Chemical Characterization of Red Pigments Used in Funerary Practices in Northeastern Patagonia (Chubut, Argentina) During the Late Holocene

Abstract

In this study, we present the chemical characterization of red pigment samples and their associated sediments that were collected from three human burial sites in Northeastern Patagonia. Regarding their chronology, the La Azucena 1 site (880 ¹⁴C years BP) corresponds to the period prior to European contact while the Loma Torta and Rawson sites date to periods following contact. These burials were discovered fortuitously. In the case of the La Azucena 1 site it was due to the impact of environmental conditions typical of this region, such as sparse vegetation and the effects of rain and wind, while for the Rawson and Loma Torta sites the burials suffered severe anthropic impact. Analysis of the red pigments and their sediments by a combination of analytical techniques using XRF, XRD, and ATR-FTIR revealed hematite as the chromophore responsible for the red color, together with large amounts of quartz in all the samples. The diffractogram of the red pigment from the La Azucena I site showed notable differences compared to those from the Loma Torta and Rawson sites, with calcite (CaCO₃) and anorthite (Na_0.45Ca_0.56)(Al_1.55Si_21.5O₈) as accompanying minerals and the presence of cristobalite, a high-temperature polymorph of silica (SiO₂), which were not identified in the sediment sample. This suggests that minerals identified in this sample are characteristic of the pigment material rather than of the sediment where the bone remains were found.

1. Introduction

Bioarchaeological research in the Chubut Province, specifically on the northern coast and the lower valley of the Chubut River, indicates that the hunter-gatherers inhabiting those areas buried their dead in elevated locations, such as river or marine terraces and hills, to protect the bodies from tidal or riverine floods [1]. A notable feature of this region is the abundant mortuary record, whose oldest evidence is an individual of 6070 ¹⁴C years BP. The remaining burial chronology ranges between 3000 ¹⁴C BP and after the XVI Century, when Europeans arrived in Patagonia. Among the identified burials, two types are distinguished: primary burials, where the bodies were interred only once, and secondary burials, where bone remains were later relocated or reburied. There were registered single, double and also collective or multiple burials where several individuals were interred in the same space.

Within this record, it is important to note that some Late Holocene sites (1000–800 ¹⁴C years BP) were associated with rich funerary goods, including red pigments, lithic artifacts (projectile points, stone balls), textile fibers, beads made of shell and of non-local rocks (malachite, turquoise, and serpentine), as well as brass or bronze items of extra-regional origin [1,2]. Later, during the Post-contact period (400–200 ¹⁴C BP), European manufactures such as glass beads and metal objects were introduced into the indigenous territory. The presence of these exotic goods suggests that the Patagonian groups participated in a wide trade network of products from different regions and cultural origins, probably linked to an increase in social and political complexity [2]. The burial sites analyzed in this study, which range from the late pre-contact (e.g., La Azucena 1, ~880 BP) to the post-contact period (e.g., Loma Torta and Rawson, ~250 BP), provide an opportunity to examine the continuity of pigment use across these distinct temporal contexts.

The existence of these trade networks is also reflected in the distribution of obsidian artifacts across Argentinian Patagonia [3]. Recent studies in the northern and southern coast of Chubut province and the lower valley of Chubut River have identified the circulation of obsidian nodules and artifacts of at least nine major obsidian sources since the Middle Holocene [3,4,5]. The spread of these materials, which in some cases covered more than 800 km, is important evidence for intergroup contact and cultural exchange that spanned vast regions. In addition, red, white, green and yellow pigments were registered in diverse archaeological contexts -rupestrian paintings, burials and resident camps—some of them located more than 300 km from the potential source areas [6]. Therefore, in certain cases, pigments must also have been obtained through trade or direct acquisition.

Along the Chubut River basin, which runs from east to west through this province, and in other potential distribution routes for obsidian [3], natural pigmented sediments have been found [6]. The similarities in the colors of the pigments found east of the Chubut River with the region’s natural sediments suggest that the river may have functioned as a corridor for the movement of people, ideas, designs, methods, and raw materials, including pigments. However, the specific routes of this exchange have not yet been definitively established.

In Argentina, the oldest record of red pigments in human burials comes from two individuals at the Arroyo Seco 2 site in the Buenos Aires province, with radiocarbon dates of 7800 ± 115 ¹⁴C years BP and 6908 ± 76 ¹⁴C years BP [7]. In Patagonia, the oldest date for red ochre in Chubut is 4760 ± 54 ¹⁴C years BP, from the Fabiana Elizabeth burial site [8]. Other sites in the region, such as San Pablo 4 in the Valdés Peninsula, have radiocarbon dates of 2920 ± 90 ¹⁴C years BP and 2780 ± 90 ¹⁴C years BP [9], suggesting that the practice of using red ochre was well-established since the Middle Holocene.

Further south, between the Santa Cruz River and the Strait of Magellan, the presence of ochre has been documented in multiple burials from the Early Late Holocene (4000–3500 ¹⁴C years BP) at sites such as Cerro Sota, Cueva Lago Sofía 1, Orejas de Burro 1, and Río Bote 1 [10,11,12]. Chemical analyses of the pigments at these sites identified hematite and goethite by Raman spectroscopy and Micro X-ray fluorescence (μ-XRF) at Río Bote 1 [12] while hematite combined with gypsum were characterized by X-ray diffraction (XRD) at Orejas de Burro 1 [13].

In this work, we present the results of the first chemical characterization of red pigment samples and their associated sediments from three human burial sites in Northeastern Patagonia, one from Valdés Peninsula (La Azucena 1) and two from the lower Chubut River valley (Loma Torta and Rawson) (Figure 1).

The pigment and sediment samples were analysed by a combination of analytical techniques, such as X-ray fluorescence spectroscopy using a portable system (pXRF), Attenuated Total Reflection Fourier-transform infrared spectroscopy (ATR-FTIR), and X-ray diffraction (XRD). pXRF [14] and ATR-FTIR [15] allowed us to obtain elemental and molecular information, respectively, in a non-destructive way and without any preparation of the samples while XRD [16] identified crystalline minerals and complemented FTIR results. Comparison of the results of the analysis of pigment samples and their associated sediments helped us to evaluate potential taphonomic contamination of the pigments. There are not any quarries of pigments in the vicinity of these burials, the nearest are located more than 100 km from them. In an effort to identify potential sources of pigment that might have been used by the hunter-gatherer groups, the results were compared with those of naturally colored red sediment samples from our reference database [6] to determine potential pigment sources.

2. Materials and Methods

2.1. Human Burial Sites

La Azucena 1 is a primary burial of two middle-aged females located in the southern part of the Valdés Peninsula. The site is situated in an erosion depression between currently active dunes; however, geomorphological studies and taphonomic evidence, such as root marks on the bones, determined that the dunes were stabilized by vegetation both before and after the burial [17]. Individual 1 was dated in 880 ± 50 ¹⁴C years BP and individual 2 in 860 ± 70 ¹⁴C years BP (LP-3831) [1], which indicates that both women would have been buried simultaneously.

Figure 2a exhibits the bones of individual 1. After processing the digital photo using DStretch^®complement Image J (1.53), the enhanced image (Figure 2b) showed the distribution of the red pigment on most of the bone fragments and the surrounding sand. This fact, along with the absence of signs of defleshing on the skeletons, rules out this practice and subsequent painting of the bones—a practice recorded at burials in the lower Colorado River basin [18]. Instead, it has been suggested that the bodies were buried in leather bundles, which likely contained a considerable amount of red pigment. The degradation of the bundle and human tissues would have facilitated the transfer of pigment to the bones and surrounding sand by contact [1]. The red pigment sample (1) was obtained from the bones of individual 1, where it was adhered, while sediment (S-1) associated with the bones was also collected. Additionally, a short-twisted fiber of guanaco hair, red in color, and a flat grinding stone with red stains on one face were found at the foot of the burial. This suggests that the stone was associated with the preparation of the pigment used in the funerary ritual.

In the lower Chubut River valley, unlike in the Valdés Peninsula, the identified burial sites are mostly double and multiple. In some cases, it has not been determined whether the bodies were buried simultaneously or if these locations served as cemeteries that were repeatedly used by ancient populations for burials [1].

The multiple burial site of Loma Torta is located in the town of Gaiman, on the Loma Torta hill, at an elevation of 110 m above sea level (Figure 3a). It was discovered in 2006 due to the frequent passage of motorcycles using this hill and others as part of a local endurance circuit [19]. At the site, remains of 12 individuals were recovered (including adult males and females, as well as infants), mostly fragmented due to significant alteration by natural factors (water and wind erosion, roots, and rodent burrows and galleries) and anthropogenic factors (motorcycle traffic and evidence of looting). With the exception of three individuals (referred to as 1, 2, and 3), buried in Sector 1 (Figure 3a) and exhibiting a degree of skeletal integrity, the remaining individuals were found in an ossuary state, suggesting the possible presence of additional human remains in the area. Among the burials, both primary and secondary interments were identified.

Associated with the skeletal remains, red ochre, shell beads, and lithic artifacts were found; however, no evidence of funerary bundles, textiles, or leather materials was recovered [19]. In this context, a lump of red pigment (2) was analyzed, which was found inside the external auditory canal (base of the skull) of individual 3, probably female (Figure 3c). This individual was radiocarbon dated to 250 ± 70 ¹⁴C years BP, and the arrangement of the bones suggested that this was a secondary burial. The reddish-brown sediment (S-2) surrounding the body was collected and also analyzed. It is hypothesized that the individual was interred in a funerary bundle with red pigment, and that the form of the lump resulted from the deposition of organic matter, either human or from other sources.

The Rawson burial site is located in a present sand quarry in the city of Rawson, 3 km from the coastline, and was discovered during gas pipeline construction works. This multiple burial site has been heavily impacted by human activity, including construction works, looting, and illegal excavations. As a result, most of the skeletal remains were found in an ossuary state, intermixed with archaeological materials (lithic artifacts, charcoal, faunal remains, and ceramics) and modern debris, as the site had also been used as a refuse dump. Fourteen individuals were identified (10 infants, three adults, and one subadult), of which only four (individuals 1 to 4) showed relatively skeletal integrity and all corresponded to primary burials. Radiocarbon dates for individuals 1 and 2 determined modern ages (less than 200 ¹⁴C years BP), while individual 3 was dated to 270 ± 60 ¹⁴C years BP [1].

The site stands out for the richness of its funerary assemblages, especially those associated with individuals 3 (subadult) and 4 (infant), suggesting a differentiation in social status. Among the materials found in this context were red ochre, guanaco leather and textile fragments, shell beads, malachite, turquoise, bronze, and glass. Notably, a ceremonial bronze axe was recovered, whose design, dimensions and metal composition are linked to the “Santa María culture,” developed during the Late Holocene (1000–500 ¹⁴C years BP) in the Calchaquí valleys of northwestern Argentina [19]. Since Patagonian hunter-gatherers did not practice metal smelting, the presence of the bronze axe is significant because it reinforces the hypothesis of interregional relationships within an exchange network during the Late Holocene [2]. Among the remains, the skull of individual 8, an infant of ca. 4 years was covered with abundant dark red pigment mixed with plant fibers (Figure 4). Pigment sample (3) was obtained from both sides of the frontal bone while sediment surrounding the skull (S-3) was also collected.

2.2. Samples Description

The pigment and sediment samples under study were collected from the three burial sites described in 2.1. by one of the authors (JGO) and Silvia Dahinten. They are part of the Bioarchaeological and Bioanthropological Collection of the Institute of Diversity and Southern Evolution (CONICET-IDEAus), within the National Patagonian Center in Puerto Madryn (CCT CENPAT), Chubut province, Argentina. In the laboratory, pigment samples 1 and 3 were obtained by scrapping the bones with a plastic spatula while sample 2 was collected with a metal clamp. Sediment samples were collected adjacent to the skeletal remains at the burial sites. All samples were stored in plastic bags until chemical analysis.

Table 1. Description of red pigments and associated sediments from the three burial sites.

Burial Site, Geoposition	Sample	Color, Munsell
La Azucena 1	1	Red (10R 5/8)
42°50′43″ S/64°09′82″ W	S1	Reddish brown (2.5YR 4/4)
Loma Torta	2	Red (10R 5/8)
43°16′36.7″ S/65°0.5′51.5″ S	S2	Reddish brown (2.5YR 4/4)
Rawson	3	Dark red (10R 3/6)
43°17′19.6″ S/65°05′51.5″ W	S3	Brown (7.5YR 5/7)

describes the origin of pigment and sediment samples characterized in this study and their color. The sediments retain the same number as the corresponding pigment sample, preceded by the prefix “S”. The color of the samples was assigned according to the Munsell^® Soil Color Charts [20].

2.3. Analytical Methods

Infrared spectra were obtained using a Nicolet iS50 FTIR spectrometer with a diamond single-bounce ATR accessory (Thermo Fisher Scientific Inc., Waltham, MA, USA). Samples were grinded in a mortar with pestle to obtain a homogeneous powder. For each sample, 64 scans were recorded in the 4000–400 cm⁻¹ spectral range in the reflectance mode with a 4 cm⁻¹ resolution. Spectral data were collected with the Omnic v9.2 (Thermo Fisher Scientific Inc., Waltham, MA, USA) software without post-run processing. The spectrum of air was used as background. X-ray fluorescence spectra were performed with a Bruker Tracer III-SD (Bruker, Bilerica, MA, USA) portable analyzer. The samples were applied on the beryllium window of a Silicon Drift detector and measured without any previous preparation. The spectra were registered at 40 kV of voltage and 15 μA current, with an acquisition time of 100 s with S1PXRF (Bruker) software. The XRF data were analyzed with Artax 7.4.0. program (Bruker, Bilerica, MA, USA). Some spectra exhibited the presence of argon (Ar), a component of the atmosphere, as well as a broad and weak peak of nickel (Ni) from the instrument components. X-ray diffractograms of the samples were obtained with a Rigaku Denki diffractometer D/max-IIIC (Rigaku Holdings Corporation, Tokyo, Japan) with Cu (λKα = 1.5405 Å) radiation in a Bragg–Brentano geometry. Diffraction patterns were collected from 3 to 60° 2Ѳ with a step size of 0.01° and recording velocity of 2°/min. The Powder Diffraction Files (PDF cards) from the software MDI/JADE 7 (Livermore, CA, 94550 USA) were used for evaluating the diffraction data. Prior to the analysis, samples were grinded in a porcelain mortar with pestle to give a fine homogeneous powder.

3. Results

3.1. XRF

The XRF spectra of pigments 1–3 showed the presence of iron peaks as the predominant element (Figure 5). Pigment 1 also showed a minor peak of calcium that could suggest the presence of calcite. On the other hand, pigment 3 revealed additional peaks of silicon, potassium, calcium, manganese, and titanium that could be attributed to accessory minerals due to the natural origin of the pigment or to components of the sediment associated with the skull. The XRF spectra of the three sediments (S1–S3) showed similar elemental compositions with iron and calcium as the predominant elements together with minor peaks of silicon, potassium, titanium, and manganese. Comparison of the XRF spectra of pigments 1 and 2 and their respective sediment samples (S-1 and S-2) showed differences in their elemental compositions. In particular, pigment 2 revealed iron as the only element detected suggesting a high content of the red chromophore in the sample. This could be related to the characteristics of the pigment sample, which was recovered as a red lump from the base of the skull of individual 3 (Figure 3c).

3.2. XRD

Hematite (α-Fe₂O₃) peaks were identified in all the diffractograms, in accordance with the presence of iron in the XRF spectra of all the samples. Nevertheless, the pigment samples revealed a higher intensity of signals than their corresponding sediments (Figure 6). This difference was especially evident in the diffractograms of red pigments 1 (Figure 6a) and 2 (Figure 6b) from the remains of the La Azucena 1 and Loma Torta sites, respectively. The identification of hematite in the three pigment samples indicates that this mineral is the chromophore responsible for the red color and that its presence in the sediments may be due to their contact with the pigment. In addition, pigment samples 1 and 2 showed high-intensity peaks at 33.2° and 35.7° (2θ). The value of full width at half maximum (FWHM) for both peaks was 0.35° indicating a fairly high crystalline hematite [21]. This is consistent with the FWHM range between 0.26 and 0.48° for the hematite peak at 33.4° in natural hematite of different crystallite sizes [22]. The identical FWHM values suggest that the hematite in pigments 1 and 2 shares a similar crystallographic quality, which could indicate a comparable geological origin or a similar thermal history during its formation or potential pre-treatment before use.

The presence of high intensity peaks of quartz (SiO₂), especially around 26° (2θ), is a common characteristic of all the samples.

The diffractogram of pigment sample 1 from the La Azucena 1 site at Peninsula Valdés showed notable differences compared to those pigments from the lower valley of the Chubut River (samples 2 and 3) with quartz (SiO₂), calcite (CaCO₃), and anorthite (Na_0.45Ca_0.56)(Al_1.55Si_21.5O₈) as accompanying minerals together with cristobalite, a high-temperature polymorph of silica (SiO₂) present in siliceous volcanic rocks [23]. The presence of cristobalite in this context suggests a natural origin rather than anthropogenic heating. This interpretation is supported by regional geological evidence, as documented in mineralogical databases [24]. Furthermore, we have previously detected cristobalite in a sediment collected in the center of Chubut province near the Chubut River [6]. Additionally, Massaferro et al. [25] reported the presence of cristobalite in an ochre sediment collected from the Río Negro steppe in Patagonia. Pigment sample 1 showed a different mineralogical composition compared to its sediment (S-1), which was composed mainly of quartz and albite Na(AlSi₂O₈) and minor hematite. This strongly suggests that the minerals identified in sample 1 are characteristic of the pigment material rather than of the sediment where the bone remains were found.

The diffractogram of pigment sample 2 (Figure 6c) showed hematite and quartz as the main components and a very low content of albite, which was detected in sediment S-2 together with quartz (Figure 6d). Therefore, it is highly probable that albite is part of the sediment rather than of pigment 2 itself. It is interesting to note that pigment sample 3 and sediment S-3 from the Rawson site contained similar mineralogical compositions in agreement with XRF results. Albite was identified as the main accompanying clay mineral, along with quartz and minor content of montmorillonite (Figure 6e,f). This suggests that these minerals are more likely components of the sediment instead of pigment 3. On the other hand, the presence of hematite in sediment S-3 could be related to its contact with the red pigment of the skull.

3.3. FTIR ATR

The presence of hematite was confirmed in the infrared spectra of all the samples through the observation of bands at 460–470 and 530 cm⁻¹, which are associated with the Fe-O stretching characteristic of hematite [26]. These bands are more intense in the pigment samples compared to their respective sediments (Figure 7) and overlap with bands from the Si-O-Al and Si-O vibrations of aluminosilicates [26]. The characteristic doublet of quartz at 780 and 800 cm⁻¹ [27] was also observed, especially visible in the infrared spectrum of pigment 3 (Figure 7c). The bands at 1070 cm⁻¹ and 1100 cm⁻¹ observed in the spectra of samples 2 and S-2, respectively (Figure 7b), were also associated with the presence of quartz [27,28].

Aluminosilicate bands were observed in all the samples, with a few differences. A broad band in the region of 3300 cm⁻¹ and a weak one at 1630 cm⁻¹, characteristic of hydroxyl groups present in aluminosilicate minerals [29], was observed in the spectra of samples 2, 3, and their respective sediments from the lower valley of Chubut River. Additionally, in samples S-2, S-3, and 3, bands corresponding to aluminosilicate vibrations at 915–918 cm⁻¹ and 1006–1010 cm⁻¹ were also observed [27,28]. These bands were more intense in the sediments than in the pigment samples. Moreover, the presence of albite detected by XRD in sediment samples S-2 and S-3 is consistent with the presence of bands at 530, 580, and 650 cm⁻¹ [30].

The infrared spectra of samples 1 and S-1 from Valdés Peninsula showed bands associated with quartz (700, 780, 800, 1090–1100, and 1170 cm⁻¹) [26] and calcite (715, 880, and 1420–1430 cm⁻¹) [30]. The absence of calcite peaks in the XRD analysis of S-1 (Figure 6b) may be due to the high concentration of quartz in the sediment, which masks minor components in the powder diffractogram. Nevertheless, the XRF spectrum of S-1 revealed calcium peaks in accordance with the presence of calcite, as detected by ATR-FTIR. Additionally, pigment 1 exhibited bands at 1010, 1640, and 3360 cm⁻¹, indicative of aluminosilicates [28,29] (Figure 7a) that are absent in sediment S-1. In contrast, S-1 displayed bands at 1000–1100 cm⁻¹ corresponding to Si-O-Si vibrations, as well as bands at 590 and 650 cm⁻¹, which are compatible with the presence of albite [30], a mineral identified in the powder diffractogram of this sample (Figure 6d).

Table 2. XRD, XRF, and ATR-FTIR results of the red pigments (1–3) and associated sediments (S-1-S-3) from the three burial sites.

Sample	XRD	XRF	ATR-FTIR
1	Quartz, cristobalite, anorthite, calcite, hematite	Fe, Ca	Quartz, aluminosilicates, calcite, hematite
S-1	Quartz, albite, hematite (minor)	Si, K, Ca, Ti, Mn, Fe	Quartz, aluminosilicates, calcite, hematite
2	Quartz, hematite, albite (minor)	Fe	Quartz, aluminosilicates, hematite
S-2	Quartz, albite, hematite (minor)	Si, K, Ca, Ti, Mn, Fe	Quartz, aluminosilicates
3	Quartz, albite, montmorillonite, hematite	Si, K, Ca, Ti, Mn, Fe	Quartz, aluminosilicates, hematite
S-3	Quartz, albite, montmorillonite, hematite (minor)	Si, K, Ca, Ti, Mn, Fe	Quartz, aluminosilicates

summarizes the results obtained by application of a multi-analytical non-destructive approach to study the composition of the pigment and sediment samples. XDR and ATR-FTIR provided specific information concerning the minerals present in the samples and revealed major differences in the mineralogical composition of pigment sample 1 and its associated sediment (S-1).

The chemical composition of pigment samples 1–3 was compared with 13 samples of natural red sediments collected along the Chubut River [6]. Although all reference samples contained hematite, none exhibited the same composition of accompanying minerals, even those sharing the same hue on the Munsell color scale. For example, pigments 1 and 2 share the same hue on the Munsell color chart (10R 5/8) as a natural pigment collected at Los Altares, a site located about 300 km from Loma Torta and 400 km from La Azucena 1 [8]. This natural red sediment was composed of hematite, quartz, albite, hematite, and kaolinite; however, clay mineral kaolinite was not detected in the analyzed pigment samples. This suggests that it is unlikely that the natural sediment sample was the source of the red pigment used in the burials at Loma Torta and La Azucena 1.

4. Discussion

The three pigment samples (1–3) associated with the studied burial sites contained hematite as the chromophore. X-ray diffraction analysis recorded a greater quantity and intensity of peaks for this mineral in the pigment samples compared to the sediment samples that were in contact with the bone remains. Similarly, FTIR-ATR analysis confirmed well-defined hematite bands in the pigment samples. Regardless of the location of the burial sites (Valdés Peninsula and lower Chubut River valley) or the type of burial (primary or secondary, individual or multiple), the practice of incorporating red pigments along with other objects in the burials is repeated. Previous studies highlight the use of hematite in funerary contexts in Patagonia since pre-Hispanic times. This practice of applying red ochre to bodies and sediments, along with the presence of funerary goods such as necklace beads made from various raw materials and other items, including projectile points, stone balls, and personal ornaments, has been repeatedly observed in burials from the area. It held symbolic value and it was common to bury the bodies with their belongings or with items that would help them survive in the afterlife [1].

An important point raised by this study concerns the chronological period of the sampled burials, which range from the late pre-contact period (La Azucena 1, ~880 BP) to the post-contact era (Loma Torta and Rawson, ~250 BP). Our results indicate a continuity in the choice of hematite as the red chromophore across these different periods. Despite the significant socio-cultural changes documented for the region, including the incorporation of European goods, the technological tradition of selecting an iron oxide-based pigment for funerary rites persisted. This suggests a deeply ingrained cultural practice. However, the notable mineralogical differences—such as the unique presence of cristobalite and anorthite in the La Azucena 1 pigment—indicate variability that could be attributed to the use of different source materials or to variations in preparation methods over time. Therefore, while the central practice shows continuity, the specific characteristics of the pigments involve changing access to sources or an evolution of technical knowledge.

Regarding the potential sources of red pigments used in the burials of Valdés Peninsula (La Azucena 1) and the lower Chubut River valley (Rawson and Loma Torta), these have not been identified to date. The natural sediments exhibited a different mineralogical composition, suggesting that the hunter-gatherers must have obtained the pigments from other sites not yet considered or by trade or exchange. As a result of these observations, a hypothesis emerges regarding the possible use of prior purification technologies for the pigment, such as leveraging different solubility or hydration properties and differential flotation of components in water, or other mechanical processes such as sieving or gravity classification, in cases where the components may have different densities. Clays, for instance, could have been separated from other components based on particle size and sedimentation velocity, given that hematite, quartz, and aluminosilicates are generally heavier and would sediment more rapidly. Sieving could also have been effective, as kaolinite is less dense and, after grinding, would form a finer powder than the other components. It is also possible that the ancient hunter-gatherers combined several of these methods. For example, grinding the mineral mixture, sieving to remove larger particles, mixing with water to settle and separate finer particles such as kaolinite, and allowing it to rest and settle multiple times to enhance purity.

While the non-destructive approach employed here (ATR-FTIR, XRD, pXRF) has successfully identified the main components of the pigments and established a fundamental distinction from the sediment matrix, the question of the precise geological source of the raw materials remains open. Future research, if sample quantity permits, would greatly benefit from the application of highly sensitive techniques such as Inductively Coupled Plasma Mass Spectrometry or XRF in addition to portable devices. The analysis of trace element patterns could provide a unique geochemical fingerprint, enabling the precise sourcing of the pigments and offering deeper insights into trade routes or material selection practices.