Bridging wide bodies of water has always been one of the most ambitious tasks in engineering. The longest bridges in the world attract public attention in part because they look improbable - narrow roads suspended above huge volumes of wind and water - and in part because the engineering decisions behind them have to confront problems that do not arise at smaller scales. Understanding why these bridges take the forms they do means understanding how structural options change with length.

For spans of up to about fifty metres, the simple beam bridge is almost always adequate. A horizontal deck rests on piers, and loads are carried in bending. Beyond fifty metres, bending becomes inefficient, and engineers turn to arches, trusses or cable-stayed designs. An arch transfers load into horizontal thrust at its ends, which the surrounding ground must absorb. A cable-stayed bridge hangs its deck from straight cables leading to one or more towers, with all forces meeting in the tower's base. Both designs can span several hundred metres.

For the longest spans - a kilometre or more - the dominant type is the suspension bridge. A main cable runs from anchor to anchor, passing over the tops of two towers, and drops vertical suspenders down to the deck below. Unlike a cable-stayed design, a suspension bridge transfers most of its load through tension along the main cable into massive concrete or rock anchors at either end. The Akashi Kaikyo Bridge in Japan, completed in 1998, has a central span of 1,991 metres and uses this principle. Its main cables, each about a metre in diameter, are composed of more than thirty-five thousand individual steel wires.

Every long suspension bridge has to deal with the behaviour of air in motion. Wind passing around a flat deck produces alternating vortices above and below, which can shake the deck at a frequency that matches its natural oscillation. The famous collapse of the Tacoma Narrows Bridge in 1940 was a consequence of this effect, now called aeroelastic flutter. Modern decks are designed with box-like cross-sections that disrupt vortex formation, and wind-tunnel testing of scale models is a standard part of design. Dr. Kenji Mori, a senior engineer at the Honshu-Shikoku Bridge Authority, has described the wind analysis for the Akashi Kaikyo Bridge as more demanding than the static load analysis.

Longer spans are theoretically possible but economically daunting. The Messina Strait, between Sicily and the Italian mainland, has been studied for decades as the site of a potential bridge more than three kilometres in a single span. Engineering designs exist, but the cost, estimated at more than ten billion euros, and concerns about seismic activity in the region have prevented construction. A separate proposal for a bridge across the Bering Strait has received attention but remains speculative.

A different way of dealing with wide waters is to combine a bridge with one or more tunnels. The Oresund link between Denmark and Sweden, opened in 2000, includes a bridge, an artificial island and an underwater tunnel. The tunnel section allows aircraft to approach Copenhagen airport without clearance problems, while the bridge section takes the road and rail crossing over a strait where a full-length bridge would have been too tall. Hybrid designs of this kind are becoming common, especially where shipping lanes or airports impose constraints that no single structure can satisfy.

Maintenance is a hidden element of every long bridge. Steel cables, exposed to salt air, lose strength unless they are continuously protected, and inspections often require specialised equipment. Dr. Mori notes that the cost of maintaining a long suspension bridge over its first fifty years is typically larger than the cost of its original construction, a fact that is rarely emphasised in public debates about whether a new bridge should be built. Failure of maintenance on an older bridge can be dramatic: the Morandi Bridge in Genoa, Italy, collapsed in 2018 after a cable-stayed segment gave way, killing more than forty people. Subsequent investigations pointed to corrosion in pre-stressed concrete that had been known about for years but not repaired.

The engineering of the longest bridges, then, is a balance between what can be built, what can be paid for, and what can be kept working through decades of exposure to wind, water and use. Their apparent simplicity, when seen from the shore, hides a web of decisions whose consequences do not fully appear for many years.