The Gold Standard of Archaeology | Particle-Size Analysis, Sediment Mixing and Depositional Energy

1. Frame a sediment-formation question

Back to contents

Ask whether a layer differs from its substrate, whether a feature fill accumulated gradually, whether flood or wind input changed, or whether construction and trampling mixed sources. The required scale determines sampling interval and method. A single texture label cannot reconstruct a sequence, and a bulk sample that crosses context boundaries answers a different question from one taken within a defined microstratum.

2. Describe the deposit before reducing it to numbers

Back to contents

Record colour, bedding, sorting, clasts, inclusions, boundaries, moisture and structure in the field. Photograph and draw the sampling position. Grain-size results gain meaning from stratigraphy and fabric; laboratory percentages cannot restore a lost boundary or tell whether a coarse fragment was part of the matrix, an artefact or later root disturbance.

3. Match the measurement method to the size range

Back to contents

Dry or wet sieving measures retained mass in coarse fractions. Sedimentation methods infer fine sizes from settling behaviour. Laser diffraction estimates equivalent diameters from light scattering and depends on optical and shape assumptions. Methods overlap imperfectly and can yield different distributions for the same sample. Report the instrument, range, settings and basis rather than merging results as if they were interchangeable.

4. Make pretreatment a documented decision

Back to contents

Drying, disaggregation, removal of organic matter or carbonate, deflocculation and sonication can reveal mineral grains but can also destroy aggregates that matter archaeologically. Decide whether the target is natural particle size, stable aggregates or the deposit as encountered. Record reagents, duration and discarded material. Test sensitive samples before applying one aggressive preparation to an entire sequence.

5. Control subsampling and fraction boundaries

Back to contents

Sediment is heterogeneous, and coarse pieces settle or segregate in storage. Homogenise where appropriate, split with a reproducible method and retain enough mass for the largest expected grains. Use stated class boundaries and preserve raw bin data. Summary statistics such as median size or sorting are useful but can hide bimodal mixtures in which the same median arises from very different sources.

6. Worked example: a silty ditch fill with two inputs

Back to contents

A ditch profile contains a lower laminated silt and an upper mottled fill. Wet sieving and laser analysis show the lower layer is narrowly sorted fine silt, while the upper layer is bimodal with fine sediment and coarse sand. Micromorphology identifies intact laminae below and bioturbated aggregates above; artefacts concentrate only in the upper fill. The interpretation is a shift from quiet water deposition to mixed infilling and disturbance, not a calculation that converts mean grain size directly into flow speed.

7. Read mixtures before assigning energy

Back to contents

Coarse grains often require stronger transport than fine grains of similar density and shape, but source availability, cohesion, flocculation, vegetation, settling and winnowing complicate the relationship. A bimodal curve can signal mixed sources or processes; poor sorting can reflect rapid deposition, reworking or human dumping. Use experimental or regional analogues cautiously and keep multiple formation routes in play.

8. Changed-context transfer check: an occupation floor

Back to contents

A floor sample is finer than adjacent construction fill. The transferable principle is to compare matched contexts and integrate fabric. The new context adds trampling, sweeping, ash, plaster and imported sediment. Fine texture could reflect deliberate surfacing or repeated cleaning rather than low-energy natural deposition. Thin-section evidence and chemical indicators are needed before naming the process.

9. Know the common misconceptions

Back to contents

Grain size does not uniquely determine depositional environment or flow energy. Laser-derived equivalent diameters are not literal sieve diameters, especially for platy clay particles. Removing organics and carbonate changes the material being described. A smooth curve does not prove a single source, and a cluster algorithm does not create stratigraphic phases. Statistical precision cannot compensate for context mixing or inadequate sample mass.

10. Audit field, preparation and calculation together

Back to contents

Archive coordinates, depths, context sheets, sample photographs, storage history, dry mass, pretreatment, splitter method, instrument calibration, size bins and raw replicate data. Plot distributions and cumulative curves, not only means. Recalculate after alternative binning, exclude obvious clasts transparently and compare untreated and treated aliquots where aggregation matters. Link every interpreted change back to a visible context boundary.

11. Frequently asked questions

Back to contents

Which method is best? The method whose size range and assumptions fit the sediment and question. Can sieve and laser data be combined? Only with an explicit overlap and comparability procedure. Does finer mean calmer water? Sometimes, but cohesion, source and post-depositional mixing can reverse that simple reading. How many samples are enough? Enough to resolve within-context variation and the boundaries being tested; no universal count exists.

12. The gold standard keeps the sediment history larger than the histogram

Back to contents

Strong particle-size work begins in the profile, uses a method suited to the material and preserves preparation and raw distributions. It treats energy as one explanation among source, sorting, human action and later mixing. The gold standard turns a histogram into archaeological evidence by showing how the measured particles relate to the deposit that was actually excavated.