Formation ages of 14C-dated human teeth

[Erik Marsh]

Hi everyone – 

I'm looking at radiocarbon dates run on teeth, and I'd like to incorporate the age of tooth formation into a Bayesian model, following the method in Millard et al. 2020 (which uses tooth ages from AlQahtani et al. 2010). I've found a much more detailed chart that tracks four stages of tooth development (White et al. 2012; image below): 1. beginning of mineralization, 2. crown completion, 3. eruption, and 4. root completion.

The tables in AlQahtani et al. 2010 give ages for clinical emergence and alveolar eruption. The ages are a little different, and I'm not sure how they line up with the four stages of development in White et al. Anyway, since I'm no dental anthropologist, the important question here is: 

For a radiocarbon-dated tooth, what age range should I be looking at?

It seems like a uniform span is best here, but Millard et al. opt for an average with a standard deviation.

In the published dates I'm looking at, I don't have details, for example, if they specifically dated the enamel or the roots.  Is one part of the tooth more standard when it comes to taking? I have to imagine most people send in an entire tooth to the lab, since they're hard to cut.

Sometimes all I have is "molar" or "pre-molar" – what I want is a reliable range that covers multiple teeth when necessary.

Any ideas on the best age range for dated teeth? Thanks!

Erik

[Helene Agerskov Rose]

Hi Eric

Teeth are great, but also complicated and I always recommend Simon Hillson's 'Dental Anthropology' (1996) as the best starting point. Working with legacy data I would try contacting the lab to hear what section of the tooth was sampled, otherwise you would have to use the entire growth span. Bearing in mind that with increasing age, teeth will form smaller amounts of secondary dentine in the pulp chamber and root canals. This is not an issue with sub/adult individuals. 

Tooth eruption is more variable than formation, but still with considerable variation around a mean age. In my opinion the approach of Millard et al. incorporates that variation well. Along the same lines, we have used the calendar offset from tooth formation (based on AlQahtani et al. 2010) to age-at-death as prior in a chronological model of a multiple grave (Rose et al. 2018). I am curious as to why you think a uniform span would be better, considering that the distributions in the plot from White et al. are shaped as battleships?

Best regards,

Helene

Rose, H. A., Meadows, J., & Bjerregaard, M. (2018). High-Resolution Dating of a Medieval Multiple Grave. Radiocarbon, 60(5), 1547-1559.

[MILLARD, ANDREW R.]

Hi Eric,

 

I quite like the Millard et al 2020 approach 😉. Putting hard boundaries on the range is difficult for multiple reasons:

1. The ages from AlQahtani et al (or  any other scheme) are averages for when a tooth reaches a specific stage and our individuals may vary. From older work by Moorrees et al, the standard deviation of age-of-formation is about 10% of age-since-conception, and the variation is log-normally distributed in age-since-conception.
2. We sampled a slice of dentine immediately below the crown. This sampling approach cuts across the growth lines, and incorporates dentine that grew over a longer period than the dates for stages would indicate. The different ages are not contributing equally to the sample.
3. The age used for the dating model is the average age of formation of the dentine sampled, not the range of ages of formation.

 

The tables in AlQahtani et al use 13 formation stages rather than the four in the White et al diagram, of which only three are formation stages and one is eruption. They are also based on a much larger, modern dataset rather than material that is 50 years old. I’d go with AlQahtani every time.

 

For dates on a tooth the age of formation of the dentine is important. You don’t care whether the tooth was erupted or not. Dating the carbonate in enamel is unreliable and rarely attempted. If the lab used a collagen extraction they were dating dentine.

 

If whole teeth were dated then I would take the midpoint of the ages from stages R¼ and R½ as a rough midpoint of the dentine formation, and allow an uncertainty of 10% of age-since-conception, rounded up to a whole year. If you only know the tooth is a premolar, this is not much of a problem as the two premolars have almost parallel development. If you only know it was a molar, this is more problematic as the midpoint ages are 6, 10 and 18 years. Here I might go for a uniform distribution on 4 to 22, as the biggest uncertainty comes which tooth it is and that is uniform across M1, M2 and M3.

 

As Helene says, secondary dentine may also contribute in older adults. Ours were all under 25 so we didn’t worry about it. Even so it may not have a large effect. If you imagine it accumulates linearly with age and contributes up to 20% of the dentine, then even in a 70-year old the average additional age would be no more than 7 years. For younger individuals, if you are rounding your final 95% HPD ranges outwards to the nearest 5 years, this is almost negligible.

 

 

Best wishes

Andrew

--

Dr. Andrew Millard

Associate Professor of Archaeology,

Durham University, UK

Email: A.R.M...@durham.ac.uk 

Personal page: https://www.dur.ac.uk/directory/profile/?id=160

Scottish Soldiers Project: https://www.dur.ac.uk/scottishsoldiers

Dunbar 1650 MOOC: https://www.futurelearn.com/courses/battle-of-dunbar-1650

 

From: ox...@googlegroups.com <ox...@googlegroups.com> On Behalf Of Helene Agerskov Rose
Sent: 16 August 2022 07:56
To: OxCal <ox...@googlegroups.com>
Subject: Re: Formation ages of 14C-dated human teeth

 

[EXTERNAL EMAIL]

--
You received this message because you are subscribed to the Google Groups "OxCal" group.
To unsubscribe from this group and stop receiving emails from it, send an email to oxcal+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/oxcal/600e111a-3d60-4649-b82a-f7b02cfb3da2n%40googlegroups.com.

[Erik Marsh]

Helene and Andrew – what wonderful answers, thanks so much! Now I know what I'm looking for, dentine formation, and it's clear that AlQahtani is a much better guide. In the mean time, I'm trying to find out if anyone knows which molars were dated.

Erik

[Erik Marsh]

So I've been able to track down some more details. Does the following look okay?

So for unidentified molars, I see the logic in a uniform range of 4–22 years. Is this range meant to include early 1st molars and late 3rd molars? I can imagine an argument for using 6–18, based on the median ages.

For 1st molars, I'll use 6±1 years.

For canines, 8.5±1 years.

For 2nd molars, 10.5±1 years.

What about accounting for secondary dentine in a 1st molar from a 45–55 year-old? 

What's the math there, maybe add 5±1 and use 11±2 years??

Thanks again!

Erik

[MILLARD, ANDREW R.]

Hi Eric,

 

Yes the logic for 4-22 is that this covers both the uncertainty on which tooth and the variability of the teeth. If you are using M1 6±1, M2 10.5±1 and M3 18±2 then 4-22 covers everything in their 2-sigma ranges.

 

Secondary dentine is harder because I haven’t found any good data on the percentage of the dentine that it constitutes at a particular age, or how variable that is. I’d be tempted to add 5±2 for a 45–55 year-old, as I don’t think it is many 10s of percent. But I’d also try running the model without it as I doubt that the results are very sensitive to what you choose for this.

To view this discussion on the web visit https://groups.google.com/d/msgid/oxcal/ea92cc79-9e3f-450a-a2f5-14455b57a885n%40googlegroups.com.