4

I'm in the process of writing figure legends but realized I don't actually know what to include or what to exclude. I wrote them based on intuition and tried to copy what I've already seen but I was wondering if anyone had good suggestions or references for making figure legends for papers and reports in the biological sciences.

Improve this question

2 Answers 2

Reset to default
5

In general, the figure should contain a title which concisely captures what it's showing: 'Effect of manipulation A on measure B', or 'Level of A influences property of B', or 'B vanishes when A is larger than X', or some such.

After that you would go on describe what the figure shows, as precisely as possible - go panel by panel and say what's on the axes (and in which units), what denotes statistical significance in the image (e.g. gray shaded areas), what the colors mean, that sort of thing. You can also write what the main statistical finding is, using just words (i.e. no p-values or similar).

Those are the basics. However, take into account that some journals allow unsubscribed readers to see the title, abstract and figures of their papers. If this is the case, try to imagine what it's like to be able to read just the abstract and see the figure with legends, and build from there. See if an experienced reader can infer what your specific contribution is from this information only (in fact, many people will anyway just skim over the figures before deciding whether to read the paper). This often means that the figure legend should also include a mini-recap of the experimental setup or specific manipulations next to the finding it depicts, but the exact contents will vary depending on what you're showing.

Improve this answer
1
  • 1
    I would also that you can sometimes point the attention of the reader to a specific part or feature of the figure, e.g. "Nuclear membranes at the bottom of the figure appear to be elongated and distorted".
    – Bitwise
    Commented Apr 7, 2014 at 0:31
1

The chosen answer suggests going through panel by panel and stating what's on each axis and its units or what the colors mean. You shouldn't need to do this: axes labels tell you what is on the axes and that's where the units should go (not in the caption). As for what the colors mean, that's ideally the purpose of a legend.

I have a post about this that goes into more detail about avoiding long-winded, procedural, and uninteresting figure captions.

Improve this answer

You must log in to answer this question.

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