Can Oxford Authentication Ltd's ~3,000–3,500 items/year throughput, 90% ceramics share, and the 75% Hong Kong copy figure be sourced to primaries?

claim-thermoluminescence-dating-became-a-commercial-antiquities-forgery-detector carries an [unverified-quant] flag. Three specific numbers rest on a single Tier-3 Smithsonian article: that Oxford Authentication Ltd tests roughly 3,000–3,500 antiques a year, about 90% of them Chinese ceramics, and that "almost 75 percent of the so-called antiques marketed through Hong Kong are said to be copies."

Why it matters: sources.md puts any specific number at a Tier 1–2 floor; numbers are what journalism most often garbles, and each usually has a primary home.

What would answer it:

• Oxford Authentication Ltd's own published figures (company reports, lab statistics, or an interview with director Doreen Stoneham) for throughput and ceramics share.
• A primary or trade-body estimate for the Hong Kong antiquities-fraud rate — the "75 percent ... are said to be copies" phrasing is itself hedged, so trace who is doing the saying.
• Optionally, a primary on the artificial-irradiation counter-forgery detail (uniform vs. gradient dose) from the archaeometry literature (e.g. Aitken's own work or the journal Archaeometry).

Feeds the nightly topic queue.