A truss is a pin-jointed collection of two-force members
A truss bridge is modelled as straight members connected at joints that are idealised as frictionless pins, with every load applied only at the joints, never along a member's length. Under that idealisation each member carries force along its own axis only — it is a two-force member, either being stretched (tension) or squeezed (compression), and never bent. This is the entire reason trusses are efficient: a straight bar loaded purely along its axis uses its full cross-section, while a beam loaded sideways only uses the material near its top and bottom fibres.
Why triangles, specifically
Fix the lengths of a triangle's three sides and its shape is fixed — there is no way to flex it without changing a side length, and pinned members cannot change length. A quadrilateral with pinned corners has no such guarantee: it can rack into a parallelogram, sliding sideways under load while every member length stays exactly the same, which means the members feel no force at all and the structure simply collapses sideways. Every rigid panel of a truss must therefore be triangulated, which is why you never see an untied square panel in a real bridge.
Solving a truss: the method of joints
Because every joint is in static equilibrium, the sum of forces at each pin is zero in both x and y. Starting from a joint with only two unknown member forces (usually a support or an end), the method of joints solves those two unknowns from the two equilibrium equations, then moves to the next joint that now has only two unknowns left, working across the whole structure:
at every joint: ΣFx = 0 ΣFy = 0 determinacy check: m + r = 2j → statically determinate (solvable by statics alone) m + r < 2j → mechanism, can collapse m + r > 2j → statically indeterminate, needs extra equations m = members, r = reaction components, j = joints
A quicker alternative for a single member deep inside a large truss is the method of sections: cut the truss with an imaginary line through no more than three unknown members, treat one side as a free body, and take moments about a point that eliminates two of the three unknowns at once — one equation, one answer, no need to march joint by joint from the support.
Reading the diagram: which members go into tension
Under a downward load in the middle of a simply supported truss, the bottom chord stretches (tension, like the bottom of a bent beam) and the top chord compresses (like the top of a bent beam). The diagonal web members alternate depending on which way they lean relative to the load path — a diagonal that would shorten if the truss sagged is in compression, one that would lengthen is in tension. Engineers choose slender cables or thin rods for tension members, since tension does not care about buckling, and thicker, stockier sections for compression members, because a slender bar in compression fails by buckling sideways long before the material itself crushes — this is Euler's buckling criterion, and it is why compression chords are visibly fatter than tension chords on a real bridge.
Truss families and why the shape changes with span
The Pratt truss puts diagonals in tension under typical loading and is cheap to build with slender diagonals; the Warren truss uses a repeating equilateral-triangle pattern with no verticals, minimising the total member count; the Howe truss is a Pratt truss upside down, putting diagonals into compression, which made sense in the 19th century when timber compression members were cheaper than iron tension rods. Longer spans add a curved top chord (a Parker or bowstring truss) to keep member forces more uniform along the span, since a straight top chord over a long simply-supported span would otherwise force the end panels to carry drastically less force than the middle ones.
Frequently asked questions
Why are triangles used instead of squares in a truss?
A triangle is the only polygon whose shape is fixed by the length of its sides alone — with pinned joints it cannot deform without a member changing length. A square with pinned corners can rack into a parallelogram under load with zero member-length change, so it carries no load unless it is braced into triangles.
How do you know if a member is in tension or compression?
Solve the joint equilibrium equations assuming every member pulls (tension). A positive result confirms tension; a negative result means the member is actually being pushed, i.e. in compression. Visually, members that sag toward the load path are usually tension, members that arch away from it are usually compression, but always check the numbers.
Is the method of joints always enough to solve a truss?
Only for a statically determinate truss, where members = 2·joints − 3. Fewer members and the structure is a mechanism that can collapse; more members and it is statically indeterminate, needing compatibility equations (or a stiffness-method solver) in addition to equilibrium.
Try it live
Everything above runs in your browser — open Bridge Design and change the parameters while it is running. Nothing is installed, nothing is uploaded, the whole model lives in one tab.
▶ Open Bridge Design simulation