The Cantilever Bridge

Sit two people on chairs facing each other, a gap between them. Each holds a stick out towards the other with the outstretched arm, and anchors a second stick down to a weight on the floor beside the chair.

Sit a third, lighter person on a board slung between the two inner stick ends.

Nobody is holding a beam across the gap, yet the load is carried.

With the load in place, have each seated person report what they feel. The outstretched arm and its stick are being squeezed. The anchor stick to the counterweight is being pulled.

Write the sign of every member on a sketch.

The top members of a cantilever arm are in tension and the bottom in compression — the exact reverse of a simply supported beam, and the reason cantilever bridges look inside-out.

Lay one continuous stiff strip across three supports, each support standing on its own kitchen scale. Zero the scales, then place a known load at mid-span of one bay.

Record all three readings and check they sum to the applied load.

You cannot get these three numbers from equilibrium alone — two equations, three unknowns. The beam's own stiffness decided the split.

Slip a thin shim under the centre support, or take one away, changing its height by one or two millimetres. Do not touch the load.

Read all three scales again.

Expect them to change substantially.

Nothing was added and nothing was removed. The structure loaded itself, purely because a support moved — and on a real bridge in a real riverbed, supports move.

Replace the continuous strip with the Gerber arrangement: two anchored arms, and a separate short suspended span resting freely on the arm ends. Load it in the same place.

Now shim the centre support again and re-read the scales.

Expect the readings to be essentially unchanged.

That is the whole patent in one measurement. With the hinges in, the reactions follow from equilibrium alone and settlement cannot induce force.

Erect the model in the real construction sequence. Add one bay to the left arm, then one to the right, alternating, so the pier is never loaded unevenly. Place the suspended span only at the very end.

Try it wrong: build one arm three bays out before starting the other.

Expect the unbalanced version to tip or need holding.

Balance is not an aesthetic; it is the construction method. Nothing ever touched the floor between the piers.

Gerber's patent is a mathematical device, not a shape. Balkenträger mit freiliegenden Stützpunkten — girder with free-standing support points — was granted in Bavaria in 1866 and first built in 1867 at Bamberg over the Regnitz and at Haßfurt over the Main. Before computers, a statically indeterminate structure had to be solved by iterative hand methods that were slow, error-prone and dependent on assumed stiffnesses. Gerber's hinges turned an unsolvable-in-practice problem into arithmetic, and made the answer independent of how much the ground moved. Both benefits mattered; the second one still does.

The Forth Bridge was built in the shadow of a disaster. The Tay Bridge collapsed in a gale on 28 December 1879, taking a train and its passengers with it. Thomas Bouch had already been commissioned to build a suspension bridge over the Forth; that design was abandoned. John Fowler and Benjamin Baker replaced it with a cantilever in steel — visibly, demonstratively, conservatively massive, with wind loading taken far more seriously than Bouch had taken it. It opened on 4 March 1890 with two main spans of 1,709 ft and a total length of 8,296 ft, and is a UNESCO World Heritage Site.

The photograph in step 1 is real and it is 1887. Baker and Fowler staged the living-model demonstration to explain the principle publicly, with two men on chairs holding sticks and bricks — and the man sitting in the middle, playing the suspended span, was Kaichi Watanabe, one of the first Japanese engineering students in Britain. He is almost always cropped out or left unnamed in the retellings. He went home and worked on Japan's railways.

Cantilever construction outlived the cantilever bridge. Steel cantilever trusses fell out of favour — they are heavy, they have enormous numbers of joints to inspect and paint, and the 1907 Quebec Bridge collapse, which killed 75 men, was a cantilever failing during erection from an underestimated dead load. But balanced cantilever construction is now completely standard in prestressed concrete: segments are cast or placed outward from each pier in matched pairs, exactly as in step 6, and the deck closes at mid-span. The bridge type became rare; the method of building without falsework became universal.

The honest limits. A cantilever arm is at its most highly stressed during construction, before the suspended span closes and before the far anchor is complete — which is when Quebec failed, twice. The anchor arms must be held down, so the end piers work in uplift and are often ballasted with concrete or kentledge. Steel cantilevers carry a large permanent maintenance cost in joints and paint. And the hinges that make the mathematics easy are also mechanical bearings: they move, they wear, they need inspecting, and a seized hinge quietly turns the structure back into the indeterminate one Gerber was avoiding.