Introduction

When ancient metal artefacts (copper, zinc, tin, lead, iron etc.) are incorporated in archaeological deposits and exposed to fluctuating humidity and temperatures, they develop a thin layer of corrosion by-products called patina, i.e. metal salt products on their outermost surface. The chemical processes responsible for the formation of these products are well differentiated, through the function of exogenous (i.e. environmental conditions, soil reaction, humidity, attachable minerals) and endogenous (alloy elaboration, the original minerals and the metallurgical technology) agents (Neff et al. 2005; Sandu et al. 2008, 2012; Keheyan and Lanterna 2021). In marine environments or in sediments located near the sea, the conditions are much more aggressive due to the dissolved salts, predominantly sodium (Na+) and chloride (Cl−) ions, interaction with the metal. Those processes significantly affect the corrosion of copper (Scott 2000) and iron (Selwyn et al. 1999; Moore 2015). By-products of metal oxidisation have the capacity to preserve various macroscopically visual remains of plants (incl. wood), their products, textiles, insects, and other organic matter (Shishlina et al. 2003; Petrova 2006; Haneca et al. 2012) once the corrosion proceeds more rapidly than the biological deterioration (Watson 1988). When all those categories of organic remains are exposed to corroding metal, the material of organic origin becomes embedded in the corrosion products. Copper is especially conducive to this process due to its toxicity protecting organic matter from being decayed bacterially (Moret 1998; Peška et al. 2006; Robbiola et al. 2011). In some cases, changing humidity conditions catalyse the replacement of original organic matter with metal salts, i.e. fossilisation by mineralisation (Jacomet 2007). Sometimes still recognisable elements of plant microstructure can be observed and can be used for identification. The surrounding environment also contains microscopic remains (i.e. pollen), which can be trapped within the patina of artefacts and thus preserved (Ganor et al. 2009).

In archaeological sites with dry soil conditions, the micro- and macrofossils captured in the by-products of metal corrosion allow tracing of organic products, despite the poor organic preservation. The detailed analysis of organic remains from such contexts may be an important source for the further archaeological interpretation of the preserved artefacts and the overall state of preservation of the sites (van Os et al. 2012). In settlement contexts with dry soil conditions, the chance of finding plant remains preserved through charring and/or mineralization is much greater than in funerary contexts with the same site conditions, unless the remains were the subject of cremation (Šoštarić et al. 2016). Therefore, the analysis of corrosion by-products with inclusions of ancient organic matter can significantly contribute to the understanding of the burial ritual and can give information on the surrounding environment, where no other organic materials are preserved. Different aspects of this topic have already been targeted by archaeological science: pollen as part of honey beverages (Rösch 1999); various microscopic remains (Quinn 2008; Ganor et al. 2009); macroscopically observable imprints, textiles, animal, and plant remains (Peška et al. 2006; Kozáková et al. 2019; Andonova et al. 2023), as well as mineralized wood, identifiable through X-Ray sub-micron tomography (Haneca et al. 2012).

The current paper presents the results of pilot archaeobotanical analyses of two ancient cemeteries from the Black Sea coast of Bulgaria, where thirty-seven graves were considered (Tables 1 and 2), with the aim of discussing the potential and limitations of these studies for the archaeological and archaeobotanical research.

Table 1 Varna cemetery, archaeobotanical results. No pollen preservation in sampled artefacts 2429, 2684, 1048, 2658, 2673, 2459, 2635, 2612, 2636. NPP=non-pollen palynomorphs
Full size table

Study area and archaeological context

Varna cemetery

The cemetery lies to the west of the city of Varna, approx. 400 m north of modern Varna Lake, which in the past was a bay of the Black Sea (Fig. 1). The cemetery dates to the Chalcolithic period, starting at 4590 cal bc and ending at 4340 cal bc (Krauß et al. 2017; Higham et al. 2018). The site covers 7,500 m2, with 315 graves excavated. It is one of the most remarkable sites in the European prehistory, holding the world’s earliest large-scale assemblage of gold artefacts (Ivanov 1988). In addition to the gold-work, the grave goods included numerous copper and flint artefacts, as well as beads made of different minerals, marble vessels and figurines, implements of bone and antler, over 12,000 Dentalium shells and about 1,100 imported Spondylus shell ornaments, i.e. bracelets, necklaces, and appliqués (Slavchev 2010). The numerous copper artefacts from the Varna cemetery show traces of patina and in some of them visible preservation of plant tissue.

Fig. 1
figure 1

Study area: a indicated within Europe as a rectangle and b within south-eastern Europe as a rectangle. c region of the Bulgarian Black Sea coast with the study sites. (Google Earth accessed on 22 November 2022)

Full size image

Messambria necropolis

Messambria, nowadays Nessebar, was a Greek colony founded at the end of the 6th/beginning of the 5th century bc (Preshlenov 2003). It is situated on the Nessebar peninsula of the Black Sea coast (Fig. 1). In the past, the peninsula was considerably bigger, but sea transgressions have decreased its surface area significantly (Gyuzelev 2008).

The necropolis of Messambria lies to the west of the ancient town, on the mainland immediately west of the peninsula isthmus. It has an area of about 15,000 m2, with more than 1,300 graves excavated, covering the time span from the 6th century bc to the 19th century ad (Bozkova et al. 2011). Depending on the time epoch, grave structures are comprised of pits with or without stones, graves with bricks or tiles, and cist graves. Offerings vary according to the specifics for the period and the social status of the deceased. Some graves contained scarce mortuary gifts while others conveyed extreme prestige yielding archaeologically rich material (Bozkova et al. 2011).

Materials and methods

The plant remains, subject of this study, were preserved through processes of metal corrosion of copper, bronze, or iron (Fig. 2), mostly by coating or replacement of the original organic matter by metal oxides and/or salts; in a few cases, imprints of the original material were observed in the corrosion layer. Observation and sampling took place under a stereo microscope (up to 64×). Macrofossils were analysed under a reflected light microscope (up to 400×) and selected samples were observed under a scanning electron microscope (SEM).

Fig. 2
figure 2

Organic remains from the studied structures; a–c Viburnum wood, d–e plant fibre textile, f–g Juncus epidermis fragment, h almonds on a strigilla (from Hristova 2015, p 126, Fig. 4f), i cluster of Vitis pollen

Full size image

After removing all visible sediments, pollen samples were taken from surfaces with inclusions of organic matter or spots with a visibly thicker layer of patina. Before sampling the copper/bronze artefacts, the upper ca. 0.5–1 mm of patina was scraped away with a clean scalpel. Similarly, in the case of the iron artefacts (nails and strigillas) the upper brittle corrosion layer was removed. The purpose of this step was the minimisation of possible contamination. Before pollen extraction by standard acetolysis (Faegri and Iversen 1989), the samples were rinsed twice in distilled water. Pollen samples coming from copper/bronze patina were usually of very low volume (Table 1), due to the limited patina surfaces on those metals. Therefore, the total processed amount of each sample was analysed. From iron corrosion samples, at least seven slides (24 × 32 mm coverslip) were counted. The percentages of each pollen taxon were calculated based on the total pollen sum of all samples containing pollen. This is possible because the studied structures are from the same historical period and summarized results from all graves can be successfully used for palaeoecological interpretations (Tables 1 and 2; Fig. 3). The pollen sum consists of arboreal (AP) and nonarboreal (NAP, herbs) pollen.

Table 2 Messambria necropolis, archaeobotanical results. No pollen preservation in sampled artefacts from graves 966, 977, 1078, 1083, 1189, 1266, 1295
Full size table
Fig. 3
figure 3

Ecological interpretation of the palynological assemblages from a Varna cemetery, b Messambria necropolis (AP = arboreal pollen, NAP = non-arboreal pollen). The datasets are given in Tables 1 and 2

Full size image

Results

Plant remains from Varna cemetery

Macroremains

Twelve artefacts yielded macrobotanical finds (Table 1). Wood was mostly retrieved from handle openings of copper implements. The dominant identifiable wood was Cornus sp., but in one artefact Viburnum wood was also determined (Fig. 2a–c). In some cases, excellent preservation of textile and singe fibres of plant origin were observed on the same studied copper tools (Fig. 2d–e); no further identifications of those fibres were possible, as no additional plant tissues other than the bast fibres were found (see Bergfjord et al. 2010). In artefact 2373 an imprint of Juncus sp. epidermis was recorded (Fig. 2f–g).

Pollen

Of 18 samples studied, pollen occurred in nine. The analysis revealed a total number of 33 pollen taxa: 14 from trees and shrubs, and 19 of NAP. The pollen of grasses and herbs is dominant—mainly representatives of Asteraceae such as Crepis/Taraxacum-type, Artemisia, Matricaria/Achillea-type, Centaurea jacea-type, but also Chenopodiaceae and Rosaceae. In the AP group, the pollen of Quercus, Carpinus betulus, Corylus, Fraxinus excelsior-type and Tilia prevails, while Pinus, Ulmus, Acer, Vitis and Viburnum have minor importance. Pollen tetrads of Daphne pollen occurred in one of the samples (Table 1).

Plant remains from Messambria necropolis

Macroremains

Macroremains preserved in by-products of metal corrosion were found in seven Hellenistic (end of the 4th–1st century bc) graves: four cists and three burial pits (Table 2). Wood preserved in connection with fragments of bronze funeral wreaths was identified in three graves (1189, 1266, 1285) as pine (Pinus sp.), while in two others (1106, 135) as coniferous wood.

The almond fruits (Amygdalus communis, Fig. 2h) were found in two cist graves (878, 1091) and are preserved in a mineralized state. They were arranged on the iron strigillas as part of the grave gifts. On a wreath in grave 1106 and 1189, parts of thread made from plant bast fibres were preserved.

Pollen

Sixteen samples were analysed, and pollen was preserved in nine of them coming from seven burials (Table 2). In total 44 pollen taxa were identified—14 trees and shrubs and 30 taxa of herbaceous plants. Quercus, Carpinus betulus and Corylus dominate the AP, while Carpinus orientalis, Acer, Fraxinus excelsior-type, F. ornus, Ulmus, Pinus, and Betula show minor significance. Main herbaceous pollen types are Poaceae, Chenopodiaceae, and Brassicaceae. Some weeds and ruderals including Plantago, Rumex, Urtica, Centaurea cyanus, Cirsium and Polygonum aviculare were also identified, as well as Cerealia-type pollen, being representative of cereal crops. Groups of pollen grains, which were still attached in tetrads, were established—Vitis (Fig. 2i), Rosaceae, Trifolium, and Cistus. Remarkably, grave 1116 had no visible organic preservation, i.e. the skeleton was completely decomposed, but pollen was retrieved from iron nails. In grave 1106 in addition to iron nails, pollen was attached to wooden fragments and thread from a bronze wreath.

Discussion

Preservation of the plant material

Macrofossils

In both study sites, wood plant remains were associated mostly with copper or bronze patina. Most of the textiles were also preserved due to the contact with the toxic copper/bronze. Observed imprints were rarely identifiable, except for a Juncus epidermis fragment. Wood tissues were usually completely replaced by the metal salts (Fig. 2b). For many wood fragments, smaller than 2 mm, it was possible to recognize the structure of wood, but no precise identification was feasible. Almond fruits arranged on the strigillas were mineralised thanks to contact with the iron surface (Fig. 2h).

Pollen

At both study sites, a certain portion of the analysed metal corrosion samples contained no preserved pollen (Tables 1 and 2). In both sites, pollen concentration was rather low and the preservation very variable. An obvious reason for the low pollen concentration is the extremely small volumes of the samples taken from the patina due to the nature of the corrosion crusts (ca. 0.1 g in average). On the other hand, cleaning the outer layer of the metal artefacts led to removal of certain parts of the corrosion surfaces and eventually of the pollen grains embedded therein. The procedure for sample processing, acetolysis, and several washings with distilled water also results in loss of pollen. Through the varying embedding conditions at each site, the formation of patina and its structure also differs, which leads to varying pollen preservation. Along with copper/bronze, iron artefacts also yielded preserved pollen. The iron artefacts form a layer of corrosion by-products, made of iron oxides, oxy-hydroxides, chlorides and/or carbonates, which is well crystallised and more compact compared to the more external zones of the corrosion system. These compact layers form the so-called “dense product layer” (Neff et al. 2005), and they probably are more suitable for preserving the pollen grains and other microscopic particles.

Reconstructing the burial ritual

Varna cemetery

The presence of Daphne pollen in one of the graves in Varna cemetery is significant. Considering the pollination biology of the genus, it is unlikely that the pollen would accidentally get into the studied burial. Most probably, the plant was an intentional element of the ritual, but whether it was used because of its aesthetic value, particular symbolic meaning, or its chemical properties (Moshiashvili et al. 2020), is hard to define. As the flowering season of Daphne is April–June (Kozhuharov 1992), spring to early summer could be considered as the most probable time of the burial, if we assume that fresh flowers were used. The main wood used for the handles of the copper axes was that of Cornus sp., most probably of Cornus mas considering the wood properties suitable for handles of implements (Gale and Cutler 2000). An exception is one artefact where the rare and very durable wood of Viburnum sp. was used. This wood has very good properties for handles and general tools, and is well known as being utilised in prehistory, even if it rather rarely occurs in the forest vegetation. The identified imprints of Juncus sp. (Fig. 2f–g) point to matting or basketry on which the metal artefact was laid (for recent discussion of Juncus uses see Andonova et al. 2023).

On the surface of some artefacts, remains of textiles of plant bast fibres were observed (Fig. 2d–e). Different plant sources could have been used to produce these fibres, such as flax and nettle. Although no further identification was possible, their presence indicates the importance of plants for the production of yarn and textiles. Here they either represent fabric or a bag in which grave goods were wrapped, or some other type of funeral textile.

Messambria necropolis

The presence of almonds arranged on strigillas (Fig. 2h) suggests their significance for the rite (Hristova 2015). There are available data for the presence of almond at the necropolis of another nearby Greek colony of the same period named Apollonia (Popova 2011). Its presence could be related to the Greek influence on the West Black Sea coast during Classical and Hellenistic times and eventually the introduction of this plant into cultivation in the study region.

The clusters of pollen of Vitis, Cistus, Rosaceae and Trifolium identified in some of the graves indicate the use of flowers, or parts of those plants in the burial ritual. Finding clusters of pollen grains can be explained only by the presence of whole flowers during the funerals (Clarke and Hamilton 2000). Additionally, their presence gives some hints on the possible season when the burial occurred. Considering the flowering season this could be defined broadly as being between May and August (Kozhuharov 1992) in the cases where fresh flowers or branches were involved in the ritual. However, it cannot be excluded that dry plants and plant products played a role in these activities. Being that vine flowers are not typically valued for their aesthetic quality, it is more likely that twigs were used in the funeral ritual, which led to the concentration of pollen preserved in the grave. Vitis pollen is an element of the pollen diagrams from the Black Sea area, but in low percentages (Tonkov et al. 2011). The Vitis finds in the burial structures may have held symbolic meaning. Vine branches occur in some sites like Chatalka, Dragodan, and Jhitosvyat (Bujukliev 1986; Theodossiev and Manov 1993; Popova 2002), which were associated with the cult of Dionysus. Such branches were a symbol of the chthonic and vegetative functions of the Thracian Heros (Horseman) (Theodossiev and Manov 1993). Evidence for grapevine in graves is known also from other parts of Europe like in Zahrádka, Czech Republic (Šálková et al. 2015), Italy (Rottoli and Castiglioni 2011), France (Preiss et al. 2005), etc.

Funeral wreaths are typical findings in the Hellenistic period graves. The wood fragments found, deriving from these funeral wreaths, belong to pine and conifers. This type of wood was probably intentionally used for the funeral wreaths, especially considering the fact that pine was believed to be apotropaic and symbol of immortality (Gale and Cutler 2000).

Evidence of the past environment in the burial place

Varna cemetery

The pollen data indicate xerophytic and halophytic grass communities of Poaceae, Asteraceae (Artemisia, Crepis/Taraxacum-type, Matricaria/Achillea-type, Centaurea), and Chenopodiaceae in the surrounding areas. Mixed oak forests, including Carpinus betulus, Tilia, Corylus, Ulmus and other deciduous trees, grew on the surrounding hills. Human activities in the region are reflected by cereal pollen, including Triticum, and anthropogenic indicators such as Plantago, Urtica, Centaurea cyanus, and Polygonum aviculare. Cereal pollen could also be found attached to cereal grains. The cereal pollen may therefore also point towards the use of grains as grave goods. Filipova-Marinova et al. (2014) noted similar results for the Late Chalcolithic in the area of Varna Lake—degradation of the mixed oak forests and significant pollen percentages of cultivated plants.

Messambria necropolis

The pollen analysis of Messambria necropolis shows the existence of oak forest with the presence of hazel, hornbeam, elm, and other deciduous trees in the close vicinity during the Hellenistic period. The inclusion of pollen and wood of pine in the samples suggests there were likely pine stands on hilly areas nearby. Looking at the potential natural vegetation in the study area it seems that some relict of forests of black pine can be found to the south of the peninsula, on the Black Sea coast. It cannot be excluded that similar habitats existed closer to Messambria in the past and became extinct due to increasing human impact. However, pine pollen can be subjected to distant transport, especially in open landscape exposed to long-term human impact like the one around Messambria. Also, coniferous wood can be easily imported, particularly in a harbour area where it can be brought in on ships. Dominant herbaceous pollen taxa, i.e. Poaceae, Chenopodiaceae and several representatives of Asteraceae most probably reflect the widespread coastal xerophytic and halophytic grass communities. The pollen spectra indicate that the area was also under significant anthropogenic influence, considering the high abundance of pollen belonging to open vegetation and elements of synanthropic vegetation like Rumex, Plantago, Urtica, Centaurea cyanus, Cirsium-type and cereal pollen grains. These results are in line with the outcomes of palynological studies from the Black Sea shelf (Atanassova 2005).

Conclusions

The archaeobotanical analyses of plant remains from funerary contexts, preserved thanks to corrosion by-products, provide rich evidence, which cannot be obtained from dry soil deposits, without these special preservation conditions. The results of the current study demonstrate an approach that is promising for detailed reconstruction of burial rituals and the detection of plants and other organic materials associated with them. The finds revealed perishable elements of the ritual like the organic grave inventory, i.e. wooden artefacts, matting, textiles, flowers, etc. Thus, they significantly enriched the information gained and improved the understanding of the burial rites.

Further, this archaeobotanical study contributed to the reconstruction of the vegetation and environment near the burials, showing in both cases vegetation of a rather open character, exposed to anthropogenic impact in the surroundings of both necropolises. In Varna the pollen assemblages point to landscape with a forest-steppe character, while in Messambria, oak woodland rich in undergrowth of shrubs played a more significant role. These results are consistent with the evidence from the off-site pollen analyses in the study area, corresponding to both periods considered.

This study gives an example of the potential of metal surfaces for organic preservation in dry soil conditions, and demonstrates that such organic remains can be a valuable source of information both on the structures from which the metal artefacts come, but also on their surrounding environment. When excavating and conserving metal artefacts from such a type of site, careful observation and sampling of possible organic residues is recommended. For optimal results, sampling strategy should be discussed with the respective specialists dealing with the artefacts, such as excavators, conservators, and museum curators. This would allow the most efficient and representative sampling, and spare any damage to the analysed artefacts.