How Much Will a Shelf Sag?

Structurally, a shelf is a beam spanning between two supports. How far it deflects depends on four things: the load, the span, the stiffness of the material (its modulus of elasticity, E) and the shape of its cross-section (the second moment of area, I). The product EI is the beam's bending stiffness.

For a rectangular shelf, I equals the depth front to back multiplied by the cube of the thickness, divided by 12. That cube is why thickness dominates: a shelf 1 in thick is (1 / 0.75)³ = 2.37 times stiffer than one 3/4 in thick, while making it a third deeper only makes it a third stiffer.

Standard beam tables, such as the American Wood Council's Design Aid No. 6, give the maximum deflection for common load and support cases. Two load cases cover most shelves: a load spread evenly along the span (a row of books) and a concentrated load at the centre (a single heavy object). Deflections from the two can be added.

The support condition matters as much as the load. A shelf resting on pins, cleats or a loose dado is close to simply supported: the ends can rotate. A shelf glued and screwed rigidly into a stiff carcass approaches a fixed-end beam, which deflects one fifth as much under a spread load. Real joints fall somewhere in between, so the simply supported case is the cautious estimate.

The USDA Forest Products Laboratory's Wood Handbook tabulates the modulus of elasticity of many species for clear, straight-grained wood at 12% moisture content. Northern red oak, for example, is listed at 12,500 MPa (1.82 million psi) and eastern white pine at 8,500 MPa (1.24 million psi). The handbook notes that these values come from centre-loaded bending tests and include some shear deflection; they can be increased by about 10% to remove it. Using them as published is slightly conservative.

Panels are listed separately. The handbook's composite-materials chapter gives 3.59 GPa for medium-density fibreboard, a range of 2.76 to 4.14 GPa for particleboard and 6.96 to 8.55 GPa for plywood. Plywood stiffness depends on the veneer layup and on whether the face grain runs along the span, so a manufacturer's published value is better where one exists.

Clear-wood values are averages. Knots, sloping grain, higher moisture content and elevated temperature all reduce stiffness, and individual boards vary around the average.

Take a 36 in span, 10 in deep, 3/4 in thick shelf on pins carrying 60 lb of books spread along its length. In northern red oak, beam theory gives an initial sag of 0.057 in, or L/630. That is 0.0191 in per foot of span.

The same shelf in MDF, at 3.59 GPa, sags 0.199 in (L/181) — 3.5 times as much, exactly the ratio of the two moduli. Making the oak shelf 1 in thick instead would cut its sag to 0.024 in, and if the oak shelf could be fully fixed at both ends, the sag would be 0.011 in.

Gluing a solid-wood strip to the front edge, especially one taller than the shelf is thick, is a common way to stiffen a shelf without making it look heavy. The strip and the shelf bend together as one section, so their stiffness is calculated as a composite. First find the combined neutral axis: the stiffness-weighted average height of the parts. Then add each part's own second moment of area to its area multiplied by the square of its distance from that shared axis. That second term, from the parallel-axis theorem, is why a strip hanging below the shelf adds so much: it puts material far from the neutral axis.

When the strip and the shelf are different materials, each part is weighted by its own E. A hardwood edging on an MDF shelf therefore contributes more than its size suggests.

For the MDF shelf above, a 1 1/2 in tall by 3/4 in thick northern red oak strip, top flush with the shelf, brings the initial sag from 0.199 in down to 0.048 in (L/745), less than the plain oak shelf.

Wood loaded for a long time continues to deform after the initial elastic deflection. The Wood Handbook describes this as creep and states that, at typical design levels and use environments, after several years the additional deformation may approximately equal the initial deflection. It adds that ordinary changes in temperature and humidity increase creep, and that green wood can creep several times more as it dries under load.

Structural design accounts for this with a long-term factor. The American Wood Council summarises the National Design Specification's approach as an expected long-term deflection of about 1.5 times the initial deflection for seasoned lumber and 2.0 times for unseasoned lumber. Books left on a shelf for years are exactly this kind of sustained load.

For the MDF shelf, the long-term sag is 0.299 in with a 1.5 factor and 0.398 in with 2.0. These factors come from solid-wood guidance; wood-based panels are not covered by the figures quoted here, so a cautious factor is sensible for them.

None of the sources cited here sets a deflection limit for furniture shelves. A widely used woodworking guideline comes from the Sagulator calculator published by Woodbin: the eye notices a deflection of about 1/32 in per running foot of span, and because shelves sag further over time, 0.02 in per foot is suggested as the target for initial sag. Expressed as span ratios, those are L/384 and L/600.

The calculator on this site passes a shelf only if both hold: initial sag at or below 0.02 in per foot and long-term sag at or below 1/32 in per foot. The oak shelf passes (0.0191 in/ft initial, 0.0286 in/ft long term). The plain MDF shelf does not (0.0664 in/ft initial); with the oak edging it does (0.0161 in/ft initial, 0.0242 in/ft long term). These are appearance guidelines, not safety limits.

Which limit controls depends on the creep factor. Passing both means initial sag at or below the smaller of 0.02 in/ft and (1/32 in/ft) divided by the creep factor. With a factor of 1.5 that is still 0.02 in/ft; above 1.5625, including the 2.0 factor, the long-term limit governs and the allowable initial sag falls to 1/64 in/ft at 2.0.

In rough order of effectiveness for a given shelf:

Beam theory assumes a straight, uniform shelf of one material behaving elastically, with small deflections. Shear deformation, which matters more for short, deep members, is ignored. The support conditions are idealised, and material values are averages for clear wood at 12% moisture content.

Most importantly, deflection is not strength. A shelf can be stiff enough to look flat and still be overloaded at its pins, brackets, screws or joints. The calculation says nothing about breaking load, fastener capacity or the cabinet that holds the shelf.

To run these numbers for your own shelf, including an edging strip and a creep factor, use the Shelf Sag Calculator.