Public-domain print source, as identified by the HathiTrust exports. Original OCR provenance and rights statements are retained with the conversion inputs.
Google-digitized HathiTrust OCR; University of Iowa copy. SHA-256: cf377ae4f00f54d34fa33c0bad6c2016704cd9f27e78e5431aeecdf20600090b
Google-digitized HathiTrust OCR; Harvard University copy. SHA-256: 10200e9c45ca2cf4a25ea683146691f23a71d2f127b7eec3c9990faf6fc22a61
Explanatory notes only, pp. 30–140. The introduction, Greek running text and separate textual appendix (pp. 141–160) are excluded. Greek readings naturally discussed within notes are retained.
Iowa is the base OCR. Harvard is collated entry by entry. Disagreements are not automatically emended: base readings appear in sic, with alternate readings in collation.json. Identical OCR may still be wrong. Line-end hyphenation is mechanically joined; no spelling modernization is applied.
Source note numbers remain in n. Corresp reuses the existing Wecklein 1885-to-Smyth crosswalk. Content corroboration against the 1888 Greek page text is recorded per entry; otherwise the alignment is explicitly provisional. This is not a newly verified complete 1888 crosswalk.
f. is expanded to one following line; ff. anchors only its starting line. Broad source ranges remain ranges. Repeated source numbers are combined into one addressable entry. Physical pages remain pb milestones within notes.
Greek script is marked as mentioned. Recognizable abbreviated citations are marked bibl; unverified cross-work line correspondences are not invented. Further quotation/gloss tagging and scan proofreading remain necessary.
Select an entry by its printed source line or range.