The Gold Standard · How X Works
What passes through the mesh can rewrite the assemblage
An excavated deposit is not the same thing as the finds that reach a tray. Mesh size, processing method, sample volume and sorting effort act as filters. This guide makes those filters measurable so small bones, beads, seeds and fragments do not disappear silently from interpretation.
Choose a reading route
- Start from first principles — define the unit before interpreting a number.
- Follow the worked example — see the method under controlled conditions.
- Test transfer — change the context and check what still holds.
- Audit evidence — inspect records, assumptions and failure points.
- Use the practical checklist — apply the idea to new work.
Chapter contents
Foundations: Chapters 1–3
Method: Chapters 4–6
Transfer and limits: Chapters 7–9
Connections and use: Chapters 10–12
1. Recovery is part of observation
Archaeological recovery is a measurement process. Excavation exposes material; sampling selects sediment; sieving separates size fractions; sorting recognises items; recording turns them into data. Loss can occur at every stage.
A bag labelled “finds from context 214” is therefore not a transparent sample of the past. It is the surviving output of a chain of decisions. The gold standard is to document that chain well enough for another researcher to understand which materials had a realistic chance of recovery.
First principle
Interpret abundance only after describing how material entered the assemblage.
2. Mesh size changes what exists in the dataset
A 10 mm mesh retains objects with different dimensions from a 4 mm or 2 mm mesh. Shape matters too: a long narrow fragment may pass diagonally, while a round item of similar mass may remain. Wet sieving can break or expose material differently from dry sieving.
Smaller mesh is not automatically better. It raises processing time, residue volume and recognition demands. The appropriate design depends on expected remains, research questions, sediment and conservation limits. What matters is a justified method and a record of its consequences.
| Decision | Potential gain | Potential cost |
|---|---|---|
| Smaller mesh | More tiny bones, beads and plant remains | More sediment and sorting time |
| Wet processing | Cleaner small finds in some sediments | Fragile material may be affected |
| Larger sample volume | Better chance of rare material | Higher processing burden |
← Previous · Contents · Next →
3. Define the recovery fraction
A recovery fraction can describe the share of observed items in a size band, the proportion of a bulk sample processed, or the proportion retained at successive meshes. The label is meaningless unless the denominator is named.
For nested sieves, one transparent measure is retained items at a mesh divided by all items recovered across the recorded mesh series. For sample processing, report litres or kilograms processed divided by the bulk sample collected. Do not mix those two fractions.
- Record the collected sample volume.
- Record the processed volume and any subsampling.
- Record mesh sizes and sequence.
- Keep residues by fraction when feasible.
- State the counting unit: item, fragment, weight or identified specimen.
← Previous · Contents · Next →
4. Worked example: nested recovery
A fictional 20-litre sediment sample is washed through nested 10 mm, 4 mm and 2 mm sieves. Sorting recovers 24 items at 10 mm, 61 at 4 mm and 118 at 2 mm: 203 items in total.
| Fraction | Recovered items | Share of recovered total |
|---|---|---|
| 10 mm | 24 | 11.8% |
| 4–10 mm | 61 | 30.0% |
| 2–4 mm | 118 | 58.1% |
| Total ≥2 mm | 203 | 100% |
If only the 10 mm fraction had been retained, nearly nine-tenths of the counted assemblage would be absent. That does not mean every site will show the same pattern. It demonstrates why a coarse-recovered assemblage cannot be compared uncritically with a fine-sieved one.
Composition can shift as well as count. Fish vertebrae or tiny beads may cluster in the fine fraction, while large mammal fragments dominate the coarse fraction.
← Previous · Contents · Next →
5. Design a recovery experiment
Before committing an entire excavation to one method, run a controlled pilot. Select comparable sediment, measure equal volumes, process with planned meshes, standardise drying and sorting time, and classify finds using the same rules. If samples are spatially heterogeneous, use multiple paired or split samples rather than one dramatic example.
The pilot should answer operational questions: how much extra residue is produced, which categories increase, how recognition error changes and whether fragile material survives. Preserve enough metadata to repeat the comparison.
Avoid false precision
Counts from small pilots describe the tested sediment, not a universal recovery rate. Use them to design the project, not to promise exact correction factors for every context.
← Previous · Contents · Next →
6. Normalise without erasing context
Counts per litre can help compare unequal processed volumes, but they do not correct for depositional density, sediment type, fragmentation or spatial clustering. Weight can stabilise some fragment counts but may hide numerous small items. Number of identified specimens depends on taxonomic and anatomical recognition.
Report more than one useful denominator when interpretation depends on it: items per litre, weight per litre, and the proportion identified. Keep contextual information alongside normalised values. A method that makes contexts comparable should not detach finds from stratigraphy.
| Measure | Useful for | Main caution |
|---|---|---|
| Count per litre | Density under equal volume | Fragmentation inflates counts |
| Weight per litre | Bulk material | Heavy pieces dominate |
| Identified specimens | Taxonomic pattern | Recognition varies by fraction |
← Previous · Contents · Next →
7. Transfer test: a shell midden
Transfer the method to a shell midden containing large shell, tiny fish bone and charred seed. A coarse mesh may represent shell mass reasonably while badly under-representing fish and plant remains. One recovery rule cannot be assumed adequate for every material class.
A defensible design might combine measured bulk sampling, a coarse fraction for large shell, a fine heavy residue for small bone and an appropriate flotation route for plant remains. Each pathway requires its own volume and mesh record.
Transfer succeeds when the analyst can trace every reported category back to a recovery opportunity.
← Previous · Contents · Next →
8. Misconceptions and limits
- “Hand collection shows the important finds.” Importance is not proportional to size or visibility.
- “Fine mesh removes bias.” It changes bias and may introduce breakage, sorting fatigue or recognition limits.
- “Counts can be corrected later.” Missing fractions rarely support reliable retrospective correction.
- “One mesh size is comparable across sediments.” Clay, gravel and organic deposits behave differently.
- “More items mean more past activity.” Recovery, fragmentation and deposition also drive counts.
Some sites cannot be sampled destructively or processed with water. Conservation requirements, permits and cultural responsibilities constrain technique. The correct response is to document the limit and narrow the claim.
← Previous · Contents · Next →
9. Recovery audit
Audit field sheets against bags, sample registers and laboratory logs. Check whether litres collected equal litres processed, whether a flot or residue is missing, whether mesh labels changed, and whether sorting was partial. Photograph unusual fractions and record who sorted them.
| Audit point | Evidence | Failure signal |
|---|---|---|
| Volume | Sample sheet and container marks | Unknown or estimated after processing |
| Mesh | Equipment log and labels | “Fine” without millimetres |
| Completeness | Residue disposition | Unsorted fraction discarded |
| Recognition | Reference collection and checks | Category absent only in one sorter |
A recovered assemblage is trustworthy when its absences have plausible methodological explanations, not when the spreadsheet simply has no blank cells.
← Previous · Contents · Next →
10. Questions readers often ask
Can results from different mesh sizes be combined?
Only with clear stratification or a validated calibration design. Keep fractions and methods visible; do not pool first and explain later.
Should every sample be processed completely?
Not always, but subsampling must be designed, measured and recorded. Arbitrary partial sorting creates an unknown denominator.
What should happen to residues?
Retention depends on research value, policy, storage and permits. Record any disposal decision and which fractions remain available.
← Previous · Contents · Next →
11. Reading routes and ecosystem connections
Start with the Gold Standard of Archaeology. Connect recovery to context and stratigraphy, to sediment and site formation, and to reproducible archaeological experiments.
Use the companion guide on assemblage integrity and mixed deposits after checking recovery. Continue through learning that transfers into life, The Gold Standard and the How X Works hub.
← Previous · Contents · Next →
12. Sources and practical checklist
Authoritative guidance includes Historic England’s notes on sampling and sieving for animal bones, its current Environmental Archaeology appendix, and Managing Lithic Sites. These sources show why sample choice, mesh and research aim must be planned together.
- State the research question.
- Measure collected and processed volume.
- Name every mesh in millimetres.
- Keep fractions distinct.
- Standardise sorting and recording.
- Pilot where methods may change composition.
- Report recovery limits before interpreting absence.