Lab Notebook Concepts
TL;DR
A witness signature, or countersignature, is a second signature on a lab notebook entry from someone other than the author, confirming they have read and understood the work it records. It corroborates what was done and when, and it is usually required by a lab's own SOPs rather than by any single regulation.
A witness signature is a second signature on a lab notebook entry, added by someone who is not the author, confirming that they have read the entry and understood what it records. Labs also call it a countersignature. On a bound notebook the line usually reads "read and understood by," which sums up the job well: the witness confirms that this record existed, in this form, on this date. Whether the science holds up is a separate question.
The practice comes from patent law. Under the old US first-to-invent system, a dated page signed by a non-inventor witness was evidence of when an idea was conceived and how diligently it was pursued.
A witness signature is also different from an approval. An approval says the work or the result is accepted, usually by a supervisor or QA, while a witness only vouches for the entry as a record. Some labs combine the two in one step, so it helps to write down which one yours means.
Someone who understands the technique but didn't do the work. For intellectual property purposes the witness also shouldn't be a co-inventor, since an inventor can't corroborate their own invention. For quality purposes they need enough hands-on experience to spot an obvious gap, like a missing lot number for the primary antibody or a dilution that doesn't match the protocol.
In practice that's usually a colleague from the next bay.
Usually no regulation requires one directly. Witness signatures are mostly required by a lab's own SOPs or by company policy. GLP rules such as 21 CFR 58.130(e) require entries to be dated and signed by the person who made them and changes to be explained, and many organizations add a witness step on top of that for studies conducted under GLP. In academic labs the requirement tends to come from the technology transfer office, for work that might lead to a patent.
The patent side has changed. The America Invents Act moved US patents to a first-inventor-to-file system for applications filed from March 16, 2013, so a witnessed page carries less weight for priority than it used to. But witnessed notebooks still matter in derivation proceedings and ownership disputes, or whenever a company needs to prove when it had a result.
It follows the paper routine, with fewer places for an entry to get stuck. A typical sequence looks like this:
In IGOR, the author signs a notebook entry in the electronic lab notebook and then assigns witnesses, and a Super Admin sets the required witness count per project team or department. A witness who rejects an entry has to leave a comment. The author corrects it and re-signs, and once approved the entry is permanently locked to protect data integrity.
Mostly because nobody can see the queue. On paper, a scientist finishes a week of transfection experiments, signs and dates each page, and hands the notebook to a colleague, where it sits in an in-tray for a month. Then the witness signs fifteen pages in one go, all with the same date.
Annoying, right?
It also wastes the one routine moment when a second person reads the record while the experiment is still fresh. That is when a missing incubation time or an unlabeled gel image can still be fixed from memory. Six months later, at an audit or while the patent application is being drafted, nobody remembers.
So decide which entries need a witness, whether that's everything or only the entries behind a patent filing or a regulatory submission. Set a turnaround target, for example ten working days from signing, and write both into an SOP. Pairing people across projects helps too, because each scientist then has a regular witness who knows their techniques. In IGOR, pending witness requests appear in Mission Control, color-coded by status and age, so you never lose track of outstanding witnessing requests.
Yes, a supervisor can witness as long as they understand the work and aren't a co-inventor on it. Some labs keep supervisor approval as a separate step, though, and in that case a peer from the same department is the better witness.
They are less central for US patent priority since first-inventor-to-file took effect for applications filed from March 16, 2013. But witnessed records still help in derivation proceedings and ownership disputes, and they come up in due diligence, so many companies and technology transfer offices still require them.
The witness checks that the entry is complete enough for someone else to follow. That usually means reagents with lot numbers and gel images or plots that link back to their raw files. If a dilution doesn't match the protocol, a short rejection comment asking for the correction is all it takes.
Yes. A witness can review and sign an ELN entry from anywhere, as long as they use their own electronic signature and the system records who signed and when. For regulated work, that signature has to meet the same controls as any other electronic signature, for example those in 21 CFR Part 11.