Research practice · illustrated guide
Documenting a pocket watch: photographs, numbers and evidence
Build a useful identification record without confusing a signature with authorship, a case number with a movement number, or a running watch with a serviced watch.
Start with an observation, not a conclusion
A good watch record begins with what can be observed safely. Write ‘name on dial’ rather than ‘made by’ until you have evidence connecting that name to manufacture. A retailer's signature, movement signature and case maker's mark can refer to different participants in the same object. Do not collapse them into one author simply because a database expects a single name.
ICOM's Object ID standard offers a useful starting framework: photographs, a short description, and nine categories covering type, materials and techniques, measurements, markings, distinguishing features, title, subject, date or period, and maker. Its purpose is identification and possible recovery, not mechanical diagnosis or authentication. Adapt it to watch research by retaining separate fields for the movement, case, dial and accessories.
- Assign your own collection identifier; do not use a presumed factory number as the only key.
- Record who made each observation and when.
- Keep ‘unknown’, ‘not examined’ and ‘illegible’ distinct.
- Preserve any family account as an attributed account, not a verified fact.
Make an exterior photographic baseline
Photograph the watch front and back, then add details of signatures, damage, unusual features and accessories. Use an uncluttered background, diffuse illumination and a view that does not distort the object. Keep original files before cropping or annotating them. Where a scale can be included without contact or risk, record it; do not derive precise dimensions from a photograph without a reliable scale.
The photograph of Quare's silver clock-watch illustrates the limits of an exterior view. The dial arrangement and decorated case are visible, but the photograph cannot certify that the movement is original or that a striking mechanism works. A useful caption identifies the object, view, institution and collection number and says what is not shown. If a case resists opening, leave it closed: forcing it to obtain a photograph may damage the hinge, bezel, dial or movement.
- Keep a dated front and back view.
- Record pre-existing dents, cracks and missing pieces without retouching them away.
- Photograph labels, watch papers and packaging separately; retain their association with the watch.
- Ask a qualified professional before opening an unfamiliar, damaged or fragile case.
Separate numbers and copy inscriptions exactly
Create separate entries for movement number, case number, reference or model number, and any institutional identifier. Numbers in different locations need not match, and matching numbers are not sufficient proof that all parts began together. Copy the digits and the location; retain punctuation, letter prefixes and unclear readings. A photograph of the inscription is more useful than a transcription alone.
Service scratches inside a case are not automatically serial numbers. They may be repairers' codes, dates or job identifiers. Do not erase them to improve legibility elsewhere. For a hallmark, record the complete group of punches rather than choosing a date letter in isolation. Country, assay office, letter cycle and object type all affect interpretation. Link a proposed reading to its reference and state whether the punch is incomplete or worn.
Build an evidence table for dating and attribution
Keep the observed evidence separate from its interpretation. A museum catalogue may date a watch approximately; a case hallmark may date assay; a factory table may estimate movement manufacture; a receipt may date sale. Those dates answer different questions. If they disagree, document the disagreement before proposing a re-casing, replacement dial or later assembly.
Use the exact wording of a source, including ‘circa’, ‘attributed to’ and ‘possibly’. Cleveland's engraved egg-watch record distinguishes Denis Martinot as designer from decoration attributed to Theodor de Bry. Removing the qualifier would make the copied record less accurate. Museum cataloguing, as described by Collections Trust's Spectrum, is an ongoing process of gathering and cross-referencing knowledge, not a requirement to conceal incomplete research until every question is answered.
Record condition without turning it into a diagnosis
Describe visible case wear, dial defects, hand damage and missing elements separately. Distinguish an observed action from a claim about health: ‘seconds hand moved during observation’ does not mean accurate, recently serviced or safe to continue using. A condition baseline documents what was seen, by whom and under what circumstances. The condition of hidden surfaces remains unexamined until a safe professional inspection takes place.
Preserve the record before treatment and keep a later treatment log alongside it. Identify replaced parts and retained originals if known. Prioritise conservation of evidence over cosmetic improvement: polishing may remove marks and opening a compromised watch can cause harm. This guide is a documentation workflow, not instructions for cleaning, lubricating or repairing a movement.
Keep the record usable and secure
Store original photographs, edited copies and documents with stable names linked to your identifier. Keep a readable export of the record as well as the app entry; a backup held only on the same device does not protect against loss of that device. ICOM explicitly includes keeping the constituted documentation in a secure place as part of Object ID.
When sharing a record publicly, consider whether full serial numbers, private addresses, purchase documents or ownership details should be withheld. Keep the full version securely for your own identification and insurance needs. Record the source URL and consultation date for web references, and retain enough citation information to find the claim again if a link changes.
Evidence and its limits
| Evidence | What it can support | What it cannot establish alone |
|---|---|---|
| Dial signature | A name present on this dial | Who made every component |
| Case hallmark | A proposed assay context when correctly read | Exact movement manufacture date |
| Movement serial table | An estimate under that table's assumptions | Case originality or sale date |
| Museum accession | Identity within that collection | Current public display or operation |
| Observed running | Movement during the observation | Serviced condition or accuracy |
Questions about the evidence
Should case and movement numbers match?
Not necessarily. Record them separately and research the maker and case system. Neither matching nor different numbers alone settles originality.
Can I identify the maker from the dial?
The dial records a signature, which may identify a retailer, maker or later replacement. Corroborate it with the movement, case and documentary evidence.
Is a photograph enough for a condition report?
No. It records visible surfaces under particular conditions. Add a written description and explicitly identify anything not examined.
Should I wind it before documenting it?
Do not operate a fragile or unassessed mechanism merely to complete the record. Ask an appropriately qualified professional about safe examination.
Sources & references
- https://icom.museum/en/resources/standards-guidelines/objectid/
- https://collectionstrust.org.uk/spectrum/procedures/cataloguing-spectrum-5-0/
- https://www.metmuseum.org/art/collection/search/194088
- https://www.clevelandart.org/art/1929.994
Where sources disagree, the disagreement is stated rather than resolved silently.