Friday, 1 November 2013

Frayer and RussellI (1987) and Lozano et al. (2013): I Have a Bone Skeletal Element Tooth to Pick with You OR When Does A Shallow Groove Become A Myth's Shallow Grave?


The other day I mentioned a recent PLoS ONE article with the unlikely title of "Toothpicking and Periodontal Disease in a Neanderthal Specimen from Cova Foradà Site (Valencia, Spain)," by M. Lozano, M.E. Subirà, J. Aparicio, C. Lorenzo, and  G. Gómez-Merino. PLoS ONE 8(10): e76852. doi:10.1371/journal.pone.0076852.

Minute scratches on two Neanderthal teeth. Claimed to be toothpick attrition. Or, in other words, Middle Palaeolithic (MP) oral hygiene or self-dentistry.

It's yet another item on the pendulous list of MP myths that deserves to be aired out and disposed of. In the process I'll point out the epistemological shortcomings of their arguments. Won't that be "just dandy!"*

OK. So, seeing the Lozano et al. paper's title reminded me of the time, many years ago . . . [Just how many years ago was it, Rob?]

 . . . twenty-five years ago, in my first semester as a doctoral student at Cal, I came across a then-recent paper in the American Journal of Physical Anthropology (AJPA). That paper has stuck in my craw [rather like a toothpick!] ever since.
Frayer, D.W. and M.D. Russell, "Artificial grooves on the Krapina neanderthal teeth." AJPA 74:393–405, 1987.
I could never put my finger on the reason I thought the piece was flawed. It was, if not the first, then it was one of the earliest claims for toothpick use in the MP. So, for you, fearless reader, to be equipped to be critical of Lozano et al.'s (2013) claims, I think it may be worth your while to spend some time with me, Frayer and Russell (1987), and some mid-twentieth-century dental anthropologists.

According to Frayer and Russell, 14 teeth from the site of Krapina bear unnatural grooves. They argue that the grooves must have resulted from habitually thrusting a small, cylindrical, wooden object—a toothpick, in other words—between two teeth such that attrition of the dental tissue left a "troughlike" appearance.
Fig. 1. Mesial surface of tooth 82, a lower left M1, showing a well-developed artificial groove. The scale is approximately three times natural size. Arrows indicate cementum buildup along lower border (Frayer and Russell 1987).**
It's a pity that to illustrate the grooves, Frayer and Russell published next-to-useless photographs. The closest they get to a decent illustration is reproduced above. And it's almost completely unhelpful. The pointy ends of two arrows in the photo below indicate a [sub-linear, sub-horizontal] area of cementum buildup immediately beneath a "troughlike" groove at the cementoenamel junction. If I'm not mistaken, one can see what must be the upper margin of the "groove" on this lower left M1's mesial surface. In this view it appears as a faint, intermittent, sub-linear, sub-horizontal, darkening on the tooth's surface. The pairs of blue and yellow arrows point out the visible portions of the groove's upper margin. The red arrows point to the lingual and buccal ends of what I take to be the most prominent points on the abnormal cementum growth [which is the line described by the illuminated upper portion and the lower portion, which is in shadow].

MAKE A NOTE OF THIS. In the above illustration, that the abnormal cementum growth isn't parallel with the "artificial groove." This suggests to me that the abnormal cementum growth is not in any way related to the much more horizontal upper margin of the "groove."

To argue that the Krapina dental lesions [i.e. premortem tissue damage] were the result of Neanderthal dentistry, in their paper Frayer and Russell cite earlier scholarship on interproximal grooving, which North American anthropologists have documented and which some have provisionally identified as being due to toothpick use—Ubelaker et al. (1969), for example.
D.H. Ubelaker, T.W. Phenice, AND W.M. Bass, “Artificial Interproximal Grooving of the Teeth in American Indians.” AJPA 30:145—150, 1969.
Ubelaker et al.'s photo illustration is far superior to that of Frayer and Russell, and clearly points up an enormous difference between the grooves observed on archaeological North Americans and those on the MP Krapina teeth. Look at the beautifully crisp reproductions in the triptych below. See? In the left-side panel an arrow points to the rightmost tooth, sporting*** the rather obvious cylindrical attrition that took place during the individual's life. Two more examples appear in the middle and right-side panels.
From Ubelaker et al. (1969).
The following are two versions of a beautiful, clear pic from an Arikara (Aboriginal Native Americans) thanatocoenose. The photo immediately below illustrates the cavernous groove on the left tooth. 

From H.E. Berryman, D.W. Owsley, and A.M. Henderson, “Non-carious Interproximal Grooves in Arikara Indian Dentitions.” AJPA 50:209-212, 1979.
In the version below, I've highlighted in red the size of the cylinder that would have been needed to create [at least] the most recent extent of the groove. Roughly in dashed yellow the original, interproximal surface profile of the affected tooth. Although neither Ubelaker et al. nor Berryman et al. give us a scale, I think it's possible to estimate that a groove like this is at least an order of magnitude larger than the lesions reported by Frayer and Russell (1987). 

From H.E. Berryman, D.W. Owsley, and A.M. Henderson, “Non-carious Interproximal Grooves in Arikara Indian Dentitions.” AJPA 50:209-212, 1979.
The latest crew trying to argue for MP dentistry—Lozano et al. (2013)—pop off the high-res beauties you see below. Nevertheless, in the pair immediately below it's not really clear what's being pointed to, much as was the case with Frayer and Russel's images of the MP lesions. But this image is not intended to document the lesions: it's meant to provide the anatomical context for the close-ups reproduced further down.

From Lozano et al. (2013)
Lozano et al. do provide clear, documentary, images of the Cova Foradà lesions. In the photomicrographs and SEM images reproduced below, the grooving is very visible. However, when you consider the scale,  you notice that A's groove is about 2 mm buccal--lingual and about 1 mm inferior--superior.

In B, the SEM of A's groove, you see, even more clearly, the lesion's topography. But you also see something else. The "groove" isn't cylindrical, it's irregular, and it has a narrowing approximately half-way. This sort of "waisting" is what you might expect if the cause of the attrition was a cylindrical object rubbed against the tissue repeatedly, but never at quite the same angle. Likewise C and D, below. These grooves on the Neanderthal teeth from Cova Foradà are tiny compared with the veritable tunnels seen in the modern human examples from Berryman et al. and Ubelaker et al., above.

From Lozano et al. (2013).
I'm intrigued by the SEM photomicrographs, B and D, above. One of the criteria that all observers agree on is that of minute striae on the attritional surface that are presumed to evince the presence of grit when the cylindrical object was being moved back and forth within the attritional groove. A quasi-cylindrical morphology would be only other observation that might persuade one to think the dental attrition had been caused by a cylindrical object.

So, now look at the SEM of B enlarged to show the maximum possible detail without the image becoming pixelated. The upper and lower margins of the 'groove' are irregular and they don't mirror one another, which you might expect if an elongated cylindrical object were the cause. Next, I see very faint sub-linear lines within the 'groove' that are more or less horizontal on the right side and which sweep upward to the left of the narrowing that occurs about a third of the way from the left edge of the image. Finally, look at the upper 'groove' margin at the the extreme right of the photo. It bends radically upward. The lower margin, by contrast, is more or less linear, and angles downward from the narrowest part of the 'groove' at about 30° below the horizontal. If, indeed the apparent striae are in fact surface features of the 'groove' notice how on the right half they are sub-horizontal and appear to 'bump' into the lower and upper margins. If those were truly striae associated with the action of a cylindrical object mixed with grit it's difficult to see how so much attrition could have occurred in the places where it appears that the striae dead-end at the 'groove' margins.

So, the impression is that this attritional feature is highly irregular, and it's difficult to imagine how it might have been created by a cylindrical wood or bone probe.

The attritional feature in D is also poor support for the hypothesis that a toothpick created it. Look closely at the right part of the photo, beginning just above the letter D. I don't see evidence of grinding or polishing. I do see laminar features that appear to have been broken irregularly, together with some evidence of striae in the upper, central portion of the feature. Again, I have trouble seeing this as the result of tooth picking.

Okay. So, there's inexplicable and dissimilar morphology in the two 'grooves' from Cova Foradà. What do we do about it? More comparison, I think. Let's have a closer look at the 14 specimens from Krapina.

As I pointed out earlier, no good photographs appear in Frayer and Russell. However, they [sort of] make up for for the illustrative shortcomings. They describe each tooth in proper, physical-anthropological, dental-anatomical detail. Their descriptions often include measurements of the overall length, depth, and vertical extent of each "troughlike" groove [although, maddeningly, not consistently].

So, in an effort to use the Frayer and Russell data for comparison with the Lozano et al. observations, I dug in. With the following photo of the entire Krapina 'grooved' sample as a reference, I read and recorded the authors' verbal descriptions, and made a little table of the observations. My table appears following the dental array that you see below.
Fig. 2. Artificially grooved teeth in the Krapina dental sample. A: Mesial of 32. B: Distal of 32. C: Mesial of 82. D: Distal of 3. E: Mesial of 5. F: Mesial of 60. G: Labial of 60. H: Distal of 60. I: Distal of 35. J: Distal of 42. K: Distal of40. L: Mesial of 167. M: Mesial of 172. N: Distal of 172. O: Mesial of 162. P: Mesial of 175. Q: Mesial of 179. R: Mesial of ramus 63. All teeth approximately natural size. Arrows indicate location of the artificial groove (Frayer and Russell 1987).*
Observations published in Frayer and Russell (1987). 
Frayer and Russell's observations are virtually identical to Lozano et al.'s, in that both groups recorded significantly smaller, and morphologically less-regular lesions than those we've seen in modern people. As you can see, the Krapina data set is also highly variable on every measure—morphology, placement, length, width, and depth. Indeed, it's so variable as to be almost completely lacking in internal consistency.

These data don't lend themselves to sorting. So, I sorted them by which tooth surface was affected. The teeth in the red box are the distal surfaces; the blue are lingual surfaces; the green encompasses the mesial surfaces. Of the 18 lesions described on 14 teeth, 11 are described straightforwardly as "grooves," or "channels." Some of the other lesions are described ambiguosly and enigmatically, compelling me to question the liklihood that a toothpick was responsible.

Notable anomalous observations, which I believe would rule out the toothpick hypothesis.
Tooth 162's mesial lesion is a "triangular polished surface."
Tooth 3's lesion is described both as a "shallow . . . groove," and as a "triangular" "polished area."
Tooth 167's lesion is described as a "wedge-shaped" "polished area." 
Tooth 5's lesion is described as "oval" and "polished." 
Tooth 172's lesion starts on the mesial surface and continues onto the lingual surface.
Tooth 60's "groove" extends from the mesial surface, wraps around the lingual surface and ends on the distal surface. 
Tooth 35's lesion is described as a "rectangular" "concavity."
Thus, of the 14 Krapina teeth described in Frayer and Russell, the seven listed above can be dismissed as evidence of MP dentistry. That leaves the seven good candidates—82, 32, 63, 175, 179, 40, and 42—all but one are described as less than 0.5 mm deep. Average lesion length is 5.2 mm (s.d. 1.7 mm) and average inferior--superior dimension is 1.5 mm (s.d. 0.7 mm).

I don't know about you. But, to leave such tiny lesions on these teeth, I'm thinking that the toothpick would have to be, on average,  less than 0.5 mm in diameter. As it turns out, today's flat toothpicks are manufactured to 1/64 of an inch thick (0.39 mm). That just tells us that the heights of Krapina's lesions are in the same neighourhood as today's toothpicks. It's unlikely that the Neanderthals were availing themselves of manufactured wooden toothpicks! However, the observations made on the seven 'good' 'grooves' means we can't rule out teeny-tiny, ad hoc slivers of wood or bone used as dental probes.

We're not yet able to quash the narrative that the Krapina attritional features were made by dental probes/toothpicks. What to do? What to do? What can we use to compare with the Krapina data? The North American studies referred to above recorded no measurements of traditional Native American interproximal grooving. So, we're left with just a hunch that the modern examples and the MP 'grooves' have fundamentally different etiologies. 

You know I wouldn't leave it there. 

After further research I was pleasantly surprised to find another Frayer paper, published in 1991. In it, he measures some teeth from the Central European Upper Palaeolithic site of Neuessing. All told, there are 10 interproximal lesions with measurements of the inferior--superior height and the horizontal extent. Alas, as Frayer notes in the text, he failed to record depth. Never mind. We'll make do. The eight teeth are all from the skull of a single individual. 

Data drawn from D.W. Frayer, “On the Etiology of lnterproximal Grooves.” AJPA 85:299—304, 1991.
Granted this is a pitifully small dataset. Moreover, the specimens all came from one adult male, which might explain the regularities noted hereafter. Despite the small sample, Frayer bases his entire argument on these same 10 data points. As before with the Krapina teeth, the author wrote descriptions, and didn't provide the observations in tabular form. I've supplied one—above. The Neuessing lesions are remarkably consistent in their two-dimensions: length averaged 5.9 mm (s.d. 1.5 mm), and superior--inferior height averaged 2.5 mm (s.d. 0.4 mm). 

OK. The Krapina and Neuessing lesion's average length area almost identical, although Krapina's is much more variable, at 5.2 mm (s.d. 1.7 mm). Don't let that fool you into thinking that there must be something to what Frayer and Russell are arguing for the Krapina teeth. That's because the length is naturally constrained by the lingual--buccal dimension, and by the curvature of the tooth surface on which the lesion occurs. Thus, I would expect almost any repeated horizontal lesion that you could possibly find on a tooth would have more or less the same length. It's the height that matters. The height provides a proxy for the cross-sectional diameter of the hypothesized toothpick. What do the Krapina and Neuessing 'groove' heights tell us? 

When the heights of the Krapina and Neuessing lesions are compared, the picture is very different. Based on the seven Krapina teeth that we didn't rule out, above, the groove heights average 1.8 mm (s.d. 1.2 mm). It's a far more variable population than that of the Neuessing individual, which has an average height of 2.5 mm (s.d. 0.4 mm). Thus, the Neuessing heights are significantly larger and less variable. This, to me, is better evidence of the essential differences between the MP and the modern samples.

The missing depth observations for the Neuessing grooves is a tragic shortcoming of Frayer's 1991 paper. I'd bet you dollars to donuts that the Neuessing grooves would be uniformly deeper than the MP ones, effectively cinching victory for the Subversive Archaeologist. 

Hey. Do you know someone at the Anthropologische Staatssammlung, München? That's where the Neuessing individual was examined for Frayer's 1991 article. How hard could it be to snatch 10 observations? Let me know!

You know. We might never be able to rule out the toothpick hypothesis. The best we can do is point out the qualitative and quantitative differences between the modern and the MP samples. 


I'm thinkin' that this is another one of those places where the MP archaeologist observe something that faintly resembles the results of behaviours or activities known to exist among modern people. That leads, ipso facto-style to the claim that the MP artifacts were created by identical processes.

On the one hand, we have the modern human examples of interproximal cementum attrition caused, some would argue, by dental probes/toothpicks. On the other hand, we have lesions on MP teeth that in some respects resembles those known to occur in relatively recent modern human populations. As with all archaeological knowledge, when musing upon the toothpick we are employing inductive analogical reasoning. 

Even just to make the claim that dental probes were the cause of premortem interproximal lesions on modern human teeth, you need to be able to show that you can unequivocally point to relevant similarities between the result of passing a wooden or bone probe repeatedly between proximate teeth and the lesion you're trying to explain. This would be a relational analogy. Relational analogies are the underpinning of jillions of well-accepted knowledge claims and theories of past processes in geology, astronomy, and, of course, archaeology and palaeoanthropology. To solve the riddle of the toothpick we need observations gained from experimental dental attrition studies, to see if there are any groove morphologies or measurements that would allow us to argue persuasively for, or against, toothpickery. Need input!

As I read the literature, the only author that states unequivocally that he can explain interproximal lesions as being the result of wooden probes is D.W. Frayer, himself. Ubelaker et al. didn't go so far. Berryman et al. didn't go so far. Neither did P.D. Schulz, "Task Activity and Anterior Tooth Grooving in Prehistoric California Indians." AJPA 46:87–92, 1977. What are the odds that Frayer is correct? Hmmm.

You and I, Dear Reader, are thus left to wonder as to the explanation for the similarities and differences in the size and shape of the two sets of observations—that of the MP and that of the modern humans. I have managed to tease out some [I think] important and statistically significant differences between Frayer's Krapina and his Neuessing observations. And don't let's forget that we undermined the Lozano et al. toothpick claim in the process.

So, the next time you see the whole dentistry business brought up when someone is extolling the Neanderthal brain, you'll [I hope] smile to yourself, and bide your time. Or you'll jump up and down and say those claims aren't founded in actualistic/experimental research. And some day you might even see these matters settled. You might even settle them yourself. I would be very pleased and happy for you.

So. Get to it! 
* noun
something that is very good or impressive
"That was a dandy of a game."
**The keen-eyed reader will notice the text that I have lined through and the absence of any sort of useful scale. This illustration is a perfect object lesson in how not to present observations in a publication.
*** verb
wear or display (a distinctive or noticeable item).
"he was sporting a huge handlebar mustache"


SA announces new posts on the Subversive Archaeologist's facebook page (mirrored on Rob Gargett's news feed), on Robert H. Gargett's page, Rob Gargett's twitter account, and his Google+ page. A few of you have already signed up to receive email when I post. Others have subscribed to the blog's RSS feeds. You can also become a 'member' of the blog through Google Friend Connect. Thank you for your continued patronage. You're the reason I do this.

No comments:

Post a Comment

Thanks for visiting!

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