arms along the body, and hands near the pelvis. The skull lay on the nape. The workers damaged the facial part, as well as the bones of the feet. A flint “trapeze” was found near an elbow of the deceased. Several Paludinae shells lay near the left foot, to the right of the top part of the right femur and lower ribs. Also, about 30 such shells, as well as fragments of Unio shells, were in the filling of the pit above the tibia. A few perforated teeth of Cyprinidae fish lay to the right of the pelvic bones, below the pelvis and over the ribs. About a dozen of the same ornaments were found under the shinbones at the very feet, three of them – “in one row.” It is specified that in total 14 intact and 4 fragments of such items were discovered at the site (Bodianskyi, unpublished, 1954, pp. 124–134).

Bodianskyi collected also the human bones from the destroyed graves, which were scattered on the surface of the quarry. In particular, he described the one almost complete skull as “burial No. 3.” It was transferred to the Anthropology Museum of Moscow State University, where Konduktorova examined it (Konduktorova, 1973, Figure 12, Table I).

Grave goods from burial No. 2 are currently housed in the scientific repository of the IA NASU. There is the broken thin, long, strongly patinated, diagonally truncated flint blade, mentioned by Bodianskyi as a “trapeze” (Figure 5(3)), 9 intact and 10 fragmented pendants made of cyprinid pharyngeal teeth (Figure 5(4–22)). They were perforated in the manner described for the ornaments from burial No. 1 at Vasylivka II. Holes are present on all intact items, and their traces are found on six of the nine fragmented teeth broken along the cutting line. Based on the “fresh” appearance of the fractures, some pendants were fragmented during or after the excavation (Haskevych, 2020, pp. 157–159).

No finds from the cemetery have been radiocarbon dated. However, the arrangement of the burials in one row, the extended supine position of the skeletons, and the presence of paired graves clearly indicate a late Mesolithic or very early sub-Neolithic date.

In Europe beyond the North-Pontic area, most Stone Age personal ornaments, made of cyprinid teeth, have been found at Mesolithic and Neolithic sites in the Danube catchment. These items are particularly abundant in Late Mesolithic burials in the Iron Gates reach of the Danube. The sites of Vlasac, Schela Cladovei, Kula, and Lepenski Vir yielded both unmodified and perforated items. Their number sometimes reaches several hundred in one burial. For example, at Vlasac, 310 perforated, 255 unmodified, and 77 fragmented teeth were in burial H2 of a woman 30–40 years old, and 252 perforated, 294 unmodified, and 155 fragmented teeth were found in burial H297 of a child. Radiocarbon dates for these burials fall around 6700–6500 BC adjusted for the freshwater reservoir effect (Cristiani & Borić, 2012). The morphology, typology, techniques, functional analysis, chronology, and archaeological context of the finds have been published in detail (Borić & Cristiani, 2016, pp. 95–99; Borić et al., 2014; Cristiani & Borić, 2012; Cristiani, Živaljević, & Borić, 2014; Mărgărit, Radu, Boroneanț, & Bonsall, 2018). This obviates the need to consider these issues here. It should be only emphasised that the holes were made by cutting on the anterior surface of the tooth neck, i.e., in the manner described earlier for the ornaments from Ukraine. By the same token, more than 300 km from the Iron Gates, in western Montenegro, one cyprinid tooth perforated in such a way was found in the Mesolithic layer of the Vrbička Cave site (Figure 2), dated to the late 8th–early 7th millennium BC (Borić & Cristiani, 2016, p. 99).

In addition to the aforementioned facts, ornaments made of pharyngeal teeth were found at three Late Mesolithic sites on the Upper Danube in Germany, almost 1,000 km west of the Iron Gates (Figure 2). Holes are missing on all 12 teeth from the secondary burial of skulls at Hohlenstein-Stadel cave, ¹⁴C dated to 7029–6499 cal. BC (hereinafter, all ¹⁴C dates have been calibrated using OxCal v 4.4.4 software (Bronk Ramsey, 2009) and the IntCal20 atmospheric curve (Reimer et al., 2020) and are given with a 95.4% confidence level). The hole was not cut but drilled in the lateral surface of the neck of seven teeth from the burial in Falkensteinhöhle site as well as nine teeth from cultural layers 9–11 at the Burghöhle Dietfurt site (Gehlen, 2016, pp. 832–838; Rigaud, Vanhaeren, Queffelec, Le Bourdon, & d’Errico, 2013). This distinguishes them from the pendants of the Iron Gates. The tradition of decorating clothes with pharyngeal teeth of fish survived on the Upper Danube until the Neolithic. This is evidenced by 20 such finds without holes from the burial of a child within the Linear Band Pottery culture settlement of Eichendorf-Aufhausen (about 5300–5200 BC) in Lower Bavaria (Gehlen, 2016, pp. 840–841).

6.2 The Engraved Bone Plates (“Bracelets”)

According to a drawing published by Telegin, and a laconic description in Bodianskyi’s field report, the bone plates (“bracelets”) from Vasylivka II had thinned rounded (oval) ends with one round hole and were decorated with geometrical patterns including meander, zigzag, rhombuses, and triangles. The decoration was applied with both straight and indented cut lines (Figure 6(1–7)). As Gokhman wrote, all the plates were found in burial No. 16 next to the remains of a mature male (Gokhman, 1966, p. 96). In the Rapids region, their closest analogues are two bone items from the site of Kizlevyi Ostriv 5 (Tuboltsev, 2003, p. 40). In 1942, Bodianskyi explored it in the southern part of the now flooded Kizlevyi Island (approximately 48°13′5.07″ N, 35°12′4.21″ E), located in the middle of the Dnipro in front of the Tiahynka gully, about 3 km north of Vasylivka II (Figure 1). Since Bodianskyi was concerned of being considered a collaborator of the German occupiers, he never made public information about this site. Also, there are no field records about it. Thus, the context, where the materials were found, is unknown. They were published after the death of Bodianskyi (Kotova & Tuboltsev, 2013; Tuboltsev, 2003). The collection, which he stored in his home, contains two fragmented bone items of an elongated oval shape with holes at the ends. According to published descriptions and drawings (Kotova & Tuboltsev, 2013, p. 50, Figure 13: 10, 11; Tuboltsev, 2003, p. 38, Figure 6: 16, 18), both were decorated with an ornament consisting of rhombic and triangular zones, as well as straight bands filled with rows of separate subtriangular notches or engraved lines with notches applied to them (Figure 6(8 and 9)).

Figure 6

Bone plates from the Dnipro Rapids sites: 1–7 – Vasylivka II (after Telegin, 1991, Figure 15), 8, 9 – Kizlevyi Ostriv 5 (after Tuboltsev, 2003, p. 44, Figure 6: 16, 18).

The only ¹⁴C date from Kizlevyi Ostriv 5 (Ki-7999) 6740 ± 90 BP was measured on an animal bone (Kotova, 2003, p. 122, Table 2) and falls into the range of 5796–5481 cal. BC. However, the flint collection from the site, which consisted of some 5,000 chipped stone artefacts, looks heterogeneous. It is possible that it also included earlier, Mesolithic, materials (Haskevych, 2020, p. 162).

Beyond the North-Pontic area, analogues to the “bracelets” from Vasylivka II are found far to the northwest, in central Poland. They were discovered in burials within the settlements of the Early Chalcolithic Brześć Kujawski culture, dated to 4350–4000/3900 BC (Czerniak & Pyzel, 2019). There, the majority of the deceased of both sexes and all ages were interred flexed on their side, with their head to the south. Decorated plates with one or two holes at each end, made of cattle ribs, are one of the culture-defining features. Most of them were found at the settlements of Krusza Zamkowa 3, Brześć Kujawski 4, and Racot 18 (Figure 2). The largest number in a single burial was 18. Usually, they lie on the hands of the deceased (e.g., Bednarczyk, Czerniak, & Kośko, 1980; Czerniak, 2002, pp. 14–19, Figures 6 and 7).

Researchers commonly note the duality of the Brześć Kujawski culture, which shows the farmer Danube tradition for settlements and burials rites and significant reminiscences of the traditions of the north European hunter-gatherers of the Maglemosian circle for personal ornaments and some tools (Czerniak, 1980, p. 171; 2007, p. 238; 2012, p. 162). However, there is no clear evidence of the genesis of the Brześć Kujawski bone plates among the local foragers of the southern Baltic and Scandinavia, since only single decorated bone plates were found on a few sites (Ralswiek-Augustenhof, Kraghedgård, Ostorf), dated to the same or a later time (Bogucki, 2008, pp. 56–58; Czerniak, 2007, p. 246, Figure 11: 15, 16; Hartz, Lübke, & Terberger, 2007, pp. 585, 589, Figure 15; Lübke, Lüth, & Terberger, 2007, p. 310).

Other analogues for the “bracelets” from Vasylivka II were found far to the east – in the middle cultural layers (2A and 2B) at the Varfolomeevskaya site, located in the steppe Volga-Ural interfluves (Yudin, 2004). Based on a large series of radiocarbon measurements, the excavator dates these layers to about 6200–5200 BC (Yudin, 2014). He interprets the decorated bone plates from there as tools for ceramic production as well as personal ornaments – pendants (Yudin, 2004, pp. 93–104, 120–124). The elongated oval shape and the presence of holes at the end bring them closer to the items under discussion. However, their decoration is not so similar. Likewise, some morphological and stylistic likeness with the plates from Vasylivka II can be seen in the engraved bone and antler items from the Late Mesolithic site of Razdorskaya 2 on the Lower Don. A series of ¹⁴C measurements date this site to the 7th millennium BC. These finds were interpreted as “pendants,” given the presence of two holes at one end (Gorelik, Tsybrii, & Tsybrii, 2014, pp. 255, 268–269, Figure 15).

6.3 The “Mediterranean (?)” Spiral Shells

The mention by Bodianskyi of “necklaces made of specific spiral Mediterranean (?) shells,” found in Vasylivka II, drew the attention only of Kotova, who literally cited these words (Kotova, 2010, p. 88). To date, there is no other information about finds of shells of Mediterranean molluscs and personal ornaments made of them at the Mesolithic and Neolithic sites in the Dnipro Rapids. However, such artefacts are quite widely represented at synchronous sites in South, South-Eastern, and Central Europe. Most often these are the shells of Spondylus, Columbella rustica, Cyclope neritea (Grünberg, 2013, pp. 247–248). The last two species could be called “spiral.” Their finds are known in the Early Mesolithic layers of Cuina Turcului rock shelter as well as in Late Mesolithic contexts at Ostrovul Banului and Vlasac, all in the Iron Gates area (Boroneanţ, 1973, Planche IV: 9; Cristiani & Borić, 2012, p. 3465). The discovery of ornaments made of such shells at a distance of more than 400–500 km from a possible source of origin (the southeastern coast of the Adriatic and northern shores of the Aegean Sea) indicates an advanced exchange of them over very long distances (Cristiani & Borić, 2012, p. 3467). Thus, their diffusion in the Dnipro Rapids region from the same source(s) as those from the Iron Gates also looks quite probable, but it remains unproven now because of the temporal or irrevocable loss of the most grave goods from Vasylivka II.

6.4 The “Vessel” Made of a Terrapin’s Carapace

The “treated” dorsal shell of a terrapin from Vasylivka II was mentioned by Gokhman (1966, p. 96) as goods from grave No. 27, where a mature male was interred. Analogues of it were found in several other burial grounds in the Dnipro Rapids region. For example, they accompanied extended supine skeletons in burials No. 2 and No. 5 at the Mykilske I (Nenasytets) burial ground, located across the river opposite Vasylivka II (Bodianskyi, 1951, p. 168). It has not been dated by radiocarbon, but its Late Mesolithic or early sub-Neolithic age is evident given the funerary tradition and the absence of pottery. Eight scutes of one (?) terrapin shell were found in a burial within the sub-Neolithic settlement of Ihren V (“Horodok”), 27.4 km north of Vasylivka II (Figure 1). Individual scutes were found in burials No. 33 and No. 35 at the Vovnihy II burial ground, as well as in the collective grave pit “A” at the Mykilske II burial ground (Telegin, 1991, pp. 49, 51). The last two cemeteries are dated to the late sub-Neolithic (Hedges et al. 1995b, p. 427; Lillie, 1998, Table I; Kotova, 2018, p. 57).

In remote areas, the shells of European pond turtle (Emys orbicularis) are known in the Mesolithic burials of the forest zone of Europe, in Sweden (Skateholm II), Germany (Bad Dürrenberg), Poland (Dudka), and Latvia (Zvejnieki). However, only in the burial of a woman and a newborn in Bad Dürrenberg, three unbroken carapaces were laid in. In the other cases mentioned, individual scutes of terrapins were found (Grünberg, 2013, pp. 244–245; Gumiński & Bugajska, 2016, pp. 496–501).

6.5 The Unio Shells

On the Dnipro Rapids, in addition to Vasylivka II, numerous unmodified shells of Unio river molluscs are among the finds from the Mesolithic Vasylivka I and Vasylivka III cemeteries (Telegin, 1982, pp. 205–210), as well as the sub-Neolithic cemeteries of Vovnihy II, Mykilske II, Nenasytets, Vilnianka, and Yasynovate I (Telegin, 1991). In synchronous burials located outside the region, such shells are ubiquitous, but most often they occurred in burial grounds of the Baltic zone (Grünberg, 2013, pp. 246–248, Figure 8).

6.6 The Deer (Elk) Incisors

Bodianskyi and Gokhman reported discovering in Vasylivka II unmodified incisors of a “deer” and “elk,” respectively. According to Gokhman (1966, p. 96), they all accompanied the skeleton of a mature male with bone “bracelets” in burial No. 16. Analogues of them are absent among burial goods of the other Mesolithic cemeteries in the region (Telegin, 1982, pp. 205–210). Also, such items are not mentioned in Telegin’s book on the sub-Neolithic burial grounds of the Mariupol type (Telegin, 1991). It should be noted that one of the most typical finds from almost all local sub-Neolithic necropolises is a pendant made of a rudimentary reindeer tooth with a drilled hole (Telegin, 1991, p. 16, Figure 13, Tabe 1). However, such ornaments are absent from Vasylivka II, which distinguishes it from the rest of the burial grounds, which was first stressed by Bodianskyi (unpublished, 1953, p. 24).

Beyond the region, beads and pendants made of deer and elk incisors are very numerous in Mesolithic and sub-Neolithic burials in northern Europe. Much less often, unmodified teeth of these animals are found there (Grünberg, 2013, pp. 233–236, Figure 2a). For example, in the Olenij Ostrov burial ground on lake Onega, the finds of untreated elk incisors are characterised by the words “as an exception,” while pendants made of such incisors amount “up to 4,500 items” (Gurina, 1956, pp. 17, 138).

7.1 Craniology

Gokhman mentioned the remains of at least 17 individuals from Vasylivka II in the collection stored at the MAE (the Kunstkamera). In particular, he characterised the long bones of 12 individuals, as well as skulls of 11 males and five females (Gokhman, 1958; 1966, pp. 96–131). While studying the skulls from Vasylivka II, he first distinguished two variants – dolichocephalic and meso-brachycephalic – of the Late Mesolithic and sub-Neolithic populations in the Dnipro region.

The dolichocephalic individuals (male skulls No. 11, 19, and 23, and female skull No. 18) have a very high neurocranium, a highly developed supraorbital ridge, a straight medium-wide forehead, and a slightly protruding rounded occiput. The face is moderately tall and wide, mesognathic or orthognathic. Its horizontal profile is very sharp. Orbits are low. The nasal opening is of narrow or medium width and moderate height. The nose bridge is straight, relatively narrow, and high. The nose protrusion is strong (Gokhman, 1966, p. 129).

The meso-brachycephalic individuals (male skulls No. 10, 12, 16, 17, and 21 and female skulls No. 8, 22, 24, and 25) have a high or medium-high (for brachycephalic skulls) neurocranium of medium or very considerable width. The forehead is very wide, moderately retreating, with well-developed superciliary arches. The occiput is angular, less often rounded. The face is very broad, medium-high, orthognathic or mesognathic, sometimes with mild alveolar prognathism. The horizontal profiling is weak in both the top and the middle. The orbits are very wide and low. The piriform opening is wide or medium-sized, sometimes narrow on female skulls. The nose bridge is wide and rather low. The angle of the nose is wide. The dorsum profile of the nasal bones is concave (Gokhman, 1966, p. 130).

Gokhman noted a combination of the characteristics of both variants in the dolichocephalic male skull No. 20, which has a very high neurocranium, very wide cheekbones, and a significant flattening of the face (Gokhman, 1966, p. 131).

In general, both described variants belong to the massive late Cro-Magnon type. The presence of dissimilar skulls in the one burial ground may suggest the group of Vasylivka II was mixed. The presence of individuals of different craniological groups in paired burials (for example, No. 23 and No. 24) confirms this too (Gokhman, 1966, p. 129). But, based on the available radiocarbon dates, the dolichocephalic individual No. 18 is later than three meso-brachycephalic individuals No. 21, 22, 27 (Tables 1 and 2).

Gokhman compared the average craniological indicators for Vasylivka II with the indicators for the earlier Mesolithic burial grounds of Vasylivka I and Vasylivka III, as well as for the later sub-Neolithic burial grounds of Vovnihy II and Vilnianka. From the results, he concluded that the sub-Neolithic series stand between the Early Mesolithic and the series from Vasylivka II. The reason for this is the numerical superiority and expressive characteristics of the meso-brachycephalic individuals in Vasylivka II (Gokhman 1966, pp. 174–181). Thus, the striking uniqueness of the group from the cemetery allowed him to assert that “the fact of the emergence of a new population in the Dnipro Rapids–North Azov Sea region at the turn of the Mesolithic and Neolithic may be considered proven” (Gokhman, 1966, p. 182).

The genesis of the dolichocephalic variant of Vasylivka II on the basis of some local Early Mesolithic populations is generally accepted (Potekhina, 1999, pp. 131–133), and the issue of the genesis of the meso-brachycephalic variant has not yet been unambiguously resolved. Gokhman and Konduktorova saw its origins in the unspecified forest-steppe regions of Eastern Europe. However, they did not compare the Vasylivka II series with the series from other burial grounds outside the Dnipro-Azov area. Instead, they and Georgiy Debets combined all skulls from this burial ground with the skulls from all sub-Neolithic cemeteries of Mariupol type examined up to that time. In their opinion, this combined series mostly resembled the skulls from the Epipaleolithic burial grounds of Afalu-bu-Rummel and Taforalt in North Africa, four skulls of the Ertebølle culture in Denmark published to that time, as well as some skulls from the Olenij Ostrov burial ground in Karelia (Debets, 1966, p. 22; Gokhman, 1966, pp. 174–184; Konduktorova, 1973, pp. 48–49).

In the 1970s, the first results of anthropological studies of the Stone Age population in the Iron Gates became available. The dead from the Danube Valley sites were divided into several types and subtypes (see review: Mikić, 1981, pp. 44–65; 1992). Among them of special interest are rare, very robust relict Upper Palaeolithic individuals resembling the Brno-Předmostí ancient European race, and relatively numerous more gracile forms closest to the Cro-Magnon race. The latter is regarded as an indigenous population, while the former, represented only at Vlasac (type A-I), were considered migrants from elsewhere. The skulls of individuals from burials 55 and 78a are mentioned as the most representative of this type (Mikić, 1981, p. 47; 1992, p. 35). It is noteworthy that in burial No. 78a, seven perforated cyprinid pharyngeal teeth and one spiral shell of the Mediterranean mollusc Cyclope neritea were found (Cristiani et al., 2014, p. 294, Table 1). Unfortunately, no one has intentionally compared the metric indicators of both the mentioned types and the two craniological variants from the Dnipro Rapids. Instead, the combined Meso-Neolithic series from the Iron Gates sites was compared with the combined series from other regions. As a result, it was found that the series from the Mariupol type cemeteries in Ukraine showed the greatest similarity to it (Alekseev, 1979, p. 50; Zoffmann, 2000, p. 75).

For the first time, Inna Potekhina (1999) applied principal component analysis and factor analysis to materials from several sub-Neolithic burial grounds from the Dnipro region as well as the combined series from the Mariupol-type cemeteries. She concluded that there had been significant morphological isolation of the Vasylivka II group when compared to the rest of the local Mesolithic and sub-Neolithic populations (Potekhina, 1999, pp. 90–92, 133–134). In addition, in the series from the burial grounds of Yasynovate I and Vilnianka as well as the series from the entire region, she identified two main craniological variants, which generally correspond to the two variants of Vasylivka II identified by Gokhman. Her comparison of the combined Neolithic series from the Dnipro region and the skulls from remote territories showed its greatest similarity to the Mesolithic burial ground of Popovo in the Arkhangelsk region in the Russian Federation (Figure 2), which belongs to the Kunda culture. Five radiocarbon measurements on human bones (Oshibkina, 2007, p. 44) date this cemetery to 9265–8346 cal. BC and 7025–6086 cal. BC (without correcting for the freshwater reservoir effect). According to Potekhina (1999, p. 147), the territory near the Lacha, Vozha, and Beloe lakes, where Popovo is located, could be the place of origin of the meso-brachycephalic variant from whence it spread to the Dnipro region. She also pointed out possible northwesterly contacts of the meso-brachycephalic population of Ukraine, based on the similarity of their skulls to those from the Vedbæk and Skateholm burial grounds in the Western Baltic (Potekhina, 1999, p. 142).

7.2 Odontometry, Dental, and Jaw Pathology

In the late 1950s, Gokhman recorded that another peculiarity of the individuals from Vasylivka II is the frequent presence of torus mandibularis. This osseous pathology is present on 3 of the 16 skulls examined by him. They are adult males of the meso-brachycephalic (No. 12, 27) and mixed (No. 20) variants. In contrast, torus mandibularis is recorded on only 1% of skulls from the other local burial grounds studied by him (Gokhman, 1966, pp. 110, 177).

On a global scale, torus mandibularis shows a strong clinal distribution with the highest frequencies in northern latitudes. Although torus mandibularis has some hereditary component, it has a strong environmental component of variance. How the latter influences torus expression remains enigmatic (Scott, Schomberg, Swenson, Adams, & Pilloud, 2016).

Based on data presented by Mirjana Roksandić (1999, pp. 255–260), in the Iron Gates, torus mandibularis occurs on 63.6% of lower jaws examined from Padina, 45.4% from Vlasac, 40.0% from Hajdučka Vodenica, and 19.6% from Lepenski Vir (Haskevych, 2020, p. 172). It is noteworthy that at Vlasac, perforated fish teeth were found in burials No. 14, 60, 74, and 78a (Cristiani et al., 2014, p. 294, Table 1), where individuals with torus mandibularis (Roksandić, 1999, pp. 256, 258) were interred. In addition, individual No. 78a is one of the individuals referred to the possibly intrusive robust craniological type A-I.

Aleksander Zubov examined the size and shape characteristics of 138 human molars from Vasylivka II. He noted the relatively large size of the teeth in the series and included them in the “macrodont” group, which also contained some past and modern equatorial and mongoloid populations (Zubov, 1968). Subsequently, based on these data, Potekhina stated that the teeth from Vasylivka II are larger than those from the sub-Neolithic cemetery of Vovnihy II (Potekhina, 1999, pp. 75–86).

In the late 1980s, Jacobs (1994a) investigated the odontological characteristics of the deceased from the Dnipro Rapids cemeteries, in particular Vasylivka II. However, he did not provide data from individual sites but included Vasylivka II in a combined “Neolithic” series. His generalised conclusion is an increase in dental and gnathic dimensions between the Mesolithic and sub-Neolithic cemetery samples. Since other instances of the Neolithic transition demonstrate the opposite trend, Jacobs suggested a connection of the observed dynamics with possible gene flow into the Dnipro region, which increased the overall robustness of the skeletons, including skulls, jaws, and teeth. Searching for the origin of this gene flow, he excluded the territories of the north of Eastern Europe because the Mesolithic people from Olenij Ostrov had smaller teeth. He also noted that according to the available data on the teeth from the Iron Gates Mesolithic site of Vlasac, their size was larger than that of teeth from the Ukrainian Mesolithic, but not in the order likely to produce the Ukrainian sub-Neolithic dental increase.

In the mid-1990s, Lillie examined human remains from Vasylivka II and noted the presence of dental calculus in half of the studied individuals interred there, and the absence of caries on their teeth. This made it possible to characterize their nutrition as rich in proteins, and not in starch found in cereals, roots, and seeds (Lillie, 1996, pp. 138–140). He also concluded that the pathological markers evident on the skeletal materials indicate that the subsistence economy remained relatively stable during the Mesolithic and Neolithic in Ukraine. The absence of caries and the consistent occurrence of calculus and enamel hypoplasias confirm the consumption of a primarily meat-oriented/high-protein diet during these periods.

7.3 Body Size and Shape

These characteristics are estimated from skeletal dimensions through calculations based on the indices computed from the limb bones. The preservation of the postcranial parts of the skeletons in Vasylivka II is poor. Only tubular bones have survived among most of them. They are all very large and robust, but with a slightly smoothed relief. In measuring them, Gokhman calculated the stature of eight adult men in the range of 167.0–175.0 cm, and five women in the range 158.0–166.7 cm. Therefore, the average stature of males was 172.7 cm and females was 162.7 cm (Gokhman, 1966, pp. 118–127, Table 22).

Comparison of the results provided by Gokhman with the calculated stature of the Mesolithic people from different parts of Europe shows the greatest similarity of individuals from Vasylivka II with the populations from the Eastern Baltic and Karelia, as well as the inhabitants of the Iron Gates. For example, the average height of males from Schela Cladovei was 182 cm and of females was 165 cm (Boroneanţ et al., 1999, p. 389). In another publication, the mean statures of 12 males and five females from this site are referred to as 174.9 cm and 163.8 cm, respectively (Niskanen, Ruff, Holt, Sládek, & Berner, 2018, p. 77). For the Vlasac population, a mean stature of 172.5 cm was calculated from the remains of 14 males and 159.5 cm for 12 females (Formicola & Giannecchini, 1999, Table 4a). A similar average (171.5 cm) was calculated for 17 Late Mesolithic males from Latvia. However, the females from there were somewhat shorter (156.0 cm) than those in the Dnipro Rapids region, based on the bones of six individuals (Gerhards, 2005, Table 1). Bones of 34 males and 17 females from Olenij Ostrov indicated the average stature of 173.1 and 162.6 cm, respectively (Formicola & Giannecchini, 1999, Table 4a). Unlike the sites mentioned, bone lengths, and hence stature, of the Mesolithic people of both sexes from the Mediterranean basin (El Collado, Arene Candide, Uzzo, and Molara), Western Europe (Muge, Téviec, and Hœdic), and Southern Scandinavia (Vedbæk, Zealand, and Skateholm) were equally significantly smaller (Formicola & Giannecchini, 1999, Table 4a, 4b; Jacobs, 1993, pp. 318–320).

It should be noted that data derived using different equations have been compared earlier. However, even taking into account the possible error, it is obvious that the stature of the Mesolithic population of the Iron Gates was more similar to that of the Dnipro basin inhabitants compared with the territories to the south, west, and north of the Danube.

7.4 Stable Isotope Analysis

In the early 1990s, Jacobs was the first to provide stable isotope data on the bone collagen from Vasylivka II. As a result, he obtained the δ¹³C ratios of three individuals, the radiocarbon ages of which were also measured. In addition, Jacobs found that the barium concentration in human bones from Vasylivka II was almost twice that of bones from the earlier, neighbouring cemetery of Vasylivka III. However, he did not publish any specific δ¹³C and Ba ratios, as well as sequence numbers of samples and burials (Jacobs, 1993, 1994a,b). In his opinion, these results show a decrease in the consumption of animal protein and an increase in the consumption of plants, particularly cereal grains (Jacobs, 1993, pp. 313, 322). As another possible food source, he proposed C₄ plants, e.g., “wild millet” and “lambs quarters” or other species of the genus Chenopodium, growing in the northern Black Sea area (Jacobs, 1994a, p. 14; 1994b, pp. 57–58). Matching this with the increase in the robustness of the skeletons, Jakobs concluded that the simultaneous intensification of physical activity and a plant-based diet might be a consequence of the adoption to a significant extent of certain aspects of a typically “Neolithic” subsistence already in the 7th millennium BC (Jacobs, 1993, 1994a,b).

In 1996, Lillie published the precise generalised range of δ¹³C ratios (from −19.5 to −20.4‰), previously announced by Jacobs. Lillie suggested that these values correspond to those expected for humans consuming a mixture of the flesh of herbivores feeding on C₃ plants and C₃ plants themselves (Lillie, 1996, p. 136). He emphasised that this conclusion agrees with the results of his odontological studies.

In 2003, Lillie (2003, p. 7) briefly mentioned the average results of a new series of the δ¹³C and δ¹⁵N values measured on human bones from Vasylivka II. Three years later, in co-authorship with Jacobs (by then deceased), Lillie presented details of these new measurements on 14 individuals; the δ¹³C ratios varied from –20.14 to –21.84‰, and δ¹⁵N from 12.35 to 14.74‰ (Lillie & Jacobs, 2006, pp. 881, 883, Table 1). It is noteworthy that they include the results of the analyses on samples with the same museum numbers as three previously measured samples. In all three cases, the newly obtained δ¹³C ratios became more negative. Only one of them differs by c. 0.1‰, which is within the typical replicate measurement error. The difference between the old and new ratios for the two others is approximately 0.4 and 2‰ (Table 2). However, the previous results are not mentioned in the article. Thus, the reason for these shifts remains unexplained. Another important feature of the article is the change of sequence numbers of the burials (they are also skeletons and skulls) when compared to those previously given by Gokhman. Thanks to both Gokhman and Lillie reporting the same museum numbers, the lists of the burials analysed could be harmonised. This is confirmed by the correspondence of the sex and age determinations, published by the respective with a 40-year gap (Table 2).

The wider range of the new results allowed Lillie to modify previous ideas about the diet of the Vasylivka II group. He believed that the higher nitrogen values (compared to the previous results for Mesolithic burials in the region) could indicate a slightly higher input from freshwater fish, or the exploitation of fish species from higher trophic levels of the riverine ecosystem. In turn, the broader spread of δ¹³C ratios reflects the mixed diet, which includes a wide spectrum of terrestrial resources of both animal and plant origin, as well as different freshwater resources. However, on average the δ¹³C values of Vasylivka II are more positive than evidenced elsewhere in the vicinity (Figure 7), whether from earlier, contemporaneous or later cemeteries (Lillie & Jacobs, 2006, pp. 882–884).

Figure 7

Scatterplot of δ¹⁵N and δ¹³C isotope values of human bones from: 1 – Vasylivka II cemetery (according to data from Lillie and Jacobs, 2006, p. 881, Table 1); 2–9 – Mesolithic (2, 3) and subNeolithic (4–9) cemeteries in the Dnipro region (according to data from: Lillie & Richards, 2000, p. 966, Table 1; Lillie et al. 2009, pp. 260, 261, Tables 3–5; 2011, p. 62, Table 1; Lillie, Henderson, Budd, & Potekhina, 2016, p. 747, Table 2; Mathieson et al., 2018, Supplementary Table 6; Budd, Potekhina, & Lillie, 2020, p. 8, Table 2); 10–19 – Mesolithic and Neolithic sites in the Iron Gates region (Schela Cladovei – according to data from Bonsall et al., 1997, p. 65, Table 4; all other – according to data collected in Jovanović et al., 2019, online resource 2 and 3 with all linked references). Illustration by D. Haskevych. Legend: 1 – Vasylivka II, 2 – Vasylivka IIІ, 3 – Mar’ivka, 4 – Deriivka I, 5 – Mykilske ІІ, 6 – Vasylivka V; 7 – Yasynovate I; 8 – Vilnianka; 9 – Vovnihy II; 10 – Hajdučka Vodenica (M); 11 – Hajdučka Vodenica (TEN); 12 – Icoana (TEN); 13 – Lepenski Vir (M); 14 – Lepenski Vir (TEN); 15 – Lepenski Vir (E-MN); 16 – Padina (M); 17 – Padina (TEN); 18 – Schela Cladovei (M); 19 – Vlasac (M). Numbers – Vasylivka II burial numbers (after Gokhman, 1966). Dash line – the conditional boundary of the “Dnipro” and “Danube” values of δ¹³C. Abbreviations: M – Mesolithic (before 6200 cal. BC); TEN – Transition to Early Neolithic (6200–5900 cal. BC); E-MN – Early-Middle Neolithic (5900–5500 cal. BC).

Potekhina stated that the average barium concentration in the human bones measured by Jacobs from Vasylivka II was “230,” which contrasts sharply with the value of “136” measured by him for the Vasylivka III burial ground. Probably, the numbers cited are indexes of log (Ba/Ca) “−2.30” and “−1.36,” respectively. As the source of the information, she referred to the article by Jacobs, where these data are absent (Potekhina, 2005, p. 169). Thus, the actual source remained unspecified.

Subsequently, Lillie and co-authors reinterpreted the Vasylivka II subsistence as being divergent from the general trend for the region. Despite acknowledging that an analysis of δ¹³C and δ¹⁵N ratios is not intended to be used to distinguish local from nonlocal individuals, they nevertheless suggested that the source of carbon in the diet of the Vasylivka II group could have differed from the main sources throughout the Dnipro Basin. Hence, it appears that this population “had recently moved into the region and had not resided for a length of time sufficient to allow their stable isotope ratios to reach equilibrium with the local environmental ranges” (Lillie et al., 2012, p. 86).

Finally, in 2018, new stable isotope values from Vasylivka II were published within the scope of an extensive ancient DNA study. The adult dolichocephalic female from burial 18 produced ratios of −20.09‰ for δ¹³C and 13.71‰ for δ¹⁵N, C:N = 3.44 (Mathieson et al., 2018, Supplementary Table 6). Comparing them with the values obtained by Lillie for the same skeleton shows the δ¹³C ratio shifted by only 0.12‰, and the δ¹⁵N ratio became more positive by 0.91‰ (Table 2).

In some earlier studies, where isotopic values for Vasylivka II were reported, the results from Ukraine were compared with those for the Iron Gates Mesolithic sites (Lillie, 2003, p. 7; Lillie & Jacobs, 2006, p. 883). The main reason for this is the leading rates of studying the palaeodiet of the population in the Middle Danube, where stable C- and N-isotope data for a large series of Mesolithic human bones along with the bones of fish and animals were published already in mid-1990s (Bonsall et al., 1997, 2000). Values of about −20 to −19‰ for δ¹³C and values of about 14 to 16‰ for δ¹⁵N were reported for humans from the sites of Vlasac, Lepenski Vir, and Schela Cladovei (Figure 7). According to these data, throughout the Mesolithic, the local Danube groups would have had access to two isotopically distinct major protein sources – riverine food and terrestrial herbivores. However, the bulk of the protein was thought to be derived from riverine fish, including anadromous species (Bonsall et al., 1997, 2000, pp. 120–122; 2004, p. 297; Borić, Grupe, Peters, & Mikić, 2004). Based on this, the difference in the isotopic values of the bones of the Mesolithic inhabitants of the Iron Gates and the Meso-Neolithic population in the Dnipro Rapids region was explained by the consumption of more terrestrial herbivores in Ukraine (Lillie & Richards, 2000, pp. 966–968) and/or as a reflection of some peculiarity of the freshwater ecosystem of the Dnipro catchment when compared to that of the Danube (Lillie et al., 2003, p. 748) due to the possible input of groundwater carbon (Lillie & Jacobs 2006, p. 882). On the other hand, it was noted that the relatively positive δ¹³C and δ¹⁵N values of several individuals from Vasylivka II were closer to the Iron Gates Mesolithic values (Lillie, 2003, p. 7), which were linked with a more abundant freshwater diet and exploitation of fish species of higher trophic levels (Lillie & Jacobs, 2006, p. 883; Lillie et al., 2011, p. 63).

New stable isotopic studies including δ³⁴S analysis have supplemented and clarified the data on the subsistence of the Mesolithic Iron Gates populations (Bonsall et al., 2015; Jovanović et al., 2019). The local Late Mesolithic humans consumed more anadromous fish, which may explain some shift in the δ¹³C and δ¹⁵N ratios towards “marine” values. In addition, the possible eating of dogs is discussed there as a reason influencing the results of the analysis (Bonsall, 2008; Jovanović et al., 2019). However, this does not change the previous findings regarding the differences/similarities of the diet of the Danube and Dnipro Mesolithic individuals.

7.5 DNA

To date, DNA was extracted from only two Vasylivka II individuals – the mature dolichocephalic woman No. 18 and the mature meso-brachycephalic man No. 21. They both belong to the U5b2 mtDNA subclade. In terms of the generally accepted subdivision into groups by genetic relatedness to a hypothesised set of ancestral populations (Haak et al., 2015; Lazaridis et al., 2014), these results are interpreted as representatives of the “Eastern” hunter-gatherers (EHG), according to the supervised ADMIXTURE analysis, that modelled each ancient individual as a mixture of populations represented by four main clusters, regarded as European Mesolithic. Similar results were also obtained for the chronologically older Early Mesolithic in the Dnipro Rapids. However, the local younger sub-Neolithic burial grounds having generally EHG affiliation demonstrate frequent, although not massive, admixtures of the “Western” hunter-gatherers (WHG) component (Mathieson et al., 2018, Figure 1: D).

In contrast, according to the aforementioned model, the Mesolithic population of the Iron Gates turned out to be a typical WHG with rare admixtures of EHG components (Mathieson et al., 2018, Figure 1: D). It should be noted that the bearers of the same subclade U5b2, which was recorded in Vasylivka II, are among them. Thus, in the Iron Gates area, they were found in synchronous burials at Schela Cladovei (male M95/2), Vlasac (female H232), and Hajdučka Vodenica (male No. 21 and No. 31). In addition, the man from Vasylivka II and the man from Hajdučka Vodenica (No. 31) are of the same Y-DNA subclade R1b1a (González-Fortes et al., 2017, p. 4, Table 1; Mathieson et al., 2018, Supplement, Table 1). It is noteworthy that 27 fish teeth, including nine perforated ones, were found in burial H232 at Vlasac (Borić et al., 2014, p. 14, Table 3), which has been shown to have some EHG traits (Mathieson et al., 2018, Figure 1: D).