A study examines how agriculture influences mortality due to poor air quality in the United States.
Image credit: Jason D. Hill.
Edited by Richard G. Klein, Stanford University, Stanford, CA, and approved July 25, 2019 (received for review June 28, 2019)
Significance
We found a previously unknown way to produce birch tar. Instead of creating cognitively demanding structures (underground or in containers), this method consists of simply burning bark close to cobbles in a hearth. The tar is deposited on the stones and can be scraped off for use. This approach to interpreting early tar resolves the mystery of the associated and still not understood early technical complexity and provides a “discoverable” pathway to one of the earliest pyrotechnologies. These results have implications for our interpretation of birch tar in the archaeological record: Birch tar from early archaeological contexts alone can no longer indicate the presence of modern cognition and/or cultural behaviors in Neanderthals.
Abstract
Birch tar production by Neanderthals—used for hafting tools—has been interpreted as one of the earliest manifestations of modern cultural behavior. This is because birch tar production per se was assumed to require a cognitively demanding setup, in which birch bark is heated in anaerobic conditions, a setup whose inherent complexity was thought to require modern levels of cognition and cultural transmission. Here we demonstrate that recognizable amounts of birch tar were likely a relatively frequent byproduct of burning birch bark (a natural tinder) under common, i.e., aerobic, conditions. We show that when birch bark burns close to a vertical to subvertical hard surface, such as an adjacent stone, birch tar is naturally deposited and can be easily scraped off the surface. The burning of birch bark near suitable surfaces provides useable quantities of birch tar in a single work session (3 h; including birch bark procurement). Chemical analysis of the resulting tar showed typical markers present in archaeological tar. Mechanical tests verify the tar’s suitability for hafting and for hafted tools use. Given that similarly sized stones as in our experiment are frequently found in archaeological contexts associated with Neanderthals, the cognitively undemanding connection between burning birch bark and the production of birch tar would have been readily discoverable multiple times. Thus, the presence of birch tar alone cannot indicate the presence of modern cognition and/or cultural behaviors in Neanderthals.
Early birch tar (henceforth tar) production by Neanderthals has been interpreted as a marker of complex technology (1), high planning depth (2) and enhanced cognitive capacity (3). It is known from the Middle Paleolithic sites of Campitello (∼200 ka [4]), Königsaue (∼84 to 40 ka [5]), and possibly Inden–Altdorf (∼120 ka [6]), leading some to argue that Neanderthals were the first to create complex production of adhesive (2, 7). The potential implications of early tar contrast with the absence of direct archaeological data on the techniques used for early tar making (8). Most interpretations are based on experimental setups involving containers and intentionally created reducing conditions (e.g., refs. 9⇓⇓–12) and sometimes elaborate experimentation (13). For example, useable quantities of tar can be produced if bark is indirectly heated in an earthen oven-like structure, a construction known as a clay castle or raised structure (14). Among aceramic techniques, raised structures come closest to techniques using ceramics or metal containers in terms of tar yield (compare refs. 2 and 10). Tar can also be produced by covering bark rolls entirely with ash and embers, with or without a fire on top (2, 13, 15). The tar is then collected from the bottom of such structures, either with a receptacle or in the windings of the bark roll itself. Tar can also be produced from bark in shallow pits, the burning end pointing downward, and the tar collected at the base (16). However, all of these experimental techniques have relied on one main assumption, that tar can only be produced by dry distillation in reducing environments (where the lack of oxygen prevents the tar’s immediate combustion). This idea likely goes back to the discovery of tar distillation apparatuses from the Bronze Age of Italy (e.g., ref. 17). The 2-container method is well-documented from Roman times, and medieval texts describe tar distillation (12). The first experimenters trying to replicate early tar making (e.g., refs. 10 and 18) adopted the assumption of the necessity for reducing conditions. The result has been that all proposed experimental techniques to date require a degree of complexity that is unlikely to have come about by chance. Moreover, these already specialized techniques presuppose the knowledge and expectation of the technique’s outcome: i.e., that tar (as a useful material) can be produced intentionally using the applied (complex) technique. But where, then, did this knowledge come from in the first place? If an easier, more intuitive and more likely technique were to produce sufficient amounts of tar, then complex techniques of tar production might have been unnecessary. In turn, the identification of tar in the archaeological record would cease to be a proxy for technological and cultural complexity. Accordingly, we investigated whether such an alternative, uncomplex, and readily discoverable method of tar production exists.
Results
We conducted systematic experiments using readily occurring, open-air conditions. First, we tested whether tar forms during burning while birch bark is still attached to wood in the pore space between bark and stem—when branches are only partially lit or when detached bark is placed onto embers. No tar formation was observed in these conditions, highlighting that birch tar is not a byproduct of open-air birchwood fires.
We then tested whether tar would form when burning birch (Betula pendula) bark alone, i.e., detached from the wood. This situation would likely have been frequent in the past, as birch bark is 1) a natural tinder (19) (burning well, even when wet) and 2) readily collectible both from trees or (even easier) from forest floors, where the birch bark tends to stay recognizable and useable for some time after the wooden core has already rotted away.
Burning birch bark on a stone surface did not yield recognizable amounts of tar. Then, we burned birch bark at the side of a stone, i.e., next to a subvertical hard surface. In this situation, a black shiny deposit formed at the interface between stone and flames (we used a variety of different river cobbles with flat surfaces during successive runs, including quartz, limestone, and silt-stone). Burning birch bark next to such stones would likely have occurred frequently in the past, and also, we obtained similar results using bone instead of stone. The adhering substance was immediately sticky to the touch and remained sticky when scraped from the warm surface. From this we deduce that our first goal was already reached: tar production can be an accidental, and indeed even a likely, outcome of everyday activities for any group building fires with birch.
From this observation, we established a minimally complex protocol that would accumulate the sticky material across multiple burning events. Each event repeated the basic technique, i.e., a birch bark piece (which naturally rolls) was lit and burned beside a river cobble (the flames of the burning bark were measured at 600 to 700 °C using a thermocouple). The cobble was put on the ground to provide a flattish, rounded surface slightly overhanging the burning bark, in our case forming angles between ∼60° and ∼80° with the ground (Fig. 1 A and B). After repeating this burning procedure 2 or 3 times, the stone was picked up, and the (black and sticky) material (henceforth tar) was scraped from the cobble with a stone tool (a small flake independently produced, a skill that can be taken for granted in Neanderthals) before the process was repeated (Fig. 1C).
Fig. 1.
Experimental birch tar making with the condensation technique. (A) Schematic drawing of the experimental setup: a cobble (1) with an inclined surface overhanging a piece of birch bark (2) is used as support for the condensation of birch tar directly above the burning bark (3). (B) Photo taken during experimentation using the setup shown in A. (C) Photo of the cobble surface where tar can be scraped off and the stone tool used for scraping. (D) Photo of a 0.62-g piece of tar produced in a single 3-h session (including bark collection).
One can produce tar with bark cut from living birch trees and/or dead bark picked up from the forest floor. To assess the productivity of the latter method in terms of tar yield, we collected dead birch bark (as it is more easily accumulated than fresh bark) from a 20-m-long transect with a breadth of 4 m in a birch forest (total area of 80 m2). This yielded 600 g of dead bark in a collection time of 27 min. When burned with the condensation technique, each 100 g of the dead bark resulted in 0.18 g of tar (as averaged from 3 measurements: 0.11, 0.08, and 0.13 g of tar from 57.7, 59.5, and 60.2 g of bark, respectively). In another experiment, we obtained 0.1 g of tar from dead bark approximately every 25 min, using 1 cobble at the time (a total of 0.62 g was made during this experiment; see Fig. 1D). While the tar yield by our condensation technique is 5 1/2 times lower than with the bark-burned-under-ash-and-embers technique—the arguably “simplest” of the more complex reducing conditions techniques (2)—it nonetheless produced useable amounts of tar (see also below) in a comparable amount of time. (Tar making with the condensation technique was filmed and can be seen in Movie S1.)
To assess the actual suitability of our condensation method for tar production (and the amounts produced in sensible time frames) we used our tar to slot-haft a Baltic flint flake and performed 2 experiments using the tool, first in a controlled, robot-aided setting (scraping wood) (Fig. 2A) and, subsequently, in an actualistic one (bone defleshing) (Fig. 2B). For these experiments, we used 0.6 g of pure (unmixed) tar that was produced by the condensation method (accumulated in a single 3-h session, including raw material collection time). The hafting arrangement consisted of a stone bit inserted and fixed by the birch tar (heated with a flame and dripped onto the haft) into a wooden cylinder (31.5 mm in diameter, 75 mm length) with a 12-mm-deep slot on 1 end (7 mm wide).
Fig. 2.
Analysis of birch tar produced by the condensation technique. (A) experimental setup using the robot arm for wood scraping under controlled conditions. (B) Actualistic defleshing experiment using the same hafted tool as in A. (C) Three photos taken of a single sample at different moments during a lap shear test, (Left) 93.3 MPa before plastic deformation of the tar; (Middle) 90.7 MPa at the beginning of plastic deformation; and (Right) after failure of the tar. (D) Chromatogram of tar produced with the condensation technique showing biomarkers and markers of heat treatment: 1 = lupa-2,20 (29)-diene; 2 = α-betuline I; 3 = lupa-2,20 (29)-dien-28-ol; 4 = lupeol; 5 = betulin. RT, retention time.
We programmed the robot arm (KUKA LWR 14) to drag the hafted tool over a wooden panel under constant vertical load (Fig. 2A). After each stroke the arm repositioned the tool through the air to the same starting point (Movie S2). We chose a stroke length of 19 cm and downward force of 100 N and a working angle of 60°. The whole robot experiment took ∼19 min with a total of 170 strokes pulled by the robot arm; we did not observe any weakening of the adhesive connection between the stone and its haft.
Following this, the same hafted tool was reused in a manual cutting experiment, defleshing an ∼30-cm-long calf femur fragment (Bos spec.). The aim was to remove the remaining meat and periosteum. Transversal scraping and longitudinal cutting motions as well as hacking were performed using the full pressure needed to complete defleshing. The experimental goal was reached at ∼20 min, after which we cleaned and critically inspected the haft. We observed no detachment or weakening of the tar connecting the stone tool to its handle (the experiment was filmed and can be seen in Movie S3). Thus, tar produced with the condensation technique is perfectly useable under both controlled laboratory and real-world working conditions, while presenting the expected adhesive properties.
While these tests illustrate the suitability of tar produced with our condensation method for hafting, it remained unknown how it would perform relative to tar produced with other, more complex techniques. We therefore conducted a lap shear test (Fig. 2C) according to a protocol proposed for testing archaeological adhesives (following the ASTM D1002 guidelines but modified to use wooden laps instead of aluminum, following the reasons given by ref. 20). For this test we produced another 0.3 g of tar using our condensation method (again from dead bark collected from the floor of a birchwood forest). The mean of 10 lap shear tests resulted in a strength of 1.145 MPa +0.403 – 0.438 (as calculated from 10 tests; SI Appendix). This strength value is >3 times above the only published lap shear values obtained from birch tar produced with the 2-container method (0.32 MPa +0.19 – 0.18 from ref. 21) and agrees with strength values measured on pine pitch (with average values ranging from 0.37 to 1.77 MPa according to pretreatment [21]), being only slightly inferior to compound adhesives based on beeswax, conifer resin, and ochre (with average values ranging from 1.27 to >3 MPa; ref. 20). Thus, pure birch tar produced in our aerated environment has similar adhesive properties to other natural adhesives, and it actually outperformed birch tar produced in a reducing environment using a more complex technique.
Gas chromatography–mass spectrometry (GC–MS) was conducted to test if birch tar produced under air using the condensation method contains similar molecular markers as anaerobically made tar and known archaeological birch tar (Fig. 2D). Sample extraction and analytical conditions were performed following protocols established for birch tar analysis (22). GC–MS has previously been used to identify Paleolithic tar as being birch bark tar through the presence of characteristic pentacyclic triterpenes, in particular, betulin and lupeol (23), and their degradation markers that may indicate heat treatment (e.g., lupa-2,20[29]diene, lupa-2,20 [29]-dien-28-ol, and allobetul-2-en). This has been the case of the 2 oldest GC–MS-analyzed examples of birch tar from Campitello and Königsaue (24, 25). Analyzing our own experimental tar, we also identified both biomarkers (betulin and lupeol) in addition to 3 degraded markers (lupa-2,20 [29]-diene, α-betuline I, and lupa-2,20 [29]-dien-28-ol) that have been described to indicate heat treatment in previous experimental studies (22) and that were found in archaeological tar samples (26). Thus, tar produced in oxygenized environments with our experimental setup provides a molecular signature of birch bark tar. Future chromatographic analyses of archaeological tar should shed further light on the similarities and dissimilarities of tars produced in different environments.
Discussion and Conclusion
Birch tar production has long been thought to take place under oxygen exclusion only (2), i.e., in technologically complex and/or unlikely settings. Reducing environments allow the preservation of several chemical components that might otherwise burn off (5). This has led many researchers to propose birch tar production using heating systems that create anaerobic conditions in containers or underground (e.g., “clay castle,” eggshell, ash mounts, ceramic containers, etc.; refs. 2, 8, and 27). However, we found that useable amounts of birch tar form in fully oxygenized environments, simply as a redisposition onto a surface, in what we call the condensation method. The underlying chemical process is likely dry distillation as for techniques using reducing conditions because the tar’s chemical components transit by a gaseous phase before condensing on the surface. Whether it is oxygen depletion due to the nearby combustion or simply slow oxidation–reaction kinetics that prevent the tar from burning off cannot be decided without further analyses.
Thus, although our experiments do not elucidate the chemistry associated with tar production by our condensation method, they show that the creation of anaerobic systems as described by previous authors (see, e.g., ref. 5) are not necessary for tar making. The identification of birch tar at archaeological sites can no longer be considered as a proxy for human (complex, cultural) behavior as previously assumed (e.g., refs. 3, 14, and 28). In other words, our finding changes textbook thinking (29, 30) about what tar production is a smoking gun of.
As our results show, tar production does not require complex cognition, nor high planning depth, and it can derive from the simple juxtaposition of 2 everyday objects for Neanderthals (birch bark and stone or bone surfaces) derived from fire making/tending. While some parts (fire making/tending—see the current debate on whether Neanderthals were able to make fire [31, 32]—and perhaps hafting in itself) may or may not be good indicators of complex, modern human-like cognition, the condensation technique itself is not: a mere repetition of bringing 2 objects in close proximity and gathering of a resource is well within the cognitive power even of nonhuman great apes (33, 34). So, the natural (instead of cultural) intelligence of Neanderthals may have sufficed for the condensation method to 1) be innovated, possibly even multiple times, and 2) be preserved in populations via a process of “socially mediated serial reinnovations” (35). The latter is clearly a case of minimal culture (36). However, because minimal culture is very widespread in the animal kingdom, it is not only within the assumable abilities of Neanderthals, but also even for the earliest of hominins (37).
A more distinctive question, however, is whether birch tar making can be used as a proxy for Neanderthal’s ability to show “cumulative culture” (38). In cumulative culture, cultural transmission, via actual copying of techniques rather than socially mediated reinnovation, over time necessarily leads to culture-dependent traits (39)—traits that cannot or are unlikely to be reinnovated. Arguably, this is not the case for tar production using a method as simple as the condensation method (see above).
This finding is important because modern human culture itself relies on culture-dependent traits, and it is currently debated which hominins (and when, and how often) had such culture-dependent traits (37, 40, 41). To throw light on these uncertainties loaded with implications for human evolution, we need a data-driven approach to archaeological finds to determine which and when they show signs of having been culture-dependent. As for tar production, the presence of tar in the archaeological record alone can no longer count as a secure case for culture-dependent traits in hominins, as the condensation method we describe even seems to be the likely method—potentially always serially reinnovated rather than copied—by which Neanderthals produced tar.
A future perspective that would allow further light to be shed on this hypothesis is comparing known artifacts associated with Paleolithic birch tar with the material produced by our own experimental tar making. Indeed, at Inden–Altdorf a sandstone cobble covered in a black tar (not yet confirmed to derive from birch) was found. Although the cobble is currently interpreted as a recipient for collecting tar in an underground structure (6), we note the striking similarity with the tar-covered cobbles we produced with our own condensation technique (compare Fig. 1 B and C with figure 3 in ref. 42). Thus, for now, the available archaeological data do not contradict our hypothesis, and we predict that future detailed analysis of new finds should strengthen our interpretations of early birch tar making.
Our findings do not necessarily lead to the conclusion that Neanderthals were not able to conduct complex procedures, nor that they were not capable of abstract thinking or high planning depths. In fact, Neanderthal modernity has been convincingly argued for based on a whole suite of behaviors (e.g., ref. 1). We merely note that, in archaeological science in general, arguing for abstract concepts like modernity or complex cognition in past populations should not rely solely on highly interpretative models of the production pathways of specific material finds. It should rather rely on the interpretation of the actually performed steps, as proven by direct archaeological data. If this is not possible, as in the case of birch tar, where direct evidence of the technique used by Neanderthals is still missing, our results highlight that the only viable interpretation of the implications of material remains is to admit the simplest possible pathway by which they can be produced. It is therefore no longer possible to use early birch tar making as proxy for complex cultural behaviors in Neanderthals.
Methods
Robot Arm.
An industrial robot arm (KUKA LWR 14) was programmed to drag the tool with straight strokes of 19 cm in length over a wooden panel. After each stroke the tool was repositioned through the air to the same starting point. The downward force was kept at 100 N, and the working angle between underground and hafting was kept at 60°. In total, 170 strokes were executed, which corresponds to a duration of ∼19 min.
Manual Cutting Experiments.
An ∼30-cm-long calf femur, purchased from a local butcher after preliminary removal of meat, was subjected to scraping and cutting motions by a 39-y-old, 75-kg city-dwelling male (R.I.), with the aim of removing the rest of the meat and the periosteum in the shortest time possible. Initially, a longitudinal cut along the periosteum was made using the tool longitudinally, followed by scraping motions using the tool transversally. To finally detach the periosteum, some mildly violent hacking motion was necessary, especially since the tool’s edge had been dulled by the previous experiment.
Mechanical Testing.
Lap shear tests were performed using an Instron 4502 universal test machine with kardanic suspended tensile grips, where laps were mounted vertically and pulled apart with a speed of 1 mm min−1. Laps were cut and precision-ground from 4-mm-thick Populus spec. polywood measuring 100 × 25.5 mm. The 25.5 × 12.5-mm measuring contact zones (319 mm2) were abraded with 100-grit sandpaper. Tests were repeated 10 times.
Chemical Analysis.
Sample preparation and GC and GC–MS analyses were performed using the method described by refs. 22 and 43. Briefly, the sample was ground and then extracted in HPLC–grade dichloromethane (1 mg mL−1). GC and GC–MS analyses were performed using an Agilent Technologies 7890B GC System series chromatograph including Agilent Technologies Capillary Flow-Technology Three-Way Splitter Kit coupled to an Agilent Technologies 5977A MSD and FID.
Acknowledgments
V. Schmid, S. Wolf, and F. Lauxmann contributed to the experiments, and we are indebted to B. Schürch. We also thank R. L. Kelly for his comments on the manuscript.
Footnotes
- ↵1To whom correspondence may be addressed. Email: patrick.schmidt{at}uni-tuebingen.de.
Author contributions: P.S. and C.T. designed research; P.S., M.B., M.R., R.I., J.P., L.R., and C.T. performed research; K.G.N. and L.R. contributed new reagents/analytic tools; P.S. and C.T. analyzed data; and P.S. and C.T. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1911137116/-/DCSupplemental.
Published under the PNAS license.
References
Birch tar production does not prove Neanderthal behavioral complexity
Article Classifications
- Social Sciences
- Anthropology
Sign up for Article Alerts
Jump to section
You May Also be Interested in
Researchers report a unique quasicrystal discovered in the remnants of the first nuclear bomb detonation.
Image credit: Luca Bindi and Paul J. Steinhardt.
A study estimates that there are 50 billion birds in the world, with the majority in palearctic and nearctic realms.
Image credit: Corey T. Callaghan.
People are increasingly dividing themselves into social and political factions. Models can hint at how it happens—and maybe offer ways to mitigate it.
Image credit: Dave Cutler (artist).
A computer visualization of the fossilized plant Asteroxylon mackiei could potentially offer clues as to how modern plants emerged.
Image credit: Matt Humpage (artist).