230Th dating of flowstone from Ignatievskaya Cave, Russia: Age constraints of rock art and paleoclimate inferences

Abstract Paleolithic antiquity of parietal art in Ignatievskaya cave, Southern Ural, is supported by its subject (Late Pleistocene animals) as well as by paleontological and palynological data, and 14C dates from cultural layers associated with artistic activity (17.8−16.3 cal ka BP; association is established by finds of ochre in these layers). However, three 14C dates of charcoal motifs yielded younger, Holocene ages (7.4−6.0 cal ka BP). In this study, we constrain the age of parietal art in the cave by 230Th dating of flowstone that brackets the paintings. Flowstone did not form in the cave between c. 78 and 10 ka BP, due to widespread permafrost in northern Eurasia at that time. Our 230Th dates do not support the middle Holocene age of art in Ignatievskaya cave and are consistent with its Upper Paleolithic antiquity instead.

Bronze and Iron Ages were found in the cave, of which only Paleolithic materials were associated with a clearly identifiable, albeit thin, cultural layer; younger finds were collected either from the cave floor or found in loose sediments (Petrin, 1992;Shirokov & Petrin, 2013). The cultural layer also contained abundant fragments of charcoal and rare clumps of ochre, which establishes a link between this layer and the red paintings (Shirokov & Petrin, 2013).
Three early conventional 14 C dates of charcoal and bones from the cultural layer yielded Upper Paleolithic ages  i.e., thousands of years [ka] before 1950 C.E.; common era; Petrin, 1992). The chronology was recently refined by accelerator mass spectrometry (AMS) dating of charcoal, with four dates constraining the time of its accumulation to 17.4-16.3 cal ka BP and one sample (from the deepest position) yielding an older age of 28.3-27.8 cal ka BP see Supporting Information S4). The dates are consistent with paleontological evidence from the cultural layer, represented by Late Pleistocene fauna (Petrin, 1992) and pollen characteristic of uniform, forest-free vegetation of a periglacial steppe (Smirnov et al., 1990).
Although these lines of evidence, as well as the similarity with the Shulgan-Tash cave, make a seemingly strong case for the Upper Paleolithic antiquity of the parietal art in Ignatievskaya cave, some ambiguity remains. First, the similarity between these two decorated Southern Ural caves is limited, due to the abundance of black (charcoal) paintings in Ignatievskaya cave, which are all but absent in Shulgan-Tash. Second, stylistically, figurative paintings in the two caves are quite distinct. Some researchers suggested that styles of figurative paintings in Ignatievskaya cave are not characteristic of the Paleolithic "naturalism" (Formozov, 1998;2000;Zhitenev, 2016).
Others countered that the "Ignatievskaya" style is not consistent with any post-Paleolithic art known in Southern Ural either (Shirokov, 2006). Third, three 14 C determinations made directly on the charcoal paintings returned Mesolithic rather than Upper Paleolithic dates (8.9-6.6 cal ka BP; Shirokov et al., 2003;Steelman et al., 2002).
In this study, we provide independent constraints on the age of paintings in Ignatievskaya cave employing 230 Th dating of flowstone that formed before and after the pigment was applied. This approach has recently established itself as a valuable and robust tool in the field of cave art dating (Hellstrom, 2012;Hoffmann et al., 2016;Pike et al., 2012;Pike, 2017).

| Materials
In this paper, we present results of the geochronological study of eight samples of speleothems (flowstone and a stalagmite) from Ignatievskaya cave. Five samples have identifiable temporal relationships with parietal art, whereas another three were used to build confidence in 230 Th dating results and characterize the speleothem/ paleoclimate system.

| Sampling
Samples of flowstone were collected as 8-mm-diameter cores using hand-held drilling equipment (Spötl & Mattey, 2012). In the laboratory, the cores were cut longitudinally with a precision diamond saw (Buehler ISOMET). Cut surfaces were polished to better reveal growth zoning and enable precise milling of subsamples. Subsamples for 230 Th dating (6-200 mg) were milled in the laboratory in a clean air laminar-flow hood.

| 230 Th dating
Samples were spiked with a mixed 229 Th-233 U-236 U tracer before chemical separation of U and Th using an Fe coprecipitation procedure, followed by chemical purification on anion exchange columns (Edwards et al., 1987;Shen et al., 2012). Procedural chemistry blanks were typically less than 100 ag for 230 Th and less than 1 fg for 234 U.
U and Th isotopic ratios and concentrations were determined at the University of Minnesota, MN, using the latest protocols on a Thermo

| 230 Th age modeling
For specimens with multiple 230 Th dates, age modeling was performed using OxCal 4.4 (Bronk Ramsey, 2008;Bronk Ramsey & Lee, 2013) to assess the time of the initiation of speleothem growth.

| Calibration of 14 C ages
All radiocarbon ages referred to in this paper were taken from publications. For consistency, the original 14 C ages were recalibrated using the program OxCal 4.4 (Bronk Ramsey, 2008) and the IntCal20.14c calibration curve (Reimerr et al., 2020).   (Rudenko, 1914). In 1960In -1961, in addition to the Holoceneage artifacts, five Upper Paleolithic stone tools were found in archaeological pits dug in the near-entrance part of the cave (Bader, 1980). In 1980, archaeologists V. Petrin, S. Chairkin, and V. Shirokov discovered Paleolithic parietal art in the cave. During a comprehensive 6-year-long cave research program that followed, five pits were excavated in different parts of the cave (Figure 2). The thin Paleolithic cultural layer ("visiting horizon") was identified at a depth of 0.05 to 0.25 m in pits I−IV, excavated in the deep aphotic part of the cave (Petrin, 1992). In addition to stone and bone tools and decorations made of bone, the layer contained abundant small fragments of charcoal and ochre as well as animal bones (Shirokov & Petrin, 2013). Archaeological finds were scarce in the interior of the cave (only 28 artifacts were found in the Large Hall) but abundant in the Entrance Hall and the Low Passage (more than 1,300 artifacts).
Today's inventory comprises about 180 locations with paintings or their fragments (Shirokov & Petrin, 2013). Both red (ochre) and black (charcoal) paintings are present. The motifs can be subdivided into nonfigurative (lines, dots, etc.) and figurative ones, among which both realistic (mammoth, horse) and composite and nonrealistic zoomorphic figures are known.
The question of the temporal relationships between the black and red paintings in the cave remains moot. At one location, in the Large Hall, a red line was found to be drawn across a black line, indicating that this particular black motif is older than the red one; however, it remains unknown how much older it is and whether this applies to other black and red paintings in the cave (Shirokov & Petrin, 2013).

| SPELEOTHEMS IN IGNATIEVSKAYA CAVE
Speleothem is a collective term describing mineral deposits formed in caves by chemical precipitation from flowing, dripping, ponded, or seeping water (White, 2019). Compared with many of its Franco-Cantabrian counterparts, Ignatievskaya cave is poorly decorated with speleothems; most of the halls and passages of the cave are devoid of them.
The term flowstone refers to one of the common types of speleothems, forming via precipitation of calcite from a thin film of water flowing along inclined surfaces (cave walls, ceilings, and floors).
Typically, flowstone has a sheet-like appearance (i.e., width and length of the deposit are much greater than its thickness) and consists of superimposed layers of minute calcite crystals growing with their c-axes perpendicular to the sheet (White, 2019).
Many flowstone occurrences in Ignatievskaya cave (some of which have been sampled in this study) have a somewhat less common "runnel" morphology (Figures S1-1). Water, from which the flowstone was deposited, did not wet the entire surface. Rather, it formed 1 to 2-cm-wide runnels, whose lengths range from tens of centimeters to several meters. Deposition of calcite was largely restricted to these runnels. Initial layers were deposited as sub-mmthin lamina. As precipitation continued, small calcite "trails" formed on inclined surfaces of cave walls began to channel the flow.
New calcite layers enveloped previously formed ones, resulting in a gradual thickening of the calcite, eventually forming inverted ridge-shaped deposits, whose length is significantly greater than their width and thickness. The height of the ridges varies from a few to 10−15 mm. The cave wall on the sides of such ridges is either barren or coated with a sub-mm-thin flowstone.

| Sampling strategy
The sampling strategy was specifically designed with conservation in mind keeping damage and disturbance to a minimum. Places where paintings and speleothems occur together are not common in the cave, and those parts of the cave where such relationships can be observed are relatively small. Accordingly, we started with a thorough examination of all locations where spatiotemporal relationships between flowstone and paintings could be assessed.
Sampling for 230 Th dating involved drilling seven mini-cores (8 mm in diameter) from the flowstone. To avoid damage to paintings underlying the flowstone, in all but one case, our cores were taken outside of, but fairly close to, the painted areas. The obvious downside to this approach is that the paint layer is only visible in one of our cores, but in the interest of preservation, we deemed this absolutely necessary. The spatiotemporal relationships between paintings and flowstone, therefore, needed to be ascertained by other means, such as detailed observations on the flowstones and paintings.
Our observations in the cave confirm that speleothems are not abundant and, in many cases, exhibit clearly identifiable functional relationships. For example, we observed that small stalagmites on the floor of the Far Hall are part of the same speleothem-forming system as runnel flowstones on its ceiling ( Figures S1-2). This field determination was subsequently confirmed by 230 Th dating. Stalagmites are thicker and allow fine-scale sampling as compared with thin runnel flowstones overgrowing paintings. These speleothems were sampled to provide more a robust reconstruction of the speleothem growth history in the cave.

| Speleothems in Far Hall
In Ignatievskaya cave, the best-preserved paintings are found in Far Hall. To clarify spatiotemporal relationships between paintings and speleothems, the entire surface of this hall was studied in detail.
The Far Hall is square in plan view (12 × 12 m in size). Its ceiling dips 12-15°toward the east, following the dip of bedding. The floor dips at a somewhat steeper angle due to a massive cone of the surface-derived debris that extends from a fracture on the western side of the hall. Large slabs of the limestone, detached from the ceiling along the bedding, are abundant on the floor. The eastern and southern walls of the hall are nearly vertical. In contrast, the ceiling and floor nearly merge on the western and northern sides of the hall, meaning that vertical walls are not really present in this area. The height of the hall is 1-5 m.
The surface of the eastern wall of the Far Hall is uneven and exposes a series of bedding planes enlarged by karst dissolution; this surface is not a good "canvas" for painting. In addition, it is mantled by a relatively thick (several cm) flowstone issuing from bedding planes. The upper few-mm-thin layers of the flowstone contain abundant soot, giving rise to its black color. Therefore, even if some paintings existed on this wall, they would not be recognizable anymore.
F I G U R E 2 A plan view of Ignatievskaya cave, showing positions of samples and location of archaeological pits, from which 14 C dates were obtained (Smirnov et al., 1990;Petrin, 1992;Dublyansky et al., 2021). The insert shows location of the Ignatievskaya cave in Eurasia [Color figure can be viewed at wileyonlinelibrary.com] All paintings in the Far Hall are located on its ceiling and southern wall. On the surfaces carrying rock paintings, the flowstone occurs in two varieties: calcite runnels and (sub)mm-thin layers of translucent calcite forming small patches of several cm 2 in size. These varieties are closely related and exhibit gradual transitions between each other.

| Paintings and speleothems on the ceiling of Far Hall
The ceiling is devoid of uniform blackening; instead, it shows abundant vermiculations. It features figures consisting of either red (ochre) or black (charcoal) lines. Many motifs appear to be overgrown by runnel flowstone so that the paint is not visible where flowstone is present ( Figure 3a). To assess the possibility that the paintings were made after the flowstone formed, and that the pigment was washed away from protruding surfaces of flowstone runnels, we examined closely the entire surface of the ceiling using optical magnification. We found that flowstone invariably overgrows the paintings-red as well as black ( Figure 3). This is obvious from onlapping relationships, where the paint is visible through thinning-out layers of semitranslucent calcite on the sides of runnels, or where the runnels are generally thin

| 230 Th dating of flowstones
Results of the 230 Th dating of speleothems associated with the paintings are shown in Table 1; detailed results on all dated samples are given in Table S2 (Figure 4c), but there is no evidence of a hiatus between the layers. Being too thin to be dated separately, the older layer was sampled together with the younger yellowish one; another date was obtained on the outer part of the younger layer. As the mixture produced an age that is only slightly older than the outer part of the flowstone (c. 8.9 ka and 7.6 ka BP, respectively), we conclude that the earliest translucent layer also has an early Holocene age.

Flowstone (eastern wall of Far Hall)
Core IGN5 was drilled through thick flowstone deposited on the eastern wall of the Far Hall. A c. 1-cm-thick layer of old flowstone (193 ka BP) is present at the base, which is separated from a c. 7-cmthick overgrowth by a clearly identifiable hiatus. Five samples taken above the hiatus returned stratigraphically consistent ages ranging from 9.9 to 5.6 ka BP (Figure 8).

Stalagmite (Large Hall)
Stalagmite IGN was provided to us by biologist N. Erokhin, who  Where multiple dates were obtained on a given sample, the ages are in a stratigraphic order, which supports the notion of closedsystem behavior of U and Th in these speleothems.
When applying 230 Th dating to speleothems overgrowing paintings, we are interested in the time when the flowstone began to form (terminus ante quem). One limitation of our dating approach in Ignatievskaya cave is that the postpaint flowstone is thin (2−3 mm). In two cases, the entire thickness of the flowstone layers present in the core had to be milled to obtain enough material for dating (Figures 5b and 7c).
Such bulk ages represent a minimum estimate of the flowstone age.
We were able to constrain the time of initiation of calcite deposition for sample IGN3, for which two layers were dated

| Comparison of 230 Th and 14 C dates from the cultural layer
Several radiocarbon dates were reported from the cultural layer found in archaeological pits II, III, and IV excavated in the cave (Figure 2; Table S4). These 14 C dates are shown in Figure 9 along with the results of the 230 Th dating of the postpaint flowstone. We note here that the cultural layer in Ignatievskaya cave is related to the paintings due to clumps of ochre found in this layer (Petrin, 1992, p. 96). Assuming that these 14 C dates correspond to the time of artistic activity in the cave, the 230 Th dates are consistent with (i.e., younger than) the 14 C dates.
Dating of the cultural layer in Ignatievskaya cave is challenging. Furthermore, the sediment in Ignatievskaya cave was subject to some mixing. Archaeologists working in the cave in the 1980s reported that many Paleolithic artifacts (stone tools, decorations, bones) were found directly on the cave floor (Petrin, 1992). Also, bones of spotted hyena (Crocuta c. spelaea), a species that extirpated in Northern Eurasia at around c. 40 ka (Stuart & Lister, 2014), were found on the cave floor.

Conditions in
These bones, which yielded (uncalibrated) 14 C ages of c. 40.2, c.
44.3, >40.1, and >62.3 ka BP (Stuart & Lister, 2014), were most likely derived from deeper sediment layers, where they are abundant (Smirnov et al., 1990). Sediment mixing may have started already in the Paleolithic. For example, two bone samples from a well-defined cultural layer in Pit II (depth: 0.06-0.08 m and 0.12-0.14 m) returned 14 C ages ranging from c. 25.6 to c. 24.4 cal ka BP   | 541 consistent with the new 230 Th dates (Figure 9). Black paintings (c. 9.0−6.5 cal ka BP) appear younger than the flowstone that overgrew both the black (min. 9.6 ka BP) and red paintings (min. 9.7 ka BP) in the same hall.
We note that none of our 230 Th dates refer exactly to the same motif from which the 14 C dates of Steelman et al. (2002) were obtained. In fact, only one of our samples, IGN3, constrains the antiquity of the black painting; others, IGN2 and IGN6, provide terminus ante quem for the red (ochre) paintings. We note, however, that the red motif Bicorn (our sample IGN2) is located very close to the black motif Mammoth dated by Steelman et al. (2002), and that both figures, red and black, are overgrown by flowstone runnels identical to the one sampled as IGN2 (Figure 3a).
In view of the new 230 Th dates, the 14 C ages of black paintings appear to be erroneous (too young). At the time of publication, Steelman et al. (2002) acknowledged that their dates appear "unexpectedly recent" (p. 347). They carefully evaluated the data and discarded many possible reasons for obtaining 14 C ages younger than the "true" age of black paintings. Furthermore, they offered four explanations of the inconsistency, three of which questioned the presumed Paleolithic antiquity of paintings and one suggested a mixing of two painting episodes: "There was repainting of an older image, such that younger overpainting and original painting charcoal combines to give a meaningless radiocarbon age, more recent that the initial painting event" (p. 348). As it is known, in the 19th century, the Far Hall was used for years as a hermitage by Orthodox monk Ignatiy, and after his death, it was frequented by his followers. Visits to the hall involving candles and torches continued through the 20th and into the early 21th century. Given that paintings are readily visible and accessible on the ceiling of the hall, the explanation offered by Steelman and coauthors is not unreasonable.

| Artistic activity in Ignatievskaya cave: Paleoclimatic context
Available geochronological information from Ignatievskaya cave is summarized in Figure 9 in the context of northern hemisphere paleoclimate. The 14 C dates from the cultural layer indicate that the cave was visited by an ancient man during deglaciation that followed the last glacial maximum (LGM  (Panova & Bykova, 1992). Twenty-six charcoal fragments were examined to determine the type of wood that was used by Paleolithic man to make torches (Shiyatov, 1992). Most of the fragments were from the pine (Pinus) and only three were from alder (Alnus). The paleobotanical evidence indicates a sparse vegetation of the periglacial type, characteristic of a dry and strongly continental climate. The landscape most likely featured cold-steppe and forest-steppe, with small patches of forests.

| Permafrost
The period of Paleolithic cave visits, indicated by 14 C dates of the cultural layer, broadly coincides with the "Last Permafrost Maximum" (c. 25-17 ka; Vandenberghe et al., 2014). At that time, the southern boundary of equilibrium permafrost in this part of Eurasia was at about 48°N. Ignatievskaya cave (53.5°N), thus, resided in the area of permanently frozen ground.
It has been shown that no stalagmite deposition occurred between c. 73 and 7.7 ka (i.e., during most of MIS 4 and the entire MIS 3 and 2) in two other Southern Ural caves, Shulgan-Tash and Victoria (Dublyansky et al., 2018). This is consistent with the interruption of speleothem growth in Ignatievskaya cave between c. 78 and 10 ka. Permafrost conditions are known to have interrupted speleothem growth in Siberian caves (Vaks et al., 2013). Similarly, the gap in speleothem growth in Ignatievskaya cave is attributed to permafrost in Southern Ural. As Ignatievskaya cave is located c. 200 km to the north of Shulgan-Tash and Victoria caves, at a similar elevation, it is reasonable to assume that at this time, the cave was also located in the zone of continuous permafrost, where the thickness of perennially frozen ground reached c. 200 m and the mean annual air temperature was between −5°C and −3°C (Velichko et al., 2002;Vandenberghe et al., 2014). The conclusion reached for Shulgan-Tash cave that "… during the time of artistic activity … cave was an inhospitable place with below-zero air temperatures throughout the year and no dripping or running water. The paintings were made by Paleolithic artists on cave walls that were dry; our data indicate that the dryness was due to the freezing temperatures" (Dublyansky et al., 2018, p. 5) is also valid for Ignatievskaya cave.

230
Th dates obtained on flowstone that formed before and after the red and black paintings in Ignatievskaya cave constrain the age of the artistic activity between c. 78 and c. 10 ka. This long period of time during which speleothems did not form reflects cold climatic conditions and the development of permafrost in Southern Ural. This interpretation is further supported by the available paleozoological and palynological data from the cultural layer in the cave, featuring species adapted to harsh cold climates.
The 14 C dates obtained from bones and charcoal of the cultural layer in the deep, aphotic zone of the cave Petrin, 1992;Smirnov et al., 1990) are broadly consistent with the 230 Th dating results. They place the cave visits (and, by extension, the creation of cave art) at c. 18.3−15.8 cal ka BP, that is, the late glacial period.
Three 14 C dates obtained directly from charcoal paintings in the Far Hall of the cave (Steelman et al., 2002) appear too young and are inconsistent with the 230 Th dates. These 14 C dates are younger than the flowstone that formed after both red and black paintings in the hall were made.