Reading a Soil's Atterberg Limits Like a Geotechnical Engineer

Four numbers, one story about soil behavior

A fine-grained soil doesn't have a single "strength" the way steel does. Depending on how much water is mixed into it, the exact same clay can behave like a brittle solid, a moldable putty, or a viscous liquid you can't even stand on. Atterberg limits are the boundary water contents where a soil crosses from one of those states into the next, and once you have four of them — Liquid limit LL, Plastic limit PL, Shrinkage limit SL, and the soil's actual Natural water content w — you can predict almost everything a geotechnical engineer needs to know before specifying a foundation, a slope, or a fill.

This isn't a laboratory curiosity. A soil sample that looks identical to the eye can be a well-behaved, low-shrink-swell clay or a soil that heaves foundations every wet season, and the only way to tell the difference reliably is to run these four numbers through the classification math.

The definitions, stated precisely

Liquid limit (LL) is the water content at which the soil transitions from plastic to liquid behavior — above it, the soil flows under its own weight. Plastic limit (PL) is the water content at which it transitions from plastic to semi-solid — below it, the soil crumbles instead of deforming when rolled into a thread. Shrinkage limit (SL) is the water content below which further drying causes no further volume change — the soil has fully shrunk.

From LL and PL alone you get the single most useful derived number in geotechnical classification, the Plasticity index PI (%):

PI = LL - PL

PI measures the width of the plastic range — how much water content a soil can tolerate while staying moldable rather than flowing or crumbling. A wide plastic range (high PI) generally means a soil that's more clay-dominated and more prone to volume change; a narrow one (low PI) leans toward silt-like behavior.

A worked example: classifying a clay from a boring log

Say a boring log reports these four values for a sample recovered a few meters below grade:

LL = 45%
PL = 22%
w  = 30%   (natural water content, in situ)
SL = 15%

First, the plasticity index:

PI = LL - PL = 45 - 22 = 23%

A PI of 23% already tells you this is a moderately plastic material — well above the roughly 7% threshold where soils are considered non-plastic to slightly plastic, but not into the very-high-plasticity range that starts causing serious shrink-swell problems.

Next, where does the in-situ soil actually sit within that 23-point plastic range? That's what Liquidity index LI and Consistency index CI are for:

LI = (w - PL) / PI = (30 - 22) / 23 = 8 / 23 = 0.348
CI = (LL - w) / PI = (45 - 30) / 23 = 15 / 23 = 0.652

Notice LI + CI = 1.000 exactly — that's not a coincidence, it's algebraically guaranteed, and it's a useful sanity check whenever you're computing these by hand. LI = 0.348 places the soil roughly a third of the way from the plastic limit toward the liquid limit: firm to medium-stiff plastic behavior, not soft, not brittle. The Natural-state verdict for a sample like this reads as "plastic, firm consistency" rather than "near-liquid" (LI approaching 1) or "over-consolidated/brittle" (LI below 0, which would mean w is actually below PL).

Where the soil sits today: liquidity and consistency index

The distinction between LI and CI is worth being precise about, because they're mirror images that get confused constantly. CI describes how far the soil is from its liquid limit — high CI (close to 1) means the soil is dry relative to its plastic range and behaves stiffly. LI describes how far the soil is from its plastic limit — high LI (close to 1) means the soil is wet relative to its plastic range and behaves softly, with LI above 1.0 flagging a sensitive or quick soil that can lose most of its strength if disturbed. Since they always sum to one, you only ever need to compute one directly and the other falls out for free — but reporting both, as this sample's tool output does, avoids the mental subtraction error of assuming "low CI means low LI," which is backwards.

The A-line and why it matters

LL and PI together, plotted on the Casagrande plasticity chart, are what actually assign a USCS symbol — CL, CH, ML, MH — and the line that separates clays from silts on that chart is the empirical A-line:

PI_A-line = 0.73 * (LL - 20)

For this sample:

A-line PI (%) = 0.73 * (45 - 20) = 0.73 * 25 = 18.25%

The actual PI is 23%, which sits above the A-line value of 18.25% — soils that plot above the A-line classify as clays (C), while soils that plot below it classify as silts (M), regardless of how "clayey" the sample looks to the naked eye. Combined with LL = 45% being under the 50% threshold that separates low from high plasticity, this sample's Plasticity-chart class comes out as CL: a low-to-medium plasticity clay. Push the same LL past 50% with the same relative PI and the classification jumps to CH — a high-plasticity clay with materially worse shrink-swell behavior, which is exactly why the 50% LL threshold gets so much attention in foundation design.

It's worth sitting with how much the classification result depends on exactly where the point falls relative to the A-line, not just on which "family" the soil looks like. Two samples with nearly identical PI can land on opposite sides of the line depending on LL alone, and that single symbol difference — CL versus ML — changes the entire downstream engineering treatment: allowable bearing pressure assumptions, compaction specifications, frost susceptibility expectations.

Common mistakes in the field

The most frequent classification error isn't in the Atterberg testing itself, it's forgetting to check the Shrinkage limit SL against the natural water content before making a swelling-risk call. In this example, w = 30% is far above SL = 15%, meaning the soil is nowhere near its fully-shrunk state — there's real room for further volume loss if it dries out, which is exactly the geometry that damages slab foundations during a dry season. A soil report that only quotes LL, PL, and PI without checking SL against current field moisture is missing half of the shrink-swell risk picture.

The second common mistake is trusting a single Atterberg test from one boring across an entire site. Fine-grained soil plasticity can vary meaningfully over short lateral distances, and a Plasticity-chart class determined from one sample shouldn't be extrapolated across a large footprint without additional borings, especially near any observed change in soil color or texture in the log.

Try it yourself

Running LL, PL, w, and SL through the full set of derived indices by hand is easy to get right once and easy to get subtly wrong under time pressure — a swapped LI/CI formula or a mis-plotted A-line point produces a classification that looks plausible but is wrong. You can enter your own boring-log values and get PI, LI, CI, the A-line PI, the resulting Plasticity-chart class, and the Natural-state verdict instantly in the Atterberg Limits & Plasticity Index Simulator.

For related subsurface calculations, the site also has an aquifer pump test tool based on the Theis solution and a Winkler foundation beam calculator that both build on the same fine-grained soil characterization.