Grass and Grass Phytolith Analysis

The family Poaceae is the fifth largest flowering plant family, with around 11,500 species in over 750 genera. Field identification typically uses vegetative characters (leaf blade shape, ligule type, sheath opening) that are not preserved on clothing fragments. Forensic identification of Poaceae material from exhibits relies on the floret and its bracts, which are more durable and carry the most diagnostic characters.

• Lemma surface and venation: the number of veins, their spacing, and the presence of ridges, tubercles, or hairs on the lemma surface are consistent within species. In grasses with distinctive awns, the awn insertion point and curvature are also diagnostic.
• Glumes: the outer bracts at the base of each spikelet. Upper and lower glume shape, surface texture, and relative length are diagnostic at genus level in many tribes.
• Caryopsis surface: pericarp texture (smooth, papillate, striate), hilum shape (punctiform or linear), and overall grain shape (elliptic, obovate, dorsally compressed) are useful when the grain is recovered in casework.
• Trichomes on glumes and lemmas: long, silicified hairs called prickle hairs or macro-hairs occur in some genera and contribute to surface attachment to clothing; their density and morphology are also diagnostic.

Standard references for European casework include Hubbard's 'Grasses of the British Isles', the Flora Europaea, and Soreng et al.'s global grass checklist. For tropical and subtropical genera, Kew's GrassBase database offers morphological descriptions and distribution data accessible to practitioners globally.

Grass phytoliths form in epidermal cells, particularly in the long cells, short cells, and bulliform cells of the leaf blade. Different cell types produce different silica shapes, and different grass taxa have different proportions of each cell type, so the resulting phytolith assemblage is taxonomically informative. The key morphotypes and their associations are:

Because most forensic scenes involve multiple grass species, the analyst recovers an assemblage of morphotypes rather than a single type. The assemblage is characterised by the percentage of each morphotype and compared to a reference dataset for the region. A scene dominated by saddle-shaped phytoliths points toward warm-season arid or semi-arid grassland; a scene dominated by elongate and rondel types points toward cool-season temperate turf. This ecological signal is often informative even when no single species can be identified.

Phytolith extraction follows a well-standardised protocol developed from archaeobotanical methods and adapted for forensic use. The procedure begins with wet-sieving to remove particles above 250 micrometres, which includes most mineral grains, organic fragments, and root debris that would obscure the phytoliths on a finished slide.

1. Carbonate removal
   The sieved sample is treated with dilute hydrochloric acid to dissolve calcium carbonate, which would otherwise form a mineral crust over phytoliths during heavy-liquid separation.
2. Organic matter removal
   Hydrogen peroxide or Schulze's reagent (nitric acid plus potassium chlorate) oxidises organic matter. This step is critical: unoxidised organic debris darkens the slide and blocks transmitted light, making morphotype identification impossible.
3. Heavy-liquid flotation
   The cleaned residue is suspended in zinc bromide or sodium polytungstate adjusted to a specific gravity of 2.3 g/cm3. Phytoliths (density approximately 2.1-2.2) float; mineral grains (density above 2.5) sink. The floating fraction is collected, washed, and dried.
4. Slide preparation and counting
   The phytolith fraction is mounted in Entellan or Canada balsam on glass slides. A minimum of 200-300 phytoliths are identified and counted per sample to give statistically reliable percentage estimates of each morphotype.

Unlike some forensic trace comparisons, where a single particle (a hair, a glass fragment) is matched to a source, phytolith evidence works at the assemblage level. The question is not whether a specific phytolith came from a specific plant but whether the assemblage in a soil sample recovered from the exhibit is statistically similar to the assemblage in soil from the reference location. This distinction matters for how the evidence is reported and how it is explained to a court.

Comparison methods range from simple qualitative assessment (the dominant morphotypes are the same between exhibit and scene) to quantitative multivariate approaches such as principal components analysis or cluster analysis applied to morphotype percentage data. The latter are more defensible under cross-examination because they provide an objective measure of similarity rather than relying on the expert's subjective assessment of which assemblages look the same.

Soil phytolith profiles are spatially specific for the same reason that pollen profiles are: grassland vegetation composition varies over tens to hundreds of metres in response to soil moisture, drainage, land management history, and local disturbance. A field that has been under improved ryegrass management for 30 years has a different phytolith profile from an adjacent unimproved meadow even if both look like 'grass' to a non-specialist. This spatial specificity is what gives phytolith evidence its forensic value.

Grass fragments transfer to clothing through brushing contact: even walking across a grass field at head height transfers floret fragments to hair, collar, and sleeves. Grass florets with awns and prickle hairs transfer more efficiently and persist longer, and some of the most forensically useful genera, including Hordeum, Bromus, and Stipa, have prominent awns that actively penetrate fabric. The physical mechanism is directional: awns are designed to anchor into soil or animal coat to facilitate burial of the seed, and fabric acts similarly to animal hair as a substrate.

Phytolith evidence transferred via soil works differently. When soil from a specific location contaminates a boot sole, vehicle tyre, or clothing, the attached soil carries the phytolith assemblage of that location. Because phytoliths are very small and lodge between fabric fibres and into soil crumbs on shoe soles, they can persist for days to weeks even after visible soil has been cleaned away. This makes phytolith evidence useful in cases where a suspect claims to have cleaned footwear or clothing, and where no macroscopic soil is visible but microscopic analysis of seams and stitching recovers residual silica bodies.

The evidential value of a phytolith assemblage match depends on how distinctive the scene profile is relative to background levels in the wider area. An assemblage dominated by elongate Pooideae morphotypes in a temperate country with widespread ryegrass agriculture provides much weaker linkage evidence than one showing an unusual combination of Panicoideae and Chloridoideae morphotypes in a temperate region where tropical grass introductions are rare and localised.

The botanist's responsibility is to characterise the regional picture: how common is this assemblage type in the area, and how many other locations might produce the same profile? This requires survey data from multiple sites in the region, not just the crime scene and the suspect's known locations. Without a regional baseline, the comparison says only that two samples look similar, not that the similarity is unusual or informative.

Reference collections for phytolith identification include the International Phytolith Society reference database, the work of Piperno (2006, AltaMira Press) for tropical taxa, and national-level datasets compiled by archaeobotanical research groups. The forensic botanist should compile their own regional reference slides from known-location grass samples when undertaking casework in a new geographic area, both to improve identification accuracy and to document the local phytolith baseline.