Sunday 11 March 2012

Why No 14C Dates for Blombos Cave's MSA? Not Lobbing Aspersions. Just Sayin' ...

If I were a chess player I probably wouldn't be trying this gambit. But I'm not. So, bear with me.
Worked bone, stone and ochre from Blombos Cave (Wikipedia)
I need help with a question that's been nagging me for years. I'm curious to know if any of the bone artifacts from the Middle Stone Age (MSA) layers at Blombos Cave have been directly dated using AMS 14C. From my reading it appears not. In fact, it looks as if 14C was abandoned in favour of luminescence techniques once they had excavated deeper than those layers identified as Late Stone Age, the earliest of which were dated to give-or-take 39 ka BP. Everything below that is deemed to be MSA, and organic materials such as charcoal and bone were passed over in favour of grains of sand or burned flints in those strata. 
     Remember that, for most of us, MSA is synonymous with the Middle Palaeolithic in the rest of the world, and, for better or worse, it's exclusively associated with the Neanderthals and their ilk, for which the jury is still out as to their cognitive equivalency with us modern types. Finding what are clearly modern human artifacts at Blombas Cave and elsewhere on the order of 30 to 50 ka earlier than anywhere else in the world has stunned and amazed scientists from Barrow to Burbank. But it's never sat well with me. 
Location of Blombos Cave, South Africa (Credit)
     You and I know that 14C is perfectly capable of accurately gauging the age of organic materials until at least 50 ka, notwithstanding the need for calibration that corrects for environmental  and other effects. Why then do we have only luminescence age determinations for the sediments in which the Blombos Cave bone artifacts and charcoal were deposited? Henshilwood et al. (2002) provide some insight, although I'm not certain they realize that by doing so they've left themselves exposed to (at a minimum) questions about their decision.
     That (to me) curious decision is explained in what amounts to a throwaway comment, which I will quote here

In radiocarbon terms, the MSA at BBC is of infinite age (Vogel, personal communication).
The MSA levels are being dated using luminescence techniques: single-grain laser luminescence (SGLL), single aliquot optically stimulated luminescence (OSL and IRSL), multiple aliquot OSL on sediments and also TL of burnt lithics and electron spin resonance (ESR) of teeth (Henshilwood et al. 2002:638). 
Being the skeptical type, I was intrigued by what seemed to me to be such a weak citation as to the inefficacy of 14C beyond 39 ka--the date of the oldest LSA at Blombos. Just a 'Vogel pers. comm.' No reams of empirical evidence. No other justification. So I endeavoured to discover by what authority Vogel had made such a pronouncement.
     No doubt some among you will think it naïve of me, or worse, that I'm poorly informed and ill-prepared to be taken seriously by the palaeoanthropological establishment. Nevertheless, I had no prior knowledge of Vogel's reputation in the radiocarbon world and in the South African archaeological community. His work includes a 1997 paper in Radiocarbon in which he attempted to calibrate 14C dates with U/Th dates within the same stalagmite from a South African cave. That work was superseded a few years later by the more widely cited Fairbanks et al. (2005), who honed the 14C calibration curve back to 50 ka using pristine corals from around the globe. Their findings are that 14C underestimates calendar years such that 45 RCYBP works out to 48,934 calendar years (give or take 500). By this means the calendar date of 39,200 BP from the lowest LSA level at Blombos would have been produced by a radiocarbon age of about 34 ka RCYBP (try it yourself by clicking here to go to Fairbanks's calibration calculator).
     It would seem, therefore, that despite Vogel's pronouncement, cited in Henshilwood et al. (2002), there is no physical limitation on dating organic material that is older than 39 ka (i.e 34,000 RCYBP), as long as its age doesn't exceed 45 RCYBP. That would, theoretically, allow the excavator of Blombos Cave to extend use of 14C for at least a further 11,000 RCY beyond the earliest LSA from Blombos Cave. Surely some of the MSA materials could be presumed to date from this 11,000-year window.
     Directly dating the bone and charcoal from the upper MSA strata may prove nothing. However, knowing that there's no theoretical limit on the use of 14C for that 11,000 year period leaves wide open the question as to why Chris Henshilwood hasn't attempted to date, directly, some of his bone artifacts or charcoal from the MSA layers. It's just possible that they would yield dates far younger than those produced by luminescence techniques (about which I've had a certain amount to say in previous efforts here at the Subversive Archaeologist).
From Henshilwood et al. (2002)
     After all, the MSA levels in Blombos (at least those illustrated in Henshilwood et al. [2002], shown above) appear to be quite shallow--the sort of depth that you could imagine accumulating in far less than the 35 ka that the luminescence dates would have you believe it took for them to accumulate--and it's within reason to suspect that they could easily have been deposited in the 11 RCY or so before 39,200 BP.
     And so. For what it's worth, I'm issuing a challenge from this lofty perch of mine. Chris, try dating some of your MSA bone using good, old-fashioned, AMS 14C and see what you get. A fair few of us are curious to know the result. 


  1. Hi Rob, interesting piece. Well outside the date range that i have experience of. However when i first used OSL I was lucky enough to be able to check the results against C14 dates, and found a very strong correlation. I like the method myself, but as you say, there's no reason not to check the dates you are on about, and it would also be interesting to see what method produced the tightest error ranges and highest degree of confidence. As i say interesting... watch this space?

  2. Thanks, Stuart. Were your OSL dates from an open-air site? I think that may be the difference, if there is one, between close correlations like you experienced and those that might be found in caves and rockshelters (about which I've bored you and everyone else before now). Definitely, watch this space!

  3. There are two main issues in play with dating bone of that age, collagen preservation and the measurement itself. I'll start with the latter. Calibration against a curve isn't the issue. What counts as 'infinite age' depends on the AMS instrument doing the measurement and to what extent they can reduce it's background activity to zero. In 2002, it's likely that most labs would have been running backgrounds maybe as high as 40-45ka BP. (Backgrounds or blanks are standards known to be infinitely old, usually of the same material as the unknowns you're running, and pretreated the same way; for wood or charcoal it might be Queets Wood or USGS coal, for bone it might be a Pleistocene mammal, for carbonates a geologic calcite, etc). Then outside of the how the machine is tuned, the material and the pre-treatment dictate how clean of contaminants the blanks get. That background 14C is subtracted from the unknowns, but it also affects the precision of the measurement. What gets tricky then is when you have poor backgrounds. It becomes difficult to distinguish a radiocarbon activity of say 45ka BP from the background. So it might be of infinite age, even though you can go and calibrate 45000+/-750 BP on Fairbank's curve. Nowadays there are AMS devices at a few labs that routinely run at backgrounds lower than 50ka ... though getting that on bone is more difficult than wood, which in turn is more difficult than calcite. Bone blanks I've processed in the last couple of years seem to run between 46 and 51ka, averaging around 48ka. So maybe, yes, the bone artifacts from Blombos could be dated with newer devices, say at UC Irvine, ETH Zurich, or Oxford. Possibly Belfast, I forget what they're running.

    But there has to be collagen in the bones in the first place. I don't know what the organic preservation is like there, but assuming it's a limestone-ish cave, that tends to be pretty poor for collagen preservation (pH too high). They look well preserved in the photo, but bones that are well preserved for the zooarchaeologist are often not well preserved for dating. If we assume we could get 1 gram of bone, and it's only 0.1% collagen after doing ultrafiltration or similar (which is low) you're running a sample that's 1mg. Combusted and assuming 40% C for collagen, you're down to 0.4 mg C. That's not super small, but it could start to eat in to the precision, since that's size dependent. If you had worse preservation, or a smaller sample to start with, you will definitely have lower precision. And that brings you back to the background. When the precision is poor, even a clean blank may not allow distinguishing infinite age from an actual 47ka age. And with a small unknown you need a comparably small blank and the background are usually higher, maybe 43-45ka ... and misery ensues.

    But none of that is to say it couldn't or shouldn't be tried, because it's obviously an important question, and having multiple lines of dating is key. The advantage of dating the bone tools instead of the sediments is that they are evidence of hominin behavior (assuming they are tools ... but that critique is your department, pal). At any rate, dating bone is hard, and older bone is harder yet.

  4. Hey, Spawn.
    Can I bring your extremely helpful background on background into the foreground and put it up as a guest post? I think it's really important to have such clarity in these matters, and I'm not certain everybody in the 'community' is aware of the niceties (should I have been?). On another note, as far as I'm aware there's plenty of charcoal. I suppose it's possible that similar chemical drawbacks ensue, but maybe not?

  5. Rob you are quite correct, all the OSL stuff I have done has been from open air sites. I see what you mean about caves, but mind you, my sites were all in Ireland, hardly famed for it's high level of sunshine ;)
    I shall leave this one to the experts I think!

  6. Thanks for clarifying. As you can probably guess, I don't dispute the physics of OSL, any more than I do that of 14C. It's always been the case that the error comes in the application.

  7. Have they done any detailed analysis of the stratigraphy beyond drawing the section? A Harris matrix for example. It might help clarify the deposition sequence and which layer is to be dated.


Thanks for visiting!

Note: only a member of this blog may post a comment.