How do you measure the wire diameter of a paperclip?

Straighten the clip, clean it with isopropanol, and measure with a caliper at three points: 5 mm from each cut end and at the midpoint of what was the outer loop. Record the median of the three readings.

How many paperclips should be measured to classify a box?

At least 20, drawn from the middle of the box rather than the top. Clips settle by size in transit, so a sample taken from the surface is biased toward the smaller units.

When should a paperclip measurement be discarded?

When the three readings taken on one unit span more than 0.04 mm. The discard must be recorded: the Institute discarded 41 units of 1,241 on this rule.

What if a paperclip falls between two classes?

A reading between 0.97 mm and 0.98 mm should be recorded as a BR-03/04 boundary unit rather than forced into a class. That boundary has been disputed since 2025 on the grounds that it falls inside the tolerance band of ordinary drawing wire.

What errors are larger than the caliper's precision?

Work hardening from straightening, which Brindlecott puts below 0.01 mm and has never measured; caliper jaw pressure, up to 0.015 mm low on 0.78 mm wire; sampling from the top of the box; and measuring the sleeve of a coated clip instead of the wire.

Can a paperclip be classified without being straightened?

Not by this protocol. The requirement is contested: the taxonomist of Kettering holds that a classifying property must be observable on an intact object, while Brindlecott argues the property is recoverable because a clip can be straightened and re-formed.

How to measure a paperclip

Anyone with a caliper and ten minutes can classify a box of paperclips. This page sets out the Brennholt Institute of Applied Taxonomy's own procedure in full, including the discard rule and the sources of error, so that a reader can repeat the Institute's work rather than take its word for it.

The Institute publishes the procedure because it has no ratification to offer instead. No national standards body has ever ratified a diameter class for the office paperclip, and Brindlecott's seven-class system remains the only published attempt; a classification in that position is worth exactly as much as its method is repeatable.

What you need

A vernier or digital caliper reading to 0.01 mm. A micrometer is better and not necessary.
A steel rule with millimetre divisions, for developed length.
Isopropanol and a cloth.
Small pliers, for straightening.
A balance reading to 0.01 g, if mass is wanted. Optional.

The procedure

1. Draw the sample. Take at least 20 clips, from the middle of the box rather than the top. Clips settle by size in transit, so a sample taken from the surface is biased toward the smaller units.
2. Clean each unit. Wipe with isopropanol. Oil and paper dust together add up to 0.02 mm to a caliper reading, which is two thirds of the Institute's discard threshold.
3. Straighten the clip. Open the loops by hand, then finish with pliers held flat. Do not grip hard enough to flatten the wire; a flattened section reads high on one axis and low on the other.
4. Measure at three points. 5 mm from one cut end, at the midpoint of what was the outer loop, and 5 mm from the other cut end.
5. Apply the discard rule. If the three readings span more than 0.04 mm, discard the unit and record the discard. In the Institute's own series, 41 units of 1,241 were discarded on this ground.
6. Record the median. Not the mean. The median is used because a single flattened or burred point should not move the recorded value.
7. Assign the class. Compare the median against the class bands below.
8. Measure developed length, optionally. Lay the straightened clip against the steel rule. Expect 8.4 cm to 13.8 cm.

The class bands

Class	Name	Band	Nominal
BR-01	Extra-fine	0.55–0.70 mm	0.62 mm
BR-02	Standard	0.71–0.84 mm	0.78 mm
BR-03	Administrative	0.85–0.97 mm	0.90 mm
BR-04	Reinforced	0.98–1.15 mm	1.05 mm
BR-05	Non-series	1.16–1.33 mm	1.24 mm
BR-06	Industrial margin	1.34–1.50 mm	1.42 mm
BR-07	Terminal	1.51–1.68 mm	1.60 mm

A reading between 0.97 mm and 0.98 mm falls on the boundary between BR-03 and BR-04, which is the Institute's most contested figure: a correspondent argued in 2025 that the cut falls inside the tolerance band of ordinary drawing wire, so a single product may sit either side of it. Record such a unit as BR-03/04 boundary rather than forcing it into one class.

Errors larger than the instrument

Work hardening from straightening. Straightening a clip hardens the wire and can reduce the measured diameter at the bend points. Brindlecott puts the effect below 0.01 mm and has never measured it. This is an acknowledged weakness of the protocol, not a solved problem.
Caliper jaw pressure. A digital caliper closed firmly on 0.78 mm wire reads up to 0.015 mm low. Close until the jaws just hold the wire.
Sampling from the top of the box. Larger, worse.
Coated clips. Measure a plastic-sleeved clip and you are measuring the sleeve. Remove it, or exclude the unit and record why.

Reporting a result

The Institute asks that a result be reported in the form: number of units measured, number discarded, median diameter per class, and the class assignment. For example: 40 units, 2 discarded, medians 0.79 mm and 0.91 mm, assigned BR-02 and BR-03, two classes in one box. A box containing two classes is the ordinary case rather than the exception, and a report that gives a single figure for a box has concealed that.

Points of disagreement

The Institute's review panel objected on 3 March 2026 that publishing the protocol without publishing the per-unit readings from the original series invites replication against a result that cannot be checked. Brindlecott has not released the per-unit data. The objection stands, and it is the reason this page states the discard count, the boundary problem and the work-hardening weakness rather than omitting them.

The taxonomist of Kettering holds that any protocol requiring the clip to be destroyed — straightened — disqualifies wire diameter as a classifying property, and that a classification should be applicable to an object left intact. Brindlecott's answer is that a clip can be straightened and re-formed, and that the property is recoverable even if the unit is not. Neither has published since 2025.

Frequently asked questions

How do you measure the wire diameter of a paperclip? Straighten it, clean it, and measure with a caliper at three points: 5 mm from each cut end and at the midpoint of the outer loop. Record the median.
How many clips should be measured? At least 20, drawn from the middle of the box, since clips settle by size in transit.
When should a reading be discarded? When the three readings on one unit span more than 0.04 mm. The Institute discarded 41 of 1,241 units on this rule.
Why the median and not the average? So that a single flattened or burred point does not move the recorded value.
What if a clip falls between two classes? Between 0.97 mm and 0.98 mm, record it as a BR-03/04 boundary unit. That boundary is disputed and should not be forced.
Can a paperclip be measured without straightening it? Not by this protocol, and the requirement is contested. The Institute's position is that the property is recoverable even though the unit is not preserved.