Soil Colour and Munsell Notation

Albert Henry Munsell, an American artist and teacher, published his colour order system in 1905 to replace subjective pigment names such as "olive drab" with a three-dimensional space where any colour has a unique address given by hue, value, and chroma.

The soil-specific chart appeared decades later, when pedologists adopted the system for field description because consistency across thousands of soil surveys required something better than vernacular colour terms. The current edition covers hue pages from 2.5R to 5G, which spans the realistic range of mineral and organic soils. Each page is a grid of physical chips at stepped value and chroma positions.

To use the chart, the analyst holds an open page next to the sample and moves the sample along the grid until the chip and sample look identical under illuminant conditions specified by the protocol. The notation is read directly from the chip's printed position. It takes seconds once the analyst is trained, and produces a code that is transferable to any other user of the same edition of the chart.

Perceived colour shifts with the light source. Forensic protocols from Kenneth Pye's laboratory practice and standards from the US Soil Survey require natural north-facing daylight or a calibrated artificial source close to D65 (6500 K colour temperature), away from direct sunlight and coloured surfaces.

• Moisture state: read both moist (field-moist or rewetted) and air-dry. Record which state was used. Moist readings better represent the condition in which a questioned sample was found at a scene.
• Sample preparation: remove visible organic debris and large stones. Crush gently to a 2 mm fraction if clods are present. The colour of a lump and the colour of its matrix can differ.
• Illumination: north-facing daylight is the classical standard. Munsell-calibrated lamp boxes are acceptable when daylight is unavailable, but the source must be documented in the case notes.
• Background: hold the chart against a neutral grey or white surface. Coloured surroundings distort the perceived hue through simultaneous contrast.

Studies by Kirsch and Shields (1990s), and more recent work by Viscarra Rossel and colleagues, show that trained pedologists agree on hue page about 80-90% of the time on unambiguous samples but diverge more on samples that fall between chips. The problem is partly perceptual (individual variation in colour vision) and partly physical (the discrete chip grid cannot represent continuous variation).

In a forensic context this matters for two reasons. First, if the original examination and the defence re-examination are done under different conditions, any difference in notation could reflect procedure rather than a genuine soil difference. Second, when a match is reported, the uncertainty around the reading must be acknowledged: the correct statement is that the samples fall within the same Munsell range under the specified conditions, not that they are the same colour to arbitrary precision.

A contact or portable spectrophotometer pressed against a prepared soil surface measures reflectance at each wavelength from roughly 400 nm to 700 nm. The resulting curve is a colour fingerprint for that sample. Software converts it to CIE L*a*b* coordinates, and the Munsell notation follows from a standard transformation. The entire reading takes less than a second.

The reading is fully documented numerically, reproducible on the same instrument, and independent of the examiner. Statistical comparison of full reflectance curves using root mean square difference or discriminant analysis can distinguish samples that share a Munsell code but differ in the fine shape of their spectrum. Viscarra Rossel, Walvoort, McBratney and colleagues demonstrated this discrimination potential in agricultural pedology; the same principle applies to forensic sample sets.

In casework, colour serves two distinct functions. The first is elimination: if the questioned soil is 5YR 4/6 (yellowish red) and the proposed source is 10YR 6/2 (light brownish grey), the difference exceeds any plausible reading error and no further analysis is needed to exclude that source.

The second function is inclusion for further testing. When a questioned and reference sample share a Munsell code within the known variability of the reading, that is not a match in the individualising sense. It is a reason to look harder at particle size, mineralogy, chemistry, and biology. Colour keeps a sample in the running; the other properties have to close the argument.

The evidential weight of a colour agreement depends on how distinctive that colour is in the local terrain. A 10YR 4/3 dark brown is common across agricultural topsoils through much of temperate Europe and North America. Matching it gives little discrimination. A 10R 3/6 (dusky red) is rare, found only in certain lateritic or terra rossa parent materials. Matching it on footwear is a strong signal worth pursuing.

The central limitation is non-uniqueness. Many unrelated soils share colours, particularly in the common 7.5YR-10YR range at moderate value and chroma, and a colour match is not an identification. Forensic geologists present colour as one line of evidence within a multidimensional comparison precisely because the phrase "same colour" can imply more certainty than the data supports.

• Mixing: soils on footwear or clothing are rarely pure. A mix of topsoil and sub-soil can average to a colour that does not match either source cleanly.
• Alteration during crime: blood, oil, or chemical contamination at the scene changes the colour of any deposited soil before the sample is collected.
• Time since deposition: oxidation, drying, and biological action on the garment surface can shift colour over days. The questioned sample may no longer match its source because it has aged.
• Sub-2 mm vs. field heterogeneity: colour of the fine earth fraction can differ from visible mottles in the bulk sample. A single reading may not represent the variability in the source area.

Kenneth Pye and Jacqueline Blott, in their foundational work on forensic soil comparison, are explicit on this point. Colour is fast, cheap, and non-destructive, making it a sensible first step. But the conclusion of a forensic comparison comes from the full suite of analyses, with colour contributing its part of the overall weight rather than carrying the case alone.