Carl Friedrich Gauss: Werke and Individual Papers (Modern LaTeX Working Drafts)

Author record for Carl Friedrich Gauss in the Modern LaTeX Editions of Mathematics Manuscripts project. This page presents selected Werke volumes and individual passages as modern LaTeX working reader PDFs, with source TeX, component PDFs, compile outputs, source references, and audit reports preserved in ZIP artifacts.

Status at a glance.

• Reader PDFs: Substantial working surface: eight broad front-facing readers are present, totaling more than 3,000 pages: Werke Bands I, I alternate, II, III, VI, VII, XI Part I, and an individual-papers grouping.
• Band II source-checked cumulative: Limited active coverage: this update adds paired original-language and English readers for Werke Band II from the actual beginning through printed page 303, including the start of Theoria residuorum biquadraticorum. Commentatio secunda, Articles 24-29, the accepted forward cumulative material, the first Latin Nachlass section, De nexu [IX]-[X], and Dedekind remarks to De nexu through printed page 303.
• Current repair incorporation: Broad working surface: the cumulative rebuild includes repaired and source-checkable supplements for Disquisitiones Arithmeticae gaps, Band I alternate expansion, Band II quadratic reciprocity and Seeber material, Band III AGM formulas, Band VI chapters 8/10/11/13, Band VII chapter/article repairs, and Band XI Part I recovered material.
• Source artifacts: Substantial working surface: TeX sources, component PDFs, compile logs, source-page references, scan witnesses, render images, and audit reports are preserved in artifact ZIPs for inspection and continuation.
• Proofreading state: Limited active coverage: machine-assisted working drafts needing source comparison, theorem/reference checking, and mathematical proofreading.

Source-quality refresh. A local 2026-06-14 audit found that future Gauss continuation should use the coherent GDZ Werke source PDFs plus IIIF manifests across Bands I-XII as the baseline. Older IA/Rich/Google PDFs and broad old reader PDFs are fallback/provenance unless a specific tranche is re-audited. The old reader PDFs should not be treated as globally completed editions.

Quality note. This is not a critical edition and not a certified critical edition. Treat each package as working draft/source-witness material unless a later page-level audit explicitly promotes the range.

Use of this record. The top-level PDFs are intended for direct reading and download. ZIP files are for reconstruction, repair, and provenance. Large raw preservation remains on the main project record rather than being re-uploaded here.

Corrections and contributions. Source comparisons, LaTeX fixes, translation corrections, and reader-facing issue reports can be suggested through the GitHub mirror via issues or pull requests: https://github.com/KokunoYumeto/modern-latex-manuscripts.

Added 2026-08-14 — Werke Band IV (Geodäsie), first complete transcription pass. This is an addition; every file previously in this record is unchanged.

486 page-fragments covering printed pages 6–492, transcribed page by page from the GDZ scan (PPN236005081) and assembled into a 415-page reader. The source is 1-bit bitonal with no text layer, so the usual coverage check — fragment letters against the scan's own OCR — could not be run on it; completeness rests instead on the printed-page-number check and on assembly reporting no duplicate or missing page numbers.

What this is not. A first pass, not a verified edition. It has not been collated against the scan page by page, and no English translation of it exists yet. One page (printed 365) is absent: the transcription service refused it repeatedly under a content filter, on Latin text about terrestrial magnetism, and it is logged for a later pass rather than guessed at.

Defects found and recorded rather than smoothed over. Some plates in this volume carry no printed folio — the page number is written on them by hand — and two were misread from the cursive, one placing a leaf 380 pages out of position until the duplicate -page-number check caught it. A recurring glyph is transcribed as \bar{b} and marked uncertain: it is a b with a stroke through the ascender, printed in contrast to plain b, and whether the printer's sort is a barred b, a crossed b or a Fraktur b is not settled.

Emphasis in the Göttingen Werke is italic, not the letterspacing (Sperrung) used by Weber's Vieweg setting; the transcription instructions were written separately for this corpus so that the Weber convention was not carried over and markup invented that the printer never used.

How much to trust this. This is a first-pass transcription. It is structurally far better than OCR — the mathematics is marked up rather than mangled, page boundaries are recorded, and the volume compiles — but in per-character fidelity it is not yet reliably better than good OCR, and it should not be read as an edition.

Defect classes measured on this corpus, none of which a successful compile detects: whole printed passages dropped; pages filed against the wrong printed number; emphasis invented where the page sets plain roman, and lost where the page sets it; occasional fabricated words; and glyphs guessed at rather than identified. These are found only by a second pass that reads the scan again, and that pass has not been run on this volume.

Use it as a searchable, buildable starting point for correction — which is what it is — and consult the scan for anything that matters.