SciELO - Scientific Electronic Library Online

 
vol.113 issue3-4Determining the relationship between invasive alien species density and a country's socio-economic statusNo 'free ride' for African women: a comparison of head-loading versus back-loading among Xhosa women author indexsubject indexarticles search
Home Pagealphabetic serial listing  

Services on Demand

Article

Indicators

Related links

  • On index processCited by Google
  • On index processSimilars in Google

Share


South African Journal of Science

On-line version ISSN 1996-7489
Print version ISSN 0038-2353

S. Afr. j. sci. vol.113 n.3-4 Pretoria Mar./Apr. 2017

http://dx.doi.org/10.17159/sajs.2017/20160085 

RESEARCH ARTICLE

 

The age of fossil StW573 ('Little Foot'): An alternative interpretation of 26Al/10Be burial data

 

 

Jan D. KramersI; Paul H.G.M. DirksII

IDepartment of Geology, University of Johannesburg, Johannesburg, South Africa
IIDepartment of Geosciences, College of Science and Engineering, James Cook University, Townsville, Queensland, Australia

Correspondence

 

 


ABSTRACT

Following the publication (Granger DE et al., Nature 2015;522:85-88) of an 26Al/10Be burial isochron age of 3.67±0.16 Ma for the sediments encasing hominin fossil StW573 ('Little Foot'), we consider data on chert samples presented in that publication to explore alternative age interpretations. 10Be and 26Al concentrations determined on individual chert fragments within the sediments were calculated back in time, and data from one of these fragments point to a maximum age of 2.8 Ma for the sediment package and therefore also for the fossil. An alternative hypothesis is explored, which involves re-deposition and mixing of sediment that had previously collected over time in an upper chamber, which has since been eroded. We show that it is possible for such a scenario to yield ultimately an isochron indicating an apparent age much older than the depositional age of the sediments around the fossil. A possible scenario for deposition of StW573 in Member 2 would involve the formation of an opening between the Silberberg Grotto and an upper chamber. Not only could such an opening have acted as a death trap, but it could also have disturbed the sedimentological balance in the cave, allowing unconsolidated sediment to be washed into the Silberberg Grotto. This two-staged burial model would thus allow a younger age for the fossil, consistent with the sedimentology of the deposit. This alternative age is also not in contradiction to available faunal and palaeomagnetic data.
SIGNIFICANCE:
Data on chert samples taken close to StW573 impose a maximum age for the fossil of 2.8 Ma - younger than the 3.67 Ma originally reported. We propose and explore a two-stage burial scenario to resolve the inconsistency and to reopen the discussion on the age of fossil StW573.

Keywords: burial dating; cosmogenic nuclides; muons; Australopithecus prometheus; burial history


 

 

Introduction

In a recent contribution, Granger et al.1 present 10Be and 26Al data on quartz from Member 2 sediments in the Silberberg Grotto in Sterkfontein Cave, South Africa, encasing StW573 ('Little Foot')2, a complete skeleton referred to as Australopithecus prometheus1. The apparent burial isochron date of 3.67±0.16 Ma is interpreted as the age of StW573.

Almost since the discovery of StW573, its age has been a subject of controversy. Based on the concept of a laterally continuous stratigraphy for the Sterkfontein Formation3 and a palaeomagnetic fit, an age of about 3.3 Ma was first proposed4. A subsequent review5 of mainly faunal data suggested a much younger age range of between 1.5 Ma and 2.5 Ma for Member 4 at Sterkfontein as well as for the sediments encasing StW573. In a response6, the lower age limit for Member 4 was firmly placed at ca 2 Ma (a limit since confirmed by an U-Pb age on its capping flowstone7) but the concerns5 regarding an age older than 3.0 Ma for StW573 were not fully dispelled. Cosmogenic 26Al/10Be burial dating8 then indicated that quartz in the sediments around the fossil had been underground for 4.17±0.35 Ma (later recalculated1 to 3.94±0.20 Ma). Because of the possibility that the quartz was reworked from previous higher levels in the cave system, this date can be regarded as a maximum age.1,7 In contrast, U-Pb dates9 of ca 2.2 Ma on CaCO3 speleothem units from below and above StW573 are minimum ages, as the dated units are not stratigraphic flowstones, but fracture fillings10. The new burial isochron date1 fits within these age brackets.

The cave deposits encasing StW573 form the northwest flank of a sediment cone that occupies most of the Silberberg Grotto area and has its apex in the eastern part of the Grotto where the ceiling is highest.11 The deposits consist mainly of matrix-supported breccia units composed of coarse-grained chert clasts and dolomite blocks, set in a muddy sand matrix that is mostly well calcified.10 Stratification of the breccia units is illustrated by the presence of a finer-grained, weakly consolidated (possibly decalcified), clay-rich sand layer (Unit B2a in Bruxelles et al.10) near the base of the dated sedimentary pile, and variations in size (Unit B310) and frequency (Unit B2b10) of chert and dolomite clasts, indicative of several fining upward cycles.10 Fossil StW573 is embedded within Unit B2b and positioned along the stratigraphic top of this unit.10 The layering, matrix-supported nature of the breccia deposits and clast size variations suggest that a succession of sheet-like, sand-rich debris flows deposited the composite package over a (geologically speaking) short period of time in a process similar to the deposition of the A. sediba skeletons at the Malapa site.12 The age of the fossil must be similar to the depositional age of the breccia units as it is complete, articulated13, and fully incorporated within the breccia.

In order to date StW573, Granger et al.1 applied the burial isochron method14-16 to the host breccia deposits. At the surface, quartz accumulates 10Be and 26Al by spallation reactions caused by neutrons, in turn produced in the atmosphere by cosmic rays. As the quartz is buried (whether in an alluvial or glaciogenic sediment, or a cave) it is shielded from neutrons and 10Be and 26Al production all but ceases (there is a much lower production rate at depth, caused by muons, discussed below). Because 26Al decays faster than 10Be, the 26Al/10Be abundance ratio decreases with time, allowing a burial age to be determined if their pre-burial ratio is known (this ratio can be calculated from models, or directly measured).

The data presented by Granger et al.1 are of high quality and the interpretation appears flawless. However, the discussion on faunal data is not closed17 and in view of the great importance of the age of StW573 in the timeline of hominin evolution, we have re-examined the data on which the 3.67±0.16 Ma burial isochron date is based, and we present an alternative interpretation that is consistent with the data, but indicative of a younger age.

 

Chert samples, in-situ steady-state concentrations and a maximum age

In order for a burial isochron to be useful, two conditions have to be met. First, the samples must have had, at the time of burial, a primary spread in 10Be and 26Al concentrations that show a correlation with each other. Second, all samples must have been buried at the same time and share the same post-burial history. The cosmogenic isotopes then decay in proportion to their abundance, so that the correlation in 26Al versus 10Be space persists along an isochron for which the slope decreases with time.

A primary spread of 10Be and 26Al concentrations with a correlation can result if (1) samples are mixtures of surface derived and previously buried quartz grains14, or (2) material with varying residence times at the surface is sampled. In the former case, a linear correlation is expected14, whereas the latter situation results in a gentle convex-up curve18 as a result of 26Al decaying faster than 10Be. In Granger et al.1, which is the first application of burial isochron dating to cave chronology, both factors contribute to the spread of data. Six quartz separates from mainly surface-derived bulk sediment samples (ST1, ST2, ST3, ST8, ST9 and composite STM2 dark) have relatively high 10Be and 26Al concentrations. Further, three chert fragments (M2CA, M2CB and M2CC) taken from the breccia immediately adjacent to the fossil and a composite (STM2 light, consisting of chert grains from bulk sediment samples ST1 and ST2) have low 10Be and 26Al concentrations. These are considered by Granger et al.1 to have been derived from higher levels in the cave, at a few metres below the surface. Together, these two sample populations define the slope of the isochron and thus the age. The chert samples yielded data with exquisite precision1, and can provide more information than just the definition of an isochron by regression. To examine this aspect it is first necessary to discuss the concentration of cosmogenic nuclides produced at depth.

Cosmogenic nuclide production rates decrease rapidly with depth under the surface, as neutron penetration in soil and rock is limited to ca 2.5 m. However, even after deeper burial, quartz still accumulates cosmogenic nuclides as a consequence of the action of muons19,20, and this subsurface nuclide production ultimately determines the concentrations of these nuclides at depth. At constant depth underground (e.g. zero erosion), the 10Be and 26Al concentrations of quartz derived from the surface or higher in the cave converge to in-situ steady-state or secular equilibrium (also known as saturation) values, as the abundance of each isotope is adjusted so that its loss by radioactive decay balances its production rate P (in 106 atoms g-1Ma-1), which depends on depth underground. Steady state is achieved after about 10 Ma, and the relation between P values and the in-situ concentrations (atoms g-1) is then:

where λ10 and λ26 are the respective decay constants; λ10=0.4988±0.0050 Ma-1 and λ26=0.9794±0.0230 Ma-1.1 In the more general situation in which erosion occurs, secular equilibrium is not achieved, as the shielding by overburden is steadily reduced and the nuclide production thus increases with time. For any given present-day depth, in-situ produced cosmogenic nuclide concentrations are always lower under erosion than they would be at zero erosion.

In addition to a burial age, a 26Al versus10Be isochron diagram can also yield the 26Al and 10Be concentrations produced in situ after burial.1 These post-burial concentrations must plot on the isochron and also on a line through the origin of the diagram with a slope corresponding to the 26Al/10Be abundance ratio of post-burial production (given by Equation 1 as ca 4.1 in steady state, although it is slightly depth dependent). This is dotted line 's' in Figure 1a, and its intersection with the isochron, P/, should represent the post-burial produced abundances: (8.5±1.3)x104 atoms/g of 26Al and (2.1±0.3)x104 atoms/g of 10Be (1σ uncertainties given, plotted with 2σ error bars).1 We note that the uncertainty of P/ is probably underestimated. The reason is that two of the samples analysed by Granger et al.1 to define the isochron and the P/ values are composites: 'STM2-light' consists of chert fragments from samples ST1 and ST2, and 'STM2-dark' of soil-derived, iron-oxide coated, rounded quartz grains from samples ST1, ST2, ST8 and ST9. The use of aggregates from multiple samples is common practice in measuring erosion rates, especially when the average value for a whole catchment is sought.21 However, in a test for collinearity, this use is inappropriate as it hides any heterogeneity that may have existed.

The in-situ cosmogenic nuclide production rates at the StW573 site can also be calculated directly, using the equations of Heisinger et al.19,20 and the empirically determined cross sections and probability factors reported by Balco et al.15 The present depth below surface of StW573 is 23 m,3,11 and the Silberberg Grotto above it was entirely filled by a flowstone boss and overlying breccia until this was mined out in the early 20th century. The bulk rock density used is 2.5 g/cm3, which (with a density of 2.85 g/cm3 for dolomite) allows for about 12% porosity if air-filled, or 8% if water-filled. An altitude correction (1500 m) was applied to the component of cosmogenic nuclides produced by stopped negative muons (accounting for between 6% and 10% of production at 23 m depth). A latitude correction, if made, would account for the deflection of cosmic rays by the earth's magnetic field at low latitudes and reduce the calculated nuclide production rates.18,22 However, fast muons that produce the bulk of 10Be and 26Al at the given depth are yielded by cosmic rays with energies >20 GeV, which are unlikely to be deflected significantly. For stopped negative muons there is probably a latitude effect although it is difficult to quantify.22 No latitude correction was applied, and the in-situ 10Be and 26Al concentrations obtained may be overestimated by up to a few per cent as a result.

With the above parameters, steady-state in-situ abundances of 0.01 65x106 atoms/g 10Be and 0.0678x106 atoms/g 26Al are calculated if erosion is zero (grey diamond calci in Figure 1a). If an erosion rate of 5 m/Ma (the minimum rate determined by Granger et al.1) is assumed, present-day in-situ abundances of 0.01 20x106 atoms/g 10Be and 0.0 5 7 7x106 atoms/g 26Al result (grey diamond calc2 in Figure 1a). If the erosion rate was 5 m/Ma and a cave chamber had existed above the Silberberg Grotto in the past (as discussed below), calculated values would be intermediate between calci and calc2. The higher the erosion rate, the lower the in-situ concentrations. The calculated results are, respectively, within and close to the 2σ (95% confidence) uncertainty limits of point P/ derived from the isochron regression, which are underestimated, as discussed above. While there is thus no real contradiction between the calculated values and P/, the difference is nevertheless significant when reconstructing the isotope abundances of chert samples through time, as shown below.

For quartz derived from the surface or from higher levels in a cave, convergence towards the in-situ produced 10Be and 26Al concentrations in quartz at any depth is given by the decay law:

where f=0 denotes the present. Equation 2 can be used to calculate 10Be and 26Al concentrations in the past (with t being negative) and future (with t being positive) based on the measured concentrations at present. If erosion is non-zero, the process can be modelled by dividing it into time steps, each with its own depth-specific production rates.

As seen in Figure 1a, none of the chert samples have in-situ equilibrium 10Be and 26Al concentrations, in accord with the assessment1 that they are derived from higher levels in the cave. Black arrows show their convergence on P/ and grey arrows on calc2. Each sample has its own individual set of decay paths (calculated by using Equation 2 forward and backward in time from their measured concentrations), reflecting their individual histories within the cave. Further, although the concentrations of P/ and calc2 are not very different, the decay paths for the two points differ markedly from each other.

Sample M2CA plots significantly below the isochron, and was not included in the regression of Granger et al.1 as it was considered reworked, i.e. to come from a previous burial location in the cave system. The back-correction for sample M2CA using production rate values for the PI abundances (black arrows and symbols in Figure 1b), yields a curve that lies significantly (well outside 2σ, i.e. 95% confidence) below the surface production curve even at 5 Ma. This value is considered a likely maximum age for cave systems to have opened in the Cradle of Humankind UNESCO heritage site21,23, as suggested by the absence of older fossils in the area17,24. A derivation - even from a few metres below the surface (which would allow a 26Al/10Be ratio range down to ~4.5) -is impossible for sample M2CA, because the absolute 10Be abundance several million years ago would then be much lower. This mismatch suggests that the centre values for PI as derived from the isochron regression are inaccurate.

If M2CA is back-corrected using the parameters for calc2 (i.e. a surface erosion rate of 5 m/Ma), the problem of its previous burial history is solved. A marginal match with near-surface abundances is achieved upward of 4.6 Ma (grey symbols and line in Figure 1b) and there is a good fit with an initial burial age of ca 5 Ma. The 10Be and 26Al concentrations of calc2 correspond to the approximate upper limit for effective in-situ production rates that can provide a realistic back-correction for this sample. This result also indicates that a cave system existed at the Sterkfontein locality as early as ca 5 Ma ago, and that material reworked from this system was ultimately deposited in the Silberberg Grotto.

As the three chert samples (M2CA, M2CB and M2CC) were taken close to each other1 (and to StW573), the same in-situ 10Be and 26Al production rates must have applied to all three after the sediments encasing the fossil were deposited. Using the parameters for calc2 to examine the past of the other chert samples is, therefore, a realistic approach. Sample M2CB yields an upper age limit for the deposit that has implications for the maximum age of StW573. In Figure 1c, the back-corrected 10Be and 26Al concentrations for this sample are shown together with the surface production curve. 26Al/10Be ratios cannot plot above this curve (the 'forbidden zone'). The back-corrected values for M2CB using calc2 production rates cross this limiting curve at 2.5 Ma, and lie within the forbidden zone outside 2σ (95% confidence) limits for ages over 2.8 Ma (grey symbols in Figure 1c). Values for 3.6 Ma clearly lie far in the forbidden zone. Sample M2CC is uninformative: it plots so close to the in-situ values that, in back-correcting, its error limits expand to include all possibilities. Values for the composite chert sample STM2-light cross the surface production curve at 3.2 Ma and move beyond 2σ uncertainty limits at 3.6 Ma (grey symbols in Figure 1d). While these values for STM2-light seem less restrictive, it must be noted that this sample is a composite and probably heterogeneous, so components of it would likely yield lower maximum ages than its bulk. Because StW573 was deposited in the Silberberg Grotto as an articulated skeleton13, the individual either died in situ or not long before deposition. This places a maximum age constraint of ca 2.8 Ma on the fossil. The use of 95% confidence limits boosts confidence in this result.

With the recent advances in precision and accuracy of measurements of low concentrations of 10Be and 26Al in quartz1 as well as a firmer basis for calculating their production rates at depth15, the approach taken here holds promise to be useful for reconstructing the geological history of cave systems.

 

Exploring a two-stage burial scenario

The maximum age for the breccia deposit encasing StW573, as determined above, appears to contradict the burial isochron date of Granger et al.1, even if the uncertainty of the latter was underestimated through the use of composite samples. This problem may be resolved by proposing that this breccia deposit contains material that was earlier buried in a chamber at a higher level in the cave system, i.e. it is a secondary deposit. An example of such a secondary deposit in Sterkfontein Cave occurs in the Name Chamber, which contains material from Member 5 (mainly) and Member 4, derived from former higher cave levels now exposed in the open excavation.25,26 As discussed above, the breccia surrounding StW573 contains chert fragments that are derived from various levels in the cave, going back as far as about 5 Ma, indicating that these sediments were reworked. A present-day example in the Cradle of Humankind of such a two-level cave (with a potential death trap) is Gladysvale.27

The deposits of Member 4 and 5, now exposed in the surface excavation pit, accumulated in a cave chamber between ca 2.5 and 1.4 Ma.17 This chamber was de-roofed as a result of erosion, estimated at a rate of ca 5 m/Ma,1 (rendering the land surface about 14 m higher at 2.8 Ma than today), and roof collapse. Figure 2a shows the position of this chamber (approximately delineated by the extent of the current excavation pit) relative to the Silberberg Grotto. Immediately south of the open excavation a large block of dolomite occurs that shows a dip of ca 30° S (Figure 2b), while the strata at Sterkfontein generally dip 25-30° NNW. This block lies above the east end of the Silberberg Grotto (Figure 2a) where the apex of its sediment cone is located.11 It was noted by Robinson25 as 'collapsed dolomite' but received no attention after that. This block is most likely part of a cave roof that collapsed into a void, thus documenting that a cave chamber once existed above the present Silberberg Grotto. The evidence does not allow determination of whether this chamber formed part of the large cave holding Members 4, 5 and 6, or was separate from it; but the second possibility cannot be excluded.

However, whilst a two-staged burial scenario is thus not inconsistent with the geological evidence, it must be assessed whether such a scenario could possibly result in a cosmogenic isotope array resembling an isochron. To do so, we calculated the 10Be and 26Al concentration data of individual samples back in time, as done for the chert samples. In Figure 3, the black symbols (here with 2σ, i.e. 95% confidence, error ellipses) and solid curves show the back-correction to 2.8 Ma for sediment samples and STM2-light, calculated using Equation 2 and applying present day in-situ abundances corresponding to calc2 of Figure 1a. Sample ST7 of Granger et al.1, taken at the surface and indicating an erosion rate of 5-6 m/Ma, is shown for comparison.

Although STM2-light is a composite sample, its average 26Al and 10Be concentrations at 2.8 Ma provide the best estimate of what in-situ accumulated cosmogenic nuclide abundances in such a previous higher level cave system could have been; at 2.8 Ma they plot just below the surface production curve (Figure 3). Long-term accumulation of 26Al and 10Be under shielding and with a low erosion rate (as indicated by ST7) must lead to a lower 26Al/10Be ratio in the sample than at the surface, as a result of the more rapid decay of 26Al compared to 10Be. Shielding could have many physical forms, such as overburden, or a position in a cave with a small opening. Notwithstanding the lack of constraints on actual cave configurations, cosmogenic nuclide accumulation under shielding conditions can be estimated. Various scenarios based on surface production data of Granger et al.1, with material residing in a covered position experiencing a shielding factor that decreases from ca 99% to ca 95% over a period of ca 2 Ma, can yield 26Al and 10Be concentrations similar to those of 'STM2-light at 2.8 Ma' in Figure 3. This is in accord with the assessment of Granger et al.1 that STM2-light constitutes chert debris from a higher level in the cave system. As it occurs thoroughly intermingled with material originally derived from the surface (samples ST1 and ST2), it is reasonable to conclude that the latter could also have resided at this higher level in the cave system.

Before first burial, all surface-derived samples must have had 10Be and 26Al concentrations plotting on the surface production curve. Given the rather large 2σ uncertainties of the back-corrected concentrations at 2.8 Ma for most samples, most of the additional correction times calculated to bring each sample back to the surface production curve also have large uncertainties. This can be illustrated by considering the varying distances from individual error ellipses to the surface production curve. For each surface-derived sample, the minimum correction time needed to intersect the surface production curve is estimated by back-correcting the point on its error ellipse closest to the surface production curve beyond 2.8 Ma, using Equation 2 (grey dot-dash curves and symbols in Figure 3). The production values corresponding to in-situ 10Be and 26Al steady-state concentrations of 'STM2-light at 2.8 Ma' were used for this as a best estimate. The correction times are listed for each sample in Figure 3. In a two-stage burial model, these represent the minimum residence times in the upper chamber before the samples were redeposited into their current position. It can be seen that the minimum residence times vary from 0 to 0.5 Ma (in a similar manner the maximum potential residence times can be calculated, which for all samples are >1 Ma). Note that the heterogeneity of the samples is highlighted by the surface curve intersection for composite sample STM2-dark, which reflects a higher apparent erosion rate (shorter surface residence time) than any of the bulk samples from which it was derived, indicating that the individual samples are mixtures of grain populations with different surface residence times. Interestingly, all apparent minimum pre-burial erosion rates are much lower than the erosion rate measured for today using sample ST7.1 This difference may reflect either lower true erosion rates21,28 or higher chemical erosion factors29 in the past, with more of the dolomite being removed by dissolution at the surface as a consequence of a more humid climate30.

This analysis demonstrates that an apparent isochron age of 3.67±0.16 Ma can be obtained for a secondary deposit which was laid down at a much younger age (2.8 Ma in our example), but which reworked surface-derived material that had accumulated in an upper chamber over a period as long as 1 Ma (2.8-3.8 Ma) or possibly even longer. At the same time, this observation points to a way of testing the two-staged burial hypothesis. The data array of Granger et al.1 is technically an isochron (meaning that any scatter of the data can be the result of analytical uncertainty) because of the rather large error limits of the data on the surface-derived samples. As shown by the chert data, it should now be possible to obtain greater precision for surface-derived samples as well. If an array with greater precision on the data from surface-derived samples (and no composites) still qualifies as an isochron, then the two-staged burial hypothesis is incorrect. If there is significant scatter, it is correct.

 

Discussion

While we have shown that the isochron of Granger et al.1 can be compatible with a two-stage burial scenario, the question remains as to how fossil StW573 could be younger than 2.8 million years old and be embedded in sediments that have been underground for (on average) over 3.5 Ma. In assessing possible models that fulfil the constraints imposed by the cosmogenic isotopes, our interpretation must also be consistent with the broader faunal content of sediments in the Silberberg Grotto, and palaeomagnetic results obtained from the flowstones within them (whether intrusive or stratigraphic).

To reconstruct plausible burial scenarios for StW573, it is important to assess the facies associations of the sediments surrounding the fossil, as described by Bruxelles et al.10 These sediments are composed of surface-derived rubble, sand and mud as well as dolomite and chert fragments of varying sizes that are thoroughly mixed together.1,10 The deposits occur as a series of layers that consist of matrix-supported breccia in which angular chert and dolomite blocks are embedded in a muddy, fine- to coarse-grained sandstone matrix with no internal structure. The clasts display a degree of grading, with variable clast sizes and clast densities across layers. The clastic sequence displays no evidence of suspension flow (e.g. cross-bedding, matrix grading, erosional channels) or standing water (e.g. mud drapes), although shelf stones show that the grotto was filled with water at times after its deposition.11

The deposits around StW573 have been described as the proximal to medial part of a talus cone.11,31 The sedimentary features summarised above are consistent with the deposits being a series of sheet-like debris flows, i.e. mixtures of water, mud, sand and breccia blocks with the internal strength and ability to carry blocks (and bodies) in the matrix32,33, yet producing preferred orientation of clasts34. These debris flows would have moved down the slope of a talus cone from an entry point, presumably within the roof to the eastern corner of the Silberberg Grotto.11 The debris flow deposits display variable composition, reflecting variations in water content, provenance sediment and flow rates, but each layer was probably deposited rapidly, as demonstrated by Unit B2b10, which envelops the fully articulated skeleton of StW573 and preserves complex body configurations of otherwise delicate elements, such as the clasped hand13. The rate of accumulation of the sequence as a whole cannot be determined from the sedimentology, and the isochron, being 'un-sharp', cannot constrain this aspect with any degree of confidence.

The fossil assemblage in the Silberberg Grotto preferentially comprises animals with climbing proclivities (i.e. primates and carnivores), and conspicuously lacks evidence of predator damage.11,17 The taphonomic data indicate that many faunal remains are from individuals that entered the Silberberg Grotto on their own and were then unable to escape35, i.e. the grotto acted as a death trap. In contrast to Member 4, which contains many hominin remains, the only hominin fossils in Member 2 are the remains of StW573, and thus the occurrence that led to a hominid entering the Silberberg Grotto appears to be rare.35 Fossil StW573 lies embedded in Unit B2b and is thought to have been preserved in the death position,10,13 implying that the individual died while being entombed in the debris flow, or shortly before.

When taken together, evidence suggests that StW573 ventured into an upper cave and wandered, or fell, into the Silberberg Grotto where it died and was buried. The reasons for entering the upper chamber could be many (e.g. to search for water, security, shelter), and it is plausible that the individual (like other animals in the Member 2 deposit) was unaware of the presence of the death trap, because they were unfamiliar with the cave system, or the death trap had recently formed (e.g. because part of the roof of the Silberberg Grotto had opened). Live animals falling into a death trap in such a situation can be accompanied, preceded or followed by unconsolidated sediment material that has been lying in the upper cave for hundreds of thousands of years. Erosion and re-deposition of sediment accumulations in the upper chamber would be even more likely if a passageway between the upper chamber and the Silberberg Grotto below had opened suddenly. Such a transient passageway would have disturbed the depositional environment in the upper chamber, allowing erosion, and could have created the death trap. Thus, the age for StW573 could be much younger than the cosmogenic burial age of the sediments that are now associated with the fossil in the Silberberg Grotto.

The assumption that unconsolidated sediment can be preserved in an upper chamber needs further comment. The sediment record of caves in the Cradle of Humankind site shows a significant bias towards fully lithified (i.e. calcified) sediments composed of coarser-grained, more permeable material indurated with calcite cement. In contrast, finer-grained, muddy, and less permeable material is less likely to be strongly indurated and lithified, and therefore less likely to be preserved. Yet some caves, such as the nearby Rising Star Cave36, are known to have contained large volumes of mostly unconsolidated sediment, much of which has been eroded in response to water movement through the cave. Other examples of poorly consolidated sediment accumulations in caves include the upper flowstone-bounded units of Gladysvale with ages of up to 0.5 Ma,27 and parts of the Member 2 deposits in the Silberberg Grotto itself (e.g. unit B2a underneath StW57310). Therefore, it should not come as a surprise that unconsolidated sediment may have existed for hundreds of thousands of years in an upper chamber above the Silberberg Grotto, before being washed down.

How does the burial scenario for StW573 fit with other dating constraints for sediments in the Silberberg Grotto? Palaeomagnetic work done on CaCO3 units and associated siltstone material in the deposit hosting StW573 shows reverse polarity in the units below the fossil, and normal polarity at the level of the fossil and above.37 The CaCO3 units around StW573 (F2 to F4) are not flowstones, but intrusive fracture fillings.7,10,37 According to Bruxelles et al.10, the lower unit (F1) is also intrusive based on the presence of a void immediately above it in which botryoidal CaCO3 has formed, but no real evidence is presented that F1 is not a flowstone. If the F1 unit is a stratigraphic flowstone, then the part of the deposit below StW573 would be placed in a reverse-polarity period. Given a minimum age of ca 2.2 Ma,7,9 based on the age of the intrusive CaCO3 units, and in case of a maximum age of 2.8 Ma as discussed above, the Matuyama C2r.2r Chron (2.58-2.16 Ma) would then be the only candidate37. If all CaCO3 units are intrusive, the palaeomagnetic data have no bearing on the age of the fossil.

The fauna in the deposits of the Silberberg Grotto is largely a subset of that in Member 4 of Sterkfontein and is not highly diagnostic for age.5,17 The fauna includes two taxa of extinct hunting hyena, Chasmaporthetis nitidula and Ch. silberbergi38, and in the former, a similarity in primitive dentition to Ch. australis from the lower Pliocene fossil deposit of Langebaanweg is noted - 'although it is not clear at this stage that the two are conspecific'38. However, both taxa also occur in Member 4 of Sterkfontein17, as well as in Member 1 of the Swartkrans site24. Member 4 has been reliably dated to between ca 2.6 Ma and 2.0 Ma by U-Pb on flowstones7; and at Swartkrans, 26Al/10Be burial ages from Member 1 sediments concur with U-Pb ages of flowstones between ca 2.2 and 1.8 Ma39,40. On the other hand, the extinct colobine monkey Cercopithecoides williamsi, found in the Silberberg Grotto deposits17 as well as in Member 4 and Swartkrans Member 15,24, is noted as not having been reported from reliably dated sites older than 2.5 Ma5. In summary, no contradiction arises from these faunal data in the case of an age <2.8 Ma for the deposit encasing StW 573.

 

Conclusion

Cosmogenic 10Be and 26Al data on chert fragments from a cave deposit can impose constraints on the age of that deposit. In the case of the sediments encasing StW573, such data1 indicate that this deposit was formed no earlier than 2.8 Ma, even if its components had been underground for (variably) longer periods, yielding an isochron age of 3.67±0.16 Ma.1 The younger age is not in conflict with faunal studies5,17,24, palaeomagnetic work37 and U-Pb dating7,9. The apparent contradiction can be resolved by invoking a two-stage burial scenario, which is geologically realistic. This scenario can ultimately yield an isochron-like data array even if primary burial ages differ among samples. It requires (1) an upper cave level environment in which sediment accumulated over time, and (2) events in which the accumulated sediment matter, including chert fragments derived from within the cave, dropped to a deeper level in the form of debris flows and was chaotically mingled. Because the fossil was incorporated as an articulated skeleton, it cannot be older than the deposit, and the individual must, therefore, have fallen into the lower cave either on its own, or incorporated in a debris flow. As the two-stage burial scenario can reconcile the indicated 2.8 Ma maximum age for the fossil with the much older isochron date, it deserves serious consideration.

 

Acknowledgements

We thank Darryl Granger for open discussions, fair comment and constructive criticism on earlier versions of this manuscript. Four anonymous reviewers are thanked for critical comments that led to improvements in the manuscript. J.D.K. thanks the South African National Research Foundation for incentive funding (rated researchers) and P.H.G.M.D. acknowledges funding received from the Australian Research Council (DP140104282) in support of this research.

Authors' contributions

J.D.K. contributed the considerations and calculations relating to cosmogenic nuclide systematics. P.H.G.M.D. contributed the sedi-mentological and taphonomical review and arguments. Both authors wrote their respective parts of the manuscript.

 

References

1. Granger DE, Gibbon RJ, Kuman K, Clarke RJ, Bruxelles L, Caffee MW. New cosmogenic burial ages for Sterkfontein Member 2 Australopithecus and Member 5 Oldowan. Nature. 2015;522:85-88. http://dx.doi.org/10.1038/nature14268        [ Links ]

2. Clarke RJ. First ever discovery of a well-preserved skull and associated skeleton of Australopithecus. S Afr J Sci. 1998;94:460-463.         [ Links ]

3. Partridge TC, Watt IB. The stratigraphy of the Sterkfontein hominid deposit and its relationship to the underground cave system. Palaeont Afr. 1991;28:35-40.         [ Links ]

4. Partridge TC, Shaw J, Heslop D, Clarke RJ. The new hominid skeleton from Sterkfontein, South Africa: Age and preliminary assessment. J Quat Sci. 1999;14:192-198. http://dx.doi.org/10.1002/(SICI)1099-1417(199907)14:4<293::AID-JQS471>3.0.CO;2-X        [ Links ]

5. Berger LR, Lacruz R, De Ruiter DJ. Revised age estimates of Australopithecus-bearing deposits at Sterkfontein, South Africa. Am J Phys Anthropol. 2002;119:192-197. http://dx.doi.org/10.1002/ajpa.10156        [ Links ]

6. Clarke RJ. On the unrealistic 'Revised age estimates' for Sterkfontein. S Afr J Sci. 2002;98:415-418.         [ Links ]

7. Pickering R, Kramers JD. Re-appraisal of the stratigraphy and determinations of new U-Pb dates for the Sterkfontein hominin site, South Africa. J Hum Evol. 2010;59:70-86. http://dx.doi.org/10.1016/j.jhevol.2010.03.014        [ Links ]

8. Partridge TC, Granger DE, Caffee MW, Clarke RJ. Lower Pliocene hominid remains from Sterkfontein. Science. 2003;300:607-612. http://dx.doi.org/10.1126/science.1081651        [ Links ]

9. Walker J, Cliff RA, Latham AG. U-Pb isotopic age of the StW 573 hominid from Sterkfontein, South Africa. Science. 2006;314:1592-1594. http://dx.doi.org/10.1126/science.1132916        [ Links ]

10. Bruxelles L, Clarke RJ, Maire R, Ortega R, Stratford D. Stratigraphic analysis of the Sterkfontein StW 573 Australopithecus skeleton and implications for its age. J Hum Evol. 2014;70:36-48. http://dx.doi.org/10.1016/j.jhevol.2014.02.014        [ Links ]

11. Clarke RJ. A deeper understanding of the stratigraphy of Sterkfontein fossil hominid site. Trans R Soc S Afr. 2006;61:111-120. http://dx.doi.org/10.1080/00359190609519960        [ Links ]

12. Dirks PHGM, Kibii JM, Kuhn BF, Steininger C, Churchill SE, Kramers JD, et al. Geological setting and age of Australopithecus sediba from southern Africa. Science. 2010;328:205-208. http://dx.doi.org/10.1126/science.1184950        [ Links ]

13. Clarke RJ. Newly revealed information on the Sterkfontein Member 2 Australopithecus skeleton. S Afr J Sci. 2002;98:523-526.         [ Links ]

14. Balco G, Rovey CWII. An isochron method for cosmogenic-nuclide dating of buried soils and sediments. Am J Sci. 2008;308:1083-1114. http://dx.doi.org/10.2475/10.2008.02        [ Links ]

15. Balco G, Soreghan GS, Sweet DE, Marra KR, Bierman PR. Cosmogenic-nuclide burial ages for Pleistocene sedimentary fill in Unaweep Canyon, Colorado, USA. Quat Geochronol. 2013;18:149-157. http://dx.doi.org/10.1016/j.quageo.2013.02.002        [ Links ]

16. Erlanger ED, Granger DE, Gibbon RJ. Rock uplift rates in South Africa from isochron burial dating of fluvial and marine terraces. Geology. 2012;40:1019-1022. http://dx.doi.org/10.1130/G33172.1        [ Links ]

17. Reynolds SC, Kibii JM. Sterkfontein at 75: Review of palaeoenvironments, fauna, dating and archaeology from the hominin site of Sterkfontein (Gauteng Province, South Africa). Palaeont Afr. 2011;46:59-88.         [ Links ]

18. Lal D. Cosmic ray labeling of erosion surfaces: In situ nuclide production rates and erosion models. Earth Planet Sci Lett. 1991;104:424-439. http://dx.doi.org/10.1016/0012-821X(91)90220-C        [ Links ]

19. Heisinger B, Lal D, Jull AJT, Kubik P Ivy-Ochs S, Neumaier S, et al. Production of selected cosmogenic radionuclides by muons 1: Fast muons. Earth Planet Sci Lett. 2002;200:345-355. http://dx.doi.org/10.1016/S0012-821X(02)00640-4        [ Links ]

20. Heisinger B, Lal D, Jull AJT, Kubik P Ivy-Ochs S, Knie K, et al. Production of selected radionuclides by muons 2: Capture of negative muons. Earth Planet Sci Lett. 2002;200:357-369. http://dx.doi.org/10.1016/S0012-821X(02)00641-6        [ Links ]

21. Dirks PHGM, Placzek C, Fink D, Dosseto A, Roberts EM. Using 10Be dating to reconstruct the evolving landscape in the Cradle of Humankind, South Africa. J Hum Evol. 2016;96:19-34. http://dx.doi.org/10.1016/j.jhevol.2016.03.002        [ Links ]

22. Desilets D, Zreda M. On scaling cosmogenic nuclide production rates for altitude and latitude using cosmic-ray measurements. Earth Planet Sci Lett. 2001;193:213-225. http://dx.doi.org/10.1016/S0012-821X(01)00477-0        [ Links ]

23. Partridge TC. Geomorphological dating of cave opening at Makapansgat, Sterkfontein, Swartkrans and Taung. Nature. 1973;246:75-79. http://dx.doi.org/10.1038/246075a0        [ Links ]

24. McKee JK. Further chronological seriations of southern African Pliocene and Pleistocene mammalian fauna assemblages. Palaeont Afr. 1995;32:11-16.         [ Links ]

25. Robinson JT. Australopithecines and artefacts at Sterkfontein. Part One: Sterkfontein stratigraphy and the significance of the extension site. S Afr Archaeol Bull. 1962;17:87-107. http://dx.doi.org/10.2307/3886942        [ Links ]

26. Val A, Stratford DJ. The macrovertebrate fossil assemblage from the Name Chamber, Sterkfontein: Taxonomy, taphonomy and implications for site formation processes. Palaeont Afr. 2015;50:1-17.         [ Links ]

27. Pickering R, Hancock PJ, Lee-Thorp JA, Grün R, Mortimer GE, McCulloch M, et al. Stratigraphy, U-Th chronology, and paleoenvironments at Gladysvale Cave: Insights into the climatic control of South African hominin-bearing cave deposits. J Hum Evol. 2007;53:602-619. http://dx.doi.org/10.1016/j.jhevol.2007.02.005        [ Links ]

28. Dirks PHGM, Berger LR. Hominin-bearing caves and landscape dynamics in the Cradle of Humankind, South Africa. J Afr Earth Sci. 2013;78:109-131. http://dx.doi.org/10.1016/j.jafrearsci.2012.09.012        [ Links ]

29. Riebe CS, Granger DE. Quantifying effects of deep and near-surface chemical erosion on cosmogenic nuclides in soils, saprolite, and sediment. Earth Surf Process Landforms. 2013;38:523-533. http://dx.doi.org/10.1002/esp.3339        [ Links ]

30. Bobe R, Behrensmeyer AK. The expansion of grassland ecosystems in Africa in relation to mammalian evolution and the origin of the genus Homo. Palaeogeogr Palaeocl. 2004;207:399-420. http://dx.doi.org/10.1016/j.palaeo.2003.09.033        [ Links ]

31. Stratford D, Grab S, Pickering TR. The stratigraphy and formation history of fossil- and artefact-bearing sediments in the Milner Hall, Sterkfontein Cave, South Africa: New interpretations and implications for palaeoanthropology and archaeology. J Afr Earth Sci. 2014;96:155-167. http://dx.doi.org/10.1016/j.jafrearsci.2014.04.002        [ Links ]

32. Coussot P Meunier M. Recognition, classification and mechanical description of debris flows. Earth Sci Rev. 1996;40:209-227. http://dx.doi.org/10.1016/0012-8252(95)00065-8        [ Links ]

33. Iverson RM. The physics of debris flows. Rev Geophys. 1997;35:245-296. http://dx.doi.org/10.1029/97RG00426        [ Links ]

34. Major JJ. Pebble orientation on large, experimental debris-flow deposits. Sed Geol. 1998;117:151-164. http://dx.doi.org/10.1016/S0037-0738(98)00014-1        [ Links ]

35. Pickering TR, Clarke RJ, Heaton JL. The context of Stw573, an early hominid skull and skeleton from Sterkfontein Member 2: Taphonomy and paleoenvironment. J Hum Evol. 2004;46:279-297. http://dx.doi.org/10.1016/j.jhevol.2003.12.001        [ Links ]

36. Dirks PHGM, Berger LR, Roberts EM, Kramers JD, Hawks J, Randolph- Quinney PS, et al. Geological and taphonomic context for the new hominin species Homo naledi from the Dinaledi Chamber, South Africa. eLIFE. 2015;4, Art. e09561, 37 pages. http://dx.doi.org/10.7554/eLife.09561        [ Links ]

37. Herries AIR, Shaw J. Palaeomagnetic analysis of the Sterkfontein palaeocave deposits: Implications for the age of the hominin fossils and stone tool industries. J Hum Evol. 2011; 60:523-539. http://dx.doi.org/10.1016/j.jhevol.2010.09.001        [ Links ]

38. Turner A. Further remains of carnivore (Mammalia) from the Sterkfontein hominid site. Palaeont Afr. 1997;34:115-126.         [ Links ]

39. Gibbon RJ, Pickering TR, Sutton MB, Heaton JL, Kuman K, Clarke RJ, et al. Cosmogenic nuclide burial dating of hominin-bearing Pleistocene cave deposits at Swartkrans, South Africa. Quat Geochron. 2014;24:10-15. http://dx.doi.org/10.1016/j.quageo.2014.07.004        [ Links ]

40. Pickering R, Kramers JD, Hancox PJ, De Ruiter DJ, Woodhead JD. Contemporary flowstone development links early hominin bearing cave deposits in South Africa. Earth Planet Sci Lett. 2011;306:23-32. http://dx.doi.org/10.1016/j.epsl.2011.03.019        [ Links ]

41. Martini JEJ, Wipplinger PE, Moen HFG, Keyser A. Contribution to the speleology of Sterkfontein Cave, Gauteng Province, South Africa. Int J Speleo. 2003;32:43-69. http://dx.doi.org/10.5038/1827-806X.32.L4        [ Links ]

 

 

Correspondence:
Jan Kramers
jkramers@uj.ac.za

Received: 18 Mar. 2016
Revised: 16 July 2016
Accepted: 21 Sep. 2016

 

 

FUNDING: National Research Foundation (South Africa); Australian Research Council

Creative Commons License All the contents of this journal, except where otherwise noted, is licensed under a Creative Commons Attribution License