Read the Reviews, Brian

Oh dear, Dr John is trying CPR on the corpse of the Glacial Theory again, this time by raising doubts about Clarke and Kirkland's Salisbury Plain River Sediment Study. 

Brian blog post from today https://brian-mountainman.blogspot.com/2026/03/probable-bias-in-zircon-apatite.html 

The paper is at https://doi.org/10.1038/s43247-025-03105-3 

Peer reviews at : https://static-content.springer.com/esm/art%3A10.1038%2Fs43247-025-03105-3/MediaObjects/43247_2025_3105_MOESM1_ESM.pdf

Even the toughest critic in the peer-review file — Reviewer #1 — hammered the exact point the blog post keeps circling: the dominant Laurentian zircon signature is exactly what you'd expect from the eroded Paleogene cover (Thanet Formation and London Basin strata) that once blanketed Salisbury Plain. That background signal is pre-Pleistocene recycling, not fresh glacial detritus. The authors kept the Stonehenge framing after revision, and the reviewers accepted it because the study's real punchline is the absence of anything extra on top of that baseline: no prominent Darriwilian (~464 Ma) Welsh peak, despite 550 grains analysed. One lone outlier doesn't save the glacial hypothesis; it underscores how clean the negative result is.

The specific methodological concerns raised in the blog post were already addressed in the peer review process — which would be apparent to anyone who had read the review file carefully. Working through the list:

• Handpicking bias: The grains weren't handpicked. They were bulk-mounted into epoxy discs and randomly selected using automated TIMA mineralogy at Curtin's John de Laeter Centre. Reviewer 2 explicitly asked about recovery rates and grain selection; the authors provided automated mineralogy maps showing zircon and apatite abundance across all samples.
• Sample size: Reviewer 1 raised exactly this concern, pointedly questioning whether four samples was sufficient for a Nature-family journal. The authors demonstrated that all four samples are statistically indistinguishable (KS test P>0.05), that 550 zircon and 250 apatite grains were analysed across four samples, with 401 concordant zircon ages forming the primary provenance dataset and that ~120 grains is the accepted threshold for statistically meaningful provenance interpretation. The inter-sample consistency across independent catchments is itself strong evidence the signal is real and regional.
• Hydraulic sorting and grain size bias: The Frantz separator and heavy liquid separation steps are described in the methods. Reviewer 2 interrogated the recovery rates specifically. The authors' response — that the minerals are relatively abundant in the samples (zircon 1.5–56 wt%, apatite 3.5–13 wt%) — addresses the concern that a tiny exotic population might be dominating the signal.
• The Paleogene cover issue: Most strikingly, the blog post misses the deepest challenge to the paper — one that Reviewer 1 raised so forcefully they initially declined to recommend publication. The Laurentian signal is entirely consistent with the former Paleogene cover, and that's actually the authors' own conclusion. The glacial framing survives not because the Laurentian signal is surprising, but because of what's absent from it.
• The apatite evidence: The blog post doesn't mention the apatite data at all. The complete absence of old Laurentian apatite — despite its abundance in the Laurentian basement — is independent corroboration of deep, multi-cycle sedimentary recycling. As Reviewer 2 (Gary O'Sullivan, Trinity College Dublin) noted, it would be impossible to deliver old zircon via first-cycle glacial transport without also delivering old apatite. The zircon-poor Chalk makes this the ideal null detector, and the apatite result is a second, chemically independent line pointing the same way.

This paper isn't where the Glacial Theory dies — it's just the latest nail. The glacial hypothesis has been losing ground on multiple independent fronts for years.  That cumulative burden is where the argument is effectively over — this paper simply adds one more count to an already lengthy indictment.