Maximum shear stress theory predicts that a ductile material will begin to yield when the largest shear stress inside it reaches the same value that causes yielding in a simple tension test. First formalized by the French engineer Henri Tresca in the 1860s, it remains one of the two most widely taught yield criteria in mechanical and civil engineering. The theory is straightforward, often conservative, and works well for metals and other ductile solids, though it has quirks and limitations that matter in both design and computation.
How the Theory Works
Every loaded object experiences stresses in multiple directions at once. At any point inside that object you can identify three “principal stresses,” which are the normal stresses acting along three mutually perpendicular axes where shear happens to be zero. Maximum shear stress theory says yielding starts when half the difference between the largest and smallest of those principal stresses equals the shear stress at which the material yields in a uniaxial tension test. In a simple tension bar, the maximum shear stress turns out to be exactly half the applied tensile stress, so the criterion boils down to a comparison: if the biggest gap between any two principal stresses reaches the tensile yield strength of the material, permanent deformation begins.
This makes the theory easy to apply in practice. You work out the principal stresses at the most highly loaded point in your part, subtract the smallest from the largest, and compare the result to the material’s known yield strength. If the difference meets or exceeds that yield strength, you predict failure. Because the criterion ignores the intermediate principal stress entirely, it tends to give a slightly pessimistic answer, predicting yielding a bit earlier than some experiments show. For safety-critical design, that built-in conservatism is often considered a feature rather than a flaw.
Tresca’s Experiments and the Birth of the Criterion
The theory carries Henri-Édouard Tresca’s name because he was the first person to articulate both the concept and the experimental evidence behind it. Starting in 1864, Tresca submitted a series of memoirs to the French Academy of Sciences describing experiments on punching, rolling, forging, stamping, and extruding various metals. He observed that solids under very high pressures behaved as though they could flow, a phenomenon he called the “fluidity” of solids. Out of those observations he proposed a mechanical theory that introduced, for the first time, what we now call a yield criterion and a plastic flow rule.1Comptes Rendus. Mécanique. About Tresca’s Memoirs on the fluidity of solids (1864–1870)
What made Tresca’s work groundbreaking was not just the data but the leap from observation to generalized rule. Before his memoirs, engineers knew that metals could be permanently deformed, but there was no unified criterion telling you when that deformation would start under an arbitrary combination of loads. Tresca’s answer, that the onset of yielding is governed by the maximum shear stress, gave designers a single number to check. The simplicity of that idea is a big reason the criterion survived the arrival of more sophisticated alternatives.
Why the Theory Fits Ductile Materials
Ductile materials, such as mild steel, aluminum, copper, and most structural alloys, fail by slipping along internal planes. At the atomic level, rows of atoms slide past one another through a mechanism driven by defects in the crystal lattice called dislocations. This slip is fundamentally a shearing process: atoms shift sideways relative to their neighbors. Because yielding in ductile metals is governed by shear at the microscopic level, a macroscopic criterion based on maximum shear stress captures the physics reasonably well.2Progress in Materials Science. Dislocation modelling of shear in f.c.c. crystals
The fit is not perfect for every ductile material, though. Real metals show a range of ratios between their yield shear stress and their yield tensile stress. In an idealized Tresca world, that ratio would be exactly 0.5, because the maximum shear stress in uniaxial tension is half the applied stress. Measured values for actual ductile materials fall in a band of roughly 0.42 to 0.75, which means the simple Tresca criterion is a good approximation for many metals but can underpredict the shear strength of materials at the higher end of that range.3Applied Mechanics and Materials. The Modified Maximum Shear Stress Failure Theory of Ductile Material Modified versions of the theory have been proposed to fit the full spread of ductile behavior across all loading directions, adjusting the failure boundary so it can accommodate materials whose shear-to-tensile yield ratio strays from the textbook 0.5 value.
Brittle materials, by contrast, tend to fail by cracking rather than shearing. Their tensile strength is often much lower than their compressive strength, a trait called the strength-differential effect. Maximum shear stress theory does not account for that asymmetry and is generally not applied to ceramics, concrete, cast iron, or rock in their brittle regimes. Other criteria, like Mohr-Coulomb, were developed specifically to handle materials where tension and compression behave very differently.4Materials & Design. Mohr-Coulomb criterion with circular failure envelope, extended to materials with strength-differential effect
Tresca Versus Von Mises
Almost every engineering textbook pairs the Tresca criterion with its main rival, the von Mises criterion. Where Tresca checks only the largest shear stress, von Mises combines all three principal stress differences into a single equivalent stress. The two criteria agree exactly under uniaxial tension and under pure shear, but they diverge under more complex loading states. Tresca’s yield boundary, when you plot it in principal stress space, forms a hexagon. Von Mises’ boundary is an ellipse that circumscribes that hexagon. Because the Tresca hexagon fits inside the von Mises ellipse, there are load combinations where Tresca predicts yielding but von Mises does not. This is why Tresca is conventionally described as the more conservative of the two: it predicts failure at equal or lower loads.
The practical difference between the two predictions is modest. At most, the von Mises equivalent stress exceeds the Tresca equivalent stress by about 15 percent, which occurs under a specific equibiaxial stress state. For many real-world loading scenarios the two criteria give answers within a few percent of each other, which is why some codes of practice allow engineers to use whichever is more convenient. Von Mises tends to be preferred in computational work because its smooth elliptical surface is easier to handle mathematically, while Tresca’s hexagonal corners create numerical headaches.
When Von Mises Can Actually Be More Conservative
The textbook statement that Tresca is always more conservative than von Mises holds true in a deterministic sense: for any single, perfectly known stress state, the Tresca equivalent stress is always at least as large as the von Mises equivalent stress. But real engineering involves uncertainty. Material properties vary from batch to batch, loads fluctuate, and geometric tolerances introduce stress variations that a single calculation does not capture.
When you account for that variability, the picture changes. A probabilistic analysis shows that there is a non-negligible chance the von Mises criterion actually yields a smaller equivalent stress than Tresca for a given set of uncertain inputs, making von Mises the more conservative prediction in that particular draw from the probability distribution.5Journal of Applied Mechanics. Tresca Versus Von Mises: Which Failure Criterion is More Conservative in a Probabilistic Context? This is not a contradiction of the deterministic math; it arises because the two criteria weight the principal stresses differently, so random scatter in those stresses can swap which criterion gives the larger answer on any individual sample. The takeaway for engineers is that simply saying “Tresca is always safe-side” is an oversimplification when real scatter in loads or material properties is present. For safety-critical probabilistic design, checking both criteria is the cautious path.
Insensitivity to Hydrostatic Pressure
One important property of maximum shear stress theory is that it does not respond to hydrostatic pressure, meaning a uniform squeeze or pull applied equally in all three directions. If you submerge a block of steel at the bottom of the ocean, the enormous hydrostatic pressure compresses it uniformly but does not create any shear stress at all, because shear stress depends on the difference between principal stresses. Since hydrostatic pressure adds the same amount to every principal stress, those differences stay unchanged and the Tresca criterion remains unaffected.
This pressure-insensitivity has been formally demonstrated for all yield criteria developed for standard ductile materials, meaning those that are homogeneous, isotropic, and lack a strength-differential effect.6The Romanian Journal of Technical Sciences. Applied Mechanics. Considerations concerning yield criteria insensitive to hydrostatic pressure For most metals under ordinary conditions, the assumption holds well. It breaks down in materials like soils, rocks, and some polymers, where compressive mean stress genuinely changes the shear strength. That is a key reason geotechnical engineers and rock mechanics specialists reach for pressure-sensitive criteria like Mohr-Coulomb or Drucker-Prager, even though they may still use the Tresca criterion in specific circumstances.
Tresca in Geotechnical Engineering
Despite being developed for metals, the Tresca criterion has a second life in soil mechanics. Clay soils loaded under undrained conditions, where water trapped in the pore spaces cannot escape quickly enough, behave as if their shear strength is independent of the applied confining pressure, at least over the timescale of the loading event. That makes the Tresca criterion a natural fit: the failure load depends only on the maximum shear stress, not on how much overburden pressure the soil is under.7Computers and Geotechnics. Flow rule effects in the Tresca model
The criterion is used regularly to compute bearing capacity of foundations, slope stability in cuttings and embankments, and short-term collapse loads for excavations, all situations where clays are loaded fast enough that the undrained assumption is reasonable. Choosing the right flow rule to pair with the Tresca yield surface matters, though. Different assumptions about how plastic strains develop once the soil starts yielding can change the computed failure load, even though the yield criterion itself stays the same.8Computers and Geotechnics. Flow rule effects in the Tresca model Engineers working with finite-element models of clay need to be aware that the flow rule choice is not just a numerical convenience; it can shift the answer by a meaningful margin.
Experimental Evidence Under Biaxial Loading
One of the sharper tests of any yield criterion is how well it predicts behavior when a material is stretched or compressed in two directions at once. Biaxial loading creates stress states that sit between the corners of the Tresca hexagon, right where the hexagonal and elliptical yield surfaces diverge most. Experiments on expanding shells of Ti-6Al-4V, a titanium alloy widely used in aerospace and biomedical applications, found that plastic instabilities under biaxial stretching were better reproduced by a Tresca yield surface than by a von Mises one.9EPJ Web of Conferences. Tresca criterion for plasticity and viscoplasticity: Application to localization in biaxial loading conditions
In those experiments, shear bands, which are narrow zones of intense deformation, developed at times that matched predictions from a Tresca-based elastoplastic model. A von Mises model, by comparison, smoothed over the localization and did not capture the banding pattern as accurately. When viscoplastic effects were added (accounting for the material’s sensitivity to the rate of loading), the Tresca model still produced localization, but it shifted to a later time and took the form of diffuse necking rather than sharp shear bands.10EPJ Web of Conferences. Tresca criterion for plasticity and viscoplasticity: Application to localization in biaxial loading conditions These results illustrate a broader point: the Tresca criterion’s angular yield surface can sometimes capture abrupt transitions in deformation patterns that a smoother surface misses.
Computational Challenges
For all its conceptual simplicity, the Tresca criterion is surprisingly awkward to implement in computer simulations. The hexagonal yield surface has sharp corners where two flat faces of the hexagon meet. At those corners the “direction” of plastic flow is not uniquely defined, because two faces give two different outward normals. Numerical algorithms that step through the loading history need to know which direction plastic strain should evolve, and ambiguity at the corners causes convergence problems, slower run times, and sometimes outright failure of the solver.
Various techniques have been developed to work around these issues. Some approaches round off the corners slightly so the surface becomes smooth enough for standard algorithms. Others reformulate the yield function in the space of principal stresses to avoid the corners entirely, allowing the algorithm to track which face of the hexagon is active and handle corner transitions explicitly.11Computer Methods in Applied Mechanics and Engineering. A new computational model for Tresca plasticity at finite strains with an optimal parametrization in the principal space Von Mises does not have this problem at all, because its elliptical yield surface is smooth everywhere. This computational convenience is a major reason why von Mises dominates commercial finite-element codes. Engineers who specifically need Tresca, for instance in undrained clay analyses or when regulatory codes mandate it, often have to use specialized material routines rather than off-the-shelf options.
Anisotropy and the Limits of Isotropy
Both Tresca and von Mises assume the material is isotropic, meaning its strength is the same in every direction. Real sheet metals, rolled plates, and forged components are not isotropic. The rolling or forging process aligns the crystal grains into a preferred texture, so the material yields more easily in some directions than others. For metals with hexagonal close-packed crystal structures, like titanium and magnesium alloys, or even some body-centered cubic sheet metals, the asymmetry goes further: the yield stress in tension can differ from the yield stress in compression along the same direction.
To handle this, researchers have developed orthotropic yield criteria that extend the isotropic framework to account for directional differences. One approach starts from an isotropic yield function and introduces linear transformations of the stress tensor that stretch and rotate it to capture the material’s anisotropy. When two such transformations are applied, the resulting criterion can represent both the variation of yield stress with direction and the tension-compression asymmetry with high accuracy for both hexagonal close-packed and body-centered cubic metals.12AIP Conference Proceedings. Orthotropic yield Criteria for modeling the combined effects of anisotropy and strength differential effects in sheet metals These advanced criteria are descendants of the Tresca and von Mises ideas but are far more complex, and they are essential for accurate simulation of processes like deep drawing, stamping, and hydroforming where the sheet’s directional properties control whether it wrinkles, tears, or forms cleanly.
When To Use It and When To Reach for Something Else
Maximum shear stress theory earns its place through a combination of physical transparency and safe-side conservatism. If you are sizing a steel shaft, checking a pressure vessel wall, or estimating the collapse load of a clay foundation under rapid loading, Tresca gives a quick, defensible answer that leans toward safety. Many design codes in mechanical and structural engineering still reference it explicitly, and some international pressure-vessel standards define allowable stresses using the Tresca criterion rather than von Mises.
The criterion runs into trouble in a few specific scenarios. Materials that are strongly anisotropic need the directional extensions described above. Brittle materials that crack rather than yield require a fundamentally different approach. High-confinement situations in soils and rocks, where the mean pressure genuinely changes the shear strength, call for pressure-sensitive criteria. And in computational settings where millions of stress evaluations are needed per time step, the hexagonal corners make Tresca substantially more expensive than von Mises, sometimes enough to tip the choice toward the smoother criterion purely on practical grounds.
For hand calculations and back-of-the-envelope checks, though, maximum shear stress theory remains hard to beat. Its logic maps directly onto the physical mechanism of yielding in ductile metals, it requires only the two extreme principal stresses, and its conservatism gives a built-in safety margin that many engineers appreciate. A century and a half after Tresca watched metal flow under his presses, his criterion is still the first one students learn and the one many practicing engineers reach for when they need a fast, trustworthy answer.