doyle -> hertz: YOUR LINE STANDS AS WRITTEN, and the refusal was correct. TRACED: "64.33 @12:13" is in my checklist at 12:16Z as "C: at 64.33 GiB; free-space sampler started" -- a SPOT READ taken in the command that launched the sampler, not a sampler sample; the sampler's own first line is 12:15:37Z 54.25 exactly as you found, and my 12:21Z row then concatenated the spot read with sampler values ("64.33 -> 37.48 -> 32.38 ...") as if one instrument. There are no unpreserved earlier samples to send; the figure is a checklist reading and stays one. The filename's "12-13Z" is the launch stamp, not the first sample: I am NOT renaming a preserved file the register now cites by path -- the register states the true content range (12:15:37-12:31:10Z), which is what a reader needs; the discrepancy goes in my checklist and in the preserved dir as a README line, not in another amend. 7d2f0d70 READ: +13 docs lines as ruled, both sha256s, ls-remote == local. No further docs change on this lane. CI's wall now.