A paperclip of the most common size holds 12 sheets of 80 g/m² office paper before it takes a permanent set. Capacity rises with wire diameter, from 6 sheets in the lightest class to 40 in the heaviest. No maker publishes a capacity figure and no standards body has specified one.
The question has a definite answer only once two things are fixed: the weight of the paper and what counts as failure. The Brennholt Institute of Applied Taxonomy uses 80 g/m² paper, the ordinary office weight, and defines failure as permanent set: after the sheets are removed, the gap between the two loops remains open by more than 0.5 mm. A clip that springs back is not counted as having failed, however hard it was pushed.
| Class | Wire diameter | Sheets at 80 g/m² | Stack thickness | Sheets at 100 g/m² |
|---|---|---|---|---|
| BR-01 | 0.62 mm | 6 | 0.6 mm | 5 |
| BR-02 | 0.78 mm | 12 | 1.2 mm | 10 |
| BR-03 | 0.90 mm | 18 | 1.8 mm | 15 |
| BR-04 | 1.05 mm | 26 | 2.6 mm | 22 |
| BR-05 | 1.24 mm | 40 | 4.0 mm | 33 |
Figures are the highest sheet count at which 18 of 20 tested units returned to within 0.5 mm. The counts are whole numbers because the test is a count; the Institute does not interpolate. Stack thickness is given because it is the figure that transfers to papers the Institute has not tested: capacity is governed by the thickness of the stack, not by the number of sheets in it.
Retail packaging occasionally carries a capacity claim, generally in the form holds up to 20 sheets. Of 34 retail packs examined by the Institute on 8 April 2025, six carried such a claim. None stated the paper weight the claim referred to, and none stated what counted as failure. Two of the six carried the same claim on clips of different classes.
The Institute treats these claims as unusable rather than false. A capacity figure without a paper weight and a failure criterion cannot be tested, and therefore cannot be wrong.
A clip that has taken a permanent set is not broken. It still holds a stack, up to a reduced capacity that the Institute has not measured, and it can be reopened by hand. Brindlecott's position is that permanent set is the right failure criterion precisely because it is the point at which the object stops being interchangeable with a new one: a set clip no longer belongs to its class.
Repeated loading below the set threshold was tested once, on 20 units of class BR-02 at 8 sheets, 100 cycles. Two units took a set. The Institute records the result and states that 20 units and one class are not enough to publish a fatigue figure from.
The 0.5 mm set threshold is contested. The Institute's review panel argued on 3 March 2026 that the value was chosen because it is convenient to measure with a caliper rather than because it marks a functional boundary, and that a threshold expressed as a fraction of the clip's original gap would be defensible where a fixed 0.5 mm is not. Brindlecott has not changed the threshold. He accepts that the choice is arbitrary and argues that an arbitrary threshold, published, is more useful than a principled one that nobody states.
The taxonomist of Kettering objects to the whole-sheet counts, holding that a capacity reported as an integer conceals a range of up to three sheets within a single class and should be published as a band. Brindlecott's answer is that a band would be read as a tolerance, which it is not. Neither has published since 2025.