I am working on resubmitting a paper to a journal and am working on my response to reviewers. I need to refer to a bunch of line numbers in my response to reviewers and it is frustrating because every time I make a change to my manuscript the line numbers change a bit.

For instance, say a reviewer writes: "you should use the word 'correlate' rather than 'co-occur' in line 243" I would then reply. "'co-occur' was replaced with 'correlate' in line 245". Note the line number is now higher because I added a sentence to my paper. This is fine, but say I have a whole bunch of these lines, and then I decide to add one more sentence to the beginning of my paper. Right now, I would have to go through my response and update all of the line numbers. I feel like there must be a better way to keep track of these. Maybe with some sort of dynamic field or something in word or open office. Does anyone have a suggestion?

  • 3
    Admittedly not very sophisticated, but in the past I've dealt with this by just putting in placeholders (e.g. "???") for line numbers and going back and filling them in once everything else is done and I won't be making any more edits. I felt like the time required to figure out and set up something automatic would be more than searching for "???" in my response and then filling in the correct number. Jan 2, 2015 at 2:46
  • 3
    One option suggested by another user was to send the revised manuscript and a latexdiff of the original document. inf.unibz.it/dis/wiki/lib/exe/fetch.php?media=latex:3.png
    – Davidmh
    Jan 2, 2015 at 8:30
  • 1
    I generally just quote the referee's suggestions verbatim, with each suggestion followed by my response to that suggestion. For simple word replacements like your example, my actual response is the single word "Done."
    – JeffE
    Oct 25, 2016 at 14:01

6 Answers 6


This is always a pain, but I find it's not actually hard to deal with, just somewhat tedious. My method for doing so is rather low-tech: I keep a copy of the old document to be able to track the old line numbers, then leave all of the line numbers in the response as XXX until I am done with the revision. Then, for each entry, you just cross-index in both places and write both line numbers, e.g.,

'co-occur' was replaced with 'correlate' in line 245 [prev. 243]

For minor wording/grammar edits, however, I usually don't actually respond in such detail, but simplify both my life and the reviewers by just saying something like: "thank you, fixed"

  • I agree, avoiding referring to the line numbers at all is a simpler solution. Convert the references into structural or context ones. "In the third paragraph of this chapter, 'leaned to dunce the Whiskey Foxtrot Tango' should be corrected to "learned to dance the Whiskey Tango Foxtrot'."
    – keshlam
    Jan 2, 2015 at 8:05
  • 5
    My advisor just copies the reviewer's comment then writes a short statement explaining what was fixed, no line numbers in our response. Something like Reviewer: "Paragraph 3 Line 2, Word X should be Y", Us:"Word X was changed to Y".
    – user137
    Jan 2, 2015 at 16:13

Just refer to the line numbers mentioned in the comments. There is no need to go through the pain of also checking where the changes occur in the new document. If you work in Word you can keep the track changes and provide both a file with all the changes visible and one which is a clean version of the new version. If you work in LaTeX you could possibly put all changes in, for example, bold (since bold is usually not used anywhere else in a manuscript, still retaining a version without any highlighted changes. There are a few attempts for revision tools in LaTeX as well but as far as I have seen require intervention with a script or external software.

In any case, I do not know of any journals, including "my own", that would require such extensive reporting so as to keep track of both old and new line numbers for minute changes. This does not mean they do not exists but is is usually way over what will be required.

When it comes to small changes such as rephrasing of the type you mention, it would be more than sufficient to provide a letter itemizing all changes and just say "changed 'to occur' on l. 243 to 'correlate'". In fact your can also consider lumping all your made minor changes into one statement to the fact that you have made the changes and then follow up on the ones where you have decided not to do so or have done something differently.

  • 1
    +1 for "usually way over what will be required". At least in the sub-fields of CS that I am familiar with, even what you suggest in your third paragraph is "way over what will be required", given that the length of what authors may respond to reviewers (e.g. by word count) is usually very restricted and only leaves enough space to address a selection of the most important review comments at all. Jan 2, 2015 at 13:57

My preference where the journal allows is to avoid making references to page and line numbers. It's a time consuming process and can generally be avoided.

  • If the change is really simple (e.g., fixing a typo; changing one word), I'll just acknowledge that the change has been made.
  • If a sentence or paragraph has been substantially reworded, I'll first paste a copy of the original paragraph and then a copy of the updated paragraph into my response document, possibly highlighting new text in bold.
  • If text has been added to a paragraph, I'll paste the updated paragraph into the response document with added text in bold.

The general logic of the approach is to make the reviewer's task as simple as possible. The reviewer generally wont want to have to perform complex cross-referencing of responses with page and line numbers in the manuscript. Instead, I think it's easier for the reviewer to make such responses self-contained in the response document. If the reviewer really wants to check that the changes have actually been made, there's enough text in the response document to enable a quick search on the manuscript PDF to find the location.


Just asked this question to myself and wanted to add my solution here.

In latex, you can use a combination of the packages lineo, hyperrefand xr to dynamically generate line numbers.

In the manuscript:

  • use lineo to number your lines (\linenumbers)
  • enable hyperref
  • wherever needed, use \linelabel{lne:label1} to mark the line.
    • You could for instance add \linelabel{lne:fromLine} at the beggining of the sentence you want to refer to and \linelabel{lne:toLIne} at the end.

In the answer to the reviewers:

  • use xrto link to the manuscript and get all the labels you defined there.
  • Wherever needed use \lineref{lne:label1} to link to the needed line.

Some lecture on lineo: https://texblog.org/2012/02/08/adding-line-numbers-to-documents/ Some lecture on xr: https://texblog.org/2016/08/23/adding-references-from-an-external-file/

  • 1
    Nice example of Latex code golf. I'm sure it was entertaining but it's not necessary for such a trivial issue. Just list the new line number.
    – Cape Code
    Nov 22, 2018 at 11:44
  • 1
    That's your opinion, I actually think it's quite useful, and maybe it will be for somebody else:) Especially if you have a lot of comments and last-minute changes, not having to rewrite all the line numbers can prove to be a considerable save of time.
    – Giezi
    Nov 22, 2018 at 18:40
  • 1
    That is an nice example of using a 15 pound sledge hammer to crack a nut...
    – Solar Mike
    Nov 23, 2018 at 9:57
  • 1
    Given that the OP mentioned Word and OpenOffice, a solution in LaTeX is probably not of any use to them. Otherwise I think this would be a pretty good answer.
    – David Z
    Nov 24, 2018 at 8:30

When I revise a paper, I like to use a \revised command, which e.g. changes the color to blue. When submitting a new version, you just need to change the command and remove it only in the final version.

When you now changed your document, you see your changes in blue and can use their line numbers, which will not change when you change the color back to black. This may be a bit faster than looking in the final document for the places where you changed something.


Also, depending on the field you are into and the confidence with technology of the reviewer, consider using GitHub or even a wiki: they have a version-control systems and you can actually add comments coding-style. Startups like Authorea or Penflip are trying to solve the problem of collaborating on these kind of documents (it's not just the review, but the whole writing papers).

I'm not affiliated with any of these projects, I just like collaborative tools an think they are important, especially in academia.

  • 11
    I don't think there is currently any field where you can have any kind of confidence about the reviewers' general technical abilities.
    – ff524
    Jan 2, 2015 at 10:40
  • It should be an AND/OR then :-). I was betting that in CS and Math people would at least be accustomed to code, Latex and some useful tools (like GitHub).
    – Aubrey
    Jan 2, 2015 at 11:18
  • 2
    @Aubrey: This may well differ by field, but the way I know blinded peer review processes, authors never learn who reviewed their submission, and (in double-blind reviews) reviewers can only ever find out whose submission they reviewed if they happen to come across the published version of the paper later on. Just as the reviews are usually sent to the editor (preferrably via some online submission platform), who then forwards the reviews to the authors, the responses by authors are sent to the editor, who passes them on to the appropriate reviewers. Jan 2, 2015 at 14:35
  • 3
    @Aubrey any major journal uses an online editorial software that handles the communication between authors and reviewers anonymously. Sure, in field with a small community you can try to guess people based on grammar errors or choice of word, but it is still speculative.
    – Cape Code
    Jan 2, 2015 at 14:47
  • 1
    @Aubrey: No, not necessarily; it may be directly applicable in some fields, as you already note in your answer, and even in fields where review blindness is generally more enforced, there are different venues, different submission types, and connected to those, a sliding scale of how ensuring review blindness and use of such possibly non-anonymous tools are weighted. Jan 2, 2015 at 14:51

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .