Direct evidence for processing Isatis tinctoria L., a non-nutritional plant, 32–34,000 years ago
PMCID: PMC12063890
PMID: 40343940
Abstract
Recovering evidence for the intentional use of plants in the Palaeolithic is challenging due to their perishable nature as, unlike chipped stone or bone artefacts, plant remains are rarely preserved. This has created a paradigm for the Palaeolithic in which plants seldom feature, resulting in a partial and skewed perspective; in fact, plants were as essential to human life then as they are today. Here, we combine morphological and spectroscopic analyses (µ-Raman, µ-FTIR) to provide robust multiscale physical and biomolecular evidence for the deliberate pounding and grinding of Isatis tinctoria L. leaves 34–32,000 years ago. The leaf epidermis fragments were found entrapped in the topography of the used surface of unmodified pebbles, in association with use-wear traces. Although their bitter taste renders them essentially inedible, the leaves have well-recognised medicinal properties and contain indigotin precursors, the chromophore responsible for the blue colour of woad, a plant-based dye that is insoluble in water. We used a stringent approach to contamination control and biomolecular analysis to provide evidence for a new perspective on human behaviour, and the applied technical and ecological knowledge that is likely to have prevailed in the Upper Palaeolithic. Whether this plant was used as a colourant, as medicine, or indeed for both remains unknown, but offers a new perspective on the fascinating possibilities of non-edible plant use.
Full Text
Modern humans (Homo sapiens) first appear in the archaeological record around 300,000 years ago, in Africa. Most of the evidence for their cognitive and technological abilities is based on recovered assemblages of chipped stone artefacts and animal bones since these endure far longer in the archaeological record than plants [1]. Accordingly, the Palaeolithic narrative centres primarily on animal hunting and stone tool manufacture. Perishable materials, the so-called “missing majority” [2], notably plants for which there is growing evidence for their use as food [3–5], string and cordage [6], weaving [7] and medicine [8], are largely missing¸ creating a partial narrative. There is therefore, a need to identify and demonstrate the use of plants and the roles they played in a wide range of activities, many of which may still be unknown. Ultimately, this will contribute to a broader understanding of Palaeolithic behaviour.
Finding direct traces of the deliberate exploitation of plant materials is complex since macroscopic un-charred plant remains rarely survive through Palaeolithic timescales and, even when they do, their presence can be argued as natural deposits [9] or the result of bioturbation or other taphonomic events [10]. Therefore, demonstrating use of plants during human evolution requires not only conditions that are favourable to the long-term survival of plant remains, but also needs to be exhaustive, as the finds must indicate human involvement, and also offer stringent evidence to negate natural occurrence and modern contamination [11]. A lack of awareness of the potential for complex plant processing can compromise evidence for this during excavation and post-excavation processes [12]. Finally, the recent application of high-resolution imaging and biomolecular techniques to the investigation of biogenic residues from the micro to the nanoscale [13–16], is now expanding the potential for detection of a new range of evidence for the use of plants and the likely behaviour associated with this. Use of these techniques can feed into a better understanding of the complexity of human-plant relationships and expand perception of plant processing. Each time a new piece of secure information emerges it provides an extraordinary insight into our prehistoric human past and contributes to a growing awareness that our Palaeolithic ancestors had sophisticated ecological knowledge of plants.
While plant processing, including cooking and grinding, has been recognised for some time for the Upper Palaeolithic [3,17–19], the evidence for modification of plant materials has, until recently, been largely restricted to identification of macro remains, notably carbonised materials and wood [20,21], likely food items [4,5] or pitch although this is generally identified by biomolecular analysis [22–26]. Plants represent an infinite resource that provide humans not only with food and raw materials but also, due to their biomolecular complexity, with medicines, poisons, flavours, aromas, hallucinogens and dyes [8]. The ingestion of non-nutritional plants containing medicinal secondary metabolites was identified in 47,000-year-old Neanderthal dental calculus [27,28], while tentative evidence for poison 40,000 years ago was recovered from Border Cave, South Africa [29]. Medicinal plants are reported from a number of Palaeolithic sites in the Caucasus [30]; however, it is challenging to demonstrate that these were ingested and/or intentionally processed [8]. To date, there is no evidence for the extraction of dyes from organic materials in the Palaeolithic; the known colourants (red, yellow, black and white) are all pigments of mineral origin apart from charcoal [31]. They are highly resistant to ageing, with little apparent degradation and known to have been used in Palaeolithic art and for other purposes. For example, ochre is known in various applications such as tanning leather or skin [32], as a preservative [33,34], as insect deterrent and as skin protection [35].
One source of evidence for the use and intentional processing of plants comes from Ground Stone Tools (GST) [3,17–19]. The use-wear traces on their utilized surfaces and the extraction of micro-residues trapped within their surface are compared to modern reference materials to reconstruct the type of materials being worked. These residues can be extracted, imaged and chemo-profiled [12,16,19].
Here, we share the robust data derived from the chemical-physical analyses that led to the identification of the indigotin molecule, which has not previously been recognised in a Palaeolithic context. We also investigate the processes that may have led to the presence of the blue residues on the stone pebbles. To address this, we examined possible accidental or deliberate reasons for their occurrence as part of the mechanical processing of leaves, including whether the effect of releasing the precursors of indigotin – a well-documented practice in both ethnographic and historical records [36,37] – could have been deliberate. This paper is therefore focused on the recovery and identification of indigotin as an identifying feature of Isatis tinctoria L., extracted from the working surface of these archaeological GSTs. Finally, we provide an exhaustive explanation of the method used for extraction of the residues to ensure the samples were free of contamination during this process. While we present some preliminary use-wear observations to support the actual use of these pebbles, an exhaustive functional analysis will be the object of a separate study.
Indigotin derives from the processing of indigoid plants, in particular species within the Isatis, Indigofera, and Persicaria genera. When the leaves are crushed the precursors are released and the fermentation process begins [38–41]. Among these, I. tinctoria, a non-nutritional plant, is the only species native to Central Asia [42,43], whereas all the other species are of tropical origin and do not naturally grow in cool environments [36]. Metabarcoding techniques applied to the sediments from Aghitu 3 cave in Armenia (39–24,000 years old) identified Isatis sp., hence providing genetic evidence that the genus was present during this period in the Caucasus [14].
I. tinctoria, colloquially known as dyer’s woad, is a biennial plant of the Brassicaceae family that grows in dry and sunny locations. This plant is a source of indoxyl glycosides – precursors of the blue chromophore indigotin and its red isomer indirubin – naturally contained in cell vacuoles of the leaf epidermis, together with a minor quantity of flavonoids [40,41,44]. It is also well known for its medicinal properties due to the significant presence of further indole alkaloids in its roots and leaves [39,45,46].
We developed an innovative and robust research design that integrates synchrotron-based micro-CT scanning (to demonstrate the porosity of rocks enabling retention of residues), microscopy at various resolutions (to document the presence of residues and their correlation with modifications to the stone surface caused by mechanical processing), and spectroscopy techniques, in particular Raman and FTIR (Fourier transform infrared) (to characterise residues and the chemical profile of their constituent molecules). We demonstrate that I. tinctoria was pounded at Dzudzuana by Homo sapiens, 34,000 years ago, using unmodified river pebbles. Unmodified pebbles are recognised as tools in the Upper Palaeolithic, although to date they have been associated with the processing of edible plants [3,16,18,19,47] and inorganic materials such as ochre [48,49]. We do not claim to understand why this plant was being ground up, however we acknowledge that it is well known today for its medicinal and blue dye applications [36,38]. By demonstrating that plant processing was not restricted to edible plants, our research provides a new dimension to human behaviour and contributes significantly to the growing body of evidence highlighting the complexity of human-plant interactions in the Upper Palaeolithic.
Dzudzuana cave, Imereti (Western Georgia), is an Upper Palaeolithic site (Fig 1) comprising three discrete Palaeolithic units spanning ca. 20,000 years of occupation, the lowest unit (D) dating to 34.5–32.2 cal. BP [50] (Fig 1). The archaeological material from this unit comprised chipped stone assemblages, faunal remains, bone tools including a needle, as well as engraved and pierced animal teeth likely to have been used as ornamental artefacts [50]. In addition, the findings included six river pebbles which are the focus of the present study (Fig 2; S1 Fig and S1 Table). Five of these pebbles provide clear evidence of use in the form of functionally active areas and associated micro-residues.
Left: Location of Dzudzuana cave (modified after Google Earth) and cave entrance (Image by LL). Right: map of the cave with the excavated area inside the red circles (redrawn and modified from Bar-Yosef et al. 2011, p. 334, Fig 2 [50]), red squares show where the six pebbles were retrieved from. Excavations in squares G-7, G-8 and I-8 reached the full depth of the archaeological sequence. These are also the squares where the stone pebbles were retrieved.
During this period, several climatic fluctuations occurred; however, during the human occupation, the climate at Dzudzuana cave was generally mild, with species such as wild grape (Vitis sylvestris), hazel (Corylus sp.), oak (Quercus sp.), beech (Fagus sylvatica), birch (Betula sp.) and Artemisia sp. all present in the Caucasus [50]. Moreover, several Brassicaceae species, among which Isatis sp., was detected by metabarcoding [14]. The present study of residues adhering to GSTs supported by preliminary use-wear analysis, demonstrates the purposeful use of these stone pebbles, the archaeological integrity of the material recovered and its accurate identification (Methods, S1 and S3 Figs, Table 2).
Here we report on the methodological sequence during the different stages of the Dzudzuana stone pebbles analysis. The sequential sampling procedure applied is summarized in Table 2 (Methods) and follows the procedures previously established and published by the authors [11,12,16,18,51].
Six stone tools were identified during several archaeological excavation campaigns (2002–2007, details are reported in S1 Table). The pebbles were retrieved in different squares (G-7, G-8, I-8; Fig 1) from Unit D, excavated by 5 or 10 cm levels. Three pebbles (Dzu S1 to Dzu S3) derive from the deepest portion of Unit D, excavated in 2002, whereas three samples (Dzu S4 to Dzu S6) come from the uppermost part of the same unit, excavated in 2007. Only two pebbles are complete (Dzu S3 and Dzu S4). Dzu S1, Dzu S2, Dzu S5, and Dzu S6 are fragments of larger pebbles; Dzu S5 still retains part of the cortical surface and Dzu S6 is a mid-segment of a pebble. During field-work after their retrieval, each pebble was summarily rinsed in running water. Each stone was inventoried in the log book of the excavation and labelled according to museum accession number, and finally placed individually in a labelled paper bag. The stones were then taken to the Georgian National Museum (GNM, Tbilisi).
The paper bags containing the stones pebbles were dusted on the outside, and the sweepings were retained (see Methods, Table 2). Then, each pebble was carefully pulled out, very softly dusted with a clean brush, and the collected dust was also retained for cross-control.
Once dry, selected areas of the stone surfaces, comprising visually evident functionally active areas, were moulded (see Fig 2 for the placement of the moulds) following a well-established procedure [12,18,52–54]. Moulding is a long-accepted method used widely by functional analysis specialists to obtain detailed negative copies of the surface, especially useful in contexts where necessary facilities or equipment to conduct full analyses are unavailable [47,53–56]. The moulding material is a silicone-based dental impression polymer, polyvinyl siloxane (PVS) (Coltene®; Speedex, Universal Activator, with linear shrinkage = -0.2%) [56].
Moulds of dorsal and lateral sides of the stones, seemingly without evidence of use, were also taken for cross reference. Two successive moulds were taken at each location, with the first removing any remaining surface sediment while the second recovered residues entrapped deep within the stone cracks as per the established method [11,12,18,51,54] (Fig 3, S4 Fig). Each mould was immediately bagged individually. Due to the PVS peel-off effect, the moulds can also dislodge residues entrapped within the pockets (pits or cracks) of the coarse stone surfaces [12,16,18] allowing the recovery of different micro-residues entrapped within these (S2 and S4 Figs).
At DAIS, moulds were initially examined under a fume hood. The mould observation revealed the presence of some residues still adhering to the stones in both the first and second moulds, suggesting that they remained entrapped deep inside the pockets on the surface of the pebbles (S2 Fig). The first mould removed the residues nearest the surface, while the second mould extracted the deeper ones. Therefore, the residues extracted with the second mould had been protected from exposure to outside elements, and it is for this reason that we examined only their internal (2nd) moulds.
The 2nd moulds from the five relevant pebbles (Dzu S1, Dzu S2, Dzu S3, Dzu S5 and Dzu S6) were subsampled and cut in half. One half was immediately stored for analysis of wear-traces, while each other half was individually placed in a test-tube with 40 ml of ultrapure water. The tubes, set in an ice bath, were sonicated using an inserted ultrasound probe (UP-200S Hielscher Ultrasonics GmbH, Germany) operating at 100 W for 10 min (0.5 cycle, 80% amplitude) for cycles of 3 minutes (for a total of 12 minutes). The ultrasound probe was carefully cleaned before and after each sonication. The obtained samples were then processed according to the established procedures [12,57], and the extracted micro residues underwent imaging and chemo-profiling (Methods, Table 1).
Different control samples were collected during the sampling process at GNM and a stringent procedure was followed to avoid possible contamination (Methods, Table 2). During all stages of the processing, Petri dish traps were placed on the desk, on the shelves, and in the fume hood, until the stone pebbles sampling was finalized. Then, all the control samples were sealed in vials and shipped to DAIS, where they underwent optical microscopy for blue micro-residues. Control samples of dust swept from the surface of the paper bags and stone pebble surfaces were also collected.
The analysis of thin sections from the archaeological stone pebbles indicated that these are mafic igneous rock. However, since the available fragments from some of the archaeological pebbles were too small to obtain significant data on the overall stone structure, the full petrographic analysis (being invasive) was performed on pebbles collected from the adjacent Nikrisi River that crossed the limestone massifs of the Zemo Imereti Plateau [58]. The analyses of the thin sections from three river pebbles confirmed that these were epidosite and micritic limestones, consistent with the petrographic analysis of the archaeological pebbles.
X-ray Computed μ-Tomography aided with synchrotron brilliance (SR-μCT) was used to analyse the porosity of the stone pebbles collected in the river. The SR-μCT showed significant internal porosity on these igneous rocks (Fig 3 panels a-c). Moreover, microscopic study of their surface texture, investigated by means of optical microscopy (OM) and scanning electron microscopy (SEM), demonstrate that the stone is rich in irregularities and in particular microcavities where residues can become trapped (Fig 3 panels e-f).
Based on an extensive experimental programme, we can confirm that micrometric residues can become entrapped in this type of pores during the pounding of the leaves [59,60] (S1 Video).
Selected moulds from Dzu S1, Dzu S2, Dzu S3, Dzu S5, and Dzu S6 (Fig 2) were analysed according to established use-wear procedures [11] (S2 Fig), by means of OMFig 3g-3i), SEM (S3 Fig panels e,f), Laser Scanning Confocal Microscopy (LSCM, S3 Fig panels a-d) and confocal profilometry (Fig 3 panels l-n) [59]. The most common features observed are polished and flattened areas, scratches and striations (Fig 3g-3l, S3 Fig).
The solution obtained through sonication of the 2nd moulds from five of the six the analysed stone tools was processed to extract micro-residues according to Cagnato and Ponce [57] (S2 Table, S4 Fig). Being aware of potential contamination issues and false positives, we deliberately decided to analyse the micro-residues from the 2nd moulds, considered less prone to contamination. The morphological analysis of the micro-residues was carried out using OM, LSCM and SEM. We recovered 67 individual micro-residues (both coloured and non-coloured) that we were able to identify as of plant-based origin. Of these, 25 individual micro-residues were blue (Figs 4, 5 panels a1-a2, b1-b2, c1-2 and 6a, S5 Fig), while 42 were non-coloured (Fig 6 panels b, c, S6 Fig). Here, we focus on the results of the physical-chemical characterization of the 25 micro-residues that showed a blue colour.
OM, LSCM and SEM images (a, b, c), and µ-Raman spectra (d). Panel a: blue micro-fragments, retrieved from Dzu S5, exhibiting non-coloured regions (OM). Panels b-c: vascular elements retrieved from Dzu S6; b: vascular tissue with aligned pits (OM) and a zoomed region as observed by SEM, retrieved from Dzu S5; c: fusiform (spindle-shape) structure tapering at each end (LSCM) and the nodes displayed in the insert (OM) retrieved from Dzu S5. Panel d: µ-Raman spectra of the micro-residues shown in the micrographs (a: spectrum of a non-coloured region of the blue fragments; b and c: spectra of non-coloured micro-residues), compared with non-coloured I. tinctoria fragments obtained from the processing of modern leaves (Ref3). The spectra have been normalised and the luminescence background has been removed. Characteristic bands of constituting molecules such as cellulose (C), and characteristic signals of its polymorphs (CI, CII), pectin (P), lignin (L) and indigotin (Ind) are indicated by their µ-Raman shift. Details of the spectra are reported in S14 Fig.
On the basis of their anatomical structure, the observed residues are identified as small fragments, primarily of leaf epidermis. Esau [61] demonstrates that the structure of the stem, leaf, and root is quite similar, whereas the arrangement of vascular and ground tissues (parenchyma and collenchyma) is distinctive being permeated by a network of interconnected veins that diffuse extensively across the mesophyll and may also be organised into bundles. In the leaf, the outer layer of epidermis forms trichomes [62], a specific type of hair, that can be highly distinctive in its shape and surface ornamentation. I. tinctoria trichomes are horn-shaped and have surface bumps, making them clearly identifiable and distinguishable from those of other indigo-bearing plants [63]. Such trichomes were observed in the archaeological residues and identified by comparison with those observed on experimentally processed leaves (Fig 4 panels e, f; S8 Fig; S1 Video).
Although highly fragmented (widths ranging between 10–25 μm), the micro-residues that we observed had different morphologies, some appearing rounded or polygonal in section and others elongated with characteristic features (namely structural defects [64]) known as cross markings and dislocations [65–67] (Fig 4a-4c). These features were observed on the archaeological fragments, and on the residues obtained through replicative experiments for the extraction of indigotin from I. tinctoria (S7 Fig).
The small size of the recovered micro-residues made it impossible to apply conventional molecular techniques (such as HPLC or GC-MS) [68]. Hence, vibrational micro-spectroscopies, notably µ-Raman and µ-FTIR, were considered as the most suitable chemo-profiling techniques to characterise the blue coloured residues and their composition. These methods permitted micrometric-resolved, non-invasive and non-destructive analyses [68].
The blue colour on the fragments was consistently identified as indigotin by μ-Raman analysis, through comparison with two reference substances, a commercial indigotin reference standard (PhytoLab®, reported as Ref1 in Fig 5) and the blue micro-residues obtained from the processing of modern reference I. tinctoria (Fig 5d, spectrum Ref 2 and S9 Fig). The spectra of the experimental blue fragments positively matched the 25 archaeological samples investigated, all showing the typical intense µ-Raman peaks of indigotin reference standard at 1575, 545, and 253 cm-1 (Fig 5 d and S10 Fig) [68–70].
Indigotin is the main constituent of the indigo dyes and woad that derives from the hydrolysis of the glycoside precursors (primarily Isatan A and Isatan B) in the form of a dark blue crystalline secondary compound. These precursors are glycosidic by-products of the light phase of photosynthesis occurring in the chloroplasts that are stored in the vacuoles of the mesophyll cells of the leaf epidermis. Hence, only these compounds are naturally present in the indigo-bearing plants and they are released from the cells once the leaf is broken down [40,41,44,45]. The chromophore indigotin is not directly available nor visible in the plant during its natural cycle. When the leaf epidermis is pounded, tiny fragments can become entrapped within the stone porosity, in turn the precursors are released and start their transformation through enzymatic hydrolysis, oxidation, and dimerization of two indoxyl molecules (S6 Fig).
The micro-residues are characterised by cellulose (C), pectin (P), and lignin (L) with distinctive Raman signals at 378 and 1098 cm-1 (C), 850 cm-1 (P) and 1604 cm-1 (L) [71–73]. The presence of these biomolecules is reported as a diagnostic composition component for phloem elongated cell walls (including sclerenchyma fibres) [71,72]. Cellulose (C) is the polysaccharide with the highest concentration of plant fibres [71] and its prominent bands at 378 and 1098 cm-1 are observed in all the archaeological samples (Fig 6 panel d). Namely, the signal at 378 cm1 corresponds to a specific cellulose, that is polymorph I (CI), found in native cellulose fibres [74,75], whereas the peak at 577 cm-1 is typical of cellulose polymorph II (CII), differing in anhydro-glucopyranose unit conformations [74,76]. Interestingly, a minor occurrence of polymorph II, along with (CI) signals, has been reported as proof of biodeterioration induced in unprocessed plant fibres by ageing processes [72]. Moreover, the presence of lignin (L) and pectin (P) polymers, which are diagnostic features of both leaf and phloem tissues, were also observed in the µ-Raman spectra of the archaeological residues. These bands are not present in the spectra of cotton seed pod hairs [71,72,77]. The signal at 1604 cm-1 is attributed to (L) [71–73], a high-weight polyphenol that accumulates in supporting tissues during plant growth. The low intensity of the (L) peak reflects the low concentration of this polymer in the leaf cells [71]. The presence of the 850 cm-1 band, assigned to (P) [73], a heteropolymer formed by monosaccharides and uronic acids linked by ester bonds, is another diagnostic peak for phloem residues. The middle lamellae, which cement together elementary individual cells into bundles in plant phloem, are rich in these polysaccharides [73].
Our analyses provide distinctive evidence that I. tinctoria residues are characterised by detectable signatures of both lignin (L) and pectin (P) along with prominent CI peaks (see spectrum listed as Ref3 in Fig 6 d and S12 Fig). The presence of these bands, along with their relative observed intensities, matches the data obtained from the chemical profiling of the archaeological micro-residues (Figs 6 and 7 b1-2), reflecting a comparable biochemical composition [71,72]. The antiquity of these fragments is confirmed by the additional presence of the cellulose polymorph (CII) (see Fig 6 d spectra a, b and S13 Fig), absent in modern residues, which is coupled with the biodeterioration of the fibres of plant origin [72].
BF-OM view (a1, b1, c1), P-OM view (a2, b2, c2). Panel d: µ-Raman spectra of a representative archaeological blue micro-residues (Fig 7 d, red line a1-2, reprised from Fig 6 d, line a) compared with blue and non-coloured jeans fibres (Fig 7 d, black c1-2 and grey b1-2 lines, respectively). µ-Raman spectra are presented after normalisation and luminescence background removal. Panel e: reflectance µ-FTIR spectra of the above-mentioned samples. Vibrational bands characteristic of cellulose polymorph II (CII), pectin (P) and lignin (L) are reported. Details of the spectra are reported in S16 Fig.
Natural and synthetic indigotin cannot be chemically differentiated [68,69] (see also S11 Fig). Therefore, the presence of indigotin in the archaeological samples alone cannot exclude the possibility of modern contamination. As a result, we studied both morphological features and chemical signatures to exclude contamination.
I. tinctoria is not the only indigo-bearing plant, therefore we considered all the most common ones to check for modern contamination. In particular, we tested Indigofera tinctoria L., Polygonum tinctorium Aiton, Isatis indigotica Fort. and Baphicacanthus cusia Bremek, all of tropical origin [36]. Ecological and historical reports make it highly unlikely they were present in the Caucasus under the cool conditions occurring during the Upper Palaeolithic, therefore their occurrence in the archaeological context would indicate modern contamination. The trichome morphology of I. tinctoria is highly characteristic and resembles those observed in the GST residues (Fig 4 panels e,f; S8 Fig) [62,78]. Today, indigotin is widely used to colour cloth, notably in the dyeing of blue jeans, that we included in our control samples (Methods, Table 2). This common item of clothing is usually manufactured from cotton (Gossypium sp.) [68,79], a seed pod fibre native to tropical and subtropical regions, with numerous species grown commercially today. From an archaeological perspective, cotton is unlikely to be present in the context since the earliest genotyped record of Gossypium worldwide is from the site of Qasr Ibrim (Nubia, southern Egypt), Middle Nile valley and dates to around 5500 BC [80]. Cotton seed pod hairs (commonly named cotton fibres) are included in the category of unicellular trichomes; these show a smooth surface without specific ‘ornamentation’ and are ribbon-like, kidney-shaped, and appear flattish in section [81,82]. They vary in diameter and form regular twists or bends along their length, a feature known as convolution [82,83] (see Fig 7 b1-b2, c1-c2; S17 Fig).
Cotton fibres from blue and non-coloured jeans fabric were analysed using both µ-Raman and µ-FTIR spectroscopies. The recorded spectra were compared with the archaeological micro-residues (Fig 7 panels d, e). The cotton fibres sampled from jeans fabrics are composed of at least 90% cellulose, natively present as (CI), as reflected by their vibrational spectra [71,72,77] (grey profiles b1-2 in Fig 7 panels d, e). To the best of our knowledge there are no reports of lignin and pectin signals in the vibrational spectra of cotton fibres [71,72,77].
Furthermore, the µ-FTIR spectrum of the archaeological blue micro-residues exhibit a peak at 1732 cm-1, assigned to C=O stretching of ester groups (Fig 7 panel e, red line spectrum a1-2, S15 Fig). The presence of this signal, attributed to pectin [77] (as also confirmed by Raman), matches the spectrum of the experimentally processed modern I. tinctoria epidermis fragments (S12 Fig panel b), that is not observed in the jeans cotton sample (Fig 7 panel e, black line c1-2 and grey profile b1-2, respectively). The intensity of this peak is reported to be strengthened by the presence of oxycellulose in aged cellulose cell walls [77]. As far as lignin (L) is concerned, its occurrence cannot be inferred by FTIR analysis due to the presence of the broad and intense water absorption band at approximately 1630 cm-1, which overlaps with lignin signals [77] (Fig 7 panel e).
The presence of (P) and (L) in the Dzudzuana micro-residues (Figs 6 panel d and 7 panel d-e red lines a1-2, S13 and S15 Figs) confirms that the residues are not cotton, since these polymers are not present in cotton fibres. All these analyses conducted indicate that the indigotin signal measured on the 25 archaeological blue residues derives from I. tinctoria.
Recovering evidence for prehistoric use of plant secondary compounds has barely begun. Although a significant number of plants naturally contain useful secondary metabolites, many require complex processing to access these and make them usable [8]. Even in the case of food, some nutritious plants require extensive leaching, roasting and/or pounding to eliminate toxins, while extraction of useful medicinal secondary compounds demands a deep knowledge of plants since many can be both medicinal and poisonous, with only the correct processing and dosage making the difference [8]. The processing of plants cannot be simply assumed or ignored as it formed part of the complex tapestry of Palaeolithic life. Nonetheless, to demonstrate the use of plants, exhaustive analytical studies are required and studies such as those presented here, is one way to achieve this.
By considering the antiquity of the blue micro-residues under study, excluding contamination was given high priority [11,51]. Therefore, we applied a stringent procedure during both the sampling activities in the museum and then during lab-processing. All the control-tests proved negative for the presence of indigotin. Also, modern garments such as blue jeans were tested as possible contaminants and our approach proved crucial to discriminate archaeological blue micro-residues from cotton, a putative source of modern contamination.
The secondary compound identified, indigotin, is not present as such in the indigo-bearing plant and it results from the enzymatic hydrolysis of its precursors, when they are released by pounding the leaves. These are by-products of the photosynthesis; hence they are naturally contained in the cells of the leaves [40–42,44,45]. This first mechanical process is propaedeutic to the following radical cross-coupling of indoxyl, with either isatin (derived from indoxyl oxidation) or a second indoxyl unit, resulting in the formation of indirubin and indigotin chromophores respectively [36,40,44,84,85], in the form of a blue powder insoluble in water and other polar solvents.
Our experiments demonstrated that this part of the chaîne opératoire is common to all the indigo-bearing plants, Indigofera tinctoria L. (true indigo), Polygonum tinctorium Aiton (syn. Persicaria tinctoria, Japanese indigo), Baphycacantus cusia Bremek (common conehead), Isatis indigotica Fort. (Chinese woad), and Isatis tinctoria (woad). However, the latter is the only species naturally present across western Eurasia [36]. The paleogenomic signature indicates the presence of Isatis sp. at the Palaeolithic site of Aghitu-3 in Armenia (Caucasus) around 36.000 years ago [14], thus making I. tinctoria consistent with the source of the identified compound, indigotin, at Dzudzuana. Since the release of these precursors can occur only when the leaves are broken down, the presence of blue fragments found entrapped in the crevices of the Dzudzuana GSTs is unlikely to be accidental.
This study presents an innovative research design, a stringent sampling procedure, and the integrated methodological approach that has led to the identification of indigotin, a secondary compound not naturally present in the plant. In addition, the presence of this molecule provides evidence for the exclusion of contamination. However, there is no direct archaeological evidence for how the I. tinctoria plant may have been used. This observation is unsurprising, as plant remains are infrequently preserved in Palaeolithic contexts, making the recovery of direct evidence for the intentional and specific use of plants highly improbable. Indeed, throughout the entirety of the Palaeolithic record, documentation of the deliberate exploitation of particular plant species is exceedingly rare [3–6,14,20,22–25,27,28,86].
Many plants have extensive medicinal properties including anti-inflammatory, anti-tumour, antimicrobial, antiviral, analgesic, and antioxidant [38,46] and there is extensive evidence for the use of these, not only in human traditional medicine, but also across the animal kingdom. All animals, and even insects, self-medicate [27]. In the case of chimpanzees, they sometimes prepare the plants prior to consumption [87]; Sumatran orang-utans are known to apply to wounds a mashed concoction of Fibraurea tinctoria leaves [88,89]. Today, the roots of I. tinctoria and other indigo-bearing plants are used in medicine because they contain flavonoids and the leaves contain indigoid-precursor molecules that have preservative, antiseptic, repellent, and protective properties [39,40,45,46]. It is therefore entirely within the behavioural context of humans, from all Palaeolithic periods, to use plants to self-medicate.
I. tinctoria is also known as a source of indigotin, a well-established blue chromophore obtained by the oxidation of precursors naturally present in the cells of the leaves. The use of I. tinctoria to obtain a blue hue is well known, and this knowledge extends into later prehistory [36,40,90–92]. The use of this plant has been recorded as dye since Egyptian times, the earliest written source being the Papyrus Graecus Holmensis (also known as the Stockholm papyrus, retrieved in the XIX century) [93]. However, while the possibility exists that I. tinctoria was transformed into woad dye and used during the Early Upper Palaeolithic, there is currently no archaeological evidence for this.
However, and more broadly, colour was a part of Upper Palaeolithic life and the use of mineral pigments is well known, in particular in rock art where red, yellow, white and black are present across the Eurasian continent and the Indonesian archipelago from around 40,000 years ago [94–96]. Blue is a relatively rare colour in nature and to the best of our knowledge, blue pigment (mineral-based) from Palaeolithic contexts has only been reported for Siberian figurines [97].
Plant-based dyes are first reported from the Natufian Epipalaeolithic layers of Kebara cave (~13,500–11,650), where red was extracted from Rubia tinctoria roots [92,98]. Seeds of I. tinctoria were first recorded in 5000-year-old Neolithic sites in France (e.g. Bouches-du-Rhône) [90,91], but to date, there is no archaeological evidence for the use of I. tinctoria prior to the Iron Age.
All the techniques and the equipment employed are summarized in Table 1.
Tomographic reconstructions, including phase retrieval pre-processing, were carried out using the STP (SYRMEP Tomo Project) software [99]. Different values of the ring removal filter and of δ/β ratio for phase retrieval correction were used, depending on the characteristics of the samples. The 32-bit.tiff reconstructed image stacks were then converted to 16-bit in order to process the datasets with available imaging software like Fiji (ImageJ), Dragonfly, CTAn and VGStudioMax.
A variety of microscopy techniques (see Table 1) were employed to inspect the archaeological tools, the reference artefacts, and residues.
μ-Raman and μ-FTIR measurements were always performed before SEM observations since metallic coating of the micro-residues or moulds’ surface interferes with these spectroscopic analyses. The µ-Raman and μ-FTIR were used to characterise both the blue and non-coloured biogenic micro residues, as well as reference materials (for spectra collection parameters see S3 and S4 Tables), since they are non-invasive and non-destructive and guarantee micrometric spatial resolution. Furthermore, vibrational spectroscopies enabled the collection of information on the molecular composition of the leaf epidermis fragments’ matrix in which the dye is adsorbed.
Raman spectroscopy is known as a powerful method for the identification of dyes, in particular when the laser exciting light is in resonance or pre-resonance with the electronic transitions of the dyestuff [100]. After initial tests performed with different exciting lines (514, 633 and 785 nm), including an extended spectral range (100–4000 cm-1), it was decided that the best conditions to characterise both dye and micro-residues was to excite them at 785 nm, focusing on the most intense vibrational modes between 200 and 1800 cm-1. Furthermore, preliminary investigation (live mode by means of oscilloscope) was conducted on at least three different spots of each single micro-residue. Subsequently, the most significant region was selected for longer inspection, ensuring a good signal-to-noise ratio of the spectrum.
The measurements were carried out with a μ-Raman Renishaw InVia spectrometer (see Table 1) and a 4 cm-1 spectral resolution was achieved by this setup. The spectrometer was coupled with a Leica DM LM optical microscope. Spectra were collected in a backscattering configuration, through a 50× objective (NA: 0.75) and with power on samples below 150 mW, which prevents fibres matrix degradation; on the other hand, when laser was focused on a blue dyed region, power was kept below 15 mW to prevent indigotin photooxidation.
FTIR spectroscopy is a complementary vibrational technique and was used to corroborate and, eventually, add information to the μ-Raman results [101]. Indeed, FTIR does not suffer from photo-induced luminescence emission interference as Raman does. µ-FTIR measurements were performed with a Bruker Lumos II FT-IR microscope, equipped with a LN-MCT (liquid nitrogen cooled – HgCdTe) high performance detector. Spectral resolution was set to 4 cm-1. Measurements were carried out in reflectance mode, to ensure a contactless approach, whereas background contribution was collected by the reflected radiation impinging on a gold film. Basic data manipulation (Fourier transform of interferograms) was performed by Opus 8.8 software and the obtained spectra plotted by means of OriginPro 2021 software.
Collecting the leaves of I. tinctoria at different time periods was necessary as the content of the precursors and the relative final compound, indigotin, vary across the year [102]. The leaves are richer in compounds during spring, but the content can also vary according to the external climatic conditions (humidity and drought). The plants were sourced in the rosette stage in the Site of Community Importance (SIC) and Special Area of Conservation (ZSC) of Marano di Valpolicella (IT3210002, Verona, IT). The identification was made by LL under the guidance of Daniele Zanini, Scientific Director of the Orto Botanico di Novezzina (Ferrara di Monte Baldo, Verona, IT).
Because indigotin is a chromophore obtained from pounded leaves, a series of replicative field experiments were carried out [37,103] at Corte Badin (Verona Italy). Details of the processing procedure can be seen in the video (see S1 Video). In order to perform the replicative experiments, the selected stone pebble was cleaned in the lab by soaking it in a beaker for 30 minutes in different solutions: HCl (5%) and, H₂O₂ (5%). Each treatment was followed by sonication in ultrapure water for 30 minutes. Moulds of the working surfaces were then taken at time zero (T₀). Experiments were carried out at different times of the year according to the I. tinctoria life cycle [37,102]: in early spring (before flowering), summer, and early autumn. 800 g of leaves, cut from the first year (rosette stage), were immediately processed using the stone pebble in a wooden mortar for approximately 30–40 minutes. The pestle was reused in the subsequent replicative experiments, without being washed, and stored in a sealed bag until the next use. The mechanical processing was characterised by a combination of vertical action (such as pounding) and horizontal linear motion that shredded the leaves, with these actions repeated until a poultice was formed. All steps of the process were photographed and video-recorded (S1 File; S1 Video). Our experiments have shown that the poultice tends to overflow from the direct contact area and adhere to the hand operating the stone, meaning the residues can easily spread across multiple areas, not necessarily limited to those in direct contact with the working area (S1 Video). Fragments of the leaves remained trapped in the uneven surface topography of the stone.
To exclude the presence of other indigo-bearing plants, we tested the most common ones [36], and in particular Indigofera tinctoria L., Polygonum tinctorium Aiton, Isatis indigotica Fort. and Baphicacanthus cusia Bremek. The plants were sourced from a Chinese Pharmacy shop in Hangzhou, China, and were collected from regions in subtropical to tropical climates namely Zhejiang Province, Hunan Province, Anui Province, and Guangdong Province respectively.
Indigotin is widely used to colour textiles, in particular in the dyeing of blue jeans, a common item of clothing usually manufactured from cotton [68], a seed pod hair. In particular, upland cotton (Gossypium hirsutum [79,81,82]) accounts for 90% of modern production [104]. Thus, to ensure that the presence of archaeological micro-residues is not the result of modern contamination, cotton fibres from blue and non-coloured jeans fabric were sourced. Morphological and vibrational analyses performed on these fibres were compared to the archaeological micro-residues.
Normalised Raman spectra of natural woad dye extracted from modern I. tinctoria leaves according to hot-water extraction (see S1 File), compared with synthetic indigotin standard. Both were excited at 785 nm and they are shown after luminescence background removal.
Normalised Raman spectra of archaeological non-coloured micro fragments from Dzu S6 m3 (a) and Dzu S1 m7 (b), reprised from Fig 6, excited at 785 nm and after luminescence background removal. Characteristic bands of cellulose (C), including its polymorphs I (CI) and II (CII), pectin (P) and lignin (L) are indicated by their Raman shift.
µ-Raman spectra of blue and non-coloured archaeological fragments. (a): spectrum of a non-coloured region of a blue residue; spectra b and c of non-coloured residues; spectra Ref3 of non-coloured I. tinctoria fibre obtained from the processing of modern leaves (reported as reference). (b and c): zoomed region of interest of the spectra reported in (a). This figure supports Fig 6 d of the main text.
µ-Raman spectra of blue and modern fibres. (a): spectrum a1-a2 of blue archaeological residues, while spectrum b1-b2 is a modern non-coloured jeans fibres reported for comparison. (b and c): zoomed region of interest of the spectra reported in (a). This figure supports Fig 7 d of the main text.
Sections
"[{\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref001\", \"pone.0321262.ref002\", \"pone.0321262.ref003\", \"pone.0321262.ref005\", \"pone.0321262.ref006\", \"pone.0321262.ref007\", \"pone.0321262.ref008\"], \"section\": \"Introduction\", \"text\": \"Modern humans (Homo sapiens) first appear in the archaeological record around 300,000 years ago, in Africa. Most of the evidence for their cognitive and technological abilities is based on recovered assemblages of chipped stone artefacts and animal bones since these endure far longer in the archaeological record than plants [1]. Accordingly, the Palaeolithic narrative centres primarily on animal hunting and stone tool manufacture. Perishable materials, the so-called \\u201cmissing majority\\u201d [2], notably plants for which there is growing evidence for their use as food [3\\u20135], string and cordage [6], weaving [7] and medicine [8], are largely missing\\u00b8 creating a partial narrative. There is therefore, a need to identify and demonstrate the use of plants and the roles they played in a wide range of activities, many of which may still be unknown. Ultimately, this will contribute to a broader understanding of Palaeolithic behaviour.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref009\", \"pone.0321262.ref010\", \"pone.0321262.ref011\", \"pone.0321262.ref012\", \"pone.0321262.ref013\", \"pone.0321262.ref016\"], \"section\": \"Introduction\", \"text\": \"Finding direct traces of the deliberate exploitation of plant materials is complex since macroscopic un-charred plant remains rarely survive through Palaeolithic timescales and, even when they do, their presence can be argued as natural deposits [9] or the result of bioturbation or other taphonomic events [10]. Therefore, demonstrating use of plants during human evolution requires not only conditions that are favourable to the long-term survival of plant remains, but also needs to be exhaustive, as the finds must indicate human involvement, and also offer stringent evidence to negate natural occurrence and modern contamination [11]. A lack of awareness of the potential for complex plant processing can compromise evidence for this during excavation and post-excavation processes [12]. Finally, the recent application of high-resolution imaging and biomolecular techniques to the investigation of biogenic residues from the micro to the nanoscale [13\\u201316], is now expanding the potential for detection of a new range of evidence for the use of plants and the likely behaviour associated with this. Use of these techniques can feed into a better understanding of the complexity of human-plant relationships and expand perception of plant processing. Each time a new piece of secure information emerges it provides an extraordinary insight into our prehistoric human past and contributes to a growing awareness that our Palaeolithic ancestors had sophisticated ecological knowledge of plants.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref003\", \"pone.0321262.ref017\", \"pone.0321262.ref019\", \"pone.0321262.ref020\", \"pone.0321262.ref021\", \"pone.0321262.ref004\", \"pone.0321262.ref005\", \"pone.0321262.ref022\", \"pone.0321262.ref026\", \"pone.0321262.ref008\", \"pone.0321262.ref027\", \"pone.0321262.ref028\", \"pone.0321262.ref029\", \"pone.0321262.ref030\", \"pone.0321262.ref008\", \"pone.0321262.ref031\", \"pone.0321262.ref032\", \"pone.0321262.ref033\", \"pone.0321262.ref034\", \"pone.0321262.ref035\"], \"section\": \"Introduction\", \"text\": \"While plant processing, including cooking and grinding, has been recognised for some time for the Upper Palaeolithic [3,17\\u201319], the evidence for modification of plant materials has, until recently, been largely restricted to identification of macro remains, notably carbonised materials and wood [20,21], likely food items [4,5] or pitch although this is generally identified by biomolecular analysis [22\\u201326]. Plants represent an infinite resource that provide humans not only with food and raw materials but also, due to their biomolecular complexity, with medicines, poisons, flavours, aromas, hallucinogens and dyes [8]. The ingestion of non-nutritional plants containing medicinal secondary metabolites was identified in 47,000-year-old Neanderthal dental calculus [27,28], while tentative evidence for poison 40,000 years ago was recovered from Border Cave, South Africa [29]. Medicinal plants are reported from a number of Palaeolithic sites in the Caucasus [30]; however, it is challenging to demonstrate that these were ingested and/or intentionally processed [8]. To date, there is no evidence for the extraction of dyes from organic materials in the Palaeolithic; the known colourants (red, yellow, black and white) are all pigments of mineral origin apart from charcoal [31]. They are highly resistant to ageing, with little apparent degradation and known to have been used in Palaeolithic art and for other purposes. For example, ochre is known in various applications such as tanning leather or skin [32], as a preservative [33,34], as insect deterrent and as skin protection [35].\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref003\", \"pone.0321262.ref017\", \"pone.0321262.ref019\", \"pone.0321262.ref012\", \"pone.0321262.ref016\", \"pone.0321262.ref019\"], \"section\": \"Introduction\", \"text\": \"One source of evidence for the use and intentional processing of plants comes from Ground Stone Tools (GST) [3,17\\u201319]. The use-wear traces on their utilized surfaces and the extraction of micro-residues trapped within their surface are compared to modern reference materials to reconstruct the type of materials being worked. These residues can be extracted, imaged and chemo-profiled [12,16,19].\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref036\", \"pone.0321262.ref037\"], \"section\": \"Introduction\", \"text\": \"Here, we share the robust data derived from the chemical-physical analyses that led to the identification of the indigotin molecule, which has not previously been recognised in a Palaeolithic context. We also investigate the processes that may have led to the presence of the blue residues on the stone pebbles. To address this, we examined possible accidental or deliberate reasons for their occurrence as part of the mechanical processing of leaves, including whether the effect of releasing the precursors of indigotin - a well-documented practice in both ethnographic and historical records [36,37] - could have been deliberate. This paper is therefore focused on the recovery and identification of indigotin as an identifying feature of Isatis tinctoria L., extracted from the working surface of these archaeological GSTs. Finally, we provide an exhaustive explanation of the method used for extraction of the residues to ensure the samples were free of contamination during this process. While we present some preliminary use-wear observations to support the actual use of these pebbles, an exhaustive functional analysis will be the object of a separate study.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref038\", \"pone.0321262.ref041\", \"pone.0321262.ref042\", \"pone.0321262.ref043\", \"pone.0321262.ref036\", \"pone.0321262.ref014\"], \"section\": \"Introduction\", \"text\": \"Indigotin derives from the processing of indigoid plants, in particular species within the Isatis, Indigofera, and Persicaria genera. When the leaves are crushed the precursors are released and the fermentation process begins [38\\u201341]. Among these, I. tinctoria, a non-nutritional plant, is the only species native to Central Asia [42,43], whereas all the other species are of tropical origin and do not naturally grow in cool\\u00a0environments [36]. Metabarcoding techniques applied to the sediments from Aghitu 3 cave in Armenia (39\\u201324,000 years old) identified Isatis sp., hence providing genetic evidence that the genus was present during this period in the Caucasus [14].\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref040\", \"pone.0321262.ref041\", \"pone.0321262.ref044\", \"pone.0321262.ref039\", \"pone.0321262.ref045\", \"pone.0321262.ref046\"], \"section\": \"Introduction\", \"text\": \"I. tinctoria, colloquially known as dyer\\u2019s woad, is a biennial plant of the Brassicaceae family that grows in dry and sunny locations. This plant is a source of indoxyl glycosides - precursors of the blue chromophore indigotin and its red isomer indirubin - naturally contained in cell vacuoles of the leaf epidermis, together with a minor quantity of flavonoids [40,41,44]. It is also well known for its medicinal properties due to the significant presence of further indole alkaloids in its roots and leaves [39,45,46].\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref003\", \"pone.0321262.ref016\", \"pone.0321262.ref018\", \"pone.0321262.ref019\", \"pone.0321262.ref047\", \"pone.0321262.ref048\", \"pone.0321262.ref049\", \"pone.0321262.ref036\", \"pone.0321262.ref038\"], \"section\": \"Introduction\", \"text\": \"We developed an innovative and robust research design that integrates synchrotron-based micro-CT scanning (to demonstrate the porosity of rocks enabling retention of residues), microscopy at various resolutions (to document the presence of residues and their correlation with modifications to the stone surface caused by mechanical processing), and spectroscopy techniques, in particular Raman and FTIR (Fourier transform infrared)\\u00a0(to characterise residues and the chemical profile of their constituent molecules). We demonstrate that I. tinctoria was pounded at Dzudzuana by Homo sapiens, 34,000 years ago, using unmodified river pebbles. Unmodified pebbles are recognised as tools in the Upper Palaeolithic, although to date they have been associated with the processing of edible plants [3,16,18,19,47] and inorganic materials such as ochre [48,49]. We do not claim to understand why this plant was being ground up, however we acknowledge that it is well known today for its medicinal and blue dye applications [36,38]. By demonstrating that plant processing was not restricted to edible plants, our research provides a new dimension to human behaviour and contributes significantly to the growing body of evidence highlighting the complexity of human-plant interactions in the Upper Palaeolithic.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.g001\", \"pone.0321262.ref050\", \"pone.0321262.g001\", \"pone.0321262.ref050\", \"pone.0321262.g002\", \"pone.0321262.s005\", \"pone.0321262.s001\"], \"section\": \"Materials\", \"text\": \"Dzudzuana cave, Imereti (Western Georgia), is an Upper Palaeolithic site (Fig 1) comprising three discrete Palaeolithic units spanning ca. 20,000 years of occupation, the lowest unit (D) dating to 34.5\\u201332.2 cal. BP [50] (Fig 1). The archaeological material from this unit comprised chipped stone assemblages, faunal remains, bone tools including a needle, as well as engraved and pierced animal teeth likely to have been used as ornamental artefacts [50]. In addition, the findings included six river pebbles which are the focus of the present study (Fig 2; S1 Fig and S1 Table). Five of these pebbles provide clear evidence of use in the form of functionally active areas and associated micro-residues.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.g002\", \"pone.0321262.ref050\"], \"section\": \"Area under study.\", \"text\": \"Left: Location of Dzudzuana cave (modified after Google Earth) and cave entrance (Image by LL). Right: map of the cave with the excavated area inside the red circles (redrawn and modified from Bar-Yosef et al. 2011, p. 334, Fig 2\\u00a0[50]), red squares show where the six pebbles were retrieved from. Excavations in squares G-7, G-8 and I-8 reached the full depth of the archaeological sequence. These are also the squares where the stone pebbles were retrieved.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref050\", \"pone.0321262.ref014\", \"pone.0321262.s005\", \"pone.0321262.s007\", \"pone.0321262.t002\"], \"section\": \"Materials\", \"text\": \"During this period, several climatic fluctuations occurred; however, during the human occupation, the climate at Dzudzuana cave was generally mild, with species such as wild grape (Vitis sylvestris), hazel (Corylus sp.), oak (Quercus sp.), beech (Fagus sylvatica), birch (Betula sp.) and Artemisia sp. all present in the Caucasus [50]. Moreover, several Brassicaceae species, among which Isatis sp., was detected by metabarcoding [14]. The present study of residues adhering to GSTs supported by preliminary use-wear analysis, demonstrates the purposeful use of these stone pebbles, the archaeological integrity of the material recovered and its accurate identification (Methods, S1 and S3 Figs, Table 2).\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.t002\", \"pone.0321262.ref011\", \"pone.0321262.ref012\", \"pone.0321262.ref016\", \"pone.0321262.ref018\", \"pone.0321262.ref051\"], \"section\": \"Stone pebbles biography and taphonomic history\", \"text\": \"Here we report on the methodological sequence during the different stages of the Dzudzuana stone pebbles analysis. The sequential sampling procedure applied is summarized in Table 2 (Methods) and follows the procedures previously established and published by the authors [11,12,16,18,51].\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.s001\", \"pone.0321262.g001\"], \"section\": \"Processing during field-work\", \"text\": \"Six stone tools were identified during several archaeological excavation campaigns (2002\\u20132007, details are reported in S1 Table). The pebbles were retrieved in different squares (G-7, G-8, I-8; Fig 1) from Unit D, excavated by 5 or 10\\u2009cm levels. Three pebbles (Dzu S1 to Dzu S3) derive from the deepest portion of Unit D, excavated in 2002, whereas three samples (Dzu S4 to Dzu S6) come from the uppermost part of the same unit, excavated in 2007. Only two pebbles are complete (Dzu S3 and Dzu S4). Dzu S1, Dzu S2, Dzu S5, and Dzu S6 are fragments of larger pebbles; Dzu S5 still retains part of the cortical surface and Dzu S6 is a mid-segment of a pebble. During field-work after their retrieval, each pebble was summarily rinsed in running water. Each stone was inventoried in the log book of the excavation and labelled according to museum accession number, and finally placed individually in a labelled paper bag. The stones were then taken to the Georgian National Museum (GNM, Tbilisi).\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.t002\"], \"section\": \"Sampling in the Georgian National Museum\", \"text\": \"The paper bags containing the stones pebbles were dusted on the outside, and the sweepings were retained (see Methods, Table 2). Then, each pebble was carefully pulled out, very softly dusted with a clean brush, and the collected dust was also retained for cross-control.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.g002\", \"pone.0321262.ref012\", \"pone.0321262.ref018\", \"pone.0321262.ref052\", \"pone.0321262.ref054\", \"pone.0321262.ref047\", \"pone.0321262.ref053\", \"pone.0321262.ref056\", \"pone.0321262.ref056\"], \"section\": \"Laboratory of Palynology of the Palaeoanthropology and Paleobiology research institute of the GNM\", \"text\": \"Once dry, selected areas of the stone surfaces, comprising visually evident functionally active areas, were moulded (see Fig 2 for the placement of the moulds) following a well-established procedure [12,18,52\\u201354]. Moulding is a long-accepted method used widely by functional analysis specialists to obtain detailed negative copies of the surface, especially useful in contexts where necessary facilities or equipment to conduct full analyses are unavailable [47,53\\u201356]. The moulding material is a silicone-based dental impression polymer, polyvinyl siloxane (PVS) (Coltene\\u00ae; Speedex, Universal Activator, with linear shrinkage = -0.2%) [56].\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref011\", \"pone.0321262.ref012\", \"pone.0321262.ref018\", \"pone.0321262.ref051\", \"pone.0321262.ref054\", \"pone.0321262.g003\", \"pone.0321262.s010\", \"pone.0321262.ref012\", \"pone.0321262.ref016\", \"pone.0321262.ref018\", \"pone.0321262.s006\", \"pone.0321262.s010\"], \"section\": \"Laboratory of Palynology of the Palaeoanthropology and Paleobiology research institute of the GNM\", \"text\": \"Moulds of dorsal and lateral sides of the stones, seemingly without evidence of use, were also taken for cross reference. Two successive moulds were taken at each location, with the first removing any remaining surface sediment while the second recovered residues entrapped deep within the stone cracks as per the established method [11,12,18,51,54] (Fig 3, S4 Fig). Each mould was immediately bagged individually. Due to the PVS peel-off effect, the moulds can also dislodge residues entrapped within the pockets (pits or cracks) of the coarse stone surfaces [12,16,18] allowing the recovery of different micro-residues entrapped within these (S2 and S4 Figs).\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.s006\"], \"section\": \"Processing at DAIS under lab-controlled conditions\", \"text\": \"At DAIS, moulds were initially examined under a fume hood. The mould observation revealed the presence of some residues still adhering to the stones in both the first and second moulds, suggesting that they remained entrapped deep inside the pockets on the surface of the pebbles (S2 Fig). The first mould removed the residues nearest the surface, while the second mould extracted the deeper ones. Therefore, the residues extracted with the second mould had been protected from exposure to outside elements, and it is for this reason that we examined only their internal (2nd) moulds.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref012\", \"pone.0321262.ref057\", \"pone.0321262.t001\"], \"section\": \"Processing at DAIS under lab-controlled conditions\", \"text\": \"The 2nd moulds from the five relevant pebbles (Dzu S1, Dzu S2, Dzu S3, Dzu S5 and Dzu S6) were subsampled and cut in half. One half was immediately stored for analysis of wear-traces, while each other half was individually placed in a test-tube with 40\\u2009ml of ultrapure water. The tubes, set in an ice bath, were sonicated using an inserted ultrasound probe (UP-200S Hielscher Ultrasonics GmbH, Germany) operating at 100 W for 10\\u2009min (0.5 cycle, 80% amplitude) for cycles of 3 minutes (for a total of 12 minutes). The ultrasound probe was carefully cleaned before and after each sonication. The obtained samples were then processed according to the established procedures [12,57], and the extracted micro residues underwent imaging and chemo-profiling (Methods, Table 1).\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.t002\"], \"section\": \"Control samples\", \"text\": \"Different control samples were collected during the sampling process at GNM and a stringent procedure was followed to avoid possible contamination (Methods, Table 2). During all stages of the processing, Petri dish traps were placed on the desk, on the shelves, and in the fume hood, until the stone pebbles sampling was finalized. Then, all the control samples were sealed in vials and shipped to DAIS, where they underwent optical microscopy for blue micro-residues. Control samples of dust swept from the surface of the paper bags and stone pebble surfaces were also collected.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref058\"], \"section\": \"Results\", \"text\": \"The analysis of thin sections from the archaeological stone pebbles indicated that these are mafic igneous rock. However, since the available fragments from some of the archaeological pebbles were too small to obtain significant data on the overall stone structure, the full petrographic analysis (being invasive) was performed on pebbles collected from the adjacent Nikrisi River that crossed the limestone massifs of the Zemo Imereti Plateau [58]. The analyses of the thin sections from three river pebbles confirmed that these were epidosite and micritic limestones, consistent with the petrographic analysis of the archaeological pebbles.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.g003\", \"pone.0321262.g003\"], \"section\": \"Results\", \"text\": \"X-ray Computed \\u03bc-Tomography aided with synchrotron brilliance (SR-\\u03bcCT) was used to analyse the porosity of the stone pebbles collected in the river. The SR-\\u03bcCT showed significant internal porosity on these igneous rocks (Fig 3 panels a-c). Moreover, microscopic study of their surface texture, investigated by means of optical microscopy (OM) and scanning electron microscopy (SEM), demonstrate that the stone is rich in irregularities and in particular microcavities where residues can become trapped (Fig 3 panels e-f).\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref059\", \"pone.0321262.ref060\", \"pone.0321262.s023\"], \"section\": \"Results\", \"text\": \"Based on an extensive experimental programme, we can confirm that micrometric residues can become entrapped in this type of pores during the pounding of the leaves [59,60] (S1 Video).\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.g002\", \"pone.0321262.ref011\", \"pone.0321262.s006\", \"pone.0321262.g003\", \"pone.0321262.g003\", \"pone.0321262.s007\", \"pone.0321262.s007\", \"pone.0321262.g003\", \"pone.0321262.ref059\", \"pone.0321262.g003\", \"pone.0321262.g003\", \"pone.0321262.s007\"], \"section\": \"Results\", \"text\": \"Selected moulds from Dzu S1, Dzu S2, Dzu S3, Dzu S5, and Dzu S6 (Fig 2) were analysed according to established use-wear procedures [11] (S2 Fig), by means of OMFig 3g-3i), SEM (S3 Fig panels e,f), Laser Scanning Confocal Microscopy (LSCM, S3 Fig panels a-d) and confocal profilometry (Fig 3 panels l-n) [59]. The most common features observed are polished and flattened areas, scratches and striations (Fig 3g-3l, S3 Fig).\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref057\", \"pone.0321262.s002\", \"pone.0321262.s010\", \"pone.0321262.g004\", \"pone.0321262.g005\", \"pone.0321262.s011\", \"pone.0321262.g006\", \"pone.0321262.s012\"], \"section\": \"The micro-residues morphology\", \"text\": \"The solution obtained through sonication of the 2nd moulds from five of the six the analysed stone tools was processed to extract micro-residues according to Cagnato and Ponce [57] (S2 Table, S4 Fig). Being aware of potential contamination issues and false positives, we deliberately decided to analyse the micro-residues from the 2nd moulds, considered less prone to contamination. The morphological analysis of the micro-residues was carried out using OM, LSCM and SEM. We recovered 67 individual micro-residues (both coloured and non-coloured) that we were able to identify as of plant-based origin. Of these, 25 individual micro-residues were blue (Figs 4, 5 panels a1-a2, b1-b2, c1-2 and 6a, S5 Fig), while 42 were non-coloured (Fig 6 panels b, c, S6 Fig). Here, we focus on the results of the physical-chemical characterization of the 25 micro-residues that showed a blue colour.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.s018\"], \"section\": \"Micro-residues retrieved from the Dzudzuana pebbles.\", \"text\": \"OM, LSCM and SEM images (a, b, c), and \\u00b5-Raman spectra (d). Panel a: blue micro-fragments, retrieved from Dzu S5, exhibiting non-coloured regions (OM). Panels b-c: vascular elements retrieved from Dzu S6; b: vascular tissue with aligned pits (OM) and a zoomed region as observed by SEM, retrieved from Dzu S5; c: fusiform (spindle-shape) structure tapering at each end (LSCM) and the nodes displayed in the insert (OM) retrieved from Dzu S5. Panel d: \\u00b5-Raman spectra of the micro-residues shown in the micrographs (a: spectrum of a non-coloured region of the blue fragments; b and c: spectra of non-coloured micro-residues), compared with non-coloured I. tinctoria fragments obtained from the processing of modern leaves (Ref3). The spectra have been normalised and the luminescence background has been removed. Characteristic bands of constituting molecules such as cellulose (C), and characteristic signals of its polymorphs (CI, CII), pectin (P), lignin (L) and indigotin (Ind) are indicated by their \\u00b5-Raman shift. Details of the spectra are reported in S14 Fig.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref061\", \"pone.0321262.ref062\", \"pone.0321262.ref063\", \"pone.0321262.g004\", \"pone.0321262.s014\", \"pone.0321262.s023\"], \"section\": \"The micro-residues morphology\", \"text\": \"On the basis of their anatomical structure, the observed residues are identified as small fragments, primarily of leaf epidermis. Esau [61] demonstrates that the structure of the stem, leaf, and root is quite similar, whereas the arrangement of vascular and ground tissues (parenchyma and collenchyma) is distinctive being permeated by a network of interconnected veins that diffuse extensively across the mesophyll and may also be organised into bundles. In the leaf, the outer layer of epidermis forms trichomes [62], a specific type of hair, that can be highly distinctive in its shape and surface ornamentation. I. tinctoria trichomes are horn-shaped and have surface bumps, making them clearly identifiable and distinguishable from those of other indigo-bearing plants [63]. Such trichomes were observed in the archaeological residues and identified by comparison with those observed on experimentally processed leaves (Fig 4 panels e, f; S8 Fig; S1 Video).\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref064\", \"pone.0321262.ref065\", \"pone.0321262.ref067\", \"pone.0321262.g004\", \"pone.0321262.g004\", \"pone.0321262.s013\"], \"section\": \"The micro-residues morphology\", \"text\": \"Although highly fragmented (widths ranging between 10\\u201325 \\u03bcm), the micro-residues that we observed had different morphologies, some appearing rounded or polygonal in section and others elongated with characteristic features (namely structural defects [64]) known as cross markings and dislocations [65\\u201367] (Fig 4a-4c). These features were observed on the archaeological fragments, and on the residues obtained through replicative experiments for the extraction of indigotin from I. tinctoria (S7 Fig).\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref068\", \"pone.0321262.ref068\"], \"section\": \"Chemo-profiling the micro-residues\", \"text\": \"The small size of the recovered micro-residues made it impossible to apply conventional molecular techniques (such as HPLC or GC-MS) [68]. Hence, vibrational micro-spectroscopies, notably \\u00b5-Raman and \\u00b5-FTIR, were considered as the most suitable chemo-profiling techniques to characterise the blue coloured residues and their composition. These methods permitted micrometric-resolved, non-invasive and non-destructive analyses [68].\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.g005\", \"pone.0321262.g005\", \"pone.0321262.s015\", \"pone.0321262.g005\", \"pone.0321262.s016\", \"pone.0321262.ref068\", \"pone.0321262.ref070\"], \"section\": \"Chemo-profiling the micro-residues\", \"text\": \"The blue colour on the fragments was consistently identified as indigotin by \\u03bc-Raman analysis, through comparison with two reference substances, a commercial indigotin reference standard (PhytoLab\\u00ae, reported as Ref1 in Fig 5) and the blue micro-residues obtained from the processing of modern reference I. tinctoria (Fig 5d, spectrum Ref 2 and S9 Fig). The spectra of the experimental blue fragments positively matched the 25 archaeological samples investigated, all showing the typical intense \\u00b5-Raman peaks of indigotin reference standard at 1575, 545, and 253\\u2009cm-1 (Fig 5 d and S10 Fig) [68\\u201370].\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref040\", \"pone.0321262.ref041\", \"pone.0321262.ref044\", \"pone.0321262.ref045\", \"pone.0321262.s012\"], \"section\": \"Chemo-profiling the micro-residues\", \"text\": \"Indigotin is the main constituent of the indigo dyes and woad that derives from the hydrolysis of the glycoside precursors (primarily Isatan A and Isatan B) in the form of a dark blue crystalline secondary compound. These precursors are glycosidic by-products of the light phase of photosynthesis occurring in the chloroplasts that are stored in the vacuoles of the mesophyll cells of the leaf epidermis. Hence, only these compounds are naturally present in the indigo-bearing plants and they are released from the cells once the leaf is broken down [40,41,44,45]. The chromophore indigotin is not directly available nor visible in the plant during its natural cycle. When the leaf epidermis is pounded, tiny fragments can become entrapped within the stone porosity, in turn the precursors are released and start their transformation through enzymatic hydrolysis, oxidation, and dimerization of two indoxyl molecules (S6 Fig).\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref071\", \"pone.0321262.ref073\", \"pone.0321262.ref071\", \"pone.0321262.ref072\", \"pone.0321262.ref071\", \"pone.0321262.g006\", \"pone.0321262.ref074\", \"pone.0321262.ref075\", \"pone.0321262.ref074\", \"pone.0321262.ref076\", \"pone.0321262.ref072\", \"pone.0321262.ref071\", \"pone.0321262.ref072\", \"pone.0321262.ref077\", \"pone.0321262.ref071\", \"pone.0321262.ref073\", \"pone.0321262.ref071\", \"pone.0321262.ref073\", \"pone.0321262.ref073\"], \"section\": \"Chemo-profiling the micro-residues\", \"text\": \"The micro-residues are characterised by cellulose (C), pectin (P), and lignin (L) with distinctive Raman signals at 378 and 1098\\u2009cm-1 (C), 850\\u2009cm-1 (P) and 1604\\u2009cm-1 (L) [71\\u201373]. The presence of these biomolecules is reported as a diagnostic composition component for phloem elongated cell walls (including sclerenchyma fibres) [71,72]. Cellulose (C) is the polysaccharide with the highest concentration of plant fibres [71] and its prominent bands at 378 and 1098\\u2009cm-1 are observed in all the archaeological samples (Fig 6 panel d). Namely, the signal at 378\\u2009cm1 corresponds to a specific cellulose, that is polymorph I (CI), found in native cellulose fibres [74,75], whereas the peak at 577\\u2009cm-1 is typical of cellulose polymorph II (CII), differing in anhydro-glucopyranose unit conformations [74,76]. Interestingly, a minor occurrence of polymorph II, along with (CI) signals, has been reported as proof of biodeterioration induced in unprocessed plant fibres by ageing processes [72]. Moreover, the presence of lignin (L) and pectin (P) polymers, which are diagnostic features of both leaf and phloem tissues, were also observed in the \\u00b5-Raman spectra of the archaeological residues. These bands are not present in the spectra of cotton seed pod hairs [71,72,77]. The signal at 1604\\u2009cm-1 is attributed to (L) [71\\u201373], a high-weight polyphenol that accumulates in supporting tissues during plant growth. The low intensity of the (L) peak reflects the low concentration of this polymer in the leaf cells [71]. The presence of the 850\\u2009cm-1 band, assigned to (P) [73], a heteropolymer formed by monosaccharides and uronic acids linked by ester bonds, is another diagnostic peak for phloem residues. The middle lamellae, which cement together elementary individual cells into bundles in plant phloem, are rich in these polysaccharides [73].\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.g006\", \"pone.0321262.g006\", \"pone.0321262.g007\", \"pone.0321262.ref071\", \"pone.0321262.ref072\", \"pone.0321262.g006\", \"pone.0321262.s017\", \"pone.0321262.ref072\"], \"section\": \"Chemo-profiling the micro-residues\", \"text\": \"Our analyses provide distinctive evidence that I. tinctoria residues are characterised by detectable signatures of both lignin (L) and pectin (P) along with prominent CI peaks (see spectrum listed as Ref3 in Fig 6 d and S12 Fig). The presence of these bands, along with their relative observed intensities, matches the data obtained from the chemical profiling of the archaeological micro-residues (Figs 6 and 7 b1-2), reflecting a comparable biochemical composition [71,72]. The antiquity of these fragments is confirmed by the additional presence of the cellulose polymorph (CII) (see Fig 6 d spectra a, b and S13 Fig), absent in modern residues, which is coupled with the biodeterioration of the fibres of plant origin [72].\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.g007\", \"pone.0321262.g006\", \"pone.0321262.g007\"], \"section\": \"Comparison of blue archaeological micro-residues (a1-a2) with modern non-coloured (b1-b2) and blue (c1-c2) jeans fibres.\", \"text\": \"BF-OM view (a1, b1, c1), P-OM view (a2, b2, c2). Panel d: \\u00b5-Raman spectra of a representative archaeological blue micro-residues (Fig 7 d, red line a1-2, reprised from Fig 6 d, line a) compared with blue and non-coloured jeans fibres (Fig 7 d, black c1-2 and grey b1-2 lines, respectively). \\u00b5-Raman spectra are presented after normalisation and luminescence background removal. Panel e: reflectance \\u00b5-FTIR spectra of the above-mentioned samples. Vibrational bands characteristic of cellulose polymorph II (CII), pectin (P) and lignin (L) are reported. Details of the spectra are reported in S16 Fig.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref068\", \"pone.0321262.ref069\", \"pone.0321262.s015\"], \"section\": \"Testing for blue modern sources of contamination\", \"text\": \"Natural and synthetic indigotin cannot be chemically differentiated [68,69] (see also S11 Fig). Therefore, the presence of indigotin in the archaeological samples alone cannot exclude the possibility of modern contamination. As a result, we studied both morphological features and chemical signatures to exclude contamination.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref036\", \"pone.0321262.g004\", \"pone.0321262.s014\", \"pone.0321262.ref062\", \"pone.0321262.ref078\", \"pone.0321262.t002\", \"pone.0321262.ref068\", \"pone.0321262.ref079\", \"pone.0321262.ref080\", \"pone.0321262.ref081\", \"pone.0321262.ref082\", \"pone.0321262.ref082\", \"pone.0321262.ref083\", \"pone.0321262.g007\", \"pone.0321262.s021\"], \"section\": \"Testing for blue modern sources of contamination\", \"text\": \"I. tinctoria is not the only indigo-bearing plant, therefore we considered all the most common ones to check for modern contamination. In particular, we tested Indigofera tinctoria L., Polygonum tinctorium Aiton, Isatis indigotica Fort. and Baphicacanthus cusia Bremek, all of tropical origin [36]. Ecological and historical reports make it highly unlikely they were present in the Caucasus under the cool conditions occurring during the Upper Palaeolithic, therefore their occurrence in the archaeological context would indicate modern contamination. The trichome morphology of I. tinctoria is highly characteristic and resembles those observed in the GST residues (Fig 4 panels e,f; S8 Fig) [62,78]. Today, indigotin is widely used to colour cloth, notably in the dyeing of blue jeans, that we included in our control samples (Methods, Table 2). This common item of clothing is usually manufactured from cotton (Gossypium sp.) [68,79], a seed pod fibre native to tropical and subtropical regions, with numerous species grown commercially today. From an archaeological perspective, cotton is unlikely to be present in the context since the earliest genotyped record of Gossypium worldwide is from the site of Qasr Ibrim (Nubia, southern Egypt), Middle Nile valley and dates to around 5500 BC [80]. Cotton seed pod hairs (commonly named cotton fibres) are included in the category of unicellular trichomes; these show a smooth surface without specific \\u2018ornamentation\\u2019 and are ribbon-like, kidney-shaped, and appear flattish in section [81,82]. They vary in diameter and form regular twists or bends along their length, a feature known as convolution [82,83] (see Fig 7 b1-b2, c1-c2; S17 Fig).\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.g007\", \"pone.0321262.ref071\", \"pone.0321262.ref072\", \"pone.0321262.ref077\", \"pone.0321262.g007\", \"pone.0321262.ref071\", \"pone.0321262.ref072\", \"pone.0321262.ref077\"], \"section\": \"Testing for blue modern sources of contamination\", \"text\": \"Cotton fibres from blue and non-coloured jeans fabric were analysed using both \\u00b5-Raman and \\u00b5-FTIR spectroscopies. The recorded spectra were compared with the archaeological micro-residues (Fig 7 panels d, e). The cotton fibres sampled from jeans fabrics are composed of at least 90% cellulose, natively present as (CI), as reflected by their vibrational spectra [71,72,77] (grey profiles b1-2 in Fig 7 panels d, e). To the best of our knowledge there are no reports of lignin and pectin signals in the vibrational spectra of cotton fibres [71,72,77].\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.g007\", \"pone.0321262.s019\", \"pone.0321262.ref077\", \"pone.0321262.s017\", \"pone.0321262.g007\", \"pone.0321262.ref077\", \"pone.0321262.ref077\", \"pone.0321262.g007\"], \"section\": \"Testing for blue modern sources of contamination\", \"text\": \"Furthermore, the \\u00b5-FTIR spectrum of the archaeological blue micro-residues exhibit a peak at 1732\\u2009cm-1, assigned to C=O stretching of ester groups (Fig 7 panel e, red line spectrum a1-2, S15 Fig). The presence of this signal, attributed to pectin [77] (as also confirmed by Raman), matches the spectrum of the experimentally processed modern I. tinctoria epidermis fragments (S12 Fig panel b), that is not observed in the jeans cotton sample (Fig 7 panel e, black line c1-2 and grey profile b1-2, respectively). The intensity of this peak is reported to be strengthened by the presence of oxycellulose in aged cellulose cell walls [77]. As far as lignin (L) is concerned, its occurrence cannot be inferred by FTIR analysis due to the presence of the broad and intense water absorption band at approximately 1630\\u2009cm-1, which overlaps with lignin signals [77] (Fig 7 panel e).\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.g006\", \"pone.0321262.s017\", \"pone.0321262.s019\"], \"section\": \"Testing for blue modern sources of contamination\", \"text\": \"The presence of (P) and (L) in the Dzudzuana micro-residues (Figs 6 panel d and 7 panel d-e red lines a1-2, S13 and S15 Figs) confirms that the residues are not cotton, since these polymers are not present in cotton fibres. All these analyses conducted indicate that the indigotin signal measured on the 25 archaeological blue residues derives from I. tinctoria.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref008\", \"pone.0321262.ref008\"], \"section\": \"Discussion\", \"text\": \"Recovering evidence for prehistoric use of plant secondary compounds has barely begun. Although a significant number of plants naturally contain useful secondary metabolites, many require complex processing to access these and make them usable [8]. Even in the case of food, some nutritious plants require extensive leaching, roasting and/or pounding to eliminate toxins, while extraction of useful medicinal secondary compounds demands a deep knowledge of plants since many can be both medicinal and poisonous, with only the correct processing and dosage making the difference [8]. The processing of plants cannot be simply assumed or ignored as it formed part of the complex tapestry of Palaeolithic life. Nonetheless, to demonstrate the use of plants, exhaustive analytical studies are required and studies such as those presented here, is one way to achieve this.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref011\", \"pone.0321262.ref051\"], \"section\": \"Discussion\", \"text\": \"By considering the antiquity of the blue micro-residues under study, excluding contamination was given high priority [11,51]. Therefore, we applied a stringent procedure during both the sampling activities in the museum and then during lab-processing. All the control-tests proved negative for the presence of indigotin. Also, modern garments such as blue jeans were tested as possible contaminants and our approach proved crucial to discriminate archaeological blue micro-residues from cotton, a putative source of modern contamination.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref040\", \"pone.0321262.ref042\", \"pone.0321262.ref044\", \"pone.0321262.ref045\", \"pone.0321262.ref036\", \"pone.0321262.ref040\", \"pone.0321262.ref044\", \"pone.0321262.ref084\", \"pone.0321262.ref085\"], \"section\": \"Discussion\", \"text\": \"The secondary compound identified, indigotin, is not present as such in the indigo-bearing plant and it results from the enzymatic hydrolysis of its precursors, when they are released by pounding the leaves. These are by-products of the photosynthesis; hence they are naturally contained in the cells of the leaves [40\\u201342,44,45]. This first mechanical process is propaedeutic to the following radical cross-coupling of indoxyl, with either isatin (derived from indoxyl oxidation) or a second indoxyl unit, resulting in the formation of indirubin and indigotin chromophores respectively [36,40,44,84,85], in the form of a blue powder insoluble in water and other polar solvents.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref036\", \"pone.0321262.ref014\"], \"section\": \"Discussion\", \"text\": \"Our experiments demonstrated that this part of the cha\\u00eene op\\u00e9ratoire is common to all the indigo-bearing plants, Indigofera tinctoria L. (true indigo), Polygonum tinctorium Aiton (syn. Persicaria tinctoria, Japanese indigo), Baphycacantus cusia Bremek (common conehead), Isatis indigotica Fort. (Chinese woad), and Isatis tinctoria (woad). However, the latter is the only species naturally present across western Eurasia [36]. The paleogenomic signature indicates the presence of Isatis sp. at the Palaeolithic site of Aghitu-3 in Armenia (Caucasus) around 36.000 years ago [14], thus making I. tinctoria consistent with the source of the identified compound, indigotin, at Dzudzuana. Since the release of these precursors can occur only when the leaves are broken down, the presence of blue fragments found entrapped in the crevices of the Dzudzuana GSTs is unlikely to be accidental.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref003\", \"pone.0321262.ref006\", \"pone.0321262.ref014\", \"pone.0321262.ref020\", \"pone.0321262.ref022\", \"pone.0321262.ref025\", \"pone.0321262.ref027\", \"pone.0321262.ref028\", \"pone.0321262.ref086\"], \"section\": \"Discussion\", \"text\": \"This study presents an innovative research design, a stringent sampling procedure, and the integrated methodological approach that has led to the identification of indigotin, a secondary compound not naturally present in the plant. In addition, the presence of this molecule provides evidence for the exclusion of contamination. However, there is no direct archaeological evidence for how the I. tinctoria plant may have been used. This observation is unsurprising, as plant remains are infrequently preserved in Palaeolithic contexts, making the recovery of direct evidence for the intentional and specific use of plants highly improbable. Indeed, throughout the entirety of the Palaeolithic record, documentation of the deliberate exploitation of particular plant species is exceedingly rare [3\\u20136,14,20,22\\u201325,27,28,86].\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref038\", \"pone.0321262.ref046\", \"pone.0321262.ref027\", \"pone.0321262.ref087\", \"pone.0321262.ref088\", \"pone.0321262.ref089\", \"pone.0321262.ref039\", \"pone.0321262.ref040\", \"pone.0321262.ref045\", \"pone.0321262.ref046\"], \"section\": \"Discussion\", \"text\": \"Many plants have extensive medicinal properties including anti-inflammatory, anti-tumour, antimicrobial, antiviral, analgesic, and antioxidant [38,46] and there is extensive evidence for the use of these, not only in human traditional medicine, but also across the animal kingdom. All animals, and even insects, self-medicate [27]. In the case of chimpanzees, they sometimes prepare the plants prior to consumption [87]; Sumatran orang-utans are known to apply to wounds a mashed concoction of Fibraurea tinctoria leaves [88,89]. Today, the roots of I. tinctoria and other indigo-bearing plants are used in medicine because they contain flavonoids and the leaves contain indigoid-precursor molecules that have preservative, antiseptic, repellent, and protective properties [39,40,45,46]. It is therefore entirely within the behavioural context of humans, from all Palaeolithic periods, to use plants to self-medicate.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref036\", \"pone.0321262.ref040\", \"pone.0321262.ref090\", \"pone.0321262.ref092\", \"pone.0321262.ref093\"], \"section\": \"Discussion\", \"text\": \"I. tinctoria is also known as a source of indigotin, a well-established blue chromophore obtained by the oxidation of precursors naturally present in the cells of the leaves. The use of I. tinctoria to obtain a blue hue is well known, and this knowledge extends into later prehistory [36,40,90\\u201392]. The use of this plant has been recorded as dye since Egyptian times, the earliest written source being the Papyrus Graecus Holmensis (also known as the Stockholm papyrus, retrieved in the XIX century) [93]. However, while the possibility exists that I. tinctoria\\u00a0was transformed into woad dye\\u00a0and used during the Early Upper Palaeolithic, there is currently no archaeological evidence for this.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref094\", \"pone.0321262.ref096\", \"pone.0321262.ref097\"], \"section\": \"Discussion\", \"text\": \"However, and more broadly, colour was a part of Upper Palaeolithic life and the use of mineral pigments is well known, in particular in rock art where red, yellow, white and black are present across the Eurasian continent and the Indonesian archipelago from around 40,000 years ago [94\\u201396]. Blue is a relatively rare colour in nature and to the best of our knowledge, blue pigment (mineral-based) from Palaeolithic contexts has only been reported for Siberian figurines [97].\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref092\", \"pone.0321262.ref098\", \"pone.0321262.ref090\", \"pone.0321262.ref091\"], \"section\": \"Discussion\", \"text\": \"Plant-based dyes are first reported from the Natufian Epipalaeolithic layers of Kebara cave (~13,500\\u201311,650), where red was extracted from Rubia tinctoria roots [92,98]. Seeds of I. tinctoria were first recorded in 5000-year-old Neolithic sites in France (e.g. Bouches-du-Rh\\u00f4ne) [90,91], but to date, there is no archaeological evidence for the use of I. tinctoria prior to the Iron Age.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.t001\"], \"section\": \"Methods\", \"text\": \"All the techniques and the equipment employed are summarized in Table 1.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref099\"], \"section\": \"Structural characteristics of the stone pebbles\", \"text\": \"Tomographic reconstructions, including phase retrieval pre-processing, were carried out using the STP (SYRMEP Tomo Project) software [99]. Different values of the ring removal filter and of \\u03b4/\\u03b2 ratio for phase retrieval correction were used, depending on the characteristics of the samples. The 32-bit.tiff reconstructed image stacks were then converted to 16-bit in order to process the datasets with available imaging software like Fiji (ImageJ), Dragonfly, CTAn and VGStudioMax.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.t001\"], \"section\": \"Stone pebbles surface microtopography\", \"text\": \"A variety of microscopy techniques (see Table 1) were employed to inspect the archaeological tools, the reference artefacts, and residues.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.s003\", \"pone.0321262.s004\"], \"section\": \"Micro-residues chemo-profiling\", \"text\": \"\\u03bc-Raman and \\u03bc-FTIR measurements were always performed before SEM observations since metallic coating of the micro-residues or moulds\\u2019 surface interferes with these spectroscopic analyses. The \\u00b5-Raman and \\u03bc-FTIR were used to characterise both the blue and non-coloured biogenic micro residues, as well as reference materials (for spectra collection parameters see S3 and S4 Tables), since they are non-invasive and non-destructive and guarantee micrometric spatial resolution. Furthermore, vibrational spectroscopies enabled the collection of information on the molecular composition of the leaf epidermis fragments\\u2019 matrix in which the dye is adsorbed.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref100\"], \"section\": \"\\u00b5-Raman\", \"text\": \"Raman spectroscopy is known as a powerful method for the identification of dyes, in particular when the laser exciting light is in resonance or pre-resonance with the electronic transitions of the dyestuff [100]. After initial tests performed with different exciting lines (514, 633 and 785\\u2009nm), including an extended spectral range (100\\u20134000\\u2009cm-1), it was decided that the best conditions to characterise both dye and micro-residues was to excite them at 785\\u2009nm, focusing on the most intense vibrational modes between 200 and 1800\\u2009cm-1. Furthermore, preliminary investigation (live mode by means of oscilloscope) was conducted on at least three different spots of each single micro-residue. Subsequently, the most significant region was selected for longer inspection, ensuring a good signal-to-noise ratio of the spectrum.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.t001\"], \"section\": \"\\u00b5-Raman\", \"text\": \"The measurements were carried out with a \\u03bc-Raman Renishaw InVia spectrometer (see Table 1) and a 4\\u2009cm-1 spectral resolution was achieved by this setup. The spectrometer was coupled with a Leica DM LM optical microscope. Spectra were collected in a backscattering configuration, through a 50\\u00d7 objective (NA: 0.75) and with power on samples below 150 mW, which prevents fibres matrix degradation; on the other hand, when laser was focused on a blue dyed region, power was kept below 15 mW to prevent indigotin photooxidation.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref101\"], \"section\": \"\\u00b5-FTIR\", \"text\": \"FTIR spectroscopy is a complementary vibrational technique and was used to corroborate and, eventually, add information to the \\u03bc-Raman results [101]. Indeed, FTIR does not suffer from photo-induced luminescence emission interference as Raman does. \\u00b5-FTIR measurements were performed with a Bruker Lumos II FT-IR microscope, equipped with a LN-MCT (liquid nitrogen cooled - HgCdTe) high performance detector. Spectral resolution was set to 4\\u2009cm-1. Measurements were carried out in reflectance mode, to ensure a contactless approach, whereas background contribution was collected by the reflected radiation impinging on a gold film. Basic data manipulation (Fourier transform of interferograms) was performed by Opus 8.8 software and the obtained spectra plotted by means of OriginPro 2021 software.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref102\"], \"section\": \"Isatis tinctoria.\", \"text\": \"Collecting the leaves of I. tinctoria at different time periods was necessary as the content of the precursors and the relative final compound, indigotin, vary across the year [102]. The leaves are richer in compounds during spring, but the content can also vary according to the external climatic conditions (humidity and drought). The plants were sourced in the rosette stage in the Site of Community Importance (SIC) and Special Area of Conservation (ZSC) of Marano di Valpolicella (IT3210002, Verona, IT). The identification was made by LL under the guidance of Daniele Zanini, Scientific Director of the Orto Botanico di Novezzina (Ferrara di Monte Baldo, Verona, IT).\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref037\", \"pone.0321262.ref103\", \"pone.0321262.s023\", \"pone.0321262.ref037\", \"pone.0321262.ref102\", \"pone.0321262.s022\", \"pone.0321262.s023\", \"pone.0321262.s023\"], \"section\": \"Isatis tinctoria.\", \"text\": \"Because indigotin is a chromophore obtained from pounded leaves, a series of replicative field experiments were carried out [37,103] at Corte Badin (Verona Italy). Details of the processing procedure can be seen in the video (see S1 Video). In order to perform the replicative experiments, the selected stone pebble was cleaned in the lab by soaking it in a beaker for 30 minutes in different solutions: HCl (5%) and, H\\u2082O\\u2082 (5%). Each treatment was followed by sonication in ultrapure water for 30 minutes. Moulds of the working surfaces were then taken at time zero (T\\u2080). Experiments were carried out at different times of the year according to the I. tinctoria life cycle [37,102]: in early spring (before flowering), summer, and early autumn. 800\\u2009g of leaves, cut from the first year (rosette stage), were immediately processed using the stone pebble in a wooden mortar for approximately 30\\u201340 minutes. The pestle was reused in the subsequent replicative experiments, without being washed, and stored in a sealed bag until the next use. The mechanical processing was characterised by a combination of vertical action (such as pounding) and horizontal linear motion that shredded the leaves, with these actions repeated until a poultice was formed. All steps of the process were photographed and video-recorded (S1 File; S1 Video). Our experiments have shown that the poultice tends to overflow from the direct contact area and adhere to the hand operating the stone, meaning the residues can easily spread across multiple areas, not necessarily limited to those in direct contact with the working area (S1 Video). Fragments of the leaves remained trapped in the uneven surface topography of the stone.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref036\"], \"section\": \"Indigo-bearing plants\", \"text\": \"To exclude the presence of other indigo-bearing plants, we tested the most common ones [36], and in particular Indigofera tinctoria L., Polygonum tinctorium Aiton, Isatis indigotica Fort. and Baphicacanthus cusia Bremek. The plants were sourced from a Chinese Pharmacy shop in Hangzhou, China, and were collected from regions in subtropical to tropical climates namely Zhejiang Province, Hunan Province, Anui Province, and Guangdong Province respectively.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.ref068\", \"pone.0321262.ref079\", \"pone.0321262.ref081\", \"pone.0321262.ref082\", \"pone.0321262.ref104\"], \"section\": \"Jeans fibres (Gossypium sp.).\", \"text\": \"Indigotin is widely used to colour textiles, in particular in the dyeing of blue jeans, a common item of clothing usually manufactured from cotton [68], a seed pod hair. In particular, upland cotton (Gossypium hirsutum [79,81,82]) accounts for 90% of modern production [104]. Thus, to ensure that the presence of archaeological micro-residues is not the result of modern contamination, cotton fibres from blue and non-coloured jeans fabric were sourced. Morphological and vibrational analyses performed on these fibres were compared to the archaeological micro-residues.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.s022\"], \"section\": \"\\u00b5-Raman spectra comparison between natural woad and synthetic indigotin.\", \"text\": \"Normalised Raman spectra of natural woad dye extracted from modern I. tinctoria leaves according to hot-water extraction (see S1 File), compared with synthetic indigotin standard. Both were excited at 785\\u2009nm and they are shown after luminescence background removal.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.g006\"], \"section\": \"\\u00b5-Raman spectra of non-coloured archaeological micro residues\", \"text\": \"Normalised Raman spectra of archaeological non-coloured micro fragments from Dzu S6 m3 (a) and Dzu S1 m7 (b), reprised from Fig 6, excited at 785\\u2009nm and after luminescence background removal. Characteristic bands of cellulose (C), including its polymorphs I (CI) and II (CII), pectin (P) and lignin (L) are indicated by their Raman shift.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.g006\"], \"section\": \"\\u00b5-Raman spectra of archaeological micro residues (reprised from Fig 6 panel d).\", \"text\": \"\\u00b5-Raman spectra of blue and non-coloured archaeological fragments. (a): spectrum of a non-coloured region of a blue residue; spectra b and c of non-coloured residues; spectra Ref3 of non-coloured I. tinctoria fibre obtained from the processing of modern leaves (reported as reference). (b and c): zoomed region of interest of the spectra reported in (a). This figure supports Fig 6 d of the main text.\"}, {\"pmc\": \"PMC12063890\", \"pmid\": \"40343940\", \"reference_ids\": [\"pone.0321262.g007\"], \"section\": \"\\u00b5-Raman spectra of blue archaeological residues compared with jeans fibres one (reprised from Fig 7 panel d).\", \"text\": \"\\u00b5-Raman spectra of blue and modern fibres. (a): spectrum a1-a2 of blue archaeological residues, while spectrum b1-b2 is a modern non-coloured jeans fibres reported for comparison. (b and c): zoomed region of interest of the spectra reported in (a). This figure supports Fig 7 d of the main text.\"}]"
Metadata
"{}"