Continuation of fishing subsistence in the Ukrainian Neolithic: diet isotope studies at Yasinovatka, Dnieper Rapids

Abstract

Yasinovatka is one of around 30 prehistoric cemetery sites of fisher-hunter-foragers located along the Dnieper River in southern Ukraine. Dating to c. 5540–4930 cal BC, the skeletal remains at Yasinovatka suggest that around sixty-eight individuals were interred at the cemetery, during three broad phases of interment: A-type burials (c. 5540–4930 cal BC), Ƃ1 pit burials (c. 5550–4750 cal BC), and Ƃ2 pit burials (c. 4980–4460 cal BC). The burials are characterized, in part, by the inclusion of a number of Mariupol-type plates of boar tusk, in addition to deer tooth pendants, Unio shells, knife-like flint blades, Cyprinidae teeth, sherds of Neolithic pottery, and significant deposits of ochre in the later burial pits. Here we analyse δ¹³C and δ¹⁵N values for 50 human bone collagen samples from the site. The majority of the isotope results show a fisher-hunter-forager population reliant predominantly on freshwater aquatic proteins, which is in keeping with previous dietary isotope studies in the area. Two individuals however have δ¹⁵N values that are clearly depleted when compared with the main population; these reflect dietary protein intakes based on plant and animal terrestrial resources rather than the predominant focus on aquatic resources. Notably, the δ¹³C values of these anomalous individuals are not enriched compared with the fauna samples analysed from the region; this supports the possibility that they were incomers to the area, potentially from a nearby agrarian population.

Introduction and background

The precise chronology of the Dnieper Rapids region has received detailed review (Gokhman 1966, Telegin 1968, 1973, Telegin and Potekhina 1987, Telegin et al. 2002, also Biagi et al. 2007), and periodization in Ukraine has been refined with the integration of recent AMS radiocarbon dates (Lillie et al. 2009, 2017; Nikitin et al. 2010, 2017). Currently, phases of burial activity are represented for the Epipalaeolithic (c. 10,200–8000 cal BC), Mesolithic (c. 7300 cal BC), Neolithic (c. 5500–5000 cal BC), and Eneolithic periods (c. 4400 cal BC) (Lillie et al. 2009; Telegin et al. 2002), but, given the heterogeneity in terms of prehistoric activity in Ukraine, refining the exact boundaries of the periods remains a complex issue.

One of the key factors when studying social developments across the Late-Glacial into the middle part of the current Interglacial period, i.e. the Epipalaeolithic through to the Eneolithic periods in Ukraine, is how the Neolithic period is defined; and indeed this is the case in other areas of Eastern Europe. In the ‘Western’ school of thought, the debate surrounding the characterization of the Neolithic period is on-going (Pluciennik 1998, Zvelebil 1998, Thomas 2002, Anthony 2007, Fernández-Domínguez and Reynolds 2017); but the term is often associated with the beginnings of food production as opposed to food extraction (Price 2000; Anthony 2007) (amongst other variables such as pottery, sedentism, demographic shifts, evidence for increasing social complexity etc. (Barker 2009, Sherratt 2016)).

In Ukraine, the onset of the Neolithic period has often been defined by the introduction of pottery (although this occurs in purely Mesolithic contexts, with the Bug-Dniester culture of southwestern Ukraine having ceramics at c. 6200calBC (Anthony 2007:149), and it has its origins to the northeast in the Middle Volga River Valley where the Elshanka culture of the Samara region produced ceramics using the coiling method (ibid. 2007)), alongside diversification in lithic technologies, and a shift in observed burial rites (cf. Gronenborn 2003; Jacobs 1993; Lillie 1998; Telegin et al. 2003; Telegin and Titova 1993). The underlying assumption driving this characterization is that a change in technical culture reflects a change in social practices (cf. Motuzaite Matuzeviciute 2014:137). Recent research has shown that these changes are heterogeneous across Ukraine, and that the rate at which changes occur can be protracted (e.g. Anthony 2007). Notably though, irrespective of the socio-technical aspects, the characterization of Neolithic sites in Ukraine generally lacks the major socio-economic transformations of a food economy that transitions from the extraction of resources to the production of resources; a quantifier that plays a distinct role in the identification of Neolithic sites in North-western Europe.

In Ukraine, one of the dominant factors characterizing the research framework to date is the ubiquitous evidence for a continued reliance on aquatic proteins well into the Neolithic period. Whilst domesticates are integrated into Bug-Dniester subsistence strategies from the end of the VIIth millennium BC in southwestern Ukraine, the earliest reliable isotopic evidence for a reliance on terrestrial domesticates further east, in the Dnieper region, only occurs at the onset of the Eneolithic period, with the Trypillia farming culture, and in the Dnieper River region at the site of Molyukhov Bugor at c. 4000 calBC (Lillie et al. 2011).

This pattern of continued dependence on aquatic proteins into the Neolithic period is rarely observed in NW Europe, if observed at all (e.g. debate in Richards et al. 2003, Milner et al. 2004 and references therein), but, by contrast, it is seen elsewhere in Eastern Europe, and in the Baltic region there is clear evidence for continuity across the Mesolithic-Neolithic transition (e.g. Eriksson 2006; Eriksson et al. 2003). Elsewhere in the Eurasian steppe zone (in northeastern Kazakhstan) Svyatko et al. (2015) have documented the fact that a continued reliance on freshwater resources occurs into the Eneolithic and Bronze Age periods. One potential factor that might explain this subsistence choice in Ukraine is the possibility of a Black Sea deluge event at c. 7300-7200 BP (Ryan et al. 2003; Yanchilina et al. 2017); indeed the potential effects of this flood event on prehistoric populations is examined in other archaeological research (Turney and Brown 2007; Bikoulis 2015). (It should be noted that the occurrence of a catastrophic Black Sea flooding event remains widely debated in the academic literature (see Yanchilina et al. 2017, Goldberg et al. 2016). However, if a flooding event did occur, it is possible that land that hitherto would have provided a fertile environment for agrarian or pastoral practices was no longer viable (as the result of inundation and/or salinification processes, Yanchilina et al. 2017), thereby potentially creating increased competition for established aquatic resources.

Given that existing dietary studies show a continuation of the consumption of aquatic proteins, the core aims of this project are to characterize the nature of diet at Yasinovatka, and also to test whether there are any isotopic differences in human diet between the two (probable) cultural groups (the Surskaja A-type burials and the Dnieper Donets Mariupol-type; Ƃ-pit burials). Lillie (1998) has previously noted that the concentration of cemeteries in and around the Dnieper Rapids, at a point where a stable and predictable resource, such as migratory and other fish are easily harvested, may well reflect the establishment of ancestral rights of access to this food source.

Yasinovatka: Boar tusk plates, deer tooth pendants, and ochre burials

Yasinovatka is situated on the high left bank of the river Dnieper (Fig. 1). The stage Ƃ cemetery area is a sub-rectangular pit, which was highlighted during excavation by red staining associated with the ochre applied to Neolithic burials in the region (Fig. 2). The cemetery contained 68 individuals; 51 adults (36 males and 15 females), four adolescents and nine children (Telegin and Potekhina 1987). In general, individuals were interred in the extended supine position, however some variation occurs. A number of skeletons were severely contracted, which is indicative of swaddling/wrapping of the body prior to burial (Telegin and Potekhina 1987). Amongst the burial inventory from Yasinovatka are a number of Mariupol-type plates of boar tusk, deer tooth pendants, Unio shells, knife-like flint blades, Cyprinidae teeth, sherds of Neolithic pottery, and a poorly preserved bone awl. It is of some note that no pottery was found associated with the stratigraphically earlier A-type grave pits (Fig. 3).

Fig. 1

The Dnieper Rapids region, showing locations of cemeteries studied by Lillie (1998). 1—Dereivka I and II, 2—Vasilyevka V, III and II, 3—Nikolskoye, 4—Yasinovatka. ▲—Mesolithic, ∆—Neolithic

Fig. 2

Yasinovatka Cemetery Grave pit Б (1 and 2) (after Telegin and Potekhina 1987)

Fig. 3

Yasinovatka Cemetery plan (a) and profile (b). Legend: I—Stones; II & III—Outlines of Grave Pits; IV—Black Soil; V—Subsoil; VI—Pit Fill; VII—Loess (after Telegin and Potekhina 1987)

The cemetery at Yasinovatka has three phases of interment; A-type, Ƃ1, and Ƃ2 (Telegin and Potekhina 1987). The earliest phase of interment is attributed to the latest stage of the Neolithic Surskaja cultural group (c. 5100–5000 cal BC) (Kotova 2010), as the A-type phase is characterized by individual and paired oval burials pits, the presence of deer tooth pendants and fish teeth (Telegin and Potekhina 1987). Burials from the Ƃ-pit (Ƃ1 and Ƃ2) are attributed to the Dnieper Donets (Mariupol-type cemeteries) cultural group (c.7000–4000 cal BC) (Telegin et al. 2002); as the Ƃ-pit is characterized by collective burial (Fig. 3), the presence of pottery sherds and fragmented vessels, and the use of red ochre.

A notable burial at the site is individual number 45, an 18–25 year old male who was buried with 11 plates made from boar tusk enamel (Fig. 4). Telegin and Potekhina (1987:57) note that the position of the boar tusk plates in the grave suggests they were sewn onto the sleeves of a garment, with the enamel on the outside.

Fig. 4

Boar tusk plates that were probably sewn on to the clothing of individual 45 at Yasinovatka (after Telegin and Potekhina 1987:58, Fig. 30)

The artefactual and funerary differences observed between the stage A burials and the stage Ƃ burials (see Figs. 2 and 3; Table 1) indicate that there is a shift away from the earlier (essentially Mesolithic) fisher-hunter-forager mode of interment in individual or paired grave pits, towards a more communal mode of interment, which is perhaps aimed at reinforcing the group identity. Interestingly, this mode of interment is often associated more with farming societies, wherein it is interpreted as representing an attempt to reinforce social ties to the collective (Meyer et al. 2012, Pearson 1999).

Table 1 Burial data for phasing, culture affiliation, absolute dating and associated artefact inventories for Yasinovatka

As this shift is occurring at a time when increasing contacts with incoming agro-pastoralists is in evidence, we are potentially seeing a societal reaction against the new contacts and/or trade with neighbouring groups, or against the integration of immigrants from these new groups into the established fisher-hunter-forager communities in and around the Dnieper region (or a combination of these factors).

Interestingly, the stratigraphic evidence (Fig. 3) supports the probability that the Ƃ-pit cut through and disturbed the existing A-type burials, and during the excavations, the Ƃ burials were found to be in a greater state of disarticulation, perhaps suggesting a similar situation to that in evidence further afield in the chambered tombs of north-western Europe, where the dead are regularly repositioned within the tomb as new interments are made (Telegin and Potekhina 1987; Whittle et al. 2011; Salanova et al. 2018; Chambon et al. 2017). These observations suggest that between stages A to Ƃ1there is some degree of socio-political shift/re-orientation in the ritual articulation of the group. The reasons and meaning behind this is difficult to disentangle given the general continuities in material culture, although the appearance of ceramics, changes in lithic inventories and the use of ochre in association with the Ƃ group may be the key indicators of the vectors for change, in that it is the heightened contacts with incoming agro-pastoralists that stimulates a reaction in these groups.

However, in addition to this, it is entirely possible that whilst the different culture group nomenclature is a modern construct, these could actually be highlighting an even more substantial situation wherein the Surskaya group is actually replaced by the Dnieper-Donets culture group at the population level, and whilst the cemetery continues in use, the population that inters its dead in the Ƃ stages is in fact a completely new fisher-hunter-forager population that has moved into this region and displaced the indigenous/earlier groups. Obviously, at this stage of the analysis the above scenarios are somewhat hypothetical, but the application of DNA analyses to the cemeteries in this region could provide fundamental insights into population dynamics and shifts in the origins of the groups that are using these burial grounds.

Chronology and diet in the Dnieper Rapids region: radiocarbon dating and stable isotope studies

The precise nature of chronological developments in the region is complex, and a number of robust reviews are present in the literature (Telegin et al. 2002; Lillie et al. 2009). At Yasinovatka, the absolute chronology is based on 10 radiocarbon ages of skeletal collagen. The radiocarbon dates from the site initially place the earliest evidence for occupation at c. 6465 ± 60 BP (OxA-6163) (5476–5271 cal BC at 95%, using OxCal 4.2.2 and IntCal13), with subsequent layers dated to 6121 ± 34 BP (OxA-17,500) (5209–4959 cal BC at 95%, using OxCal 4.2.2 and IntCal13) (Fig. 5, modelled dates).

Fig. 5

Modelled radiocarbon ages of human and animal skeletal collagen from Yasinovatka (OxCal v4.3.2 (2017), 95%, IntCal13 (Reimer et al. 2013)). Source: Lillie 1998, Lillie et al. 2009

Existing radiocarbon dating at the site of Yasinovatka (and at proximal sites in the region), has produced evidence to support the presence of a radiocarbon freshwater reservoir effect; via the dating of human bone, red deer teeth, and fish teeth (Lillie et al. 2009). In oversimplified terms, a freshwater reservoir effect can cause an anomalous ¹⁴C offset in the radiocarbon age of a sample that is either from a freshwater context, or that has consumed freshwater resources; specifically it can produce a ¹⁴C age that is artificially ‘older’ than its true radiocarbon age (Ascough et al. 2005, Philippsen 2013). At Yasinovatka, the radiocarbon ages suggest a potential reservoir effect of c. 470 years (see Fig. 6), with surrounding Neolithic sites in the Dnieper Rapids region producing radiocarbon offsets of c. 100-400 years (Lillie et al. 2009:259–62), but the degree of offset is not systematic as the degree of freshwater resource consumption, as indicated by the δ¹⁵N values on these individuals, can indicate an offset closer to the terrestrial baseline where the values are lowered (i.e. values towards 11–12‰ will produce a limited offset in radiocarbon years), whilst elevated values can cause the degree of offset to increase to the 470 years attested by individual 54 at Yasinovatka. Therefore, any human individuals that produce elevated δ¹⁵N values should potentially be considered to have the reservoir effect, and any radiocarbon ages on these individuals should be examined carefully.

Fig. 6

Calibrated radiocarbon ages, showing a reservoir effect for burial 54 at Yasinovatka (OxCal v4.3.2 (2017), 95%, IntCal13 (Reimer et al. 2013)). Source: Lillie et al. (2009:261)

To date, c. 170 human, fauna, and fish bone collagen (or dentine) samples have been analysed for diet isotope studies from the Epipalaeolithic to Eneolithic cemeteries in the Dnieper region (Lillie and Richards 2000; Lillie et al. 2003, 2011; Lillie and Jacobs 2006). The prevailing conclusion of these studies demonstrates variability in access to dietary proteins across these periods, with the notable factor that the consumption of aquatic proteins is distinct during every phase in the study period (Lillie 1998; Lillie and Richards 2000; Lillie et al. 2003; Lillie and Jacobs 2006; Lillie and Budd 2011; Lillie et al. 2011). Previous dietary isotope studies into the human populations at Yasinovatka highlighted depleted δ¹³C values − 22.4 ± 0.3‰, and elevated δ¹⁵N values 14.2 ± 0.3‰, reflecting a diet that incorporates significant quantities of aquatic proteins. The human samples from Yasinovatka have produced the most enriched δ¹⁵N values of all sites across the region; being commensurate with Vasilyevka II (Mesolithic), and Nikolskoye (Neolithic).

There are only limited fauna samples from prehistoric sites in the Dnieper Rapids region, which limits the material available for isotope analysis (Lillie et al. 2011). Fewer than twenty wild or domesticated fauna samples have been isotopically analysed from the sites in the study region; and of these samples, only four originate from the sites clustered around Yasinovatka (red deer and Bovidae sp.). At − 19.5 ± 0.6‰ the δ¹³C values suggest a terrestrial C₃ plant signal of c. -20.5 ± 1‰, and the δ¹⁵N values at 6.9 ± 0.8‰ (Lillie et al. 2011), indicating a terrestrial plant signal of anything from c. 1–4‰, assuming a trophic enrichment of 5–6‰ (Hedges and Reynard 2007; O'Connell et al. 2012; Fernandes et al. 2014). Whilst this dataset is admittedly limited, if humans at Yasinovatka or proximal sites were consuming a combination of terrestrial plants and fauna we would anticipate δ¹³C values of c. -19 ± 0.5‰, and δ¹⁵N values of c. 10 ± 1.5‰.

Methods: stable isotope analysis (δ¹³C and δ¹⁵N)

Dietary stable isotope analysis is a method now routinely applied in archaeological studies to provide direct evidence for human and animal diets. Many robust reviews and critiques of the methodology exist in the literature (e.g. Hedges and Reynard 2007; Makarewicz and Sealy 2015; Styring et al. 2015), and therefore extensive details of the principles of the method and its application are not repeated here. Briefly, the analysis of carbon and nitrogen ratios from bone collagen allows for a direct assessment of prehistoric dietary pathways, particularly in the last c. 10 years of the individual’s life (Schwarcz and Schoeninger 1991). Stable carbon isotope ratios (δ¹³C) measured from bone collagen can allow us to distinguish between dietary protein from marine, terrestrial and, under some circumstances, freshwater resources (Schwarcz and Schoeninger 1991; Cerling et al. 1997; Richards 2002). In addition, δ¹³C values are influenced by the composition of the local vegetation.

Nitrogen stable isotope ratios (δ¹⁵N) are used to establish the trophic level of an organism in the food web, with δ¹⁵N enrichment of c. 3–6‰ (e.g. Δ¹⁵N_diet-body) observed as one progresses up the food chain (Schoeninger and DeNiro 1984; Minagawa and Wada 1984; Hedges and Reynard 2007). The isotope measurements of human and animal bone collagen are biased towards the protein component of ingested foods, which can be used to infer potential sources of dietary carbon, alongside providing information regarding ecological niches, vegetation patterning (C₃ vs. C₄ species present) and habitat (Schoeninger and DeNiro 1984; Lee-Thorp and Van der Merwe 1987).

In this study, due to the location of the cemetery and existing dietary isotope studies of sites in the region (Lillie and Richards 2000; Lillie et al. 2011) we anticipate the inclusion of freshwater aquatic protein in the diets of the humans studied. The investigation of aquatic proteins in human diet is often more complex than those related to terrestrial resources, and carbon and nitrogen isotope values from bone collagen with an aquatic signal vary considerably between sites (Katzenberg and Weber 1999; Lillie et al. 2011; Svyatko et al. 2016; Robson et al. 2016; Drucker et al. 2018). In simple terms, bone collagen values from freshwater aquatic species often exhibit δ¹³C ratios that are more depleted than terrestrial animal bone collagen. Robson et al. (2016) analysed bone collagen of freshwater and brackish species (pike, perch, and zander), which produced δ¹³C of − 24.2 to − 19.3‰, and δ¹⁵N values of 5 to 10‰. Nitrogen isotope values from freshwater species tend to be more elevated than their terrestrial counterparts (freshwater fish collagen at the Iron Gates produces δ¹⁵N values of 10 ± 1‰ (Bonsall et al. 2004)), due to the longer length of the freshwater foodweb. However, some aquatic species, particularly cyprinids, produce noticeably lower bone collagen δ¹⁵N values, c. 6 ± 1‰ (Naito et al. 2013; Schmölcke et al. 2015).

Materials

Fifty human bone collagen samples were analysed at the Light Stable Isotope Laboratory at Oxford University. Bone collagen was extracted using a modified version of the Longin (1971) method, using acid and H₂O (MilliQ) washes, freeze-drying and isotope measurement on an Elemental Analyser linked to a continuous flow Sercon dual inlet mass spectrometer. Known-value standards are included in order to calibrate (and drift correct) the readings taken from the unknown (bone) samples. Isotopic ratios are calculated with reference to in-house standards, which for this project included alanine (information regarding other in-house standards, such IAEA standards, were not provided by the laboratory). δ¹³C measurements are reported on the VPDB scale and δ¹⁵N is reported with reference to AIR.

Quality control parameters are used in stable isotope analysis to ensure the accuracy of the data produced. These include a collagen yield of 1% at 5 mg or above, and a C:N ratio of between 2.9 and 3.6 (DeNiro and Epstein 1978; DeNiro 1985; Brock et al. 2010). The %C and %N of the collagen samples is also calculated as this provides a reliable indicator of sample preservation (Ambrose 1990; Van Klinken 1999; Sealy et al. 2014). For bone collagen, the ideal range for %C is c. 40–48%, and for %N is approximately c. 12–17% – it should be noted these ranges are approximate, with individual studies (e.g. Ambrose 1990; Van Klinken 1999; Sealy et al. 2014) producing slightly different ranges. The purpose of these parameters is to ensure that sufficient organic material (e.g. mainly collagenous protein) was preserved, and that this material retains an acceptable in vivo signal (as assessed by the C:N and collagen yield). Statistical analyses were performed in the open source statistics package R (http://www.r-project.org/).

Stable isotope results

The δ¹³C and δ¹⁵N values are presented in Table 2. Of the 50 samples processed, 21 failed the quality control parameters, with 14 failing due to low or no collagen yield, and 7 failing due to C:N ratios outside of the boundary 2:9–3:6 (DeNiro and Epstein 1978; Brock et al. 2010). These failed samples are listed below, but not included in statistical testing or interpretation. Table 2 also details human δ¹³C and δ¹⁵N values from previous dietary studies at Yasinovatka (Lillie and Richards 2000:966, Lillie et al. 2011:62). These isotope values are integrated into the analysis below.

Table 2 δ¹³C and δ¹⁵N values of human bone collagen samples at Yasinovatka, where ‘–‘denotes that no collagen yield was achieved. Results in italics = C:N ratios outside of the 2.9-3.6 boundary

The majority of the δ¹³C and δ¹⁵N results, at − 22.9 ± 0.9 and 13.8 ± 1.5‰, clearly demonstrate that the majority of the individuals at Yasinovatka were reliant on aquatic protein as the mainstay of their diet (Fig. 7).

Fig. 7

Individual human, faunal, and fish δ¹³C and δ¹⁵N ratios for Yasinovatka, alongside averaged human and fauna δ¹³C and δ¹⁵N ratios from the nearby contemporary sites of Dereivka, Nikolskoye, Vasilyevka V, Vyasivok, and the Crimean site of Fat’ma Koba. Note: this graph includes human δ¹³C and δ¹⁵N data from Yasinovatka from a previous study (Lillie and Richards 2000:966, 2011: 62)

Two notable exceptions occur, burial 45 who produced a δ¹³C value of − 23.6‰ and a δ¹⁵N value of 11.4‰, and burial 19, who produced a δ¹³C value of − 22.4‰ and a δ¹⁵N value of 7.4‰. The δ¹⁵N values of these individuals strongly support the consumption of C₃ terrestrial plants and animals, with burial 19 (female, aged 20–25) most likely consuming less animal protein than burial 45 (male, aged 18–25), as evidenced by her lower δ¹⁵N value. Interestingly, their δ¹³C values are noticeably depleted compared with the (admittedly limited) fauna samples analysed from this region. This lends support to the possibility that these individuals were not consuming terrestrial plants and animals from this study region.

Statistical testing

The existing published isotope data from Yasinovatka (Lillie and Richards 2000; Lillie et al. 2011; Webster 1996) is integrated into the new dataset for statistical testing. Mann Whitney U tests (at alpha = 0.05) were undertaken to investigate possible differences in δ¹³C and δ¹⁵N values between the A-type and Б-pit burials (Mann Whitney U is used as the residuals of the data are non-normally distributed). The difference in δ¹³C values between the A-type and Б pit burials is not significant at p = 0.238, neither is the difference in δ¹⁵N values at p = 0.352. Of the 38 individuals that produced reliable isotope data, only 16 are securely sexed, and of those only three are identified as female. As a result, it is not possible to investigate dietary isotope difference by sex.

Of the A-type Surskaja burials, only select individuals were interred with grave goods, which consisted mainly of deer tooth pendants and fish teeth (Table 1). The difference in δ¹³C values between A-type individuals with grave goods and A-type individuals without, is not significant at p = 0.984, neither is the difference in δ¹⁵N values p = 0.103. Finally, individuals with grave goods and those without were tested without reference to A-type or Б pit burials, and no difference is evident (with δ¹³C at p = 0.787, and δ¹⁵N at p = 0.158). Tukey’s outlier method was applied to the δ¹³C and δ¹⁵N dataset, and identified two individuals with divergent δ¹⁵N values - burial 45 at 11.4‰ and burial 19 at 7.4‰.

Discussion

Overall, the most striking conclusion evident from the human δ¹³C and δ¹⁵N dataset is that it demonstrates a clear, continued, and heavy reliance on aquatic proteins into the Neolithic period, regardless of whether individuals show burial affinities with the Surskaja or Dnieper Donets cultural groups. Interestingly, two individuals stand out as being isotopically distinct in terms of their δ¹⁵N values from the rest of the cemetery population, a young female found in close proximity to a deer tooth pendent (burial 19, δ¹⁵N = 7.4‰), and the highly decorated young male with boar tusk plates sewn onto his clothing (burial 45, δ¹⁵N = 11.4‰).

The isotope values for burial 19 could conceivably reflect the consumption of a plant dominated diet, as both her carbon and nitrogen isotopes mirror the fauna values measured at the Neolithic and Eneolithic phases at Molyukhov Bugor (Lillie et al. 2011). To date no other human sample in the study region has produced comparatively depleted δ¹⁵N values (Lillie et al. 2011), although her value is in keeping with agragrian communities such as Trypillia (e.g. Lillie et al. 2017). Burial 19 has previously been analysed for dietary isotope analysis, and the original depleted δ¹⁵N value led the authors to suggest the possibility of a misidentified faunal sample (Lillie et al. 2011; 64). For this study the bone sample was obtained directly from the individual’s skull, and therefore the potential for misidentification is avoided.

The other outlier, burial 45, is slightly more complex. His δ¹⁵N value is easily in-keeping with a diet based on terrestrial animal protein with some contribution from terrestrial plants, but the δ¹³C value is distinctly depleted compared with the fauna analysed from the site. One possibility is that individual was consuming a combination of terrestrial fauna and selected aquatic remains from the region. The boar tusk plates that were sewn onto this individuals clothing may represent signifiers of his success as a hunter, and as such, provide an indication of at least one of the primary elements of his diet?

Support for the suggestion that aquatic resources may well have been important in the diet of individual 45 is found in the fish samples from the Neolithic sites of Dereivka (n = 1) and Vil’nyanka (n = 1), which have δ¹³C values of − 25.5 and − 26‰, and δ¹⁵N values of 10.5 and 10‰ respectively, i.e. values that would be commensurate with a limited contribution of aquatic proteins for individual 45. However, an alternative possibility is that the depleted δ¹³C values represent hunting and foraging of terrestrial species in woodland environments. For example, La Villette, Cuiry, Maizy, and Bercy are Neolithic sites located in close proximity to the River Seine within the Paris basin whose terrestrial wild and domesticate species have produced notably depleted δ¹³C values. Red and roe deer species have δ¹³C values of c. -24.7‰, aurochs c. -24‰, and Bos sp. of c. -23.2‰ (see Goude and Fontugne 2016); demonstrating depleted terrestrial δ¹³C values are possible alongside freshwater/brackish river systems.

At the nearby Neolithic site of Dereivka (Fig. 1), two male individuals produced δ¹⁵N values that were similarly divergent when compared with the majority of the cemetery population (with δ¹⁵N values at c.10.5 and 9.9‰, compared with a site average of 12 ± 1‰, see Lillie et al. 2003, 2011). Burial 49, the individual with the lowest δ¹⁵N value at 9.9‰, also represents the richest burial at Dereivka, and was interred with deer and fish tooth pendents, alongside a Mariupol plate that shows a distinct artefactual affinity with the 11 boar tusk ornaments associated with burial 45 at Yasinovatka.

As the majority of the adults analysed at Yasinovatka exhibit elevated δ¹⁵N values, we can assess the presence of a reservoir effect given that earlier radiocarbon studies have shown the existence of an offset from one of the Yasinovatka burials (Lillie et al. 2009:261). For burial 54, with a δ¹⁵N value of 14.1‰, we observe a chronological difference of c. 470 years between the individual (human) and the deer tooth pendent that was dated. Several studies have investigated the relationship between dietary offsets and radiocarbon ages, variably applying regression and Bayesian modelling to calculate ¹⁴C offsets using δ¹⁵N values (Naito et al. 2010, Ramsey et al. 2014, Schulting et al. 2014, Sayle et al. 2016). At Yasinovatka, whilst investigations into the reservoir effect are currently limited to one burial, we can tentatively use the observed difference of c.470 years as a primary baseline for comparison. Assuming a very simple linear step-wise regression correction factor, we can use the δ¹⁵N value of burial 54 (e.g. 14.1‰) as the upper end point for fish consumption, which has a corresponding ¹⁴C offset of 470 years, and assume a δ¹⁵N value of 10‰ to reflect the consumption of C₃ terrestrial proteins only (and therefore no reservoir offset). Using this basic approach, we can speculate ¹⁴C offsets of between c.240–470 years for individuals with δ¹⁵N values of 12 to 14‰. Interestingly, if we apply this ¹⁴C correction factor to the dated individuals that have elevated δ¹⁵N values, the oldest burials at the site are individual 45 (boar tusk plate burial) and individual 19. Both of these individuals exhibit δ¹³C and δ¹⁵N values that are in-keeping with a C₃ terrestrial diet.

It is, of course, difficult to accurately evaluate whether these individuals do in fact represent the oldest burials at the site, as research into the ¹⁴C reservoir offset at Yasinovatka is limited, and the ¹⁴C offset as applied here is a basic linear model. Obviously, the data indicates that an extended research program is necessary if we are to further disentangle the freshwater radiocarbon reservoir effect at Yasinovatka in more detail and with greater accuracy.

Overall, the combined research to date strongly supports the presence of fishing, hunting, and foraging/gathering activities in the Dnieper Rapids region from the Epipalaeolithic sites well into the Neolithic period. The degree to which these resources were exploited was probably dependant on a number of factors, including individual dietary preference, but also socio-economic factors, and potentially, political factors e.g. Rouja (1998). Ancient DNA analysis and craniometrics from individuals in the region support the possibility that human populations occupying the Dnieper Rapids were not necessarily indigenous to the region, with the presence of non-locals present from the Mesolithic period onwards (Lillie and Budd 2011, Lillie et al. 2012, Mathieson et al. 2018).

A preliminary, and indeed tentative suggestion, is that the individuals with comparatively depleted δ¹⁵N values and associated decorative Mariupol-type boar tusk plates, perhaps represent the presence of non-local human groups either imposing themselves on the indigenous communities of the Dnieper Rapids region, or incomers to the region who represent marriage partners from other groups who would have cemented ties and rights of access to the rich resource base of the Dnieper Rapids region. Given the lack of evidence for inter-personal violence and the nature of the burials afforded to these individuals the latter scenario may well be the more realistic proposition at this stage. The re-use of the cemetery site at Yasinovatka, potentially with burial phases from at least two distinct cultural groups, supports the idea that access to the resources in this region was considered important across the Mesolithic and Neolithic periods, with the superimposition of the later cultures burials asserting ancestral legitimisation and also mirroring similar activities throughout prehistoric Europe (e.g. Zvelebil 1993; Lynch 2000).

Conclusion

At Yasinovatka the dietary isotope studies demonstrate that the majority of the population had a strong reliance on aquatic protein for the mainstay of their diet; thus highlighting the continued importance of fishing as a mainstay of subsistence strategies into the Neolithic period in the Dnieper Rapids region of Ukraine. The presence of two anomalous δ¹⁵N values in the dataset is of interest, particularly as individual 45 shows close artefactual and isotopic affinities with similarly isotopically divergent individuals studied at the site of Dereivka. The lower δ¹⁵N values for individual 19 at Yasinovatka could be indicating that some degree of population movement between the groups exploiting the region around the Dnieper is occurring across the Mesolithic to earlier Neolithic periods.

It is important to note that fishing was pivotal in terms of subsistence strategies from 10,200calBC onwards in this region, and that despite assertions that climate influenced the timing and adoption of agriculture in Ukraine, regional variation and asynchronicity must be taken into account. In the Dnieper Region we see continuity in fisher-hunter-forager lifeways through to at least c. 3500calBC, when areas of Ukraine to the west of the Dnieper have already developed some degree of pastoralism, with associated agricultural activity.

The fact that the major river valleys offered indigenous groups the opportunity to exploit a reliable suite of fish, animal and plant resources for in excess of 6500 years mirrors the exceptional situation occurring in the Baltic region, and provides an opportunity for long-term studies of food extraction societies at a time when farming is seen to be becoming the mainstay of subsistence in Neolithic Europe. The specific environmental context and significant areal extent of Ukraine in terms of its north-south and west-east climate, ecological and human activity zones offers a unique context for the study of prehistoric societies in the European mainland.