Introduction

The first Neolithic culture to emerge on Polish territory during the second half of the 6th millennium BCE is the Linear Pottery Culture (abbreviated as LBK from German). It is now clear that the people of this cultural group brought knowledge of plant cultivation to Central Europe, with cultivated species originating from foreign sources in the Middle East1,2. However, the definitive list of cultivated species remains under debate, as several species considered first crops were not uniformly present across the earliest Neolithic ecumene1, and losses of certain species were observed2, likely due to challenges in adapting to the climatic conditions of the new regions3,4. Additionally, our knowledge of the spectrum of plant species known from Poland used by the first farmers may have been biased due to a lack of archaeological sites with representative collections of macroscopic plant remains. Other factors affecting the understanding of Early Neolithic plant collections include post-depositional disturbances, which have led to mixed assemblages that are difficult to discern during archaeological fieldwork5.

The archaeological excavation of the Early Neolithic LBK settlement in Biskupice (site 18) began with the aim of gathering multiple samples to create representative plant assemblages. The site’s special conditions offer archaeological features representing only one cultural phase associated with LBK occupation, significantly reducing the problem of mixing cultural contexts. Situated in the Carpathian Foothills, this location was not initially considered typical for this cultural unit. However, recent decades have revealed that it is not a unique site in the foothills zone, as other sites demonstrate the penetration of this first stable settlement type into mountainous regions6.

The main objective of the research, which is to document a list of cultivated plants used by the first Neolithic groups settled in the foothills, is divided into three closely aligned specific objectives, which aim to address the following questions: (a) Was there a wide or narrow spectrum of cultivated plants?; (b) Were wild plants used?; (c) Can this plant assemblage be used to reconstruct agrarian techniques?

The site in Biskupice is located in the Wieliczka Foothills (312 m.a.s.l.) belonging to the belt of the Carpathian Foothills. It is situated on a loess-mantled hummock and delimited to the north and south by two streams7. During two archaeological excavation campaigns (2020 and 2021), at least five LBK household units were unearthed (Fig. 1). Originally rectangular, these houses were oriented predominantly along a north–south axis. Houses 1 and 3 are similar in length, as indicated by the distribution of elongated pits marking their boundaries, while House 2 is half the length but comparable in width (6 m × 11 m). House 4 and alleged House 5 were only partially excavated. There were also several pits that were likely associated with some household units, which differed in size, shape and archaeological findings density (Figs. 1 and 2). All of these features yielded a significant number of artifacts, including pottery fragments, flints, sporadic obsidian artefacts and fragments of daub. Based on a detailed analysis of artifacts from House 2, particularly in connection with the unique face vessel found in it7, the chronology of the settlement was established to the late phase of the LBK, known as the Želiezovce phase8.

Fig. 1
figure 1

Plan of the archaeological excavations at Biskupice 18 showing detected features coming from several houses. X-PAD Ultimate https://geomax-positioning.com modified in CorelDRAW Graphics Suite 17 https://www.coreldraw.com.

Full size image
Fig. 2
figure 2

Selected archaeological features from Biskupice 18.

Full size image

The Biskupice site provided the largest dataset of macroscopic plant remains for LBK sites in Poland, which were also utilized for radiocarbon dating to verify not only the chronology of the site but also the presence of specific plants during the Early Neolithic (Supplementary Tables 1–2).

Results

Charred plant remains from soil samples

A total of 1648 samples were analysed from Biskupice, with 1012 (61%) containing macroscopic plant remains, yielding 11,704 specimens. Due to poor preservation, 1240 could not be identified taxonomically. Among charred plant remains, 25 species, seven genera and five families were identified, with plant density per liter varying from 0.2 to 17.7 specimens (Table 1, Supplementary Table 2).

Table 1 Macroscopic plant remains from archaeological site of Biskupice 18: summary of the results (for details: see Supplementary Table 2).
Full size table

In House 1, 1499 specimens were found, primarily in feature 9, with emmer wheat (Triticum dicoccon) predominating among cultivated plants. Herbaceous wild plants included fat hen (Chenopodium t. album), black bindweed (Fallopia convolvulus), soft brome (Bromus hordeaceus), and numerous caryopses of Bromus sp. House 2, the smallest but most abundant in plant remains, had 4445 specimens dominated by emmer wheat, einkorn (Triticum monococcum), a single specimen of bread wheat (Triticum aestivum) and barley (Hordeum vulgare), although due to its preservation state it is not possible to determine whether it is naked or hulled barley. Wild plants were mainly F. convolvulus, Ch. t. album, and B. hordeaceus. House 3 yielded 511 remains, mainly from feature 82, with emmer wheat predominant and some einkorn. Wild plants included F. convolvulus, Polygonum lapathifolium, and Ch. t. album. House 4 had only 120 specimens in feature 138, with emmer and einkorn prevalent among cultivated plants. Wild herbaceous plants found were Ch. t. album, F. convolvulus, P. aviculare, Sambucus cf. ebulus, and some wild grasses (Poaceae) (Table 1, Supplementary Table 2, Fig. 3a, b).

Fig. 3
figure 3

(a) Share of cultivated plants from Houses 1–4 from Biskupice 18. (b) Share of wild plants from Houses 1–4 from Biskupice 18. (c) Share of cultivated and wild plants from selected archaeological features from Biskupice 18. (d) Normal distribution plot of the archaeobotanical samples from elongated pits. (e) Results of Principal Component Analysis (PCA) conducted on archaeobotanical samples collected from elongated pits connected to Houses 1–4 to observe qualitative differences between plant macro-remains. (f) Results of PCA conducted on selected materials from archaeobotanical samples collected from elongated pits connected to Houses 1–4.

Full size image

Other features near Houses 2 and 3 (Fig. 1 and 2d) showed varying plant densities. Feature 18 had the highest number of remains (1483; 17.7 sps/L), with emmer and einkorn common among cultivated plants. Wild herbaceous plants included Fallopia convolvulus, Bromus sp., Chenopodium t. album, and Fragaria sp./Potentilla sp. Feature 150, despite its depth, had only 51 specimens (0.3 sps/L), with a similar spectrum of wild plants: F. convolvulus, Ch. t. album, and Bromus sp. Feature 135 contained mainly cereal grains (182 remains; 1.2 sps/L). Features 22, 26, and 69 stood out for their high numbers of remains, differing in plant density and composition. Feature 22 exhibited a composition similar to feature 18. Feature 26 contained abundant cereal grains and wild plants. Circular pit 69, mainly in its lower parts, contained destroyed cereal grains and wild plants like F. convolvulus and Ch. t. album. Feature 21 had a notable concentration with 220 specimens (10.5 sps/L), including the site’s only flax seed (Linum usitatissimum) (Table 1, Supplementary Table 2, Fig. 3c).

Results of statistical analysis

Regarding the distribution of plant macro-remains in the samples collected from the elongated pits, the plot showed that the analyzed variables are not normally distributed and that there are differences among houses (Fig. 3d). This might be caused by a large number of empty samples in the elongated pits connected to House 3 and a relatively high number of macro-remains in pit 24 (House 2). On the other hand, the results of the PCA conducted for all cultivated and wild plants showed that Principal Component 1 (PC1) is correlated with variables such as Cerealia caryopsis, Cerealia chaff, Poaceae large, Fallopia convolvulus, Bromus, and Polygonaceae. It explains 14% of the variance, while Principal Component 2 (PC2) explains 7% of the variance and is positively correlated with weeds such as Chenopodium album, Bromus, and Poaceae large (Fig. 3e). The results indicate significant taxonomical similarity among the analysed samples. However, samples from elongated pits associated with House 2, particularly pit 24, are clearly separated on the diagram, likely due to the significantly higher number of macro-remains in this pit. This confirms a previous suggestion of a considerable discrepancy in the quantity of carpological materials between House 2 and the elongated pits associated with other houses, possibly explained by a higher proportion of wheat cereals compared to weed macro-remains in the samples.

To mitigate the influence of cereals, a second PCA was conducted. In the PCA plot of legumes and wild plants, Principal Component 1 (PC1) explains 10.3% of the variance and shows positive correlations with Silene sp., Fabaceae, and Poaceae, and a negative correlation with Bromus. Principal Component 2 (PC2) explains 9.3% of the variance and is positively correlated with Chenopodium t. album and negatively correlated with Polygonaceae (Fig. 3f). In this analysis, samples were grouped based on their affiliation to specific features. A modest pattern was observed where labels of groups of features assigned to the same house tended to cluster together, although most samples appeared to be distributed across the plot. This may be explained by a general similarity in the weed composition in all elongated pits. However, there is slightly higher similarity among the infills of elongated pits belonging to particular household units.

Spatial distribution of plants – spatial variability in plant uses within the site

The archaeobotanical material from all houses displayed a similar taxonomic composition of cultivated and wild herbaceous plants, although their relative frequencies varied among houses (Figs. 3a–c, 4). Features associated with House 2 contained the highest number of cereal grains and chaff, whereas House 1 exhibited the highest ratio of wild herbaceous plants to cultivated ones. House 2 contained more herbs compared to Houses 3 and 4, although the latter is only partially preserved. Various parts of plant remains were discovered in elongated pits and other types of pits, likely serving economic functions such as storing food supplies. However, it is possible that plant remains reflect household activities such as food processing or post-embedding work. The presence of chaff, spikelet forks, rachis, straw, legumes, and weeds indicates grain processing during preparation for consumption9,10. Chaff remains may also imply the preservation of glume wheats in spikelets11.

Fig. 4
figure 4

Spatial distribution of selected plant macro-remains from Houses 2 and 3 from Biskupice 18. X-PAD Ultimate https://geomax-positioning.com modified in CorelDRAW Graphics Suite 17 https://www.coreldraw.com.

Full size image

Only Houses 2 and 3 enable observation of the spatial distribution of plant remains within these households and their associated pits (Fig. 4). Cereal remains (caryopses and chaff), were present in all elongated pits of both houses, as well as in the majority of post-holes in House 2. A similar pattern is observed for wild herbaceous plants. There is a significant disparity in the number of plant specimens recovered, with almost eight times more specimens found in House 2 compared to House 3. However, the analysis of pottery in relation to the radiocarbon dating results (Fig. 5) does not indicate that these differences stem from chronological disparities between the two households. Any chronological differences observed are relatively minor, possibly spanning one to two generations, without reflecting significant stylistic changes. Thus, it is likely that these discrepancies can be attributed to differences in the cultural layers deposited in the elongated pits, bearing material connected to food or textile production or representing discard areas. Furthermore, pit 26, initially linked to House 3 based on feature distributions, may also relate to House 2 upon plant density analysis (Fig. 4).

Fig. 5
figure 5

Results of the radiocarbon dating (uncal. BP) coming from plant macro-remains from Houses 2 and 3 from Biskupice 18.

Full size image

For both houses, the elongated pits on the eastern side showed distinctly different dynamics of filling formation compared to those on the western side. This implies different activities in the vicinity of both parts of the settlement. Pits on the western side of both houses, with exception of pit 20, exhibit a similar pattern, with the highest plant density observed not in the bottom parts, but in a few layers of the upper levels (Supplementary Fig. 1). Conversely, elongated pits 25 and 126 on the eastern side show the highest density levels at the bottom and in their uppermost parts. This suggests that the formation dynamics of their fillings, in terms of plant remains, varied depending on pit location, with pits on the same side likely filled simultaneously due to similar or even simultaneous processes (Fig. 3d).

Pits like 18, 22, and 69 show a high concentration of plant remains, especially cereals and wild herbs, per liter of sediment (Table 1). This suggests specific functions, possibly related to intensified food processing or varied waste disposal practices. In contrast, post-holes exhibit fewer plant specimens, with some empty and only a few containing significant taxa, albeit including cultivated and wild herbs.

Plant remains from daub

Ten taxa were identified from daub fragments, predominantly from House 2 and four unrelated pits (nos. 15, 69, 136, and 150; Supplementary Table 3). Among cereals, the remains of emmer were particularly notable, represented by imprints and charred or dried spikelets, spikelet forks, glumes, and caryopses. Some of the remains belonged to one or both of the hulled wheat species (T. dicoccon or T. monococcum). The least abundant were imprints of barley grains. Many undetermined remains belonging to Cerealia indet. were observed, including leaf and stalk fragments and pieces of grains. Also, specimens classified as cereals or wild grasses appeared.

Wild herbaceous plants were very scarce and were documented only in feature 15. Several caryopses of wild grasses were preserved, as well as a grain of Bromus sp., a grain of Bromus secalinus, a seed of Chenopodium t. album, and remains representing two families, Polygonaceae and Caryophyllaceae. Daub pieces from feature 150 differed from others as they contained only remains of small fragments of charcoal, preserved in samples from different depths. Prominent among these were mainly remnants of oak (Quercus sp.) and single specimens of wood from Scots pine (Pinus sylvestris).

Pollen samples

Only seven samples from two archaeological features contained pollen grains (Fig. 6). In feature 126 (House 3), pollen grains were present in the bottom part of its filling, but their frequency is limited to a few dozen, and thus they do not adequately represent the vegetation. The results of palynological studies confirm the presence of only two arboreal taxa: Pinus sylvestris and Picea abies. Among herbaceous taxa, pollen grains of Poaceae, Cyperaceae, Cichorioideae, Centaurea cyanus and cereals (Cerealia, Triticum type, and Secale cereale) were determined. In addition, spores of ferns, hornworts Anthoceros and Phaeoceros and liverwort Riccia were present.

Fig. 6
figure 6

Results of the pollen analysis from archaeological features from Biskupice 18.

Full size image

In feature 9 (House 1) pollen grains are numerous only in a single sample (Fig. 6). The pollen spectrum is characterized by a predominance of arboreal pollen from Pinus sylvestris, Betula, Salix, Quercus, Ulmus, Picea, Populus, Alnus and Sorbus type. Herbaceous plants reaching a considerable value of 22% are represented by Cyperaceae, Filipendula, Poaceae, Anthemis and Chenopodiaceae.

Discussion

Crop assemblage – the choice of emmer wheat

Emmer grains, totaling 1019 specimens (Table 1; Fig. 7b,c), due to the sampling strategy from the most abundant assemblage of this wheat in southern Poland. In other sites of the LBK culture, only a maximum of a few dozen grains were found, as seen in Brzezie 17 and Gwoździec 26,12. The second wheat species found in Biskupice was einkorn, mostly preserved as caryopsis (Table 1; Fig. 7b:a). Grains and chaff of this wheat were also documented in southern Poland, albeit in lower quantities than emmer12,13 and in Ludwinów 7, in the Kujawy region14. The taxonomic composition of cultivated plants identified at Biskupice, based on charred plant remains from the pits and daub, suggests that emmer wheat played the most significant role in LBK cultivation. It is possible that einkorn wheat was also sown in smaller quantities. This pattern is consistent with data obtained from the analysis of charred remains from other sites in Poland12. However, in other areas, emmer does not always prevail over einkorn; at the Vráble site in Slovakia, dated to the Želiezovce phase, both cereal species were found in similar quantities15. In contrast, in the easternmost areas of the LBK in Europe, such as Ukraine and Moldova, einkorn usually prevailed over emmer16,17, which is a similar pattern to Early Neolithic sites of Bulgaria4. This difference may be attributed to variations in climate, as einkorn wheat is less productive than emmer, but more resistant to low temperatures4,18. The frequent rains during the Atlantic phase and high humidity in Central Europe19 may have influenced farmers’ choice of cereal species.

Fig. 7
figure 7

(a) Relative frequency of wild and cultivated plants from Houses 1–4 at site Biskupice 18 (excluded: Cerealia due to their predominance; bread wheat (Triticum aestivum due to AMS 14 C dating). (b) Charred remains of cultivated plants from Biskupice 18: a Caryopsis of Triticum monococcum (einkorn), b Glume base of Triticum dicoccon (emmer wheat), c Caryopsis of Triticum dicoccon (emmer), d Triticum sp. (wheat), e Fabaceae indet., large (probably Pisum sp., pea). (c) Charred remains of wild plants from Biskupice 18: a Polygonum sp. (knotweed), b caryopsis of Echinochloa crus-galli (cockspur grass), (c) Fallopia convolvulus (black bindweed), d Chenopodium t. album (fat hen), (e) Bromus sterilis (sterile brome). Photos: Krzysztof Stachowicz (7b and c).

Full size image

Remains of barley were scarce, but they appeared as charred and in daub fragments (Table 1). Its presence was documented in southern Poland at seven sites12,13. Given that barley remains are often found as single specimens and its role in Early Neolithic cultivation is sometimes debated, with some considering it a weed4,20, one grain of barley underwent radiocarbon dating. This was particularly significant as two barley grains found at Vráble were of younger origin, suggesting they were intrusive from younger layers15. The dating of barley from Biskupice, correlating with other LBK dates, confirms its presence during that period (6210 ± 40 uncal BP; Supplementary Table 1).

Bread wheat, identified in Biskupice only as singular specimens, was determined to be younger through radiocarbon dating (1205 ± 30 uncal BP; Supplementary Table 1), not surprising given its increased cultivation in early Medieval Poland21. Bread wheat was found at LBK sites in Poland12 and other European locations such as Bulgaria4,22 and Hungary23, but not confirmed by their direct dating. Also, its presence at Vráble is considered doubtful due to the absence of chaff15. Similarly, the presence of broomcorn millet (Panicum miliaceum), identified as a single specimen in pit 119 of House 1, has been questioned to be of Neolithic origin due to recent studies on its history and expansion to Europe24,25,26. While attempts were made to radiocarbon date it, insufficient carbon content hindered dating. Millet remains were found in six archaeological sites with LBK plant assemblages from southern Poland12, but were not directly dated. On the other hand, radiocarbon dating of millet grains from Vráble and Kamyane (Ukraine) indicated a younger chronology15,17, demonstrating the potential for intrusive millet remains. Therefore, the Neolithic chronology of millet remains uncertain, with the oldest dates suggesting its confirmed presence in Europe from the second millennium BC26.

Other cultivated plants found in Biskupice are remarkably scarce. A single seed of flax was discovered. Similarly, this taxon is not abundantly represented at three other LBK sites in southern Poland12. Flax was also sporadically present in the northernmost areas of the LBK in Poland, specifically in Kuyavia14. Its scarcity at LBK sites is further confirmed by its limited occurrence in Vráble15. Among legumes, seeds probably of peas (cf. Pisum sativum, Table 1; Fig. 7b:e) were found. They were uncommon at other sites in southern Poland, being found at only one site12. In contrast, peas were more commonly detected in Vráble15.

In summary, thanks to the extensive sampling efforts across various types of features, it was possible to determine the first crops used in this settlement and to examine differences of plant compositions in different archaeological contexts. The agricultural system of the Neolithic settlement in Biskupice focused on cultivating two cereal species: emmer and einkorn wheat, with barley playing a minor role, likely as crop admixtures. Emmer’s predominance is evident across all households. Pollen analysis, though limited, suggests the presence of plots at some distance ofthe households used for cereals cultivation (Cerealia, Triticum type, and Secale cereale—possibly sporadic weed21,27). Alternatively, pollen presence might also indicate cereal processing close to the houses, although in this case, the accumulation of cereal straw should result in a higher frequency of cultivated crop pollen. Although legumes and flax were sporadic, their dietary role is challenging to determine. Legumes, rich in amino acids, and dietary fiber, possibly served as meat substitutes, while flax was primary an oil supplier21. Typically, Early Neolithic sites exhibit a low number of plant remains found, and greater accumulations of plant remains appear only exceptionally28,29. In LBK plant assemblages in Poland, Central and eastern Europe, a very low density of specimens per litre of sediment is common6,12,13,14, 17. For example, in the case of the Early LBK settlement in Gwoździec 2, located also in the foothills, the density of cereal grains in the elongated pits did not exceed 0.316. This scarcity of specimens is often associated with a low number of samples collected. However, in Biskupice, significant effort was made to sample all parts of the pit’s fillings, yielding thousands of macroscopic plant remains. The relatively low number of cultivated species is not due to inadequate sampling, but rather it confirms the farmers’ deliberate choices. This, in turn, confirms a general overview of the LBK crops observed across the LBK world compared to the earlier Neolithic cultures responsible for its formative phase (Supplementary Table 4). A longer list of crops observed at Polish sites12, including spelt wheat (Triticum spelta), broomcorn millet (Panicum miliaceum), and barley (Hordeum vulgare), differs from what is known from other LBK areas30 (Supplementary Table 4), particularly in comparison to western regions, and this discrepancy might be due to taphonomic problems. However, the radiocarbon dating of barley from Biskupice confirms its presence as early as in the LBK period, which adds important data for crop variability. The choice of using emmer as the main cereal seems to be deliberate and was either due to its better adaptation to the climatic conditions north of the Carpathian Mountains or to people’s preference for its use, such as for flour. The presence of barley, which is better adapted to continental climates and winter frosts, may be the first sign of its use in a region with slightly different climatic conditions, such as the foothills surrounding Biskupice. The scarcity of legumes and flax might suggest that people struggled to introduce these plants in the foothills, or it may indicate their minor role at the LBK sites. Alternatively, this underrepresentation could be due to taphonomic processes, preservation biases, or differences in agricultural practices and dietary preferences.

Wild plants – food supplies and environmental indicators

The most abundant wild herbaceous plants were the diaspores of black bindweed (Fallopia convolvulus), totaling 1057 fruits (Table 1; Fig. 7c:c), a common weed in cereal crops and ruderal habitats31. F. convolvulus was radiocarbon dated, confirming its Early Neolithic chronology (Supplementary Table 1). It was identified at nine LBK sites in southern Poland12,31. In Vráble, it ranks among the three most represented wild herbaceous plants, but with only 12 specimens found15. This species was also found in the easternmost areas of the LBK17.

Fat hen (Chenopodium t. album) was highly prevalent at Biskupice, with 913 seeds found (Table 1; Fig. 7c:d). Primarily thriving in ruderal habitats, it can also be found in cultivated fields. Known for its prolificacy and long seed lifespan31, it is a common find at archaeological sites6,12, 17, 32. In the northern LBK regions, its presence suggests significant crop-like utilization33. In Vráble, it dominates as the most abundant wild herbaceous plant, with 34 seeds15, and was also documented in the eastern LBK areas17.

A significant collection of bromegrass (Bromus sp.) fragments, totaling 986 grains, was documented at Biskupice (Table 1). While species determination was limited, identified species include Bromus cf. arvensis, B. hordeaceus, B. secalinus, and B. sterilis (Fig. 7c:e). Typically thriving in ruderal areas, with the exception of B. hordeaceus, which also inhabits meadows34, bromegrass is frequently encountered in LBK sites across southern Poland12. Widely regarded as a common weed in modern winter crops, the Bromus genus encompasses nearly 100 species35. In Poland, B. hordeaceus prevails as the most widespread species with occurrences in diverse habitats36.

Cockspur grass (Echinochloa crus-galli; Table 1; Fig. 7c,b) was common in Biskupice, recognized as a weed in cereal and root crops today. Its origin from Southeast Asia distinguishes it from other wild plants associated with founder crops, which typically originate from West Asia1. Belonging to the Panicoidae subfamily, it is linked with broomcorn millet and its spread in Europe’s segetal flora has been usually tied to its migration routes37. However, archaeological evidence in Poland dates back to the Early Neolithic31, with sporadic findings in the easternmost LBK sites17, Vráble15 and other Central European sites4. A grain from Biskupice was dated and confirmed its association with the LBK (6020 ± 35 uncal BP; Supplementary Table 1). This raises questions about its connection with broomcorn millet spread, especially considering recent research indicating the absence of confirmed millet remains in European Neolithic contexts26. Therefore, the presence of this species in the LBK layers is highly significant and contributes to its history in Polish flora, suggesting that it appeared earlier than millet and was already part of the natural vegetation in the foothills by the sixth millennium BC. The routes and means by which it reached Central Europe would certainly require further investigation.

Other herbaceous wild plants found at Biskupice included various species of Polygonum (Table 1; Fig. 7c:a) and small crabgrass (Digitaria ischaemum). These plants exhibit high plasticity, thriving in diverse soil types30. Additionally, catchweed bedstraw (Galium aparine), hempnettle (Galeopsis tetrahit), wild turnip (Brassica cf. rapa), black nightshade (Solanum cf. nigrum; Fig. 7c:d), dwarf elder (Sambucus ebulus), and were recorded, with some species also found in meadows and forests31. Similarly, field sorrel (Rumex acetosella) can be found in both ruderal and meadow areas. The herbaceous plants in Biskupice likely originated from various habitats. Most of them, came from segetal weeds and ruderal habitats, belong to the Stellarietea mediae class (e.g., Chenopodium t. album, Fallopia convolvulus, Galium aparine, Bromus arvensis, and Polygonum aviculare). Some plants currently occur in weed communities of root-crops cultivation from the order Polygono-Chenopodietalia (e.g., E. crus-galli, P. lapathipholium) and Panico-Setarion aliance (e.g., Digitaria ischaemum). In root-crops communities, Ch. t. album, and P. persicaria could also grow31,38 (Supplementary Table 5).

Associated with ruderal habitats is dwarf elder (Sambucus cf. ebulus), which falls within the Artemisietea vulgaris class. The LBK’s penetration range might have also included sandy grasslands of the Corynephoretalia canescentis, where field sorrel (Rumex acetosella) could have thrived. Most species found at the site require soil rich in mineral salts (trophism index Tr4) or moderate (Tr3) to develop, while only a few can grow in poor soils (Tr2), such as Digitaria ischaemum and Rumex acetosella. Galium aparine, Echinochloa crus-galli, and Polygonum lapathifolium might originate from very rich areas (Tr5)38,39 (Supplementary Table 5). In Biskupice, there is not a distinct plant assemblage directly associated with food storage, hindering observations on the relationship between cultivated and wild plants for inferring plot characteristics based on the assumption that the weeds grew alongside crops28,31. Pit 18 potentially served as a storage pit due to the significant cereal accumulation, but the high number of wild plants, possibly originating from various habitats like arable fields, ruderal areas, meadows, and forests, suggests it may have also functioned as a trash area near the food processing zone. On the other hand, all plants exhibit similar preservation states and are uniformly distributed across various layers of the pit (Supplementary Fig. 2), indicating they may represent a store pit or might have originated from nearby cereal stores. The presence of both emmer and einkorn occurring at many sites within the same feature suggests they were sown together as a mixture. This is confirmed by their similar soil requirements and the length of their vegetation period21. Barley was likely sown separately due to its vegetation period and soil requirements. However, the small proportion of barley grain in samples with wheat may be remnants of grain from the previous year, which survived in the soil and sprouted in a plot sown with a different species21. The presence of common weeds like fat hen, black bindweed, and various Polygonum and Bromus species suggests diverse cereal cultivation methods and field preparations31. Their presence, along with legumes and flax typically sown in spring (Supplementary Table 4:sa), hints at spring cultivation practices4,31, although the long flowering period of fat hen and black bindweed suggests that autumn sowing cannot be ruled out4,15, 41. The taxonomic composition of weeds occurring together with the grain does not provide a definitive answer and may suggest both spring and winter cultivation of cereals. Spring species include, for example, Fallopia convolvulus and Galeopsis tetrahit, while winter species include Bromus hordeaceus and B. secalinus31 (Supplementary Table 5). Interestingly, the same wild herbs prevail in other features from Biskupice, indicating similar arable field conditions near settlements.

A general affinity in the weed composition observed in all elongated pits by PCA may indicate quite similar agricultural practices among the household units. In detail, the results of PCA (Fig. 3e) of the wild plants suggest that, at least in some cases, Fallopia convolvulus, Bromus and Polygonaceae, together with grasses, were associated with cereals, likely growing in the same fields or being brought to the settlement together with the harvest, as observed by PC1. Additionally, PC2 shows a weak positive correlation between F. convolvulus, Bromus, and Chenopodium album (Spearman’s rs 0.3–0.36), suggesting that these plants may have grown in the same fields or at least been stored (and further charred) together.

The palynological analysis revealed that only two features retained preserved pollen grains and spores, indicating rapid coverage to enable their preservation. Interestingly, the presence of ferns, hornworts Anthoceros and Phaeoceros, and liverwort Riccia spores in elongated pit 126 suggests that this pit remained open for a period. Because hornworts are associated with moist, sparsely vegetated, or bare soil, their presence may indicate ongoing management practices at the site, including agriculture that regularly disturbs or purifies the soil. Some species, like Anthoceros agrestis, are now connected with agricultural landscapes, thriving in moist field soils, therefore they can be linked with the field plots42,43.

The prevalent herbs might have been also consumed in prehistory and historic periods33,40. For instance, fat hen could have served as a cereal replacement, with both its seeds and green parts potentially used for meals12. Other wild herbaceous plants like Bromus secalinus, B. hordeaceus, Fallopia convolvulus, and Echinochloa crus-galli, along with Polygonum lapathifolium and Rumex acetosella, have been found in the intestines of European bog bodies40,44. Although these findings represent a younger chronology, from the Early Iron Age and Middle Ages (twelfth century), they confirm their economic significance and potential use in the diet, which might provide insights for the Neolithic era.

Furthermore, gathered fruits are scarce in Biskupice, with only a few hazelnut fragments (Corylus avellana) and possible wild strawberry fruits (Fragaria sp. or Potentilla sp.) (Table 1), which might have been collected from local woodlands. Seeds of black elder (S. nigrum) were also present. Sambucus fruits are edible and could have been collected on purpose for food. Generally, gathered plants are found in low quantities at LBK sites, especially those with only charred remains12. However, sites with preserved plant assemblages in both charred and uncharred forms tend to exhibit a higher variety of gathered plants, mostly preserved as uncharred45.

Finally, both cultivated plants and wild herbs were utilized as temper in daub. The presence of wood fragments (charcoal) in daub suggests they are remnants of wooden post structures coated with clay, similar to examples from nearby site Brzezie 1746.

Conclusions

The study of macroscopic plant remains from Biskupice 18, the largest Early Neolithic assemblage in Poland, sheds light on plant usage and the agricultural practices of the first Neolithic farmers in the foothills during the late phase of the LBK. The analysis of the dataset, based on representative sampling at the site, reveals a diverse spectrum of cultivated plants, although narrower compared to initial founder crops known from southern Europe and West Asia. Emmer wheat emerges as the predominant crop, followed by einkorn wheat and barley, with evidence suggesting the cultivation of flax and leguminous plants. Pollen analysis hints at nearby cereal cultivation plots or just confirms crop processing at the site. Wild herbaceous plants, including black bindweed, fat hen, and various bromegrass species, are prevalent, possibly indicating their role as weeds in cereal crops. However, the dietary significance of certain wild herbs cannot be overlooked, reflecting the complexity of Neolithic subsistence strategies, with indications of dietary diversity and potential economic significance beyond agricultural purposes. The spatial distribution of plant remains within households and associated pits suggests diverse agricultural activities such as grain processing and food storage, with certain features possibly primarily serving as storage pits with their secondary function as trash areas. Thanks to radiocarbon dating of selected plant species, it was possible to track the history of cultivated and wild plants used and grown in the vicinity of the first stable settlements in the foothill zone of southern Poland.

Materials and methods

Archaeobotanical samples

There are three main sample types that were processed: 1) soil samples and 2) daub samples used for macro-remains analysis, and 3) soil samples used for palynological studies. The highest number of samples were obtained in the first category.

The soil samples of standardized volume (3 L) were systematically taken from all archaeological features (pits, post-holes, and elongated pits). They were taken in a so-called “grid,” arranged horizontally, on average, every 10 cm, and vertically, from each layer representing 10 cm of depth. The number of samples depended on the size and depth of the features; more samples were taken from large pits than from smaller ones (post-holes). The number of samples taken was also influenced by the state of preservation of the features caused by erosion, ploughing, or other factors such as damage by drains (e.g., features 4, 69, 135). A mixed strategy was used during sampling. In the first excavation season, in 2020, material was taken in a grid, every 10 cm of depth (features 1–106), while in the second season, in 2021, the sampling followed the above-mentioned procedures for the first halves of the features. For the second halves, samples were collected only from specific layers such as clusters of daub, loess layers, burnt layers, and bottoms of features. Post-holes were explored plastically, and due to their shallow depth, practically the entire half of their fill was sampled (approximately 3 L). Altogether, 1647 samples were analysed for macroscopic plant remains.

Macroscopic plant remains were wet sieved using the flotation method with sieves of 0.5 mm and 1 mm mesh diameter at the laboratory of the W. Szafer Institute of Botany PAS (IB PAS). The resulting plant material consisted mostly of charred specimens and a few uncharred diaspores belonging likely to a contemporary seed bank. The latter group was excluded from the analysis. The remains of fruits and seeds were taxonomically identified on the basis of morphological features visible under a stereoscopic microscope at magnifications of 10–60 × . They were identified using keys, atlases and other publications21,47,48,49,50,51, following nomenclature according to Mirek and co-authors53. A comparative collection of modern diaspores, as well as a collection of fossil floras belonging to the National Biodiversity Collection of Modern and Fossil Organisms of the IB PAS was used.

The quantitative analysis was performed arbitrarily counting each specimen as one, regardless of whether it was preserved whole or in fragments. In the case of cereal grains, in addition to whole grains, the proximal and apex end parts were counted and then treated as one specimen. The number of remains representing different parts of the cereals (e.g., glume, glume base, rachis internodes and spikelet forks) was given separately. Seeds and fruits preserved as incomplete were counted as one specimen. Cereal grains (caryopses) were identified at the species and genus level, while poorly preserved pieces of grain were classified as indeterminate cereals (Cerealia indet.). The indeterminate category included fragments of plant remains; due to their poor state of preservation, they were not classified to any taxonomic rank.

In Biskupice, the wild plants primarily consist of weeds and ruderals, prompting questions about reconstructing past plant communities. Reconstruction typically relies on actualism principles21 and references to contemporary phytosociological classifications38,39, 54, 55. Plant communities are identified based on the presence of characteristic species, whose remains indicate areas frequented by prehistoric settlements.

In the features found in Biskupice in general there was a scarcity of daub fragments. Each piece was collected and measured with a total station. Finally, there were only 133 daub samples coming from eight features. The individual samples contained between 1 and 36 fragments of small size, burnt or dried to varying degrees, ranging in colour from light yellow through orange to black. It should be noted that different colours could occur within a single piece. In many of the pieces of clay, imprints, relatively burnt or dried, very fragile fragments of plant tissues and organs were found arranged disorderly, both on the surface and within the lumps. The degree of identification of plant remains preserved in daub is determined on one hand by the structure of the plant organ (e.g., the sculpturing of the surface or its absence), the quality of the plant imprint, as well as the possibility of distinguishing a particular remain from other specimens, and, on the other hand, by the plasticity of the material and the granulometric composition of the mineral fraction added to the clay21.

Daub fragments were observed on the surface and then broken up. The test surface was cleaned mechanically. Analyses were carried out with a stereoscopic microscope using magnifications of up to 40 × . For the determination of charred wood, a reflected light microscope was used using magnifications of up to 200 × . Preserved plant remains and their imprints were determined using methods appropriate for archaeobotany by observing morphological features of fruits and seeds and chaff fragments of cultivated and wild grasses (see above). Charcoal fragments were determined on the basis of anatomical structures visible in three sections of wood: transverse, longitudinal tangential and longitudinal radial56.

In addition to qualitative data, quantitative data are provided in Supplementary Table 3. Two measures were used. For specimens that could be distinguished within a single fragment, their absolute numbers are given. If the preserved remains formed accumulations, so that their number could not be presented and only their presence was indicated in the table (with the symbol #).

Some of the remains, due to poorly preserved distinctive features, were described with two alternatives as either emmer or einkorn (Triticum dicoccon or T. monococcum). A considerable number of specimens were generally included in the unclassified cereals (Cerealia indet.). Less well-preserved remains were designated as cultivated or wild grasses (Cerealia indet. or Poaceae indet.). Structural impressions, legible only in small daub fragments from features 136 and 150, were also not taxonomically defined.

Pollen samples

29 samples of 2 cm3 of fresh, mostly mineral material were used for palynological studies. Sediment was collected to the plastic bags and metal containers from the exposed walls of the archaeological features 9, 24, 26, 68, 82, 83, 126, 135, 136, 150, 153 and 157 during archaeological excavations in 2021. Sample processing included: HCl-, KOH- treatment, sieving in ultrasonic cleaner (5 µm screen); 40% HF-treatment and Erdtman’s acetolysis. Prior to acetolysis, one Lycopodium tablet with a known number of spores was added to each sample for estimation of the pollen and spores concentration. The residues were mounted in concentrated glicerol57,58. The taxonomic affiliation of the plant micro-remains was determined under Nikon Eclipse 600 microscope with 400 × and 1000 × magnification, using specialized keys59 and the reference palynological collection of the National Biodiversity Collection of Recent and Fossil Organisms at W. Szafer Institute of Botany, Polish Academy of Sciences, Kraków. Due to the extremely low pollen concentration, sporomorphs present on the surface of 2–10 microscopic slides were determined. The percentage values of taxa presented in the pollen diagram drawn on using POLPAL software60 were calculated based on the sum of arboreal [AP] and non-arboreal pollen [NAP] grains of terrestrial plants. The percentage pollen values of local taxa were determined on the basis of the total sum [AP + NAP = 100%) increased by the number of the respective taxa (Fig. 4).

Finds of pollen and spores in situ at dry archaeological sites are rare, as sporomorphs are sensitive to oxygen and require constantly moist conditions or rapid coverage by other sediment for long-term survival. The presence of pollen grains indicates temporarily existing conditions suitable for their preservation.

Radiocarbon dating

Some of the identified plant remains were selected for radiocarbon dating method performed at the Radiocarbon Laboratory in Poznań. They include mostly cereal grains, but also some wild plants were chosen to confirm their Neolithic chronology. Accelerator mass spectrometry (AMS) was used for dating, following standard chemical pre-treatment methods61. Calibration was performed using OxCal ver. 4.4 software62 with a 14C IntCal20 calibration curve63.

Radiocarbon dating results, conducted on cereal grains and diaspores of wild herbaceous plants, confirmed the settlement’s existence towards the end of the 6th millennium BCE (Supplementary Table 1; Fig. 5). However, the plateau effect is evident, with the majority of dating results falling within a broad range spanning from 5200 to 5000 cal. BCE. This broad range hampers the ability to discern chronological differences between archaeological features and houses. The only exception is the date obtained for bread wheat, which turned out to be later (1205 ± 30 uncal BP, Supplementary Table 1). This grain originated from a post-hole, indicating post-depositional disturbances.

Statistical analyses

To assess the distribution of the number of macro-remains in the archaeobotanical samples collected from the elongated pits, a normal distribution plot was prepared (Fig. 3d). Additionally, to further investigate quantitative differences in the dataset, two Principal Component Analyses (PCA) were conducted. To increase the statistical reliability of the analysis, T. dicoccon and T. monococcum were grouped into one category, and other plant macro-remains were grouped at the genus level (e.g., Bromus) or family level (e.g., Polygonaceae). Both analyses considered variables that appeared in the samples at least three times. Moreover, only samples that yielded at least three macro-remains were included in the analysis. In the first PCA (Fig. 3e), both cultivated and wild plants were analyzed. PCA is sensitive to outliers and can also be influenced by the scale of the data. For the subsequent PCA, cereals, which constitute over 76% of all macro-remains, were excluded from the analysis, and all data were additionally normalized (Fig. 3f).

For feature 18, Hierarchical Cluster Analysis (HCA) was conducted based on Ward’s Method. This analysis aimed to examine the taxonomic similarity between individual archaeobotanical samples and thus demonstrate the concentration of similar botanical macro-remains in certain parts of the feature or, on the contrary, indicate the mixed character of the filling of the analyzed pit (Supplementary Fig. 2b). Four clusters were identified based on the dendrogram, which were then plotted on the plan (Supplementary Fig. 2a). The depth at which samples were collected was marked on the plan using colours (black-purple-green gradient). Clusters identified as a result of the HCA analysis seem to describe four types of archaeobotanical samples. Among them, cluster I stands out the most, containing samples with a significant number of wheat and Bromus caryopses. These samples are concentrated at different depths in the central part of the pit. On the other hand, samples located at the boundary of the pit (cluster III) contained moderate amounts of wheat caryopses and diversified weed diaspores. Interestingly, almost all samples with a relatively small number of archaeobotanical macro-remains (cluster IV) were located in the lowest parts of the pit. When interpreting the described situation, attention should be paid to the fact that in the case of pit 18, the variation in the depth of the samples is not very significant, as, in fact, only a 40-cm bottom part of the pit is preserved, as the upper part of the feature has already eroded.

Taphonomical observations of the pits based on the plant macro-remains

To compare the dynamics of filling the elongated pits in Houses 2 and 3, we compiled values expressing the density of carpological material (number of specimens per litre of sediment). The analysis included the density of Early Neolithic archaeobotanical materials most commonly found at the Biskupice settlement, such as caryopsis and chaff of wheat, Chenopodium t. album, Bromus, Fallopia convolvulus, Polygonum, and large Poaceae, as well as cumulative values for these macro-remains (Supplementary Fig. 1). The units analyzed were 10–20-cm mechanical layers identified during field research. Depths reached by the examined pits and the mechanical levels within them with a high density of archaeobotanical material were marked on the illustration. The 0–40-cm layers corresponds to the plow horizon, and were excluded.