Collection of Natural Histories and Travel WritingPublic Domain English, German, French, Spanish, Portuguese and Italian

How the list was built

• Project Gutenberg: every English, German, French, Spanish, Portuguese or Italian text in the PG catalogue (October 2026), scored on LoC subject headings and LoC classes QK/QL (2 points each), PG bookshelves (1–2), LoC classes QH and G (1) and title keywords (1). Works scoring below 2 and works catalogued as fiction are left out.
• Deutsches Textarchiv: the genres Reiseliteratur (in Belletristik and Gebrauchsliteratur), Naturgeschichte, Zoologie, Biologie and Geologie in full. Geographie is split by title: titles naming a journey count as travel, the rest as natural history. Works in mixed genres (Naturwissenschaft, Populärwissenschaft, Bericht, Ethnologie and others) were sorted by the LLM described under BHL; this brings in Humboldt's Kosmos, Brehm and Bölsche.
• Wikisource (German): the categories Reisebericht, Zoologie (with Entomologie, Ichthyologie, Malakologie, Ornithologie), Botanik, Mykologie and Biologie, without encyclopedia entries. Most entries are magazine articles, many from Die Gartenlaube.
• Wikisource (English, French, Spanish, Italian and Portuguese): the categories for travel, exploration, natural history, zoology, botany, ornithology and entomology with their subcategories (Spanish: Relatos de viajes, Zoología, Botánica, Ciencias naturales; Italian: Diari di viaggio, Storia naturale, Zoologia, Botanica, Scienze naturali), and every text whose Wikidata item has the genre travel literature or nature writing. These categories also hold poems, fiction and encyclopedia entries, so the LLM kept only natural histories and travel accounts. Many entries are articles from Popular Science Monthly, The Zoologist and Le Tour du monde.
• Biodiversity Heritage Library: English, German, French, Spanish, Portuguese and Italian monographs up to 1929 from BHL's data export whose subject headings include natural history, nature study, or description and travel, voyages and exploration, leaving out seed and nursery catalogues, field notebooks, periodicals, textbooks and juvenile works. BHL is mostly scientific literature, so the same LLM then kept only natural histories and travel accounts and dropped technical reports, taxonomic monographs, expedition results and proceedings. Links open BHL's page viewer at the first volume. Unlike the other libraries, which offer proofread transcriptions, BHL has page scans with uncorrected OCR text; it is serviceable for 19th-century English in roman type but often garbled for blackletter, Fraktur and older prints.
• Internet Archive: texts dated up to 1929 in English, German, French, Spanish, Portuguese or Italian whose subject headings name natural history, naturalists, description and travel, voyages, exploration or scientific expeditions, without IA's copies of BHL and Gutenberg and without manuscript ships' logbooks. Libraries scanned many works more than once, so scans and volumes with the same author and title are one entry, linked to the most read copy. The same LLM as for BHL then kept natural histories and travel accounts. Like BHL, these are page scans with uncorrected OCR.
• Gallica: books up to 1929 that the BnF catalogue lists as freely readable in Gallica, with subject headings for voyages, description and travel or natural history, or titles naming a voyage, excursion or natural history, sorted by the same LLM. Gallica's own search refuses scripts, so the BnF catalogue's SRU interface was used instead.
• Biblioteca Nacional de España, Universidad Complutense de Madrid, Biblioteca Nacional de Portugal: printed texts up to 1929 in the six languages whose title names a journey, voyage, expedition, natural history, animals or plants, or whose subject headings name travel or natural history, found through Europeana, because the BNE's own services refuse scripted requests. BNE entries open the Biblioteca Digital Hispánica's viewer. Portuguese entries open the Biblioteca Nacional Digital. Complutense entries open their scans in HathiTrust rather than Google Books, which made them but often shows only snippets outside the US; HathiTrust shows these pre-1880 books in full everywhere. Copies are merged and the same LLM sorts them as for the Internet Archive. Some BNE records name their author only by an identifier that could not be resolved, so these works show no author.
• Darwin Online and Wallace Online (ed. John van Wyhe): every English, German and French edition of Darwin's and Wallace's books up to 1929, from their bibliographical databases. The travel narratives (the Beagle journal, the Amazon, the Malay Archipelago) count as natural history and travel, the other books on evolution, geology and the habits of plants and animals as natural history. Left out are the barnacle monographs, the specialists' parts of The Zoology of the Voyage of H.M.S. Beagle, letters and autobiographies, and Wallace's writings on spiritualism and politics.
• edition humboldt digital (BBAW, CC BY-SA): Alexander von Humboldt's travel journals of the American journey 1799–1804, the European journeys 1790–1805 and the Russian-Siberian journey 1829, his correspondence with Ehrenberg, Kunth, Hooker, the Schlagintweit and Schomburgk brothers and others (one entry per correspondence, each letter linked), and his notes for the new edition of the Géographie des plantes, from the edition's TEI data. The journals switch between French and German; both are listed where each makes up at least a fifth of the text. The edition's administrative documents and modern research articles are left out.
• EcoCor: the 16 of its 213 texts that are natural history or travel writing rather than fiction, labelled by an LLM and checked by hand. Each links to EcoCor's plant and animal annotations.
• Genres: the same LLM then gave every work one genre from a fixed list: travel account, voyage or expedition, sketches of a region, hunting, sport and mountaineering, naturalist's travels, natural history, nature essays and observation, letters and journals. Travel works it called a guidebook, topography or local history, emigration or promotion tract, devotional book, technical scientific work, history or biography, fiction or poetry, or periodical were left out, because the broad subject heading "Description and travel" also covers these. Natural-history works came in through precise subject headings, so they were left out only as fiction, devotional books, promotion or guidebooks; technical works, regional descriptions and natural-history periodicals stay and show their genre (periodicals are hidden by default, see the checkbox). Works from the curated sources (the DTA's genres, Humboldt, Darwin and Wallace, EcoCor) were kept whatever their genre. A genre adds its category to a work but never removes one, so a naturalist's voyage stays natural history. Genres are a reading aid and can be wrong.
• Places come from LoC subject headings (Gutenberg), or from the title, description or subject headings, read by an LLM (qwen38-27b on Uni Passau's innkube cluster). They were geocoded with OpenStreetMap's Nominatim. Broad regions such as "Africa, Central" or "Orient" sit at an approximate centre.
• Keywords were written by the same LLM from each work's title, author, date and subjects. They are a reading aid and can be wrong.
• Publication years for Gutenberg works come, in this order, from Wikidata (items carrying the Gutenberg ebook number; the date of the original work is preferred), from an Open Library search on title and author, or from the ebook's own title page (the "Original publication" line of newer ebooks, or the transcribed title page). A year is kept only if it fits the author's life (up to 60 years after death, for posthumous editions) and is before 1971, which leaves out modern reprints. Where none of these gives a year and the author's birth and death years are known, the middle of the author's life stands in as an estimate, marked "c." and "est."; otherwise the entry shows the author's birth year, marked "auth. b.". Open Library sometimes has a misdated early record, so a few years are too early.
• Volumes and editions of one work in the same library and language are merged into one entry ("3 volumes", "2 versions"): same author and the same title once volume numbers are removed, nearly the same title, or the same work on Wikidata. Translations that Wikidata ties to the same work are linked.
• Works found in more than one library are listed in each and tagged "also in". The scan libraries are the exception: an Internet Archive, Gallica, BNE or Complutense copy of a work that another source already has is not listed separately; that source's entry is tagged "also in" that library instead.
• Download the data and scripts: gutenberg_nature_travel.csv · bhl_nature_travel.csv · wikisource_nature_travel.csv · ecocor_nature_travel.csv · ehd_nature_travel.csv · dwo_nature_travel.csv · ia_nature_travel.csv · bnf_nature_travel.csv · europeana_nature_travel.csv · gutendex_nature_travel.py · bhl_nature_travel.py · wikisource_nature_travel.py · ecocor_nature_travel.py · ehd_nature_travel.py · dwo_nature_travel.py · ia_nature_travel.py · bnf_nature_travel.py · europeana_nature_travel.py · build_overview.py · genres.py · years.py · geolocate.py · keywords.py · group.py · keyword_types.py · compact.py.