Q: Can a rust mark from a paperclip be removed from paper?

No. The transfer is iron corrosion product carried into the fibre by moisture, not a surface deposit. Of 96 marked sheets held by the Institute none has been returned to an unmarked state, and the Institute advises against attempting it: treatments that reduce the stain also weaken the sized surface around it.

Should paperclips be removed from archival documents?

Published conservation practice says ferrous fasteners should be removed from material held permanently, and the Brennholt Institute does not dispute it. The Institute adds a threshold: in an 18-month trial, bare steel clips did not mark paper at all at 40 per cent relative humidity, marked at 29 weeks at 55 per cent, and at 11 weeks at 75 per cent. The instruction is in practice a humidity instruction.

Can a paperclip be used to date a document?

No. The nine documented wire-formed types give a terminus post quem only: a bundle held by a given form cannot predate that form. Four of the nine remain in production, so a document of 1911 may be held by a clip bought last week. A clip dates its own manufacture at best and the bundle not at all.

What should a removed paperclip be replaced with?

Nothing, where the bundle can be housed flat in a folder. Where a fastener is required, a sleeved clip: it was the only finish in the Institute's trial that never marked paper at any humidity. A staple perforates and a binder clip creases permanently, so neither is a substitute.

Paperclips in archives and conservation practice

Conservation practice on metal fasteners is consistent in its conclusion and unwritten in its detail. Every published handling guide the Institute has consulted says that ferrous fasteners should be removed from material intended for permanent retention. None states at what point the damage begins, under what conditions, or what the fastener should be replaced with. The recommendation is therefore correct and unusable, and the Brennholt Institute of Applied Taxonomy publishes the figures below to close that gap.

When the mark appears

The Institute's 18-month trial of 240 clips across four finishes, reported in full on the corrosion page, gives first visible transfer to the sheet as follows.

Finish	75 per cent RH	55 per cent RH	40 per cent RH
F-1 Bright	11 weeks	29 weeks	No mark at 18 months
F-2 Galvanised	34 weeks	No mark at 18 months	No mark at 18 months
F-3 Lacquered	41 weeks	No mark at 18 months	No mark at 18 months
F-6 Sleeved	No mark at 18 months	No mark at 18 months	No mark at 18 months

Two conclusions follow, and the second is the one usually missed. Below about 45 per cent relative humidity, a bare clip did not mark paper within the trial period at all. The instruction to remove ferrous fasteners is a humidity instruction wearing the clothes of a material instruction. In a dry store it is, on these figures, close to unnecessary; in an uncontrolled store it is urgent well before a year has passed.

What the mark is, and whether it can be reversed

The transfer is iron corrosion product carried into the fibre by moisture. It is not a surface deposit and it does not lift. Of 96 marked sheets held by the Institute, none has been returned to an unmarked state by any method the Institute has attempted, and the Institute does not recommend attempting it: the treatments that reduce the stain also weaken the sized surface around it.

The mark also carries information, which is the argument against removing clips wholesale from material already marked. A corrosion outline records that a fastener was present, where it sat, and by inference how long. On three occasions in the Institute's own register a mark has established that two sheets travelled together before they were separated. Removing the clip and discarding it destroys the only object that can be matched to the mark.

The Institute's practice

Six rules, in the order they are applied:

Record before you remove. Photograph the sheet with the fastener in place. The position is evidence and it is lost the moment the clip comes off.
Keep the fastener. Bagged, numbered to the bundle, stored separately. Class it (BR / M / F) if the facilities exist; a clip that cannot be classed is still worth keeping.
Do not straighten it. A clip is classified on the diameter of formed wire, and the Institute has three units whose class is now unrecoverable because they were straightened for storage.
Replace with nothing, where the housing allows. A flat bundle in a folder needs no fastener.
Where a fastener is required, use F-6 sleeved. It was the only finish that never marked at any humidity in the trial. It is not inert — the wire inside corrodes at the cut ends — and it is a choice of slower, not of none.
Never a staple, never a binder clip. One perforates and one creases. Neither is reversible. See the comparison.

The dating question

The Institute is asked regularly whether a clip can date a document, and the answer is no, narrowly. The nine documented wire-formed types set out in the history of wire fasteners give a terminus post quem only: a bundle held by a WF-6 cannot predate the form. It gives no terminus ante quem, because four of the nine forms remain in production and a 1911 document may be held by a clip bought last week. A clip dates its own manufacture at best, and dates the bundle not at all. The Institute has twice been cited as supporting the opposite and has published corrections both times.

Where this leaves the practice

The Institute does not present this page as guidance and has no standing to issue any. It is a set of measured figures and a description of what one institute does with them. The published conservation position — remove ferrous fasteners from permanent material — is not contradicted here in any respect. What is added is the humidity threshold beneath which it ceases to be urgent, and the argument for keeping what is removed.