Identify a piece

Read the piece before you search

Begin with the jewelry in hand. Notice its form, stones, colors, construction, and overall composition, then compare those observations with several verified archive records.

Search the archive

Five steps

Move from broad observations to close comparison

Use the strongest details first. Leave uncertain attributes open instead of forcing a match.

  1. 01

    Start with the whole piece

    Name the object type and notice its proportions, symmetry, and whether it belongs to a set.

    If the form is unclear, keep the search broad.

  2. 02

    Read the stones and color

    Notice dominant colors, stone shapes, finishes, and arrangements such as navettes, chatons, cabochons, or pearls.

    Distinct combinations are often more useful than a single familiar color.

  3. 03

    Inspect the reverse

    Compare settings, metalwork, joins, links, pin or clip hardware, backing shapes, and how decorative elements are attached.

    Construction can be more informative than the decoration on the front.

  4. 04

    Notice the composition

    Look at motifs and design language: flowers, leaves, stars, crosses, geometric forms, sprays, clusters, or asymmetric layouts.

    Consider how the entire design is assembled, not just one recognizable element.

  5. 05

    Compare several records

    Search with one or two strong observations, open several promising records, and compare the full evidence set.

    A convincing identification usually has several independent details in agreement.

Useful observations

Build a short evidence set

  • Object form and approximate proportions
  • Primary and secondary colors
  • Stone shapes, finishes, and arrangement
  • Motif or recognizable subject
  • Metal color, settings, and visible construction
  • Closure, pin, clip, or linking mechanism
  • Repairs, replacements, or missing components

Archive fields are clues, not a checklist

A blank archive field means the detail was not recorded or remains unknown; it does not automatically mean the feature is absent.

Ready to compare?

Start broad, then narrow

Choose one or two observations you trust. Add another filter only when the result set is still too large.