Research Eagle

Figure requirements

Pen-and-ink illustration of a researcher photographing colleagues at work outdoors

Figures are the most common cause of a manuscript being returned after acceptance, which is the most frustrating point at which to discover a problem. The requirements are specific, mechanical and published in advance, and almost nobody reads them until production asks.

The five checks that catch most submissions

1. Resolution

The usual minimums are 300 dpi for photographs and halftones, 600 to 1200 dpi for line art, at final printed size. That last phrase is what trips people up: a figure at 300 dpi when 4 inches wide is 150 dpi when printed at 8 inches. Resolution is meaningless without the intended dimensions.

Resampling upward does not help. Enlarging a 150 dpi image to 300 dpi in an editor adds pixels without adding information, and production teams can see it. Export again from the source application at the size you need.

2. File format

TIFF and EPS remain the common requirements, with PDF accepted by many publishers for vector figures. PNG and JPEG are frequently not accepted for final figures even when the submission system will let you upload them, because the system accepts them at review stage and rejects them at production.

Where your figure is vector in origin, a chart, a diagram, a flowchart, keep it vector. Exporting a vector plot to a bitmap discards resolution you already had.

3. Colour mode

RGB for online-only journals, CMYK where there is a print edition. This matters visually: saturated RGB colours, particularly bright greens and blues, shift noticeably when converted to CMYK, and the conversion happens whether or not you do it. Doing it yourself means you see the result first.

Check also whether colour figures carry a charge. Several journals still charge for colour in print while providing colour online at no cost, in which case your figure needs to remain readable in greyscale.

4. Captions and numbering

  • Figures cited in the text in sequential order, with no gaps.
  • Captions below figures, above tables, at most publishers.
  • Captions supplied as editable text, not embedded in the image file.
  • Every abbreviation in the figure defined in its caption, because figures are reproduced standalone.
  • A caption that makes the figure interpretable without the main text.

5. Fonts and legibility

Embed all fonts, or convert text to outlines. A missing font substitutes at production and the substitution is often catastrophic for a carefully spaced axis label. Minimum text size after reduction is typically 6 to 8 point, which means axis labels set at 8 point in a figure that will be reduced by half are illegible.

Design the figure at the size it will be printed. Most legibility failures are the result of designing at screen size and letting production shrink it.

Things that are increasingly checked

  • Image manipulation. Adjustments must be applied to the whole image, not selectively. Splicing lanes from different gels without a visible divider is a research integrity issue, not a formatting one.
  • AI-generated imagery. Prohibited outright by several publishers, including enhancement of real data images.
  • Accessibility. Do not encode meaning in colour alone. Add shape, pattern or direct labelling so the figure survives greyscale printing and the most common forms of colour vision deficiency.
  • Permissions. A figure reproduced or adapted from another paper needs permission and a credit line, including where you are the author of the original.

A pre-submission figure checklist

  1. Every figure at or above the required resolution at final printed width.
  2. Every figure in an accepted format, vector kept vector.
  3. Colour mode matching the journal’s requirement.
  4. Count within the journal’s maximum.
  5. All cited in text, in order.
  6. Captions editable, below the figure, abbreviations defined.
  7. Fonts embedded, text legible at final size.
  8. Readable in greyscale.
  9. Permissions obtained for anything reproduced.
Scroll to Top