Short answer
By architectural paint research: taking tiny samples that contain every layer of paint down to the substrate and examining them in cross-section under a microscope, often with ultraviolet fluorescence and chemical analysis, to identify the sequence of schemes and their pigments. Scraping back layers on site is quicker but unreliable, because old paint discolours and thin layers are easily missed — and even a lab result has to be interpreted, since pigments and binders change colour with age.
The traditional method was to scrape through later paint on site until an early layer appeared, then match it by eye. It is fast, but it fails in predictable ways. Thin layers or glazes are easily cut through unnoticed; dirt layers between schemes can be mistaken for paint; and the colour exposed is the colour of an aged surface, not the colour that was applied. Linseed-oil paints yellow and darken when kept from light, varnishes brown, and some pigments fade or change. A creamy yellow revealed under three later coats may have been a cool white when it was new. Matching it faithfully reproduces the ageing, not the original scheme.
In laboratory paint research a sample a few millimetres across is taken with a scalpel, set in resin, cut and polished so that every layer can be seen edge-on under a microscope. The stack reads like a timeline: primer, successive finish coats, dirt layers that mark years of exposure, and new schemes. Ultraviolet light makes binders and varnishes fluoresce differently, helping to separate layers and date them. Polarised-light microscopy and infrared spectroscopy identify pigments and binders — which matters both for dating (some pigments only became available at known times) and for understanding how a colour has changed. The technique grew from work in museums on easel paintings and was applied to buildings from the mid-twentieth century.
Pigment identification is often more useful than colour matching. Finding a synthetic pigment in a layer sets the earliest date that layer could have been applied.
Turning a sample into a paint specification is interpretation. The researcher estimates the original appearance from the unaged parts of the layer, the pigment and binder identified, and knowledge of how those materials change. The result is often expressed as a reference in a colour-order system, such as Munsell notation, and then matched by a paint maker. A modern paint of the same colour may still look different because historic lead-based or distemper finishes had a different surface texture, gloss and translucency. Conservation projects therefore often specify the paint type as well as the colour, and sometimes choose to present a later scheme if it matters more to the building's story.
For a listed building, the Planning (Listed Buildings and Conservation Areas) Act 1990 requires consent for works that alter it in a way that affects its character as a building of special architectural or historic interest. Whether a repaint does that depends on the building and the change: repainting in a matching colour is unlikely to, while stripping historic paint or changing a significant decorative scheme may. In conservation areas the ordinary right to repaint can be withdrawn by an article 4 direction. The local planning authority's conservation officer is the person to ask; nothing on this page is legal or planning advice.
Why: The scheme was matched to aged, discoloured paint exposed by scraping.
Fix: Commission cross-section analysis and have the original appearance estimated from pigment and binder evidence.
Why: Historic finishes had different gloss, texture and translucency from modern acrylics.
Fix: Specify an appropriate paint type as well as the colour, and trial it on site.
Each statement is labelled by kind — established fact, a standard’s requirement, observed market data, a convention, or Colourwise’s own interpretation or analysis — with the strength of the evidence behind it.
FactModerate evidence
Historic paint analysis uses sampling and laboratory examination — including optical, fluorescence and polarised-light microscopy and Fourier-transform infrared spectroscopy — to identify paint layers, pigments and binders.
Caveat: From a secondary summary; the museum or heritage-body source it cites could not be reached for this review.
FactModerate evidence
On-site 'scratch and match' is regarded as unreliable because low magnification makes it easy to miss or misidentify layers and it does not account for paint ageing and discoloration.
Caveat: From a secondary summary; the museum or heritage-body source it cites could not be reached for this review.
StandardStrong evidence
In England and Wales, works that alter a listed building in any manner affecting its character as a building of special architectural or historic interest require authorisation under the 1990 Act.
Source: Planning (Listed Buildings and Conservation Areas) Act 1990, section 7
Colourwise interpretationModerate evidence
A colour matched directly from an exposed early paint layer is likely to reproduce that layer's aged appearance — often yellower and darker for oil paints — rather than its original colour.
Based on: Follows from the documented discoloration of aged paint layers that makes scrape-and-match unreliable, combined with the well-known yellowing of linseed oil binders kept out of light.
Reviewed 29 September 2026. Colourwise summarises its sources in its own words and does not reproduce standards text or proprietary colour data. Spotted an error? Tell us.