Skip to content
The Encyclopedia of Horology

Research practice · illustrated guide

Reading museum catalogues and auction listings critically

Compare object identity, attribution, dating, condition and photographs across collection records and sales descriptions without upgrading uncertainty into fact.

Identify the exact object before comparing claims

Begin with the institution's collection number or an auction's sale date and lot number, not merely a title such as ‘gold pocket watch’. Generic titles recur, and one manufacturer may have produced superficially similar watches for decades. Save the page address, consultation date and any visible distinguishing details. For a museum object, retain the accession number even if the title later changes.

Cleveland's egg watch is collection number 1929.994; the two Met clock watches illustrated in this research section are 17.190.1496a, b and 17.190.1447. These identifiers keep three specific objects separate from their broad categories. Do not join a sales history to one of them simply because the case shape or maker name resembles a photograph.

Read authorship as roles plus qualifiers

A record can identify a designer, movement maker, decorator and retailer independently. The Cleveland API gives Denis Martinot the role ‘designed by’ and Theodor de Bry ‘decorated by’ with the qualifier ‘attributed to’. A shortened caption saying ‘by Martinot and de Bry’ loses both role and uncertainty. Copy the full relationship before composing your own summary.

Auction houses and museums do not necessarily use every attribution term identically. Where a claim matters, consult that institution's explanatory wording or the sale's catalogue conditions. Do not invent numerical probabilities for ‘possibly’, ‘workshop of’ or ‘attributed to’. Record whose opinion is being reported, and distinguish it from an independently demonstrated signature or archival match.

Distinguish production, assay, sale and later intervention

A date may refer to a whole object, a movement, a case, an inscription or a period inferred from style. ‘Circa 1690’ remains approximate. A range such as the Met's ca. 1736–95 is not permission to choose a midpoint as a manufacture year. Preserve the range and explain the source of the proposed date if the catalogue supplies it.

A later case, dial, repair or presentation inscription can coexist with an earlier movement. That does not automatically make the object fraudulent, but it changes the historical question. Ask which component each source describes and whether the source actually examined that component. Keep an observation, the interpretation and the conclusion in separate fields so a later researcher can reconsider the evidence.

Use photographs as evidence with boundaries

An exact object photograph is preferable to a picture chosen only because a search title contains the maker's name. Check the visible number, signature, case decoration or other distinctive feature against the source page. Record which view it shows. A front view cannot establish escapement geometry, the underside of a balance or the originality of parts hidden behind a plate.

Images can also conceal condition through lighting, cropping or the omission of a view. A reflective case can look markedly different under soft illumination. Do not infer a restoration or defect from one tonal difference alone. When using a contextual image rather than the exact object, say so explicitly. Technical schematics should be labelled as explanatory drawings, not photographs or reconstructions of a particular movement.

  • Match the photo to the source object, not just its category.
  • Keep the institution's credit and source-page link.
  • Retain the stated licence separately from whether permission has actually been confirmed.
  • Do not treat access to a downloadable image as proof of unrestricted reuse.

Treat condition descriptions as scoped observations

A catalogue description and a condition report serve different purposes. A sales listing may describe apparent operation at one moment without guaranteeing service history or continued running. Read any limits on inspection before paraphrasing it. Ask whether the movement was opened, whether every function was tested and whether the examiner had appropriate expertise. If the text does not answer, record that limitation rather than assume a test occurred.

Do not turn ‘running’ into ‘excellent mechanical condition’ or a cosmetic grade into a reliability prediction. Document case, dial, movement and accessories separately. For a proposed purchase or treatment, consult an appropriately qualified specialist about the particular watch; a general comparison table cannot authenticate it or replace physical examination.

Separate holdings from the plan for a visit

An online collection record establishes what the institution catalogues, not what it will display on the day of your visit. Objects can be stored, lent, conserved or moved. A collection catalogue may remain available long after an exhibition ends. Use current official visitor information and ask about the exact accession number where viewing one object is important.

A research contact or library listing is not an entitlement to handle watches or access storage. Ask about appointments, facilities and restrictions. Retain a dated note of the reply and avoid copying old admission prices or opening times into a supposedly permanent object description. Collection history and visit planning should remain linked but distinct.

Build a bibliography that another reader can follow

Record author, full title, edition, publisher, year and the relevant page or catalogue entry when verified. The Met's 2015 European Clocks and Watches catalogue describes 54 selected objects; the institution's publication page supplies authorship, extent and access information. Watkins's fourth-edition Mechanical Watches bibliography helps locate historical literature, but a bibliographic listing is not independent proof of an object's attribution.

A citation should support a particular statement rather than serve as decoration below a page. If you consulted only a publisher announcement or API record, say so instead of implying that you read a complete monograph. When two sites repeat the same statement, trace their underlying source before calling them independent confirmation. Leave unresolved disagreements visible and update the record when stronger evidence appears.

Evidence and its limits

Qualitative research comparison; not an authenticity score
Record typeMost useful forImportant limitation
Museum object recordAccession, catalogued material, qualified datingNot proof of current display or operation
Scholarly collection catalogueExtended analysis and component descriptionsCheck edition and specific entry
Auction listingA dated offering and stated descriptionSeller's claims need scrutiny; inspect terms
Condition reportThe examiner's scoped observationsNot necessarily full testing or service history
Publisher announcementPublication metadata and stated scopeNot evidence that the book was read
Search resultFinding a possible sourceSnippet is not a substitute for the record

Questions about the evidence

Does ‘attributed to’ mean the maker is confirmed?

No. Preserve the qualifier and role exactly, identify the institution making the attribution and consult the supporting record.

Are museum watches always on display?

No. A collection record documents holdings, not a guarantee of public display. Check current official information before a visit.

Can I reuse any auction photograph?

No. Availability online does not establish usage rights. Retain the credit and source and seek permission where the licence has not been confirmed.

Are multiple websites independent evidence?

Not if they copy the same catalogue or listing. Trace the underlying evidence before treating repeated claims as corroboration.