Bearing Capacity, and How a Load Spreads Into What Carries It

The division, and the thing it is compared against

Bearing is load divided by contact area, and the contact area is normally the unknown being sized. That much is arithmetic. Everything that matters is in what the resulting pressure is compared against, and that limit comes from whatever is softer — the ground under a footing, the timber under a beam end, the grout under a baseplate.

The comparison has two distinct forms and they are not interchangeable. A STRENGTH limit asks whether the material will fail: the soil will shear, the timber will crush, the grout will spall. A SERVICEABILITY limit asks whether the movement before failure is acceptable. Both produce a pressure, and for most foundations on most soils the serviceability limit is the smaller of the two by a wide margin.

Three terms, and which of them matters depends on the soil

Terzaghi's bearing capacity expression is the sum of three contributions: one from the soil's cohesion, one from the weight of soil surcharging the footing at its founding depth, and one from the width of the footing itself. Each carries its own bearing capacity factor, and all three factors depend on the friction angle.

Which term dominates tells you what to do about a foundation that does not work. On a clay with meaningful cohesion and little friction, the cohesion term carries nearly everything and widening the footing buys almost nothing — the answer is depth, or a different foundation. On a clean sand with no cohesion, the cohesion term is zero and the other two carry the load, so BOTH going deeper and going wider help.

The factors rise steeply with friction angle — steeply enough that a few degrees of optimism in the soil parameters moves the ultimate capacity by tens of percent. This is the same sensitivity that makes the friction angle the most consequential input on the lateral-pressure pages, and for the same reason: it appears inside an exponential rather than as a multiplier.

A wider footing settles MORE at the same pressure

This is the result that surprises people and it follows from how stress spreads. A footing does not load a thin slice of soil; it loads a bulb extending roughly one and a half to two times its own width below it. A footing twice as wide therefore stresses soil twice as deep.

So two footings carrying the same PRESSURE — not the same load, the same pressure — do not settle the same. The wider one compresses a thicker layer of soil and settles more. Sizing a large footing to the same allowable pressure that worked for a small one on the same site is a standard way to produce differential settlement between a column and a wall.

It also explains why site investigation depth is a function of foundation size. A borehole taken to two metres tells you about a strip footing and nothing about a raft, because the raft's stress bulb reaches soil that was never sampled. The classic failure is a firm crust over soft clay: a small footing sits happily on the crust while a large one punches its stress bulb straight through it into material nobody looked at.

Bearing length in timber, masonry and steel

The same division governs a beam end, and the limit changes character. Timber loaded PERPENDICULAR to the grain crushes at a small fraction of its parallel-to-grain strength, because the load is squashing the cell structure sideways rather than compressing it along its length. A bearing length that satisfies the beam's bending design can still crush the plate it sits on.

Masonry has its own version: a concentrated load on a wall spreads through the masonry at a rate the codes define, so the bearing stress depends on the length of bearing AND on how far the load has dispersed by the time it reaches the section being checked. Too short a bearing concentrates it; too close to an end, and there is no masonry on one side to disperse into.

Steel baseplates sit on grout and the grout sits on concrete, so there are two bearing checks in series with different allowable pressures — and the concrete's is increased where the plate is small relative to the pedestal, because the surrounding concrete confines the loaded patch. Every one of these pages says which interface it is checking, since the answer differs at each.

Where it fails: eccentricity, layers and the load that is not central

The division assumes the load acts at the centre of the contact area, and eccentric loads do not. The standard treatment reduces the footing to an EFFECTIVE AREA centred on the load — Meyerhof's method — which is a smaller area than the real one, so the pressure is higher than load over gross area suggests. Where the eccentricity exceeds a sixth of the width, part of the base lifts off entirely and the pressure over what remains rises sharply.

Layered ground breaks the closed-form solutions outright. A strong layer over a weak one can punch through; a weak layer over a strong one may be fine at small widths and fail at large ones. No single bearing capacity figure describes a layered profile, and a site investigation that stopped at the first firm stratum has not found out.

Crane mats and scaffold baseplates add a further case: the load is temporary, mobile and applied to ground that was never prepared as a foundation. The pressure calculation is identical; the allowable value is not, because the material is whatever happens to be there and may be fill, a service trench backfill, or a basement slab of unknown thickness.

The alternatives: test it, or design to settlement directly

The most direct alternative is a PLATE LOAD TEST — load a plate on the actual ground and measure what it does. Its limitation is exactly the stress-bulb argument above: a small plate tests a shallow bulb, so its result cannot simply be scaled to a large footing on a layered site.

In practice most shallow foundations on granular soils are designed from in-situ test correlations — SPT blow counts or CPT cone resistance — against SETTLEMENT rather than against capacity, precisely because settlement governs. That inverts the logic of this page: rather than computing an ultimate capacity and dividing by a factor, the design picks a tolerable settlement and finds the pressure that produces it.

The pages here compute pressures and capacities in closed form, which is the right tool for checking a proportion or sizing a first pass. Where the ground is layered, the loads are eccentric, or the structure is sensitive to differential movement, the answer is a geotechnical report and not a calculator — and every one of these pages says so.