Two clays can look alike on site and behave very differently once wet. The Atterberg limits (the liquid and plastic limits) and the plasticity index derived from them are the simplest laboratory measure of that difference. They are used to classify fine soils, flag swell potential and check fill against a specification. Below we explain how we test them in our soil laboratory and how to read them. Our technical lead’s doctoral research was on the statistical modelling of the plasticity of Cypriot clayey soils.

What LL, PL and PI mean

As it dries, a fine soil passes from a liquid to a plastic state, in which it can be moulded without cracking, and then becomes semi-solid and solid. The Atterberg limits are the water contents at these boundaries.

Term	Definition	What it tells you
Liquid limit (LL)	Water content at the change from plastic to liquid	How much water the fines hold before they flow
Plastic limit (PL)	Water content below which a rolled thread crumbles	Lower boundary of plastic behaviour
Plasticity index, PI = LL − PL	Range of water content over which the soil is plastic	Main classification value
Consistency index, Ic = (LL − w)/PI	Position of the natural water content w between the limits	Ic ≥ 1: w at or below PL, stiff; Ic near 0: w near LL, very soft
Liquidity index, IL = (w − PL)/PI	The same position, measured from PL	IL = 1 − Ic

TS EN ISO 14688-2:2018 defines both indices in this form. They rely on the natural water content, so the sample must reach us sealed and undried; our guide to Proctor, CBR and water content tests explains how to bag it.

How we determine the limits

We follow TS EN ISO 17892-12:2018+A2:2022, testing the fraction finer than about 0.4 mm (ASTM D4318-26 uses the 425 µm sieve):

Liquid limit: the standard allows the fall cone method, where a standard cone sinks a set depth into the soil paste, or the Casagrande method, where a groove cut in the soil in a brass cup closes after 25 blows. We use the Casagrande cup and read LL from several points over a range of water contents.
Plastic limit: a soil thread is rolled by hand on a glass plate until it crumbles at about 3 mm diameter; the determination is repeated and averaged.
Report: LL, PL, PI, the method, the preparation and the percentage finer than 0.4 mm.

The two methods do not give identical results, and the standard itself acknowledges operator judgement and some variability.

Sample preparation: why drying matters

Where possible we start from the natural water content, soaking and mixing the soil until uniform. Drying first is quicker but can change the limits of some soils: ASTM D4318-26 makes wet preparation the default, and ASTM warns that oven-drying lowers the liquid limit of organic soils dramatically. Two Cyprus materials need particular care:

Carbonate-rich marls: break lumps down by soaking and gentle working; hard grinding can crush weak carbonate grains into artificial fines.
Weathered or mixed material: marl, clay bands and cemented fragments in one bag give limits that belong to no layer. Sample and bag each layer separately.

The plasticity chart and classification

A fine soil is named by plotting PI against LL. The A-line, PI = 0.73 (LL − 20), separates clays above it from silts below it. TS EN ISO 14688-2:2018 classifies fine soils by plasticity, and specifications often use the Unified Soil Classification System of ASTM D2487-17(2025), which splits the chart at LL = 50:

Symbol	Position on the chart	Soil
CL	LL < 50, PI > 7, on or above the A-line	Lean clay
CL-ML	PI 4 to 7, on or above the A-line	Silty clay
ML	LL < 50, PI < 4 or below the A-line	Silt
CH	LL ≥ 50, on or above the A-line	Fat clay
MH	LL ≥ 50, below the A-line	Elastic silt

Grading comes first: plasticity names the fines, but their amount decides whether the soil is fine- or coarse-grained. In the USCS a soil is fine-grained when 50% or more passes 75 µm; a sand or gravel with over 12% fines takes the name of its fines, such as SC (clayey sand), and 5–12% gives a dual symbol. ISO sets the fines boundary at 0.063 mm. We wash clayey soils when grading to TS EN ISO 17892-4:2016, because dry sieving leaves clay on coarse grains.

Plasticity and swelling: indicators, not measurements

Smectite-rich clays combine high LL and PI with large volume changes; see Ground in Cyprus: marl and expansive clay. Classic references relate index values to swell:

Holtz and Gibbs (1956, Transactions of the ASCE, vol. 121) linked expansion to colloid content, PI and shrinkage limit: PI below 18 low, 15–28 medium, 25–41 high, above 35 very high. The ranges overlap; PI alone does not decide the class.
Seed, Woodward and Lundgren (1962, Journal of the Soil Mechanics and Foundations Division, ASCE, vol. 88) related the swell of clays compacted at standard Proctor optimum, under a 1 psi (6.9 kPa) surcharge, to clay content and activity.
Linear shrinkage to BS 1377-2:2022 measures how much a soil bar prepared near its liquid limit shortens on drying; we often add it for clays.
The limits describe only the fraction below 0.4 mm, so UK housing guidance (NHBC Standards, Chapter 4.2) multiplies PI by the proportion passing 425 µm, giving a ‘modified plasticity index’.

These come from particular soils and test conditions; none is universally valid, and none measures swell or swelling pressure. For compacted fill, the swell recorded in the soaked CBR test (TS EN 13286-47:2021) is a direct measurement; for natural clay below foundations, the geotechnical engineer decides whether swell tests are needed.

Cypriot clays vary: report ranges, not one value

Clayey and marly units in Cyprus change over short distances in clay and carbonate content and in weathering, so one liquid limit describes a sample, not a layer. We recommend several tests per layer, from different points and depths, and report:

every result with its depth and description,
per layer: number of tests, minimum, maximum, mean and, where numbers allow, standard deviation,
all results on one plasticity chart, so that separate layers and outliers stand out; outliers are checked and retested where material remains.

Two statistical points matter. PI is the difference of two measurements, so its relative uncertainty exceeds that of LL. And published correlations, regional ones included, hold only for the soils and data range they came from. Choosing a characteristic or design value is the project geotechnical engineer’s judgement; our report gives them the full data set.

NP results, methylene blue and specification limits

If the plastic limit cannot be determined, or equals or exceeds the liquid limit, the soil is reported as non-plastic (NP), as in ASTM D4318-26. NP is a valid result, common for sandy and crushed-rock fines.

NP fines can still give a high methylene blue value (TS EN 933-9:2022; see Aggregate testing in Cyprus: grading and LA). The limits show how the fines behave with water as a whole, while methylene blue, on the 0/2 mm fraction, responds to clay minerals even in small amounts. A high MB value with NP fines therefore suggests some clay, for example as grain coatings or rock infill, and a fresh sample from elsewhere in the source is worth testing.

Limits on LL or PI for fill, sub-base and base come from the project specification, not the test standard. Tell us which applies and we will report against it.

How we can help

We run Atterberg limits, linear shrinkage, water content, grading and classification in our soil laboratory, on samples you bring or that we take during site investigation and geotechnical drilling. Grading plus Atterberg limits usually needs 3–5 kg of fine soil, or 10–15 kg if gravelly. Send the specification through our contact page and we will agree the programme.

Frequently asked questions

Is a high plasticity index a bad result?

Not in itself; it depends on the use. Under light buildings and slabs, and in fill and sub-base, high plasticity usually calls for attention. Acceptance is set by the specification and the project engineer.

Why did two laboratories report different liquid limits?

Compare the method (fall cone or Casagrande) and the preparation (natural or dried) on both reports first. Sample differences and operator judgement also play a part.

Can the plasticity index tell me how much the ground will swell?

No. PI, linear shrinkage and fines content indicate potential. Actual swell also depends on the water content at construction, the load and site water conditions, and must be measured if the engineer needs it.