On: 10,000-Year-Old Stone Tool Site Discovered in Suburban Seattle

[ars]

As far as I know you can't radiocarbon date anything that has been soaked in water.

That would explain the discrepancy between 3,500 and 10,000 years.

[jschwartzi → ars]

I'm curious why you can't? I'm guessing that seawater permeation would alter the isotope distribution, but I also wonder if the alteration is predictable in some fashion. Does seawater affect the decay rate in some fashion?

[ars → jschwartzi]

Yes, that's exactly it. You can sometimes adjust if the water was constant and you have some other clock you can calibrate to (for example the bottom of the ocean at a specific depth).

But if you have water that came and went then the errors are too large.

[jccooper → ars]

Uh, no. Radiocarbon dating is routinely done on such things as shells, shipwrecks, peat, mud, and groundwater.

[ars → jccooper]

It only sometimes works if the water is constant, and not replenished. And even then you need an external clock to calibrate each site.

If it's things like floodwater then the errors are too large.

Water does not have the same isotope ratio as air, and water can also move carbon in and out of the sample.

They have procedures for adjusting the dating for marine samples, but it's only valid in certain cases.